Preparation of solid-state samples of a transition metal coordination compound for synchrotron radiation photoemission studies

Article abstract of DOI:10.1016/S0368-2048(02)00272-4

The aim of this research was to identify a sample preparation method suitable for the study of transition metal complexes by photoemission spectroscopy with synchrotron radiation as the X-ray source, even in the case where the compound is not evaporable. Solid-phase samples of W(CO)₄(dppe) [dppe=1,2-bis(diphenylphosphino)ethane] were prepared according to different methods and their synchrotron radiation XPS spectra measured. The spectra acquired from samples prepared by spin coating show core level peaks only slightly broader than the spectrum recorded from UHV evaporated samples. Moreover, for these samples the reproducibility of the binding energy values is excellent. The dependence of the spin coating technique on parameters such as solvent and solution concentration, spinning speed and support material was studied. The same preparation method also allowed the acquisition of valence band spectra, the main peaks of which were clearly resolved. The results suggest that use of the spin coating technique may allow us to reveal small differences in core level binding energies due to ligand substitution.

Full text of DOI:10.1016/S0368-2048(02)00272-4

Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
www.elsevier.com/locate/elspec
Preparation of solid-state samples of a transition metal coordination
compound for synchrotron radiation photoemission studies
b
b
c
Corrado Crottia , , Erica Farnetti , Teresa Celestino , Stefano Fontana
*
aCNR—Istituto Struttura della Materia, Sede di Trieste and Consorzio Interuniversitaro Nazionale per la Scienza e la Tecnologia dei
Materiali (INSTM), S.S. 14, Km 163.5, 34012 Basovizza, Trieste, Italy
bDip. Scienze Chimiche, Univ. di Trieste-V. Giorgieri 1, 34127 Trieste, Italy
cSincrotrone Trieste S.C.p.A.-S.S. 14, Km 163.5, 34012 Basovizza, Trieste, Italy
Received 16 August 2002; received in revised form 21 October 2002; accepted 22 October 2002
Abstract
The aim of this research was to identify a sample preparation method suitable for the study of transition metal complexes
by photoemission spectroscopy with synchrotron radiation as the X-ray source, even in the case where the compound is not
evaporable. Solid-phase samples of W(CO)4(dppe) [dppe51,2-bis(diphenylphosphino)ethane] were prepared according to
different methods and their synchrotron radiation XPS spectra measured. The spectra acquired from samples prepared by
spin coating show core level peaks only slightly broader than the spectrum recorded from UHV evaporated samples.
Moreover, for these samples the reproducibility of the binding energy values is excellent. The dependence of the spin coating
technique on parameters such as solvent and solution concentration, spinning speed and support material was studied. The
same preparation method also allowed the acquisition of valence band spectra, the main peaks of which were clearly
resolved. The results suggest that use of the spin coating technique may allow us to reveal small differences in core level
binding energies due to ligand substitution.
2002 Elsevier Science B.V. All rights reserved.
Keywords: Synchrotron radiation; Sample preparation; Transition metal complexes; Spin coating; Photoemission spectroscopy; Tungsten
carbonyl phosphine
1. Introduction
chemical environment around the metal atom, which
is the site where the catalytic reaction takes place.
This variation may affect both the steric distribution
of the ligands around the metal and the electron
densities on the metal and the ligands. The clearest
demonstration of this dependence is given by the fact
that small variations of the structure of a ligand can
turn an efficient catalyst into an inert complex.
Photoemission spectroscopy, particularly photo-
emission from core levels (XPS), provides an im-
portant contribution to the evaluation of the electron
Transition
metal
coordination
compounds
(TMCCs) play a relevant role in many applications
of modern chemistry, such as homogeneous catalysis
and bioinorganic chemistry. Their importance is
mainly due to the endless possibilities of varying the
*
Corresponding author. Tel.: 139-040-922-1091; fax: 139-040-
226-870.
E-mail address: corrado.crotti@ism.cnr.it (C. Crotti).
0368-2048/02/$ – see front matter
PII: S0368-2048(02)00272-4
2002 Elsevier Science B.V. All rights reserved.

142
C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
density both on the metal and on the ligands, and
indeed its application to TMCCs has been well
known for the last 30 years [1]. However, in order to
detect very small differences in core level binding
energies (BEs), which can be associated with small
variations in the structure of the complex, it is
necessary to acquire XPS spectra with a very high
energy resolution (¯0.1 eV).
problem of differential charging on solid, noncon-
ducting samples, such as evaporation of thin gold
layers on TMCC samples [7], or the use of a low
kinetic energy electron flood gun [4,6]. Unfortuna-
tely, neither of these methods is a general solution,
since they often lead to sample decomposition.
Several articles have reported XPS measurements of
TMCC where the samples were prepared by brushing
the thin sample powder on a conducting support [8]
or dipping the support in a solution of the sample
[1b]. However, these techniques can only be applied
for traditional X-ray source experiments, and they
cannot solve the differential charging problem when
synchrotron light is used, as we show in the present
paper.
In 1976, Bancroft et al. published a paper on the
solid-state broadening effect for XPS studies of tin
organometallic compounds [9]. They clearly demon-
strated that, even with a traditional X-ray source, the
only method for acquiring XPS spectra from solid-
state TMCC samples with characteristics similar to
those obtained from gas-phase samples is the in situ
UHV evaporation of the compound on a conducting
support. Analogous results were reported by Lich-
tenberger et al. for chromium, molybdenum and
tungsten complexes [10]. Unfortunately, this tech-
nique is feasible only for evaporable compounds.
Several recent publications report comparisons of
XPS measurements on solid-state TMCC samples,
where the authors rationalize quite small differences
in BE (¯0.1 eV) without taking into account the
poor overall resolution of their experimental system
(¯0.8 eV for nonmonochromatized sources, plus
¯1.0 eV due to differential charging), even if the
obtained data appear to be reproducible over time
and measurement sets.
The use of synchrotrons as X-ray sources has
given new life to photoemission spectroscopy, allow-
ing incomparable energy resolution and photon
energy tunability within a very wide continuous
range. However, up to now the application of
synchrotron radiation XPS to TMCCs has been
limited to complexes which can be evaporated to the
gas phase [2], and these are very few compared with
the total of TMCCs. The inherent insulating nature
of these compounds causes broadening and shifting
of their photoemission peaks when they are acquired
from solid-state samples. Moreover, the shift de-
pends on the distance between the molecule and the
conducting support, i.e. the layer in which the
photoemitting molecule is lying, thus giving rise to
differential charging [3]. Such problems were much
less severe in early XPS spectrometers, which used
X-ray sources delivering a high divergence beam,
because the X-ray flux density was considerably
more uniform than in modern systems, the mono-
chromatized sources of which deliver smaller spots
[4]. Moreover, in the XPS process itself, there are
often sufficient stray, low-energy electrons from
Bremsstrahlung and other processes to partially
neutralize the charge centres. On the other hand,
when a monochromator is used, we have a relatively
electron-clean environment around the sample [5],
and only recently some monochromatized spectrome-
ters have been modified to minimize charge broaden-
ing [6]. However, problems due to differential
charging become unacceptable if synchrotron radia-
tion is used as the source, owing to its brightness and
the dependence of the photon flux on time.
In the present paper we report a study of the
dependence of broadening effects on the sample
preparation method. We acquired synchrotron radia-
tion XPS spectra from samples of a TMCC prepared
by various techniques, either traditional (brushing
and dipping) or more innovative (spin coating), and
compared the results with data published in the
literature for gas-phase samples or nonsynchrotron
radiation X-ray sources. The main goal was to
identify a procedure which would minimize differen-
tial charging, thus allowing the acquisition of
synchrotron radiation XPS spectra from solid-state
To obtain useful information, and particularly to
exploit the high-resolution characteristics of
synchrotron light in the application to solid-state
TMCC samples, a charge neutralizing technique is
required that completely eliminates this differential
charging on the surface.
Several systems have been applied to solve the

C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
143
samples even for nonevaporable TMCCs. With the
same technique it would also be feasible to acquire
XPS spectra using monochromatized X-ray sources
in order to measure core level BE with higher
resolution.
washed with methanol and dried in vacuo. Yield
76%. 1H NMR (CDCl3, 25 8C): d 7.6–7.3 (m, 20H,
Ar), 2.59 (m, 4H, CH2); 31P NMR (CDCl3, 25 8C): d
140.6 (s, JPW 231.7 Hz); 13C NMR (CDCl3, 25 8C):
d 208.2 (dd, CO trans P, JCPtrans 25.4 Hz, JCPcis 6.9
Hz), 201.3 (t, CO trans CO, JCP 6.9 Hz), 136.4–
128.7 (Ar), 29.9 (m, CH2). IR (Nujol) 2011, 1909,
1886, 1875 cm21 (nCO).
2. Experimental
2.1. General
2.4. XPS measurements
All the reactions and manipulations were routinely
performed under an argon atmosphere using standard
Schlenk tube techniques. W(CO)6 (Strem), 1,2-bis-
(diphenylphosphino)ethane (dppe) (Strem), bis(2-
methoxyethyl)ether (diglyme) and all other chemi-
cals were reagent grade and were used as received
from the commercial suppliers. The compound
W(CO)4(dppe) was prepared according to the pro-
cedure reported in the literature [11] with slight
modifications (vide infra).
The samples for XPS were prepared by the
following procedures:
(a) brushing: ¯50 mg of a very thin powder of the
compound was spread onto a clean gold foil, and
then rubbed with a quartz rod until a very thin
layer intimately pressed on the gold surface was
obtained;
(b) dipping: ¯2 mg of the compound was dissolved
in 10 ml of acetone; a clean gold foil was dipped
into the solution and then left to dry in air;
(c) spin coating: a solution of the compound at the
desired concentration was prepared in the sol-
vent of choice (acetone, chloroform, mesitylene
or dichloromethane); a few drops of the solution
were then placed on the surface of the support
and the spin coater was started for 5 min, at the
appropriate spinning rate, until complete evapo-
ration of the solvent. The spin coater used was a
Model P6708 from Specialty Coating Systems
Inc.;
(d) UHV evaporation: thin films of the compound
were deposited on a gold foil surface that had
been previously cleaned by Ar ion sputtering;
evaporation of the compound occurred from a
quartz crucible heated at 420 K inside the UHV
preparation chamber and the film thickness was
determined by a quartz crystal monitor. Com-
pound integrity was verified by off-line evapora-
tion on a thin silicon wafer and subsequent
comparison of its infrared spectrum with analo-
gous spectra obtained on solution and spin
coated specimens. Samples for XPS measure-
ments were prepared by progressive evaporation
of small amounts of compound, and the deposi-
tions were precisely controlled by monitoring the
2.2. Instrumental
1H, 13C and 31P NMR spectra were recorded on a
JEOL EX400 spectrometer operating at 399.77,
1
100.54 and 161.82 MHz, respectively. H chemical
shifts are reported relative to tetramethylsilane; 13C
chemical shifts are reported relative to the solvent
peak (for CDCl3 d 77.0); 31P chemical shifts are
reported relative to 85% H3PO4 as external standard,
with downfield shifts taken as positive. Infrared
spectra were recorded in Nujol mull on a FT-IR
Perkin-Elmer System 2000 spectrometer.
2.3. Preparation of W(CO)4(dppe)
Tungsten hexacarbonyl (1.0 g, 3.0 mmol) and
dppe (1.2 g, 3.0 mmol) were added into a Schlenk
tube containing 5 ml of diglyme. The reaction
mixture was heated at reflux for 45 min, during
which time the hexacarbonyl sublimed on the
Schlenk walls was periodically returned to the
solution by vigorous shaking. The brown solution
obtained was cooled to room temperature, yielding a
light-yellow solid. Addition of 5 ml of methanol to
complete the precipitation of the product was fol-
lowed by filtration: the solid thus obtained was

144
C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
W 4f, P 2p and Au 4f core level intensities for
each evaporation. The pressure in the preparation
chamber was ,2.031029 mbar during prepara-
tion.
2.6. Data analysis
For both XPS and UPS spectra, the sum of several
(8–16) rapid, single scan spectra (¯85 s/scan), each
recorded on a fresh sample area (¯0.1 mm2), gave a
final spectrum with a higher signal/noise ratio and
careful control of the homogeneity of the compound
distribution on the sample. Spectra were fitted with a
Gaussian–Lorentzian line shape using a nonlinear
least-squares procedure, modified to allow the vari-
ation of both Gaussian and Lorentzian components
for each doublet of peaks present in the spectrum.
For each sample, W 4f and P 2p binding energies
were referenced to the Au 4f7 / 2 peak at 84.00 eV,
thus eliminating any error due to fluctuation of
photon energy or electron kinetic energy.
Each sample was then introduced into the ex-
perimental chamber through a load-lock chamber.
During all measurements, the experimental chamber
pressure was no higher than 2.0310210 mbar. No
evaporation of the compound from the gold support
occurred at room temperature in UHV. The sample
holder on the manipulator allowed sample cooling at
110 K, as well as gentle warming by a resistance
placed behind the sample.
All XPS and UPS spectra were recorded by an
angle-integrated Omicron EA 125 HR energy ana-
lyzer mounted in the experimental chamber of the
VUV-Photoemission beam-line [12] at the Elettra
Synchrotron Facility (Trieste-I). The synchrotron
radiation beam was normal incident with respect to
the surface; photoelectrons were collected with the
analyzer axis at 458 to the substrate surface normal.
The photon energy was fixed for all core level
spectra at 201.31 eV, which was the best compromise
in order to maximize the photoionization cross-sec-
tions for both W 4f and P 2p core levels [13]. To
ensure high energy resolution and to minimize
sample decomposition, the spectra were acquired
with a pass energy of 5 eV, and both beam-line slits
closed at 10 mm each, thus decreasing the photon
flux on the sample. The resulting overall instrumen-
tal resolution (photons plus electrons) was 0.10 eV.
2.7. AFM images
AFM images were recorded with a AFM micro-
scope designed and built by the A. Cricenti group in
the ISM laboratories, Rome [15]. The maximum
lateral resolution was about 20 nm. In order to
optimize the AFM measurements, the images were
recorded on samples prepared by spin coating on a
steel support mechanically polished down to 0.1 mm
roughness.
3. Results and discussion
3.1. Choice of compound
2.5. UPS measurements
In order to fully evaluate the effects of the
preparation method on the quality of the XPS
spectra, a careful choice of the model compound had
to be made. We chose to carry out the tests on
W(CO)4(dppe) samples for the following reasons: (i)
W 4f are very narrow core levels, therefore they are
particularly suitable for the evaluation of the
broadening caused by differential charging; (ii) the P
2p core levels are also useful, because phosphorus-
based ligands are widely used in organometallic
chemistry and the possibility to visualize their spin–
orbit splitting offers a further indication of reduced
differential charging; (iii) the room temperature
vapor pressure of this compound is low enough to
avoid any evaporation in UHV, but the compound
Valence level spectra of the sample prepared by
the spin coating procedure were acquired under the
same conditions as for the XPS spectra, but with the
photon energy set at 80.25 eV, in order to compare
our results with the spectra acquired for analogous
complexes in the gas phase and already published in
the literature [14], and at 49.18 eV, to detect peak
intensity changes induced by cross-section varia-
tions. Two spectra were recorded under the same
conditions (at 80.25 and 49.18 eV) from the gold
supports, as they appear before deposition of the
compound, in order to acquire the background
spectra to be subtracted from the sample spectra.

C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
145
can be evaporated by rather moderate heating with-
3.3. W 4f measurements
out thermal decomposition, thus allowing analysis of
solid-state, gas-phase and evaporated samples; (iv)
the compound is soluble in many common organic
solvents and its solutions are quite stable, thus
allowing good flexibility of applications; (v) many
physico-chemical data on analogous complexes are
available in the literature [16]; (vi) several papers
have been published on the photoemission spec-
troscopy of phosphine carbonyl complexes of
molybdenum and tungsten, particularly because of
the interest in the effect of variation of the phosphine
ligand [14,17,18] and the nature of the phosphorus–
metal bond [19].
The W 4f XPS spectra of the compound
W(CO)4(dppe), acquired for samples prepared by
different methods, are shown in Fig. 1. The doublet
peaks due to the original compound can be observed
at about 163 and 165 eV kinetic energy (KE). The
second doublet at much lower KE (158–160 eV) is
related to a molecular species with the tungsten atom
in a higher oxidation state and is produced by
decomposition of the original compound during
sample preparation. Its intensity depends on the
sample preparation procedure, but it does not change
upon beam exposition.
On the other hand, the compound was shown to
decompose upon beam exposition, thus leading to a
third, very large doublet at about 162 and 164 eV. In
order to minimize the latter decomposition, we
reduced both the photon flux on the sample and the
acquisition time for each single point of the scan,
thus obtaining very quick single scans.
3.2. Comparison of the different preparation
methods
The preparation methods were chosen from tradi-
tionally used XPS techniques (brushing and dipping)
and more innovative methods (spin coating). The
latter is commonly used in the deposition of photo-
resists on semiconductors during the preparation of
electronic devices, and the underlying physics has
been studied [20]. However, its use for the prepara-
tion of samples for photoemission spectroscopy is
mainly limited to conjugated polymers [21], whereas
its application to the preparation of TMCC samples
is still rather uncommon and has not been thoroughly
investigated [22]. The last method tested was UHV
evaporation, since this technique is known to give
the best control of the homogeneity and thickness of
the deposited material [9]. Therefore, the samples
produced in this way ensured minimization of the
differential charging phenomena, and the photo-
emission spectra should be considered as the optimal
results obtainable from solid-state samples.
For each sample, the quality of the spectrum was
evaluated on the basis of the BE of the core level
peaks, since a higher BE indicates higher charging,
and of the Gaussian FWHM (G ), since a less
homogeneous deposition of the molecules on the
surface—upon a statistical distribution—leads to a
broader peak, i.e. a higher G. Moreover, a full
comparison with the results reported by other authors
is meaningful only on the basis of G, whereas BEs
are more affected by differences in the experimental
setup.
The spectra obtained by this procedure were
analyzed by peak fitting, thus allowing the identifica-
tion of the components contributing to the total
spectrum, and a more precise measurement of the
KE, intensity and width of the single components.
An example of the fit is shown in Fig. 2, which
shows the W 4f core level spectrum obtained from
the sample prepared by spin coating from a CH2Cl2
solution. Three components can be identified: the
first and most intense is due to the W 4f atoms in the
original complex, W(CO)4(dppe) [W 4f5 / 2 KE5
163.143(6) eV, G50.680(14) eV]. The second,
broader component [W 4f5 / 2 KE5162.176(282) eV,
G51.265(633) eV] is related to the species origina-
ting from sample decomposition upon beam exposi-
tion, whereas the third component [W 4f5 / 2 KE5
158.546(75) eV, G51.301(138) eV] is related to
decomposition of the compound during sample prep-
aration. The intensity of the second component is
only 10% of that of the signal of the original
compound. However, when several scans are ac-
quired on the same spot of the sample (thus leading
to progressive compound decomposition upon beam
exposition), we observe a reduction of the com-
ponent relative to the original compound and a
corresponding increase of the second component,
which at the same time becomes much broader. With

146
C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
Fig. 1. W 4f core level spectra of W(CO)4(dppe) samples prepared by different methods: (a) UHV evaporation; (b) spin coating from a
CH2Cl2 solution; (c) spin coating from a CHCl3 solution; (d) spin coating from a (CH3)2CO solution; (e) spin coating from a
1,3,5-C6H3(CH3)3 solution; (f) dipping into a (CH3)2CO solution; (g) brushing.
this experimental setup (which delivers ¯1.631011
photons/s), the W(CO)4(dppe) lifetime, as measured
by monitoring the reduction in intensity of the peaks
related to the original compound upon beam exposi-
tion, was about 9 min.
For all three components we optimized the values
of the lorentzian width (50.063 eV), the branching
ratio (51.333) and the spin–orbit splitting (52.151
eV). These results are reasonably in agreement with
the data reported in the literature [14,23], therefore
the same values were used for the peak fitting of all
the other W 4f spectra.
The results of the fit for the spectra obtained from
samples prepared by different methods (brushing,
dipping and spin coating) and by spin coating from
solutions in different solvents (mesitylene, acetone,
chloroform and dichloromethane) are summarized in
Table 1. For each preparation are listed the W 4f5 / 2
KE obtained by peak fitting, the W 4f7 / 2 BE derived
from the W 4f7 / 2 KE and referenced to the Au 4f7 / 2
BE at 84.00 eV and G; the 2s fitting error is also
reported in parentheses for each value. In the same
table, we also list for comparison the corresponding
values for two other preparation methods reported in
the literature: a spectrum obtained by brushing the
same compound, but using a monochromatized X-
ray gun instead of synchrotron radiation [17], and a
spectrum obtained with synchrotron radiation for
cis-W(CO)4(PMe3)2 in the gas phase [14].
The W 4f spectra reported in Fig. 1 clearly show
that brushing and dipping are unsuitable for the
preparation of samples for synchrotron radiation
XPS, because of their broadness and the uncertain
determination of BE. Moreover, the bad quality of
the spectra makes any peak fitting meaningless,
therefore the corresponding values are missing from
Table 1. However, it should be emphasized that
samples prepared by the same methods would be
suitable for XPS with a traditional X-ray source,
although with a lower resolution.

C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
147
Fig. 2. Experimental W 4f core level spectra of W(CO)4(dppe) samples prepared by spin coating from a CH2Cl2 solution and its fit: (a)
overall fitted spectrum; (b) spectrum of the first component; (c) spectrum of the second component; (d) spectrum of the third component.
The residual shown is the difference between the experimental and overall fitted spectra.
Spectra obtained from spin coated samples (Fig.
1(b)–(e)) are strongly dependent on the solvent used
for the solution, both for the BE and FWHM of the
peaks related to the original compound and for the
amount of decomposed compound at lower KE. As
shown in Table 1, the best results are obtained with
more volatile solvents [i.e. CH2Cl2 .(CH3)2CO¯
CHCl3]. On the other hand, the good BE peak fitting
result for the sample from mesitylene solution is not
fully reliable for the detection of the simultaneous
presence of other overlapped signals. The best results
from the spin coated sample (i.e. from CH2Cl2
solution) are clearly not comparable to those from
gas-phase samples, where the FWHM is so narrow to
allow the resolution of the vibrational fine structure
due to W–CO stretching and the detection of
broadening due to ligand field splitting [14]. How-
ever, spin coated samples approach the results for
UHV evaporated samples, with a very small shift
towards higher BE and a more pronounced increase
of G. Moreover, the same spin coated samples give
far more resolved spectra than those reported from
samples prepared by brushing and with a mono-
chromatized X-ray gun [17].
The results obtained for the spin coating technique
appear to be dependent on other parameters in
addition to the solvent: the spinning rate, which has
to be as fast as possible to minimize the thickness of
the deposited compound on the surface [20], treat-
ment and material of the support on which to deposit
the compound (see below in the discussion of the
AFM results) and the concentration of the starting
solution. Tests carried out to explore the conse-
quences of sample cooling at 110 K, as well as
gentle warming at 390 K, did not show significant
effects on W 4f BE and G. Only a slight increase in

148
C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
Table 1
W 4f KEs, BEsa and Gaussian widths (G ) for W(CO)4(dppe) samples prepared by different methods (in eV)
Preparation method
W 4f5 / 2 KE
163.264(8)
W 4f7 / 2 BE
31.028(8)
G
UHV evaporation
Spin coating from a
CH2Cl2 solution
Spin coating from a
CHCl3 solution
Spin coating from a
(CH3)2CO solution
Spin coating from a 1,3,5-
C6H3(CH3)3 solution
Dipping in a (CH3)2CO
solution
0.493(11)
163.143(6)
162.994(13)
163.010(8)
163.259(28)
31.118(6)
31.312(13)
31.305(8)
30.996(28)
0.680(15)
0.751(30)
0.765(19)
0.836(71)
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
Brushing
Brushing; monochromatized
X-ray gunb
Gas phase synchrotron
radiation XPSc
n.a.
n.a.
31.2
1.5
36.400(20)
0.320(5)
a BEs are referenced to the Au 4f7 / 2 BE at 84.00 eV.
b Ref. [17].
c cis-W(CO)4(PMe3)2 [14].
peak intensity was observed as a result of sample
warming, probably due to partial degassing of the
surface from small co-adsorbed molecules (i.e. H2O,
CO, CO2, solvent, etc.).
and exit beam-line slit widths increased from 10 to
20 mm, thus increasing the photon flux on the sample
by a factor of ¯5. Although a greater number of
photons accelerates decomposition of the compound,
a comparison between spectra at higher and lower
photon fluxes gives a useful indication of the amount
of differential charging present on the sample, since
charging phenomena are roughly proportional to the
number of photons impinging on the sample.
The results reported in Table 2 show that the W 4f
BE decreases on increasing the dilution of the
starting solution, until a concentration of 0.127
mmol/l is reached. Further dilution does not give a
subsequent improvement. The effect on G is not as
The influence of the concentration of the starting
solution on the quality of the W 4f spectra was
investigated in the case of samples prepared by spin
coating from dichloromethane solutions. The results
of the peak fitting of such spectra are listed in Table
2, where the W 4f5 / 2 KE, W 4f7 / 2 BE and G
corresponding to solutions with concentrations be-
tween 0.058 and 0.469 mmol/l are reported. For
each sample obtained from a different concentration,
a second spectrum was acquired with both entrance
Table 2
W 4f KEs, BEsa and Gaussian widths (G ) for W(CO)4(dppe) samples prepared by spin coating from CH2Cl2 solution at different
concentrations and at different slit openings
Conc.
Slit
(mmol/l)
10/10 mm
20/20 mm
W 4f5 / 2 KE
(eV)
W 4f7 / 2 BE
(eV)
G
W 4f5 / 2 KE
(eV)
W 4f7 / 2 BE
(eV)
G
(eV)
(eV)
0.469
0.234
0.127
0.058
162.618(5)
162.680(4)
162.760(4)
162.743(9)
31.290(5)
31.195(4)
31.104(4)
31.129(9)
0.676(9)
162.482(12)
162.602(10)
162.775(11)
162.728(14)
31.426(5)
31.273(10)
31.089(11)
31.144(14)
0.905(39)
0.891(45)
0.816(47)
0.659(43)
0.697(13)
0.666(10)
0.655(23)
a BEs are referenced to the Au 4f7 / 2 BE at 84.00 eV.

C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
149
evident. The W 4f BE dependence on the con-
centration is clearly related to the thickness of the
layer deposited on the support. Confirmation for this
comes from a comparison of the results obtained at
10/10 and 20/20 mm. For both 0.469 and 0.234
mmol/l solutions, a greater photon flux results in an
increase in W 4f BE and G; that is a strong evidence
of differential charging occurring on the sample. On
the other hand, there is no significant effect on W 4f
BE of the slit opening for both 0.127 and 0.058
mmol/l solutions, thus confirming that differential
charging is minimized in these cases.
It should be pointed out that the W 4f BE obtained
for the 0.127 mmol/l concentration is very close to
the value shown in Table 1, although the two W 4f
KEs are different. The reason for this lies in the use
of slightly different photon energies for the two
experimental data sets, which are compensated by
referencing to Au 4f7 / 2 BE at 84.00 eV. As far as the
precision of the measured BE is concerned, it is
important to underline that the value reported in
parentheses is the 2s error caused by the uncertainty
in the fitting procedure. The generally small value,
particularly for the W 4f core level, is related to the
high precision available for the determination of the
KE of the peak, on the basis of the experimental data
set. On the other hand, the reproducibility in the
acquisition of the experimental result is the parame-
ter we need to maximize in order to allow the use of
this technique to detect the BE shift for two different
complexes. We have found that, using the spin
coating technique under the most optimal conditions,
the W 4f spectra are reproducible within 620 meV.
3.4. P 2p measurements
Analogous XPS spectra of the P 2p core levels
recorded from samples prepared using the same
techniques as in Fig. 1 are shown in Fig. 3. In this
case the peaks related to decomposed compounds are
Fig. 3. P 2p core level spectra of W(CO)4(dppe) samples prepared by different methods: (a) UHV evaporation; (b) spin coating from a
CH2Cl2 solution; (c) spin coating from a CHCl3 solution; (d) spin coating from a (CH3)2CO solution; (e) spin coating from a
1,3,5-C6H3(CH3)3 solution; (f) dipping into a (CH3)2CO solution; (g) brushing.

150
C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
not as separated as in the case of the W 4f core level,
but their presence leads to an overall broadening of
the doublet centered around 65 eV. The broad
shoulder at about 68 eV in Fig. 3(a), obtained from a
sample prepared by UHV evaporation, is assigned to
a gold Auger transition, made more evident by the
low coverage of the gold surface from compound
deposition.
component to the compound decomposed during
sample preparation. The best values obtained by
peak fitting for the lorentzian width (50.120 eV), the
branching ratio (52.000) and the spin–orbit splitting
(50.860 eV) are in agreement with analogous data
reported in the literature [24], therefore the same
values were used for the peak fitting of all the other
P 2p spectra.
As in the case of W 4f, an example of a fitted
spectrum is shown in Fig. 4, with the P 2p core level
peak fitting of the spectrum obtained from the
sample prepared by spin coating from a CH2Cl2
solution. Also for P 2p peaks, the spectrum can be
reproduced by a three-component fitting, with the
most intense component [P 2p1 / 2 KE564.655(88)
eV, G50.603(105) eV] being correlated with the
original compound. Although the fitting errors for
the other two components [P 2p1 / 2 KE5
63.948(10.782) eV, G51.174(7.966) eV and P 2p1 / 2
KE564.262(1.345) eV, G50.752(1.338) eV] are
much greater, on the basis of their intensities we can
assign the second component to the compound
decomposed under beam exposition and the third
The results of peak fitting for all the P 2p spectra
obtained from different preparation methods are
listed in Table 3. As in Table 1, we also list for
comparison the corresponding values published for
W(CO)4(dppe), with a monochromatized X-ray gun
[17], and for cis-W(CO)4(PMe3)2 with synchrotron
radiation in the gas phase [14].
Fig. 3 shows that the conclusions drawn for the W
4f core level spectra are confirmed for the P 2p
peaks: brushing, dipping and spin coating from a
mesitylene solution are not suitable techniques for
the preparation of samples for synchrotron radiation
XPS. Spin coated samples lead to much more
defined spectra, but only the use of a CH2Cl2
solution allows us to approach the shape of the peaks
Fig. 4. Experimental P 2p core level spectra of W(CO)4(dppe) samples prepared by spin coating from a CH2Cl2 solution and its simulation
by fitting: (a) overall fitted spectrum; (b) spectrum of the first component; (c) spectrum of the second component; (d) spectrum of the third
component. The residual shown is the difference between the experimental and overall fitted spectra.

C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
151
Table 3
P 2p KEs, BEsa and Gaussian widths (G ) for W(CO)4(dppe) samples prepared by different methods (in eV)
Preparation method
P 2p1 / 2 KE
64.693(6)
P 2p3 / 2 BE
130.869(6)
G
UHV evaporation
Spin coating from a
CH2Cl2 solution
Spin coating from a
CHCl3 solution
Spin coating from a
(CH3)2CO solution
Spin coating from a 1,3,5-
C6H3(CH3)3 solution
Dipping in a (CH3)2CO
solution
0.592(18)
0.603(105)
0.767(239)
0.615(112)
0.939(63)
64.655(88)
64.394(408)
65.545(181)
64.539(29)
130.876(88)
131.182(408)
131.040(181)
130.986(29)
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
Brushing
Brushing; monochromatized
X-ray gunb
Gas-phase synchrotron
radiation XPSc
n.a.
n.a.
131.4
n.a.
136.760(80)
0.500(30)
a BEs are referenced to the Au 4f7 / 2 BE at 84.00 eV.
b Ref. [17].
c cis-W(CO)4(PMe3)2 [14].
obtainable by UHV evaporation. Moreover, the evi-
dent splitting between the P 2p1 / 2 and P 2p3 / 2
components is further evidence of the very good
resolution [24], although, for this core level, the
peaks related to decomposed complexes are more
overlapped with those of the original compound, thus
increasing the overall broadening. One more conse-
quence of the overlapped peaks is the increased
difficulty in the peak fitting procedure, which leads
to greater 2s errors. Spin coated samples from a
CH2Cl2 solution give, in this case, the same P 2p BE
and G of UHV evaporation, and the lowest G of the
spin coated samples. The P 2p G is very close to the
W 4f G reported in Tables 1 and 2, which is
reasonable, since the Gaussian FWHM derives from
the instrumental resolution and from the residual
differential charging, and these are the same for both
core levels. The P 2p spectrum from the CH2Cl2 spin
coated sample is really very good when compared
with that recorded from an analogous gas-phase
sample [14], both in overall appearance (Fig. 4), and
G (Table 3). No experimental data have been re-
ported for the P 2p FWHM from samples prepared
by brushing and with a monochromatized X-ray gun
[17], but it seems reasonable to presume a FWHM
close to that reported for W 4f (¯1.5 eV), and
therefore much higher than the 0.6 eV we report.
3.5. AFM images
In order to observe the real distribution of the
molecules on the substrate, we acquired AFM im-
ages of a sample obtained by spin coating deposition
from a dilute dichloromethane solution on a carefully
polished steel substrate. Two images are shown in
Fig. 5, acquired before (a) and after (b) the deposi-
tion of the compound. A comparison of the two
images allows us to reveal the disappearance of the
long diagonal stripes (Fig. 5(a)), which are the
residual structure due to the mechanical polishing of
the surface. After compound deposition, the surface
appears more uniform with the exception of small
grains of compounds, one of which is shown in the
A–A9 section, the height of which ranges between 4
and 12 nm. We observed a clear decrease of both the
height and number of grains as the concentration of
the spin coated solution decreased. The regular
surface upon deposition as shown in Fig. 5(b),
compared with the structures in the naked support
shown in Fig. 5(a), is good evidence of the suitability
of the spin coating technique. The core level peak
broadening in comparison with the spectrum ob-
tained by UHV evaporation arises from the residual
inhomogeneity of the molecular distribution, and
mostly from the presence of small sample grains and

152
C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
Fig. 5. AFM images of a stainless steel plate acquired (a) before and (b) after the spin coating deposition of W(CO)4(dppe) from a dilute
CH2Cl2 solution.
from the irregularity of the metallic support surface.
The number and size of sample particles are mini-
mized by an increase in dilution, and would probably
be further reduced by using suitable microfilters on
the solution before spin coating. The problem of the
irregularity of the metallic support has to be mini-
mized by control of the support polishing. The
support shown in Fig. 5(a) underwent careful polish-
ing down to 0.1 mm roughness, and, indeed, no
structures evidenced in the A–A9 section exceed 12
nm. However, since a W(CO)4(dppe) molecule has a
necessary for the use of the spin coater. Moreover,
stainless steel allows good and rather easy polishing
of the surface. A final gold coating ensures a high
chemical inactivity towards the sample, although it is
well known that a gold surface may also react with
the phosphine ligand [26]. Moreover, the Au 4f core
levels are very suitable as a reference for other core
level BEs.
3.6. UPS measurements
˚
size of 10.538.6311.2 A [based on Mo(CO)4(dppe)
Valence level spectra of the sample prepared by
the best spin coating procedure were acquired at
49.18 and 80.25 eV photon energy. After subtraction
of a Gaussian tail as background caused by sec-
ondary electrons, the Au valence level spectra were
subtracted from the W(CO)4(dppe) spectra until the
Fermi edge completely disappeared. The final spectra
are shown in Fig. 6, with the single components
obtained by peak fitting. The BE and overall FWHM
structural data [25], as no data are available for
W(CO)4(dppe)], even the shallow valleys shown in
Fig. 5(a) allow the formation of at least 13–14 layers
before being completely filled by the sample.
Therefore, these observations introduce the prob-
lem of the choice of the material for the support. The
use of carefully polished stainless steel plates fulfills
the need for a rigid and flat support, which is

C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
153
Fig. 6. Valence level spectra of a W(CO)4(dppe) sample prepared by spin coating from a dilute CH2Cl2 solution, with 49.18 eV (a) or 80.25
eV (b) photon energy. The single peaks obtained by fitting are shown as dotted lines.
resulting from the fitting procedure for both photon
energies are listed in Table 4. An overview spectrum
of the valence band for gas-phase samples of
W(CO)4(dppe) with synchrotron radiation is not
available in the literature, but we list in Table 4 some
gas-phase data reported for cis-W(CO)4(PMe3)2
with synchrotron radiation [14], and for
W(CO)4(dppm)
[dppm51,2-bis(diphenylphos-

154
C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
Table 4
Valence band BEs and Gaussian widths (G ) for W(CO)4(dppe) samples prepared by spin coating from a dilute CH2Cl2 solution (in eV)
compared with literature data
Peak
Assignment
hn 5 49.18 eV
hn 5 80.25 eV
Literature data
BE
G
Da
BE
G
Da
BE
G
Da
11
10
9
8
7
Satellite 2
Satellite 1
C 2s
Valence shake-up
4s (CO)
20.42
17.85
16.08
14.23
11.96
2.42
2.23
1.69
1.84
1.10
17.95
15.37
13.61
11.76
9.49
21.17
18.22
18.06
14.50
12.30
3.00
1.20
4.30
2.39
1.32
18.27
15.32
15.16
11.60
9.40
26.14b
23.60b
21.86b
18.66b
16.50b
2.33b
2.13b
1.35b
2.80b
1.00b
19.01
16.47
14.73
11.53
9.37
6
5s11p (CO)
1s (C–H)
s (P–C)
9.71
7.48
6.11
5.21
4.56
2.47
3.04
2.24
1.03
0.93
0.93
0.92
7.24
5.01
3.63
2.73
2.08
–
10.24
8.47
7.45
6.24
4.96
2.90
2.74
2.50
1.62
1.39
1.16
1.14
7.34
5.57
4.55
3.34
2.06
–
14.47c
12.67c
11.22c
10.62c
9.74c
n.a.
n.a.
n.a.
n.a.
n.a.
0.50b
7.34
5.54
4.09
3.49
2.61
–
5
4
3
2
1
s (P–C)
s (W–P)
p (Phenyl)
W 5d
7.13b
a D5(BEpeak 2BEW 5d).
b cis-W(CO)4(PMe3)2 [14].
c Fig. 2(b) of Ref. [18f] for W(CO)4(dppm).
phino)methane] with a He lamp [18f]. However, the
BE values from gas-phase samples are not readily
comparable with solid-phase samples, because, for
the latter, it is necessary to also take into account the
work function F (which is not readily calculable for
our system), whereas F 5 0 for a gas. Therefore, we
have also listed in the same table the BE difference
between each peak and the W 5d peak [D5
(BEpeak 2BEW 5d)], since the W 5d peak is one of
the most separated and recognizable peaks. The peak
fitting was started on the basis of the peaks whose
presence was clearly evident from the shape of the
overall spectrum, thus leading to a first, rather good
agreement. Only during the refining procedure, and
particularly by comparison with the gas-phase re-
sults, did we realize the necessity of adding more
peaks (namely peaks 3 and 9), thus obtaining better
agreement between real and fitted spectra, and with
the data reported in the literature.
they are reproducible with a change in photon
energy. Moreover, the agreement with the reported
gas-phase results is rather good, as the D com-
parisons show for most of the peaks. Also, a visual
comparison between our 80.25 eV spectrum and the
He lamp spectrum of W(CO)4(dppm) [18f] confirms
the reliability of the spectrum. In any case, it should
be taken into account that we are looking at a
spectrum of a solid-state sample prepared in air: the
presence of small co-adsorbed molecules (although
partially subtracted together with the background)
and the intrinsic broadening due to solid-state effects
hardly make it comparable to the resolution obtained
in the gas phase. As a matter of fact, only a few
examples of TMCC valence band spectra from solid-
state samples can be found in the literature [10], and,
as far as we know, none of the reported samples
were prepared in air. On the other hand, we found
papers reporting the valence band spectra from UHV
evaporated samples, since preparation in air had
given unacceptable results [27]. Nevertheless, the
spectrum we obtained from the spin coating sample
is suitable to identify most of the peaks in the
valence band of W(CO)4(dppe).
Although for the UPS spectra the fitting procedure
gave poorer agreement and larger fitting errors than
for the core level spectra, the results shown in Fig. 6
and Table 4 are surprisingly encouraging. After
subtraction of the secondary electrons Gaussian
baseline and of the gold background, the structures
in the valence band are rich and well defined, and
In order to shed light on the decomposition
mechanism on beam exposition, a series of scans

C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
155
corresponding peak. However, it appears that the
phenyl rings of the phosphine (peak 2) are not
significantly affected by decomposition.
4. Conclusions
Solid-state samples of W(CO)4(dppe) were pre-
pared by different methods in order to record
synchrotron radiation photoemission spectra with
very low broadening caused by differential charging.
For each preparation method, the XPS spectra of the
W 4f and P 2p core levels were acquired and their
BE and Gaussian FWHM measured. Samples pre-
pared by brushing or by dipping gave unacceptable
spectra, whereas those obtained by the spin coating
preparation method showed much better resolution.
We found that the best results are obtained by spin
coating at a high spinning rate and from dilute
solutions in low boiling solvents. AFM images of the
metallic support before and after sample spin coating
show significant homogeneity in the distribution of
the molecules, and stress the importance of a starting
support with very low roughness. Spectra obtained
by spin coating under optimal conditions gave core
level BE and Gaussian FWHMs very close to those
acquired by the best known method for the prepara-
tion of solid-state samples, e.g. evaporation of a thin
film of the compound in UHV.
The sample prepared by the same method also
gave UPS spectra that were suitable for identifying
the main characteristics of the valence band. This
was confirmed by comparison with the gas-phase
spectra of analogous compounds, although solid-state
effects introduce unavoidable peak broadening.
However, the availability of solid-state samples for
synchrotron radiation UPS opens up several possi-
bilities in the field of advanced techniques in the
study of transition metal coordination compounds,
such as resonant photoemission [2].
Fig. 7. Sequence of valence level spectra of a W(CO)4(dppe)
sample prepared by spin coating from a dilute CH2Cl2 solution,
acquired on the same sample spot, with 80.25 eV photon energy
(top spectrum, first scan; bottom spectrum, last scan).
was acquired on the same sample spot, each one
lasting about 70 s for a photon flux of ¯5.131011
photons/s. The sequence of single scans (Fig. 7)
shows a strong decrease in intensity in the region
between 8 and 12 eV, which corresponds to peaks 5,
6 and 7, as confirmed by peak fitting of the single
scans. The same fit showed a general broadening of
all peaks, but particularly of peak 1. Identification of
peaks 6 and 7 as CO-based orbitals [peak 6, 5s11p
(CO); peak 7, 4s (CO)] allows us to propose an
hypothesis for sample decomposition based on the
loss or the transformation of the carbonyls. Such
decomposition can be caused by interaction of the
compound either with high energy photons or with
electrons photoemitted from vicinal molecules, as
has already been observed for other metal carbonyls
under synchrotron radiation exposure [28]. The
broadening of peak 1 (i.e. W 5d orbitals) can be
easily explained by the formation of several W-
containing species with different coordination en-
vironments, as observed for the broadening of the W
4f core level. It is not as easy to formulate any
hypothesis about the fate of the coordination of the
phosphine ligand, because of little evidence of the
Therefore, spin coating has been shown to be a
suitable method for the preparation of samples for
synchrotron radiation photoemission spectroscopy,
and most likely even for transition metal complexes
which cannot be evaporated. It seems therefore that
the spectra so obtained could be used to detect core

156
C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
Antonini, M. Capozi, S. Rinaldi, M. Luce, P. Perfetti, K.C.
Prince, C. Astaldi, M. Zacchigna, L. Romanzin, A. Savoia,
Nucl. Instr. Methods A 364 (1995) 374.
level BE shifts related to small modifications in the
structure of the compound.
[13] J.J. Yeh, I. Lindau, At. Data Nucl. Data Tables 32 (1985) 1.
[14] J. Wu, G.M. Bancroft, R.J. Puddephatt, Y.F. Hu, X. Li, K.H.
Tan, Inorg. Chem. 38 (1999) 4688.
Acknowledgements
[15] (a) A. Cricenti, R. Generosi, Rev. Sci. Instr. 66 (1995) 2843;
(b) A. Cricenti, G. De Stasio, R. Generosi, M.A. Scarselli, P.
Perfetti, M.T. Ciotti, D. Mercanti, J. Vac. Sci. Technol. B 14
(1996) 1395;
(c) A. Cricenti, M.A. Scarselli, R. Paleari, A. Mosca, J. Vac.
Sci. Technol. B 12 (1994) 1494.
[16] (a) G. Boxhoorn, J.M. Ernsting, D.J. Stufkens, A. Oskam,
Thermochim. Acta 42 (1980) 315;
We are deeply grateful to all colleagues, research-
ers and technicians of the Istituto Struttura della
Materia (National Council for Research) for their
suggestions, support and encouragement. We thank
Sincrotrone Trieste S.C.p.A. for financial support
through a Basic Research Project and for the pur-
chase of the spin coater instrument.
(b) G. Al-Takhin, J.A. Connor, G. Pilcher, H.A. Skinner, J.
Organomet. Chem. 265 (1984) 263.
[17] B.J. Brisdon, G.F. Griffin, J. Pierce, R.A. Walton, J. Or-
ganomet. Chem. 219 (1981) 53.
References
[18] (a) H. Daamen, A. Oskam, D.J. Stufkens, Inorg. Chim. Acta
38 (1980) 71;
(b) B.E. Bursten, D.J. Darensbourgh, G.E. Kellogg, D.L.
Lichtenberger, Inorg. Chem. 23 (1984) 4361;
(c) G.M. Bancroft, L. Dignard-Bailey, R.J. Puddephatt,
Inorg. Chem. 23 (1984) 2369;
(d) D.L. Lichtenberger, G.E. Kellogg, G.H. Landis, J. Chem.
Phys. 83 (1985) 2759;
(e) R.J. Puddephatt, L. Dignard-Bailey, G.M. Bancroft,
Inorg. Chim. Acta 96 (1985) L91;
(f) G.M. Bancroft, L. Dignard-Bailey, R.J. Puddephatt,
Inorg. Chem. 25 (1986) 3675;
[1] (a) C.D. Cook, K.Y. Wan, U. Gelius, K. Hamrin, G. Johan-
sson, E. Olsson, H. Siegbahn, C. Nordling, K. Siegbahn, J.
Am. Chem. Soc. 93 (1971) 1904;
(b) L. Wang, J.R. Sowa Jr., C. Wang, R.S. Lu, P.G. Gassman,
T.C. Flood, Organometallics 15 (1996) 4240;
(c) N.E. Gruhn, D.L. Lichtenberger, in: A.B.P. Lever, E.I.
Solomon (Eds.), Inorganic Electronic Structure and Spec-
troscopy, Vol. II, Wiley, New York, 1999, pp. 553–574, and
references therein.
[2] (a) J.C. Green, Acc. Chem. Res. 27 (1994) 131;
(b) X. Li, G.M. Bancroft, R.J. Puddephatt, Acc. Chem. Res.
30 (1997) 213, and references therein.
[3] (a) D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis,
Vol. 1, Wiley, New York, 1990, Appendix 2;
(b) B.T. Tielsch, J.E. Fulghum, Surf. Interface Anal. 24
(1996) 422;
(c) B.T. Tielsch, J.E. Fulghum, D.J. Surman, Surf. Interface
Anal. 24 (1996) 459.
[4] P.E. Larson, M.A. Kelly, J. Vac. Sci. Technol. A 16 (1998)
3483.
[5] T.L. Barr, Modern ESCA, CRC Press, Boca Raton, FL,
1994, Chapter 7.
[6] D.M. Sarno, L.J. Matienzo, W.E. Jones Jr., Inorg. Chem. 40
(2001) 6308.
[7] T. Ogama, J. Vac. Sci. Technol. A 14 (1996) 1309.
[8] (a) S.J.N. Burgmayer, H.L. Kaufmann, G. Fortunato, P. Hug,
B. Fischer, Inorg. Chem. 38 (1999) 2607;
(g) D.L. Lichtenberger, Acc. Chem. Res. 20 (1987) 379;
(h) D.L. Lichtenberger, M.E. Jatcko, Inorg. Chem. 31
(1992) 451.
[19] D.L. Lichtenberger, M.E. Jatcko, J. Coord. Chem. 32 (1994)
79.
[20] (a) D.P. Birnie III, R.N. Vogt, M.N. Orr, J.R. Schifko,
Microelectron. Eng. 29 (1995) 189;
(b) D.P. Birnie III, M. Manley, Phys. Fluids 9 (1997) 870;
(c) D.P. Birnie III, J. Non-Cryst. Solids 218 (1997) 174.
[21] (a) Th. Kugler, A. Andersson, Loglund, A.B. Holmes, X. Li,
W.R. Salaneck, Synth. Met. 100 (1999) 163;
(b) Th. Kugler, W.R. Salaneck, H. Rost, A.B. Holmes,
Chem. Phys. Lett. 310 (1999) 391;
(c) G. Greczynski, M. Falham, W.R. Salaneck, Chem. Phys.
Lett. 321 (2000) 379;
(d) G. Greczynski, Th. Kugler, W.R. Salaneck, Curr. Appl.
Phys. 1 (2001) 98;
(b) J.V. Ortega, B. Hong, S. Ghosal, J.C. Hemminger, B.
Breedlove, C.P. Kubiak, Inorg. Chem. 38 (1999) 5102.
[9] G.M. Bancroft, I. Adams, H. Lampe, T.K. Sham, J. Electron
Spectrosc. Relat. Phenom. 9 (1976) 191.
[10] (a) D.L. Lichtenberger, J.G. Kristofzski, J. Am. Chem. Soc.
109 (1987) 3458;
(b) D.L. Lichtenberger, C.D. Ray, F. Stepniak, Y. Chen, J.H.
Weaver, J. Am. Chem. Soc. 114 (1992) 10492.
[11] K.R. Birdwhistell, Inorg. Synth. 29 (1992) 141.
[12] C. Quaresima, C. Ottaviani, M. Matteucci, C. Crotti, A.
(e) G. Greczynski, Th. Kugler, M. Keil, W. Osikowicz, M.
Falham, W.R. Salaneck, J. Electron Spectrosc. Relat.
Phenom. 121 (2001) 1;
(f) M. Onoda, K. Tada, H. Nakayama, Synth. Met. 102
(1999) 975.
[22] G. Polzonetti, G. Iucci, C. Furlani, M.V. Russo, A. Furlani,
G. Infante, G. Paolucci, B. Brena, D. Cocco, Chem. Phys.
Lett. 267 (1997) 384.
[23] D.M. Riffe, G.K. Wertheim, P.H. Citrin, Phys. Rev. Lett. 63
(1989) 1976.

C. Crotti et al. / Journal of Electron Spectroscopy and Related Phenomena 128 (2003) 141–157
[24] C. Ottaviani, M. Capozi, C. Quaresima, M. Matteucci, C.
157
[26] L.J. Matienzo, S.O. Grim, Anal. Chem. 46 (1974) 2052.
Crotti, P. Perfetti, C. Astaldi, M. Zacchigna, C. Comicioli,
M. Evans, K.C. Prince, J. Electron Spectrosc. Relat. Phenom.
76 (1995) 139.
[27] D.L. Lichtenberger, K.W. Nebesny, C.D. Ray, D.R. Huffman,
L.D. Lamb, Chem. Phys. Lett. 176 (1991) 203.
[28] R. Zanoni, M.N. Piancastelli, J. McKinley, G. Margaritondo,
Appl. Phys. Lett. 55 (1989) 1020.
[25] I. Bernal, G.M. Reisner, G.R. Dobson, C.B. Dobson, Inorg.
Chim. Acta 121 (1986) 199.