Dihydrogen with frequency of motion near the 1H Larmor frequency. Solid-state structures and solution NMR spectroscopy of osmium complexes trans-[Os(H··H)X(PPh2CH2CH 2PPh2)2]+ (X = Cl, Br)

Article abstract of DOI:10.1021/ja9529044

A high-yield route to the new complexes OsBr₂(dppe)₂ and trans-OsHBr(dppe)₂ starting from (NH₄)₂-[OsBr₆] is described. The new 5-coordinate complexes [OsX(dppe)₂]PF₆ (X = Cl (80s) and X = Br (90s)) are prepared by reaction of cis-OsX₂(dppe)₂ with NaPF₆. Complexes 80s and 90s consist of distorted trigonal bipyramidal cations with Y -shaped equatorial planes. They react in CH₂Cl₂ with H₂ or HD (1 atm) to give complexes trans-[Os(H··H)X(dppe)₂]PF₆ (X = Cl (10sPF₆), X = Br (20sPF₆)) or trans-[Os(H··D)X(dppe)₂]PF₆, respectively. The last complexes have J(H,D) = 13.9 and 13.7 Hz, respectively. The BF₄^- salts of these complexes, 10sBF₄ and 20sBF₄, respectively, are prepared by reacting trans-OsHX(dppe)₂ with HBF₄·Et₂O or DBF₄·Et₂O. These complexes are characterized by NMR, IR, and FAB MS. The single-crystal X-Yay and neutron diffraction studies of 10sPF₆ revealed an elongated H··H ligand with d_HH = 1.11(6) (X-ray) or 1.22(3) A?(neutron) occupying one site in an octahedral complex. The X-ray diffraction study of 20sBF₄ produced a similar structure with d_HH = 1.13(8) A?. One fluorine of the anion in each structure is positioned near the acidic H₂ ligand. A linear relationship between d_HH and J(H,D) for many dihydrogen complexes is used to indicate that complexes 10s and 20s have H-H distances of about 1.2 A? in solution. Plots of ln(T_l) versus inverse temperature for 10s and 20s are distorted from the usual V shape, suggesting that the rotational frequency of the H₂ ligand is near that of the Larmor frequency. Therefore the d_HH for 10s is between the values of 1.04 and 1.31 A? calculated from the T_l (min) for fast and slow spinning H₂, respectively. The chloride ligand in trans-[Os(H··H)Cl(L)₂]⁺ buffers the effect of changing the cis ligands L from dppe to depe to dcpe so that there is little change in the H-H distance. Complexes 10s and 20s have pK_a values of 7.4 and 5.4, respectively, while trans-[Os(H₂)H(dppe)₂]⁺ is much less acidic with a pK_a of 13.6. These pK_a values and some E_1/2 values are used to show that 10s and 20s are dihydride-like even though they have relatively short H-H distances. Properties of trans-[Os(H··H)Cl(depe)₂]BF₄ (30s) are also reported.

Full text of DOI:10.1021/ja9529044

5396
J. Am. Chem. Soc. 1996, 118, 5396-5407
1
Dihydrogen with Frequency of Motion Near the H Larmor
Frequency. Solid-State Structures and Solution NMR
Spectroscopy of Osmium Complexes
trans-[Os(H‚‚H)X(PPh2CH2CH2PPh2)2]+ (X ) Cl, Br)
Patricia A. Maltby,† Marcel Schlaf,† Martin Steinbeck,† Alan J. Lough,†
Robert H. Morris,*,† Wim T. Klooster,‡ Thomas F. Koetzle,‡ and
Ramesh C. Srivastava‡,§
Contribution from the Departments of Chemistry, UniVersity of Toronto, 80 St. George Street,
Toronto, Ontario M5S 3H6, Canada, and BrookhaVen National Laboratory,
Upton, New York 11973
ReceiVed August 23, 1995. ReVised Manuscript ReceiVed February 5, 1996X
Abstract: A high-yield route to the new complexes OsBr2(dppe)2 and trans-OsHBr(dppe)2 starting from (NH4)2-
[OsBr6] is described. The new 5-coordinate complexes [OsX(dppe)2]PF6 (X ) Cl (8Os) and X ) Br (9Os)) are
prepared by reaction of cis-OsX2(dppe)2 with NaPF6. Complexes 8Os and 9Os consist of distorted trigonal bipyramidal
cations with “Y”-shaped equatorial planes. They react in CH2Cl2 with H2 or HD (1 atm) to give complexes trans-
[Os(H‚‚H)X(dppe)2]PF6 (X ) Cl (1OsPF6), X ) Br (2OsPF6)) or trans-[Os(H‚‚D)X(dppe)2]PF6, respectively. The
-
last complexes have J(H,D) ) 13.9 and 13.7 Hz, respectively. The BF4 salts of these complexes, 1OsBF4 and
2OsBF4, respectively, are prepared by reacting trans-OsHX(dppe)2 with HBF4‚Et2O or DBF4‚Et2O. These complexes
are characterized by NMR, IR, and FAB MS. The single-crystal X-ray and neutron diffraction studies of 1OsPF6
revealed an elongated H‚‚H ligand with dHH ) 1.11(6) (X-ray) or 1.22(3) Å (neutron) occupying one site in an
octahedral complex. The X-ray diffraction study of 2OsBF4 produced a similar structure with dHH ) 1.13(8) Å.
One fluorine of the anion in each structure is positioned near the acidic H2 ligand. A linear relationship between
dHH and J(H,D) for many dihydrogen complexes is used to indicate that complexes 1Os and 2Os have H-H distances
of about 1.2 Å in solution. Plots of ln(T1) versus inverse temperature for 1Os and 2Os are distorted from the usual
“V” shape, suggesting that the rotational frequency of the H2 ligand is near that of the Larmor frequency. Therefore
the dHH for 1Os is between the values of 1.04 and 1.31 Å calculated from the T1(min) for fast and slow spinning H2,
respectively. The chloride ligand in trans-[Os(H‚‚H)Cl(L)2]+ buffers the effect of changing the cis ligands L from
dppe to depe to dcpe so that there is little change in the H-H distance. Complexes 1Os and 2Os have pKa values
of 7.4 and 5.4, respectively, while trans-[Os(H2)H(dppe)2]+ is much less acidic with a pKa of 13.6. These pKa
values and some E1/2 values are used to show that 1Os and 2Os are dihydride-like even though they have relatively
short H-H distances. Properties of trans-[Os(H‚‚H)Cl(depe)2]BF4 (3Os) are also reported.
Introduction
smaller H-H bond energy in 1Ru compared to the one in 5Ru.
Here we establish the H-H distances in the solid-state and in
solution for the corresponding osmium(II) complexes trans-[Os-
(H‚‚H)XL2]+ (L ) dppe, X ) Cl (1Os), X ) Br (2Os), L )
depe, X ) Cl (3Os)). During the course of this work Mezzetti
et al.3 prepared and characterized in solution the related
complexes trans-[Os(H‚‚H)Cl(dcpe)2]BPh4 (4Os) and the os-
mium(IV) trihydride complex [OsH3(dcpe)2]BPh4 (7Os).1 Also
the single-crystal neutron diffraction study of trans-[Os(H‚‚H)-
(OAc)(en)2]PF6 was reported.4 These complexes provide useful
comparisons to those of the present work.
Determining the H-H distance in solution by use of the T1-
(min) method is complicated by the motion of the dihydrogen
ligand. Dihydrogen ligands with rotational frequencies much
greater than the proton Larmor frequency (2πν, where ν is the
NMR spectrometer frequency) have dipolar relaxation rates
which are four times slower than ligands with the same H-H
distances but with rotational frequencies much less than 2πν.
It is already established by use of T1(min) and J(H,D) values
that the complex trans-[Os(H2)H(dppe)2]BF4 (5Os) in solution
Recently we studied the differences in the trans influence of
chloride compared to hydride on the properties of the dihydrogen
ligand in the complexes trans-[Ru(H2)XL2]+ (X ) Cl, L )
dppe1 (1Ru), depe (3Ru); X ) H, L ) dppe (5Ru), depe
(6Ru)).2 Although the T1 and J(H,D) NMR data clearly
indicated that the dihydrogen ligand was elongated in the
chloride complex 1Ru relative to the hydride complex, 5Ru,
the H-H bond length in the former complex could not be
unambiguously established in solution by NMR methods or in
the solid-state for 1RuPF6 by single-crystal X-ray analysis.
However complex 1Ru has a much lower pKa value of 6
compared to the value of 15 for 5Ru; this was related to a
† University of Toronto.
‡ Brookhaven National Laboratory.
§ Permanent address: Department of Physics, Indian Institute of Technol-
ogy, Kanpur 208016, India.
X Abstract published in AdVance ACS Abstracts, May 15, 1996.
(1) Ligand abbreviations: PR2CH2CH2PR2, R ) phenyl (dppe), ethyl
(depe), cyclohexyl (dcpe). Complex numbering: the anion is indicated (e.g.,
1OsPF6) when solid-state properties are discussed. We have no evidence
that solution properties are influenced by the anion. However the solution
NMR properties refer to 1OsPF6, 2OsPF6, and 3OsBF4 unless specified
otherwise.
(3) Mezzetti, A.; Del Zotto, A.; Rigo, P.; Farnetti, E. J. Chem. Soc.,
Dalton Trans. 1991, 1525-1530.
(4) Hasegawa, T.; Li, Z.; Parkin, S.; Hope, H.; McMullan, R. K.; Koetzle,
T. F.; Taube, H. J. Am. Chem. Soc. 1994, 116, 4352-4356.
(2) Chin, B.; Lough, A. J.; Morris, R. H.; Schweitzer, C. T.; D’Agostino,
C. Inorg. Chem. 1994, 33, 6278-6288.
S0002-7863(95)02904-0 CCC: $12.00 © 1996 American Chemical Society

1
Dihydrogen with Frequency Near H Larmor Frequency
J. Am. Chem. Soc., Vol. 118, No. 23, 1996 5397
crude material is pure enough for subsequent reactions. However it
can be recrystallized in a dinitrogen atmosphere. cis-OsBr2(dppe)2
(0.044 g), acetone (7 mL), and benzene (3 mL) were heated, and the
warm solution was filtered through a Kimwipe tissue suspended in a
Pasteur pipette. The volume was reduced by one-half and the pale
yellow solution transferred to a vial. An equal volume of hexanes
was carefully added, and diffusion was allowed to occur overnight.
The pale-yellow microcrystalline solid was filtered from the pale-
yellow liquor and dried in Vacuo. Anal. Calcd for C52H48OsP4Br2:
C, 54.46; H, 4.22. Found: C, 54.57; H, 4.30. FAB MS: calcd for
C52H48192OsP479Br2 1146.1. FAB MS obsd (2-nitrophenyl octyl ether
matrix): 1146.5 (M+), 1067.5 (M+ - Br), 987.6 (M+ - H,2Br), 985.7
has a rapidly spinning dihydrogen ligand with an H-H distance
of 0.99 ( 0.01 Å.5 It has an H-H distance of 0.99 Å according
to a solid-state NMR study.6 The complexes trans-[Os(H‚‚H)-
(OAc)(en)2]PF6 and [Ru(C5Me5)(H‚‚H)(dppm)]BF4 are estab-
lished to have H2 ligands with slow rotation rates (rate , ν).7
Recently Jalon et al. have shown that [Nb(C5H4SiMe3)2(HD)-
(PMe2Ph)]+ has a slowly rotating, elongated H‚‚D ligand.8 For
the complex trans-[Os(H‚‚H)H(depe)2]+ (6Os),9,10 which con-
tains the depe ligand which is more electron-donating than dppe,
a rapid equilibrium has been proposed such that the observed
NMR properties are the weighted average of the those of the
dihydrogen hydride, [Os(H2)H(L)2]+, and trihydride ([OsH3-
(L)2]+) forms.5 Such a proposition was required to explain the
solvent-dependent T1(min) values and the solvent- and temper-
ature-dependent J(H,D) coupling of [Os(H‚‚D)D(depe)2]+. In
a communication of the present work, the complex 3Os was
also proposed to have such a rapid equilibrium.11 Here we
provide the first evidence for complexes of a dihydrogen ligand
that have a frequency of motion near the 1H Larmor resonance
frequency.
2
(M+ - 3H,2Br). NMR δ(31P, CH2Cl2): 1.2 (t, J(P,P) 5.4), -1.3 (t).
Preparation of [OsCl(dppe)2]PF6 (8Os). cis-OsCl2(dppe)2 (300 mg,
0.27 mmol) and NaPF6 (100 mg, 0.8 mmol) were dissolved in a mixture
of 15 mL of THF and 15 mL of CH2Cl2. The reaction was stirred at
room temperature for 48 h to give a dark-brown solution. The reaction
mixture was evaporated to dryness and redissolved in 15 mL of CH2-
Cl2. The dark-brown solution was filtered through Celite/cotton wool
to remove the excess NaPF6 and the NaCl formed in the reaction.
This reaction solution can be used to prepare 1Os as described below.
Dark-red crystals of 8Os suitable for X-ray diffraction analysis15
were obtained within 5 d by slow diffusion of diethyl ether into a
CH2Cl2 solution of 8Os at room temperature. NMR δ(31P, 121.5 MHz,
CH2Cl2): 48.4 (s), 26.0 (s), -143 ppm (1J(P,F) 700 Hz).
Experimental Section
All operations were conducted under a purified nitrogen or argon
atmosphere using vacuum line and glovebox techniques. All solvents
were dried over appropriate reagents and distilled under N2 before use.
Deuterated solvents were dried over Linde-type 4 Å molecular sieves
and degassed with several evacuate/refill cycles prior to use. Reagent-
grade chemicals were used as purchased from Aldrich Chemical Co.,
Inc., unless otherwise stated. Phosphine ligands were purchased from
Strem Chemical Co. or Digital Speciality Chemicals Ltd. The
complexes (NH4)2[OsBr6],12 cis-OsCl2(dppe)2,13 trans-[Os(H2)H(dppe)2]-
BF4 (5Os),5,14 and OsH2(dppe)2,5,14 were prepared as described previ-
ously.
Preparation of [OsBr(dppe)2]PF6 (9Os). This was prepared from
cis-OsBr2(dppe)2 by the same method used for 8Os. NMR δ(31P,
CH2Cl2): 51.1 (br), 23.5 (br), -143 ppm (1J(P,F) 700 Hz). The reaction
solution was used to prepare 2OsPF6 as described below.
Preparation of trans-OsHCl(dppe)2 (10Os). trans-[OsHCl-
(dppe)2]‚C6H6 was first prepared by Chatt and Hayter in 40% yield by
the reaction of cis-OsCl2(dppe)2 and excess LiAlH4. The ν(OsH) mode
was reported at 2046 cm-1 16
We obtained a mixture of OsH2(dppe)2
.
and trans-OsHCl(dppe)2 by following their method. The hydridochloro
complex was prepared from 5Os by the displacement of dihydrogen
by chloride as described below.
Microanalytical results and FAB MS (NBA matrix), IR, and 1H and
31P NMR spectra were obtained as previously described.2 31P NMR
were proton decoupled, except for one experiment involving trans-
[Os(H‚‚D)Cl(dppe)2]BF4. 1H NMR spectra were simulated by use of
the FIRSTORDER simulation program written by T. Burrow. T1
measurements were made at 500, 400, and 300 MHz using the
inversion-recovery method. The concentration of the sample must
be the same for a comparison to be made of results obtained at different
magnetic fields; therefore, the sample was sealed in a 5 mm NMR
tube for the measurements. Pulses were calibrated to exactly 90° and
180° at all reported temperatures. The temperatures were calibrated
by use of the methanol method.
Complex 5Os (0.105 g, 0.092 mmol) was dissolved in acetone (10
mL), and then LiCl (0.025 g, 0.60 mmol) was added under N2. The
mixture was stirred, and after 45 min, a pale-yellow precipitate had
formed. After 3 h a 31P NMR spectrum of the reaction liquor indicated
that no 5Os remained in solution and the reaction was therefore
complete. The acetone was removed under vacuum, and the benzene-
soluble portion was isolated by filtration through Celite. The benzene
was removed under vacuum. Acetone (5 mL) was added, and the
mixture was cooled. The pale-yellow solid was collected by filtration
and dried under vacuum (0.070 g, 74%). Anal. Calcd for C52H49-
ClOsP4: C, 61.02; H, 4.83. Found: C, 60.59; H, 4.74. FAB MS:
calcd for C52H4935Cl192OsP4 1024, observed 1025 (MH+ ) 1Os) and
987.7 (M+ - H,Cl). NMR δ(1H, C6D6, 200 MHz): 7.6-6.8 (m, 40
H, Ph), 2.57 (m, 4 H, CH2), 2.00 (m, 4 H, CH2), -20.40 (qnt, 1H,
2J(H,P) ) 15.4 Hz, OsH). NMR δ(31P, C6D6): 30.5 (s). IR (Nujol):
2124 cm-1 (w, ν(OsH)).
Preparation of trans-OsHBr(dppe)2 (11Os). The following se-
quence of chemicals was added to a 1 L two-neck round-bottom flask:
(NH4)2[OsBr6] (1.21 g, 1.72 mmol), dppe (1.50 g, 3.76 mmol), 250
mL of EtOH, and 250 mL of MeOH. A glass stopper and a condenser
fitted with a gas inlet were attached, and the gray slurry was degassed
using several vacuum and argon refill cycles. The reaction mixture
was refluxed for 19 h under Ar, and a pale-yellow solid in a yellow
liquor was formed. The mixture was allowed to cool and the solvent
volume reduced by one-half. The solid was filtered, washed with
MeOH (15 mL), and dried in Vacuo; 1.007 g (55% yield) of trans-
OsHBr(dppe)2 was obtained. FAB MS: calcd for C52H49192OsP479Br
1068, obsd 1069 (MH+ ) 2Os), 988 (overlapping 988 (M+ - H, Br)
and 990 (MH+ - Br) envelopes). IR(Nujol): 2174 (w, ν(OsH)), 1952,
1988, 1811, 1732 cm-1 (br, vw, δ(CH) overtones). NMR δ(31P, CH2-
Cl2): 28.7 (s). NMR δ(1H, CD2Cl2): 7.35-6.95 (m, 40H, Ph), 2.66
(m, 4H, CH2), 2.10 (m, 4H, CH2), -20.38 (qnt, 1H, J(H,P) ) 15.3 Hz,
OsH).
Preparation of cis-OsBr2(dppe)2. (NH4)2[OsBr6] (0.267 g, 0.378
mmol), dppe (0.608 g, 1.53 mmol), 70 mL of EtOH, and 60 mL of
MeOH were added to a 250 mL two-neck round-bottom flask.
A
condenser was attached, and the dark slurry was purged by use of
several vacuum/Ar cycles. The mixture was refluxed under Ar for 33
h; a pale-yellow solid in a yellow liquor was observed. The volume
was reduced to 100 mL, and the pale-yellow solid was filtered from
the solution. The solid was washed with MeOH (10 mL) and dried in
Vacuo. cis-OsBr2(dppe)2 (0.410 g) was obtained (92% yield). This
(5) Earl, K. A.; Jia, G.; Maltby, P. A.; Morris, R. H. J. Am. Chem. Soc.
1991, 113, 3027-3039.
(6) Wisniewski, L. L.; Zilm, K. W. Private communication.
(7) Klooster, W. T.; Koetzle, T. F.; Jia, G.; Fong, T. P.; Morris, R. H.;
Albinati, A. J. Am. Chem. Soc. 1994, 116, 7677-7681.
(8) Jalon, F. A.; Otero, A.; Manzano, B. R.; Villasenor, E.; Chaudret,
B. J. Am. Chem. Soc. 1995, 117, 10123-10124.
(9) The H‚‚H or H‚‚D notation signifies the probable presence of an
elongated (>1.0 Å) H2 or HD ligand with J(H,D) < 25 Hz. See ref 10.
(10) Jessop, P. G.; Morris, R. H. Coord. Chem. ReV. 1992, 121, 155-
284.
(11) Cappellani, E. P.; Maltby, P. A.; Morris, R. H.; Schweitzer, C. T.;
Steele, M. R. Inorg. Chem. 1989, 28, 4437-4438.
(12) Dwyer, F. P.; Hogarth, J. W. Inorg. Synth. 1957, 5, 204.
(13) Chatt, J.; Hayter, R. G. J. Chem. Soc. 1961, 896-904.
(14) Cappellani, E. P.; Drouin, S. D.; Jia, G.; Maltby, P. A.; Morris, R.
H.; Petroff, A.; Schweitzer, C. T.; Xu, W. Manuscript in preparation.
(15) Lough, A. J.; Morris, R. H.; Schlaf, M. Acta Crystallogr., Sect. C,
in press.
(16) Chatt, J.; Hayter, R. G. J. Chem. Soc. 1961, 2605-2611.

5398 J. Am. Chem. Soc., Vol. 118, No. 23, 1996
Maltby et al.
Preparation of trans-OsHCl(depe)2 (12Os). The complex was
prepared using a modified version of Chatt and Hayter’s method; NMR
data were not given when this complex was first described.13 A slurry
of LiAlH4 (0.073 g, 1.9 mmol) in ca. 10 mL of THF was added to a
stirred pale-yellow solution of cis-OsCl2(depe)2 (0.400 g, 0.59 mmol)
in 15 mL of THF. The gray slurry was refluxed for 25 min and allowed
to cool. Excess LiAlH4 was neutralized by the controlled addition of
7 mL of ethanol (dihydrogen gas is evolved at this step). After solvent
evaporation, the gray residue was treated with a total of 45 mL of THF
and the bulk of the lithium salts were removed by filtration through
THF-saturated Celite. The pale-yellow filtrate volume was reduced
to dryness, NaCl (0.049 g, 0.84 mmol) and acetone (30 mL) were added,
and the milky mixture was refluxed for 2.5 h. The trans-OsHCl(depe)2
product was isolated in the following manner. Acetone was removed
under vacuum, and the portion soluble in benzene (45 mL in total)
was filtered through benzene-saturated Celite. The pale yellow filtrate
volume was reduced to dryness and a minimal volume of acetone (e5
mL) added. The yellow solution was cooled for 1 day, and the resulting
white needles were removed from the yellow liquor. A total of 0.185
g (47% yield) were obtained. NMR δ(1H, benzene-d6): -22.8 (qnt,
J(H,P) ) 15.3 Hz). NMR δ(31P, acetone): 33.4 (s). IR (Nujol): 2050
cm-1 (m, ν(OsH)).
was removed, and the solid was washed with diethyl ether by two
agitation/decantation cycles and then dried under vacuum. This was
dissolved in CD2Cl2 or acetone-d6. Method 2. Complex 8Os in CD2-
Cl2 was reacted with HD gas (generated from NaH and D2O); this gave
a purer solution than method 1. NMR δ(1H, CD2Cl2, 500 MHz):
-11.58 (1:1:1 t of qnt, 1H, J(H,D) ) 13.9 ( 0.2, J(H,P) ) 11.7, Os-
(H‚‚D)). NMR δ(31P, CD2Cl2): 20.9 (s).
Preparation of trans-[Os(H‚‚H)Br(dppe)2]BF4 (2OsBF4). trans-
OsHBr(dppe)2 (0.393 g, 0.359 mmol) was partially dissolved in 5 mL
of CH2Cl2. Approximately 90 µL of HBF4‚Et2O was added in portions
over a 2-3 min period with stirring. The pale-yellow suspension turned
dark brown at first but then became pale brown within 10 min. The
solution was layered with diethyl ether, and the precipitation of large,
clear, pale-brown crystals started immediately. Yield: ≈80%. NMR
δ(1H, CD2Cl2, 400 MHz): -11.45 (br qnt, 2J(H, P) ) 11.5 Hz, OsH2).
NMR δ(31P, CH2Cl2, 121.47 MHz): 17.6 (s).
Preparation of trans-[Os(H‚‚H)Br(dppe)2]PF6 (2OsPF6). This
complex was prepared by starting from 9Os in a fashion similar to the
preparation of 1OsPF6. NMR δ(1H, CD2Cl2, 400 MHz): -11.46 (qnt,
2J(H,P) ) 11.9 Hz, OsH2). NMR δ(31P, CH2Cl2, 121.47 MHz): 17.6
(s), -143 ppm (1J(P,F) 700 Hz). T1 data are provided in Table 3.
Observation of trans-[Os(H‚‚D)Br(dppe)2]+ (2Os-d1). Complex
9Os in CD2Cl2 was reacted with HD gas as above. NMR δ(1H, CD2-
Cl2, 293 K, 400 MHz): -11.6 (1:1:1 t of qnt, 2J(H,P) 11.9 Hz, 1J(H,D)
13.75).
Preparation of trans-[Os(H‚‚H)Cl(depe)2]BF4 (3Os). trans-OsH-
Cl(depe)2 (0.100 g, 0.156 mmol) was partially dissolved in 7 mL of
diethyl ether. Under 1 atm of dinitrogen a diluted HBF4‚Et2O solution
was added dropwise with stirring until a white precipitate formed. The
mixture was cooled and the liquor removed by decantation. Fresh
diethyl ether was added, and the decantation process was repeated. The
white solid was filtered from cold diethyl ether and dried in Vacuo
(0.074 g, 65% yield). The product was crystallized from acetone by
diffusion of diethyl ether vapor. Anal. Calcd for C20H50BF4ClOsP4:
C, 33.05; H, 6.93. Found: C, 32.98; H, 7.00. NMR δ(1H, acetone-
d6, 400 MHz): 2.45 (m, 4 H, PCH2CH2P), 2.31 (m, 4 H, PCH2CH2P),
2.0-1.75 (m, 16 H, CH2), 1.18 (m, 12H, CH3), 1.09 (m, 12H, CH3),
-13.81 (qnt, 2H, 2J(H,P) ) 11.7 Hz, OsH2). NMR δ(31P, acetone, 81
MHz): 25.8 (s). IR (Nujol): 1050 cm-1 (s, BF4-).
Preparation of trans-[Os(H‚‚H)Cl(dppe)2]BF4 (1OsBF4). trans-
OsHCl(dppe)2, (10Os) (0.060 g, 0.059 mmol) was suspended in 15
mL of diethyl ether under 1 atm of dinitrogen. HBF4‚Et2O (∼7 µL,
excess) was added dropwise with stirring. The mixture was stirred for
10 min and then cooled. The liquor was removed from the cream-
colored solid by decantation and replaced with fresh diethyl ether (15-
20 mL). The slurry was stirred and allowed to cool, and the solvent
containing any excess HBF4‚Et2O was removed by decantation. The
washing cycle was repeated several times, and the product was then
filtered and dried under vacuum (0.060 g, 92%). The compound is
soluble in CH2Cl2 but has a low solubility in THF or acetone. FAB
MS: calcd for C52H5035Cl192OsP4 1025.2, obsd 1025.3 (M+), 1023.3
(M+ - 2H), 989.3 (M+ - H,Cl), and 987.3 (M+ -3H,Cl). NMR δ(1H,
CD2Cl2, 400 MHz): 7.4-6.9 (m, 40 H, Ph), 2.83 (m, 4 H, CH2), 2.23
2
(m, 4 H, CH2), -11.45 (qnt, 2H, J(H, P) ) 11.6 Hz, OsH2). NMR
δ(31P, CH2Cl2, 121.47 MHz): 20.8 (s). NMR δ(31P, THF, 121.47
MHz): 23.1 (s). T1 data are provided in Table 3 (see below). The
NMR data previously reported11 for the complex trans-[Os(H‚‚H)Cl-
(dppe)2]BF4 are in error. NMR signals emanating from the complex
trans-[Os(CH3CN)H(dppe)2]+, formed from 5Os and acetonitrile present
in the solvents as a contaminant, were erroneously assigned to the trans-
[Os(H‚‚H)Cl(dppe)2]+ complex.
Acetone solutions of 3Os turn a wine-red color after several days at
room temperature; broad resonances are observed in the 1H NMR
spectrum: δ(1H, acetone-d6, 200 MHz) 8.3, 6.4, 2.7, -0.7, -4.3, -10.4,
-24.6.
pKa Determination of 1Os. A sample of the complex (26 mg, 0.02
mmol) was dissolved in 4 mL of CH2Cl2. The 31P NMR spectrum
was a single peak at 20.75. An aliquot solution of PPhMe2 (30 mg) in
CH2Cl2 (10 mL) was prepared. One equivalent (1 mL, 0.02 mmol of
PPhMe2) of this solution was added to the solution of the complex and
the 31P NMR acquired immediately using a gated decoupling 31P{1H}
routine with delay times of 10 s between pulses. An NMR spectrum
recorded 1 h later did not show any change in spectral appearance and
relative integral values. A second equivalent of the phosphine solution
was added and the 31P NMR recorded with the relative integral values
Clear colorless crystals of 1OsBF4 suitable for X-ray analysis were
obtained by slow vapor diffusion of diethyl ether into a CH2Cl2 solution
of the complex over a 14 h period under N2. Details of this structure,
which was not completely solved because of disorder problems, can
be found elsewhere.17 The monoclinic cell dimensions at 173 K were
found to be a ) 15.808(2) Å, b ) 16.967(3) Å, c ) 17.614(4) Å, and
â ) 102.44(1)° in the space group P21/n with Z ) 4, formula wt
1111.32, and V ) 4613.4(15) Å3.
Preparation of trans-[Os(H‚‚H)Cl(dppe)2]PF6 (1OsPF6). Dihy-
drogen gas was bubbled through a dark-brown solution of 8Os in
CH2Cl2. Within 2 min, the solution cleared and the color changed to
a light green. The solution was layered with an equal volume of di-
ethyl ether. Within 1 week at room temperature X-ray-grade
light-green crystals formed coincident with decolorization of the
mother liquor. Overall yield starting from cis-OsCl2(dppe)2 to 8Os
and then 1OsPF6 was 85%. The structure was determined by single-
crystal X-ray and neutron diffraction studies (see below). The NMR
data are identical to those of 1BF4 (see above) apart from the 31P
showing a clear shift of the equilibrium toward protonated phosphine
+
HPPhMe2
. Over time some trans-[Os(PPhMe2)Cl(dppe)2]PF6 is
formed by substitution of the H‚‚H with free phosphine. This reaction
is slow and does not influence the acid-base equilibrium. Observed
NMR shifts: PPhMe2, -46.08 (s); [HPPhMe2]+, -0.35 (d, J(H,P) )
515.2 Hz); trans-[Os(H‚‚H)Cl(dppe)2]PF6, +20.75 (s); trans-OsHCl-
(dppe)2, 30.41 (s); trans-[Os(PPhMe2)Cl(dppe)2]PF6, -63.35 (s); trans-
[Os(PPhMe2)Cl(dppe)2][PF6], 22.45 (s). For PPhMe2 and [HPPhMe2]+
the reported shifts represent extreme values for complete deprotonation/
protonation. In the presence of both the protonated and unprotonated
form the peaks coalesce to a single broad signal. The observed shift
is the weighted average of the two extreme values, i.e. ppm(obsd) )
x ppm(PPhMe2) + y ppm([HPPhMe2]+) with x and y as the relative
concentrations.
pKa Determination of 2Os. 11Os (50 mg, 0.047 mmol) and P(p-
CH3OC6H4)3 (15 mg, 0.04 mmol) are dissolved in 5 mL of CH2Cl2.
An 85% solution of HBF4‚Et2O (8 µL, 0.05 mmol) in Et2O was added
and the 31P NMR spectrum acquired immediately using a gated
decoupling 31P{1H} routine with delay times of 10 s between pulses.
-
resonance of the PF6 anion which appears as a septet at -143 ppm,
1J(P,F) 700 Hz.
Observation of trans-[Os(H‚‚D)Cl(dppe)2]+ (1Os-d1). Method 1.
trans-OsHCl(dppe)2 (10Os) (0.035 g, 0.034 mmol) was suspended,
under 1 atm of argon, in 13 mL of diethyl ether containing 0.07 mL of
D2O. Approximately 120 µL of HBF4‚Et2O was then added to produce
DBF4 in situ. The mixture was stirred for 25 min. After 5 min the
suspension color had turned from pale yellow to white. The liquor
(17) Maltby, P. A. Ph.D. Thesis, University of Toronto, 1993.

1
Dihydrogen with Frequency Near H Larmor Frequency
J. Am. Chem. Soc., Vol. 118, No. 23, 1996 5399
Table 1. Crystal Data, Details of Intensity Measurements, and Structure Refinements for trans-[Os(H‚‚H)X(PPh2CH2CH2PPh2)2]Y (X ) Cl, Y
) PF6 (1Os) and X ) Br, Y ) BF4 (2Os))
X-ray, 1Os
neutron, 1Os
X-ray, 2Os
empirical formula
fw
C52H50ClF6OsP5
1169.4
C52H50ClF6OsP5
1169.4
C52H50BrOsP4BF4
1155.7
crystal system
space group
cell dimensions
a, Å
monoclinic
P21/c
monoclinic
P21/c
monoclinic
P21/n
15.054(3)
17.295(3)
15.142(8)
17.283(5)
15.690(2)
17.259(4)
b, Å
c, Å
18.481(4)
101.37(3)
4717(1)
18.47(3)
101.5(1)
4736(5)
17.508(2)
102.51(2)
4629(1)
â, deg
V, Å3
T, K
110
4
1.647
29.9
20.0(5)
4
1.640
1.896
150
4
1.658
38.1
Z
d
calcd, g cm-3
µ, cm-1
radiation
Mo KRj
neutrons, λ ) 1.163 95(10) Å
Mo KRj
2θ range, deg
scan type
scan speed
crystal size, mm
no. of reflns measd
no. of reflns used
R(F) [Rw(F)]a
R(F2) [Rw(F2)]b
goodness of fitc
min, max peak in ∆F map, e Å3
2.0-54.0
5.0-108.0
4.0-54.0
ω:2θ
ω:2θ
ω:2θ
variable; 6-60 deg/min in ω
0.6 × 0.4 × 0.15
8524
variable step scans
variable; 6-60 deg/min in ω
0.42 × 0.41 × 0.39
10007
2.5 × 1.8 × 1.3
6521
6612 (F > 6.0σ(F))
0.0297 [0.0368]
6062
d
7881 (F > 4.0σ(F))
0.034 [0.036]
0.28 [0.15]
1.85
+0.95, -0.99
1.46d
1.16
+1.52, -1.89
e
a R(F) ) Σ∆/Σ|Fo|; Rw ) Σ∆w1/2/Σ|Fo|w1/2, where ∆ ) ||Fo| - |Fc|| and w-1 ) σ2(F) + 0.0001F2 for 10sPF6 or w-1 ) σ2(F) + 0.0002F2 for
2
4
2
20sBF4. b R(F2) ) Σ∆/ΣFo ; Rw(F2) ) [Σw∆2/Σw2Fo ]1/2 where ∆ ) |Fo - Fc2|, w-1 ) σ2(F2). c S ) [Σ(w∆2)/(n - p)]1/2, where n is the number
of refln used and p is the number of parameters refined. d For comparison, the R(F) and Rw(F) values for 2407 reflections with (F > 3.0σ(F)) are
0.102 and 0.064, respectively, with goodness of fit 1.80. e Minimum and maximum residuals were -5.3 and +8.9% of the maximum peak for the
Os. These are associated with disordered phenyl rings.
An NMR spectrum recorded 1 h later did not show any change in
spectral appearance and relative integral values. Observed NMR
shifts: P(p-CH3OC6H4)3, -10.88 (s); [HP(p-OCH3C6H4)3 ]+, 1.81 (d,
J(H,P) ) 511 Hz); trans-OsHBr(dppe)2, 28.65 (s); trans-[Os(H‚‚H)-
Br(dppe)2]PF6, 17.52 (s).
port H6S of the high flux beam reactor at Brookhaven National
Laboratory. The neutron beam, monochromated by Ge (220) planes
in transmission geometry, was of wavelength 1.163 95(10) Å as
calibrated against a KBr crystal (a0 ) 6.6000 Å at 295 K). The sample
temperature was maintained at 20.0 ( 0.5 K during the experiment,
and unit cell dimensions were determined by least-squares fit of sin2-
(θ) values for 32 reflections in the range 41 < 2θ < 53°. Intensity
data were obtained over one quadrant of reciprocal space by means of
ω-2θ scans. The intensities of two reflections were monitored
periodically during the data collection and showed no systematic
variations throughout. Integrated intensities I0 and variances σ2(I0) were
derived from the scan profiles. Lorentz factors were applied, as well
as an absorption correction. Transmission factors were in the range
0.683-0.789. Averaging over 155 symmetry-related pairs of reflections
resulted in an internal agreement factor of 0.072 and yielded 6113
independent observations. Further details are given in Table 1. Initial
coordinates were obtained from the X-ray structure. Least-squares
refinements were carried out by a full-matrix procedure,20 minimizing
Σw(Fo2 - k2Fc2)2 using all independent data minus 51 reflections, which
were affected by the powder lines from the aluminum container. The
final model included the scale factor k and positional and anisotropic
displacement parameters âij for all atoms, except for four P atoms and
one disordered phenyl group, which were treated isotropically. Also,
for some of the atoms, the cross terms âij were set to zero, since these
caused the displacement tensors to go nonpositive definite, and in any
case, the âij values in question did not differ significantly from zero.
The isotropic type I extinction parameter21 was omitted, since it failed
to assume a significant value. This gave a total of 916 variable
No substitution reaction or coalescence of the phosphine signals is
observed in the reaction of P(p-CH3OC6H4)3 with 2Os.
X-ray Structural Characterization of 1OsPF6 and 2OsBF4. A
green needle of 1OsPF6 and a pale-brown crystal of 2OsBF4 were
mounted on glass fibers in the air, placed on a Siemens P4 diffracto-
meter, and cooled in a stream of cold nitrogen for data collection.
A summary of selected crystallographic data are given in Table 1.
Mo KRj radiation (λ ) 0.710 73 Å) was used. For each compound the
intensities of three standard reflections measured every 97 reflections
showed no decay. The data were corrected for Lorentz and polarization
effects and for absorption18 (minimum and maximum transmission
coefficients were 0.3785 and 0.8039 for 1OsPF6 and 0.5432 and 0.6438
for 2OsBF4).
The structures were solved and refined using the SHELXTL\PC19
package. All non-hydrogen atoms were refined with anisotropic thermal
parameters to minimize ∑w(Fo - Fc)2, where w-1 ) σ2(F) + 0.0001F2
for 1OsPF6 and w-1 ) σ2(F) + 0.0002F2 for 2OsBF4. Hydrogen atoms
were included in calculated positions and treated as riding atoms. An
overall hydrogen atom thermal parameter was refined to 0.029(2) Å2
for 1OsPF6 and 0.038(2) Å2 for 2OsBF4. The positions of the
dihydrogen atoms were determined from difference electron density
maps and were refined with isotropic thermal parameters. Selected
bond lengths and angles are given in Table 2. The structure of the
cation 1OsPF6 is shown in Figure 1.
2
parameters. The refinement converged with fit indices R(Fo ) ) 0.276,
2
Rw(Fo ) ) 0.151, and Rw(Fo) ) 0.081, S ) 1.46, based on 6062
Neutron Structure Determination. Crystals were obtained by slow
diffusion of Et2O into a saturated solution of 1OsPF6 in CH2Cl2. A
light-green specimen of volume 6.0 mm3 was mounted on an aluminum
pin with halocarbon grease and sealed under a helium atmosphere inside
an aluminum container. This container was placed in a closed-cycle
helium refrigerator and mounted on the four-circle diffractometer at
reflections. A rigid-body analysis22 was done to correct the H-H
distance for the librational motion of the H2 ligand. The whole molecule
except for the H2 ligand was one rigid body with the H2 ligand forming
the other rigid body. The librational axis was from the Os to the
(20) Lundgren, J.-O. Crystallographic Computer Programs, Report UUIC-
B13-4-05, Institute of Chemistry, University of Uppsala, Uppsala, Sweden,
1982.
(18) Sheldrick, G. M. SHELXA-90 absorption correction program,
University of Go¨ttingen, Germany.
(19) Sheldrick, G. M. SHELXTL-PC, Siemens Analytical X-ray Instru-
ments Inc., Madison, WI.
(21) Becker, P. J.; Coppens, P. Acta Crystallogr. 1974, A30, 129-144.
(22) Craven, B. M. Program EKRT, University of Pittsburgh, 1992.

5400 J. Am. Chem. Soc., Vol. 118, No. 23, 1996
Maltby et al.
Table 2. Selected Bond Lengths (Å) and Angles (deg) for
trans-[Os(H‚‚H)X(PPh2CH2CH2PPh2)2]Y (X ) Cl, Y ) PF6
(1OsPF6) and X ) Br, Y ) BF4 (2OsBF4))
X-ray, 1OsPF6 neutron, 1OsPF6 X-ray, 2OsBF4
Os-HOs1
Os-HOs2
HOs1-HOs2
Os-P1
1.41(5)
1.67(5)
1.11(6)
1.59(2)
1.57(2)
1.15(3)
2.39(1)
2.39(1)
2.37(1)
2.37(1)
2.44(1)
1.85(1)
1.83(1)
1.84(1)
1.84(1)
1.83(1)
1.85(1)
1.90(2)
1.78(2)
1.83(2)
1.84(1)
1.83(1)
1.83(1)
1.52(1)
1.48(2)
1.52(6)
1.64(6)
1.13(8)
2.391(1)
2.379(1)
2.382(1)
2.368(1)
2.420(1)
1.860(5)
1.819(5)
1.832(5)
1.832(5)
1.820(4)
1.834(5)
1.843(5)
1.797(6)
1.827(5)
1.819(5)
1.825(5)
1.819(4)
1.523(6)
1.502(6)
2.414(1)
2.372(1)
2.398(1)
2.369(1)
2.560(1)
1.849(5)
1.836(5)
1.827(5)
1.828(5)
1.816(5)
1.829(5)
1.840(4)
1.819(6)
1.820(5)
1.832(5)
1.825(5)
1.828(5)
1.521(7)
1.518(7)
Os-P2
Os-P3
Os-P4
Os-Cl (Br)
P1-C1
P1-C11
P1-C21
P2-C2
P2-C31
P2-C41
P3-C3
P3-C51
P3-C61
P4-C4
Figure 1. Molecular structure of the cation 1OsPF6 as determined by
single-crystal X-ray diffraction.
P4-C71
P4-C81
C1-C4
midpoint of HOs1 and HOs2. Under these assumptions the dHH
lengthens from 1.15(3) to 1.22(3) Å. The core structure of 1OsPF6 is
shown in Figure 2. Bond lengths and angles are listed in Table 2.
C2-C3
Results and Discussion
Cl(Br)-Os-P1
Cl(Br)-Os-P2
Cl(Br)-Os-P3
Cl(Br)-Os-P4
P1-Os-P2
92.5(1)
84.6(1)
86.4(1)
95.3(1)
98.9(1)
178.7(1)
79.3(1)
81.7(1)
178.2(1)
100.1(1)
158(2)
160(2)
69(2)
103(2)
87(2)
104(2)
112(2)
78(2)
92(2)
92.7(4)
84.9(4)
85.6(4)
95.2(4)
99.7(4)
178.1(5)
78.9(4)
91.7(1)
85.5(1)
85.6(1)
95.7(1)
98.5(1)
177.3(1)
80.8(1)
81.7(1)
178.6(1)
99.2(1)
151(2)
166(2)
70(2)
102(2)
76(2)
96(2)
113(2)
81(2)
103(2)
Starting Materials. The starting materials for the preparation
of the osmium dihydrogen complexes are the complexes trans-
OsHXL2 (L ) dppe, X ) Cl (10Os), Br (11Os), L ) depe, X
) Cl (12Os)) or the new 5-coordinate complexes [OsX(dppe)2]+
(X ) Cl (8Os), Br (9Os)). Complexes 10Os and 12Os were
first reported by Chatt and Hayter16 without full spectroscopic
characterization, but the method of synthesis of 10Os was
difficult to reproduce. Complex 10Os is obtained in a pure
form by reacting trans-[Os(H2)H(dppe)2]BF4 (5Os) with LiCl.
The synthesis of the new complexes cis-OsBr2(dppe)2, and 11Os
directly from (NH4)2[OsBr6] represents a significant improve-
ment in the route to dppe complexes of Os(II) in terms of
number of steps and overall yield.
P1-Os-P3
P1-Os-P4
P2-Os-P3
81.1(4)
P2-Os-P4
178.6(5)
100.3(4)
157.4(9)
159.9(10)
71.8(9)
102.7(9)
81.8(8)
104.8(9)
110.1(9)
78.8(9)
P3-Os-P4
Cl(Br)-Os-HOs1
Cl(Br)-Os-HOs2
P1-Os-HOs1
P1-Os-HOs2
P2-Os-HOs1
P2-Os-HOs2
P3-Os-HOs1
P3-Os-HOs2
P4-Os-HOs1
P4-Os-HOs2
HOs1-Os-HOs2
Os-P1-C1
The complexes 8Os and 9Os are prepared from the dihalide
complexes cis-OsCl2(dppe)213 or cis-OsBr2(dppe)2, respectively,
as in eq 1 (X ) Cl or Br). Complex 8Os forms air-stable, red
97.7(8)
75.6(9)
76(2)
41(2)
83(2)
42(3)
42.6(12)
108.4(6)
119.0(6)
117.1(6)
106.1(6)
101.4(7)
103.0(7)
107.7(6)
119.6(6)
119.8(6)
104.4(7)
100.7(6)
102.1(7)
109.8(7)
119.1(7)
119.5(7)
104.5(8)
95.4(8)
105.0(8)
104.2(6)
119.3(6)
121.4(6)
104.5(6)
101.5(6)
103.3(7)
113.7(7)
109.0(8)
108(1)
107.9(1)
118.9(2)
116.8(1)
106.3(2)
100.7(2)
104.4(2)
107.1(1)
119.9(1)
119.9(2)
103.8(2)
100.3(2)
103.0(2)
108.2(2)
116.7(2)
119.8(2)
105.2(3)
99.8(2)
105.0(2)
103.4(1)
120.5(1)
120.2(2)
105.0(2)
102.2(2)
103.1(2)
112.6(3)
106.7(4)
110.6(3)
112.0(3)
108.2(2)
121.9(2)
116.2(2)
102.9(2)
102.5(2)
102.7(2)
107.2(2)
121.8(2)
119.1(2)
102.6(2)
100.5(2)
102.4(2)
108.3(2)
116.7(2)
123.2(2)
104.1(2)
99.1(2)
102.7(2)
103.8(1)
119.1(1)
120.4(1)
105.2(2)
102.4(2)
103.8(2)
112.9(3)
107.2(3)
111.5(3)
110.7(3)
cis-OsX2(dppe)2 + NaPF6 f [OsX(dppe)2]PF6 + NaX (1)
Os-P1-C11
Os-P1-C21
C1-P1-C11
C1-P1-C21
C11-P1-C21
Os-P2-C2
Os-P2-C31
Os-P2-C41
C2-P2-C31
C2-P2-C41
C31-P2-C41
Os-P3-C3
Os-P3-C51
Os-P3-C61
C3-P3-C51
C3-P3-C61
C51-P3-C61
Os-P4-C4
crystals. The cation of 8Os has a distorted trigonal bipyramidal
structure in the solid state15 which is almost identical to that of
the ruthenium(II) analogue (8Ru) reported recently.2
Synthesis of the Complexes trans-[Os(H‚‚H)X(L)2]+. These
dihydrogen complexes are prepared by reacting complexes 8Os
or 9Os with H2 (1 atm) for 10 min at 20 °C in CH2Cl2 as in eq
2 (X ) Cl (1OsPF6), X ) Br (2OsPF6)). They are also made
by reacting complexes 10Os, 11Os, or 12Os with HBF4 in ether
under Ar or N2 as in eq 3 (X ) Cl, L ) dppe (1OsBF4); X )
Br, L ) dppe (2OsBF4); X ) Cl, L ) depe (3OsBF4)).
[OsX(dppe)2]PF6+ H2 f trans-[Os(H‚‚H)X(dppe)2]PF6
(2)
Os-P4-C81
Os-P4-C71
C4-P4-C71
C4-P4-C81
C71-P4-C81
P1-C1-C4
trans-OsHXL2 + HBF4‚Et2O f trans-[Os(H‚‚H)X(L)2]BF4
(3)
Reaction 2 produces 1OsPF6 and 2OsPF6 in excellent yield
as light-green or pale-brown crystals, respectively. Reaction 3
produces cream-colored 1OsBF4, pale-brown 2OsBF4, or white
3OsBF4. These forms of 1Os or 2Os are soluble in CH2Cl2
and stable for at least 1 week under 1 atm of N2 at 20 °C.
P2-C2-C3
P3-C3-C2
P4-C4-C1
110.3(7)

1
Dihydrogen with Frequency Near H Larmor Frequency
J. Am. Chem. Soc., Vol. 118, No. 23, 1996 5401
Figure 2. Structure of the core of 1OsPF6 as determined by neutron
diffraction.
Complex 3OsBF4 dissolves in acetone, THF, and CH2Cl2 to
give colorless solutions; however, these begin to decompose
after 1 d. The dihydrogen complexes 1Os and 2Os also form
in the gas phase during the FAB MS analysis of the complexes
10Os and 11Os; this must involve protonation reactions like
eq 3 except that the nitrobenzyl alcohol solvent used in this
analysis is a much weaker acid than HBF4.
The IR spectra (Nujol mull) of 1OsPF6 and 1OsBF4 are
identical and do not show any peaks in the region 3000-1800
cm-1 that could be assigned to Os-H or H-H stretches. The
IR spectrum of 2OsBF4 also does not display such absorptions.
Solid-State Structure of trans-[Os(H‚‚H)Cl(dppe)2]PF6
(1OsPF6). The X-ray structure analysis shows that there is an
elongated dihydrogen ligand with dHH of 1.11(6) Å located trans
to the chloride ligand on an octahedral osmium center. The
structure from neutron diffraction at 20 K refined to give an
H-H distance of 1.15(3) Å; when a correction is applied to
account for the librational motion of this ligand, the H-H
distance becomes 1.22(3) Å.23 Neutron diffraction studies of
dihydrogen complexes reveal that a wide range of distances is
possible. Complexes with H-H distances close to the present
value are [Ru(C5Me5)(H2)(dppm)]BF4 (1.10(2) Å),7 cis-
IrCl2(H2)H(PiPr3)2 (1.11(3) Å),24 [Os(H‚‚H)(OAc)(en)2]PF6
(1.34(2) Å),4 and Re(H‚‚H)H5(Ptol3)2 (1.357(7) Å).25 Therefore
complex 1Os fills in the gap in this apparent continuum of
distances.
The orientation of the H2 group is staggered with respect to
the P-Os-P bonds. The H atoms are between pairs of P atoms
which have angles P-Os-P of about 100° and not between
those with P-Os-P of about 80°, the bite angle of the dppe
ligands. The X-ray and neutron studies provided identical
orientations within experimental error. This is similar to the
orientation of the H‚‚H group in [Os(H‚‚H)(OAc)(en)2]+ where
the hydrogens are between pairs of nitrogens not linked by the
ethylene backbones.4 The staggered conformation was calcu-
lated to be 2.5 kcal mol-1 lower in energy than the eclipsed
conformation for this latter complex.26 When the H2 unit
eclipses a trans-P-M-P grouping, this P-M-P angle becomes
less than 180° and bent away from the H2 ligand as in the case
of trans-[Fe(H2)H(dppe)2]BF4, where the P-Fe-P angle is
163.0(1)°,27 and in the case of trans-Mo(H2)(CO)(dppe)2, where
the P-Mo-P angle is 174.2(1)°.28
Figure 3. Final difference electron density map in the BrOsH2 plane
of 2OsBF4. Contours are shown every 0.1 e Å3. The contours on Os
climb in the center to 1.5 e Å3 but are truncated at 0.7 e Å3.
In the X-ray structure the dihydrogen ligand appears to be
unsymmetrically bonded to the Os with Os-H distances of 1.41-
(5) and 1.67(5) Å. However this is an artifact of the experiment
because the neutron study reveals essentially equivalent Os-H
distances of 1.57(2) and 1.59(2) Å. The H2 ligand in trans-
Re(H2)Cl(PMePh2)4 appeared to be unsymmetrically bonded in
the refined model from X-ray diffraction;29 this might also be
an artifact. The complex 5OsBF4 has been shown in a neutron
diffraction study to have a symmetrical triangular OsH2 unit
with Os-H distances of 1.71(4) and 1.77(3) Å and an H-H
distance of g0.79 Å.30,31 The Os-H distances are shorter for
1OsPF6 relative to 5OsBF4 because of the lower trans influence
of chloride versus hydride.
The four P atoms of 1OsPF6 are coplanar with each other
and the Os atom. The Os-P distances (2.38(1) Å on average)
are systematically longer than those in 5OsBF4 (2.343(2) Å on
average)30,32 because of increased steric repulsions due to the
chloride ligand in 1OsPF6 compared to the hydride ligand in
5OsBF4.
-
The PF6 anion inserts fluorine F(5) into the H2 binding
pocket defined by four phenyl rings. The two H‚‚F distances
are 2.68(4) Å (F5-HOs1) and 2.72(4) Å (F5-HOs2). There
is an electrostatic attraction between the fluoride (δ-) and acidic
dihydrogen (δ+) groups. However the distances are longer than
the sum of van der Waals radii for hydrogen and fluorine (2.55
Å). A related interaction between a chloride ligand and the
hydrogen of a dihydrogen ligand in Ir(H2)(H)Cl2(PiPr3)2 has
already been reported.24
Solid-State Structure of trans-[Os(H‚‚H)Br(dppe)2]BF4,
2OsBF4. The pale-brown crystals of this complex were
isomorphous with those of 1Ru.2 In contrast to those of 1Ru,
data quality was sufficient to allow the refinement of the
positions of the hydrogens bonded to osmium in the X-ray
structure determination at 150 K. The cation is octahedral with
the dihydrogen ligand positioned trans to the bromide. Figure
3 shows a difference electron density map in the BrOsH2 plane.
(23) The libration of the H2 ligand creates banana-shaped smears of
nuclear density at the ends of the H-H vector. When this shape is modeled
by a thermal ellipsoid, an artificially short d(HH) is obtained. In order to
verify the correction, calculations were also performed using Maverick and
Trueblood’s THMA14 program which yielded a value of 1.24 Å. See ref
24.
(24) Albinati, A.; Bakhmutov, V. I.; Caulton, K. G.; Clot, E.; Eckert, J.;
Eisenstein, O.; Gusev, D. G.; Grushin, V. V.; Hauger, B. E.; Klooster, W.
T.; Koetzle, T. F.; McMullan, R. K.; O’Loughlin, T. J.; Pelissier, M.; Ricci,
J. S.; Sigalas, M. P.; Vymenits, A. B. J. Am. Chem. Soc. 1993, 115, 7300-
7312.
(25) Brammer, L.; Howard, J. A. K.; Johnson, O.; Koetzle, T. F.; Spencer,
J. L.; Stringer, A. M. J. Chem. Soc., Chem. Commun. 1991, 241-243.
(26) Craw, J. S.; Bacskay, G. B.; Hush, N. S. J. Am. Chem. Soc. 1994,
116, 5937-5948.
(27) Morris, R. H.; Sawyer, J. F.; Shiralian, M.; Zubkowski, J. J. Am.
Chem. Soc. 1985, 107, 5581-2.
(28) Kubas, G. J.; Burns, C. J.; Eckert, J.; Johnson, S. W.; Larson, A.
C.; Vergamini, P. J.; Unkefer, C. J.; Khalsa, G. R. K.; Jackson, S. A.;
Eisenstein, O. J. Am. Chem. Soc. 1993, 115, 569-581.
(29) Cotton, F. A.; Luck, R. L. Inorg. Chem. 1991, 30, 767-774.
(30) Maltby, P. A.; Morris, R. H.; Klooster, W. T.; Koetzle, T. F.; Ricci,
J. S.; Albinati, A. Manuscript in preparation.
(31) The H-H distance still needs to be corrected for the effects of the
large amplitude of libration of the H2 ligand.
(32) Farrar, D. H.; Maltby, P. A.; Morris, R. H. Acta Crystallogr., Sect.
C 1992, C48, 28-31.

5402 J. Am. Chem. Soc., Vol. 118, No. 23, 1996
Maltby et al.
Table 3. Observed T1 Values of the Dihydrogen Ligands in the Complexes trans-[Os(H‚‚H)X(L)2]+ and the Methylene Protons of 1Osa
1Os, 300 MHz
T1(obsd), ms
1Os, 400 MHz
T1(obsd), ms
1Os, 500 MHz
T1(obsd), ms
1Os, 400 MHzb
T, K
T, K
T, K
T, K
T1(obsd) [T1(calcd)] ms
296
275
251
63
53
41
294
273
250
229
79
66
57
54
300
273
258
242
231
226
210
189
88
73
68
67
70
70
80
99
294
272
251
0.37 [0.37]
0..31 [0.29]
0.25 [0.24]
226
201
39
46
207
185
176
60
74
81
229
207
186
0.22 [0.23]
0.25 [0.26]
0.37 [0.41]
176
70
2Os, 400 MHz
3Os, 400 MHz
T, K
T1(obsd), ms
T, K
T1(obsd), ms
T, K
T1(obsd), ms
T, K
T1(obsd), ms
293
272
251
76
64
57
229
208
186
57
66
87
295
268
248
217
159
113
227
207
187
79
60
60
a Observed values reproducible to ( 2 ms; errors at 293 K are ( 2 ms, and at 180 K, ( 5 ms. The solvent is CD2Cl2 for all complexes except
for 3Os where acetone-d6 was used. b T1(obsd) values are for the methylene protons of the dppe backbone of 1Os. The T1(calcd) values were
obtained by use of the parameters τ0 and Ea for 1Os as listed in Table 4.
The refined model gave an H-H distance of 1.13(8) Å and
Os-H distances of 1.64(6) and 1.52(6) Å as indicated by the
triangle in Figure 3. Within the 3σ level of the estimated
standard deviations of atomic positions, a more symmetrical
arrangement, in which the hydrogen atoms are positioned in
the centers of the contours, is also conceivable. In this case
the H-H distance lengthens to about 1.4(1) Å. Both H-H
distances fall in the range calculated from the 1H NMR T1(min)
measurements on the dihydrogen ligand (see below). The H2
orientation is similar to the one in 1OsPF6. The BF4 anion is
located with F(2) inserted into the H2 binding pocket with an
F(2) to H(2) distance of 2.6(1) Å.
The 2.320(1) Å Os-Cl bond length in 1OsPF6 and the 2.560-
(1) Å Os-Br bond length in 2OsBF4 are similar to those of
other Os(II) halide complexes.33 This suggests that the H2
ligands in these complexes do not have a high trans influence.
Both the Os-Cl and the Os-Br bonds are tilted toward P(2)
and P(3), as in the structure of 1Ru, because of steric repulsions
with phenyl rings on P(1) and P(4).
NMR Properties of 1Os, 2Os, and 3Os in Solution. The
31P{1H} spectra of these dihydrogen complexes in CD2Cl2 or
acetone-d6 solution at 295 K are singlets; the spectrum of 1Os
remains a singlet down to 176 K at 121.5 MHz. Therefore these
complexes are regular octahedral or there is a low-energy
exchange process which makes the P atoms equivalent.
Figure 4. Plot of T1 of the dihydrogen nuclei on a log scale versus
inverse temperature for 1Os observed at 300 (2), 400 (9), and 500
([) MHz. The smooth curves are from calculated values based on dHH
-
The high-field region of the 1H NMR spectra of these
complexes in CD2Cl2 at 295 K displays a quintet at -11.44
ppm with |J(H,P)| ) 11.7 Hz for 1Os, at -11.46 ppm with
|J(H,P)| ) 11.9 Hz for 2Os, and at -14.0 ppm with |J(H,P)| )
11.8 Hz for 3. These quintets become less well resolved as the
temperature of the samples is lowered until they become broad
singlets at 400 MHz below 200 K. The chemical shift of 1Os
in CD2Cl2 is temperature dependent and is given by δ(H‚‚H)
) -11.72 + 9.42 × 10-4(T), where T is in K. Simulation of
the line shapes shows that the J(H,P) coupling is not affected
by temperature and that the broadening at low temperature is
due to the short T2 expected for the H‚‚H ligand. Therefore
the H atoms on osmium are chemically equivalent from room
temperature to low temperature and they remain coupled to four
equivalent P atoms.
(fast spin) ) 1.04 Å or dHH(slow spin) ) 1.31 Å and the correlation
time of the molecule given by the τ0 and Ea parameters of the CH2
protons of the dppe backbone (see Table 4).
tion of complex were used at each spectrometer frequency; this
is crucial in order to fit the data as described below. The data
for complex 3Os fit the usual “V”-shaped curve when ln T1 is
plotted against 1/T, whereas those for 1Os and 2Os deviate from
the V shape (Figure 4). The temperature-dependent T1 values
1
can be calculated by use of the equation for H-1H dipolar
relaxation with a suitable spectral density function and a
temperature-dependent correlation time, τc ) τoeEa/RT, as
reported previously.34 The parameters which determine the
correlation time of the molecule, τo and Ea, should be the same
as those of the corresponding ruthenium complexes as long as
the concentrations of the solutions are the same (Table 4).
Indeed the data for 3Os can be fit with approximately the same
parameters as those of 3Ru2 along with either a dHH of 1.06 Å
The T1 values of the H atoms on Os were measured as a
function of temperature, T (175-300 K), by the inversion
recovery method (Table 3). Samples with the same concentra-
(33) Orpen, A. G.; Brammer, L.; Allen, F. H.; Kennard, O.; Watson, D.
G.; Taylor, R. J. Chem. Soc., Dalton Trans. 1989, S1.
(34) Bautista, M. T.; Earl, K. A.; Maltby, P. A.; Morris, R. H.;
Schweitzer, C. T.; Sella, A. J. Am. Chem. Soc. 1988, 110, 7031-7036.

1
Dihydrogen with Frequency Near H Larmor Frequency
J. Am. Chem. Soc., Vol. 118, No. 23, 1996 5403
J(ω) ) S2τc/(1 + ω2τc2) + (1 - S2)(τ′)/(1 + ω2(τ′)2)
Table 4. Parameters Used To Fit the Temperature Dependence of
the T1 Values of Complexes 1Os, 2Os, and 3Os (Table 3) and 1Ru,
3Ru, 5Os, and 6Os
(5)
function has a factor S2 which ranges from 1 (no internal motion)
dHH, Å
minc
T(min), T1(min)a
Ea,
spinning
rateb
to 0.25 (fast spinning) while 1/τ′ ) 1/τc + 1/τH . By use of
2
complex
K
ms
τo, ps kcal/mol
corrd
dHH ) 1.22 Å from the neutron study, we determined that S2 is
approximately 0.6, which suggests the existence of an inter-
mediate frequency of motion of the H2. However, eq 5 with S2
) 0.6 did not fit well to all the temperature-dependent T1 data.
A more complex physical model needs to be developed to
analyze this and future data for this interesting case of
intermediate motion of the H2 ligand.36b
1Os
230
230
240
238
238
210
210
210
233
203
53 ( 1 1.1e
53 ( 1 1.1e
25 ( 1 1.8
57 ( 1 1.7
57 ( 1 1.7
60 ( 1 0.34
60 ( 1 0.34
28 ( 1 0.37
40 ( 1 0.90
80 ( 1 0.19
2.5e
2.5e
2.4
2.3
2.3
2.6
2.6
2.7
2.6
2.9
fastf
slowf
fast
1.04
1.31
0.92
1.05
1.33
1.06
1.34
0.94
0.98
1.08
1.35
0.94
1.08
1.36
1.10
1.38
0.96
1.02
1Ru
2Os
fastf
slowf
fast
3Os
slow
fast
3Ru
5Os
6Os
Therefore, the flattened shape of the ln T1 versus 1/T plots is
an indicator that the motion or motions of the H2 ligand have
frequencies similar to that of the overall motion of the molecular
complex. In such a situation the actual dHH will be between
the values calculated from the T1(min) value for “fast spinning”
and “slow spinning”. These limiting values are listed in Table
4 for 1Os and 2Os.
Preparation and Properties of the Deuterated Complexes.
The complex trans-[Os(H‚‚D)Cl(dppe)2]+ (1Os-d1) was pre-
pared by reacting 10Os with HBF4‚Et2O in excess D2O/ether
or by reacting 8Os with HD gas (generated by reacting D2O
with NaH). Some 1Os contaminated the first sample. The 1H
NMR spectrum observed at 298 K is a 1:1:1 triplet of quintets
at 0.013 ppm upfield of the resonances for 1Os with J(H,P) )
11.7 Hz and J(H,D) ) 13.9 ( 0.2 Hz for 1Os-d1 in CD2Cl2
and J(H,P) ) 11.7 Hz and J(H,D) ) 14.4 ( 0.2 Hz for 1Os-d1
in acetone-d6. The J(H,D) values are also temperature depend-
ent, decreasing with decreasing temperature (Table 5). The
isotopomers 1Os and 1Os-d1 appear as two closely spaced
singlets at 20.8 and 20.7 ppm, respectively, in the 31P{1H} NMR
spectrum. Complex 2Os-d1 gives a very similar 1H NMR
spectrum with J(H,P) 11.9 Hz and J(H,D) 13.7 Hz.
fast
fastg
(1.12)g (1.17)g
a The observed minimum T1 value, T1(min), for the dihydrogen ligand
(at temperature, T(min), and 400 MHz) obtained by direct observation
and fitting to the temperature-dependent T1 equation using a temperature
dependent correlation time τc ) τoeE /RT 34
.
b A fast rotational frequency
a
of the H2 unit is .400 MHz, while a slow frequency is ,400 MHz.
c Obtained from fitting the T1 data. d Corrected for the relaxation
contributions from neighboring protons.35 e Parameters obtained by
fitting the T1 data of CH2 groups of the dppe backbone (Table 3 and
unpublished data for 300 and 500 MHz). f Actually an intermediate
motion of the H2 of frequency ≈ 400 MHz is suspected (see text).
g Actually fast-spinning (H2) and static (H)2 forms are proposed to be
in rapid equilibrium.
for a fast spinning ligand or 1.34 Å for a slow spinning ligand;
the calculated T1 values are listed in Table 3, and the calculated
dHH values are listed in Table 4. These are the minimum values
of dHH; to obtain the actual values the relaxation rates have to
be corrected for the contributions of the ethyl group protons by
the method of Desrosiers et al.35 The resulting distances, which
are slightly longer, are also listed in Table 4.
The unusual temperature dependence of the T1 data at three
different spectrometer frequencies for the dihydrogen ligand of
complex 1Os is demonstrated in Figure 4. The smooth curves
are calculated using dHH(slow spinning) ) 1.31Å or dHH(fast
spinning) ) 1.04 Å and the standard spectral density functions.
For these calculations the parameters τ0 ) 1.08 ps and Ea )2.5
kcal mol-1 which give the correlation times for the molecule at
each temperature (Table 4) were determined from the T1 values
of the methylene hydrogens of the dppe backbone (Table 3).
These were similar to the correlation times calculated by use
of the parameters for 1Ru (Table 4). For a given spectrometer
frequency the observed T1 values of the dihydrogen ligand of
1Os are shorter than expected at both high temperatures and
low temperatures (Figure 4). The deviations are caused by an
H2 correlation time or rotational frequency which is near the
correlation time or tumbling frequency of the complex. We
attempted to model the dipolar relaxation by assuming that the
correlation time of the H2 ligand can be expressed in the
The complex 3Os is enriched in deuterium by slow exchange
with acetone-d6. 3Os-d1 gives a pseudo-septet at -13.83
ppm, 0.016 ppm upfield of and superimposed on the (H‚‚H)
1
signal for 3Os. The H resonances of 3Os-d1 were observed
by use of an inversion recovery experiment with a delay time
of 150 ms at 400 MHz to null out the resonances of 3Os (see
refs 2, 37, and 38 for similar experiments). The null time
corresponds to a T1 for 3Os of 217 ms ()150/ln(2)), which is
in agreement with other determinations of T1 (see Table 3).
The observed intensities of the lines of the (H‚‚D) septet
are 0.7:4.8:11:14:11:4.8:0.7, close to the expected 1:5:11:14:
11:5:1 pattern if J(H,D) ≈ J(H,P). The outer peak separation
of 73.2 Hz corresponds to a J(H,D) coupling of 13.1 Hz,
assuming that J(H,P) for 3Os-d1 is identical to the 3Os value
of 11.8 Hz.
Correspondence of J(H,D) and dHH. We find that a plot
of J(H,D) versus dHH for complexes whose H-H distances have
been determined by single-crystal neutron diffraction, X-ray
diffraction, or solid-state NMR techniques (Table 6) gives a
straight line:
exponential form τH ) τroteE /RT. We used the spectral density
rot
2
function for the case of dipolar relaxation of H2 perpendicular
to the axis of rotation (eq 4)34 and varied dHH, τrot and Erot.
J(ω) ) 0.25τc/(1 + ω2τc2) +
dHH ) -0.0167J(H,D) + 1.42
(6)
0.75(τc/(2 + τH /τc))/(1 + ω2(τc/(2 + τH /τc))2) (4)
2
2
One complex whose parameters deviate from this plot is Ir-
(H2)HCl2(PiPr3)2, which may not have the same hydrogen-
chlorine interaction in solution that it does in the crystal as has
been discussed previously.7 Equation 6 probably does not hold
for J(H,D) < 3 Hz because cis-dihydride complexes with dHH
Values of dHH )1.06 Å, τrot ) 0.3 ps, and Erot ) 1.5 kcal mol-1
gave a reasonable fit to the 500 MHz data but did not fit all the
400 and 300 MHz data. We also tried the “model-free” spectral
density function suggested by Lipari and Szabo (eq 5).36a The
(37) Chinn, M. S.; Heinekey, D. M.; Payne, N. G.; Sofield, C. D.
Organometallics 1989, 8, 1824-1826.
(38) 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,
4876-4887.
(35) Desrosiers, P. J.; Cai, L.; Lin, Z.; Richards, R.; Halpern, J. J. Am.
Chem. Soc. 1991, 113, 4173-4184.
(36) (a) Lipari, G.; Szabo, A. J. Am. Chem. Soc. 1982, 104, 4559-4570.
(b) This model should include a treatment of the H2 libration.

5404 J. Am. Chem. Soc., Vol. 118, No. 23, 1996
Maltby et al.
Table 5. J(H,D) Values for [Os(H‚‚D)Cl(dppe)2]+ as a Function of
some complexes trans-[M(CO)X(L)2]+ of Table 7 which
decrease with this order of ligands. It is also of interest to note
that ν(CO) values are lower when CO is trans to chloride than
when trans to hydride. In each series of complexes with the
same metal and trans ligand there is an apparent increase in the
H-H distance (except possibly for the Ru complexes) and a
decrease in the J(H,D) coupling of the trans-[M(HD)X(L)2]+
isotopomers with this order of ligands. Such a trend would be
expected since an increase in π-basicity of the metal along this
series would result in increased M(dπ) f H2(σ*) back-bonding.
Associated with a lengthening of the H-H bond is an increase
in the coupling, J(H,P), which in turn probably indicates an
increase in hydride character of the H‚‚H unit.
The most interesting effect of halide versus hydride is the
comparison of H-H distances in the series 1Os-4Os versus
5Os-7Os. In the former series the character of the H‚‚H unit
changes only slightly while in the latter there is a large structural
change on going from a spinning dihydrogen ligand in 5Os to
a dihydride unit in 7Os. It appears that the halide ligand buffers
the effect of the phosphine ligands so that the H-H bond only
lengthens a little when the electron-donor strength of the
phosphine ligand increases from dppe to dcpe. As noted above,
the Os-P bonds in 1Os are longer than those in 5Os and this
might reduce the transmission of electronic influences from the
substituents on phosphorus to the metal. A result of this
buffering effect is that, although there is an increase in H-H
distance on going from trans-[Os(H2)H(dppe)2]+ to trans-[Os-
(H‚‚H)Cl(dppe)2]+, there is a decrease for the corresponding
pair of dcpe complexes. This contrasts with the results for
ruthenium where a change of the trans ligand from hydride to
chloride always results in an increase in H-H distance and in
the hydride character of the dihydrogen ligand.2 It is possible
that the splitting of the H-H bond in complex 4Os and in trans-
[Os(H‚‚H)Cl(en)2]+ (see Table 6) is arrested because of first
coordination sphere repulsions with the chloride ligand. This
might also be the explanation for why replacement of the
chloride in the dihydrogen complex, Ir(H2)ClH2(PPri2)2,39 with
hydride to give the pentahydride, IrH5(PPri3)2,40 results in the
breaking of the H-H bond. However a more likely explanation
in this case is that octahedral Ir(H2)H3(PPri3)2 is expected to be
high in energy because it is forced to have trans hydride
ligands.41
1
Solvent and Temperature from 500 MHz H NMR Spectra
1Os-d1 in CD2Cl2
T, K J(H,D), Hz
1Os-d1 in acetone-d6
T, K J(H,D), Hz
253
263
273
283
298
308
13.6
13.8
13.9
13.7
13.9
14.2
253
273
298
308
13.8
14.2
14.4
14.5
Table 6. H-H Distances of Dihydrogen Complexes in the Solid
1
State and J(H,D) from Solution H NMR Studies
J(H,D),
Hz
complex
dHH, Å methoda
ref
Cr(H2)(CO)3(PiPr3)2
[Fe(H2)H(dmpe)2]BPh4
[Fe(H2)H(dppe)2]BPh4
Mo(H2)(CO)(dppe)2
W(H2)(CO)3(PiPr3)2
[Ir(H2)H(bq)(PPh3)]+
[Ru(H2)Cp(CO)(PCy3)]+
[Ru(H2)Cp(dmpe)]+
[Ru(H2)Cp(dppm)]+
0.85
0.86
0.88
0.88
0.89
0.94
0.97
1.02
1.01
s
X
s
s
s
s
s
s
X
n
n
X
n
n
X
35
31
32
34
33.5
29.5
28
45
46, 47
48, 49
28
50, 51
49, 52
49, 53
49, 53
54
7
24
this work
this work
4
22
21.9
20.9
12
13.7
13.9
9.1
7.2
43
[Ru(H‚‚H)Cp*(dppm)]BF4 1.10(2)
cis-Ir(H‚‚H)HCl2(PiPr3)2
[Os(H‚‚H)Br(dppe)2]PF6
[Os(H‚‚H)Cl(dppe)2]PF6
1.11(3)
1.13(8)
1.22(3)
[Os(H‚‚H)(en)2(OAc)]PF6 1.34(2)
[Os(H‚‚H)Cl(en)2]PF6
1.39(10)
0.74
4
HD(g)
a n ) neutron diffraction, X ) X-ray diffraction, s ) solid-state
NMR.
of 2 Å have J(H,H) values of at least 10 Hz which correspond
to J(H,D) of up to 2 Hz. Therefore the plot should curve to
higher dHH at small J(H,D) values.
Equation 6 appears to provide the best estimate of dHH in
solution for a dihydrogen complex when J(H,D) is known. As
mentioned above, the T1 method gives ambiguous values of dHH
which depend on the interpretation of the motion of the H2
ligand. However an analysis of the dHH values at the fast and
slow spin limits of 64 complexes which also have measured
J(H,D) values shows that eq 6 could apply to all of these (see
supporting information). In several cases it can be decided
whether the complexes have fast-spinning or slow-spinning H2
ligands. In general the complexes with fast-spinning H2 ligands
have hydride as the trans ligand. There is no obvious correlation
apparent as yet on the nature of the trans ligand for the
complexes which have slow-spinning H2 ligands. The dHH
values calculated by use of eq 6 for complexes 1Os-6Os and
their ruthenium analogs are listed in Table 7. It appears that
complexes 1Os and 2Os have H-H distances in solution which
are between the fast- and slow-spinning limits. The rectangular
bar in Figure 5 shows these extreme dHH values for 1Os as well
as the limits of J(H,D) observed at high and low temperature.
Therefore the dHH for 1Os in solution is likely to be about 1.20
Å as indicated by the intersection of this bar with the line from
the J(H,D) values and the point from the neutron-derived solid-
state distance. Similarly the dHH for 2Os must also be around
1.2 Å.
It is also possible that 1Os-4Os exist as mixtures of
dihydrogen complexes and dihydride complexes which are in
rapid equilibrium as has been proposed for 6Os.5 Then the
observed NMR properties of the complexes (δ(H2), δ(P), J(H,P),
J(H,D), T1(H2)) would be the weighted average of those of the
1
two forms. The low-temperature (180 K) H and 31P spectra
do not give evidence for two species. The increase in J(H,D)
of 1Os-d1 with increasing temperature, as is also observed for
4Os-d1,5 provides evidence for such a rapid equilibrium where
the equilibrium shifts more from the Os(H)(D) species toward
the Os(HD) one with increasing temperature. It is difficult to
explain the increase in J(H,D) with a single species since an
increase in thermal motion of an elongated H‚‚D ligand might
be expected to result in a decrease in J(H,D) with increasing
temperature as observed for the complex [Ru(C5Me5)(H‚‚H)-
(dppm)]+.7 Other possible evidence for a rapid equilibrium is
the observation of the temperature dependence of the chemical
shift of the dihydrogen ligand and the observation of disorder
in one dppe ligand in the X-ray structure of 1OsBF4 which might
Comparison of trans-[M(H2)X(L)2]+ Spectral Data and
H-H Distances in Solution. Table 7 outlines the changes that
occur when the trans ligand is changed from halide to hydride
on osmium and ruthenium for three chelating diphosphine
ligands. The bidentate ligands push more electron density onto
the metal as dppe (R ) Ph) < depe (R ) Et) < dcpe (R ) Cy).
This is clearly indicated by the C-O stretching frequencies for
(39) Mediati, M.; Tachibana, G. N.; Jensen, C. M. Inorg. Chem. 1992,
31, 1827-1832.
(40) Garlaschelli, L.; Khan, S. I.; Bau, R.; Longoni, G.; Koetzle, T. F.
J. Am. Chem. Soc. 1985, 107, 7212-7213.
(41) Eisenstein, O. Personal communication.

1
Dihydrogen with Frequency Near H Larmor Frequency
J. Am. Chem. Soc., Vol. 118, No. 23, 1996 5405
Table 7. Spectroscopic Properties and H-H Distances of Dihydrogen Complexes of the Type trans-[M(H‚‚H)XL2]+
T1 (min) at
400 MHz, ms
trans-[M(CO)XL2]+
ν(CO) cm-1
M
H2
X
R on P
dHH from T1, Å
1J(H,D), Hz
dHH from J(H,D), Å
2J(H,P), Hz
1Os
2Os
H‚‚H
H‚‚H
H‚‚H
H‚‚H
H2
H‚‚H
(H)2
H2
H2
H‚‚H
H2
H2
H2
Cl
Br
Cl
Cl
H
H
H
Cl
Cl
Cl
H
H
H
Ph
Ph
Et
Cy
Ph
Et
Cy
Ph
Et
Cy
Ph
Et
53
57
60
50
40
1.08-1.35
1.08-1.36
1.1-1.4
1.1-1.4
1.02
1.2-1.5
>1.6
0.94-1.2
0.96-1.2
0.94-1.2
0.90
0.85
0.84
13.9a
13.7
13.1
10.5
25.5
11.5a
1.19
1.19
1.20
1.24
0.99
1.23
11.7
11.7
11.7
12.8
5
1936b
3Os
4Osc
5Os
1917d
2003e
6Os
80
5.5
1974f
7Osc
1Rui
3Rui
4Ruc
5Rul
6Rul
7Ruc
175g
25
10.4g
7.4
1975,h 1875h
1946j
25.9
25.2
16
32.0
32.3
31.5
0.99
1.00
1.15
0.88
0.88
0.89
28
20
20
16
7.2
8
1940k
1987m
1958f
Cy
15
a J(H,D) solvent and temperature dependent. b Reference 55. c The T1 values from ref 3 had to be scaled from 80 to 400 MHz to provide this
comparison. The calculation of H-H distances for these complexes has larger uncertainty because of the less accurate 80 MHz T1 data and the
problem that the bidentate ligands have several protons that can contribute to the relaxation of the metal-bonded hydrogens. d Reference 56. e Reference
5. f Reference 57. g Values averaged with those of the terminal hydride due to rapid intramolecular H-atom exchange. h Mixed ν(OsH) and ν(CO)
modes. Reference 3. i Reference 2. j Reference 58. k Reference 59. l Reference 38. m Reference 60.
be explained by the cocrystallization of dihydride and dihydro-
gen isomers.17 The anomalously short T1 values at low
temperatures (Figure 4, 400 and 500 MHz) are not explained
by such an equilibrium since lower temperatures should result
in a shift in the postulated equilibrium to the Os(H)2 form which
would have a longer T1.
The actual description of complexes 1Os-3Os is probably
neither an elongated dihydrogen complex with a fixed H-H
distance nor a rapid equilibrium but instead complexes with
two H atoms moving rapidly in flat potential wells with shallow
minima at the crystallographically determined positions where
the H-H separation is about 1.2 Å.
Acidity of 1Os and 2Os. Complex 1Os is not deprotonated
by PEtPh2 in CH2Cl2. This contrasts with 1Ru which is partially
+
deprotonated by PEtPh2 and is in equilibrium with HPEtPh2
(pKa 4.9) and 10Ru.2 Therefore 1Os is less acidic than 1Ru,
Figure 5. Plot of dHH versus J(H,D) for dihydrogen complexes. The
filled squares are the data from Table 6 except for those of the iridium
complex. The rectangular bar represents the limits of dHH determined
for the T1 data for 1Os (see Table 4).
which has a pKa of 6.0. The osmium complex is partially
deprotonated by PMe2Ph and in equilibrium with HPMe2Ph+
(pKa 6.5) and 10Os (eq 7, K ) 0.12, 293 K). Therefore the
trans-[Os(H‚‚H)Cl(dppe)2]+ + PMe2Ph {K
\}
trans-[Os(H‚‚H)Br(dppe)2]+ + PR3 {K
2Os
\}
1Os
trans-OsHCl(dppe)2 + HPMe2Ph+ (7)
+
trans-OsHBr(dppe)2 + HPR3 (8)
10Os
11Os
In summary, pKa values for complexes trans-[Os(H‚‚H)X-
(dppe)2]+ increase with the trans X-ligand as Br- (5.4) < Cl-
(7.4) < H- (13.6). The same trend is observed in the enthalpy
of protonation of the complexes Os(C5H5)(PPh3)2X to give
hydrides [OsH(C5H5)X(PPh3)2]+ where -∆HHM increases with
X- in the order I- (14.1 kcal mol-1) < Br- (16.3) < Cl- (19.7)
< H- (37.3). Note that the trend for the halides is opposite to
that expected on the basis of the electronegativity of the
substituent. Angelici has explained this effect of X- on the
acidity of osmium monohydride or dihydride complexes in terms
pKa of 1Os is determined to be 7.4 ( 0.3. This is much more
acidic than the complex with hydride trans to dihydrogens the
pKa of 5Os is 13.6 ( 0.3.42 This large difference in acidity
can be attributed to the stronger H-H interaction in 5Os and
the stronger electron-donor ability of hydride versus halide (see
below).
Complex 2Os is partly deprotonated by P(p-CH3OC6H4)3 and
is in equilibrium with HP(p-CH3OC6H4)3+ (pKa 4.57) (eq 8, R
) p-CH3OC6H4, K ) 0.155, 293 K). The pKa of 2Os is
therefore 5.4(0.3. The fact that 2Os is more acidic than 1Os
is consistent with the fact that 1Os is not deprotonated at all by
P(p-CH3OC6H4)3.
of the donor abilities of the X- ligands as reflected in their
43
gas-phase proton affinities which correlate exactly with -∆HHM
.
The electrochemical potentials E°(Os3+/Os2+) of 10Os and
11Os are -0.14 and -0.11 V vs Fc+/Fc, respectively.14,38 The
(42) Cappellani, E. P.; Drouin, S. D.; Jia, G.; Maltby, P. A.; Morris, R.
H.; Schweitzer, C. T. J. Am. Chem. Soc. 1994, 116, 3375-3388.
(43) Angelici, R. J. Acc. Chem. Res. 1994, 28, 51-60.

5406 J. Am. Chem. Soc., Vol. 118, No. 23, 1996
Maltby et al.
Table 8. Dependence of Acidity (pKa) and M-H Bond
Dissociation Energy on the H-H Distance of Dihydrogen and
Dihydride Complexes
Conclusions
We propose that the new octahedral complexes trans-[Os-
(H‚‚H)X(L)2]+ (L ) dppe, X ) Cl, Br, L ) depe, X ) Cl),
have elongated dihydrogen ligands in the range of 1.1-1.22 Å
in the solid-state and in solution. The distances are established
in the solid-state by X-ray diffraction for the dppe complexes
and neutron diffraction for 1OsPF6. Therefore these complexes
are classified as having stretched or elongated dihydrogen
ligands, arrested along the pathway to homolytic splitting or
oxidative addition.10 However it is likely that the hydrogen
atoms are moving rapidly in a flat potential surface above the
OsP4 plane and that this elongated separation is at a shallow
minimum on this surface. The temperature and solvent
dependence of J(H,D) noted for 1Os-d1 might be evidence for
the flat potential energy well and large excursions from
equilibrium positions of the H and D.
The distances in solution are more difficult to obtain. An
approximate dHH can be obtained from the coupling J(H,D) by
use of eq 6. The method of determining the T1(min) value of
the 1H nuclei of the dihydrogen ligand provides upper and lower
limits for dHH. However the correct spectral density function
to describe the motion of the H‚‚H ligand is needed to derive
a reliable dHH value. We describe for the first time how the
deviations in T1 values of the H2 ligand from those expected
on the basis of the correlation time of the molecule signal the
case of hindered spinning of the H‚‚H ligand. However a
suitable spectral density function to define the motion of the
H2 ligand and allow a fit to all the T1 data has not yet been
found.
The hindered motion when dihydrogen is trans to Cl and Br
in 1Os-3Os contrasts with the fast spinning when dihydrogen
is trans to hydride in complexes trans-[M(H2)HL2]+ (M ) Fe,
Ru, Os).38 The H-H distances of the latter complexes are short
(in the range 0.85-1.0 Å) and the M-H2 interaction is weak
because of the high trans influence of the hydride which
weakens Os r H2 σ-bonding and the strong ligand field splitting
effect of hydride which lowers the Os(dπ) electron energies and
decreases Os f π back-donation. Craw and Hush26 have
explained the stretching of the dihydrogen ligands in the series
of complexes trans-[Os(H‚‚H)L(NH3)4]n+ as a function of a
spectrochemical parameter for L which expresses this ligand
field splitting concept. This explains why the weaker field
halide ligands cause longer dHH distances and smaller J(H,D)
couplings in such complexes relative to the hydride ligand.
We note that there is a “buffering effect” of the trans halide
on the transmission of electron density from the cis chelating
phosphines to the dihydrogen ligand via Os(dπ) f H2(σ*) back-
donation. For the case of chloride trans to dihydrogen, there is
a moderate increase in dHH on going from the less donating
dppe ligand (1.22 A) in 1Os to the more donating dcpe (1.3 Å)
in 4Os, while in the case of hydride trans to dihydrogen, the
dppe complex 5Os has dHH of 1.0 Å and the dcpe complex 7Os
has no H-H bonding (dHH > 1.6 Å). This effect is attributed
to a weakening of Os-P bonding by the halide ligand which is
reflected in the longer Os-P bond distances of 1Os and 2Os
compared to 5Os.
d
HH, Å
∆HBDE,
complex
from J(H,D) pKa kcal/mol
ref
[Os(H2)H(dppe)2]+, 5Os
[Os(H‚‚H)Cl(dppe)2]+, 1Os
[Os(H‚‚H)Br(dppe)2]+, 2Os
[Ru(H2)H(dppe)2]+, 5Ru
[Ru(H2)Cl(dppe)2]+, 1Ru
[RuH3(dppfc)2]+ a
1.0
1.2
1.2
0.9
1.0
>1.8
1.1
>1.8
13.6 80 ( 1 42
7.4 73 ( 1 this work
5.4 71 ( 1 this work
15.0 82 ( 2 42
6.0 71 ( 3
4.4 58 ( 2
2
2
[Cp*Ru(H‚‚H)(dppm)]+
[Cp*RuH2(dppm)]+
9.2 73 ( 2 42, 61
8.8 72 ( 2 42, 61
a dppfc ) 1,1′-bis(diphenylphosphino)ferrocene.
pKa and E° values can be used to calculate the energy required
to remove an H atom from 1Os and 2Os, ∆HBDE ) 73 ( 1 and
71 ( 1 kcal mol-1, respectively.42 By contrast the comparable
energy for 5Os is 80 ( 1 kcal mol-1 42
Again these H atom
.
abstraction energies parallel those calculated from -∆HHM for
the hydride complexes [OsHX(C5H5)(PPh3)]: 63.2 (Br-), 66.4
(Cl-), 73.6 (H-) kcal mol-1 44
The unusually large BDE value
.
for 5Os might be a result of some H-H bonding as well as
Os-H bonding since this complex has the shortest H-H
distance (1.0 Å versus about 1.2 Å for 1Os and 2Os and >1.6
Å for [OsH2(C5H5)(PPh3)]+).
Some ∆HBDE values that have been determined so far are
listed in Table 8. There is a trend developing such that ,as the
H-H bond stretches past 1.0 Å, the energies drop to values
expected for metal dihydrides. Therefore, even though the
hydrogens are still close together at 1.1-1.2 Å, they have the
acidity of dihydrides or, in other words, their acidity is not
influenced by H-H bonding. In the few instances where it
has been determined, osmium hydride complexes are less acidic
than comparable ruthenium hydride complexes.42 This is also
true in a comparison of 1Os and trans-[Ru(H‚‚H)Cl(dppe)2]+
(Table 8) with pKa values of 7.4 and 6.0, respectively. The
complexes trans-[M(H2)H(L)2]+ are the exception to this rule
where the ruthenium complexes are less acidic than the osmium
complexes because of the stronger H-H bonds in the former.42
(44) Wang, D.; Angelici, R. J. J. Am. Chem. Soc. 1996, 118, 935-942.
(45) Kubas, G. J.; Nelson, J. E.; Bryan, J. C.; Eckert, J.; Wisniewski,
L.; Zilm, K. Inorg. Chem. 1994, 33, 2954-2960.
(46) Baker, M. V.; Field, L. D.; Young, D. J. J. Chem. Soc., Chem.
Commun. 1988, 546-548.
(47) Hills, A.; Hughes, D. L.; Jimeneztenorio, M.; Leigh, G. J.; Rowley,
A. T. J. Chem. Soc., Dalton Trans. 1993, 3041-3049.
(48) Ricci, J. S.; Koetzle, T. F.; Bautista, M. T.; Hofstede, T. M.; Morris,
R. H.; Sawyer, J. F. J. Am. Chem. Soc. 1989, 111, 8823-8827.
(49) Zilm, K. W.; Millar, J. M. AdV. Magn. Opt. Reson. 1990, 15, 163-
200.
(50) Kubas, G. J.; Unkefer, C. J.; Swanson, B. I.; Fukushima, E. J. Am.
Chem. Soc. 1986, 108, 7000-7009.
(51) Zilm, K. W.; Merrill, R. A.; Kummer, M. W.; Kubas, G. J. J. Am.
Chem. Soc. 1986, 108, 7837-7839.
(52) Crabtree, R. H.; Lavin, M.; Bonneviot, L. J. Am. Chem. Soc. 1986,
108, 4032-4037.
(53) Chinn, M. S.; Heinekey, D. M. J. Am. Chem. Soc. 1990, 112, 5166-
5175.
(54) Heinekey, D. M.; Oldham, W. J. Chem. ReV. 1993, 93, 913-926.
(55) Smith, G.; Cole-Hamilton, D. J.; Thorton-Pett, M.; Hursthouse, M.
B. J. Chem. Soc., Dalton Trans. 1985, 387-393.
(56) Mezzetti, A.; Del Zotto, A.; Rigo, P. J. Chem. Soc., Dalton Trans.
1990, 2515-2520.
There is clearly a continuum of dHH of dihydrogen complexes
ranging from 0.85 to 1.4 Å (Table 6). The osmium complexes
of Table 6 make an interesting series of structures with seven
donor atoms on the metal starting with octahedral Os(II)
complexes with short dHH and proceeding to structures which
are intermediate in the oxidative addition to Os(IV), 7-coordinate
complexes. It is likely that other complexes in Table 6 and
others in the literature have unsymmetrical T1 versus 1/T
behavior because of hindered motion of the H‚‚H ligand. It is
(57) Bancroft, G. M.; Mays, M. J.; Prater, B. E.; Stefanini, F. P. J. Chem.
Soc. (A) 1970, 2146-2149.
(58) Smith, G.; Cole-Hamilton, D. J. J. Chem. Soc., Dalton Trans. 1984,
1203-1208.
(59) Mezzetti, A.; Del Zotto, A.; Rigo, P.; Pahor, N. B. J. Chem. Soc.,
Dalton Trans. 1989, 1045-1052.
(60) Smith, G.; Sutcliffe, L. H.; Cole-Hamilton, D. J. J. Chem. Soc.,
Dalton Trans. 1984, 1209-1214.
(61) Smith, K.-T.; Rømming, C.; Tilset, M. J. Am. Chem. Soc. 1993,
115, 8681-8689.

1
Dihydrogen with Frequency Near H Larmor Frequency
J. Am. Chem. Soc., Vol. 118, No. 23, 1996 5407
important to report all temperature-dependent T1 data as well
as the T1(min) value so that such effects can be detected.
The complexes 1Os (X ) Cl) and 2Os (X ) Br) are much
more acidic than complex 5Os (X ) H). The trend in acidity
as a function of ligand is mirrored by that of the hydride
complexes [OsH(C5H5)X(PPh3)2]+. This suggests that hydrogen-
hydrogen bonding in these complexes with elongated dihydrogen
ligands does not influence the acidity, except perhaps for 5Os.
DE-AC02-76CH00016 with the U.S. DOE, Office of Basic
Energy Sciences, Division of Chemical Sciences.
Supporting Information Available: Table of J(H,D) and
T1 values of 64 dihydrogen complexes, their classification as
fast, intermediate, or slow spinning, and a plot showing this
classification, and full details of the X-ray crystal structures of
1OsPF6 and 2OsBF4 and the neutron diffraction analysis of
1OsPF6 (49 pages). This material is contained in many libraries
on microfiche, immediately follows this article in the microfilm
version of this journal, can be ordered from the ACS, and can
be downloaded from the Internet; see any current masthead page
for ordering information and Internet access instructions.
Acknowledgment. We are grateful to NSERC Canada and
the donors of the Petroleum Research Fund as administered by
the American Chemical Society for research grants and Johnson
Matthey PLC for a loan of osmium salts. We thank Nick Plavac
for some NMR results and Curt Koehler for technical assistance.
The neutron diffraction study was carried out under Contract
JA9529044