Synthesis and characterization of trans-[Os(en)2py(H)]2+ and related studies

Article abstract of DOI:10.1021/ic030091b

trans-[Os(en)₂pyH](Otf)₂, 2, is recovered from an acidic solution of trans-[Os(en)₂py(H₂)](OTf)₂, 1, which has been subjected to one electron oxidation. The structures of both 1 and 2 have been determined by single crystal X-ray analysis. In cyclic voltammetry, 2 shows a one electron oxidation wave at 0.95 V and a one electron reduction wave at -1.2 V, neither accompanied by a signal for the reverse process. Reduction of 2 by Zn/Hg in methanol solution leads to quantitative formation of [Os(en)₂(py)H₂)]²⁺, predominantly in the trans-form. In aqueous solution, species 2 reacts rapidly with N-methylacridium ion, [MAH]⁺, by hydride transfer. One electron chemical oxidation of 2 to the corresponding Os(IV) is slower than that of 1 to 2 owing to the increase in coordination number when Os(IV) is produced. Treatment of 1, or the cis-form, 1′, in DMSO by sodium t-butoxide produces mainly the corresponding isomers of the monohydrides of Os^II, that derived from 1′ is deep red in color while the trans-monohydride is colorless. Both react with [MAH]⁺ to form [MAH]₂, and both disappear rapidly in acetone or acetonitrile, presumably by reducing the solvents. Reaction of trans-[Os(NH₃)₄(H₂)H₂O] (BPh₄)₂, 4, in acetone-d₆ as solvent with either CH₃CHO or styrene leads to hydrogenation of the substrate. Reactions which compete with trans-[Os(en)₂(n²H₂)(CF₃ SO)₃]CF₃SO₃ release of substrate from the trans-complex before isomerization to the cis-form, required for hydrogenation to occur, result in the trans-derivative of the added solute. When H₂C = CH - CH₂ - SCH₃ is the substrate, binding takes place at sulfur. Complete conversion to the cis-substrate isomer is observed, without hydrogenation occurring even though contact between dihydrogen and the double bond is possible.

Full text of DOI:10.1021/ic030091b

Inorg. Chem. 2003, 42, 3815−3821
Synthesis and Characterization of trans-[Os(en)2py(H)]2+ and Related
Studies
J. Scott McQueen, Noriharu Nagao, Todd Eberspacher, Z. W. Li, and Henry Taube*
Department of Chemistry, Stanford UniVersity, Stanford, California 94305
Received March 12, 2003
trans-[Os(en)2pyH](Otf)2, 2, is recovered from an acidic solution of trans-[Os(en)2py(H )](OTf)2, 1, which has been
2
subjected to one electron oxidation. The structures of both 1 and 2 have been determined by single crystal X-ray
analysis. In cyclic voltammetry, 2 shows a one electron oxidation wave at 0.95 V and a one electron reduction
wave at −1.2 V, neither accompanied by a signal for the reverse process. Reduction of 2 by Zn/Hg in methanol
solution leads to quantitative formation of [Os(en)2(py)H )]2+, predominantly in the trans-form. In aqueous solution,
2
species 2 reacts rapidly with N-methylacridium ion, [MAH]+, by hydride transfer. One electron chemical oxidation
of 2 to the corresponding Os(IV) is slower than that of 1 to 2 owing to the increase in coordination number when
Os(IV) is produced. Treatment of 1, or the cis-form, 1′, in DMSO by sodium t-butoxide produces mainly the
II
corresponding isomers of the monohydrides of Os , that derived from 1′ is deep red in color while the trans-
monohydride is colorless. Both react with [MAH]+ to form [MAH]2, and both disappear rapidly in acetone or acetonitrile,
presumably by reducing the solvents. Reaction of trans-[Os(NH ) (H )H O](BPh4)2, 4, in acetone-d6 as solvent with
3 4
2
2
either CH CHO or styrene leads to hydrogenation of the substrate. Reactions which compete with trans-[Os(en)2(η2-
3
H )(CF SO)3]CF SO release of substrate from the trans-complex before isomerization to the cis-form, required for
2
3
3
3
hydrogenation to occur, result in the trans-derivative of the added solute. When H CdCHsCH sSCH is the
2
2
3
substrate, binding takes place at sulfur. Complete conversion to the cis-substrate isomer is observed, without
hydrogenation occurring even though contact between dihydrogen and the double bond is possible.
Introduction
quantitatively by the oxidant having been consumed to form
the equivalent amount of the final product which is the
complex of [Fe(CN)6]4- with the monohydride of OsIV. No
evidence for the formation of an intermediate dihydrogen
or monohydrido complex of OsIII having been observed, and
no other evidence for the stability of this intermediate in
other oxidations having been encountered, the view devel-
oped that monohydrido complexes of OsIII with amines as
co-ligands were fugitive in nature. This was bolstered by
fact that at this time only two cases of one electron oxidation
of dihydrogen complexes were known, that introduced above,
and that reported earlier4 as the net one electron oxidation
of [Re(PMePh2)4(H2)Cl]. The present work rectifies the
misapprehension about the stability of the products resulting
from the one electron oxidation of dihydrogen complexes
of OsII amines.
Numerous examples of dihydrogen complexes are de-
scribed in the literature, the deprotonation of which is favored
by electron withdrawing co-ligands or an increase in the
oxidation state of the metal. An example1 of the latter, closely
related to the present study, is the one electron oxidation of
[Os(NH3)5H2]2+, in which the dihydrogen is very difficult
to deprotonate, to yield [Os(NH3)5(H2)]3+ which in water as
a solvent is moderately acidic. Oxidation in this example
was achieved by cyclic voltammetry, and no effort was made
to isolate the oxidized product. Later work in this laboratory
focused on the two electron oxidation of dihydrogen
complexes of osmiumamines to form monohydrides of
OsIV,2,3 all of which have a very low affinity for protons. In
one study, the complex of [Os(en)2(H2)]2+ with [Fe(CN)6]4-
in aqueous solution was titrated with an oxidant, and the
optical density was measured at stages in the overall two
electron oxidation.3 It was found to be accounted for
Experimental Section
General Procedures. N-Methylacridinium iodide, [MAH]I, was
synthesized according to a published procedure.7 Acetone, aceto-
(1) Harman, D. W.; Taube, H. J. Am. Chem. Soc. 1990, 112, 2261.
(2) Li, Z. W.; Yeh, A.; Taube, H. J. Am. Chem. Soc. 1993, 115, 10384.
(3) Li, Z. W.; Taube, H. Inorg. Chem. 1994, 33, 2874.
(4) Cotton, F. A.; Luck, R. L. Inorg. Chem. 1989, 28, 2181.
Inorganic Chemistry, Vol. 42, No. 12, 2003 3815
10.1021/ic030091b CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/22/2003

McQueen et al.
nitrile, and DMSO were purified by vacuum distillation from CaH2,
diethyl ether by distillation from sodium/benzophenone ketyl, and
MeOH by distillation under nitrogen from Mg turnings. Water was
degassed by sparging with argon prior to use. All reactions were
carried out under an inert atmosphere using either a nitrogen filled
glovebox or standard Schlenk techniques.
in 1 (H1 and H2) and the hydride ligand in 2 (H1) were included in
the model as unrestrained isotropic atoms and refined to conver-
gence.
trans-[Os(en)2py(H2)](OTf)2, 1. The synthesis of this compound
has been reported.5 In the present work, pyridine (60 drops) was
added to a solution of [Os(en)2H2O(H2)](OTf)2 (1.50 g) in acetone
(40 mL) and the mixture kept at room temperature while being
stirred. The yellow powder that formed was collected, washed with
Infrared spectra were recorded on a Mattson Infinity 60AR FTIR
spectrometer. 1H NMR spectra were recorded on a Varian XL-400
spectrometer. Electrochemical measurements were performed using
a BAS-50 potentiostat and a platinum working electrode, a platinum
counter electrode, and a silver wire reference electrode. Unless
otherwise noted, electrochemical experiments were conducted with
a scan rate of 200 mV/s. All measured potentials are referenced to
the ferrocene/ferricenium couple. UV-vis spectra were recorded
on a Hewlett-Packard 8452A or 8453 diode array spectrophotometer
using 1 cm cells. Magnetic susceptibility measurements were made
by the Evans method8 using a 400 MHz instrument and a 5 mm
NMR tube microcell assembly purchased from Wilmad Glass
Company. The general procedure was to dissolve the solute in a
d4-MeOH solution containing CH2Cl2 as the reference. The refer-
ence solution was placed in a spherical capillary bulb attached to
a Teflon sleeve and lowered into a 5 mm NMR tube containing
the d4-MeOH/d2-CH2Cl2 solution. Measurements were carried out
at temperatures between 190 and 310 K.
1
acetone and ether, and dried. Yield ) 1.20 g (70%). HMR (d6-
DMSO) δ -9.25 (s, 2H), 1.73 (s, 2H), 1.91 (s, 2H), 4.05 (s, 2H),
5.68 (s, 2H), 7.65 (t, J ) 6.3 Hz, 2H), 8.05 (t, J ) 7.5 Hz, 1H),
8.84 (d,J ) 5.7 Hz, 2H). No major peaks in the NMR spectrum
were unaccounted for. UV-vis, 3.30 × 10-4 M in H2O, 1.00 ×
10-2 M HOTf (λmax, nm; (ꢀ, M-1 cm-1)): 202 (6.4 × 103); 242
(5.3 × 103); 340 (3.7 × 103).
There was no precedent in the literature for obtaining crystals
of 1 suitable for structure determination by X-ray diffraction. A
difficulty encountered is the isomerization5 of 1 to the cis-form,
1′, which we circumvented by using methanol as a solvent and
reducing the temperature. The methanol solution of 1 (0.04 M) was
kept at -10 °C, and diethyl ether was allowed to diffuse into the
mother liquor from the vapor phase. After 5 days, well formed
crystals appeared. IR (KBr) ) 2020 cm-1
.
trans-[Os(en)2py(H)](OTf)2, 2. Method 1. Ferricenium triflate,
[Fc]OTf (1.2 mL of 0.14 M solution), was added dropwise to a
methanol solution of 1 (90 mg, 0.13 mmol). The blue color of Fc+
immediately dissipates resulting in a yellow solution. Following
addition of 1.2 mL of a 0.14 M methanolic solution of Na(Ot-Bu),
the resulting solution was concentrated to 3 mL, and ether was
added to precipitate a yellow powder. This was removed by
filtration, washed with ether, and dried in vacuo. Yield ) 81 mg
(90%).
Crystal Structure Determinations for trans-[Os(en)2py(H2)]-
(OTf)2 (1) and trans-[Os(en)2py(H)](OTf)2 (2). Single crystals for
the [OTf]- salt of trans-[Os(en)2py(H2)]2+ are, at -104 ( 2 °C,
triclinic, space group P1h (No. 2) with a ) 10.0245(3) Å, b )
10.3111(3) Å, c ) 12.4820(3) Å, R ) 94.670(1)°, â ) 111.545-
(2)°, γ ) 111.446(2)°, V ) 1081(1) Å3, and Z ) 2 {dcalcd ) 2.117
g‚cm-1; µa(Mo KR) ) 6.18 mm-1}. Single crystals for the [OTf]-
salt of trans-[Os(en)2py(H)]2+ are, at -90 ( 1 °C, triclinic, space
group P1h (No. 2) with a ) 10.021(2) Å, b ) 10.330(2) Å, c )
12.470(2) Å, R ) 94.81(3)°, â ) 111.03(3)°, γ ) 111.25(3)°, V )
1088(2) Å3, and Z ) 2 {dcalcd ) 2.101 g‚cm-1; µa(Mo KR) ) 6.14
mm-1}. Totals of 5918(3676) (1) and 5970(3721) (2) independent
(unique) reflections having 2θ (Mo KR) < 52.0° (1) or 51.9° (2)
[the equivalent of 0.60 (1) and 1.00 (2) limiting Cu KR spheres]
were collected on a Siemens SMART-CCD diffractometer and
absorption corrected using the Siemens SADABS program. Both
structures were solved using “heavy-atom” Patterson techniques
with the Siemens SHELXTL-PC software package. The resulting
structural parameters have been refined to convergence {R (based
on F2) ) 0.0644 (1) and 0.0686 (2) for 3632 (1) and 3652 (2)
independent absorption-corrected reflections having 2θ(Mo KR) <
52.0 (1) or 51.9 (2) and I > 0}, {R1 (unweighted, based on F) )
0.0253 (1) and 0.0279 (2) for 3502 (1) and 3412 (2) independent
absorption-corrected reflections having 2θ(Mo KR) < 52.0 (1) or
51.9 (2) and I > 4σ(I)} using counter-weighted full-matrix least-
squares techniques and structural models which incorporated
anisotropic thermal parameters for all full-occupancy non-hydrogen
atoms.
Method 2. A solution of Ag(OTf) in MeOH (2 mL, 0.871 mmol)
was added to a solution of trans-[Os(en)2py(H2)](OTf)2 in MeOH
(30 mL, 0.870 mmol), and the mixture was stirred for 3 h in
darkness. The precipitate, Ag metal, was collected by filtration,
washed with MeOH, and then dried in a vacuum (yield 92.7 mg,
98.6%). Sodium tert-butoxide (84.5 mg, 0.879 mmol) was added
to the filtrate, and the liquid was evaporated to a volume of ∼5
mL. Ether (∼50 mL) was added causing precipitation of a yellow
powder which was collected, washed with ether, and then dried in
vacuo. Yield 425 mg (71%). Anal. Calcd for (C11H22N5S2F6O6-
Os): C, 19.2; H, 3.2; N, 10.2. Found: C, 19.2, H, 3.1, N, 10.0. IR
(KBr) νOsH ) 2020 cm-1. µeff ) 1.8 µB. Crystals of 2 were grown
following the procedure used for 1.
trans-[Os(en)2py(H)](OTf), 3. Sodium tert-butoxide (1.2 equiv)
was added to a d6-DMSO solution of 1 causing the solution to turn
faint pink. 1H NMR (d6-DMSO): 8.55 (d, 2H, J ) 4 Hz), 7.77 (t,
1 H, J ) 6 Hz), 7.65 (t, 2H, J ) 7 Hz), 2.0-2.6 (m, 16 H), -3.64
(s, 1H). Resonances for t-BuOH are seen at 1.20 and 2.60 ppm.
Despite its pink hue, no band maxima were observed in the visible
range. The pink color may is attributable to a trace of the cis-isomer,
which is the more stable form.
With the exception of the hydrogen atom(s) of the dihydrogen
(1) and the hydride (2), all hydrogen atoms were included in the
final refinement cycles at idealized positions with the isotropic
thermal parameter fixed at 1.2 times the equivalent isotropic thermal
parameter of the atom to which it is bonded. The dihydrogen ligand
cis-[Os(en)2py(H)](OTf), 3′. The same procedure was followed
as for 3 but with the cis-form of the dihydrogen complex, prepared
as previously described.5 The resultant solution is deep red. 1H NMR
(d6-DMSO): 8.92 (d, 2 H, J ) 6 Hz), 7.47 (t, 1 H, J ) 7 Hz), 6.94
(t, 2 H, J ) 6 Hz), 2.3-2.6 (m, 8 H), -4.02 (s, 1 H). Resonances
for t-BuOH are seen at 1.20 and 2.60 ppm. UV-vis (DMSO): 408
(5) Li, Z. W.; Taube, H. J. Am. Chem. Soc. 1994, 116, 9506.
(6) Li, Z. W.; Taube, H. J. Am. Chem. Soc. 1991, 113, 8946.
(7) (a) Hembre, R. T.; McQueen, J. S. J. Am. Chem. Soc. 1994, 116, 2141.
(b) Hembre, R. T.; McQueen, J. S. Angew. Chem. 1997, 36, 65.
(8) As described by: Crawford, T. H.; Swanson, J. J. Chem. Educ. 1971,
48, 382.
nm, ꢀ ) 3800 M-1 cm-1; 501 nm, ꢀ ) 3943 M-1 cm-1
.
trans-[Os(NH3)4(H2O)(H2)] (BPh4)2, 4. This compound was
prepared as previously reported.6
3816 Inorganic Chemistry, Vol. 42, No. 12, 2003

trans-[Os(en)2py(H)]2+ Studies
observed also for trans-[Os(en)2OAc(H2)]2+ and [Os(dppe)-
Cl(H2)]+.9,10
The Os-H bond lengths found for 1, 1.61(6) and 1.48(7)
Å, are close to those observed for [Os(en2)OAc(H2)]+
(1.59(1) and 1.60(1) Å), but the H-H distances differ
markedly. For the acetato complex, a distance of 1.34(1) Å
has been measured by neutron diffraction,9 that observed for
1 is 0.85(8) Å by X-ray diffraction, and that suggested by
the empirical correlation of JH-D with H-H distance is 1.15
Å. As expected, replacement of a σ-donor ligand by one that
is also a π-acceptor leads to a contraction of the H-H
distances. This culminates in instability of the metal-
dihydrogen bond as illustrated by the release of dihydrogen
when a co-ligand is a strong π-acid, e.g., CO, HCN,
isonitriles, or pyrazimium ion.11
Figure 1. Perspective drawing of the trans-[Os(en)2py(H2)]2+ cation
present in crystalline trans-[Os(en)2py(H2)](OTf)2 (1). All non-hydrogen
atoms are represented by thermal vibration ellipsoids drawn to encompass
50% of their electron density; hydrogen atoms are represented by small
spheres.
The four amine nitrogens attached to the metal center are
coplanar with a maximum atomic displacement of 0.066(2)
Å from the least-squares mean plane through the five atoms.
As expected, the pyridine ring is nearly perpendicular (87.9-
(1)°) to the equatorial plane already described. The Os-N
average12a distance, 2.142(4,5,7,4) Å, is close to that found
in trans-[Os(en)2(OAc)(H2)]+ which is reported9 as 2.13(5)
Å. The out-of-plane bending of the ethylene carbon backbone
is evidenced by the N-C-C-N dihedral angles in both
ligands of 54°, similar to that observed for trans-[Os(en)2-
(OAc)(H2)]+.
Oxidative Titration of 1. An aqueous solution of 1 (4.0 × 10-3
M) was titrated with a solution of ferricenium triflate of the same
concentration. The ensuing reaction is rapid, and the end point was
indicated by the persistence of absorbance maxima at 251 nm,
attributable to ferricenium ion.
Reduction of [MAH]I with trans-[Os(en)2py(H)](OTf)2. An
aqueous solution (0.5 mL, 0.03 M) of 2 was added to an equimolar
solution of [MAH]I and the resulting solution stirred for 1 h in the
dark. A white solid was removed from the resulting yellow solution
by filtration which was washed with H2O and dried in vacuo. NMR
Although the preparation of 1 has been described in the
literature,5 the absorption spectrum has not been reported.
We observe an intense absorption band at 340 nm and assign
it to a πd f π* (pyridine) transition. This is shifted to higher
energy as compared to that in [Ru(NH3)5(py)]2+, a result of
the replacement of an NH3 ligand by the electron withdraw-
ing ligand η2-H2. The absorbance at 242 nm is the π f π*
transition for pyridine, and that at 202 nm we attribute to a
πd f σ* (H2) transition. Many of the dihydrogen complexes
of the OsII amine series show a well defined absorption band
of high intensity in this energy region, not attributable to
absorption by co-ligands.13
1
analysis of the solid in CDCl3 confirms it as being MAH2. H
NMR: δ 7.19 (t), 7.15 (d), 6.91 (t), 6.87 (t), 3.86 (s, 3H). UV-vis
(CH3CN): 349, 395, 411.
Reduction of [MAH]I with trans-[Os(en)2py(H)]OTf. A solu-
tion of [MAH]I (5.8 mg, 0.018 mmol) was added to an NMR tube
containing 0.50 mL of 3.6 × 10-2 M in d6-DMSO. The solution
darkened, and the white solid which forms was filtered off, washed
with H2O, and dried in vacuo. The 1H NMR spectrum of the white
1
solid dissolved in CDCl3 confirms the formation of (MAH)2. H
NMR: δ 7.12 (t), 6.71 (t), 6.48 (d), 3.92 (s, 2H), 2.99 (s, 6H).
Synthesis and Properties of [Os(en)2py(H)](OTf)2, 2.
The titration of 1 in methanol with either [Fc]OTf or AgOTf
shows that one proton is released for each equivalent of
oxidant consumed; no gas evolution is observed. No 1H NMR
signals are observed for 2 in solution except for two broad
signals at 20 and 25 ppm. An IR band at 2020-1 cm is
attributed to the Os-H stretch which replaces the dihydrogen
signal of 1 at 2292 cm-1. All other bands are almost identical
to those observed for 1. This suggests that the electron
Reduction of 2 in MeOH or H2O. Zn/Hg amalgam (∼300 mg)
was added to an aqueous solution of 2, and the mixture was stirred
for 12 h in darkness. The solvent was removed in vacuo and the
1
residue extracted with d4-MeOH. H NMR analysis yielded a 9/1
mixture of 1 and the cis-isomer, 1′. Reformation of the dihydrogen
complexes of Os(II) was quantitative as demonstrated by conducting
the reduction of 2 in d4-MeOH containing an internal standard. The
reduction of 2 is slow enough so that the trans-isomer, if formed
as the initial product, can isomerize to the cis.5
Reduction of 2 in MeCN. The procedure already described was
followed, but in this case, the residue was extracted with CD3CN.
Analysis by 1H NMR shows that three products, trans-[Os-
(en)2MeCN(H2)]2+/1′/1, are formed in a 90:8:2 ratio, characterized
by dihydrogen resonances at δ -9.31, -7.76, and -8.78 for the
three products, respectively.
(9) Hasagawa, T.; Li, Z. W.; Parkin, S.; McMillan, R. K.; Koetzle, T. F.;
Taube, H. J. Am. Chem. Soc. 1994, 116, 4352.
(10) Maltby, P. A.; Schlaf, M.; Steinbeck, M.; Lough, A. J.; Morris, R.
H.; Klooster, W. T.; Koetzle, T.; Srivastava, R. C. J. Am. Chem. Soc..
1996, 118, 5396.
(11) Li, Z. W. Ph.D. Thesis, Stanford University, 1993, pp 61-62
(12) The first number in parentheses following an average value of a bond
length is the root-mean-square estimated standard deviation of an
individual datum. The second and third numbers are the average and
maximum deviations from the average value, respectively. The fourth
number represents the number of individual measurements which are
included in the average.
Results and Discussion
The molecular structure of 1 as determined by X-ray
diffraction is shown in Figure 1. The H-H axis lies between
the N atoms of the two chelate rings in an orientation
(13) Reference 11, p 147 and pp 136-139.
Inorganic Chemistry, Vol. 42, No. 12, 2003 3817

McQueen et al.
densities on the metal atoms in [Os(en)2py(H2)]2+ and [Os-
(en)2py(H)]2+ are almost the same, an accidental result of
dihydrogen acting as an electron withdrawing ligand on OsII,
compensating for the electron donating capacity of the
hydride anion (H-) when the metal is in the higher oxidation
state.
of H- leaves a vacant site on the metal, which can be filled
by the solvent so that the signal at -0.34 V can reasonably
be assigned to the [Os(en)2(py)CH3CN]3+/2+ couple. An
alternative reaction mode analogous to that of [Os(NH3)5-
(H)]2+ with acetone, to yield the isopropyl alcohol complex
of [Os(NH)5]3+, is that the product of the addition of H- to
CH3CN remains attached to the metal.
A question that arises is whether 2 is stable. As a solid, 2
is robust and can be stored for extended periods of time
without change. When it is dissolved, the question arises as
to whether it is stable with respect to disproportionation to
1 and a monohydride of OsIV in the thermodynamic sense,
or whether it is only kinetically stable and appears in the
present system because the one electron oxidation of 1 is
much more rapid than that of 2. The equilibrium issue cannot
be answered on the basis of the present information. What
is known is that, at moderate acidities (1 × 10-3 to 1 M), 1
and the OsIV monohydride coexist for long periods of time
without evidence of reaction. This is not surprising in view
of the fact that one electron oxidation of 2, unless the driving
force is very great, is expected to be slow because an increase
in coordination number accompanies the oxidation.2,3 On this
account, regardless of the state of the equilibrium, the
disproportionation of 2 is expected to be slow. The oxidation
of 2 to the corresponding monohydride of OsIV is found to
be much slower than is oxidation of 1 to 2, in keeping with
the aforementioned deep seated structural change accompa-
nying the former. The product of the second stage has a
magnetic moment of 2.6 µB, close to the values observed
for other OsIV(en)2 monohydrides which have been encoun-
tered.2 The quantitative formation of 2 by the use of one
electron oxidants raises the question of why the oxidation
of [OsII(en)2(H2)(Fe(CN)6)]2-, as mentioned in the Introduc-
tion, takes such a different course from that of 1. The reaction
of O2 with [OsII(en)2(H2)(Fe(CN)6)]2- has been observed to
be 500 times faster than it is with the corresponding complex
with [Ru(CN)6]4- replacing [Fe(CN)6]4-.14 The major dif-
ference between the two anions is that [Fe(CN)6]4- is
oxidized at a much lower potential than [Ru(CN)6]4- so that
[Fe(CN)6]4- can mediate in the redox reactions at issue. Thus,
it is an atypical ligand, and it can be expected that stable (at
least kinetically stable) OsIII monohydride complexes of the
present series will be found with a wide range of co-ligands.
Electrochemical Studies. Cyclic voltammetry of 1 in dry
acetonitrile solution yields an irreversible one electron
oxidation wave at +0.65 V (vs Fc+/Fc couple) when a
switching potential of 1.0 V is maintained. When repeated
scans from 1.0 to -1.5 V are taken, a small reduction wave
is observed at -1.1 V. In cyclic voltammetry, 2 undergoes
one electron oxidation at +0.95 V and one electron reduction
at -1.2 V, the complementary waves being absent. While
the latter potential is close to that observed for 1 on repeated
scans, it is probably not ascribable to an accumulation of 2.
When multiple scans are taken with 2, nearly reversible
waves are seen at -0.34 and -0.77 V.
When methanol is the solvent for 2, no subsidiary waves
appear even on repeated scans, which is behavior consistent
with the attribution made for such waves in the reducible
solvent, CH3CN. An oxidation wave is observed at 0.74 V
and a reduction wave at -0.82 V, neither accompanied by
a signal for the complementary processes. When the medium
is acidified (0.10 M HOTf), there is no change in the cathodic
wave, and no hint of the appearance of complementary
oxidation wave. Acidification does shift the anodic wave to
0.64 V, but again, there is no hint of a complementary
process. It is difficult to see how protons could affect the
oxidation of 2 to the corresponding OsIV species, and it is
likely that what is being observed is an effect of the anion:
it, rather than H2O, is acting as the seventh ligand.
When cyclic voltammetry is performed on [Os(NH3)5-
(H2)]2+ in water, a one electron anodic signal is observed at
-0.12 V, without a signal in this potential range for the
complementary reduction. However, in 1 M H+, a fully
reversible signal for the [Os(NH3)5(H2)]3+/2+ couple is
observed, (E1/2 ) -0.14 V) showing that [Os(NH3)5(H2)]3+
is not a strong acid. Because on the addition of acid with 2
as the subject of study, there is no hint of reduction to [Os-
(en)2py(H2)]3+, we can conclude that at the prevailing
concentration, it is almost completely deprotonated in water,
the latter is a much stronger acid than [Os(NH3)5(H2)]3+, a
result which is in harmony with the much reduced basicity
of the co-ligand pyridine compared to ammonia or an amine.
There is an interesting difference between the behavior
of [Os(NH3)5H]2+ which is formed on one electron oxidation
of [Os(NH3)5(H2)]2+ in unacidified water and that of species
2. After [Os(NH3)5(H)]2+ is formed, it undergoes a two
electron irreversible oxidation at 0.33 V which leads to [Os-
(NH3)5H2O]3+ and H+, corresponding to the oxidation of H-.
In sharp contrast, trans-[Os(en)2py(H)]2+ undergoes one
electron oxidation to a monohydride of OsIV; i.e., the hydride
remains intact. The difference is less ascribable to the
difference in four NH3 compared to two en species as co-
ligands, than to that between py and NH3. Because of the
electron withdrawing power of py, more electron density is
withdrawn from the hydride ligand in the case of 2, making
it less susceptible to oxidation.
In the context of this, it is of interest to note that the one
electron oxidation of a coordinated hydride by an external
reagent has also been encountered.2 The action of a one
electron oxidant on the monohydride of OsIV yields OsIII:
[Os(en)2H]3+ ) [Os(en)2]3+ + H+ + e-
OsIV and the external oxidant cooperate in converting H- to
H+.
In a following section, 2 is shown to be an active H-
transfer agent, as is the case also for [Os(NH3)5(H)]2+.1 Loss
Magnetic Susceptibility Measurements on 2. The NMR
samples were prepared as outlined in the subsection General
(14) Reference 11, p 147.
3818 Inorganic Chemistry, Vol. 42, No. 12, 2003

trans-[Os(en)2py(H)]2+ Studies
hydride. In the cis-isomer, such labilizing effects commonly
are less pronounced in line with the behavior we observe.
When the reaction is conducted in H2O or methanol, cis-
[Os(en)2py(H2)]2+ is produced in 90% yield. The important
difference likely is that, in solvents where disassociable
protons are abundant, reduction is rapid, and the isomeriza-
tion occurs after reduction.
Use of 2 as a Reducing Agent. The reducing capabilities
of 2 were investigated using the organic oxidant N-methyl-
acridinium ion (MAH+). This substrate is readily reduced
by either electron transfer or hydride transfer, the products
[MAH]2 and MAH2, respectively, being easily identified.
When an aqueous solution of 2 is added to an equimolar
solution of MAH+, the hydride transfer product MAH2 is
formed nearly quantitatively, a small amount (3%) of the
electron transfer product being formed. When the same
reaction is run in acetonitrile, formation of MAH2 is
quantitative. The facile hydride transfer by a monohydride
of OsIII has precedent in the quantitative reduction of acetone
to 2-propanol by [Os(NH3)5(H2)]3+ (≡[Os(NH3)5H]2+ + H+).
Figure 2. Perspective drawing of the trans-[Os(en)2py(H)]2+ cation present
in crystalline trans-[Os(en)2py(H)](OTf)2 (2). All non-hydrogen atoms are
represented by thermal vibration ellipsoids drawn to encompass 50% of
their electron density; hydrogen atoms are represented by small spheres.
Procedures, with CH2Cl2 being used as an internal reference.
The data were treated as described in ref 8, resulting in a
value of 1.8 µB at 300 K, as expected for one unpaired
electron.
In principle, [Os(en)2py(H)]2+ can react with [Os(en)2py-
(H2)]2+ by transfer of a H atom (equivalent to a proton and
electron) from the latter to the former. If rapid enough, its
1
occurrence could be observed by line broadening of the H
Structure of 2 in the Solid State. Analysis of the results
of the X-ray diffraction study shows the amine ligands to
be in a plane with hydride and pyridine occupying the
remaining positions (Figure 2). The pyridine N-Os bond
length in 2 is 2.183(4) Å compared to 2.135(4) Å in 1 while
the amine N-Os average12 bond lengths compare as 2.118-
(4,6,11,4) and 2.142(4,5,7,4) Å, respectively. The contraction
of the pyridine N-Os bond length for the metal in the lower
oxidation state is ascribable to a greater contribution by back-
bonding to the interaction for OsII compared to OsIII, with
this extra component more than compensating for the greater
radius of OsII.15 In contrast, with the amines acting as simple
σ-donors, the bond distances in the two oxidation states show
the more usual relationship. The Os-H bond length in 2 is
1.69(5) Å to be compared with the average12 of Os-H bond
lengths in 1 as 1.55(7,7,7,2) Å. Apparently, the shrinkage
of the electron cloud of H- occasioned by the addition of a
proton more than compensates for the increased radius of
the metal ion. In this context, comparisons of these distances
when pyridine is replaced by a simple σ-donor ligand, which
results in a much weaker H-H bond, would be of interest.
The remaining distances and angles for 1 and 2 are very
similar.
NMR signals observed for 1. With the concentrations of the
reactants at the millimolar level, no line broadening was
observed in an experiment performed to look for this effect.
A methanolic solution of 1 and MAH+, at millimolar
concentrations and kept overnight, shows no evidence of
reduction of MAH+, but almost complete conversion to the
cis-isomer does occur.
Properties of the trans- (3) and cis- (3′) Forms of [Os-
(en)2py(H)]+. Qualitative experiments17 show that the con-
version of η2-H2 to the deuterated form in [OsII(en)2L(η2-
H2)] complexes in D2O is accelerated when base is added to
the medium, and that it takes place much more rapidly when
the L is a π-acid such as pyridine than when it is a simple
σ-donor (trans-[Os(en)2OH(H2)]+ shows no exchange in 1.0
M OD- even after many hours). Though pyridine as co-
ligand enhances the acidity of the coordinated η2-H2, 1 still
is very weak acid, but it can be deprotonated by the action
of a strong base in a nonprotic solvent. On the addition of
1
Na[t-BuO] to a solution of 1 in d6-DMSO, the H NMR
signal at -7.8 ppm disappears, and a new signal correspond-
ing to a single proton appears at -3.6 ppm. We assign this
peak to trans-[Os(en)2H(py)]+. The cis-isomer, 3′, of the
monohydride has also been produced by a similar procedure;
Reduction of 2. When a solution of 2 in acetonitrile is
stirred over Zn/Hg for 12 h in the dark, trans-[Os(en)2(CH3-
CN)(H2)]2+ is formed in 90% yield. Along with this product,
free pyridine appears as does a small amount of cis-[Os-
(en)2py(H2)]2+. The reduction of 2 in acetonitrile is expected
to be slow, because H+ must be gathered from proton-labile
impurities to convert H- to H2, affording time for partial
conversion of 2 to the cis-isomer, 2′. We attribute the
replacement of pyridine to the labilizing effect of the trans-
1
it is characterized by a H NMR signal at -4.0 ppm. No
isomerization of either form is observed on the time scale
of our observations, several hours.
Attempts to isolate solid forms of either isomer by the
addition of ether to solutions in DMSO failed despite carrying
out the operations in a controlled atmosphere box. When
the solid, colorless when prepared from the trans-isomer and
(16) Luetkens, M. L.; Elcesser, W. L.; Huffman, J. C.; Settleberger, A. P.
J. Chem. Soc., Chem. Commun. 1983, 1072.
(17) Li, Z. W. Unpublished results.
(15) Gress, M. E.; Creutz, C.; Quicksall, C. O. Inorg. Chem. 1981, 20,
1522.
Inorganic Chemistry, Vol. 42, No. 12, 2003 3819

McQueen et al.
red when prepared from the cis-isomer, is dissolved in d4-
CH3OH for 1H NMR measurements, the hydride resonances
disappear. Both hydrides decompose slowly in d6-DMSO
solution; the decomposition is obvious in the case of the cis-
isomer because of its color. The changes were also followed
by 1H NMR spectroscopy. No new signals appear on
decomposition indicating that the ligands remain attached
to the metal, which has been oxidized to a paramagnetic state.
Both 3 and 3′ are highly reactive reducing agents, as shown
by the prompt disappearance of the hydride NMR signals
when acetonitrile or acetone is added to the solution.
Presumably, reaction with solvent also accounts for their
disappearance in the DMSO. Species 3 reacts rapidly with
MAH+ to form [MAH]2.
Much of the difference in behavior of the monohydrides
of OsII and OsIII as reducing agents can be understood without
reference to the oxidant, on the basis of their properties alone.
While the extent of covalency in the bond between the metal
and hydride, which would be much less for the metal in the
2+ than in the 3+ state, would suggest that the former, being
more like NaH, would be even more disposed to hydride
transfer. This argument leaves out of consideration that the
simple loss of an electron from 3 produces a stable product,
the monohydride of OsIII, and because of its lower Franck-
Condon barrier, a one electron change is favored over a two
electron change. As to the monohydride of OsIII where
hydride transfer is observed, again by considering its
alternative, the loss of an electron, a low energy path is
accessible only when it is accompanied by a change in
coordination number, and it is thus inherently slow. The
circumstances governing the choice of mechanism thus are
rather specific to the present system and do not lend
themselves to simple generalization.
The striking difference in the color of 3 and 3′, and more
particularly why 3 is colorless, merits consideration. The
change from H2 to H- as co-ligand can hardly cause the πd
f π* transition, metal to pyridine, to disappear. Replacing
η2-H2, which is electron withdrawing, by H-, which is
electron releasing, would raise the πd level relative to the
π* of pyridine, and if the resulting energy difference were
small enough, the transition would appear in the NIR and
be in a range which was not examined in our work. Another
source of charge transfer absorption is interligand, which
leads to a striking expected difference between the cis- and
trans-forms. In the cis-geometry, direct overlap of the
σ-orbital of H- with a lobe of the π* orbital of pyridine is
possible and would account for one of the absorption bands,
while in the trans-disposition, such overlap is nonexistent
because of the σ-character of the filled orbital in hydride
ion in relation to the π*-orbital of pyridine. This interaction
can explain the appearance of two absorption bands for the
cis-isomer in the visible region. By raising the π*-level
sufficiently the πd-π*absorption, suggested as appearing
in the NIR region for the trans-form, now appears in the
visible. It is to be noted that, in lacking π-character, H- is
unique among mononuclear anions.
mediate stage suggesting that initial attack by a proton does
not take place on the bound hydride. When DMSO solutions
of 3 and 3′ are diluted with H2O, yellow solutions result;
the change is particularly striking in the case of 3′ where
the red color is dissipated in less than a minute, being
replaced by new bands at 362 and 431 nm, which, within
90 min themselves, are replaced by bands at 336 and 393
nm, these corresponding to the formation of cis-[Os(en)2py-
(H2)]2+, the product expected upon protonation of 3′. Two
alternatives for the intermediate stage are the following: first,
protonation of the π-cloud of the pyridine, made a possibility
by the massive transfer of the πd-electrons of the metal to
the π-acid, or direct protonation of the πd-electrons of the
metal. Of the two, the former seems to be the more likely
because of the depletion of the πd electron density by the
π-acid.
The behavior of 3 is quite different in that the dihydrogen
complex is not reconstituted on protonation. The yellow color
which appears on reaction with water, attributable to an
absorption band at 438 nm (ꢀ ∼ 1.04 × 102 M-1 cm-1),
reaches a maximum in 10 min and then slowly fades. The
product of this change has not been identified. Protonation
of the π-electron cloud of the pyridine is not useful for the
conversion of H- to H2, but the resulting species could serve
as an intermediate stage in reduction of pyridine with
concomitant oxidation of the metal. Protonation with strong
acid leads to the liberation of H2, without reconstituting the
dihydrogen complex.
Hydrogenation by Dihydrogen Complexes of Osmium-
am(m)ines. In contrast to the action of the monohydrides
of OsII and OsIII, 2 and 3, respectively, as described, under
similar conditions, trans-[Os(en2)py(H2)]2+, 1, shows no
evidence of reaction with (MAH)+ even after many hours
of exposure. This outcome is not surprising considering that
the dihydrogen complex is not a good one electron reducing
agent, nor is transfer of H- from it to a substrate expected
to be a kinetically facile process. The immediate product of
such a reaction is [Os(en)2py(H)]3+, which can be regarded
as a monohydride of OsIV, but which requires an additional
ligand to be stable.2 As the experiments to be described show,
with suitable modification of composition, the dihydrogen
complexes OsIIam(m)ines can act as hydrogenation agents.
Most of the experiments were done18 with a solid of
composition trans-[Os(NH3)4(H2)(H2O)](BPh4)2, 4, in acetone-
d6 as solvent. Hydrogenation activity depends on access by
the substrate to a coordination site on the metal atom, a
condition which is satisfied by the tetraammine dihydrogen
-
residue in solution whether, H2O, BPh4 , or d6-acetone, each
of which is readily replaceable occupies the sixth position,
but not satisfied by pyridine.
Despite H2 initially being trans to the accessible site, in a
solution containing the dihydrogen complex at the 0.01 M
level, with CH3CHO at 0.46 M, after 2 h free CH3CH2OH
is detected at a yield corresponding to the consumption of
48% of the dihydrogen complex. The reduction of aldehyde
The reactions with H2O show unexpected complications
in that conversion to the final products involves an inter-
(18) See ref 11, pp 106-110.
3820 Inorganic Chemistry, Vol. 42, No. 12, 2003

trans-[Os(en)2py(H)]2+ Studies
is competitive with the formation of trans-[Os(NH3)4(η2-CH3-
CHO)L]2+, where L is presumably d6-acetone, in 41% yield.
Ten minutes after the start of the reaction, a strong signal
for η2-H2 is observed at -10.64 ppm, compared to that
observed before the addition of aldehyde, -13.38 ppm.
Taking account of the composition of the system, because
the shift is in the region characteristic of oxygen donors, we
attribute the signal to η1-aldehyde, this intermediate then
collapsing to trans-[Os(NH3)4(η2-CH3CHO)L]2+ with release
of H2, or rearranging to the cis-form, following which,
reduction of the aldehyde occurs. No signal for the osmium-
containing product of the latter path is observed, and we infer
that the OsII species, which after hydrogenation loses the
stabilizing action of η2-H2 as a π-acid, undergoes oxidative
addition to a paramagnetic product by some component of
the mixture. It should be mentioned that π-acids such as
pyridine and CN- are known to favor isomerization of the
trans- to the cis-form, and to facilitate the release of H2,
and thus, η1-CH3CHO is expected to act in a similar fashion.
When styrene was used as the substrate, present at the
0.020 M level, with the dihydrogen complex at 0.010 M,
after 8 h, free ethylbenzene corresponding to 12% of the
dihydrogen complex was detected. The only osmiumammine
product observed, comprising 50% of that introduced, was
trans-[Os(NH3)4(η2-alkene)L]2+. On starting with trans-[Os-
(en)2(η2-H2)(CF3SO3)]CF3SO3, no hydrogenation was ob-
served, and the only osmiumamine product detected was cis-
[en2(η2-alkene)2]2+, in 70% yield. It is to be noted that rates
of isomerization can be quite different for (en)2 as compared
to (NH3)4 as co-ligands.
condition for hydrogenation to occur, is not a sufficient
condition. When H2CdCHsCH2sSCH3 is used as a sub-
strate in reaction with trans-[Os(en)2(η2-H2)(CF3SO3)]CF3SO3,
quantitative formation of the cis-dihydrogen complex with
the substrate attached to the metal at the sulfur site is
observed, and there is no subsequent reduction of the alkene
function even though the cis-geometry is compatible with
close contact between the dihydrogen and the alkene group.
Thus, it is demonstrated that, for hydrogenation of an olefin
to occur, both substrate and H2 must be attached to the metal.
With 2-vinyl pyridine, 2 h after the start of the reaction, free
2-ethylpyridine had appeared corresponding to 4% of the
dihydrogen complex introduced; the bulk of the dihydrogen
complex appeared in the cis-form, with substrate bound at
N. With the passage of time, the amount of ethylpyridine
did not increase. In this case, the preferential binding to the
heteroatom compared to the alkene site is less favored than
it is for the sulfur derivative, and hydrogenation takes place
for a minor path in which alkene rather than N binds to the
metal while H2 is still attached.
Acknowledgment. This work was supported by the
National Science Foundation under Grant NSF-CHE 9727416.
Dr. Fred Hollander of the University of CaliforniasBerkeley
collected the X-ray crystallographic data.
Supporting Information Available: Details pertaining to the
X-ray crystallographic structural determination of 1 and 2 are
available in CIF format. This material is available free of charge
via the Internet at http://pubs.acs.org.
The following experiments show that contact of (η2-H2)
with an alkene function, though presumably a necessary
IC030091B
Inorganic Chemistry, Vol. 42, No. 12, 2003 3821