Donor properties of diphosphine ligands in tungsten carbonyl complexes: Synchrotron radiation XPS measurements and DFT calculations

Article abstract of DOI:10.1021/om0495978

Synchrotron radiation XPS measurements of W 4f and P 2p core level binding energies in the series W(CO)₄(P-P) (P-P = dppm (1), dppe (2), dppp (3), dppb (4), dmpe (5), F-dppe (6); dppm = bis(diphenylphosphino)methane, dppe = 1,2-bis(diphenylphos

Full text of DOI:10.1021/om0495978

Organometallics 2004, 23, 5219-5225
5219
Don or P r op er ties of Dip h osp h in e Liga n d s in Tu n gsten
Ca r bon yl Com p lexes: Syn ch r otr on Ra d ia tion XP S
Mea su r em en ts a n d DF T Ca lcu la tion s
Corrado Crotti,*,†,‡ Erica Farnetti,§ Teresa Celestino,§ Mauro Stener,‡,§ and
Stefano Fontana|
CNR, Istituto Struttura della Materia, Sede di Trieste, S.S.14, Km163.5,
34012 Basovizza (Trieste), Italy, Consorzio Interuniversitario Nazionale per la Scienza e la
Tecnologia dei Materiali (INSTM) and Dip. Scienze Chimiche, Universita` di Trieste,
V. Giorgieri 1, 34127 Trieste, Italy, and Sincrotrone Trieste S.C.p.A., S.S.14, Km163.5,
34012 Basovizza (Trieste), Italy
Received J une 4, 2004
Synchrotron radiation XPS measurements of W 4f and P 2p core level binding energies in
the series W(CO)4(P-P) (P-P ) dppm (1), dppe (2), dppp (3), dppb (4), dmpe (5), F-dppe (6);
dppm ) bis(diphenylphosphino)methane, dppe ) 1,2-bis(diphenylphosphino)ethane, dppp
) 1,3-bis(diphenylphosphino)propane, dppb ) 1,4-bis(diphenylphosphino)butane, dmpe )
1,2-bis(dimethylphosphino)ethane, F-dppe ) 1,2-bis(bis(pentafluorophenyl)phosphino)ethane)
are reported. The results are interpreted in terms of effects of the chelate ring size and of
the nature of substituents of the P atoms. The trend of XPS data show an excellent agreement
with the results of DFT calculations, obtained by the ∆SCF approach. Further analysis of
the Kohn-Sham eigenvalues calculated at the ground-state level has assessed the role played
by the initial state effects.
In tr od u ction
electronic and steric factors appear to be strictly cor-
related. A crucial effect on the reactivity of metal-
diphosphine compounds appears to be played by the
chain length connecting the two phosphorus atoms,
which has been interpreted in some cases as favoring a
certain geometry of an intermediate of the catalytic
cycle1d,6 and in others as allowing partial dissociation
of the chelate ring, thus providing an available coordi-
nation site;7 however, the variation of the metal-
phosphorus bond nature as a consequence of the diphos-
phine chain length variation should be also taken into
account. An attempt to correlate the steric effectssand
particularly the chelate effectssof bidentate phosphines
with the metal-phosphorus bond nature has been made
by Lichtenberger and J atcko,8 who employed gas-phase
Recent literature reports on homogeneous transition-
metal-based catalysts employing diphosphine ligands
provide increasing evidence that the electronic and
steric properties of the phosphine play a key role in
determining the catalytic properties of the complexes.1
For monodentate phosphines a number of parameters
have been established2 to describe steric and electronic
properties, even if some controversial points still re-
main, such as the extent of metal-phosphorus π-back-
bonding.2c,3 On the other hand, despite some remarkable
recent reports,1i,4 stereoelectronic effects of diphosphine
ligands5 are not equally well understood, partly because
* To whom correspondence should be addressed. E-mail:
corrado.crotti@ism.cnr.it.
† Istituto Struttura della Materia.
(3) (a) J oerg, S.; Drago, R. S.; Sales, J . Organometallics 1998, 17,
589. (b) Wang, S. P.; Richmond, M. G.; Schwartz, M. J . Am. Chem.
Soc. 1992, 114, 7595. (c) Frenking, G.; Wichmann, K.; Froehlich, N.;
Grobe, J .; Golla, W.; Van, D. L.; Krebs, B. Laege, M. Organometallics
2002, 21, 2921.
(4) (a) Dierkes, P.; van Leeuwen, P. W. N. M. J . Chem. Soc., Dalton
Trans. 1999, 1519. (b) van Leeuwen, P. W. N. M.; Kamer, P. C. J .;
Reek, J . N. H.; Dierkes, P. Chem. Rev. 2000, 100, 2741.
(5) (a) Casey, C. P.; Whiteker, G. T. Isr. J . Chem. 1990, 30, 299. (b)
Casey, C. P.; Whiteker, G. T.; Melville, M. G.; Petrovich, L. M.; Gavney,
J . A., J r.; Powell, D. R. J . Am. Chem. Soc. 1992, 114, 5535. (c) Hirota,
M.; Sakakibara, K.; Komatsuzaki, T.; Akai, I. Comput. Chem. 1991,
15, 241. (d) White, D.; Taverner, B. C.; Coville, N. J .; Wade, P. W. J .
Organomet. Chem. 1995, 495, 41.
(6) (a) Kranenburg, N.; Kamer, P. C. J .; van Leeuwen, P. W. N. M.;
Vogt, D.; Keim, W. J . Chem. Soc. Chem. Commun. 1995, 2177. (b)
Drent, E.; van Broekhoven, J . A. M.; Doyle, M. J . J . Organomet. Chem.
1991, 417, 235. (c) Petrosyan, V. S.; Permin, A. B.; Bogdashkina, V. I.;
Krut’ko, D. P. J . Organomet. Chem. 1985, 292, 303.
‡ INSTM.
§ Dip. Scienze Chimiche, Universita` di Trieste.
| Sincrotrone Trieste.
(1) (a) van der Veen, L. A.; Boele, M. D. K.; Bregman, F. R.; Kamer,
P. C. J .; van Leeuwen, P. W. N. M.; Goubitz, K.; Fraanje, J .; Schenk,
H,; Bo, C. J . Am. Chem. Soc. 1998, 120, 11616. (b) Kamer, P. C. J .;
van Leeuwen, P. W. N. M.; Reek, J . N. H. Acc. Chem. Res. 2001, 34,
895. (c) Daura-Oller, E.; Poblet, J . M.; Bo, C. Dalton 2003, 92. (d) Casey,
C. P.; Paulsen, E. L.; Beuttenmueller, E. W.; Proft, B. R.; Petrovich,
L. M.; Matter, B. A.; Powell, D. R. J . Am. Chem. Soc. 1997, 119, 11817.
(e) Casey, C. P.; Paulsen, E. L.; Beuttenmueller, E. W.; Proft, B. R.;
Matter, B. A.; Powell, D. R. J . Am. Chem. Soc. 1999, 121, 63. (f) Sesto,
B.; Consiglio, G. J . Am. Chem. Soc. 2001, 123, 4097. (g) Gonsalvi, L,;
Adams, H.; Sunley, G. J .; Ditzel, E.; Haynes, A. J . Am. Chem. Soc.
2002, 124, 13597. (h) Bianchini, C.; Lee, H. M.; Meli, A.; Oberhauser,
W.; Peruzzini, M.; Vizza, F. Organometallics 2002, 21, 16. (i) Koide,
Y.; Bott, S. G.; Barron, A. R. Organometallics 1996, 15, 2213.
(2) (a) Tolman, C. A. Chem. Rev., 1977, 77, 313. (b) Tolman, C. A.
J . Am. Chem. Soc. 1970, 92, 2953. (c) Liu, H.-Y.; Eriks, K.; Prock, A.;
Giering, W. P. Organometallics 1990, 9, 1758. (d) Fernandez, A.; Reyes,
C.; Wilson, M. R.; Woska, D. C.; Prock, A.; Giering, W. P. Organome-
tallics 1997, 16, 342. (e) Drago, R. S.; J oerg, S. J . Am. Chem. Soc. 1996,
118, 2654. (f) Moore, S. J .; Marzilli, L. G. Inorg. Chem. 1998, 37, 5329.
(7) (a) Falcon, M.; Farnetti, E.; Marsich, N. J . Organomet. Chem.
2001, 629, 187. (b) Xu, F. Y.; Zhao, A. X.; Chien, J . C. W. Makromol.
Chem. 1993, 194, 2579. (c) Spogliarich, R.; Farnetti, E.; Kaspar, J .;
Graziani, M.; Cesarotti, E. J . Mol. Catal. 1989, 50, 19. (d) J ames, B.
R.; Mahajan, D. J . Organomet. Chem. 1985, 279, 31.
10.1021/om0495978 CCC: $27.50 © 2004 American Chemical Society
Publication on Web 09/23/2004

5220 Organometallics, Vol. 23, No. 22, 2004
Crotti et al.
UPS spectroscopy to evaluate the electronic mapping
in a series of molybdenum and tungsten diphosphine
complexes; however, the need of using evaporable
compounds severely limited the number of examples
studied. A previous investigation on a series of tungsten
mono- and diphosphine complexes by UPS spectroscopy
had been reported by Bancroft et al.;9 the results were
mainly interpreted in terms of W-P bond nature,
whereas the effect of chelation was neglected.
The longstanding application10 of photoelectron spec-
troscopy, both of core level (XPS) and of valence band
(UPS), to transition-metal coordination compounds has
gained new life by the use of synchrotrons as X-ray
sources; incomparable energy resolution and photon
energy tunability have made feasible a more precise
evaluation of the variations of the binding energies
(BEs)sand therefore of the electron densitiessdepending
on the substitution of ligands in the complex. An early
application of synchrotron radiation (SR) in UPS and
XPS spectroscopy has been reported by Bancroft and
co-workers, who showed how the high resolution can
evidence fine effects in ligand substitution within a
series of tungsten carbonyl phosphine complexes.11
However, such studies were necessarily limited to
volatile compounds of monodentate phosphines.
F igu r e 1. Structures of the six diphosphines used for the
synthesis of the compounds under investigation.
information on the donor properties of the phosphine
employed: such properties will determine whether a
complex will behave as an active catalyst rather than
an inert object.
Here we report the investigation on a series of
tungsten carbonyl complexes with chelating diphos-
phines by synchrotron radiation XPS and by DFT
calculations: the differences of measured and calculated
BEs have been used to estimate the variations of
electron densities caused by the different donor abilities
of the phosphine. The complexes under investigation are
the octahedral derivatives W(CO)4(P-P) (P-P ) dppm
(1), dppe (2), dppp (3), dppb (4), dmpe (5), F-dppe (6);
dppm ) bis(diphenylphosphino)methane, dppe ) 1,2-
bis(diphenylphosphino)ethane, dppp ) 1,3-bis(diphenyl-
phosphino)propane, dppb ) 1,4-bis(diphenylphosphino)-
butane, dmpe ) 1,2-bis(dimethylphosphino)ethane,
F-dppe ) 1,2-bis(bis(pentafluorophenyl)phosphino)-
ethane) (see Figure 1): compounds 1-4 provide a series
in which the effect of chain length variation is examined,
whereas for complexes 2, 5, and 6 the substituents on
phosphorus are changed while maintaining the same
chelate ring size.
The limitation of the feasibility of SR XPS to com-
plexes which can be evaporated in the gas phase12 has
been recently overcome by our group with the setup of
a sample preparation method which produces a thin and
homogeneous layer of a transition-metal compound on
a conducting, inert support.13 The solid-phase samples
prepared by that technique minimize the differential
charging effects due to the inherent insulating nature
of these compounds, and their synchrotron radiation
XPS spectra show a reduced broadening and shifting
and an excellent reproducibility. Using this method, it
is feasible to observe very slight differences in the BE
that would otherwise be undetectable by traditional
XPS.
The interpretation of XPS spectra can be assisted by
quantum-chemical calculations; in fact, theory may be
useful to discuss the results in terms of electronic
effects. For this reason we have calculated the core
electron binding energies by means of a method based
on the density functional theory (DFT), which has
proven14 very accurate and computationally economic
for application to large transition-metal compounds.
Although most of these complexes could be evaporated
without decomposition, we have chosen them in order
to compare our solid-state results with previous higher
resolution and accurately calibrated gas-phase studies9
and to demonstrate the feasibility of this approach to
solid-state samples of nonvolatile organometallic com-
pounds.
In our present studies we propose to verify the effect
of a variation either of the diphosphine chain length or
of the substituents on phosphorus atoms on metal and
phosphorus core level BEs and to obtain quantitative
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All the reactions and manipula-
tions were routinely performed under an argon atmosphere
by using standard Schlenk tube techniques. Bis(2-methoxy-
ethyl) ether (diglyme) and all other chemicals were reagent
grade and were used as received by commercial suppliers.
The compound W(CO)4(dppe) was prepared according to the
previously reported procedure.15
(8) Lichtenberger, D. L.; Yatcko, M. E. J . Coord. Chem. 1994, 32,
79.
(9) Bancroft, G. M.; Dignard-Bailey, L.; Puddephatt, R. J . Inorg.
Chem. 1986, 25, 3675.
(10) Cook, C. D.; Wan, K. Y.; Gelius, U.; Hamrin, K.; J ohansson,
G.; Olsson, E.; Siegbahn, H.; Nordling, C.; Siegbahn, K. J . Am. Chem.
Soc. 1971, 93, 1904.
(11) Wu, J .; Bancroft, G. M.; Puddephatt, R. J .; Hu, Y. F.; Li, X.;
Tan, K. H. Inorg. Chem. 1999, 38, 4688.
In str u m en ta tion . 1H, 13C, and 31P NMR spectra were
recorded on a J EOL EX400 spectrometer operating at 399.77,
(12) (a) Green, J . C. Acc. Chem. Res. 1994, 27, 131. (b) Li, X.;
Bancroft, G. M.; Puddephatt, R. J . Acc. Chem. Res. 1997,30, 213 and
references therein.
(13) Crotti, C.; Farnetti, E.; Celestino, T.; Fontana, S. J . Electron
Spectrosc. Relat. Phenom. 2003, 128, 141.
(14) Stener M., Lisini A.; Decleva P., J . Electron Spectrosc. Relat.
Phenom. 1994, 69, 197.
(15) Quaresima, C.; Ottaviani, C.; Matteucci, M.; Crotti, C.; Antonini,
A.; Capozi, M.; Rinaldi, S.; Luce, M.; Perfetti, P.; Prince, K. C.; Astaldi,
C.; Zacchigna, M.; Romanzin, L.; Savoia, A. Nucl. Instr. Methods Phys.
Res. A 1995, 364, 374.

Diphosphine Ligands in Tungsten Carbonyl Complexes
Organometallics, Vol. 23, No. 22, 2004 5221
1
100.54, and 161.82 MHz, respectively. H chemical shifts are
XP S Sp ectr a . The samples for XPS were prepared by the
spin-coating technique as described elsewhere:13 a few drops
of a solution of the desired complex in CH2Cl2 (0.085 mmol/L)
were placed on the carefully polished surface of a gold foil
(roughness e0.1 µm; 0.5 µm diamond suspension as polishing
medium), and the spin coater was started for 5 min, at a
spinning rate of 3000 rpm, until complete evaporation of the
solvent. The spin coater used was a Model P6708 from
Specialty Coating System Inc. The dependence of the overall
spectrum on several parameters during the sample prepara-
tion (spin speed rate, solution concentration, gold foil rough-
ness) and the necessary characteristics of the support have
already been reported in a previous paper.13 The deposition
of the complexes by this technique was shown to be homoge-
neous enough to minimize the differential charging, leading
to no drifting peaks upon changing of photon flux. The
thickness of the layer was estimated to be around a few
monolayers, but the Au 4f signal from the underlying gold foil
was always detectable and was even more intense than the
W 4f signal. Each sample was introduced into the experimental
chamber through a load-lock chamber; during all measure-
ments, the experimental chamber pressure was not higher
than 2.0 × 10-10 mbar. No evaporation of the complexes from
all the solid-state samples has been observed at room tem-
perature in UHV.
reported relative to tetramethylsilane, 13C chemical shifts are
reported relative to the solvent peak (δ 77.0 for CDCl3), and
31P chemical shifts are reported relative to 85% H3PO4 as
external standard, with downfield shifts taken as positive.
Infrared spectra were recorded in Nujol mulls on a FT-IR
Perkin-Elmer System 2000 spectrometer.
Syn th esis of th e Com plexes. (a) P r epar ation of W(CO)4-
(d p p m ). Tungsten hexacarbonyl (0.52 g, 1.4 mmol) and dppm
(0.54 g, 1.4 mmol) were added into a Schlenk tube containing
5 mL of diglyme. The reaction mixture was heated at reflux
for 75 min, during which time the hexacarbonyl that sublimed
on the Schlenk walls was periodically returned to the solution
by vigorous shaking. The yellow-brown solution obtained was
cooled to room temperature; addition of 5 mL of methanol
caused precipitation of yellow crystals, which were filtered,
washed with methanol, and dried in vacuo. Yield: 71%. 1H
NMR (CDCl3, 25 °C): δ 7.5-7.2 (m, 20H, Ar), 4.89 (t, 2H, CH2,
J HP ) 9.1 Hz); 31P NMR (CDCl3, 25 °C): δ -23.3 (s, J PW 202.2
Hz). 13C NMR (CDCl3, 25 °C): δ 210.3 (m, CO trans P), 202.8
(t, CO trans CO, J CP 7.3 Hz), 136-128 (Ar), 52.2 (t, CH2, J CP
) 14.0 Hz).
(b) P r ep a r a tion of W(CO)4(d p p p ). In a Schlenk tube
containing 5 mL of diglyme, 0.52 g of tungsten hexacarbonyl
(1.4 mmol) and 0.61 g of dppp (1.4 mmol) were added. The
reaction mixture was heated at reflux for 30 min, and then
the resulting yellow-green solution was cooled to room tem-
perature. After addition of 5 mL of methanol a pale yellow
solid was obtained, which was filtered, washed with methanol,
and dried in vacuo. Yield: 77%. 1H NMR (CDCl3, 25 °C): δ
7.5-7.3 (m, 20H, Ar), 2.56 (m, 4H, PCH2), 2.00 (m, 2H, CH2).
31P NMR (CDCl3, 25 °C): δ +0.4 (s, J PW ) 223.0 Hz). 13C NMR
(CDCl3, 25 °C): δ 205.7 (m, CO trans P), 203.1 (t, CO trans
CO, J CP ) 7.4 Hz), 138-128 (Ar), 31.1 (m, PCH2), 20.2 (s, CH2).
The XPS spectra were recorded by the angle-integrated
Omicron EA 125 HR energy analyzer mounted in the experi-
mental chamber of the VUV-photoemission beam line15 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 45° 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 sections for both W 4f and P 2p core levels.16 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 were closed at 10 µm each, thus decreasing the
photon flux on the sample. The instrumental resolution was
estimated by a Fermi edge spectrum acquired with the same
setting of the core level spectra. The resulting resolution was
never worse than 0.10 eV, with an estimated photon energy
contribution of ≈45 meV, and the remaining uncertainty was
attributed to the analyzer.
(c) P r ep a r a tion of W(CO)4(d p p b). Tungsten hexacarbo-
nyl (0.52 g, 1.4 mmol) and dppb (0.64 g, 1.4 mmol) were added
into a Schlenk tube containing 5 mL of diglyme. The reaction
mixture was heated at reflux for 35 min, the yellow-brown
solution obtained was cooled to room temperature, and addi-
tion of 5 mL of methanol caused precipitation of light yellow
crystals, which were filtered, washed with methanol, and dried
1
in vacuo. Yield: 81%. H NMR (CDCl3, 25 °C): δ 7.5-7.3 (m,
20H, Ar), 2.6 (bm, 4H, PCH2), 2.0-1.7 (m, 4H, CH2). 31P NMR
(CDCl3, 25 °C): δ 11.3 (s, J PW 230.9 Hz). 13C NMR (CDCl3, 25
°C): δ 205.4 (m, CO trans P), 202.9 (t, CO trans CO, J CP ) 7.4
Hz), 138-128 (Ar), 31.2 (m, PCH2), 23.5 (s, CH2).
XP S Da ta An a lysis. The sum of several (8 ÷ 16) quick,
single-scan spectra (∼85 s/scan), each recorded on a fresh
sample area (∼0.1 mm2), gave a final spectrum with higher
signal/noise ratio and, at the same time, a careful control of
the homogeneity of the compound distribution on the sample.
Spectra were fit with a Gaussian-Lorentzian line shape using
a nonlinear least-squares procedure, modified to allow the
variation of both Gaussian and Lorentzian components for
each doublet of peaks present in the spectrum. For each
sample, the BEs from W 4f and P 2p final spectra were
referenced to the Au 4f7/2 peak at 84.00 eV acquired before
and after the series of single scans, thus eliminating any error
due to fluctuation of photon energy or electron kinetic energy.
The BEs obtained by the fit calculation are given in the present
work with a 95% interval confidence shown in parentheses.
Such interval confidence provides an evaluation only of the
precision of the fit calculation on the basis of its standard
deviation; on the other hand, the reproducibility of the BEs
could be obtained by the comparison of several (not less than
five) samples of different batches of the same compound,
measured during different beam times, and it was within (30
meV.
(d ) P r ep a r a tion of W(CO)4(d m p e). In a Schlenk tube
containing 5 mL of diglyme were added 0.74 g of tungsten
hexacarbonyl (2.0 mmol) and 0.30 g of dmpe (2.0 mmol). The
reaction mixture was heated at 130 °C for 5 h, and then the
resulting yellow solution was cooled to room temperature. Pale
yellow crystals were obtained; precipitation was completed by
addition of pentane. The product was filtered, washed with
methanol, and dried in vacuo. Yield: 46%. 1H NMR (CDCl3,
25 °C): δ 3.64-3.50 (m, 4H, CH2), 1.60 (d, 12H, CH3, J HP
6.5 Hz). 31P NMR (CDCl3, 25 °C): δ 9.0 (s, J PW ) 222.0 Hz).
13C NMR (CDCl3, 25 °C): δ 209.1 (dd, CO trans P, J CPtrans
22.1, J CPcis ) 5.5 Hz), 201.6 (t, CO trans CO, J CP ) 7.4 Hz),
31.8 (m, CH2), 19.1 (d, CH3, J CP ) 25.8 Hz).
(e) P r ep a r a tion of W(CO)4(F -d p p e). Tungsten hexacar-
bonyl (0.35 g, 1.0 mmol) and F-dppe (0.76 g, 1.0 mmol) were
added into a Schlenk tube containing 8 mL of diglyme. The
reaction mixture was heated at reflux for 90 min, the yellow-
brown solution obtained was cooled to room temperature, and
addition of 5 mL of methanol caused precipitation of a light
gray solid, which was filtered, washed with methanol and with
)
)
DF T Ca lcu la tion s. The calculations of core electron BEs
have been performed in terms of the so-called ∆SCF approach,
1
methylene chloride, and dried in vacuo. Yield: 38%. H NMR
(CDCl3, 25 °C): δ 2.94 (m, 4H, CH2). 31P NMR (CDCl3, 25 °C):
δ +17.3 (s, J PW ) 264.4 Hz). 13C NMR data were not obtained
due to low solubility.
(16) Yeh, J . J .; Lindau, I. At. Data Nucl. Data Tables 1985, 32, 1.

5222 Organometallics, Vol. 23, No. 22, 2004
Crotti et al.
Ta ble 1. W 4f Bin d in g En er gies of W(CO)4(L-L)a
calcd
exptl
W 4f BE
W 4f BE
complex
L-L
W 4f7/2 BE (eV)
∆SCF (eV)
-ꢀ (eV)
1
2
3
4
5
6
dppm
dppe
dppp
dppb
dmpe
F-dppe
31.123(7)
31.118(6)
31.081(6)
31.136(4)
31.296(8)
31.931(90)
46.58
46.57
46.54
46.52
46.65
47.38
38.91
38.92
38.89
38.88
38.74
39.58
a
Calculated values are reported at both the ∆SCF level and as
KS eigenvalues (-ꢀ).
identification of the components contributing to the total
spectrum, and a more precise measurement of KE,
intensity, and width of the single components. The
overall spectrum has been simulated by the convolution
of three components. The first and most important
doublet, being due to the W 4f atoms in the original
complex (W 4f7/2 BE ) 31.118(6) eV), is overlapped with
two other signals correlated to two different decomposi-
tion products: the second one upon beam exposition (W
4f7/2 BE ) 31.368(451) eV) and the third one due to
oxidation during the sample preparation (W 4f7/2 BE )
37.718(353) eV). The identification of the three compo-
nents has been suggested by the observation of a
sequence of spectra acquired on the same spot, where
the progressive intensity decrease of the first component
is accompanied by the parallel increase of the second,
broader component.13 In contrast, the third component
remains unchanged upon beam exposition, but its
intensity depends on the procedure adopted during
sample preparation. The acquisition procedure of the
spectra has been optimized in order to minimize the
amount of the second component (i.e.: the decomposi-
tion upon beam exposition), which was calculated to be
less than 10% of the original complex. The whole
spectrum shows an excellent separation among the
peaks, with an unprecedented overall resolution for a
solid-state sample prepared in air.
F igu r e 2. Experimental W 4f spectrum of W(CO)4(dppe)
(2) and its simulation by fitting: (a) overall fitted spectrum;
(b) spectrum of the first component; (c) spectrum of the
second component; (d) spetrum of the third component. The
residual shown is the difference between the experimental
and overall fitted spectra.
which consists of calculating the BEs as the total energy
difference between the ground-state molecule and the ion
obtained after the core electron has been removed; thus, two
separate calculations are necessary. The BEs calculated at the
∆SCF level contains both initial and final state effects: initial
effects are of the inductive type and are derived essentially
from the electron density over the atomic site where core
ionization occurs; on the other hand, final state effects, namely
relaxation, are a consequence of the response of the electrons
to the core hole formation, which are quite strong for core
ionization. Therefore, in general BEs are modulated not only
by the electron density of a specific atomic site but also by the
capability of the electrons to screen the core hole. As in the
present work we suggest employing experimental BEs to probe
the electron density of the complexes under study, it is
important to verify that the BE trend is mainly governed by
the inductive effects. For this reason it is very useful to also
extract from the theory information pertaining to only initial
states. This can be done by simply monitoring the Kohn-Sham
(KS) eigenvalue of the core state calculated at the ground-state
level, which is not affected by relaxation. For this reason we
have inserted in Tables 1 and 2 also the KS eigenvalue for
comparison. The ∆SCF calculations have been performed by
solving the KS equations for both initial and final states,
employing the LCAO method implemented in the ADF pro-
gram.17 We have employed a TZ basis set for W, while for other
atoms a DZP set has been employed. The VWN18 exchange-
correlation potential has been employed in all calculations. The
geometries of all the complexes have been optimized for all
diphosphine ligands, and the BE calculations have been
performed on the optimized structures.
The list of the W 4f7/2 BEs for complexes 1-6,
calculated on the basis of the corresponding KEs, is
shown in Table 1 with the 95% interval confidence given
in parentheses.
Analogously, the P 2p spectrum of W(CO)4(dppe) is
shown in Figure 3. Although the P 2p separation among
the three components is much smaller than in the case
of the W 4f core level, we can assign them on the basis
of their relative intensities, with the three components
correlated to the original complex (P 2p3/2 BE ) 130.876-
(21) eV), to the compound decomposed upon beam
exposition (P 2p3/2 BE ) 131.386(796) eV), and to the
complex decomposed during the sample preparation (P
2p3/2 BE ) 131.180(233) eV), respectively. As for the W4f
case, the three components have been assigned on the
basis of the evolution of the spectrum upon beam
exposition. A further weak and broad shoulder at about
126.6 eV of BE has been identified as an Auger peak of
the gold foil underlying the sample. Also in this case,
the resolution between the peaks is nearly comparable
to that of the best spectra reported in the literature for
phosphine-containing gas-phase samples.11
Resu lts
An example of a W 4f XPS spectrum of the compound
W(CO)4(dppe) is shown in Figure 2, together with the
results of the peak fitting analysis, which allows the
The P 2p3/2 BEs for all six complexes, calculated on
the basis of the corresponding KEs, is shown in Table 2
with the 95% interval confidence given in parentheses.
(17) Baerends, E. J .; Ellis, D. E.; Roos, P. Chem. Phys. 1973, 2, 41.
(18) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J . Phys. 1980, 58, 1200.

Diphosphine Ligands in Tungsten Carbonyl Complexes
Organometallics, Vol. 23, No. 22, 2004 5223
F igu r e 4. W 4f BE chemical shifts of complexes 1-6
referenced to the W(CO)4(dppp) (3) BE: XPS experimental
measurements, ∆SCF calculations, and ground-state KS
eigenvalues.
F igu r e 3. Experimental P 2p spectrum of W(CO)4(dppe)
(2) and its simulation by fitting: (a) overall fitted spectrum;
(b) spectrum of the first component; (c) spectrum of the
second component; (d) spetrum of the third component. The
residual shown is the difference between the experimental
and overall fitted spectra.
a series of analogous molecules; therefore, they are
practically canceled if relative differences are consid-
ered. For this reason it is more convenient to discuss
the theoretical BEs in terms of shifts. With regard to
the W4f core level, the comparison among the XPS
measurements, the ∆SCF calculations, and the KS
eigenvalues shows the same trend for all three methods,
with the exception of 5, which gives a rather high ∆SCF
BE but the lowest KS eigenvalue. This finding indicates
that the dmpe ligand is less effective in screening the
core hole, and therefore, in 5 the relaxation plays a
minor role with respect to the other series terms. The
same trend is found for the P 2p BE (Table 2): in this
case the differential relaxation detected in 5 is even
more pronounced than in W, due to the vicinity of the
P atom to the methyl groups.
Ta ble 2. P 2p Bin d in g En er gies of W(CO)4(L-L)a
calcd
exptl
P 2p BE
P 2p3/2 BE
complex
L-L
P 2p3/2 BE (eV) ∆SCF (eV)
-ꢀ (eV)
1
2
3
4
5
6
dppm
dppe
dppp
dppb
dmpe
F-dppe
131.058(8)
130.876(21)
130.828(6)
130.801(14)
131.285(26)
132.378(105)
138.96
138.77
138.69
138.69
138.96
140.00
124.21
124.02
123.94
123.93
123.83
125.24
a
Calculated values are reported at both the ∆SCF level and as
KS eigenvalues (-ꢀ).
The comparison between theory and experiment of W
4f BE for all six complexes is better visualized in Figure
4, where the BE values are reported in terms of shifts
relative to W(CO)4(dppp) (3), taken as reference. The
experimental slight BE decrease on going from 1 to 3,
followed by an increase which is very pronounced on
going from 5 to 6, is very well reproduced by the ∆SCF
values, apart for a minor discrepancy in the ordering
between 3 and 4.
An analogous comparison for the P 2p core level BEs
is shown in Figure 5. The significative decrease of BE
from dppm to dppe is confirmed in both experimental
and theoretical data, whereas the differences among
compounds 2-4 are just above the uncertainty related
to both methods. As observed for W 4f data, also P 2p
BEs increase significantly for compounds 5 and 6, with
a rather good agreement between experimental and
theoretical results.
We can analyze the W4f data reported in Table 1 by
separating the effect on the BEs induced by (i) variation
of the substituents on phosphorus atoms and (ii) varia-
tion of chain length between phosphorus atoms. (i)
Comparison of BEs for compounds 2, 5, and 6 reveals
that the nature of substituents on phosphorus atoms of
R2P(CH2)2PR2 is relevant in determining BEs (R ) Ph,
Me, C6F5). (ii) The modification of the chain length only
causes a slight variation of BE, almost within the
instrumental reproducibility (Table 1, entries 1-4).
As evidenced by the results reported in Table 2, the
P2p BEs follow the same trends observed for the W 4f
data; i.e., the effect of the substitution on the phos-
phorus atoms causes more important BE differences
than does the variation of chain length, except for the
BE of dppm complex 1, which is significantly higher
than for complexes 2-4.
The results of DFT calculations of W 4f and P 2p core
BEs are also reported in Tables 1 and 2, respectively.
It is well-known14 that in DFT the calculated absolute
binding energies are affected by errors due to the
presence of spurious electronic self-interaction terms.
However, these errors remain practically constant along
Discu ssion
Photoemission spectroscopy can in principle be con-
sidered as the most suitable technique to probe elec-
tronic properties of a series of ligands as a function of

5224 Organometallics, Vol. 23, No. 22, 2004
Crotti et al.
issue are studies by Bancroft and co-workers,11,19 who
reported UPS measurements on the series W(CO)5-
(PMenPh3-n) and W(CO)4(R2P(CH2)nPR2), and the trend
in W 5d BEs and in the stabilization energy of the
phosphorus lone pair on coordination. These authors
deduced that phenylphosphines are slightly stronger
donors than methylphosphines: although they take into
account the possible differential relaxation effects, they
concluded that in this case the BE shifts are related to
the real relative molecular orbital energies.
The XPS and ∆SCF results we are reporting in
Figures 4 and 5 show a behavior similar to the W 5d
data reported by Bancroft, with the phenylphosphine
complex BE (2) lower than the corresponding meth-
ylphosphine (5) value. Indeed, dmpe is less effective in
screening the core hole, and this is the reason its
experimental and ∆SCF values are high in the series.
This effect can be ascribed to the methyl groups, which
are less effective as electron reservoirs to screen the core
hole with respect to the phenyl group. In fact, the π
electrons of the phenyl moiety are expected to be much
more polarizable and therefore able to screen the core
hole than the more rigid σ electrons of the methyl group.
F igu r e 5. P 2p BE chemical shifts of complexes 1-6
referenced to the W(CO)4(dppp) (3) BE: XPS experimental
measurements, ∆SCF calculations, and ground-state KS
eigenvalues.
However, the results obtained by taking into account
only the initial state effect (i.e., the ground-state KS
eigenvalue) indicate an opposite behavior. The trend
evidenced in Figures 4 and 5 indicates that for dmpe in
the ground state the W atom is the least positive in the
series, as a consequence of a higher electron density on
the metal: i.e., of the enhanced donor properties of the
methyl substituents. On the other hand, it is known that
relaxation effects in core level BE shifts can be much
higher than in the valence band;20 this should account
for the discrepancy between our ground-state results
and W 4f BE data as well as Bancroft’s W 5d measure-
ments. The same considerations can be applied to the
P atom, with an enhanced intensity due to the proximity
of the P atom to the substituted moieties. We wish to
point out that the very satisfactory agreement between
the trends of XPS and ∆SCF results supports the
validity of both methods, thus confirming the reliability
of KS results as well.
substituents. A number of studies are present in the
literature, which report on photoemission measure-
ments on series of coordination compounds, either on
the valence shell (UPS) or on the core levels (XPS).11,12
However, to detect the small BE variation induced by
minor modifications in the ligand structure, it is neces-
sary to turn to instrumental techniques with very high
energy resolution power, such as the SR photoemission.
In the results of SR XPS measurements we are now
reporting, W 4f and P 2p BEs in the series W(CO)4(L-
L) appear to be (i) strongly influenced by the substitu-
ents on phosphorus atoms but (ii) only slightly influ-
enced by the chain length connecting the phosphorus
atoms. The latter effect is in fact detectable only in P
2p BEs, whereas the BE shifts obtained for the W 4f
core level are comparable to instrumental reproduc-
ibility. Then, it appears that the chelate ring size has a
minor influence on the P-W bond nature of these
compounds, despite the change of natural bite angle1a
in the series from dppm (four-membered chelate ring)
to dppb (seven-membered chelate ring). It must also be
noted that such small BE chemical shifts can be
partially ascribed to the “buffer” effect of the W(CO)4
fragment, which is expected to attenuate differences in
the donating properties of phosphines by corresponding
variations in the extent of W-CO π-bond. Lichtenberger
and J atcko, who also reported8 minor changes in the
UPS data of M(CO)4(dmpm) and M(CO)4(dmpe) (M )
Mo, W; dmpm ) bis(dimethylphosphino)methane, dmpe
) bis(dimethylphosphino)ethane) interpreted such re-
sults in terms of ”phosphine twist”: i.e., the capability
of phosphines to accommodate the geometric constraints
while maintaining as much of the M-P bond as possible,
resulting in a slightly bent metal-phosphine bond.
Conversely, the pronounced effect of the substituents
on phosphorus atoms as evidenced by both experimental
and calculated values deserves a deeper discussion,
mainly regarding the relative donor capability of phenyl-
and methylphosphines. Particularly relevant to this
Finally, it is worth noting the very strong effect found
for 6 in comparison to 2, which is entirely ascribed to
the initial state effect caused by the high F electro-
negativity. Although such behavior could be easily
foreseen, the excellent agreement of predictions based
on the fluorine inductive effect, experimental measure-
ments, and theoretical calculations is a further proof of
the validity of the methods used.
However, our results suggest that it is safe to use XPS
measured BEs to extract information about electron
densities only within a homologous series of com-
pounds: as a matter of fact, the trend given by XPS and
∆SCF data is perfectly confirmed by KS eigenvalues in
the series 1-4, whereas measurements on the dmpe
compound 5 should be considered within a series of
alkyl-substituted phosphines, where analogous trends
in ∆SCF and KS calculations are expected.
(19) Puddephatt, R. J .; Dignard-Bailey, L.; Bancroft, G. M. Inorg.
Chim. Acta 1985, 96, L91.
(20) Carlson, T. A., Photoelectron and Auger Spectroscopy; Plenum:
New York, 1975; p 74.

Diphosphine Ligands in Tungsten Carbonyl Complexes
Organometallics, Vol. 23, No. 22, 2004 5225
Con clu sion s
excellent agreement in the trends, thus confirming the
feasibility of the techniques used.
In the present paper we have reported the first
example of application of SR XPS measurements on a
series of solid-state coordination compounds. In the
compounds W(CO)4(P-P), the W 4f and P 2p BE
measurements allow an evaluation of the relative donor
capability of the chelating diphosphines, as a function
of chelate ring size and substituents on phosphorus. The
chain length between the phosphorus atoms has only a
minor effect on the P and W electron density, whereas
the nature of substituents has an important influence
on both core levels. A comparison between experimental
BEs and DFT-based calculations results has shown an
Ack n ow led gm en t. We are deeply grateful to all the
colleagues and technicians of the Istituto Struttura della
Materia (National Council for Researches) for their
valuable suggestions and support. We thank the Sin-
crotrone Trieste S.C.p.A. for financial support through
a Basic Research Project and for purchasing the spin
coater instrument. M.S. is grateful to the INSTM for a
generous grant of computer time on the IBM SP4
supercomputer of CINECA (Bologna, Italy).
OM0495978