Ab Initio and Experimental Studies on the Structure and Relative Stability of the cis-Hydride-η2-Dihydrogen Complexes [{P(CH2CH2PPh2)3}M (H)(η2-H2)]+ (M = Fe, Ru)

Article abstract of DOI:10.1021/ic960736q

Ab initio calculations (DMOL method) including the estimate of the total energy and the full optimization of the geometrical parameters have been used to study the electronic structures and the coordination geometries of the model systems [{P(CH₂CH₂PH₂)₃}M (H) (L)]⁺ (M = Fe, L = H₂, C₂H₄, CO, N₂; M = Ru, L = H₂). Single crystal X-ray analyses have been carried out on the complexes [(PP₃)Fe (H)(η²-H₂)]BPh₄·0.5THF (1·0.5THF), [(PP₃)Fe (H)(CO)]BPh₄·THF(3·THF), and [(PP₃)Ru (H)(η²-H₂)]BPh ₄·0.5THF(5·0.5THF) [PP₃ = P(CH₂CH₂PPh₂)₃]. Crystal data: for 1·0.5THF, triclinic P1 (No. 2), a = 17.626(3) A?, b = 14.605(3) A?, c = 12.824(4) A?, α = 90.09(2)°, β = 103.87(2)°, γ = 107.46(2)°, Z = 2, R = 0.082; for 3·THF, triclinic P1 (No. 2), a = 12.717(2) A?, b = 14.553(1) A?, c = 17.816(2) A?, α = 72.90(1)°, β = 76.82(2)°, γ = 89.71(1)°, Z = 2, R = 0.067; for 5·0.5THF, monoclinic P2/1a (No. 14), a = 19.490(5) A?, b = 19.438(2) A?, c = 16.873(5) A?, β= 110.96(2)°, Z = 4, R = 0.074. On the basis of theoretical calculations, X-ray analyses, and multinuclear NMR studies, all of the complexes of the formula [(PP₃)M (H) (L)]BPh₄ [M = Fe, L = H₂ (1), C₂H₄ (2), CO (3), N₂ (4); M = Ru, L = H₂ (5), C₂H₄ (6)] are assigned a distorted octahedral structure where the hydride (trans to a terminal phosphorus donor) and the L ligand occupy mutually cis positions. The theoretical calculations indicate that the H₂ ligand in the η²-dihydrogen-hydride derivatives 1 and 5 is placed in the P-M-H plane (parallel orientation) and that there is an attractive interaction between the H and H₂ ligands. XPS measurements, performed on the iron complexes, show that the Fe → L back-bonding interaction plays a leading role in 3. It is concluded that the stronger metal-H₂ bond in the dihydrogen-hydride complex 1, as compared to the Ru analog 5, is due to the greater d(metal) → σ*(H-H) back-donation as well as a more efficient interaction between the terminal hydride and an H of the dihydrogen ligand. This cis effect is suggested to contribute to the relative stability of the iron complexes, which increases in the order C₂H₄ < N₂ < H₂ < CO.

Full text of DOI:10.1021/ic960736q

Inorg. Chem. 1997, 36, 1061-1069
1061
Ab Initio and Experimental Studies on the Structure and Relative Stability of the
cis-Hydride-η2-Dihydrogen Complexes [{P(CH2CH2PPh2)3}M(H)(η2-H2)]+ (M ) Fe, Ru)
Claudio Bianchini,* Dante Masi, and Maurizio Peruzzini
Istituto per lo Studio della Stereochimica ed Energetica dei Composti di Coordinazione del CNR, Via J.
Nardi 39, 50132, Firenze, Italy
Maurizio Casarin,* Chiara Maccato, and Gian Andrea Rizzi
Dipartimento di Chimica Inorganica, Metallorganica ed Analitica, Universita` di Padova, Via Loredan
4-35131, Padova, Italy
ReceiVed June 14, 1996X
Ab initio calculations (DMOL method) including the estimate of the total energy and the full optimization of the
geometrical parameters have been used to study the electronic structures and the coordination geometries of the
model systems [{P(CH2CH2PH2)3}M(H)(L)]+ (M ) Fe, L ) H2, C2H4, CO, N2; M ) Ru, L ) H2). Single
crystal X-ray analyses have been carried out on the complexes [(PP3)Fe(H)(η2-H2)]BPh4‚0.5THF (1‚0.5THF),
[(PP3)Fe(H)(CO)]BPh4‚THF (3‚THF), and [(PP3)Ru(H)(η2-H2)]BPh4‚0.5THF (5‚0.5THF) [PP3 ) P(CH2CH2PPh2)3].
Crystal data: for 1‚0.5THF, triclinic P1 (No. 2), a ) 17.626(3) Å, b ) 14.605(3) Å, c ) 12.824(4) Å, R )
90.09(2)°, â ) 103.87(2)°, γ ) 107.46(2)°, Z ) 2, R ) 0.082; for 3‚THF, triclinic P1 (No. 2), a ) 12.717(2) Å,
b ) 14.553(1) Å, c ) 17.816(2) Å, R ) 72.90(1)°, â ) 76.82(2)°, γ ) 89.71(1)°, Z ) 2, R ) 0.067; for 5‚0.5THF,
monoclinic P2/1a (No. 14), a ) 19.490(5) Å, b ) 19.438(2) Å, c ) 16.873(5) Å, â ) 110.96(2)°, Z ) 4, R )
0.074. On the basis of theoretical calculations, X-ray analyses, and multinuclear NMR studies, all of the complexes
of the formula [(PP3)M(H)(L)]BPh4 [M ) Fe, L ) H2 (1), C2H4 (2), CO (3), N2 (4); M ) Ru, L ) H2 (5), C2H4
(6)] are assigned a distorted octahedral structure where the hydride (trans to a terminal phosphorus donor) and
the L ligand occupy mutually cis positions. The theoretical calculations indicate that the H2 ligand in the η2-
dihydrogen-hydride derivatives 1 and 5 is placed in the P-M-H plane (parallel orientation) and that there is an
attractive interaction between the H and H2 ligands. XPS measurements, performed on the iron complexes, show
that the Fe f L back-bonding interaction plays a leading role in 3. It is concluded that the stronger metal-H2
bond in the dihydrogen-hydride complex 1, as compared to the Ru analog 5, is due to the greater d(metal) f
σ*(H-H) back-donation as well as a more efficient interaction between the terminal hydride and an H of the
dihydrogen ligand. This cis effect is suggested to contribute to the relative stability of the iron complexes, which
increases in the order C2H4 < N2 < H2 < CO.
Introduction
by the substrate.5 Similarly, the use of either 1 or 5 as catalyst
precursors for the hydrogen-transfer reductions of R,â-unsatur-
ated ketones results in distinct activities and chemoselectivities
just because of the stronger metal-H2 bond in the iron
derivative.6
The different strength of the M-H2 bond is the key factor
which controls and distinguishes the varied chemistry of the
two cis-hydride-η2-dihydrogen complexes [(PP3)M(H)(η2-H2)]-
BPh4 [M ) Fe, 1; Ru, 5; PP3 ) P(CH2CH2PPh2)3].1,2
(3) (a) Bianchini, C.; Meli, A.; Peruzzini, M.; Frediani, P.; Bohanna, C.;
Esteruelas, M. A.; Oro, L. A. Organometallics 1992, 11, 138. (b)
Bianchini, C.; Meli, A.; Peruzzini, M.; Frediani, P.; Bohanna, C.;
Esteruelas, M. A.; Oro, L. A. In AdVances in Catalyst Design; Graziani,
M., Rao, C. N. R., Eds.; World Scientific: Singapore, 1991; p 267.
(c) Bianchini, C.; Meli, A.; Peruzzini, M.; Vizza, F.; Zanobini, F.
Organometallics 1989, 8, 2080.
(4) See for example: (a) Andriollo, A.; Esteruelas, M. A.; Meyer, U.;
Oro, L. A.; Sa´nchez-Delgado, R. A.; Sola, E.; Valero, C.; Werner, H.
J. Am. Chem. Soc. 1989, 111, 7431. (b) Albertin, G.; Amendola, P.;
Antoniutti, S.; Ianelli, S.; Pelizzi, G.; Bordignon, E. Organometallics
1991, 10, 2876. (c) Marinelli, G.; Rachidi, I. E.-I.; Streib, W. E.;
Eisenstein, O.; Caulton, K. G. J. Am. Chem. Soc. 1989, 111, 2346.
(d) Christ, M. L.; Sabo-Etienne, S.; Chaudret, B. Organometallics
1995, 14, 1082. (e) Esteruelas, M. A.; Herrero, J.; Lo`pez, A. M.; Oro,
L. A.; Schulz, M.; Werner, H. Inorg. Chem. 1982, 31, 4013. (f)
Esteruelas, M. A.; Oro, L. A.; Valero, C. Organometallics 1992, 11,
3362. (g) Espuelas, M. A.; Esteruelas, M. A.; Lahoz, F. J.; Oro, L.
A.; Valero, C. Organometallics 1993, 12, 663. (h) Harman, W. D.;
Taube, H. J. Am. Chem. Soc. 1990, 112, 2261. (i) Lin, Y. R.; Zhou,
Y. F. J. Organomet. Chem. 1990, 381, 135. (j) Bianchini, C.; Farnetti,
E.; Frediani, P.; Graziani, M.; Peruzzini, M.; Polo, A. J. Chem. Soc.,
Chem. Commun. 1991, 1336.
The dihydrogen ligand is so strongly bound to the iron center
that when 1 is employed as a homogeneous catalyst for the
hydrogenation of 1-alkynes to alkenes, a free coordination site
for the incoming alkyne molecule is provided by detachment
of a phosphine arm, rather than by H2 loss.3 In contrast, the
hydrogenation of 1-alkynes catalyzed by 5 occurs, as is usually
observed in analogous reactions,4 by initial displacement of H2
X Abstract published in AdVance ACS Abstracts, February 1, 1997.
(1) (a) Bianchini, C.; Peruzzini, M.; Zanobini, F. J. Organomet. Chem.
1988, 354, C19. (b) Bianchini, C.; Peruzzini, M.; Polo, A.; Vacca,
A.; Zanobini, F. Gazz. Chim. Ital. 1991, 121, 543.
(2) Bianchini, C.; Perez, P. J.; Peruzzini, M.; Zanobini, F.; Vacca, A. Inorg.
Chem. 1991, 30, 279.
(5) Bianchini, C.; Bohanna, C.; Esteruelas, M. A.; Frediani, P.; Meli, A.;
Oro, L. A.; Peruzzini, M. Organometallics 1992, 11, 3837.
S0020-1669(96)00736-7 CCC: $14.00 © 1997 American Chemical Society

1062 Inorganic Chemistry, Vol. 36, No. 6, 1997
Bianchini et al.
In an initial attempt to rationalize the different metal-
dihydrogen bond strength in 1 and 5, the barrier to rotation of
the H2 ligand was determined from inelastic neutron scattering
(INS) studies of the H2 rotational energy levels.7 From this
investigation, it was concluded that the Fe center (rotational
barrier of 1.82 kcal mol-1) is a better back-donor than Ru
(rotational barrier of 1.36 kcal mol-1).8 Analogous conclusions
were also reached by Morris and co-workers from NMR studies
on isostructural η2-H2 complexes of the Fe group (Fe, Ru,
Os).9,10 Back-donation arguments alone, however, cannot
explain the differences in the reactions of the two compounds
in solution. In particular, they do not account for the fact that
H2 readily displaces the ethene ligand from [(PP3)Fe(H)(C2H4)]-
BPh4 in which the back-bonding contribution is greater than in
1, whereas H2 does not displace ethene from [(PP3)Ru(H)(C2H4)]-
BPh4 (Vide infra). In other words, the extraordinarily strong
binding of the H2 ligand to iron may not be only a function of
the donation [M r σ(H2)] and back-donation [M f σ*(H2)]
contributions.11,12 The so called “cis effect”,13 the attractive
two-electron interaction between cis σ(M-H) and σ*(H-H),
might play a key role in determining the overall stability of the
present cis-hydride-η2-dihydrogen Fe and Ru complexes. The
cis effect, in fact, not only may control the orientation of the
H2 ligand in nonclassical polyhydrides13,14 but also may influ-
ence the strength of the metal-dihydrogen bond,14 as shown
by ab initio MO calculations in the related model system cis-
[Fe(PH3)4(H)(H2)]+.14,15
In a further attempt to gain insight into the different chemistry
of 1 and 5, the solid-state structures of both compounds were
determined by X-ray diffraction analyses. The crystallographic
results, which are presented in this paper, however, do not allow
one to discuss the relative orientation of the H and H2 ligands
as the dihydrogen ligands could not satisfactorily be located.
Thus, we decided to perform a theoretical analysis of the
prototype complex cations [{P(CH2CH2PH2)3}M(H)(H2)]+ (M
) Fe, 1*; M ) Ru, 5*) through calculations which include the
estimate of the total energy and the full optimization of the
geometrical parameters. The theoretical analysis has been
extended to other members of the iron family, namely, the
cations [{P(CH2CH2PH2)3}Fe(H)(L)]+ (L ) C2H4, 2*; CO, 3*;
N2, 4*), with the aim of understanding the trend in the qualitative
stability of the actual compounds [(PP3)Fe(H)(L)]BPh4 (L )
H2, CO, C2H4, N2)1,16 with respect to the dissociation of the L
ligand, which increases in the order C2H4 < N2 < H2 < CO.
Within this context, valuable information has been provided by
XPS measurements carried out on all of the iron complexes.
The experimental and theoretical studies, taken together,
provide a concerted view of the molecular structure of 1 and 5
and the difference in their chemistry.
Experimental Section
Reagents. Tetrahydrofuran (THF) was purified by distillation under
nitrogen over LiAlH4. All of the other reagents and chemicals were
reagent grade and, unless otherwise stated, were used as received by
commercial suppliers. All reactions and manipulations were routinely
performed under a dry argon atmosphere by using standard Schlenk-
tube techniques. The solid complexes were collected on sintered glass
frits and washed with light petroleum ether (bp 40-60 °C) or n-pentane
before being dried in a stream of argon. The ligand P(CH2CH2PPh2)3
(PP3) was purchased from Pressure Co. and used without further
purification. The complexes [(PP3)Fe(H)(η2-H2)]BPh4 (1),1 [(PP3)-
Fe(H)(η1-N2)]BPh4 (4),1 [(PP3)Ru(H)(η2-H2)]BPh4 (5),2 [(PP3)Ru(H)-
(CO)]BPh4 (7),2 [(PP3)Ru(H)(η1-N2)]BPh4 (8),2 [(PP3)FeH2] (9),17 and
[(PP3)RuH2] (10)2 were prepared as described in the literature.
Elemental analyses (C, H, N) were performed using a Carlo Erba Model
1106 elemental analyzer.
Spectroscopic Measurements. Deuterated solvents for NMR
measurements (Merck and Aldrich) were dried over molecular sieves
(4 Å). 1H and 13C{1H} NMR spectra were recorded on Varian VXR
300 or Bruker AC200 spectrometers operating at 299.94 or 200.13 MHz
(1H) and 75.42 or 50.32 MHz (13C), respectively. Peak positions are
relative to tetramethylsilane and were calibrated against the residual
solvent resonance (1H) or against the deuterated solvent multiplet (13C).
13C-DEPT experiments were run on the Bruker AC200 spectrometer.
The 1H,13C-2D HETCOR NMR experiment on complex 6 was recorded
on a Bruker AVANCE DRX 500 spectrometer equipped with a 5-mm
1
triple resonance probe head for H detection and inverse detection of
the heteronucleus (inverse correlation mode, HMQC experiment) with
no sample spinning. 31P{1H} NMR spectra were recorded on either
the Varian VXR 300 or Bruker AC200 instruments operating at 121.42
and 81.01 MHz, respectively. Chemical shifts were measured relative
to external 85% H3PO4 with downfield values taken as positive.
Computer simulations of the 31P{1H} NMR spectrum of 6 was carried
out with a locally developed package containing the programs
LAOCN318 and DAVINS19 run on a Compaq Deskpro 386/25 personal
computer. The initial choices of shifts and coupling constants were
refined by iterative least-squares calculations using the experimental
digitized spectrum. The final parameters gave a satisfactory fit between
experimental and calculated spectra, the agreement factor R being less
than 1% in all cases. Infrared spectra (400-4000 cm-1) were recorded
as Nujol mulls on a Perkin-Elmer 1600 Series FT-IR spectrometer
between KBr plates.
Preparation of [(PP3)Fe(H)(η2-C2H4)]BPh4 (2). Method A. Neat
MeOSO2CF3 (65 µL, 0.57 mmol) was syringed into a well-stirred THF
(15 mL) suspension of [(PP3)FeH2] (9) (365 mg, 0.50 mmol) under an
atmosphere of dry ethene. Immediately the starting dihydride dissolved
to yield a yellow solution. Stirring was continued for 1 h, and then
solid NaBPh4 (400 mg, 1.17 mmol) and ethanol saturated with ethene
(20 mL) were added. Slow concentration under a steady stream of
ethene afforded 2 as yellow microcrystals; yield, 87%.
(6) Bianchini, C.; Farnetti, E.; Graziani, M.; Peruzzini, M.; Polo, A.
Organometallics 1993, 12, 3753.
(7) For a general review of INS techniques see: Eckert, J. Spectrochim.
Acta 1992, 29, 747.
(8) Eckert, J.; Albinati, A.; White, R. P.; Bianchini, C.; Peruzzini, M.
Inorg. Chem. 1992, 31, 4241.
(9) Eckert, J.; Blank, H.; Bautista, M. T.; Morris, R. H. Inorg. Chem.
1990, 29, 747.
(10) (a) Bautista, M. T.; Cappellani, E. P.; Drouin, S. D.; Morris, R. H.;
Schweitzer, C. T.; Sella, A.; Zubkowski, J. J. Am. Chem. Soc. 1991,
113, 4876. (b) Cappellani, E. P.; Drouin, S. D.; Jia, G.; Maltby, P.
A.; Morris, R. H. J. Am. Chem. Soc. 1994, 116, 3375.
(11) General reviews on molecular hydrogen complexes include: (a) Jessop,
P. G.; Morris, R. H. Coord. Chem. ReV. 1992, 121, 155. (b) Kubas,
G. J. Acc. Chem. Res. 1988, 21, 120. (c) Kubas, G. J. Comments Inorg.
Chem. 1988, 7, 17. (d) Crabtree, R. H.; Hamilton, D. G. AdV.
Organomet. Chem. 1988, 28, 299. (e) Heinekey, D. M.; Oldham, W.
J., Jr. Chem. ReV. 1993, 93, 913.
(12) A review on the theoretical work on nonclassical transition metal
polyhydrides has appeared: Lin, Z.; Hall, M. B. Coord. Chem. ReV.
1994, 135/136, 845.
(13) (a) Jackson, S. A.; Eisenstein, O. Inorg. Chem. 1990, 29, 3910. (b)
Riehl, J.-F.; Pe´lissier, M.; Eisenstein, O. Inorg. Chem. 1992, 31, 3344.
(c) Riehl, J.-F.; Jackson, S. A.; Pe´lissier, M.; Eisenstein, O. Bull. Soc.
Chim. Fr. 1992, 129, 221. (d) van der Sluys, L. S.; Eckert, J.;
Eisenstein, O.; Hall, J. H.; Huffmann, J. C.; Jackson, S. A.; Koetzle,
T. F.; Kubas, G. J.; Vergamini, P. J.; Caulton, K. G. J. Am. Chem.
Soc. 1990, 112, 4831.
Method B. Complex 2 was also prepared by bubbling ethene into
a THF solution (15 mL) of [(PP3)Fe(H)(η1-N2)]BPh4 (4) (200 mg, 0.16
mmol) for 4 h at 0 °C. Addition of ethanol (20 mL) and workup as
above gave 2 in 76% yield. Anal. Calcd for C68H67BFeP4: C, 75.99;
H, 6.28. Found: C, 75.63; H, 6.36. IR: ν(FesH) not observed.
(16) Stoppioni, P.; Mani, F.; Sacconi, L. Inorg. Chim. Acta 1974, 11, 227.
(17) Bianchini, C.; Laschi, F.; Peruzzini, M.; Ottaviani, M. F.; Vacca, A.;
Vizza, F.; Zanello, P. Inorg. Chem. 1990, 29, 3394.
(14) Maseras, F.; Duran, M.; Lledo´s, A.; Bertran, J. J. Am. Chem. Soc.
1992, 114, 2922.
(15) Maseras, F.; Duran, M.; Lledo´s, A.; Bertran, J. J. Am. Chem. Soc.
1991, 113, 2879.
(18) Castellano, S.; Bothner-By, A. A. J. Chem. Phys. 1964, 41, 3863.
(19) Stephenson, D. S.; Binsch, G. J. Magn. Reson. 1980, 37, 295.

cis-Hydride-η2-Dihydrogen Complexes
Inorganic Chemistry, Vol. 36, No. 6, 1997 1063
Table 1. Summary of Crystal Data for 1‚0.5THF, 3‚THF, and 5‚0.5THF
1‚0.5THF
3‚THF
5‚0.5THF
chem formula
fw
space group
a, Å
b, Å
c, Å
C68H69BFeO0.5P4
1084.85
P1 (No. 2)
17.626(3)
14.605(3)
12.824(4)
90.09(2)
103.87(2)
107.46(2)
3047.69
C71H71BO2P4Fe
1146.90
P1 (No. 2)
12.717(2)
14.553(1)
17.816(2)
72.90(1)
76.82(2)
89.71(1)
3058.86
C68H69BRuO0.5P4
1127.21
P2/1a (No. 14)
19.490(5)
19.438(2)
16.873(5)
90
110.96(2)
R, deg
â, deg
γ, deg
90
5969.31
V, Å3
Z
2
1.18
2
1.24
4
1.26
F(calcd), g cm-3
radiation
graphite monochromated
Cu KR, λ ) 1.5418 Å
32.85
0.082
0.087
graphite monochromated
Mo KR, λ ) 0.710 69 Å
3.91
0.067
0.069
graphite monochromated
Cu KR, λ ) 1.5418 Å
34.96
0.074
0.079
µ, cm-1
Ra
b
Rw
a R ) ∑||Fo| - |Fc||/|Fc|. b Rw ) [∑w(|Fo| - |Fc|)2/∑w(Fo)2]1/2
.
31P{1H} NMR (C2D4Cl2/C7D8 [2:1 v/v], 121.42 MHz, sealed NMR tube
prepared under C2H4): at +25 °C, broad AMQ2 spin system, δA 162.0
(br), δM 79.7 (br), δQ 70.8 (br); at -25 °C, AMQ2 spin system, δA
161.71 (dt, JAM ) 44.4 Hz, JAQ ) 22.6 Hz), δM 79.45 (dt, JMQ ) 15.4
Hz), δQ 70.38 (dd). 1H NMR (-25 °C, CD2Cl2, 299.94 MHz, sealed
marized in Table 1. Experimental data were recorded at room
temperature on either an Enraf-Nonius CAD4 diffractometer (3‚C4H8O)
or a Philips-PW1100 diffractometer (1‚0.5C4H8O and 5‚0.5C4H8O) with
an upgraded computer control (FEBO system) using graphite-mono-
chromated Mo KR radiation (3‚C4H8O) or graphite-monochromated Cu
KR radiation (1‚0.5C4H8O and 5‚0.5C4H8O). A set of 25 carefully
centered reflections was used for the centering procedure of the crystals
and for determining the lattice constants. As a general procedure, the
intensity of three standard reflections were measured periodically every
2 h for orientation and intensity control. This procedure did not reveal
an appreciable decay of intensities for 5‚0.5C4H8O, while a decay of
about 5% was noticed for the specimen of 3‚C4H8O. A more significant
decay was observed for 1‚0.5C4H8O so that two different crystals were
used for the data collection of this compound. All of the data were
corrected for Lorentz and polarization effects. Empirical correction
for the absorption effect was performed at an advanced state of the
structure refinement by using the program DIFABS.20 Atomic scat-
tering factors were those tabulated by Cromer and Waber21 with
anomalous dispersion corrections taken from ref 22. The computational
work was performed on a Digital DEC 5000/200 computer (1‚0.5C4H8O
and 5‚0.5C4H8O) or an HP Vectra/486 computer by using the SHELX-
76 program.23 The programs PARST,24 ORTEP25 , ZORTEP,26 and
SIR9227 were also used. Final atomic coordinates of all atoms and
structure factors are available as Supporting Information.
NMR tube prepared under C2H4): δC H 3.43 (br q, JHP ) 3.0 Hz, 4H),
4
δFeH -8.43 (dt, JHPM ) 63.4 Hz, JHPQ2 ) 47.7 Hz, JHPA not resolved e
1
1 Hz, 1H). The 31P and H spectra did not change down to -80 °C.
Preparation of [(PP3)Fe(H)(CO)]BPh4 (3). This compound was
prepared several years ago by Sacconi et al. by treatment of a
suspension of [(PP3)FeBr]BPh4 in benzene with solid NaBH4 under
CO.16 However, this method gives a moderate yield of 3. The new
procedures described below provide much better yields.
Method A. The hydride-carbonyl complex 3 was prepared by
substituting CO for ethene in the above described reaction of 9 with
methyl triflate. Upon addition of NaBPh4 and ethanol, the pale yellow
solution was concentrated under a stream of argon until lemon yellow
crystals of 3 precipitated; yield, 90%.
Method B. Bubbling carbon monoxide into a THF solution of 4
for 15 min at room temperature, followed by workup as described
above, gave 3 in 92% yield. Anal. Calcd for C67H63BFeOP4: C, 74.87;
H, 5.99. Found: C, 74.79; H, 6.06. IR: ν(FesH) 1890 cm-1 (m);
ν(CtO) 1927 cm-1 (vs). 31P{1H} NMR (22 °C, CD3COCD3, 121.42
MHz) AMQ2 spin system: δA 169.34 (td, JAM ) 32.3 Hz, JAQ ) 27.2
Hz), δM 85.85 (dt, JMQ ) 12.0 Hz), δQ 87.64 (dd). 1H NMR (22 °C,
[(PP3)Fe(H)(η2-H2)]BPh4‚0.5C4H8O (1‚0.5C4H8O). Pale yellow
crystals of 1‚0.5C4H8O were grown by slow diffusion of ethanol into
a diluted THF solution of 1 under a constant stream of hydrogen. The
structure was solved by heavy atom and Fourier techniques and refined
by full-matrix least-squares methods. Anisotropic thermal parameters
were used only for Fe, P, B, and C atoms of the polyphosphine ligands.
At an advanced stage of the refinement two intensities of 0.6 e-/Å3 at
the two cis positions of the octahedron around iron were considered
potential H atoms. Although sound spectroscopic evidence shows the
presence of an H2 molecule at the position trans to the unique P atom
of the PP3 ligand, the two H components could not be resolved.
However, the refinement of the H2 ligand as a single hydride was
relatively successful. In fact, the observed behavior was very similar
to that pertaining to the classical hydride ligand undoubtely present in
the cis position. All of the phenyl rings were treated as rigid bodies
CD3COCD3, 299.94 MHz) δFeH -12.78 (tdd, JHPQ ) 59.4 Hz, JHPM
48.9 Hz, JHPA ) 17.7 Hz, 1H).
)
Preparation of [(PP3)Ru(H)(η2-C2H4)]BPh4 (6). Method A. Neat
MeOSO2CF3 (65 µL, 0.57 mmol) was syringed into a well-stirred THF
(15 mL) suspension of [(PP3)RuH2] (10) (400 mg, 0.52 mmol) under
an atmosphere of prepurified ethene. Immediately the starting dihydride
dissolved to yield a colorless solution. Stirring was continued for 1 h,
and then solid NaBPh4 (400 mg, 1.17 mmol) and ethanol saturated
with ethene (20 mL) were added. Slow concentration under a steady
stream of ethene afforded 6 as off-white microcrystals; yield 90%.
Method B. The (hydride)ethene complex 6 was prepared by
bubbling ethene into a THF solution (10 mL) of either [(PP3)Ru(H)-
(η2-H2)]BPh4 (5) (200 mg, 0.18 mmol) or [(PP3)Ru(H)(η1-N2)]BPh4
(8) (200 mg, 0.16 mmol) for 30 min at room temperature. Upon
addition of ethanol (20 mL) and concentration under ethene, off-white
crystals of 6 precipitated; yield 90%. Anal. Calcd for C68H67BRuP4:
C, 72.92; H, 6.03. Found: C, 72.84; H, 5.95. IR: ν(RusH) 1972
cm-1 (w br). 31P{1H} NMR (20 °C, CD2Cl2, 121.42 MHz), AMN2
spin system: δA 144.16 (dt, JAM ) 19.5 Hz, JAN ) 10.5 Hz), δM 55.51,
δQ 54.86 (second order MN2 multiplet, JMN ) 13.0 Hz). 1H NMR (20
°C, CD2Cl2, 299.94 MHz) δC2H4 3.58 (br s, H), δRuH -8.85 (dtd, JHPA
) 20.1 Hz, JHPM ) 48.3 Hz, JHPN ) 20.9 Hz, 1H). 13C{1H} NMR (20
°C, CD2Cl2, 75.45 MHz): δC H 51.69 (dq, JCPA ) 7.5 Hz, JCPM ≈ JCPN
(20) Walker, N.; Stuart, D. Acta Crystallogr. 1965, 31, 104.
(21) Cromer, D.; Waber, J. Acta Crystallogr. 1983, 39A, 158.
(22) International Tables of X-ray Crystallography; Kynoch Press: Bir-
mingham, U.K. 1974; Vol. 4.
(23) Sheldrick, G. M. SHELX-76 Program for Crystal Structure Determi-
nation; University of Cambridge: Cambridge, U.K., 1976.
(24) Nardelli, M. Comput. Chem. 1983, 7, 95.
(25) Johnson, C. K. Report ORNL-5138; Oak Ridge National Laboratory:
Oak Ridge, TN, 1976.
(26) Zsolnai, L.; Pritzkow, H. ZORTEP, University of Heidelberg: Ger-
many, 1994.
2
4
) 1.3 Hz).
X-ray Diffraction Studies. A summary of crystal and intensity data
for the compounds 1‚0.5C4H8O, 3‚C4H8O, and 5‚0.5C4H8O are sum-
(27) Altomare, A.; Burla, M.; Camalli, G.; Cascarano, G.; Giacovazzo, C.;
Guagliandi, A.; Polidori, G. J. Appl. Crystallogr. 1994, 27, 435.

1064 Inorganic Chemistry, Vol. 36, No. 6, 1997
Bianchini et al.
Table 2. Selected Molecular Properties for H2, C2H4, CO, and N2: Binding Energies (BE), Bond Lengths (BL), Bond Angles (BA), Dipole
Moments (µ), First Ionization Energy (IE), and Selected Vibrational Frequencies (ωc)a
molecule
-BEb (eV)
4.57 (4.49)b
BL (Å)
0.766 (0.742)c
BA (deg)
180
µe (D)
IEf (eV)
ωcg (cm-1
)
H2
C2H4
0
0
16.20 (15.43) 4179 (4401)
11.19 (10.51) 1407 (1443) ν12(CH2); 1664 (1623) ν2(CC);
26.52 (23.07)b 1.09 (1.09)d (C-H)
1.32 (1.34)d (C-C);
3062 (2990) ν11(CH); 3167 (3106)
ν9(CH); 3143 (3272) ν5(CH)
116.8 (117.8)d (HCH)
CO
N2
12.84 (11.11)b 1.129 (1.128)c
180
180
0.217 (0.122)h 14.17 (14.00) 2159 (2169)
11.19 (9.76)b
1.098 (1.097)
0
16.08 (15.58) 2398 (2359)
a Experimental values in parentheses. b Reference 48. c Reference 43. d Reference 28b. e Reference 49. f Experimental IEs50 are compared with
Slater's Transition State IEs obtained (TSIE) by running spin-polarized calculations.51 g Vibrational frequencies from references 43 and 52. h The
computed CO electric dipole moment has the correct orientation (in the sense C-O+).
with D6h symmetry (CsC ) 1.39 Å), and the hydrogen atoms were
introduced at calculated positions (CsH ) 1.08 Å). The difference
map also revealed the presence of a molecule of THF solvent which
was attributed a population factor of 0.5. The final difference map
was featureless.
[(PP3)Ru(H)(η2-H2)]BPh4‚0.5C4H8O (5‚0.5C4H8O). Pale yellow
crystals of 5‚0.5C4H8O were grown by slow diffusion of ethanol into
a diluted THF solution of 5 under a constant stream of hydrogen. The
structure was solved by heavy atom and Fourier techniques and refined
by full-matrix least-squares methods. Anisotropic thermal parameters
were used only for Ru, P, B, and C atoms of the polyphosphine ligands.
All of the phenyl rings were treated as rigid bodies with D6h symmetry
(CsC ) 1.39 Å), and the hydrogen atoms were introduced at calculated
positions (CsH ) 1.08 Å). However, the hydride ligands, not located
on ∆F maps, were not included in the model. The difference map
also revealed the presence of a molecule of THF solvent which was
attributed a population factor of 0.5.
[(PP3)Fe(H)(CO)]BPh4‚C4H8O (3‚C4H8O). A thin lemon yellow
parallelepiped of 3‚C4H8O, suitable for an X-ray diffraction study, was
slowly grown from a diluted THF solution of 3 layered with ethanol
and kept under a positive pressure of argon. The structure was solved
by direct methods using the SIR92 program27 and all of the non-
hydrogen atoms were located through a series of Fo Fourier maps.
Refinement was done by full-matrix least-squares calculations, initially
with isotropic thermal parameters, and then, during the least-squares
refinement, Fe, P, and C atoms, except the carbonyl carbon atom, were
allotted anisotropic thermal parameters. At the last stage of the
refinement, an intensity of 0.85 e-/Å3, near the metal atom, was lo-
cated in the difference map and successfully refined in the least-
squares cycles as the hydride ligand, with free positional and isotropic
thermal parameters. The difference map also revealed the presence of
a molecule of THF solvent which was successfully retained in the
model.
XPS Measurements. A Perkin-Elmer Φ 5600ci spectrometer with
standard Mg KR radiation (1253.6 eV) was used for the XPS analyses.
The working pressure was less than 1.8 × 10-9 mbar. The spectrometer
was calibrated by assuming the binding energy (BE) of the Au 4f7/2
line at 83.9 eV with respect to the Fermi level. As an internal reference
for the peak positions the C 1s peak of hydrocarbon contamination
has been assumed at 284.8 eV. The standard deviation in the BE values
of the XPS lines is 0.10 eV. After a Shirley-type background
subtraction, the raw spectra were fitted using a nonlinear least-squares
fitting program adopting Gaussian-Lorentzian peak shapes for all of
the peaks. The powder samples were pressed onto a grooved copper
plate and immediately introduced into the analysis chamber. No flood
gun was used to avoid sample charging, in order to minimize the
possibility of sample decomposition due to the electron beam.
Computational Details. All calculations have been run by using
the DMOL method,28 an ab initio total energy numerical method where
the local density functional (LDF) Kohn-Sham equations are solved
for systems with a finite size providing energy eigenvalues, eigenvec-
tors, and charge distribution and allowing the analytic evaluation of
energy gradients (force calculations). Exact numerical LDF spherical-
atom atomic orbitals (NAOs) are used as basis functions. Because of
the quality of these orbitals, basis set superposition effects are minimized
and a good description of even weak bonding interactions is possible.28a
It has already been shown29 that a double numerical basis set (DN),
obtained by increasing the functions of the neutral atom by the atomic
valence functions of the corresponding 2+ ions, provides good results
and that DN bond lengths (BLs) and DN bond angles (BAs) in the
geometry optimization of molecular systems usually converge within
∼0.01 Å and 1°, respectively. In order to ensure a sufficiently high
variational flexibility, the following NAOs obtained from LDF calcula-
tions relative to free atoms/ions have been employed: (a) Ru, the 1s-
5s NAOs of the neutral ruthenium atom, the 4d-5p NAOs of Ru2+
;
(b) Fe, the 1s-4s NAOs of the neutral iron atom, the 3d-4p NAOs of
Fe2+; (c) P, the 1s-3p NAOs of the neutral phosphorus atom, the 3s-
3d NAOs of P2+; (d) O, the 1s-2p NAOs of the neutral oxygen atom,
the 2s-2p NAOs of O2+ and two sets of 3d, 2p, and 1s NAOs generated
from two hydrogenic calculations using Z ) 5 and Z ) 7; (e) N, the
1s-2p NAOs of the neutral nitrogen atom, the 2s-2p NAOs of N2+
and two sets of 3d, 2p, and 1s NAOs generated from two hydrogenic
calculations using Z ) 5 and Z ) 7; (f) C, the 1s-2p NAOs of the
neutral carbon atom, the 2s-2p NAOs of C2+ and two sets of 3d, 2p,
and 1s NAOs generated from two hydrogenic calculations using Z )
5 and Z ) 7; (g) H, the 1s NAO of the neutral hydrogen atom and two
sets of 1s and 2p NAOs generated from two hydrogenic calculations
using Z ) 1.3 and Z ) 4. During the calculations, the Ru 1s-3d NAOs,
the Fe 1s-2p NAOs and the P, O, N, and C 1s NAOs have been kept
frozen in a fully occupied configuration, allowing their exclusion from
the variational space. Detailed information about the numerical
integration scheme is reported in ref 28a. Here, it is sufficient to specify
that a number e 1000 sample points/atom has been used. The
integration points are spherically distributed around each atomic site.
Radial points are taken from the nucleus to an outer distance of 10
bohr. The number of radial points (NR) in the selected range is designed
to scale up with increasing atomic number (Z): NR ) 14(Z + 2)1/3. As
for the l value of the one-center expansion about each nucleus is
concerned, a value of l one greater than that in the atomic basis set has
been found to provide sufficient precision.28a Here, we have adopted
the following degrees of angular truncation: l ) 3 for Fe and Ru and
l ) 2 for P, O, N, C, and H. The use of sets (d) ÷ (g) for the subsequent
free molecules C2H4, CO, N2, and H2 has been found to be adequate to
correctly reproduce a series of molecular properties as bond lengths
(BL), bond angles (BA), vibrational frequencies (ν), dipole moments
(µ), and ionization energies (IEs) (Table 2).
As it is shown in ref 30, the main discrepancy between theory and
experiment usually regards the molecular atomization energy (AE)
which, on the other hand, is a well-known drawback of the local density
approximation (LDA).31 In ref 30, it was also pointed out that a definite
improvement can be obtained by using the Becke 1988 version of a
gradient corrected exchange functional.32 Since our main interest in
the present theoretical analysis was the electronic and molecular
structure of the prototype molecular ions rather than their overall AEs,
we have limited our investigation to the use of the LDA. All
(29) (a) Casarin, M.; Tondello, E.; Vittadini, A. Surf. Sci. 1994, 303, 125,
and references cited therein. (b) Casarin, M.; Tondello, E.; Vittadini,
A. Surf. Sci. 1994, 307/309, 1182. (c) Casarin, M.; Tondello, E.;
Vittadini, A. Inorg. Chim. Acta 1995, 235, 151.
(30) Casarin, M.; Favero, G.; Tondello E.; Vittadini A. Surf. Sci. 1994,
317, 422.
(28) (a) Delley B. J. Chem. Phys. 1990, 92, 508. (b) Delley B. J. Chem.
Phys. 1991, 94, 7245.
(31) Ziegler T. Chem. ReV. 1991, 91, 651.
(32) Becke, A. D. J. Chem. Phys. 1988, 88, 2547.

cis-Hydride-η2-Dihydrogen Complexes
Inorganic Chemistry, Vol. 36, No. 6, 1997 1065
1
Scheme 1
at -25 °C. At this temperature, the H NMR spectrum of 2
exhibits a high-field multiplet which is assigned to the terminal
hydride ligand. Interestingly, this resonance appears as a doublet
of triplets (δFeH -8.43, JHPM ) 63.4 Hz, JHPQ ) 47.7 Hz) with
no apparent coupling to the cis disposed bridgehead phosphorus
atom. This unique feature of the proton NMR spectrum of 2 is
not observed in the spectrum of the ruthenium complex 6 where
a canonical doublet of triplets of doublets centered at -8.85
ppm is displayed.2,33 A broad signal (2, δ 3.43; 6, δ 3.58)
corresponding to four protons and featuring the shape of an
unresolved quartet in the iron derivative (JHPM ≈ JHPN 3.0 Hz)
is assigned to the four equivalent protons of the metal
coordinated ethene molecule, which consistently should freely
rotate about the M-C2H4 axis. This assignment is supported
not only by the chemical shift values34 but also by the results
of a 2D-1H,13C-NMR HETCOR experiment for complex 6. In
fact, the 13C{1H} NMR spectrum of 6 contains a doublet of
quartets at 51.69 ppm (JCPA ) 7.5 Hz, JCPM ≈ JCPN ) 1.3 Hz)
which inverts in a DEPT experiment and thus can unequivocally
be assigned to the π-ethene carbon atoms. As a matter of fact,
from an examination of the cross-peaks network (HMQC 2D-
NMR experiment), one can correlate this carbon resonance with
the broad singlet at 3.58 ppm. The low stability of 2 in all
common NMR solvents did not allow us to record an informa-
tive 13C NMR spectrum.
X-ray Diffraction Studies. X-ray diffraction analyses were
carried out on the iron and the ruthenium dihydrogen-hydride
complexes after recrystallization from THF/ethanol mixtures
which gave the hemisolvate adducts 1‚0.5THF and 5‚0.5THF.
An X-ray structure determination was undertaken also on the
iron(II) (hydride)carbonyl complex 3 after the compound was
recrystallized from THF/ethanol solution to give 3‚THF.
The crystal structure of the two nonclassical polyhydrides
consists of discrete [(PP3)MH3]+ cations (M ) Fe, Ru),
tetraphenylborate anions, and clathrated THF molecules in a
1:1:0.5 ratio. Selected bond lengths and angles are given in
Table 3, while ORTEP drawings of the complex cations are
presented in Figure 1 (Fe) and Figure 2 (Ru).
A stoichiometric ratio of 1:1:1 features the crystal structure
of the carbonyl complex 3‚THF which contains [(PP3)Fe(H)-
(CO)]+ cations, tetraphenylborate anions and interspersed THF
solvent molecules in the lattice. A perspective view of the
complex cation is shown in Figure 3, while selected metrical
data are provided in Table 3.
All complex cations exhibit distorted octahedral structures
with the (PP3)M fragment adopting a pseudo C2V symmetry.
Indeed, distortions from the idealized octahedral geometry are
invariably encountered in transition metal complexes of tripodal
polyphosphines and are generally determined by the steric
constraints imposed by this type of chelating ligands.33
In the present complexes, the largest deviation from the
octahedral geometry is associated with the opening of the
P(1)-M-P(2) angle which ranges from 153.0(2)° (3‚THF) to
154.1(1)° (1‚0.5THF). This kind of deformation is a common
feature of six-coordinate PP3 metal complexes irrespective of
the transition metal.33 Among the several structures which are
relevant to this study, it is worth mentioning the nonclassical
trihydride [(PP3)OsH3]BPh4 (11) [157.0(1)°]35 and the iron
derivative [(PP3)Fe{η2-O,O′-O2CCH2CH2CtCH}]BPh4 [159.7-
(1)°].36 A still more relevant bending with respect to the ideal
calculations have been performed on a workstation IBM 6000/550 at
the Inorganic Chemistry Department of the University of Padova.
Results and Discussion
Synthesis and Spectroscopic Characterization of [(PP3)-
M(H)(η2-C2H4)]BPh4 (M ) Fe, 2; Ru, 6). The reaction of
MeOSO2CF3 with a THF solution of the dihydride complexes
[(PP3)MH2] (M ) Fe, 9; Ru, 10) under an atmosphere of ethene
(Scheme 1) is an excellent method for the synthesis of the new
octahedral cis-hydride-ethene complexes [(PP3)M(H)(η2-
C2H4)]+ (M ) Fe, 2; Ru, 6) which can be isolated as
tetraphenylborate salts upon addition of NaBPh4. Alternatively,
compounds 2 and 6 can be prepared by bubbling ethene into
THF solutions of [(PP3)M(H)(L)]BPh4 complexes [M ) Ru, L
) N2 (8), H2 (5); M ) Fe, L ) N2 (4)].
The two (hydride)ethene complexes 2 and 6 exhibit a totally
different stability in solution. In particular, while the ruthenium
derivative is air-stable and does not eliminate ethene under ar-
gon, dihydrogen, or dinitrogen, the iron derivative is air- and
moisture-sensitive and readily decomposes to give the para-
magnetic dark-violet five-coordinate derivative [(PP3)FeH]BPh4.17
Moreover, the iron derivative readily undergoes in solution the
displacement of ethene by carbon monoxide, dinitrogen, or
dihydrogen to form the known [(PP3)Fe(H)(L)]BPh4 compounds
(L ) CO, N2, and H2),1,16 whereas the ethene ligand in the Ru
derivative 6 is displaced by CO to give the (hydride)carbonyl
complex [(PP3)Ru(H)(CO)]BPh4 (7) but not by N2 or H2.2
The IR spectrum of the iron derivative does not display any
absorption ascribable to a ν(Fe-H) vibration. In contrast, a
weak band at 1972 cm-1 in the IR spectrum of 6 is assigned to
ν(Ru-H). In keeping with an octahedral coordination geometry
around the metal centers, both 2 and 6 exhibit an AMQ2 splitting
pattern in their 31P{1H} NMR spectra.1,2,17,33 The high-field
components in the spectrum of 6 are slightly perturbed by
second-order effects. The Ru derivative is stereochemically
rigid on the NMR time scale, while the iron complex features
a dynamic behavior as shown by the broadening of the
phosphorus resonances at room temperature. Unfortunately, the
observation of the fast-exchange spectrum was not possible due
to extensive decomposition of 2 in CD4Cl2/toluene-d8 solution
at temperatures higher than 40 °C. Above this temperature, a
deep violet solution was invariably obtained with no signal
detectable in the 31P NMR spectrum over a spectral width of
100 000 Hz, which suggests the conversion to the above
mentioned paramagnetic five-coordinate [(PP3)FeH]+ species.
In contrast, the scrambling process which makes equivalent the
terminal phosphorus nuclei of the PP3 ligand, is frozen already
(33) For other examples of octahedral complexes of PP3 see: (a) Bianchini,
C.; Masi, D.; Meli, A.; Peruzzini, M.; Zanobini, F. J. Am. Chem. Soc.
1988, 110, 6411. (b) Bianchini, C.; Meli, A.; Peruzzini, M.; Ramirez,
J. A.; Vacca, A.; Vizza, F.; Zanobini, F. Organometallics 1989, 8,
337. (c) Bianchini, C.; Masi, D.; Meli, A.; Peruzzini, M.; Ramirez, J.
A.; Vacca, A.; Zanobini, F. Organometallics 1989, 8, 2179. (d) Di
Vaira, M.; Peruzzini, M.; Stoppioni, P. Inorg. Chem. 1991, 30, 1001.
(e) Bianchini, C.; Linn, K.; Masi, D.; Mealli, C.; Peruzzini, M.;
Zanobini, F. Inorg. Chem. 1992, 31, 4036. (f) Barbaro, P.; Bianchini,
C.; Peruzzini, M.; Polo, A.; Zanobini, F. Inorg. Chim. Acta 1990, 220,
5. (g) Bianchini, C.; Frediani, P.; Masi, D.; Peruzzini, M.; Zanobini,
F. Organometallics 1994, 13, 4616. (h) Jia, G.; Drouin, S.; Jessop, P.
G.; Lough, A. J.; Morris, R. H. Organometallics 1993, 12, 906.
(34) See for example: (a) Werner, H.; Schwab, P.; Wolf, J. Chem. Ber.
1992, 125, 2641. (b) Barbaro, P.; Bianchini, C.; Meli, A.; Peruzzini,
M.; Vacca, A.; Vizza, F. Organometallics 1991, 10, 2227.
(35) Bianchini, C.; Linn, K.; Masi, D.; Peruzzini, M.; Polo, A.; Vacca, A.;
Zanobini, F. Inorg. Chem. 1993, 32, 2366.

1066 Inorganic Chemistry, Vol. 36, No. 6, 1997
Bianchini et al.
Table 3. Selected Bond Distances (Å) and Angles (deg) for
1‚0.5THF, 3‚THF, and 5‚0.5THF
1‚0.5THFa
3‚THF
5‚0.5THF
Fe-P(1)
2.208(3)
2.217(3)
2.223(3)
2.150(3)
1.59(7)
1.6(1)
2.222(3)
2.216(3)
2.236(4)
2.181(4)
1.44(10)
Fe-P(2)
Fe-P(3)
Fe-P(4)
Fe-H(1)
Fe-H(2)
Ru-P(1)
2.350(3)
2.334(4)
2.369(3)
2.249(3)
Ru-P(2)
Ru-P(3)
Ru-P(4)
Figure 2. ORTEP drawing of the complex cation [(PP3)Ru(H)-
(η2-H2)]+. Only the ipso carbons of the phenyl rings are shown for the
sake of clarity.
Fe-C(7)
1.69(1)
1.17(2)
153.0(2)
107.7(2)
85.8(1)
103.8(2)
84.6(1)
86.6(1)
77(4)
C(7)-O(1)
P(1)-Fe-P(2)
P(1)-Fe-P(3)
P(1)-Fe-P(4)
P(2)-Fe-P(3)
P(2)-Fe-P(4)
P(3)-Fe-P(4)
H(1)-Fe-P(1)
H(1)-Fe-P(2)
H(1)-Fe-P(3)
H(1)-Fe-P(4)
H(2)-Fe-P(1)
H(2)-Fe-P(2)
H(2)-Fe-P(3)
H(2)-Fe-P(4)
H(1)-Fe-H(2)
H(1)-Fe-C(7)
Fe-C(7)-O
C(7)-Fe-P(1)
C(7)-Fe-P(2)
C(7)-Fe-P(3)
C(7)-Fe-P(4)
P(1)-Ru-P(2)
P(1)-Ru-P(3)
P(1)-Ru-P(4)
P(2)-Ru-P(3)
P(2)-Ru-P(4)
P(3)-Ru-P(4)
154.1(1)
101.2(1)
86.0(1)
102.8(2)
85.6(1)
87.2(1)
92(3)
96(3)
96(3)
176(2)
83(3)
72(3)
76(3)
166(4)
80(4)
167(4)
81(4)
96(4)
95(4)
174(1)
97.4(5)
89.8(5)
99.0(5)
172.9(5)
Figure 3. ORTEP drawing of the complex cation [(PP3)Fe(H)(CO)]+.
Only the ipso carbons of the phenyl rings are shown for the sake of
clarity.
hydride and a nonclassical H2 ligand must be located mutually
cis in the equatorial plane of the octahedron defined also by
the bridgehead P4 atom and by one terminal PPh2 group (P3).
In the structure of 3‚THF, a terminal carbonyl ligand occupies
the position trans to a bridgehead P4 atom which is taken by
the nonclassical ligand in both 1‚0.5THF and 5‚0.5THF.
However, although the presence of the hydride ligands was
apparent from the difference Fourier maps in both 1‚0.5THF
and 5‚0.5THF, the refinement procedures were unsuccessful.11,38
We could not precisely locate the hydride ligands in the structure
of the ruthenium complex and could only approximate their
position in the iron complex 1‚0.5THF. A similar situation was
encountered in the structure of 11,35 where a single unresolved
electron density trans to the bridgehead P4 atom in the Fourier
difference map was treated as a single hydride (H(1) in Table
3), although there is a clearcut spectroscopic evidence pointing
to presence of a molecular hydrogen ligand in this position.1,2
The Fe-H bond distances found in the polyhydride 1‚0.5THF
are similar to each other [1.59(7) Vs 1.6(1) Å] and falls within
153.9(1)
100.3(1)
84.3(1)
99.1(1)
83.5(1)
84.5(1)
a In the (dihydrogen)hydride complex 1‚0.5THF, H(1) represents the
two hydrogen atoms of the H2 ligand. Crystallographically, the two H
atoms appear as a unique unresolved peak and have been left out of
the ORTEP diagram (Figure 1).
the range reported by Orpen and co-workers [dFe-H ) 1.609
av
Å, σ ) 0.004].39 A similar relationship between the Os-H
and Os-H2 bond lengths has already been observed in 11.
Interestingly, the Fe-H bond distance in the hydride-carbonyl
complex 3‚THF [1.44(10) Å] is significantly shorter than those
pertaining to the classical and nonclassical hydrides in 1‚0.5THF,
but matches well with the value reported for the hydride-
carbonyl complex 12 [1.43(9) Å].37
In agreement with the strong trans influence exerted by the
terminal hydride ligand, the P3 atom trans to the hydride exhibits
the longest M-P separation.33,35 The other Fe-P and Ru-P
bond distances are within the range found for other phosphine
Figure 1. ORTEP drawing of the complex cation [(PP3)Fe(H)(η2-H2)]+.
Only the ipso carbons of the phenyl rings are shown for the sake of
clarity.
value of 180° has been found in the hydride-carbonyl complex
[(NP3)Fe(H)(CO)]BPh4 (12) [142.1(1)°] [NP3 ) N(CH2CH2-
PPh2)3].37
In keeping with the molecular structure of 1135 as well as
the spectroscopic characterization of 11 and 5,2 a terminal
(36) Linn, K.; Masi, D.; Mealli, C.; Bianchini, C.; Peruzzini, M. Acta
Crystallogr. 1992, C48, 2220.
(37) George, T. A.; Rose, D. J.; Chang, Y.; Chen, Q.; Zubieta, J. Inorg.
Chem. 1995, 34, 1295.
(38) Bau, R.; Teller, R. G.; Kirtley, S. W.; Koetzle, T. F. Acc. Chem. Res.
1979, 12, 176.
(39) Orpen, A. G.; Brammer, L.; Allen, F. H.; Kennard, O.; Watson, D.
G.; Taylor, R. J. Chem. Soc., Dalton Trans. 1989, S1.

cis-Hydride-η2-Dihydrogen Complexes
Inorganic Chemistry, Vol. 36, No. 6, 1997 1067
Table 4. Binding Energy Values (eV) for [(PP3)Fe(H)(L)]BPh4 (L
) H2, 1; C2H4, 2; CO, 3; N2, 4)
compound
Fe 2p3/2
P 2p
[(PP3)Fe(H)(η2-H2)]BPh4
[(PP3)Fe(H)(η2-C2H4)]BPh4
[(PP3)Fe(H)(CO)]BPh4
[(PP3)Fe(H)(η1-N2)]BPh4
708.2
708.4
709.1
708.4
131.5
131.7
131.8
131.7
complexes of iron(II)36,40 and ruthenium(II)33e,g stabilized by
tripodal polyphosphine ligands.
XPS Measurements. Table 4 reports the binding energies
of Fe 2p3/2 and P 2p peaks for complexes 1-4. As expected
for polyphosphine organometallic compounds, the values of the
binding energies are significantly different from those of
molecular systems where a d6 metal ion is coordinated by “hard”
ligands such as O, N, or Cl and has a formal oxidation number
of +2.41 Actually, the lower binding energies we have measured
are consistent with an oxidation state closer to 0 or +1. Such
a difference is attributed to the large number of phosphine donor
atoms which increase the electron density around the metal
center.
Comparable binding energy values of the Fe 2p3/2 peak have
been found for compounds 1, 2, and 4 (Table 4), whereas this
peak lies at significantly higher binding energy in 3. The
observed trend of binding energies can be related to the different
contribution of back-donation from occupied Fe 3d t2g-like levels
(despite the absence of any symmetry element in the structure
of the prototype molecular ions, it is still possible to pick up
the eg and t2g character of Fe 3d based AOs) into the virtual
levels of the sixth ligand L (L ) H2, C2H4, CO, N2), and points
out that this interaction is very strong when L ) CO.
A further point to consider is that the Fe 2p3/2 band shape of
the carbonyl complex 3 is different from those observed for 1,
2, or 4. There is, in fact, a strong shake up of the structure,
indicating that final state effects cannot be neglected in the
photoionization process. Therefore, simple considerations based
on the electron density calculated on the metal may lead to
wrong previsions about the binding energy values.
XPS data are not reported for the Ru complexes as the Ru
3d peak extensively overlaps with the very intense C 1s peak;
the binding energies thus cannot precisely be determined.
Theoretical Results. Ball and stick representations of the
complex cations [{P(CH2CH2PH2)3}Fe(H)(H2)]+ (L ) H2, 1*;
C2H4, 2*; CO, 3*; N2, 4*) are reported in Figure 4 ([{P(CH2-
CH2PH2)3}Ru(H)(H2)]+, 5*, is isostructural with 1*). Unlike
the carbonyl 3 and dinitrogen 4 derivatives, two conformers
are possible for the dihydrogen and ethene complexes 1 and 2:
one with the L ligand parallel (|) to the M-Hi bond (Hi )
terminal hydride), the other with L perpendicular ( ) to it. Both
conformations have been taken into account for the Fe com-
plexes 1* and 2*, while only the H2 derivative 5* has been
investigated for Ru.
From a perusal of Table 5, one may readily infer that only
slight variations of the geometrical parameters are exhibited by
the [{P(CH2CH2PPH2)3}Fe(H)]+ fragment along the investigated
series, which is in good agreement with crystallographic data
for 1 and 3. An analogous agreement is obtained for 5*| and
5* (Table 6). The largest deviation between theory and
experiment has systematically been found for the M-Pe
distances as a consequence of the substitution of the actual
phenyl groups with hydrogen atoms. Consistently, the agree-
Figure 4. Ball and stick representations of the complex cations
[{P(CH2CH2PH2)3}Fe(H)(L)]+ (L ) H2, 1*; C2H4, 2*; CO, 3*; N2, 4*).
ment between experimental and theoretical Fe-Pa bond lengths
is much better.
Selected gross atomic charges (Q) and overlap populations
(OPs) obtained from the Mulliken42 population analysis are
reported in Tables 7 and 8. An inspection of these tables clearly
indicates that the metal center carries a significant amount of
electronic charge as a consequence of the strong donation from
the P lone pairs (it is worth mentioning that the overall positive
charge carried by the complex cations is almost completely
localized over the H atoms of the model PH2 groups). In
agreement with the XPS measurements (Table 4), the negative
charge localized over the Fe atom in 1* is the largest in the
series, most probably due to the weak back-donation from the
occupied 3d Fe t2g-like AOs (dxy, dxz, dyz in our framework)
into the high-lying H2 σ* MO. The poor back-donation into
the σ* H2 level is thus consistent with the presence of an intact
H2 ligand, although we observe a significant decrease of the
H-H OP upon H2 coordination (Table 7). Most importantly,
the total energy corresponding to the optimized molecular
structure of 1*| is lower than that of 1* by about 1.5 kcal mol-1
(a rotational barrier of 1.82 kcal mol-1 was determined for the
H2 ligand by the INS studies).8 A reasonable explanation of
the higher stability of 1| is offered by the weak but not negligible
bonding interaction between Hi and H2 (OP ) 0.047e). As a
whole, the present results are in excellent agreement with those
obtained by Bertra´n and co-workers for the model compound
cis-[Fe(PH3)4(H)(H2)]+.14,15 Also in this model system, the
conformer where the H2 ligand is placed in the P-Fe-H plane
(parallel orientation) is more stable than the perpendicular
isomer. More importantly, the Fe-H2 bond is much stronger
in the parallel isomer and is also asymmetric due to an attractive
cis effect between the terminal hydride and the H2 ligand. The
cis effect is suggested to be the factor which stabilizes the
parallel isomer.13 Different bonding distances between the metal
center and the hydrogens of the H2 ligand have experimentally
been found in Fe(H)2(H2)(PEtPh2)3 authenticated by neutron
diffraction.13d In the present theoretical analysis, we do not find
asymmetry in the Fe-η2-H2 moiety (Table 5), most likely due
to the limitations of the applied methodology, and the only
evidence for the cis effect is provided by the bonding interaction
between the terminal hydride Hi and one of the hydrogens of
H2. On the other hand, the existence of a cis effect is supported
by the low-energy intramolecular exchange between the hydride
and the molecular hydrogen in 1 [∆Gq(300 K) ) 12 ( 1 kcal
mol-1].1,11a This exchange has been interpreted on both
experimental1,2 and theoretical grounds15 as occurring through
(40) Bianchini, C.; Laschi, F.; Masi, D.; Ottaviani, F. M.; Pastor, A.;
Peruzzini, M.; Zanello, P.; Zanobini, F. J. Am. Chem. Soc. 1993, 115,
2723.
(41) Siedle, A. R.; Newmark, R. A.; Korba, G. A.; Pignolet, L. H.; Boyle,
P. D. Inorg. Chem. 1988, 27, 1593.
(42) Mulliken, R. S. J. Chem. Phys. 1955, 23, 1833.

1068 Inorganic Chemistry, Vol. 36, No. 6, 1997
Table 5. Selected Bond Lengths (BL) for 1*|, 1* , 2*|, 2* , 3*, and 4* (When Available Experimental Values Are Given in Parentheses)a
1* 2*
Bianchini et al.
BL (Å)
BL (Å)
3*
4*
1*|
1*
2*|
2*
BL (Å)
BL (Å)
Fe-H1
Fe-H2
Fe-Hi
H2-H1
1.56 (1.6)
1.56 (1.6)
1.51 (1.59)
0.95
1.57
Fe-C1
2.06
2.07
1.49
1.39
1.09
2.16
2.15
1.84
1.84
2.06
2.08
1.52
1.40
1.10
2.14
2.14
1.83
1.83
Fe-C
C-O
Fe-Hi
1.75 (1.69)
1.15 (1.17)
1.51 (1.44)
Fe-N
N-N
Fe-Hi
1.80
1.11
1.51
1.57
1.51
0.93
Fe-C2
Fe-Hi
C2-C1
C-H(C2H4)
Fe-Pa
Fe-Peb
Pa-Cb
Pe-Cb
2.14 (2.15)
2.15 (2.22)
1.84 (1.82)
1.85 (1.85)
2.13 (2.15)
2.15 (2.22)
1.84 (1.82)
1.85 (1.85)
2.17 (2.18)
2.14 (2.23)
1.83 (1.82)
1.84 (1.84)
2.15
2.16
1.83
1.84
a Key: Pa denotes the bridgehead phosphorus atom of PP3 and Pe indicates the equatorial phosphorus atoms of PP3; Hi denotes the terminal
hydride ligand. b Mean value.
Table 6. Selected Bond Lengths (BL) for 5*| and 5*
( ); 0.033e (|)) to the Fe-C2H4 bonding, clearly showing that
back-donation, if on the one hand it plays a leading role in the
(Experimental Values When Available Are Given in Parentheses)a
Fe-C2H4 interaction, on the other hand is the main origin of
the C1-C2 bond weakening.
BL (Å)
BL (Å)
5*|
5*
5*|
5*
Unlike the C1-C2 OP in 2*, the C-O one in 3* increases
Ru-H1
Ru-H2
Ru-Hi
H2-H1
1.77
1.77
1.63
0.89
1.74
1.75
1.64
0.91
Ru-Pa
Ru-Peb
Pa-Cb
Pe-Cb
2.26 (2.25)
2.31 (2.35)
1.83 (1.84)
1.85 (1.86)
2.28 (2.25)
2.31 (2.35)
1.84 (1.84)
1.85 (1.86)
upon coordination (Table 7). Nevertheless, the C-O stretching
frequency decreases in going from the free ligand (2143 cm-1 43
)
to the coordinated one (1927 cm-1). Such a variation of the
C-O stretching frequency upon coordination perfectly matches
our theoretical outcomes which give the values of 2159 and
2044 cm-1 for ν(C-O) in the free and coordinated molecules,
respectively. These results are also in accord with those
obtained for CO chemisorbed on CuCl,44 confirming that
between donation and back-donation interactions, the latter
influences the C-O stretching frequency much more than the
former. In the present case the C-O σ (π) OP moves from
0.054e (0.436e) to 0.254e (0.399e) upon coordination. The
increase of the σ OP is rather close to that computed for CO on
CuCl (0.200e Vs 0.189e),44 while the decrease of the π OP is
twice as large as that computed for CO on CuCl (-0.037e Vs
-0.018e).44 Accordingly, the red shift of the CO stretching
frequency in 3 is significantly larger than that of CO chemi-
sorbed on CuCl.44 Interestingly, the red shift measured for
ν(CO) in the iron derivative 2 (∆ν ) 216 cm-1) is larger than
that of the Ru analog [(PP3)RuH(CO)]BPh42 (∆ν ) 161 cm-1).
In a sense, this finding confirms that, in the present family of
complexes, iron is better suited than ruthenium for π-back-
bonding interactions. In contrast, steric effects might control
the relative stability of ethene complexes 2 and 6, which
increases in the order Fe < Ru. Due to its smaller size, iron
penetrates the cavity of the tripodal ligand more deeply than
ruthenium and thus any bonding interaction with molecules of
a certain size may be stereochemically disfavored (indeed,
ethene is the only alkene that coordinates the metal center in
the [(PP3)Fe(H)]+ system).45
a Key: Pa denotes the bridgehead phosphorus atom of PP3 and Pe
indicates the equatorial phosphorus atoms of PP3; Hi denotes the
terminal hydride ligand. b Mean value.
an open direct transfer mechanism. In particular, Jackson and
Eisenstein have proposed that the nascent bond between the H
and H2 ligands is a factor that might lower the activation energy
of intramolecular H exchange.13a
As in the isostructural Fe complex 1*, the total energy
corresponding to the optimized molecular structure of the η2-
H2 Ru derivative 5*| is more negative by 1.5 kcal mol-1 than
5* (INS studies give a value of 1.36 kcal mol-1 for the barrier
to rotation of the H2 ligand).8 The origin of such a ∆E is
probably the same proposed for 1* as we still observe an
appreciable, even though smaller, bonding interaction between
Hi and H2 (OP ) 0.032e) in the Ru complex. Within this
context, it may be a useful reminder to mention that the main
difference between the electronic structures of 1* and 5* is the
stronger Fe-H2 interaction in the former complex. Actually,
the Fe-H2 OP is dramatically larger than the Ru-H2 one.
Furthermore, the M-H2 distance is definitely shorter in 1* than
in 5*. Bertra´n and co-workers suggest that the origin of the
interaction between the cis hydride and dihydrogen ligands is
electrostatic in nature (due to the negative and positive charges
supported by the H and H2 ligands, respectively).14,15 In this
picture, the higher electronegativity of Ru as compared to Fe,
decreasing the negative charge on the terminal hydride, would
actually reduce the cis effect and ultimately account for the
weaker Ru-H2 bond as well as the slightly higher energy barrier
to the H/H2 exchange determined by NMR spectroscopy for 5
[∆Gq(300 K) ) 13 ( 1 kcal mol-1].2,11a
From the analysis of the theoretical data pertaining to the
ethene complex 2*, one may readily infer that (i) unlike the
η2-H2 derivative 1*, the parallel conformer results less stable
than the perpendicular one by ca. 4.5 kcal mol-1, most likely
due to steric hindrance; and (ii) independent of the assumed
conformation, there is a relevant decrease of the C1-C2 OP in
going from the free ligand to the coordinated one (Table 7).
Further insight into the nature of the Fe-C2H4 interaction can
be obtained by considering the OPs between the Fe 3d and 2pz
AOs of C1 and C2. The Fe 3d t2g-like orbitals participate much
more (0.065e ( ); 0.069e (|)) than the Fe 3d eg-like ones (0.032e
As for the dinitrogen complex 4*, it is noteworthy that,
despite a rather large Fe-N OP (indicative of a quite strong
interaction between the metal center and the dinitrogen ligand),
the N-N OP is substantially unaffected upon coordination. Two
factors can concur to determine this finding: (i) the poor M f
N2 back-donation; (ii) the nonbonding character of the N2-based
2σu MO that is responsible of the N2 f M donation.
In conclusion, the experimental and theoretical observation
that the Fe f L back-bonding interaction in the carbonyl
(43) Huber, K. P.; Herzberg, G. Molecular Spectra and Molecular
Structures. Constants of Diatomic Molecules, Van Nostrand: New
York, 1979; Vol. 4.
(44) Casarin, M.; Favero, G.; Tondello, E.; Vittadini, A. Surf. Sci. 1994,
317, 422.
(45) Bianchini, C.; Peruzzini, M. Unpublished results.

cis-Hydride-η2-Dihydrogen Complexes
Inorganic Chemistry, Vol. 36, No. 6, 1997 1069
Table 7
(A) Effective Atomic Charges for 1*|, 1* , 2*|, 2* , 3* and 4* from Mulliken Population Analysis
Fe
Pa
Pe
Hi
H1
H2
C1
C2
C
O
N1
N2
1*|
1*
2*|
2*
3*
4*
-0.72
-0.72
-0.53
-0.52
-0.58
-0.48
0.11
0.10
0.11
0.10
0.10
0.10
-0.19
-0.20
-0.19
-0.17
-0.20
-0.21
0.06
0.09
0.08
0.06
0.08
0.07
0.13
0.12
0.10
0.11
-0.08
-0.18
-0.19
-0.26
0.23
0.05
-0.18
-0.05
0.06
-0.07
C2H4
CO
-0.10
-0.10
(B) Selected Overlap Populations (e × 10-3) for 1*|, 1* , 2*|, 2* , 3*, and 4*
Fe-H1
Fe-H2
H2-H1
Fe-C1
Fe-C2
C2-C1
Fe-C
C-O
Fe-N1
N1-N2
1*|
1*
164
155
167
164
132
133
2*|
2*
130
117
145
125
423
418
3*
4*
H2free
C2H4free
COfree
N2free
372
653
490
273
541
545
419
716
Table 8
displacement by N2, such as the different entropy of binding of
H2 Vs N246 or weaker Fe-P bond strengths, which would favor
the decoordination of a phosphine arm of PP3 over H2
elimination, can safely be ruled out. The Fe-P bond distances
in 1 fall in the usual range for iron-phosphine complexes and
are even shorter than those determined in [Fe(η2-H2)H(Ph2PCH2-
CH2PPh2)2]BPh4 which readily undergoes the displacement of
H2 by N2.47 Similarly, entropy arguments alone do not justify
the stability of the dihydrogen-hydride Fe complex under a
nitrogen atmosphere as they should equally affect the chemistry
of the Ru complex.
(A) Effective Atomic Charges for 5*| and 5* from Mulliken
Populations Analysis
Ru
Pa
Pe
Hi
H1
H2
5*|
5*
-0.61
-0.60
0.10
0.10
-0.23
-0.24
0.03
0.04
0.11
0.10
0.08
0.109
(B) Selected Overlap Populations (e × 10-3) for 5*| for 5*
Ru-H1
Ru-H2
H2-H1
5*|
5*
115
114
98
103
200
202
Acknowledgment. We thank Mr. Febrizio Zanobini for his
technical assistance.
complex is much more efficient than those in the H2, N2, and
C2H4 derivatives 1, 2, and 4, confirms that the relative stability
of these four complexes is not exclusively dictated by π-back-
donation, and suggests that an important role may be played
by the cis effect operative only when L is H2.
Supporting Information Available: Tables of crystallographic data
for 1‚0.5THF, 3‚THF, and 5‚0.5THF and of final positional parameters
and isotropic and anisotropic displacement parameters for all atoms
for 1‚0.5THF, 3‚THF, and 5‚0.5THF (17 pages). Ordering information
is given on any current masthead page.
Conclusions
IC960736Q
An ab initio study on the model systems [{P(CH2CH2PH2)3}-
M(H)(L)]+ (M ) Fe, L ) H2, C2H4, CO, N2; M ) Ru, L )
H2), X-ray analyses, and XPS and NMR measurements on actual
complexes with the tripodal tetraphosphine PP3, taken altogether,
have confirmed the peculiar effects of the hydride ligand on
the bonding between a cis H2 ligand and the metal center.
Though conclusive evidence accounting for the extrastability
of the (dihydrogen)hydride Fe complex has not been obtained,
both the theoretical and experimental studies are consistent with
greater d(metal) f σ*(H-H) back-donation and cis effect in
the Fe complex as compared to the Ru analog. Other factors
that may account for the stability of the Fe complex toward H2
(46) Gonzales, A. A.; Hoff, C. D. Inorg. Chem. 1989, 28, 4295.
(47) Ricci, J. S.; Koetzle, T. F.; Bautista, M. T.; Hofstede, T. M.; Morris,
R. H.; Sawyer, J. F. J. Am. Chem. Soc. 1989, 111, 8823.
(48) Seminario, J. M. Chem. Phys. Lett. 1993, 206, 547.
(49) CRC Handbook of Chemistry and Physics, 64th ed.; CRC Press: Boca
Raton, FL, 1984; p E-92.
(50) Rabalais, J. W. Principles of UltraViolet Photoelectron Spectroscopy;
J. Wiley Interscience: New York, 1977.
(51) Slater, J. C. Quantum Theory of Molecules and Solids. The Self-
Consistent Field For Molecules and Solids; McGraw-Hill: New York,
1974; Vol. 4.
(52) Herzberg, G. Infrared and Raman Spectra; Van Nostrand: New York,
1945, p 173.