LOW TEMPERATURE DEPOSITION OF METAL CARBIDE FILMS
FROM SINGLE SOURCE PRECURSORS
by
RODRIGO R. RUBIANO
Submitted to the
Department of Nuclear Engineering and Department of Material Science and Engineering
in Partial Fulfillment of the Requirements for the Degrees of
Master of Science in Nuclear Engineering,
Bachelor of Science in Nuclear Engineering, and
Bachelor of Science in Material Science and Engineering
at the
Massachusetts Institute of Technology
February 1994
Rodrigo R. Rubiano 1994
All rights reserved
The author hereby grants to MIT and os Alamos National Laboratory permission to reproduce and to distribute
copies of this thesis document in whole or in part.
Signature of Author-
I -. 1 I
Iq
IiT
f
Department of Nuclear Engineering
Department of Material Science and Engineering
December 10, 1993
Certified by -
091
Uday B. Pal
Chipman Professor, Material Science and Engineering
Thesis Supervisor
Certified by Ronald Ballinger
Professor, Nuclear Engineering
Departmental hesis Reader
Accepted by Professor Allan Henry, Chairman
Departmental Committee on Graduate Students
Department of Nclear Engineering
Accepted by -
MAS7,H TF77 !NrT!
,. ; J--;F
MAR
2 1994
Science
AVOW*
Professor Samuel M. Allen
Chairman, Undergraduate Tesis Committee
Department of Material Science and Engineering
2
3
LOW TEMPERATURE DEPOSITION
OF METAL CARBIDE FILMS
FROM SINGLE SOURCE PRECURSORS
by
RODRIGO R. RUBIANO
Submitted to the
Department of Nuclear Engineering and
Department of Material Science and Engineering
on December 10, 1993
in partial ulfillment of the requirements
for the Degrees of
Master of Science in Nuclear Engineering,
Bachelor of Science in Nuclear Engineering, and
Bachelor of Science in Material Science and Engineering.
ABSTRACT
Stable zirconium carbide and titanium carbide thin films have been deposited
from single source organometallic precursors, tetraneopentyl zirconium and
tetraneopentyl titanium respectively, at substrate temperatures above 500 'C.
Materials deposited above this temperature are cystalline by X-ray diffraction.
A metal to carbon ratio of 12 is observed by Auger electron spectroscopy
depth profiling for both systems. X-ray photoelectron spectroscopy (XPS)
indicates the metal is in only one phase - a metal carbide. he observed
spectra correspond well to spectra for metal carbide standards. Carbon XPS
4
reveals three phases in the zirconium system and reveals only two phases in
the titanium system. A carbidic and a hydrocarbon species are present in both
metal carbide coatings. The third carbon phase present only in the zirconium
carbide films has yet to be identified. Elastic recoil detection finds,a large, up
to 16%, hydrogen content in the thin metal carbide films. Except for the third
unknown carbon species observed in zirconium carbide thin films, similar
results are obtained for the deposition of zirconium carbide and titanium
carbide.
Thesis Supervisor:
Dr. Uday B. Pal
Title: John Chipman Assistant Professor, Material Science and Engineering
5
TABLE OF CONTENU
1. INTRODUCTION ........................................................................
12
I.I. M etal Carbides .....................................................................
1.2. Conventional Chemical Vapor Deposition of Metal
Carbides .............................................................................
1.3. Organometallic Chemical Vapor Deposition of Metal
Carbides .............................................................................
1.4. Previous Efforts and Current Goals .......................................
12
2. EXPERIM ENTAL .......................................................................
2.1. Precursor Synthesis ..............................................................
2.2. Deposition Reactor System ....................................................
2.3. Film Characterization Equipment ..........................................
17
17
17
20
3. RESULTS /DISCUSSION ............................................................
3.1. Zirconium Carbide Deposition ..............................................
3.1.1. Film Stability ............................................................
3.1.2. X-Ray Diffraction .....................................................
3.1.3. Film Morphology ......................................................
3.1.4. Film Thickness and Average Growth Rate ...................
21
21
21
24
26
28
13
14
16
3.1.5. AES Analysis ............................................................
29
3.1.6. XPS Analysis ............................................................
3.1.7. ERD Analysis ...........................................................
3.2. Titanium Carbide Deposition .................................................
3.2. 1. General ....................................................................
3.2.2. Film Stability ............................................................
3.2.3. X-Ray Diffraction .....................................................
3.2.4. Film M orphology ......................................................
3.2.5. Film Thickness and Average Growth Rate ...................
3.2.6. AES Analysis ............................................................
36
43
45
45
45
46
49
49
49
3.2.7. XPS Analysis ............................................................
54
3.2.8. ERD Analysis ...........................................................
60
6
4. CONCLUSION ............................................................................
63
4.1. Comparison of ZrC and TiC Systems in Present Work ............. 63
4.2. Differences from Previous Efforts ......................................... 63
5. FUTURE W ORK .........................................................................
64
APPENDIX 1. ZRC STANDARDS .................................................... 66
REFERENCES .................................................................................
71
7
LIST OF FIGURES
Figure 21
Schematic of OMCVD reactor system.
Figure 3 la
Scanning electron micrograph of a ZrC film deposited at
390 'C and 1-2 Torr on an A1203 substrate.
Figure 3.1b
Scanning electron micrograph of a ZrC film deposited at
570 'C and 10-4 Torr on a Si(l 10) substrate after exposure
to ambient conditions for one month.
Figure 32
XRD of a ZrC film grown on a Si(l IO) substrate at 750 C
and 5xIO-5 Torr.
Figure 3.3a
Scanning electron micrograph of a ZrC film deposited at
570 'C and 10-4 Torr on a Si(l 10) substrate.
Figure 3.3b
Cross-sectional micrograph of a ZrC film deposited at
570 'C and 10-4 Torr on a Si(I 10) substrate.
Figure 34
AES survey spectrum of ZrC thin film deposited at 570 OC
and 10-4 Torr. Surface layer contamination was removed by
Ar sputtering for 60 minutes.
Figure 35
AES depth profile of ZrC thin film deposited at 570 'C and
10-4 Torr.
Figure 36
AES depth profile of ZrC thin film deposited at 750 'C and
10-4 Torr.
Figure 37
XPS survey spectrum of ZrC thin film deposited at 570 C
and 10-4 Torn Surface layer contamination was removed by
Ar sputtering for 20 minutes.
Figure 38
Zirconium XPS for a ZrC film grown at 570 'C and
10-4 Torr after sputtering.
Figure 39
Carbon XPS for a ZrC film grown at 570 'C and 10-4 Torr
after sputtering.
8
Figure 310
Carbon XPS for a ZrC film grown at 750 'C and 10-4 Torr
after sputtering.
Figure 31
Deconvoluted carbon XPS for a ZrC film grown at 570 C
and 10-4 Torr after sputtering.
Figure 312
Deconvoluted carbon XPS for a ZrC film grown at 750 C
and 10-4 Torr after sputtering.
Figure 313
Deconvoluted carbon XPS for a ZrC film grown at 570 C
and annealed at 880 'C for 2 hours.
Figure 314
XRD of a TiC film grown on a Si(l 10) substrate at 570 OC
and IxIO-5 Torr.
Figure 315
XRD of a TiC film grown on a Si(I 10) substrate at 300 C
and 10-4 Torr.
Figure 316
AES survey spectrum of TiC thin film deposited at 570 C
and lxIO-5 Torr after sputtering.
Figure 317
AES depth profile of TiC thin film deposited at 570 'C and
lxIO-5 Torr.
Figure 318
AES depth profile of TiC thin film deposited at 300 C and
10-4 Torr.
Figure 319
XPS survey spectrum of TiC thin film deposited at 570 C
and 10-4 Torr after sputtering.
Figure 320
Titanium XPS for a TiC film grown at 570 'C and
jx10-5 Torr after sputtering.
Figure 321
Titanium XPS for a TiC film grown at 300 C and 10-4 Torr
after sputtering.
Figure 322
Carbon XPS for a TiC film grown at 570 'C and
lxIO-5 Torr after sputtering.
Figure 323
Carbon XPS for a TiC film grown at 300 'C and 10-4 Torr
after sputtering.
9
Figure 324
Deconvoluted carbon XPS for a TiC film grown at 570 C
and IxIO-5 Torr after sputtering.
Figure 325
Deconvoluted carbon XPS for a TiC film grown at 300 C
and 10-4 Torr after sputtering.
10
LIST OF TABLES
Table 31
ZrC Deposition Process Using ZrNP4 Precursor.
Table 32
AES and ERD Results of ZrC from ZrNp4-
Table 33
AES Relative Sensitivity Factors as a Function of Electron
Beam Voltage.
Table 34
TiC De
Table 35
AES and ERD Results of TiC from TiNP4.
''
Process Using TiNP4 Precursor.
11
ACKNOWLEDGMENTS
I owe a great deal of thanks to all of the helpful folk at the Los
Alamos National Laboratory (LANL) and at MIT who participated in this
thesis project. Without their knowledge and experience, much of this work
could not have been accomplished. Certain individuals went out of their
way in ensuring this venture was successful. They deserve special thanks.
David Smith of the Isotopes and Nuclear Chemistry group (INC-1)
provided invaluable guidance to my research. His ideas and enthusiasm
were the driving forces of this project.
Bob Springer of the Coatings and Polymer Chemistry group
(MST-7) gave a helping hand in the AES and XPS characterization and data
reduction. Without his insight and constant questions, this project would
have been too easy to complete.
Joe Laia, group leader of MST-7, gave me a job and provided the
laboratory equipment and monetary support to complete this work.
Dr. Uday Pal and Dr. Ronald Ballinger, thesis advisors at MIT,
played a critical role in the writing of the thesis. Their comments were
greatly appreciated.
Lastly, but most important, I want to thank all of my friends and
family. I owe my sanity to Shari Schuchmann, Prof. Fred McGarry, the
Fiji Class of 1992, John Moalli, Genevieve Hodgson, the JerkyBoys and Gal
of 306A, and the MIT Football team. The largest debt, though, is too my
family. Mom and Dad shaped me into who I am today and will be
tomorrow. I owe them everything. My brothers and sister made sure I
didn't forget that somebody was always there to listen and comfort.
Thank you.
2
1. INTRODUCTION
1. 1.
Metal
Carbides
Group IVB and VB carbides are of interest for a number of
applications that take advantage of their refractory properties; extremely
high melting points, hardness, and chemical stability. 14 Their fields of use
are broad and include corrosion-resistant materials and materials for the
chemical industry; materials for nuclear energy production; materials for
the manufacture of aircraft and rockets; materials for electrical and radio
purposes; and super-hard and wear-resistant materials.5
In the nuclear industry, zirconium carbide's low neutron-capture
cross section ((Yc= 02 barns) allows it to be used as a coating for graphite
matrices containing UC and ThC as fuel elements in power reactors.6
Also, ZrC may be substituted for SiC in the TRISO fuel particles of the
high-temperature gas-cooled reactor (HTGR). This could have two major
advantages. The first is to increase the temperature to which the graphite
fuel body containing the particles can be heated during the fabrication
procedure to produce a fuel body matrix with a higher then-nal
conductivity. Ibis improves heat transfer within the reactor core and can
yield a higher coolant temperature for a given coated particle operating
temperatures The second is the potential to increase the operating
temperature of the coated particles themselves. This could increase
attainable coolant temperatures still further. With its high temperature
capability and good neutronic properties, ZrC seems to be the logical
choice of coat material for consideration as a replacement for SC.7 ZrC is
also used for electrical and radio purposes. It has been incorporated in
thermionic cathodes for electronic devices working in portable equipment
3
under conditions of low vacuum and intensive ionic bombardment. This
refractory carbide is also resistant to molten metals and can be used as
crucibles and boats for melting metals and as an inner coating for tubes
conveying liquid metals.5 Finally, ZrC has great hardness and high
abrasive power and serves well to polish other materials.8
In Japan and the U.S., titanium carbide has been identified as a
material of interest for the "first wall" coating for the fusion reactor. It is
also being studied as a low-friction thermal barrier coating for cylinder
walls in the adiabatic diesel engine.9 Titanium carbide also serves as a
protective material when handling molten metals and is commonly used in
parts of pumps for transferring liquid sodium. TiC likewise plays an
important role in the manufacture of aircraft and rockets. Its high melting
point, hardness, heat resistance and strength at high temperatures make it
an attractive material for gas-turbine blades in jet aircraft engines. TiC is
also used in friction disks for aircraft construction and serves as a
protective coating for rocket components. In the electrode industry, TiC is
a constituent of electrodes for the underwater oxy-electric cutting of steel
and is used in thermocouple electrodes for measuring temperatures up to
2000 C in reducing and neutral atmospheres and in vacuum.10
1.2. Conventional Chemical Vapor Deposition of Metal Carbides
Established chemical vapor deposition (CVD) routes to metal
carbides rely on the chemical reaction of metal halides with methane and
hydrogen mixtures at temperatures in excess of 1000 'C. The high
processing temperatures arise from the thermally robust nature of the
metal halide compounds, a result of the strength of the metal halide bonds.
Due to elevated deposition temperatures, thermally sensitive substrates can
4
not be coated using metal-halide processes. This is not a serious drawback,
though. Many of the applications for these carbide materials focus on their
use in high temperature environments; therefore, limitations due to the
elevated deposition temperatures do not exist in most instances.
One of the disadvantages of conventional metal-halide based CVD is
the production of corrosive gases (e.g., HCl, HF) in the processes." The
presence of these gases necessitates the design of expensive and complicated
deposition systems. Also, in many CVD processes the presence of a
halogen is detrimental. It can promote corrosion of the film either during
deposition, or if halogen atoms become incorporated into the film, after the
deposition b exposure of the finished product to air, water, or extreme
temperatures
In some cases, conventional CVD processes do not yield
high-quality materials or consistently reproducible materials. An added
limitation to the present technology is the extremely low volatility of the
metal halide precursors.13 For example, ZrC14remains a powder at low
temperatures and intermediate pressures. This makes it difficult to
introduce the precursor into the CVD system. Without a reliable method
for introducing a powder precursor into the heat zone of a CVD system,
inconsistent films are deposited.7 An exception to the volatility limitation
is TiC14 a liquid at room temperature.
1.3. Organornetallic Chemical Vapor Deposition of Metal Carbides
Organometallic complexes as CVD precursors are a simple and
powerful method for producing coatings at low temperatures, eliminating
deleterious byproducts, and removing the halide from the process
completely. Many organornetallic complexes have attractive decomposition
temperatures, in the range of 100 - 200 'C. Reduced processing
15
temperatures, made possible by the reduced thermal stability of the
complexes, open up the possibility of coating thermally sensitive materials.
Also, low-temperature depositions can yield materials not stable at higher
temperatures as well as metastable materials and phases. Lowering the
temperature that solid materials can be synthesized at not only increases the
energy efficiency of the process but also makes it possible to use the
material in applications where higher temperatures lead to undesirable
consequences such as interlayer atomic diffusion, decohesion of overlayers
due to mismatches in thermal expansion coefficients, or temperatureinduced changes in the shape or the crystallinity of the substrate or nearby
structures.
The organometallic CVD (OMCVD) process provides other
advantages than just a benign deposition environment. With an
organometallic precursor it is possible to assemble the desired components
of a composite material into a single precursor (e.g., Nb(C6H5CH3)2 for
NbC, Hf[CH2C(CH3)314for HfQ14,15enhancing process control and
providing a low-temperature route to complex composite materials. These
complexes are "tailored" for the production of the metal carbide films
since they only contain the metal, carbon, and hydrogen. Many examples of
these MRn complexes are known for zirconium, hafnium, niobium, and
tantalum where R is an alkyl ligand, carbocyclic ligand such as benzene or
cyclopentadiene, or some combination. Most can be prepared, in high
yield, "one-pot" syntheses starting from the binary metal halide and readily
available alkali metal reagents using standard techniques.16
One disadvantage of OMCVD is its low deposition rate relative to a
conventional CVD process. In OMCVD, typical deposition rates are on the
order of I gm/hr.17,18 For example, the deposition of TiC from the
6
organornetallic Ti[CH2C(CH3)314 at 355 'C is 0.3 gm/hr.9 In
conventional CVD, the deposition of TiC from the reaction of TiC14and
CC14 at 1600 'C is
120 pm/hr.19 A second disadvantage of OMCVD is the
incorporation of residual carbon and hydrogen in the films. This stems
directly from the large carbon-to-metal ratio of the organometallic
precursors.
1.4. Previous Efforts and Current Goals
Previous work has found that deposition of metal carbides from
organornetallic precursors generally yields amorphous materials that
contain some amount of residual carbon.13,15,17,20
Claims of no observable
residual carbon in the deposited material have been made in a few
instances.13,21Generally the characterization of the deposited materials is
not complete and in some cases only a brief comment is made about the
deposition of'a carbide from an organometallic precursor. The focus of
this work is on the deposition of zirconium carbide and titanium carbide
from zirconium and titanium organometallics, specifically the deposition of
ZrC and TiC from tetraneopentyl zirconium (Zr[CH2C(CH3)314)and
tetraneopentyl titanium (Ti[CH2QCH3)314). A thorough characterization
of the deposited materials is presented using several analytical techniques,
including X-ray diffraction, scanning electron microscopy, Auger electron
spectroscopy, X-ray photoelectron spectroscopy, and elastic recoil
detection. The results from these two processes are compared directly and
to previous efforts in the deposition of similar metal carbides from
organometallic precursors9,11,14to understand the details of the deposition
process. An increased understanding of the factors important in
organometallic CVD will lead to a reliable process for metal carbides.
7
2. EXPERIMENTAL
2 I. Precursor Synthesis
Tetraneopentyl zirconium (ZrNP4) and tetraneopentyl titanium (TiNP4)
were prepared by literature procedures.22,23 The compounds were prepared
by the reaction of neopentyllithium with the anhydrous metal chlorides in nhexane (TiC14) or n-hexane/ether (ZrCI4) per Equation 2.1.
4 LiCH2C(CH3)3 + MC14 ----- > M[CH2C(CH3)314
4 LiCl
(2.1)
The only variation introduced was the use of Ti(O-n-BU)4 as the starting
titanium reagent in the synthesis of the tetraneopentyl complex. This afforded
an 80% yield of TiNP4, an increase from the reported 50% yield using
Ti(OCH2CH3)4 as the titanium reagent. The precursor complexes were
sublimed twice at 10-4 Torr to assure adequate purity. The organometallic
complexes were characterized by NMR, IR, and mass spectroscopy.
2.2. Deposition Reactor System
All films were deposited in a horizontal, hot-walled chemical vapor
deposition (CVD) reactor (Figure 21). The base pressure of the system is
approximately 10-6 Torr. This is achieved with a 330 liter/second
turbornolecular pump. An alternate pumping manifold maintained
pressures of 10-3 Torr using a mechanical pump. Attached to the system is
a 200 amu quadrupole mass spectrometer. The pressure is monitored by
an MKS electronic manometer, a Pirani gauge, or a cold cathode ionization
18
-0
0
.r
.9.
tm
C
cn
a)
12
(n
-0
6
Ccn
2
a,
U)
a-
E
m
0 1-1
zm
E
CD
U)
CD
u
m
>1
(n
0
E
:3
t5
CZ
a)
cc
LLCD
.C
U)
E
M
0
02
O
0
0
.9
a,
CZ
:RCL
0
u
0
C)
0)
a)
-C
cu
0)
C
0
E
ao
-0
0Cu
T
C)
.a2
1
.N
C
.0
CD
'D
0
i:
m
u
.z:2
.2-.j
'D
8
aE
:3
CL
-ro
.C
U
m
5
CD
E
0
-6t
cnC
",m
2E
E
CT)
C.)
(n
17
Cj
a)
:3
.21
U-
9
gauge. An in-line liquid nitrogen trap was placed between the deposition
zone and the pumping system. The cold trap was not used when mass
spectra were taken.
Films were deposited on sapphire, Si(I IO) and Si(100) substrates.
Sapphire substrates were cleaned with methanol, trichloroethane a second
methanol rinse, and blown dry with nitrogen. Three inch silicon wafers
were broken into approximately 0.5 cm squares and blown clean with
nitrogen. The substrates were mounted on a 381 cm (1.5 inch) diameter
boron nitride (BN) stage with silver mounting paste. The stage rested on
the bottom of the 762 cm (3-inch) diameter quartz tube approximately
cm from the front of the 10.16 cm 4 inch) diameter clam-shell furnace.
The substrates were mounted perpendicular to the flow through the
reactor.
he substrate temperature was monitored by a thermocouple
embedded in the BN stage. The glass walls of the inlet side of the reactor
was maintained at a temperature of 60 - 75 'C using resistive heating
straps. This prevented the condensation of the precursor.
he precursor
vessel was heated using an oil bath to 50 - 60 'C to increase the vapor
pressure of the compound. Typical deposition runs were I - 2 hours in
length. Standard film thicknesses obtained were I - 2 gm. Deposition
pressures using the turbomolecular pump were 5x10-5 to 10-4 Torr.
Pressures of 1-2 Torr were obtained using the mechanical pump system or
by throttling a bellows valve located between the reactor and the turbo
pump. Deposited films were cooled to room temperature in vacuo.
20
2.3. Film Characterization Equipment
Scanning electron micrographs were obtained in a Hitachi S520-LB
scanning electron microscope (SEM). Auger electron spectroscopy (AES)
and X-ray electron spectroscopy (XPS) were performed with a Physical
Electronics Model 545 and Model 548 cylindrical mirror system. Electron
beam conditions for AES were normally
beam current in a spot size of
keV and less than I gA total
gm. For analysis of the titanium carbide
films, an electron beam voltage of 3 keV was used. AES depth profiling
was performed using 35 keV Ar ions. Argon pressure, measured before
bleeding the gas into the analysis chamber, was approximately was
1 x 10-4 Torr at a 25 mA emission, resulting in a total system pressure of
8 x 10-9 Torr. Typical ion beam raster area for each sample was a
2 mm x 2 nu-n square. Sensitivity factors used for chemical composition
of ZrC films were 014 for carbon (graphite), 016 for zirconium, and
0.42 for oxygen.27 For the titanium carbide films, analyzed with a 3 keV
electron beam, the sensitivity factors are 021 for carbon (graphite),
0.45 for titanium, and 052 for oxygen. X-ray photoelectron spectra were
obtained using a Mg Koc source (hv = 1253.6 eV) at 15 keV and 200 watts.
Narrow peak scans were obtained with a band pass energy of 25 eV and
analyzer resolution of 16%. Prior to all XPS scans, the sample was
sputtered with Ar to remove any surface contaminants, especially
adventitious carbon and oxygen. The established von Cittert deconvolution
procedures24,were employed to remove X-ray satellite peaks, X-ray line
widths, and analyzer broadening from the spectra. Elastic recoil detection
(ERD) with a 20 MeV He+ beam was performed using the 3 MV tandem
accelerator at the Ion Beam Materials Laboratory at Los Alamos National
2
Laboratory. A kapton foil (C22H1205N2) was used as a standard with a
known hydrogen content for ERD analysis. Hydrogen content
measurements are accurate to within 2 atomic percent. X-ray diffraction
was utilized in determining cystallinity. Deposited film thickness values
were obtained from cross-sectional view in the SEM. Morphological
analysis was conducted with an optical microscope and the SEM. The
elemental composition was determined from depth profiling with the AES
system and XPS analysis determined the chemical species of the constituents
in the material. Elastic recoil detection (ERD) was used to determine the
amount of hydrogen incorporated in the materials during deposition.
3. RESULTS /DISCUSSION
3.1. Zirconium Carbide Deposition
3.1.1. Film Stability
Thin films deposited at temperatures from 300 to 570 'C did
not adhere to single-crystal sapphire (A1203) substrates.
he deposited
material cracks and de-laminates shortly after removal from the CVD
reactor (Figure 3.1a). Highly reflective, silvery/gray films can be
deposited on Si(I 10) and Si(100) from zirconium tetraneopentyl (ZrNP4 at
temperatures from 300 to 750 'C and pressure from 12 to 10-4 Torr
(Table 31). On silicon substrates, thin films prepared at substrate
temperatures below 500 'C are not stable upon exposure to the ambient
atmosphere and adherence is poor. The deposited material cracks, crazes,
and peels from the substrate shortly after removal from the CVD reactor,
exhibiting behavior similar to films deposited on sapphire. However, thin
22
Figure Ma
Scanning electron
micrograph of the surface of a ZrC film
Figure 3.1b Scanning electron
micrograph of the surface of a ZrC film
deposited at 390 'C and 10-2 Torr on an
deposited at 570 'C and 10-4 Torr on a
A1203 substrate.
Si(I 10) substrate after exposure to ambient
conditions for one month.
23
Table 31
ZrC Deposition Process Using ZrNP4 Precursor
Substrate
Deposition
Growth Rate
Substrate;
Temperature
Pressure
(pm/hr)
Stability
(OC)
(Torr)
(Time to degrade)
2xlO-2 a
A1203; minutes
570
Si(I 10); months
570
IxIO-4
b
1.0
Si(l 10); months
5 70
2 I O-2 c
0.5
Si(I 10); months
570
5xlO-5
d
0.6
Si(IOO); months
400
IxIO-4
b
0.5 e
Si(I 10); minutes
400
IxIO-2 c
0.2
SiO 10); hours
300
3xIO-5 b
SKI IO); minutes
a.
750
5xlO-5
Mechanical pump in use.
b.
C.
d.
e.
Turbomolecular pump in use.
Pressure regulated with bellows valve on turbomolecular pump.
Annealed at 880 'C for 2 hours immediately following deposition.
One of four films was partially intact for SEM analysis.
b
0.3
SiG 10); months
films deposited above 550 C on silicon substrates appear to be stable, even
when exposed to ambient atmosphere for extended periods of time. In few
instances, exposure to atmospheric conditions for months led to these films
cracking and de-laminating off the Si substrates (Figure 3.1b). Presently,
no specific information about the reactivity of the thin film material is
available. It is presumed that this macroscopic degradation is a result of
24
strain relief brought about by reaction with water vapor in the atmosphere.
The reaction is presumed to be oxidation of partially decomposed or
undecomposed precursor incorporated in the thin film during deposition.
3.1.2. X-Ray Diffraction
Films deposited above 550 C are crystalline by XRD
(Figure 32). 'Me results show some orientational preference along the
ZrC(l I 1). This orientational preference is observed for all three
substrates at deposition temperatures of 570 C and 750 C and pressures of
10-2 and 10-4 Torr. The growth temperature has a strong influence on the
structure of the deposited film. At low temperatures, surface diffusion is
slow relative to the arrival rate of the precursor, and an amorphous film is
deposited. This behavior is observed for the deposition of ZrC from
ZrNP4 at 300 'C. At high temperatures, surface diffusion is fast relative to
the incoming-, flux, allowing the adsorbed species to diffuse to step growth
sites, and single crystalline layers are formed. At intermediate
temperatures, nucleation occurs at many different points on the surface.
Adsorbed species diffuse to the islands, which grow and for
a
polycrystalline film.25 This behavior is observed for the deposition of ZrC
from ZrNP4 at 570 and 750 'C. An approximate particle size of 40 A can
be obtained
rom the Scherrer formula26 for the depositions at 570 'C.
Increasing the silicon substrate temperature to 750 'C results in an increase
in the particle size to 75 A. The Scherrer formula is used to estimate the
particle size of very small crystals from the measured width of their
diffraction curves.
25
4
4
0
LO
C)
cn
LO
00
u
0
LO
CD
CD
r-
-P
(Z
4
cn
cil
CD
C-
tD
CM
-
aj
M
w
0
1-11
C
c"I
c
cu
0
C14
C14
CI-4
1.=1
1-1
cn
Ln
clu
0
4
0C-4
c
44
u
4
C'n
4-4
0
m C)
<=)C=)
O"
.
xLn
LO
C)
CD
C\j
LO
qw
CD
m
Ul)
clu
CD
m
LO
lw
CD
CNJ
LO
CD
17
M
clu
V"
1;
CD
(M
(m
CD
(squnoo) 44;suaquj
m
Q)
4
z
01
.H
L4
26
3.1.3. Film Morphology
Optical microscopy shows a smooth, featureless surface for all
deposited material. Scanning electron micrographs do not reveal any grain
structure. Particulates are observed on some of the films' surface. At very
high magnification X10k), some surface roughness is observed (Figure
3.3a).
his is attributed to a three-dimensional island growth process.25 In
three-dimensional island growth, small clusters are nucleated directly on
the substrate surface and grow into islands of the film material.
eventually coalesce to form a continuous film.
hey
his growth mode takes
place when the film atoms are more strongly bonded to each other than to
the substrate. Small clusters whose dimensions are on the order of the film
thickness are observed on the surface of the thin films (Figure 3.3a).
These clusters are incorporated in the film as they could not be removed by
simply blowing on the film surface with nitrogen.
he origin and
formation process of these clusters is not known at this time. However, a
possible explanation is that the clusters are deposited material fragments
that spalled off the coated reactor walls and were incorporated in the film.
To investigate this hypothesis, the reactor interior walls were cleaned. Te
material on the reactor wall was oxidized by heating the quartz tube in air
to a temperature exceeding 700 'C for several hours. The white residue
remaining could then be easily scraped from the reactor wall. A
subsequent deposition resulted in a film with an apparent diminished
number of clusters. A second hypothesis is that these clusters are
fragments of precursor that did not sublime but were carried in solid form
to the film surface. This prediction was tested by inserting a glass frit in
the pathway between the precursor vessel and the substrates just before the
hot zone of the reactor. The resulting film again showed a greatly reduced
27
Figure 3.3a Scanning electron
micrograph of the surface of a ZrC film
micrograph of a ZrC film deposited at
deposited at 570 'C and 10-4 Torr on a
570 'C and 10-4 Torr on a Si(I IO)
Si(I 10) substrate.
substrate. The gray material in the
Figure 3.3b
Cross-sectional
micrograph is the ZrC film adhered to the Si
substrate on the right.
28
number of particulates on the surface. It is clear that the cleaning of the
reactor tube aided in reducing the number of clusters observed. However,
the benefit of the frit's presence is not completely clear as this experiment
was conducted with the cleaned reactor tube. Thus, the source and growth
process of these clusters is still not explicitly known. Cross-sectional SEM
micrographs (Figure 3.3b) do not show any grain or columnar structure
even at high magnification. The roughness observed is a result of the
fracture surface.
3.1.4. Film Thickness and Average Growth Rate
Typical film thicknesses are 12 gm as measured in crosssectional micrographs (Figure 3.3b). The average growth rate at
10-4 Torr and substrate temperature of 570 'C is approximately 1.0 gm/hr.
Decreasing the substrate temperature to 400 'C results in a drop in the
average growth rate to 0.5 gm/hr. Increasing the substrate temperature to
750 'C results in a decrease in the average growth rate to 03 gm/hr. This
second observation is likely due to a decrease in the amount of precursor
reaching the substrate due to the increased deposition on the reactor walls
in front of the BN stage. At this temperature, the dark fringe in front of
the stage (signifying ZrC coating on the reactor wall) extended further
away from the stage than at lower temperatures. This is a simplistic
argument. A 'better understanding of the change in average growth rate
with temperature requires a more involved study of the OMCVD process
in which adsorption, decomposition, diffusion, and desorption of the
precursor and its products are analyzed.
Increasing the deposition pressure to 1-2 Torr reduces the average
growth rate b roughly a factor of 2 from the rates observed at 10-4 Torr.
29
There is no observed difference in average growth rate between the
Si(110) and Si(100) substrates. The increase in pressure results in a
diminished sublimation rate of the organometallic and leads to a decrease in
growth rate. At the higher pressures, fluctuations in the growth rate are
seen for a fixed substrate temperature. This is attributed to variations in
the sublimation rate of the solid precursor.
3.1.5. AES Analysis
Survey AES spectra of the deposited material show Zr C and
0 as the only observable constituents of the thin film. Figure 34 is a
typical AES survey spectra of a ZrC film deposited at 570 C and
10-4 Torr. Surface contamination is removed by sputtering with argon
ions immediately before the spectra is acquired. AES depth profiling of
the films reveals a very uniform composition. The depth profile presented
in Figure 35 shows a constant ratio of zirconium to carbon throughout the
bulk of the
aterial for a thin film deposited at 570 C and 10-4 Torr. The
zirconium t carbon ratio remains constant at approximately 12. Residual
oxygen is observed in all deposited materials and is typically 2 at.% for
depositions at temperatures greater than 500 'C. Depositions at lower
temperatures result in films with a greater oxygen content, as high as
10 at %. The oxygen profile by AES depth profiling is indicative of a
diffusion process. It is believed that upon exposure to ambient atmosphere,
oxygen quickly diffuses into and chemically reacts with the material
prepared at lower temperatures. Mis provides further evidence for the
hypothesis that films prepared below 500 'C are unstable due to oxidation
with the water vapor in the atmosphere.
30
4
4
0
E-4
LO
4
44 44
0
$4
4-4
0
4
u
4
4
Q)
3p/Np
31
C)
-
0LO
CD
4
4
0
E
C)
C)
Ln
v
M
C)
C)
C)
lw
u
N
Ln
4J
'o Z
CD
C)
W)
m
00
0cn
c,
0
10
4.)
li
U
In
0U')d
CN E
I I
1
4
I
I
II
0C
C)
CN
co Z))
4-4
0
i
I
I
I
C)
LO
F a)
I
4-4
0
4
124
4J
I
C)
I
C)
C
I
C)
C)
LO
a
4
I
L
C)
C)
Ul)
0
0
0
IT
0
0
U')
m
0
0
C)
m
1 I
C)
0
U-)
N
0
0
0
N
HdcfV ',llsu;)lul
Ln
a0
U')
0O
C)
0C)
LO
C)
C)
ra
4
a)
>
32
Increasing the substrate temperature to 750 'C does not alter the
composition of the material (Figure 36). Also, the two hour anneal at
880 'C of a film deposited at 570 'C and 10-4 Torr does not change the
Zr:C ratio or the residual oxygen content of the film. The 880 'C anneal
of the ZrC material has a two-fold purpose. First, it was presumed that
some partially decomposed precursor was incorporated into the film. The
anneal serves to fish
the chemical decomposition of the ZrNP4 and
possibly change the relative amounts of carbide and residual carbon in the
film (See XPS Analysis section). The second purpose of the anneal is to
thermally drive residual hydrogen from the thin film.
The composition of the films deposited at various temperatures and
pressures is presented in Table 32. The listed compositions were
calculated using sensitivity factors for carbon (graphite), zirconium, and
oxygen found in the open literature.27
The atomic concentration of element X in the sample is expressed as:
Cx = Ix / Sx dx) / I (la / Sx dx)
where
(3.1)
is the peak-to-peak Auger amplitude from the sample, Sx is the
relative sensitivity between the element and silver, the summation is over
one peak per each element in the -sample and dx is the scale factor (which is
the same for all peaks in these experiments and thus cancels out). The
sensitivity factors, Sx employed in these calculations are from elemental
standards. Table 33 lists the sensitivity factors used in these calculations.
Since the listed sensitivity factors correspond to elemental standards, some
error is introduced in the composition calculations of the OMCVD ZrC
33
0CC)D
I
i
m
4
0
E
00
00
LO
a
00
80
N
(Tj
u
0
a
aI
IM
0M
00
00
0
N
00
0 A4
" It
a) Q)
a
a
0
t-00
-
N
o
0
M,
U')
jU
aa
r:0
44
u
a
r-
.m
0)
0)
'a
00
0
E
a
5
-r-)a)I
r_
m
u.r-
M
44 4
0 a)
0a
0
4
I-
4i
aaa
I0
0a
aa
a
m
Le)
a
2
00
0
=
a
*
aa
12
0
m
Q)
-
4
CL
(A
a
O
C)
CD
C)
C)
C)
C)
CD
4r
VI
m
C)
CO
U')
C14
C)
C14
LO
C)
C)
CD
CD
CD
C)
lgdd-V'.(Ilsu;)Iul
CL4
a4 a4
0
:-
44
0 x
m0
a0
A ---
c
0)
4
C)
0
aC)
LO
C)
34
Table 32
Substrate
AES and ERD Results of ZrC from ZrNP4
Substrate
Pressure
Composition
Hydrogen
Temperature
(Torr)
(AES Depth
Content
Profile)
(ERD)
(OC)
SiG 10)
SiO 00)
400
10-4
400
10-2
Zrl.OC2.100.27a
570
10-4
Zrl.OC2.000.06
570
10-2
Zrl.OC2.100.07
750
10-4
Zrl.OC2.000.07
19%
570
10-4
Zrl.OC2.000.05
15%
570
10-4
Zrl.OC2.200.07
8%
880'C Anneal
a. Film changed color from silver to dark green before introduced to
AES/XPS system. This reaction accounts for the high oxygen content.
films.
his is especially true for the carbon concentration values as three
carbon species are present in the deposited materials (See XPS Analysis
section). A better method is to use a laboratory-specific reference - a
measurement of the pure material with the same equipment under the same
conditions.28 This requires obtaining a ZrC standard.
his method was
investigated. The ZrC standards used in the AES calibration proved to be
contaminated with graphite (See Appendix 1), and sensitivity factors could
not be extracted.
35
Table 33
AES Relative Sensitivity Factors as a Function of Electron
Beam Voltage
ELEMENT
Carbon
Oxygen
Zirconium
Titanium
RELATIVE SENSITIVIY VALUE
3 keV
5 keV
0.21
0.14
0.52
0.42
0.22
0.16
0.45
0.33
In one experiment the substrate holder was placed at the center of the
furnace, doubling the distance from the front of the hot zone to the
substrates. There was speculation that increasing the residence time of the
precursor could change the composition of the deposited material. The
reasoning for this relies on the assumption that gas phase decomposition of
the organornetallic is occurring in the deposition process. The increased
time in the hot zone would allow the precursor to more fully decompose
and possibly result in a film less carbon rich. AES depth profiling,
however, showed a very thin film with no change in the relative
composition of the deposited material. This experiment showed that the
film composition is not dependent on the substrate location in the reactor
tube and gas phase reactions do not play a significant role at pressures of
10-4 Torr. It was also apparent that most of the deposition took place on
the reactor walls in the vicinity of the original location of the substrate
holder. Thus it is noted that the substrate placement for most efficient film
growth at 570 'C is at the original location near the front of the hot zone.
36
3.1-6. XPS Analysis
Survey XPS spectra of the thin fms show Zr, C, and
as the
only observable elements (Figure 37). Zirconium XPS of the deposited
material indicate that Zr is in a single phase - a metal carbide (Figure 38).
The observed spectra correspond well to spectra observed for ZrC
standards: Zr 3d5/2, 180.0 eV; Zr 3d3/2, 182.5 eV. For several of the
films, a small shoulder is evident on the higher binding energy side of the
Zr 3d3/2 feature for ZrC. This is attributed to
accounts for the 2
rO2 in the thin film and
oxygen content typically observed by AES.
Carbon XP spectra show two features with a possible weak shoulder
to the low binding energy side of the C Is region (Figure 39 and 310).
These spectra were directly compared to the ZrC standards' XP spectra
(Appendix 1). The feature at 285 eV can be assigned to hydrocarbon or
C-C bonds.
he second feature at 283.7 eV is -1 eV higher in energy than
expected for the carbidic carbon of ZrC. Upon deconvolution (in which
X-ray satellite peaks, X-ray line widths, and analyzer broadening are
removed from the spectra), it becomes evident that the C XPS is made up
of 3 features (Figure 311 and 312): hydrocarbon or C-C bonds at
285 eV, carbidic carbon of ZrC at 282.7 eV, and a new unassigned C Is
feature at 283.6 eV. The exact nature of this third carbon peak is not
known at this time. It may correspond to a zirconium carbon bond that is
similar to the starting organometallic complex; Zr--C-, Zr-CHn- or
Zr=CH-. The energy shift from the observed carbidic carbon binding
energy is too large to be attributed to simply an amorphous carbide phase.
The deposition conditions were varied to determine the
interrelationship of the three carbon species observed. Increasing the
37
EA
a)
-P
-4
4i
Ln
4J
4
0
)
a i
En
4-4
LO
>1
Q)
4
z
to
W
14
Wx
.,I
4.
38
4-4
0
41
N
4
Q)
4-)
4
4
4
0
E-4
roIr
$4
0
4
144
u
4
4.4
0
0
C,
u
4
.1-4
N
CIO
C4
C,
14
a)
z
Z=
r-
5:
z
4=
c
fI
WN
4z
Z
4
0
v
.1-4
ri.
39
C!
4
-P
oc
r-
4i
4I
0
Ln
00
4i
4-4
u
oc
oc
4.4
WN
m
40
4
w
_U
ct:
00
C
CO
C
CN
co
C
cla, 'a
z
LO
m
co
W
CD
4J
C
0
r-
m
4
co
4-4
co
co
C
U
$-4
4
51
0
44
CN
$4
U
C)
c
C)
00
a)
O
00
r-
0
U')
Ln
(3)N
LO
41
a)
4i
::I
00
cu
Ln
4i
44
oc
C7,
Wx
VI
.1-4
4.
42
4
w
4-)
44
(O
$4
w
0
E-4
0
1-1
0
00
f11
Ln
Q,
'2
-e
I
u
I,-
-J,
---l
44
4
V)
.a)
L
It
- I
(n
4
4J
92
6
w
00
14
0
.5
c
'!>
wr
0
.s
10
44
u
-a
En
rz O
co
10
00
u
Q) O
r-I
A
-I
x
---:i
a
;61
8
'V
J,i
a
0
C
I
0C,
11
'r
-P
z
1-4
0
I-
1
f
0C
(a)N
c
c
c
0
e
c
1=
,4
a)
$-4
43
deposition pressure from 10-4 to 1-2 Torr has no effect on the carbon
make up of the deposited material. Variation in the distance of the
substrate from the front of the hot zone does not dramatically effect the
observed carbon XPS. Increasing the substrate temperature to 750 C
results in an increase in the carbidic content with a decrease in the
unidentified C Is feature (Figure 312). The annealing at 880 'C for
2 hours of a film deposited at 570 'C results in an increase in the relative
amount of the unknown carbon species with an apparent decrease in the
hydrocarbon or C-C bond species (Figure 313). There is no observable
variation in the carbidic content in the annealed thin film. At this time, no
conclusions can be drawn from the observed behavior of the carbon species
with deposition conditions.
3.1.7. ERD Analysis
Elastic recoil detection is used to determine the hydrogen
content in the thin films. Earlier efforts have observed residual hydrogen
in materials deposited from organornetallic precursors.29,30 ERD reveals a
large amount of hydrogen in the deposited material. The hydrogen content
of the ZrC films deposited at 570 'C and 10-4 Torr is
-1.1 x 1022 atoms/cm3, roughly 16% H. Increasing the deposition
temperature to 750 'C does not alter the hydrogen content. Annealing at
880 'C for 2 hours a film deposited at 570 'C results in a two fold decrease
in the hydrogen content to 5.0 x 1021 atoms/cm3 or -8% H. 'Me nature of
the hydrogen in the thin films is not known. From the zirconium XPS
there is no evidence for the presence of ZrH2. Therefore, the hydrogen
may be associated with the excess carbon in the material as hydrocarbon
44
N
$4
co
co
.0
Ln
4i
44
14
fo
u
Q)
4i
z
0
0
WN
45
fragments or hydrogen may be a dopant in the ZrC lattice occupying
carbon defect sites.29
3.2. Titanium Carbide Deposition
3.2. 1. General
'Me deposition of TiC from TiNP4 was studied using the same
conditions as in the ZrC experiments.
It was of interest to compare the
zirconium carbide results with previous work13. Depositions of TiC using
parameters described
300 C and
10-4
Torr ) in previous efforts9 were
also conducted to compare results from the two deposition systems. Thus,
thin films of TiC were deposited on Si(100) and Si(I 10) at pressure of
10-4
Torr and substrate temperatures of 300 and 570 'C.
he precursor
was prepared in the same fashion as the ZrNP4. During the depositions,
the bright-yellow TiNP4 precursor in the quartz holder darkened to a black
color, signifying a change of valence state of titanium. This is due to the
light sensitivity of TiNP4. In the zirconium carbide depositions, no color
change in the precursor was noted as ZrNP4 does not react upon exposure
to a light source.
3.2.2. Film Stability
Titanium carbide films deposited at 570 'C and 10-4 Torr on
Si(I 10) and Si(100) were silver-colored and mirror-bright. The films
exhibit good adhesion to the silicon substrates and appear stable upon
contact with ambient conditions. No apparent changes in color and no
cracking or crazing of the deposited material is evident even after months
of exposure. However, films deposited at 300 'C and 10-4 Torr are not
46
stable upon prolonged exposure to the ambient atmosphere. The color of
the film changes from a metallic silvery/gray to a red/blue-green tinge
after approximately one week of exposure. This color change is a result of
an oxidation reaction of the film with water vapor in the atmosphere No
cracking or de-lamination of these films is evident. Table 34 summarizes
the two experiments and their respective film stability results.
Table 34
TiC Deposition Process Using TiNP4 Precursor
Substrate
Deposition
Growth Rate
Substrate;
Temperature
Pressure
(gm/hr)
Stability
(OC)
(Torr)
570
I 10-4
0.5
SiG 10) NA
300
5xI0-5
0.2
SiG IO); days
(Time to degradA
NA = not applicable; film did not degrade.
3.2.3. X-Ray Diffraction
At substrate temperatures of 570 'C, thin films are deposited
on Si(I 00) and Si(I IO) substrates that exhibit weak reflections in X-ray
diffraction (Figure 314) that correspond to crystalline TiC.31 The small
peaks at 20 = 35.9' and 41.7' correspond to diffraction from the (III) and
(200) planes of TiC, respectively. Material deposited at 300 'C is
amorphous by X-ray diffraction (Figure 315). The weak reflections near
47
4
4
0
Ln
Ln
fo
4
-P
U)
cn
CL)
cu
C)
C-
CO
C31
cu
M
r.
0
r.
cu
0
X4
0
LO
44
E-4
m
44
0
O
M C:) 0
CD
Ln
VI"
.
X
n
'IT
0
0
'IT
CD 0
Ln
0
m
m
0
Ln
cu
0
0
0
0
Ln
0
Ln
cu
It"
0
(squnoo) 44sua4uj
0
4
I-t
Htr
44
48
D
4
4
0
4-J
tt
4
4J
U)
U)
U)
C)
.
0
CD
CD
CD
C-
CT)
CD
(a
4-J
0
CD
.C=
0
4
C) C\J
CD
m
r
44
u
.H
44
0
0
0
117
U)
4
Cr) C CD )
0 0
I".
m 0
) O C 0
0
11.7N 0 m 0
N - - - -
0
0
17 Cj
0 0 0 0
(squno:)) Ajisuaquj
z
m
H
;L4
49
28 = 40' evident in Figure 315 do not correspond to diffraction from any
crystallographic orientation of TiC.
3.2.4. Film Morphology
Optical microscopy shows a smooth, featureless surface for
TiC material deposited at 570 and 300 'C. Scanning electron micrographs
do not reveal any grain structure. Similar to the ZrC thin films, some
surface roughness was observed at very high magnification and is
attributed to a three-dimensional island growth process.25 No large
clusters were observed on the surface of the film.
3.2.5. Film Thickness and Average Growth Rate
The film thickness for the 570 'C experiment ranges from
1.0 gm to 0.5 ptm as measured in cross-sectional micrographs.
At the
lower temperature of 300 'C, the observed growth rate decreased to
approximately 02 ptm/hr. A thorough analysis of the average growth rate
is not possible due to the limited number of experiments conducted with
TiNP4.
3.2.6. AES Analysis
Survey AES spectra of the deposited material show Ti, C, and
0 as the only observable constituents of the thin film. Figure 316 is an
AES spectrum of a TiC film deposited at 570 'C and 10-4 Torr after
removal of surface contamination. Auger depth profiling of the film
reveals a uniform composition. The depth profile in Figure 317 shows a
constant ratio of titanium to carbon throughout the bulk of the material.
Even though the average peak-to-peak height for each element fluctuates as
50
0
4
4J
44
4
4
0
cc
E-4
O
LO
4J
4J
0
0
4
44
u
A,
E--4
44
4
>
U)
4
ap/mp
no
w
51
00e13
-
a
w
4
0
E-4
05
a00
co
00
0rl-
u
Ln
4-)
EMEA
C)
C)
CD
co
m
4 -4
.1.4 0)
LO a
0
aCD
C)
LO
0
ri
99
4;
s
0CD p
44
u (z
E-4a
.1-4 (1)
C)
44 it
0 w
a)
>
00
0cn
a:,L
Lrl
(i-
00
CD
Y:
-4
.,j
4-4
0
4 X
04 a-'
4-)
CN
U)
CD
CL
CD
C)
O
0
c
14
-,L
CD
00
Ln
m
00
0
m
00
LO
C14
00
0
0C)
LO
04
HdaV',il!su3lul
00
2
0CD
Ln
C)
CO
53
C)
C)
CD
co
;
4
w
0
C)
Cl
C
r.-
u
i
C
C
C)
,-r
co
4-1
tt 4-j
'L"
1,
0
Q)
C)
C)
i
0
O
Ul)
111;1
0-
i
A4
10
dF
O
CD
C)
'IT
(i
Y
4.4
1
4;
E
E-4 Q)
0
U
04
I
0
,>O
44
Ito#
C)
C)
C)
0
>
Cl)
44
*11
18
0
z6%,I
c
0
C)
C)
4-)
O
C)
Q
co
C
CN
Z
r_
0)
W
Q.
W
I-I
A
'T
0
I
,A
4
Z
C)
C)
U')
04
C)
O
C)
C14
00
Ul)
C)
C)
C)
Hdjv 1psualuj
C)
C)
Ul)
Z
1
C)
Q
4
54
Table 35
Substrate
AES and ERD Results of TiC from TiNP4
Substrate
Pressure
Composition
Hydrogen
Temperature
(Torr)
(AES Depth Profile)
Content
(OC)
Si(i 10)
(ERD)
300
10-4
TiLOC2.700.31
--
570
10-4
Ti .OC2.600.02
13%
3.2.7. XPS Analysis
Survey XPS spectra of the thin film show Ti, C, and
as the
only observable elements (Figure 319). For the TiC film deposited at
570 'C, the titanium XPS (Figure 320) indicates that titanium is present
only in a single phase - the metal carbide. The observed values for the
Ti 2pi/2 and Ti 2P3/2peaks are 461.4 eV and 455.4 eV, respectively.
These values correspond well to results from previous work: Ti 2pl/2,
461.2 eV; Ti 2P3/2, 455.1 eV.9 For the film deposited at 300 'C, the
titanium is single phase and the observed values of the Ti 2p peaks also
correspond
ell to previous efforts. The peaks are broader (Figure 321)
than those corresponding to the 570 'C TiC film. This peak broadening is
due to the amorphous nature of the 300 'C TiC film.
Carbon XPS spectra exhibit only two states for TiC films deposited
at 570 and 300 'C (Figures 322 and 323). Carbon associated with
hydrocarbon or C-C bonds is assigned to the feature at 285.1 eV. The
lower binding energy C Is band at 282.5 eV is assigned to the carbidic
carbon of TiC. No third unassignable feature is observed. Upon
55
U)
Q)
4
FE
4
4-4
fla
4
4
0
E--4
"a,
I
0
u
44
u
44
0
4
4-)
u
>4
4
CL4
X
0%
1--i
,4
a)
4
.1-4
4.
56
4
c
.17
-7:
-I;_
-1
Ln
i
j
-4
A
- -- l
fo
4
4
0
ELn
I
;I-'_
w
t
i'I
T
i2
1.
-
C,
IC
11
I
w
0
w
"
u
C
rLn
-W
fa
.s
-
.-
-1 _;,
zIr
-
0
0
m
4
r-T
10
-4-4
u
-
`4
IC
S,1f
4
0
44
-
,cc
M
10
IT
14
WN
57
4
a)
I
-W
U)
IT
oc
>
4i
0
4
t4
E--4
4
0
44
U)
FE
00
IC
.14
.1i
CT.
58
4
O
-W
4i
U)
4
4
0
Ln
LO
rz
44
oc
u
--A
4
0
44
x
C
0
4
4
59
$-4
0)
4i
4-)
t
IZ-1
I
10
r-
I
C
i
i
-
r1
I
:c
-C
-
-
E--4
oc
C
z-m
1
Q
I-
-
W
00114, 0
w
CD
0
M
.S
-c
-2
I-
co
0
4
-
E--4
-1-1
k
tt
4
0
4-4
c
-:E
cl
co
0
.Q
4
M
u
i;
4;
-"t-
a,
1
5Z
12
cc
I-
rn
N
,4
a)
4
(a)K
60
deconvolution, it is clearly evident that only two carbon phases are present
(Figures 324 and 325). An interesting characteristic in Figures 322 and
3.23 is the ratio of hydrocarbon to carbidic carbon. The carbon XPS of
the lower temperature film shows a relatively larger amount of
hydrocarbon. This is expected considering that the lower temperature film
is amorphous by XRD. It has been argued that an amorphous TiC film can
be deposited if organic molecules are trapped in the material. The
presence of organic molecular fragments and hydrocarbon clusters reduces
the molecular migration for Ti and C. Hence, this limits the process of
nucleation ad growth of crystallites.9
3.2.8. ERD Analysis
Elastic recoil detection reveals a large amount of hydrogen in
the material deposited at 570 'C and 10-4 Torr.
'he hydrogen content is
-1.2 x 1022 atoms/cm3, roughly 13% H. As with the ZrC material, the
hydrogen may be associated with the excess carbon in the material as
hydrocarbon fragments or hydrogen may be a dopant in the TiC lattice
occupying carbon defect sites.29 The film deposited at 300 'C was not
characterized by ERD.
6
N
4
4i
44
4
4
0
E-4
Ln
C
Ln
4i
0
4
44
4J
E-4
i
4J
4
0 c
4-4
oc
4
u
14
4
WNI
62
4
0
4i
I
,1
44
(a
4
4
i
I--
0
-IC
ls-
' i
I
I
4J
-k
0
-
a,
'O
tr
>
It
v
.0
I-
m
u
-I--
I",
00
w
0
.
w
u
JE 4
1-4 0)
-4 4J
44
U 04
.1-4 En
.S
c
.8
Im
0L-
4
x
c
j
1-
co
0
-
x
5
0
1%,
u
4J
0
C"I
0
u
11
Ln
N
,4
a)
4
E
I
(3)N
C:
44 -1-1
63
4. CONCLUSION
4.1. Comparison of ZrC and TiC Systems in Present Work
Similar results have been obtained in our system for the deposition
of ZrC from ZrNP4 and TiC from TiNP4. Deposition temperatures
greater than 500 'C are necessary to grow stable thin films. Above this
temperature, the deposited materials are crystalline by XRD. A constant
metal to carbon ratio of 1:2 is observed for both materials by AES depth
profile. 'Me residual oxygen content remains low in films deposited above
500 'C for both materials. XPS reveals carbidic and graphitic or
hydrocarbon species in both the ZrC and TiC films. However, a third
unknown carbon species is present only in the ZrC material. The titanium
and zirconium XPS indicate the metals are single phase. The observed
spectra correspond well to spectra observed for carbide standards (for
ZrQ and to spectra reported in previous work (for TiC). ERD finds a
large, up to 16 %, hydrogen content for both the TiC and ZrC materials.
The large hydrogen content decreases the density of the metal carbide
material. This can lead to problems in some applications, especially in
coatings for nuclear fuels where its effectiveness as a barrier to diffusion
of fission products decreases as density decreases. Further work is
required in alleviating this deficiency.
4.2. Differences from Previous Efforts
These results are essentially consistent with the findings of the
previous efforts with some significant differences observed, particularly a
third unidentified carbon state in the ZrC system. A further difference is
the incorporation of residual carbon. In these experiments, a metal to
64
carbon ratio of 1:2 is determined by AES. In previous work with TiC a
metal to carbon ratio of 1:0.95 is reported.9 In previous work, the
presented XPS spectra tend to show a larger amount of carbon than that
reported. This could be a result of a lack of or inadequate deconvolution
of the carbon XPS spectra. It is believed that a more thorough XPS
analysis performed in this project is the likely source for the larger carbon
content relative to results from previous work. A number of studies are
currently being pursued at Los Alamos National Laboratory in an effort to
further understand these differences.
5. FUTURE WORK
Several studies are necessary to further understand the mechanisms
responsible for the presence of a third unknown carbon species in the ZrC
material and for the large amounts of free carbon and hydrogen in both
materials. The following are recommended studies
- Mass spectral analysis of the cracking gases given off during the
deposition.
• Surface chemical studies of the deposition process.
• Parametric study of the effects on the deposited material from variations
of the deposition conditions.
• Use of molecular hydrogen as a carrier gas to lower the total residual
carbon content in the thin films
• Deposition of other metal carbides (CrC and HfC from CrNP4 and
HfNp4) to help understand the differences in the ZrC and TiC systems.
65
A better understanding of the differences in the OMCVD studies will
result from the additional information and will lead to a new, reliable lowtemperature process for metal carbides.
66
APPENDIX 1. ZRC STANDARDS
Zirconium carbide standards were obtained in an effort to determine
the Auger sensitivity factors for zirconium and carbidic carbon in their
compound form. The sensitivity factors are required to determine an
accurate composition of the ZrC films deposited via OMCVD. The use of
published sensitivy factors for zirconium and carbon introduces error in
the composition calculation as these values correspond to elemental
standards.
Two different ZrC standards were analyzed. An arc-melted ZrC
standard was provided by a Nuclear Materials and Technology group
(NMT-1) at Los Alamos National Laboratory. A second ZrC standard was
produced via physical vapor deposition (PVD) from a purchased ZrC
target. The PVD ZrC standard consisted of a ZrC film deposited on a
Si(100) substrate by magnetron sputtering. A unifon-n composition for
both standards was observed by AES depth profiling. XPS was conducted
on the standards to obtain reference spectra. The zirconium XPS indicated
the presence of only one phase of the metal - the metal carbide
(Figures A.1 and A.2). In both standards, the carbon XPS showed the
presence of two species; a carbidic carbon and hydrocarbon or graphite
(Figures A.3 and A.4). Because of the hydrocarbon or graphite
contamination, accurate sensitivity factors could not be determined.
However, the XPS spectra for both zirconium and carbon proved useful by
providing accurate reference binding energy values.
hese spectra were
directly compared to XPS spectra for the OMCVD deposited ZrC films.
The results of these comparisons are presented in the XPS Analysis section
of the Zirconium Carbide Deposition chapter.
67
4J
rz
0
1--i
4
44
7x
.4i
>
--j
(n
t9
-
u4
6
N
lu
-c
m;4
-11,
,u
10
.s
t"
0)
4J
1--i
Q)
F.
.S
co
l
I
u
4
44
0
OC
0
Ol
WN
J
68
4
a)
4-J
4-4
-------
r-
-
0
i
En
0
ii
i
4
0is,
I
-
I-
u
-4
U)
-I
>1 i
4i
-
-I--
-Z !i::-
1:1 -I
0e)
--
Q) En
-W
.H (n
(n cu
0
aQ) 4-)r.
4
-I,-
-Sl
11
I-
4i
U)
u
t
44
-
0,
0
!s
U)
-4
a4
x
F
0
u
_
(a),
I
4
L4
Ln
69
C
C
co
CL
(n
0
co
E
C
O
CN
C4
co
cn
co
C
N
7
E
C
co
0
cn
co
co
C
EL
x
C:
0
cu
L)
co
Ln
MN
to
Im
70
LO
4
"I
4-)
44
C
72
04
C)
4-)
>i 4J
00
Q)
4J
U) 4-)
4
oc
oc
4-J
U)
4
0
4
MN
:1
71
REFERENCES
1. P. Schwarzkopf, P. Kieffer, W.Leszynski and F. Benesovsky,
Refractory Hard Metals (Macmillan, New York, 1953).
2. E.K. Storms, 'Me Refractory Carbides (Academic, New York, 1967).
3. L.E. Toth, Transition Metal Carbides and Nitrides (Academic, New
York, 197 1).
4. G.V.Samsonov, Ed., Refractory Carbides (Consultants Bureau, New
York, 1974).
5. G.V. Samsonov, Te Properties of Refractory Compounds
(Metallurgizdat, 1962).
6. A. Strasser, Nuclea
5, 353 1960).
7. C.M. Hollabaugh, L.A. Wahman, R.D. Reiswig, R.W. hite, and P.
Wagner, Nuclear Techn
, 35, 527-535, (September 1977).
8. T.Ya. Kosolapova, Carbides (Plenum Press, New York, 1971).
9. A.E. Kaloyeros, W.S. Williams, C.M. Alloca, D.M. Pollina, and G.S
Girolami, Advanced Ceramic Materials, 2 [3A], 257-263, 1987).
10. G.V. Samsonov and P.S. Kislyi, High Temperature Nonmetallic
Thermocouples and Terminals (Izd. "Naukova Dumka," Kiev, 1965).
11. C.F. Powell, J.H. Oxley and J.M. Blocher, ed., Vapor Deposition
(Wiley, New York, 1966).
12. D.C. Smith, J.R. Laia, "Organometallic Chemical Vapor Deposition",
Los Alamos National Laboratory FY 1991 Progress Report, Los
Alamos National
Laboratory, November 10, 1991.
1
13. G.S. Girolami and J.E. Gozum, MRS Symp. Proc., 168, 319 1990),
and references therein.
14. E.J. Cukauskas, R.T. Holm, A.D. Berry, R. Kaplan, and M.L.H.
Green, IEEE Trans. Magnet., 23, 999-1002 1987).
72
15. P.D. Stupic, L.K. Cheatham, J.J. Grahm, and A.R. Barron, MRS
168, 363-367 1990).
.9
16. D.C. Smith, R.R. Rubiano, M.D. Healy, R.W. Springer, MRS Symp.
Proc., 282, 643-649 1993).
17. F. Maury, F. Ossola, Thin Solid Films, 207, 82-89 1992).
18. F. Maury, F. Ossola, F. Schuster, Surface and Coatin2s Technoloa,
54155, 204-210
1992).
19. C. Jiang, T. Goto, T. Hirai, J. of Mat. Sci, 25, 1086-1093 1990).
20. G. Huchet and F. Teyssandier, in Proceedins of the Eleventh
International Conference on Chemical Vor Deposition, edited by
K.E. Spear and G.W. Cullen (The Electrochemical Society,
Pennington, NJ 1990), pp. 703-709.
21. N.M. Rutherford, C.E. Larson, and R.L. Jackson, MRS Symp. Proc.,
1319 439
1989).
22. G. Girolarni, J. Jensen, D. Pollina, W. Williams, A. Kaloyeros, and C.
Allocca, J. Am. Chem. Soc., 109, 1579-1580 1987).
23. P. Davidson, M. Lappert, and R. Pearce, Journal of
Ch
, 57, 269-277 1973).
r2anornetallic
24. Deconvolution routines were written at Los Alamos National
Laboratory.
25. J.L. Vossen and W. Kem, Eds., Thin Film Processes 11(Academic,
New York, 1991).
26. B.D. Cullity, Elements of X-ray Diffraction (Addison-Wesley,
Readings, 1978).
27. L.E. Davis, N.C. MacDonald, P.W. Palmberg, G.E. Riach, R.E.
Weber, Handbook of Auger Electron S12ectroscQpy(Physical
Electronics Industries, Inc., Eden Prairie, Minnesota, 1978).
28. J.M. Slaughter, W. Weber, G. Guntherodt, and C.M. Falco, MRS
Bulletin, XV11 No. 12, 39 (December 1992).
73
29. A.E. Kaloyeros, W.S. Williams, F.C. Brown, A.E. Greene, and J.B.
Woodhouse, Phys. Rev.
37, 771 1988).
30. R. Fix, R.G. Gordon, D.M. Hoffman, Chem. Mater 3 1138 1991).
31. G.S. Girolami, J.A. Jensen, D.M. Pollina, W.S. Williams, A.E.
Kaloyeros, and C.M. Alloca, J. Am. Chem Soc., 109, 1579 1987).