MAY/JUNE 1999
Volume 43 • Number 3
The Journal of
IMAGING SCIENCE
and
TECHNOLOGY
IS&T
The Society for Imaging Science and Technology
EDITORIAL STAFF
M.R.V. Sahyun, Editor
IS&T
7003 Kilworth Lane, Springfield, VA 22151
715-836-4175
E-mail: sahyunm@uwec.edu
FAX: 703-642-9094
Pamela Forness, Managing Editor
The Society for Imaging Science and Technology
7003 Kilworth Lane, Springfield, VA 22151
703-642-9090; FAX: 703-642-9094
E-mail: pam@imaging.org
Vivian Walworth, Editor Emeritus
Eric Hanson, Associate Editor
Akihiro Hirano, Associate Editor (deceased)
Michael M. Shahin, Associate Editor
Arthur Soucemarianadin, Associate Editor
Mark Spitler, Associate Editor
David S. Weiss, Associate Editor
David R. Whitcomb, Associate Editor
This publication is available in microform and annually
on CD-Rom.
Papers published in this journal are covered in BECITM,
INSPEC, Chemical Abstracts, and Science Citation
Index.
Address remittances, orders for subscriptions and single
copies, claims for missing numbers, and notices of
change of address to IS&T at 7003 Kilworth Lane,
Springfield, VA 22151. 703-642-9090; FAX: 703-6429094; E-mail: info@imaging.org.
The Society is not responsible for the accuracy of
statements made by authors and does not necessarily
subscribe to their views.
Copyright © 1999, The Society for Imaging Science
and Technology. Copying of materials in this journal
for internal or personal use, or the internal or personal
use of specific clients, beyond the fair use provisions
granted by the U.S. Copyright Law is authorized by
IS&T subject to payment of copying fees. The Transactional Reporting Service base fee for this journal
should be paid directly to the Copyright Clearance
Center (CCC), Customer Service, (508) 750-8400, 222
Rosewood Drive, Danvers, MA 01923 or check CCC
Online at http://www.copyright.com. Other copying for
republication, resale, advertising or promotion, or any
form of systematic or multiple reproduction of any
material in this journal is prohibited except with permission of the publisher.
Library of Congress Catalog Card No. 59-52172
Printed in the U.S.A.
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY (ISSN:1062-3701) is published bimonthly
by The Society for Imaging Science and Technology,
7003 Kilworth Lane, Springfield, VA 22151. Periodicals postage paid at Springfield, VA and at additional mailing offices. Printed at Imperial Printing
Company, Saint Joseph, Michigan.
Society members may receive this journal as part of
their membership. Thirty dollars ($30.00) of the membership dues is allocated to this subscription. IS&T
members may refuse this subscription by written
request. Subscriptions to domestic non-members of
the Society, $120.00 per year; single copies, $25.00
each. Foreign subscription rate, US $135.00 per year.
POSTMASTER: Send address changes to JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY, 7003 Kilworth Lane, Springfield, VA 22151.
Guide for Authors
Scope. THE JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY is dedicated to the advancement
of knowledge in the imaging sciences, in practical applications of such knowledge, and in
related fields of study. The pages of this journal are open to reports of new theoretical or
experimental results and to comprehensive reviews.
Submission of Manuscripts. Send manuscripts to Pamela Forness, Managing Editor, IS&T,
7003 Kilworth Lane, Springfield, VA 22151. Submit only original manuscripts not previously
published and not currently submitted for publication elsewhere. Prior publication does not
refer to conference abstracts, paper summaries or non-reviewed proceedings, but it is expected
that Journal articles will expand in scope, the presentation, and dissemination of such
preliminary communication. Please include keywords on your title and abstract page.
Editorial Process. All manuscripts submitted are subject to peer review. If a manuscript
appears better suited to publication in the JOURNAL OF ELECTRONIC IMAGING, published jointly
by IS&T and SPIE, the editor will make that recommendation to both the editor of that journal
and the author. The author(s) will receive the necessary forms, confirmation, reviewers' reports,
notification of acceptance (or rejection), and tentative date of publication from the Editor. The
author will receive page proofs and further instructions directly from the JIST production team.
All subsequent correspondence about the paper should be addressed to the Managing Editor.
Manuscript Preparation. Manuscript should be printed double-spaced, with all pages
numbered. The original manuscript and three duplicates are required, with a legible set of
illustrations to accompany each copy. The illustrations included with the original manuscript
must be of quality suitable for reproduction or available in digital form. Single-spaced
manuscripts, as well as those submitted in “two column” format, will be returned to the
author(s) for reformatting.
Title and Abstract Pages. Include on the title page, page one, the address and affiliation
of each author. Include on page two an abstract of no more than 200 words stating briefly
the objectives, methods, results, and conclusions of the work. List several keywords describing
the major topics discussed in the paper.
Style. The journal will generally follow the style specified in the AIP Style Manual, published
by the American Institute of Physics, available from Springer Verlag, New York, 1-201-348-4505.
Equations. Number equations consecutively, with Arabic numbers in parentheses at the right
margin.
Illustrations. Number all figures consecutively and type captions double-spaced on a separate
page or pages. Figures should be presented in such form that they would remain legible
when reduced, usually to single column width (3.3 in., 8.4 cm). Recommended font for figure
labels is Helvetica, sized to appear as 8-point type after reduction. Recommended size for
original art is at least 2 times final size. Lines must be at least 1 point.
Color. Authors may either submit color separations or be billed by the publisher for the
cost of their preparation. Digital submission of color figures should be in CMYK EPS or
TIFF format if possible. Additional costs associated with reproduction of color illustrations
will be charged to the author or the author’s supporting institution. Additional color illustrations may be submitted as supplementary material for posting on the IS&T website and
for inclusion in the CD-ROM version of the Journal.
References. Number references sequentially as the citations appear in superscript form in the
text. Type references on pages separate from the text pages, using the following format:
Journal papers: Author(s) (first and middle initials, last name), title of article (optional), journal name (in italics),
volume (bold), first page number, year (in parentheses).
Books: Author(s) (first and middle initials, last name), title (in italics), publisher, city, year, page reference.
Examples:
1. H. P. Le, Progress and trends in ink-jet printing technology, J. Imaging Sci. Technol. 42, 46 (1998).
2. E. M. Williams, The Physics and Technology of Xerographic Processes, John Wiley and Sons, New York,
1984, p. 30.
3. G. K. Starkweather, Printing technologies for images, gray scale, and color, Proc. SPIE 1458, 120 (1991).
4. L. T. Creagh-Dexter, Applications in commercial printing for hot melt ink-jets, in Proc. IS&T’s 10th Int’l.
Congress on Adv. in Non-Impact Printing Technologies, IS&T, Springfield, VA 1994, pp. 446–448.
Page charges. To help support expenses of publication the author’s sponsoring institution
will be billed at $80 per printed page. Such payment is not a condition for publication, and
in appropriate circumstances page charges are waived. Requests for waivers must be made
in writing to the managing editor, prior to acceptance of the paper and preferably at the
time of submission.
Additional charges. The author will be held responsible for the cost of extensive alterations
and/or for the costs of materials if the paper is withdrawn after preparation for publication.
Manuscripts in Digital Form. Following acceptance of a manuscript, the author will be
required to submit the final version of the text and illustrations both as hardcopy and as
digital files on diskette. The preferred formats are: MS Word (RTF) or ascii text; Graphics
should be submitted as separate eps or tiff files. Files can be compressed if needed. Zip
disks are accepted and may be returned to you upon request.
Vol. 43, No. 3, May/June
1999
i
CODEN: JIMTEG 43(3) 201–307 (1999)
ISSN: 1062-3701
May/June 1999
Volume 43, Number 3
Journal of
IMAGING SCIENCE
and
TECHNOLOGY
Official publication of IS&T—The Society for Imaging Science and Technology
CONTENTS
v
From the Editor
M. R. V. Sahyun
201
Electron Trapping in N,N¢-bis(1,2-dimethylpropyl)-1,4,5,8-Naphthalenetetracarboxylic Diimide Doped Poly(styrene)
P. M. Borsenberger, W. T. Gruenbaum, E. H. Magin, and S. A. Visser and D. E. Schildkraut
206
Hole Transport in Doubly-Doped Polystyrene
S. Heun and P. M. Borsenberger
213
Thermally Stimulated Luminescence in Molecularly Doped Polymers
A. Kadashchuk, N. Ostapenko, V. Zaika and P. M. Borsenberger
220
Recent Advances in Charge Transport in Random Organic Solids: The Case of Conjugated Polymers
and Discotic Liquid Crystals
D. Hertel, A. Ochse, V. I. Arkhipov, and H. Bässler
228
Suitable Definition of Drift Mobility
Akiko Hirao, Takayuki Tsukamoto and Hideyuki Nishizawa
233
Transient Space-Charge-Limited Current Measurements of Mobility in a Luminescent Polymer
J. C. Scott, S. Ramos and G. G. Malliaras
237
Carrier Transport in Molecularly Diluted Liquid Crystalline Photoconductor
Kensuke Kurotaki and Jun-ichi Hanna
242
Effect of Metal Contact Fabrication and Nature of the Metal Contact on the Evolution in the Charge Injection
Efficiency of Evaporated Metal Contacts on a Molecularly Doped Polymer
Andronique Ioannidis, John S. Facci, and Martin A. Abkowitz
248
Photoconduction Mechanism in Single-Layer Photoconductor with Metal-Free Phthalocyanine
Kazuki Kubo, Toshio Kobayashi, Suguru Nagae, and Takamitsu Fujimoto
Contents continued
ii
Journal of Imaging Science and Technology
Contents continued
254
Extrinsic Photocarrier Generation Mechanism in a Dual-Layer Organic System
Minoru Umeda
261
Photocarrier Generation in Polysilane Films Doped With and Without Fullerene
Yoshikazu Nakayama, Akira Saito, Tatsuo Fujii, and Seiji Akita
266
Sensitized and Intrinsic Carrier Generation in Phenethylperylene/Tritolylamine Thin Film Structures
Zoran D. Popovic, Robin Cowdery, Iltaf M. Khan, Ah-Mee Hor, and Joshua Goodman
270
Image Resolution in Liquid Development for Electrophotography
Inan Chens
274
A Study of Non-Uniform Charging by Charging Roller with DC Voltage
Masami Kadonaga, Tomomi Katoh and Tomoko Takahashi
280
Silsesquioxane Sol-Gel Materials as Overcoats for Organic Photoreceptors
D. S. Weiss, W. T. Ferrar, J. R. Corvan, L. G. Parton, and G. Miller
288
Effects of Silica Additive Concentration on Toner Adhesion, Cohesion, Transfer, and Image Quality
B. Gadys, D. J. Quesnel, D. S. Rimai, S. Leone, and P. Alexandrovich
295
Effect of Adsorption of Long Chain Alcohol Molecules on Silica Particles on Toner Charging
Kock-Yee Laws, and Ihor W. Tarnawskyj
300
Effect of Alcohol Grafting on the Charging Characteristics of Silicas in Xerographic Toner
Kock-Yee Law, and Ihor W. Tarnawskyj
DEPATMENTS
iii
Calendar
306
Business Directory
Calendar
IS&T Meetings
October 17–22, 1999—NIP15: The l5th International
Congress on Digital Printing Technologies, General Chair: Michael Lee, The Caribe Royal Resort Suites,
Lake Buena Vista, Florida
November 16–19, 1999—7th Color Imaging Conference— Color Science, Systems & Applications, cosponsored by the Society for Information Display;
General Co-chairs: Jack Holm (IS&T) and T odd
Newman (SID), The SunBurst Resort Hotel, Scottsdale,
Arizona
September 10–14, 2000— International Symposium
on Silver Halide Technologies, co-sponsored by
SPSTJ, General Co-chairs: Rene DeKeyzer, Gary House,
Melville Sahyun, and Tadaaki Tani, Resort Hotel MontGabriel, Montreal (St. Adele), Quebec, Canada
November 6–10, 2000— 8th Color Imaging Conference—Color Science, Systems, and Applications,
co-sponsored by the Society for Information Display, The
SunBurst Resort Hotel, Scotsdale, Arizona
January 22–28, 2000—IS&T/SPIE Electronic Imaging: Science and Technology, General Co-chairs: John
McCann (IS&T) and Giordano Beretta (SPIE), San Jose
Convention Center, San Jose, California
For more details, contact IS&T at
703-642-9090; FAX: 703-642-9094;
E-mail: info@imaging.org;
or visit us at www.imaging.org
January 31–February 2, 2000— 11th International
Symposium on Photofinishing Technology, General
Co-chairs: Steven Howe and Daniel English, co-located
with the PMA Exhibition, Las Vegas, Nevada
For a more complete listing of other imaging conferences
March 26–29, 2000—The PICS Conference, (IS&T’s
53rd Annual Spring Conference), General chair: Jim
Milch, The Portland Marriott Hotel, Portland, Oregon
• Visit IS&T’s website: www.imaging.org
• See the Other Meetings column in the member
newsletter, IS&T Reporter
• Request a printout via e-mail: info@imaging.org or
fax: (703) 642-9094
Vol. 43, No. 3, May/June
1999
iii
From the Guest Editor
This issue of the Journal is a memorial to the late Dr .
Paul M. Borsenberger who passed away on July 17, 1998.
After receiving a Ph.D. from Stanford in materials science he joined the Eastman Kodak Company where he
rose to the rank of Senior ResearchAssociate. A prolific
researcher, Dr. Borsenberger published about one hundred scientific papers in areas related to the photoconductivity of disordered solids and the application of these
materials to electrophotography. He also co-authored,
with Guest Editor, two books, “Organic Photoreceptors
for Imaging Systems” and “Organic Photoreceptors for
Xerography”. It is especially noteworthy that in 1995
he was ranked in the top ten of the Science Watch
“Roundup of hot papers and scientists”. This was the
first time a physical science researcher had been so honored. His research involved extensive collaborative efforts. Of approximately sixty colaborations, thirty
percent were from academic institutions and companies
other than Eastman Kodak.
The response to our request for papers for a Memorial issue of the Journal has been outstanding. Many
others wrote expressing their desire to honor Dr .
Borsenberger’s memory, but that they would not be able
to submit a paper. It is our intention that this issue be
a tribute to Dr . Borsenberger from all of us who have
been influenced by his kindness, good nature, and extensive knowldge.
iv
Journal of Imaging Science and Technology
We received far more papers than could be accommodated in a single Journal issue so that some must be
published in a later issue. Also, we decided to limit the
subject matter for this memorial issue to the science
and technology of electrophotography. Other papers have
been diverted to other Journal issues and to the Journal of Electronic Imaging where they will appear with
a dedicatory statement to Dr. Borsenberger.
The organization of papers in the memorial issue is
by subject matter, beginning with carrier transport in
disordered media. Dr. Borsenberger is a co-author on
several of these papers. Following are papers concerning carrier generation, general electrophotographic technology and, finally, toner technologies.
The Guest Editor would like to thank the Journal
Editor and staff for their cooperation in compiling this
memorial issue. W e invite the Journal readership to
enjoy the wonderful science and technology of the papers, which appear in this and in part of an upcoming
issue, as a memorial to Dr. Paul M. Borsenberger.
David S. Weiss
Eastman Kodak Company
Guest Editor
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Electron Trapping in N,N′-bis(1,2-dimethylpropyl)-1,4,5,8Naphthalenetetracarboxylic Diimide Doped Poly(styrene)
P. M. Borsenberger, W. T. Gruenbaum,▲ E. H. Magin, and S. A. Visser*
Office Imaging Division, Eastman Kodak Company, Rochester, New York
D. E. Schildkraut
Imaging Research and Advanced Development, Eastman Kodak Company, Rochester, New York
Electron mobilities have been measured in N,N′-bis(1,2-dimethylpropyl)-1,4,5,8-naphthalenetetracarboxylic diimide doped poly(styrene)
containing a series of acceptor traps: 4-(cyanocarboethoxymethylidene)-2-methyl-1,4-naphthoquinone (MNQ), 3,5-dimethyl-3 ′,5′diisopropyl-4,4′-diphenoquinone (DPQ), 4 H-1,1-dioxo-2,6-di-tert-butyl-4-(dicyanomethylidene)thiopyran (TBS), N,N ′-dicyano-2-tertbutyl-9,10-anthraquinonediimine (DCAQ), and 4 H-1,1-dioxo-4-dicyanomethylidene-2-p-tolyl-6-phenylthiopyran (PTS). From reduction potential measurements, the trap depths of MNQ, DPQ, TBS, DCAQ, and PTS are 0.19, 0.19, 0.20, 0.35, and 0.40 eV
, respectively.
The mobilities decrease with increasing trap depth and trap concentration. The results are discussed within the framework of th e
Hoesterey–Letson formalism and the recent simulations of Wolf and co-workers and Borsenberger and co-workers.
Journal of Imaging Science and Technology 43: 201–205 (1999)
Introduction
Molecularly doped polymers contain an electron donor or
acceptor molecule in a polymer host. Hole or electron transport occurs by charge transfer between adjacent donor or
acceptor molecules, respectively. This can be described as
a one-electron oxidation or reduction process between neutral molecules and their charged derivatives.1–4 Due to their
widespread use as xerographic photoreceptors,5–11 there is
considerable interest in transport phenomena in these
materials. The mobilities are very low, strongly field and
temperature dependent, as well as dependent on the
dopant molecule, dopant concentration and the polymer
host. For a review, see Borsenberger and Weiss.11
Many recent studies have been described by a formalism based on disorder, due to Bässler and coworkers. 12–17
The formalism is premised on the argument that transport occurs by hopping though a manifold of localized
states that are distributed in energy . The key parameter is σ, the energy width of the hopping site manifold
or density-of-states (DOS). The principal predictions are
the field and temperature dependencies of the mobility
where simulations predict lnµ ∝ βE1/2 and –(T 0/T)2 relationships.18 Here, E is the field, T temperature, and β
and T 0 are coefficients that increase with decreasing
temperature and decreasing field, respectively. These results agree with those reported for a wide range of m olecularly doped polymers, as well as pendant and main
chain polymers and vapor-deposited molecular glasses.19
Recently, Wolf and co-workers20 and Borsenberger and
co-workers21 extended the formalism to include effects
Original manuscript received April 2, 1998
▲ IS&T Member
* Corresponding author
© 1999, IS&T—The Society for Imaging Science and Technology
of trapping. In the treatments of W olf and co-workers
and Borsenberger and co-workers, traps are considered
to be neutral when empty and charged when occupied.
The site energies are taken as the total energy of the
dopant molecule with a hole or excess electron on that
site, relative to that of the uncharged molecule. The
simulations show that the presence of a distribution of
shallow traps, offset from the intrinsic DOS by energy
Et, does not change the basic phenomenology of transport, as revealed by the temperature and field dependencies of the mobility . The characteristic ln µ ∝ βE1/2
and –(T 0/T) 2 relationships are retained. The effect of
trapping can be quantitatively accounted for by the replacement of σ with an effective width σeff. Relating the
trap-controlled mobility to the trap-free mobility by an
expression due to Hoesterey and Letson22
µ(c) = µ(c = 0) f–1 = µ(c = 0) {1 + c[exp(Et/kT)]}–1
(1)
yields a relationship between σeff2 and the trap depth and
the logarithm of the trap concentration c as
(σeff/σ)2 = 1 + (3kT/2σ)2(Et/kT + ln c).
(2)
Here, µ(c) is the trap-controlled mobility, µ(c = 0) the trapfree mobility, f a term that describes the increase of the
transit time by the time spent by a carrier in traps, σ the
width of the DOS in the absence of traps, and k
Boltzmann’s constant. For c[exp(Et/kT)] >> 1, Eq. 1 predicts the mobility scales with trap concentration as c–1.
For a series of arylamine donor molecules doped with traps
of different depths, the work of V eres and Juhasz 23 and
Borsenberger and co-workers 21,24–26 yield c–1.0 to c–1.5. The
results further show that the coefficient of the concentration dependence increases with increasing trap depth. For
hole trapping, the results thus suggest that while the
Hoesterey–Letson formalism may provide a meaningful
201
TABLE I. Molecular Weights, Reduction Potentials, and Trap
Depths of Compounds Used in This Study
Compound
NTDI
MNQ
DPQ
TBS
DCAQ
PTS
M(g/mole)
406
267
296
304
312
358
ERED(V)
–0.596
–0.403
–0.403
–0.392
–0.245
–0.194
Et(eV)
0.19
0.19
0.20
0.35
0.40
The reduction potentials were measured in dichloromethane versus
saturated calomel by Osteryoung square wave voltammetry. A detailed
description of the technique is given in Ref. 26. The uncertainties in
the potentials are estimated as ±0.004 V.
description for shallow traps, it may not hold for traps of
moderate depth. For electron trapping of N,N ′-bis(1,2dimethylpropyl)-1,4,5,8-naphthalenetetracarboxylic
diimide (NTDI), the only literature reference is the work
of Borsenberger and co-workers.27 In agreement with the
simulations of Wolf and co-workers, a plot of ( σeff/σ)2 versus trap depth showed a linear dependence. The slope,
however, was considerably lower than predicted. The dependence on trap concentration was not described.
To further investigate electron trapping in these materials, we have extended our earlier work with NTDI
to include the effects of trap concentration. The traps
were 4-(cyanocarboethoxymethylidene)-2-methyl-1,4naphthoquinone (MNQ), 3,5-dimethyl-3′,5′-diisopropyl4,4′-diphenoquinone (DPQ), 4 H-1,1-dioxo-2,6-di-tertbutyl-4-(dicyanomethylidene) thiopyran (TBS), N,N ′d i c y a n o - 2 -t e r t - b u t y l - 9 , 1 0 - a n t h r a q u i n o n e d i i m i n e
(DCAQ), and 4 H-1,1-dioxo-4-dicyanomethylidene-2-ptolyl-6-phenylthiopyran (PTS). The polymer host was
poly(styrene). From reduction potential measurements,
the trap depths of MNQ, DPQ, TBS, DCAQ, and PTS
are 0.19, 0.19, 0.20, 0.35, and 0.40 eV, respectively.
Experimental
Figure 1 shows the molecular structures of NTDI, MNQ,
DPQ, TBS, DCAQ, and PTS. Samples were prepared by
dissolving different ratios in dichloromethane, then coating the solutions on Ni-coated poly(ethylene terephthalate) substrates that had previously been coated with a
0.30 µm layer of α-Se. The solids concentration of the coating solutions was 10%. All samples contained 40 wt%
NTDI, equivalent to 1.0 × 10–3 moles of NTDI/cm 3. The
trap concentrations are expressed as the mole fraction of
the traps to NTDI and correspond to the parameter c in
the work of Wolf and co-workers20 and Borsenberger and
co-workers.21 The molecular weights, reduction potentials, and trap depths are summarized in T able I. A detailed description of the techniques used for thereduction
potential measurements has been given in Ref. 25. From
cross-section photomicrographs, thicknesses of the
doped polymer layers were between 10 and 12 µm.
The mobilities were measured by conventional timeof-flight photocurrent transient techniques. For a review
of the method, see Melnyk and Pai. 28 In brief, the displacement of a sheet of electrons, created in the α-Se
layer by 3 ns exposures of 440 nm radiation, is timeresolved. The exposures were derived from an N 2 pumped dye laser. The photocurrent transients were
measured with a transient digitizer. The mobilities were
derived from the conventional expression, µ = L2/t 0V,
where L is the sample thickness, t0 the transit time, and
V the applied potential. All measurements were made
at room temperature.
202
Journal of Imaging Science and Technology
Figure 1. The molecular structures of molecules used in this study.
A more detailed description of the techniques used for
sample preparation and the mobility measurements has
been given in our earlier work.20,21,24-27
Results
For NTDI, the photocurrent transients are similar to
those reported for a wide range of acceptor doped polymers. The transients feature an initial spike of very
short duration, a plateau of variable temporal length,
and a long tail. Plateaus were observed over the range
of fields investigated. The width of the tail can be described by the tail-broadening parameter W, defined as
W = (t 1/2 – t 0)/t1/2, where t1/2 is the time for the photocur rent to decay to one-half its value att0. Values of W were
weakly field dependent, increasing with increasing field.
At 3.6 × 105 V/cm, W was approximately 0.42. The features for NTDI containing MNQ, DPQ, TBS, DCAQ, or
PTS were substantially different. The presence of PTS
at concentrations of a few multiples of 10 –7 erodes the
transients. The initial spike is suppressed, the plateaus
are less well defined, and W is increased. For concentrations in excess of 10 –6, transit times can be resolved
only from double logarithmic transients. For concentrations in excess of 10 –3, however, the transients closely
resemble those of NTDI, although with transit times
that are very long, frequently in excess of a few s. For
NTDI containing DCAQ, the transients were degraded
only for concentrations in excess of approximately 10–5.
Transit times could be resolved from double linear transients at all concentrations. For NTDI containing MNQ,
DPQ, or TBS, the transients were unchanged except for
concentrations in excess of a few multiples of 10 –4. As
with DCAQ, transit times could be derived from double
linear transients for all concentrations.
Visser, et al.
TABLE II. NTDI Electron Trapping Parameters
Trap
molecule
Et(eV)
MNQ
DPQ
TBS
DCAQ
PTS
0.19
0.19
0.20
0.35
0.40
c1/2EXP
3.0 ×
2.5 ×
3.5 ×
2.0 ×
1.5 ×
10–3
10–3
10–3
10–5
10–4
c1/2CAL
5.5 ×
5.5 ×
3.6 ×
1.2 ×
1.6 ×
10–4
10–4
10–4
10–6
10–7
n
0.89
0.91
0.96
1.33
1.54
The parameter n is derived from the relationship µ(c) ∝ µ(c = 0)c–n.
Figure 3. The mobility versus MNQ concentration. The field was
3.6 × 105 V/cm. The dashed line is the mobility in the absence of
MNQ. The arrow indicates the trap concentration for which the
trap-free mobility is reduced by a factor of two.
Figure 2. The field dependencies of the mobilities for NTDI and
NTDI containing MNQ, DPQ, TBS, DCAQ, and PTS. The trap
concentrations were 10–2.
For both NTDI and NTDI containing MNQ, DPQ, TBS,
DCAQ, and PTS, the field dependencies of the mobility
can be described as ln µ ∝ βE 1/2. Here, β is a coefficient
that is weakly dependent on trap depth. Figure 2 shows
the room temperature results for trap concentrations of
10–2. Figures 3 through 7 show the room temperature
mobilities versus trap concentration. For MNQ, DPQ,
TBS, and DCAQ, the mobilities were derived from transients in double linear current versus time representation. For PTS, however, it was necessary to use double
logarithmic transients. Table II summarizes values of
c1/2 and n derived from the results in Figs. 3 through 7.
Here, c 1/2 is the trap concentration where the mobility is
reduced by a factor of two from its trap-free value. The
parameter n is derived from the relationship µ(c) ∝ µ(c
= 0)c –n under the condition c[exp(Et/kT)] >> 1.
Discussion
First, we discuss the temporal features of the photocur rent transients. Provided the trapping lifetime is well in
excess of the transit time, a carrier does not interact with
trapping centers during its transit and the transients remain unaffected. As the trap concentration is further increased, the trapping lifetime eventually becomes
comparable to the transit time. Consider the case where
the number of traps is within a factor of two of the num-
Electron Trapping in ....Doped Poly(styrene)
Figure 4. The mobility versus DPQ concentration. The field was
3.6 × 105 V/cm. The dashed line is the mobility in the absence of
DPQ. The arrow indicates the trap concentration for which the
trap-free mobility is decreased by a factor of two.
ber of jumps a carrier makes upon traversing the sample
and the trap depth is such that it takes several multiples
of the trap-free transit time to escape thermally
. For these
conditions, some of the carriers will traverse the thickness without trapping, some will have single trapping
events, and some will have multiple trapping events. This
regime is usually described as trap-perturbed and char -
Vol. 43, No. 3, May/June 1999
203
Figure 5. The mobility versus TBS concentration. The field was
3.6 × 105 V/cm. The dashed line is the mobility in the absence of
TBS. The arrow indicates the trap concentration for which the
trap-free mobility is decreased by a factor of two.
Figure 7. The mobility versus PTS concentration. The field was
3.6 × 105 V/cm. The dashed line is the mobility in the absence of
PTS. The arrow indicates the trap concentration for which the
trap-free mobility is decreased by a factor of two.
Figure 6. The mobility versus DCAQ concentration. The field
was 3.6 × 105 V/cm. The dashed line is the mobility in the absence of DCAQ. The arrow indicates the trap concentration for
which the trap-free mobility is decreased by a factor of two.
Figure 8. The concentration c1/2 at which the mobility is decreased
by a factor of two from its trap-free value versus trap depth. The
dashed line was calculated from Eq. 3.
acterized by a wide dispersion of transit times.As the concentration is further increased, all carriers experience
multiple trapping events during their transit. Under these
conditions, the dispersion of transit times is considerably
reduced and the transients more closely resemble those
in the absence of traps, although featuring transit times
that are substantially longer. This regime is usually described as trap-controlled. At very high trap concentrations, trap-to-trap hopping occurs with the result that the
mobility increases with trap concentration. Depending on
the trap depth, the trap-to-trap regime may or may not be
observed. The results observed with NTDI containing
204
Journal of Imaging Science and Technology
Visser, et al.
the trapping factor in Eq. 1 to deviate from a product of
the trap concentration and an exponential term that contains the trap depth.
Figure 9. The dependence of the mobility on trap concentration
versus trap depth. The parameter n is derived from the relationship µ(c) ∝ µ(c = 0) c–n.
MNQ, DPQ, TBS, DCAQ, and PTS are in accord with these
arguments. The results clearly show that the trap depth
and concentration play a major role in the temporal features of the transients.
Next, we discuss the Hoesterey–Letson formalism.
This is perhaps the simplest approach to trapping. It is
based on a multiple trapping argument and premised
on the early work of Shockley and Read 29 and Bube. 30
The model assumes a discrete trap depth and does not
include effects related to disorder . The model leads to
two basic predictions. First, the concentration at which
the mobility is decreased by a factor of two is
c1/2 = exp(-Et/kT)
(3)
Figure 8 shows results derived from Figs. 3 to 7. The results are clearly not in accord with Eq. 3. The formalism
underestimates the onset of the trap-controlled regime,
particularly for deep traps. For MNQ, DPQ, and TBS, traps
with depths of 0.19, 0.19, and 0.20 eV , the discrepancies
are approximately a factor of five. For PTS, a 0.40 eV trap,
the discrepancy is a factor of 103. The second prediction is
that for c[exp(Et/kT)] >> 1, the mobility scales with trap
concentration as c–1. Figure 9 shows the results obtained
from the data in Figs. 3 to 7. Contrary to predictions, the
dependence of the mobility on trap concentration is clearly
dependent on trap depth. While the multiple trapping assumption must eventually break down with increasing trap
depth, this cannot account for coefficients of less than unity
.
The most likely explanation is related to the width of the
intrinsic and trap manifolds. The Hoesterey–Letson for malism is based on a discrete trap depth, an assumption
which is likely unrealistic for disordered molecular solids.
Physically, this assumption neglects the opening of new
relaxation pathways for an ensemble of carriers due to
the additional states at the tail of the DOS. This causes
Electron Trapping in ....Doped Poly(styrene)
Concluding Remarks
The results of this study show that the field dependencies
of the mobility for NTDI containing traps with depths between 0.19 and 0.40 eV agree with the simulations of W
olf
and co-workers.20 The characteristic lnµ ∝ βE1/2 dependencies were observed for all trap depths and concentrations.
There was no evidence of dependencies of the form ln µ ∝
βE, as predicted for the deep trapping regime.21 The use of
the Hoesterey–Letson formalism to describe the effects of
trap concentration leads to significant discrepancies concerning the onset of the trap-controlled regime and the
dependence of the mobility on trap concentration at high
concentrations. A similar effect has been previously reported for hole trapping in a series of arylamine doped
polymers.23–26 It is our speculation that the discrepancies
are due to the neglect of disorder in the derivation of the
trapping factor that describes the time spent by carriers
in traps.
References
1. G. Pfister, Phys. Rev. B 16, 3676 (1977)
2. J. Mort and G. Pfister, Polym. Plast. Technol. Eng. 12, 89 (1979).
3. D. M. Pai, J. F. Yanus, M. Stolka, D. Renfer, and W. W. Limburg, Philos.
Mag. B 48, 505 (1983).
4. J. S. Facci and M. Stolka, Philos. Mag. B 54, 1 (1986).
5. A. R. Melnyk and D. M. Pai, Proc. SPIE 1253, 141 (1990).
6. D. M. Pai, in Frontiers of Polymer Research, P. N. Prasad and J. K.
Nigam, Eds., Plenum Press, New York, 1991, p. 315.
7. D. M. Pai and B. E. Springett, Rev. Mod. Phys. 65, 163 (1993).
8. K.-Y. Law, Chem. Rev. 93, 449 (1993).
9. M. Stolka and J. Mort, in Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed., John Wiley and Sons, New York, 1994, p. 245.
10. M. Stolka, in Special Polymers for Electronics and Optoelectronics, J.
A. Chilton and M. T. Goosey, Eds., Chapman and Hall, London, 1995,
p. 284.
11. P. M. Borsenberger and D. S. Weiss, in Organic Photoreceptors for
Xerography, Marcel Dekker, Inc., New York, 1998.
12. H. Bässler, Phys. Status Solidi (b) 175, 15 (1993), and references
therein.
13. H. Bässler, Int. J. Mod. Phys. B 8, 847 (1994).
14. H. Bässler, in Disorder Effects on Relaxation Processes, R. Richert
and A. Blumen, Eds., Springer-Verlag, Berlin, 1994, p. 485.
15. H. Bässler, Mol. Cryst. Liq. Cryst. 252, 11 (1994).
16. B. Hartenstein and H. Bässler, J. Non-Cryst. Solids 190, 112 (1995).
17. B. Hartenstein, H. Bässler, A. Jakobs, and K. W. Kehr, Phys. Rev. B.
54, 8574 (1996).
18. P. M. Borsenberger, L. Pautmeier and H. Bässler, J. Chem. Phys. 94,
5447 (1991).
19. P. M. Borsenberger, E. H. Magin, M. Van der Auweraer, and F. C. De
Schryver, Phys. stat. sol. (b) 140, 9 (1993).
20. U. Wolf, H. Bässler, P. M. Borsenberger, and W. T. Gruenbaum, Chem.
Phys. 222, 259 (1997).
21. P. M. Borsenberger, W. T. Gruenbaum, U. Wolf, and H. Bässler, submitted to Chem. Phys.
22. D. C. Hoesterey and G. M. Letson, Phys. Chem. Solids 24, 1609 (1963).
23. J. Veres and C. Juhasz, Philos. Mag. B 75, 377 (1997).
24. P. M. Borsenberger, E. H. Magin and S. A. Visser, submitted to Jpn. J.
Appl. Phys.
25. P. M. Borsenberger, W. T. Gruenbaum, E. H. Magin, S. A. Visser, and
D. E. Schildkraut, submitted to J. Polym. Sci.: Part B: Polym. Phys.
26. P. M. Borsenberger, W. T. Gruenbaum, E. H. Magin, S. A. Visser, and
D. E. Schildkraut, submitted to J. Imaging Sci. Technol.
27. P. M. Borsenberger, W. T. Gruenbaum, E. H. Magin, and S. A. Visser,
submitted to Phys. stat. sol.
28. A. Melnyk and D. M. Pai, in Physical Methods of Chemistry, B. W.
Rossiter and R. C. Baetzold, Eds., J. Wiley and Sons, New York, 1993,
2nd ed., Vol. 8, 1993, p. 321.
29. W. Shockley and W. T. Read, Jr., Phys. Rev. 87, 835 (1952).
30. R. H. Bube, in Photoconductivity of Solids, John Wiley and Sons, New
York, 1960, p. 68.
Vol. 43, No. 3, May/June 1999
205
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Hole Transport in Doubly-Doped Polystyrene
S. Heun* and P. M. Borsenberger
Office Imaging Division, Eastman Kodak Company, Rochester, New York
Hole mobilities have been measured in polystyrene films doped with a mixture of two very similar triphenylamine derivatives at
an overall concentration of 20% by weight. The data were analyzed in the framework of a formalism based on disorder due to
Bässler and coworkers. The key predictions of the formalism concern the field and temperature dependencies of the mobility
from which the key parameters σ, the energy width of the hopping site manifold, and∑, the positional disorder parameter, can be
determined. The experimental results from this study are in good agreement with the predictions of the formalism, although
some of them are difficult to explain on arguments based on disorder
. This is especially true for the energetic disorder parameter
σ that was found to be smaller for films that contained 5% of the higher oxidation potential compound (an antitrap thus) compared to polystyrene layers doped with 20% of the pure compounds. But even the trap case cannot be fully explained on the basis
of trapping arguments or the dipolar disorder model. Therefore, the results confirm earlier findings that doubly-doped polymers
show some other contribution to charge transport than merely disorder related effects.
Journal of Imaging Science and Technology 43: 206–212 (1999)
Introduction
Molecularly doped polymers are essentially isolating
materials capable of transporting injected or photogenerated charges by charge transfer between adjacent
dopant sites. The low molecular weight dopants are donor molecules in the case of hole transporting and acceptor molecules for electron transporting layers. Due
to the widespread use of these materials as transport
layers for xerographic photoreceptors, 1,2 and their potential use in electroluminescent, 3,4 photorefractive, 5,6
and photovoltaic devices, 7 there has been considerable
interest in charge transport phenomena in these materials in the past decade. Many recent studies have been
described by a formalism based on disorder , due to
Bässler and coworkers.8–10 In the disorder formalism, it
is assumed that charge propagation occurs by hopping
through a manifold of localized states with super -imposed energetic and positional disorder. The key assumptions of the formalism are:
1. the distributions of site energies and distances are
Gaussians,
2. the jump rates can be described by an expression due
to Miller and Abrahams, 11
3. electron-phonon coupling is sufficiently weak to
render polaronic effects negligible, yet strong enough
to guarantee coupling to the heat bath, and
4. the process is incoherent, characterized by loss of
phase memory after each jump.
In the Miller–Abrahams expression, downward jumps
in energy are assumed not to be impeded by an energy
matching condition for dissipating the excess electronic
energy. The expression also implies that these jumps
are not accelerated by the electric field. The key predictions of the disorder formalism are:
1. nondispersive transients over a wide range of temperatures with a transition to dispersive transport at a
critical temperature,
2. a field dependence of the mobility of log µ ∝ E 1/2 over
a wide range of fields, followed by a saturation and
eventual increase in mobility with decreasing field
at very low fields, and
3. a temperature dependence of log µ ∝ T–2.
These predictions agree with experimental results reported for a wide range of both donor and acceptor doped
polymers,12–17 main chain and pendant group polymers,18–20
as well as vapor deposited molecular glasses.21–25
A key parameter of the disorder formalism is σ, the
(Gaussian) energy width of the hopping site manifold,
or density-of-states (DOS). There is strong evidence that
the overall width of the DOS in common molecularly
doped polymers or molecular glasses is determined by a
dipolar component that depends on the concentration
and dipole moments of all dopant molecules and the
matrix, 26–28 and a van der W aals component that increases with increasing dilution due to an increase in
structural randomness:13,27
σ = √(σ d2 + σ vdW2).
Original manuscript received November 18, 1998
* Corresponding author; Present address: Océ Technologies B.V., Group
Research & Technology, 5900 MA Venlo, The Netherlands.
© 1999, IS&T—The Society for Imaging Science and Technology
206
(1)
The dipolar contribution can be described by an expression due to Young29:
σ d = 7.04c1/2p/(a2 ∈),
(2)
where c is the dopant concentration, p the dipole moment, a the average intersite distance (in Å), and ∈ the
dielectric constant. Knowing the dipole moments and
the van der Waals contribution thus allows predicting
the energetic disorder parameter.
One of the most important limitations of the formalism is that it does not address charge transport in the
presence of a second transport molecule that is offset
from the primary dopant in transport energy. This has
led to the recent interest in trapping effects in molecularly doped polymers. 30,31 Here, adding a second transport molecule with a lower oxidation potential in a
concentration in excess of a critical trap concentration
cc caused an apparent change of the overall width of the
DOS according to:
(σ eff/σ)2 = 1 + (3kT/2σ)2(E t/kT + lnc),
%TTB in PS
µ0/(cm2/Vs)
σ/eV
Σ
µ(RT)/(cm2/Vs)
0
1.8 × 10–3
0.102
2.5 × 10–4
3.8
2.4 × 10–6
5
6.6 × 10–4
0.105
2.3 × 10–4
3.9
4.4 × 10–7
15
7.4 × 10–4
0.096
1.6 × 10–4
4.1
1.1 × 10–6
20
2.9 × 10–3
0.102
2.4 × 10–4
3.9
2.6 × 10–6
C/(cm/V)1/2
(3)
where σeff is the disorder parameter as derived from experiments, σ the width of the DOS in the absence of traps,
and Et the energetic displacement between the trap and
the intrinsic DOS. With this adjustment, the normal field
and temperature dependencies as predicted by the for malism could be recovered. Eq. 3 however , does not address changes in the energy width due to concomittant
changes in the dipolar or van der Waals contributions as
predicted by the dipolar disorder model from Eqs. 1 and
2. These effects become especially important at high relative trap concentrations where the trap limit does not
apply due to charge transport occuring by hopping within
the trap manifold. Moreover, Eq. 3 assumes that the only
significant difference between the trap and the primary
charge transport molecule is the difference in oxidation
potential which does not need to be the case.
A good way to investigate these effects is provided by
the so called double-doping experiments where two wellknown charge transport materials are dissolved in a polymer matrix together. Due to their widespread use in
xerographic photoreceptors, triphenylamine derivatives
are probably the best investigated class of hole transport materials. With one exception (p-EFTP,32–34), agreement of experimental data with predictions of the
formalism has led to the conclusion that charge transport in triphenylamine doped polymers can be explained
by simple disorder-controlled hopping without invoking
polaronic effects. Two members of this class are N,N´,N´´,
N´´´-tetrakis(4-methylphenyl)-(1,1´-biphenyl)-4,4´-diamine
(TTB) and N,N´-diphenyl-N,N´-bis(3-methylphenyl)-(1,1´biphenyl)-4,4´-diamine (TPD).35,36 Their oxidation potentials were measured by cyclovoltammetry in
dichloromethane against an Ag/AgCl standard electrode
and found to be 0.628V for TTB and 0.716V for TPD. 37
TTB is thus 88meV easier to oxidize than TPD and should
act as a trap in doubly-doped polymer layers. The structural similarity of both compounds (see Fig. 1) provides
the chance that the difference in oxidation potential is in
fact the only difference of relevance. The charge transport parameters for polystyrene layers doped with 20%
by weight of either TTB or TPD were indeed very similar38 with a σ of 0.102 eV for both compounds (see also
Table I). The small difference in dipole moment (1.56
Debye for TTB and 1.52 Debye for TPD39), was not recovered at that concentration. With the apolar polystyrene
(PS) as the polymer host, the dipolar and van der Waals
contributions to the overall width of the DOS should
therefore remain constant for films that contain the same
overall concentration of both compounds. This article
presents the results of charge transport experiments in
such doubly-doped layers.
Hole Transport in Doubly-Doped Polystyrene
TABLE I. Transport Parameters and Room Temperature Mobility for
Polystyrene Doped with 20% of One or Two Triarylamine Derivatives
N,N´,N´´, N´´´-tetrakis(4-methylphenyl)-(1,1´-biphenyl)4,4´-diamine (TTB)
N,N´-diphenyl-N,N´-bis(3-methylphenyl)-(1,1´biphenyl)-4,4´-diamine (TPD)
Figure 1. Molecular Structures of TTB and TPD.
Experimental
The molecular structures of TTB and TPD are illustrated
in Fig. 1. Both compounds were synthesized by a modified Ullmann condensation from N,N´-di(4-methylphenyl)-(1,1´-biphenyl)-4,4´-diamine and p-iodotoluene, and N,N´-diphenyl-(1,1´-biphenyl)-4,4´-diamine and
m-iodotoluene, respectively. 40 They were purified by
chromatography on silica gel. The PS was obtained from
Sinclair Koppers as Dylene 8X (Mw = 200000 g/mol) and
used as received. Samples were prepared by dissolving
the appropriate ratios of TTB, TPD, and PS in chloroform, then coating the resulting solution on polymer
substrates that had been previously coated with a semitransparent conducting layer . The overall triphenylamine concentration was 20% by weight, comprised
of 5% TTB and 15% TPD (“5/15”) and 15% TTB and 5%
TPD (“15/5”), respectively. Due to the difference in molecular weight between the two compounds, this translates to molar concentrations of 4.8 mol-% TTB in the 5/
15- and 5.2 mol-% TPD in the 15/5-coating. The films
were dried at room temperature on a covered coating
block in a humidity-controlled laminar air-flow-hood for
45 min, then for another 45 min in an oven at 80 °C.
Samples prepared that way were amorphous and showed
Vol. 43, No. 3, May/June 1999 207
Asorbance
Asorbance
Wavelength /nm
Figure 3. Sample configuration.
Wavelength /nm
Figure 2. UV/Vis-spectra of polystyrene films doped with 20%
triphenylamine: (a) 20% TTB and 20% TPD; (b) 5% TTB / 15%
TPD and 15% TTB / 5% TPD.
no tendency to crystallize over a period of several
months. From cross section photomicrographs and capacitance measurements, the film thicknesses were determined as between 8.9 and 10.4 µm. Figure 2 shows a
comparison of the UV/V is-spectra of the doubly-doped
films with those of the pure molecularly doped polymers
from Ref. 38. Due to the relatively high dopant concentration in conjunction with the film thicknesses and the
UV absorption of the conductive substrate, these spectra are not well resolved. Nevertheless, they provide a
means to check for the most important premise of
double-doping experiments: Additivity of the density of
states and lack of non-linear intermolecular interactions
between the two dopant molecules.As the spectra show,
the red-shifted TTB absorption appears in both spectra
according to the chosen concentration without any indication of non-linear effects. This has also been supported
by low-temperature luminescence spectroscopy where
excitation of the TPD molecules caused a rapid energy
41
Comtransfer to the TTB states even in the 5% sample.
munication between the different dopant molecules is
thus not a problem, as one would have guessed from
their similar molecular structures.
For hole injection, a 0.3 µm layer of α-Se was vapor
deposited on the free surface of the doped polymer films,
followed by a 300 Å Au layer. The sample configuration
is illustrated in Fig. 3. The mobilities were measured
by conventional time-of-flight techniques that have been
208
Journal of Imaging Science and Technology
described in detail elsewhere.42 Here, the displacement
of a sheet of holes generated in the α-Se layer is timeresolved. The exposures were of 3 ns 440 nm radiation
derived from a nitrogen-pumped dye laser (Laser Sciences, Inc., model 337). They were filtered such that the
total charge injected into the sample was less than 2 to
3 × 10–2 CV, where C is the sample capacitance (typically 55 pF) and V the applied voltage. The transients
were measured with a transient digitizer (T ektronix
Model 2301). The transit times were derived from the
intersection of asymptotes to the plateau and trailing
edge of the photocurrent transients in double linear
current versus time representation. With this arrangement, transit times as short as a couple of hundred nanoseconds could be resolved. The mobilities were
determined from the conventional relationship, µ = L2/
(t0•V), where L is the thickness and t 0 the transit time.
The principal limitations of the technique are dielectric
breakdown at high fields and signal-to-noise consider ations at low fields and low temperatures.All measurements were done in an environmental chamber
(Despatch model 900) with a temperature uncertainty
at the sample mount of ±0.2K.
Results
Figure 4 shows typical photocurrent transients measured at room temperature for both TTB concentrations.
Shown are the features usually observed in molecularly
doped polymers: a short initial spike, a plateau of variable length, and a long tail. The plateaus are not well
resolved because the relaxation of charge carriers is
slow, indicating a high degree of disorder in these diluted systems. The dispersion of the transients increases
with decreasing temperature, yet a transition to the
dispersive transport regime was not observed. Figure 5
shows the field dependencies of the mobility at differ ent temperatures for the film that contained 5% TTB
Heun and Borsenberger
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
Hole Transport in Doubly-Doped Polystyrene
(4)
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
2
 2 σ 2

  σ 

µ = µ 0 exp −
− ∑2  E  .
exp C 





kT
kT
3




 
○
○
Analysis
The key parameters of the disorder formalism are the
energetic (or diagonal) disorder σ and the positional (or
off-diagonal) disorder Σ. Both represent the width of
Gaussian distributions that are a consequence of the
structural disorder in amorphous systems. The parameters can be derived from the field and temperature
dependencies of the charge carrier mobilities. For high
fields and Σ ≥ 1.5, the formalism43 predicts:
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
and 15% TPD. The results follow a log µ ∝ E 1/2 relationship over a wide range of fields with a change of sign in
the slope near 315 K (42°C). At low fields, the field dependency shows a departure from the characteristic log
µ ∝ E 1/2 relationship, a feature that has been observed
with 20% TPD doped PS as well. 38 Figure 6 shows the
extrapolated zero-field mobilities plotted semilogarithmically versus T–2 as suggested by the Gaussian disor der model. Despite the presence of a trap, the zero-field
mobilities follow a T–2 dependence almost perfectly, no
additional activation energy needs to be invoked. The
same is true for the 15/5 sample; Figures 7 and 8 show
the results. Here, the change of sign in the slope occurs
at 280 K (7°C).
○
○
○
○
○
○
○
○
○
○
○
○
Figure 4. Room temperature photocurrent transients.
○
○
○
○
○
○
○
○
○
○
Figure 5. Log µ versus E1/2, parametric in temperature, for
5% TTB/15% TPD doped PS.
Figure 6. Log µ (E = 0) versus T –2 for the 5/15 film. The zerofield values of the mobility were obtained by the extrapolation of the data in Fig. 5 to E = 0.
Vol. 43, No. 3, May/June 1999 209
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
○
Figure 7. Log µ versus E1/2, parametric in temperature, for
15% TTB / 5% TPD doped PS.
Figure 8. Log µ (E = 0) versus T –2 for the 15/5 film. The zerofield values of the mobility were obtained by the extrapolation
of the data in Fig. 5 to E = 0.
Here, µ 0 is the mobility at zero field and infinite temperature and C an empirical constant, given as 2.9 ×
10 –4 (cm/V)1/2 for an intersite distance of 6 Å. Eq. 4 is
5
V/
valid only for fields in excess of a few multiples of 10
cm and the temperature range of Tg > T > Tc, where Tg
is the glass transition temperature and T c the
nondispersive to dispersive transition temperature. At
low fields, the disorder formalism predicts that the mobility saturates with decreasing field or , in the case of
large Σ, passes through a minimum, then increases as
the field is further reduced. 44
From Eq. 4, σ and µ0 can be determined from the slope
and intercept of a plot of log µ (E=0) versus T –2. The parameter Σ can be determined from a plot of the slopes
β = δ ln (µ/µ 0) / δ E 1/2 of the high field region of the logµ
versus E1/2 plots, against (σ/kT)2. β versus (σ/kT)2 should
be linear with a slope of 2.9 × 10 –4 (cm/V) 1/2. The parameter Σ is obtained from the β = 0 intercept where
(σ/kT) 2 = Σ 2. From the data in Fig. 6,µ 0 = 6.6 × 10 –4 cm 2/
Vs and σ = 0.105 eV for 5% TTB and 15% TPD (“5/15”)
in PS. For the film with 5% of the antitrap (15% TTB
and 5% TPD, “15/5”, see Fig. 8), the values areµ 0 = 7.4
× 10–4 cm 2/Vs and σ = 0.096 eV. Figure 9 shows the plots
of β versus (σ/kT) 2 for both films. The data giveC = 2.3
× 10–4 (cm/V) 1/2 and Σ = 3.9 for the 5/15, and C = 1.6 ×
10 –4 (cm/V) 1/2 and Σ = 4.1 for the 15/5 sample. The results are summarized in Table I, together with the results from Ref. 38. The room temperature mobilities
in Table I were measured at 24 °C for a field strength
of 2.6 × 105 V/cm.
Discussion
From the results illustrated in the preceding section, it is
clear that the field and temperature dependencies of the
mobility for both films are in agreement with Eq. 4. This
shows that the disorder formalism provides a reasonable
framework for data interpretation as if only one transport
material was present. The data interpretation is self-consistent to the point that the experimental and calculated
values for the temperature where the change in sign of the
slope of the field dependence of the mobility occurs are in
good agreement. From the temperature dependencies of the
mobility, it can be concluded that no polaronic effects need
to be invoked to account for the primary experimental observations of relevance. This is in accordance with the results obtained for polystyrene doped with 20% of the pure
compounds38 as well as for their vapor deposited glasses. 45
Both studies had also shown that one of the few differences
in charge transport behavior between TTB and TPD is the
field dependence of the mobility at low fields. Only TTB
showed the saturation and eventual increase of the mobility in the low field region as predicted by the simulations
the disorder formalism is based upon. For TPD, the mobilities dropped with decreasing field strength. Because both
doubly-doped polystyrenes of this study contained significant amounts of TPD, the measured low field behavior is in
accordance with what had to be expected. The room temperature mobility dropped by a factor of 5 upon doping the
TPD film with 5% of the trap TTB, while the influence of
the antitrap TPD on the TTB mobility was considerably
210
Journal of Imaging Science and Technology
Heun and Borsenberger
0.5
1.5
a
b
1.0
(dlnµ/d√E) / 10–3√(cm/V)
(dlnµ/d√E) / 10–3√(cm/V)
0.0
0.5
0.0
–0.5
–1.0
–0.5
10
12
14
16
18
20
22
(σ/kl)2
–1.5
8
10
12
14
16
18
(σ/kT)2
Figure 9. Slopes β versus (σ/kT)2 for both films. A value of 0.105 eV was used as σ for the 5/15 film (Graph (a)), a value of 0.096
eV as σ for the 15/5 film (Graph (b)).
smaller (about a factor of 2). The prefactor mobility on the
other hand is almost the same for both doubly-doped layers, indicating quite probably that the donor molecule that
is present in the highest concentration dominates the transport behavior at zero-field and infinite temperature.
The most important result of this study concerns the
energy widths of the doubly-doped films. T o our knowledge, a smaller width of the density of states for doublydoped layers compared to those of the doped polymers of
the pure compounds has never been reported. Also, the
observed increase in σ from 0.102 eV to 0.105 eV upon
replacing 5% of the TPD-molecules with the shallow trap
TTB is unreasonably small. Unfortunately, despite the selfconsistency of the data interpretation these results are
not easily understood. From the UV/V is-absorption and
the low temperature luminescence spectra it could have
been expected that the overall width of the DOS was the
sum of two Gaussians of σ = 0.102 eV set off from each
other by an energy of Et = 88 meV and weighed by the
concentrations in which the two compounds were mixed.
Such an approach has been suggested by Pautmeier,Scott
and Schein46 and applied to doubly-doped layers of two
hydrazones doped into polycarbonate at an overall concentration of 50%. However, determining the apparent σs
from time-of-flight experiments for a number of concentrations and recovering the difference in oxidation potential from the results did not lead to one constant offset
energy. The authors took that as an indication that the
disorder formalism fails and that polarons play a significant role in these systems. Polaronic contributions to
charge transport however, have been shown to be at least
negligibly small in TTB and TPD 38,45 and have not been
recovered from the temperature dependence of the mobility in this study either. Moreover, it is not clear whether
simply weighing the contributions to the overall width of
the DOS by the molar ratios is sufficient to predict an over
all width of the DOS since the effects described by Eqs. 1
Hole Transport in Doubly-Doped Polystyrene
and 2 are neglected using such a procedure. Unfortunately
,
the UV/Vis-spectra from Fig. 2 are not resolved well enough
to permit an estimate of the change in bandwidth for TTB
in the doubly-doped films.
Regardless of these influences should the sum of two
Gaussians offset from each other by some constant energy
Et render a distribution that is broader than theoriginal,
and, even more importantly , should the width of the
overall density of states show a symmetrical dependence
on concentration with a maximum at equal concentration of both donor molecules. Because it does not matter whether sites are added at the upper or lower end of
the energetic scale, the described treatment predicts an
energetic disorder parameter of σ = 0.109 eV for both
concentrations investigated in this article, which is in
obvious disagreement with the experimental results.
Another way to interpret the data is given by the trapping arguments from Refs. 30 and 31, and the dipolar
disorder model. The simulations in Ref. 30 predict that
the required critical trap concentration cc for a trap
depth of Et = 88 meV is about 10%, which is identical to
the maximum concentration for trap limited transport.
Therefore, Eq. 3 does not apply to relative trap concentrations of 25% in the 5/15 and 75% in the 15/5 film.
(Calculating the apparent widths of the DOS according
to Eq. 3 anyway, renders σ eff = 0.116 eV for 5% and even
σ eff = 0.122 eV for 15% TTB.) For concentrations in excess of the trap limit, hopping within the trap manifold
rather than conventional trapping should dominate. In
that case however, the samples should behave like films
that contain 5 or 15% TTB, respectively. Due to the similar dipole moments of both transport molecules and the
identical overall dopant concentration, the dipolar contribution to the overall width of the DOS can be assumed
to be constant for both TTB concentrations, whereas the
van der W aals components for triphenylamine doped
polystyrene should be σvdW = 0.116 eV for a concentra-
Vol. 43, No. 3, May/June 1999 211
tion of 5% and σvdW = 0.104 eV for a concentration of
15%.27 According to Eq. 1, this yields widths of the DOS
of σ = 0.116 eV and σ = 0.104 eV, respectively, using the
dipolar contribution as calculated from Eq. 2 orσ = 0.121
eV and σ = 0.109 eV for a dipolar component computed
from the experimental width of σ = 0.102 eV from Ref.
38 and a van der Waals component27 of σ vdW = 0.096 eV
for 20% triphenylamine doped PS. Here again, the width
of the DOS should increase with decreasing TTB concentration as long as it stays above the trap limit, which
disagrees with the experimental results. It should be
noted that even the experimental value of σ = 0.105 eV
for the 5/15 film cannot be explained on the basis of such
an approach because the broadening due to dilution of
the main transport molecule is too small. Therefore this
treatment does not lead to a satisfactory explanation of
the results either, although they may suggest that the
van der Waals component is not as independent of the
nature of the matrix (here: PS in comparison with PS
doped with a certain percentage of triphenylamine derivative) as previously suggested.27
The same is true for the positional disorder parameter Σ. The disorder formalism assumes that positional
disorder is due to a distribution of intersite distances
and mutual orientations,43 both of which influence the
coupling between dopant molecules and therefore affect hopping transport. Because the molecules of this
study are of very similar size and geometry and render films of very similar positional disorder when doped
into PS alone, it seems reasonable to assume that intermolecular distances and orientations remain the
same in the doubly-doped films. If both dopants par ticipated in the charge transport process, Σ should
therefore stay constant as well. If however the ener getics precluded the antitrap from contributing,
intersite distances between the actual hopping sites
would increase with increasing dilution. Again both
doubly-doped polymers should show an increase in Σ.
Both predictions made on the basis of the definition of
positional disorder in the formalism are in disagreement with the experimental results.
Overall, it is the doubly-doped polystyrene with 5%
of the antitrap that renders the most inexplicable results. The trends in the doubly-doped layers with 5%
TTB do at least go in the right direction, with a broader
σ, the lowest room-temperature and zero-field mobilities, and a positional disorder parameter Σ that is close
to the one of polystyrene doped with 20% TPD. All this
is not the case for the mirror concentration, even the constant C for this sample is significantly farther removed
from the predicted value of 2.90 (cm/V)1/2 than for the other
compositions. The latter suggests that there is some other
effect aside from energetic and positional disorder that
influences charge transport in doubly-doped polymers
despite the agreement with the predicted field and temperature dependencies of the mobility.
Summary
Hole tranport has been measured in polystyrene layers
doped with two mixtures of TTB and TPD at an overall
concentration of 20% by weight. The mobilities follow
the field and temperature dependencies as predicted by
the formalism based on disorder due to Bässler and coworkers, but the results for the key parameters ener getic and positional disorder are inconsistent with
existing models to predict them. Because polaronic contributions to charge transport can be excluded in this
212
Journal of Imaging Science and Technology
class of materials, the origin of the deviation from pure
disorder controlled hopping could not be determined.
Acknowledgment. We thank M. Detty for the synthesis of TPD, B. Henne for measuring the oxidation potentials, and R. H. Young and H. Bässler for many
helpful discussions. S. Heun would also like to thank
the Office Imaging Division of Eastman Kodak Company
for a financial grant and the Deutsche Forschungsgemeinschaft for a graduate scholarship.
References
1. D. M. Pai and B. E. Springett, Rev. Mod. Phys. 65, 163 (1993).
2. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for Xerography, Marcel Dekker, New York, 1998.
3. H. Antoniadis, B. R. Hsieh, M. A. Abkowitz, and M. Stolka, Appl. Phys.
Lett. 62, 3167 (1993).
4. J. Pommerehne, H. Vestweber, W. Guss, R. F. Mahrt, H. Bässler, M.
Porsch, and J. Daub, Adv. Mat. 7, 551 (1995).
5. W. E. Moerner and S. M. Silence, Chem. Rev. 94, 127 (1994).
6. G. G. Malliaras, V. V. Krasnikov, H. J. Bolink, and G. Hadziioannou,
Appl. Phys. Lett. 66, 1038 (1995).
7. C. W. Tang, Appl. Phys. Lett. 46, 183 (1986).
8. H. Bässler, Phys. stat. sol. (b) 175, 15 (1993), and references therein.
9. H. Bässler, Int. J. Mod. Phys. B 8, 847 (1994).
10. H. Bässler, Mol. Cryst. Liq. Cryst. 252, 11 (1994).
11. A. Miller and E. Abrahams, Phys. Rev. 120, 745 (1960).
12. P. M. Borsenberger, E. H. Magin, M. v. d. Auweraer, and F. C. De
Schryver, Phys. stat. sol. (a) 140, 9 (1993), and references therein.
13. P. M. Borsenberger, E. H. Magin and S. Heun, Macromol. Symp. 116,
51 (1997).
14. P. M. Borsenberger and D. S. Weiss, J. Imaging Sci. Technol. 41, 185
(1997).
15. P. M. Borsenberger, J. R. Cowdery-Corvan, E. H. Magin, and J. A.
Sinicropi, Thin Solid Films 307, 215 (1997).
16. P. M. Borsenberger, H.-C. Kan, E. H. Magin, and W. B. Vreeland, J.
Imaging Sci. Technol. 39, 6 (1995).
17. P. M. Borsenberger and W. T. Gruenbaum, J. Polym. Sci. Polym. Phys.
34, 575 (1996).
18. M. A. Abkowitz, M. J. Rice and M. Stolka, Phil. Mag. B 61, 25 (1990).
19. M. Gailberger and H. Bässler, Phys. Rev. B 44, 8643 (1991).
20. H. Bässler, P. M. Borsenberger and R. J. Perry, J. Polym. Sci. Polym.
Phys. 32, 1677 (1994).
21. P. M. Borsenberger, E. H. Magin, M. v. d. Auweraer, and F. C. De
Schryver, Phys. stat. sol. (b) 186, 217 (1994).
22. P. M. Borsenberger, W. T. Gruenbaum, E. H. Magin, and L. J. Rossi,
Jpn. J. Appl. Phys. 35, 6135 (1996).
23. P. M. Borsenberger, E. H. Magin and J. Shi, Physica B 217, 212 (1996).
24. P. M. Borsenberger, W. T. Gruenbaum and E. H. Magin, Jpn. J. Appl.
Phys. 35, 2698 (1996).
25. J. A. Sinicropi, J. R. Cowdery-Corvan, E. H. Magin, and P. M.
Borsenberger, Chem. Phys. 218, 331 (1997).
26. P. M. Borsenberger and H. Bässler, Phys. stat. sol. (b) 170, 291 (1992).
27. P. M. Borsenberger, W. T. Gruenbaum, E. H. Magin, and L. J. Sorriero,
Chem. Phys. 195, 435 (1995).
28. W. T. Gruenbaum, E. H. Magin, and P. M. Borsenberger, Jpn. J. Appl.
Phys. 35, 2704 (1996).
29. R. H. Young, Phil. Mag. B 72, 435 (1995).
30. U. Wolf, H. Bässler, P. M. Borsenberger, and W. T. Gruenbaum, Chem.
Phys. 222, 259 (1997).
31. P. M. Borsenberger, E. H. Magin and S. A. Visser, Jpn. J. Appl. Phys.
37, 1945 (1998).
32. G. Verbeek, M. van der Auweraer, F. C. De Schryver, C. Geelen, D.
Terrell,and S. De Meuter, Chem. Phys. Lett. 188, 85 (1992).
33. B. Hartenstein, H. Bässler, S. Heun, P. M. Borsenberger, M. van der
Auweraer, and F. C. De Schryver, Chem. Phys. 191, 321 (1995).
34. S. Heun, H. Bässler and P. M. Borsenberger, Chem. Phys. 200, 265
(1995).
35. M. Stolka, J. F. Yanus and D. M. Pai, J. Phys. Chem. 88, 4707 (1984).
36. M. Abkowitz and D. M. Pai, Phil. Mag. B 53, 193 (1986).
37. B. Henne, priv. Comm.
38. S. Heun and P. M. Borsenberger, Physica B 216, 43 (1995).
39. P. M. Borsenberger and J. J. Fitzgerald, J. Phys. Chem. 97, 4815 (1993).
40. F. Ullmann, Ber. 36 2382 (1903).
41. S. Heun, Thesis, Marburg 1995.
42. A. R. Melnyk and D. M. Pai, Physical Methods of Chemistry, 2nd ed., B.
W. Rossiter and R. C. Baetzold, Eds., Vol. 8, John Wiley and Sons,
New York, New York, 1993.
43. H. Bässler, Phys. stat. sol. (b) 175, 15 (1993).
44. L. Pautmeier, R. Richert and and H. Bässler, Synth. Met. 37, 271 (1990).
45. S. Heun and P. M. Borsenberger, Chem. Phys. 200, 245 (1995).
46. L. Th. Pautmeier, J. C. Scott and L. B. Schein, Chem. Phys. Lett. 197,
568 (1992).
Heun and Borsenberger
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Thermally Stimulated Luminescence in Molecularly Doped Polymers
A. Kadashchuk,* N. Ostapenko, V. Zaika
Institute of Physics, National Academy of Sciences of Ukraine, Kiev, Ukraine
P. M. Borsenberger
Office Imaging Division, Eastman Kodak Company, Rochester, New York, USA
The low-temperature thermally stimulated luminescence (TSL) technique has been applied for the first time for probing the
energetic disorder of localized states in molecularly doped polymers (MDPs) as substituted triphenylamines doped into poly(styr
ene)
(PS). TSL of both the neat MDPs, as tri- p-tolylamine (TTA) and tri-p-anisylamine (TAA) doped PS, and the doubly doped polymers, as TTA doped PS containing small concentrations of di-p -anisyl-p-tolylamine (DAT), was studied. The results are described
in terms of the Gaussian disorder model and the energetic relaxation of photogenerated charge carriers, that provide reasonable
understanding of all observed trends in the TSL.Analysis of both the energetic position of the TSLpeak maximum and the shape
of its high-energy wing allowed extraction of a parameter characterizing the energetic disorder in MDPs, which agreed well with
the width of the density-of-states determined from charge transport measurements. The effect of extrinsic trapping because of
DAT on TSL properties can be reasonably interpreted in terms of the effective energetic disorder and the TSLresults are in good
agreement with those obtained earlier by charge transport studies.
Journal of Imaging Science and Technology 43: 213–219 (1999)
Introduction
In recent years, much attention has been paid to charge
carrier transport phenomena in molecularly doped polymers (MDPs) because of their practical importance. 1
Several applications could be mentioned: organic electrophotographic photoreceptors,1,2 electro-luminescent 3
and electrophotovoltaic 4 devices, and photorefractive
media,5 etc. MDP typically consists of charge-transporting molecules (strong electron donors or acceptors capable
of transporting charge carriers) randomly dispersed in
an electrically inert polymer host acting as a binder
. It
is generally accepted that hole or electron transport
occurs by thermally activated hopping between adjacent donor or acceptor molecules, respectively.
MDPs represent examples of amorphous systems
where charge transport has most often been described
within the framework of the Gaussian disorder model 6
of Bässler. This model is based on disorder controlled
hopping of charge carriers through a manifold of localized states with a Gaussian density-of-states (DOS) distribution. The distribution of site energies is often
caused by energetic disorder and is characterized by the
width of the DOS, σ (parameter of the model). Most studies of charge transport in MDP have been based on conventional photocurrent transient measurements and
performed on systems where charge trapping could be
neglected. Recently, Wolf and co-workers 7 and
Borsenberger and co-workers 8 studied the effect of ex-
Original manuscript received November 2, 1998
* corresponding author; E-mail: kadash@iop.kiev.ua
© 1999, IS&T—The Society for Imaging Science and Technology
trinsic traps on the transport properties of charge carriers in doubly doped polymers, and extended the disorder
formalism to include the effect of trapping. It was shown
that the effect of shallow traps can be accounted for by
the replacement of σ with an effective width, σeff, dependent on both trap depth and concentration, the basic phenomenology of transport remaining unaltered.7,8
The aim of the present work is to explore the use of
low-temperature thermally stimulated luminescence
(TSL) spectroscopy for probing the energetic disorder of
localized states in MDPs (both the neat systems and
those containing extrinsic shallow traps). Recently, the
applicability of the TSL method for evaluation of the
energetic disorder parameter in carbazole pendant group
polymers9 and σ-conjugated polysilylenes10,11 has been
demonstrated. It is known 12–14 that the DOS of disor dered organic systems is not subject to direct optical
probing, therefore, the only methods suitable for such
purpose are those that are based on thermally induced
transitions among the manifold of carrier states. 9-11,13
Our approach is based on the assumption that the lowest energy portion of an energetically disordered manifold of localized states at very low temperatures may
manifest itself in TSL as trapping centers for charge carriers. Actually, at the low temperature limit, kT << σ, the
charge carrier hopping occurs toward states of lower and
lower energy, until, on reaching the band tail, the concentration of such states is so low that further hops are
impossible and, consequently, the charge carriers finally
become trapped in the tail sites, which are the local
minima of the site energy. Therefore, by analyzing the
trap distribution function, one may estimate the shape
of the deepest part of the DOS distribution and, consequently, the parameter of the energetic disorder, σ.
This approach is conceptually similar to that employed
earlier by Bässler 13,15 using the thermally stimulated
213
current (TSC) technique for studying the width of the
DOS distribution in disordered tetracene layers by analyzing the shape of the high-energy wing of a TSC peak.
In this connection it should be mentioned that TSC technique has attracted increasing attention recently in the
hope of obtaining important information on the DOS distribution in the bandgap of disordered solids. 16-20 However, as previously shown,16,17 a basic limitation for the
application of the TSC method for DOS probing, inhopping transport systems with very low charge mobility ,
seemed to be due to that the TSC peak can be irrelevant
to direct detrapping. This is because it is influenced by
transport process, i.e., it is determined by the increasing mobility of carriers with increasing temperature (socalled “transport peak”) and, therefore, its position on
a temperature scale is field and thickness dependent.16
On the other hand, the main advantage of applying
the TSL method for systems with very low charge mobility, in contrast to other techniques, is that TSL, if
caused by geminate recombination, is not influenced by
the transport processes and is directly related to
detrapping. 16–18 The geminate character of the TSL can
be proved by quenching the TSL intensity with an applied electric field. 9,16,18 Of special relevance for disor dered polymer systems is the employment of the method
of TSL with temperature modulation, the so-called fractional TSL, 21–25 which is a useful tool for determining
trap depths when different groups of traps are not well
separated in energy, or are continuously distributed
(which, in our experience, is a very typical situation for
amorphous polymers), and it allows the analysis of the
trap spectra even when they are complex. The fractional
TSL technique (also called the fractional glow technique), that has long been proposed by Gobrecht and
Hofmann, 21 is an extension of the initial rise method
and is based on cycling the sample with a large number
of small temperature oscillations superimposed on a
uniform heating. The main reason for applying this
method is that the usual quantitative evaluation of the
TSL glow curves is very inaccurate, or even impossible
if the traps are continuously distributed in depth or if
the trap spectrum consists of several types of traps with
discrete but very close lying activation energies. In this
case, the glow peaks fuse one into another so that many
different traps may produce a composite peak and may
be scarcely (or even not at all) indicated by individual
glow maxima. The width and symmetry of such composite glow peaks are no longer suitable for ordinary
derivation of trapping parameters.21 It is recognized21-23
that the method of fractional TSL avoids the disadvantages of the common glow curve methods, is characterized by greater accuracy and a high resolving power ,
and it does not require a knowledge of the frequency
factors and retrapping probabilities.21, 22
In the present work, we report the first TSL study of
some substituted triphenylamines doped into
poly(styrene) (both the neat trap-free MDPs and trapcontaining systems). The TSL behaviors of the MDPs
studied are interpreted in terms of the Gaussian disor der model and the energetic relaxation of photogenerated
charge carriers within a manifold of states of Gaussian
distribution.
Experimental
The molecular structures of the compounds used in this
study are illustrated in Fig. 1. Charge transporting
molecules (CTM): tri- p-tolylamine (TTA) and tri- panisylamine (TAA), as well as molecules of di-p -anisylp-tolylamine (DAT) acting as a weak hole trap were
214
Journal of Imaging Science and Technology
Figure 1. The molecular structures of tri- p-tolylamine (TTA),
tri-p-anisylamine (TAA), di-p-anisyl-p-tolylamine (DAT), and
poly(styrene) (PS).
supplied by Eastman Kodak Company . Poly(styrene)
(PS) was provided by the Institute of Chemistry, Acad.
of Sci. of Ukraine. All materials were used as received.
Samples were prepared by dissolving the appropriate
ratios of CTM and PS in dichloromethane, then the
resulting solutions were cast on a metal substrate. The
films were dried for 3 h at 40 o C in air, then at room
temperature for 2 h in vacuo. All samples of the neat
MDP contained 40% TTA (or TAA) with respect to PS.
The CTM concentrations are expressed as the wt.% of
the total, which is equivalent to the mass ratio. In the
case of doubly doped polymers, the concentration of the
DAT ranged from 0.05 to 4 wt.% with respect to TT A.
The solids concentration of the coating solution was
10%.
TSL measurements were carried out with an automatic equipment for optical thermoactivated spectroscopy over a wide temperature range from 4.2 K to 350 K
with an accuracy better than 0.1 K. The polymer samples
were mounted in a holder of the optical helium cryostat
and, after cooling, they were irradiated with UV light.
For excitation, the light from a high-pressure 500 W
mercury lamp was used. After terminating the excitation the luminescence signal was detected with a cooled
photomultiplier operated in photon-counting mode. TSL
measurements were performed in two different regimes;
under the uniform heating with the rate β = 0.15 K/s
and in the fractional heating regime.
Our fractional TSL technique, and the procedure of
data processing, were similar to that described in Refs.
22, 25. The main difference was in the temperature
range extension from the commonly used liquid-nitrogen temperature down to the liquid-helium temperature.
The mean activation energy <E> is determined during
each temperature cycle as
<E> = –d [ln I (T)]/d (1/kT),
(1)
where I is the intensity of the thermoluminescence, T
is the temperature in the measuring cycle, and k is the
Boltzmann constant. A trap distribution function, H (E),
can be determined in arbitrary units as:22
H (E) ~ I (T)/(d < E>/dT).
(2)
Kadashchuk, et al.
Figure 2. (a) TSL glow curve under excitation with unfiltered light of Hg lamp for 30 s at 4.2 K for TT
A:PS (Curve 1) and TAA:PS
(Curve 2); Curves 1' and 1'’ represent deconvolutions using Gaussian functions. The temperature dependence of the mean activation energy <E> as obtained by fractional TSL (inset); extrapolation used an empirical expression (4) to give by the solid line in
the inset. (b) TSL peak of TT A:PS obtained after 30 min exposure to additional IR irradiation (Curve 1), after standard UV
excitation for 30 s. The transparency band of the IR-filter used in experiment was 900-4500 nm.
The frequency factor at the maximum of TSL peak, S
is given by the formula:
S = <Em>β/kT m2 exp (<Em>/kTm),
(3)
where Tm and <E m> are the temperature and the activation energy of the maximum of the TSL peak, respectively. All measurements were done in the helium
atmosphere.
Results
Trap-Free Polymer Systems. It was found that the
neat TTA-doped PS, as well as systems containing extrinsic traps, showed very strong thermoluminescence
induced by UV-radiation at liquid helium temperature.
Figure 2(a) presents a typical TSL glow curve for the
neat TTA (40%) doped PS film (TT A:PS) (Curve 1). As
one can see, the low-temperature TSLunder UV-excitation at 4.2 K arises immediately after sample heating
has begun to heat up and extends to ~150 K. The existence of a quasi-continuous trap distribution in this sys-
tem has been found. The mean activation energies, <E>,
as revealed by the fractional TSL [Fig. 2(a), inset], linearly increase with temperature according to the following empirical formula (in eV)
<E>(T) = 0.0032 × T – 0.08
(4)
These results indicate the lack of the charge-carrier
deep trapping in studied the MDP and on the presence
of a large concentration of shallow localized states capable of charge carrier capturing at 4.2 K.
The TSL glow curve of TTA:PS is evidently composed
of two peaks: low-temperature with maximum at Tm ≅
35 K and high-temperature peak atTm ≅ 75–78 K. Curves
1′ and 1 ′′ in Fig. 2(a) present the separation the TSL
glow curve into two Gaussians. It should be noted that
the Gaussian function used for the approximation of the
high-temperature TSL peak [Fig. 2(a), Curve 1 ′] has a
width of about 25 K. In energy terms this is equivalent
to 0.08 eV because of Eq. 4, (25 × 0.0032 = 0.08). The
activation energy and frequency factor in the maximum
Thermally Stimulated Luminescence in Molecularly Doped Polymers
Vol. 43, No. 3, May/June 1999
215
of the above-mentioned TSL peaks are <Em> = 0.032 eV
and S = 2 × 10 3 s–1 (for the peak at 35 K), and < Em> =
0.16 ÷ 0.17 eV and S = 3 × 109 s–1 (for the peak at 75 K),
respectively.
It should be pointed out that the low- and high-temperature TSL peaks of TTA:PS most likely have somewhat different physical origins as supported by their
different frequency factors. Frequency factors of order
3 × 109 s–1 for the high-temperature peak is quite typical
for amorphous photoconducting polymers: a comparable
value was obtained for the main TSL peak of polyvinylcarbazole (PVK) 25 (10 8 s –1) and poly(methylphenylsilylene) (PMPSi) 10 (10 10 s –1), while the anomalous low
value of order 10 3 s–1 for the low-temperature peak at
35 K could be a result of recombination of an electron
and hole (probably closely situated) by charge tunneling. It is a well-known fact that a low S-value is often
associated with under -barrier tunnel transitions. The
detailed study of this problem will be presented elsewhere. We note that a similar (but not always so pronounced) low-temperature feature at T ≤ 50 K is
inherent for all polymers we have studied, including
PMPSi and PVK-type polymers, and it seems to be of
secondary importance in the present study . Therefore,
in this work we will focus our attention mainly on the
high-temperature peak, which will be referred to hereafter as the main TSL peak.
We have found a very interesting and useful method
of separation of the high-temperature TSL peak by the
additional exposure of a sample to IR-irradiation at
4.2 K following the conventional UV -excitation (“IRcleansing” effect). This effect was reported earlier in
the TSL study of PMPSi10 and was explained in terms
of the energetic relaxation of photogenerated charge
carriers within the Gaussian shaped DOS. IR-excitation to the highest portion of the DOS involves an increase in the number of new sites a carrier visited at
4.2 K and, consequently , leads to an increase in the
probability of reaching lower energy tail states. Figure 2(b) presents the TSL glow curve of TT A:PS obtained as a result of the additional IR-cleansing for 30
min at 4.2 K. Such IR-exposure leads to near complete
cleansing of the high-temperature TSL peak; and although the broadness of the peak is not changed, its
maximum is slightly shifted towards higher temperature and only a weak shoulder persists instead of the
previously strong low-temperature peak. It should be
noted that no sample heating occurred in the process,
and the sample was immersed in liquid helium. The
cleansed TSL peak of TTA:PS can be successfully fitted by a Gaussian with the half-width about 0.08 eV
[solid curve in Fig. 2(b)].
It is of obvious interest to compare the above results
with TSL data of another trap-free MDP system. Curve
2 in Fig. 2(a) presents TSL glow curve for the neat TAA
(40%) doped PS (T AA:PS) film. As one can see, the
TAA:PS has similar TSL behavior, except that the corresponding TSL peaks are notably shifted towards
higher temperatures within respect to that of TTA:PS.
It was found that the main TSL peak of TAA:PS is centered at Tm ≈ 110 K and the activation energy in the
peak maximum comprises <Em> = 0.24 eV.
Trap-Containing Polymer Systems. TTA doped PS
containing di-p-anisyl-p-tolylamine (DAT) provides a
good example of MDP where transport is trap-affected.8
DAT is a shallow trap with depth8 of 0.15 eV because of
the lower potential of ionization relative to TT A. TSL
study of MDPs containing other shallow traps has re-
216
Journal of Imaging Science and Technology
vealed qualitatively similar behaviors and will be presented elsewhere. Figure 3(a) presents TSL glow curves
of TTA:PS containing different concentration of DAT: c
= 0%, 0.05%, 0.24%, 1%, 4% (Curves 1, 2, 3, 4, 5, respectively). All curves in Fig. 3(a) are normalized to the
maximum intensity of the high-temperature (main) TSL
peak. As one can see, even a small concentration of DA
T
traps exerts a rather considerable effect on the TSL.
The most characteristic property is the considerable
shift of the main TSL peak towards higher temperatures
with increasing trap concentration. This peak shifts
from 75 K at zero DAT concentration to ≈115 K for c =
4% (Fig. 3(a), Curve 5).
It is worth noting that the effect of DAT traps can be
seen more clearly after the IR-cleansing of the main TSL
peak [Fig. 3(b)]. Relevant TSL glow curves detected after the additional exposure of the sample to IR-radiation for 30 min at 4.2 K are presented in Fig. 3(b) (all
curves are normalized at the maximum intensity). The
dependence of <E> on temperature (as measured by the
fractional TSL technique) coincided well with that given
by Eq. 4 for DA T concentration c ≤ 0.24%, however, a
slight deviation it was found for systems containing DA
T
with c ≥ 1%. The activation energy at the maximum of
the high-temperature TSL peaks was equal to 0.17, 0.2,
0.215, 0.264 eV, for DAT concentration c = 0%, 0.24%,
1%, 4%, respectively.
Analysis and Discussion
Let us first consider the trap-free MDP systems.
Charge transport in TTA:PS and TAA:PS has been extensively studied and described in terms of the Gaussian
disorder model. 26–29 The width of the DOS distribution
for the low-polarity material TTA (40%) doped PS, σ =
0.083 eV, was determined by Borsenberger 26–28 and
Young. 29 On the other hand, the degree of energetic disorder for TAA (40%) doped PS has been found27,28 as notably larger, σ = 0.107 eV , due to the considerable
contribution of the dipolar component into the total
width (the dipole moment of TAA is 2 D28). It was shown
that polaronic effects need not be invoked to explain
charge transport properties in the above MDPs.
It is assumed that, just as in the case of earlier studied PMPSi, 10,11 the main TSL peak of TTA:PS is associated with thermal release of charge carriers occupying
the intrinsic tail states with the DOS distribution of a
Gaussian shape
 E2 
H ( E ) ~ exp − 2 
 2σ 
(5)
where E is the energy of the localized state with respect
to the DOS center. It is supposed that a trapped charge
carrier, to become mobile again, should be thermally
excited to the level of the so-called transport energy ,30
E* , which normally is located bellow the center of the
DOS. To consider the transport energy position, the following criterion30 as a first approximation seems to be
a reasonable choice for the case of the low-temperature
range relevant to the TSL experiment. For the Gaussian
shaped DOS, of a system with sixfold coordination, E *
could be defined by the condition:30
E*
∫ f ( E )dE = 1 / 6, (numerically : E* = σ )
−∞
(6)
Kadashchuk, et al.
Figure 3. (a) TSL glow curves of TTA:PS containing different concentration of DAT: c = 0%, 0.05%, 0.24%, 1%, 4% (Curves 1, 2, 3, 4, 5,
respectively). All curves are normalized at the maximum intensity of the high-temperature TSL peak; (b) TSL peaks of the same
systems obtained after exposition to additional IR irradiation. Experimental conditions were the same as presented in Fig. 2.
where f(E) is a Gaussian function. Thus, the experimentally measured thermal release energy, <E>, in the TSL
data analysis is identified with the energy of a localized state E, with respect to the transport energy position, E*:
<E>= E – E*
(7)
We should emphasize that taking into consideration
the above empirical linear dependence (Eq. 4) and Eq. 2
from the theory 22 of the fractional TSL, the temperature profile of the TSL peak,I(T), should reflect the trap
distribution function, H(E). The results of Gaussian
analysis of the high-temperature wing of the TSL peak
of TTA:PS, i.e., ln(I(T)) versus E2, made by converting
the temperature scale to the trap energy scale using the
empirical calibration (Eq. 4) as well as Eqs. 6 and 7, is
presented in Fig. 4, Curve 1. The following conclusions
can be drawn from the above analysis:
(i) the high-energy wing of TSL peak can be well approximated by a Gaussian distribution (cf. Eq. 5);
(ii) the width of this distribution, σ’, formally deter mined by taking its inclination angle tangent, yields
the value σ’ = 0.082 eV (Fig. 4, solid line 1).
This value agrees well with the width of the DOS, σ =
0.083 eV,7,8,26-29 obtained by the charge transport measurements. For comparison purposes, the Gaussian
analysis of the same peak when ignoring the consider ation of transport energy below the center of the DOS
(assuming that < E> = E in contrast to Eq. 7) is presented by Curve 2 in Fig. 4. As one can note, neglecting
of the transport energy concept leads to smaller value
of σ’ = 0.074 eV estimated from the slope of the solid
line 2 in Fig. 4.
As was shown earlier , 10 the disorder parameter of
amorphous photoconducting systems could be evaluated
from TSL data also by alternative means, namely by
analyzing the activation energy value in the maximum
of TSL peak, <E m>. Actually, a charge carrier after its
photogeneration occurring at the TSL excitation at 4.2
K, is subjected either to geminate recombination or to
trapping by shallow localized states. The latter process
Thermally Stimulated Luminescence in Molecularly Doped Polymers
Vol. 43, No. 3, May/June 1999
217
Figure 4. Gaussian analysis of the high-energy wing of the TSLpeak of TTA:PS (see text
for explanations) when using Eq. 7 (Curve 1),
and with the assumption E = < E> ignoring
the transport energy concept (Curve 2). Solid
lines 1 and 2 are extrapolations withσ’ = 0.082
and 0.074 eV, respectively.
could be well described by the theory developed by
Movaghar 31 and B ässler 14 for energetic relaxation of
particles within the Gaussian shaped manifold of states.
According to this theory, the energetic relaxation of randomly generated particles in the zero-temperature limit
occurs through a nonactivated downward hopping with
the decay in the energy on the level of the time scale
given by the formula:14, 31
ER = σ [δ ln ln (t/t0)] 1/2
(8)
where ER is the mean energy of the charge carrier packet
relative to the center of DOS distribution, δ is the dimensionality of the lattice (usually taken as 3), 14,31 t is
the time, and t0 is the dwell time of a carrier at a lattice
site without disorder (for a charge carrier, t0 is accepted
as 10–13 s).6 Hence, the parameter σ could be estimated
from the assumption that the experimentally accessible
activation energy <Em> corresponds to the mean energy
of the relaxed charge-carrier packet, ER (given by Eq. 8).
Actually, assuming σ = 0.083 eV and using the expression (8) for the experimental time scale of 10 2 s, we obtain the value ER = 0.27 eV. The experimental value of
the activation energy of the maximum of the TSL peak
within respect to the center of the DOS (given as E =
<E> + E* due to Eq. 7) is 0.253 eV in reasonable agreement with the above calculated value. As one can see,
the inclusion of the transport energy by means of Eq. 7 is
of importance for the analysis of the energetic position of
the TSL peak maximum and leads to better coincidence
between calculated and experimental values (note that
experimental value without considering the transport
energy according to Eq. 4 is only <Em> = 0.17 eV).
Using this line of reasoning, the relationship between
the experimentally measured <Em> and the value of ER
calculated by Eq. 8 could be expressed as follows: <Em>
= ER – E* = ER – σ. It follows, that a rather simple means
of evaluation of the σ-value from the experimentally
accessible value of <Em> is:
σ' =
< Em >
[3 ln ln(t / t ) ] − 1
0
1/ 2
,
(9)
According to Eq. 9 one obtains σ’ = 0.075 eV for the
TTA:PS system.
The TSL properties of the T AA:PS system could be
218
Journal of Imaging Science and Technology
Figure 5. Correlation between the activation energy in the
TSL peak maximum, <Em>, and σ obtained from charge transport measurements for TT A:PS, 7 PMPSi, 10 TAA:PS, 27 and
PVK:TNF7.
well interpreted in a similar manner , but taking into
account the fact that degree of energetic disorder in this
system is larger than that for TTA:PS (σ = 0.107 and 0.083
eV, respectively).27,28 According to Eq. 8, <Em> value is expected to be also larger, as was observed experimentally:
<Em> = 0.24 eV. The disorder parameter for TAA:PS can
be estimated by Eq. 9 which gives σ’ = 0.106 eV.
Kadashchuk, et al.
It should be emphasized, that the above-mentioned
interpretation of TSL data presumes a linear relationship between the activation energy in the TSL peak
maximum and the degree of energetic disorder in amorphous systems. Such behavior has been observed and it
seems not to depend on the type of polymer . Figure 5
presents collected charge transport and TSL data for
some different polymer systems, which are plotted as
<E m> versus σ. As one can see, there is a striking correlation between those data and the experimental points
are close to falling on a straight line.
Finally, let us consider TSL behaviors of TTA:PS containing extrinsic traps caused by DAT. Transport properties of these MDPs have been studied recently 8 by
Monte Carlo simulations and mobility measurements.
The principal conclusion was that the system behaves
as if the addition of traps has widened the DOS and the
basic features of transport can be modeled in terms of
the disorder concept using σeff as the essential material
parameter. TSL data presented in Fig. 3 agrees well with
such a notion. The characteristic high-temperature shift
of the main TSL peak with increasing trap concentration is most likely indicative of an increase in the σ eff,
while no new features were observed in the TSL glow
curve in comparison with the trap-free system. Thus,
the general picture of the TSL behavior of TTA:PS containing varying concentrations of DA T is undistinguished from the neat TT A:PS. Using measured < Em>
values, one can estimate the energetic disorder parameters by means of Eq. 9 as σ’ = 0.075, 0.088, 0.095, and
0.116 eV for concentrations of DA T equal to 0, 0.24, 1,
and 4%, respectively. These values agree reasonably
with parameters σ eff = 0.083, 0.095, 0.101, and 0.107 eV
earlier obtained 8 from the charge transport measurements for the same DA T concentrations. It should be
noted that the high-temperature TSL peak, which manifested a strong shift towards higher temperatures with
an increase of DA T concentration, does not reflect directly the value of DAT trap depth, because at c(DAT) =
4% the value of <E m> is 0.264 eV that is far larger than
the trap depth of 0.15 eV. The origin of this peak is due
to superimposition of energetic disorder and trapping
effects, and it can be characterized by the effective disorder parameter.
Conclusion
The interpretation of low-temperature TSL of MDPs as
associated with charge carrier thermal release from intrinsic tail states of the DOS distribution is suggested.
Such an approach, based on the Gaussian disorder
model, provides a reasonable understanding of all observed trends in the TSL. According to this, the shape
of the high-temperature wing of the TSL peak and the
energetic position of its maximum could be explained
consistently incorporating the concept of the transport
energy. Gaussian analysis of the high-temperature wing
of the TSL peak of TTA:PS yielded the width of states
profile for localized charge carriers equal to σ’ = 0.082
eV (Fig. 4, Curve 1). It was shown that the position of
the TSL peak maximum is in accord with the predictions of the theory for non-activated energetic relaxation
of photogenerated carriers within a Gaussian-shaped
manifold of localized states, and, therefore, no additional
features of the DOS in the gap are necessary for the
existence of the low-temperature TSL peak. The activation energy in the TSLpeak maximum of TTA:PS, which
under the consideration of the transport energy concept
is equal to 0.253 eV , agreed satisfactorily with calculated value of 0.27 eV . TSL of TAA:PS could be inter preted in similar manner taking into account the larger
degree of energetic disorder because of the considerable
dipolar disorder in comparison with the former system.
The effect of extrinsic traps in TT A:PS containing
varying concentrations of DA T on TSL properties can
be interpreted in terms of the effective energetic disorder. The origin of the TSL peak for such doubly doped
MDPs is due to superimposition of energetic disorder
and trapping effects, and it can be characterized by the
effective disorder parameter.
Acknowledgments. The research described in this article was made possible in part by Award No. UE1-326
of the U.S. Civilian Research & Development Foundation for the Independent States of the Former Soviet
Union (CRDF). Discussions with H. B ässler, M.
Abkowitz and R. Young are gratefully acknowledged.
References
1. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for Imaging Systems, Dekker, New York, 1993.
2. K.-Y. Law, Chem. Rev. 93, 449 (1993).
3. C. Adachi, K. Nagai and N. Tamoto, Appl. Phys. Lett. 66, 2679 (1995).
4. C. W. Tang, S. A. Van Slyke and C. H. Chen, Appl. Phys. Lett. 48, 183
(1988).
5. W. E. Moerner and S. M. Silence, Chem. Rev. 94, 127 (1994).
6. H. Bässler, Phys. stat. sol. (b) 175, 15 (1993) and references therein.
7. U. Wolf, H. Bässler, P. M. Borsenberger and W. T. Gruenbaum, Chem.
Phys. 222, 259 (1997).
8. P. M. Borsenberger, W. T Gruenbaum, U. Wolf, and H Bässler, Chem.
Phys. 234, 277 (1998).
9. A. K. Kadashchuk, N. I. Ostapenko and N. V. Lukashenko, Adv. Mater.
Opt. Electr. 7, 99 (1997).
10. A. Kadashchuk, N. Ostapenko, V. Zaika, and S. Nespùrek, Chem.
Phys. 234, 285 (1998).
11. S. Nespùrek, A. Kadashchuk, N. Ostapenko, and V. Zaika, Mol. Cryst.
Liq. Cryst. 324, 95 (1998).
12. M. Pope and C. Swenberg, Electronic Processes in Organic Crystals, Clarendon Press, Oxford, New York, 1982, p. 821.
13. H. Bässler, Phys. stat. sol. (b) 107, 9 (1981).
14. H. Bässler, in Disordered Effect on Relaxational Processes, SpringerVerlag, R. Richter and A. Blumen, Eds., Berlin, Heidelberg 1994, p. 585.
15. R. Eiermann, W. Hofberger and H. Bässler, J. Non-Cryst. Solids 28,
415 (1978).
16. I. Glowacki and J. Ulanski, J. Appl. Phys. 78, 1019 (1995).
17. J. Ulanski, J. Sielski, I. Glowacki, and M. Kryszewski, IEEE Trans.
Electr. Insul. 27, 714 (1992).
18. I. Glowacki, J. Ulanski, J. K. Jeszka, and M. Kryszewski, Mater. Sci.
13, 83 (1987).
19. J. W. Stasiak and T. J. Storch, Proc. SPIE 3144, 72 (1997).
20. S. D. Baranovskii, M. Zhu, T. Faber, F. Hansel, and P. Thomas, Phys.
Rev. 55, 16226 (1997).
21. H. Gorbecht and D. Hofmann, J. Phys. Chem. Solids 27, 509 (1966).
22. I. A. Tale, Phys. stat. sol. (a) 66, 65 (1981).
23. H. L. Oczkowski, Acta Phys. Pol. A 76, 649 (1989).
24. A. K. Kadashchuk, N. I. Ostapenko, Yu. A. Skryshevskii, V. I. Sugakov,
and T. O. Susukolova, Mol. Cryst. Liq. Cryst. 201, 166 (1991).
25. P. I. Butlers, I. A. Tale, J. Pospísil, and S. Nespùrek, Prog. Coll. Polym.
Sci. 78, 93 (1988).
26. P. M. Borsenberger, W. T. Gruenbaum, E. H. Magin, and L. J. Sorriero,
Chem. Phys. 195, 435 (1995).
27. P. M. Borsenberger and M. B. O’Regan, Chem. Phys. 200, 257 (1995).
28. P. M. Borsenberger, E. H. Magin, M. B. O’Regan, and J. A. Sinicropi,
J. Polym. Sci. B: Polym. Phys. 34, 317 (1996).
29. R. H. Young, J. Chem. Phys. 103, 6749 (1995).
30. B. Hartenstein and H. Bässler, J. Non-Cryst. Solids 190, 112 (1995).
31. B. Movaghar, B. Ries and M. Grunewald, Phys. Rev. B 34, 5574
(1986).
Thermally Stimulated Luminescence in Molecularly Doped Polymers
Vol. 43, No. 3, May/June 1999
219
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Recent Advances in Charge Transport in Random Organic Solids:
The Case of Conjugated Polymers and Discotic Liquid Crystals
D. Hertel,▲ A. Ochse, V. I. Arkhipov, and H. Bässler
Institut für Physikalische Chemie und Zentrum für Materialwissenschaften, Philipps-Universität Marburg, D-35032 Marburg
The purpose of the present work is to delineate similarities as well as differences concerning the charge transporting propertie s
of π-conjugated polymers as well as of discotic liquid crystals. Materials investigated are a (i) ladder -type poly-paraphenylene
(LPPP), (ii) a member of the phenylenevinylene family, and (iii) several asymmetrically substituted triphenylenes. The disorder
formalism explains the field and temperature dependence of the mobility adequately provided that the disorder , which is controlled by the sample topology and random dipolar electric fields, is sufficiently large. Differences are noted in highly order ed
LPPP and in symmetric discotic crystals. It is conjectured that finite size effects in the case of LPPP and dynamic effects in the
liquid crystals overcompensate the effect of energetic disorder.
Journal of Imaging Science and Technology 43: 220–227 (1999)
Introduction
The disorder formalism has turned out to be a powerful tool to rationalize charge carrier motion in random
organic solids.1–3 The elementary event can be described
by either a redox process or a hopping process, depending upon the terminology of either chemists or physicists, among chemically identical but physically
different moieties, such as molecules or sub-units of a
polymer. Structural disorder in a glassy system implies
that their self-energies, for instance the van der W
aals
energies of molecular anions or cations in a polarizable medium, are distributed, as are the transfer integrals controlling charge or excitation exchange among
the moieties. Unambiguous signatures are the
inhomogenous line widths in optical absorption or fluorescence spectra,4 the broadening of diffraction features
of electron micrographs of a molecular glass5 or the distribution of intrinsic localized charge states in a polymer.6 It is obvious that the roughness of the energy
landscape in which charge carriers migrate has to depend on the shape of the molecules. If they carry polar
functionalities their interaction energy will vary more
strongly upon molecular displacement than apolar and
more spherical molecules because the variations of the
local electric field are larger.
A microscopic picture of charge carrier hopping in a
random organic system was developed both via Monte
Carlo (MC) simulations7 and analytic effective medium
theory.8 One of its essential features is relaxation.9 Upon
generating an ensemble of charge carriers by a light
flash the charge carrier tends to settle in the tail states
Original manuscript received November 23, 1998
▲ IS&T Member
© 1999, IS&T—The Society for Imaging Science and Technology
220
of the distribution of the hopping sites. Therefore transport must slow down in time and the equilibrium energy to which the ensemble relaxes must decrease as
the temperature decreases. For this reason activation
energy for hopping must increase at lower temperatures
and transport must become dispersive.
There is abundant evidence that the above-mentioned
concept is able to recover to basic experimental obser vations related to charge transport in organic glasses
or molecularly doped polymers such as
(i) the ln µ versus T –2 type temperature dependence
of the mobility,
(ii) the ln µ versus E1/2 type field dependence,
(iii) the temperature dependence of the slope of the ln
µ versus E1/2 dependence,
(iv) the effect of polar functionalities of either transport
or matrix molecules and
(v) the increasing dispersion of time of flight signals
at lower temperatures.1–3
The key parameters of the formalism are the standard
deviation σ of the (Gaussian) distribution of states
(DOS), which can be split into a van der W aals component and a polar contribution, and the positional disorder parameter.10–12 It has been recognized, however, that
in experiments the ln µ versus E1/2 type field dependence
is obeyed at much lower electric fields than those that
Monte Carlo simulations predicted. Recently , major
progress concerning this problem was made by introducing correlation among the hopping sites. 13,14 From
this, the length scale of the hopping process is extended.
It appears that the amended version of the previous
hopping formalism provides an adequate basis of experimental analysis.
The aim of this article is to delineate application as
well as the limitations for advanced organic transport
materials focussing on π-conjugated polymers and
discotic liquid crystals. Recent interest was caused by
the observation of comparatively large hole mobilities,
Me R R'
n
R'Me R
1.5
photocurrent [a. u.]
Hole Mobility in π-Conjugated Polymers. Two π-conjugated polymers have been selected for the present
study. Methyl-substituted ladder-type poly(para)phenylene (MeLPPP) synthesized by U. Scherf at the Max
Planck Institute for Polymer Research in Mainz employing a Suzuki reaction,15 is a material which exhibits an
extraordinary weak disorder, both in liquid and solid
solution and in film, manifested by the small
inhomogenously line broadening in both absorption and
fluorescence.16 The reason is the planarization of the
skeleton due to covalent bridging. Films, typically 1µm
thick, are prepared by spin-coating onto an indium tin
oxide glass slide. Before evaporating a typically 150 nm
thick aluminium top contact, the samples were stored
for 12 h under reduced pressure of 10 –6 mbar. Before
the experiment one sample—referred to as Sample A—
was kept at 150°C under vacuum for at least 4 hr. The
widths of the S1 → S0 0–0 fluorescence bands of Sample
B is 375 cm–1 (fwhm) and of Sample A 600 cm–1 (fwhm);
the corresponding standard deviations of Gaussian
bands are 160 cm–1 and 300 cm–1. For Sample B there is
no indication of defect emission while in Sample A part
of the fluorescence is emitted from dimers or aggregates. 16 For comparison, the standard deviation of the
S1 → S0 0–0 absorption band in PPV17 is 650 cm–1.
The second material is poly(N-phenylimino-1,4-phenylene-1,2-ethenylene-1,4-(2,5-dioctoxy)-phenylene-1,2ethenyl-ene-1,4-phenylene) (PAPPV) which was
synthesized, by means of a Horner reaction, in the group
of Prof. Hörhold at the University of Jena.18 The PAPPV
samples were prepared in the same mannner.
Transient photocurrent was generated by an optical
parametric oscillator driven at 450 nm (MeLPPP) and
470 nm (PAPPV) using the ITO/polymer interface as a
charge generation layer. Typical photocurrent transients
of MeLPPP at different applied fields are illustrated in
Fig. 1. The signals are normalized to the charge carrier
transit time ttr. After a fast initial spike, the current
exhibits a well-established plateau region and then
slowly tails off. The TOF signal at 3× 105 V/cm does not
show the inital spike due to the limited time resolution
of our apparatus. The TOF signals of MeLPPP are
nondispersive and feature characteristics of Gaussian
charge transport which is in marked contrast to conjugated polymers of the PPV family.19–21 If the spreading
of the tail in the TOF signals is due to thermal diffusion as expected for Gaussian transport, the dispersion
w should decrease by a factor of 2 if the applied voltage
is increased by a factor of 5 at a given temperature. It is
apparent from Fig. 1 that the transients in MeLPPP do
not meet this condition. The shape of the TOF signals
is approximately independent of the applied field, i.e.,
they bear out universality . Therefore, the dispersion
cannot be accounted for by thermal diffusion although
its presence indicates disorder.
Although the samples show the same characteristics
in the shape of the TOF signals, they differ markedly in
Me R R'
R'
R:
n-decyl
R': -n-hexyl
1.0
0.5
0.0
0
1
2
3
t/ttr
Figure 1. Photocurrent transients of MeLPPP normalised to
the transit time t tr at T = 243 K and applied fields of E = 6 ×
10 4 V/cm (solid line) and E = 3 × 10 5 V/cm (line with solid
circles). The inset shows the chemical structure of the MeLPPP
.
T [K]
423
393
363
333
303
273
243
213
183
153
-3
10
2
Experiment
2.0
µ [cm /Vs]
which are important for the design of organic light emitting diodes. The main questions, which we address, are
the role of charge dislocation in a π-conjugated main
chain polymer and the effect of static and dynamic disorder in the discotic liquid crystals. W e will describe
the key observations, delineate both the success as well
as the required extensions of the disorder formalism and
give an outline of new conceptional approaches which
one could envisage.
-4
sample A
annealed at 150 °C
10
200
300
400
1/2
500
600
1/2
E [(V/cm) ]
Figure 2. Electric field dependence of the mobility µ of the
MeLPPP Sample A at various temperatures. Data are plotted on
a log µ versus E1/2 scale.
their temperature and field dependent behaviour. The
field dependence of the hole mobility in SampleA ranges
from 7 × 10-5 cm2/Vs at 153K to 1.6 × 10–3 cm2/Vs at 423K
as shown in Fig. 2. For fields above 4 × 104 V/cm, the
mobility follows the square root of the electric field
which is an ubiquitous feature of molecularly doped
polymers or organic glasses. The observed room temperature mobilities of approximately 1 × 10–3 cm2/Vs are
orders of magnitude higher than in PPV and its derivatives 19,20 and even higher than in the recently investigated polyfluorene. 22 In Fig. 3, the temperature
Recent Advances in Charge Transport in Random Organic Solids...
Vol. 43, No. 3, May/June 1999
221
sample B
sample A
10
-3
-3
2
2
µ [cm /Vs]
µ [cm /Vs]
10
-4
10
-4
3x10
5
10
15
20
25
2
30
35
2
3
4
5
-2
(1000/T) [K ]
6
7
-1
1000/T [K ]
Figure 3. The logarithm of the mobility versus 1/T 2 for MeLPPP Sample A (left) solid circles: E = 1.4 × 10 5 V/cm, open squares: E
= 8 × 104 V/cm and the Arrhenius plot of the hole mobility of MeLPPP Sample B (right). The symbols refer to E = 8 × 10 4 V/cm
(solid triangles), E = 6 × 104 V/cm (open circles) and E = 4 × 10 4 V/cm (solid squares).
  2σ  2 
 σ  2

2
µ = µ 0 exp −
  exp C   − ∑  E




 3kT 
 kT

222
Journal of Imaging Science and Technology
(1)
393
363
303
273
213
153
123
sample B
2
µ [cm /Vs]
dependence of the logarithm of the mobility for MeLPPP
is plotted versus 1/T 2 (left) for Sample A and versus 1/T
(right) for Sample B. The weak temperature dependence
as well as the negligible electric field dependence of µ
in Sample B (Fig. 4)23 are unique and unexpected as far
as charge transport of disordered organic solids is concerned. The transport behaviour is similar to that of
molecular crystals rather than that of molecularly doped
polymers (MDPs). On the other hand the mobility in the
MeLPPP film is 2 to 3 orders of magnitude less than in
molecular crystals. It cannot be due to trapping, since
the low temperature asymptote in the Arrhenius plot
translates into an activation energy of only 22 meV ( E
= 4 × 10 4 V/cm), which is less than the variance of the
distribution of singlet states24 and at least one order of
magnitude less than the activation energies that determine the trap controlled transport of holes in PPV .19,20
Because arguments that invoke traps are unable to explain the low mobilities, we will discuss these low mobilities as inherent features of MeLPPP.
The temperature dependent mobility (left in Fig. 3)
as well as the field dependent mobility of SampleA (Fig.
2) on the other hand, fulfill the relations predicted by
the disorder formalism. One can therefore quantify the
degree of disorder on the charge transport by calculating the width of the DOS σ according to Eq. 1, µ is the
mobility, µ0 is the mobility at zero field, σ is the ener getic disorder parameter (the width of the DOS), C an
empirical constant and Σ is the quantity that describes
the positional disorder.7
K
K
K
K
K
K
K
-3
10
4x10
-4
200
300
E
400
1/2
500
600
1/2
[(V/cm) ]
Figure 4. The electric field dependence of the hole mobility of
MeLPPP Sample B at different temperatures. Data are plotted on a log µ versus E1/2 scale.
The zero field mobilities at various temperatures can
be extracted from Fig. 2 and plotted versus 1/ T 2. The
slope of the resulting straight line yields σ ≈ 50 meV.
Hertel, et al.
can be compared with the transients in Ref. 26. The resulting width of the DOS σ is 58 meV. From the field
dependence of the mobility in PAPPV, which is similar
to that of Sample A of MeLPPP (Fig. 2),24 we are able to
obtain a σ of 52 meV. This value is in reasonable good
agreement with the value of 58 meV extracted from the
dispersive transient.
MC simulations also provide an estimate of the temperature at which the transition from nondispersive
transport to dispersive transport should occur, if the energetic disorder is known:26
-2
10
OC8H17
N
H17C8O
n
2
µ [cm /Vs]
-3
10
2
 σ 
σˆ 2 = 
 = 44.8 + 6.7 log d
 kTc 
Tc~153K
-4
10
10
20
30
40
50
2
60
70
80
-2
(1000/T) [K ]
Figure 5. The temperature dependence of the mobility µ of
PAPPV for applied fields of E =7 × 104 V/cm (solid circles) and
E = 6.4 × 10 4 V/cm (open squares). The log µ is plotted versus
1/T2. The arrow marks the transition from the nondispersive
to the dispersive transport regime. The inset shows the chemical structure of the PAPPV.
Typical values of σ for MDPs 25 are 80 to 100 meV . Although the width of the DOS in MeLPPP is reduced compared with MDPs, the absolute value of the mobility is
at least one order of magnitude lower than in MDPs with
the highest mobilities, e.g., for the triphenylamine TAPC
in polystyrene.1
In MeLPPP, the rigid backbone gives rise to improved
structural order as compared with PPV-type conjugated
polymers. In order to further elucidate the principle of
charge transport of the PPV-type polymers we have investigated a amino derivative of PPV (PAPPV). In Fig.
5 the temperature dependence of the mobility is shown.
It can be considered as a conventional conjugated polymer with respect to its electronic properties. The broad,
structurless absorption and photoluminescence spectra
of the PAPPV support this notion. As we have shown
recently,24 the increased disorder in P APPV leads to a
broadening in the tails of the TOF signals as compared
with MeLPPP. At temperatures below 153 K the broadening of the tails of the photocurrent transients leads
to compeletely dispersive TOF signals because the
charge carriers do not attain their dynamic equlibrium
before they reach the electrode. An analysis of the dispersive TOF signals in terms of the Scher -Montroll
theory failed as the slopes of the tangents on the cur rent decay in double logarithmic representation do not
add up to two.26 This result is not unexpected since the
dispersion in the Scher-Montroll theory originates from
the positional disorder in the sample, whereas in conjugated polymers the energetic disorder should be much
more important for the charge transport.
An alternative explanation for the disappearance of a
clearly indicated transit time is given by MC simulation work where the dispersion was correlated with enAPPV
ergetic disorder.26 At 153 K the initial slope of the P
transient in a double logarithmic scale is –0.18, which
(2)
where σ is the width of the DOS, Tc the transition temperature and d the sample thickness in cm. The disor der parameter σ of 52 meV corresponds to a dispersive
transport regime below T c = 142 K. This is only 1 1 K
below the measured transition temperature of 153 K
(Fig. 5).
Discotic Liquid Crystals
Discotic liquid crystals, which form hexagonal columnar mesophases, are of interest for their high hole mobilities. Typical representatives of this class of materials
consist of triphenylene cores symmetrically substituted
by six aliphatic ethers. Mobilities parallel to the stack
axis of 10 –4 cm 2 /Vs are reported for hexahexyloxytriphenylene (H6T). 27,28 For hexapentyloxytri-phenylene (H5T) 29,30 and for hexabutyloxytriphenylene
(H4T) 31 values of 10–3 cm 2/Vs and 10–2 cm 2/Vs have been
reported, respectively. Perpendicular to the columns, the
mobility is two (H5T) to three (H6T) orders of magnitude lower. The relatively high mobilities in the direction of the columns in the liquid crystalline phase have
been explained by liquid-like self-healing of structural
defects on a time scale faster than hopping of charge
carriers. H6T and H5T form the conventional discotic
hexagonal ordered phases (D ho) featuring a two-dimensional hexagonal intercolumnar order and uncorrelated
one-dimensional intracolumnar order. The higher hole
mobility in H4T has been attributed to improved order
in the discotic hexagonal plastic phase ( Dhp) in which
the centers of every disc form a three-dimensional ar ray and the only degree of freedom of the discs is their
rotation about the plane axis, their positions in adjacent columns being correlated (see Fig. 6). The existence of the Dhp-phase was first recognized by Glüsen and
co-workers.32,33 in H4T and subsequently also found in a
few related materials. In all mentioned hexaalkyloxytriphenylenes the hole mobility turned out to be almost field and temperature independent.
In the present work we studied hole transport in
discotic liquid crystals based upon the hexaalkyloxytriphenylenes, in which one of the six ether groups is
replaced by an ester substituent. 34 This leads to an extended range of the mesophase and suppresses crystallization. This extends the temperature range of
transient photoconduction measurements. Four derivatives of H5T , i.e., the ester of pivalic acid (termed
pivaloate), the ester of cyclohexane carbonic acid
(cyclohexanoate), the ester of 4-cyanobenzoic acid
(cyanobenzoate) and the ester of 4-nitrobenzoic acid
(nitrobenzoate) and a derivative of H4T, the ester of 4pentenoic acid (pentenoate), were studied (Fig. 6).
Pivaloate and pentenoate show the D hp-phase within
Recent Advances in Charge Transport in Random Organic Solids...
Vol. 43, No. 3, May/June 1999
223
Dho-phase:
O
Dhp-phase:
CN
(1)
10
O
(2)
-4
NO2
160°C
120°C
60°C
20°C
-20°C
-60°C
10
-5
10
-6
2
(3)
µ [cm /Vs]
O
O
(4)
O
(5)
Figure 6. At the left side the type of molecular displacements
and rotations in the hexagonal ordered phase D ho and the hexagonal plastic phase Dhp is illustrated. In the bottom the chemical structure of the hexaalkyloxytriphenylenes is shown. R´ =
R = C 5H 11 for H5T, R´ = R = C4H9 for H4T. In the ester substituted compounds R´ is one of the ester substitutes whose chemical structure is shown at the right for (1): cyanobenzoate, (2):
nitrobenzoate, (3): cyclohexanoate, (4): pivaloate and (5)
pentenoate.
50
100
150
E
224
Journal of Imaging Science and Technology
250
300
350
1/2
[(V/cm) ]
Figure 7. Field dependence of the hole mobility in pentenoate
at several temperatures.
T [K]
2
µ [cm /Vs]
500 400
certain temperature ranges. The other materials reveal
only the D ho-phase between the clearing temperature
and glass transition.
The starting material for the synthesis is
pentaalkyloxytriphenylene alcohol synthesized via the
biphenyl route as described in Refs. 35 and 36. Subsequently, the esters were prepared by esterifying the
triphenylene alcohols with acid chloride as described in
Ref. 33. The substances were filled in 30 µm thick
sandwhich cells at a temperature above the clearing
temperature. To orient the discotic substances
homeotropically the samples were cooled down from the
isotropic phase at a very slow rate, implying an orientation of the columns perpendicular to the substrate.
A standard time-of-flight (TOF) setup was used for
measuring the photocurrent transients. The sample was
excited with the frequency tripled output of a Nd-YAG
laser at 355 nm where the penetration of the light is 1.5
µm. The TOF-signals were non-dispersive. The litera31
ture data for hole mobility in H4T have been confirmed.
Unlike H4T and H5T, the mobility of the ester derivatives is field dependent, following a ln µ versus E1/2 law,
with increasing slope at decreasing temperature. Figure 7 shows the mobility of pentenoate plotted against
√E for several temperatures as an example. The other
esters behave similarly.
In contrast with the hexaalkyloxytriphenylenes the
mobility is temperature dependent. The temperature dependence deviates from the Arrhenius law and rather
follows a ln µ versus T–2 relation (Fig. 8) maintaining a
constant slope at the glass transition temperature. At
the transition between the Dhp-phase and the Dho-phase
a slight decrease of the slope, (|ln dµ/dT –2 |), for
pivaloate and pentenoate is noted. It is more pronounced
in pivaloate (Fig. 9).
There are some unambiguous signatures of disorder
dominated charge carrier hopping as the prevailing
transport mechanism7,9:
200
1/2
10
-3
10
-4
300
225
250
cyclohexanoate
nitrobenzoate
cyanobenzoate
H5T
Ti
10
-5
10
-6
Tg
Tg
10
-7
4
6
8
10
12
2
2
14
16
2
18
20
22
2
1000 /T [1000 /K ]
Figure 8. Temperature dependence of the hole mobility in
cyclohexanoate (E = 2.9 × 104 V/cm), cyanobenzoate (E = 105 V/cm)
and nitrobenzoate (E = 6 × 104 V/cm) in log µ versus T–2 representation. Clearing temperatures Ti and glass temperatures Tg are
marked. The mobility of H5T is also plotted for comparison.
(i) The TOF signals of the compounds carrying an estersubstituent broaden upon decreasing temperature.
(ii) The temperature dependence of their hole mobilities
obeys a ln µ versus T–2 law and
(iii) the field dependence of their mobilities follows a ln
µ a E 1/2 law, with decreasing slope at increasing
temperature.
Hertel, et al.
T [K]
300
250
500 400
200
TABLE I. List of the Energetic Order Parameters σ and Positional Order Parameters Σ of the Hopping Sites Derived from the Analysis of
µ(E,T ) Data.
Compound
10
-2
10
-3
10
-4
10
-5
10
-6
pentenoate
pivaloate
H5T
H4T
Pivalote
Pentenoate
Cyclohexanoate
Cyanobenzoate
Nitrobenzoate
σ (meV)
Σ
84
104
108
124
127
1.8
1.8
2.5
2.6
3.0
Tt
2
µ [cm /Vs]
Ti
Tg
10
-7
4
6
8
10 12 14 16 18 20 22 24 26
2
2
2
2
1000 /T [1000 /K ]
Figure 9. Temperature dependence of the hole mobility in
pivaloate (E = 4.5 × 104 V/cm) and pentenoate (E = 3.2 × 10 4 V/
cm) in log µ versus T –2 representation. Clearing and glass temperatures T i and Tg and also Tt, the phase transition temperature between D ho- and D hp-phase, are marked by straight lines
(pentenoate) or dashed lines (pivaloate). The mobilities of H4T
and of H5T are also plotted for comparison.
The µ(E,T) dependence of all examined esters is the
same as has been observed in molecularly doped polymers, it is in accordance with Eq. 1.
The value of the energetic gaussian widthσ is around
100 meV, which is also a typical value for molecularly
doped polymers. The lowest values are found in pivaloate
and pentenoate, that accordingly also feature the highest mobilities. The highest values of σ are obtained for
cyanobenzoate and nitrobenzoate whose group dipole
moments are higher. Recall that the group dipole moment of the ester group is 1.95 D and that the moments
of the cyano group and of the nitro group attached to a
phenyl ring are 4.18 D and 4.22 D, respectively. The effect is well documented in the literature on molecularly
doped polymers and molecular glasses. 10–12,37,38 It has
been ascribed to the increase of the fluctuation of the
random electric fields in the vicinity of positionally disordered molecules carrying dipole moments.A summary
of the energetic as well as the positional disorder parameters is given in Table I.
Discussion
It is apparent that certain aspects of charge carrier
motion in π-conjugated polymers and discotic liquid crystals are in accordance with the predictions of the disorder formalism for Gaussian shaped DOS while others
require conceptional modifications. The successful interpretation of the hole mobilities in discotic liquid crystals carrying ester substituents in terms of the disorder
formalism implies the absence of coherence effects. It is
obvious that it is sufficient to consider nearest neighbour
hopping only. Apparently, the variance of the intermo-
lecular spacing and, in the hexagonal plastic phase, even
the rotation of molecules carrying polar substituents is
sufficient to destroy any possible coherence effects and
to localize a charge carrier . Extrapolating the linear
portion of ln µ versus T–2 plots towards very high temperature yields values of µ0 ranging from 0.03 cm 2/Vs
(pivaloate, pentenoate and cyclohexanoate) to around 5
× 10–3 cm2/Vs which is less than the value for molecular
crystals. At higher temperatures the slope of lnµ versus
T –2 plots tends to decrease. This effect is most pronounced with pivaloate at the transition from the hexagonal plastic Dhp-phase to the Dho-phase. A similar effect
has been found with molecularly doped polymers and
molecular glasses above the glass transition temperature. It has been ascribed to the onset of dynamic disor
der39 yielding smaller values of µ as expected on the
premise of a constant static disorder potential.
It is remarkable that the hole mobilities in symmetric H4T and H5T are temperature independent while
those of the asymmetric esters follow a ln µ versus T–2
dependence. Further, the mobility of H4T more or less
agrees with the 1/T → 0 intercept of the ln µ versus 1/T–2
plot of pentenoate, i.e., of derivatized H4T . This suggests that the main reason of the disorder is the presence of random potential fluctuations caused by polar
functionality. This is confirmed by the variation of the
energetic order parameter σ with the group dipole moments and the reduced disorder evident from the temperature and field independence of the hole mobility of
unsubstituted H4T. On the other hand the rather low and
temperature independent value of µ in the discotic phase
of unsubstituted H4T as compared with molecular crystals, indicates that the limiting process for non-activated
hopping is likely to be associated with the rotation of
the disc-like molecules that prevent optimum electronic
overlap rather than due to energetic disorder.
As far as conjugated polymers are concerned the disorder formalism provides an adequate description of the
hole mobility in P APPV and the MeLPPP after heat
treatment which may introduce aggregates that act as
physical traps. However, trap-free and only weakly disordered MeLPPP behave differently as far as the temperature and the field dependence is concerned. In order
to explain the low mobility of MeLPPP, arguments have
to be invoked which are yet not included in the disorder
model. Charge transport in MDPs occurs by transfer of
charges from molecule to molecule, each being different
in energy due to disorder . This is reflected in the dependence of charge transfer on temperature and electric field. In these cases the mean free path of the carrier
is equal to the intermolecular distance. For conjugated
polymers, that may not necessarily be true. A polymer
like MeLPPP consists of arrays of subunits which are
disordered and separated by topological defects, these
segments being more or less electronically decoupled.
The length of the segments is subjected to a statistical
Recent Advances in Charge Transport in Random Organic Solids...
Vol. 43, No. 3, May/June 1999
225
distribution resulting in inhomogeneous broadening of
the absorption and fluorescence spectra. The effective
conjugation length for MeLPPP is 6.5 ± 0.5 nm, equivalent to 14.5 ± 1.5 phenylene units. 40 Charge transport
in conjugated polymers occurs both by migration between the segments of the same chain (on-chain transport) and by hopping between adjacent chains
(inter-chain transport).
A weak temperature dependence of the mobility requires a low activation energy of rate-limiting carrier
jumps. For a system of point-like hopping sites this implies weak energy and/or strong positional disorder. Under the condition σ < kT the mobility must be practically
independent of the temperature and reveal a weak field
dependence within the whole range of variation of these
two parameters. However, the above condition is not fulfilled even for weakly disordered conjugated polymers
such as MeLPPP especially at lower temperatures where
anomalously weak T- and F-dependencies of the mobility are still observed. A strong positional disorder implies a broad distribution of inter-site distances that is
possible only in a diluted hopping system. Making a
jump over the distance ∆x along the field F a carrier
gains the electrostatic energy ∆E = eF∆x . If this energy
is higher than both kT and σ, carriers will jump mostly
along the field direction and such jumps will require no
activation. Under these conditions both the field and
temperature dependences of the transit time must saturate and longest non-activated jumps along the field
direction will play a role of rate-limiting steps. In a hopping system with both positional and energy disorder ,
the regime of T-independent mobility sets on at a sufficiently strong field: F > σ/e∆x . For conjugated polymers
with typical inter-site distance of 0.6 nm and σ ranging
from 50 to 100 meV this estimate yieldsF ranging from
8 × 10 5 to 1.5 × 10 6 V/cm. This is at least one order of
magnitude higher than the experimentally observed
onset of the regime of weak field and temperature dependence of the mobility in MeLPPP. Therefore the traditional version of the disorder model cannot account
for the lack of F- and T-dependence of the mobility in
weakly disordered conjugated polymers.
These materials consist of arrays of coupled subunits
(conjugated segments) which are positionally and orientationally disordered. Conjugated segments, that belong to the same polymer chain, are separated by
topological defects. Charge carriers occupy extended
states and are therefore mobile within segments while
charge transfer between different segments occurs via
tunneling jumps. The length of segments ranges from 5
to 10 nm in different materials that is much longer than
the inter-segmental distance of typically 0.6 nm. Under these conditions, carriers may cross the longest part
of the total distance moving in the conduction band of
conjugated segments and make much fewer inter -segmental tunneling jumps compared to what is predicted
by the model of a random network of point-like hopping
sites.
The variation of carrier electrostatic energy in the
field of 105 V/cm on the length of 6 nm is 0.06 eV, which
is larger than the thermal energy of 0.025 eV at the
room temperature. Under these circumstances carriers
are localized within a field-induced potential well close
to the low energy ends of segments independent of the
point at which the carrier has entered the segments and
carriers gain much more electrostatic energy by travelling along a segment than that which they can gain by
making a jump between sites. This is essentially equiva-
226
Journal of Imaging Science and Technology
lent to the field-induced positional disorder. On the one
hand, field-assisted localization makes inter-segmental
jumps of carriers against the field direction difficult even
at moderate fields. Under such conditions carrier jumps
from dead ends of segments along the field direction will
be the rate-limiting steps. On the other hand, one may
expect a much larger localization radius for tunneling
jumps within a segment compared to that for inter-segmental jumps. This can strongly enhance long carrier
jumps in the forward direction and suppress the effect
of energic disorder. The intra-segmental contribution
to the gain of electrostatic energy makes rate-limiting
carrier tunneling jumps easier onto shallower segments
along the field direction and effectively reduces the activation energy of the hopping drift mobility. Moreover,
if the intra-segmental gain of energy at a given external field exceeds the width of the DOS, carrier hopping may need no further thermal activation and both
the field and the temperature dependencies of the mobility practically vanish. Quantitative description of
charge transport in disordered hopping systems with
finite size of hopping sites will be given in a future work
of the authors including consideration of the orientational disorder caused by random orientations of conjugated segments.
Acknowledgement. It is a pleasure to acknowledge the
contribution of H. H. Hörhold, A. Kettner, J. Kopitzke and
J. H. Wendorff. This work was supported by the Deutsche
Forschungsgemeinschaft (Sonderforschungsbereich 383
and 436 Rus 113/9314).
References
1. P. M. Borsenberger, L. Pautmeier and H. Bässler, J. Chem. Phys. 94,
5447 (1991).
2. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for Imaging Systems, Marcel Dekker, New York, 1992.
3. M. A. Abkowitz, H. Bässler and M. Stolka, Phil. Mag. B 63, 201 (1991).
4. R. I. Personov, in Spectroscopy and excitation dynamics of condensed
molecular systems, V. M. Argranovitch and R. M. Hochstrasser, Eds.,
North Holland, Amsterdam, 1983, p. 555.
5. R. Eiermann, G. M. Parkinson, H. Bässler, and J. M. Thomas, J. Phys.
Chem. 86, 313 (1982).
6. A. Kadashchuk, N. Ostapenko, V. Zaiko, and S. Nêspùrek, Chem.
Phys. 234, 285 (1998).
7. H. Bässler, Phys. stat. sol. (b) 15, 175 (1993).
8. B. Movaghar, M. Grünwald, B. Ries, H. Bässler, and D. Würtz, Phys.
Rev. B 33, 5545 (1986).
9. H. Bässler, in Disorder Effects on Relaxational Processes, R. Richert
and A. Blumen, Eds., Springer-Verlag, Berlin, Heidelberg 1994, p.
485.
10. P. M. Borsenberger and J. J. Fitzgerald, J. Phys. Chem. 97, 4815
(1993).
11. P. M. Borsenberger and H. Bässler, J. Chem. Phys . 95, 5327 (1991).
12. P. M. Borsenberger, W. T. Gruenbaum, E. H. Magin, and L. J. Sorriero,
Chem. Phys. 195, 435 (1995).
13. Yu. N. Gartstein and E. M. Conwell, Chem. Phys. 245, 351 (1995).
14. D. H. Dunlap, P. E. Parris and V. M. Kenkre, Phys. Rev. Lett . 77, 542
(1996).
15. U. Scherf, A. Bohnen and K. Müllen, Makromol. Chem . 193, 1127
(1992).
16. R. F. Mahrt, T. Pauck, U. Lemmer, U. Siegner, M. Hopmeier, R.
Henning, H. Bässler, E. O. Göbel, P. Haring Bolivar, G. Wegmann, H.
Kurz, U. Scherf, and K. Müllen, Phys. Rev. B 54, 1759 (1996).
17. S. Heun, R. F. Mahrt, A. Greiner, U. Lemmer, H. Bässler, D. A. Halliday,
D. D. C. Bradley, P. L. Burns, and A. B. Holmes, J. Phys. Condens.
Matter 5, 247 (1993).
18. H. Rost, A. Teuschel, S. Pfeiffer, and H. H. Hörhold, Synth. Met. 84,
269 (1997).
19. H. Meyer, D. Haarer, H. Naarmann, and H. H. Hörhold, Phys. Rev. B
52, 2587 (1995).
20. E. Lebedev, Th. Dittrich, V. Petrova-Koch, S. Karg, and W. Brütting,
Appl. Phys. Lett. 71, 2686 (1997).
21. P. W. M. Blom and M. C. J. M. Vissenberg, Phys. Rev. Lett. 80, 3819
(1998).
Hertel, et al.
22. M. Redecker, D. D. C. Bradley, M. Inbasekaran, and E. P. Woo, Appl.
Phys. Lett. 73, 1565 (1998).
23. D. Hertel, U. Scherf and H. Bässler, Adv. Mat. 10, 1119 (1998).
24. D. Hertel, U. Scherf, H. H. Hörhold, and H. Bässler, J. Chem. Phys.
110 (1999).
25. P. M. Borsenberger, E. H. Magin, M. Van der Auweraer, and F. C. De
Schryver, Phys. stat. sol. (a) 140, 9 (1993).
26. P. M. Borsenberger, L. Pautmeier and H. Bässler, Phys. Rev. B 46,
12145 (1992).
27. N. Boden, R. J. Bushby, J. Clements, B. Movaghar, K. J. Donovan,
and T. Kreouzis, Phys. Rev. B 52, 13274 (1995).
28. N. Boden, R. J. Bushby, J. Clements, K. Donovan, B. Movaghar, and
T. Kreouzis, Phys. Rev. B 58, 3063 (1998).
29. D. Adam, F. Closs, T. Frey, D. Funhoff, D. Haarer, H. Ringsdorf, P.
Schumacher, and H. Siemensmeyer, Phys. Rev. Lett . 70, 457 (1993).
30. D. Adam, D. Haarer, F. Closs, T. Frey, D.Funhoff, K. Siemensmeyer,
P. Schumacher, and H. Ringsdorf, Ber. Bunsenges. Phys. Chem. 97,
1366 (1993).
31. J. Simmerer, Adv. Mat. 8, 815 (1996).
32. B. Glüsen, W. Heitz, A. Kettner, and J. H. Wendorff, Liq. Cryst. 20,
627 (1996).
33. B. Glüsen, A. Kettner and J. H. Wendorff, Mol. Cryst. Liq. Cryst. 303,
115 (1997).
34. A. Ochse, A. Kettner, J. Kopitzke, J. H. Wendorff, and H. Bässler,
Phys. Chem. Chem. Phys. 1, 1757 (1999).
35. N. Boden, R. C. Borner, R. J. Bushby, and A. N. Cammidge, J. Chem.
Soc., Chem. Commun. 1994, 465 (1994).
36. P. Henderson, H. Ringsdorf and P. Schuhmacher, Liq. Cryst. 18, 191
(1995).
37. P. M. Borsenberger and M. B. O´Regan, Chem. Phys. 200, 257 (1995).
38. P. M. Borsenberger, E. H. Magin and J. Shi, Physica B 217, 212 (1996).
39. P. M. Borsenberger, L. Pautmeier and H. Bässler, J. Chem. Phys . 95,
1258 (1991).
40. J. Grimme, M. Kreyenschmidt, F. Uckert, K. Müllen, and U. Scherf,
Adv. Mat. 7, 292 (1995).
Recent Advances in Charge Transport in Random Organic Solids...
Vol. 43, No. 3, May/June 1999
227
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Suitable Definition of Drift Mobility
A. Hirao, T. Tsukamoto and H. Nishizawa
Materials and Devices Research Laboratories, Research and Development Center, Toshiba Corporation, Kawasaki, Japan
Measurement of the drift mobility defined as the proportionality constant of the mean velocity (v) to the electric field strength (E) is
necessary for understanding of carrier transport. However , it is difficult to obtain v from the time-of-flight transients. Thus
the
velocity obtained from the transit time has been analyzed instead of the mean velocity
. The most common measurement of the mobility
(µk_ex) is obtained from the time derived from the intersection of the asymptotes to the plateau and tail of the transients. Because the
long tail of photocurrent transients for molecularly doped polymers indicates anomalous dispersion of carrier transit times, th
e difference between v/E and µk_ex is not negligible. Recently , a theoretical photocurrent transients equation (PTE) has been introduced.
Fitting of the PTE to nondispersive transients gives the values of v and the diffusion coefficient ( D) simultaneously. In this article,
using the PTE, both the mobility (µk_cal), obtained from a kink in the photocurrent transient, and the tail-broadening parameter (Wcal)
were derived as functions of v, D and sample thickness. We have tried to explain the anomalous behavior of µk_ex and the tail-broadening parameter (Wex) in order to verify the PTE. The dependences ofµk_ex and Wcal on the electric field and the sample thickness satisfactory agreed with those of µk_ex and Wex. These verify the PTE and suggest that fitting of the PTE to photocurrent transients is suitable
way to obtain the drift mobility.
Journal of Imaging Science and Technology 43: 228–232 (1999)
Introduction
Molecularly doped polymers (MDPs), in which guest
charge transport molecules are dispersed in the host
polymer matrix, are widely used for a photoconductor,1–
3
photorefractive devices, 4,5 organic electroluminescent
devices,6 and the synapse bond devices.7 The fundamental processes of these devices involve carrier transport
phenomena. Carrier transport also plays an important
role in carrier injection. 8 Hence, an understanding of
the carrier transport mechanism is a matter of great
significance for science and technology.
Carrier transport has been evaluated by several methods. 9 The most common method is measurement of the
photocurrent transients by time-of-flight (TOF). 1–3 In
this measurement, the displacement current when a
carrier packet moves in a sample is measured. Therefore the number of arrived carriers cannot be simply
detected, and a detailed analysis of the photocurrent
transients is necessary to obtain the drift mobility .1,3
However, in the conventional method, the drift mobility µk_ex is derived from the time t 0 defined by intersection of the asymptotes to the plateau and tail of the
transient. The tail of the current does not decrease linearly. Hence, t 0 shows variation that originates in the
choice of the tangent line. Therefore, µk_ex contains ambiguity. In addition to this, the velocity obtained fromt0
is not the mean velocity. This situation is an obstacle to
understanding the carrier transport mechanism.
Original manuscript received November 19, 1998
© 1999, IS&T—The Society for Imaging Science and Technology
228
Recently, a method of obtaining both the velocity ( v)
and the diffusion coefficient (D) from the photocurrent
transient has been developed.10,11 In Refs. 10 and 11, the
theoretical photocurrent transients equation (PTE) was
proposed. It was based on the fact that a carrier packet
drifts at a constant velocity and spreads by diffusion.
The top electrode is assumed to act as a reflecting and
partially absorbing wall and the counter electrode is
assumed to act as an absorbing wall. By fitting the PTE
to a photocurrent transient, v and D can be obtained
simultaneously. The obtained velocity is independent of
thickness and shows no negative field dependence in a
low electric field. Thus the fitting method is a suitable
way to obtain the mobility . 12 However, the method of
obtaining the mobility by the fitting of the PTE to photocurrent transients has not been widely used yet. The
reasons for this are (1) the traditional method is fairly
straightforward and (2) this fitting method has not been
generally verified.
In this article, we have tried to explain the anomalous behavior of µk_ex and the tail-broadening parameter13
(W ex) using the PTE. In the following section, the different definitions of the drift mobilities are reviewed. Then,
µ k_cal and W cal as functions of v, D and L that derived
from the PTE are reviewed, where µk_cal is the mobility
obtained from a kink in the photocurrent transients and
W cal is the tail-broadening parameter . Finally, the dependences of µk_cal and Wcal on the electric field, the temperature, and the thickness are analyzed and discussed.
Definitions of Drift Mobilities1–3
As shown in Fig. 1, typical non-dispersive transients
for MDPs have an initial spike, a plateau of variable
duration, and a long tail. The initial spike has been explained as the results of trapping at sites with waiting
J (t ) = −
+
Figure 1. Time-of-flight photocurrent transient signal and some
definitions of the transit time. The broken line is the experimentally measured photocurrent, and the solid line is the current
obtained by calculating Eq. 1 using the parameters obtained by
fitting. The inset shows a structural molecular formula for charge
transporting molecule, DDB.
times that are comparable to the transit time,3 the thermalization of the carrier packet within the density of
states (DOS),3 or forward diffusion of carriers at the illuminated surface.10,11 The plateau region indicates the
displacement of a carrier packet at constant velocity and
diffusion. 3 The long tail shows the breakdown of
Einstein’s relationship relating the mobility and the
diffusion coefficient.3
By analyzing the shape of the photocurrent transients,
we can obtain a few kinds of mobilities in accordance
with different definitions of the transit times. Pautmeier
and co-workers have shown the Monte Carlo simulation
results14 for mobilities derived from transit times:
(a) the time t 0 defined by the intersection of the
asymptotes to the plateau and tail of the transient,
(b) the time t 1/2 for the photocurrent to decay to one-half
its value at t0, and
(c) the ensemble average arrival time.
The logarithm of the mobility from method (a) shows
proportionality with the square root of the electric field.
On the other hand, the logarithms of the mobilities from
methods (b) and (c) are proportional to the square root
of the field in only high electric field regions. However,
the experimental results that differ from the above-mentioned simulation results have been reported.1,3
The time t 0 corresponds to the time when the cur rent begins to decrease rapidly . Because the carrier
packet is generated as a sheet, the time t 0 is approximately the arrival time of the earliest carriers at the
counter electrode. The earliest carrier is thought to be
transported by a combination of drift and forward diffusion. 10,11 At the time t 0, the sum of the drift length
and the diffusion length representing the earliest carriers is equal to the sample thickness. Because the diffusion length is proportional to t 0.5, the contribution of
the diffusion length increases with decreasing the
sample thickness. Actually, the µ k_ex has sometimes
shown thickness dependence.1,3
The experimental procedure for measuring of the average velocity using the PTE is expressed as follows.10,11
Suitable Definition of Drift Mobility
en0
2L
 ( L − vt) 2 
 v 2 t 2 
D
exp−
 − exp−

πt 
4 Dt 

 4 Dt 
 L − vt  
en0 v   vt 
erf 
 + erf 
 ,
2 L   2 Dt 
 2 Dt  
(1)
where n0 is the number of holes, L is the thickness of
the MDP, and erf(x) is the error function. The fitting of
Eq. 1 to the transients gives the values of D, v, and n0.
The mobility µa and the transit time ta obtained by fitting are written as µ a ≡ v/E and t a ≡ L/v.
As illustrated in Fig. 1 with a solid line, Eq. 1 was
successfully fitted to experimental photocurrent transients whose shapes were non-dispersive over a temperature range from 260 K to 330 K and over electric
field range from 1 to 3.6 MV/cm. The sample used in
this measurement was DDB doped bisphenol-A-polycarbonate with a ratio of 0.26 in molar units, where DDB
is a charge transporting molecules shown in Fig. 1 as
an inset.
The mobility of an MDP depends on electric field and
temperature, as well as on the structure of donor and
acceptor functionality. These dependences have been
described by the disorder formalism.15 The disorder formalism was developed by Bässler,15 and was verified by
Borsenberger and co-workers. 1,3 According to Refs. 10
through 12, the dependence of µa on electric field and
temperature is also described in Eq. 2:
µ (T , E) =
  σ  2

  2σ  2 

2
µ 0 exp −
  exp C 
 − ∑  E ,
  kT 


  3kT  
(2)
where σ is the width of the DOS, ∑ is a parameter that
describes the degree of positional disorder , µ 0 is the
prefactor mobility, and C is an empirical constant, typically 2.9 × 10–4 (cm/V)1/2. The values of D also show a similar dependence on temperature and electric field as10,11
D(T , E) =
  T  2

  T 2

D0 exp − 1   exp CD  1  − ∆  E ,
  T 


  T  
(3)
where D 0, T1, ∆, and CD are constants. The parameters
of a DDB doped bisphenol-A-polycarbonate with a ratio
of 0.26 in molar units are µ0 = 1.10 × 10–2cm2/Vs, σ = 0.115
eV, ∑ = 2.89, C = 2.90 × 10–4 (cm/V)1/2, D0 = 6.30 × 10–2 cm/s,
T1 = 896 K, ∆ = 6.26, and CD = 1.35 × 10–3 (cm/V)1/2. To
deduce the values of µk_cal and Wcal, the parameters in
Eqs. 2 and 3 will be used in a later section.
Expressions for µk_cal and Wcal as Functions of v, D
and L
In this section, both µ k_cal and W cal as functions of v, D
and L derived by using the PTE are reviewed. 16 First,
the definition of the time t0 must be derived. In graphical way, the time t 0 is derived from the intersection of
the asymptotes to the plateau and tail of the transients.
However, the tail of the TOF current does not decrease
linearly. Hence the choice of the tangent line of the tail
has variation and t 0 contains ambiguity. The definition
of t 0_cal as the intersection between the tangent line at
Vol. 43, No. 3, May/June 1999
229
Figure. 2. Dependence of µ of the DDB-doped polymer on the
square root of the applied electric field. The solid circles show
µk_ex at T = 285 K. The µa is shown by the open circles and the
solid line. The values of µk_cal shown as the dashed line are calculated from Eq. 4.
plateau and the tangent line at ta is reasonable. In this
case, t 0_cal is expressed as a function of v, D, and L:
t0 _ cal

3D 
πD
 1 − Lv − 2 Lv 

L 
= ⋅
.
D 
v

1 −


2 Lv 
L
D
1 −
.
v
Lv 
(4)
(5)
The tail-broadening parameter W is defined13 as
W=
t1 / 2 − t0
.
t1 / 2
(6)
The parameter W as the function ofD, v, and L is obtained
by inserting Eqs. 4 and 5 into Eq. 6 as
Wcal
1 D 


2  Lv 
=
D

1 −


Lv 
2
πD
Lv
.
D 

1 −


2 Lv 
+
(7)
In the case of Lv >> D, we obtain from Eq. 7:
WE =
πD
.
vL
(8)
If Einstein’s relationship relating mobility and diffusion
holds, Eq. 8 reduces to
230
Journal of Imaging Science and Technology
WE =
πkT
.
eLE
(9)
Previous reports17,18 have shown anomalous behaviors of
Wex that were not similar to those of WE. The dependence
of Wcal on E, T, and L is discussed in a later section.
The expression of t1/2_cal for the photocurrent to decay to
one-half its value at plateau16 was found to be
t1 / 2 _ cal =
Figure 3. Dependence of µk_cal on sample thickness at T = 303 K
and E = 2.5 × 105 V/cm. The µa is shown by the arrow.
Comparison of Mobility Behaviors
Let us compare behaviors of mobilities:µa, µk_ex, and µk_cal.
The graphical analysis of TOF transients givesµk_ex and
the PTE analysis give v (= µ a E) and D. The µk_cal can be
expressed as a function of E, L, and T by substituting
Eqs. 2 and 3 with the above-mentioned parameters of
DDB-doped polymer.
The dependences of µk_ex, µ k_cal and µ a on the square
root of the electric field is shown in Fig. 2, where L =
7.6 × 10 –4 cm and T = 285 K. The logarithm of µk_ex is
shown as solid circles which is larger than µa, shown as
open circles. The difference betweenµk_ex and µa increases
as the electric field decreases. The field dependence of
µk_cal shown as the dashed line is similar to that of µk_ex,
thus t0 can be approximated by Eq. 4. The negative field
dependence of µk at low electric field originates in a large
diffusion coefficient of MDPs. 10,11 The time t0 that gives
µk_ex is close to the arrival time of the earliest carriers at
the counter electrode. The carrier packet spreads with
time by diffusion, thus the arrival time of the earliest
carriers is shorter than the transit time ta. If the applied electric field is small, the carriers will spend a long
time to transit the layer. Under this condition, the difference between t 0 and ta is large. Therefore the ratio of
µk to µ a is large.
The slope of µk_ex approaches that of µa in the high electric field region. This suggests that the disorder formalism parameters σ and ∑ obtained from µ k_ex in the high
electric field region show small deviation from those
obtained from µ a.
Figure 3 shows the thickness dependence of the mobility where T = 300 K and E = 2.5 × 104 V/cm. The logarithm of µk_cal shows thickness dependence and is different
Hirao, et al.
Figure 5. W versus temperature. The solid circles show Wex obtained at E = 3.0 × 105 V/cm and L = 7.6 × 10–4 cm. The solid line
is Wcal. The dashed line shows WE.
Figure 4. W versus E. The solid circles show Wex of DDB-doped
polymer at T = 303 K and L = 7.6 × 10–4 cm. The field dependence
of Wcal of DDB-doped polymer shown by the solid line is similar
to that of Wex. The dashed line shows WE.
from log µa shown by the arrow. The difference between
µa and µk_cal is particularly marked when the sample is
thin. This dependence agreed adequately with those of
µk_ex in many studies. 1,3 These unexpected behaviors are
not due to the experimental error but due to the fact that
t0_cal is a function of E and L as shown in Eqs. 2, 3, and 4.
To discuss the dependence of µk_cal as an actual characteristic value of a substance, the thickness of the sample
should be sufficiently large in this system.
Behavior of the Tail-broadening Parameter W
Borsenberger and co-workers have experimentally and
numerically described the dependence of W ex on various
parameters in detail.17,18 The anomalous dependence of
W ex on the electric field, the temperature, and the thickness has been reproduced by their detailed Monte Carlo
simulations. In this section, the values of Wcal are calculated by substituting experimentally obtained values
of v and D into Eq. 7 in the same way with µ k_cal. As a
result, the reproduction of Wex by Wcal has been achieved
as follows.
Figure 4 shows the electric field dependence ofWcal as
the solid line. It agrees adequately with the Wex, which
is indicated by solid circles. In this case, theW decrease
monotonously with the electric field. The W sometimes
increases with the electric field. This dependence is reproduced with different values of v and D.
The Wex also shows the temperature dependence as
solid circles in Fig. 5 as well as Wcal as solid line. The
temperature dependence of Wcal is similar to that of Wex.
We also analyze Borsenberger and B ässler’s data by
the PTE. Figure 5 in Ref. 18 indicates that W depends
on the temperature and is independent of the concentration of the charge transporting molecule when the
electric field and the sample thickness are constant. The
substitution of the data of this figure into Eq. 7 gives D/v
at E = 2.0 × 105 V/cm and L = 10.0 µm. The obtained values of D/v are used for the calculation of W cal as functions of L, shown in Fig. 6 in Ref. 18. Figure 6 shows
the thickness dependence of Wcal as a solid line with that
of W ex re-plotted from Ref. 18 as symbols. The depen-
Suitable Definition of Drift Mobility
Figure 6. The thickness dependence of Wex of TAPC-doped polystyrene, parameteric in T where TAPC is 1,1-bis(di-4-tolylaminophenyl)cyclohexane. The data of Wex are re-plotted of Fig. 6 in
Ref. 18. The field was 2.0 × 105 V/cm. The temperature dependence of Wex at E = 2.0 × 105 V/cm and L = 10 µm was also shown in
Ref. 18. The substitution of the data into Eq. 7 gives D/v at E =
2.0 × 105 V/cm and L = 10 µm. The obtained values are used for
the calculation of Wcal dependence on L shown by the solid line.
The dependence of Wcal is similar to that of Wex.
dence of W cal is similar to that of W ex except for the thin
thickness region. When the sample is thin, the plateau
of TOF transients disappear because the earliest carriers
arrive at the counter electrode immediately. In this case,
the comparison between Wex and Wcal is meaningless.
In Figs. 4 and 5, WE of Eq. 9 is shown as the dashed
line. Figure 4 shows monotonous decrease in WE over
all the electric field region; however , such behavior of
W E has not yet been reported. The W E is proportional to
T0.5. Figure 5 shows W ex is not proportional to T0.5. These
results are a reminder that Einstein’s relationship does
not hold in the MDPs.
Because the values of µ k_cal agreed adequately with
those of µ k_ex, µa is a suitable definition of the drift mobility. The values of W cal are similar to those of W ex in
Figs. 4, 5 and 6. Hence the expressions of t0 (Eq. 4) and
t1/2 (Eq. 5) as functions of v, D and L are suitable, and
Vol. 43, No. 3, May/June 1999
231
the values of v and D that substituted into Eqs. 4 and 5
describes well the carrier transport of the MDPs. These
results verify the description of charge transport in
MDPs by the PTE.
Summary
The mobility (µk_cal) obtained from a kink in the photocurrent transient and the tail-broadening parameter
(W cal) as functions of v, D and L have been derived from
the PTE based on the fact that a carrier packet drifts at
a constant velocity and is spread by diffusion. The dependence of µk_cal and Wcal on electric field, temperature,
and sample thickness have been investigated by substituting the experimentally obtained v, and D. The dependences of µk_cal and Wcal agreed adequately with those
of µk_ex and W ex. Our analysis also shows if the sample is
sufficiently thick and the electric field is sufficiently
high, the deviation of µk from µa calculated from v is
small for purposes of interpreting the dependences of
the mobility in terms of the disorder formalism. These
results suggest that the PTE describe the photocurrent
transients in MDPs adequately, and the analysis by the
PTE is suitable for measuring the carrier transport in
MDPs.
232
Journal of Imaging Science and Technology
References
1. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptor for Imaging Systems, Marcel Dekker, New York, 1993.
2. L. B. Schein, Electrophotography and Development Physics, 2nd ed.,
Springer, New York, 1992,
3. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for Xerography, Marcel Dekker, New York, 1998.
4. K. Sutter and P. Günter, J. Opt. Soc. Am. B 7, 2274 (1990).
5. S. Ducharme, J. C. Scott, R. J. Twieg, and W. E. Moerner, Phys. Rev.
Lett. 66, 1846 (1991).
6. J. Kido, K. Hongawa, K. Okuyama, and K. Nagai, Appl. Phys. Lett.
64, 815 (1994).
7. H. Körner and G. Mahler, Phys. Rev. B 48, 2335 (1993).
8. E. M. Conwell and M. W. Wu, Appl. Phys. Lett. 70, 1867 (1997).
9. E. A. Silinsh, Organic Molecular Crystals, Springer-Verlag, Berlin,
1980, p. 36.
10. A. Hirao, H. Nishizawa and M. Sugiuchi, Phys. Rev. Lett. 75, 1787
(1995).
11. A. Hirao and H. Nishizawa, Phys. Rev. B 54, 4755 (1996).
12. A. Hirao and H. Nishizawa, Phys. Rev. B 56, R2904 (1997).
13. L. B. Schein, Philos. Mag. B 65, 795 (1992).
14. L. T. Pautmeier, R. Richert and H. Bässler, Philos. Mag. B 63, 587
(1991).
15. H. Bässler, Phys. stat. sol. (b) 175, 15 (1993).
16. A. Hirao, T. Tsukamoto and H. Nishizawa, Phys. Rev. B 59, in press,
No. 19.
17. P. M. Borsenberger, L. T. Pautmeier and H. Bässler, Phys. Rev. B 48,
3066 (1992).
18. P. M. Borsenberger and H. Bässler, J. Appl. Phys. 75, 967 (1994).
Hirao, et al.
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Transient Space-Charge-Limited Current Measurements of Mobility in a
Luminescent Polymer
J. C. Scott,* S. Ramos and G. G. Malliaras
IBM Research Division, Almaden Research Center, San Jose California, and Center on Polymer Interfaces and Macromolecular Assemblies (CPIMA)
Transient time-of-flight methods with voltage pulse injection have been adapted to determine the field dependent mobility of
holes in the electroluminescent polymer , MEH-PPV. The time-resolved current response confirms that gold forms an Ohmic
contact to MEH-PPV and that there are very few traps in the polymer
. These conclusions are in agreement with earlier interpretation of steady-state current-voltage measurements.
Journal of Imaging Science and Technology 43: 233–236 (1999)
Introduction
Charge carrier mobility plays a pivotal role in the operation of organic light emitting diodes (OLEDs).1 When
charge carrier injection at the electrodes is Ohmic, the
mobilities of electrons and holes dictate the operating
voltage of the device, and hence its power efficiency
. The
dynamic response of an OLED is controlled by the rate
at which electron and hole densities accumulate to the
levels which ensure efficient recombination. 2 The relative mobilities of electrons and holes also play a role,
albeit less important than their relative injection rates,
in determining the quantum efficiency.3 Because of various experimental difficulties, there are relatively few
direct measurements of electron and hole mobilities in
materials of interest for OLEDs.4,5 It is the purpose of this
article to present results on one such material, the luminescent polymer poly(2-methoxy-5-(2’-ethyl)hexoxy-phenylene vinylene) (MEH-PPV). The data were obtained from
time-of-flight transients in the s pace-charge-limited regime using a novel extension of a standard experimental
technique that we hope may be useful for other materials.
We6,7,8 and others9,10 have inferred charge carrier mobilities from steady state current measurements in
single carrier devices. The assumptions behind this
method of determination are that
(i) the current is space-charge-limited, i.e., that the
injecting contact is Ohmic,
(ii) the transport is trap-free, or at least that the
trapped charge density makes a negligible
contribution to the space-charge field,
(iii) the mobility has a particular field dependence,
usually taken to be the “Poole-Frenkel like” form
µ ~ exp E , and
(iv) the influence of the minority carrier may be safely
neglected.
Original manuscript received October 27, 1998
∗ Corresponding author, e-mail address: jcscott@almaden.ibm.com
© 1999, IS&T—The Society for Imaging Science and Technology
In view of the central role of mobility in OLEDs it is
important to provide independent measurement as well
as to test the validity of these assumptions. Several
other methods have been suggested for the determination of mobility, including transient electroluminescence.11,12 This approach has the difficulty ofdistinguishing
the relative roles of electrons and holes, because by definition both signs of carrier must be injected into the device. Moreover, if the contacts limit the current it has been
shown2 that the delay in emission is not related to the
transit time(s) of the carrier(s) but rather to the time
required to accumulate sufficient charge in the recombination zone. The most direct method to obtain mobilities in organic semiconductors is well recognized to be
transient time-of-flight (TOF) measurement. 13 This is
typically performed in the “space-charge-free” limit using photogeneration of carriers. However, the photoinduced TOF technique has been successfully applied to
conjugated polymers5,14,15,16 and other luminescent OLED
materials4,17 in only a few cases. The difficulty here appears to be the necessity of preparing relatively thick (several microns), trap-free samples in order to observe
non-dispersive transients with a well defined transit time.
In this article, we report TOF data in which the
charges are injected by applying a voltage step to an
Ohmic contact, the so-called transient space-charge limited current (SCLC) technique.18 This method has been
applied to organic semiconductors in the past, 19,20 typically using thick samples and therefore relatively high
voltages. Here, we describe an experimental technique
which permits the use of samples of order 100 nm in thickness, and voltages to below 1 V . Thus the same sample
preparation techniques can be used as for OLEDs, and
one does not need to worry that different materials processing may introduce, for example, different trapping
behavior. Our new method is used to measure the mobility of holes in MEH-PPV. At the same time the data
confirm that gold forms an Ohmic contact, and may per
mit a quantitative evaluation of charge trapping.
This article is arranged as follows. In the next section, we give the details of the experimental method.
233
Then we present the results and discuss their implications for the behavior of OLEDs. Finally we summarize
our conclusions and present some ideas for the further
extension of the transient SCLC approach.
Experimental
The transient SCLC technique is well established and
is conceptually extremely simple. The sample, thickness
L, is prepared in a sandwich geometry between two electrodes, at least one of which is Ohmic for the carrier of
interest, i.e., it is capable of injecting and maintaining
a space-charge-limited current. One applies a step
change in voltage (V) and measures the time dependence
of the resulting current response. For sufficiently low
trap density and long trapping times, the current initially rises from a non-zero initial (post step) level to a
value above its steady state value.18 In the case of a field
independent mobility, the maximum in the current occurs at a time t0 = 0.79tT, where t T = L2/µV is the transit
time for carriers in a fieldE = V/L. The analysis is more
difficult when the mobility is field dependent, but a similar maximum is observed at the time when the first car
riers reach the far electrode.
To date, the technique has been used mostly for relatively thick samples, such that the capacitance is low and
the transit time, even at relatively high voltages, is long
compared to the RC time constant of the circuit. (The relevant resistance is that in series with the sample, usually
the input impedance of the detector electronics.) Thus the
capacitive charging current has decayed long before the
current maximum associated with the transit of the first
carriers. The circuit that is used in this case is extremely
simple: merely a voltage source, the sample and the cur rent detector in series. 19 As the transit time becomes
shorter at higher voltages and in thinner samples, it is
necessary to “cancel out” the capacitive part of the response. Helfrich and Mark 21 described a bridge circuit
which accomplishes this. It employed a floating voltage
source and a single-ended amplifier as detector . In this
work we used a different configuration of bridge that takes
advantage of modern electronic equipment and offers several advantages over the earlier designs.
The circuit, still very simple, is shown in the inset to
Fig. 1. Here the voltage source is referenced to ground,
and can therefore be any commercial pulse generator
with an appropriate rise time, pulse length, repetition
rate, and voltage and current capability. The sample is
placed in one arm of the bridge and the variable capacitor (C) adjusted to equal that of the sample. W e find
that a separate auxiliary low-level sinusoidal source is
very helpful for balancing the bridge. A small variable
resistor (R) is included in the tuning arm in order to
compensate for any resistance in the leads to the sample.
The matched resistors (R1) in the other two arms of the
bridge are selected so that the high frequency impedance of the entire bridge circuit is 50 Ohms and therefore matched to the cable from the pulse generator
. Care
is taken to equalize the lengths of all cables in the arms
of the bridge. The short-time detection limit of this circuit is due to residual imbalances of the bridge which
we believe are caused by the frequency dependent dielectric constant of the sample. W e have been able to
reduce this instrumental “dead-time” to <2 µs. The transient charge carrier current is detected using a commercial differential amplifier. Instruments with bandwidths
3
are readily availup to 1 MHz and with gains of over 10
able. Signal averaging techniques may also be used by
repetitively pulsing the sample with a well-defined duty
cycle. In this way, one may first investigate the behav-
234
Journal of Imaging Science and Technology
Figure 1. Typical transient current response to a step change
in applied voltage. Some experimental details are given in the
legend. The inset shows the bridge circuit used to minimize
the influence of capacitive charging current, and to match the
circuit to the transmission cable from the pulse generator.
ior of a “well-rested” sample, separating the pulses by
many minutes, and then explore the effects due to accumulation of trapped charge as the repetition period
becomes short compared to the detrapping time.
In this article we present only data obtained with long
times between pulses such that we are observing the
transient behavior in a sample initially free of trapped
charge. A typical response is shown in Fig. 1. The sample
used here was 384 nm thick, with a gold anode and aluminum cathode. The sample was prepared in the same
manner that we use for MEH-PPV light emitting diodes,
as previously described.22 For the data of Fig. 1, a voltage step from 0 to 8.5 V was applied, with a rise time of
10 ns. The current has the form expected for trap-free
space-charge limited behavior . It starts (after the
deadtime) at a non-zero value and rises to a clear maximum at 54 µs. It then settles to a virtually time independent steady-state value. The presence of the current
maximum immediately establishes 18 that the trapping
time is long compared to the transit time. The constancy
of the current for times later than about 2tT reveals that
the trap density is low.
In Fig. 2, we plot the time of the current maximum as
a function of the initial applied voltage, corrected for
the built-in voltage which arises because of the differ ence in work-functions of the electrodes. 6 As expected,
the transit time decreases as the voltage increases, until finally we can no longer detect it due to the instrumental dead-time.
Discussion
Extraction of the mobility from the time of the current
maximum is complicated by the fact that the mobility
depends on electric field. As discussed by Many and
Rakavy18 for the case of field independent mobility, the
current maximum occurs when the first holes reach the
cathode. As the carriers cross the sample, the field at
the leading edge increases until at t0 it is 2 V/L at the
cathode. It finally settles to EC = 1.5V/L in the steady
state, at which time the position dependence of the field
has the familiar E ~ x 1/2. When the mobility is an increasing function of electric field (as for the PooleFrenkel form) the charge density in the low field region
Scott, et al.
Figure 2. Transit time as a function of applied voltage for two
sample of different thickness, as indicated.
near the anode is higher than the field independent case.
Thus near the anode the field varies more steeply than
square-root of distance, and less steeply near the cathode. We may therefore state the limits of the steady state
cathode field, and correspondingly of the maximum field
at the transient leading edge, as V/L < EC < 1.5 V/L.
Thus when we derive a field dependent mobility from
the fastest transit time, the reader should remember
that we are giving a mobility averaged over a range of
fields somewhat larger than the average field in the
sample. (A more accurate treatment requires numerical analysis 23,24 and is the subject of ongoing study. Details will be published at a later date.)
With these caveats, we now present the averaged mobility values obtained from the current maxima according to the expression
µ (V / L) = 0.79 L2 / t0 V .
(1)
The results are plotted in Fig. 3, as ln( µ) versus E1/2,
where E = ( V – Vbi)/L is the average field across the
sample. The values of the hole mobility obtained from
the time-of-flight compare well with those obtained previously from analysis of the steady state SCLC by ourselves3,8 in MEH-PPV and others 9 in similar polymers.
Writing the mobility in the form
µ = µ 0 exp E / E0 ,
(2)
we can extract the zero-field mobility µ0 = 2 × 10–7 cm2/
Vs, and characteristic field E0 = 3.1 MV/m.
In Fig. 4, we compare explicitly the experimental
steady-state SCLC, determined as the limiting behavior of the current transient, with that expected from the
measured mobility, namely23
J = (9 / 8)εε 0 µ 0 exp(0.89 V / E0 L )V 2 / L3 .
(3)
The values differ by about a factor of two, which may
be partly accounted for by the effects of field dependent
mobility but which may also reflect a small degree of trapping. The overall field dependence obtained from transient and from steady-state measurements is similar. The
general agreement once again confirms that gold forms
an Ohmic contact to MEH-PPV and that hole transport
is close to the trap-free space-charge-limited regime.
Figure 3. Mean mobility (see text for explanation of “mean”)
as a function of electric field, plotted according to the form of
Eq. 2. The samples are the same as in Fig. 3.
In this first report describing the experimental
method, we have presented detailed data on only a few
samples, each with Au/Al electrodes. We have additional
data on samples of different thickness and with other
electrode materials. The results, which are in agreement
with those presented here, will be given in a future publication.
The transient SCLC technique permits, in principle,
the determination of the trapping time for the charge
carriers crossing the sample. Indeed, by careful signal averaging techniques we are able to detect a drop in the hole
current of typically <20 % which occurs with a characteristic time on the order of milliseconds. However, before
we interpret this behavior as unequivocally due to trapping, it is necessary to evaluate another potential mechanism, namely ionic motion. If there is a (small)
concentration of mobile anions or cations in the sample25,26
they will drift under the influence of the applied elec tric
field towards the electrode, setting up a dipole layer(s)
and screening the field in the bulk of the material. Thus
although the ion current itself may not be detectable,
its effect, through changes in the electric field profile,
may be observed in the electronic (here hole) current.
Conclusions
We have described an experimental technique which extends the well-established methodology of transient spacecharge limited currents into a regime of sample thickness
and voltage that is particularly appropriate for the study
of materials used in organic light-emitting diodes. Experimental data obtained on MEH-PPV confirm clearly that
gold forms an Ohmic anode for this luminescent polymer
and that the transport of holes is in the trap-free spacecharge-limited regime. The mobility , obtained from the
time-of-flight, is found to depend on electric field, with a
behavior that is well approximated by the Poole-Frenkel
form. The steady-state space-charge-limited current predicted from the mobility agrees well with that measured
directly.
It will be interesting to extend this method to electron transport in MEH-PPV, to bipolar devices and to
other materials of interest. Such work is under way. It
is worth commenting that we do not yet have useful electron data because of the difficulty of preparing reliable
Transient Space-Charge-Limited Current Measurements of Mobility....
Vol. 43, No. 3, May/June 1999 235
is supported by the NSF-MRSEC program under grant
number DMR-9400354.
References
Figure 4. Comparison of the measured steady state spacecharge-limited current (J DC) with that predicted from the timeof-flight mobility ( J MOB), plotted according to Eq. 3. Sample,
189 nm in thickness, is one of those from Figs. 2 and 3.
and reproducible electrodes for the injection of electrons
and the blockage of holes.
Transient SCLC measurement, using bridge circuits
similar to the one discussed above, have the potential for
determining additional important parameters of the
materials used in OLEDs and other organic electronic
devices. For example, trapping has a clear signature in
the current decay following the voltage step; detrapping
can be explored through variations of the pulse length
and duty cycle; the field dependence of detrapping can
be probed by applying a reverse bias during the“resting”
time; and the effect of non-Ohmic electrode injection will
be seen in early time behavior of the current and the suppression of the space-charge induced maximum.
Acknowledgments. We wish to acknowledge several
discussions with Dr. M. A. Abkowitz who contributed to
our understanding of the results of this paper. CPIMA
236
Journal of Imaging Science and Technology
1. For a review of OLEDs, see S. Miyata and H. S. Nalwa, Eds. Organic
Electroluminescent Materials and Devices , Gordon and Breach,
Amsterdam, 1997.
2. V. R. Nikitenko, Y.-K. Tak and H. Bässler, J. Appl. Phys. 84, 2334–
2340 (1998).
3. G. G. Malliaras and J. C. Scott, J. Appl. Phys. 83, 5399–5403 (1998).
4. R. G. Kepler, P. M. Beeson, S. J. Jacobs, R. A. Anderson, M. B.
Sinclair, V. S. Valencia and P. A Cahill, Appl. Phys. Lett. 66, 3618
(1995).
5. E. Lebedev, Th. Dittrich, V. Petrova-Koch, S. Karg, and W. Brütting,
Appl. Phys. Lett. 71, 2686–2688 (1997).
6. G. G. Malliaras, J. R. Salem, P. J. Brock, and J. C. Scott, Phys. Rev.
B 58, 1311–1314 (1998).
7. J. C. Scott, G. G. Malliaras, J. R. Salem, P. J. Brock, L. Bozano, and
S. A. Carter, Proc. SPIE 3476, 111–122 (1998).
8. L. Bozano, S. A. Carter, J. C. Scott, G. G. Malliaras, and P. J. Brock,
Appl. Phys. Lett. 74, 1132–1134 (1999).
9. P. W. M. Blom, M. J. M. de Jong and M. G. van Munster, Phys. Rev. B
55, R656–R659 (1997).
10. J. Pommerehne, A. Selz, K. Book, F. Koch, U. Zimmermann, Chr.
Unterlechner, J. H. Wendorff, W. Heitz, and H. Bässler, Macromol.
30, 8270–8277 (1997).
11. C. Hosokawa, H. Tokailin, H. Higashi, and T. Kusumoto, Appl. Phys.
Lett. 63, 1322–1324 (1993).
12. P. Ranke, I. Bleyl, J. Simmerer, D. Haarer, A. Bacher, and H. W.
Schmidt, Appl. Phys. Lett. 71, 1332–1334 (1997).
13. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for Electrophotography, Marcel Dekker, New York, 1993.
14. M. Gailberger and H. Bässler, Phys. Rev. B 44, 8643 (1991).
15. H. Meyer, D. Haarer, H. Naarman, and H.H. Hörhold, Phys. Rev. B
52, 2587 (1995).
16. M. Redecker, D. D. C. Bradley, M. Inbasekaran, and E. P. Woo, Appl.
Phys. Lett. 73, 1565–1567 (1998).
17. L.-B. Lin, S. A. Jenekhe, R. H. Young, and P. M. Borsenberger, Appl.
Phys. Lett. 70, 2052–2054 (1997).
18. A. Many and G. Rakavy, Phys. Rev. 126, 1980–1988 (1962).
19. M. A. Abkowitz and D. M. Pai, Philos. Mag. B 53, 193–216 (1986); M.
A. Abkowitz, J. S. Facci and M. Stolka, Chem. Phys. 177, 783–792
(1993).
20. C. Giebeler, H. Antoniadis, D. D. C. Bradley, and Y. Shirota, Appl.
Phys. Lett. 72, 2448–2451 (1998).
21. W. Helfrich and P. Mark, Z. Phys. 166, 370 (1964).
22. J. C. Scott, S. A. Carter, S. Karg, and M. Angelopoulos, Synth. Metals
85, 1197 (1997).
23. P. N. Murgatroyd, J. Phys. D: Appl. Phys. 3, 151–156 (1970).
24. R. H. Young, Phil. Mag. Lett. 70, 331–333 (1994).
25. F. Huang, A. G. MacDiarmid and B. R. Hsieh, Appl. Phys. Lett. 71,
2415–2417 (1997).
26. D. Zou, M. Yahiro and T. Tsutsui, Appl. Phys. Lett. 72, 2484–2486
(1998).
Scott, et al.
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Carrier Transport in Molecularly Diluted Liquid Crystalline Photoconductor
K. Kurotaki and J.-I. Hanna
Imaging Science and Engineering Laboratory, Tokyo Institute of Technology, Nagatsuta Midori-ku, Yokohama 226-8503, Japan
The carrier transport properties of a molecularly diluted smectic liquid crystalline photoconductor , 2-(4'-octylphenyl)-6dodecyloxynaphthalene (8-PNP-O12) and 2-(4'-hexyoxy)-6-octypbiphenyl (6O-BP-8) system, were investigated by time-of-flight tech nique, in order to clarify the nature of electronic conduction in the liquid crystalline mesophases. The mobility in the dilute d liquid
crystals was ambipolar, independent of both electric field and temperature in SmA and SmB phase as in the pure 8-PNP-O12, and
continuously reduced with an increase in the diluent concentration. The reduction, however , remained within a small range of on e
third of that of pure material even in 60 mol%. The carrier transport in the diluted liquid crystals was described by the relat
ion of a µ/
ρ2 ∝ exp(-2 ρ/ρ0), where µ is the mobility, ρ the average hopping distance, and ρ0 a wavefunction decay constant of molecular orbital,
indicating the 2-dimentional random hopping mechanism. The fairly largeρ0 of 2.3 ~2.4 A characterizes a fast mobility gently decreasing with an increase in the diluent concentration. The molecular ordering within a smectic layer did not affect the carrier tra nsport
properties at all except the initial difference of the mobility, as far as comparison of those in SmAand SmB phases were concerned. In
addition, the effect of self-organization of hopping site is discussed in terms of carrier transport in disorded materials system.
Journal of Imaging Science and Technology 43: 237–241 (1999)
Introduction
In these past two decades, organic photoconductors including molecularly doped polymers were well established for xerographic applications and have been
increasing their importance as industrial materials. 1
This is due to the increasing demand in non-impact
printing technologies for computer outputs, i.e., the
emergence of a new business of laser printers and to
the recent increasing activity in organic light emitting
diodes.2 All these are backed up by the unique nature of
organic materials, i.e., variety of materials and their
feasibility of designing and manufacturing new materials for increasing requirements in photoelectrical properties, in addition to feasibility of preparing large-area
and uniform thin films in low cost.
In contrast to outstanding advances in their practical
application, the nature of carrier transport properties
in organic disordered systems have remained to be fully
understood for a long time. In this decade, however ,
there was significant advance in its theoretical under standing, which owes to establishment of analytical
method for abstracting the essence of carrier transport
properties in various disordered systems, i.e., the disorder formalism proposed by Bässler3 and to the recognition of the importance of carrier-dipole interaction in
the carrier transport. 4,5 This must have never been
brought without earnest and steady efforts to under -
Original manuscript received December 11, 1998
© 1999, IS&T—The Society for Imaging Science and Technology
stand the nature made by the late Dr. Borsenberger and
his co-workers. 1 Now, our understanding is coming up
to the origin of the specific nature in the disordered car
rier transport on the theoretical basis.6–9
At the same time, this understanding provides us with
a guiding principal and a new idea for upgrading the
present properties of organic photoconductors. Indeed,
Bosenberger and co-workers demonstrated that the hole
mobility is improved up to 10 –3 cm2/Vs even in conventional molecularly doped polymers by appropriate choice
of a polymer matrix and a carrier transport material.10
On the other hand, there was a coincidence in the different direction with his demonstration. That is, it is
the discovery of a fast electronic conduction in discotic
and smectic liquid crystalline mesophases, which is
characterized by a high mobility over 10 -3 cm2/Vs independent of electric field and temperature. 11–15 This can
be an alternative way to take for upgrading photoconductive properties outside of the conventional organic
photoconductors.
The liquid crystalline photoconductors exhibit some
kind of crystal-like self-organizing molecular alignments
and liquid-like fluidity.16 This unique feature provides
us with a good basis of their practical application to
large-area electronic devices. Thus, liquid crystalline
photoconductors is a promising material for making a
break-through in device applications of organic
photoconductors currently limited by electrical properties of the materials.17 Indeed, the conventional organic
photoconductors have been practically used in xerographic drums and more recently in organic light emitting diodes as mentioned above, but their electrical
properties characterized by a small mobility of 10 –5 ~
10–6 cm 2/Vs that depends on electric field and temperature are rather poor from electronic materials point of
237
n – C8H17
On – C12H25
2-(4′-Octylphenyl)-6-dodecyloxynaphthalene (80-PNP-O12)
n – C6H13O
n – C8H17
4-Hexyoxyl-4′-octylbiphenyl (6O-PB-8)
Figure 1. Chemical structures of 8-PNP-O12 and 6O-BP-8.
Figure 2. Phase diagram of 6O-BP-8 and 8-PNP-O12 system.
view. This inferior property, however, is covered well in
the present application by specific device structures and
performance: in xerographic applications, the drums are
illuminated at a high electric field of 105 ~ 106 V/cm and
accessed at a very slow frequency of 0.1~1 Hz (6 ~ 60
ppm); on the other hand, the cell thickness is thinned
down to less than 0.1 µm in order to establish a high
electric field of 105 ~ 106 V/cm and a fast response in the
light emitting diodes.2
There is another reason why the liquid crystalline
photoconductor is an important material deserving our
attention. That is, it is likely that their carrier transport characteristics provide us with a new insight into
the nature of carrier transport in disordered systems
ever established. This is because liquid crystalline materials are a unique materials system just between or dered and disordered materials, with which we can test
and confirm our established understanding: the liquid
crystals exhibit a variety of molecular alignment, where
each molecule is thermally fluctuated; for example, in
smectic liquid crystals where all the molecules sit in
layers and are oriented in a direction with thermal fluctuation, the molecular alignment in the layer are or dered from “liquid-like” in SmA phase to “almost crystal”
in SmE phases, and so on. 16
In this work, we have investigated the carrier transport in molecularly diluted liquid crystalline
photoconductors in order to characterize their carrier
transport properties by transient photocurrent measurements. Here, we describe the effect of molecular dilution on the carrier transport in liquid crystalline
photoconductors elucidated and discuss self-organization in the molecular system in comparison with the conventional diluted carrier transport system, i.e.,
molecularly doped polymers.
Experimental
A photoconductive smectic liquid crystal, 2-(4'octylpheny)-6-dodecyloxynaphthalene (8-PNP-O12) was
prepared by a cross coupling reaction of corresponding
benzene and naphthalene derivatives with a palladium
catalyst described elsewhere 14 and purified by recrystallization from n-hexane. 8-PNP-O12 exhibits phase
transitions from Crystal to SmB phase at 79 °C, from
SmB to SmA phase at 100°C, and from SmA to isotropic
phase at 121°C. A diluent liquid crystal, 2-(4'-hexyloxy)6-octypbiphenyl (6O-BP-8) was prepared similarly from
corresponding benzene derivatives, which exhibit SmE
phase between 30°C and 49°C, and SmB phase between
49°C and 83°C. The chemical structures of these mate-
238
Journal of Imaging Science and Technology
rials are shown in Fig. 1. These two smectic liquid crystals were miscible without phase separation in the
present experimental conditions up to ~60 mol%.
The liquid crystal cells were prepared by capillaryfilling the mixed liquid crystals in isotropic phase into
the glass cells made of indium-tin oxide coated glass
plates (electrode area: 16 mm2) and a silica spacer. The
resulting molecular orientation was parallel to the electrode surface in terms of longitudinal axis of the liquid
crystal molecule, i.e., homogeneous alignment. This was
never affected by the electric field applied.
The conventional time-of-flight setup, equipped with
a N 2-laser and digital oscilloscope, was used in order to
measure transient photocurrents. The mixed liquid crystals exhibited optical absorption at 337 nm high enough
to ensure the one carrier condition for carrier transit.
The transit time was determined with an inflection point
in double logarithmic plots of the transient photocur rents as a function of time.
Result and Discussion
6O-BP-8 used as a diluent is a calamitic (rod-like) liquid
crystal and has similarity to a host material, 8-PNP-O12
in its chemical structure. This is beneficial to the suppression of phase separation and the maintenance of liquid
crystalline phases in a wide concentration range, which
enables us to establish a wide variation of the hopping
distance while keeping their microscopic molecular circumstances unchanged. Thus, it exhibited complete compatibility when mixed with 8-PNP-O12 and gave a liquid
crystalline mixture in the whole range of concentration
studied up to ~60 mol%. The mixture exhibited a new
smectic phase at the lower temperature region, which
is never seen in pure 8-PNP-O12, probably SmE. SmB and
SmA phases were well maintained in the lower concentrations less than 80 mol%. SmA phase was sensitive to
6O-BP-8 concentration and its temperature range was
decreased with an increase in the concentration and finally disappeared more than 80 mol%, as shown in Fig. 2.
To clarify true effects of the dilution on the carrier
transport, the transient photocurrent measurement was
focused on SmB, SmA and isotropic phases that appear
in pure 8-PNP-O12. 14 The photocurrents tended to be
reduced as the diluent concentration was increased. As
is expected from the ambipolar nature of carrier transport in pure 8-PNP-O12 reported previously, the mixed
liquid crystals also exhibited the ambipolar carrier
Kurotaki and Hanna
Temperature (°C)
Figure 4. Carrier mobility in 50 mol% diluted 8-PNP-O12 with
6O-BP-8 as a function of temperature. The squares and circles
indicate positive carriers and negative carriers, respectively .
The mobility was measured at 1 × 10 4 V/cm.
Figure 3. Transient photocurrents of 50 mol% diluted 8-PNPO12 with 6O-BP-8 in SmB phase illuminated with a N 2-laser
pulse of 337 nm; (a) For positive carriers; (b) for negative carriers. The cell thickness was 15 µm at 80°C (SmB phase).
transport basically. In SmA and isotropic phases, however, the transient signals for negative carriers were
too small to determine the transient time when the
diluent concentration was more than 40 mol%.
Figure 3 shows typical transient photocurrents in
SmB phase of 50 mol% diluted 8-PNP-O12. The transient photocurrents were non-dispersive irrespective of
the carrier sign. There was initial delay for positive car
riers probably due to carrier trapping at surface states
on the ITO electrode and initial decay for negative carriers. The µ/τt plot as a function of V/ d2 gave a welldefined line starting from zero, where µ is the mobility
of carriers, τt the transit time, V the applied voltage,
and d, a cell thickness, indicating that the mobility does
not depend on the applied electric field. The carrier
mobility was determined from a slope of the line and
was 6 × 10–4 cm 2/Vs irrespective of carrier signs. This
mobility was one-third of that in pure 8-PNP-O12. This
transport is reasonably attributed to the electronic conduction as discussed previously because of high viscosity in the present mixed liquid crystals. In SmA and
isotropic phases, the mobility for positive carriers was
determined to be 1.1 × 10 –4 cm2/Vs and 5 × 10–4 cm2/Vs,
respectively. These mobilities were corresponding to 40
~ 50% of those in the pure 8-PNP-O12, while the mobility could not be determined for negative carriers because
of small signals as described.
The mobilities at 1 × 10 4 V/cm are plotted as a function of temperature in Fig. 4. In smectic phases, the
Figure 5. Carrier mobility of diluted 8-PNP-O12 as a function of 6O-BP-8 concentration in SmB (80°C) and SmA (110°C)
1 × 10 4 V/cm.
mobility was independent of temperature, while
Arrhenius type of temperature dependence was observed
in isotropic phase. The ionic conduction is most likely
in isotropic phase as is the case of pure 8-PNP-O12.14
The present experimental results were summarized
in Fig. 5, where the mobilities in SmA and SmB phases
were plotted as a function of the 6O-BP-8 concentration
in 8-PNP-O12. The mobilities were almost equal for
positive and negative carriers and gave no difference in
the figure. In fact, the mobility was continuously decreased with an increase in the 6O-BP-8 concentration
irrespective of the phases, but its reduction remained
within one-third of corresponding mobilities of pure 8PNP-O12.
In smectic phases, all the molecules align in layers
with an orientation in one direction. The distance between the smectic layers is about 36A, which is very far
compared with 4A for the average distance between liquid crystalline molecules within a smectic layer. Therefore, it is reasonable that the photogenerated carriers
hop among the molecules in the smectic layer to reach
the counter electrode when the electric field is applied
parallel to the layer. This is the case of the present experiments, i.e., the homogeneous alignment.
Carrier Transport in Molecularly Diluted Liquid Crystalline Photoconductor
Vol. 43, No. 3, May/June 1999 239
Figure 6. Concentration dependence of carrier mobility in
molecularly diluted 8-PNP-O12 and typical molecularly doped
polymers.
Let us consider the effect of dilution semi-quantitatively. The HOMO level of 6O-BP-8 is higher than that
of 8-PNP-O12, because of small molecular-orbital of biphenyl moiety. Therefore, 6O-BP-8 is expected to be electronically inactive as an impurity as far as 6-PNP-O12
molecules contribute as the major hopping passway .
Therefore, 6O-BP-8 can be an ideal diluent for 8-PNPO12. The average distance between 8-PNP-O12 molecules for each concentration can be calculated and
range from 4A to 6.8A in the range of 0~60 mol%. This
is very different from more than 8A in the molecularly
doped polymers. It should be noted that the reduction
of mobility remained within 40~50% in spite of this large
change in the hopping distance.
Assuming 2-D random hopping of carriers in the smectic layer, the experimental data were plotted as a function of the average hopping distance caluculated,
according to the relation, 18 µ ∝ ρ 2exp (-2 ρ/ ρ0) as is the
case of the carrier transport in disordered systems,
where µ is the mobility, ρ the average hopping distance,
and ρ0 a wavefunction decay constant of molecular or bital. In fact, in the fluid media exhibiting a high mobility over 10 –3 cm 2 /Vs, it is reasonable that the
contribution of translational molecular motion is negligible to the carrier transport or hopping, because of a
very short estimated residence time of ~10 –9 s at each
hopping site. As shown in Fig. 6 accompanied by representative results for the molecularly doped polymers, 19–21
log µ/ρ2 gives a linear relation with ρ. This result indicates that the carrier transport in smectic mesophases
can be described by the 2-D random hopping model
within the smectic layers. It is clear from Fig. 6 that in
the diluted liquid crystals, the average hopping distance
of 4~6.8A is fairly small compared with these of molecularly doped polymers of > 8A. This results from a closed
packing of the molecules due to the self-organization in
the mesophase. In addition, the slopes of lines for the
diluted liquid crystals are gentler than those of the molecularly doped polymers, indicating larger ρ 0 in the diluted liquid crystals. In fact, ρ0 was determined to be
2.3A and 2.4A for SmB and SmA phases, respectively.
These values are fairly large comaperd with a typical
value of 1 ~ 2A in the molecularly doped polymers, which
characterizes the fast carrier transport and explains a
small reduction of mobility , i.e., 50 ~ 60%, when the
diluent concentration was increased.
240
Journal of Imaging Science and Technology
As far as molecular ordering within the smectic layer
is concerned, there is a big difference between SmA and
SmB phases: in SmA phase, all the molecules sit at random in the smectic layer; on the other hand, the molecules sit in hexatic order in SmB phase. Therefore, it
is likely that microscopic circumstance of molecules in
the smectic layer is different in these phases in terms
of energetic and spatial disorder . In the pure 8-PNPO12, however, there is no difference in the carrier transport behaviors except a fairly big difference in the
mobility of one-order of magnitude. This is true in the
present diluted liquid crystals as well. As described, no
difference is observed even in the ρ0 characterizing the
hoping site. This indicates that the basic physical process is the same in SmA and SmB phases in terms of
determining hopping rate. Therefore, the mobility difference between these mesophases including those in
the pure 8-PNP-O12 is attributed to the difference in
the pre-factor mobility, µ0.
Disorder formalism as described in the following equation, where µ is the mobility, µ0 a prefactor mobility, σ
the density of states, ∑ a parameter characterizing the
degree of positional disorder, k the Boltzman constant,
E the electric field, T the temperatrure, and C an empirical constant of 2.9 × 10–4 (cm/V) 1/2, specifies the the
carrier transport characteristics in individual materials systems.
For ∑ ≥ 1.5,
µ(σ, ∑, E, T) = µ0 exp [–2σ/3kT)2] exp [C{σ/kT)2 – ∑2}E1/2]
For ∑ < 1.5,
µ(σ, ∑, E, T) = µ0 exp [–2σ/3kT)2] exp [C{σ/kT)2 – 1.52}E1/2]
In this formalism, if the disorder manifolds become
zero, the temperature and electric field dependences disappear as is the case of the molecular crystals. In the
liquid crystals however, the molecular alignment is not
fixed and thermally fluctuated, so that there exist the
disorder to any appreciable extent. Therefore, we cannot understand these unique carrier features of liquid
crystalline photoconductors in the framework of disor der formalism. There is a need for ample experimental
data to describe a total view of the carrier transport
before its formalism.
Conclusion
The electronic carrier transport properties in a diluted
liquid crystalline photoconductor, i.e., 6O-BP-8 and 8PNP-O12 system, was investigated by transient photocurrent measurements. The mobility was ambipolar ,
independent of electric field and temperature, and continuously reduced with an increase in the diluent concentration studied up to 60 mol%, but the reduction
remained within the range of one-third of that of pure
8-PNP-O12. In isotropic phases, however, the transport
was ionic whose mobility was on the order of 10 –5 cm2/
Vs and depended on temperature.
It is revealed that the carrier transport characteristics in the mesophase can be described by a 2-D random
hopping model, which is characterized by a relatively
small average hopping distance, of 4~7A and a large ρ0
of 2.4A. This explained an apparently gentle dependence
of the mobility on the diluent concentration.
There remain many interesting experimental results
to be explained, including no mobility dependence on
Kurotaki and Hanna
electric field and temperature, and no effect of molecular ordering in intra-smectic layer on the mobility. The
liquid crystalline system is a very unique materials system enjoying both macroscopic molecular alignment and
microscopic molecular disorder. The experimental and
theoretical understanding of carrier transport in this
system gives us a new insight into the understanding
of that in conventional disordered systems and re-inforce
its framework of our understanding.
Acknowledgements. We thank M. Funahashi for guiding material preparation and TOF measurements. This
work was partly supported by Grant-in-Aid for Scientific Research on Priority Area on Basic Research from
Ministry of Education, Sport, and Culture of Japan.
References
1. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for
Xerography, Marcel Dekker, Inc. New York, 1998.
2. C. W. Tang and S. A. van Slyke, Appl. Phys. Lett . 51, 12 (1987).
3. H. Bässler, Phys. stat. sol. (b) 175, 15 (1993).
4. P. M. Bosenberger and H. Bässler, J. Chem. Phys . 95, 5327 (1991).
5. M. Sugiuchi and H. Nishizawa, J. Imag. Sci. Technol. 37, 245
(1993).
6. H. Nishizawa, M. Sugiuchi and T. Uenohara, Proc. Mat. Res. Soc .
227, 33 (1992).
7. A. Dieckman, H. B. H. Bässler, and P. M. Borsenberger, J. Chem.
Phys . 99, 8136 (1993).
8. A. Hirao and H. Nishizawa, Phys. Rev. B 56, R2904 (1997).
9. D. H. Dunlap, P. E. Parris and V. M. Kenkre, Phys. Rev. Lett . 77,
542 (1996).
10. P. M. Borsenberger, W. T. Gruebaum, L. J. Serriero, and N.
Zumbulyadis, Jpn. J. Appl. Phys . 34, L1597 (1995).
11. D. Adam, F. Closs, T. Frey, D. Funhoff, D. Haarer, H. Ringsdorf, P.
Schumacher, and K. Siemensmeyer, Phys. Rev. Lett. 70, 457 (1993).
12. D . A d a m , P. S c h u h m a c h e r, J . S i m m e r e r, L . H a u s s l i n g , K .
Siemensmeyer, K. H. Etzbach, H. Ringsdorf, and D. Haarer, Nature
371, 141 (1994).
13. M. Funahashi and J. Hanna, Phys. Rev. Lett . 78, 2184 (1997).
14. M. Funahashi and J. Hanna, Appl. Phys. Lett. 71, 602 (1997).
15. M. Funahashi and J. Hanna, Appl. Phys. Lett . 73, 3733 (1998).
16. Handbook of Liquid Crystals, D. Emus, J. W. Goodby, G. W. Spiess,
and H. W. VillP, Eds., Wiley-VCH, 1998.
17. A. Miller and A. Abrahams, Phys. Rev. 120, 745 (1960).
18. K. Kogo, T. Gouda, M. Funahashi, and J. Hanna, Appl. Phys. Lett .
73, 1595 (1998).
19. G. Pfister, Phys. Rev. B 16, 3676 (1977).
20. J. X. Mack. L. B. Schein and A. Peled, Phys. Rev. B 11, 39 (1989).
21. G. Pfister, Phys. Rev. B 16, 3676 (1977).
Carrier Transport in Molecularly Diluted Liquid Crystalline Photoconductor
Vol. 43, No. 3, May/June 1999 241
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Effect of Metal Contact Fabrication on the Charge Injection Efficiency of
Evaporated Metal Contacts on a Molecularly Doped Polymer
A. Ioannidis,▲ J. S. Facci and M. A. Abkowitz
Center for Photoinduced Charge Transfer, University of Rochester, Rochester, New York and
Wilson Center for Research and Technology, Xerox Corp., Webster, New York
We previously reported that contact injection efficiencies are amenable to direct measurement in thin trap-free hole transport
polymer
films by a technique combining time-of-flight bulk mobility measurements with steady state current densities measured at the ntact
co
under test. In the present article, films of a trap-free molecularly doped polymer , TPD/polycarbonate, are solution coated onto a
carbon-filled polymer substrate that is demonstrably ohmic for hole injection. Thermally evaporated Au and Ag as well as liquid Hg
form the top contacts. Field dependent contact injection efficiencies are computed from the combined measurements and monitored
over time. A persistent pattern in the evolution of contact injection efficiency with time is revealed. Invariably contact inje
ction
efficiencies evolve by orders of magnitude from initially blocking to ohmic or to an equilibrium value dependent on the nature of the
metal. For thermally evaporated Au contacts, coating studies suggest that the slow stage in the observed two-stage evolution of
contact formation represents a process of recovery from damage to the transport layer ’s surface caused by the accumulating hot Au
atoms. Such a process is not observed for substrateAu. Comparisons of the evolution in injection efficiencies of evaporatedAg contacts
with substrate Ag, as well as of liquid Hg contacts, demonstrate that the initially blocking nature of the contact and a fast evolution
process are not associated with recovery of the interface from thermal damage but are probably a more general aspect of contact
formation.
Journal of Imaging Science and Technology 43: 242–247 (1999)
Introduction
Fundamental questions concerning metal/organic interfaces are receiving much attention, stimulated by the
wide application of organic films in electrophotography1
and in the rapidly expanding field of organic electronic
devices.2 In these applications, the manner in which electrical contact is made to the molecular material is ultimately critical to device operation. Contacts of metals
to conventional semiconductors and insulators are interpreted 3 within the framework of band theory . However, it is difficult to rationalize the application of models
developed for band-type materials to the case of disor dered molecular materials (e.g., polymers and small
molecules) where carriers are localized and transport
involves discrete hopping within a distribution of energy states. An issue of current interest is the form of
the correlation of charge injection into disordered or ganic materials with the interfacial energy barrier as
estimated from the relative workfunctions of the interface materials. However, “workfunctions” of molecular
materials are either inferred from electrochemically
determined oxidation potentials or measured microscopically in ultra-high vacuum.4 These may not relate
directly to barriers formed under less pristine conditions
for example those typically encountered using solution
coating methods. Furthermore, it is unclear how band-
type treatments of energy barriers, such as those incorporating the Richardson-Schottky theory of thermionic
emission or the Fowler-Nordheim tunneling model,5 can
be simply recast to apply to disordered molecular materials. Models of injection and charge transport geared
specifically to hopping systems have in fact been recently
proposed by Conwell and coworkers 6,7 and by Bässler
and coworkers. 8–11
The problem of separating charge transport and interface effects on the injection current is clearly solved
in the case of trap-free molecularly doped materials. The
present study is based on a method previously described12,13 for obtaining the injection efficiency of a contact on a trap-free unipolar transport medium, e.g., a
molecularly doped polymer (MDP). The contact injection efficiency of an evaporated-Au/MDP interface is
probed directly by comparing the small signal time-offlight (TOF) behavior with field dependent steady state
current measurements. The trap-free MDP is a glassy
solid solution of an electroactive triarylamine derivative TPD, (shown below) in an inert polycarbonate matrix. By inert it is meant that charge transport does not
occur through states associated with the polymer, even
though the mobility can vary by orders of magnitude
depending on the nature of the polymer.14
Original manuscript received October 29, 1998
▲ IS&T Member
© 1999, IS&T—The Society for Imaging Science and Technology
242
N
CH3
N
CH3
The metal contact is deposited by thermal evaporation,
which has been the method of choice for metal contact for
mation in organic devices such as Light Emitting Diodes
(LEDs) or Thin Film Transistors (TFTs).15 In the following we analyze the evolution of hole injection efficiency
from an evaporated Au contact on the MDPas it increases
over time from emission limited or blocking to the maximum efficiency which in the case of Au is an ohmic contact. This time dependent process is not observed when
the MDP is coated onto a preformedAu substrate which
is initially ohmic. We describe injection evolution for the
evaporated Au contact in detail for a series of samples,
characterizing the time and temperature dependence of
the phenomenon to elucidate the nature of contact efficiency recovery process. It proves fruitful to extend these
measurements to Ag and liquid Hg contacts.
Previous analysis16 of the kinetics of the evaporatedAu/MDP interface formation revealed a non-linear behavior, composed of a rapid increase in injection
efficiency at early times (hours) followed by a much
slower increase over a period of weeks. The results were
represented by two exponentials with associated time
constants, and thus two main processes were distinguished. The rate of the evolution at early times was
temperature dependent and an Arrhenius plot yielded
an activation energy of ca. 0.3 eV . In order to investigate what mechanisms may be responsible for these two
evolution processes, it is helpful to regard a sample as
a three-component system, comprised of the MDP , the
Au and their interface. A major mechanism postulated
to operate with metal contacts on organic thin films is
the diffusion of metal atoms arriving from the vapor
phase into the transport film as recently reported.17 For
example, metal diffusion is observed in perylenetetracarboxylic dianhydride (PTCDA) thin films coated
with reactive metal contacts (Al, T i, In, Sn) but is not
observed in the case of the least reactive metalsAu and
Ag. For the MDP the possibility of Au penetration and
diffusion was investigated 16 by transmission electron
microscopy (TEM). Results of these studies showed the
interface to be abrupt to within 1 nm and invariant over
time, i.e., no Au/MDP interpenetration was observed
over a period of at least one month.
The present work describes the effect of Au contact
fabrication on hole injection efficiency. Two distinct time
dependent processes governing the evolution of the contact injection efficiency immediately after fabrication
are identified. To test the generality of this result, comparisons with other metal contacts, e.g.,Ag and Hg, have
also been carried out. The slow or long-term process is
correlated with metal evaporation conditions. The underlying mechanism is discussed. On the other hand the
rapid or short term process is shown to be less sensitive
toward the detailed conditions of metal fabrication and
appears to be a persistent feature of the MDP/metal interface formation. Manipulations geared toward investigating the role played by the organic surface in the
evolution of hole injection efficiency and the generality
of the observations for other organic materials will be
reported elsewhere.
Experimental
The sample configurations used are shown in Fig. 1(a)
and an illustration of the graphical method used to determine hole transit time and hole drift mobility from
the knee of the TOF transient photodischarge curve is
presented in Fig. 1(b). Molecularly doped polymer films
were solution coated onto a carbon-filled polymer con-
Figure 1. (a) Schematic diagram of the experimental sample
configuration. 40 wt% films of TPD/polycarbonate are coated
onto a MystR ® substrate which is ohmic for injection into the
TPD layer. The Al top contact is used for obtaining TOF results and the test contacts are vapor deposited Au or Ag as
well as liquid Hg fitted onto the same film as the Al contact.
(b) Graphical method for obtaining the transit time of carriers
from the knee of the small signal TOF photodischarge curve.
tact (MystR®) to thicknesses of 20–30 µm from a 4 wt%
methylene chloride solution of TPD and polycarbonate (40/
60 wt%). Films were slowly dried in a local atmosphere
saturated with methylene chloride, cured for 30 min in a
convection oven over a gradient of temperatures ending
in 110°C, and finally allowed to cool to room temperature
before evaporating the top metal contacts. All metal contacts were evaporated by resistive heating of the metal
source, producing films of 220–250 Å in thickness.
The small signal hole drift mobility, µ, is obtained by
measuring the time ( t tr ) required for a photoinduced
charge packet to transit the sample thickness in a conventional time-of-flight (TOF) experiment, such that18
µ = d / E ttr
(1)
where d is the sample thickness and E = V/d is the average electric field. The TOF experimental arrangement has
been described previously.12,13,19
The measurement of µ enables the calculation of the
trap-free space charge limited current, the maximum
current that may be sustained by the bulk, according to
Child’s Law,5
Effect of Metal Contact Fabrication....on a Molecularly Doped Polymer
Vol. 43, No. 3, May/June 1999
243
contact between successive J Au measurements in the
same film, remains invariant with time, ensuring that
any changes in J Au are not due to a change in the bulk
transport property of the film. The injection current from
the bottom MystR ® contact which is ohmic for hole injection is also monitored periodically and serves as another control measurement. The contact ohmicity of
MystR ® is illustrated in the inset of Fig. 2, which demonstrates that the calculated trap-free space-charge limited current densities JTFSCLC (open squares) coincide with
the measured steady state dark injection current (solid
line) from the MystR® substrate Jm. An equivalent measure of injection efficiency for a contact under test in
the present transport system can therefore be obtained
by direct comparison of the injection current at the test
contact to that at the MystR® contact Jm, viz.,
Injection Efficiency = JAu / Jm = JAu / JTFSCLC.
(4)
Injection efficiency is computed at a common field, 1× 105
V/cm.
Scanning electron microscopy (SEM) and x-ray diffraction (XRD) measurements on 220 Å thick Au films deposited at 10 Å/s were obtained as a function of time
following deposition, allowing the contact to age under
ambient conditions. The latter analyses were done in
parallel with injection efficiency measurements.
Figure 2. Temporal evolution of the hole injection current from
a 220 Å Au contact evaporated at 10 Å/sec onto a thin film of 40
wt% TPD/polycarbonate. The latter are compared with Jm which
represents the level of ohmic injection current. Note that the top
Au contact evolves from emission limited to ohmic over the course
of the analysis (15 min to 30 days). Inset: comparison of J TFSCLC
(open squares), the trap-free space-charge limited current density calculated from hole drift mobility in 40 wt% TPD/polycar bonate with Jm, the steady state dark injection current from the
MystR® substrate. The coincidence of the two current densities
demonstrates that the substrate is ohmic for injection into 40
wt% TPD/polycarbonate.
JTFSCLC = (9/8) (ε εo µ E2 / d),
(2)
where ε is the relative dielectric constant andεo is the permittivity of vacuum. A measured current density that coincides with the calculated J TFSCLC is the necessary and
sufficient condition to classify a material as trap-free.
However, in the case of TPD/polycarbonate, there is also
well known experimental confirmation as xerographic
charging/discharging measurements consistently reveal no
hole trapping at either short or very long time scales. The
ratio of the injected current density from an evaporated
metal contact against J TFSCLC is defined to be the contact
injection efficiency (illustrated henceforth for Au), i.e.,
Injection Efficiency = JAu / JTFSCLC.
(3)
Details of the rationale for this quantitative determination of injection efficiency have been previously discussed. 12,13,19 The hole drift mobility, monitored at the Al
244
Journal of Imaging Science and Technology
Results and Discussion
Hole injection currents from evaporated Au and Ag contacts, as well as liquid Hg, were obtained as a function
of field and time under ambient conditions. Current density versus field data for all metal/MDP samples show
no hysteresis and results were reproducible for several
sample sets. In the case ofAu, measurements were performed under a variety of metal evaporation conditions.
Typical J Au versus field data parametric in time are
shown in Fig. 2 for a film of 40 wt% TPD/polycarbonate
on MystR ® at 23 °C. The figure shows the evolution in
J Au, compared to the time invariant Jm curve.
SEM and XRD measurements were performed in or der to investigate the possibility that the morphology
or surface texture of the metal film itself is changing
with time, potentially affecting the contact workfunction
or actual contact area. The XRD results show two peaks
of equal amplitude corresponding toAu(111) and Au(222)
and both their amplitude their relative ratios remain
invariant from 1 h to 2 weeks after Au deposition. Note
that this corresponds to a time span encompassing both
evolution processes. Scanning electron microscopy results were obtained over the same time span at 300 nm,
600 nm and 1 µm resolutions and show a cracked Au
film morphology. SEM results reveal a porous film in
which the density and size of cracks does not, however,
change in time. Therefore, the evolution in injection efficiency cannot be readily assigned to changes in metal
film structure.
A further possibility related to metal fabrication is
that during the metal evaporation, the energetic Au atoms or the accumulation of a hot Au layer on the MDP
may in some way damage the MDP surface. Such damage may be repaired over time by polymer chain or small
molecule diffusion which could act to replace damaged
surface molecules and indeed the time scale of the slow
evolution is not inconsistent with such a mechanism. 20
Accordingly, a systematic variation of metal evaporation conditions was performed. A comparison of four
evaporation conditions is shown in Figs. 3(a)–3(d). The
Ioannidis, et al.
Figure 3. Temporal evolution of the injection efficiency at 1.0 × 105 V/cm of evaporated Au contacts on 40 wt% TPD/polycarbonate as
a function of Au deposition conditions. All Au contacts are 220 Å. Panel A: Au is deposited in two steps, 50 Å and 170 Å, at 10 Å/sec.
Panel B: 11 layers of 20 Å each at 10 Å/sec. Panel C: 11 layers at 2 Å/sec. Panel D: a single continuous step at 2 Å/sec. In all cases the
injection efficiency is initially blocking.
Au contacts shown in Figs. 3(a)–3(c) are fabricated by an
incremental or sequenced deposition process. The deposition of each layer is separated by 1–2 min during which
the source is not heated and a shutter covers the MDP
surface. Specifically the Au contact in Fig. 3(a) is deposited in two stages, a 50 Å and a 170 Å layer at 10 Å/s; the
Au contact of Fig. 3(b) is composed of 1 1 20 Å depositions at 10 Å/sec and finally theAu contact of Fig 3(c) is
composed of 11 20 Å depositions at 2 Å/s. Note that all
layered depositions result in a dramatic reduction in
the time scale of the long-term evolution process. Each
of the evaporated Au contacts deposited under these
conditions achieves contact ohmicity in under 20 h. This
is significantly shorter than the ~800 h required to
achieve contact ohmicity for a continuous Au deposition
at 10 Å/s as depicted in Fig. 2. Decreasing the rate of Au
deposition from 10 Å/s to the 2 Å/s in a continuous deposition of the Au contact does not significantly reduce the
time required to achieve contact ohmicity from 800 h as
suggested by Fig. 3(d), i.e., no change is noted in the long
term process. These manipulations of metal evaporation
conditions therefore suggest that damage inflicted of the
MDP surface is thermally induced by the arrival of hot
Au atoms during a typical continuous evaporation and this
damage can be virtually eliminated if Au is deposited in
multiple layers separated by short cool-down periods.
Especially notable in Fig. 3 is that in no case is the
initial blocking nature of the interface affected nor the
rapid initial rise in injection efficiency . Therefore, in
agreement with the prior kinetic studies,16 these results
indicate strongly that there are indeed two distinct processes governing the evolution in contact behavior. However, in order to more fully understand the nature of
this behavior, the behavior of contacts which are fabricated under conditions that ensure the interface suffers minimal thermal damage was investigated. Toward
this end the injection behavior using pre-formed metal
substrates was investigated. In this case the metal is
evaporated onto glass and the MDP film is then solution coated onto the metal in the usual manner.
Injection efficiency results for Au substrate contacts
showed these contacts were always ohmic at the minimum measurement time of ca. 3 h. The minimum measurement time is established by the sample preparation
steps of solution casting, evaporation, curing and evaporation of the top contact.
Effect of Metal Contact Fabrication....on a Molecularly Doped Polymer
Vol. 43, No. 3, May/June 1999
245
Figure 4. Comparison of the temporal evolution of the injection efficiency J Ag/JTFSCLC for a freshly deposited Ag substrate (a) with the
injection efficiency JAg/Jm of an evaporated Ag top contact (b).
In contrast with the above results, Fig. 4(a) presents
injection efficiency versus time for a sample ofAl/MDP/
Ag-substrate. In the latter case injection efficiency is
computed from the TFSCLC obtained from TOF mobility measurements. The injection efficiency for the Agsubstrate sample is monitored beginning 3 h after Ag
substrate is contacted with the polymer casting solution.
As distinct from the case of Au substrate contact, an evolution in injection efficiency is observed stronglysuggesting a contact forming process independent of the effects
of interfacial damage arising from metal deposition.
Figure 4(b) shows the injection efficiency (J Ag/Jm)
versus time for a comparable MDP film with an evaporated Ag top contact that was deposited layer-by-layer.
It is anticipated from the results depicted in Figs. 3(a)–
3(c), that thermally induced interfacial damage fromAg
contacts fabricated under these conditions is minimized.
Comparison of Figs. 4(a) and 4(b) indicate that the evolution of injection efficiency in both the Ag substrate
and evaporated Ag contact are comparable. This comparison of injection from a Ag substrate to that from a
Ag layered top contact suggests that the layer -by-layer
deposition eliminates most, but not all of the effect ofmetal
deposition. Note that steady state injection efficiency
of 0.6 is also similar in both cases. Ag contacts, unlike
Au, never become ohmic for injection of holes into the
MDP. This is consistent with the workfunction of Ag
which is ca. 0.5 eV less than that ofAu.21 A further comparison in time scales between Fig. 4(b) and layeredAu
contacts in Figs. 3(a)–3(c) indicates that the “fast” process is markedly slower for the evaporatedAg (top) contact. On this basis Ag was chosen as the substrate metal
in order to maximize the chance of observing any relaxation behavior because measurements cannot be per formed until after the MDP film has been cured.
246
Journal of Imaging Science and Technology
Figure 5. Temporal evolution of the injection efficiency J Hg/Jm
for a liquid Hg droplet contact made to a 40 wt% TPD/polycar bonate film. The area of the liquid contact is defined by a Teflon
template containing a 0.316 cm2 hole.
Finally, Fig. 5 depicts injection efficiency versus time
for a Hg/MDP/MystR® sample, showing that contacts of
a liquid Hg droplet of well defined area made to the
sample surface also give rise to an evolution in injection behavior. This contact is particularly interesting
Ioannidis, et al.
in that any possibility of thermal damage to the inter face is precluded. In addition, it is also possible to begin
measurements of current density almost immediately after top contact formation. Taken together, these results
indicate time dependent contact formation processes
operate with a variety of metals under a variety of fabrication conditions. In particular a persistent “fast” contact formation process is operational, independent of the
nature of the top contact and the manner in which it is
fabricated.
Conclusion
The time dependence of the injection efficiency from evaporated Au into a trap-free molecularly doped polymer has
been observed for various metals under different fabrication conditions. Differently prepared Au, Ag and Hg contacts all show an initially emission limited hole injection
efficiency into the MDP and this efficiency increases over
time. Two processes governing the evolution in efficiency
can be distinguished for evaporated Au and Ag contacts,
consistent with earlier analyses of the kinetics of injection evolution from evaporatedAu. In the case ofAu, whose
workfunction is near that of the MDP (~5.5 eV) the injection current becomes ohmic for both substrates and evaporated top contacts. No change in metal morphology or
surface texture over time was detected.A systematic variation of the conditions of metal evaporation shows that a
slow, long-term component of the evolution is due to the
method of evaporation and can be virtually eliminated by
performing a layer -by-layer metal deposition. This par ticular solution to the fabrication problem indicates that
the MDP surface is thermally damaged during a typically
rapid metal deposition of energetic Au accumulated continually on the sample surface. The long-term evolution
process would then reflect recovery of the MDP surface.
This recovery may be due to polymer chain motions that
act to replace damaged segments at the surface or diffusion of the molecular dopant, TPD, that would restore a
surface concentration of TPD depleted by sublimation
during the heating of the MDP surface. On the other hand
we have distinguished an early time or rapid component
of the evolution in injection efficiency which is determined
to be largely independent of a degradative process of kinetic origin.
Acknowledgments. The authors gratefully acknowledge
the technical assistance of H. Freitas and J. Czerniawski
for TEM measurements and S. Ingham for X-ray diffraction measurements. This work was supported by a Science and Technology Center Grant CHE-91-20001.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors For Imaging Systems, Marcel Dekker, New York, 1993.
M. Pope, Mol. Cryst. Liq. Cryst. 223, 1 (1993).
L. J. Brillson, Surface Science Reports 2, 123 (1982).
M. Fujihira, Forces in Scanning Probe Methods, Vol. 286, Kluwer Academic Publishers, Dordrecht, 1995.
M. A. Lampert and P. Mark, Current Injection in Solids, Academic Press,
New York, 1970.
E. M. Conwell, Handbook of Organic Conductive Molecules and Polymers, Vol. 4, John Wiley and Sons, New York, 1997.
M.-W. Wu and E. M. Conwell, Chem. Phys. Lett. 266 (1997).
H. Bässler, Phys. stat. sol. 175, 15 (1993).
B. Movaghar, M. Grunewald, B. Ries, H. Bässler, and D. Wurtz, Phys.
Rev. B 33, 5545 (1986).
P. M. Borsenberger, L. Pautmeier and H. Bässler, J. Chem. Phys. 94,
5447 (1991).
U. Wolf, H. Bässler, P. M. Borsenberger, and W. T. Gruenbaum, Chem.
Phys. 222, 259 (1997).
M. A. Abkowitz and D. M. Pai, Phil. Mag. B 53, 193 (1986).
M. A. Abkowitz, J. S. Facci and M. Stolka, Appl. Phys. Lett. 63, 1892
(1993).
W. T. Gruenbaum, L.-B. Lin and P. M. Borsenberger, Phys. stat. sol. B
204, 729 (1997).
N. C. Greenham and R. H. Friend, Solid State Physics 49, 1 (1995).
A. Ioannidis, J. S. Facci and M. A. Abkowitz, J. Appl. Phys. 84, 1439
(1998).
Y. Hirose, A. Kahn, V. Aristov, P. Soukiassian, V. Bulovic, and S. R.
Forrest, Phys. Rev. B 54, 13748 (1996).
A. R. Melnyk and D. M. Pai, Determination of Electronic Optical Properties, Vol. 8, John Wiley and Sons, New York, 1993.
M. A. Abkowitz, J. S. Facci and J. Rehm, J. Appl. Phys. 83, 2670 (1998).
R. P. Wool, Polymer Interfaces, Hanser-Verlag, Munich, 1995.
H. B. Michaelson, J. Appl. Phys. 48, 4729 (1977).
Effect of Metal Contact Fabrication....on a Molecularly Doped Polymer
Vol. 43, No. 3, May/June 1999
247
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Photoconduction Mechanism in Single-Layer Photoconductor with
Metal-Free Phthalocyanine
K. Kubo, T. Kobayashi, S. Nagae, and T. Fujimoto
Advanced Technology R&D Center, Mitsubishi Electric Corporation, Hyogo, Japan
The photoconduction mechanism in a metal-free phthalocyanine pigment dispersed in a polymer matrix was investigated. The
charging potential started to decay remarkably after a threshold light exposure. The threshold exposure increased as the initial
potential increased and as the thickness of the photoconductive layer decreased. This result may indicate that the threshold
exposure depends on the quantity of charge. The temperature dependence of the threshold exposure was also investigated. The
threshold exposure decreases with increasing temperature. The activation energy was estimated to be 0.049eV at an electric
field of 4.5 × 105V/cm. This value is almost equal to that of the photogeneration process in phthalocyanine. The photoinduced
decay rate after the induction period increased and the activation energy decreased with increasing field intensity . The anticipated field dependent phenomenon was not found in these results. Therefore, we think there is a possibility that the mechanism
is different from the prevalent trap theory.
Journal of Imaging Science and Technology 43: 248–253 (1999)
Introduction
Organic photoconductors are useful as electrophotographic photoreceptors in copy machines and laser printers. The electrophotographic process comprises
charging, exposing, developing, transfer and fixing
steps. The photoreceptor forms electrostatic latent images in the charging and exposing processes, then electrically adsorbs toners in the developing process, and
puts them on paper in the transfer and fixing process.
Most photoreceptors are negatively charged in an electrophotographic system. In general, corona discharge
devices are available to charge photoreceptors. These
devices generate ozone as by-product and the amount
of ozone is greater with negatively charged photoreceptors than with positively charged photoreceptors.1
The single-layer photoconductor consisting of a phthalocyanine pigment dispersed in a polymer matrix2 is
a positively charged photoreceptor. This photoconductor
is more eco-friendly than negatively charged ones. It
comprises a single layer that has two functions: photocarrier generation and carrier transport. Its structure
is so simple that it can be easily manufactured. Fur thermore, the single-layer photoconductor is a high
gamma photoreceptor.3 It can sharpen the edge of electrostatic latent images in a laser printer and create
hardcopies printed with high resolution and high quality. Therefore, such a photoconductor is a very promising electrophotographic photoreceptor.
The unique high gamma property occurs because the
charging potential starts to decay remarkably after a
Original manuscript received July 30, 1998
© 1999, IS&T—The Society for Imaging Science and Technology
248
certain threshold level of light exposure that we call the
“induction effect”.2,4 Several mechanisms have been proposed to explain this phenomenon, for example, the
simple trap model 5,6 and the structural trap model. 7,8
However, the induction mechanism is not completely
clear at this time.
In this study, we investigated the photoconduction in
the single-layer photoconductor with a metal-free phthalocyanine pigment to clarify the properties and
mechanism of the induction. The exposure in the induction period was measured at various initial charging
potentials for photoconductor samples of various thicknesses and at various temperatures. Then the results
were analyzed to estimate the activation energy in the
induction process. Subsequently, the decay rates after
the induction period were measured at various temperatures, and we estimated the activation energy in the
charge transport process. The mechanisms of the
photoconduction and the induction effect are discussed
on the basis of these results.
Experimental
Sample Preparation. The photoconductor consisted of
an x-type metal-free phthalocyanine pigment dispersed
in a resin matrix, which was a mixture of polyester and
melamine polymers. The composition of the phthalocyanine and resin was 23 and 77 wt%, respectively . The
samples were prepared as follows. First, a phthalocyanine pigment was deflocculated and dispersed in the
resin with an organic solvent in a paint shaker. Then, a
photoconductive layer was formed on a conductive substrate by dip coating, and the sample was air-dried and
cured. The samples for the photoinduced decay measurements were formed on aluminum plates with oxide coating. A sandwich-type cell, in which the photoconductive
layer was sandwiched between indium tin oxide (ITO)
F i g u r e 1 . Photoinduced decay curve of single-layer
photoconductor with metal-free phthalocyanine; λ = 780 nm, I
=1.9 µw/cm2. Sample thickness = 16.8 µm.
Figure 2. Photocurrent curve of photoconductor with metalfree phthalocyanine; λ = 780 nm, I = 1.5 µw/cm2. Sample thickness = 18.0 µm, bias = 400v.
on a glass substrate and vacuum-deposited gold (Au),
was used for the photocurrent measurements.
Measurements. The photoinduced decay was measured
by evaluating the electrophotographic properties of the
photoconductors (GENTEC Cynthia59). The samples
were positively charged with a corona discharge device
and irradiated by monochromatic light ( λ = 780 nm).
The photocurrent was measured with a picoammeter
and a voltage power supply , and a cryostat controlled
the sample’s temperature. The thickness of the photoconductive layer was evaluated with a thickness meter
and determined as an average value at several points.
Results
Induction Effect. Figure 1 shows a photoinduced decay curve (PIDC), where the charging potential is plotted against the exposure time. The charging potential
is nearly constant at the beginning of exposure, but decays remarkably when the total light exposure reached
a certain threshold. This phenomenon is the induction
effect.
The threshold exposure necessary to start decay was
estimated from the product of time t in Fig. 1 and the
light intensity. Time t is designated as the intersection of two dotted lines extrapolated from the line before and after the start of significant decay . The
threshold exposure was found to be almost independent of light intensity.
Figure 2 shows the photocurrent as a function of exposure time, measured in the sandwich-type cell with
constant applied voltage. Light intensity, sample thickness and voltage were similar to those of the PIDC in
Fig. 1. The induction effect in the photocurrent was not
found on the same time scale as in the PIDC. This indicates that the induction effect is a unique phenomenon
in the PIDC, perhaps because constant voltage is applied to the photoconductor during the photocurrent
measurement or because a strong charge could be momentarily injected into the photoconductor in the sandwich-cell configuration.
Potential Dependence of the Threshold Exposure
Figure 3 shows the threshold exposure against the initial charging potential in 7. 0, 11.6 and 16.8 µm thick
Photoconduction Mechanism in Single-Layer Photoconductor......
Figure 3. Plots of threshold exposure of photoconductor with
metal-free phthalocyanine against initial charging potential.
samples. The threshold exposure increases almost linearly with increasing potential. The number of photons estimated from the threshold exposure was
approximately 10 11–10 12/cm 2. It was the same order as
that of the charge carriers for charging up to the initial potential.
If we compare this result for samples of different thickness, it is clear that the thinner sample needs a greater
threshold exposure at the same potential. These results
are consistent with the relationship between the potential and the quantity of charge.
Figure 4 shows plots of the threshold exposure in Fig.
3 against the electric field intensity. The threshold exposure is larger in the thicker sample at the identical
electric field and at the identical charge density. However, the threshold exposure is not directly proportional
to the thickness. This may indicate that charge traps in
a portion of the photoconductive layer have an influence on the threshold exposure.
Temperature Dependence of the Threshold
Exposure
Figure 5 shows the threshold exposure in the PIDC as a
function of temperature. The dependence on the tem-
Vol. 43, No. 3, May/June 1999
249
Figure 4. Plots of threshold exposure against electric field
intensity of photoconductor with metal-free phthalocyanine.
Figure 6. Arrhenius plot of reciprocal of threshold exposure
of photoconductor with metal-free phthalocyanine; initial potential = 522v; sample thickness = 11.6 µm.
rier generation efficiency η (T) is generally expressed
as
η (T) = η0
•
exp(–E0/kT),
(2)
where η0 is constant, E0 is the activation energy, k the
Boltzmann constant and T the temperature. 9
The combination of Eqs. 1 and 2 leads to
1/L (T) = C exp (–E/kT).
Figure 5. Temperature dependence of threshold exposure of
photoconductor with metal-free phthalocyanine. Initial potential = 522v; sample thickness = 11.6 µm.
perature suggests that the induction effect has a relationship with a thermally activated process, such as
charge generation, charge transport or carrier trapping.
The dependence, however, is negative, so that it is difficult to get information about the induction process from
the plots.
Therefore, we attempted to analyze the temperature
dependence of the threshold exposure. The existence of
the threshold exposure suggests that a certain constant
amount of photogenerated carriers is necessary before
the potential begins to decay . The certain constant
amount of photogenerated carriers F is proportional to
the product of threshold exposure L and carrier generation efficiency η. L and η are functions of temperature
T and are designated as L (T) and η (T). Then, F is expressed as
F=A
•
L (T)
•
η (T),
(1)
where A is a coefficient. The coefficient A is constant if
the thermally activated process in induction is the only
photocarrier generation process. If the induction process includes another thermally activated process, the
expression may include an exponential term. The car -
250
Journal of Imaging Science and Technology
(3)
Here C is a constant coefficient, and E is the activation
energy for the induction process (E = E0 or maybe E0 +
Eother).
Figure 6 is the Arrhenius plot of the reciprocal of the
threshold exposure shown in Fig. 5. The activation energy E is estimated to be 0.049eV, where the initial potential is 522V and the electric field is 4.5× 105V/cm. As
to the photocarrier generation in x-type metal-free phthalocyanine, Popovic9 obtained approximately 0.045eV
as the activation energy at 4.5 × 10 5V/cm. The good
agreement between these two values may indicate that
the thermally activated process in the induction period
is simply the photocarrier generation process.
Figure 7 shows the values of the activation energy in
Eq. 3 at several initial potentials. The values were 0.0670.049eV at a potential range of 290-522V and at a field
range of 2.5–4.5 × 10 5V/cm. As the initial potential increases, the activation energy decreases. The decrease
in the activation energy should accelerate the induction
process and decrease the threshold exposure; however,
the threshold exposure increases with an increase in
the initial potential, as shown in Fig. 3. Therefore, we
think that the thermally excited process is not the only
factor causing the induction effect.
Photoinduced Decay Rates After the Induction
The potential in the PIDC decays after the induction
period. We examined the temperature dependence of the
photoinduced decay rate at various initial potentials and
obtained information about the charge transport process. Figure 8 shows the Arrhenius plots of the decay
rates. Each decay rate was the largest decay rate measured in each PIDC and was estimated from the maximum in the first derivative of the PIDC. The decay rates
Kubo, et al.
Figure 7. Activation energy of induction effect as a function
of initial charging potential in photoconductor with metal-free
phthalocyanine; sample thickness = 11.6 µm.
increase as the temperature rises. This result indicates
that the charge transport process is positively dependent on the temperature. The activation energy of the
decay is 0.17 – 0.16eV at a potential range of 300– 522V
,
with an electric field range of 2.6 × 105 – 4.5 × 105V/cm.
Next, we obtained the activation energy of the charge
transport process from the measurements of the photocurrent. The charge transport process in the photocurrent measurement is presumed to be same as the process
in the photoinduced decay. The activation energy estimated from the photocurrent was 0.29 – 0.18eV at an
electric field range of 0.1 1 to 0.82 × 10 5V/cm. Figure 9
shows the activation energy values obtained from the
photocurrent and the photoinduced decay as a function
of field intensity. The activation energy of the charge
transport process is dependent on the electric field at a
low field range of 104 V/cm and only slightly dependent
at a high field range greater than 105 V/cm. As a whole,
the value decreases with increasing field intensity.
Discussion
Thus far, induction effect has been described as the phenomenon that causes an S-shaped PIDC for the negatively charged photoreceptors.5–8 The photoreceptor with
phthalocyanine is positively charged and the PIDC
seems to fall monotonically rather than be S-shaped
(Fig. 1). The threshold exposure is almost independent
of light intensity and the photoconductor has a high
gamma property. The mechanism of the induction effect in the photoconductor with phthalocyanine may not
always be the same as the mechanism that has been
suggested for negatively charged photoconductors.
The difference between the photoinduced decay and
the photocurrent results indicate that both charging carriers and photoinduced carriers play important roles in
the induction effect (Figs. 1 and 2). Moreover , the distribution and the quantity of charge seem to be important factors as well.
The increase in threshold exposure with increasing
potential indicates that the induction period increases
as the electric field intensity and the quantity of charge
increases (Fig. 3). The thickness dependence of the
threshold exposure approximately supports this result.
Figure 4 shows the existence of a factor such as charge
trapping that occurs during the induction period. How-
Photoconduction Mechanism in Single-Layer Photoconductor......
Figure 8. Temperature dependence of photoinduced decay
rates of photoconductor with metal-free phthalocyanine;
sample thickness = 11.6 µm.
Figure 9. Activation energies for photoinduced decay and photocurrent of photoconductor with metal-free phthalocyanine.
ever, a simple trap cannot be considered to cause the
induction effect because field intensity should reduce
the depth of the simple trap according to the PoolFrenkel effect,10 whereas the threshold exposure only
increases with increasing field intensity . As such, the
induction process occurs a little below the surface of the
photoconductor because the threshold exposure was not
found to be proportional to thickness. The existence of
a trap seems to complicate the explanation of the induction effect.
The induction process apparently includes a temperature-dependent process (Fig. 5), including the carrier
generation process. It is important to determine whether
or not there is any other temperature-dependent process. The comparison of the estimated activation energy
in the induction process with the value reported in the
photogeneration process indicates that there is little
possibility of including any other temperature-dependent process in the induction process (Fig. 6). Moreover
,
a decrease in the activation energy with increasing field
intensity is ordinarily a field-dependent change (Fig. 7).
The potential decay process after the induction period should reflect the charge transport process in the
photoconductor. The decay rates increase with increasing field intensity (Fig. 8), and the activation energy in
Vol. 43, No. 3, May/June 1999
251
the transport process decreases with increasing field intensity (Fig. 9). Compared with the activation energy of
the induction process, that of the transport process is
larger and the basis of each process is thought to be
different.
Several theories have been suggested to explain the
mechanism of the induction effect.5–8 One theory is based
on the simple trap model such that carrier traps are
filled with photogenerated carriers during the induction period, then the charging potential decays. 5 The
number of traps is related to the length of the induction
period in the PIDC. 5,6 The theory, however, cannot explain all of the properties of the induction effect as described above.
Another theory is based on structural traps. 7,8 The
structural traps are traps or dead ends in a random
network formed by the disordered dispersion of photoconductive particles or aggregates. Borsenberger and coworkers explain the S-shaped PIDC of their aggregate
organic photoconductor in terms of field-dependent trapping by structural traps, but their explanation is restricted to the electron-dominated photoinduced
discharge.8 This idea is similar to the disorder model
suggested by Bässler to explain the conductivity in molecularly doped polymers.11 That model is based on negative differential resistance against the field intensity of
the charge transport and has been experimentally demonstrated in molecularly doped polymers and evaporated
organic compounds.12,13 The negative differential resistance against the electric field is caused by field-induced
localization of charge in traps or dead ends. According
to the structural trap model, the induction effect can be
explained as the phenomenon occurring when the charge
is captured in traps or in dead ends in a random network and moves to the substrate with a gradual reduction in the local field intensity. The potential dependence
of the threshold exposure can be explained by the fieldinduced localization in traps or dead ends. However, the
negative resistance of the drift mobility against the field
intensity has not been demonstrated in the photoconductor with phthalocyanine at a practical field range. 5
In addition, the photoconductor is a positively charged
photoreceptor and the induction effect in a positively
charged photoreceptor has never been explained with
structural traps.
S-shaped dark decay was measured in the photoconductor with phthalocyanine. S-shaped dark decay did
not appear in the photoconductor that was not charged or
light-exposed, but after a repetition of charging and dark
decay or after a repetition of charging and photoinduced
decay, the S-shaped dark decay eventually appeared.
Moreover, we found that the dark decay rate increased
with increasing field intensity with a maximum at a low
field range. This means the dark decay rate also decreases with increasing field intensity at some higher
field range. These phenomena are believed to be caused
by traps and trapped carriers in the photoconductor. We
think the traps may be structural traps. The property
of dark decay is still under investigation.
We infer that the induction effect in PIDC of the
photoconductor with phthalocyanine is caused by another mechanism. Traps exist in the whole region of the
photoconductive layer. If photogenerated carriers fill all
traps in the induction period, the potential would then
decay. The potential did not decay much during this period, so the induction phenomenon seems to take place
specifically in the upper, sub-surface region of the layer.
This is supported by the report that the induction effect
is also measured in a double-layer photoconductor that
252
Journal of Imaging Science and Technology
has an upper charge generation layer including phthalocyanine.14
Let us suppose that the origin of the induction effect
begins in the structural traps. In the induction period,
enough photogenerated carriers would be generated to
neutralize field-dependent structural traps, and charge
carriers would start to move to the substrate. Because
the quantity of the charge would not change, the electric field intensity from the part where the carriers exist to the substrate would not change and the structural
traps would exist throughout the period. Therefore, the
structural traps should be related to both the induction
process and the charge transport process.
However, the activation energy in the induction effect was almost equal to that of the carrier generation
(Fig. 6) and does not show a field-dependent change as
predicted by the structural trap model (Fig. 7). Furthermore, the difference of the activation energy estimated
between the induction process and the charge transport
process indicates that the origins of these processes are
different from each other (Figs. 7 and 9). It seems unreasonable to conclude that the structural traps cause
the induction effect.
The charge transport process after the induction period is also a thermally activated process, which may
be an intermolecular or inter-particle hopping process.
The activation energy in the photoconduction process
was dependent on the field intensity in the low fields
and somewhat dependent in the high fields.
From the experimental results, we think that the
quantity and the distribution of the charge from corona
charging and the photogenerated carriers are important
in the induction process. The potential dependence of
the threshold exposure may result from the change of
the quantity of the charge due to corona charging. The
charging carriers are slightly distributed inside the photoconductive layer and the carriers exist in the binder.
This idea seems to be supported by a volume charge capacitor model, where charging carriers are distributed
inside the photoconductive layer in the photoconductor
with ZnO,15 and by the report of the depth of charge
injection in the metal/polymer contacts, where the depth
can be up to several microns from the surface.16 Accordingly, we consider the following explanation for the induction effect.
First, charging carriers are distributed slightly inside
the photoconductive layer and the carriers exist in the
binder polymer or at the interface between the binder
and phthalocyanine particles. Then, at the beginning of
the photoinduced decay, the charge needs to move from
charge holders, such as the binder, to a conductive material, such as phthalocyanine, for the charge transport.
At that time, the minus charge of the photogenerated
carriers neutralizes the charging carriers around the
phthalocyanine particles and a hole remains in the particle, so that it is a conductive material. Consequently,
the positive charge from corona charging is taken into
phthalocyanine as a hole. From this point, we present
two different mechanisms as follows.
One mechanism is based on the change of the local
electric field. At first, the space charge field formed by
charging carriers interrupts hole transport through phthalocyanine particles. Then, as the charge is taken
into phthalocyanines, the electric field toward the substrate becomes stronger and holes can move to the substrate. Otherwise, holes move by diffusing up to the
edge of the charging region. Some photocarriers are
necessary to change the local field intensity or to diffuse the carriers.
Kubo, et al.
The other mechanism is based on a production of the
charge transport paths. The photogenerated electrons in
phthalocyanines are given to electron-accepting materials,
such as oxygen, and the electron-accepting materials facilitate charge transport paths. When the photogenerated
electron neutralizes the charging carriers around the phthalocyanine, the electron-accepting materials cannot
accept electrons and do not facilitate or spread the charge
transport paths, so that the charge does not move and
the potential does not decay.
These mechanisms qualitatively explain the induction
effect but are not well demonstrated. Therefore, to further clarify the mechanism of the induction effect, more
information about the phenomenon is necessary. In addition, it is important to discuss how the charge is distributed in the photoconductor , how photogenerated
carriers move and what material keeps charges,
photogenerated holes and electrons.
Conclusion
The photoconduction mechanism in the photoconductor
with a metal-free phthalocyanine pigment dispersed in
a polymer matrix was investigated. We obtained the following results.
1. The threshold exposure in the induction effect
increased with increasing initial potential and with
decreasing thickness of the photoconductive layer.
2. The activation energy in the induction effect was
estimated to be 0.049eV at the electric field 4.5 ×
10 5 V/cm, which is nearly equal to that of the
photocarrier generation.
3. The activation energy in the induction effect was
estimated to be 0.067–0.049eV at a field range of 2.5–
4.5 × 105V/cm. The values decreased with increasing
field intensity.
Photoconduction Mechanism in Single-Layer Photoconductor......
4. The activation energy of the charge transport process
was estimated to be 0.29 – 0.18eV at an electric field
range of 0.11 to 0.82 × 105V/cm and 0.17 – 0.16eV at
an electric field range of 2.6 × 10 5 – 4.5 × 105V/cm.
The activation energy of the charge transport process
decreases with increasing field intensity.
From these results, we infer the possibility of a different mechanism from the prevailing trap theories. The
suggested mechanisms pay attention to the quantity and
the distribution of charging carriers and photogenerated
carriers. We think it is necessary to study the movement of charge carriers in a microscopic area to more
clearly elucidate the induction effect.
References
1. E. M.Williams, The Physics and Technology of Xerographic Processes,
John Wiley and Sons, Inc., New York, 1984, p. 51.
2. W. F. Berg and K. Hauffe, Current Problems in Electrophotography,
Walter de Gruyter, New York, 1972, p. 287.
3. J. Decker,K. Fukae, S. Johnson, S. Kaieda, and I. Yoshida, Proc. IS&T’s
7th Intl. Congress on Advances in Non-Impact Printing Technol. Vol.1,
IS&T, Springfield, VA, 1991, p. 328.
4. H. Ueda and T. Noda, Minolta Techno Report, 4, 21 (1987).
5. A. Omote, Y. Itoh and S. Tsuchiya, J. Imag. Sci. Technol. 39, 271 (1995).
6. K. Kitamura and H. Kokado, Soc. J. Electrophot. Jpn 20(2), 60 (1982).
7. K. Oka, Soc. J. Electrophot. Jpn. 37(1), 53 (1998).
8. P. M. Borsenberger, A. Chowdry, D. C. Hoesterey, and W. Mey, J. Appl.
Phys. 49(11), 5555, (1978).
9. Z. D. Popovic, J. Chem. Phys. 76 (5), 2714 (1982).
10. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors For Imaging Systems, Marcel Dekker, Inc., New York, 1993, p. 154.
11. H. Bässler, Phys. stat. sol. (b) 175, 15 (1993).
12. R. Young, J. Chem. Phys. 103(15), 6749 (1995) .
13. A. Ioannidis and J. P. Dodelet, J. Phys. Chem. B 101, 891, (1997).
14. T. Suzuki and Y. Takahashi, Proc. IS&T’s 13th Int’l. Conf. on Digital
Printing Technol. IS&T, Springfield, VA, 1997, p. 279
15. J. A. Amick, R.C.A. Rev. 20, 770 (1959).
16. T. J. Fabish and C. B. Duke, J. Appl. Phys. 48, 4256 (1977).
Vol. 43, No. 3, May/June 1999
253
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Extrinsic Photocarrier Generation Mechanism in a Dual-Layer Organic
System
M. Umeda
Research and Development Center, Ricoh Co., Ltd., Yokohama, Japan
Extrinsic photocarrier generation in a dual-layered organic device consisting of fluorenone-bisazo-pigment-based carrier genera
tion layer (CGL) in combination with triphenylamine-derivative-based carrier transport layer (CTL) was investigated in order to
elucidate the carrier generation pathways accurately and to improve the carrier generation efficiency. As a result, photocarrier
generation was proven to occur via the following reaction pathways; (i) exciton produced by photon absorption in the bulk of the
CGL diffuses to the CGL/CTL interface, (ii) the exciton makes photoinduced electron transfer to generate a geminate pair at the
interface which competes with the deactivation of the exciton, and (iii) the geminate pair dissociates into free carriers that
compete with geminate recombination. The overall photocarrier generation efficiency was expressed by using these reaction
rates. The efficiency was then simplified to an expression that involves four elementary processes of (i) photoinduced electron
transfer, (ii) deactivation of the azo excited state, (iii) dissociation of geminate pair and (iv) geminate recombination. The reactions, except for the azo excited state deactivation, were considered to occur between the neighboring two molecules and to obey
the electron transfer theory. The method to improve the efficiency is discussed based on electron transfer theory.
Journal of Imaging Science and Technology 43: 254–260 (1999)
Introduction
Organic photoreceptors have been extensively utilized for
electrophotographic processes in photocopiers and laser beam printers.1 The reason behind this wide acceptance
is based on their high sensitivity that is equal to or higher
than that of inorganic photoreceptors.2,3 Furthermore, the
spectral sensitivity of organic photoreceptors is preferably
controlled, in contrast to that of selenium, zinc oxide, cadmium sulfide and amorphous silicon, in the visible or near
infrared region. This is the advantage of organic materials.
However, scientifically speaking, the mechanism thought
to be responsible for the high efficiency of carrier
photogeneration has not been determined with certainty.
In organic photoreceptors, a light-to-electrical energy
conversion efficiency larger than 0.5 is remarkable 4
when the values are compared to the efficiency of other
organic devices. 5 Generally, carrier photogeneration efficiency of inorganic materials exceeds that of organic
materials, because free carriers are directly generated
by photon absorption in inorganic materials (crystals),6
whereas photocarriers are generated via multi-step reactions in organic materials,7 where an exciton produced
by photon absorption generates a geminate hole-electron pair and then dissociates into free carriers. 8 The
tightly bound states of hole and electron, exciton and
geminate pair, are difficult to separate against the coulomb energy. In addition, a multi-step reaction usually
reduces the overall efficiency . For these reasons, adequate comprehension of the carrier photogeneration
Original manuscript received October 7, 1998
© 1999, IS&T—The Society for Imaging Science and Technology
254
mechanism with high efficiency photoreceptors is needed
to understand organic and optoelectronic devices.
There seems to be two efficient methods to generate
free carriers in organic materials. One possibility is the
autoionization of a higher excited state which directly
dissociates before it relaxes to the exciton (the lowest excited state). The other possibility is to use a catalyst that
diminishes the activation energy of the exciton dissociation reaction. Previously, we investigated the photocarrier
generation mechanism of highly-sensitive dual-layered
photoreceptors which consist of an azo-pig ment-based
carrier generation layer (CGL) in combination with a carrier transport layer (CTL). A series of studies led us to
conclude that (i) photocarriers are generated at the interface between the CGL and the CTL, 9,10 and that (ii)
the photocarrier generation at the interface is via photoinduced electron transfer (ET)11,12 that can be described
by the Marcus theory.13 These results implied that the
carrier transport material (CTM) used in the CTL catalytically interacts to the photoexcited azo pigment to
produce photocarriers with high efficiency.
In such extrinsic photocarrier generation mechanism,
one of the major steps is the photoinduced ET between
the azo pigment and the CTM. However, the overall reaction pathways and elementary processes are not sufficiently comprehended. Thus, to improve the overall
efficiency, a kinetic investigation based on the elementary processes involved is necessary. In this article, we
focused our attention on a highly sensitive layered photoreceptor as shown in Fig. 1. First, the extrinsic
photocarrier generation is overviewed and divided into
elementary processes. Next, the overall efficiency is
expressed by employing rate constants of each elementary process. Finally, methods to improve the kinetic
efficiency are discussed.
Figure 1. A dual-layered configuration and chemical structures
of the fluorenone bisazo pigment (lower) and CTM (upper) with
ionization potentials12 used in this study.
Overview of Extrinsic Photocarrier Generation9,12
The plots in Fig. 2 show the relationships between the
quantum efficiency and the electric field of the layered
photoreceptors (see Fig. 1) and the single-layer CGLs
(without CTL),14 and the relationship between the CTLinduced azo photoluminescence (PL) quenching efficiency
and the electric field. 12 The fact that the quantum efficiency of the layered photoreceptor is much higher than
that of the single-layer CGL indicates that the carrier generation is sensitized by the CTL. The sensitized carrier
generation obviously takes place at the CGL/CTL interface where azo pigment exists with the CTM.9,10
A relationship between the CGL thickness and its
maximum absorbance at 583 nm was measured for
single-layer CGLs. A linear relationship that obeys
Lambert’s law was observed.14 This relation means that
the photon absorption occurs in the bulk of the CGL in
the layered photoreceptor. According to these results,
the photon absorption and the carrier generation occur
at different sites. Generally, photoexcited states migrate
in the solid state (exciton diffusion). 5,15 It is therefore
suggested that the photoexcited state produced in the
bulk of the CGL diffuses to the CGL/CTL interface and
dissociates into free carriers.
As shown in Fig. 2, CTL-induced azo PL quenching is
observed. This phenomenon is a result of the photoinduced ET from the IP level of the CTM in the ground
12
state to the IP level of the photoexcited azo compound.
The product of the ET should be a geminate hole-electron pair. In Fig. 2, the magnitude of the azo PLquenching efficiency is independent of the external electric
field, and the xerographic quantum efficiency (i.e., the
overall photocarrier generation efficiency) of the layered
Figure 2. The electric field dependence of the quantum efficiency
of layered photoreceptor and single-layer CGLs using fluorenone
bisazo pigment,14 and CTL-induced azo PLquenching efficiency.12
For the electric field dependence of the quantum efficiencies of
layered photoreceptors, CGL thicknesses were 0.09 µm (◆ ), 0.20
µm (❍), 0.30 µm (∆) and 0.49 µm (■ ), and CTL thicknesses were
about 22 µm. Illumination intensity at 580 nm was 3.8 × 1015
photons s–1 m–2. For single layer CGL, thicknesses were 0.27 µm
(▼) and 0.15 µm (■), and intensity at 580 nm was 8.3 × 1015 photons s–1 m–2. The solid circle (●) data were obtained with the 0.27
µm single layer CGL sample using a positive surface charge. For
the electric field dependence of the azo PL quenching efficiency
(◆) at a layered CGL/CTL structure, CGL thickness was 0.15µm
and CTL thickness was 1.9 µm. Excitation light power was 1.0 W
m-2 at 514.5 nm.
photoreceptor approaches the PL quenching efficiency
at high electric fields. Based on these observations, the
geminate hole-electron pair generated via the photoinduced ET dissociates into free carriers due to an external electric field.16 The dissociation efficiency equals ~1
at electric fields of larger than 3 × 107 V m–1.
According to these considerations, photocarrier generation pathways for the extrinsic mechanism17 are schematically shown in Fig. 3. The reacting pathways are
completely different from those in phthalocyanines represented by the intrinsic carrier generation mechanism.18,19
Experimental
Samples. Chemical structures of the azo pigment 20 and
CTM21 used in this study are shown in Fig. 1. A tetrahydrofuran (THF) dispersion containing a poly(vinyl butyral)
and the azo pigment in a weight ratio of 4:10 was prepared using a ball mill. The dispersion was applied to the
surface of an aluminized polyester film using a blade and
then dried for three minutes at 120°C under atmosphere
to form the CGL with a 0.16–0.3 µm thickness. A THF
solution containing a polycarbonate and the CTM in a
weight ratio of 10:9 was applied to the surface of the CGL
using a blade and then dried for 20 min at 120 °C under
atmosphere to form a CTL about 20 µm thick. The solid
content of the solution was kept at 20 wt.%. Thus, the
desired layered photoreceptors were prepared (see Fig. 1).
A single-layer CGL was prepared on a surface of
nonfluorescent quartz glass for luminescence lifetime
measurement. The above-prepared dispersion was applied to one side of the quartz glass using a dipping process to form a CGL of 0.15 µm thick.
Extrinsic Photocarrier Generation Mechanism in a Dual-Layer Organic System
Vol. 43, No. 3, May/June 1999
255
Wavelength / nm
Figure 3. A schematic illustration of the photocarrier generation at the interface in the layered photoreceptor. A and T respectively denote azo compound and CTM molecule, and A*
means a photoexcited azo compound.
Measurements. The quantum efficiency of photocarrier
generation in the layered photoreceptors was measured
by a xerographic technique. 22 This method enabled us to
measure the surface potential on the photoreceptor activated by a negative corona charge under illumination. The
application of corona charge and subsequent exposure of
the photoreceptor to light were conducted in an electrostatic paper analyzer (Kawaguchi Electric Works; Tokyo,
Japan, SP-428). Monochromatic light was applied to the
CGL of the layered photoreceptor through an optically
transparent CTL. The photoinduced discharge curve leads
to a quantum efficiency, φ, which represents the number
of surface charges removed by each absorbed photon.23
φ(F) =
C dVs
eI dt
(1)
F
Here, C is the capacitance of the photoreceptor per unit
area, e is the electronic charge, I is the incident light intensity in photons per second and per unit area, Vs is the
surface potential, and F is the electric field.
Photoluminescence lifetime of the azo pigment was
measured by a picosecond fluorescence lifetime measurement system (Hamamatsu Photonics; Hamamatsu, Japan, C4780), that includes a streak camera (Hamamatsu
Photonics; Hamamatsu, Japan, C4334) and a laser diode (635 nm-4 pJ/pulse) for sample excitation. The laser beam irradiated the sample from the CGL side and
the luminescence of the azo compound from the same
side was measured using a glass fiber in combination
with a 660 nm short-wavelength cut-off filter to eliminate the excitation light.
The absorption spectrum of the CGL having a CTL
which was dissociated at the electrode, and the spectrum of a single independent CTL, serving as a refer ence sample, were obtained using a spectrophotometer
(Shimadzu, Kyoto, Japan; UV-3100). Thicknesses of the
photoreceptors were measured by a surface profile measuring system (Sloan, Santa Barbara, CA; Dektak IIA).
All measurements were carried out at 25±2°C.
256
Journal of Imaging Science and Technology
Figure 4. Dependence of the quantum efficiency on illumination wavelength at layered photoreceptor and absorption spectrum of the CGL. The CTL thickness was 20.3µm and the CGL
thickness was 0.3 µm. Illumination intensity was 3.1 × 10 15
photons s –1 m–2 and electric field was 4 × 107 V m–1. Absorption
shorter than 440 nm is attributed to the CTL.
Results and Discussion
Adequate Comprehension of Photocarrier Generation Process. In order to understand the reaction pathways accurately, we now describe the remaining
processes which are still uncertain from the above section. Namely, the photoexcited state of the azo compound
that participates in the photoinduced ET, exciton diffusion in the CGL and bimolecular recombination of free
carriers are characterized.
Figure 4 shows quantum efficiencies of the layered
photoreceptor with varying illumination wavelengths
and absorption spectrum of the CGL. The quantum efficiency remains a constant value with respect to the
excitation wavelength. This indicates that photocarrier
generation occurs from a fixed energy state, which appears to be due to relaxation from higher excited states
at the azo pigment. 10,24 It is therefore evident that the
photocarriers are not generated via autoionization.
Figure 5 shows fluorescence intensity decay of the azo
compound in the single-layer CGL. The deconvoluted
lifetime is 409 ps, which gave a χ 2 value of 1.14. This
short lifetime indicates that the luminescence of the azo
compound is not generated from the triplet excited state,
but from the singlet excited state. The singlet excited
state of the azo compound was identified as a Frenkel
exciton according to an electroabsorption study.25
If the exciton produced in the bulk of the CGL dissociates at the moment when the exciton reaches to the
CGL/CTL interface, the quantum efficiency for
photocarrier generation will be expressed by Eq. 2 in
the case where d » α -1 and d » L.15,26
1
φ −1 = φ o−1 
+ 1
α ⋅L 
(2)
Here, d is the thickness of the CGL, α is the absorption
coefficient, and L is the exciton diffusion length.
Umeda
Figure 5. Decay profile of the PL from the azo compound in the
CGL. The data points are the observed decay and the solid curve
is the fit of one component, 409 ps. The excited light pulse response is also presented by the dashed-dotted line. The lower
portion of the figure shows the plot of the weighted residuals
resulting from the fitting.
Figure 6 shows the reciprocal quantum efficiency of
the layered photoreceptor versus the reciprocal absorption coefficient of the CGL. Because the quantum efficiencies were obtained at the high electric fields, thequantum
efficiency is taken as the ET efficiency. According to Fig.
6, the experimental data does not fit Eq. 2.
In the case where α -1 » L » d, the quantum efficiency
will be independent of α .26 Thus, the result in Fig. 6 represents L » d. Moreover, the magnitude of the quantum
efficiency of the layered photoreceptors where the CGL
thickness varies between 0.05-0.61 µm, is independent
of the CGL thickness (this result is in part also shown
in Fig. 2). 14 It is therefore concluded that the exciton
diffusion is much faster than the subsequent processes
and is not a rate-determining step throughout the
photocarrier generation process. In fact, the CTM penetrates into the CGL during the wet-overcoating operation of the CTL; this results in an azo pigment particle
to be surrounded by the CTMs. 14,27 Based on the CGL
structure, the exciton in the azo particle can efficiently
reach the azo/CTM interface with a substantially short
diffusion length.
Next, the relationship between the neutralization rate
of the surface charge of the photoreceptor and an incident light intensity was measured in order to under stand the bimolecular recombination process of the free
carriers. In the case where the recombination rate of
free carriers is negligibly small, the neutralization rate
of the surface charge is proportional to the incident light
intensity and can represent the exciton creation rate.
Conversely, in the case where bimolecular recombination of free carriers predominantly occurs, the neutralizing rate will be proportional to the square root of the
incident light intensity. The number of holes, N, that
neutralize surface charges per second and per unit area
can be expressed by
N=
C dVs
e dt
.
F
(3)
Figure 6. Reciprocal photocarrier generation efficiency versus
reciprocal absorption coefficient of the CGL at the layered photoreceptor. Plots were taken from the data in Fig. 4.
Figure 7. The neutralizing rate for surface charges N versus illuminating rate of incident photons I at the layered photoreceptor.
The CTL thickness was 20.3 µm and the CGL thickness was 0.16
µm. The electric fields were at 2 × 107 (● ) and 4 × 106 (■ ) V m-1.
The relationship between N and incident light intensity,
I, is shown in Fig. 7. The incline of the graph is 1. It is
therefore deduced that the bimolecular recombination rate
is negligibly small. 10 Accordingly, free carriers, once generated at the interface, predominantly reach to the top of
the photoreceptor to cancel out the surface charges and a
substantial bimolecular recombination of free carriers does
not occur.
Kinetics of Photocarrier Generation. The above observed results are summarized as follows: Higher
photoexcited state of the azo pigment relaxes to the lowest excited state of exciton. The exciton diffuses to the azo/
CTM interface to make ET which competes with the deactivation to the ground state. The ET at the interface gen-
Extrinsic Photocarrier Generation Mechanism in a Dual-Layer Organic System
Vol. 43, No. 3, May/June 1999
257
Similarly, the overall recombination rate constant, kr, is
given by
kr =
Figure 8. A schematic of extrinsic carrier generation pathways
in the layered photoreceptor containing azo pigment.
(
kdiff ket kgr
k-diff k-et + kdiss + kgr
)
.
(6)
The quantum efficiency of the overall photocarrier
generation is defined by32
φ=
ks
.
ks + k r + k L
(7)
Substituting Eqs. 5 and 6 into Eq. 7, one obtains
φ=
Kdiff ket ⋅
Kdiff ket
kdiss + kgr
k-et + kdiss + kgr
⋅
+ kL
kdiss
.
k-et + kdiss + kgr (8)
Here, Kdiff = kdiff/k-diff.
Improvement of Photocarrier Generation Efficiency.
The expression of photocarrier generation efficiency, Eq.
8, can be simplified and understood by using the Marcus
theory.11 The Marcus expression of the electron transfer is
given by13
 (λ − ∆E) 2 
,
ket = k0 exp −
4λkB T 


Figure 9. Energy gap dependence of ket/k-et. The arrow indicates
the energy gap of the present system.
erates a geminate pair. The geminate pair dissociates into
free carriers or undergoes geminate recombination. The
free carriers neutralize the surface charges of the photoreceptor and never undergo bimolecular recombination.
In general, the relaxation from higher excited state
to the lowest excited state (i.e., S1) occurs within a picosecond and the reaction is irreversible. 28 Thus, the relaxation process will be neglected. In addition,
photoinduced ET generally accompanies back electron
transfer.29,30 As a result of these processes, the extrinsic
photocarrier generation is schematically represented in
Fig. 8.
Because the photoinduced discharge of the surface potential is emission limited 31 under the experimental conditions,9,10 a steady state is postulated to the intermediates
of (A…T)* and (A–…T+). Therefore, the overall photocarrier
generation rate constant, ks, is expressed as
ks =
kdiff ket kdiss
. (4)
k-diff k-et + k-diff kdiss + k-diff kgr + ket kdiss + ket kgr
Because the exciton diffusion rate is fast enough in
the overall process, k diff » k et and kdiff » (k diss+kgr) are presumed. Thus, Eq. 4 is simplified to
ks =
258
(
kdiff ket kdiss
k-diff k-et + kdiss + kgr
)
.
Journal of Imaging Science and Technology
(5)
(9)
where k 0 is the preexponential factor , λ is the total
reorganization energy, and ∆E is the energy gap that is
defined as the IP difference between the azo compound
and CTM.11
∆E = IP(azo) - IP(CTM)
(10)
For a reversible ET system, the preexponential factor, k0,
is the same for the two-rate constants.33,34 Thus, a ratio of
the forward ET rate constant, ket (energy gap is ∆E), to
the back ET rate constant, k-et (energy gap is - ∆E), is expressed by35
 ∆E 
ket
= exp
.
k− et
 kB T 
(11)
Figure 9 shows the relationship between the energy
gap and the ratio of the rate constants based on Eq. 11.
In the case where ∆E is 0.10 eV, k et/k-et is calculated to
be 50. This value of the ratio means that back ET rate
is negligibly small when the energy gap is larger than
that value. It is, therefore, concluded that the back ET,
which decreases the number of geminate pairs, will be
ignored when we appropriately control the energy difference between the HOMOs of the CGM and the CTM.
Because the energy gap of the present system is 0.47
eV (see Fig. 1), the magnitude of ket/k -et is calculated to
be 8.9 × 107. This obviously reveals that the ET to generate geminate pair takes place predominantly , while
substantial back ET does not occur.
In the case where the energy gap is set to exothermic
(∆E > 0) in order to enhance the geminate-pair genera-
Umeda
2). The expression of the dissociation rate constant as a
function of the electric field is not presently proposed. Electric-field-dependent rate constant will be studied based
on electron-transfer reaction.38
Consequently, the photocarrier generation efficiency of
the extrinsic system was expressed by kinetic rate constants based on the reaction pathways represented in Fig.
8. The expression of the efficiency was simplified; there
remained four rate constants in the reduced formula. In
the formula, attention was focused on the three reactions
that originate in ET and the method to enhance the efficiency was discussed based on the ET theory.
Figure 10. Schematic energy diagram of the simplified extrinsic
photocarrier generation expressed by four elementary processes;
(a) geminate hole-electron pair generation process: the photoinduced ET (charge separation) competes with the deactivation of
the excited state; (b) free carrier generation process: the geminate-pair dissociation (charge shift) competes with the geminate
recombination (charge recombination).
tion efficiency, based on Eq. 1 1, results in k et > k-et. At
high electric fields, the geminate-pair dissociation efficiency (~1.0) exceeds the photoinduced ET efficiency
(0.56) as shown in Fig. 2. If we presume k et « (k diss+k gr),
then k-et « (kdiss+k gr) will stand.
Furthermore, the exciton diffusion process is recognized as equilibrium (K diff = 1), because the exciton diffusion is much faster than the subsequent processes.
Under these conditions, Eq. 8 will be reduced as
φ=
kdiss
ket
⋅
.
ket + kL kdiss + kgr
(12)
The overall efficiency is thus expressed by four rate constants. The first term of Eq. 12 represents geminate-pair
formation efficiency, and the second term is geminate-pair
dissociation efficiency to cause free carriers. The photoinduced ET that obeys the Marcus theory 13 occurs between
the neighboring azo and CTM molecules. 36 Similarly, the
following reactions of geminate recombination and geminate pair dissociation must take place between two neighboring molecules. Figure 10 demonstrates an energy
scheme involving the four processes. According to this
scheme, three reactions are considered as ET;29,30 the photoinduced ET at the interface is the charge separation
(A*…T → A–…T+), the geminate recombination is equivalent
to the charge recombination (A –…T+ → A + T), and geminate pair dissociation corresponds to the charge shift (T +
+ T → T + T+).37 Thus, these three reactions obey Eq. 9.
To improve the former geminate-pair production efficiency, k et enhancement based on charge separation using Eqs. 9 and 10 will be practical. This is because the
ET efficiency is not influenced by the electric field (see
Fig. 2) and the equations are independent of electric
field. In addition, to employ a carrier generation material bearing a long lifetime of photoexcited state is also
effective. To enhance the latter efficiency , k gr which is
regarded as a rate constant of charge recombination
could be discussed based on Eq. 9 with an energy gap of
∆E = IP(CTM) - EA(azo).
(13)
The remaining kdiss, is considered to be the charge shift
with an energy gap of ∆E = 0. However, the dissociation
efficiency strongly depends on the electric field (see Fig.
Conclusions
This article described the extrinsic photocarrier generation in the dual-layered photoreceptor including the
bisazo pigment in the CGL. First, the photocarrier generation pathways were briefly overviewed and then precisely investigated. Next, the overall efficiency of the
carrier photogeneration was expressed by using the rate
constants of the elementary processes represented. Finally, the methods to enhance the efficiency were discussed based on the simplified expression. The results
are summarized as follows.
1. The extrinsic photocarrier generation was proven to
occur via following reaction pathways; the exciton
produced by photon absorption in the bulk of the CGL
diffuses to the CGL/CTL interface, the exciton makes
photoinduced ET to generate geminate pair at the
interface which competes with the deactivation of the
exciton, and the geminate pair dissociates into free
carriers which competes with the geminate recombination. The bimolecular recombination rate of free
carriers was known to be negligibly small.
2. The overall photocarrier generation efficiency was
expressed by using the above-mentioned reaction
rates; and the overall efficiency was simplified to the
product of ET efficiency and the geminate-pair
dissociation efficiency. The reduced expression
involves four elementary processes of the photoinduced ET, deactivation of the azo excited state, the
dissociation of geminate pair and the geminate
recombination. The reactions except for the deactivation
were considered to occur between the two neighboring
molecules and to obey the electron transfer theory .
The method to improve the efficiency was discussed
based on the energy gap for electron transfer.
However, the electric-field-dependent expression of geminate-pair-dissociation rate is still unknown. In the future,
we plan to study the electric-field-dependent dissociation
from the viewpoint of the electron-transfer reaction.
Acknowledgment. The author acknowledges T. Niimi for
his technical support of the PL lifetime measurement.
Appendix: Derivation of Eq. 4
Because the photoinduced discharges were measured under the emission-limited condition, the concentration of
the intermediate in Fig. 8, (A *…T), is presumed to be a
steady state. Thus, a concentration change in terms of rate
is given as
[
]=
d ( A * L T)
dt
[
]
[(
kdiff [ A *][T] − ( k− diff + ket ) ( A * L T ) + k− et A – L T +
Extrinsic Photocarrier Generation Mechanism in a Dual-Layer Organic System
)]
(A1)
= 0.
Vol. 43, No. 3, May/June 1999
259
For the other intermediate, (A-…T+),
[(
–
d A LT
+
dt
6.
)] =
7.
] (
[
5.
ket ( A * L T ) − k − et + kdiss + kgr
)[(A – LT+ )] = 0.
(A2)
9.
The overall photocarrier generation rate is expressed by
ks[A*] [T] = kdiss [(A–LT+)].
(A3)
From Eqs. A1 and A3, the product, [A*][T], could be eliminated as
[(
 kdiff + kdiss

+ k− et  A – L T +

ks


)] = (k
− diff
[
8.
]
+ ket ) ( A * L T) . (A4)
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
From Eqs. A2 and A4, one obtains
22.
ket ( kdiff kdiss + k− et ks )
ks ( k-diff + ket )
[(A
(k
–
− et
L T+
)] =
+ kdiss + kgr
23.
) [( A
–
LT
+
)].
(A5)
Thus, the concentration term, [(A-…T+)], will be eliminated
from Eq. A5, then ks expression is given as
ks =
k-diff k-et
kdiff ket kdiss
. (4)
+ k-diff kdiss + k-diff kgr + ket kdiss + ket kgr
24.
25.
26.
27.
28.
29.
30.
31.
References and Notes
T. Kawamoto, Electrophotography 27, 320 (1988); 28, 171 (1989); 29,
156 (1990); 30, 222 (1991); 31, 134 (1992); 32, 124 (1993); 33, 149
(1994); 34, 96 (1995).
2. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for Xerography, Marcel Dekker, New York, 1998, Chap. 6.
3. K.-Y. Law, in Handbook of Organic Conductive Molecules and Polymers,
Vol. 1, H. S. Nalwa, Ed., John Wiley & Sons, Chichester, 1997, Chap. 10.
4. M. Umeda, Trends in Physical Chemistry, Vol. 7, Reseach Trends,
Trivandrum, 1997, p. 107.
1.
260
Journal of Imaging Science and Technology
32.
33.
34.
35.
36.
37.
38.
J. Simon and J.-J. André, Molecular Semiconductors , Springer-Verlag,
Berlin, 1985.
R. H. Bube, Electronic Properties of Crystalline Solids, Academic Press,
New York, 1974.
A. V. Patsis and D. A. Seanor, Photoconductivity in Polymers, Technomic,
Westport, 1976.
P. M. Borsenberger and D. S. Weiss, in Handbook of Imaging Materials, A. S. Diamond, Ed., Marcel Dekker, New York, 1991, Chap. 9.
M. Umeda, T. Niimi, and M. Hashimoto, Jpn. J. Appl. Phys. 29, 2746
(1990).
M. Umeda and M. Hashimoto, J. Appl. Phys. 72, 117 (1992).
M. Umeda, T. Shimada, T. Aruga, T. Niimi, and M. Sasaki, J. Phys. Chem.
97, 8531 (1993).
T. Niimi and M. Umeda, J. Appl. Phys. 74, 465 (1993).
R. A. Marcus, J. Chem. Phys. 24, 966 (1956); Ann. Rev. Phys. Chem.
15, 155 (1964).
M. Umeda and T. Niimi, J. Imag. Sci. Technol. 38, 281 (1994).
B. J. Mulder, Philips Res. Rep. Suppl. 4, 44 (1968).
M. Umeda, Proc. 10th International Congress on Advances in NIP
Technologies, IS&T, Springfield, VA, 1994, p. 239.
M. Umeda, Nippon Kagaku Kaishi, 1996, 932 (1996).
Z. D. Popovic, J. Chem. Phys. 78, 1552 (1983).
A.-M. Hor and Z. D. Popovic, Proc. 7th International Congress on Advances in NIP Technologies, Vol. 1, IS&T, Springfield, VA, 1991, p. 293.
M. Hashimoto, Electrophotography 25, 230 (1986).
M. Sasaki, U. S. Patent 4 892 949, 1990; Chem. Abstr. 100, 112236w
(1984).
A. R. Melnyk and D. M. Pai, in Physical Methods of Chemistry, 2nd ed.,
Vol. 8, B. W. Rossiter and R. C. Baetzold, Eds., John Wiley & Sons,
New York, 1993, Chap. 5.
J. Mort and G. Pfister, in Electronic Properties of Polymers, J. Mort and
G. Pfister, Eds., John Wiley & Sons, New York, 1982, p. 215.
P. M. Borsenberger and D. C. Hoesterey, J. Appl. Phys. 51, 4248 (1980).
M. Umeda and M. Yokoyama, Jpn. J. Appl. Phys. 34, L44 (1995).
J. H. Perlstein and P. M. Borsenberger, in Extended Linear Chain Compounds, 2nd ed., J. S. Miller, Ed., Plenum Press, New York, 1982, p.
339.
T. Niimi and M. Umeda, J. Appl. Phys. 76, 1269 (1994).
N. J. Turro, Modern Molecular Photochemistry, Benjamin/Cummings
Publishing, Menlo Park, California, 1978, Chap. 6.
Photoinduced Electron Transfer , M. A. Fox and M. Chanon, Eds.,
Elsevier, Amsterdam, 1988, Parts A-D.
G. J. Kavarnos, Fundamentals of Photoinduced Electron Transfer, VCH
Publishers, New York, 1993.
J. Mort and I. Chen, in Applied Solid State Science, Vol. 5, R. Wolf, Ed.,
Academic Press, New York, 1975, p. 83.
M. D. Tabak and P. J. Warter, Phys. Rev. 173, 899 (1968).
D. Rehm and A. Weller, Isr. J. Chem. 8, 259 (1970).
T. Kakitani and N. Mataga, J. Phys. Chem. 89, 4752 (1985).
T. Kakitani and N. Mataga, J. Phys. Chem. 91, 6277 (1987).
M. Umeda and T. Niimi, Jpn. J. Appl. Phys. 33, L1789 (1994).
The charge shift between the azo compounds (A– + A → A + A–) occurs
to a considerable extent.
M. Umeda, M. Ohta, and M. Yokoyama, Extended Abstracts (The 44th
Spring Meeting) Vol. 3, The Japan Society of Applied Physics and Related Societies, Tokyo, 1997, p. 1182.
Umeda
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Photocarrier Generation in Polysilane Films Doped With and Without
Fullerene
Y. Nakayama,▲ A. Saito, T. Fujii, and S. Akita
Department of Physics and Electronics, Osaka Prefecture University, Osaka, Japan
The photocarrier generation kinetics in poly(methylphenylsilane) films with and without C 60 has been studied by measuring
accurate subgap absorption spectra, absorption spectra contributing to photocurrent, and the normalized photoconductivity
. The
photoconduction spectrum of polysilane has 0.1 eV higher onset than the absorption. The photocarriers are not generated at the
electrode, but in the bulk. These results suggest that the photocarriers are more likely photogenerated free-holes in a disorde
red
system than photogenerated charged-polarons. The analysis based on a disorder model successfully explains the zero-temperature normalized photoconductivity. Doping of C 60 sensitizes efficiently the photoconduction of polysilane in the low photon-energy region where C 60 has optical absorption. This sensitization is suppressed at low temperatures. In the low temperature
region, photogenerated holes are trapped in the tail states of polysilanes in a C
60 doped sample and consequently, the Fermi level
moves to increase the dark and photoconductivity. The trapped holes recombine with electrons from C60 at temperatures higher
than 130 K.
Journal of Imaging Science and Technology 43: 261–265 (1999)
Introduction
Polysilanes are organic materials with mobility as high
as 10-4 cm2/Vs even in unoriented films.1 Oriented films
show one order of magnitude higher mobility along the
orientation and one order of magnitude less mobility perpendicular to the orientation as compared with unoriented
films.2 A possible mobility of 0.1 – 1 cm 2/Vs has been
pointed out for oriented films. 3,4 This anisotropy is
caused by the difference in the hopping distance between
the intrachain hopping and interchain hopping.2
The polysilane of poly(methylphenylsilane) (PMPS)
shows photoconductivity and doping of C 60 in PMPS
sensitizes the photoconductivity. 5,6 It is generally accepted that C 60 incorporated in the polysilane accepts
a photogenerated electron to leave a hole in the polymer chain. 5,7 A possibility of energy transfer followed
by the electron transfer from the polysilane to C60, leaving a hole in the polymer chain has also been proposed
.6
It has been reported that PMPS shows a photoconduction spectrum with onset coincident with the optical absorption edge.8,9 This measurement was done for
a sandwich structure sample with a transparent electrode through which the PMPS film was exposed to
light. The result has been explained by a model in
which an exciton is photogenerated in the film, migrates toward the transparent electrode and dissociates, by an acceptor -like effect of the electrode , to
provide a free hole.8,9 However, this explanation is not
conclusive.
Original manuscript received November 2, 1998
▲ IS&T Member
© 1999, IS&T—The Society for Imaging Science and Technology
It is well known that in molecular crystals, an exciton is produced by photon energies more than the exciton binding energy by 0.5 – 1.0 eV , whereupon it
undergoes dissociation to produce a free hole or a free
electron.10,11 However, the situation is different in amorphous materials. The photoconduction begins almost at
the optical absorption edge, which is the same in the
case of PMPS. There are two models to explain this behavior. One is the disorder model, 12 that energetic disorder of hopping sites lowers the exciton binding energy
.
The other is the charged-polaron excitation model,10 that
photoexcitation yields charged polarons contributing to
the photocurrent.
In this article, we explore the kinetics of photo generation of carriers in PMPS films by measuring accurate subgap optical absorption spectra, subgap absorption spectra contributing to photocurrent, and the
normalized photoconductivity. We also investigate PMPS
films doped with fullerene C60.
Experiment
Polysilane of the PMPS-type used had molecular weight
of ca. 130,000. Samples of self-supporting thick films
and coated thin films were prepared. The self-supporting films with thickness of 40µm were prepared by peeling off films that were cast from a toluene solution of
PMPS on mica substrates. The oriented film was also
prepared by mechanically stretching the self-supporting film. The thin films were deposited onto indium tin
oxide (ITO) coated glass plates by spin casting. The resulting thickness was 250 nm. The fullerene doping in
the polysilane films was carried out by dissolving PMPS
and C 60 in toluene in weight ratio of 25:1. The C 60 powder used was prepared by dc arc discharge using graphite electrodes and had a purity of at least 90%.
261
Results
Figure 1 shows the spectra of α measured by PDS, ηα
measured by CPM, their ratio η and the normalized
photoconductivity σp/eG=ηµτ at room temperature for
PMPS films with and without C60. Here G is the excitation rate per cm3, and µ and τ are the mobility and lifetime of carriers, respectively . The absolute value of η
was not determined and is assumed for simplicity, to be
unity at a low photon energy.
For the PMPS film, the PDS measurement reveals a
sharp band tail with exponential rather than Gaussian
distribution. The Urbach energy is estimated to be 40
meV. It also indicates quite low values of α in a low hν
region, i.e., a low density of states (DOS) in the midgap.
The CPM absorption starts to rise at the point B of 3.35
eV which is 0.1 eV higher than the onset A of the PDS
curve. As the photon energy increases toward 3.35 eV
the η value decreases, which indicates the change of
photoconduction mechanism at the onset B. This is
clearly reflected by the variation ofσp/eG measured without the acceptor -like effect of electrodes which has a
minimum C coincident with the onset B. It is believed
that the photoconductivity is due to impurities and/or
defects at low photon energies and is dominated by the
band-to-band transition at hν > 3.35 eV. The increase
in σp/eG with increasing photon energy from the point
C is caused by the increase in µ and/or τ of free holes,
because the value of η does not increase. Because the
262
Journal of Imaging Science and Technology
α,ηα (cm–1)
10 6
10 4
10 2
η
10 0
10 0
with C60
10–2
σp/eG (cm2/V)
For the thick samples coplanar A1 electrodes with a
gap of 100 µm were prepared to avoid the acceptor-like
effect of electrodes for photocarrier generation, especially in the case of oriented films that were designed
for measurement of the photocurrent as a function of
the orientation. On the other hand, on the thin samples,
an Al electrode was deposited as a counter electrode to
the ITO transparent electrode to form a sandwich cell.
Photocurrent measurements were performed with a
xenon lamp as a light source, for samples kept in a temperature controlled cryostat. The parameters were the
photon energy hν, temperature T, and electric field. The
constant photocurrent method (CPM)13 was also used to
measure the subgap absorption contributing to the
photoconduction. In this method the photon flux was
measured as a function of hν, keeping photocurrent constant, namely maintaining the quasi-Fermi level for
holes constant. The CPM signal or the inverse of the
measured photon flux is proportional to a product of the
optical absorption coefficient α and the quantum yield
η . Because uniform photocarrier-generation in the bulk
is a key principle, the thin samples were used for the
CPM measurement. The thin samples were also used
for the measurement of the electric field dependence,
while the thick samples were used for other photocur rent measurements. The applied electric field was 10 4
V/cm, except for the measurement of electric field dependence. The data were taken in a pointwise manner
to exclude any distortion due to residual internal space
charge accumulation.
Photothermal deflection spectroscopy (PDS)14 was performed to measure the spectrum of α in the subgap region. The samples used for this measurement were 500
nm thick films coated on quartz. The spectrum was nor
malized to the data in the band gap region measured by
a conventional transmission method. The deflection medium used was 1,4 butanediol which does not dissolve
PMPS and has a large change of refractive index with
temperature (dn/dT=7 × 10 –4 deg–1).
10–10
10–12
1.5
with C60
2.5
3.5
Photon Energy (eV)
Figure 1. Spectra of the optical absorption coefficient α measured by PDS, the optical absorption contributing the photocurrent ηα measured by CPM, their ratioη and the normalized
photoconductivity σp/eG for PMPS films doped with and without C 60. A and B denote the onset for PDS and CPM, respectively, and C denotes the minimum of σp/eG. The value of η is
set to be unity at a low photon energy. For CPM samples of a
sandwich structure were used, while for the measurement of
normalized photoconductivity samples with coplanar electrodes
were used. The samples were unoriented.
current density or the carrier transport energy level in
the measurement of σp becomes high at a high photon
energy, increase in µ is most probable.
The doping of C60 modifies the optical absorption spectrum of the polysilane at hν < 3.45 eV. It is found that
the relative value of η is high around hν = 1.9 eV and
shows a steep decrease with increasing hν from 3.4 eV.
The energy 1.9 eV corresponds to the energy gap15 of C60
clusters. The decrease of η is reflected in the decrease
in σ p/eG. The sensitization of photoconductivity by C 60
is effective at photon energies up to 3.65 eV . The
photocarrier generation mechanism changes to be the
same as for undoped PMPS beyond this energy.
Figure 2 shows the temperature dependence of σp/eG
measured for PMPS films. Photon energies of 3.4 ~ 3.55
eV were used in this light measurement. Photoconductivity is observed even at a low temperature ofT = 15 K
and rises steeply at T > 200 K. However, it decreases at
T > 240 K. This is because of photodegradation of the
Nakayama, et al.
Figure 2. Temperature dependence of the normalized photoconductivity σp/eG for oriented and unoriented PMPS samples
with coplanar electrodes. The photocurrent is measured along
the orientation for the oriented film.
Figure 3. Temperature dependence of the normalized photoconductivity σ p/eG, dark current Id and photocurrent Ip for C60
doped PMPS samples with coplanar electrodes. The samples
were unoriented.
sample. The dotted line is an expected curve. In Fig. 2,
the data for the oriented film of PMPS are also plotted.
This measurement was done using rather weak light to
avoid the photodegradation so that the data are scattered. The value of σp/eG begins to rise at 100 K, which
is half the value observed for the unoriented film.
Figure 3 shows the temperature dependence of dark
current Ιd, photocurrent Ιp and σp/eG measured for the
C60 doped PMPS film on 3.54 eV excitation. The value of
σp/eG at 15 K is almost equal to the value for PMPS
alone, but its temperature dependence is different. At
15 K the Ιd value is ~ 10–13 A, which is comparable to the
noise level, and the Ιp value is ten times higher than the
Ιd value. The values of Ιd and Ιp, both increase with increasing temperature up to 130 K. However, this is not
a real temperature dependence. We have confirmed that
the dark and photocurrent increase to their respective
saturated values with time even when the sample is kept
at a constant temperature of 77 K. In Fig. 3, at 130 K,
the Ιd value begins to decrease and becomes equal to the
noise level at T > 200 K. On the other hand, theΙp value
reaches a value forty times higher than the Ιp value at
T > 200 K. In this temperature region, Ιp is not ther mally activated. A dip observed in Ιd and Ιp at 160 K is
not artifactual but reproducible.
and the lowest unoccupied molecular orbital (LUMO) of
C60 are assumed to form bands because the optical absorption spectra show broad bands.
The comparison between the CPM and the PDS data
for PMPS alone clearly indicates that the onset of the
photocarrier generation due to the band-to-band transition is 0.1 eV higher than that of optical absorption. The
correspondence of this onset with the minimum of the
normalized photoconductivity measured in the cell with
Al coplanar-electrodes confirms that this behavior for the
photocarrier generation is characteristic of PMPS films.
The evidence is inconsistent with the model proposed by
Kepler and Soos8,9 where the photocarrier generation in
the sandwich cell is caused at the electrode, which acts
as an acceptor to extract electrons from excitons and provide free holes.
Possible processes for photocarrier generation with
low or no excess energy are a disorder model12 for polymer materials and a model 10 of charged polaron excitation. In the disorder model for polymer materials, the
energetic disorder of hopping sites in the polymer matrix lowers the dissociation energy of excitons to yield
free carriers. In the charged polaron excitation model,
the lowest optical transition in a one-dimensional polymer is from the ground state to the relaxed excited state
which is the coupling of electrons with distortions in
the polymer backbone by electron-phonon interaction.
This photoexcitation yields charged (positive and negative) polarons that promptly contribute to the photocurrent without any additional energy. It is believed that
the disorder model is more likely for PMPS because of
the 0.1 eV excess energy. Later we will discuss the generation process of free carriers based on another approach16 of the disorder model which has been developed
in the field of inorganic amorphous semiconductors. The
basic concepts in the disorder model for the polymer
materials and the inorganic material are the same.
As shown in Fig. 1, the η value decreases with increasing photon energy even after the photocarrier-gen-
Discussion
With respect to the photoconduction spectra in Fig. 1,
the free carrier generation at energies less than the absorption edge of PMPS for samples with and without
C60 is not caused by exciton formation. Its origin in the
sample without C60 must be other impurities and/or defects. This generation process is denoted by (1) in Fig.
4. For the C60 doped sample, as denoted by (2) and (3) in
Fig. 4, C 60 is photoexcited, and subsequently , an electron transfers from the highest occupied molecular or bital (HOMO) band of polysilane to C 60 to leave a hole
in the polysilane. These processes are essentially the
same as reported by Kepler and Cahill.6 In Fig. 4, HOMO
Photocarrier Generation In Polysilane Films Doped With And Without Fullerene
Vol. 43, No. 3, May/June 1999
263
lifetime17 τ ≈ 1 ns of the photoluminescence using the
relation16 of τ = τ02 ν0.
It is clear from Eqs. 1 and 2 that due to inequality ν0
> τ0–1, the probability of diffusion is not negligible for
the first steps. After each hop, the average concentration of accessible states decreases and the distances R
and r increase. The geminate recombination probability is small when R < Rc, where R c = (a/2)ln( ν0τ0) is the
characteristic length. It reaches a maximum near R =
R c and then decreases with increasing R. The survival
probability of pairs is given by
φ ( R) = A( Rc R) ,
β
Figure 4. Scheme of generation processes of photocarriers in
polysilane with and without C 60.
eration mechanism changes from (1) to (4) in Fig. 4. Possible origins are exciton quenching and charge trapping
and recombination.
In Fig. 1 the normalized photoconductivity exhibits
an increase at photon energies above 3.65 eV for films
doped with and without C60. This energy is higher than
the peak energy of the optical absorption, indicating that
the charges are photoexcited to states higher than the
DOS peak. According to both the disorder models, those
charges easily become free from the electron-hole pairs.
The result that the sensitization of photoconduction due
to C60 disappears at high photon energies may be caused
by the low electric field, because the sensitization has a
strong electric field dependence.6
Because of the lack of thermal energy , the value of
σ p/eG is ~ 10-12 cm-2/V at T = 15 K, and carriers cannot hop
up in energy. This value is close to the zero-temperature photoconductivity estimated for hydrogenated
amorphous Si by the disorder model.16
We will modify the disorder model of inorganic semiconductors in order to apply it to organic materials with
a high rate of geminate recombination. Let us consider
an electron-hole pair generated at or just below the
mobility edge of an amorphous semiconductor at zero temperature. The pair is generated quite close together because of the exponential decay their overlap integralwith
distance. In this analysis the fate of one electron-hole
pair is described assuming that the electron is fixed in
space. The hole can take part in two competing processes:
it can hop down in energy because T = 0 K to the nearest localized state of the tail, at distancer, with the rate
ν d (r) = ν 0 exp( −2r a) ,
(1)
or it can recombine with the hole at a rate
ν r (r) = τ 0 −1 exp( −2 R a) ,
(2)
where R is the electron-hole separation and a is the localization radius of the electron. The prefactor ν 0 is the
phonon frequency (≈10 12 s–1) and τ0 is the dipole radiation lifetime (≈3 × 10 –11 s) which is estimated from the
264
Journal of Imaging Science and Technology
(3)
with A = 3.0, where β = 1.5 is adopted instead of β = 1
used for the inorganic case by taking into account the
high geminate-recombination probability. The nongeminate recombination that contributes to the photoconductivity appears when R becomes about half the average
carrier separation 0.5n0–1/3/2 (> R c) where n0 is the steady
state electron (or hole) concentration under the generation rate G at T = 0 K. Assuming that the band tails
have the DOS correspond to an exponential distribution with the width ε0, we have the expression for the
normalized photoconductivity given by
σp
eG
=
ea 2
ln(ν 0τ 0 )
4ε 0
[
]
1.5
L0.5 ,
(4)
where L = n0–1/3/a and is the solution of the equation
{
[
}
1.5 −1
]
L = ln 3Gτ 0 a3 L ln(ν 0τ 0 )
.
(5)
Using the experimental value G = 1019 cm–3s –1, the experimentally determined value ε 0 = 0.04 eV (from Fig.
1) and the reasonable material parametersν0τ0 = 30 and
a = 1 nm for an electron (the value for an electron might
be larger than that for a hole, and the larger value would
be effective for the recombination), we obtain σp/eG = 2
× 10 –12 cm2/V for the PMPS from Eq. 4. The estimated
value of σp/eG is in agreement with the experimental
result in Fig. 2. Furthermore, this theory predicts that
the temperature Tr where the photoconductivity begins
to rise (the transport energy crosses the zero-temperature Fermi level) is 3ε0/2kL, where k is the Boltzmann’s
constant. The parameters for PMPS give T r = 30 K. It
can be seen in Fig. 2 that the photoconductivity begins
to rise around 30 K, however the sharp rise occurs at
100 K and 200 K. The rise at rather high temperatures
has been observed for the case of hydrogenated amor phous SiN. 18 This could be related to carrier transport
properties such as mobility. The oriented polysilane film
that should have higher mobility 2 than the unoriented
polysilane film has lower value of T r.
It has been confirmed that the photocurrent varies in
a superlinear fashion with the electric field. This can
be explained by the electric field distortion not only of
the Coulombic potential field between an electron-hole
pair, but also of the potential between nearest hopping
sites.
The unusual temperature dependence of dark and photocurrent observed for the C 60 doped PMPS film is explained as follows. The photogenerated electrons in the
polysilane transfer to C60 as denoted by (5) in Fig. 4 and
holes can stay long in the polysilane because of the spatial separation from electrons. At a low temperature,
holes are trapped in the tail states. The accumulation
Nakayama, et al.
of the trapped holes shifts their quasi-Fermi level toward, or into, the HOMO band and then the dark and
photocurrent become larger and larger with time. The
measurement was carried out with increasing temperature so that the variation at 15 K < T < 130 K reflects
the accumulation of trapped holes. The pointwise-manner detection of data could not remove the trapped holes
in this case. As temperature rises beyond 130 K, the
trapped holes in the polysilane and the electrons trapped
at the lowest level in the LUMO band of C 60 are thermally excited to recombine each other, by which the dark
current decreases and the ratio of the photocurrent to
the dark current becomes large.
We have confirmed that the accumulation of trapped
holes does not occur by the excitation using the photon
energy of 2.25 eV is difficult, indicating that process 3
in Fig. 4 is not efficient at a low temperature, although
it contributes to high photoconduction at room temperature. These results eliminate the possibility 6 of the energy transfer from PMPS to C60 followed by process 3 at
a low temperature. However, it might be possible at high
temperature and could cause the dip of Id and Ip curves
at 160 K. This inference is not conclusive and is subject
to further study.
Conclusion
The photoconduction spectrum of PMPS film has an
onset 0.1 eV higher than that of the optical absorption.
This photoconduction is not caused by free carriers created from excitons at the electrodes. For the free-car rier generation in the bulk, the disorder model is more
likely than the charged polaron model. Analysis based
on the disorder model has successfully explained the
zero-temperature photoconductivity. Doping of C 60 efficiently sensitizes the photoconduction of PMPS in the
spectral region where C 60 has optical absorption. This
sensitization is suppressed at low temperatures. In the
low temperature region, photogenerated holes are
trapped in the tail states of PMPS in C 60 doped films.
The accumulation of trapped holes moves the Fermi level
to increase the dark and photoconductivity . The holes
trapped in PMPS and the electrons trapped in C 60 are
thermalized to recombine each other at temperatures
higher than 130 K.
Acknowledgments. This work was supported by the
Grant-in-Aid for Science Research (C) from the Ministry of Education, Science, Sports and Culture of Japan.
References
1. R. G. Kepler, J. M. Zeigler, L. A. Harrah, and S. R. Kurtz, Photocarrier
generation and transport in σ-bonded polysilanes, Phys. Rev . B35,
2818 (1987).
2. Y. Nakayama, K. Hirooka and R. West, Electric conduction in oriented
polysilane films, Solid State Commun. 100, 759 (1996).
3, Y. Nakayama, A. Saito, K. Hirooka, and R. West, Carrier transport in
oriented polysilane films, Proc. of IS&T’s 13th Int. Conf. on Digital
Printing Tech. , IS&T, Springfield, VA, 1997, p. 207.
4. Y. Nakayama, A. Saito, S. Ninomiya, S. Akita, M. Aramata, and R.
West, Hole drift mobility along silicon chains in polysilane films, Proc.
3rd Int. Conf. on Imaging Science and Hardcopy, Chongqing, China
1998, p. 47.
5. Y. Wang, R. West and C. H. Yuan, Fullerene-doped polysilane
photoconductor, J. Am. Chem. Soc. 115, 3844 (1993).
6. R. G. Kepler and P. A. Cahill, Photoinduced charge transfer and charge
carrier generation in polysilane films containing C60 molecules, Appl.
Phys. Lett. 63, 1552 (1993).
7. C. H. Lee, G. Yu, D. Moses, K. Pakbaz, C. Zhang, N. S. Sariciftci, A.
J. Heeger, and F. Wudl, Sensitization of the photoconductivity of conducting polymers by C60: Photoinduced electron transfer, Phys. Rev .
B 48, 15425 (1993)
8. R. G. Kepler and Z. G. Soos, Electronic excitations of poly(methylphenylsilane) films, Phys. Rev. B 43, 12530 (1991).
9. R. G. Kepler and Z. G. Soos, The role of excitons in charge carrier
production in polysilanes, Primary Photoexcitations in Conjugated
Polymers: Molecular Exciton versus Semiconductor Band Model , N.
S. Sariciftci, Ed., World Scientific, 1997, p. 363.
10. A. J. Heeger, Nature of the primary photoexcitations in poly(arylenevinylenes): bound neutral excitons or charged polaron pairs, Primary
Photoexcitations in Conjugated Polymers: Molecular Exciton versus
Semiconductor Band Model, N. S. Sariciftci, Ed., World Scientific,
1997, p. 20.
11. H. Bässler, Excitons in conjugated polymers, Primary Photoexcitations
in Conjugated Polymers: Molecular Exciton versus Semiconductor
Band Model, N. S. Sariciftci, Ed., World Scientific, 1997, p. 51.
12. U. Albrecht and H. Bässler, Yield of geminate pair dissociation in an
energetically random hopping system, Chem. Phys. Lett. 235, 389
(1995).
13. H. G. Grimmeiss and L-A. Ledebo, Spectral distribution of photoionization cross sections by photoconductivity measurements, J. Appl.
Phys. 46, 2155 (1975).
14. W. B. Jackson, N. M. Amer, A. C. Boccara and D. Fournier, Photothermal deflection spectroscopy and detection, Appl. Optics 20, 1333
(1981).
15. S. Saito and A. Oshiyama, Cohesive mechanism and energy bands
of solids C 60, Phys. Rev. Lett. 66, 2637 (1991).
16. B. I. Shklovskii, H. Fritzsche and S. D. Baranovskii, Recombination
and photoconductivity in amorphous semiconductors at low temperature, J. Non-Cryst. Solids 114, 325 (1989).
17. S. Aihara, N. Kamata, W. Ishizawa, M. Umeda, A. Nishibori, D.
Terunuma, and K. Yamada, Efficient intermolecular energy transfer
between polysilanes revealed by time-resolved photoluminescence,
Jpn. J. Appl. Phys. 37, 4412 (1998).
18. Y. Nakayama, P. Stradins and H. Fritzsche, Metastable centers and
photoconduction in a-SiN x:H, J. Non-Cryst. Solids 164–166, 1061
(1993).
Photocarrier Generation In Polysilane Films Doped With And Without Fullerene
Vol. 43, No. 3, May/June 1999
265
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Sensitized and Intrinsic Carrier Generation in
Phenethylperylene/Tritolylamine Thin Film Structures
Z. D. Popovic,*,▲ R. Cowdery,† I. M. Khan,‡ A.-M. Hor,*,▲ and J. Goodman‡
*Xerox Research Centre of Canada, Mississauga, Ontario, Canada
†
Eastman Kodak Company, Rochester, New York
‡
Center for Photoinduced Charge Transfer, University of Rochester, Rochester, New York
The photoconductivity mechanism was investigated for vacuum-evaporated phenethylperylene (PPEI) films deposited on a thin
polycarbonate film doped with varying concentrations of tritolylamine (TT A) and subsequently exposed to methylene chloride
vapors. Compared to structures without TT A in the polycarbonate layer , the presence of TT A leads to an increase of carrier
generation efficiency and strong quenching of perylene fluorescence indicating a surface-sensitized carrier generation process.
Fluorescence quenching measurements on samples with and without TT
A show a linear correlation between fluorescence quenching
and carrier generation at high fields. In the presence of TT A, significant photoconductivity is observed long before the appear ance of fluorescence quenching. A marked change of curvature (inflection point) in carrier generation accompanies the appear ance of fluorescence quenching at fields in excess of 100 MV/m. These results demonstrate that in samples containing TTA, two
different carrier generation mechanisms are operating simultaneously . At low fields, carrier generation is dominated by the
sensitized component. At high fields, although the sensitized component saturates, the intrinsic component causes a further
increase in overall carrier generation. The experimental results are consistent with the notion that the intrinsic photoconduct
ivity component originates from direct dissociation of the fluorescent first excited singlet state into free carriers.
Journal of Imaging Science and Technology 43: 266–269 (1999)
Introduction
The mechanisms of charge carrier generation in photoconductive solids have been the subject of extensive experimental and theoretical studies. Both intrinsically
and extrinsically controlled processes in solid state systems have been described. In intrinsic photoconductors,
carrier generation is an intrinsic property of the bulk
material and the presence of uncontrolled impurities can
adversely affect photoconductive properties. In extrinsic photoconductors, pure materials usually show no or
very small photoconductive response. Only the addition
of suitable sensitizers makes the photoconductive response significant. Photoconductors can be further divided into surface-sensitized or bulk-sensitized types,
depending on the sentisizer location.
Both surface- and bulk-sensitized carrier generation
have been well documented in the scientific literature.
In a number of inorganic1 and organic systems,2 increased
photoconductive response has been obtained by adding
suitable dopants to the material bulk. Surface sensitization of photoconductivity has been observed in anthracene, perylene, and dibenzanthracene crystals. 3
Phthalocyanines also exhibit increased photoconductivity in the presence of sensitizers. 4 In addition, surface
Original manuscript received November 23, 1998
▲ IS&T Member
© 1999, IS&T—The Society for Imaging Science and Technology
266
sensitization is responsible for efficient carrier generation in aggregate type photoconductors 5 where carrier
generation was shown to originate at the interface of the
filamentary crystallized dye and the amorphous hole
transport layer phases. It was also observed in thin films
of benzimidazole perylene 3,4,9,10-tetracarboxylic-acid
(BZ perylene) overcoated with tetraphenyldiamine (TPD)/
polycarbonate hole transport layer6 and in azo pigments
in contact with a number of hole transport molecules.7
In the present work the carrier generation mechanism
is investigated for vapor-deposited thin films of perylene
bis(phenethylimide) (PPEI or phenethylperylene) on a
polycarbonate polymer layer doped with varying concentrations of tritolylamine (TT A) hole transport molecules and then exposed to methylene chloride vapors.
The structures of these materials are given in Fig. 1.
Recently it has been shown that excitons in PPEI are
strongly quenched by electron donor molecules, 8 leading to the conclusion that exciton diffusion lengths are
very long, possibly exceeding 1 µm. In this work it is
shown that TTA strongly quenches phenethyl perylene
fluorescence. In addition, by measuring the electric field
induced fluorescence quenching and carrier generation
using a delayed field collection technique, it is shown
that carrier generation has two components: (i) extrinsic, involving exciton dissociation by the TTA at the pigment surface, and (ii) intrinsic, originating from direct
dissociation of excitons into free carriers.
Experimental
The phenethylperylene pigment [Fig. 1(a)] was synthesized by cyclizing perylene tetracarboxylic dianhydride
Figure 1. Structures of the molecules used in this work: (a)
phenethylperylene, and (b) tritolylamine.
with an excess of phenethylamine, as described in Ref.
9. Tritolylamine [Fig. 1(b)] was synthesized as described
in Ref. 10. Makrolon 5705 polycarbonate polymer was
purchased from Mobay Chemicals and used as received.
The sample preparation procedure was as follows.A thin
layer of hexamethyldisilazane (HMDS, Olin Microelectronics Materials) was first spin coated on a NESA glass
substrate, rotating at 5000 rpm to form a barrier layer
of HMDS. A doped polycarbonate layer solution, containing tritolylamine (1% to 20% by weight) and Makrolon
binder to form a 3% weight solids solution in a
dichloromethane/1,1,2-trichloroethane solvent system,
was then spin coated on the HMDS barrier layer by
pipetting 2 mL of the 3% solids solution onto the substrate (in the static mode), then ramping to 500 rpm for
10 s, followed by 2000 rpm for 1 min to form a 1
µm
charge transport layer. The perylene was then vapor
deposited onto the charge transport layer to a thickness
of 0.1 µm and solvent treated with dichloromethane vapors to undergo a polymorphic conversion to the
photoactive form. At the end, a thin Al electrode was
vacuum evaporated to form a sandwich cell. This led to
highly reproducible samples as demonstrated in Fig. 2.
The difference in absorbance around 400 nm is due to
different concentrations of TTA that has an absorption
edge in that wavelength region.
The electric field in the pigment layer was calculated
from the applied voltage, Vappl, as
Eappl = (CVappl)/(Aε0εr),
(1)
where A = 0.4 cm2 is the cell area, εr = 5.3 is the relative
dielectric constant of phenethylperylene,11 and the other
symbols have their usual meanings. It is interesting to
note that Eq. 1 is generally valid for a plane capacitor
with any number of layers with different dielectric constants. In order to calculate the electric field in any specific layer, it is only necessary to know the total sample
capacitance, sample area, and the dielectric constant of
that particular layer. The dielectric constants of other
layers are not important. This is a direct consequence
of the continuity of the dielectric displacement vector
Figure 2. Absorption spectra of phenethylperylene films obtained by vacuum evaporation and subsequent exposure to
methylene chloride vapors. Spectra A, B, C, and D were obtained at concentrations of TT A in the polycarbonate matrix of
0%, 1%, 5%, and 20% by weight, respectively . At 465 nm absorbance of all samples was about 1 ± 0.1. For comparison the spectra are normalized to the same value at 465 nm for all samples.
across interfaces dividing materials of different dielectric constants.
Carrier generation was measured using a delayed field
collection technique and the electric field induced
quenching of total fluorescence as described in detail in
Ref. 12. Samples were illuminated with 532 nm, 5 ns
pulses from a frequency-doubled Nd-Y AG laser. In order to prevent carrier injection, the samples were biased in a unipolar fashion and measured at low
repetition rates of about 0.1 Hz. Reproducible measurements were only obtained when erase light pulses were
applied to shorted samples prior to each biased measurement. Time-resolved fluorescence with no applied
sample bias was measured using single-photon counting with picosecond 590 nm light pulses for sample excitation, 670 nm light detection and the experimental
setup described in Ref. 13. T ime resolved electric field
induced fluorescence quenching measurements were
also attempted,13 but these were not possible due to unipolar sample bias, which led to charge accumulation and
quick electric breakdown in the sample.
Results and Discussion
Figure 3 shows the fluorescence spectra of four samples
with 0%, 1%, 5%, and 20% TT A by weight in the polycarbonate layer, which will be referred to as samplesA,
B, C, and D, respectively. For these measurements the
samples were excited with 550 nm light illuminating
the evaporated pigment film from the glass substrate
side. As the concentration of TTA increases, the fluorescence dramatically decreases but the shape of the fluorescence spectra does not appear to change significantly
.
Only a small shift to shorter wavelengths is observed
when the concentration of TT A increases. The ratio of
peak fluorescence for 0% TTA to 20% TTA is about 14.
This significant decrease of fluorescence can only be
explained in terms of a charge transfer reaction induced
by the presence of TT A. Exposure to methylene chloride
vapors, which is necessary to induce the phenethyl-
Sensitized and Intrinsic Carrier Generation in Phenethylperylene/Tritolylamine ...
Vol. 43, No. 3, May/June 1999
267
Figure 3. Fluorescence spectra of Samples A, B, C, and D.
The presence of TT A in the polycarbonate matrix leads to a
significant fluorescence decrease.
perylene conversion to photoactive form, also enables
diffusion of TTA molecules so that they come into intimate contact with the pigment molecules.
As
phenethylperylene is only sparingly soluble in methylene chloride, most likely, TTA molecules are not incorporated into perylene crystal lattice, but instead diffuse
along the grain boundaries of polycrystalline pigment
film. The conclusion that fluorescence quenching induced by the presence of TT A by charge transfer reaction mechanism is also supported by photoconductivity
measurements (presented later).
Figure 4 shows the time-resolved fluorescence decay
for Samples A and D containing 0% and 20% TTA in polycarbonate with illumination from the glass side. As the
concentration of TTA in the Makrolon layer increases
the decay becomes significantly faster . In the sample
with 20% TTA the lifetime of the dominant component
is only about 30 ps, which is close to the detection limit
of our instrumentation. In contrast, the dominant decay component for 0% TTA sample is about 600 ps. Surface quenching by TTA is very efficient and leads to a
decrease in fluorescence lifetime by a factor of about 20
with almost complete disappearance of the longer lived
tail of fluorescence observed in samples with 0% TT A.
This is consistent with the assumption that long-lived
fluorescence observed in samples with no TT A corresponds to trapped excitons, most probably located on
the surface of grains comprising the polycrystalline pigment film. Twenty percent TTA in Makrolon almost completely suppresses the long-lived fluorescence
component, which decreases by about two orders of magnitude (Fig. 4). This is an indication that the remaining
fluorescence predominantly originates from mobile excitons that decay radiatively before reaching the TT A
quenching sites. This conclusion will be important later
for the interpretation of electric field induced fluorescence quenching in samples containing 20% TT A. It is
also consistent with the blue shift of the fluorescence
peak as TTA concentration increases (Fig. 3).
The quantitative measure of electric field induced fluorescence quenching is fluorescence quenching efficiency,
Φ(E), defined by
Φ(E) = [I f(0) - I f(E)]/I f(0),
268
Journal of Imaging Science and Technology
(2)
Figure 4. Time resolved fluorescence decays for Samples A
(0% TTA) and D (20% TT A) with illumination from the glass
side. Curve I shows the instrument response function.
where If(E) is integrated fluorescence at field E applied
to the sample. Figure 5(a) shows fluorescence quenching
as a function of the electric field for Samples A and D. In
the case of Sample A, which does not contain TTA in the
polymer film, fluorescence quenching can be detected at
about 60 MV/m and gradually increases with increasing
applied field. For Sample D, with 20% of TTA in the polymer film, no fluorescence quenching is observed until the
electric field reaches 100 MV/m. Figure 5( b) shows the
plot of fluorescence quenching as a function of relative
photoresponse measured by the delayed collection field
method. Assuming that fluorescence quenching is caused
by exciton dissociation into carriers by the electric field,
a linear relationship is expected between the carrier generation efficiency, η(E), and fluorescence quenching:12
Φ (E) = [η(E) – η(0)]/[1 – η(0)],
(3)
where η(0) corresponds to a part of carrier generation
that is not connected to electric field—induced fluorescence quenching. In the case of samples with 20% TTA,
η(0) is assumed to represent saturation value of the sensitized carrier generation.
Let us introduce the relative photoresponse,R, which
is defined as
R(E) = ∆V/I light ,
(4)
where ∆V is a voltage drop induced on a sample by a
light pulse of energy I light. The carrier generation efficiency is proportional to relative photoresponse,
η(E) = C R(E) ,
(5)
and by combining Eqs. 3 and 5 at high fields we obtain,
Φ(E) = C R(E) /[1 – η(0)] – η(0)/[1 – η(0)].
(6)
It therefore follows that at high fields a linear correlation is expected between relative photoresponse and
fluorescence quenching [Fig. 3(b)]. The slope and intercept of this linear plot determine constant C, which enables rescaling of relative photoresponse, R, to quantum
efficiency, η (Eq. 5).12
Popovic, et al.
Fluorescence
Quenching Efficiency
Fluorescence
Quenching Efficiency
Carrier Generation
Efficiency
Figure 5. (a) Fluorescence quenching, Φ, as a function of electric field for Samples A (open circles) and D (solid circles); (b)
fluorescence quenching plotted as a function of relative
photoresponse, R. Straight lines are best, least square fits to
high field data; (c) rescaled carrier generation efficiency, η, for
Samples A and D.
Rescaled data are shown in Fig. 5(c). At low fields,
the sample without TT A shows much smaller carrier
generation efficiency than the sample with 20% TT A,
although the latter samples do not show detectable electric field induced fluorescence quenching. These results
can only be interpreted as a surface-sensitized carrier
generation by TTA. The shape of carrier generation efficiency curve at high fields is very interesting. It shows
a tendency to saturate, but around 100 MV/m an inflection point appears and change of curvature occurs. This
coincides with the appearance of fluorescence quenching around the same field and indicates a change in the
dominant carrier generation mechanism. In the discussion of time-resolved fluorescence decays [Figs. 4(a) and
4(b)], we concluded that in the presence of TT A, fluorescence must originate from mobile intrinsic excitons
that decay radiatively before reaching TT A quenching
sites. It therefore follows that the continuing increase
in carrier generation in samples with TT A originates
from carriers generated by the direct dissociation of the
intrinsic mobile singlet excitons.
Direct dissociation of the first excited singlet state at
high applied fields is not surprising. Electric fields lead
to gradients in the valence and conduction energy levels. If they are strong enough, it is energetically possible that the bound singlet energy levels match the
energy of electron-hole pairs at some distance accessible
by tunneling. For example, a field of 100 MV/m will produce an energy change of 0.2 eV at 2 nm, which may be
enough to lead to excited state quenching by tunneling,
to separated electron and hole pair states.
It is interesting to compare values for carrier generation efficiency determined in this work for samples without TTA to xerographic measurements by Magin and
Borsenberger 11 on thin evaporated films of
phenethylperylene. For polycrystalline samples, which
should be similar to ours, they measured a quantum
efficiency of 10% at a field of 50 MV/m. At this same
field we obtained a quantum efficiency of 4%, which is
in reasonable agreement considering the differences in
the sample preparation procedures.
Conclusions
The relative photoresponse and fluorescence quenching
were measured in thin films of phenethylperylene pigment induced by the presence of hole transport molecule
TTA and by the electric field. The results clearly demonstrate that TTA is a surface sensitizer greatly enhancing photoconductivity when compared to samples that
do not contain TTA. Electric field induced fluorescence
quenching measurements, combined with relative
photoresponse measurements, indicate that in samples
with TTA a change of carrier generation mechanism
occurs at high fields. Although the sensitized carrier
generation saturates, further increase of carrier generation is observed originating from direct dissociation
of the intrinsic mobile excitons into free carriers. Car rier generation from the first excited singlet state at
high fields should be a universal property of many materials. High enough fields will lead to sufficient gradients in conduction and valence energy levels to enable
electron tunneling from the bound excited states into
separated carriers.
Acknowledgments. We want to thank D. W eiss for
fruitful discussions, and E. Magin and J. Sinicropi for
their help in sample preparation.
References
1. A. Rose, Concepts in Photoconductivity and Allied Problems,
Interscience Publishers, New York 1963, pp. 42-47.
2. H. Meier, Organic Semiconductors, Verlag Chemie, Weinheim, 1974,
pp. 352, 355.
3. H. Meier, Organic Semiconductors, Verlag Chemie, Weinheim, 1974,
pp. 350–352.
4. R. Loutfy and R. Menzel, J. Chem. Soc. 102, 4967 (1980); R. O Loutfy
and C. K. Hsiao, Photogr. Sci. Eng. 24, 165 (1980).
5. P. M. Borsenberger, A. Chowdry, D. C. Hoesterey, and W. May, J.
Appl. Phys. 49, 5555 (1978).
6. Z. D. Popovic, A. Hor and R. O. Loutfy, Chem. Phys. 127, 451 (1988).
7. M. Umeda, T. Shimada, T. Aruga, T. Niimi, and M. Sasaki, J. Phys.
Chem. 97, 8531 (1993).
8. B. A. Gregg, J. Sprague and M. W. Peterson, J. Phys. Chem. B 101,
5362 (1997).
9. P. M. Borsenberger, M. T. Regan and W. J. Staudenmayer, U.S. Patents 4,578,334 and 4,618,560 (1986).
10. C. J. Fox and W. A. Light, U.S. Patent 3,706,554 (1970).
11. E. H. Magin and P. M. Borsenberger, Proc. IS&T’s Eight Int’l. Congress on Advances in Non-Impact Printing Technologies, IS&T, Springfield, VA, 1992, p. 243.
12. Z. D. Popovic, J. Chem. Phys. 78, 1552 (1983).
13. Z. D. Popovic, M. I. Khan, S. J. Atherton, A. Hor, and J. L. Goodman,
J. Phys. Chem. B 102, 657 (1998).
Sensitized and Intrinsic Carrier Generation in Phenethylperylene/Tritolylamine ...
Vol. 43, No. 3, May/June 1999
269
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Image Resolution in Liquid Development for Electrophotography
I. Chen▲
Wilson Center for Research and Technology, Xerox Corporation, Webster, New York
The charge transport model of liquid immersion development (LID) is extended for consideration of line-pair images. The transfer of modulation from the image-wise charge distribution on the photoreceptor surface to the developed toner mass distribution
is examined as a function of line-width. The line-width dependence of the modulation transfer is found to be weaker than that
previously expected for electrostatic images, even at line-widths of small multiples of the receptor thickness. This indicates a
favorable image resolution for LID that can be attributed to a new reason that is inherent in the electrophoretic development
process, and is independent of the toner size or the development gap size.
Journal of Imaging Science and Technology 43: 270–273 (1999)
Introduction
Development of electrostatic images with toners dispersed in non-polar liquids, i.e., liquid immersion development (LID), has been generally perceived to possess
the potential of achieving the high image resolution demanded by high quality color prints due to the smaller
toner size (∼1 µm) and the feasibility of narrower development gap (<100 µm).1,2 Although laboratory models as
well as commercial printers based on LID3,4 have demonstrated high quality color prints as expected, there is no
quantitative analysis of the attributes responsible for the
high image quality. Recent mathematical models of LID
in the literature,5–7 though quite extensive in taking into
consideration the space-charge effects, are all limited to
the treatment of solid-area images, and hence, do not provide information on the image resolution power associated with the electrophoretic development process. In this
article, the charge transport model of LID is extended
for consideration of line-pair images. The transfer of
modulation from the image-wise charge distribution on
the photoreceptor surface to the developed toner mass
distribution is examined as a function of line-width. The
line-width dependence of the modulation transfer is found
to be weaker than that previously expected for electrostatic images, even at line-widths of small multiples of
the receptor thickness. This indicates that there is a new
reason for a favorable image resolution for LID, that is
inherent in the electrophoretic development process, and
is independent of the previously suggested processes,
namely, the small toner size or the small development
gap.
The mathematical procedure for calculating the toner
deposition in LID of line-pair images is described in the
Original manuscript received February 12, 1998
▲ IS&T Member
© 1999, IS&T—The Society for Imaging Science and Technology
270
next section. This is followed by the presentation and
discussion of the major numerical results. A comparison with previous treatments of image resolution for
electrostatic images suggests the physical reason for the
favorable image resolution.
Mathematical Procedure
As in the previous work, 6 the development zone is represented by the plane-parallel geometry shown in Fig.
1. The electrophoresis in the liquid developer (or ink), 0
≤ z ≤ Li, is described by the continuity equations for the
charge densities ρ(x, z),
∂ρ/∂t = ± ∇(µρE)
(1)
where ρ stands for the charge densities of toners ρ t and
counter-ions ρp, and µ stands for the corresponding mobilities µt and µp. In inks of practical interest, the density of co-ions is usually so small that it is neglected for
simplicity in this discussion. Without loss of generality,
the toners are assumed to be negatively- and the
counter-ions positively-charged. The field E(x, z) can be
calculated from the potential V(x, z), that can be obtained from the solution of Poisson’s equation,
∇E = − ∇ 2 V = (ρt + ρp)/εi
(2)
with the boundary condition that the discontinuity in the
normal (z) component of electric displacements at z = 0 is
equal to the charge density Qs(x) at the receptor surface,
εiEiz(x, 0) – εrErz(x, 0) = Q s(x)
(3)
where εi and εr are the permittivities, and Eiz and E rz
are the normal components of fields in the ink and receptor, respectively. The receptor is assumed to be spacecharge-free during the development. Other boundary
conditions are that the receptor is grounded at the substrate (z = −Lr), i.e., V(x, −Lr) = 0, and that a bias voltage Vb is applied at the development electrode at z = Li,
Z
Vb
Li
X
0
Developed Mass
Liquid
Develper
(Ink)
Qs
Receptor
–Lr
Figure 1. Schematic geometry and coordinate system for the
mathematical model of liquid immersion development.
i.e., V(x, Li) =Vb. The initial conditions are that both the
toner and the counter -ion charge distributions in the
ink are uniform, having the values −ρt(x, z) = ρp(x, z) = ρo.
The toner deposition results from the arrival of toner
current Jzt(0) at the receptor surface. This is represented
by the time rate of change of toner chargeQt(x, t) at the
surface z = 0 as,
Figure 2. Time evolution of developed mass/area, M(x,t), calculated from the deposit toner charge Q t(x, t), at three positions in a sinusoidal line-pair, the peak (x = w/2), the node (x =
w) and the valley ( x = 3w/2), for two line-widths w. The normalized units for mass/area, M o ≡ –Qo/qm, and time, to ≡ εi/µ tρ o,
are defined in the text.
∂Qt(x, t)/∂t = – Jzt(0) = – [µ tρtEz]z=0
mensional Poisson equation has to be solved by the finite difference method at each time interval t > 0.
(4)
with the initial value, Q t(x, 0) = 0. The initial charge
distribution at the receptor surface is that of the linepair image, given by a sinusoidal function of line-width
w, (i.e., period 2w), and amplitude Qo,
Q s(x, 0) = (Qo/2)[1 + sin (πx/w)]
(5)
where the charge polarity is assumed to be positive, Qo
> 0, to be consistent with the assumption of negative
toners. It is assumed that toners and/or counter -ions
arriving at the receptor surface come in close contact
with, and modify, the surface charge density Qs according to the time rate of change,
∂Qs(x, t)/∂t = Jzp(0) − Jzt(0) = [µpρpEz]z=0 − [µtρ tEz]z=0
(6)
Because the ink layer is neutral and space-chargefree at t = 0, Eq. 2 reduces to Laplace equation whose
solution can be expressed in a closed form as,1,8
V(x, z) = U0(z) + U1(z)sin(πx/w)
(7)
where, in the receptor, –L r ≤ z ≤ 0,
U 0(z)= (Vb + Q oD i/2)(D r + z/εr)/(Dr + Di)
(8a)
U 1(z) = (wQo/2π)sinh[π(Lr + z)/w]/[εrcosh(πLr/w)
+ εicoth(πLi/w)sinh(πL r/w)]
(9a)
and in the ink, 0 ≤ z ≤ Li.
U0(z)= [(Vb + QoD i/2)D r + (Vb − QoD r/2)z/εi]/(Dr + Di) (8b)
U1(z)= (wQo/2π)sinh[π(Li − z)/w]/
[ εrcoth(πLr/w)sinh(πLi/w) + εicosh(πLi/w)]
(9b)
with D r = Lr/εr, and Di = Li/εi,
Using the above initial values and time derivatives,
the time evolution of the charge distributions can be
calculated numerically. A feature that differs from the
previous case of solid-area images is that the two-di-
Image Resolution in Liquid Development for Electrophotography
Results and Discussion
For a constant toner charge-to-mass ratio q m, the distribution of developed toner mass, M(x, t) can be obtained from the total toner charge at the receptor
surface, Qt(x,t) of Eq. 4, as
M(x, t) = Qt(x, t)/q m
(10)
Figure 2 shows the time evolution of M(x, t) at three
positions in a line-pair: the peak (at the center of on
line, x = w/2), the node (at the geometrical border between on and off lines, x = w) and the valley (at the
center of off line, x = 3w/2), for two line-widths w = 1
and 8, in multiples of the receptor thicknessLr. The ink
layer thickness is Li = 2Lr; the permittivities of ink and
receptor are equal, εi = εr; the toner and counter-ion mobilities are equal, µt = µp. The initial toner (or counterion) charge density is ρo = Qo/Lr. The developed mass M
is given in units of Mo ≡ –Q o/q m where Qo is the initial
charge amplitude introduced in Eq. 5. The abscissa,
time, is in units of to ≡ εi/µtρ o (which can be recognized
as the dielectric relaxation time of the ink). For typical
values of Lr= 25 µm, qm = –250 µC/g, Qo = 50 nC/cm2, εi = 2
× 10 –13 F/cm, and µ t = 10–4 cm2/Vs, the units have values
of Mo = 0.2 mg/cm2 and t o = 10–4 s.
The development is seen to approach saturation in
about 100 to. The asymptotic spatial distributions ofM(x,
t) for the above two line-pair images are shown in Fig.
3. For comparison, the initial surface charge distribution, Q s(x,0) of Eq. 5, is also shown (in units of Q o).
The effects of reduced line-width can be seen in Fig. 2
as the slightly slower build up of the developed mass at
the peak (x = w/2) and a larger deposition at the valley
(x = 3w/2). The incomplete neutralization, seen in Fig.
3 near the peak, is a consequence of space-charge-limited transport, and the deposition near the valley , (especially for the narrower line), is caused by the fringe
fields. The combined effects lead to a smaller developed
Vol. 43, No. 3, May/June 1999
271
Developed Mass
Dev. Mass Contrast, MC
Figure 3. Spatial distributions of the developed mass/area,
M(x, t) at t = 100 t o, for two line-widths w = 8 and 1 in multiples of the receptor thickness Lr. The initial surface charge
distribution Q so= Qs(x,0) is also shown for comparison.
Figure 4. Developed mass contrast, M C(w) versus line-width
w, for four values of ink-layer thickness Li. w and Li are in
units of the receptor thickness L r. The contrast is normalized
to the maximum value Q o/q m.
MC(w) ≡ M(w/2, t∞) − M(3w/2, t∞)
(11)
Figure 4 shows the developed mass contrastMC (in units
of –Qo/qm) as a function of line-width, for four values of ink
layer thickness, Li. The change of MC with line-width is
seen to increase slightly as the ink layer thickness increases from Li = 1, to 8 Lr. Similar results are obtained
with other values of parameters within the range of practical interest (e.g., with µt ≠ µp and εi ≠ εr). The decrease of
contrast with decreasing line-width represents the loss of
modulation transfer, resulting in failure to resolve fine
lines. The loss of modulation transfer at small line-widths
is common in many imaging processes, and can occur in
various stages—in exposure, development and/or transfer. A striking observation from Fig. 4 is the smallness of
this resolution loss compared to what is previously expected for electrostatic imaging process.1,2
Although many different methods have been used to
develop electrostatic latent images,2,9,10 the image resolution has only been discussed by examining how well
the modulation in the image-wise charge distribution
is retained in the surface voltage or field distributions.1,2
The developed mass distribution is assumed to be proportional to the surface voltage or field from the imagewise charge distribution. Thus, for example, for the
sinusoidal line charge pattern Q s (x, 0) of Eq. 5, the surface voltage V(x, 0), and the normal component of field
Ez(x, z) can be calculated from Eqs. 7, 8(b), and 9(b).
The mass contrast MV that is proportional to the modulation of the surface voltage would be,
MV(w) ∝ V(w/2, 0) – V(3w/2, 0) =
(wQo/π)/[εrcoth(πLr/w) + εdcoth(πLd/w)]
(12)
and the mass contrast that is based on the modulation
of the normal field at z is,
ME(w) ∝ E z(w/2, z) – E z(3w/2, z) = Qocosh[π(Ld − z)/w]/
[εrcoth(πLr/w)sinh(πL d/w) + εdcosh(πL d/w)] (13)
272
Journal of Imaging Science and Technology
Dev. Mass Contrast, MV
mass contrast MC for the narrower line, where M C is
defined as the difference in the asymptotic values,M(x,
t∞), at the peak and the valley,
Figure 5. Developed mass contrast based on surface voltage,
M V, Eq. 12, versus line-width w, calculated for four values of
development gap Ld. w and Ld are in units of the receptor thickness Lr, and M V is normalized to its maximum as unity.
where the subscript i for the ink layer is replaced by d
for the development gap. The line-width dependence of
MV and ME, for the geometrical conditions similar to that
of Fig. 4 are shown in Figs. 5 and 6, respectively , and
discussed below. The contrasts are in units normalized
to the maximum as unity (i.e., voltage modulation in units
of QoLr/εr, and the field modulation in units of Qo/εr).
The decrease in the voltage-based mass contrast M V
due to the line-width reduction is seen in Fig. 5 to be
much more pronounced than that for the mass contrast
MC shown in Fig. 4. The curves corresponding to the
case of development gap L i or Ld = 2Lr, from each of the
three figures, are reproduced for comparison in Fig. 7.
The variation of the field-based mass contrast ME calculated with the field at z = 0.2 Lr above the receptor
surface, as shown in Fig. 6, has an appearance differ ent from that of MV in Fig. 5. For the development gap
Chen
Dev. Mass Contrast, ME
Dev. Mass Contrast
Figure 6. Developed mass contrast based on normal field,ME ,
Eq. 13, versus line-width w, calculated at a distance z = 0.2
above the surface, for four values of development gap Ld. w, z
and L d are in units of the receptor thickness Lr, and M E is normalized to its maximum as unity.
Figure 7. Comparison of developed mass contrasts calculated
from deposit toner charge M C, Eq. 11, from voltage modulation
M V, Eq. 12, and from field modulation M E, Eq. 13, for the case
of development gap Li or L d = 2L r, (i.e., the reproduction of one
curve each from Figs. 4, 5 and 6).
greater than the receptor thickness, Ld > L r, a maximum
in contrast ME appears at a line-width w ≈ 2L r. This is
the well-known account for the poor development of
solid-area as well as very fine line electrostatic images
with a large development gap (e.g., cascade development). Furthermore, contrary to the case of voltagebased MV, the field-based ME is larger for the smaller
development gap (Fig. 6). In fact, based on the latter
feature, Schaffert has predicted a higher resolution of
LID than dry-toner development because of the feasibility of smaller gap in LID.1,2
Returning to Fig. 4, the similar increase of contrast
with decreasing gap (or ink layer thickness) is seen for
the mass contrast MC, which is calculated directly from
the electrophoretic motion of charged species. In other
words, the LID mass contrast MC is more similar to the
field-based contrast ME, rather than to the voltage-based
contrast M V. This suggests a physical reason for the
weaker line-width dependence of MC, shown in Figs. 4
and 7, namely, the importance of local field variation
due to the space-charge-perturbed electrophoretic motion in LID. It should be noted that the contrasts shown
in Figs. 5 and 6 are calculated with the voltage and the
field in a space-charge-free development gap, from the
solutions of Laplace equation, Eqs. 7, 8(b), and 9(b),
based on the initial image-wise charge distribution.
However, the gap is not space-charge-free during most
of LID time. The fields used in the calculation of curves
in Fig. 4 are the self-consistent fields from the solutions
of Poisson’s equation, based on the instantaneous distributions of charges in the gap (or ink) and the neutralized image-wise charge on the receptor surface.
It is expected that image resolution can be described
by the voltage-based mass contrastMV, if the toner deposition is generation-limited. That is, the electrostatic
force from the image-wise charge is used mostly to create free toners, which arrive at the receptor surface almost instantaneously because of the high mobility, e.g.,
in the dry media. On the other hand, the toner deposition in LID is transport- and space-charge-limited, because of the large amounts of toner and counter -ion
charge that move simultaneously in the more viscous
and lower mobility liquid media. The local field that
determines the transport is influenced by the local
charge densities (of toners and counter-ions in the gap)
as much as by the charge on the receptor.
Image Resolution in Liquid Development for Electrophotography
Summary and Conclusions
The charge transport model of liquid immersion development is extended and applied to line-pair images of various line-widths. The developed mass distributions in the
direction perpendicular to the line are calculated from the
toner charge deposited on the receptor , and the peak-tovalley contrasts are investigated as a function of linewidth. The decrease of modulation transfer with the
line-width is found to be much less serious than that previously expected for electrostatic imaging. This indicates
a favorable image resolution which can be attributed to a
new reason inherent in electrophoretic motions of toners
and counter-ions in LID, hence, non-existant in dry powder development. The consideration of the contributions
from toner and counter -ion space charges and the timevarying receptor surface charge to the electric fields that
drive the development, is suggested as the physical reason for the difference.
Acknowledgment. The author wishes to thank Dr . J.
Mort for valuable discussions on the subject of this work.
References
1. R. M. Schaffert, Photogr. Sci. Eng . 6, 197 (1962).
2. R. M. Schaffert, Electrophotography, Focal Press, London, 1975.
3. M. Omodani, M. Fujita, M. Ozawa, and M. Ohta, IS&T’s NIP13: Intl.
Conference on Digital Printing Technologies , IS&T, Springfield, VA,
1997, p. 820
4. Y. Niv, IS&T’s 10th Intl. Congress on Advances in Non-Impact Printing Technol ., IS&T, Springfield, VA, 1994, p. 196
5. G. Bartscher and J. Breithaupt, J. Imaging Sci. Technol . 40, 441
(1996), and references therein.
6. I. Chen, J. Imaging Sci. Technol. 39, 473 (1995).
7. I. Chen, J. Mort, M. A. Machonkin, J. R. Larson, and F. Bonsignore, J.
Appl. Phys . 80, 6796, (1996)
8. I. Chen, Photogr. Sci. Eng. 26, 153 (1982).
9. E. M. Williams, The Physics and Technology of Xerographic Processes, John Wiley and Sons, New York, 1984.
10. L. B. Schein, Electrophotography and Development Physics, SpringerVerlag, Berlin, 1988.
Vol. 43, No. 3, May/June 1999
273
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
A Study of Non-Uniform Charging by Charging Roller with DC Voltage
M. Kadonaga, T. Katoh and T. Takahashi
Research and Development Center, Ricoh Company, Ltd., Yokohama, Japan
Model experiment and numerical simulation are carried out in order to make clear the generating mechanism of non-uniform
charging by charging roller with DC voltage. From the experiments using metal roller and polyethylene terephthalete (PET) sheet ,
the periodic charging patterns can be recognized on the PET sheet by cascade development with toners after the sheet is charged
. The
period and the size of the patterns become large as the applied voltage increases. The shapes of the patterns are different bet ween
positive and negative charging. The characteristic of the patterns is similar to that of separating discharges on anelectrified insulating sheet. The results show that the charging patterns are made as follows. (1)Abnormal discharge occurs between the charging
roller and the PET sheet, and large amount of charge is deposited on the sheet. (2) The charge is forced to move due to surface
discharge. On the other hand, two-dimensional electrostatic simulation of the charging roller is carried out with considerationof
the surface discharge as well as the abnormal discharge. From the simulation, the periodic charging patterns can be generated
on the PET sheet and may verify the new model of generating non-uniform charging patterns as proposed above.
Journal of Imaging Science and Technology 43: 274–279 (1999)
Introduction
A charging roller system is one of the contact charging
devices of electrophotographic machines. Recently it has
become popular because of extremely less ozone emission than corona charging devices.1 Figure 1 shows a schematic diagram of the charging roller system with DC
voltage. The system consists of the photoconductor(OPC),
the charging roller and the DC power supply . The electric discharge happens between the roller and the OPC,
and thus the surface of the OPC is charged. It is difficult to obtain uniform charging by roller with DC voltage and the periodic charging patterns can often be
observed on the OPC, especially by a single-layer roller
with low resistance. Because the periodic charging patterns degrade the quality of printing image, they must
be eliminated and uniform charging is desired.
In order to obtain the uniform charging, AC voltage
is often superimposed on DC voltage. However , with
DC+AC voltage, a great deal of discharge happens in
the vicinity of the nip between the OPC and the roller.
Ozone is generated by discharge and it degrades the
OPC which is very sensitive to the ozone.
When various kinds of proper material are used for
the elastic layer and the surface layer of the charging
roller, non-uniform charging can be suppressed even
though with only DC voltage the OPC degradation rarely
happens due to less ozone emission. However, the generating mechanism of such non-uniform charging patterns on the OPC is not still clear . In order to make
Original manuscript received June 9, 1998
© 1999, IS&T—The Society for Imaging Science and Technology
274
clear the generating mechanism of non-uniform charging by charging roller with DC voltage, model experiment and numerical simulation are carried out.
Objective
The objective is to make clear the generating mechanism of the non-uniform charging patterns by a charging roller with DC voltage.
Experiments
Experimental Setup. Experimental setup model of the
charging roller is constructed as shown in Fig. 2. The
metal roller is used due to its extremely low resistance,
so as to obtain large and clear periodic charging patterns. PET (polyethylene terephthalete) sheet with 25
micron thickness, back coated with aluminum, is used
instead of OPC because of its similar electrostatic char-
DC Power
Supply
Vp
Surface Layer
Elastic Layer
Center Shaft
Charging
Roller
OPC
Figure 1. Schematic diagram of the charging roller system
with DC voltage.
Roller Holder
Metal Roller
PET Sheet
X-Stage
Moving Direction
Vp DC
Figure 2. Schematic diagram of the experimental setup.
Experimental Results. A series of non-uniform charging patterns for various negative DC voltages is shown
in Fig. 3. The PET sheet moves upwards at the speed of
2.4mm/s as the arrow shows in Fig. 3(a). The diameter
of the roller (D) is 12mm. The generating pattern looks
like a stripe or a round shape. The size (width) and the
period of the pattern in the moving direction becomes
large as the applied voltage (Vp) increases. Along the
length of the roller , non-uniformity also exists. When
Vp is small, discharge may happen simultaneously along
the length of the roller and the patterns are stripes. On
the other hand, when Vp is large, discharge may happen one by one, and the patterns become isolated and
are deposited in interleaving parallel rows. As the applied voltage increases, the discharge gap becomes
longer and the discharge may become more unstable.
The experiments are also carried out under the condition that VL = 40 mm/s and D = 40 mm, and the re sults
resemble those shown in Fig. 3. As the roller is a conductor and the PET sheet is an insulator, discharge phenomenon in this experiment may be independent of VL
and D. The discharge gap is not effected much even
though the roller diameter is changed when the applied
voltage is the same. The results shown in this paper
are all carried out under the condition thatVL = 2.4 mm/
s and D = 12 mm.
By the conventional electrostatic surface potential
meter, TREK MODEL344, such non-uniform charging
potential distribution is not able to be measured. If the
conventional meter is used, only average potential value
Figure 3. The series of the non-uniform charging patterns for
various negative DC voltages.
Surface Potential (V)
acteristic. The PET sheet is set on thex-stage and moves
at the speed VL under the roller biased with DC voltage.
The roller is rotated simultaneously as the PET sheet
moves due to the surface friction. After the PET sheet
is charged, the non-uniform charging patterns can be
observed by cascade development with toners. When
negative voltage is biased on the roller, positive toners
are used to develop, and vice-versa.
On the other hand, the surface potential distribution,
before development with toner , is measured by two
methods. The first method is by the conventional electrostatic surface potential meter (TREK MODEL 344,
Medinam, New York, USA), whose resolution is more
than 1 mm. The second method is the high resolution
electrostatic surface potential meter using a scanning
electrostatic force microscope,2 whose resolution is about
30 µm.
Applied Voltage(V)
Figure 4. The relationship between the applied voltage and average potential value of the PET sheet with negative charging.
is obtained because of its course resolution. Figure 4
shows the relationship between the applied voltage and
the average surface potential obtained by the TREK
MODEL344. The straight line in Fig. 4 shows the ideal
charging characteristic deduced from Eq. 1.3 The experimental results are deviated from the ideal charging
characteristic.
Vs = Vp – 312 – 6.2(d/ε′) – (7737.6d/ε′)1/2
where
A Study of Non-Uniform Charging by Charging Roller with DC Voltage
d:
ε′:
Vp:
Vs:
(1)
Thickness of the PET sheet [m]
Relative dielectric constant of PET
The applied voltage [V]
The ideal surface potential value [V]
Vol. 43, No. 3, May/June 1999
275
Surface Potential (V)
Figure 5. The potential distribution of the PET sheet along the moving direction in the case of negative charging.
(a) Vp = +1000V
(b) Vp = +1200V
(c) Vp = +1400V
Figure 6. The series of the non-uniform charging patterns for
various positive DC voltages.
Figure 5 shows the potential distribution along the
moving direction measured by the high resolution electrostatic surface potential meter. The surface potential
is measured at intervals of 32 µm and it is enough to
show the non-uniform charging distribution corresponding to Fig. 3. It is recognized from Figs. 3, 4 and 5 that
unstable discharge may happen in the gap between the
roller and the PET sheet.
Figure 6 shows a series of the non-uniform charging
patterns for various positive DC voltages. The pattern
looks like a tree-like shape with fine notches at the lower
edge of each pattern and is completely different from
that of negative charging. However , the characteristic
of the patterns are similar to that of separating discharge patterns 4 and the Lichtenberg’ s figures. 5 Compared with these discharges, charge on the PET sheet
may be forced to move because of surface discharge.
From these experimental results, two effects should
be considered in modeling the generating mechanism
of non-uniform charging patterns. The first effect should
be the abnormal discharge between the roller and PET
sheet. The second effect should be the surface discharge
on the PET sheet.
276
Journal of Imaging Science and Technology
Non-Uniform Charging Model
A new model for generating the non-uniform charging
patterns along the moving direction is proposed. As the
non-uniformity exists in two-dimensional, three-dimensional consideration is desired. Such non-uniformity in
both directions may be due to abnormal discharge. If
non-uniformity in the moving direction can be eliminated, non-uniformity along the length of the roller will
disappear simultaneously. Therefore, this article focuses
on the non-uniform charging along the moving direction. Figure 7 shows the time sequence of this model
near the entrance of the nip between the roller and the
PET sheet. First, strong discharge occurs in the large
gap between the roller and the PET sheet. Large
amounts of charge are deposited on the PET sheet by
abnormal discharge [Fig. 7(a)]. The potential of the PET
sheet, where the charge is deposited, is so high that the
surface discharge takes place [Fig. 7(b)]. Because the
potential of the discharge point on the PET sheet remains high, following discharge cannot take place for a
while [Fig. 7(c)]. As the sheet moves and the highly
charged part is far enough from the discharge point, the
potential of the sheet at the same position decreases
and then discharge occurs again [Fig. 7(d)]. 6 The same
steps repeat successively, and finally non-uniform charging patterns can be generated on the PET sheet.
Numerical Simulation
Simulation Model. Even though numerical simulations of one-dimensional analysis for the charging roller
are very popular,3,7,8 they cannot simulate this new model
of generating non-uniform charging patterns. In order
to certify this new model, two-dimensional simulation
of the charging roller is carried out according to the following four steps.
Step 1: Calculation of the electrical field to obtain
the potential distribution around the roller and
the PET sheet. Boundary fitting coordinate
mesh is used to fit the roller shape. Poisson’ s
equation is solved by the finite difference
method considering the movement of the PET
sheet.
Step 2: Consideration of the discharge between the
roller and the PET sheet. Using Paschen’ s law for
the breakdown voltage (Vpa= 312 + 6.2 × 106 • g),
the discharge gap can be derived. In the case of
normal discharge, when the discharge occurs at
the gap g[m], the amount of charge ( dQ)
Kadonaga, et al.
Charge Density (C/m2)
Discharge Gap (µm)
Figure 8. The relationship between the discharge gap and discharge density in negative charging.
slightly larger than Vpa and it cannot stop when Vg becomes equal to Vpa. The voltage drop ofVg becomes larger
than that of normal discharge because of large amount
of dQ. This equation is only empirical, but the large
value of dQ is deduced from Eqs. 2 and 4 when the discharge gap g is large.
∆V = Vg – β • g2
where
d:
g:
ε′:
ε0:
β:
Vg:
∆V:
Figure 7. Model of generating non-uniform charging patterns.
deposited on the sheet is estimated from Eqs. 2
and 3.9 Equation 3 indicates that the discharge
occurs when the voltage across the gap ( Vg) is
slightly larger than V pa and stops when Vg
becomes equal to Vpa.
dQ = (d + ε′
•
g)
•
ε0
•
∆V/(d
•
g)
∆V = Vg – (312 + 6.2 × 106 • g)
(2)
(3)
On the other hand, in the case of abnormal discharge,
Eq. 4 may be used instead of Eq. 3. This equation and
the value of β are determined so as to fit the experimental results shown in Fig. 8 explained later . Equation 4
indicates that abnormal discharge occurs when Vg is
(4)
Thickness of the PET sheet [m]
Discharge gap [m]
Relative dielectric constant of PET
Dielectric constant of air
The parameter for the abnormal discharge
Voltage across the discharge gap
Voltage drop of Vg due to the discharge
Figure 8 shows the relationship between the discharge
gap g and discharge density by one abnormal discharge.
The values of triangles are estimated from Fig. 4 with
the assumption that the charge would be deposited in
one mesh by one abnormal discharge. The solid line in
Fig. 8 shows the relationship obtained with Eqs. 2 and
4 when the parameter for the abnormal dischargeβ is 1
× 10 12. The value of β may depend mainly on the material and the resistance of the roller.
Step 3: Consideration of the surface discharge. The
electrical field after the abnormal discharge is
recalculated because the abnormal discharge
disturbs the electrical field. Next, the electrical field strength E is estimated at every point
on the sheet. If the value of the electrical field
strength is larger than the Elimit, which is the
breakdown field strength of the surface discharge, surface discharge may happen and
charge (dq) is forced to move in the direction
according to the electrical field around the
point.
dq/dt = α • E (in case of E > Elimit)
(5)
In this study, Elimit= 3.5 × 10 6[V/m] and α = 4 × 10 –7 are
used. The value of Elimit is determined in order to simulate the pitch of the patterns as close as possible to the
A Study of Non-Uniform Charging by Charging Roller with DC Voltage
Vol. 43, No. 3, May/June 1999
277
Figure 9. Uniform charging (Vp =
1600 V).
Figure 10. Non-uniform charging.
experimental result. The parameter α acts something
like a resistivity against the surface discharge current.
The value of α has no effect on the final discharge pattern if it is sufficiently small, because Step 3 should be
repeated until surface discharge does not occur at any
point on the PET sheet. After Step 3, the electrical field
at time T is obtained.
Step 4: Time is advanced (T = T + dt). The PET sheet
moves forwards, then back to Step 1 and repeats the same steps successively.
278
Journal of Imaging Science and Technology
Simulation Results
Numerical results are shown in Figs. 9 and 10 that indicate the charge distribution on the PET sheet in gray
scale. Figure 9 shows the uniform charging if the abnormal discharge does not occur (using Eqs. 2 and 3).
This is the ideal charging by the charging roller. Figure
10 shows the non-uniform charging patterns when the
abnormal discharge and the surface discharge take place
(using Eqs. 2, 4 and 5). The bright portions indicate
highly charged places. (a) is the case of Vp = 1000 V, (b)
Vp = 1200 V, (c) Vp = 1400 V (d) Vp = 1600 V and (e) Vp =
Kadonaga, et al.
Charge Density (C/m2)
X (mm)
Figure 11. The charge distribution on the PET sheet obtained
by calculation.
1800 V. Non-uniform periodic charging patterns can be
recognized on the PET sheet. Figure 11 shows the charge
distribution on the PET sheet for various applied voltages. Figure 12 shows the comparison of the charging
characteristic result between experiment and calculation. The calculation result shows good agreement with
the experimental result. The comparison of the period
of the patterns between the experiment from Fig. 3 and
calculation from Fig. 10 is shown in Fig. 13. The period
of calculation shows smaller value than that of experiment. This discrepancy is probably due to the differ ence of the positioning of the patterns between 2-D and
1-D. In Fig. 3, patterns are deposited in interleaving
parallel rows and the period may become larger than
that in one-dimensional even parallel rows. Even though
there is discrepancy between experiment and calculation, size and period of the charging patterns become
large as applied voltage is increased. This calculation
result may certify the new model of generating non-uniform charging patterns proposed in this article.
Conclusion
The experiment to clarify the mechanism of the nonuniform charging by charging roller with DC voltage is
carried out together with the numerical simulation. The
abnormal discharge and the surface discharge on the
OPC may cause the non-uniform charging patterns. Considering the abnormal discharge and the surface discharge, the relationship between the pattern and the
applied voltage is simulated numerically and shows good
agreement with the experimental result.
Figure 12. The comparison of the charging characteristic result between experiment (negative charging) and calculation.
Figure 13. The comparison of the period of the patterns between experiment (negative charging) and calculation.
References
1. J. Araya, N. Koitabashi, S. Nakamura, and H. Hirabayashi, US Patent
5,164,779 (1992).
2. J. Takahashi and T. Katoh, T. IEE Japan, 117-E, 594 (1997)
3. S. Nakamura, H. Kisu, J. Araya, and K. Okuda, Electrophotography
30, 302 (1991). ( in Japanese )
4. X. Ji, Y. Takahashi, Y. Komai, and S. Kobayashi, J. Electrostatics 23,
381 (1989).
5. F. H. Merrill and A. von Hippel, J. Appl. Phys. 10, 873 (1939)
6. H. Hirakawa and Y. Murata, Proc. 1995 Annual Meeting of The Institute of Electrostatics Japan, Tokyo, Japan, 1995, p.133. (in Japanese)
7. H. Kawamoto and H. Satoh, J. Imaging Sci. 38(4), 383 (1994).
8. M. Kadonaga, Proc. Japan Hardcopy’ 95, The Society of Electrophotography of Japan, Tokyo, Japan, 1995, July, p. 55. (in Japanese)
9. R. M. Schaffert, Electrophotography, Focal Press, Stoneham, MA,
1975.
A Study of Non-Uniform Charging by Charging Roller with DC Voltage
Vol. 43, No. 3, May/June 1999
279
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Silsesquioxane Sol-Gel Materials as Overcoats for Organic Photoreceptors
D. S. Weiss,▲ W. T. Ferrar, J. R. Corvan, L. G. Parton, and G. Miller
Office Imaging Division, Eastman Kodak Company, Rochester, New York
Organic photoreceptors for use in electrophotographic processes may be overcoated to impart wear and scratch resistance, to protect
the surface from corona generated chemicals, and to improve the efficiency of electrophotographic process steps, such as toner
transfer
and cleaning. In this article, we will discuss the use of silsesquioxanes prepared by the sol-gel process as photoreceptor over coats.
Included in our discussion will be the chemistry and procedures involved in overcoat fabrication, the methods used to determine the
effects of the overcoat on the physical and electrophotographic characteristics of the photoreceptor
, and the effects of the silsesquioxane
chemical structure on these characteristics.
Journal of Imaging Science and Technology 43: 280–287 (1999)
Introduction
Organic photoreceptors are comprised of one or more
polymer-based thin layers coated on a substrate. The
substrate is typically an aluminum drum or a flexible
polymeric web with a conductive metal coating for a
ground connection. 1 For many applications, the
photorecptor architecture has separate charge generation (CGL) and hole transporting charge transport layers (CTL). When a negative surface potential is desired
the charge generation layer is adjacent to the conductive material. For positive charging applications the
positions of the CTL and CGL may be reversed. These
layers are usually prepared by solvent coating technologies such as dip or ring coating for drums and hopper
coating for webs.
In the electrophotographic process the photoreceptor
is subjected to a variety of physical and chemical abuses
that may determine its productive lifetime. The surface
of an organic photoreceptor surface is relatively soft so
that cleaning, whether by blade or brush, causes scratching and abrasive wear . Unintended contact of the sur face with hard objects, such as staples or paper clips, may
result in scratches that necessitate immediate photoreceptor replacement. The photoreceptor surface is also
relatively permeable, and its components are reactive toward the ozone and nitrogen oxides generated during
corona charging. Thus, after extended exposures to these
chemicals the electrophotographic characteristics may degrade to the point where image defects become objectionable and the photoreceptor must be replaced. Because of
these and other factors, the lifetime limit of organic photoreceptors is on the order of a hundred thousand imag-
Original manuscript received November 9, 1998
▲ IS&T Member
© 1999, IS&T—The Society for Imaging Science and Technology
280
ing cycles as contrasted with a million or more obtained
with the much harder amorphous silicon and arsenic
triselenide photoreceptors. Thus, there have been extensive efforts over the years to make organic photoreceptors less susceptible to these undesirable effects. One
approach has been to overcoat the photoreceptor surface
with a material that is tough, chemically impervious, and
inert. Many materials have been developed, but relatively
few have actually been commercialized. Some examples
are electrically insulating overcoats from organic or silicon 2 based polymers; somewhat conducting overcoats
from organic polymers doped with charge transport materials or semiconductive particles; very thin overcoats
of refractory materials such as diamond-like carbon (Diamond 4 from HDS Inc.) or aluminum nitride;3 and somewhat insulating overcoats of inorganic glassy materials
(Ultrashield from Optical T echnologies Corp.4). In this
article, we will discuss examples of the latter in which
the overcoat is a silsesquioxane organosilicone polymer
prepared by the sol-gel process. In a previous paper 5 we
discussed some of the history , chemistry, and development of silsesquioxane coatings. In this article we will
review and expand on this material and follow with discussions of how such photoreceptor overcoats may be
characterized, and how the silsesquioxane chemical composition affects these characteristics as they relate to electrophotographic performance.
Silsesquioxane Polymers
Silsesquioxanes are silicon-based polymers where the
monomer unit has the structure R−SiO1.5. The monomer
structure may also be pictured as R −Si(O~)3 where the
oxygens are bonded to silicon atoms of other monomer
units to produce a highly crosslinked polymeric structure. The physical characteristics of polysilsesquioxanes
are a combination of those of silica glass and organic polymers, and as such , are ideally suited for use as protective overcoats for photoreceptors. Silica glasses are
usually prepared by high temperature melt processes.
Figure 1. Preparation of a silsesquioxane via the sol-gel process involves hydrolysis and partial polymerization of trialkoxysilanes
to form a sol, followed by thermal curing to produce a highly crosslinked gel.
However, silsesquioxanes can be prepared by the sol-gel
process6 at relatively low temperatures. Figure 1 shows
the relevant chemistry. In this process, the precursor
monomer R−Si(OH)3, is allowed to partially polymerize
to produce a colloidal suspension or sol of hydroxylated
silsesquioxane that can be stored for subsequent use. In
practice, the monomer is typically produced in-situ by
the hydrolysis of an alkyltrialkoxysilane, R −Si(OR’) 3.
When desired, the sol is thermally cured whereupon extensive crosslinking occurs in the gel phase of the reaction. Silsesquioxane polymers prepared in this manner
are sometimes called sol-gels.
The use of the relatively low-temperature sol-gel process to prepare thin polymeric overcoats stems from two
similar US patents, H. A. Clark to Dow Corning Corp.
(1976)7 and R. B. Frye 8 to the General Electric Corp.
(1978). Principal applications are as abrasion resistant
overcoats for plastic items such as lenses. In the Dow
Corning process an acidic dispersion of colloidal silica
is reacted with hydroxylated silsesquioxane. The latter
is generated in-situ by the aqueous hydrolysis of an
alkyltrialkoxysilane. Methyl is the preferred silicon substituent as it produces the toughest coatings. A typical
procedure is to mix glacial acetic acid with aqueous
acidic colloidal silica (15 nm particles) followed by the
addition of methyltrimethoxysilane. The final pH is
adjusted to 3–6. This mixture undergoes partial condensation for several days and constitutes the “sol” phase
of the procedure. The use of stronger acids significantly
shortens the shelf life. Gelation is catalyzed by the addition of salts (sodium acetate or benzyltrimethylammonium
acetate) and/or excess acid followed by preparing the
coating and curing the overcoated material at elevated
temperatures (50–150°C). The General Electric process
differs in that the hydrolysis and partial condensation
of the colloidal silica and alkyl or aryltrialkoxysilane is
carried out under weakly basic conditions.A typical procedure is to add an aqueous low-alkali silica dispersion
such as DuPont’ s Ludox LX to a solution of the
trialkoxysilane. After 1 day at room temperature at a
pH > 7.2, hydrolysis is complete and partial condensation has occurred. The sol is stabilized and the %solids
adjusted by the addition of iso-butyl alcohol. The sol is
coated onto a substrate with or without the addition of
a flow control agent (such as a polysiloxane polyether
copolymer, General Electric SF-1066 or Dow Corning
DC190, for example), which also reduces stress induced
cracking in the final coating. In addition, a salt catalyst such as sodium acetate may be added to accelerate
curing of the coated sol. Without catalyst the final curing is carried out for 1 h at 120°C.
The physical, chemical, and electrical characteristics
of silsesquioxane polymers can be controlled in many
ways including the chemical structure of the R group,
the incorporation of copolymers or polymer blends, the
addition of filler particles which may or may not
crosslink into the polymer backbone, the addition of
salts, the variation of the cure conditions and catalyst
to influence the extent of crosslinking, and the overcoat
thickness.
Silsesquioxane Overcoated Photoreceptors
The physical, chemical, photo- and dark-electrical characteristics of overcoated photoreceptors must satisfy the
requirements of the process for which they are intended
while increasing the process lifetime of the photoreceptor. A successful overcoat must balance several characteristics. The overcoat must be tough enough to
minimize wear and scratches, but not be so brittle that
cracking occurs. The bulk conductivity must be high
enough to avoid an undesirable residual surface potential in the electrophotographic process, but not be so
high that undesirable spreading of the electrostatic latent image occurs. The surface must also be highly insulating to avoid unwanted spreading of the latent
image charge9 as might occur from the accumulation of
surface salts with liquid development 10 or corona generated chemicals.11,12 Furthermore, the overcoat characteristics must remain favorable over the range of
temperatures, humidities, corona exposures, and cleaning conditions experienced in the electrophotographic
process.
Examples of silsesquioxane overcoated photoreceptors
may be found in the patent literature. Schank (Xerox
Corp.) reported the use of Dow Corning Vestar Q9-6503
and General Electric SHP-1000 and 1010 in the fabrication of silsesquioxane overcoats of crosslinked siloxanol
and colloidal silica. 13 An acrylic polymer, General Electric SHP-200 was used to form an adhesive interlayer
Silsesquioxane Sol-Gel Materials as Overcoats for Organic Photoreceptors
Vol. 43, No. 3, May/June 1999
281
TABLE I. Overcoat Properties and the Methodologies Used in
Their Determination
Crosslinking
Hardness
Brittleness
Scratch resistance
Surface conductivity
Bulk conductivity
Chemical permeability
29Si NMR: T 2 (–60 ppm)/T 3 (–70 ppm)
Nanoindentation
ANSI wedge brittleness test
Diamond stylus scratch with AFM imaging
Electrostatic image spreading (RH dependence)
Electrophotographic residual potential
(RH dependence)
Corona gas exposure and electrical-only
electrophotographic cycling (RH dependence)
and ammonia gas was used as the condensation catalyst. Overcoat thicknesses less than 0.5 µm were difficult to apply and those > 5 µm had a tendency to crack
and were difficult to cure. A later patent suggested the
incorporation of a small proportion (2–10%) of quater nary ammonium salts, (RO)3Si(CH2)3NR’ 4+ Cl- monomer
unit, to act as a catalyst and to increase the overcoat
conductivity to allow for increased overcoat thickness.14
The preferred major component was methyltrialkoxysilane for maximum hardness and it was suggested that
resins added as plasticizers and lubricants were beneficial. Further refinements were to incorporate monomer
units where the Si−alkyl substituent contained an electron acceptor moiety such as nitrile or chlorene.15 In the
absence of ammonia, typical curing conditions were
100–140°C. In a patent assigned to the Japan Atomic
Energy Research Institute, Shindengen Electric Manufacturing Co., and Yamanashi Electronics Co., improvements were made to reduce the humidity sensitivity of
the conductivity.16 The key feature was to incorporate
component alkyltrialkoxysilane mono mers such that
neither acid or catalyst salts were necessary to effect
crosslinking. A typical formulation consisted of adding
water to a mixture of methyltrimethoxys ilane, β-(3,4epoxycyclohexyl)-ethyltrimethoxysilane, and γ-aminopropyltriethoxysilane to effect the hydrolysis. A coating
solution was made by the addition of excess ethanol. The
solution was coated onto an organic photoreceptor and
cured at 80°C for 1 h.
We have investigated Ultrashield (Optical T echnologies Corp.) as well as our own formulation overcoats on
Kodak photoreceptors. For the latter, a typical procedure was to add glacial acetic acid and water to the
mixture of trialkoxysilanes and stir for 24 h.17 The sol
is prepared by dilution with ethanol to about 20% solids followed by stirring for approximately 1 week. Just
prior to coating, addenda such as surfactants, lubricants, plasticizers, salts, etc., are introduced. Thesubstrate photoreceptor is commonly first overcoated with a
thin (0.1-0.5 µm) polymeric primer layer such as
poly(methacrylate-co-methylmethacrylate-co-methacrylic
acid). The silsesquioxane dope is coated on the photoreceptor at 10 ft/min with ramped heating to 90°C and then
cured at 80°C for 24 h to complete the gel process. Overcoats were prepared in the range of 1-5-µm thickness and
the resulting photoreceptor package evaluated as described below.
Results and Discussion
Overcoat Characterization. In preparing overcoated
photoreceptors, the factors that we controlled were composition, cure conditions, and thickness. Table I shows
some overcoat characteristics that are relevant to electrophotographic performance and gives a brief indication of how each characteristic was quantified. Details
282
Journal of Imaging Science and Technology
TABLE II. Some Monomer Units Used in the Formulation of Silsesquioxane Overcoats for Organic Photoreceptors and the
Overcoat Characteristics that are Affected
Methyltrimethoxysilane
Propyltrimethoxysilane
3-Aminopropyltrimethoxysilane
3-Glycidoxypropyltrimethoxysilane
Lithium iodide
Hardness and scratch resistance
“Organic” character-plasticity
Extent of cure and bulk conductivity
Crosslinking and bulk conductivity
Bulk conductivity
of these techniques and some representative examples
will be presented. In Table II we show some commonly
used compositional elements and the overcoat characteristics that they influence. Other addenda such as
lubricants, plasticizers, crosslinking agents, and fillers
will also affect the overcoat characteristics.
Crosslinking. The extent of condensation of each Si
in the polymer matrix is defined as “ Tx” where “ x” is
how many of the three “−OH” units of the monomer, RSi(OH)3, have been converted into “−O−Si−” crosslinks. 18
The method of choice for this determination is29Si NMR.
In sol formation monomer hydrolysis occurs followed by
condensation. The extent of condensation increases as
a function of time and temperature. Figure 2 shows the
high resolution spectrum of a sol in ethanol/water aged
for three weeks at ambient temperature. Resonances for
T1 (–52 ppm), T 2 (–59 ppm), and T 3 (–70 ppm) Si are
observed. No resonance is observed for noncondensed
Si (T0). For this sol the fraction of Si-OH moieties which
have undergone condensation is 0.83. Solid state 29Si
NMR is used for analysis of the coated and cured solgel. Figure 2 also shows the solid state spectrum of a
sample with an overcoat cured at 80°C for 24 h. Only T 2
(–58 ppm) and T3 (–68 ppm) are observed (condensation
fraction of 0.92). To characterize the extent of crosslinking in cured samples we use the T2/T 3 ratio. In general,
the extent of crosslinking was found to increase with
increasing temperature and increasing overcoat thickness: T 2/T3: 0.38 (1 µm), 0.35 (3 µm), 0.30 (5 µm) for a
material with a propyltrimethoxysilane/methyltrimethoxysilane ratio of three. The extent of crosslinking was
reduced when the propyl/methyl ratio was increased.
The thickness effect may occur because the thicker overcoats retain solvent for a longer time than thinner over
coats allowing more extensive crosslinking. The effect
of the propyl organic substituent may be to sterically
retard polymerization.
Hardness. Hardness was determined by nanoindentation. Qualitative measurements were carried out by
measuring the residual indentation depth (nm) in the
photoreceptor surface after contact with a pyramidal indenter under a specified load. The effects of overcoat
thickness and curing on hardness and crosslinking are
shown in Figure 3 for the Ultrashield material. With an
8 g load, a nonovercoated photoreceptor had an indentation depth of 2.8 µm. This decreased to 2.3 µm with a
7 µm uncured overcoat and to 1.5 when the overcoat
was cured. The extent of crosslinking decreased with
thickness (as mentioned above) and increased with curing. We found that in practice, measurements of scratch
resistance and brittleness were more likely to correlate
with actual photoreceptor damage occurring in the electrophotographic process.
Weiss, et al.
Brittleness. Brittleness was determined by the
American National Standards Institute PH 1.31 Brittleness of Photographic Film, Method B, “W edge Brittleness Test”. Film samples were equilibrated at 70°F and
15% RH for 24 h. The ends of a test strip were inserted
(coated side out) through a “wedge” that consisted of a
metal block with a large opening (1 in) on one side tapering to a narrow slot (0.06 inch opening) on the other
side. One of the film ends was clamped to the wedge
near the narrow opening and the other end brought
through the narrow opening to create a film loop in the
large opening. To carry out the test, the free film end
was rapidly pulled through the wedge. Thus, the diameter of the loop rapidly decreases as the film end is
pulled. The film was then examined visually for over coat cracking and the diameter of the loop in inches at
failure was recorded. The larger the diameter the more
brittle the sample.
Figure 2. 29Si NMR spectra of silsesquioxane sol in ethanol/
water [top spectrum T 1 /T 2 / T 3 (0.06/0.81/1.00)] and
silsesquioxane gel [bottom spectrum] T 2/T3 (0.34/1.00)].
Scratch Resistance. Photoreceptors were scratched
with a diamond stylus (4-8 g load is typical) and the
scratches visualized with AFM. Figure 4 shows the
AFMs of scratches produced on an overcoated and
nonovercoated photoreceptor with an 8 g load. The overcoat in this test was 3 µm thick and had a 60% propyl
content. At this loading, overcoat cracking is observed.
The load appropriate to the stresses the photoreceptor
will actually experience in the electrophotographic process is what should be used in this test. T o quantify
the scratch depth, the peak-to-valley heights at three
positions along the AFM image of the scratch are averaged. Figure 5 shows the scratch data for the
overcoated and nonovercoated photoreceptors in Fig .
4 as a function of loading. W ithout an overcoat the
scratch depth is large and increases with load. The
overcoated photoreceptor shows little scratching at the
4 and 6 g loads but catastrophic cracking at the 8 g
load. At the high load, the scratch depth of the
overcoated and nonovercoated photoreceptors are the
same and the overcoat offers no scratch protection. Figure 6 shows scratch (4 g load) and brittleness data for
photoreceptors with 4 µm overcoats with increasing
amounts of n-propyltrimethoxysilane relative to
methyltrimethoxysilane. Brittleness, plotted as the reciprocal of the diameter (inches) at failure, decreases
gradually with increasing propyl content while the
Figure 3. Effects of curing (80°C for 2
h) and overcoat thickness on the hardness (residual indentation with an 8 g
load) and crosslinking (29Si NMR) of the
silsesquioxane overcoat.
Silsesquioxane Sol-Gel Materials as Overcoats for Organic Photoreceptors
Vol. 43, No. 3, May/June 1999
283
80
60
S. D.
40 µm
µm
8.58
4.29
0.00
20
80
60
40
µm
0
20
0
Figure 5. Scratch depths (peak-to-valley) for the samples
shown in Fig. 4 with 4, 6, and 8 g loads on the stylus.
80
60
S. D.
40 µm
µm
8.58
4.29
0.00
20
80
60
40
µm
0
20
0
Figure 4. AFMs of scratches produced on an overcoated (3
µm overcoat with 60% propyl content) (top) and nonovercoated
(bottom) photoreceptor with a 2.5 µm diamond stylus with
an 8 g load.
scratch depth is low and relatively invariant except at
100% propyl content.
Bulk Conductivity. The bulk conductivity of the overcoat will affect the electrophotographic characteristics of
the photoreceptor. If highly resistive , a residual potential will be observed in charge/expose electrophotographic
cycling. If highly conductive , there will be no residual
potential but latent image degradation can occur . This
will be discussed in the following section. T o determine
the electrophotographic residual potential overcoated
photoreceptors were exercised for 5k of charge/expose/
erase cycles. Figure 7 shows the results of a study on the
effects of overcoat composition (0% and 60% propyl) and
thickness on brittleness, residual potential, and cure.
Overcoats with 0% propyl are highly cured at all thicknesses, brittleness increases with thickness, as does the
residual potential. The latter indicates the overcoat is
highly resistive. When the overcoat contains 60% propyl
the characteristics are very different. These overcoats are
much less cured and the cure increases with thickness.
Furthermore, brittleness and the residual potentials are
low and independent of thickness.
Surface Conductivity. Surface conductivity is determined with the lateral image spreading technique described previously.9 The photoreceptor is corona charged
and then exposed through a 2.5 mm slit to produce a
284
Journal of Imaging Science and Technology
Figure 6. Effect of % propyl content on scratch resistance
(peak-to-valley scratch depth in microns) and brittleness (reciprocal of the diameter of the circle in inches, where overcoat
cracking is first observed). Data is for a 4 µm overcoat.
“square well” surface potential pattern. The latent image shape is recorded by moving the photoreceptor past
a high-resolution surface voltmeter probe. The image
shape is determined as a function of time and the temporal characteristics are then related to either a “surface
resistance” or an image width determined at a specified
time after exposure. Figure 8 shows the time dependent
changes in latent image shape for an overcoated photoreceptor. This data can be fit to the theory with an effective surface resistance of 3× 1015 Ω/square. Figure 9 shows
the effects of humidity and corona gas exposure on the
effective surface resistivity. At elevated humidity and
with corona exposure, the image spreads much faster .
We believe this is due to the production of acid by the
interactions between the nitrogen oxide gases produced
in the corona and water present in the atmosphere and
Weiss, et al.
Figure 7. Effects of silsesquioxane overcoat composition and thickness on brittleness (reciprocal of the diameter of the circle in
inches, where overcoat cracking is first observed), residual potential (ratio of the residual surface potential after 5 k electricalonly cycles to the initial surface potential), and cure (T 2/T 3 by solid state 29Si NMR).
Figure 8. Electrostatic image shape
(normalized surface potential versus
position) for an overcoated photoreceptor. The exposure (centered at 0 cm)
was through a 2.5 mm slit. The image
shape was determined at 5, 150, 600,
and 1800 s after the exposure.
Silsesquioxane Sol-Gel Materials as Overcoats for Organic Photoreceptors
Vol. 43, No. 3, May/June 1999
285
Figure 9. Effective surface resistivity as monitored by timedependent changes in the width of the electrostatic image obtained by exposing the photoreceptor through a 2.5 mm slit
(see Fig. 8). The image width was obtained as the distance
(mm) between tangents drawn to the image edges at the sur face potential in the unexposed regions 100 s after the exposure. Corona exposures were obtained by positioning the
sample in front of the corona charger and exposing it to the
effluent gasses for the indicated times. This was carried out
at the indicated relative humidity.
on the photoreceptor surface. This acid makes the photoreceptor surface conductive.
Chemical Permeability. In addition to influencing the
surface conductivity some photoreceptors are damaged
upon exposure to corona produced gases. 19 In this test,
photoreceptors are exposed to the gases from a corona
charger during electrical-only electrophotographic cycling. Figure 10 shows that the photoreceptor without
the overcoat exhibits decreased charge acceptance. However, the overcoated photoreceptor (Ultrashield from
Optical Technologies Corp.) is completely stable under
the same conditions.
Conclusions
The formulation and fabrication of silsesquioxane over coats on organic photoreceptors is a complex process. Fur
thermore, the desired overcoat characteristics must be
determined with reference to the electrophotographic
process and the temperature and humidity conditions,
for which the photoreceptor is intended. In this article,
we have described testing methods for several important
overcoat characteristics: hardness, brittleness, scratch resistance, surface conductivity, bulk conductivity, and
chemical permeability; and, we have presented representative results to show how these characteristics are influenced by formulation composition, coating and curing
conditions, and overcoat thickness. Space limitations
have prevented a complete discussion of all these factors, but we include here some general observations with
respect to these characteristics.
Extensive crosslinking maximizes hardness and
scratch resistance, decreases the bulk conductivity and
its RH dependence, and increases brittleness. Thicker
overcoats and higher coating/curing temperatures increase crosslinking. Increased crosslinking is also re-
286
Journal of Imaging Science and Technology
Figure 10. The effect of a silsesquioxane overcoat on the positive corona charging characteristics of an organic photoreceptor. The photoreceptors (nonovercoated-top figure and
overcoated-bottom figure) were charge/expose exercised under
conditions where the gasses from the corona were not vented
but allowed to build up in the apparatus.
lated to a high methyl content, the presence of amines,
and the presence of glycidoxy substituents. Post coating cure has a minor effect on crosslinking, but is necessary for optimum performance. Increased “organic”
content helps to overcome the brittleness associated with
a highly crosslinked material having only methyl substituents, but hardness and scratch resistance are degraded with increasing organic content. The inclusion
of a low lattice energy salt, such as LiI, and a complexing
agent to prevent surface “blooming” increases the overcoat bulk conductivity, especially at low RH. However ,
incorporated salts make the overcoat more sensitive to
humidity and acidic corona effluents. Decreased over coat thickness reduces the electrophotographic residual
potential at the expense of decreased photoreceptor
scratch resistance.
Silsesquioxane overcoats have been successfully used
to extend the lifetimes of organic photoreceptors in the
electrophotographic process. We have shown that there
are many important physical and chemical overcoat
characteristics and that these may be affected by the
composition, cure conditions, and thickness. Through
careful manipulation of these factors it is possible to
Weiss, et al.
optimize the silsesquioxane overcoat to the demands of
the specific electrophotographic process for which it is
intended.
References
1. P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for Xerography, Marcel Dekker, Inc., New York, 1998.
2. K. K. Kochelev, V. I. Zhylina, G. E. Khots, O. K. Kocheleva, and V. V.
Sleptsov, IS&T’s NIP12: International Conf. on Digital Printing Technol.,
IS&T, Springfield, VA, 1996, p. 483.
3. X. S. Miao, Y. C. Chan, C. K. H. Wong, D. P. Webb, W. W. Lam, K. M.
Leung, and D. S. Chiu, J. Electronic Materials 26, 387 (1997); X. S.
Miao, Y. C. Chan and E. Y. B. Pun, Thin Solid Films 315, 123 (1998).
4. L. Cornelius, R & R News, July, 1994, p. 34.
5. D. S. Weiss, W. T. Ferrar and R. Cowdery-Corvan, Proc. IS&T’s NIP14:
International Conference Digital Printing Technologies, IS&T, Springfield, VA, 1998, p. 520.
6. L. L. Hench and J. K. West, Chem. Rev. 90, 33(1990).
7. H. A. Clark, U. S. Patent 3,986,997 (1976).
8. R. B. Frye, U. S. Patent 4,277,287 (1978).
9. D. S. Weiss, J. R. Cowdery, W. T. Ferrar, and R. H. Young, J. Imaging
Sci. Technol. 40, 322(1996).
10. I. Chen, J. Mort, M. A. Machonkin, and J. R. Larson, J. Imaging Sci.
Technol. 40, 431(1996).
11. E. J. Yarmchuck and G. E. Keefe, J. Appl. Phys. 66, 5435(1989).
12. T. Kobayashi, T. Saito, S. Aratani, S., Suzuki, and T. Iwayanagi, J. Imaging Sci. Technol. 39, 485(1995).
13. R. L. Schank, U. S. Patent 4,439,509 (1984).
14. R. L. Schank, U. S. Patent 4,595,602 (1986).
15. R. L. Schank, U. S. Patent 4,923,775 (1990).
16. M. Kumakura, I. Kaetsu, M. Horigome, T. Isomura, T. Yomeyama, and
T. Murata, U. S. Patent 4,912,000 (1990).
17. W. T. Ferrar, J. R. Cowdery-Corvan, E. T. Miskinis, C. Newell, D. S.
Rimai, L. J. Sorriero, J. A. Sinicropi, D. S. Weiss, and N. Zumbulyadis,
U. S. Patent 5,731,117 (1998).
18. R. H. Glaser, G. L. Wilkes and C. E. Bronnimann, J. Non-Crystalline
Solids 113, 73(1989).
19. D. S. Weiss, J. Imaging Sci. Technol. 34, 132 (1990).
Silsesquioxane Sol-Gel Materials as Overcoats for Organic Photoreceptors
Vol. 43, No. 3, May/June 1999
287
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Effects of Silica Additive Concentration on Toner Adhesion, Cohesion,
Transfer, and Image Quality
B. Gady,▲ D. J. Quesnel,* D. S. Rimai,† S. Leone,† and P. Alexandrovich†
Eastman Kodak Company, Rochester, New York
This article discusses the effect of silica concentration on transfer of nominal 8.5 µm diameter surface-treated toners, with the
silica concentration on the surface of the toner varying between 0 and 2% by weight of toner. In essence, it was found that, while
transfer efficiency increased with increasing silica concentration, resolution decreased and dot structure after transfer was
degraded. Toner adhesion measurements, performed using an ultracentrifuge, were found to correlate well with the transfer
efficiency measurements. Analysis of the results suggests that the adhesion and cohesion of toner is dominated by van der Waals
interactions. However, electrostatic forces associated with the charge on the toner become more significant with increasing silica
concentration with the two types of interactions becoming comparable when the silica concentration reached 2%.
Journal of Imaging Science and Technology 43: 288–294 (1999)
Introduction
It is well established that the adhesional properties of
toner particles affect transfer . 1–3 Numerous methods
have been employed to reduce toner-to-photoconductor
adhesion in order to both improve transfer and facilitate cleaning. For example, surface treatments such as
zinc stearate4 and Teflon have been demonstrated to significantly reduce toner adhesion.5 In two component developer systems, addition of third-component particulate
addenda to the toners have shown marked effects on
toner adhesion and improved the toner flow as well. 6,7
Third-component addenda such as silica8 are a particularly efficient means to reduce toner adhesion both to
itself and to photoconductors. Indeed, over the past few
years, the use of particulate addenda has enabled the
mean volume weighted average diameter of toner par ticles in commercially available electrophotography to
decrease from about 12 µm to approximately 8.5 µm.
Although the mechanism is not fully understood, it has
been shown that particles having diameters in the range
of tens of nanometers located on the surface of the toner
particles affects the adhesive forces to non-toner sur faces and the cohesive forces between toner particles.
The mechanism is presumably by the particulate addenda serving as asperities that reduce adhesion by
roughening the surface, preventing intimate contact
between the toner and the adherent surface or other
toner particles.
The reduction of cohesion between toner particles can
introduce new problems during transfer. As the images,
Original manuscript received August 5, 1998
comprised of collections of charged toner particles, are
transferred to the receiver , the repulsive electrostatic
forces between toner particles can cause the images to
fly apart. This effect is most severe in halftone dot images where the halftone dots literally can explode once
they leave the stabilizing influence of the latent image
charge pattern in the photoconductor. While dot explosions can occur in non-treated toner systems, Rimai and
Sreekumar 9 have observed that the use of submicrometer particulate addenda can aggravate the dot explosion problem, presumably by reducing the cohesion
between toner particles and thereby accentuating the
electrostatic repulsion between those particles. Alternatively, Tombs has proposed10 that when transfer is accomplished using an electrically biased transfer nip, dot
explosion may be caused by transfer of some of the toner
particles and halftone dots across the air gap in the
prenip region due to high electrostatic fields. The sur face forces must overwhelm the electrostatic repulsion
between the like sign charged toner particles in order
to keep the dots from exploding.
Turning now to the physics of adhesion, the adhesion
of particles to a compliant substrate such as polyurethane is well described11 by the JKR theory of adhesion.12
According to that theory, the force FS needed to remove
a particle of radius R from a substrate is given by
FS = −
3
w AπR
2
(1)
where wA is the thermodynamic work of adhesion and is
related to the surface energies γP and γS of the particle
and substrate, respectively, as well as their interfacial
energy γPS by
▲ IS&T Member
* current affiliation, University of Rochester, Rochester, New York
† current affiliation, NexPress Solutions LLC., Rochester, New York
© 1999, IS&T—The Society for Imaging Science and Technology
288
w A = γP + γS – γPS.
(2)
It is apparent from Eq. 1 that the JKR theory predicts
that the force needed to remove a particle from a sub-
strate is independent of the Young’s modulus of the substrate. Yet experimentally, the forces depend on the
moduli of the substrate. The role of the elastic modulus
in controlling particle adhesion can be understood by recognizing that particles are not perfect spheres as required
by the JKR theory. Rather, they have asperities, and as
shown by Fuller and T abor, 13 and more recently by
Schaefer and coworkers,14 the engulfment of the asperities into the substrate governs the removal force. Soft
photoconductors impede transfer by promoting particle
engulfment, as discussed by Mastrangelo. This effectively
serves to diminish the beneficial effect of the silica. Accordingly, the amount of silica, which effectively serves
as asperities on the surface of a toner particle, should
significantly affect the size of the removal force, especially for photoconductors that do not show substantial
particle engulfment. In principle then, the addition of
silica should facilitate transfer. However, as previously
shown,9 the addition of submicrometer addenda can also
enhance dot explosion. Indeed, dot explosion can occur
whether due to the reduced adhesion permitting toner
particles to transfer in the prenip region, or simply to a
decrease in the interparticle cohesiveness.
The purpose of this study is to address several questions relevant to transfer. Among these questions are:
1. Is the use of particulate addenda really necessary?
2. How does the amount of the surface treatment affect
transfer?
3. Does the amount of surface treatment affect
resolution or dot integrity?
4. How large are the forces holding the toner particles
to the photoconductor? Are these forces predominately due to van der W aals or electrostatic
interactions?
5. Can a sufficiently large electrostatic field be exerted
on toner particles and image structures, such as
halftone dots, to allow them to jump across air gaps
in the prenip region?
These questions were addressed in this study using
image quality attributes as well as more fundamental
measurements including transfer efficiency metrics and
adhesion-force measurements.
Experiment
In this study, the transfer efficiency (percent of toner
transferred divided by the amount of residual toner the
photoconductor plus the amount of toner transferred),
dot structure, and resolution of electrostatically transferred images were determined for a series of nominal
8.5 µm volume averaged diameter ground, cyan, toner
particles. In addition, the force needed to remove the
particles from a photoconductor was measured using a
Beckman LM 70 ultracentrifuge.
Two series of toners were used in this study
. The first
consisted of a ground polyester with between 0% and
2% Aerosil R972 (produced by Degussa, Inc.,
http:\\www.degussa.com) silica particles, by weight,
added to the surface of the toner particles. These particles have an average diameter , as reported by
DeGussa, of approximately 16 nm although SEM micrographs show agglomerates in the range of 60 nm.
The second series was quite similar except the toner
particles also contained a silicone release agent. The
volume-weighted average diameter of the toner, as determined using a Coulter Multisizer , was approximately 8.6 mm for the toner without the silicone
additive and approximately 8.1 mm for the siliconecontaining toner.
An electrophotographic developer was created by mixing the toner with a carrier comprising hard ferrite par
ticles. The carrier particles had a volume-weighted
diameter of approximately 30 mm. The initial toner concentration in the developer was approximately 6%. The
toner charge was determined using an apparatus containing two planar electrodes spaced approximately 1
cm apart. Approximately 0.1 g of developer was deposited on one electrode, located above, but in close proximity to, a donut-shaped segmented series of magnets
with alternating polarity. An electrometer was connected
to the upper electrode. The electrodes were biased in
such a manner as to attract the toner to the upper electrode as the magnets rotated, thereby simulating electrophotographic development. After all the toner was
stripped from the developer , the charge on the upper
electrode was determined and the mass of the toner giving rise to that charge was measured. This technique is
more fully described elsewhere. 15 The toner charge-tomass ratio was found to be approximately -37 ± 3 mC/g
for each of the toners.
Twelve grams of developer were loaded into a sumpless
development station comprising a rotating core of alternating pole magnets and a concentric stainless steel
shell. This type of station was chosen because it allowed
small amounts of developer to be used, and avoided
variations in the toner concentration and charge-to-mass
ratio associated with larger, more conventional stations.
Development was performed using the so-called “SPD”
technique, as discussed by Miskinis. 16 A commercially
available organic photoconductor was initially charged
to a predetermined negative potential using a grid-controlled DC corona charger and an electrostatic latent
image formed by contact-exposing the photoconductor
using a test target. T oner was deposited using discharged area development. The test target contained a
series of continuous-tone neutral density steps, a 150line rule 30% dot halftone pattern, and a resolution
chart. The photoconductor was then passed over the
development station where toner was deposited on the
photoconductor in an image-wise fashion. In order to
avoid complications associated with receiver variations,
the toner was electrostatically transferred directly to a
biased transfer roller having a resistivity of the order
of 109 Ω•cm. The speed of the photoconductor during
the transfer process was approximately 2.5 cm/s. The
width of the transfer nip formed between the transfer
roller and the photoconductor was approximately 3 mm.
Transfer voltages ranged between 500 and 2,500 V.
Transfer efficiency was measured using transmission
densitometry for toned optical densities on the
photoconductor between 0.1 and 1.0, using an X-Rite
model 310 densitometer with StatusA filters. The average transfer efficiency (determined by measuring the
transmission density of the transferred image and dividing by the sum of the transmission densities of the
transferred and residual densities) over the range of
optical densities was determined as a function of voltage applied to the transfer roller. The conducting layer
of the photoconductor was grounded and the maximum
transfer voltage applied was 2500 V. The transfer efficiency increased with applied transfer voltage over the
entire 0–2500 V range. The voltage, V 90%, at which the
average transfer efficiency exceeded 90% was then determined for each series of toners containing the various levels of silica mentioned above. In addition the
average transfer efficiency over both the range of toned
optical densities and the range of voltages between V80%
and 2500 V was also determined. This averaging proce-
Effects of Silica Additive Concentration on Toner Adhesion, ...Image Quality
Vol. 43, No. 3, May/June 1999
289
Figure 1. Average voltage for 90% transfer versus % silica addenda with (open symbols) and without (solid symbols) silicone adhesion additive.
dure was carried out using numerical integration of
curves fit to the data over the aforementioned range.
This method of averaging provides a measure of the “robustness” of the toner to transfer variations. Finally ,
the resolution and dot integrity were determined both
before and after transfer at an applied transfer voltage
of 1500 V. Each of these measurements was performed
with and without the addition of a silicone release agent
to the toner to promote release from the photoconductor
.
The adhesion of the toner particles to the
photoconductor was determined by developing low density patches and removing the toner in an ultracentrifuge capable of spinning at 70,000 rpm. The procedure
is as follows. The initial number of particles on the
photoconductor was established by counting, using suitable image analysis software. Next, the photoconductor
was placed in the centrifuge and spun at the desired
speed. The sample was then removed and the remaining particles counted. This process was repeated for a
series of increasing speeds. Centrifugation was per formed in a low vacuum of approximately 10 –2 torr
(roughing pump vacuum). The initial coverage was 0.5
density as measured in transmission corresponding to
a 50–60% surface coverage by the particles.
Results
The applied voltage, V90%, where the transfer efficiency
exceeds 90%, as a function of silica concentration, is
shown in Fig. 1 for the toners with and without the silicone additive. As can be seen, the voltage necessary for
90% transfer drops rapidly with increasing silica concentration for both toners. However, the effect levels off
for silica concentrations of more than 0.5% with the effect for 1% and 2% silica only incrementally larger than
that at 0.5%. Moreover, it can be seen that the use of a
silicone additive in conjunction with the silica not only
does not result in a further reduction in the voltage
needed for 90% transfer but actually reduced the effect
of the silica treatment applied without the silicone additive. The silicone additive may be acting as a liquid
bridge that actually reduces the efficiency of the silica
in separating the toner from the surface. Further studies are needed to understand this issue in more detail.
Figure 2 shows the integrated averaged transfer efficiency above 80% for each of the two silica-treated toner
series, normalized to the performance of the toner with-
290
Journal of Imaging Science and Technology
Figure 2. Normalized density averaged transfer efficiency
integrated over voltage from the voltage needed for 80% transfer to the upper bound of 2500 V with (open symbols) and without (solid symbols) silicone adhesion additive as a function of
silica content. Normalization is with respect to the integrated
density averaged transfer efficiency for the toner without silica
addenda and without silicone adhesion additive.
out silica or silicone additive. Solid symbols show the
results without silicone additive while open symbols
show the results when silicone additive is present. The
integrated averaged transfer efficiency is determined
by first averaging the measured transfer efficiency over
a range of 10 density steps from 0.1 to 1.0 for each voltage from 0 to 2500 V in steps of about 200 V. A smooth
curve is then fit to the average transfer efficiency as a
function of voltage and this curve is integrated from the
lowest voltage that produces an 80% average transfer
efficiency to the maximum voltage examined, 2500 V .
In this way, systems with sharply spiked average transfer efficiency versus applied transfer voltage will show
a lower voltage integrated average and can be distinguished from more robust systems showing a broad
maximum. It can be seen from these figures that the
integrated average transfer efficiency , a measure of
transfer robustness, despite an initial decrease, generally improves with increasing silica concentration, but
at a decreasing rate once the silica concentration exceeds 0.5% by weight of toner. These results are consistent with the voltage results shown in Fig. 1. Also in
agreement with Fig. 1, the data shows that the presence of the silicone additive reduced the integrated average transfer for all conditions.
From the data presented thus far, it may appear that
the process of transferring toner can be made more robust, although perhaps reaching a point of diminishing
returns, simply by increasing the concentration of silica
on the toner particles. However , this is not quite cor rect. Transfer is not just the removal of toner from a
photoconductor accompanied by a deposition of the toner
on a receiver. Rather, it is that process with the additional constraint that image disruption must be minimized. Image disruption was characterized in this study
by microscopically examining the halftone dot pattern
and resolution chart before and after transfer.
In this study the effect of the silica concentration on
image disruption was determined by qualitatively examining the structure of the halftone dots and measuring the resolution in line pairs per millimeter before
and after transferring the image using a 1500 V trans-
Gady et al.
(A)
Figure 4. Resolution as a function of silica concentration for
the toner with (open symbols) and without (solid symbols)
silicone.
(B)
(C)
Figure 3. Halftone dot patterns after transfer for the siliconecontaining toner with 0% 3(A), 0.5% 3(B), and 2.0% 3(C) silica.
fer bias. Before transfer , a resolution between 14 and
16 line pairs per millimeter was obtained. Moreover, the
dots were well formed, exhibited minimal satellite for mation, and, in general, appeared to reproduce the test
target quite well. However, it was found that after transfer, the dots were disrupted, with the amount of disruption and the number of satellites increasing
monotonically with increasing silica concentration. This
effect is shown in Figs. 3A-3C for the silicone-containing toner with 0, 0.5, and 2.0% silica, respectively . As
can be seen in Fig. 3(A), in the absence of silica, the
halftone dots are still fairly well formed after transfer,
although disruption and the presence of satellite toner
particles are obvious. Increasing the amount of silica to
0.5% clearly resulted in significantly more dot disruption and satellite formation, as shown in Fig. 3(B). Upon
further increasing the amount of silica to 2.0%, the dot
structure has been nearly obliterated by disruption of
the dots during transfer , as illustrated by Fig. 3(C).
Resolution also tends to decrease with increasing silica
concentration. This effect is shown in Fig. 4 for toners
both without and with the silicone additive. The reduc-
Figure 5. The percent of toner removed from the photoconductor
at 70,000 rpm as a function of silica concentration, with (open
symbols) and without (solid symbols) silicone.
tion in resolution is more severe for the toner system
containing the silicone adhesion additive.
As indicated earlier, an ultracentrifuge was used to
characterize the toner-to-photoconductor adhesion as a
function of the weight percentage of silica. Figure 5 reports the percentages of toner with silicone (open circles)
and without silicone (solid circles), that were removed
from the photoconductor at 70,000 rpm for the five levels of silica examined. W ith the exception of an initial
increase at 0.25% silica, the percent removed increases
monotonically with increasing silica content, asymptotically approaching 100% removal at or around 2% silica
by weight. The initial increase at 0.25% silica is viewed
as an anomalous point that is correlated with the atypically smooth surface morphology of this particular toner
mixture when examined by scanning electron microscopy (SEM). The presence of silicone in the toner mixtures showed no further reduction in the adhesion force,
even in the absence of the silica. These results suggest
that while the presence of silica significantly reduces
the adhesional forces, the presence of silicone does not.
The behavior of the toner -silica mixtures determined
Effects of Silica Additive Concentration on Toner Adhesion, ...Image Quality
Vol. 43, No. 3, May/June 1999
291
Figure 6. The percent removed by centrifuge as a function of
removal force for three levels of silica: 0% (solid circles); 1%
(open circles); and 2% (solid triangles); for toner without silicone adhesion additive.
by mechanical measurements in the ultracentrifuge are
essentially unchanged by the presence of silicone in contrast with the systematic changes in the adhesional behaviors inferred from the transfer measurements
mentioned earlier.
Figure 6 shows the percent of the toner (without silicone) removed from the photoconductor as a function of
the mean applied force produced by different centrifuge
speeds. Data for three silica concentrations of 0%, 1%,
and 2% are shown. The highest force corresponds to
70,000 rpm so that the end points of the curves in Fig. 6
are the 1 st , 3 rd, and 5 th data points from Fig. 5. As can
be seen, the general shapes of the curves gradually
change for increases in silica concentration. W ithout
silica, the percent removed is nearly linear with the
mean applied force over the range investigated. There
is no tendency to reach an asymptote. W ith 2% silica,
the curve rises steeply and then curves to asymptotically approach 100% particle removal as the mean applied force is increased. The result for 1% silica is
intermediate following the 0% result initially and then
rising as the centrifugation speed and hence mean force
is increased. Because there is a distribution in toner
sizes, the larger particles would be removed first. If 1%
is insufficient to coat all the particles completely , this
could be a rationalization of the behavior observed for
1% silica.
The mean applied forces reported above were calculated by assuming that the particles were spherical polyester toner with a radius of 4 µm and a mass density of
1.2 g/cm3. The removal force, PS, estimated at the 50%
removal point, was determined to be 970 nN, 580 nN,
and 39 nN for the 0%, 1%, and 2% silica-coated toner
particles, respectively.
Analysis
As shown in the previous section, transfer efficiency
improves with increasing silica concentration while dot
integrity and resolution are both degraded. Moreover ,
the force needed to detach the toner from the
photoconductor also decreases with increasing silica concentration.
As is well known, there is much debate in the literature as to whether the force of adhesion of toner par ticles to a photoconductor arises from surface forces
292
Journal of Imaging Science and Technology
such as those due to van der Waals interactions or electrostatic forces from a toner particle seeing its image
charge. Although the resolution of that debate is well
beyond the scope of this article, it is worthwhile to estimate to adhesional forces arising from both mechanisms.
Let us first assume that the uncoated toner particles
are spheres with a radius of approximately 4 µm. The
particle removal force, FS, can be calculated from JKR
theory using Eq. 1. Assuming a reasonable value of w A
= 0.05 J/m 2, the particle removal force is estimated to
be 943 nN. In light of the approximations made, this
value is in reasonable agreement with the experimentally obtained value of 970 nN.
Estimates of the electrostatic contribution to particle
adhesion are not as simple to make, owing to polarization and charge distribution effects. Although details of
this problem are presented elsewhere, 17,18 these issues
will be examined briefly. Assuming that an irregularly
shaped toner particle can be approximated as a dielectric sphere of radius R having a chargeq uniformly distributed over its surface, the electrostatic image force
of attraction, FI, between that particle and a conducting substrate is given by
FI = α
q2
4 π ε 0 ( 2 R) 2
.
(3)
When κ = 4, representing a value of the dielectric constant appropriate for toners, the value 19 of α is 1.9. In
the present situation, however, the toner is not adhered
to a conductor. Rather, the photoconductor comprises
an organic binder whose dielectric constant is similar
to that of the toner. The problem of a charged dielectric
particle adhering to a dielectric substrate has not been
solved. Moreover, even for the case of a spherical par ticle in contact with a plane, there will be a finite contact area associated with deformations of the contacting
materials due to adhesion-induced stresses. For polymeric materials, such deformations can be quite large,
with the contact radius being of the order of 10% of the
particle radius.20 For irregularly shaped toner particles,
independent sets of measurements by Eklund and coworkers 21and by Bowen and co-workers 22 both report
contact areas being of the order of 10% of the projected
cross-sectional area of the toner particles. In the case
where the dielectric constants of the two contacting materials are equal and there are no air gaps,α = 1.0. Presumably, the present case would lie between these
extremes. Using the values of charge to mass reported
earlier, (37 ± 3 µC/g, ρ = 1.2 g/cm3), it is then calculated
that FI would be in the range of 20 to 40 nN for the
present toner particles. This value is far less than the
measured force needed for detachment shown in Fig. 6.
However, as is discussed by Hays, the charge on a
toner particle may not be uniformly distributed over its
surface. In that instance, the electrostatic contribution
to the force of adhesion, FE, is related to a surface charge
density, σ, and the actual area of contact between the
particle and substrate, and AC, by
FE =
σ 2 AC
.
2ε 0
(4)
Using Eq. 4 and assuming that the contact area is approximately 10% of the cross-sectional area, one could simply solve for the charge density needed to give the
measured removal force. Upon substitution, one finds that
σ = 1.85 × 10–3 coul/m2. Using a parallel plate capacitor
Gady et al.
approximation, one finds that this charge density would
result in an electric field of approximately 2.1 × 108 V/m.
This would clearly exceed the Paschen limit in air and
would result in dielectric breakdown as the toner particle
approached the photoconductor during development.23,24
Alternatively, it is worthwhile to estimate F E within
the confines of the Paschen limit.Again, this is not simple
to do, as the Paschen limit decreases with increasing air
gap. Assuming that the particle can get to within 10 µm
of the photoconductor without discharging, the supportable field would be approximately 3.5 × 107 V/m. The attainable surface charge density would then be of the order
of 3 × 10–4 coul/m2 and FE would be of the order of 30 nN,
which is consistent with estimates of F I. Therefore, the
force of adhesion due to the presence of localized charged
patches is much smaller than those contributions attributed to van der Waals interactions.
The electric field needed to detach the particle from
the photoconductor can also be estimated. As discussed
by Hays, the force needed to detach a particle from a
substrate F D is given by
FD = βqED − γπ (2 R) 2 ED2
(5)
where ED is the electric field needed to detach the par ticle and β and γ are the polarization correction factors
with values approximately 1.6 and 0.063 for a dielectric
constant of κ = 4. Following Hays, 18 it is assumed that
detachment occurs when the electrostatic detachment
force equals or exceeds the forces driving attachment,
FD – FA ≥ 0
(6)
where FA represents the total force adhering the toner
to the photoconductor. It should be noted that Hays’ assumption that FA = FI or FA = FE is not strictly correct
due to the compliance of the contacting materials. T o
correctly solve the problem, one needs to include the
effects of mechanical deformations of the particle and
substrate, along with the forces that arise from works
of adhesion, as discussed by Johnson and co-workers.12
Naturally, these deformations will impact the solutions
to the image charge and removal force calculations due
to the substantial changes in geometry that can be
caused by surface forces. However, the present approximation is frequently used in the literature and should
suffice for the present calculations.
As before, the contributions due to polarization are difficult to precisely determine because the dielectric constant of the particle and substrate, which are in intimate
contact, are essentially the same. However, estimates can
still be made. Following Hays, it was assumed that the
second term on the right-hand side of Eq. 5 is small and
can be neglected. If polarization effects are then ignored
(α = β = 1), using the experimental removal force of 970
nN suggests that ED takes on a value in the neighbor hood of 8 × 107 V/m, which is too large a field to sustain
in air. The detachment field would be even larger if polarization were significant. Accordingly, it should not be
possible to electrostatically detach this toner from the
photoconductor without balancing surface forces, as discussed in Ref. However, the ability of toner particles to
jump air gaps is generally required, as discussed earlier,
in order to achieve good transfer due to tent poling effects such as those arising from receiver roughness, toner
stack height variations, toner particle-size polydispersity,
etc. Therefore, the transfer efficiency of such a toner that
cannot traverse an air gap is generally relatively poor ,
which was indeed, observed in earlier studies.3
Evaluating the detailed physics of the toner -tophotoconductor interaction requires some clarifying assumptions. First, assume that the toner is held to the
photoconductor principally by relatively short-range van
der Waals forces and that the role of the silica is to physically separate the toner from the photoconductor. A precise determination of the effect of the silica on the toner
detachment forces will require a detailed knowledge of
how the toner-to-photoconductor contact deforms under
the influence of the surface forces. This depends on a
number of factors such as the size and distribution of the
silica, the shape of the toner , the range of the interactions, and the compliance of the materials. However, one
may still make some order of magnitude estimates of the
detachment forces of the silica-treated toner particles.
The percent of the surface coverage of the toner by
the silica can be estimated by assuming both the toner
and silica are spherical. For the purpose of this calculation, assume the weight fraction of the silica is 1%. The
primary particle size of the silica is 16 nm diameter but
it is clustered into particles of 60 nm average diameter,
also assumed to be spherical. Using ρ = 1.75 g/cm 3 as
the mass density of the silica and ρ = 1.2 g/cm 3 as the
mass density of the toner, and knowing that the toner
has a mean diameter of 8µm, the fraction of the surface
area of the toner covered by silica clusters is 25%. For
2% silica by weight, the area coverage calculated is 50%.
These estimates are consistent with SEM micrographs
of the toner.
Again, assuming a spherical toner particle, the contact radius aJKR, estimated using JKR theory , is given
by
 6πwA R 2 
aJKR = 

E


1/ 3
(7)
where E is the Young’s modulus of polyester,25 approximately 3 Gpa. In the absence of silica, aJKR = 196 nm.
Assuming a similar contact region exists when silica is
present, it is then estimated that approximately 10 silica
particles would be in contact with the photoconductor
when the silica concentration is 2%. The separation force
FS’ is then given by
3
FS ' = n wA πr
2
(8)
where n = 10 is the number of contacts and r = 30 nm is
the radius of the silica particle clusters. Assuming that
the work of adhesion for silica to photoconductor remains
at wA = 0.05 J/m2, upon substitution it is found that FS’
≈ 70 nN. The experimentally obtained value of FS’ was
approximately 39 nN. In view of the approximations
made, the experimentally obtained value is in reasonable agreement with the estimated value. It is interesting to note that these values are also close to the
estimated contributions of the electrostatic image
charges to the total force of adhesion, suggesting that,
at this level of silica treatment, both van der W
aals and
electrostatic interactions are significant factors in determining the total force holding the toner to the
photoconductor. The applied electrostatic field needed
to effect separation of the toner from the photoconductor
was estimated, using Eq. 5, to be in the range of 3 to 6×
106 V/m, which is readily obtainable.Accordingly, transfer efficiency should be quite good in the presence of
the silica particles in agreement with the experimental
observations.
Effects of Silica Additive Concentration on Toner Adhesion, ...Image Quality
Vol. 43, No. 3, May/June 1999
293
The detachment force for the toner particles containing 1% silica was determined by the centrifuge experiments to be approximately 580 nN, or about an order of
magnitude larger than the estimated image charge contributions. In this case, the detachment field was calculated to be approximately 4.9 × 10 7 V/m, ignoring
polarization effects. This result suggests that transfer
of the toner across an air gap would not be feasible even
with this level of silica present. Rather, it is necessary
for the receiver to contact the toner, thereby supplementing the electrostatic transfer forces with surface forces.
The observed losses in dot integrity and resolution
can also be explained in terms of decreasing adhesion.
As discussed previously, the highly charged toner par ticles would tend to repel one another rather than exist
as a coherent mass, as in a dot or alpha-numeric char acter. However, at short ranges,26 i.e., less than 30 nm,
the attractive van der W aals forces dominate over the
Coulombic repulsion stabilizing the images during
transfer. While offering beneficial effects for transfer
by reducing the toner -to-photoconductor adhesion, the
presence of the nanometer-size silica particles reduces
the interparticle cohesion as well, thereby increasing
the propensity for clusters of toner particles comprising the images to fly apart during transfer. Indeed, increases in toner cohesion with aging, attributed to the
silica being engulfed by the toner particles and thereby
losing their spacer effect, was reported by Ott.8
of the van der W aals and the electrostatic forces become comparable in magnitude.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Conclusions
It was found that the transfer efficiency of an electrophotographic toner increases with an increasing concentration of nanometer -size silica particles on the
surface of the toner . However, accompanying the improved transfer efficiency is a loss of resolution and a
decrease in dot integrity . These results track with a
decrease in the adhesion of the toner to the
photoconductor, as measured with an ultracentrifuge.
The size of the removal forces measured appear consistent with estimates that assume van der Waals interactions, but, in general, appear too large to be
attributed to electrostatic interactions alone. As the
concentration of silica approaches 2%, the contributions
294
Journal of Imaging Science and Technology
19.
20.
21.
22.
23.
24.
25.
26.
N. S. Goel and P. R. Spencer, Polym. Sci. Technol. 9B, 763 (1975).
C. J. Mastrangelo, Photo. Sci. Eng. 22, 232 (1978).
D. S. Rimai and A. Chowdry, U.S. Patent #4,737,433 (1988).
E. M. Williams, Physics and Technology of Xerographic Processes ,
Wiley-Interscience, New York, 1984.
D. S. Rimai, unpublished results.
P. K. Watson, H. Mizes, A. Castellanos, and A. Pérez , Proc. 21st Annual Meeting of the Adhesion Society, R. A. Dicke, Ed., Adhesion
Society, Blacksburg, 1998, pp. 272–274.
J. M. Valverde, A. Ramos, A. Castellanos, and P. K. Watson, Proceedings of the 21st Annual Meeting of the Adhesion Society, R. A.
Dicke, Ed., Adhesion Society, Blacksburg, 1998, pp. 278–280.
M. L. Ott, in Proc. 19 th Annual Meeting of the Adhesion Society, T. C.
Ward, Ed., Adhesion Society, Blacksburg, VA, 1996, pp. 70–73.
D. S. Rimai and C. Sreekumar, unpublished results.
T. N. Tombs, private communication.
D. S. Rimai and L. P. DeMejo, Annu. Rev. Mater. Sci. 26, 21 (1996).
K. L. Johnson, K. Kendall, and A. D. Roberts, Proc. Roy. Soc. London
Ser. A, 324, 301 (1971).
K. N. G. Fuller and D. Tabor, Proc. Roy. Soc. London, Ser. A, 345,
327 (1975).
D. M. Schaefer, M. Carpenter, B. Gady, R. Reifenberger. L. P. DeMejo,
and D. S. Rimai, J. Adhesion Sci. Technol . 9, 1049 (1995).
J. C. Maher, IS&T’s Tenth International Congress on Advances in NonImpact Printing Technologies , IS&T, Springfield, VA, 1994, pp. 156–
159.
E. T. Miskinis, Proc. Sixth International Congress on Non-Impact Printing , IS&T, Springfield, VA, 1990, pp. 101–110.
D. A. Hays, in Fundamentals of Adhesion and Interfaces, D. S. Rimai,
L. P. DeMejo, and K. L. Mittal, Eds., VSP, Utrecht, 1995, pp. 61–72.
D. A. Hays, in Advances in Particle Adhesion, D. S. Rimai and L. H.
Sharpe, Eds., Gordon and Breach, Amsterdam, 1996, pp. 41–48.
D. A. Hays, in Particles on Surfaces 1: Detection, Adhesion and Removal, K. L. Mittal, Ed., Plenum Press, New York, 1988, pp. 351–
360.
D. S. Rimai, L. P. DeMejo and R. C. Bowen, J. Adhesion Sci. Technol.
8, 1333 (1994).
E. A. Eklund, W. H. Wayman, L. J. Brillson, and D. A. Hays, in IS&T’s
Tenth International Congress on Advances in Non-Impact Printing
Technologies, IS&T, Springfield, VA, 1994, pp. 142-146.
R. C. Bowen, L. P. DeMejo and D. S. Rimai, J. Adhesion 51, 191
(1995).
F. Paschen, Wied. Ann. 37, 69 (1889).
J. D. Cobine, Gaseous Conductors, Dover Publications, New York,
1957.
D. W. van Krevelen, Properties of Polymers , Elsevier, Amsterdam,
1976.
B. Gady, R. Reifenberger, D. S. Rimai, and L. P. DeMejo, Langmuir
13, 2533 (1997).
Gady et al.
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Effect of Adsorption of Long Chain Alcohol Molecules on Silica Particles
on Toner Charging
K.-Y. Law▲,* and I. W. Tarnawskyj
Xerox Corporation, Wilson Center for Research and Technology, Webster, New York
The adsorption of long chain alcohols on the surfaces of hydrophobic silicas has been studied using 1-hexadecanol on R972 silica as a
model system. Both DSC and IR spectral data suggest that 1-hexadecanol is in the gaseous state on the silica surface at concent
rations
< 4.8% by weight. At 9.4% loading, which is equivalent to half of a theoretical monolayer on the silica surface, results sugges
t that the
adsorbed 1-hexadecanol forms a hydrocarbon protective layer on the surface. The formation of the layer implies that the hydroca
rbon
chains in the adsorbed 1-hexadecanol are interacting with each other, presumably by folding the hydrocarbon chains back toward the
silica surface. As the concentration of 1-hexadecanol increases, the space occupied by the folded hydrocarbon chains is replace
d by the
added 1-hexadecanol, up to one theoretical monolayer (17%). Beyond this concentration, crystallization of 1-hexadecanol occurs. The
effect of the chain length of the adsorbed alcohol is studied at a theoretical monolayer coverage for a series of normal alcoho
ls, from C12
to C22. While protective hydrocarbon layers are formed for all the normal alcohols studied, IR spectral data suggest that the layers
formed from 1-octadecanol, 1-eicosanol and 1-docosanol are less organized. The surfaces of these modified silicas may be more hydrophilic as compared to those from 1-dodecanol, 1-tetradecanol and 1-hexadecanol. The charging properties of alcohol-treated sili cas
were studied by first blending them with 9 µm unpigmented SPAR toner at 0.5% by wt. concentration, followed by charging the
resulting toners with metal beads. The tribo data reveal that adsorption of long chain alcohols on the silica surface enhances
the
negative tribocharge of the resulting toner at both low (20%) and high (80%) relative humidity (RH). High tribo with minimal RH
sensitivity, as judged from the tribo ratio from 20% to 80% RH, are obtained when R972 is covered with one theoretical monolayer of
1-hexadecanol. The attainment of optimal charging result is shown to correlate to the proposed molecular structure of the adsor bed
alcohol layer. The important role of the hydrophobicity of the silica surface in toner charging is discussed.
Journal of Imaging Science and Technology 43: 295–299 (1999)
Introduction
Non-porous fumed silicas ranging from 5 to 50 nm in
diameter are commonly used as flow-aids in xerographic
toners.1 In practice, they are performing a dual function. They enhance the flow while controlling the charging characteristics of the toner . One of the drawbacks
of the approach is the humidity sensitivity of toner
charging, e.g., the tribo of the toner decreases as the
ambient humidity increases. W e reported earlier that
the charging of both hydrophilic and hydrophobic silicas are sensitive to humidity when they are incorporated into a toner; and expectedly hydrophilic silica is
more humidity sensitive between the two.2 The correlation between hydrophobicity of the silica surface and
the degree of humidity sensitivity in toner charging is
quite profound, however . We recently attempted to
modify the surfaces of silica particles using conventional
charge control additives (CCA). While the effect of the
added CCA on toner charging is evident, the humidity
sensitivity remains.3,4 It was concluded that a possible
solution to eliminate or reduce humidity sensitivity is
by modifying the surface of silica particles, rendering it
Original manuscript received October 6, 1998
▲ IS&T Member
* To whom correspondence should be addressed
© 1999, IS&T—The Society for Imaging Science and Technology
hydrophobic. We intuitively feel that this may be accomplished by encapsulating the silica surface with a
hydrocarbon layer such that the fluctuation of water
concentration on the surface is small as the ambient
humidity is varying.
In this work, we report the use of a hydrocarbon layer,
formed by adsorption of long chain alcohol molecules on
the surface of a hydrophobic silica, as an encapsulating
layer for the silica surface. Using 1-hexadecanol on hydrophobic silica R972 as a model, we systematically show
that the adsorbed alcohol layer enhances negative charging and reduces the humidity sensitivity of the charging
process when the alcohol-treated silica is incorporated
in an unpigmented polyester (SP AR) toner. While both
the chain length of the alcohol molecule and the concentration of the adsorbed alcohol on the silica surface are
shown to have an effect on toner charging, the overall
result can be rationalized in terms of an increase in hydrophobicity of the silica surface after the alcohol treatment. The origin of the increase is discussed.
Experimental
Materials. The hydrophobic silica used in this work
was R972 from Degussa. Long chain alcohols, 1dodecanol, 1-tetradecanol, 1-hexadecanol, 1-eicosanol
and 1-docosanol were of the highest commercial quality purchased from Aldrich Chemical Co. The coating
solvents were low boiling hydrocarbons, e.g., pentane,
hexane and cyclohexane; they were of spectral grade
295
TABLE I. Physical and Spectroscopic Data of Silicas Formed by Adsorbing Various Concentrations of 1-hexadecanol on the
Surface of R972 Silica
1-Hexadecanol concentration (% wt.)
C
H
< 0.5
Si
50.40
m.p. by DSC (degree C)
IR (cm-1)
0
Found:
0.92
—
—
4.8
Calc’d:
Found:
4.68
4.74
0.69
0.59
48.08
48.33
—
2928.4
9.1
Calc’d:
Found:
8.05
8.18
1.29
0.94
45.81
45.52
41
(very broad)
2917.7
13
Calc’d:
Found:
11.11
10.88
1.84
1.64
43.85
43.64
41
2917.3
17
Calc’d:
Found:
14.00
13.45
2.36
2.11
44.98
43.60
41
2916.9
23
Calc’d:
Found:
18.94
18.84
3.25
3.30
38.81
36.85
50
2917.0
29
Calc’d:
Found:
23.33
22.78
4.04
4.17
35.99
35.93
50
2917.0
100
Calc’d:
79.27
14.14
54
2918.1
a
(a)CH 2 stretching in the hydrocarbon chain of 1-hexadecanol.
from Fisher Scientific. The toner resin was a linear
polyester, SPAR, from Goodyear. The toner was unpigmented and was prepared by a melt-extrusion and jetting technique. It was classified to ~ 9 µm. The metal
beads for the charging experiments were made of steel
core (~ 130 µm) and were solution-coated by a carbonblack doped poly(methyl methacrylate) polymer on the
surface at a total weight loading of ~ 0.8%.
General Techniques. Infrared spectra were deter mined on a Perkin Elmer model 1750 FTIR. Differential scanning calorimetry (DSC) was performed on a
Perkin Elmer DSC7 with a 2-stage inter -cooler and a
IBM PS2 model 50z computer.
Preparation of Alcohol-Treated Silicas. A long
chain alcohol was dissolved in ~ 100 mLof a hydrocarbon solvent inside a 250 mL round bottom flask. Silica
R972 was added to the alcohol solution and the resulting suspension was subjected to an ultrasonic treatment for 2 h and then stirred overnight. The solvent
was removed on an evaporator. The white residue obtained was dried in a force-air oven overnight at 70°C.
The solid was transferred to a 4-oz bottle and rollmilled with 35 g of 1/4′′ steel shot for 30 min at a speed
of 90 ft/min, yielding a fluffy white powder , the alcohol-treated silica.
Preparation and Evaluation of Experimental
Toners. The above alcohol-treated silica (0.063 g), the
SPAR toner (12.5 g) and 125 g of 1/4 ′′ steel shot were
placed inside a 4-oz bottle and was rolled for 30 min to
prepare an experimental toner. Developer was prepared
by placing the above toner (1.25 g) and the metal beads
(60 g) inside a 2-oz bottle. The entire content was conditioned inside a humidity-controlled glove box at a constant humidity, either 20% or 80%, overnight and was
sealed. The tribo of the toner was generated by tumbling the toner and the metal beads inside the 2-oz bottle
on a roll-milled for 5 min at a speed of 90 ft/min. The
charge generated was determined by the standard blowoff technique inside a Faraday cage under lab ambient
condition. 5 Controlled experiments indicated that the
charging process was leveled off after 5 min of roll milling and the tribo value was primarily sensitive to the
ambient RH when the toner was charged up.
296
Journal of Imaging Science and Technology
Results and Discussion
Alcohol-Treated Silicas. The Concept. The surface
of hydrophobic fumed silica, such as R972, is quite polar even though it has been made hydrophobic by a
silanating agent. There are residual silanol groups as
well as siloxane groups on the surface. 6 These
functionalities are capable of forming H-bonding with
the OH group in alcohol. 7 The interaction would result
in alcohol adsorption on the silica surface.At sufficiently
high concentration, hydrocarbon chains of certain long
chain alcohols start to interact with each other by the
Van der Waals force, forming an encapsulating hydrocarbon layer. A schematic of the concept is shown in
Scheme 1.
Scheme 1.
Preparation, Characterization and Physical
Properties. Hydrophobic silica R972 was chosen as a
model for preparation and characterization studies.
Basically, alcohol adsorption takes place when R972 is
added to a hexane solution containing a long chain alcohol. After removal of the solvent, then drying, the alcohol-treated silica is prepared. Table I summarizes the
results of the effect of 1-hexadecanol concentration on
the physical and spectroscopic properties of the alcohol-treated R972 silicas. Elemental data show that there
is a good agreement between the found values and the
theoretical values based on the feed ratios. Evidence for
the occurrence of alcohol adsorption comes from the
hexane washing experiment. For instance, at 1hexadecanol concentration ≤ 17%, the alcohol molecules
are found to be immobile and cannot be washed off by
hexane. At 1-hexadecanol concentrations higher than
17%, part of the adsorbed 1-hexadecanol can be washed
off by hexane and the resulting silicas generally consist
~ 17% by weight of 1-hexadecanol. The result suggests
that up to ~ 17% of 1-hexadecanol can be tightly
adsorbed on the surface of R972 silica.
Law and Tarnawskyj
centration on the surface. W e hypothesize that the
adsorbed alcohol is in a gaseous state at these low concentrations. At concentrations ≥ 9.1%, the hydrocarbon
chains interact with each other on the silica surface, forming a hydrocarbon layer. It is important to note that at
9.1% of 1-hexadecanol, there is only half of a theoretical
monolayer of 1-hexadecanol on the surface. The fact that
the hydrocarbon chains start to interact and form a hydrocarbon layer indicates that hydrocarbon chains may
be folding back toward the silica surface (Scheme 2A).
Between 9.1 to 17%, the space occupied by the folded
hydrocarbon chains is replaced by the incoming alcohol
as the concentration of 1-hexadecanol is increasing
(Scheme 2B). At concentrations above one theoretical
monolayer, the excess alcohol may simply crystallizes on
the alcohol-treated silica surface (Scheme 2C).
Figure 1. DSC isotherms of various 1-hexadecanol treated
R972 silicas.
A similar conclusion can also be reached via the filtration experiment. For example, by adding 3 g of R972
into a 100 mL hexane solution containing 1.2 g of 1hexadecanol, one would yield an alcohol-treated silica
containing 29% of 1-hexadecanol if one prepares the
sample by the solvent evaporation method. Instead, if
one isolates the alcohol-treated silica by filtration followed by a gentle hexane wash, the silica is shown to
contain ~ 17% by wt. of 1-hexadecanol as indicated by
the weight up-take. This finding confirms that the maximum amount of 1-hexadecanol that can adsorb on the
silica surface of R972 is ~ 17%. Incidentally, in a typical
preparation of an alcohol-treated silica containing 17%
by wt. of 1-hexadecanol, 3 g of R972 and 0.6 g of 1hexadecanol are used. The total surface area in the silica
sample is 3.3 ± 0.6 × 1022 Å2 based on the BET surface
area of R972. 6 The total molecular area for 0.6 g of 1hexadecanol is estimated to be ~ 3.7 ± 0.6 × 10 22 Å2, assuming that 1-hexadecanol forms a Langmuir–Blodgett
film structure with a molecular area of 25 Å2/molecule.8
The matching of surface area between the silica and 1hexadecanol suggests that R972 can adsorb as much as
one theoretical monolayer of alcohol on the surface.
The DSC scans for various 1-hexadecanol treated R972
silicas are depicted in Fig. 1. The melting peaks and the
IR spectral data are tabulated in Table I. The results show
that melting peaks only become observable at 1hexadecanol concentrations ≥ 9.1%. Controlled experiments suggest that the absence of a melting peak at
concentrations ≤ 4.8% is not due to the low alcohol con-
Scheme 2.
The hypothesis in Scheme 2 is supported by IR spectral data that show that the hydrocarbon chains are
indeed interacting with each other when the alcohol
concentration is higher than half of a monolayer . For
example, the aliphatic C-H stretching of the methylene
group is known to be sensitive to the packing density of
the hydrocarbon chain, with a lower stretching frequency for CH 2 group in the crystalline phase. 9,10 This
frequency shift has been used to characterize the packing of the hydrocarbon chains in self-assembled monolayers and Langmuir -Blodgett films.11 In T able I, we
show that the C-H stretching frequency decreases from
2928.4 cm–1 at 4.8% of 1-hexadecanol on R972 to ~ 2918
cm-1 for 1-hexadecanol concentration ≥ 9.1%. The data
clearly suggest that 1-hexadecanol molecules are interacting with each other at concentrations ≥ 9.1% and
there is certain degree of hydrocarbon chain packing
analogous to that occurs in pure 1-hexadexanol.A simi-
Effect of Adsorption of Long Chain Alcohol Molecules on Silica Particles on Toner Charging
Vol. 43, No. 3, May/June 1999
297
TABLE II. DSC and IR Spectral Data of Silicas Formed by Adsorbing Various Alcohols on the Surface of R972 at a Theoretical Monolayer Coverage
Alcohol
1-dodecanol
1-tetradecanol
1-hexadecanol
1-octadecanol
1-eicosanol
1-docosanol
m.p. by DSC (degree C)
neat
on R972
29
40
54
62
68
72, 74
14
31
41
55
63
69
Concentration of
1-hexadecanol
IR (cm –1) a
neat
on R972
2926.1
2919.9
2918.1
2917.0
2917.8
2917.0
TABLE III. Effect of 1-Hexadecanol Adsorption on R972 on the Charging
of SPAR Toner
SPAR resin only
0% (R972 only)
4.8%
9.1%
13%
17%
23%
29%
2925.9
2919.9
2916.9
2917.9
2918.0
2917.6
(a) CH2 stretching in the hydrocarbon chain of various alcohols.
Toner tribo
20% RH
80% RH
–15.4
–24.1
–30.4
–30.4
–30.4
–32.1
–27.9
–27.5
µC/
µC/g
µC/g
µC/g
µC/g
µC/g
µC/g
µC/g
–2.2
–4.9
–8.9
–12.3
–12.8
–12.9
–11.5
–11.9
µC/g
µC/g
µC/g
µC/g
µC/g
µC/g
µC/g
µC/g
Humidity
sensitivity
a
7.0
4.92
3.41
2.47
2.36
2.49
2.43
2.31
(a) ratio of tribo value at 20% and 80% RH.
lar conclusion has also been reached by solid state C
NMR spectroscopy in a complimentary study . In that
study, in addition to the packing information, 13C NMR
data clearly show that the OH groups in the adsorbing
alcohol are interacting with the silica surface.12
13
Effect of Chain Length. By making the assumption
that each aliphatic hydrocarbon occupies ~ 25 Å 2, 8 a
series of alcohol-treated R972 silicas of varying alcohol chain length were prepared. The DSC and IR results along with the data of the pure alcohols are
tabulated in Table II. Both DSC and IR data in T able
II suggest that alcohol molecules are adsorbing on the
surface of R972 and that hydrocarbon layers are also
formed through Van der Waals interactions of the hydrocarbon chains. For 1-dodecanol, 1-tetradecanol, and
1-hexadecanol, the CH 2 stretching frequencies are either equivalent or lower than those of the pure materials, implying that a significant packing for the
hydrocarbon chains has occurred on the silica surface
as compared to the neat materials. The contrary is observed for 1-octadecanol, 1-eicosanol, and 1-decosanol;
relatively speaking, the hydrocarbon layers in these
systems are less organized on the silica surface. W e
suspect that as the chain length in these long chain
alcohols is increased, hydrophobic interaction may become more and more important, the thermal stability
gained by having the OH group interacting with the
silica surface is no longer dominating. A schematic of
the possible interactive state is given in Scheme 3. If
this molecular model is true, the surfaces of the 1octadecanol, 1-eicosanol and 1-decosanol treated silicas will be relatively hydrophilic. Incidentally, we find
that these silicas are less effective in imparting negative charging on SPAR toner and their charging is relatively humidity sensitive as compared to silicas treated
by 1-dodecanol, 1-tetradecanol and 1-hexadecanol. The
schematic provided in Scheme 3 is internally consistent with these observations.
Tribocharging of Alcohol Treated Silicas in Unpigmented SPAR Toners. The tribocharging properties
of the prepared alcohol-treated silicas were examined
by first blending them with a 9 µm unpigmented SPAR
toner and then determining their tribo values against
the metal beads at 20% and 80% RH at room temperature. The charging of the unpigmented SPAR toner and
the R972/SPAR toner were also studied as controls. The
results for the modified R972 silicas containing varying
concentrations of 1-hexadecanol are summarized in
Table III. The tribo results can qualitatively be rationalized based on the molecular models shown in Scheme
2. For example, the humidity sensitivity for the SP AR
toner and the SPAR toner containing R972 is between
298
Journal of Imaging Science and Technology
Scheme 3.
4.9 and 7. At 1-hexadecanol loading between 4.8% to
17%, we show that a hydrocarbon layer is formed due to
chain folding. In our charging study , we observe a
gradual increase in tribocharge at both 20% and 80%
RH. Simultaneously, the humidity sensitivity is also
reduced. Maximum tribo value and minimum humidity
sensitivity (ratio of the tribo values obtained at 20% and
80% RH) are obtained when the R972 silica is treated
with 17% 1-hexadecanol. This optimal performance corresponds to the adsorption of 1 monolayer of 1hexadecanol on the surface of R972. W e suggest that
adsorption of 1-hexadecanol on R972 renders the silica
surface more hydrophobic, leading to an increase in
negative charging and a reduction in humidity sensitivity in the charging process. 2 This rationalization is
supported by a water up-take study. We found that the
water uptake for R972 treated with 17% of 1hexadecanol is 0.17% from ~ 0% to 80% RH. The water
uptake is less that of the control (R972 silica), which
adsorbs 0.26% by weight of water under an identical
condition.
At 1-hexadecanol loading beyond 17%, both negative
charging and humidity sensitivity fall off from the optimal value. The results can be rationalized using the model
in Scheme 2C. For instance, when more than one monolayer of 1-hexadecanol is placed on the surface of R972,
crystallization of 1-hexadecanol occurs. We assume that
due to the more random orientation of 1-hexadecanol
molecules on the silica surface (Scheme 2C), hydrophobicity of the alcohol-modified surface will be less than
optimal, leading to the inferior charging results.
Law and Tarnawskyj
TABLE IV. Tribocharging Properties of Various Alcohol-Treated
R972 Silicas at a Monolayer Coverage
Alcohol
1-dodecanol
1-tetradecanol
1-hexadecanol
1-octadecanol
1-eicosanol
1-docosanol
Toner tribo
20% RH
80% RH
-
28.5
29.3
30.4
26.9
26.2
24.8
µC/g
µC/g
µC/g
µC/g
µC/g
µC/g
- 9.3
- 11.7
- 12.3
- 8.4
- 6.1
- 5.6
Humidity sensitivity
µC/g
µC/g
µC/g
µC/g
µC/g
µC/g
a
3.06
2.50
2.49
3.20
4.30
4.43
(a) ratio of tribo value at 20% and 80% relative humidity.
Table IV summarizes the results of the effect of the
chain length of the adsorbed alcohol on the charging in
SPAR toner. A chain length effect is observed and optimal tribos at 20% and 80% RH are obtained for 1hexadecanol. Theoretically, the number of alcohol
molecules that are on the silica surface is identical. The
variation in tribo suggests that the formed hydrocar bon layers are different among all the alcohols studied.
The increase in tribo from 1-dodecanol to 1-tetradecanol
to 1-hexadecanol seems to suggest that the protection
from the hydrocarbon layer increases as the chain length
increases in this regime. The tribo decreases as the chain
length increases for 1-octadecanol, 1-eicosanol and 1docosanol. The decrease in tribo and the increase in
humidity sensitivity on toner charging suggest that the
surfaces of these modified silicas are more hydrophilic
relative to their shorter chain analogs. We suggest that
the packing of the hydrocarbon chains for these long
chain alcohols is less ordered on the silica surface. Specifically, due to the increased hydrophobic interaction
(through the CH 2 groups) between the hydrocarbon
chains, these long chain alcohols may have been incorporated into the hydrocarbon layer with the OH group
pointing to both directions (Scheme 3). Again, hydrophobicity of the modified silica surface is a determining
factor for tribocharging and humidity sensitivity.
Conclusions
This work demonstrates that hydrocarbon layers are
formed when long chain alcohols are adsorbed on the
surface of silica particles. The formed hydrocarbon layers increase the hydrophobicity of the silica surfaces.
As a result, higher tribo values and reduction in humidity sensitivity in toner charging are obtained when
the silica is incorporated in SPAR toner. For R972 silica,
optimal charging results are obtained when one theoretical monolayer of 1-hexadecanol is adsorbed on the
silica surface. Both alcohol concentration and chain
length are shown to have an effect on toner charging.
The results can be rationalized based on the molecular
structure of the adsorbed alcohol layer on the silica sur
face. As it turns out, depending on the structure of the
adsorbing layer, the hydrophobicity of the alcohol-modified silica surface does vary. The charging performance
correlates well to the hydrophobicity of the silica sur face, the higher the hydrophobicity, the better the charging results.
Acknowledgements. The authors thank Denise Bayley
for the support and cross-evaluation of the alcoholtreated silicas described in this work and Ralph Mosher
for the use of his DSC.
References
1. R. J. Gruber and P. J. Julien, in Handbook of Imaging Materials , A. S.
Diamond, Ed., Marcel Dekker, Inc., New York, 1991, p. 159.
2. K. Y. Law and I. W. Tarnawskyj, J. Imaging Sci. Technol. 41, 550
(1997).
3. K. Y. Law, I. W. Tarnawskyj, P. J. Julien, and F. Lee, J. Imaging Sci.
Technol. 42, 459 (1998).
4. K. Y. Law and I. W. Tarnawskyj, J. Imaging Sci. Technol . 42, 579
(1998).
5. L. B. Schein, Electrophotography and Development Physics, SpringerVerlag, New York, 1988, p. 79
6. Degussa Technical Bulletin Pigments, No. 11, Basic Characteristics
of Aerosils.
7. R. K. Iler, The Chemistry of Silica, John Wiley and Sons, New York,
1979, p. 655.
8. G. L. Gaines, Insoluble Monolayers at Liquid Gas Interfaces,
Interscience, New York, 1966, p. 249.
9. R. G. Snyder, H. L. Strauss, and C. A. Elliger, J. Phys. Chem. 86,
5145 (1982).
10. R. G. Snyder, M. Maroncelli, H. L. Strauss, and V. M. Hallmark, J.
Phys. Chem. 90, 5623 (1982).
11. M. D. Porter, T. B. Bright, D. L. Allara, and C. E. D. Chidsey, J. Am.
Chem. Soc. 109, 3559 (1987).
12. Y. H. Chin, S. Kaplan, K. Y. Law and I. W. Tarnawskyj, unpublished
NMR results.
Effect of Adsorption of Long Chain Alcohol Molecules on Silica Particles on Toner Charging
Vol. 43, No. 3, May/June 1999
299
JOURNAL OF IMAGING SCIENCE AND TECHNOLOGY • Volume 43, Number 3, May/June 1999
Effect of Alcohol Grafting on the Charging Characteristics of Silicas in
Xerographic Toner
K.-Y. Law*,▲ and I. W. Tarnawskyj
Xerox Corporation, Wilson Center for Research and Technology, Webster, New York
A series hydrophobic silicas have been synthesized by reacting normal alcohols with hydrophilic fumed silica A130 (16-nm) at
high temperature. Elemental analysis, solvent washing study and IR analysis suggest that alcohols molecules are grafted onto
the silica surface. A rough estimation from the elemental data reveals that the surface area that is accessible for alcohol gra
fting
is ~ 4 % and the grafting process is independent of the chain length of the normal alcohol. Water wetting tests indicate that the
synthesized silicas are hydrophobic. Charging studies of the alcohol grafted silicas in unpigmented SP AR (a polyester) toner
reveal that they generally charge to a higher charge level at both low (20%) and high (80%) relative humidity (RH). The chargin
g
process appears to be less humidity sensitive as compared to analogous, commercially available hydrophobic silica, e.g., R972,
from Degussa. The length of the grafting group is shown to affect the charging performance and the molecular origin is proposed
and discussed. Two strategies to further improve the charging performance of alcohol-grafted silicas were also studied in this
work. We observed very similar charging levels and a reduction in humidity sensitivity when we grafted hydrophilic silicaA130
with branched alcohols. We observed very interesting humidity effect when we modified a 1-dodecanol graftedA300 silica with 1hexadecanol. Relatively speaking, the charge level of the modified silica is found to be lower than that of the unmodified graf
ted
silica at 20% RH. The opposite is observed at 80% RH. This results in a developer that has very little humidity sensitivity in
toner charging. The insensitivity of the toner charging process is rationalized.
Journal of Imaging Science and Technology 43: 300–305 (1999)
Introduction
Non-porous fumed silicas ranging from 5 to 50 nm in diameter are commonly used as flow-aids in xerographic
toners.1 In practice, they are performing a dual function.
They enhance the flow while controlling the charging char
acteristics of the toner. One of the drawbacks of the approach is the humidity sensitivity of the toner charging
process, e.g., the tribo of the toner decreases as the relative humidity (RH) in the ambient increases. We reported
earlier that both hydrophilic and hydrophobic silicas are
negative charge control additives in toner. The toner charging process is sensitive to humidity and expectedly hydrophilic silica is more humidity sensitive between the two.2
The correlation between hydrophobicity of the silica sur face and the humidity sensitivity (ratio of the charge level
at 20% RH to that at 80% RH) is quite profound. W e recently attempted to modify the charging characteristics of
silica particles by treating them with conventional charge
control additives. While the effect of the added charge control additives on toner charging is evident, the humidity
sensitivity of the charging process remains. 3,4 A possible
solution to eliminate or reduce humidity sensitivity in
Original manuscript received October 6, 1998
▲ IS&T Member
* To whom correspondence should be addressed
© 1999, IS&T—The Society for Imaging Science and Technology
300
toner charging is to modify the surface of silica particles.
It is believed that if the moisture content of the silica surface is maintained at a constant level as the ambient humidity is varying, the toner charging process may become
humidity insensitive.
Recently, we 5 reported an investigation on the use of
long chain alcohol to modify the surface of hydrophobic
silicas. The results of that study showed that, although
adsorption of long chain alcohol on the surface of silica
particles improves toner charging, the humidity sensitivity of the charging process is still far from idea. Specifically, toner charge decreases by more than a factor
of 2 as the ambient RH for the charging process increases
from 20% to 80%. In this work, we report the synthesis
of a series of novel hydrophobic silicas by grafting normal alcohols of different chain length onto the surface
of hydrophilic fumed silica A130 (from Degussa). The
alcohol-grafted silicas are found to be hydrophobic in
the water-wetting test. Charging studies reveal that the
synthesized hydrophobic silicas out-perform an analogous commercial silica, R972, from the charge level and
humidity sensitivity viewpoints. The charging perfor mance is shown to correlate to the structure of the
grafted alcohol and optimal results are observed for the
1-dodecanol grafted A130 silica. Since elemental analysis suggest that the accessible surface area for alcohol
grafting is ~ 4%, we have devised two strategies to further improve the toner charging process. One of the
approaches is to graft branched alcohol onto the silica
surface and the other approach is to “cover up” the
Scheme 1.
grafted silica surface with 1-hexadecanol. A significant
improvement in humidity sensitivity is obtained for a
1-hexadecanol modified alcohol-grafted silica. Humidity sensitivity as low as 1.28 is achieved. The origin for
the humidity insensitivity is discussed.
Experimental
Materials. The fumed silica samples (A130, A300, R972
and R812) used in this work were from Degussa. 1-Butanol,
1-octanol, 1-dodecanol, 1-hexadecanol, 1- octadecanol, and
hexadecane were the highest commercial quality pur chased from Aldrich Chemical Co. 2-Butyl-1-octanol and
2-pentyl-1-nonanol were bought from Pfaltz and Bauer .
Diundecylcarbinal was synthesized by reducing diundecyl
ketone (Pfaltz and Bauer) with lithium aluminum hydride
(Aldrich). The washing solvents were hexane and methanol; they were reagent grade from Fisher. The toner resin
was a linear polyester SP AR from Goodyear. The toner
was unpigmented and was prepared by a melt-extrusion
and jetting technique. It was classified to ~ 9µm. The metal
beads for the charging experiments were made of steel
core (~ 130 µm) and were solution-coated by a carbon-black
doped poly(methyl methacyrate) polymer on the surface
at a total weight loading of ~ 0.8%.
Preparation of Alcohol-Grafted Silicas. A hydrophilic silica A130, 3 g, was activated in a furnace at ~
600 C for 3 – 4 h. It was then transferred to a 250-mL
three-neck flask containing a mixture of 1-dodecanol (50
mL) and hexadecane (50 mL). The resulting dispersion
was stirred and was heated in an oil bath at a bath temperature of ~ 270 C under a nitrogen atmosphere overnight (16 h). The mixture was then cooled to room
temperature and the silica product was isolated by filtration. After washing the product thoroughly with
methanol and hexane, the white solid obtained was dried
in a vacuum oven at ~ 80 C overnight. A hydrophobic
silica was obtained, yield ~ 3.3 g. IR spectroscopy suggested the presence of hydrocarbon chains in the silica
sample and the hydrocarbon chains could not be removed
even after vigorous washing with methanol. The silica
product was made fluffy by roll milling it with steel shot
(35 g, 1/4” diameter) inside a 4-oz bottle for 30 min.
Different alcohol-grafted silicas could be prepared by
changing the alcohol to 1-butanol, 1-octanol, 1-hexadecanol, 1-octadecanol or other branched alcohols; or by
the use of a different hydrophilic silica, e.g., A300 from
Degussa.
Preparation and Evaluation of Experimental
Toners. The above alcohol-grafted silica (0.063 g), the
SPAR toner (12.5 g) and 125 g of 1/4" steel shot were
placed inside a 4-oz bottle and was rolled for 30 min to
prepare an experimental toner. Developer was prepared
by placing the above toner (1.25 g) and the metal beads
(60 g) inside a 2-oz bottle. The entire content was conditioned inside a humidity controlled glove box at a constant humidity, either 20% or 80%, overnight and was
sealed. The tribo of the toner was generated by tumbling the toner and the metal beads inside the 2-oz bottle
on a roll-milled for 5 min at a speed of 90 ft/min. The
charge generated was determined by the standard blowoff technique inside a Faraday cage under lab ambient
condition. 6 Controlled experiment indicated that the
charging process was leveled off after 5-minute of roll
milling and the tribo value was primarily sensitive to
the ambient RH when the toner was charged up.
Results and Discussion
Preparation of Alcohol-Grafted Silicas. Fumed
silicas are genuinely hydrophilic and the surface of these
materials is known to consist of isolated and geminal
silanol groups. 7 In this work, hydrophilic fumed silica
A130 (16 nm diameter, BET surface area 130 m2/g) was
chosen for model study . A130 was first activated in a
furnace at 600 - 700 C to remove any surface adsorbed
water. It was then reacted with an alcohol at high temperature to form a condensed product, the surface “ester”.8 Because water is the by-product of the synthesis,
for convenience, 1-hexadecane is introduced as an
azeotropic co-solvent. The procedure is analogous to that
described by Ballard and co-workers.9 A schematic of the
reaction is shown in Scheme 1.
A series of normal alcohols were studied. The silica
product was isolated by filtration and was made fluffy
by ball-milling the white solid with steel shots. IR spectral data suggest the presence of aliphatic hydrocarbon
chains on the silica surface. Evidence for the occurrence
of grafting comes from a solvent washing study, where
we observed absolutely no weight loss from these
samples after extensive and vigorous washing with
methanol. The silica products are hydrophobic as judged
by their water wetting property, which is in contrast to
the starting silica (A130). The analytical data of all the
synthesized grafted silicas are tabulated in Table I.
The analytical data for A130 was determined as control. The data shows that its silicon content (37.51%) is
Effect of Alcohol Grafting on the Charging Characteristics of Silicas in Xerographic Toners
Vol. 43, No. 3, May/June 1999 301
TABLE I. Analytical Data of Alcohol-Grafted Silicas Synthesized
From A130.
Alcohol
C
H
TABLE II. Tribocharging properties of various alcohol-grafted
A130 silicas in SPAR toner.
Si
Grafted-silica
None (A130)
1-butanol
1-octanol
1-dodecanol
1-hexadecanol
1-octadecanol
Calc’d:
Found:
Found:
Found:
Found:
Found:
Found:
—
—
1.70
4.46
5.73
6.70
7.03
—
—
< 0.5
0.53
0.70
0.78
0.94
46.74
37.51
45.16
44.09
42.82
41.25
40.87
Toner tribo
20% RH
80% RH
A130 control
–25.1 µC/g
R972 control
–26.4 µC/g
None (SPAR toner only) –15.4 µC/g
A130/1-butanol
A130/1-octanol
A130/1-dodecanol
A130/1-hexadecanol
A130/1-octadecanol
–28.6
–29.0
–32.4
–30.6
–29.2
µC/g
µC/g
µC/g
µC/g
µC/g
Humidity sensitivity a
–1.9 µC/g
–4.9 µC/g
–2.2 µC/g
13.2
5.39
7.0
–6.6 µC/g
–7.1 µC/g
–11.1 µC/g
–10.5 µC/g
–10.9 µC/g
4.33
4.08
2.92
2.91
2.68
(a) ratio of tribo value at 20% and 80% relative humidity.
significantly lower than the theoretical value of SiO 2,
46.74%. The discrepancy is by no means surprising since
A130 has a very large surface area and it is expected
that the elemental content on the surface would be different from that in the bulk, which is essentially SiO 2.
Using a water uptake experiment, we found earlier that
A130 adsorbs a considerable amount of water under
ambient conditions.2 We believe that a routine drying
in the analytical lab (e.g., 100 C at 0.02 mmHg) may not
be vigorous enough to remove all the surface water.10
The analytical data in T able I show that the carbon
content in the grafted silica increases as the chain length
of the alcohol increases. From the carbon content in the
alcohol-grafted sample, one can estimate the percentage of the surface area that is grafted by the alcohol. For
instance, the surface area for A130 is ~1.3 × 1022 Å2. 7 For
the 1-butanol grafted material, the surface area that is
covered by the butoxy groups is estimated to be ~ 0.05×
1022 Å2, if one assumes that each hydrocarbon chain only
occupies ~ 25 Å 2.11 The fraction of surface area that is
grafted by the butoxy group is ~ 4%. If one corrects for
the formula weight of the alkoxy group, the analytical
data in Table I actually suggests that the fraction of
surface area that has been grafted with alcohol is approximately 4 ± 0.7% among all the normal alcohols investigated. Admittedly, the reaction temperature for the
1-butanol and 1-octanol grafted experiments may have
been lower than the other grafting experiments because
of their lower boiling points. We assume that longer 1alcohols may be less reactive because of the steric effect. These two effects may compensate for each other
and results in a similar degree of grafting within the
series of normal alcohols studied.
Tribocharging of Alcohol-Grafted Silicas in Unpigmented SPAR Toners. The tribocharging properties of
the synthesized alcohol-grafted silicas were evaluated in
an experimental unpigmented SPAR toner. The toner was
first prepared by blending the alcohol-grafted silica with
the 9 µm SPAR toner. Developer was formulated using the
toner and the metal beads. It was then conditioned at a
constant RH, either at 20% or 80%, overnight. The toner
was charged by tumbling it with metal beads on a rollmill and the generated charge was determined using the
blow-off technique.6 The charging data for all the alcoholgrafted silicas studied in this fashion are listed in T able
II. The charging of the unpigmented SP AR toner is included as control. Also included in the controls are the
charging properties of silicasA130 and R972 from Degussa.
It is important to note that R972 is a commercially available hydrophobic silica and it was synthesized by
silanating A130 with dichlorodimethylsilane.7 The surface
areas among A130, R972 and all the synthesized alcoholgrafted silicas in Table II are therefore comparable.
The charging data of the three control toners indicate
that, (1) indeed both hydrophilic and hydrophobic silicas are negative charging agents for the SP AR toner,
(2) hydrophilic silica is less potent in toner charging,
and (3) the charging involving hydrophilic silica is more
humidity sensitive. We also reached the same conclusion when we studied the charging of silicas in styrenebutadiene toner. 2 Comparison of charging between the
alcohol-grafted silicas and that of R972 suggests that
the alcohol-grafted silicas synthesized in this work are
superior. The alcohol-grafted silicas not only impart
more negative charges at both 20% and 80% RH, they
are also less humidity sensitivity too.A detailed examination of the data shows that the negative tribo increases initially as the chain length of the grafting group
increases. A maximum charge level is attained for the
1-dodecanol grafted material and the tribo decreases as
Scheme 2.
302
Journal of Imaging Science and Technology
Law, et al.
TABLE III. Tribocharging Properties of a 1-dodecanol-Grafted
A300 Silica in SPAR Toner.
Grafted-silica
Toner tribo
20% RH
80% RH
Humidity sensitivity a
A300 control
–33.1 µC/g
R812 control
–39.1 µC/g
None (SPAR toner only) –15.4 µC/g
–6.1 µC/g
–11.4 µC/g
–2.2 µC/g
5.43
3.43
7.0
–37.3 µC/g
–15.0 µC/g
2.49
A300/1-dodecanol
(a) ratio of tribo value at 20% and 80% relative humidity.
the chain length of the graft material becomes longer .
The observation may be rationalized based on the molecular model described in Scheme 2.
Logically, there exist alkoxy groups on the silica surface after grafting. The alkoxy groups are expected to
lie on the silica surface as depicted in Scheme 2A. Since
hydrocarbon chains are hydrophobic, the grafted
alcohols thus offer some level of protection for the silica
surface from moisture. The improvement in charging
and humidity sensitivity relative toA130 is anticipated.
The data in Table II suggests that the level of protection provided by the grafted alcohols is actually better
than the hydrophobic groups in R972.
As estimated earlier, only 4% of the surface area is
grafted. According to Scheme 2A, the protection generated by the grafted material should increase as the chain
length of the alkoxy group increases. Indeed, this is observed. We observe an increase in charge level as the
chain length increases from 1-butanol to 1-dodecanol.
Presumably at some point when intra-chain hydrophobic interaction becomes important, chain folding will
occur (Schematic 2B). The folding of the hydrocarbon
chain reduces its effectiveness in protecting the silica
surface. The lower toner tribo observed for the 1hexadecanol and 1-octadecanol grafted materials may
be indicative of chain folding. In the other words, the
result suggests that chain folding becomes feasible when
the chain length is longer than 16 carbons. We also obtained evidence of chain folding in a previous investigation when we studied the adsorption of 1-hexadecanol
on the surface of silica R972.5
We have extended the grafting process to a smaller
hydrophilic silica, A300 (7 nm).7 The 1-dodecanol grafted
material was prepared analogously and the charging of
the material was studied using the protocol described
above. The data is given in Table III along with those of
the controls. The tribo values for these silicas are more
negative than those in T able II and the observation is
attributable to the surface area effect.2 In Table III, the
control hydrophobic silica is R812 and it is made hydrophobic by silanating A300 with hexamethyldisilizane. 7
While the charge levels for the 1-dodecanol graftedA300
silica and that of R812 are comparable at 20% RH, there
is a significant difference in charge level at 80% RH.
The grafted A300 silica was shown to charge significantly
better, - 15.0 versus - 11.4 µC/g. As a result, the humidity
sensitivity for the grafted material reduces substantially.
Approaches for Further Improvements in Toner
Charging. Comparing to commercially available hydrophobic silicas, the alcohol-grafted silicas synthesized so
far are more potent in imparting negative charging in
SPAR toner. Even more importantly, the charging appears
to be less humidity sensitive. However , even in the best
case, the humidity sensitivity is more than a factor of 2
and there is a need of further improvement. Since results
in Table I suggest that only 4% of the surface is accessible
for alcohol grafting and since the charging data suggest
that the charging will improve if the level of protection is
increased, we have devised two schemes to improve the
protection of the silica surface from moisture in the ambient. The results are summarized as follows.
(a)Branched alcohols. Because the accessible area for
alcohol grafting is ~ 4%, the only way to enhance the
surface protection with the same number of grafting
site is to use branched alcohols. The concept is
graphically illustrated in Scheme 3.
One can envision that there exist two types of
branches, namely V-type and Y-type. They both
should improve the surface protection. Diundecylcarbinol, 2-butyl-1-octanol and 2-pentyl-1-nonanol
were the chosen branched alcohols for this work
because of their commercial availability . Grafted
silicas were synthesized by reacting these alcohols
with A130 at high temperature. The resulting silicas
were hydrophobic and their charging properties were
studied as described previously. The data (Table IV)
shows that, although mixed results are obtained at
low humidity (20%), notable improvement in
tribocharging is observed at 80% RH as compared to
the 1-dodecanol grafted material. As a result, the
humidity sensitivity reduces to 2.34 in the case of
Scheme 3.
Effect of Alcohol Grafting on the Charging Characteristics of Silicas in Xerographic Toners
Vol. 43, No. 3, May/June 1999 303
TABLE IV. Tribocharging Properties of Various Types of Alcohol-Grafted Silicas in SPAR Toner.
Toner tribo
Grafted-silica
Type of grafting
A130 control
R972
None (SPAR only)
A130/1-dodecanol
A130/diundecylcarbinol
A130/2-butyl-1-octanol
A130/2-pentyl-1-nonanol
linear
V-type
Y-type
Y-type
20%
80%
Humidity sensitivity a
–25.1 µC/g
–26.4 µC/g
–15.4 µC/g
–1.9 µC/g
–4.9 µC/g
–2.2 µC/g
13.2
5.39
7.0
–32.4
–34.4
–31.5
–31.4
µC/g
µC/g
µC/g
µC/g
–11.1 µC/g
–12.6 µC/g
–11.4 µC/g
–13.4 mC/g
2.92
2.73
2.76
2.34
(a) ratio of tribo value at 20% and 80% relative humidity.
TABLE V. Effect of Adsorption of 1-Hexadecanol on the Surface of Alcohol-Grafted Silicas on Toner Charging.
Toner tribo
Grafted-silica
1-hexadecanol adsorption
20%
80%
Humidity sensitivity
A130/1-dodecanol
None
9.1% by wt.
–32.4 µC/g
–36.4 µC/g
–11.1 µC/g
–17.1 µC/g
2.92
2.12
A300/1-dodecanol
None
9.15%
17%
–37.3 µC/g
–29.5 µC/g
–25.4 µC/g
–15.0 µC/g
–17.1 µC/g
–19.8 µC/g
2.49
1.72
1.28
a
(a) ratio of tribo value at 20% and 80% relative humidity.
the 2-pentyl-1-nonanol grafted material. The
improvement is significant when compared to the
humidity sensitivity of R972, which is 5.39.
(b)Alcohol Adsorption. In a previous report, we 5
showed that when molecules of long chain alcohols
are adsorbed on the surface of hydrophobic silicas, a
hydrocarbon layer is formed via V an der W aals
interaction of the hydrocarbon chains. This layer acts
as a barrier , protecting the silica surface from
moisture “attack”. The resulting modified silica was
found to be more potent in terms of imparting
negative charging and the charging process is less
humidity sensitive. In this work, we attempt to
strengthen the surface protection of the alcoholgrafted silica using the same approach. Two alcoholgrafted silicas, namely the 1-dodecanol grafted A130
and the 1-dodecanol grafted A300, were examined and
1-hexadecanol was the adsorbing alcohol. These
modified silicas were prepared using the solution
coating method as detailed in a previous publication.5
The charging characteristics of the modified silicas
were studied in SP AR toner and the results are
summarized in Table V. For the 1-dodecanol grafted
A130 silica, adsorption of 9.1% of 1-hexadecanol
enhances the tribocharge at both 20% and 80% RH.
The humidity sensitivity reduces to 2.12. This is the
lowest value recorded for all the 16-nm fumed silicas
studied, including A130, R972 and the 1-dodecanol
grafted A130 silica.
Very interesting results are obtained for the 1dodecanol grafted A300 silica. After modification of the
silica surface with 1-hexadecanol (at 9.1% and 17% by
wt.), the tribocharge was found to decrease as compared
to the starting material at 20% RH, e.g., from - 37.3µC/
g to - 29.5 and - 25.4µC/g for 9.1 and 17% 1-hexadecanol
adsorption, respectively. This is unexpected, since alcohol adsorption generally increases the hydrophobicity
of the silica surface and enhances negative charging in
SPAR toner. 5 On the other hand, the charging for the 1-
304
Journal of Imaging Science and Technology
hexadecanol modified silicas is relatively insensitive to
RH. For example, the charge levels decreases from - 25.4
to - 29.5 µC/g at 20% RH to - 17.1 to - 19.8 µC/g at 80%
RH. The net result is very extraordinary, because now the
humidity sensitivity for these two new silicas is in the
range of 1.28 to 1.72. Because tribocharge at ~ 20 µC/g is
already useful for xerographic development. A humidity sensitivity of 1.28 suggests that the charging system is practically insensitive to humidity.
In our previously investigation on the adsorption of
1-hexadecanol on hydrophobic silica surface, we 5 demonstrated that it is possible to adsorb alcohol molecules
in two different directions. When the interaction between the OH group in alcohol and the silica surface is
dominant, 1-hexadecanol will be adsorbed “heads down”
onto the silica surface (Scheme 4A). On the other hand,
when hydrophobic interaction between the hydrocarbon
chains becomes important, the alcohol adsorption may
be less specific and adsorption with “heads up” will occur (Scheme 4B). The latter will decrease the hydrophobicity of the adsorbed layer, resulting in a lower charge
level.5
We suggest that, for the 1-dodecanol grafted A300
silica, adsorption of 1-hexadecanol on the silica surface
may be non-specific. This lowers the hydrophobicity of
the modified silica surface and results in a decrease in
toner charge at 20% RH. On the other hand, the hydrocarbon layer that is formed from the hydrocarbon chains
is still effective in terms of protecting the silica surface
from moisture. The humidity insensitivity observed for
the 1-hexadecanol modified grafted A300 silica is therefore the result of a lower charge level at 20% RH and a
slight charge enhancement at 80% RH.
Summary and Remarks
This work demonstrates that hydrophilic silicas can be
rendered hydrophobic by grafting alcohols on their sur faces. The surface area that is accessible for grafting is
estimated to be ~ 4%. The charging property of these alcohol-grafted silicas has been studied in unpigmented SPAR
toner. Results show that grafting alcohols onto the sur -
Law, et al.
Scheme 4
face of silica is more effective in imparting negative charging in SPAR toner as compared to commercially available
hydrophobic silicas of the same particle size. Comparatively, the alcohol-grafted silicas are found to be more negative going at both low (20%) and high (80%) relative
humidity. A reduction in humidity sensitivity in the charging process is thus obtained.
The potency in imparting negative charging in SPAR
toner is shown to correlate to the structure of the grafting group. Examination of the charging results for a
series of A130 grafted silicas suggest that there exists a
chain length effect on charging. The effectiveness in imparting negative charging increases as the chain length
increases initially and an optimal charging is obtained
for the 1-dodecanol material. For longer grafting group,
the charge level starts to decrease. The results are rationalized based on the degree of protection that is generated by the grafting group. For example, for short
grafting groups such as the butoxy group, the protection provided to the silica surface would increase as the
chain length increases. However , as the chain length
becomes longer, intra-chain hydrophobic interaction may
become dominant. This hydrophobic interaction results
in chain folding, decreasing the protection efficiency of
the grafting group. Indeed, the charge level is shown to
decrease for grafting groups with more than 16 carbons.
While alcohol grafting represents an improved method
to synthesize hydrophobic silica for toner, the charging
is still quite humidity sensitive. T wo strategies were
developed to further improve the charging performance.
Evidence is provided that the surface protection can be
enhanced via branched alcohols. We found that both the
charging value and the humidity sensitivity are improved when A130 is grafted with 2-pentyl-1-nonanol.
Alternatively, improvement can also be obtained by introducing a hydrocarbon layer via the adsorption of 1hexadecanol on the surface of the alcohol-grafted silica.
We discovered that when 1-hexadecanol is adsorbed on
the surface of a 1-dodecanol graftedA300 silica, the tribo
values at 20% RH actually is decreased. On the other
hand, the tribo value at 80% is not affected. The charge
levels at 20% and 80% are –25.4 and –19.8 µC/g, respectively, in the best case. This charge level is very
respectable for xerographic development and the charging process is practically humidity insensitive. We have
studied over 100 commercial and synthetic silicas or
mixture throughout this work, the humidity sensitivity
recorded in this work, 1.28 in SPAR toner, is the lowest
thus far.
References
R. J. Gruber and P. J. Julien, in Handbook of Imaging Materials, A. S.
Diamond, Ed., Marcel Dekker, Inc., New York, 1991, p. 159.
2. K. Y. Law and I. W. Tarnawskyj, J. Imaging Sci. Technol. 41, 550 (1997).
3. K. Y. Law, I. W. Tarnawskyj, P. J. Julien, and F. Lee, J. Imaging Sci.
Technol. 42, 459 (1998).
4. K. Y. Law and I. W. Tarnawskyj, J. Imaging Sci. Technol. 42, 579 (1998).
5. K. Y. Law and I. W. Tarnawskyj, J. Imaging Sci. Technol. 43, 299 (1999).
6. L. B. Schein, Electrophotography and Development Physics, SpringerVerlag, New York, 1988, p. 79.
7. Degussa Technical Bulletin Pigments, No. 11, Basic Characteristics of
Aerosils.
8. R. K. Iler, The Chemistry of Silica, John Wiley and Sons, New York,
1979, p. 573.
9. C. C. Ballard, E. C. Broge, R. K. Iler, D. S. St. John, and J. R. McWhorter,
J. Phys. Chem. 65, 20 (1961).
10. R. K. Iler, The Chemistry of Silica, John Wiley and Sons, New York,
1979, p. 625.
11. G. L. Gaines, Insoluble Monolayers at Liquid Gas Interfaces ,
Interscience, New York, 1966, p. 249.
1.
Effect of Alcohol Grafting on the Charging Characteristics of Silicas in Xerographic Toners
Vol. 43, No. 3, May/June 1999 305
The Journal of
Imaging Science and
Technology CD
Editor: Volume 40, 1996—Vivian Walworth
Editor: Volume 41, 1997—Melville R. V. Sahyun
The Journal of Imaging Science and
Technology is dedicated to the advancement of knowledge in the imaging sciences, in practical applications
of such knowledge, and in related fields
of study. The pages of this journal are
open to reports of new theoretical or
experimental results and to comprehensive reviews.
Vol. 40, 1996 and Vol. 41, 1997
now available on CD
Cost is as follows:
IS&T Member
Non-member
Institution
CD-ROM
$45.00
$55.00
$100.00
Vol. 42, 1998 and Vol. 43, 1999
available Spring, 2000
306
Journal of Imaging Science and Technology
The Business Directory
AUTOMATE
TECHNOLOGY OF MATERIALS
COLOR & DENSITY MEASUREMENTS
of Proofs and Calibration Sheets
The Laboratory for all your
Analytical needs
X-Y scanning stages and software for measuring sheets up to 30′′ × 40′′ with the handheld instruments that you presently use to
make measurements manually
David L. Spooner, PE
rhoMetric Associates, Ltd.
2918 N. Franklin Street
Wilmington, DE 19802-2933
(302) 764-9045• FAX (302) 764-5808
e-mail: rhomet@delanet.com
• X-ray Diffraction
• Pigment and Mineral ID & Quantification
• Electron Microprobe
• Elemental Analysis
• Small Spot Analysis
• SEM, FTIR, Thermal & Petrographic
2922 De La Vina St.
Ste. C
Santa Barbara, CA 93105
Ph: (805) 569-3382
1-800-682-2922
www.rain.org/~iyengar
Edgar B. Gutoff, Sc.D., P.E.
PAUL J. MALINARIC
consulting engineer
Consulting Chemical Engineer
194 Clark Road, Brookline, MA 02445-5848
Phone/Fax: 617-734-7081
ELECTROSTATIC CONSULTING ASSOCIATES
consulting in:
• electrostatics • static problems
• electrographics • xerographics • dielectric materials
Call to Inquire for Fax number
Telephones
556 Lowell Rd.
Groton, MA 01450
Business: 978-448-5485
Home:
978-448-5583
• Consulting in slot, slide, curtain, and roll
coating; in coating die design, and in
drying technology.
• Drying Software
• Coating Seminars
Co-authored Coating and Drying Defects (1995),
Modern Coating and Drying Technology (1992),
and The Application of SPC to Roll Products (1994).
EPPING GmbH—
PES Laboratorium
Measurement Devices for
Charge, Conductivity and Magnetic
Parameters for Powders.
R & D & E for Electrostatic Powder Physics
Contact:
Andreas J. Kuettner
Carl-Orff-Weg 7
85375 Neufahrn bei Freising
Germany
TEL +49-8165-960 35
FAX +49-8165-960 36
E-mail rhepping@t-online.de
Business Directory Ad—Six Issues
Journal of Imaging
Science and Technology
IS&T Mbrs. U.S. $100
Non-Mbrs. U.S. $150
Contact: Pam Forness
IS&T, 7003 Kilworth Lane
Springfield, VA 22151
703-642-9090: FAX: 703-642-9094
E-mail: pam@imaging.org
IS&T Recent Progress Series
Keeping up with the latest technical information is a task that becomes increasingly difficult. This is not only caused by
the large amount of information, but also
by its dispersed distribution at a variety
of conferences. The “Recent Progress” series collects, through the eyes of the Soci-
ety for Imaging Science and Technology,
technical information from several conferences and publications into a concise treatise of a subject. This series allows the professional to stay up-to-date and to find the
relevant data in the covered field quickly
and efficiently.
Available Now
• Recent Progress in Color
Management and Communication,
1998 (Mbr. $65; Non-Mbr. $75)
• Recent Progress in Ink-Jet
Technologies
1996 (Mbr. $55; Non-Mbr. $65)
• Recent Progress in Toner Technology • Recent Progress in Digital Halftoning
1997 (Mbr. $65; Non-Mbr. $75)
1995 (Mbr. $55; Non-Mbr. $65)
• Recent Progress in Color Science
1997 (Mbr. $65; Non-Mbr. $75)
• Recent Progress in Digital Halftoning II
1999 (Mbr. $65; Non-Mbr. $75)
Positions Available and
Positions Wanted
Can now be found
on the
IS&T homepage
Please visit us at
http://www.imaging.org
click on
Available soon
• Recent Progress in Ink Jet Technologies II
1999 (Mbr. $65; Non-Mbr. $75)
Plus shipping & handling: $4.50 U.S.;
$8.50 outside the U.S.A.
EMPLOYMENT
OPPORTUNITIES
Contact IS&T to order Today!
Phone: 703-642-9090
Fax: 703-642-9094
E-mail: info@imaging.org
www.imaging.org
Vol. 43, No. 1, March/April 1999
307
IS&T—The Society for Imaging Science and Technology
1998-1999 Board of Directors
President
ROBERT GRUBER, Xerox Corporation, 800 Phillips Road, W114-40D, Webster, NY 14580
Voice: 716-422-5611
FAX: 716-422-7716
e-mail: rgruber@crt.xerox.com
Executive Vice President
JOHN D. MEYER, Hewlett Packard Laboratories, 1501 Page Mill Rd., 2U-19, P.O. Box 10490, Palo Alto, CA 94304
Voice: 650-857-2580
FAX: 650-857-4320
e-mail: meyer@hpl.hp.com
Conference Vice President
WAYNE JAEGER, Tektronix, M/S 61-IRD, 26600 S. W. Parkway, Wilsonville, OR 97070-1000
Voice: 503-685-3281
FAX: 503-685-4366
e-mail: wayne.jaeger@tek.com
Publications Vice President
REINER ESCHBACH, Xerox Corporation, 800 Phillips Road, 0128-27E, Webster, New York 14580
Voice: 716-422-3261
FAX: 716-422-6117
e-mail: eschbach@wrc.xerox.com
Vice Presidents
JAMES KING, Adobe Systems Inc., 345 Park Ave., MS: W14, San Jose, CA 95110-2704
Voice: 408-536-4944
FAX: 408-536-6000
e-mail: jking@adobe.com
JAMES R. MILCH, Eastman Kodak Company Research Labs, 343 State Street, Rochester, New York 14650
Voice: 716-588-9400
FAX: 716-588-3269
e-mail: jrmilch@kodak.com
W. E. NELSON, Texas Instruments, P. O. Box 655474, MS 63, Dallas, TX 75265
Voice: 972-575-0270
FAX: 972-575-0090
e-mail: wen@msg.ti.com
SHIN OHNO, Sony Corporation, Business & Professional Systems Co., 4-14-1 Okata, Atsugi 243, Kanagawa 243-0021, Japan
Voice: 81-462-27-2373
FAX: 81-462-27-2374
e-mail: shin@avctl.cpg.sony.co.jp
MELVILLE R. V. SAHYUN, Department of Chemistry, University of Wisconsin, Eau Claire, WI 54702
Voice: 715-836-4175
FAX: 715-836-4979
e-mail: sahyunm@uwec.edu
DEREK WILSON, Coates Electrographics, Ltd., Norton Hill, Midsomer Norton, Bath, BA3 4RT, England
Voice: 44-1761-408545
FAX: 44-1761-418544
e-mail: derek.wilson@msn.coates.co.uk
Secretary
BERNICE ROGOWITZ, IBM Corp., T. J. Watson Research, P. O. Box 704, M/S H2-B62, Yorktown Heights, NY 10598-0218
Voice: 914-784-7954
FAX: 914-784-6245
e-mail: rogowtz@us.ibm.com
Treasurer
GEORGE MARSHALL, Lexmark International, Inc., 6555 Monarch Rd., Dept. 57R/031A, Boulder, CO 80301
Voice: 303-581-5052
FAX: 303-581-5097
e-mail: toner@lexmark.com
Immediate Past President
JAMES OWENS, 3 Woods End, Pittsford, NY 14534
e-mail: jcowens@post.harvard.edu
Executive Director
CALVA LEONARD, IS&T, 7003 Kilworth Lane, Springfield, VA 22151
Voice: 703-642-9090
FAX: 703-642-9094
e-mail: calva@imaging.org
CHAPTER DIRECTORS
Boston, Massachusetts (BO)
LYNNE CHAMPION, 1265 Main Street, W4-2M, Waltham, MA 02454
Korea (KO)
J.-H. KIM, Pusan National Univ., Dept. of Elec. Eng., San 30 Chang Jun-Dong, Kumjung, Pusan 609-735, Korea
Rochester, New York (RO)
JOANNE WEBER, Eastman Kodak Company, 234 Rhea Crescent, Rochester, NY 14615-1212
Tokyo, Japan (JA)
YOICHI MIYAKE, Dept. of Information and Computer Sci., Faculty of Engineering, Chiba Univ., 1-33 Yayoicho, Inage-ku, Chiba 263, Japan
Tri-State
JAMES CHUNG, Fuji Photo Film, 555 Taxter Road, Elmsford, NY 10523
308
Journal of Imaging Science and Technology
IS&T’s NIP15:
International
Conference on
Digital Printing
Technologies
The Caribe Royal Resort Suites,
Lake Buena Vista, Florida
General Chair:
Michael Lee
October 17–22, 1999
Come and meet us in Florida for IS&T's NIP15: International Conference on
Digital Printing Technologies. Over the years, the NIP Conferences have
emerged as the preeminent forum for discussion of advances and directions in
the field of non-impact and digital printing technologies. A comprehensive
program of more than 150 contributed papers from leading scientists and
engineers, is planned along with daily keynote addresses, an extensive program of 26 tutorials, a print gallery and an exhibition of digital printing products, components, materials and equipment. Following the presentations
each day, the authors will be available for one-on-one discussions. The preliminary program is now available at www.imaging.org.
The Society for Imaging Science and Technology
7003 Kilworth Lane, Springfield, VA 22151 USA
703-642-9090; FAX:703-642-9094;
Email:info@imaging.org; www.imaging.org
• Toner Based Marking: Processes
• Toner Based Marking: Materials
• Optoelectronic Imaging Materials
and Devices
• Ink-Jet Processes
• Ink-Jet Materials
• Printing Systems—Engineering/
Optimization
• Media for Digital Printing
• Print and Image Quality
• Color Science/Image Processing
• Advanced and Novel Printing
• Specialty Printing Applications
• Thermal Printing
• Liquid Toner Processes and Materials
• Textile and Fabric Printing
• Computer-to-Plate Technology
• Digital Printing Instrumentation
• Environmental Issues
7th IS&T/SID
Color Imaging Conference
Color Science, Systems and Applications
November 16–19, 1999 • The SunBurst Resort • Scottsdale, Arizona
General Co-Chairs: Jack Holm, Hewlett Packard (IS&T)
and Todd Newman, Canon Information Systems (SID)
The Color Imaging Conference is the premier technical conference for scientists and engineers working in the areas of color science, color engineering and their application to color
products and color imaging technology. 1999 marks the seventh year of this topical, annual
conference. The conference is international in nature. In previous years one third of the
participants came from outside the United States and Canada. The range of professional disciplines represented includes: digital photography, color
science, color engineering, image processing, color reproduction, prepress, display design, computer simulation, data visualization in color, psychophysics, optical physics, virtual reality, systems engineering, software applications development, and hardware development. It is the broad mix of
professional interests that is the hallmark of this conference. The focus is color-color as a critical element of the research and application efforts of
this segment of the professional community. The conference program is designed to promote interaction among the participants. The format includes invited addresses by leading specialists in various color-related fields as well as submitted papers presented in oral and poster format. We will
continue the single-session format for this year’s conference to allow participants to attend all presentations. This is the conference to meet and talk
with those people that share your interest in color, color research and its application to products. Preliminary program is now available at
www.imaging.org.
• Image Capture
• Scene Perception
• Color Management
• Color Appearance
• Color Constancy
• Standards
• Gamut Mapping
• Printing
Co-sponsored by IS&T—The Society for Imaging Science and Technology and SID—Society for Information Display
For more information contact IS&T, 703-642-9090; FAX: 703-642-9094; EMAIL: info@imaging.org; www.imaging.org
IS&T Corporate Members
The Corporate Members of your Society provide a significant amount of financial support that assists IS&T in disseminating information and
providing professional services to imaging scientists and engineers. In turn, the Society provides a number of material benefits to its Corporate
Members. For complete information on the Corporate Membership program, contact IS&T, 7003 Kilworth Lane, Springfield, VA 22151.
Sustaining Corporate Members
Applied Science Fiction
8920 Business Park Drive
Austin, TX 78759
Hewlett Packard Labs.
1501 Page Mill Road
Palo Alto, CA 94304
Tektronix, Inc.
P.O. Box 4675
Beaverton, OR 97076-4675
Eastman Kodak Company
343 State Street
Rochester, NY 14650
Lexmark International, Inc.
740 New Circle Road NW
Lexington, KY 40511
Xerox Corporation
Webster Research Center
Webster, NY 14580
Polaroid Corporation
P.O. Box 150
Cambridge, MA 02139
Supporting Corporate Members
Adobe Corporation
345 Park Avenue
San Jose, CA 95110-2704
Kodak Polychrome Graphics
401 Merritt 7
Norwalk, CT 06851
Konica Corporation
No. 1 Sakura-machi
Hino-shi, Tokyo 191 Japan
Torrey Pines Research
6359 Paseo Del Lago
Carlsbad, CA 92009
Xeikon, NV
Vredebaan 72
2640 Mortsel, Belgium
Donor Corporate Members
Agfa Division Bayer Corp.
100 Challenger Road
Ridgefield Park, NJ 07760
Ilford Imaging U.S.A. , Inc.
West 70 Century Road
Paramus, NJ 07653
Research Laboratories of Australia
7, Valetta Road, Kidman Park
S. Australia, 5025, Australia
BARCO Graphics
Tramstraat 69
B-9052 Gent, Belgium
KDY Associates
9 Townsend West
Nashua, NH 03063
Ricoh Company Ltd.
15-5, Minami-Aoyama
1-chome, Minato-ku, Tokyo 107 Japan
BASF Corporation
100 Cherry Hill Road
Parsippany, NJ 07054
Kind & Knox Gelatin, Inc.
P.O. Box 927
Sioux City, IA 51102
SKW Biosystems, Inc.
2021 Cabot Boulevard West
Langhorne, PA 19047
Canon , Inc.
Shimonaruko 3-40-2
Ohta-ku, Tokyo 146 Japan
Minolta Co., Ltd.
1-2, Sakuramachi
Takatsaki, Osaka 569 Japan
Felix Schoeller Jr. GmbH & Co. KG
Postfach 3667
D-49026 Osnabruck, Germany
Mitsubishi Electric
5-1-1 Ofuna, Kamakura
Kanagawa 247 Japan
Sharp Corporation
492 Minosho-cho
Yamatokoriyama,
Nara 639-1186 Japan
Fuji Photo Film USA, Inc.
555 Taxter Road
Elmsford, NY 10523
Monroe Electronics, Inc.
100 Housel Avenue
Lyndonville, NY 14098
Hallmark Cards, Inc.
Chemistry R & D
2501 McGee, #359
Kansas City, MO 64141-6580
Nitta Gelatin NA Inc.
201 W. Passaic Street
Rochelle Park, NJ 07662-3100
Hitachi Koki Co., Ltd.
1060 Takeda, Hitachinaka-City
Ibaraki- Pref 312 Japan
Quality Engineering Assoc.
25 Adams Street,
Burlington, MA 01803
Sony Corporation
6-7-35 Kita-shinagawa
Shinagawa, Tokyo 141 Japan
Sony Electronic Photography &
Printing
3 Paragon Drive
Montvale, NJ 07645
Trebla Chemical Company
8417 Chapin Ind. Drive
St. Louis, MO 63114