Photochemistry of thin amorphous films of Fe(CO)4PPh3 on Si(111) surfaces

Article abstract of DOI:10.1016/S0022-328X(97)00599-8

The photochemical reactions of Fe(CO)₄PPh₃ and Fe(CO)₃(PPh₃)₂ as amorphous films on silicon surfaces are presented. The mechanism of the reaction of Fe(CO)₄PPh₃ in the films has been studied in some detail. Initial CO loss leads to a thermally unstable intermediate, Fe(CO)₃PPh₃, which decomposes in the film leading to the production of iron. In thick films photoproduced CO remains trapped in the film and may react with Fe(CO)₃PPh₃, regenerating the starting material. Further evidence for this intermediate arises from experiments conducted with PPh₃ added to the film. In these films the initial photoproduct is trapped by PPh₃ yielding Fe(CO)₃(PPh₃)₂. A quantitative study of the quantum yield efficiency in these films was undertaken. The Fe(CO)₃(PPh₃)₂ films are photosensitive undergoing CO loss to yield Fe(CO)₂(PPh₃)₂, as demonstrated by its trapping, by PPh₃, to form Fe(CO)₂(PPh₃)₃. Extended photolysis of Fe(CO)₃(PPh₃)₂ films, including those containing photoproduced Fe(CO)₂(PPh₃)₃, results in the formation of iron. The surface photochemistry of both Fe(CO)₄PPh₃, and Fe(CO)₃(PPh₃)₂ are shown to be compatible with standard lithography. Patterns of 1×100 μm lines of iron oxide were easily produced on a silicon surface.

Full text of DOI:10.1016/S0022-328X(97)00599-8

Journal of Organometallic Chemistry 554 Ž1998. 63–73
ž /
ž /
Photochemistry of thin amorphous films of Fe CO PPh on Si 111
4
3
surfaces
Sharon L. Blair, Ross H. Hill )
Department of Chemistry, Simon Fraser UniÕersity, Burnaby, British Columbia, Canada, V5A 1S6
Received 9 January 1997; revised 21 April 1997
Abstract
The photochemical reactions of FeŽCO. PPh and FeŽCO. ŽPPh . as amorphous films on silicon surfaces are presented. The
4
3
3
3 2
mechanism of the reaction of FeŽCO. PPh in the films has been studied in some detail. Initial CO loss leads to a thermally unstable
4
3
intermediate, FeŽCO. PPh , which decomposes in the film leading to the production of iron. In thick films photoproduced CO remains
3
3
trapped in the film and may react with FeŽCO. PPh , regenerating the starting material. Further evidence for this intermediate arises from
3
3
experiments conducted with PPh added to the film. In these films the initial photoproduct is trapped by PPh yielding FeŽCO. ŽPPh . .
3
3
3
3 2
A quantitative study of the quantum yield efficiency in these films was undertaken. The FeŽCO. ŽPPh . films are photosensitive
3
3 2
undergoing CO loss to yield FeŽCO. ŽPPh . , as demonstrated by its trapping, by PPh , to form FeŽCO. ŽPPh . . Extended photolysis of
2
3 2
3
2
3 3
FeŽCO. ŽPPh . films, including those containing photoproduced FeŽCO. ŽPPh . , results in the formation of iron. The surface
3
3 2
2
3 3
photochemistry of both FeŽCO. PPh , and FeŽCO. ŽPPh . are shown to be compatible with standard lithography. Patterns of 1=100
4
3
3
3 2
mm lines of iron oxide were easily produced on a silicon surface. q 1998 Elsevier Science S.A.
Keywords: Photochemistry; Surface chemistry; Thin films
1
. Introduction
photochemically in the solid state to thin films of the
metal or its oxide.
Organometallic complexes have been used as precur-
Previously we have demonstrated the ability to fabri-
cate metal and metal oxide structures via photolysis of
inorganic precursor films through a lithography mask
w17,18x. In this process an amorphous film is cast by
sors to generate thin films of metals and metal oxides
by both chemical and photochemical vapor deposition
pathways w1–9x. The advantage of the photochemical
pathway has been the use of photolysis to direct the
deposition process. This is of particular importance in
applications where specific coating of a pattern is desir-
able, such as in the fabrication of microelectronics.
Using photolysis to direct the deposition allows for the
elimination of a polymer masking step in the fabrication
procedure. An additional advantage of the photochemi-
cal process is the low process temperature. Our own
research has involved the study of inorganic and
organometallic complexes in the solid state as precur-
sors to metal and metal oxide films w10–17x. In these
spin coating. This is followed by the irradiation of the
surface film, normally conducted under vacuum, to
generate the final metal films. The vacuum appears to
be necessary for the production of high quality films as
it aids in the diffusion of the photoejected ligands from
the surface. In this paper we explore the use of mixtures
and quantitative quantum yield information to elucidate
the mechanism of the photochemical reactions which
occur in the amorphous thin film phase.
The complex studied, Fe
a variety of reasons. The photochemistry of Fe
ŽCO.4 PPh , was chosen for
3
Ž
CO.5
studies, films of the precursor complexes are converted
and its substituted derivatives have been examined pre-
viously in the gas phase w19x, in solution w20x, in glasses
w21x, matrices w22,23x and bound to both polymers w24x
and surfaces w25x. This report is, however, the first
)
Corresponding author.
report of its chemistry in an amorphous thin film.
0
022-328Xr98r$19.00 q 1998 Elsevier Science S.A. All rights reserved.

6
4
S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
Precedent suggests that loss of CO, according to Eq.
, is the dominant primary photoprocess for iron
carbonyl complexes.
1
Ž .
h
n
Fe
Ž
CO.4 PPh ™ Fe
Ž
CO.3PPh qCO
1
Ž .
3
3
Reaction in the condensed phase may be expected to
result in the formation of a variety of products resultant
from the thermal chemistry of this intermediate. Clus-
ters with iron carbonyl phosphines are well-known w26–
2
9x.
In this study we have restricted our attention to
ŽCO.4 PPh . The coordination of a large phosphine,
Fe
3
Fig. 2. Ža. FTIR spectra associated with 0.56, 1.12, 1.68, 2.24 and
PPh , renders the vapor pressure of the complex suffi-
ciently low that evaporation of the starting material
during the study is inconsequential. It should be noted
3
2
2
.80 molecules of FeŽCO. ŽPPh . per A˚ on a silicon surface. Žb.
3 3 2
y1
Plot of absorbance of the nŽCO. band at 1877 cm
Fe CO PPh3 versus coverage, data from Fig. 1a.
of
Ž
. Ž
.
2
3
that the PPh ligand will have a negative impact on the
3
purity of the photoproduced films. Our concern in this
paper is, however, directed at the process investigation
rather than the purity of the photoproduced materials.
being allowed. From the single observed band we can
confirm the trans geometry of the phosphine ligands.
The absolute frequencies reported here are consistent
with the spectra reported in solution w31x and in low-
temperature matrices w21x. This similarity in the IR
2
. Results
spectra indicates that there are no strong intermolecular
interactions in the thin amorphous films. It should be
noted that the films studied here are multilayered and
that no effect due to the interface with either the silicon
surface or the vacuum is spectroscopically observed.
The plots of absorbance versus coverage are shown
in Fig. 1b and Fig. 2b. The points on these plots give an
approximate estimate of the error associated with the
measurement of the absorbance. In practice the ab-
sorbance will also depend upon the optical quality of
the film deposited. In the case of the evaporative depo-
sition the quality was variable and we believe resulted
in the large spread in the data. The data is treated as
linear. This observation indicates that the absorbance in
the FTIR may be used as an indication of the amount of
a species in the film. Throughout this work we will use
the absorbance to monitor the extent of reaction in these
systems. Similar plots were made for the additive
molecules, PPh and HCPh . Extinction coefficients at
2
.1. Spectroscopic data for the complexes
The FTIR spectra of the carbonyl regions of
Fe
and Fig. 2a and are summarized in Table 1. The symme-
try of Fe
CO.4 PPh is C and we expect three allowed
transitions associated with the CO stretching modes
2A1 and E . All of these bands are observed. For
Fe
CO. ŽPPh3.2 the trans isomer is expected to have
only one band observed
ŽCO.4 PPh and FeŽCO. ŽPPh3.2 are shown in Fig. 1a
3 3
Ž
3
3n
Ž
.
Ž
3
Ž
E in D3h symmetry while
.
either a C2n or a C structure would lead to three bands
s
3
3
selected wavelengths were calculated from these plots
and are presented in Table 1.
Electronic absorption spectra for amorphous films of
the pure compounds were obtained and the results are
summarized in Table 1. Absorption spectra in the films
are comparable to results obtained in other media w31x.
The extinction coefficients reported in Table 1 were
determined by measurements relative to the FTIR bands
for which less scattering was expected. In this way we
could use optical quality spin-coated films for the deter-
mination of these relative extinction coefficients. A
direct calibration by electronic absorption spectroscopy
Fig. 1. Ža. FTIR spectra associated with 0.21, 0.42, 0.63 and 0.84
2
molecules of FeŽCO. PPh3 per A˚ on a silicon surface. Žb. Plot of
4
y1
absorbance of the 1931 cm
band of FeŽCO. PPh versus cover-
4 3
age, data from Ža..

S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
65
Table 1
FTIR spectral data for relevant complexes on SiŽ111.
y1
Complex
Assignment
Energy wcm
x
Absorbance
Assignment
Energy wnmx
Ž´IR w A˚ moleculey1x.
2
Ž´UV wcm moly1x.
2
per monolayer
6
FeŽCO. PPh
n
2048 Ž0.031.
1973 Ž0.022.
1931 Ž0.073.
1877 Ž0.094.
0.00047
0.00034
0.0011
0.0010
d–d
CT
CT
d–d
d–d
d–d
CT
366 Ž3.78=10 .
4
3
ŽCO.
6
n
268 Ž34.4=10 .
ŽCO.
6
n
232 Ž99.3=10 .
ŽCO.
6
FeŽCO ŽPPh .
n
436 Ž1.12=10 .
3
3
2
ŽCO.
6
366 Ž5.77=10 .
6
348 Ž7.69=10 .
6
252 Ž45.0=10 .
6
CT
230 Ž72.3=10 .
a
3
FeŽCO. ŽPPh .
n
1894
2
3
ŽCO.
n
1838
1433 Ž0.0029.
ŽCO.
y5
PPh3
5.8=10
y6
7
6
42 Ž0.0031.
94 Ž0.0044.
6.2=10
y6
8.7=10
HCPh3
1495 Ž0.0116.
0.00023
a
See Ref. w30x.
is precluded by problems associated with light scatter-
ing at shorter wavelengths for poor quality films.
of carbon than phosphorus from phosphines. This sug-
gests that a significant amount of the carbon found is
associated with the phosphine in the film. Further evi-
dence for PPh being retained in the film arises from
the depth profiling experiments. In these experiments
2
.2. Photochemistry of Fe(CO) PPh3
3
4
the surface of the film contains less PPh as indicated
3
The spectroscopic changes resultant from photolysis
by the observation of 1.6% phosphorous. For longer
sputtering times the interior of the film is exposed and
Ž .
higher amounts of phosphorus 3.4% are evident. This
of thin films of Fe
ŽCO.4 PPh are shown in Fig. 3. A
3
decrease in intensity to baseline of all FTIR bands
associated with the carbonyl stretching vibrations at
y1
result is indicative of efficient loss of PPh from the
3
2
048, 1973 and 1931 cm is observed. This indicates
surface of the film. In the interior of the film, where
the loss of all starting material. No bands associated
with a CO containing photoproduct are apparent. Each
band decreases in intensity proportional to its original
intensity hence no direct evidence is found for the
formation of an intermediate. This is consistent with an
initial primary photoprocess yielding a thermally unsta-
ble intermediate. This intermediate then undergoes rapid
thermal decomposition. This evidence is consistent with
diffusion is required prior to loss of the PPh , more
3
phosphorus is retained.
Previous work w21x in other media indicated that the
expected primary photoproduct is Fe CO PPh . This
Ž .
3 3
coordinatively unsaturated system is known to be sus-
Eq.
thermally unstable Fe
decomposition as shown in Eq.
Ž
1
.
as the initial photochemical step producing
Ž
CO.3PPh which undergoes rapid
3
2 .
Ž .
Fe CO.3PPh ™Feq3COqPPh
Ž
2
Ž .
3
3
FTIR spectroscopic data following photolysis of the
compound indicates loss of all starting material. The
resultant films were analyzed by Auger electron spec-
troscopy to determine if the ejected ligands diffused out
of the film. The analysis of films produced from
Fe
PPh from the film leaving a film composed primarily
ŽCO.4 PPh are consistent with the loss of CO and
3
3
of iron, oxygen, and some residual PPh3 Table 2 . The
Ž .
surface layer is oxidized, as expected for iron, but depth
profiling with argon ions indicates a decreasing oxide
content as the film is sputtered. Independent experi-
ments indicate that sputtering causes more efficient loss
Fig. 3. FTIR spectral changes associated with the photolysis for 0,
6
0, 120, 210, 270, 330, 600 and 1365 min of a 134-monolayer film
of FeŽCO. PPh deposited on SiŽ111..
4
3

6
6
S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
Table 2
Properties of the films: Composition of initial films by FTIR, thickness of the initial films and composition of the final films by Auger analysis
Initial film
composition
Initial film
thickness Žnm.
Sputter
time Žmin.
% Fe
% P
% C
% O
FeŽCO. PPh
273
0
3
0.25
1
14.1
16.9
10.3
8.6
1.6
3.4
1.9
5.5
68.3
71.8
77.8
77.9
16.1
7.8
7.7
8.0
4
3
FeŽCO. ŽPPh .
3 2
34
3
ceptible to attack by two electron donor ligands, L, to
yield Fe
CO.3PPh L w21,32x.
In order to investigate if Fe
the thin film photoreaction, a film of codeposited
Fe
CO.4 PPh and PPh was studied. This reaction was
done with the aim of trapping the intermediate,
Fe
CO.3PPh , with PPh to form Fe CO. ŽPPh3.2. Spin
coating a chip with a solution containing Fe
amount of an additional band due to Fe
Ž
CO. ŽPPh3.3 is
2
y1
Ž
observed at 1838 cm
species at 1894 cm
Ž
the other band due to this
is obscured . The formation of
ŽCO. ŽPPh3.3 is associated with the subsequent pho-
toreactions of trans-Fe
ŽCO. ŽPPh3.2 and is treated sep-
3
3
y1
Ž
CO.3PPh was formed in
.
3
Fe
2
Ž
3
3
arately below.
Ž
Ž
3
3
3
Fe
Ž
CO.3PPh qPPh ™Fe
Ž
CO. ŽPPh3 .2
3
Ž .
Ž
CO.4 PPh3
3
3
3
and PPh results in the formation of stable amorphous
3
In order to assess the efficiency of the trapping
reaction of Eq. in competition with the decomposi-
tion of Eq. , studies were done in which the amount
films composed of the compounds. In Fig. 4 the spec-
troscopic changes resultant from photolysis of a film
3
Ž .
2
Ž .
composed of both PPh and Fe
ŽCO.4 PPh are shown.
3
3
of PPh within the film was varied. The efficiency of
3
The loss of carbonyl bands associated with
trapping can be expressed as the ratio of trans-
Fe
Ž
CO.4 PPh is observed and is accompanied by the
3
Fe
sults of several experiments indicated that when no
PPh is added to the film no trans-Fe
CO. ŽPPh3.2 is
ŽCO. ŽPPh3.2 formed to FeŽCO.4 PPh lost. The re-
y1
3
3
growth of a new absorption at 1877 cm . This band
increases in intensity, reaches a maximum and then
decays in intensity as a result of further photolysis. The
transient band at 1877 cm is assigned as due to the
production of trans-Fe
Ž
3
3
y1
observed. The exception to this occurs when thick films
are monitored at long photolysis times. In these in-
ŽCO. ŽPPh3.2. This assignment
3
stances the photoproduced PPh is sufficient to produce
3
has been confirmed by comparison with the absorption
associated with an authentic sample. This assignment is
consistent with the primary photoprocess being loss of
detectable amounts of trans-Fe
with PPh3 added the initial ratio of trans-
Fe
CO. ŽPPh3.2 formed to Fe CO.4 PPh lost is within
error of 1 regardless of the concentration of PPh within
ŽCO. ŽPPh3.2. For films
3
Ž
Ž
3
3
the axial CO, forming the Fe
in Eq. . This intermediate is trapped by PPh within
the film, according to Eq.
Fe
CO. ŽPPh3.2 , in competition with the decomposition
reaction of Eq.
ŽCO.3PPh intermediate as
3
3
1
Ž .
3
the film; as long as some trans-Fe
observed. Due to the photosensitivity of trans-
Fe
CO. ŽPPh3.2 this ratio decreases as the photolysis
continues. The photoreaction of the trans-
Fe
CO. ŽPPh3.2 ultimately results in a low steady state
concentration of trans-Fe
ŽCO. ŽPPh3.2 is
3
3 to form trans-
Ž .
Ž
3
Ž
3
2 . It is also apparent that a small
Ž .
Ž
3
Ž
CO. ŽPPh3.2
.
3
2
.3. Quantum yield studies on Fe(CO) PPh
4 3
The variation in quantum yields with reaction condi-
tions were studied. The effect of both film thickness and
additives on the quantum yield for disappearance of
Fe
ness on the quantum yield will be presented.
For a relatively thick film of Fe
CO.4 PPh , irradia-
ŽCO.4 PPh was studied. First the effect of film thick-
3
Ž
3
tion led to loss of absorption bands associated with the
starting material and no intermediate was evident, Fig.
3
. A plot of absorbance versus irradiation time was fit
Fig. 4. FTIR spectral changes associated with the photolysis for 0, 5,
to an exponential decay as shown in Fig. 5. The decay
constant can be converted to quantum efficiency and
was found to be
1
0, 15, 20, 30, and 75 min of a 27-monolayer film of FeŽCO. PPh
4 3
codeposited with PPh3 deposited on SiŽ111.. Note several additional
y4
spectra were recorded for this experiment.
Ž
3.8"0.5
.
=10
.

S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
67
3
indicates an order of magnitude increase in the quan-
tum yield as a result of conducting the photochemistry
under an air atmosphere.
The quantum yield for reaction of the films contain-
ing both Fe
from spectral changes upon photolysis. Control experi-
ments indicate that films containing both Fe
CO.4 PPh3
ŽCO.4 PPh3 and PPh3 can be determined
Ž
and PPh3 are thermally sensitive. This complication
allows only a qualitative assessment of the effect of
additives on the quantum yield for decomposition. The
spectral changes presented in Fig. 4 for a film contain-
ing 27 monolayers Fe
PPh were used to determine the quantum yield. The
ŽCO.4 PPh3 and 48 monolayers
3
quantum yield for the loss of Fe
ŽCO.4 PPh3 in films
composed in part of PPh can be calculated from an
3
exponential fit of A versus irradiation time as shown in
t
Fig. 6. The decay constant, in conjunction with light
intensity and extinction coefficient yields the quantum
yield as 0.047"0.02. This value is significantly larger
than the quantum yield in the absence of the PPh3
y1
Fig. 5. Plot of absorbance at 1931 cm
versus photolysis time for a
1
34-monolayer film of FeŽCO. PPh , the data was derived from
4
3
Fig. 3.
additive. The combination of both air and PPh results
3
in a similar increase in the quantum yield as shown in
Table 3.
The dependence of the photoefficiency of
CO.4 PPh3 on film thickness was studied in some
Fe
Ž
When an inert substance such as HCPh is added to
3
the film the photoproduct is unchanged and the quantum
yield is not significantly altered as indicated in Table 3.
Before considering the detailed interpretation of these
detail. Films of varying thickness were prepared and the
photochemistry monitored for each film. The quantum
yields are reported in Table 3. The general trend is that
as the film thickness increases the quantum yield de-
creases. A detailed interpretation of this result is pre-
sented in Section 3. Before considering the interpreta-
tion of these quantum yields the influence of additives
on the quantum yield will be presented.
results the photochemistry of Fe
presented in Section 2.4.
ŽCO. ŽPPh3.2 will be
3
2
.4. Photochemistry of Fe(CO) (PPh )
3 3 2
The effect of air on the reaction efficiency was
investigated. When films were monitored in an air
atmosphere, the spectral changes were similar to those
observed in a vacuum. The quantum efficiency of the
reaction in air was much higher than for films of
comparable thickness under vacuum. The data in Table
The observation of photoproduced trans-
ŽCO. ŽPPh3.2 resultant from the photolysis of a film
composed of both PPh3 and Fe
firmed by comparison of the FTIR spectrum obtained
with that due to an authentic sample. Prolonged photol-
Fe
3
Ž
CO.4 PPh3 was con-
Table 3
Quantum yields for decomposition of FeŽCO. PPh in thin films
4
3
Coverage
wmoleculer A˚ y2 x Žmonolayers.
Ao
Quantum yield
Žerror.a
Additives
Coverageb
wmolecules A˚ y2 x Žmonolayers.
Atmosphere
0
0
0
1
2
0
1
0
1
1
.10 Ž7.
0.0076
0.0362
0.0484
0.0862
0.147
0.0707
0.0817
0.0293
0.0984
0.0865
0.0032 Ž0.001.
0.0016 Ž0.0002.
0.0023 Ž0.0008.
0.0007 Ž0.0001.
0.00038 Ž0.00005.
0.0043 Ž0.004.
0.011 Ž0.002.
Vacuum
Vacuum
Vacuum
Vacuum
Vacuum
Air
Air
Vacuum
Air
.50 Ž33.
.66 Ž44.
.18 Ž78.
.01 Ž134.
.97 Ž64.
.12 Ž74.
.4 Ž27.
c
0.047 Ž0.02.
PPh3
PPh3
HCPh3
9.8 Ž48.
c
.35 Ž89.
.18 Ž79.
0.048 Ž0.01.
7.9 Ž39.
0.0009 Ž0.0001.
0.395 Ž20.
Vacuum
a
Error calculated from exponential decay fit.
calculated from absorbance.
Error high due to competing thermal reaction.
b
c

6
8
S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
ysis of the film composed of both PPh3 and
Fe
CO.4 PPh led to loss of all peaks including those
associated with the trans-Fe
product. It was therefore necessary to establish that
samples of trans-Fe
CO. ŽPPh3.2 were photochemi-
cally active. Films of this complex exhibit a single
Ž
3
Ž
CO. ŽPPh3.2 photo-
3
Ž
3
y1
n
Ž
CO
.
at 1877 cm . Photolysis of a thin film of
trans-Fe
Ž
CO. ŽPPh3.2 results in the decrease in inten-
3
sity of the single CO band as illustrated in Fig. 7. No
new absorption bands associated with a possible inter-
mediate species are observed. The overall reaction initi-
ated by photolysis occurs as in Eq. 4 . Further evidence
Ž .
for the loss of all ligands from the complex comes from
the Auger analysis which is consistent with the loss of
all ligands yielding a somewhat impure
surface.
Fig. 7. FTIR spectral changes associated with the photolysis for 0, 2,
5, 30, 60, 120 and 245 min of 17-monolayer film of
FeŽCO. ŽPPh . .
1
a
Table 2 iron
Ž .
3
3 2
h
n
Fe
Ž
CO. ŽPPh3 .2 ™ Feq3COq2PPh3
4
Ž .
3
nism, involving CO loss, the presence of uncoordinated
PPh3 in the film would be expected to result in the
One interpretation is that photolysis of trans-
CO. ŽPPh3.2 results in CO loss leading to the ther-
Fe
Ž
Ž
. Ž .
and the production of
3
trapping of Fe CO PPh3
2
2
mally unstable species Fe
Ž
CO. ŽPPh3.2. This would be
Ž . Ž
.
2
Fe CO PPh3 . The presence of PPh would not be
2
3
3
consistent with the absence of any observable iron
carbonyl photoproduct. The thermal decomposition of
expected to alter the products of the phosphine loss
mechanism. However, trapping of Fe CO PPh in this
Ž
.
3
3
Fe
and PPh . An alternative reactivity would be the loss of
Ž
CO. ŽPPh3.2 would lead to iron, carbon monoxide
mechanism would regenerate trans-Fe CO PPh3 re-
. Ž
Ž
.
2
3
2
3
sulting in a decrease in the observed quantum yield for
decomposition.
the PPh ligand from trans-Fe
the unstable Fe
scenarios would be expected to be affected differently
ŽCO. ŽPPh3.2 leading to
3
3
Ž
CO.3PPh discussed above. These two
3
In order to investigate the primary photoreaction of
trans-Fe CO PPh3 thin films with PPh codeposited
Ž
. Ž
.
3
2
3
by the addition of PPh to the film. In the first mecha-
3
were prepared and studied. Photolysis of these films
would be expected to yield Fe
CO. ŽPPh3.3 if the
Ž
2
initial photoreaction was CO loss. Irradiation results in
y1
the loss of intensity at 1877 cm associated with the
single absorption band of trans-Fe
Ž
CO. ŽPPh3.2. Con-
3
y1
comitant growth of a band at 1838 cm
is observed.
This band is attributed to the formation of
Fe
Fe
Ž
CO. ŽPPh3
.
. The second band associated with
2
3
y1
Ž
CO. ŽPPh3
.
, expected at 1894 cm
w30x, is ob-
2
3
y1
scured by the band at 1877 cm associated with the
starting material. The steady-state concentration of this
species was quite small. The production of
Fe
figure the continued photolysis of the film composed of
both PPh and Fe
CO.4 PPh led to the production of
ŽCO. ŽPPh3.3 is also apparent from Fig. 4. In this
2
Ž
3
3
trans-Fe
product led to the formation of Fe
evidenced by the production of a band at 1838 cm
The observation of Fe
CO. ŽPPh3.3 production from
photolysis of Fe
CO. ŽPPh3
reaction pathway of Fe
For this reason we favor the interpretation that photoly-
sis of Fe
CO. ŽPPh3.2 results in the loss of CO yielding
Fe
CO. ŽPPh3.2 according to Eq.
tively unsaturated intermediate may either decompose
Eq.
.. or, in the presence of PPh , react to form
Fe
CO. ŽPPh3.3 ŽEq.
ŽCO. ŽPPh3.2. The photolysis of this photo-
3
Ž
CO. ŽPPh3
.
as
3
.
2
y1
Ž
2
Ž
.
confirms that at least one
3
2
Ž
CO. ŽPPh3.2 must be CO loss.
3
Ž
3
Ž
5 . The coordina-
Ž .
y1
2
Fig. 6. Plot of the absorbance at 1931 cm
versus photolysis time
for a 27-monolayer film of FeŽCO. PPh3 codeposited with PPh3.
4
Ž
Ž
6
The data is derived from the experiment of Fig. 4, although data left
out of Fig. 4 is included.
3
Ž
Ž ... The FeŽCO. ŽPPh3.3 photo-
7
2
2

S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
69
Fig. 9. Optical micrograph of surface obtained by the photolithogra-
phy of FeŽCO. PPh on a siliconŽ111. surface. An internal scale is
4
3
provided by the 50-mm length of the lines.
Fe
according to Eq.
reversal of Eq. 5
Ž .
ŽCO. ŽPPh3.2. This would result from the trapping
3
Ž
7
.
which would prevent any thermal
. Quantum yields for the photodecom-
. Ž
y1
Fig. 8. Plot of the absorbance at 1877 cm
versus photolysis time
Ž
.
position of Fe CO PPh3
in films both with and
3
2
for a 14-monolayer film of FeŽCO. ŽPPh . , the data is derived from
3
3 2
without PPh were measured in order to further investi-
3
the experiment of Fig. 7.
gate these possibilities.
h
n
Fe
Ž
CO. ŽPPh3 .2 ™ Fe
Ž
CO. ŽPPh3
.
qPPh3
.5. Quantum yield studies on Fe(CO) (PPh )
3 2
8
Ž .
3
3
product is known to have a labile PPh ligand w33x. If it
3
lost a PPh it would likely decompose.
3
2
3
h
n
Fe
Ž
CO. ŽPPh3 .2 ™ Fe
Ž
CO. ŽPPh3 .2 qCO
Ž
5
.
3
2
The quantum yields for the photoreaction of
ŽCO. ŽPPh3.2 were also determined by monitoring
Fe
Fe
Ž
Ž
CO. ŽPPh3 .2 ™Feq2COq2PPh3
Ž
Ž
6
7
.
.
Fe
2
3
the decay of starting material as a function of photolysis
CO. ŽPPh3 .2 qPPh ™Fe
Ž
CO. ŽPPh3 .3
2
3
2
y1
time. In Fig. 8 the decay of the 1877 cm absorbance
The alternative dissociative photoreaction of
CO. ŽPPh3.2 is shown in Eq.
. This would not
band of Fe
ŽCO. ŽPPh3.2 is plotted against photolysis
3
Fe
Ž
Ž
8
.
time and fit to an exponential decay. The resultant
quantum yield is 0.004"0.001. Similar experiments
were conducted on samples with additives and thicker
films. The results are summarized in Table 4.
3
give rise to independently detectable products and so we
cannot rule out this pathway. However, if Eq.
adequately describes the photochemistry of this system
then we expect that the photoefficiency of the reaction
8
Ž .
The results indicate that a large increase in the
of Fe
Ž
CO. ŽPPh3.2 in films containing PPh should be
quantum yield is associated with the addition of PPh to
3
3
3
lower than of those containing no PPh . This is ex-
the film. In this system we find no quantum yield
variation associated with the film thickness. Table 4
shows that films which range in thickness from 13 to 75
nm have the same quantum yield. This result supports
3
pected to result from trapping the photogenerated
Fe
quence of reactions shown in Eqs.
ŽCO.3PPh3 intermediate. Alternatively if the se-
5 – 7 is dominant,
Ž . Ž .
then the addition of PPh would be expected to result in
the contention that Eqs.
reaction pathway.
Ž
5
.
–
Ž
7
.
represent the major
3
an increase in the quantum yield for decomposition of
Table 4
Quantum yields for reaction of FeŽCO. ŽPPh . in thin films
3
3 2
y2
Quantum yielda
b
y2
Coverage wmolecules A˚
Žmonolayers.
x
Ao
Additives
Coverage wmolecules A˚
Žmonolayers.
x
Atmosphere
0
0
0
0
.16 Ž15.
.84 Ž79.
.36 Ž34.
.28 Ž26.
0.0147
0.0788
0.0337
0.0262
0.004 Ž0.001.
0.006 Ž0.001.
Vacuum
Vacuum
Vacuum
Air
c
0.07 Ž0.04 .
PPh3
PPh3
3.0Ž14.
4.3Ž21.
c
0.012 Ž0.007 .
a
Error calculated from fit.
Calculated from absorbance.
Error high due to competing thermal reaction.
b
c

7
0
S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
2
.6. Lithography
generate Fe
leads to a similar set of reactions. Photoejection of CO
to form Fe
CO. ŽPPh3.2 may be followed by decompo-
ŽCO. ŽPPh3.2. Photolysis of FeŽCO. ŽPPh3.2
3
3
This surface photochemistry is compatible with
Ž
2
lithography allowing us to pattern surfaces. This was
demonstrated by irradiation of a surface film of
sition, reaction with photogenerated CO to reform the
starting material, or trapping by PPh3 to yield
Fe
Ž
CO.4 PPh through a lithography mask. The mask
Fe
Ž
CO. ŽPPh 3 .3
.
Finally, the photolysis of
3
2
was set up in contact with the surface negating the need
for optics. The result of this procedure is shown in Fig.
Fe
Ž
Fe
Ž
CO. ŽPPh3.3 results in CO loss to produce
2
CO.ŽPPh3.3. This photoproduct is unstable and may
9
. Clear formation of lines 1 mm by 50 mm are visible.
either decompose or react with photogenerated CO.
While this interpretation of the reaction mechanism is
relatively straightforward consideration of the efficiency
of the reaction yields further intricate details.
3
. Discussion
The quantum yield for this process has been mea-
sured for a variety of film thickness and is presented in
Table 3. Thin films have a higher quantum yield for
decomposition. The simplest interpretation of this data
is related to the reaction of uncoordinated CO with the
The photolysis of thin amorphous films of
Fe
ŽCO.4 PPh on silicon wafers results in the decompo-
3
sition of this species with no observed intermediates.
This result is somewhat surprising in view of the known
photochemistry of iron carbonyl under a variety of
conditions. For example, low-temperature matrix studies
indicate that the formation of products associated with
the loss of a single ligand are readily trapped at low
temperature w23x. While this result in not directly trans-
ferable to the film environments, related studies in
polymers indicate that such intermediates are stable
under conditions where no diffusion motion is allowed
and there are no opportunities for reaction. In this case,
one may speculate that the formation of iron–iron bonds
may provide the mechanism for decomposition although
it is not clear why no bands associated with dimers or
clusters are observed.
unsaturated intermediate, Fe
ate was demonstrated above to be sufficiently stable to
react with PPh3 yielding Fe
CO. ŽPPh3.2. It is hence
reasonable to expect that the photoproduced
Fe
CO.3PPh within the film may be able to also trap
the photogenerated CO according to Eq. 9 . Thicker
Ž .
ŽCO.3PPh . The intermedi-
3
Ž
3
Ž
3
films would hence be expected to retain more CO
leading to more efficient trapping and an overall lower
quantum yield. The quantum yield for decomposition,
D
F , may be expressed as a function of the quantum
0
yield for loss of CO, F , times the fraction which
thermally decomposes prior to trapping by CO as shown
Ž .
in Eq. 10 .
The evidence presented here supports a mechanism
summarized in Scheme 1. Initial photolysis of
Fe
Ž
CO.3PPh qCO™Fe
Ž
CO.4 qPPh3
Ž
9
.
.
3
Fe
Fe
Ž
CO.4 PPh3 results in CO loss. The product,
CO PPh , is not sufficiently stable to be observed
k2
D
0
F sF
Ž
10
Ž
.
3
3
k1 CO qk2
A plot of
spectroscopically under our conditions. This species
may decompose to iron, releasing all ligands, back react
with the photogenerated CO or, in the presence of PPh3,
Ž
F D.y1 versus wCOx is expected to yield a
straight line. In our case the source of CO is the
Scheme 1. Proposed reaction sequence for photolysis of FeŽCO. PPh in multilayer films.
4
3

S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
71
reaction itself. For thick films the quantum yield de-
creases with time, indicating a buildup of CO in the
film. For thin films this does not occur indicating that
CO diffuses out of the film and does not attain a
sufficient concentration to alter the quantum yield. No
direct measure of wCOx within the film is available
The reaction of PPh with the unsaturated intermedi-
3
ate should lead to an increase in the quantum yield. The
quantum yield variation resultant from a change in film
composition is presented in Table 3. This data clearly
indicates that the introduction of PPh3 in the film
results in an increase in the quantum yield. Unfortu-
nately the increase in thermal reactivity precludes a
quantitative study of the quantum yield. This system is
more complex to model because a change in the compo-
sition will affect not only the physical properties of the
film but also the chemistry. The qualitative result does,
however, indicate an increase in quantum yield associ-
ated with the addition of PPh3.
although an estimate for the dependence of film thick-
ness may be made.
Assuming the rate of diffusion of CO within a film is
constant then the CO concentration will be a function of
the thickness of the film. In the simplest case we can
think of the effusion of the gas from the film being
affected by the thickness. Under these conditions we
assume the proportion of CO retained, and hence wCOx
These results are consistent with the set of reactions
shown within Scheme 1. Photoinduced loss of the axial
to be related directly to film thickness. Under these
qualitative assumptions we have Eq.
Ž
11
.
where L is the
CO results in the formation of Fe
ŽCO.3PPh . This
3
thickness in monolayers of the film and c is the propor-
tionality constant.
species is thermally unstable and undergoes rapid de-
composition yielding iron films. Some recoordination of
CO occurs in thick films. In the presence of a coordinat-
1
F D
k1cL
1
s
q F
Ž
11
.
ing ligand, such as PPh , the intermediate is stabilized
3
0
0
F k
2
by the coordination of the added ligand. In the case of
PPh , the product of association is Fe .
ŽCO. ŽPPh3.2
If the steady-state concentration of CO within the
film is proportional to the film thickness then a plot of
3
3
Two key pieces of evidence must be considered at
this juncture. First, the qualitative trend established that
D
y1
Ž
F
.
versus film thickness is expected to be linear.
the quantum yield increased with increasing PPh con-
This plot is presented in Fig. 10 and the linearity is
3
y1
tent in the film. However, the addition of PPh resulted
surprising. The slope of this line, 22"2 nm , has no
3
in a 100% trapping efficiency of the initial photo-
obvious significance since we have no way to relate the
thickness to the absolute concentration of CO. It is
important to note that the intercept, 69.46"124, is
within the expected range. For an initial quantum yield
product, Fe CO.3PPh , generating Fe
Ž ŽCO. ŽPPh3.2. The
3
3
trapping efficiency did not increase with increasing
PPh content in the film. The observation of this behav-
3
ior with codeposited HCPh is particularly informative.
of 1 we expect an intercept of 1
Ž
well within our error
.
.
3
This additive is not expected to coordinate and lead to
no evident trapping of an intermediate. The addition of
HCPh3 did not result in an increase in the quantum
yield for the reaction. This latter fact allows us to rule
out an effect due to the concentration of the starting
material within the film. In view of these two factors we
can also draw some conclusions regarding the trapping
The found intercept indicates a quantum yield for CO
loss of 0.014 although a large error is associated with
this. The lower limit for this yield is 0.005. The data
does not establish an upper limit for the quantum yield.
of Fe
At low concentrations of PPh3 in the film the
CO.3PPh is effectively trapped in competition with
ŽCO.3PPh by PPh .
3 3
Fe
decomposition. This trapping results in an effective
chemical yield of 100% for trans-Fe
CO. ŽPPh3.2 for-
Ž
3
Ž
3
mation. When the amount of PPh in the film is in-
3
creased the quantum yield increases with no change in
the chemical yield. This is consistent with the PPh3
effectively competing with the CO recombination reac-
tion. This competition results in the increase in the
overall quantum yield for reaction.
4
. Conclusions
Photochemical studies of amorphous films of
Fe
on film thickness. The primary photoreaction, loss of
Ž
CO.4 PPh showed a dependence of quantum yield
Fig. 10. Plot of the reciprocal of the quantum yield for decomposition
3
of FeŽCO. PPh , in thin films versus film thickness.
4
3

7
2
S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
CO, is irreversible in thin films resulting in loss of all
ligands. In thick films, the ejected CO does not diffuse
through the film as efficiently and, as a result reacts
a silicon wafer. The solvent was allowed to evaporate
and the FTIR spectrum obtained. This process was
repeated several times yielding the spectra shown in
Fig. 1a. The area over which the film was deposited was
with the Fe
ŽCO.3PPh intermediate. This combination
3
2
results in the observation of lower quantum yields for
thicker films. When the initial film is a mixture of
measured to be 0.79 cm . Each drop of the solution
2
corresponds to a coverage of 0.21 molecules per A˚ .
Fe
PPh forming Fe
Ž
CO.4 PPh3 and PPh3 the intermediate reacts with
The corresponding calibration curve of absorbance ver-
2
y1
Ž
CO. ŽPPh3.2 which is photosensitive.
sus molecules per A˚ for the absorbance at 1931 cm
3
3
Independent photochemical studies of pure
Fe
CO. ŽPPh3.2 resulted in loss of all ligands with no
observable intermediate. This system was demonstrated
to be compatible with current lithography techniques.
However, Auger analysis indicated that the resultant
films were contaminated with phosphorus. This was
is shown in Fig. 1b. The slope of this line yields an
2
y1
Ž
absorbance of 0.073
Ž .
8.1 A˚
ing a molecular volume of approximately , this
"10%
Ž .
A˚ molecule . Assum-
3
3
corresponds to an absorbance of approximately 0.0011
y1
per monolayer for the absorption at 1931 cm . The
corresponding extinction coefficients of the other bands
were obtained from their relative intensity and are pre-
sented in Table 1.
expected, due to the low volatility of PPh . For practi-
3
cal applications the incorporation of a more volatile
ligand, such as in Fe
Ž
CO.4 PEt would be preferable.
The same treatment of a solution prepared from
3
Fe
ml
ŽCO. ŽPPh3.2 Ž0.0055 g. Ž
in methylene chloride 5.0
resulted in each drop contributing 0.56 molecules
to the film. The spectra for sequential drops are
3
.
y2
5
. Experimental
A˚
illustrated in Fig. 1a and the resultant calibration curve
The silicon wafers were obtained from Pacific Mi-
croelectronics Center, Canada. The Si 111 surface was
used in these studies and the wafers were p-type silicon
with tolerances and specifications as per SEMI Standard
M1.1.STD.5 cut to the approximate dimensions of 1
is shown in Fig. 2b. An absorbance of 0.094
Ž
"14%
.
2
y1
Ž
.
A˚ molecule
is found for the single absorbance at
y1
1877 cm . The molecular volume of this molecule is
3
Ž .
9.5 A˚
and the absorbance is 0.0010
approximately
per monolayer.
cm=1.5 cm in house. The CaF crystals were obtained
The electronic absorption spectra were obtained by
2
from Wilmad Glass.
spin-coating samples onto CaF plates and calibrated by
2
The FTIR spectra were obtained with 4 cmy1 resolu-
tion using a Bomem Michelson 120 FTIR spectro-
photometer. The samples were held in an anodized
aluminum sample mount within a NaCl faced vacuum
chamber. Electronic absorption spectra were obtained
with a HP 8452A diode array spectrometer.
comparison with FTIR obtained on these samples. The
energies of the electronic transitions are obtained di-
rectly while the extinction coefficients were obtained
relative to the calibrated intensities of the infrared tran-
sitions. Energies and extinction coefficients for both
compounds are reported in Table 1.
The photolysis beam was a 100 W high-pressure Hg
lamp in an Oriel housing equipped with condenser
lenses and filtered through a 10 cm water filter with
5.2. Photolysis of complexes as films on silicon surfaces
quartz optics. Additional filters
nm to isolate the 366 emission line were used.
Ž
band pass 350–400
All photolysis experiments were done in the same
.
manner. For example, a Si
with Fe
CO.4 PPh3 by dispensing a single drop of a
CH Cl solution containing the iron complex onto a
111 surface was prepared
Ž .
Auger spectra were obtained using a PHI double pass
CMA at 0.85 eV resolution at the Surface Physics
Laboratory, Department of Physics, Simon Fraser Uni-
versity. Film thickness was determined using a Leitz
Laborlux 12 ME S optical microscope with an interfer-
ence attachment.
Ž
2
2
spinning silicon chip. This resulted in the formation of a
uniform, amorphous coating of the chip. The chip was
then transferred to a vacuum Dewar. The system was
placed under vacuum and the FTIR spectrum obtained.
The sample was then irradiated for 60 min and another
FTIR spectrum was obtained. This procedure was re-
Ž .
peated for the following accumulated photolysis times:
120, 210, 270, 330, 600 and 1365 min. The spectra are
presented in Fig. 3.
The chromium plated quartz lithography mask was
kindly donated by IBM.
The complexes, Fe
were prepared by literature procedures w31,34x.
Ž
CO
.
PPh , Fe
Ž
CO. ŽPPh3.2
,
4
3
3
5
.1. Calibration of absorption on the surface
The intensity of the light source was measured with
an International Light IL1350 Radiometer and the pho-
y9
y1
y2
A stock solution of Fe
Ž
CO.4 PPh3
Ž
0.0134 g
.
was
ton output at 366 nm was 9.45=10
The output of the light source was found to be constant
within error 10% by periodic measurement of intensity
throughout the experiments.
E s
cm
.
prepared in CH Cl
diluted by a factor of 10 with CH Cl . A drop
ml
Ž
5 ml
.
. This solution was then
0.0045
.
of this solution was then deposited on the surface of
2
2
Ž
Ž
.
2
2

S.L. Blair, R.H. HillrJournal of Organometallic Chemistry 554 (1998) 63–73
73
w4x N.H. Dryden, R. Kumar, E. Ou, M. Rashidi, S. Roy, P.R.
Norton, R.J. Puddephatt, Chem. Mater. 3 Ž1991. 677.
w5x H.H. Gilgen, T. Cacouris, P.S. Shaw, R.R. Krchnavek, R.M.
Osgood, Appl. Phys. B 42 Ž1987. 55.
The quantum yields were calculated as described
previously w12,17x. In cases where curvature was evi-
dent in the plots only the first 10% was used to calcu-
late the quantum yield.
w6x D.K. Liu, R.J. Chin, A.L. Lai, Chem. Mater. 3 Ž1991. 13.
w7x J.S. Foord, R.B. Jackman, Chem. Phys. Lett. 112 Ž1984. 190.
w8x V. Borker, K.S. Rane, V.N. Kamat Dalal, J. Mater. Sci. 4
Ž1993. 241.
The sample was then moved to the Scanning Auger
Microprobe for elemental analysis of the surface. The
results are presented in Table 2.
In some experiments PPh3 was codeposited with
Fe CO.4 PPh3 from a single solution containing both
Ž
w9x G. Frolov, V.S. Zhigalov, S.M. Zharkov, I.R. Yarullin, Phys.
Solid State 36 1994 526.
Ž
.
w10x B.J. Palmer, A. Becalska, R.H. Hill, J. Photochem. Photobiol.
components. The result of photolysis of this sample is
shown in Fig. 4. Spectral results for the photolysis of
A: Chem. 57 Ž1991. 457.
w11x B.J. Palmer, A. Becalska, R. Hader, R.H. Hill, Polyhedron 10
Ž1991. 877.
Fe
ŽCO. ŽPPh3.2 are illustrated in Fig. 8.
3
w12x A. Becalska, R.J. Batchelor, F.W.B. Einstein, R.H. Hill, B.J.
Palmer, Inorg. Chem. 31 Ž1992. 3118.
5
.3. Lithography
w13x D.G. Bickley, R.H. Hill, C.I. Horvath, J. Photochem. Photobiol.
A: Chem. 67 Ž1992. 181.
w14x T.W.H. Ho, S.L. Blair, R.H. Hill, D.G. Bickley, J. Photochem.
A film containing Fe CO.4 PPh was spin-coated onto
Ž
3
Photobiol. A: Chem. 69 Ž1992. 229.
5 B.J. Palmer, R.H. Hill, J. Photochem. Photobiol. A: Chem. 72
a silicon
Ž
111
.
surface, as described in Section 5.2. The
w1
x
thickness of the film was measured by optical interfer-
ometry and found to be 137 nm. The sample was placed
in contact with a lithography mask and irradiated through
a 10 cm water filter and the mask using a 150 W Xe
high pressure lamp for 36 h. The sample was removed
and rinsed with methylene chloride to dissolve the
unexposed starting material. The image presented in
Fig. 7 was obtained with a Leitz Laborlux ME 2 S
optical microscope.
Ž1993. 243.
w16x S.L. Blair, W. Xia, R.H. Hill, J. Photochem. Photobiol. A:
Chem. 81 Ž1994. 183.
w17x B.J. Palmer, A. Becalska, T.W.H. Ho, R.H. Hill, J. Mater. Sci.
2
8 Ž1993. 6013.
w1 x
8 S.L. Blair, J. Hutchins, R.H. Hill, D.G. Bickley, J. Mater. Sci.
29 Ž1994. 2143.
w19x J.T. Yardley, B. Gitlin, G. Nathanson, A.M. Rosan, J. Chem.
Phys. 74 Ž1981. 370.
w20x G.L. Geoffroy, M.S. Wrighton, Organometallic Photochemistry,
Academic Press, New York, 1979.
w21x D.K. Liu, C. Brinkley, M.S. Wrighton, Organometallics 3 Ž1984.
1
449.
w22x M. Poliakoff, J.J. Turner, J. Chem. Soc., Faraday Trans. 70
Acknowledgements
Ž1974. 93.
w23x M. Poliakoff, J. Chem. Soc., Dalton Trans. Ž1974. 210.
w24x R.D. Sanner, R.G. Austin, M.S. Wrighton, W.D. Honnick, C.U.
Pittman, Inorg. Chem. 18 Ž1979. 928.
We wish to thank B. Heinrich and K. Myrtle of the
Surface Physics Laboratory, Simon Fraser University,
for assistance in obtaining the Auger spectra. We thank
w25x D. Sunil, J. Sokolov, M.H. Rafailovich, X. Duan, H.D. Gafney,
Inorg. Chem. 32 Ž1993. 4489.
NSERC Canada for financial support.
Ž .
w26x A. Shojaie, J.D. Atwood, Organometallics 4 Ž1985. 187.
w27x S.B. Butts, D.F. Shriver, J. Organomet. Chem. 169 Ž1979. 191.
w28x P.A. Dawson, B.M. Peake, B.H. Robinson, J. Simpson, Inorg.
Chem. 19 Ž1980. 465.
References
w29x S.M. Grant, A.R. Manning, Inorg. Chim. Acta 31 Ž1978. 41.
w30x T.R. Gaffney, J.A. Ibers, Inorg. Chem. 21 Ž1982. 2851.
w1x M.J. Almond, D.A. Rice, C.A. Yates, Chem. Br. 24 Ž1988.
130.
w
3
x
Ž
.
1 H.L. Conder, M.Y. Darensburg, J. Organomet. Chem. 67 1974
93.
1
w2x M.J. Rand, J. Electrochem. Soc. 120 Ž1973. 686.
w
w
2 S.K. Nayak, T.J. Burkey, J. Am. Chem. Soc. 115 1993 6391.
x Ž .
3
3
w3x R. Kumar, M. Rashidi, R.J. Puddephatt, Polyhedron 8 Ž1989.
x
Ž
.
3 H.C. Achton, A.R. Manning, Inorg. Chim. Acta 71 1983 163.
5
51.
w34x F.A. Cotton, R.V. Parish, J. Chem. Soc. Ž1960. 1440.