A neutron diffraction study of [OsClH3(PPh3) 3]: A complex containing a highly stretched dihydrogen ligand

Article abstract of DOI:10.1002/anie.200502297

(Figure Presented) Elastic bonds: The structure of [OsHCl(η ²-H₂)(PPh₃)₃] has been determined by neutron diffraction (see figure). The complex presents an unusual H...H contact of 1.48(2) A and can be regarded either as a highly stretched dihydrogen complex or as a compressed dihydride complex.

Full text of DOI:10.1002/anie.200502297

Angewandte
Chemie
Dihydrogen or Dihydride?
mediate, or stretched dihydrogen complexes II (also called
“compressed dihydride” complexes[3a,f,g]) have H H distances
ꢀ
in the range 1.1–1.5 Š.[4] Although a moderate number of
DOI: 10.1002/anie.200502297
ꢀ
stretched dihydrogen complexes have been reported, the H
A Neutron Diffraction Study of [OsClH3(PPh3)3]:
A ComplexContaining a Highly “Stretched”
Dihydrogen Ligand**
H distances in these complexes have mainly been obtained
from NMR spectroscopic data (T1 and JH,D measurements) or
from X-ray data, which are not very sensitive to H positions.[3]
ꢀ
The most reliable technique to determine H H distances
in hydride complexes is neutron diffraction. However, only a
few examples of molecules exhibiting the intermediate stage
II have been structurally characterized by this technique.[4,5]
Intramolecular “reaction trajectories” have been mapped out
before in a classical series of studies of organic molecules by
Bürgi and Dunitz using X-ray crystallography.[7a] In addition,
Crabtree et al.[7b] have used a series of X-ray and some
Muhammed Yousufuddin, Ting Bin Wen, Sax A. Mason,
Garry J. McIntyre, Guochen Jia,* and Robert Bau*
Oxidative addition of dihydrogen to a transition metal is one
of the most important steps in many catalytic processes.[1]
Some years ago, Kubas and co-workers[2] used neutron
diffraction to show for the first time that H2 could bind to a
metal in a sideways fashion as h2-H2 , now called the non-
classical dihydrogen ligand (I). Further insertion of the metal
into the dihydrogen ligand, upon continuation of the oxida-
ꢀ
neutron structures to map the lengthening of a C H bond
during metal insertion, although the variation in those C H
ꢀ
distances (ꢁ 0.1 Š) is much less pronounced than the large
distributions of H···H distances we will discuss later in Table 1.
Several authors have pointed out[3a,h,i] that the energy barrier
for moving the H···H distance from 1.00 to 1.60 Š is only
about 4 kcalmolꢀ1, which suggests that the ligand environ-
ment around the metal atom should have a significant
effect.[3c] Clearly, a more complete set of stretched dihydrogen
ligands is needed to obtain a clearer understanding of this
intriguing interaction. In this regard, we report herein a new
example of the stretched dihydrogen phenomenon in the
complex [OsClH3(PPh3)3] (1).
ꢀ
tive addition process, weakens the H H bond causing it to
lengthen (II) and finally cleave to form two terminal hydride
ꢀ
ligands (III). The weakening of the H H bond (II) is
considered to be an intermediate step along the pathway,
and it is this step that has garnered a considerable amount of
interest.[3–5]
The title complex was prepared by the reaction of
[OsCl2(PPh3)3][8a] with H2 in the presence of NEt3, as
described by Caulton and co-workers.[8b] The compound 1
was first made about 30 years ago, but was wrongly formu-
lated as [OsClH(PPh3)3].[8c] However, a careful analysis of the
1H NMR spectroscopic data by Caulton et al. showed that the
complex 1 was actually [OsClH3(PPh3)3] with a nonclassical
dihydrogen ligand, as suggested by the short relaxation time
(T1 = 26 ms) of the hydride at ꢀ408C and 300 MHz. However,
a minimum T1 value could not be determined due to the low
solubility of the Os complex at low temperatures, nor was it
possible to conclude from the NMR study if the suspected
dihydrogen ligand was trans to a chloride or a phosphane
ligand.[8b] It was also difficult to determine the JH,D coupling
constant for the isotopolog [OsClHD2(PPh)3] due to the
broadness of the hydride signal. Thus, it was impossible to
Structural studies[5] in this area have thus far provided
significant information regarding the metal–dihydrogen inter-
action along the reaction pathway. The side-on, or non-
ꢀ
classical dihydrogen complexes I, have an H H bond length
ꢀ
of the order of 0.8–1.0 Š (not much longer than the H H
distance of 0.74 Š in H2 itself[6]) while classical dihydride
complexes III (with essentially no H···H interactions) gen-
erally have H···H distances greater than 1.7 Š. The inter-
[*] M. Yousufuddin, Prof. Dr. R. Bau
Department of Chemistry
University of Southern California
Los Angeles, CA 90089 (USA)
Fax: (+1)213-740-0930
ꢀ
estimate the H D distance in the complex. In the present
study, the structure has been unequivocally determined by a
single-crystal neutron diffraction analysis.
E-mail: bau@usc.edu
The sample was initially characterized by X-ray diffrac-
tion[9] in order to determine the atomic coordinates used for
neutron diffraction data analysis. The neutron data analysis
was carried out both at room temperature and at 5 K on a
crystal measuring 1.2 ” 0.8 ” 0.3 mm3 using the VIVALDI
diffractometer at the Institut Laue–Langevin in Grenoble.[10]
Full-matrix least-squares refinement of atomic positions
yielded the following agreement factors for the low-temper-
ature data set: R(F) = 0.0831 for 3560 independent reflections
(I > 2s(I)).[11]
Dr. T. B. Wen, Prof. Dr. G. Jia
Department of Chemistry
Hong Kong University of Science and Technology
Clear Water Bay (Hong Kong)
Fax: (+852)23521594
E-mail: chjiag@ust.hk
Dr. S. A. Mason, Dr. G. J. McIntyre
Institut Laue–Langevin
6, rue Jules Howowitz, BP 156
38042 Grenoble Cedex (France)
[**] This work was funded by the American Chemical Society (grant PRF-
40715-AC3) and by the Hong Kong Research Grant Council (project
no. 601804).
Figure 1 shows the molecular sructure of 1 and its core at
5 K. Our results reveal the presence of a dihydrogen ligand
Angew. Chem. Int. Ed. 2005, 44, 7227 –7230
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7227

Communications
a phosphane ligand (H1-Os1-P3 172.5(6)8) while the dihy-
drogen ligand sits essentially trans to the chloride ligand (X-
Os1-Cl1 171.4(3)8, where X represents the center of the
dihydrogen ligand). The dihydrogen, hydride, and chloride
ligands lie in the same plane (Figure 1), which is similar to the
atomic arrangement in [IrCl2H(H···H)(PiPr3)2].[4e] Two other
stretched dihydrogen moieties previously studied by neutron
diffraction, in which the dihydrogen ligands were also located
trans to negatively charged ligands, were also found in six-
coordinate osmium complexes.[4b,c]
Speculation can be made as to the role of the chloride
ligand in influencing the geometry of complex 1. Some years
ago we reported the results of a neutron diffraction analysis of
[OsH4(PMe2Ph)3], which contains four terminal hydrides with
Figure 1. Molecular structure of [OsClH3(PPh3)3] (1, left) and its core
ꢀ
(right). The H2 H3 distance is 1.48(2) Š and the H1···H2 distance is
1.67(2) Š, as determined from this neutron study. Other notable bond
ꢀ
ꢀ
distances include: Os H2=1.59(2) and Os H3 1.61(1) Š.
[12a]
ꢀ
an average Os H bond distance of 1.659(3) Š.
The only
apparent difference between that complex and the present
complex 1 (besides the difference in phosphane ligands) is the
substitution of a chloride ligand for a hydride ligand. Perhaps
the presence of the trans chloride ligand, which is more
electronegative than hydride, promotes closer H···H inter-
actions,[3h] and this could be the driving force for formation of
the stretched dihydrogen ligand.
The main point of our communication is shown in Table 1,
which summarizes experimental data, from which a crude
approximation of a “reaction profile” can be derived, for the
specific process of oxidative addition of a metal atom into an
(H2···H3), although not of the nonclassical type originally
envisioned,[8b] but of the “stretched” variety (II), with an
H···H bond distance of 1.48(2) Š. In addition, one terminal
hydride ligand (H1) was located, plus one chloride and three
ꢀ
phosphane ligands. The Os H1 distance to the terminal
ꢀ
hydride ligand (1.62(2) Š) is similar to other Os H distances
reported in literature,[12] and this terminal hydride ligand H1
is positioned 1.67(2) Š from the H2 atom of the dihydrogen
ligand, a distance which is in the range of slightly nonbonding
interactions.[5] It should be noted that, within experimental
error, we found a barely significant increase in geometry
between the neutron structures refined at 5 K and at
295 K,[13a] which is somewhat consistent with a theoretical
study which suggests that the H···H distance in a stretched
dihydrogen ligand could be temperature dependent.[13b]
The molecule is substantially distorted from an ideal
octahedral geometry. The terminal hydride is located trans to
ꢀ
H H bond. In this table, we have included all neutron
ꢀ
diffraction data on H H complexes that were available to us,
with the exception of the last category at the bottom of the
table (“classical” hydride complexes, in which there are no
H···H bonding interactions), for which neutron results are too
numerous to be listed in full.
Table 1: A selection of H···H distances [Š] and H-M-H angles [8] determined by neutron diffraction.[a]
1
2
3
4
5
6
[c]
ꢀ
ꢀ
ꢀ
Compound
H H
H-M-H
Av. M H
“Adjusted”
av. M H
M X
“Adjusted”
M X
[b]
[b,c]
ꢀ
ꢀ
Nonclassical Complexes
(with “almost unstretched”
H···H distances)
[Mo(CO)(h2-H2)(dppe)2]
[Fe(H)(h2-H2)(dppe)2]+
[W(CO)3(h2-H2)(PiPr3)2]
[FeH2(h2-H2)(PPh2Et)3]
[Ir(H)2(I)(h2-H2)(PiPr3)2]
0.80–0.85[14e]
0.816(16)[14a]
0.82(1)[14b]
24.0–25.6
29.3(5)
25.1
29.9(4)
28.2(3)
1.922(15)
1.616(10)
1.89(1)
1.592(8)
1.756(7)
1.79
1.62
1.76
1.59
1.66
1.886(15)
1.616(10)
1.89(1)
1.538(8)
1.703(7)
1.76
1.62
1.76
1.54
1.60
0.821(10)[14d]
0.856(9)[14c]
Slightly Stretched
H···H distances
[(C5Me5)OsH2(h2-H2)(PPh3)]+
[(C5Me5)Ru(h2-H2)(dppm)]+
[IrCl2H(h2-H2)(PiPr3)2]
1.014(11)[4g]
1.10(2)[4d]
1.11(3)[4e]
35.4(4)
38(1)
42.3(9)
1.669(9)
1.67(2)
1.444(18)
1.58
1.59
1.34
1.589(9)
1.58(2)
1.347(18)
1.50
1.50
1.25
Significantly Stretched
H···H distances
[OsCl(h2-H2)(dppe)2]+
[Os(CH3CO2)(h2-H2)(en)2]+
[ReH7{P(p-tolyl)3}2]
[OsClH3(PPh3)3]
1.22(3)[4c]
1.34(2)[4b]
1.357(7)[4a]
1.48(2)[d]
1.49(4)[4f]
42.6(12)
49.6(6)
47.6(2)
55.1(8)
54.1(1)
1.58(2)
1.60(1)
1.681(4)
1.60(2)
1.64(2)
1.49
1.51
1.57
1.51
1.55
1.47(2)
1.45(1)
1.538(4)
1.42(2)
1.46(2)
1.38
1.36
1.43
1.33
1.37
[OsH5(PMe2Ph)3]+
[e]
[e]
Classical Complexes
(no H···H bonding)
[OsH4(PMe2Ph)3]
[ReH5(PMePh2)3]
[FeH2(CO)2{P(OPh)3}2]
1.840(6)[12a]
1.845(7)[15a]
2.011(3)[15b]
67.9(2)
66.8(3)
82.5(1)
1.646(3)
1.677(5)
1.525(2)
1.56
1.57
1.53
–
–
–
–
–
–
[e]
[e]
[e]
[e]
[a] Abbreviations: dppe=1,2-bis(diphenylphosphanyl)ethane; dppm=1,2-bis(diphenylphosphanyl)methane; en=ethylenediamine. [b] The
ꢀ
ꢀ
“adjusted” M H and M X distances were corrected for the differences in the atomic radii of the metal atoms; the correction term for Fe was
ꢀ
arbitrarily set to be zero. Atomic radii data from Ref. [16]. [c] M X distances involved with the H2 ligand; X is the center of the H···H bond. [d] Present
work. [e] The position of “X” (the mid-point of the H···H bond) is not meaningful for classical hydrides, in which there are no H···H interactions.
7228
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7227 –7230

Angewandte
Chemie
ꢀ
Barea, M. A. Esteruelas, A. Lledos, A. M. Lopez, E. Onate, J. I.
Tolosa, Organometallics 1998, 17, 4065; i) F. Maseras, A. Lledos,
E. Clot, O. Eisenstein, Chem. Rev. 2000, 100, 601.
Columns 1, 2, and 3 list the experimental H H distances,
ꢀ
H-M-H angles, and M H distances, respectively, sorted in the
ꢀ
order of increasing H H distance. One can discern a very
crude general trend of longer H H distances coupled with
shorter M H distances. However, one has to account for
differences in atomic radii of the metal atoms, and so in
[4] a) L. Brammer, J. A. K. Howard, O. Johnson, T. F. Koetzle, J. L.
Spencer, A. M. Stringer, J. Chem. Soc. Chem. Commun. 1991,
241; b) T. Hasegawa, Z. Li, S. Parkin, H. Hope, R. K. McMullan,
T. F. Koetzle, H. Taube, J. Am. Chem. Soc. 1994, 116, 4352;
c) P. A. Maltby, M. Schlaf, M. Steinbeck, A. J. Lough, R. H.
Morris, W. T. Klooster, T. F. Koetzle, R. C. Srivastava, J. Am.
Chem. Soc. 1996, 118, 5396; d) W. T. Klooster, T. F. Koetzle, G.
Jia, T. P. Fong, R. H. Morris, A. Albinati, J. Am. Chem. Soc.
1994, 116, 7677; e) A. Albinati, V. I. Bakhmutov, K. G. Caulton,
E. Clot, J. Eckert, O. Eisenstein, D. G. Gusev, V. V. Grushin,
B. E. Hauger, W. T. Klooster, T. F. Koetzle, R. K. McMullan, T. J.
OꢀLoughlin, M. Pelissier, J. S. Ricci, M. P. Sigalas, A. B. Vyme-
nits, J. Am. Chem. Soc. 1993, 115, 7300; f) T. J. Johnson, A.
Albinati, T. F. Koetzle, J. Ricci, O. Eisenstein, J. C. Huffman,
K. G. Caulton, Inorg. Chem. 1994, 33, 4966; g) C. L. Gross, D. M.
Young, A. J. Schultz, G. S. Girolami, J. Chem. Soc. Dalton Trans.
1997, 3081.
[5] R. Bau, M. H. Drabnis, Inorg. Chim. Acta 1997, 259, 27, and
references therein.
[6] CRC Handbook of Chemistry and Physics (Ed.: D. R. Lide),
75th ed., CRC, Ann Arbor, MI, 1994, p. 9.
[7] a) H. B. Bürgi, J. D. Dunitz, Acc. Chem. Res. 1983, 16, 153;
b) R. H. Crabtree, E. M. Holt, M. Lavin, S. M. Morehouse,
Inorg. Chem. 1985, 24, 1986.
[8] a) P. R. Hoffman, K. G. Caulton, J. Am. Chem. Soc. 1975, 97,
4221; b) G. Ferrando, K. G. Caulton, Inorg. Chem. 1999, 38,
4168; c) A. Oudeman, F. Van Rantwijk, H. J. Van Bekkum, J.
Coord. Chem. 1974, 4, 1.
ꢀ
ꢀ
ꢀ
column 4 we list “adjusted” M H distances in which the
differences in atomic radii are used as “correction factors”,
with the atomic radius of iron selected arbitrarily as a
“standard”. One can see that column 4 is somewhat
“smoother” than column 3. Finally, recognizing that one
should rather consider metal–ligand distances instead of
metal–hydrogen distances, we have provided two extra
columns in which M X distances are tabulated (X is the
midpoint of the H H bond). Once again, we give “unad-
justed” and “adjusted” values (columns 5 and 6, respectively),
in which column 6 contains the values “corrected” for differ-
ꢀ
ꢀ
ꢀ
ences in atomic radii. The differences between M H distances
(column 3) and metal–ligand distances (column 5) are not
large for true nonclassical H2 complexes (top portion of
Table 1), but the differences become more apparent where
“slightly stretched” and “significantly stretched” dihydrogen
complexes are concerned (middle portion of Table 1).
Perhaps the most meaningful comparison is between
column 1 (or column 2) and column 6. Although there is
obviously some scatter in the data, one can discern a definite
ꢀ
trend in which an increasing H H distance appears to be
correlated with a decreasing metal–ligand distance. In other
words, this is experimental evidence for the reaction pathway
I!II!III. A somewhat similar conclusion for metal insertion
[9] T. B. Wen, G. Jia, unpublished results.
[10] C. Wilkinson, J. A. Cowan, D. A. A. Myles, F. Cipriani, G. J.
McIntyre, Neutron News 2002, 13, 37.
into a C H bond has appeared earlier,[7b] but in that case only
ꢀ
[11] Crystals suitable for neutron-diffraction studies were grown by
layering a minimum amount of hexane over a dichloromethane
solution of the compound under an atmosphere of hydrogen. A
triangular prismatic single crystal with a volume of less than
0.5 mm3 was wrapped in aluminum foil in a helium-filled glove
bag and glued to a vanadium pin. The crystal was then mounted
in a He cryostat on the Very-Intense Vertical-Axis Laue
Diffractometer (“VIVALDI”) at the Institut Laue–Langevin in
Grenoble, France. Initially, data collection proceeded at room
temperature until a sufficient number of unique reflections was
obtained for structure refinement. This was a precautionary step
taken just in case the crystal cracked during the subsequent
cooling procedure. The crystal was then cooled to 5 K and data
collection repeated. Crystal data for 1 (unit cell dimensions from
the X-ray analysis[9]): C54H48ClOsP3, Mr = 834, approximate
crystal dimensions 1.2 ” 0.8 ” 0.3 mm3; orthorhombic, space
group Pca21, a = 19.258(1), b = 12.949(1), c = 17.836(1) Š, V=
ꢀ
a small spread of C H distances was observed, much more
subtle than the large spread in H H distances listed in
Table 1. One can conclude from this comparison that the
process of oxidative addition into an H H bond is intrinsically
ꢀ
ꢀ
ꢀ
different to oxidative addition into a C H bond.
Received: June 30, 2005
Revised: August 22, 2005
Published online: October 17, 2005
Keywords: hydride ligands · hydrogen · neutron diffraction ·
.
osmium · oxidative addition
4448(1) Š3, F(000) = 862, Z = 4, 1calcd = 1.245 gcmꢀ3
,
m =
[1] a) G. Kubas, Metal Dihydrogen and s-Bond Complexes: Struc-
ture, Theory, and Reactivity, Kluwer, New York, 2001; b) M.
Peruzzini, R. Poli, Recent Advances in Hydride Chemistry,
Elsevier, Amsterdam, 2001.
[2] G. J. Kubas, R. R. Ryan, B. I. Swanson, P. J. Vergamini, H. J.
Wasserman, J. Am. Chem. Soc. 1985, 107, 451.
[3] Examples of recent work: a) D. M. Heinekey, A. Lledos, J. M.
Lluch, Chem. Soc. Rev. 2004, 33, 175; b) D. G. Gusev, J. Am.
Chem. Soc. 2004, 126, 14249; c) R. Gelabert, M. Moreno, J. M.
Lluch, A. Lledos, V. Pons, D. M. Heinekey, J. Am. Chem. Soc.
2004, 126, 8813; d) P. Barrio, M. A. Esteruelas, A. Lledos, E.
Onate, J. Tomas, Organometallics 2004, 23, 3008; e) B. Eguillor,
M. A. Esteruelas, M. Olivan, E. Onate, O rganometallics 2005,
24, 1428; f) M. Vogt, V. Pons, D. M. Heinekey, Organometallics
2005, 24, 1832; g) R. Gelabert, M. Moreno, J. M. Lluch, A.
Lledos, D. M. Heinekey, J. Am. Chem. Soc. 2005, 127, 5632; h) G.
0.310 mmꢀ1
.
q = 4–728, completeness = 72%, resolution d-
(min) = 0.92 Š, white beam neutron source (l = 0.9–2.7 Š),
T= 5 K, 25311 measured data (VIVALDI instrument with
cylindrical area detector made from Gd2O3-doped BAFBR:Eu2+
image plates), of which 4552 (Rint = 0.2856) were unique.
Intensities were indexed and processed using LAUEGEN,
reflections were integrated and the background removed using
INTEGRATE + , reflections were normalized to a constant
incident wavelength using LAUENORM, and neutron data
were phased using the atomic coordinates determined from the
X-ray structure. For a full description of data collection and
treatment using VIVALDI, see: E. Ding, B. Du, E. A. Meyers,
S. G. Shore, M. Yousufuddin, R. Bau, G. J. McIntyre, Inorg.
Chem. 2005, 44, 2459. Full-matrix least-squares refinement on
F2, data to parameters ratio: 10.6:1, final R indices (I > 2s(I)):
Angew. Chem. Int. Ed. 2005, 44, 7227 –7230
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7229

Communications
R1 = 0.0831, wR2 = 0.1812; R1 = 0.1205, wR2 = 0.1984 for all
data, GOF on F2 = 1.037. CCDC-275803 contains the supple-
mentary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Center via www.ccdc.cam.ac.uk/data_request/cif.
[12] a) D. W. Hart, R. Bau, T. F. Koetzle, J. Am. Chem. Soc. 1977, 99,
7557; b) J. A. K. Howard, O. Johnson, T. F. Koetzle, J. L.
Spencer, Inorg. Chem. 1987, 26, 2930.
ꢀ
[13] a) For data at 295 K, we obtained the following results: H1 H2
ꢀ
1.71(4) and H2 H3 1.56(4) Š, R(F) = 0.1403 for 2818 independ-
ent reflections (I > 2s(I)); b) R. Gelabert, M. Moreno, J. M.
Lluch, A. Lledos, J. Am. Chem. Soc. 1997, 119, 9840.
[14] a) J. S. Ricci, T. F. Koetzle, M. T. Bautista, T. M. Hofstede, R. H.
Morris, J. F. Sawyer, J. Am. Chem. Soc. 1989, 111, 8823; b) G. J.
Kubas, R. R. Ryan, B. I. Swanson, P. J. Vergamini, H. J. Wasser-
man, J. Am. Chem. Soc. 1984, 106, 451; c) J. Eckert, C. M. Jensen,
T. F. Koetzle, T. L. Husebo, J. Nicol, P. Wu, J. Am. Chem. Soc.
1995, 117, 7271; d) L. S. van der Sluys, J. Eckert, O. Eisenstein,
J. H. Hall, J. C. Huffman, S. A. Jackson, T. F. Koetzle, G. J.
Kubas, P. J. Vergamini, K. G. Caulton, J. Am. Chem. Soc. 1990,
112, 4831; e) G. J. Kubas, C. J. Burns, J. Eckert, S. W. Johnson,
A. C. Larson, P. J. Vergamini, C. J. Unkefer, G. R. K. Khalsa,
S. A. Jackson, O. Eisenstein, J. Am. Chem. Soc. 1993, 115, 569.
[15] a) T. J. Emge, T. F. Koetzle, J. W. Bruno, K. G. Caulton, Inorg.
Chem. 1984, 23, 4012; b) N. N. Ho, R. Bau, S. A. Mason, J.
Organomet. Chem. 2003, 676, 85.
[16] J. McMurray, R. C. Fay, Chemistry, 2nd ed., Prentice Hall, New
Jersey, 1998.
7230
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7227 –7230