Preparation and surface science characterization of model Ziegler-Natta catalysts. Role of undercoordinated surface magnesium atoms in the chemisorption of TiCl4 on MgCl2 thin films

Article abstract of DOI:10.1021/jp982017y

Two new synthetic routes for the preparation of model Ziegler-Natta catalysts under UHV conditions are described. The exposure of metallic magnesium to TiCl₄ produced titanium chloride films with Ti in the 4+, 3+, 2+, and 0 oxidation states. Stable titanium chloride films could also be obtained by TiCl₄ and Mg codeposition on MgCl₂ and Au. A TiCl₄/TiCl₂ film was obtained in this case. The reaction of these systems with AlEt₃ produced model catalysts for the polymerization of both ethylene and propylene. XPS is a proper technique for the characterization of the oxidation state of Ti in a variety of titanium chloride surface species. The stability of MgCl₂ surfaces with a high concentration of undercoordinated Mg atoms was studied in UHV by Mg gas-phase deposition on a MgCl₂ multilayer film. The Mg adatoms were readily coordinated by the Cl ions diffusing from the halide bulk to the surface. Mg-containing MgCl₂ faces are thermodynamically unstable, and the fast diffusion of the Cl ions in the MgCl₂ matrix allows the recovery of the chlorine termination to lower the system surface energy. The high mobility of the chlorine ions is of central importance for the molecular level understanding of the dynamic equilibrium among the MgCl₂ surface, TiCl₄, and the electron donors used in the synthesis of high performance microporous Ziegler-Natta catalysts. The deposition of MgCl₂ in the presence of TiCl₄ was studied for the stabilization of high Miller index faces during the MgCl₂ film growth. The interaction between the two halides is too weak to influence the MgCl₂ deposition, and TiCl₄ could not be chemisorbed at 300 K.

Full text of DOI:10.1021/jp982017y

8788
J. Phys. Chem. B 1998, 102, 8788-8795
Preparation and Surface Science Characterization of Model Ziegler-Natta Catalysts. Role
of Undercoordinated Surface Magnesium Atoms in the Chemisorption of TiCl4 on MgCl2
Thin Films
Enrico Magni and Gabor A. Somorjai*
Department of Chemistry, UniVersity of California, Berkeley, California 94720-1460, and Materials Sciences
DiVision, Lawrence Berkeley National Laboratory, UniVersity of California, Berkeley, California 94720
ReceiVed: April 27, 1998; In Final Form: July 31, 1998
Two new synthetic routes for the preparation of model Ziegler-Natta catalysts under UHV conditions are
described. The exposure of metallic magnesium to TiCl4 produced titanium chloride films with Ti in the 4+,
3+, 2+, and 0 oxidation states. Stable titanium chloride films could also be obtained by TiCl4 and Mg
codeposition on MgCl2 and Au. A TiCl4/TiCl2 film was obtained in this case. The reaction of these systems
with AlEt3 produced model catalysts for the polymerization of both ethylene and propylene. XPS is a proper
technique for the characterization of the oxidation state of Ti in a variety of titanium chloride surface species.
The stability of MgCl2 surfaces with a high concentration of undercoordinated Mg atoms was studied in
UHV by Mg gas-phase deposition on a MgCl2 multilayer film. The Mg adatoms were readily coordinated
by the Cl ions diffusing from the halide bulk to the surface. Mg-containing MgCl2 faces are thermodynamically
unstable, and the fast diffusion of the Cl ions in the MgCl2 matrix allows the recovery of the chlorine termination
to lower the system surface energy. The high mobility of the chlorine ions is of central importance for the
molecular level understanding of the dynamic equilibrium among the MgCl2 surface, TiCl4, and the electron
donors used in the synthesis of high performance microporous Ziegler-Natta catalysts. The deposition of
MgCl2 in the presence of TiCl4 was studied for the stabilization of high Miller index faces during the MgCl2
film growth. The interaction between the two halides is too weak to influence the MgCl2 deposition, and
TiCl4 could not be chemisorbed at 300 K.
1. Introduction
surface. This reaction produced a titanium chloride film in
which the Ti was in the 0, 2+, 3+, and 4+ oxidation states.
The exposure of this system to AlEt3 produced an active catalyst
for the polymerization of both ethylene and propylene. Active
model Ziegler-Natta catalysts were also obtained by Mg and
TiCl4 codeposition on either MgCl2 or Au, followed by reduction
and alkylation with AlEt3. During the codeposition experiment,
a redox reaction between the deposited Mg atoms and the TiCl4
molecules impinging from the gas phase allowed for the
deposition of a titanium chloride film. This film was very
similar in composition to the one obtained by electron irradiation
induced chemical vapor deposition of TiCl4 on the same
substrates.5-8
Ziegler-Natta catalysts are used for the polymerization of
R-olefins such as ethylene and propylene. The precursor of the
active site is obtained by the chemisorption of TiCl4 on highly
disordered MgCl2. The titanium chloride is then reduced and
alkylated by the reaction with an aluminum alkyl to produce
the active center. It has been suggested that, during the
preparation of the microporous catalyst using conventional
synthetic techniques, the TiCl4 is chemisorbed on the under-
coordinated Mg atoms exposed at the lateral cuts of the MgCl2
crystallites. In particular, the tetra- and pentacoordinated Mg
atoms present at the (110) and (100) faces of MgCl2 are thought
to be the binding sites for TiCl4.1-4
The MgCl2 films grown under UHV conditions expose only
Cl atoms at the solid-vacuum interface. Gas-phase TiCl4
molecules do not chemically interact with these films.9-12 The
stability of MgCl2 surfaces containing undercoordinated Mg sites
was studied at room temperature by the gas-phase deposition
of Mg atoms on the halide surface. ISS and XPS analyses
indicated that the deposited Mg atoms were rapidly coordinated
by the Cl atoms diffusing to the surface from the subsurface
layers of the MgCl2 film. The concentration of Mg atoms at
the solid-vacuum interface was below the detection limit after
the deposition of 1016 Mg atoms/cm2.
Finally, we studied the possibility of increasing the sticking
probability of TiCl4 to MgCl2 by simultaneous exposure of the
magnesium chloride films to TiCl4 and MgCl2 molecular beams.
In this experiment, our goal was the preferential growth of faces
of MgCl2 other than the basal plane. Very little TiCl4 was
deposited under these conditions.
In this paper, we report the results of our detailed study of
the interaction between the TiCl4 molecules and the surface of
MgCl2 films prepared under controlled conditions. We specif-
ically focused our attention on the role of the surface Mg atoms.
Model Ziegler-Natta catalysts were prepared under ultra-high-
vacuum (UHV) conditions by gas-phase deposition of MgCl2
and TiCl4 on a gold foil. The surface concentration of Mg atoms
was controlled by gas-phase deposition. The titanium chloride
films thus obtained were reduced and alkylated by reaction with
AlEt3 to form the active sites. The model surfaces were
characterized by X-ray photoelectron spectroscopy (XPS), ion
scattering spectroscopy (ISS), and polymerization activity tests.
We studied the deposition of TiCl4 on metallic magnesium
and monitored by XPS the reduction of the halide at the metal
* To whom correspondence should be addressed at the Department of
Chemistry. Fax: +1 510 643 9668. E-mail: somorjai@socrates.berkeley.edu.
10.1021/jp982017y CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/13/1998

Model Ziegler-Natta Catalysts
J. Phys. Chem. B, Vol. 102, No. 44, 1998 8789
TABLE 1: Parameter Values for the Synthetic Curves Used
in the Description of the Experimental Ti 2p XPS Peaksa
2. Experimental Section
The UHV apparatus and experimental procedure used in the
present study have been described in detail elsewhere.5,6 Briefly,
the apparatus consisted of a preparation chamber, maintained
at a background pressure of 1 × 10-9 Torr by an ion pump,
coupled with a PHI 5400 Small Spot ESCA analysis system
and a reaction cell, where gases and liquids were introduced
near atmospheric pressure. The sample under study was
transferred from one section of the apparatus to the others
without exposure to air or any other contaminant through a series
of gate valves and transfer arms. The preparation chamber was
equipped with a Knudsen cell that produced a beam of MgCl2
molecules used for the growth of clean films of controlled
thickness. A second Knudsen cell allowed for the sublimation
of metallic magnesium and the production of a Mg atomic beam.
The temperature of both sources was adjusted in order to obtain
the desired flux. A leak valve connected to a directed doser
allowed for the introduction of a controlled pressure of TiCl4
during the deposition of the titanium chloride film. The analysis
chamber was provided with a non-monochromatic X-ray source,
a hemispherical electron energy analyzer having a mean radius
of 28 cm, and a differentially pumped ion gun. The electron
analyzer was operated at fixed energy resolution using a pass
energy of 8.9 or 71.5 eV as specified for each experiment
reported. The reaction cell was directly connected to the source
of AlEt3 which was introduced either in the gas or in the liquid
phase. After the reaction with AlEt3, the cell was evacuated
by a liquid-nitrogen sorption pump and a turbo molecular pump
before transferring the sample to the UHV sections of the
apparatus.
TiCl4
TiCl3
TiCl2
Ti
pass energy
(eV)
p3/2 p1/2 p3/2 p1/2 p3/2 p1/2 p3/2 p1/2
TL
TS
G
14 18 14 18 13.5 18 14 18
0.6 0.6 0.6 0.6 0.55 0.5 0.5 0.5
8.9
75 75 75 75 75
75 75 75
1.7 1.7 1.7
fwhm 1.7 1.7 1.7 1.7 1.7
71.5
G
85 85
85
1.9
85
1.9
fwhm 1.9 1.9
a G ) percent of Gaussian Contribution; fwhm ) full width at half-
maximum; TL ) tail length; TS ) tail scale.
convolution of an exponential function and a Gaussian peak.14
The shape of each synthetic peak is fully identified by the
percent Gaussian contribution (G), the full width at half-
maximum (fwhm), the tail length (TL), and the tail scale (TS).
The values of the shape parameters that characterize the eight
peaks used for the curve fitting of our Ti 2p experimental signals
are reported in Table 1. In all cases, p1/2 peaks have areas
exactly half of those of the respective p3/2 peaks.
The chemical composition of the uppermost monolayer of
the deposited films was analyzed by ISS, using He ions of 700
eV incident energy with a flux of 5 × 1012 ions/(cm2 s). To
minimize the damage to the surface under study, the incident
ion beam was rastered on the sample surface and the data
acquisition time limited to 3 min. These conditions allowed
us to obtain spectra with good signal-to-noise ratios and
negligible damage of the sample surface.
3. Results
The elemental composition and the oxidation state of the
atoms in the deposited films were monitored by XPS, using Al
KR excitation radiation (1486.6 eV) and detecting photoelectrons
at a takeoff angle of 54°. The Au 4f7/2 peak at 84.0 eV was
taken as a reference for the energy scale. The XPS peak areas
are reported after correction by the corresponding sensitivity
factors (Au 4f7/2, 3.5; Ti 2p, 2.0; Mg KLL, 2.4; Cl 2p, 0.89; C
1s, 0.30; O 1s, 0.71; Al 2s, 0.29)13 and normalization with
respect to the Au 4f7/2 peak area of the clean substrate set equal
to 100. The atomic compositions were calculated from the
integrated peak areas of XPS spectra collected at an electron
pass energy of 71.5 eV, corrected by the corresponding
sensitivity factors. We used the approximation that the surface
composition was homogeneous over the analysis volume. From
repeated experiments, we estimated the following relative
random errors in the calculation of the atomic concentrations:
Ti, 5%; Mg, 1%; Cl, 2%; Al, C, and O, 20%.
3.1. Preparation of Model Ziegler-Natta Catalysts: TiCl4
Deposition on Metallic Mg. A detailed analysis of the surface
reaction between TiCl4 and Mg is not reported in the literature.
We studied the reduction of TiCl4 on the Mg surface by gas
phase deposition of the halide on metallic magnesium. Oxygen-
free Mg surfaces can readily be obtained by sublimation of the
metal and deposition from the gas phase on a gold foil. We
prepared clean Mg films using a Knudsen cell with the Mg
source at 650 K and maintaining the Au substrate at 300 K. A
thick magnesium overlayer was deposited under these conditions
after 60 min. The lower curves in Figure 1, parts a and b,
respectively show the Mg KLL and the Mg 2p regions of the
XPS spectrum of the Mg film. They are characteristic of
metallic Mg, with the most pronounced of the Mg KLL peaks
at 1185.5 eV and the Mg 2p peak at 50.0 eV, in agreement
with the literature values.7,10,13-19
For each of the XPS spectra in the Ti 2p region, an attempt
has been made to fit the experimental curve with a series of
synthetic peaks that represent the contribution of the photo-
electron emission from atoms in different chemical environ-
ments. Care has been taken to fit the experimental curves to
the minimum number of peaks. These peaks are described as
a mixture of Gaussian and Lorentzian contributions in order to
take into consideration the effect of the instrumental error on
the peak shape characteristic of the photoemission process. A
certain degree of asymmetry is introduced to the peak shape to
account for the tail present at the high binding energy side of
the peak due to inelastic scattering of the photoelectrons during
their transport to the sample surface and to the possible presence
of shake-up features. The use of asymmetric curves allows us
to assume a linearly sloping background. The tail is described
by a non-negative term added to the high binding energy side
of the mixed Gaussian and Lorentzian peak. This term is the
The exposure of this Mg film to TiCl4 at a pressure between
1 × 10-7 and 1 × 10-6 Torr induced a fast redox reaction at
the film surface. The Ti4+ was reduced to the 3+, 2+, and 0
oxidation states and the Mg was oxidized to MgCl2. Parts a-c,
respectively, of Figure 1 show the Mg KLL, Mg 2p, and Ti 2p
regions of the XPS spectra recorded after the exposure of 10,
100, and 1000 L of TiCl4. (The langmuir (L) is a practical
unit of exposure defined as 1 × 10-6 Torr‚s.) Table 2
summarizes the atomic compositions at the film surface after
the different exposures. A small amount of O was accumulated
at the film surface due to the high reactivity of Mg with oxygen-
containing molecules present in the vacuum system during the
TiCl4 exposure. The surface reaction between Mg and TiCl4
generated a new peak in the Mg KLL and in the Mg 2p regions
of the XPS spectra, with maxima at 1178.9 and 53.0 eV. These
peaks became stronger for larger TiCl4 exposures and are due
to the formation of MgCl2.7,10,20,21

8790 J. Phys. Chem. B, Vol. 102, No. 44, 1998
Magni and Somorjai
Figure 1. Mg film deposition on polycrystalline Au followed by the exposure to TiCl4 and the reaction with AlEt3. Mg KLL, Mg 2p, and Ti 2p
regions of the XPS spectra are shown. Mg source temperature: 650 K. TiCl4 pressure: from 1 × 10-7 to 1 × 10-6 Torr. AlEt3 exposure: liquid
layer for 10 min. Substrate temperature: 300 K. (a) Mg KLL peak. Detector pass energy: 8.9 eV. (b) Mg 2p peak. Detector pass energy: 71.5 eV.
(c) Ti 2p region of the spectra and curve fitting. Detector pass energy: 8.9 eV.
TABLE 2: XPS Surface Atomic Compositions of the Mg
Film after Exposure to TiCl4 and AlEt3
of unreduced TiCl4 after the AlEt3 exposure can be estimated
by eliminating the asymmetric component in the synthetic peak
corresponding to TiCl3 and introducing the 2p3/2 and 2p1/2
components of TiCl4 in the fitting. In so doing, we calculated
that TiCl4 accounts for less than 11% of the total Ti present
after the reduction-alkylation reaction. The C uptake is
consistent with the formation of products of partial alkylation
of the titanium chloride. TiCl2Et was probably present, but the
corresponding Ti 2p3/2 peak at 456.2 eV overlaps with the peak
due to TiCl2. The appearance of the spin-orbit doublet at 455.2
and 461.2 eV is indicative of the formation of TiClEtn, where
n can be 1 and/or 2. All these features are consistent with our
previous work in which we extensively monitored by XPS the
surface reaction of a TiCl4/TiCl2 thin film with gas- and liquid-
phase AlEt3 at 300 K.7
The TiCl4/TiCl3/TiCl2 film prepared by exposure of the Mg
surface to TiCl4 became an active catalyst for the Ziegler-Natta
polymerization of R-olefins after the reduction-alkylation
reaction with the aluminum alkyl. We tested the activity of
this model catalyst by running gas-phase polymerization reac-
tions in the presence of 760 Torr of monomer at 300 K. Both
ethylene and propylene were polymerized.
3.2. Preparation of Model Ziegler-Natta Catalysts: TiCl4
and Mg Codeposition on MgCl2 and Au. Titanium chloride
films were prepared by TiCl4 and Mg codeposition on MgCl2.
In this experiment, we deposited 10 layers of MgCl2 on the Au
substrate and we exposed the film, held at 300 K, to simulta-
neous beams of TiCl4 and Mg. The halide was introduced in
the chamber at 2 × 10-7 Torr; this corresponds to a flux of 3
× 1013 TiCl4 molecules/(cm2 s). The flux of the magnesium
beam was measured, as described in section 3.3, to be 3 × 1012
Mg atoms/(cm2 s) at the target sample. Under these experi-
mental conditions, about 10 TiCl4 molecules hit the MgCl2
surface per every Mg atom that adsorbed on it (assuming unitary
sticking probability for Mg on MgCl2 at room temperature).
This large excess of TiCl4 enhanced the probability for each
Mg adatom to react with the titanium chloride. At the same
time, the gas-phase collision probability between one Mg atom
and one TiCl4 molecule was of the order of 10-4, which suggests
a maximum deposition rate for products of gas-phase reaction
of the order of 108 molecules/(cm2 s).
atomic composition (%)
model system
Mg film
Ti
Mg
100
Cl
C
O
Al
TiCl4 exposure
10 L
100 L
1000 L
AlEt3 exposure
liquid film/10 min
10
16
16
46
25
20
41
59
62
3
2
5
11
16
40
23
5
The oxidation of the Mg atoms was coupled with the
reduction of the titanium in the TiCl4 molecules sticking to the
film surface. After the exposure of 10 L of TiCl4, the surface
atomic concentration of Ti was already 10%. The Ti 2p region
of the XPS spectrum is dominated by a peak at 457.4 eV which,
together with the peak at 463.3 eV, forms the spin-orbit doublet
characteristic of Ti atoms in TiCl3.22 A second intense doublet
is present at 454.4 and 460.5 eV, due to the 2p3/2 and 2p1/2
photoelectrons from zerovalent Ti atoms, eventually alloyed with
the Mg film.6 A third weak spin-orbit doublet is present, with
maxima at 456.2 and 462.2 eV, due to the presence of a small
amount of TiCl2.5-7 The Cl to Ti atomic ratio (Table 2) was
4:1, indicating that every Cl atom liberated from the TiCl4
reduction stayed on the Mg surface forming Mg-Cl bonds. At
larger TiCl4 exposures, the TiCl2 2p3/2 and 2p1/2 peaks became
stronger and two new peaks appeared at 458.6 and 464.7 eV.
This is the spin-orbit doublet characteristic of TiCl4.5-7,22,23
At the same time, the Ti zerovalent peaks gradually disappeared.
After the exposure of 1000 L of TiCl4, the Ti atoms accounted
for 16% of the film surface atomic composition; Ti4+, Ti3+
,
and Ti2+ were present in roughly the same amount. Incidentally,
we underline the possibility to distinguish by XPS among Ti
atoms simultaneously present at the sample surface in the 4+,
3+, 2+, and 0 oxidation states. This is of central importance
for the characterization of the oxidation state of the titanium
present in the stereoselective active site for the Ziegler-Natta
polymerization.
The exposure of this titanium chloride film to a liquid layer
of AlEt3 at 300 K for 10 min produced the extensive reduction
of the TiCl4 to TiCl3 and its partial alkylation products. This
is suggested by the shape of the Ti 2p region of the XPS
spectrum (Figure 1c). Our curve fitting indicates that the 458.6
and 464.7 eV components disappeared and the lower oxidation
state peaks grew more intense. The upper limit of the amount
Figure 2 shows the Mg KLL, Cl 2p, and Ti 2p regions of the
XPS spectra acquired at different stages of the TiCl4 and Mg
codeposition. The Mg KLL and the Cl 2p peaks, initially at
1178.1 and 201.1 eV, gradually shifted to 1179.2 and 200.1 eV

Model Ziegler-Natta Catalysts
J. Phys. Chem. B, Vol. 102, No. 44, 1998 8791
Figure 4. TiCl4 and Mg codeposition on polycrystalline Au. Mg KLL
and Ti 2p regions of the XPS spectra at different TiCl4 exposures are
shown. TiCl4 pressure: 2 × 10-7 Torr. Mg flux: 3 × 1012 atoms/(cm2
s). TiCl4 to Mg molecular flux ratio at the target surface: 10:1. Substrate
temperature: 300 K. (a) Mg KLL peak. Detector pass energy: 71.5
eV. (b) Ti 2p peak. Detector pass energy: 8.9 eV.
Figure 2. TiCl4 and Mg codeposition on 10 layers of MgCl2 film. Mg
KLL, Cl 2p, and Ti 2p regions of the XPS spectra at different TiCl4
exposures are shown. TiCl4 pressure: 2 × 10-7 Torr. Mg flux: 3 ×
1012 atoms/(cm2 s). TiCl4 to Mg molecular flux ratio at the target
surface: 10:1. Substrate temperature: 300 K. Detector pass energy:
8.9 eV for the Mg KLL and the Ti 2p peaks; 71.5 eV for the Cl 2p
peak.
To test the catalytic properties of the TiCl4/TiCl2/MgCl2 film
prepared by TiCl4 and Mg codeposition, this model system was
exposed to a liquid layer of AlEt3 at 360 K for 10 min in the
reaction cell. Ethylene was introduced at 760 Torr and
polymerized on the catalyst surface at 360 K.
Titanium chloride films of similar quality were grown by
TiCl4 and Mg codeposition on polycrystalline Au. We exposed
a clean Au foil held at 300 K to a beam of 3 × 1012 Mg atoms/
(cm2 s) in a background pressure of 2 × 10-7 Torr of TiCl4.
Figure 4 shows the Mg KLL and Ti 2p XPS peaks after the
exposure of 60 and 480 L of TiCl4. The Mg peak with a
maximum around 1179.8 eV is probably due to the presence of
a chlorine defective MgCl2 or to the formation of a mixed
titanium/magnesium chloride (TiRMg1-RCl2).5,6 The actual peak
position shifted during the deposition experiment toward the
value characteristic for MgCl2. The Ti 2p region shows the
presence of Ti atoms in two oxidation states. The Ti 2p3/2 peaks
at 458.7 and at 456.4 eV are assigned to TiCl4 and to TiCl2,
respectively.
Figure 3. TiCl4 and Mg codeposition on 10 layers of MgCl2 film
monitored by XPS as a function of the deposition time. Mg flux: 3 ×
1012 atoms/(cm2 s). Substrate temperature: 300 K. Solid curves: 10:1
TiCl4 to Mg molecular flux ratio at the target surface. Dotted curves:
100:1 TiCl4 to Mg molecular flux ratio at the target surface.
3.3. Mg Deposition on MgCl2 Multilayer Films. Instabil-
ity of Undercoordinated Mg Surface Atoms. It has been
proposed that, during the preparation of the microporous catalyst,
TiCl4 chemisorbs on the tetra- and pentacoordinated Mg atoms
exposed at the (110) and (100) faces of the MgCl2 crystallites.1-4
To test this hypothesis, we attempted the preparation of MgCl2
films with a high concentration of undercoordinated Mg atoms
at the solid-vacuum interface by exposing a MgCl2 thin film
to a beam of Mg atoms.
after the exposure of 360 L of TiCl4. This 1 eV shift of the
peak maxima can be described as surface charging of the
analyzed sample. This charging effect is more intense on the
MgCl2 film than after the TiCl4 and Mg codeposition. The Ti
2p region shows the presence of two spin-orbit doublets.
Again, the presence of some surface charging induced the shift
of the peaks toward lower binding energies as the TiCl4 and
Mg exposures increased. The Ti 2p3/2 peaks with maxima at
459.2 and 456.8 eV after 24 L of TiCl4 shifted to 458.7 and
456.3 eV after 360 L of exposure. These two peaks are
respectively assigned to TiCl4 and TiCl2, in agreement with our
previous work.5,6 The same shift in the energy scale is evident
in the Mg and Cl peaks.
The atomic beam was produced by a Knudsen cell in which
a magnesium rod was maintained at 570 K. The flux of Mg
atoms was measured by the rate of deposition of magnesium
on a clean gold foil held at 300 K. The change of surface
composition was monitored by both XPS and ISS. Figure 5
shows the attenuation of the Au 4f7/2 peak and the increasing
area of the Mg KLL peak during this experiment. The peak
areas have been corrected by the respective sensitivity factors.
The shape of these curves does not reveal the presence of any
evident break, making difficult the determination of the time
corresponding to the deposition of one monolayer of magnesium.
The ISS uptake analysis showed that the Mg overlayer grew
either by the formation of three-dimensional islands or by a
Stranski-Krastanov mechanism, as perhaps suggested by the
XPS experiment. The ISS spectra recorded after the deposition
of different amounts of Mg are reported in Figure 6a and the
Au and Mg peak areas are shown in Figure 6b, as a function of
In Figure 3, the Ti, Mg, and Cl uptakes corresponding to the
above-described experiment are reported as a function of the
deposition time. In addition, the uptake curves measured during
the exposure of the 10 layers of MgCl2 film to a flux of 3 ×
1012 Mg atoms/(cm2 s) and 3 × 1014 TiCl4 molecules/(cm2 s)
are shown. The two sets of curves are not significantly distinct,
indicating that in both cases the rate of the film growth was
determined by the Mg flux, which was the limiting reagent. A
substantial amount of titanium chloride was deposited after 30
min. The surface atomic composition was 12% Ti, 18% Mg,
and 70% Cl. In all cases, the deposited films were clean with
no detectable traces of oxygen contamination.

8792 J. Phys. Chem. B, Vol. 102, No. 44, 1998
Magni and Somorjai
Figure 7. Mg deposition on 10 layers of MgCl2 film monitored by
XPS and ISS as a function of the deposition time. Mg flux: 3 × 1012
atoms/(cm2 s). Substrate temperature: 300 K. (a) Mg KLL XPS peak.
The inserts in the high side of the energy scale are multiplied by 10.
Detector pass energy: 71.5 eV. (b) ISS spectra.
Figure 5. Mg deposition on a polycrystalline Au foil monitored by
XPS as a function of the deposition time. Mg source temperature: 570
K. Substrate temperature: 300 K.
the flux of Mg atoms at the target surface from the deposition
rate, the condensation coefficient of Mg on the Au surface
should be determined. However, it is generally accepted that
this coefficient for metals on gold at 300 K is unity.26 Under
this assumption, the Mg flux at the target surface coincides with
the estimated deposition rate.
The knowledge of the flux was necessary for the precise
dosage of the Mg atoms on the magnesium chloride surface, as
described below. In the analysis of the results of this last
experiment, once again we assume a unity sticking coefficient
for Mg on MgCl2 at 300 K. It has been shown that the
condensation coefficients for Pt, Pd, and Ir deposited on NaCl
between 573 and 673 K are approximately 0.4.27 The strong
dependence of the condensation coefficient on the substrate
temperature has been studied in detail in the case of Pd
deposition on MgO.28 Coefficients close to unity have been
measured for this system when the substrate was at 430 K. The
correct determination of the sticking probability of Mg on the
MgCl2 film will be the subject of a future investigation.
The lower curves in Figure 7, parts a and b, respectively show
the Mg KLL XPS peak and the ISS spectrum corresponding to
10 layers of MgCl2. The Mg KLL peak has its maximum at
1178.4 eV, as expected for MgCl2. The ISS spectrum shows
that the film was Cl terminated, in agreement with our previous
results.9-12 We then exposed this film, held at 300 K, to the
Mg atomic beam for 5, 10, 20, and 69 min. Neither the XPS
nor the ISS spectra reported in Figure 7 showed any significant
modification of the halide surface composition after the deposi-
tion of about 4 × 1015 atoms/cm2. Only after the deposition of
1 × 1016 atoms/cm2, corresponding to about 10 layers of Mg,
a weak peak appeared in the XPS spectrum at 1186.6 eV due
to the presence of metallic magnesium. The Mg peak became
evident also in the ISS spectrum. The area of the ISS peak
was only 11% of the area of the Mg peak measured after the
deposition of Mg on Au for 30 min under the same experimental
conditions (see Figure 6a). It is worth mentioning that if the
actual Mg sticking coefficient on MgCl2 would be only 0.1,
the equivalent of about one layer of Mg should be deposited at
the halide surface. If this amount of Mg would stay at the
solid-vacuum interface, it would be easily detected by both
XPS and ISS.
Figure 6. Mg deposition on a polycrystalline Au foil monitored by
ISS as a function of the deposition time. Mg source temperature: 570
K. Substrate temperature: 300 K. (a) ISS spectra and (b) Mg and Au
peak areas.
the deposition time. After 6 min, the Au peak was about 20%
of the peak corresponding to the clean Au surface. At the same
time, the Mg peak accounted for 90% of the peak measured
after 30 min of deposition.
From the intensity of the Mg XPS peak acquired after the
deposition of a thick magnesium film (see section 3.1), we could
calculate the intensity of the Mg peak due to one monolayer of
Mg atoms at the solid-vacuum interface using the well-known
equation24
i ) I[1 - exp(-t/(λ sin æ))]
Here, i is the signal intensity due to one monolayer at the
interface, I is the signal intensity corresponding to an infinitely
thick material, t is the thickness of one monolayer, λ is the
inelastic mean free path of the detected electrons, and æ is the
takeoff angle. Considering the thickness of one monolayer to
be 2.6 Å, half of the c parameter of the hexagonal close-packed
unit cell of Mg, and assuming the inelastic mean free path for
the 1185.5 eV Mg KLL Auger electrons to be 30 Å,25 we
estimated that the Mg KLL peak due to one monolayer of
magnesium deposited on the gold substrate should have a
normalized area of approximately 13. This value is represented
by the dotted line in Figure 5 and was reached after 5.2 min of
deposition.
From the XPS and the ISS analysis we conclude that
approximately one monolayer of Mg was deposited in 6 min.
Since one monolayer of close-packed Mg atoms has a density
of 1.1 × 1015 surface atoms/cm2, we estimated a Mg deposition
rate of approximately 3 × 1012 atoms/(cm2 s). To calculate
The addition of Mg atoms from the gas phase did not produce
a MgCl2 film with a high concentration of surface Mg atoms.
Instead, two phases were formed, MgCl2 and metallic Mg. This
can be deduced from the analysis of the Mg KLL peak (Figure
7a). The formation of a highly defective, Cl-poor MgCl2-R
phase as a consequence of the deposition of Mg on the MgCl2

Model Ziegler-Natta Catalysts
J. Phys. Chem. B, Vol. 102, No. 44, 1998 8793
film can be ruled out because of the lack of a broad peak
between 1178.4 and 1186.6 eV. The peak at 1186.6 eV,
detected after long Mg exposure, is consistent with the formation
of a metallic phase immiscible with MgCl2. The weak intensity
of both this peak and the Mg peak in the ISS spectrum (Figure
7b) suggests that the metallic phase must be either embedded
in the MgCl2 film or concentrated at the Au surface.
We conclude that Mg-terminated MgCl2 surfaces are unstable
under UHV conditions. The high surface energy of this system
is the driving force for the diffusion of the Mg adatoms from
the surface into the bulk or, more probably, for the diffusion of
the subsurface Cl ions to the surface of the halide, to recover
the thermodynamically less energetic chlorine termination.
According to this second picture, the bulk MgCl2 that loses its
Cl ions generates the subsurface Mg phase.
Figure 8. TiCl4 and MgCl2 codeposition on 10 layers of MgCl2 film.
TiCl4 pressure: 5 × 10-7 Torr. MgCl2 flux: 8 × 1011 molecules/(cm2
s). TiCl4 to MgCl2 molecular flux ratio at the target surface: 100:1.
Substrate temperature: 300 K. (a) Ti 2p region of the XPS spectra
after the exposure of 150 and 450 L of TiCl4. Detector pass energy:
71.5 eV. (b) Ti 2p, Mg KLL, and Cl 2p XPS peak areas as a function
of the deposition time.
3.4. TiCl4 and MgCl2 Codeposition on Multilayer MgCl2
Films. We have previously reported that the MgCl2 films
produced by gas-phase deposition on gold foils grow via a layer-
by-layer mechanism.9,10 The lateral facets of the two-
dimensional islands growing at the solid-vacuum interface
could, in principle, present a high density of undercoordinated
Mg atoms. In fact, we could imagine these Cl-Mg-Cl one-
layer facets as segments of MgCl2 faces like, for example, (100)
or (110). If these lateral facets were bulk-terminated, we would
expect a strong interaction between a Lewis base and the surface
Mg atoms. This would cause either the preferential growth of
faces other than the basal plane or the inclusion of a substantial
amount of the Lewis base in the MgCl2 film during its growth.
To limit the lateral growth of the MgCl2 two-dimensional
islands and to favor the expansion of higher Miller index MgCl2
faces, we deposited MgCl2 in the presence of TiCl4 in the
vacuum chamber background pressure. A film of 10 layers of
MgCl2 previously deposited on a gold foil held at 300 K was
simultaneously exposed to TiCl4 and MgCl2. The MgCl2
molecular flux was such that one monolayer of the halide was
deposited in 18 min under UHV conditions. From the deposi-
tion rate, we estimated a MgCl2 molecular flux of 8 × 1011
molecules/(cm2 s) at the target surface. The TiCl4 pressure
during the codeposition experiment was 5 × 10-7 Torr,
corresponding to a flux of 7 × 1013 molecules/(cm2 s). Under
these experimental conditions, approximately 100 molecules of
TiCl4 hit the magnesium chloride film per each MgCl2 molecule.
Considering that a directed doser was used to bring the gas inlet
right in front of the target sample, the actual TiCl4 pressure at
the MgCl2 film surface could have been higher than the value
measured at the ion gauge. Parts a and b, respectively, of Figure
8 show the Ti 2p region of the XPS spectrum and the Ti 2p,
Mg KLL, and Cl 2p uptake curves. The chemical composition
of the deposited film remains substantially unchanged during
the experiment. Even after the exposure of 450 L of TiCl4, the
Ti 2p signal is just above the noise level, indicating a Ti atomic
composition of about 0.2%. This corresponds to a sticking
probability for TiCl4 on MgCl2 of the order of 10-5, to be
compared to a value close to unity for MgCl2 onto itself at room
temperature.
thermodynamically unstable at the magnesium chloride surface
under UHV conditions around room temperature. The high
surface energy related to the presence of undercoordinated Mg
atoms at the interface is the driving force for the surface
reconstruction noticed during the deposition of Mg atoms on
MgCl2 multilayer films. The diffusion of Cl ions from the
MgCl2 bulk to the surface of the film and their interaction with
the Mg atoms deposited from the gas phase are sufficiently fast
at 300 K to yield an equilibrated system in the time scale of
our experiment. However, the Mg adatoms can be stabilized
at the surface of the MgCl2 film when deposited in a background
pressure of TiCl4. We can picture the MgCl2 film with a small
concentration of undercoordinated surface Mg adatoms as a
transient state. TiCl4 molecules chemisorb on this surface by
reacting with these Mg atoms. The redox surface reaction
between the TiCl4 and the Mg adatoms is faster than the
diffusion of the Cl ions from the film bulk. A titanium chloride
film composed of TiCl4 and TiCl2 can be grown under these
conditions.
The results of our experiments are important to understand,
at the molecular level, the role of the MgCl2 as a support for
TiCl4. The chlorine ions have a high mobility in the MgCl2
structure and can rapidly diffuse to the surface to correct any
defect, such as an undercoordinated Mg surface atom. We
would then expect the (100) and (110) faces of MgCl2 to be
reconstructed under UHV conditions and expose only Cl atoms
at the interface. The long-range order implied by the suggested
chemisorption of TiCl4 on the (100) and (110) faces at the lateral
cuts of the MgCl2 crystallites would in fact not be present. The
sites of TiCl4 chemisorption should then be considered as point
defects on the MgCl2 surface. The mobility of the Cl atoms is
responsible for the interaction between MgCl2 and different
Lewis bases, including TiCl4. In the presence of a Lewis base,
the surface Cl atoms can migrate and allow for the formation
of stable surface compounds with the reactive Mg atoms. The
dynamic nature of the chemistry at the MgCl2 surface is
fundamental in the delicate equilibrium between the chemi-
sorption of TiCl4 and the different electron donors used for the
synthesis of high-performance catalysts.
4. Discussion
Our experiments show in a direct way that Mg atoms are
responsible for the chemisorption of TiCl4 at the MgCl2 surface.
Even though the actual sticking coefficient of Mg on MgCl2 at
300 K has not been determined, this should be close to unity.
Under this assumption, we have measured by ISS a Mg surface
concentration of about 10% after the deposition of the equivalent
of 10 layers of Mg on the halide film. Mg atoms are
Titanium chloride films can be grown under UHV conditions
by gas-phase codeposition of TiCl4 and Mg on either MgCl2 or
gold around 300 K. Table 3 gives the normalized areas of the
Ti 2p, Mg KLL, and Cl 2p peaks measured after different
exposures during the TiCl4 and Mg codeposition on the two
substrates, together with the surface atomic composition of the
deposited overlayer. The deposition of the titanium chloride

8794 J. Phys. Chem. B, Vol. 102, No. 44, 1998
Magni and Somorjai
TABLE 3: TiCl4 and Mg Codeposition on 10 Layers of
MgCl2 and on Polycrystalline Aua
TABLE 4: XPS Surface Atomic Compositions of the
Titanium Chloride Films Deposited by (a) TiCl4 Exposure to
Mg, (b) TiCl4 Electron-Induced Deposition on MgCl2, and
normalized
peak area
atomic
composition (%)
a
TiCl4
exposure
(L)
(c) TiCl4 and Mg Codeposition on MgCl2
atomic composition (%)
substrate
Au foil
Ti 2p Mg KLL Cl 2p Ti Mg Cl
deposition method
Ti
Mg
Cl
O
60
480
60
13.6
26.0
9.5
8.8
20.3
44.4
28.7
53.3 18
108.6 17
12
13
26
18
70
70
68
70
(a) TiCl4 exposure to Mg film
(b) TiCl4 electron-induced deposition
on MgCl2
16
27
20
0
62
73
2
0
MgCl2, 10 layers
115.8
6
360
19.7
110.7 12
(c) TiCl4 and Mg codeposition on MgCl2 12
18
70
0
a Normalized XPS peak areas and overlayer atomic compositions.
TiCl4 flux: 3 × 1014 molecules/(cm2 s). Mg flux: 3 × 1012 Mg atoms/
(cm2 s). Substrate temperature: 300 K.
a Deposition conditions are as for Figure 9.
film was slightly faster on Au than on MgCl2. Again, this can
be explained by the diffusion of the Cl ions from the film bulk
to the surface when depositing on MgCl2. In this case, some
of the Mg deposited on the film was not available for the redox
reaction with TiCl4 to form the titanium chloride film. However,
every Mg atom deposited on the Au surface was readily
available for the reaction with the TiCl4 to produce the TiCl4/
TiCl2/MgCl2 film.
The result of the TiCl4 and MgCl2 codeposition on multilayer
MgCl2 films can be interpreted with similar arguments. The
high surface energy of MgCl2 films with undercoordinated Mg
atoms at the solid-vacuum interface can be lowered by diffusion
of Cl atoms to the surface and the full coordination of the
outermost Mg atoms. The lateral facets of the two-dimensional
islands do not show the bulk termination structure, and TiCl4
cannot chemisorb on them. The result of this experiment is
substantially different from that of the TiCl4 and Mg codepo-
sition experiment, where Mg atoms were deposited from the
gas phase on the MgCl2 surface. When the MgCl2 molecules
impinge on the halide film, they probably can diffuse as MgCl2
entities on the chloride surface, until they find a two-dimensional
island to which they adhere. The Mg atom is never exposed to
the interface with a geometry favorable for TiCl4 to react with
it.
We have described in the present work the synthesis of
titanium chloride films by TiCl4 and Mg codeposition on MgCl2
and by the exposure of metallic magnesium to TiCl4. The
preparation and characterization of titanium chloride films by
electron irradiation induced chemical vapor deposition of TiCl4
on Au and on MgCl2 have been extensively described.5-8 In
all cases, a surface Mg atom appears to be the site for the TiCl4
chemisorption. The successive reaction with an aluminum alkyl
produces the active site for the Ziegler-Natta polymerization.
The oxidation states of the Ti atoms in the films deposited with
these procedures were measured by XPS. The titanium chloride
film obtained by gas-phase deposition of TiCl4 on Mg showed
a quite complicated stoichiometry, with the titanium atoms in
the oxidation states 4+, 3+, 2+, and 0. The relative amount
of titanium in the different oxidation states changed with the
TiCl4 exposure. The detailed analysis of the film prepared by
electron irradiation induced chemical vapor deposition of TiCl4
on MgCl2 thin films revealed the presence of a few layers of
TiCl2 with one monolayer of TiCl4 chemisorbed at the surface.6
The simultaneous deposition of TiCl4 and Mg on MgCl2
produced a film in which the titanium atoms were in the
oxidation states 4+ and 2+. Table 4 gives the surface atomic
composition of films prepared with the three different methods,
and Figure 9 reproduces the Ti 2p region of the corresponding
XPS spectra. Because of the different acquisition conditions,
the intensities of these spectra cannot be directly compared. The
similarity of the spectra corresponding to the films obtained by
TiCl4 and Mg codeposition and by electron irradiation induced
Figure 9. Ti 2p region of the XPS spectra of titanium chloride films
deposited at 300 K by three different methods. The curve fitting of the
experimental spectra is shown. (a) Exposure of a Mg film to 1000 L
of TiCl4. TiCl4 pressure: 1 × 10-6 Torr. Detector pass energy: 8.9
eV. (b) Electron irradiation induced TiCl4 deposition on 10 layers of
MgCl2 film. TiCl4 exposure: 240 L. TiCl4 pressure: 2 × 10-7 Torr.
Electron flux: 3 × 1014 electrons/(cm2 s). Electron energy: 1 keV.
Detector pass energy: 71.5 eV. (c) TiCl4 and Mg codeposition on 10
layers of MgCl2 film. TiCl4 exposure: 360 L. TiCl4 pressure: 2 ×
10-7 Torr. Mg flux: 3 × 1012 atoms/(cm2 s). TiCl4 to Mg molecular
flux ratio at the target surface: 10:1. Detector pass energy: 8.9 eV.
TiCl4 deposition is quite evident. In both cases, the titanium
atoms are present in oxidation states 4+ and 2+. However,
the amounts of Mg present in the films prepared with these
two methods are substantially different. While no Mg was
detected by XPS in the film obtained by electron irradiation
induced deposition, about 18% of Mg was present in the film
prepared by codeposition of TiCl4 and Mg. The redox surface
reaction between TiCl4 and Mg produced a mixture of TiCl2
and MgCl2. Because of the similarity of the crystallographic
structures of MgCl2 and TiCl2,29-31 a mixed titanium/magnesium
chloride with titanium in the 2+ oxidation state (TiRMg1-RCl2)
might be formed during the codeposition experiment. Accord-
ing to this interpretation, films b and c of Figure 9 and Table 4
would only differ in their subsurface composition, being pure
TiCl2 in film b and a mixed titanium/magnesium chloride in
film c. In both cases, TiCl4 is strongly chemisorbed at the film
surface.
It appears that the TiCl4/TiCl2/MgCl2 film is quite stable at
room temperature, being formed by means of two substantially
different synthetic methods. In particular, the TiCl4 and Mg
codeposition makes use of Mg atoms at conventional thermal
energy, as opposed to the irradiation of the MgCl2 film with
electrons of 1 keV of energy. This new chemical method for
the synthesis of titanium chloride films could allow us to prepare

Model Ziegler-Natta Catalysts
J. Phys. Chem. B, Vol. 102, No. 44, 1998 8795
titanium chloride films of uniform composition and surface
structure. Our goal is to obtain in this way a single-site
heterogeneous Ziegler-Natta catalyst.
under Contract No. DE-AC03-76SF00098. The authors also
acknowledge support from Montell USA, Inc.
References and Notes
5. Conclusions
(1) Dusseault, J. J. A.; Hsu, C. C. J. Macromol. Sci., ReV. Macromol.
Chem. Phys. 1993, C33, 103.
(2) Corradini, P.; Guerra, G. Prog. Polym. Sci. 1991, 16, 239.
(3) Barbe`, P. C.; Cecchin, G.; Noristi, L. AdV. Polym. Sci. 1986, 81,
1.
(4) Giannini, U.; Giunchi, G.; Albizzati, E.; Barbe`, P. C. Recent
AdVances in Mechanistic and Synthetic Aspects of Polymerization; Fonta-
nille, M., Guyot, A., Eds.; NATO ASI Series, Vol. 215; D. Reidel Publishing
Co.: Dordrecht, The Netherlands, 1987; p 473.
(5) Magni, E.; Somorjai, G. A. Surf. Sci. 1996, 345, 1.
(6) Magni, E.; Somorjai, G. A. J. Phys. Chem. 1996, 100, 14786.
(7) Magni, E.; Somorjai, G. A. Surf. Sci. 1997, 377, 824.
(8) Magni, E.; Somorjai, G. A. Catal. Lett. 1995, 35, 205.
(9) Magni, E.; Somorjai, G. A. Appl. Surf. Sci. 1995, 89, 187.
(10) Magni, E.; Somorjai, G. A. Surf. Sci. 1995, 341, L1078.
(11) Fairbrother, D. H.; Roberts, J. G.; Rizzi, S.; Somorjai, G. A.
Langmuir 1997, 13, 2090.
(12) Roberts, J. G.; Gierer, M.; Fairbrother, D. H.; Van Hove, M.;
Somorjai, G. A. Surf. Sci. 1998, 399, 123.
(13) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D.
Handbook of X-ray Photoelectron Spectroscopy; Physical Electronics,
Inc.: Eden Prairie, MN, 1995.
(14) The tail used for the description of the asymmetric XPS peaks has
the mathematical form
We studied the chemisorption of TiCl4 molecules on a clean
Mg surface. Titanium chloride films could be deposited in this
way with the Ti atoms in the 4+, 3+, 2+, and 0 oxidation states.
The composition of this film is quite complex and depends on
the TiCl4 exposure. Upon reaction with AlEt3, this film became
an active catalyst for the Ziegler-Natta polymerization of
ethylene and propylene. With this experiment, we proved the
XPS analysis suitable for the quantitative determination of TiCl4,
TiCl3, and TiCl2 species simultaneously present at the sample
surface.
MgCl2 surfaces with high concentrations of undercoordinated
Mg atoms are unstable under UHV conditions. The Mg adatoms
deposited from the gas phase on a multilayer MgCl2 film were
readily coordinated by the Cl ions diffusing from the halide
bulk to the surface to lower the surface energy of the system.
The MgCl2 film was predominantly Cl-terminated after the
deposition of an equivalent of 10 layers of metallic Mg. This
is an important direct evidence of the dynamic behavior of the
MgCl2 surface at 300 K and of the instability of MgCl2 faces
other than the basal plane under UHV conditions.
TS{1 - exp[-LN(2)(2(Xi - PP)/fwhm)2]} ×
exp{(-6.9/TL)[2(Xi - PP)/fwhm]}
The Mg adatoms deposited on the MgCl2 film were stabilized
at the film surface in the presence of TiCl4 in the gas phase. In
this case, the Mg adatoms reacted with the TiCl4 molecules from
the gas phase before the Cl ions diffusing from the film bulk
could coordinate to the Mg atoms at the interface. Our
experiments show conclusively that surface magnesium atoms
are responsible for the chemisorption of TiCl4 and the formation
of the precursor of the polymerization reaction active sites. The
redox surface reaction between the TiCl4 molecules and the Mg
atoms produced a mixture of TiCl2 and MgCl2 with a monolayer
of TiCl4 molecules chemisorbed at the surface. A similar TiCl4/
TiCl2/MgCl2 film was produced by TiCl4 and Mg codeposition
on a gold foil. The XPS analysis indicated that the titanium
chloride films obtained by TiCl4 and Mg codeposition were very
similar to the ones obtained by electron irradiation-induced TiCl4
deposition. We suggest that the TiCl4/TiCl2 system is highly
stable at room temperature. Both model systems became active
Ziegler-Natta catalysts after the reaction with AlEt3.
Finally, we tried to enhance the growth of high Miller index
faces of MgCl2 by depositing magnesium chloride in the
presence of a large excess of TiCl4. Titanium chloride could
not be deposited under these conditions. This is indicative of
the absence of any Mg at the growing MgCl2 film surface. This
result can be interpreted, once again, in view of the high surface
energy of MgCl2 faces with undercoordinated Mg atoms at the
solid-vacuum interface.
where Xi is the binding energy value of the ith data point, PP is the binding
energy of the peak center, and the remaining symbols are explained in the
text. The curve fitting with synthetic peaks of this mathematical form is
easily accessible using the software provided with our PHI 5400 ESCA
system.
(15) Ley, L.; McFeely, F. R.; Kowalczyk, S. P.; Jenkin, J. G.; Shirley,
D. A. Phys. ReV. B 1975, 11, 600.
(16) Steiner, P.; Reiter, F. J.; Hoechst, H.; Huefner, S.; Fuggle, J. C.
Phys. Lett. 1978, 66 A, 229.
(17) Fuggle, J. C. Surf. Sci. 1977, 69, 581.
(18) Hoogewijs, R.; Fiermans, L.; Vennik, J. J. Electron Spectrosc. Relat.
Phenom. 1977, 11, 171.
(19) Fuggle, J. C.; Watson, L. M.; Fabian, D. J.; Affrossman, S. J. Phys.
F. Met. Phys. 1975, 5, 375.
(20) Garbassi, F.; Pozzi, L. J. Electron Spectrosc. Relat. Phenom. 1979,
16, 199.
(21) Seyama, H.; Soma, M. Chem. Lett. 1981, 7, 1009.
(22) Mousty-Desbuquoit, C.; Riga, J.; Verbist, J. J. Inorg. Chem. 1987,
26, 1212.
(23) Mousty-Desbuquoit, C.; Riga, J.; Verbist, J. J. J. Chem. Phys. 1983,
79, 26.
(24) Briggs, D.; Seah, M. P. Practical Surface Analysis. Volume 1s
Auger and X-ray Photoelectron Spectroscopy, 2nd ed.; J. Wiley & Sons:
Chichester, U.K., 1990.
(25) Tanuma, S.; Powell, C. J.; Penn, D. R. Surf. Interface Anal. 1991,
17, 911.
(26) Ohring, M. The Materials Science of Thin Films; Academic Press:
London, 1992.
(27) Rupprechter, G.; Hayek, K.; Rendo´n, L.; Jose´-Yacama´n, M. Thin
Solid Films 1995, 260, 148.
(28) Henry, C. R.; Chapon, C.; Duriez, C.; Giorgio, S. Surf. Sci. 1991,
253, 177.
(29) Bassi, I. W.; Polato, F.; Calcaterra, M.; Bart, J. C. J. Z. Kristallogr.
1982, 159, 297.
(30) Baenziger, N. C.; Rundle, R. E. Acta Crystallogr. 1948, 1, 274.
(31) Gal’perin, E. L.; Sandler, R. A. Kristallografiia 1962, 7, 217.
Acknowledgment. This work was supported by the Director,
Office of Energy Research, Office of Basic Energy Sciences,
Materials Sciences Division, of the U.S. Department of Energy