Surface chemistry model of cobalt-molybdenum catalysts supported on planar [gamma]-Al203
by Timothy Dean Kirkpatrick
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in
Chemical Engineering
Montana State University
© Copyright by Timothy Dean Kirkpatrick (1986)
Abstract:
Cobalt-molybdenum catalysts supported on γ-Al2O3 are well known for their success in the
hydrotreating of petroleum feedstocks and coal liquifaction products. Their surface structure has been
studied by several methods including recent attempts by surface science techniques. Applications of the
surface science techniques are limited due to the porosity and insulating character of the γ-Al2O3
supports. This study presents an alternative approach Dy replacing the porous γ-Al2O3 supports with
planar γ-Al2O3 grown on aluminum substrates: Preparation of the model catalysts was performed
completely in situ. The γ-Al2O3 was prepared by thermally oxidizing the planar, polycrystalline
aluminum substrates. The molybdenum and cobalt were vapor deposited to approximately monolayer
coverage and subsequently oxidized. The planar model catalysts were characterized by ESCA, SIHS
and AES and their surface structure compared to that of industrial catalysts.
The ESCA Al 2p band was used to monitor the oxidation of the aluminum. Shifts in the binding energy
enabled differentiation between amorphous and crystalline Al2O3. Subsequent deposition and
oxidation of molybdenum resulted in broadening and binding energy shifts of the Al 2p and Mo 3d
bands suggesting interaction between the molybdenum and the γ-Al2O3.
Addition of cobalt onto the molybdenum layer appeared to result in an interactive phase between the
cobalt and the molybdenum. The Co 2p band indicated the cobalt was in tetrahedral coordination.
Depth profiles, obtained with SIMS, indicated the cobalt was interacting primarily with the
molybdenum. When the order of metals deposition was reversed, the cobalt appeared to interact
primarily with the γ-Al2O3 in a CoAl2O4 type phase. The AES line scans suggested that both the
cobalt and molybdenum were well dispersed as would be characteristic of a uniform monolayer
coverage.
The findings of this study indicate similarities between the surface structure of the planar model
catalysts and the industrial catalysts. SURFACE CHEMISTRY OF MODEL COBALT-MOLYBDENUM CATALYSTS
SUPPORTED ON PLANAR Y-Al3O3
by
Timothy Dean Kirkpatrick
A thesis submitted in partial fulfillment
of the requirements for the degree
of
Master of Science
in
Chemical Engineering
MONTANA STATE UNIVERSITY
BozemanrMontana
June 1986
MAIM LIB.
VJ 72
ii
APPROVAL
of a thesis submitted by
Timothy Dean Kirkpatrick
This thesis has been read by each member of the thesis
committee and has been found to be satisfactory regarding
content, English usage, format, citations, bibliographic
style, and consistency, and is ready for submission to the
College of Graduate Studies.
4-j S 6
Date
Chairperson,Graduate Committee
Approved for the Major Department
6. / m
Date 0
, J
^ead,Major Department
I
Approved for the College of Graduate Studies
Date
Graduate Dean
STATEMENT OF PERMISSION TO USE
In presenting this
the requirements for
University,
I
agree
a
thesis
in partial fulfillment of
master's
that
the
degree
Library
at Montana State
shall
make
available to borrowers under
rules
quotations from this
are allowable without special
thesis
of the Library.
it
Brief
permission, provided that accurate acknowledgment of source
is made.
Permission
for
reproduction of this
professor,
or
in
extensive
thesis
his/her
may
be
absence,
Libraries when, in the opinion
of the material is for
quotation
from
or
granted
by my major
by
Director of
the
of either, the proposed use
scholarly purposes.
Any copying or
use of the material in this thesis for financial gain shall
not be allowed without my written permission.
Signat
Date_____
z
iv
TABLE OF CONTENTS
Page
TITLE P A G E .................. ......................
i
APPROVAL ..........................................
11
STATEMENT OF PERMISSION TO U S E ............ .. . . .
TABLE OF CONTENTS
..........................
iv
LIST OF T A B L E S .................................. ..
vi
LIST OF FIGURES
...
iii
. . . .............................
ABST R A C T................ .................... . . .
vii
ix
\
INTRODUCTION.......... ...........................
I
BACKGROUND............ ..
......................
i
Importance of Co-MoZ1Y-Al2D 3 Catalysts..........
4
Surface Structure of Co-MoZy-Al2O3 Catalysts . .
Limitations of the Literature and Experimental
Methods . . . . ................................
6
Planar Model Catalysts
as, an Alternative . . . .
Application of ESCA in Catalysis Research . . . .
12
14
Application of AES in Catalysis Research . . . .
Application of SIMS in Catalysis Research .
..
17
19
RESEARCH OBJECTIVES
...............................
EXPERIMENTAL . . . . I .......... ..............
Sample Handling System
........................
4
9
22
23
23
Preparation of Planar 1Y-Al2O3 ..................
Deposition of Cobalt and Molybdenum............
Oxidation of Cobalt and M o l y b d e n u m ........ .. .
ESCA A n a l y s i s .......... ........................
25
26
29
30
AES A n a l y s i s .......... ........................
SIMS Anal y s i s .......... ...................... .
31
31
V
Page
33
RESULTS AND DISCUSSION
Growth of Y-Al2O g ..............................
Co -Mo Zy --A^O3 Model Catalyst ESCA Analysis . . .
Deposition and
Oxidation of Molybdenum. . . .
33
33
33
Deposition
Mo-CoZY-Al3O3
Deposition
Deposition
41
48
49
55
and
Model
and
and
Oxidation
Catalyst
Oxidation
Oxidation
of Cobalt . . . . . .
ESCA Analysis . . .
of Cobalt ........
of Molybdenum. . . .
Co-MoZY-Al3O3 Model Catalyst SIMS Analysis . . .
Co-MoZY-Al3O3 Model Catalyst AES Analysis . . . .
Comparison of Planar Models to Industrial
Catalysts............................
59
62
65
SUMMARY AND CONCLUSIONS . .........................
70
RECOMMENDATIONS........................ ..
72
. . . .
REFERENCES CITED ..........................
74
APPENDICES............ '...........................
78
' Appendix A- Sample Calculations ................
79.
Appendix B- Peak Binding E n e r g i e s ........ ..
83
Appendix C- X-ray Induced Auger Spectra ........
86
vi
LIST OF TABLES
Table
I.
2.
3.
4.
5.
6.
7.
8.
Page
Oxidizing Conditions for the Cobalt
and Molybdenum ........................
....
29
Peak Widths for the Co-Mo/y-Al-O?
Model Catalyst ........................
. . . .
38
Peak Areas and Area Ratios for the
Co-Mo/y-AlgOg Model Catalyst ..........
....
40
Atomic Concentrations for the
Co-Mo/y-AlgOg Model Catalyst ..........
. . . .
48
Peak Areas and Area Ratios for the
Mo-Co/y-AlgOg Model Catalyst ..........
....
54
Peak Widths for the Mo-CoZy-Al=O^
Model Catalyst ........................
....
55
Peak Binding Energies for the
Co-MoZ1Y-Al2O3 Model C a t a l y s t ..........
....
83
Peak Binding Energies for the
Mo-CoZy-Al2O3 Model Catalyst ..........
, . . .
84
vii
LIST OF FIGURES
Figure
Page
1.
ESCA Al 2p SampleS p e c t r a .....................
16
2.
Sample Handling R o d ..........................
24
3.
Al 2p ESCA Spectra Displaying Oxidation
of Aluminum Foil ...............................
34
Al 2p ESCA Spectra for the Co-MoZy-Al0O0
Model Catalyst.................... 7 . . . .
35
4.
5.
.
Mo 3d ESCA Spectra for the Co-MoZy -AI0O0
Model Catalyst After Molybdenum
2 3
Deposition and Molybdenum F o i l ................
36
6.
Mo 3d ESCA Spectra for the Co-MoZY-Al7O0
Model Catalyst After Molybdenum Oxidation
and Oxidized Molybdenum F o i l .................... 37
7.
O Is ESCA Spectra for the Co-MoZy-Al0O0
Model Catalyst.................... . 3 ........
39
Co 2p ESCA Spectra for the Co-MoZy-Al0O0
Model Catalyst After Cobalt Deposition 3 ,
and Cobalt F o i l ..............................
43
Co 2p ESCA Spectra for the Co-MoZy-Al0O0
Model Catalyst After Cobalt Oxidation2 3
and Oxidized Cobalt F o i l ...............
45
8.
9.
10.
11.
12.
13.
Al 2p ESCA Spectra for the Mo-CoZy-Al0O0
Model Catalyst.................... 7 3 . . .
.
50
Co 2p ESCA Spectra for the Mo-CoZy-Al0O0
Model Catalyst After Cobalt Deposition 3
and Cobalt F o i l ..............................
51
O Is ESCA Spectra for the Mo-CoZy-Al0O0
Model Catalyst.................... . 3 ........
52
Co 2p ESCA Spectra for the Mo-CoZy-Al0O_
Model Catalyst After Cobalt Oxidation2 3
and Oxidized Cobalt F o i l ...............
53
viii
Fxqure
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Page
Mo 3d ESCA Spectra for the Mo-Co/y-Al^O-,
Model Catalyst After Molybdenum
2 3
Deposition and Molybdenum F o i l ................
56
Mo 3d ESCA Spectra for the Mo-Co/-y-Al„0,.
Model Catalyst After Molybdenum Oxidation
and Oxidized Molybdenum F o i l ..................
58
SIMS Depth Profiles of the Co-MoZ-Y-Al-O-,
Model Catalyst.................... 7 . . . .
.
60
SIMS Survey of the Co-Mo/-y-Al-O^
Model Catalyst............ 2 T . . ...........
61
AES Line Scans of the Co-MoZy-Al-OModel Catalyst................2 .3.............
63
SEM Micrograph of the Co-MoZy-Al-OModel Catalyst Surface ........................
64
O KLL X-ray Induced Auger Spectra of the
Co-MoZy-AlgOg Model Catalyst ..................
87
Co LMM X-ray Induced Auger Spectra of the
Co-MoZy-Al-O- Model Catalyst and Cobalt
Foil . . . .....................................
88
Mo MNN X-ray Induced Auger Spectra of the
Co-MoZy-Al-O- Model Catalyst and Molybdenum
Foil . . . .....................................
89
O KLL X-ray Induced Auger Spectra of the
Mo-CoZy-AlgOg Model Catalyst ..............
90
Co LMM X-ray Induced Auger Spectra of the
Mo-CoZy-Al-O-, Model Catalyst and Cobalt
Foil . . . .............................
. .
91
ix
ABSTRACT
Cobalt-molybdenum catalysts supported on y-Al.O? are
well known for their success in the hydrotreating of
petroleum feedstocks and coal liquifaction products. Their
surface structure has been studied by several methods
including recent attempts by surface science techniques.
Applications of the surface science techniques are limited
due to the porosity and insulating character of the y-Al 0
supports. This study presents an alternative approach^by
replacing the porous y-Al 0_ supports with planar y-Al 0
grown on aluminum substrates:
2 3
Preparation of the model catalysts was performed
completely in situ. The y-Al_0_ was prepared by thermally
oxidizing the planar, polycrystalline aluminum substrates.
The molybdenum and
cobalt
were
vapor deposited to
approximately monolayer coverage and subsequently oxidized.
The planar model catalysts were characterized by ESCA, SIHS
and AES and their surface structure compared to that of
industrial catalysts.
The ESCA Al 2p band was used to monitor the oxidation
of the aluminum.
Shifts in the binding energy enabled
differentiation between amorphous and crystalline Al0O .
Subsequent deposition and oxidation of molybdenum resulted
in broadening and binding energy shifts of the Al 2p and Mo
3d bands suggesting interaction between the molybdenum and
the Y-Al0O3 .
Addition of cobalt onto the molybdenum layer appeared
to result in an interactive phase between the cobalt and
the molybdenum. The Co 2p band indicated the cobalt was in
coordination.
Depth profiles, obtained with
SIMS, indicated the cobalt was interacting primarily with
the molybdenum. When the order of metals deposition was
reversed, the cobalt appeared to interact primarily with
the Y-Al0O in a CoAl 0
type phase.
The AES line scans
suggested that both the cobalt and molybdenum were well
dispersed as would be characteristic of a uniform monolayer
coverage.
J
The findings of this study indicate similarities
between the surface structure of the planar model catalysts
and the industrial catalysts.
J
I
INTRODUCTION
Cobalt-molybdenum
catalysts
are
well
on
gamma
known
alumina (Co-Mo/y-AlgO^)
for
their
success
in
the
hydrotreating of petroleum feedstocks and coal liquifaction
products.
nitrogen
These
and
fuel
sources
sulfur,
contain
which
environmental standards.
must
elements, such as
be
removed
to
meet
The most common catalyst for this
purpose is sulfided Co-Mo/y-AlgOg.
In the past, manufacture of catalysts, such as
Co-Mo/^-AlgOg, has been regarded
more
successful recipes were acquired
by trial and error rather
than scientific
understanding
CU.
However, over the
more
interested
in
of
the
as
an art in which
processes involved
past decade scientists have become
the
design
of
catalysts
with well
defined properties.
This interest has been
spurred
application of modern surface
to
catalysis
research.
science and other techniques
These
researcher with the ability
to
structure and reaction
phenomena
Improved understanding
of
and how it
relates
to
by the advent of the
the
techniques
provide
the
study the catalyst surface
on
a microscopic level.
catalyst surface structure
catalytic activity and selectivity
will allow for the design of improved catalysts.
This idea
2
is particularly
enticing
since the activity
and
therefore the surface
for
the
Co-Mo/y-AlgOg catalyst
selectivity
structure,
of this catalyst, and
is
greatly dependent on
the preparation parameters [23.
A large volume of research on Co-Mo/y-AlgOg catalysts
has been conducted with the
its surface structure.
as
electron
goal of trying to characterize
Numerous analytical techniques such
spectroscopy
for
chemical
Mossbauer emission spectroscopy
infrared
spectroscopy
(ISS), and x-ray
these studies.
has been
While
the
the
(MES), Raman spectroscopy,
ion-scattering
diffraction
successfully
debate as to
(IR),
analysis (ESCA),
(XRD)
spectroscopy
have been employed in
application of these techniques
demonstrated,
structure
of
the
there
is still much
catalyst and how the
structure relates to activity and selectivity.
Because the
industrially
very porous insulators, some
science techniques for
adsorption
phenomena
employed.
These
used
Y-Al2O3 supports are
of
the more powerful surface
studying
the surface structure and
of
the
techniques
catalyst
include
have
secondary
not
been
ion mass
spectrometry (STMS), ultraviolet photoelectron spectroscopy
(UPS), electron energy
loss
spectroscopy (EELS) and Auger
electron spectroscopy (AES).
If model Co-MoZy-Al2O3 catalysts
planar
Y-Al3O3
surface,
the
surface
were supported on a
science techniques
3
listed above could be
used to investigate their adsorption
properties and surface structure.
The
procedures
objective
for
of
this
preparing
study
planar
is
to
model
develop
the
Co-MoZ-Y-Al2O3
catalysts and to study their surface structure to determine
if the planar models
catalysts.
are
representative of the industrial
4
BACKGROUND
Importance of Co-MoZ-v-Al^Oj Catalysts
Coal liqulfaction
consist predominantly
products
and petroleum feedstocks
of
various
hydrocarbons,
contain lesser amounts of
sulfur,
nitrogen, and oxygen in
the form of a variety
organic compounds.
of these
of
heteroatoms,
particularly
sulfur
but also
The presence
and nitrogen,
pose several problems.
The levels of both sulfur
meet EPA emission standards.
and nitrogen in fuels must
The presence of nitrogen and
sulfur in refining feedstocks can
refining operations.
The
also lead to problems in
organosulfur content of the feed
to a hydrocracker must be reduced to avoid poisoning of the
hydrocracking catalyst.
Heterocyclic nitrogen containing
compounds such as
quinolines,
strongly basic.
These
particular concern
in
pyridines and acridines are
strongly
basic
acid-catalyzed
compounds are of
reactions,
such as
catalytic cracking, since they poison the catalysts C33.
The removal
refining is
known
sulfur removal is
nitrogen removal as
and HDS procedures
of
sulfur
as
known
and
nitrogen
hydrotreating.
More specifically,
as hydrodesulfurization (HDS) and
hydrodenitrogenation
are
from fuels in
(HDN).
The HDN
basically catalytic hydrogenations
5
designed to selectively remove sulfur and nitrogen from the
various organic compounds and convert them HgS and NHg.
In general, a catalyst
the rate of a
consumed
chemical
in
the
is a substance that increases
reaction without being appreciably
process.
The
providing an alternate path
the reaction to proceed.
catalyst
does
this by
of lower activation energy for
In the hydrotreating of refinery
feedstocks where numerous reactions are possible due to the
variety of compounds present, the catalyst must selectively
catalyze the desired reaction(s).
heterocyclic
compound
is
For example, if HDS of a
desired,
the
catalyst
must
selectively promote replacement of the sulfur with hydrogen
while allowing
minimum
hydrogenation
of
the unsaturated
aromatics.
One
of
the
hydrotreating
is
more
commonly
Co-Mo/y-AlgOg.
particularly effective for HDS.
is prepared
by
used
for
catalyst
is
Industrially, the catalyst
the
porous y-AlgOg supports
with salt solutions of the metals.
The metals are calcined
to their oxides.
impregnating
The
catalysts
Prior to
insure proper activity.
are usually in the range
C23.
use, the metals are sulfided to
The composition of these catalysts
of
3-6
wt%
Co and 12-18 wt% Mo
The balance of the composition is the support.
metal composition ranges
These
correspond to approximately those
required for monolayer coverages.
6
Surface Structure of Co -Mo /y -A1„0j Catalysts
Perhaps because
removal
of
sulfur
environmental
studied of the
AlgOg.
of
their
from
petroleum
importance
of
more
in the selective
feedstocksr
sulfur
hydrotreating
Much of the
success
removal,
catalysts
and
the
the
most
has been Co-MoAy-
recent research on this catalyst
has focused on the utilization of surface science and other
techniques to characterize
Despite the number
of
the catalyst surface structure.
these investigations performed, the
nature of the surface species on this catalyst is still not
fully understood C4, 53.
Generally, the role of Co is believed to be that of a
promoter
promoting
while
Mo
is
the
active
role
of
Co
is
not
summarized by Massoth
C63,
been ascribed to (I) an
intercalation
fully
(3)
effect
The
understood.
As
promoting
increase
increase in Mo reduction;
(4) an
the
C53.
component
role of Co has
in Mo dispersion; (2) an
an increase in Hg mobility;
with
MoS2;
(5) a synergism
between MoS2 and CogSg crystallites; (6) a specific kinetic
effect; (7) a decrease in
surface segregation of
deactivation; (8) an increase in
mixed
sulfide
prevention of MoS2 crystallization.
may have some
influence
on
phases; and (9) the
All of these factors
the activity of Co-MoZy-Al2Og
catalysts [73.
Studies of the surface structure of the catalyst have
been conducted with
the
catalyst
in
both the oxidic and
7
sulfided states.
is
the
active
While
form
understanding of
the
the
sulfided form of the catalyst
used
in
oxidic
hydrotreating,
precursor
a thorough
is important C53.
Much of the concern in characterizing the surface structure
of
the
catalyst
is
in
understanding
y-AlgOg
the
interactions
between the metals and
the
support and how these
interactive
species
affect
selectivity.
Studies of the catalyst in the oxide form may
catalyst
activity
and
provide insight into those interactions.
Studies on the
have
identified
dimensional
oxide
numerous
compounds
cobalt, and/or
form
species
that
molybdenum
of
such
CoAl2O4 and Al2(MoO4 )3 C7- 123?
the catalyst to date
including:
include
as
(I)
aluminum.
CoMoO4 ,
three
oxygen,
Co3O4 , MoO3 ,
(2) two dimensional species
that involve Mo-y-A^Og and Co-Mo interactions E5, 133; and
(3) dispersed Co
and
Mo
ions
with coordination symmetry
dependent on that available on the y-AlgOg surface
[5, 11, 133.
The discrepancies in the conceptions of the structure
of Co-Mo/y-AlgOg catalysts probably result from their great
intrinsic structural complexity
which
may be sensitive to
preparation
from
the
parameters,
and
experimentally studying this catalyst system.
difficulty
of
Such studies
should be carried out in situ in order to avoid exposure of
the catalyst to the atmosphere [143.
8
Variables in catalyst preparation include calcination
conditions, percent metal loading
order of metal impregnation.
the y-AlgOg support
prior
the surface structure.
and time enhances
[113.
on the catalyst, and the
In addition, the condition of
to
metals impregnation affects
Increased calcination temperature
the
interaction
Percent metal
loading
has
observed species of both Co and
between Mo and y-AlgOg
a
marked affect on the
Mo
LI, 113.
The order of
metal impregnation has been shown to affect the interaction
of the
Co
and
the
Mo
phases
dehydration-dehydroxyIation
of
E8,
93.
the
The
y-AlgOg
state of
support also
affects the interactions of Co and Mo [73.
As with the oxide
exists as
to
sulfided form.
of Mo on the
the
It
form
exact
is
support
of the catalyst, controversy
chemical
makeup
of
the active
widely accepted that the dispersion
is
very
involves monolayer type coverage
been proposed to describe
high
[53.
and that it probably
Several models have
the structure of this monolayer,
but its characteristics are still in question:
(I)
The monolayer model assumes that the Mo monolayer in
the calcined catalyst remains essentially intact upon
sulfiding, except for replacement of some terminal
oxygen anions with sulfur anions [153.
The location
of Co is somewhat uncertain but it has been proposed
that it is located inside the y-AlgOg, thereby
stabilizing the monolayer.
9
(2)
The intercalation model assumes that Co intercalates
into the bulk of MoSg [16].
(3)
The contact synergy model supposes that the active
phases are present as separate phases of CogSg and
MoSg and the promoting effect of Co is suggested to
be the result of a contact between these phases Cl?].
(4).
More recently, the Co^Mo-S model has been proposed
which suggests that Co exists with Mo as a Co-Mo-S
surface phase.
This surface phase is suspected to
provide the necessary activity [14].
Since the active Co-Mo-S phase was proposed, numerous
studies have been performed which support this model
C18-21].
The activity
for
the
HDS of thiophene has been
shown to be directly related to the amount of Co located in
the Co-Mo-S phase [21].
The
presence of Co in this phase
appears to be enhanced
if
Mo
is
amount of Co contained
in
the
impregnated first.
Co-Mo-S
activity are both maximized when
The
phase and the HDS
the Co/Mo atomic ratio is
between 0.5 and 1.0 [21].
Limitations of the Literature and Experimental Methods
A variety of
aforementioned
spectroscopy
techniques
studies.
for
chemical
have
been
utilized in the
These
include
electron
analysis
(ESCA),
Mossbauer
10
emission spectroscopy
(MES),
Raman spectroscopy, infrared
spectroscopy (IR), ion-scattering spectroscopy (ISS), and
x-ray diffraction (XRD).
While
techniques
the
to
catalysts can
success
the
not
in
support
is
a
application
characterization
be
denied,
surface science techniques
system is not free
the
of
the
to
the
porous
these
Co-MoZ-Y-Al2O3
application
of these
study of this catalyst
significant
very
of
of
problems.
insulator.
The -Y-Al2O3
The
insulating
characteristics of the support cause electrostatic charging
of
the
sample
during
many
types
of
surface
studies.
Electrostatic charging
results
from
the sample surface due
to
acceptance and emission of
ions and/or electrons.
This charging problem can make data
the
analysis extremely difficult.
applications
of
electron
charge imbalances in
This may explain the lack of
spectroscopic
and
microscopic
techniques like Auger electron spectroscopy (AES), electron
energy loss
electron
spectroscopy
microscopy
(ELS),
(SEM
scanning or transmission
or
TEM)
and
ultraviolet
photoelectron spectroscopy (UPS) to this catalyst system.
Because of the porosity
no
well
defined
characterized by
[223.
the
surface
existing
of
these catalysts there is
which
can
be
precisely
surface science techniques
Techniques which employ ion beams, such as secondary
ion mass spectroscopy
(SIMS),
are
not beneficial because
11
the sputter beam size
of
existing
made reasonably smaller than
the
ion sources can not be
particle or pore size of
the catalyst.
The
difficulty
techniques to
of
porous
catalysts
adsorption and kinetic
level.
applying
studies
the
may
surface
explain
science
the lack of
performed on a microscopic
The adsorption of nitrous oxide gn the catalyst has
been studied with IR
E18,
19,
223.
However, the primary
purpose of these studies was to identify the phases present
on the catalyst
compounds are
surface.
more
Studies
rare.
adsorption was used in
[233. Again, however,
In
using actual reactant
one
such
study pyridine
conjunction with Raman spectroscopy
the
thrust
of
this
study was the
characterization of the catalyst surface.
Some of the more
powerful
ELS, and UPS, which may
be used for adsorption and kinetic
studies are difficult to
for
reasons
already
techniques may
techniques, such as SIMS,
utilize with the porous catalysts
stated.
provide
Utilization
information
atoms on which adsorption occurs
adsorption.
adsorb
on
activity.
For
the
example,
same
as
these
the surface
and the mode of molecular
nitrogen
surface
such
of
containing
sites
responsible
compounds
for HDS
Yet, it is known that hydrogen saturation of the
aromatic ring
occurs
whereas for HDS, C-S
before
bond
C-M
bond
scission for HDN;
scission occurs before hydrogen
12
saturation of the double bonds.
This suggests a different
adsorption phenomena for nitrogen containing compounds than
for sulfur compounds.
Kinetic
studies
basically involved
on
flow
Co-MoZY-Al2O3
systems
were determined via analyzing
reactor.
It
would
reactions as they
be
on
might best be accomplished
has been one such
the
which the mechanisms
by
to
actually study the
catalyst
surface.
This
the incorporation of static
programmed
study
have
the product streams from the
desirable
occur
SIMS with temperature
in
catalysts
which
desorption (TPD).
There
examined the desorption of
hydrogen and thiophene and mixtures of the two under vacuum
by TPD C24].
Incorporation
provide a means of
of
static
SIMS with TPD may
observing reaction intermediates on the
catalyst surface.
Planar Model Catalysts as an Alternative
An alternative
Al2O3 catalyst is to
approach
replace
Y-Al2O3 supports with a
planar aluminum.
thin
studying
the Co-MoZy-
the industrially used porous
surface oxide layer grown on
As discussed in the previous section, the
porosity and insulating
characteristics
catalysts pose difficulties in
the techniques suitable
the porous
for
Y-Al3O3
for
supports
the
of the industrial
application of many of
catalysis studies.
with
a
Replacing
well defined planar
13
surface should
eliminate
many
of
these difficulties and
provide new avenues by which adsorption and kinetic studies
of
the
catalyst
may
be
performed.
aluminum substrates also offers
the
samples
under
clean,
Utilizing
planar
the advantage of preparing
ultra
high
vacuum
(UHV)
conditions.
Growth of -Y-Al2O3
extensively
studied
in
different techniques is
on
bulk
aluminum metals has been
recent
years.
provided
by
A
Cocke,
summary
of
et al. [25].
Perhaps the simplest method is the direct thermal oxidation
of aluminum metal.
This can be accomplished by heating the
aluminum to approximately 863 K
in the presence of oxygen.
The thickness of the oxide layer may be varied depending on
the
time
allowed
for
oxidation.
application performed in this
desired to mitigate sample
For
the
type
of
study, thin oxide layers are
charging during the application
of the various surface science techniques.
The incentive
for
utilizing
planar
supports is to
provide a means of studying
the adsorption and kinetics of
hydrotreating reactions
on
Co-MoZy-Al2O3
microscopic level.
order
In
adsorption and kinetic
that
the
planar
representative of
catalyst.
A
to
studies
model
it
insure
catalysts
on a
the validity of
is necessary to verify
catalysts
are
structurally
the
surface
of
the
actual industrial
clearer
picture
of
the catalyst's surface
14
structure should
ambiguities
be
revealed
encountered
due
when
to
the elimination of
applying
surface
science
techniques to the porous catalysts.
Application of ESCA in Catalysis Research
Electron
spectroscopy
powerful technique in
for
catalysis
catalyst's surface chemistry.
irradiation
of
chemical
the
The
is a
research for studying the
The technique involves the
material
being
monoenergetic x-rays. Typically either
rays are used.
analysis
kinetic
Hg
studied
with
Koc of Al Ka x-
energies of these x-rays are
1253.6 eV and 1486.6 eV respectively.
The impinging x-rays cause
the emission of electrons
from the sample by the photoelectric effect.
These emitted
electrons have kinetic energies determined by?
KE = hv - BE - 4>s
where hv is the photon energy,
KE is the kinetic energy of
the emitted electrons,
the
electrons and (j>s is
BE
the
emitted electrons are
is
spectrometer
analyzed
Since the spectrometer work
are known, this allows
energy.
tens
Since only
of
Angstroms
energy of the
work function.
The
by their kinetic energies.
function and the photon energy
calculation of the electron binding
those electrons which originate within
of
sufficient energy to be
sensitive.
binding
the
sample
surface
escape
with
analyzed, the technique is surface
15
The binding energy of the electrons is characteristic
of
the
parent
elemental
atom
from
identification
which
is
atom.
originated.
possible
importantly, however, the binding
also influenced by the
it
chemical
with
ESCA.
Thus,
More
energy of an electron is
environment of its parent
Chemical environment may include the oxidation state
or coordination of the parent atom.
phenomena
provide
in
ESCA
known
information
on
as
the
Further details on the theory
This gives rise to the
chemical
catalyst
of
shifts
which may
surface chemistry.
ESCA may be found in the
literature [26-29].
Sample
spectra
for
aluminum,
oxidized, are provided in Figure
I.
the shift in binding
the
energy
of
both
metallic
These spectra display
Al 2p electrons upon
transition from metallic Al to crystalline Al2O3 .
width at half maximum
in
Figure
chemical
I,
may
environment
(FWHM)
also
of
The full
of the spectra, as displayed
provide
an
and
information
element.
For
about
the
example, an
increase in the FWHM of the spectra for a given element may
be an indication of
the
presence
of
more than one phase
containing that element.
The intensity of the
about the quantity of
metal
spectra may provide information
ions on the catalyst support.
The intensity is typically taken as the area under the
INTENSITY (arbitrary units)
16
Full Width at
Half Maximum
^/Intensity is
^ Area Under
Curve
_ _ _ _ _ _ _ L_ _ _ _ _ _ I_ _ _ _ _ _ _ I_ _ _ _ _ _ _ I_ _ _ _ _ _ I_ _ _ _ _ _ I_ _ _ _ _ _ I_ _ _ _ _ _ _ I- - - - - - - - - - 1- - - - - - - - - -
84
80
76
72
68
BINDING ENERGY (eV)
Figure I.
ESCA Al 2p Sample Spectra
(a) metallic; (b) oxidized
64
17
spectral curve.
This intensity
of a given spectrum may be
related to the atomic concentration
of that element by the
following equation [30];
V
s*
zlX zsX
where Ix is the intensity
factor for a
given
elements present.
and Sx is the atomic sensitivity
element.
The
Physical
summation is over all
Electronics Inc. has compiled
atomic sensitivity factors using silver as a standard C303.
The intensity ratios
atoms to spectra
of
of
the
the
Support
spectra
for the metal
may provide information
about the distribution of the
metal atoms.
smaller Co/Al intensity ratio
for one catalyst relative to
another with the same Co
Co migration into the
Co-containing
For example, a
loading may be indicating greater
support
agglomerates
lattice or the formation of
on
the
support
surface,
resulting in less dispersed coverage of the Co.
Application of AES in Catalysis Research
When an electron
inner shells of an
may
be
process.
filled
by
vacancy
atom,
either
The nonradiative,
frequently.
is
formed
in
one of the
such as during photoemission, it
a
or
radiative
or
nonradiative
Auger process, occurs most
v/
18
The Auger process
shell vacancy by
orbital.
This
an
involves
electron
the
from
readjustment
filling of an inner
a
causes
less tightly bound
the
ejection
of an
electron from the less tightly bound orbital with an energy
equal to the difference
and final states.
transition.
in
For
total
energies of the initial
example,
The initial state
consider
has
the KLL Auger
a hole in the K shell,
while the final state has two vacancies in the L shell.
Similar to ESCA, the
ejected
Auger electrons can be
analyzed by their kinetic
energies.
of
characteristic
these
atoms.
few
electrons
are
The kinetic energies
of
their parent
Again, only those electrons that originate within a
Angstroms
suffering
loss.
of
the
A
sample
thorough
utilized in, and the theory
surface
review
escape
of
the
without
equipment
of, AES is provided by Carlson
[263.
The Auger
process
impinging electron or
studied.
Its
may
photon
applicability
be
initiated
beam
in
on
by
the material being
catalysis research stems
from the fact that electron beams may be utilized.
x-rays, an electron beam
By
moving
the
finely
surface of the catalyst,
can
be
focused
either an
Unlike
focused on a small area.
electron
beam
about the
the spatial distribution of metal
atoms on the support may be
studied.
In addition, the use
19
of an electron beam
allows for high resolution photography
of the sample surface.
Auger
electron
spectroscopy
elemental identification, although
ESCA and AES are
efforts .
When
source in
ejected
employed
a
AES,
photon
shifts
Auger
this
beam
in
electrons
also
in
for
study where both
would be a duplication of
is
the
used as the excitation
kinetic
which
environment of the element may
a
provides
energies
depend
on
be observed.
the
Of the
chemical
The direction
of these chemical shifts parallel those observed in ESCA.
Application of SIMS in Catalysis Research
Surface
analysis
by
SIMS
bombarding the surface of the
high energy ions.
As
a
is
material
result
accomplished
by
being studied with
of these high energy ions
striking the sample surface, particles are removed from the
surface, some of
which
are
negative ions.
These
ions
in
the
are
form of positive and
passed
through
a mass
spectrometer where they are separated on the basis of their
mass-to-charge ratio and subsequently detected.
The technique is surface sensitive in that only those
ions which originate within the
surface
escape
the
solid
for
first few Angstroms of the
analysis.
However,
by
20
continuous bombardment of the
ions one digs deeper into
sample
with the high energy
the solid, thereby revealing the
composition beneath the surface.
Secondary ion mass spectrometry analyses are commonly
referred to
as
being
either
Operation in the static mode
of the sample surface.
static
(SSIMS) or dynamic.
provides for a lesser erosion
The erosion rate may be controlled
by varying the incident ion
beam current.
Further details
on the theory of SIMS may be found in the literature
[31, 323.
Secondary
ion
mass
spectrometry
technique in catalysis research.
to probe into the bulk
species
as
attainable.
a
function
of
into
the
solid
are
This procedure is known as depth profiling and
atoms and the support.
Since
on the support
surface.
valuable
Owing to its capability
depth
on
is
a
of the sample, the concentration of
may provide information
SSIMS
is
is
interactions between the metal
typically
preferred
to
the coverage of metal atoms
around a monolayer quantity,
avoid
rapid
destruction
of the
Electron spectroscopy for chemical analysis might
also be employed in depth
profiling studies to examine the
chemical environment of the metal atoms present beneath the
surface.
Similar to
the
electron
the ion beam can be focused
thus allowing examination
metal ions on the support.
beams employed in AES,
on a given area of the sample,
of
the
lateral distribution of
21
Similar to ESCA, SIMS may be used to characterize the
chemical structure of
a
catalyst
ions removed from the surface may
ions or clusters of ions.
unsupported
catalyst,
bulk
the
The secondary
be in the form of single
By comparing the SIMS spectra of
compounds
secondary
surface.
with
ion
SIMS
clusters
spectra
observed
of
the
may
be
identified as originating from a given chemical species.
A
review of these and other applications of SIMS in catalysis
research is provided by Brown and Vickerman [333.
: 22
RESEARCH OBJECTIVES
The objectives of this
research are essentially two­
Development
experimental
fold;
1.
of
the
preparation of the planar
model
catalysts.
procedures
for
This includes
preparation of the planar 1Y-Al2O3 by following the works of
other researchers and
developing
a suitable technique for
impregnating the planar supports with the Co and Mo.
2.
Characterization of the surface structure of the planar
model catalysts with ESCA,
findings
catalysts.
with
published
SIMS
and AES and comparing the
observations
on
the industrial
23
EXPERIMENTAL
Sample Handling System
The
samples
were
associated with the
consisted of
chamber.
a
ESCA
main
in
the
equipment.
analysis
These two chambers
Oxidation of the samples
chamber.
prepared
vacuum
system
This vacuum system
chamber
and a preparation
were sealed from one another.
was
performed in the preparation
Vapor deposition of
the
Co and Mo was performed
in the analysis chamber.
The sample handling rod,
for transport of the
shown
samples
in Figure 2, allowed
from the preparation chamber
to the analysis chamber without exposure to the atmosphere.
Titanium (Ti) foil was employed
7mm X 7mm X
Imm
aluminum
folding the Ti foil over
rod
contained
electrical
allowing for resistive
as the sample holder. The
coupons
the
were
edges
leads
of the aluminum.
to
heating.
held in place by
the
sample
The
holder
Thus, the sample holder
served as a heater for the aluminum.
A thermocouple attached
to
the
underside of the Ti
foil was used to monitor the sample temperature.
thermocouple actually measured
the sample itself, an
the
Since the
Ti temperature and not
infrared pyrometer was also employed
to monitor the surface temperature of the samples.
CONDUCTING WIRES TO
TITANIUM FOIL
- - - - - - THERMOCOUPLE LEADS TO
UNDERSIDE OF TITANIUM
FOIL
Figure 2.
Sample Handling Rod
25
Preparation of the Planar v-Al^Oj
Preparation of
the
planar
model catalysts followed
the works of Merrill C343
for the development of a y-AlgOg
surface layer.
placing
pure
Prior
to
pblycrystalline
polished to
remove
remove
residual
performed with
aluminum
surface
samples
were mechanically
imperfections
organics.
alumina
in vacuum, the 99.999%+
Mechanical
powders.
and cleaned to
polishing
was
Cleaning was performed
ultrasonically with acetone followed by methanol.
Once placed in the
and sputter cleaned with
or
subsurface
adsorbed
vacuum. the samples were annealed
ions to remove any surface
argon
oxygen
ESCA
powders on the surface.
any
and
survey
remaining AlgOg
scans were used to
verify completion of the in situ cleaning.
Oxidation
of
the
aluminum
exposing the samples to
a
for 0.5 hr at 823
The
K.
mixture
accomplished via
-4
of Og/HgO at 10
torr
Al 2p peak binding energy was
used to monitor the
oxidation
band displays
distinct
very
was
process.
shifts
Since the Al 2p
upon transition from
pure aluminum to amorphous AlgOg and finally to crystalline
AlgO3 ,
it
was
possible
crystalline AlgOg.
to
verify
the
presence
of
This was confirmed by comparison of the
various Al 2p band binding energies with values reported in
the literature [253.
The experimental conditions assured
that the crystalline AlgOg formed was y-AlgOg.
26
Thin
oxide
layers
were
desired
to
charging problems in the ensuing studies.
depth
of
secondary
electrons
approximately 20A, the presence of
avoid
Since the escape
analyzed
a
sample
by
ESCA
is
small Al 2p band for
pure aluminum provided a rough idea of the oxide thickness.
Deposition of Cobalt and Molybdenum
Deposition of
the
Co
and
Mo
on
the
-Y-Al2O3 was
attempted via two techniques.
The first technique was the
industrially
wetness
used
incipient
technique.
This
technique required removal of the oxidized samples from the
vacuum and immersing them in
an aqueous solution of cobalt
nitrate or ammonia hepta
molybdate.
several drawbacks.
was
solution
It
concentration
monolayer coverage of
difficult
required
the
This technique posed
metals
to
to
provide
on
determine the
the desired
the sample surface.
Some of the samples did not wet very well which resulted in
agglomeration of
the
Co
and/or
Mo
atoms
on the sample
surface as verified by Auger line analysis.
The
other
technique
evaporation of the Co and
was accomplished by
pure Co or Mo
the vacuum
Mo
passing
filament
chamber.
utilized
direct physical
on the sample surface.
electrical
positioned
This
was
This
current through a
close to the sample in
technique
proved superior to
incipient wetness in that more even surface coverage of the
27
metals was
obtained.
deposition was
In ■ addition,
done
in
situ,
the, physical vapor
whereas
incipient wetness
required removal of the samples from the vacuum.
Two planar model
the physical vapor
catalysts
deposition
were prepared, utilizing
technique.
The difference
between the two samples was the order of metals deposition.
As
previously
mentioned,
■coverage of both the.Co
desired.
and
Mo
approximately
monolayer
on the sample surface was
Two methods
were
employed to monitor the amount
of metals deposition.
The
first method utilized a quartz
crystal oscillator (QCO) positioned
the vacuum chamber.
sample and the
QCO,
close to the sample in
As the metals evaporated onto both the
the
The frequency drop of the
frequency
of the QCO decreased.
QCO was related to the thickness
of the deposited layer by the following equations [35];
d = CA/(pnWA )]1/3,
AF = dCp
where; d = thickness monolayer - cm
A = molecular weight of metal - g/mol
p = density of metal - g/cm3
n = number of atoms in a molecule
Wa = Avogadro's number
C = 56,497,175 Cm2Zgsec for 5 MHz quartz
crystal
AF = frequency drop of QCO
A sample calculation utilizing this equation is provided in
Appendix A.
28
The close proximity of the hot Co and Mo filaments to
the QCO tended to
Therefore, ESCA
Catalysis
increase
peak
its
areas
Research)
were
evaporation to provide an
surface concentrations.
frequency due to heating.
(see
also
Application
determined
approximation
of
of ESCA in
after
each
the Co and Mo
In order to implement this method,
the planar model catalysts
were
assumed to have a layered
type structure as shown below:
1
2
3
4
5
6
7
8
9
10
Al
This is an example of
•
Since
the
Mo deposited directly onto the
Al-O
1.91A, the average thickness
be 2.0A.
Considering
electrons
to
be
20A,
structure are 'seen' by
the
bond
distance is approximately
of
each layer was assumed to
escape
depth of the secondary
approximately
the
10
spectrometer.
layers
of
the
For the above
example, the ratio of the Mo 3d peak area to the sum of the
areas of the Al 2p,
0
Is
and Mo 3d peaks (corrected with
sensitivity factors) should equal 0.1 for a monolayer of Mo
atoms.
A sample calculation is provided in Appendix A.
29
This calculation typically showed
monolayer of
Co
or
Mo
monitored by the QCO.
had
For
lightly sputtered with
been
that more than one
deposited
while being
those cases, the samples were
Ar+
until
the
ESCA area ratios
indicated monolayer coverage of the metals.
Oxidation of Cobalt and Molybdenum
Following
deposition
metals were oxidized.
For
of
each
of
the
metals, the
example, on the Co -Mo Zy -AI2O^
(Mo deposited first) planar model, the Mo was deposited and
oxidized followed by
layer.
metals
The
are
deposition
oxidizing
shown
in
and
conditions
Table
I.
oxidation
for
of the Co
oxidation
These
conditions were
determined by trial and
error.
and the Co 2p or Mo
ESCA spectra was obtained.
3d
spectra indicated incomplete
of the
The samples were oxidized
oxidation
of
the
If the
Co or Mo,
this process was repeated with increased temperature and/,or
oxygen pressure.
Table I. Oxidizing Conditions for the Cobalt and
Molybdenum
Temperature O2ZH2O Pressure
Planar Model
Metal
(K)
(Torr) IO^
Co -Mo Zy -AI0O,
M o -Co Zy -AI0O,
Z Js
Mo
773-853
Co
Co
Mo
Time
(hr)
773-833
0.5
130.0
150.0
500.0
160.0
0.25
0.25
0.25
0.25
0.17
843-863
853
20.0
400.0
0.25
0.50
30
ESCA Analysis
The ESCA analyses of
performed
by
spectrometer
a
the planar model catalysts were
Leybold-Hereaeus
(L-H
EAlD.
(KE=12$3.6 eV) x-rays
photoelectron
Non-monochromatized
were
All ESCA data was obtained
x-ray
used
Mg
Kot
as the radiation source.
with the pressure in the vacuum
chamber at 10 ® torr or lower.
As described in the discussion on sample preparation,
ESCA data was collected at
each step in preparation of the
planar models.
Al
The Al 2p,
2s,
0
Is, Mo 3d and Co 2p
spectra were all recorded as applicable.
and Co LMM x-ray
as applicable.
pure Mo and
In
Auger spectra were also recorded
addition,
Mo
foils,
both
obtained for reference.
The
with an
Co
induced
analog
and
The 0 KLL, Mo MNN
3d
metallic
ESCA
digital
and Co 2p spectra of
and oxidized, were
data was recorded both
recorder.
The
latter was
performed by a Tektronix computer, and stored on a magnetic
disc.
The ESCA data was analyzed
computer program developed
by
Montana State University Physics
provided the capability to
with the aid of a special
XEROX
and
modified by the
Department.
smooth
This program
the data and obtain the
kinetic energy locations of
the
peak areas and FMHM of
spectra were obtained with the
the
peaks.
In addition, the
31
program.
The
program
shake up satellites
was
from
also
the
used
parent
to deconvolute the
peaks
in the Co 2p
spectra.
Auger Analysis
The Auger
line
analyses
Electronics (PHI 595) scanning
with the PHI 595
model.
was
only
Since the PHI
were
done
by
a Physical
Auger microprobe.
Analysis
performed on the Co-MoZ-y-A^O^
595
was associated with a different
vacuum system, the sample was briefly exposed to atmosphere
for transport to the PHI 595.
The line analyses
1000 X.
The pressure
or lower.
were
in
The electron
obtained
at a resolution of
-9
the vacuum chamber was 10
torr
beam
beam current was 0.2 pA.
voltage
was 5.0 KeV and the
This beam current provided for a
beam diameter of approximately 7000A.
The Auger data was
disc with the use
of
a
recorded
DEC
computer programs, developed
were used to smooth the
digitally on a magnetic
PDP 11/04 computer.
Special
by Physical Electronics Inc.,
data and account for topographical
effects on the sample surface.
SIMS Analysis
The
Hereaeus
SIMS
analyses
secondary
ion
were
mass
performed
spectrometer
by
a
Leybold-
(L-H
SSB/10)
32
employing a quadrapole mass filter.
only performed on
the
The SIMS analysis was
Co-Mo/y—AlgOg
model catalyst.
analysis was done in the PHI 595 vacuum system.
The
Argon ions
(Ar+ ) were used for the primary ion beam.
The depth profiles of
were obtained at
four
surface.
The primary
pressure
in
backfilling
the
with
the
elements
different
ion
argon
locations
energy
chamber
was
was
3.0
gas.
(Al, 0, Co, Mo)
2.8
X
The
on the sample
kV.
10
depth
The base
torr
before
profiles were
obtained with a noble gas pressure of 2.0 X 10 8 torr.
ion beam was rastered over
an
area
of 0.25 mm2.
The
The ion
2
beam current was approximately 20pA/cm .
Survey scans
were, also
bottom of the craters left
from
primary ion energy and
chamber
those employed for the
depth
rastered over a smaller
crater edges.
obtained
area
with
SIMS at the
the depth profiling.
The
pressures were the same as
profiles.
The ion beam was
to minimize effects from the
The survey scans covered 0-150 amu.
33
RESULTS AND DISCUSSION
Growth of v-AlnOj
The Al
2p
bands
in
Figure
energy shifts observed during
exposure to
10
-6
torr
O2
additional band appeared
indicating
the
display the binding
oxidation
and
at
formation
3
673
2.9
of
After additional exposure to
of the Al.
K
for
0.33
After
hr, an
eV higher binding energy
an
amorphous
IO-6
torr
oxide layer.
O2
and 823 K for
0.33 hr, the oxide band shifted 3.8 eV higher than that for
metallic Al, indicating the formation of crystalline Al3O3.
These values
are
consistent
with
Al
2p
binding energy
ranges reported in the literature C253.
C0-M0/V-AI2O3 Model Catalyst ESCA Analysis
Deposition and Oxidation
and Mo 3d
spectra
of
provided in Figures
the
4-6.
of
of
the
Al
2p
Numerical
deposition of the Mo.
higher binding
energy
line
The Al 2p
Co-MoZy-Al2O3 planar model are
binding energies are tabulated in
energy
Molybdenum:
values
of the peak
Appendix B.
The binding
decreased
0.6
eV following
The Mo 3d band is also at a slightly
compared
to
the
pure metallic Mo
band.
The FWHM of the spectra are provided in Table 2.
the Mo spectra, the FWHM is taken over the entire Mo 3d
For
INTENSITY
( arbitrary
units)
34
BINDING ENERGY (e V)
Figure 3.
Al 2p ESCA Spectra Displaying Oxidation
of Aluminum Foil
(a) cleaned aluminum foil; (b) after
exposure to 10 6 torr 0„, 673 K, 0.33 hr;
6
(c ) after additional exposure to 10
torr O2 , 873 K, 0.33 hr
+*
—
INTENSITY (arbitrary units)
35
BINDING ENERGY (eV)
Figure 4.
Al 2p ESCA Spectra for the
Co-MoZ^-Al2O3 Model Catalyst
(a) aluminum foil oxidized 10
torr O ZH2O, 823 K, 0.5 hr; (b)
after Mo deposition; (c) after
Mo oxidation; (d) after Co
deposition; (e) after Co
oxidation
36
INTENSITY (arbitrary units)
227,8
230,9
240
236
Figure 5.
232
228
224
BINDING ENERGY (eV)
220
Mo 3d ESCA Spectra for the
Co-MoZ-Y-Al-O- Model Catalyst
After MolyBdenum Deposition
and for Molybdenum Foil
(a) pure cleaned Mo foil;
(b) deposited Mo layer
INTENSITY (arbitrary units)
3 7
241
Figure 6.
237 233 229 225
BINDING ENERGY (eV)
221
Mo 3d ESCA Spectra for the
Co-Mo/f-Al-O? Model Catalyst
After MolyDdenum Oxidation
and for Oxidized Molybdenum Foil
(a) oxidized Mo layer; (b)
oxidized Mo layer after Co
deposition; (c) oxidized Mo layer
after Co oxidation; (d) oxidized
pure Mo foil
38
band.
Initially, the
amount
of
Mo deposited was greater
than a monolayer. The FWHM of
the Mo 3d band was 4.75 eV.
After sputtering to remove the
excess
Mo, the FWHM of the
Mo 3d band increased to 5.00 eV.
Table 2.
Peak Widths* (eV) for the Co-MoZy-Al7O1 Model
Catalyst
z J
_L Al 2p
IAl Oxidized I 2.19
IMo Deposited
2.21
IMo Oxidized I 2.40
ICo Deposited I 2.10
ICo Oxidized
2.42
IMo Metal
I
ICo Metal
ICo Metal
IOxidized
I 0 Is
I 3.08
I 2.65
I 3.06
I 2.47
I 2.94
—* —
—
—
—
—
Mo 3d1
Co ____
2p-,
o_L
J /,Z
—
5.00
6.50
7.00
5.63
4.75
—
—
—
—
—
—
—
—
4.38
4.81
—
—
—
2.50
3.75
—- —
—
I
I
I
I
I
I
I
I
* Full Widths at Half Maximum
1 Taken Over Entire Mo 3d Band
Mo
may
sharing the terminal oxygens
have the net effect
of
be
partially oxidized by
the
of
decreasing
O
The
W
Y-Al2O3.
interaction between the
NJ
Mo and
indicate
-k
1
>
H
These observations
This would
the oxidation state of
the Al, resulting in a shift to lower binding energy of the
Al 2p band.
The
0
Is
spectra
shown
in
supportive of an oxygen-sharing interaction.
Figure 7 are
The decreased
binding energy of the 0 Is band following deposition of the
Mo indicates
an
environment [36].
increase
in
the
valency
of the oxygen
39
BINDING ENERGY (eV)
Figure 7.
O Is ESCA Spectra for the
Co-MoZx-Al2O3 Model Catalyst
(a) oxidized Al foil; (b) after
Mo deposition; (c) after Mo
oxidation; (d) after Co deposition
(e) after Co oxidation
40
Oxidation of the Mo
shift of the Al 2p band
b&C).
However, the
of the band increased suggesting
between
the
intensity ratios of the ESCA
3.
not result in an additional
binding energy (see Fig. 4, curves
FHHM
enhanced interaction
did
Mo
and
Y-Al2O3.
The
spectra are provided in Table
The Mo/Al intensity ratio decreased following oxidation
of the Mo.
Assuming
spectra equally,
movement of Mo
this
into
the
additional oxygen affects both
observation
the
subsurface
may
be
attributed to
sites of the Y-a^20S
lattice due to counter diffusion of the Mo4n and Al+3 ions.
This also suggests increased interaction between the Mo and
Y-Al2O3 , although, the possibility of some MoO3 evaporation
can not be excluded since MoO3
evaporates at 973 K and one
atmosphere pressure.
Table 3.
Peak Areas and Area Ratios for the Co -Mo Zy -AI7O
Model Catalyst
I
I Normalized Peak Areas*
I
(IO3)
Al 2p IO Is IMo 3d ICo 2p
IAl Oxidized 134.7 142.2 I———— I————
IMo Deposited 116.5 120.2 15.53 I--IMo Oxidized 120.0 130.6 13.99 I————
ICo Deposited 116.4 121.9 12.98 16.24
ICo Oxidized 113.4 124.9 13.60 13.24
A
I
I
I
I
I
I
I
I
Peak Area |
Ratios
Mo/Al ICo/All
————
0.336
0.200
0.182 10.3821
0.259 10.2431
Area under the peak is divided by the sensitivity factors
(S) obtained from reference 30.
4.5
5Al 2p = O*11' 5O Is = °-63' 3Mo 3d = 1-2, 3Co 2p
These sensitivity factors were determined using a double
pass, cylindrical mirror analyzer.
It was assumed they
are accurate for the hemispherical analyzer used in this
study.
41
The Mo 3d band became quite broad with three distinct
peaks as shown in Figure
prior to
higher
oxidation,
binding
6.
the
Compared
entire
energy.
This
envelope
is
increased oxidation state of the Mo.
indicates the presence of more
the Mo.
to the Mo 3d band
is
expected
shifted to
due
to the
The shape of the band
than one oxidation state of
The Mo6+, Mo5+ and Mo4+ oxidation states of Mo may
all be present.
It is not clear
the
Mo
is
due
whether
to
the incomplete oxidation of
interaction
with
the
-Y-Al2O3.
As
evidenced by the changes in
the
intensity ratios, there
strong interaction between the
Mo and -Y-Al2O3.
is
Furthermore, it
band that a bulk phase
of
Al3O3 surface, as this
would
of the Mo 3d band
[37].
MoO3
Mo 3d
spectra
of
the
is obvious from the Mo 3d
is
not forming on the y-
result in a resolved doublet
Preliminary experiments indicated
difficulty in oxidizing Mo to Mo
This is evidenced by
Al 2p bands and the Mo/Al
under vacuum conditions.
multi-peak characteristic of the
oxidized
Mo
foil
shown
in Figure 6.
Preliminary experiments also indicated that oxidation of Mo
containing
samples
under
atmospheric
pressure
was more
complete.
Deposition
and
deposition of the Co,
shape but
the
entire
Oxidation
the
band
Mo
of
Cobalt;
Following
3d band maintained the same
shifted
to
a
lower binding
42
energy.
This
might
be
attributed
to an oxygen-sharing
interaction between the Co and Mo, resulting in a decreased
oxidation of the Mo.
s
The Co 2p spectra of the Co layer deposited on the
Co-Mo/y-AlgOg model catalyst and pure
Figure 8.
The
binding
higher and the FWHM
energy
is
Co foil are shown in
of
greater
the
Co 2p3/2 band is
(see Table 2) compared to
the Co foil.
The 0 Is band also shifted to a lower binding
energy
Fig.
(see
7,
curves
support an oxygen-sharing
c&d).
These observations
interaction
between
the Co and
Mo.
The Al 2p band shifted
to a lower binding energy and
narrowed upon Co deposition.
It is suspected that this is
due to a
shielding
effect
by
the
Co
layer.
A greater
fraction of the detected Al 2p electrons are from the Al-Mo
interactive
species.
Since
temperature during the
and Al as a result of
expected
to
be
the
evaporation,
Co
minimal
lattice
which
some
may
of
at
the
have
was
at
room
interaction of the Co
migration into the Al lattice is
this
previously stated, it appeared
the Mo layer,
sample
point.
However,
as
that following oxidation of
Mo
left
migrated into the y-AlgO^
some
exposed
y-AlgOg for
interaction with the Co.
After oxidation
of
the
Co
layer,
the
Mo 3d band
returned to its more characteristic doublet form with the
INTENSITY (arbitrary units)
43
792,9
778,5
BINDING ENERGY (eV)
Ficrure 8.
Co 2p ESCA Spectra for the
Co-Mo/y-Al-Og Model Catalyst
After CobaItjDeposition and
for Cobalt Foil
(a) pure cleaned Co foil;
(b) deposited Co layer
44
Mo 3d5/2 binding energy
equal
to
233.0
expected Mo 3d5/2 binding energy for
doublet is fairly
broad
and
not
eV.
This is the
Mo6+ [37].
The Mo 3d
well resolved, which is
attributed to the presence of more than one Mo6+-Containing
phase [38].
It is
difficult
with
ESCA
different Mo6+ compounds, such
their Mo 3d
same [53.
band
binding
Considering
to
as MoO3 and CoMoO4 , because
energies
the
distinguish between
are approximately the
difficulty in oxidizing the Mo
layer prior to Co deposition (see Table I), it is suspected
that the more complete oxidation
of the Co layer is
Mo.
due
An increase in
to
the
of the Mo after oxidation
interaction of the Co with the
Mo/Al intensity and a decrease in
the CoZAl intensity ratio was observed (see Table 3), which
indicates migration
layer.
of
This may also
some
Co
indicate
into
or
through the Mo
interaction of the Co with
the Mo.
The Co 2p spectra
of
oxidized
oxidized Co layer are shown in
pure Co foil and the
Figure 9. The Co 2p spectra
of oxidized Co contain several features useful in analysis.
The spin orbit splitting of
(AE) and the
shake
up
difference
satellite
at
the
between
higher
Co 2p3/2 and 2p1/2 levels
the
2p3^2 level and its
binding
energy
(AS) can
provide information on the oxidation state and coordination
of the Co [393.
INTENSITY (arbitrary units)
45
BINDING ENERGY (eV)
Figure 9.
Co 2p ESCA Spectra for the
Co-Mo/y-Al.O? Model Catalyst
After CobaItjOxidation and
for Oxidized Cobalt Foil
(a) oxidized Co layer;
(b) oxidized pure Co foil
46
For paramagnetic
been
found
to
be
cobaltous
(Co
approximately
diamagnetic cobaltic (Co3+)
approximately 15 eV C393.
been found to be 6.2 eV
2+
)
16
compounds
compounds AE has
eV,
and
whereas
for
Co metal AE is
The AS of the Co 2p3/2 level has
for CoMoO4 (Co octahedral) and 5.1
eV for CoAlgO4 (Co tetrahedral) [53.
It is also known that
Co3+ displays a much weaker satellite intensity relative to
Co3+ E393.
Therefore,
the
intensity
relative to the intensity of
the
of
the satellite
entire Co 2p3/2 band may
provide an indication of the Co oxidation state.
Analysis of the Co 2p
reveals AE to be 15.4 eV
this, the presence of
3+
Co ) is possible.
band
(see
both
for the oxidized Co foil
Fig. 9, curve b).
CoO
However,
(Co2+)
Based on
and Co3O4 (Co2+r
comparing
the
satellite
intensity of the oxidized Co foil to those in Co 2p spectra
of CoO and Co3O4 obtained by Stoch and Ungier [363 suggests
the presence of Co3O4 .
In
addition, the oxidized Co foil
appeared black in color which suggests Co3O4 .
The AS of
the
Co
2p3/2
model catalyst is 5.2 eV and
a).
These
values
indicate
tetrahedral Coordination,
Co3+ can not be discounted.
band
of the Co-MoZy-Al2O3
AE is 15.7 (see Fig. 9, curve
predominantly
although
the
Co
presence
ions in
of some
47
The absolute binding
energies
of
the Co 2p3/2 band
for Co compounds decrease in the order C o A ^ O 4)CoMoO4>Co304
C2, 53.
The Co 2p3^2 band binding energy of the
Co-MoZy-AlgO3 model catalyst
oxidized Co foil.
presence
of
This
either
might
predominantly present as
a CoAl2O4 type phase.
greater
could
be
CoAl2O4
Considering the AS, it
the SIMS depth
is
an
or
be
than that of the
indication of the
CoMoO4
surmised
type
phases.
that the Co is
tetrahedralIy coordinated Co
However,
profiles
(see
as
2+
in
the Mo 3d spectra and
later discussion) indicated
there was interaction of the Co with the Mo.
Further
substantiating
speculation
of
interaction
between the Co and Mo are the atomic surface concentrations
of the elements shown
interpreted as the
spectrometer.
Table
amount
of
Of particular
the Mo concentration
concentration
in
These values may be
each
element 'seen' by the
importance is the increase in
coupled
following
4.
with
a
oxidation
Assuming the additional oxygen
decrease
of
the
in the Co
Co
layer.
affected both equally, this
can only be explained by movement of Co into or through the
Mo layer.
The Al 2p band binding
to approximately the same
the Mo layer.
These
interaction between
energy and its FWHM increased
values as following oxidation of
observations
the
Co
and
may be an indication of
y-Al-O?.
While species
48
identification based on
these data may be
the
Al
indicating
2p
band is not possible,
a reduction of the shielding
effect caused by the deposited Co layer due to migration of
Co into the y-AlgO^.
intensity
ratio,
completely
Considering the decrease in the Co/Al
some
through
of
the
the
Mo
Co
layer,
may
have
allowing
for
migrated
direct
interaction with the Y-Al2O3.
Table 4.
Atomic Surface Concentrations for the
Co-MoZy-Al2O3 Model Catalyst
I
I
L Mo Deposited I
I Mo Oxidized
I
I Co Deposited I
I
I . Co Oxidized
I Al Oxidized
Al
0.45
0.39
0.37
0.34
0.30
* Calculated by C
0
I
I 0.55
I 0.48
I 0.56
I 0.46
I 0.55
Mo
Co I
I
I
I
I
I
I
I
I 0.13 I
I 0.07 - L
—
0.13
0.07
0.06
0.08
—
—
—
—- —
—
—
—
—
= -------11^ sX
Mo-CoZv-Al2O3 Model Catalyst ESCA Analysis
The ESCA
intensity
deposition of the Mo was
the signal pre-amplifier.
data
for
this sample following
hampered due to difficulties with
This explains the substantially
smaller peak areas following deposition
Table 5.
Because of
the
caused by this equipment
noise
This difficulty did
not
in the spectra, which was
malfunction,
peak areas following deposition
hinder
binding energies of the FWHM.
of the Mo shown in
the accuracy of the
of the Mo is questionable.
determination of the peak
49
Deposition
and
Oxidation
spectra of the Mo-CoZ^-Al2Og
10. Numerical
values
of
binding energy
again, is
peak
Al 2p
are shown in Figure
binding
energies are
The Al 2p band shifted to a lower
following
probably
The
Cobalt:
catalyst
the
tabulated in Appendix B.
of
deposition
due
to
the
of
Co
the
Co.
This,
interacting with the
surface oxygens of the iY-Al2Og.
The higher Co 2p3^2 band
binding
foil
energy
relative
binding energy of the 0
12, support an
to
Is
Co
band,
oxygen-sharing
and
the decreased
shown in Figures 11 and
interaction
of the Co with
the 1Y-Al2O3 .
Oxidation of the
Co
resulted
Co/Al intensity ratio as shown
could be attributable to
into the
Al3O3
two
lattice,
crystallites on
the
binding energy is
or
Al3O3
greater
of
Fig. 9, curve a).
The AS
is. 4.9
CoAl3O4 .
the
eV
Based on
migration of Co into
lattice.
Table 5.
(2)
the
surface.
that
which
This decrease
formation of Co3O4
The
Co 2p3/2 band
for the oxidized Co
In fact, it is higher than the
Co-MoZ-Y-Al2O3
This
a decrease of the
factors; (I) migration of Co
than
foil as shown in Figure 13.
binding energy
in
in
model catalyst (see
suggests the presence of CoAl3O4 .
also
supports
the presence of
these
observations, there was probably
the
tetrahedral sites of the -Y-Al3O3
However, the AE is only 15.5 eV suggesting the
INTENSITY (arbitrary units)
50
84
80
76
72
68
BINDING ENERGY (eV)
Figure 10.
64
Al 2p ESCA Spectra for the
Mo-CoZY-Al2O3 Model Catalyst
(a) aluminum foil oxidized 10
torr 0 /HO, 823 K, 0.5 hr;
(b) after^Co deposition;
(c ) after Co oxidation;
(d) after Mo deposition;
(e) after Mo oxidation
INTENSITY (arbitrary units)
51
BINDING ENERGY (eV)
Ficrure 11.
Co 2p ESCA Spectra for the
Mo -Co Zy -AI-O^ Model Catalyst
After Cobalt Deposition and
for Cobalt Foil
(a) deposited Co layer;
(b) cleaned pure Co foil
INTENSITY (arbitrary units)
52
542
538 534 530
526
BINDING ENERGY (eV)
Figure 12.
522
O Is ESCA Spectra for the
Mo-CoZ1Y-Al2O3 Model Catalyst
(a) oxidized Al foil; (b) after
Co deposition; (c) after Co
oxidation; (d) after Mo
deposition; (e) after Mo
oxidation
53
INTENSITY (arbitrary units)
782,9
BINDING ENERGY (eV)
Figure 13.
Co 2p ESCA Spectra for the Mo-CoZ-Y-Al7O1
Model Catalyst After Cobalt Oxidatiori
and for Oxidized Cobalt Foil
(a) oxidized pure Co foil; (b) oxidized
Co layer; (c) oxidized Co layer after
Mo deposition; (d) oxidized Co layer
after Mo oxidation
54
presence of some Co3+.
It appears that oxidation of the Co
resulted in the formation of both CoAl2O4 and Co2O4 .
The Al 2p band binding energy increased to just below
that for Y-Al2O3.
A
slight
occurred as shown in Table
enhanced interaction
6.
between
broadening
of the band also
The increased FWHM suggests
the
Co
and
Y-Al2O3.
The
increased binding energy of the Al 2p band may be explained
by migration of Co ions into the tetrahedral sites of the
Y-Al3O3 lattice, thus forcing more
octahedral sites.
greater when the
of the Al ions into the
The binding energy of Al 2p electrons is
Al
ions
are
in octahedral coordination
[403.
Table 5.
Peak Areas and Area Ratios for the Mo-CoZY-Al2O3
Model Catalyst
I Normalized Peak Areas*
J_________ (103)
IAl 2p I 0 Is IMo 3d ICo 2p
IAl Oxidized 130.5 I 39.4 I———— I————
15.87
ICo Deposited 116.1 I 19.4 I-- —
I--17.13
ICo Oxidized 135.6 I 51.1
IMo Deposited I 2.8 I 4.6 10.69 10.49
IMo Oxidized 188.3 1132.0 19.50 18.13
I
I
Peak Area |
I
Ratios
I
I MoZAl ICo/All
I--- I
I —
—
10.3651
I
-—
—
10.2001
I
I 0.246 10.1751
I 0.108 10.0921
Area under the peak is divided by the sensitivity factors
(S ) obtained from reference 30.
4.5
1.2, SCo 2p
0.63, SMo 3d
0.11, S0 Is
Al 2p
These sensitivity factors were determined using a double
pass, cylindrical mirror analyzer. It was assumed they
are accurate for the hemispherical analyzer used in this
study.
55
Deposition and Oxidation
of
Molybdenum:
The Mo 3d
spectrum, recorded following deposition of the Mo, is shown
in Figure 14.
The higher Mo 3d band binding energy and the
greater FWHM, relative
to
pure
oxidation of the deposited Mo
of the
Al
2p
and
Co
Mo foil, indicate partial
layer.
The binding energies
bands decreased, suggesting
^^2/2
oxygen-sharing interactions between
both
elements and the
Mo.
Table 6.
Peak Widths* (eV) for the Mo-CoZy-Al0O-, Model
Catalyst
^
J __________
I
I
1
—
1
—
Q
-------—
I Mo 3d1 I
I
I
I
I 5.50 I
I 5.00 I
I 4.75 I
I
CN
I O Is
I 2.69
I 2.58
I 2.69
I 3.25
I 2.61
C
U
I Al 2 d
IAl Oxidized I 2.36
ICo Deposited
2.33
ICo Oxidized I 2.49
IMo Deposited
2.63
IMo Oxidized I 2.39
IMo Metal
ICo Metal
ICo Metal
IOxidized
j/2— t
3.13
4.04
7.50
3.27
2.50
3.75
I
I
Full Widths at Half Maximum
Taken Over Entire Mo 3d Band
The
Co
2p
spectrum
particular, indicates a
and Mo.
(see
strong
This is evidenced
Fig.
due to interaction
c), in
interaction between the Co
of the Co 2p3/2 band.
new peak suggests substantial reduction
CoAl2O4
curve
by the appearance of a new peak
on the low binding energy side
Since the Co in
13,
This
of some of the Co.
is nonreducible, this is probably
between
the
Mo
and
the
Co that was
retained on the surface as Co3O4 following oxidation of the
INTENSITY (arbitrary units)
56
BINDING ENERGY (eV)
Figure 14.
Mo 3d Spectra for the M o -Co Zy -AI7OModel Catalyst After Molybdenum 2 3
Deposition and for Molybdenum Foil
(a) deposited Mo layer; (b) pure
cleaned Mo foil
57
Co layer.
The suggested
is plausible,
interaction between the Mo and Al
provided
that
resulted in penetration of
sites in the
Y-Al2O3
oxidation
some
lattice,
of
the
Co layer
of the Co into subsurface
thus leaving surface sites
available for interaction with the Mo.
Following oxidation of
the
band binding energy increased
Co oxidation and the
b&d).
FWHM
Mo
layer,
the Co 2p3y2
to slightly above that after
decreased
(see Fig. 13, curves
The AE increased to 15.9 eV while the AS remained at
4.9 eV.
There was
ratio and an
also
increase
a decrease in the Co/Al intensity
in
the
satellite intensity ratio.
These observations indicate an increase in the CoAl2O4 type
phase.
The Co that
lattice may have
had previously situated in the Y-Al2O3
diffused
leaving surface
sites
deeper
available
into
for
the lattice, thus
any
Co^
that was
retained on the surface prior to Mo oxidation [83.
Considering the
interaction of
the
Co
Co
2p
with
data,
the
it
Mo
appears
was
reduced at the
expense of an increase in the CoAl2O4 type phase.
as shown in
Figure
15,
the
compared to the Co-MoZY-Al2O3
curve a) prior to
Co
Mo
that the
However,
did oxidize more readily
model
catalyst (see Fig. 6,
deposition which may be attributable
to some interaction with the Co.
The direction of shifts
in
the x-ray induced O KLL,
Mo MNN and Co LMM Auger spectra paralleled the related ESCA
INTENSITY (arbitrary units)
58
BINDING ENERGY (eV)
Figure 15.
Mo 3d ESCA Spectra for the
Mo-Co/y-Al-O- Model Catalyst
After MolyDdenum Oxidation and
for Oxidized Molybdenum Foil
(a) oxidized Mo layer;
(b) oxidized pure Mo foil
59
binding energy shifts.
While
these spectra were omitted
from the discussion of the ESCA analysis, they are provided
in Appendix C.
Co-Mo/v-Al^Oj Model Catalyst SIMS Analysis
Depth profiles of the Al, 0, Co and Mo, obtained with
SIMS are
shown
in
Figure
16.
increased at the onset of
the
intensity steadily decreased
The
layer.
It is
not
contamination,
on
top
from
otherwise
observed in the Mo
the beginning.
suggest
of,
believed
ion intensity
analysis, while the Co
together, these two observations
Co was maintained
Mo+
or
an
Coupled
that some of the
together
that
ion
with, the Mo
this was due to surface
increase,
similar
to
that
ion intensity, would have been observed
in the Co+ ion intensity.
The Co+ ion intensity decreased more rapidly than the
Mo+ ion intensity.
some Co remaining
This
on
top
indicates
of
the
that, in addition to
Mo
layer, some Co was
dispersed in the Mo layer.
The retention of Co, either on
top of,
the
or
together
with
Mo
layer strengthens the
speculation of a Co-Mo interactive species.
A SIMS survey scan taken at
depth profiling craters is
indicates the presence of
the bottom of one of the
shown
both
in
Co
Figure 17.
The data
and Mo, which supports
the Co-AlgOg and Mo-AlgO3 interactions suggested by the
ION INTENSITY (c/s)
60
Figure 16.
SIMS Depth Profiles of the
Co-Mo/^-AlgOg Model Catalyst
ION INTENSITY (arbitrary units)
61
43 54 59 70
ATOMIC MASS UNITS
Figure 17.
SIMS Survey of the Co-Mo/^-Al_0_
Model Catalyst
^ ^
62
ESCA
results.
However,
ionization effects to
the
this
contribution
observation
can
of
matrix
not be ruled
out.
Co-Mo/-V-Al^O-J Model Catalyst AES Analysis
The AES line scans
for
provided in Figure 18.
to determine whether
The
or
the
Al,
0,
Co, and Mo are
purpose of the line scans was
not
there
was formation of bulk
phases, such as MoO^ or Co^O^, on the surface.
The line
coverage of
scans
both
indicate
the
Co
fairly homogeneous surface
and
Mo.
intensity, and the
corresponding
at 40 microns
due
was
to
decrease in Co
increase in Mo intensity
agglomeration
imbedded Al2O3 particle from sample
0 line scans also display
The
of Mo around an
polishing.
The Al and
an increase in intensity at this
same location.
The presence of
small
microcrystallites
can not be
ruled out based on these data.
Recalling the Co 2p spectra
of this sample,
of
the
Cd3O^, was indicated.
100 microns across the
presence
some
Co3+, possibly as
Since
the line analysis traversed
sample,
small crystallites may not
have been detectable.
An electron micrograph of the sample surface is shown
in Figure 19.
This micrograph, obtained at a magnification
of 1000X, is the same
obtained.
area
from which the line scans were
The surface of the sample appears fairly smooth.
6 3
MICRONS
Ficrure 18.
AES Line Scans of the
Co-Mo/"y-Al^Og Model Catalyst
(a) aluminum; (b) molybdenum;
(c ) oxygen; (d) cobalt
6 4
<— 5
4- 4
<— 3
4- 2
4—
Figure 19.
I
SEM Micrograph of the Co -Mo Zy -AI9O
Model Catalyst Surface
Note: This micrograph is from the same area
in which the AES line scans were
performed. The 5 dark lines are carbon
burns from the electron beam. Line #2
is where the data in Figure 18 was
obtained. The white patches are
imbedded alumina polishing powders.
65
The white patches in the micrograph are the Al^O^ polishing
powders imbedded in the
horizontally
through
surface.
The dark lines running
the micrograph
resulting from the AES line scans.
are
carbon
burns
The line labeled number
2 is where the data in Figure 18 was obtained.
Comparison of Planar Models to Industrial Catalysts
The literature on studies of the surface structure of
the calcined
industrial
Co-Mo/^-AlgOg
Depending on the preparation
techniques employed in
catalysts is vast.
parameters and the analytical
the
studies,
a variety of surface
species have been identified.
Since
study is to
determine
planar model catalysts are
structurally
similar
comparison
of
the
if
to
the
the
was performed
on
industrial
findings
literature is appropriate.
the
the purpose of this
of
this
study
Co-Mo/y-AlgO^
results
with
a
the
Since a more complete analyses
model
discussion is directed towards that sample.
applicable, the ESCA
catalysts,
for
catalyst, the
However, where
the Mo-Co/y-AlgOg model
catalyst are utilized.
Recalling the ESCA results of the Co-MoZ1Y-A^Og model
catalyst after oxidation of
the
Mo layer, it appears that
there is strong interaction between the Mo ions and the
Y-AlgOg. This interaction is suspected to be the result of
the Mo ions interacting with the terminal oxygens of the
Y-AlgOg.
It is widely believed on the industrial catalysts
I
66
that the Mo is highly dispersed over the -Y-Al2O3 support in
essentially a monolayer [6].
differ from
bulk
MoO3 .
Rather,
interact with the support
the Mo
in
tetrahedral
This monolayer is believed to
the
Mo
is thought to
forming a surface molybdate with
or
octahedral
surroundings [113.
While determination of the Mo coordination was not possible
with the techniques employed in
this
study,
the AES line
scans indicated the Mo was well dispersed on the surface as
in monolayer type coverage.
The Co in the
Co-MoZy-Al2O3 planar model appeared to
be distributed throughout the Mo
layer.
Based on the SIMS
depth profiling, the majority of the Co was retained on top
of, or together
with
the
between the Co and Mo.
the presence of Co
Mo layer suggesting interaction
The enhanced oxidation of the Mo in
and
the
ESCA intensity ratios further
indicated interaction between the
Co
and Mo, although the
SIMS surveys and ESCA results indicated some interaction of
the Co with the Y-Al2O3 .
The Co 2p3/2 band binding energy of the Co-MoZy-Al2O3
model catalyst is between that
of the oxidized Co foil and
the Mo-CoZy-Al2O3 model catalyst.
predominant
Co
species
for
tetrahedrally coordinated Co
the
predominant
Co
2+
species
the
Assigning Co3O^ as the
oxidized
Co
foil
and
, in a CoAl2O4 type phase, as
in
the
Mo-CoZy-Al2O3 model
67
catalyst, It
might
be
binding energy of the
assumed
that
Co-Mo/7-A ^ O g
the
Co
^ ^ / 2 band
model catalyst is the
result of a Co-Mo interactive species.
The assignment of the Co as CoAlgO^ in the
Mo-Co/y-AlgOg model catalyst would seem valid since, in the
absence of an underlying Mo
diffuse into the
-y-AlgOg
layer the Co might more freely
lattice
E83.
Contrary to this
speculation, studies on the industrial catalysts indicate a
greater tendency
for
directly to
Y-Al2O3
the
Co3O4
formation
supports
formation has been observed
with
as 2
extent
wt%.
However,
dependent on the mode
temperature.
catalyst
the
of
The
fact
indicates
most
Co
attributed to the
differences
planar models and
the
CS,
Co
123.
is added
This Co3O4
Co concentrations as low
of
Co3O4 formation is
addition and the calcination
that
of
when
the
the
Co
in
industrial
Mo-CoZy-Al2O3
is
CoAl2O4
model
may be
preparation between the
catalysts.
The Co 2p
spectra of the
Mo-CoZy-Al2O3 model catalyst did indicate
3+
the presence of some Co
prior to oxidation of the Mo
layer.
The suggestion of a Co-Mo interactive phase on the
Co-MoZy-Al2O3 model catalyst
CoMoO4 is actually forming on
is
the
indicated tetrahedral coordination
in CoMoO4 is octahedral).
not
meant
to imply that
model catalyst.
of
The AS
the Co^+ ions (Co^+
The majority of the Co
2+
ions in
68
the
Co-Mo/y-AlgOg
coordinated in a
ESCA data also
Co-Mo
catalyst
may
interactive
indicated
and the y-AlgOg,
tetrahedralIy
model
some
Co
tetrahedralIy
phase.
The SIMS and
interaction between the Co
This interaction
coordinated
be
2+
might be in the form of
situated
in
the y-AlgOg
lattice in a CoAlgO^ type phase. In addition, the presence
3+
of some Co
in the Co-Mo/y-AlgO^ planar model is not
discounted based on the
small crystallites of
ESCA
data.
Cd3O4
on
the
This
may be due to
surface, or from the
presence of Co^+ in the Co-Mo interactive phase.
Numerous investigators have suggested the presence of
Co-Mo
interactive
species
on
catalysts CS, 8, 13, 20, 413.
the bilayer is present
although the presence
not excluded.
as CoAl3O4 .
calcined industrial
Gajardo et. al. C133 reports
the presence of a Co-Mo bilayer
retained on top of the
the
in which some of the Co is
Mo
layer.
They suggest the Co in
as Co 3+ in octahedral surroundings,
of
octahedralIy coordinated Co^+ is
The remaining Co is dispersed in the Y-Al3O3
Okamoto et. al. C413 propose a similar bilayer
structure with some Co situated on top of the Mo layer.
The Co-Mo interactive
species
proposed
by Chin and
Hercules C53 is somewhat different than the bilayer models.
They propose the existence
with
the
Co
of
predominantly
a Co-Mo interaction species
present
as
tetrahedralIy
69
coordinated Co
2+
.
The key difference in their structure is
the tetrahedralIy coordinated
Co
2+
is primarily contained
in the Co-Mo interaction phase, not as CoAl2 .
It is not helpful to
speculate further, based on the
data of this study, on the
nature of the Co species in the
Co-MoZ1Y-Al2O3 model catalyst.
For example, determination
of whether the tetrahedralIy coordinated Co
CoAl2O4
or
as
a
Co-Mo
interactive
accomplished by reducing the
2+
is present as
species
might
be
model catalyst under hydrogen
and examining the Co 2p ESCA spectra.
Since CoAl2O4 is not
reducible, this may reveal the extent of CoAl2O4 formation.
This is, in fact, how Chin and Hercules determined the Co
2+
in their catalyst was tetrahedralIy coordinated in the
3+
Co-Mo interactive phase.
The presence of Co
as Co3O4
could be verified by x-ray diffraction 181.
/
SUMMARY AMD CONCLUSIONS
The objective
of
this
techniques for preparing
-Y-Al2O3 catalysts and
study
was
to establish the
planar model cobalt-molybdenum on
compare
the industrially used catalysts.
their
surface structure to
The salient findings of
this study are as follows:
1.
Growth of -Y-Al2O3 on planar aluminum was successfully
demonstrated via direct thermal oxidation.
2.
Physical vapor deposition of the cobalt and
molybdenum on the planar Y-Al2O3 was superior to the
industrially used incipient wetness technique for
these samples.
Physical vapor deposition provided
for more even surface coverage of the metals and
enabled complete in situ preparation of the planar
model catalysts.
3.
Similar to the calcined industrial catalysts, the
molybdenum deposited directly onto the planar Y-Al3O^
appeared to be well dispersed, as in monolayer
coverage, and interacting with the Y-Al2O3 .
4.
When cobalt was deposited on top of the molybdenum
layer, the cobalt appeared to be in tetrahedral
coordination and to interact primarily with the
molybdenum.
While more data is needed to
71
characterize the nature of the Co species, this
observation is similar to those made on the
calcined, industrial catalysts.
5.
While based on more limited data, there appeared to
be less interaction between the cobalt and molybdenum
when cobalt deposition preceded molybdenum.
In this
case, the Co appeared to exist primarily as CoAl2O4 .
7.2
RECOMMENDATIONS
Based on the experience gained from this experimental
work
and
the
literature
review,
the
following
recommendations are provided for future work:
1.
The STMS and AES analysis of the Co-Mo/y-AlgOg model
catalyst provided valuable information about the
metal dispersion and interaction between the metals.
These same analyses should be performed on the
Mo-CoZ-Y-AlgOg model catalyst to provide information
about the effect of metals impregnation order on the
surface structure of the model catalysts.
2.
While ESCA is a powerful technique for providing
information about the chemical environment of atoms
on the surface, future studies should incorporate a
wider variety of instruments and experimental
,
techniques in order to strengthen the findings.
Examining changes in the Co 2p ESCA spectra under a
’
reducing environment would be of great value in
cobalt species identification, since CoAlgO4 is
non reducible.
X-ray diffraction, Raman
spectroscopy and Mossbauer spectroscopy are among
the proven techniques for elucidating information
'
|
73
about the chemical environment of atoms on the
surface in studies on the industrial catalysts=.
3.
In conjunction with the above recommendation, pure
reference compounds should also be analyzed by the
same instruments to assist in identifying the
chemical species present on the surface.
74
REFERENCES CITED
75
REFERENCES CITED
1.
Foger, K. In "Catalysis"; Anderson, J .R .; Boudart, M
Eds.; Springer-Verlag: New York, 1984; Vol. 6,
Chap IV.
2.
Sahin, T . Ph.D. Thesis, Montana State University,
Bozeman, Montana, 1983.
3.
Satterfield, C.N. "Heterogeneous Catalysis in
Practice"; McGraw-Hill: New York, 1980.
4.
Sarode, P.R.; Sankar, G.; Srinivasan, A.; Vasudevan,
S ; Rao, C.N.R.; Thomas, J.M. Ancrew. Chem. Int. 1984,
4, 323.
5.
Chin, R.L.; Hercules, D.M. J. Phvs. Chem. 1982. 86,
3079.
6.
Massoth, F.E. "Advances in Catalysis", Vol. 27,
Academic Press, New York, 1978.
7.
Makovsky, L.E.; Stencel, J . M . ; Brown, F . R . ; Tischer,
R . E . ; Pollack, S .S . J. Cat a l . 1984, 89, 334.
8.
Chung, K .S . ; Massoth, F.E. J. Catal. 1980, 64, 320.
9.
Chung, K .S . ; Massoth, F.E. J. Catal. 1980, 64, 332.
10.
Lojacono, M.; Verbeek, J.L.; Schuit, G.C.A.
J. Catal. 1973, 29, 463.
11.
Zingg, D.S.; Makovsky, L.E.; Tischer, R.E.; Brown,
F . R . ; Hercules, D.M. j. P h y s . Chem. 1980, 8 4 , 2898.
12.
Medema, J.; VanStam, C.; DeB e e r , U.H.J.; Konings,
A.J.A.; Koningsberger, D.C. J. Catal. 1978, 53, 386.
13.
Gajardo, P.; Grange, P.; Delm o n , B. J. Catal.
63, 201.
1980.
76
14.
Topsoe, H.; Clausen, B.J.; Candia, R.; Wivel, C.;
Morup, S . J. Catal. 1981, 68, 433.
15.
Lipsch, J.M.J.G.; Schuit, G.C.A. J. Catal. 1969, 15,
179.
16.
Voorhoeve, R.J.H.; Stuiver, J.C.M. J. Catal. 1971, 23
243.
17.
Hagenbuch, G.; Courty, P.; Delmon, B. J. Catal. 1973,
31, 264.
18.
Topsoe, N.; Topsoe, H. J. Catal.
1982, 75, 354.
19.
Topsoe, N.; Topsoe, H .J. Catal.
1983, 84, 386.
20.
Wivel, C .; Clausen, B.J.; Candia, R .; Morup, S.;
Topsoe, H. J. Catal. 1984, 87, 497.
21.
Wivel, C.; Clausen, B.J.; Candia, R.; Morup, S.;
Topsoe, H. J. Catal, 1981, 68, 453.
22.
Topsoe, N.; Topsoe, H. J. Catal.
23.
Schrader, G.L.; Cheng, C.P. J. Phvs. Chem, 1983, 87,
3675.
24.
Yerofeyev, V.I.; Kaletchits, I.V. J. Catal. 1984, 86
55.
25.
Cocke, D.L.; Johnson, E.D.; Merrill, R.P.
Catal. Rev.-Sci. Encr. 1984, 26(2) , 163.
26.
Carlson, T.A. "Photoelectron and Auger Spectroscopy";
Plenum Press: New York, 1975; Chap. I-V.
27.
Fadley, C.S. In "Electron Spectroscopy; Theory,
Techniques and Applications"; Brundle, C.R.;
Baker, A.D., Eds.; Academic Press: New York, 1978;
Vol. 2, Chap. I.
28.
Fadley, C.S.; Hagstrom, S.B.M. In "X-ray
Spectroscopy"; Azaroff, L.V., Ed.; McGraw-Hill: New
York, 1974; Chap. VIII.
1982, 77, 293.
77
29.
Riggs, W.M.; Parker, M.J. In "Methods of Surface
Analysis"; Czanderna, A.W., Ed.; Elsevier Scientific:
New York, 1975; Vol. I, Chap. IV.
30.
"Handbook of X-ray Photoelectron Spectroscopy";
Physical Electronics Division, Perkin Elmer Corp.,
1979.
31.
McHugh, J.A. In "Methods of Surface Analysis";
Czanderna, A.W., Ed.; Elsevier Scientific:
New York, 1975; Vol. I, Chap. VI.
32.
Werner, H.W.; Morgan, A.E. In "Magnetic Field and
Related Methods of Analysis";
33.
Brown, A.; Vickerman, J.C. Surface and Interface
Analysis 1984, 6(1), I.
34.
Czanderna, K.K.; Morrissey, K.J.; PalmStrom, C.J.;
Carter, C.B.; Merrill, R.P., submitted to J . Mat.
Sci., July, 1984.
35.
Rocker, G., Ph.D. thesis, Montana State University,
Bozeman, Montana, 1982.
36.
Haber, J.; Stoch, J.; Ungier, L. J. Elec. Spec. Rel.
Phen. 1976, 9, 459.
37.
Patterson, T.A. J. Phvs♦ Chem. 1976, 80, 1700.
38.
Ratnasmy, P. J. Catal. 1975, 40, 137.
39.
Oku, M. J. Elec. Spec. Rel. Phenom. 1976, 8, 475.
40.
Nichollis, C.J.; Urch, D.S.; Kay, A.N.L. J .G .S . Chem.
Commun. 1972, 1198.
"\
Okamoto, Y.; Toshinobu, I.; Teranishe, J. J. Catal.
1980. 65, 448.
41.
78
APPENDICES
79
APPENDIX A
Sample Calculations
8 0
This is an example calculation
surface coverage of the cobalt and
quartz crystal oscillator (QCO).
for determining the
molybdenum via the
p 1® equation relating the thickness of a monolayer of
the metals to the frequency drop of the QCO are;
d = CA/(pnNA )31/3,
AF = dCp
where; d = thickness of monolayer - cm
A = molecular weight of metal - g/mol
p = density of metal - g/cm3
n = number of atoms in a molecule,
Na = Avogadro's number
C = 56,497,175 cm2/gsec for 5 MHz quartz
crystal
AF = frequency drop of QCO - Hz
For this example calculation molybdenum will be used.
A = 95.94 g/mol; p = 10.2 g/cm3 ; n = I
Na = 6.023 X IO23
d = [95.94/(10.2 X I X 6.023 X 1023)]1/3 = 2.5 X 10~8 cm
AF = 2.5 X IO"8 X 56,497,175 X 10.2 = 14.4 Hz
Therefore, when one monolayer of molybdenum atoms has
been deposited on the planar model catalyst, the frequency
of the QCO should have decreased 14.40 Hz.
81
This is an example calculation for determining the
surface coverage of the cobalt and molybdenum via the ESCA
peak areas. This example is for molybdenum deposited onto
the Y-Al2Og.
The equation
ESCA peak area is;
C
relating
"X.
3C
= -----x
EIx ZSx
atomic
concentration
to the
,.where; Ix = peak area
Sx = sensitivity factor
The planar model catalysts were assumed to have a
layered structure as shown below;
1
2
3
4
5
6
7
8
9
10
The following assumptions were made:
1. Average thickness of each layer equalled 2.0 A.
2. Escape depth of secondary electrons equalled 20 A
and therefore, 10 layers are 'seen' by the
spectrometer.
Considering these assumptions, C„
should equal 0.10 for a
monolayer of Md atoms on the Y-AT2Og surface.
82 ■
After deposition of the Mo on the Co -Mo Zy -AI7O- model
catalyst, the ESCA peak areas for each of the elements
present were;
Mo 3d - 6.64 X 103‘;
0 Is = 1.93 X IO4
Al 2p = 2.20 X IO3
The atomic sensitivity factors for these elements, as
provided
by
the
Handbook
of
X-ray
Photoelectron
Spectroscopy [30], are;
Mo 3d
"Mo
1.2; Al 2p — 0.11; 0 Is
0.63
6.64X10 /I.2
6.64X10 /1.2 + 2.20X10/0.11 + I.93X104/0„63
0.13
83
APPENDIX B
ESCA Peak Binding Energies
TABLE 7.
ESCA Peak Binding Energies (eV) for the Co-MoZ-Y-Al2 O 3 Model Catayst
Co 2p,I /Z
Mo 3d-, , Mo 3d,3 /,o
Z
.. J /L
SS1
CO_2p-j
J/Z
SS2
Al 2p
0 Is
Al Oxidized
76.07
532.28
Mo Deposited
75.56
531.65
231.07
Mo Oxidized
75.56
531.62
A
A
Co Deposited
75.11
531.06
A
A
Co Oxidized
75.40
531.23
236.17
233.00
797.41
230.97
227.84
—
—
—
—
-----—
—
—
—
—
777.90
—
780.65
787.00
—-
—
—
—
Co Metal
—
—
Co Metal
Oxidized
—
—
Mo Metal
Mo Metal
Oxidized
. ry
—
227.92
A
A
—
—
—
—
—
-----—
—
—
—
793.50
—
778.50
—
—
—
792.95
—
—
796.05
803.00
804.75
1 Shake up satellite on high binding energy side of Co 2p1/2 band
^ Shake up satellite on high binding energy side of Co 2p3/2 band
The Mo 3d bands displayed more than two peaks
781.79
786.98
ESCA Peak Binding Energies (eV) for the Mo-CoZ-Y-Al2 O 3 Model Catalyst
Mo 3dc
Mo 3d..
D /6
3/2
n
0
to
p
TABLE 8.
SS1
SS2
Co 2p J/Z
0 Is
Al Oxidized
76.19
532.48
—
—
Co Deposited
75.12
531.25
—-
--
Co Oxidized
76.02
532.15
—-
—
798.08
803.75
782.54
787.45
Mo Deposited
75.20
531.40
231.40
228.30
796.75
802.50
781.00
786.50
Mo Oxidized
74.70
532.18
237.50
234.60
798.75
804.25
782.86
787.80
230.97
227.84
—
—
—
—
—
—
—
———
—
777.90
—-
—
j
Al 2p
J-/z.
—
—
—
—-
793.50
—
778.40
—
Mo Metal
--
Mo Metal
Oxidized
—-
—
Co Metal
—
—
—
—
792.95
Co Metal
Oxidized
—
—
—
—
796.05
—
—
A
A
804.75
N -K
1 Shake up satellite on high binding energy side of Co 2p,/9 band
Shake up satellite on high binding energy side of Co 2p3/2 band
The Mo 3d bands displayed more than two peaks
780.65
787.00
86
APPENDIX C
X-ray Induced Auger Spectra
INTENSITY (arbitrary units)
8 7
756
752
748
744
740
736
BINDING ENERGY (eV)
Figure 20.
O KLL Auger Spectra for the
Co-MoZ^-Al2O3 Model Catalyst
(a) oxidized Al foil; (b) after
Mo deposition; (c) after Mo
oxidation; (d ) after Co deposition;
(e) after Co oxidation
INTENSITY (arbitrary units)
88
BINDING ENERGY (eV)
Figure 21.
Co LMM Auger Spectra for the
Co- Mo/y-Al-Og Model Catalyst
and Cobalt FGil
(a) pure cleaned Co foil; (b) Co
foil oxidized; (c) deposited Co
layer; (d) oxidized Co layer
INTENSITY (arbitrary units)
89
BINDING ENERGY (eV)
Ficrure 22.
Mo MNN Auger Spectra for the
Co-MoZ-y-Al^O^ Model Catalyst and
MolybdenumzFoil
(a) deposited Mo layer; (b) oxidized
Mo layer; (c ) oxidized Mo layer
after Co deposition; (d) oxidized Mo
layer after Co oxidation; (e) pure
cleaned Mo foil; (f) oxidized Mo
foil
INTENSITY (arbitrary units)
9 0
BINDING ENERGY (eV)
Ficrure 23.
0 KLL Auger Spectra for the
M o -Co /y -AI^O^ Model Catalyst
(a) oxidized Al foil; (b) after Co
deposition; (c ) after Co
oxidation; (d) after Mo oxidation
INTENSITY (arbitrary units)
91
BINDING ENERGY (eV)
Figure 24.
Co LMM Auger Spectra for the
Mo- CoZ-Y-Al7O- Model Catalyst
and Cobalt FOil
(a) pure cleaned Co foil; (b) Co
foil oxidized; (c) deposited Co
layer; (d) oxidized Co layer
MONTANA STATE UNIVERSITY LIBRARIES
stks N378.K637
RU
Surface chemistry model of cobalt-molybd
3 1762 00513611 2
N3 7 8
Kirk p a t r i c k , T i m o t h y D.
K637
Surface chemistry of
model cohalt-molybdenum
con.2
DATE
ISSUED TO
R378
K& 3 7
con .2