Oxidation and Protonation of Transition Metal Hydrides: Role of an Added Base as Proton Shuttle and Nature of Protonated Water in Acetonitrile

Article abstract of DOI:10.1021/ic960394o

The Cp₂Fe⁺ oxidation and the protonation of CpMoH(CO)₂L (L: PPh₃, 1; PMe3, 2) in MeCN have been investigated. In the dry solvent, the oxidation of both compounds consumes 1 mol of oxidant/mol of hydride with production of [CpMo(CO)₂L(MeCN)]⁺ (L: PPh₂, [3]⁺; PMe₃, [4]⁺) and H₂. The stoichiometry changes toward the consumption of 2 mol of oxidant in the presence of excess water when the oxidizing equivalents are added rapidly, either chemically or electrochemically. However, 1 oxidizing equiv is again sufficient to consume the hydride material completely under conditions of slow oxidation. Under comparable conditions, the more basic 2 leads to a lower [ox]/M-H stoichiometry. Protonation of 1 and 2 with HBF₄·Et₂O in dry MeCN results in rapid H₂ evolution and formation of [3]⁺ and [4]⁺, respectively, whereas the presence of excess water suppresses the H₂ evolution and gives rise to protonated water. However, this process is followed by slow and irreversible delivery of the proton back to 1 or 2 to afford [CpMoH₂(CO)₂L]⁺, which ultimately decomposes to [3]⁺ or [4]⁺ and H₂. The dihydride complex is too unstable to be isolated, even when the protonation of 1 or 2 is carried out in a noncoordinating solvent. The proton delivery is faster for the more basic 2 and slower for the less basic 1. Thus, water operates as a proton shuttle , whose speed depends on the basicity difference between the hydride complex and water. The identity of the protonated water in MeCN as [H(H₂O)₃₂]⁺ is suggested by an independent ¹H-NMR experiment in CD₃CN.

Full text of DOI:10.1021/ic960394o

5154
Inorg. Chem. 1996, 35, 5154-5162
Oxidation and Protonation of Transition Metal Hydrides: Role of an Added Base as Proton
Shuttle and Nature of Protonated Water in Acetonitrile
E. Alessandra Quadrelli, Heinz-Bernhard Kraatz, and Rinaldo Poli*
Department of Chemistry and Biochemistry, University of Maryland, College Park, Maryland 20742
ReceiVed April 9, 1996X
The Cp2Fe+ oxidation and the protonation of CpMoH(CO)2L (L: PPh3, 1; PMe3, 2) in MeCN have been
investigated. In the dry solvent, the oxidation of both compounds consumes 1 mol of oxidant/mol of hydride
with production of [CpMo(CO)2L(MeCN)]+ (L: PPh3, [3]+; PMe3, [4]+) and H2. The stoichiometry changes
toward the consumption of 2 mol of oxidant in the presence of excess water when the oxidizing equivalents are
added rapidly, either chemically or electrochemically. However, 1 oxidizing equiv is again sufficient to consume
the hydride material completely under conditions of slow oxidation. Under comparable conditions, the more
basic 2 leads to a lower [ox]/M-H stoichiometry. Protonation of 1 and 2 with HBF4‚Et2O in dry MeCN results
in rapid H2 evolution and formation of [3]+ and [4]+, respectively, whereas the presence of excess water suppresses
the H2 evolution and gives rise to protonated water. However, this process is followed by slow and irreversible
delivery of the proton back to 1 or 2 to afford [CpMoH2(CO)2L]+, which ultimately decomposes to [3]+ or [4]+
and H2. The dihydride complex is too unstable to be isolated, even when the protonation of 1 or 2 is carried out
in a noncoordinating solvent. The proton delivery is faster for the more basic 2 and slower for the less basic 1.
Thus, water operates as a “proton shuttle”, whose speed depends on the basicity difference between the hydride
complex and water. The identity of the protonated water in MeCN as [H(H2O)3]+ is suggested by an independent
1H-NMR experiment in CD3CN.
Introduction
The products of this decomposition process, e.g. [M(S)]+ and
[MH2]+, can either be stable or further decompose to other
products.
The main difference caused by the nature of the deprotonating
agent (external base, eq 2a, or unreacted metal hydride, eq 2b)
affects the number of moles of oxidizing agent consumed in
the process: 1 faraday/mol of hydride if M-H is the proton
trap (eq 5) or 2 faradays/mol of M-H if the proton is captured
by an external base (eq 6).
The oxidation of transition metal hydrides has received a great
deal of attention because of the fundamental role that the metal-
hydrogen bond plays in organometallic synthesis and catalysis1
and because of the higher reactivity that odd-electron systems,
e.g. 17-electron radicals obtained by oxidative processes, display
with respect to their 18-electron counterparts.2,3 It has been
established that, following one-electron oxidation, the acidity
of a metal hydride species is enhanced by some 20 orders of
magnitudes; e.g., the pKa decreases by a factor of ca. 20.4
Therefore, provided a sufficiently good base is present, oxidation
is usually followed by deprotonation to initiate the process of
decomposition of the 17-electron product. In certain instances,
the starting metal hydride complex is itself a sufficiently good
base to deprotonate the oxidized hydride. The various pos-
sibilities for the first steps of the oxidation process are
summarized in eqs 1-4 (ancillary ligands around the metal are
2M-H + S - 2e- f [M(S)]+ + [MH2]+
M-H + S + B - 2e- f [M(S)]+ + BH+
(5)
(6)
The nature of the reaction products differs in the two cases.
However, the dihydride product of eq 5 may or may not be
stable toward reductive elimination of H2 (eq 7) and production
of more [M(S)]+, resulting in the oxidation stoichiometry of eq
8. This possibility has precedent, for instance, for the oxidation
M-H - e- f [M-H]•+
(1)
[MH2]+ + S f [M(S)]+ + H2
(7)
(8)
[M-H]•+ + B f M• + BH+
(2a)
2M-H + 2S - 2e- f 2[M(S)]+ + H2
[M-H]•+ + M-H f M• + [MH2]+
(2b)
of RhH(dppe)2, giving [Rh(dppe)2]+ and H2, whereas the
corresponding iridium derivative gives a 1:1 mixture of [Ir-
(dppe)2]+ and [IrH2(dppe)2]+.5 Therefore, when H2 is reduc-
tively eliminated, the metal-containing product for both the 2
faraday/mol process (deprotonation by external base, eq 6) and
the 1 faraday/mol process (deprotonation by M-H starting
material, eq 8) is the same; the only difference between the
M• + S f M(S)•
(3)
(4)
M(S)• - e- f [M(S)]+
omitted for generality). Following deprotonation, the resulting
17-electron M• radical is trapped by a donor solvent molecule
S and further oxidized to afford an 18-electron [M(S)]+ species.4
(3) Organometallic radical processes; Trogler, W. C., Ed.; Elsevier:
Amsterdam, 1990; Vol. 22.
(4) Ryan, O. B.; Tilset, M.; Parker, V. D. J. Am. Chem. Soc. 1990, 112,
2618-2626.
(5) Pilloni, G.; Schiavon, G.; Zotti, G.; Zecchin, S. J. Organomet. Chem.
1977, 134, 305-318.
X Abstract published in AdVance ACS Abstracts, August 1, 1996.
(1) Crabtree, R. H.; Hamilton, D. G. AdV. Organomet. Chem. 1988, 28,
299-338.
(2) Astruc, D. Acc. Chem. Res. 1991, 24, 36-42.
S0020-1669(96)00394-1 CCC: $12.00 © 1996 American Chemical Society

Protonation of Transition Metal Hydrides
Inorganic Chemistry, Vol. 35, No. 18, 1996 5155
two pathways is the final oxidation state of the hydrogen atom:
H+ in one case and H2 in the other.
results that we have obtained show how water (or any other
base for that matter) can be expected to regulate the stoichio-
metric outcome of a metal hydride oxidation process under a
general set of conditions.
There are various reports on the oxidative chemistry of four-
legged piano stool Mo(II) and analogous W(II) hydrides under
a variety of experimental conditions (e.g. chemical vs electro-
chemical oxidation; with or without a strong base as proton
acceptor), but the observed results and especially their inter-
pretations are not always completely convincing. Kochi et al.
reported that CpMoH(CO)3 is oxidized in MeCN at a Pt
electrode to consume 1.08 ( 0.05 electrons per molybdenum
atom, producing [CpMo(CO)3]2.6 However, Tilset et al.
reported later that, under apparently identical conditions,
electrolysis of CpMoH(CO)3 required 2.1 ( 0.1 faradays/mol
to produce a solution of [CpMo(CO)3(MeCN)]+ (cf. the stoi-
chiometry of eq 6).4 Since this experiment was carried out
without added strong base, the proton is believed to be captured
Experimental Section
Materials. All operations were carried out under an atmosphere of
argon using standard glovebox and Schlenk-line techniques. Solvents
were dehydrated by standard methods (toluene and n-heptane from Na;
THF and Et2O from Na/K/benzophenone; MeCN from CaH2), deoxy-
genated, and distilled directly from the dehydrating agent under
dinitrogen prior to use. The amount of residual water in the solvents
was assessed by titration with a coulomatic Karl Fischer titrimeter
(Fisher Scientific). CpMoH(CO)3,8 CpMoH(CO)2(PPh3) (1),9 and
CpMoH(CO)2(PMe3) (2)10 were prepared as previously described.
1
Spectroscopic and Electrochemical Analyses. Samples for H-,
-
by the “medium”, this being the solvent, or PF6 from the
31P{1H}-, and 19F{1H}-NMR in thin-walled 5 mm glass tubes were
measured on Bruker WP200 and XL400 spectrometers. The 1H-NMR
spectra were calibrated against the residual proton signal of the
deuterated solvents. The 31P- and 19F-NMR spectra were calibrated
against 85% H3PO4 and trifluoroacetic acid, respectively, in capillary
tubes which were placed in a different 5 mm glass tube, containing
the same deuterated solvent used for the measurement. Cyclic
voltammograms were recorded with an EG&G 362 potentiostat
connected to a Macintosh computer through MacLab hardware/software;
the electrochemical cell was a locally modified Schlenk tube with a Pt
counter electrode sealed through uranium glass/Pyrex glass seals. The
cell was fitted with a Ag/AgCl reference electrode and a Pt (2 mm2)
working electrode. All potentials are reported vs the Fc/Fc+ couple.
Bulk electrolyses with coulometric analyses were carried out in a locally
designed double Schlenk cell with the two compartments being
connected by a medium frit at the lower end and a pressure-equalizing
open tube at the upper end. One compartment of the cell was fitted
with a Pt or Au gauze working electrode and a Ag/AgCl reference
electrode, while the other compartment was fitted with a Pt counter
electrode. For the coulometric experiments carried out on a thiol-coated
Au electrode, the electrode was first cleaned by exposure to chromic
acid and then to HF, followed by rinsing with water and immediate
immersion into an ethanolic solution of HO(CH2)nSH (n ) 2 or 4) for
ca. 3 h, according to the published procedure.11 n-Bu4NPF6 (Aldrich)
was used directly as received as the supporting electrolyte for all
electrochemical experiments. Elemental analyses were performed by
MHW Laboratories (Phoenix, AZ). The EPR spectroelectrochemical
cell was a modified version of a cell described in the literature.12 The
modification consists of replacing the aqueous reference electrode with
a silver wire pseudoreference electrode, in order to allow measurements
at subfreezing temperatures. Measurements were carried out at the
X-band microwave frequency on a Bruker ER 200 D spectrometer
upgraded to ESP 300, equipped with a cylindrical ER/4103 TM 110
cavity. Solutions of the analyte (0.1-0.2 M) were electrolyzed by
application of a constant current of 1 mA with an EG&G 362
galvanostat.
supporting electrolyte, or residual traces of water. Chemical
oxidation with Fe(phen)3 (2 equiv) also produced [CpMo-
3+
(CO)3(MeCN)]+ and gave evidence (1H-NMR) for production
of a strong acid, characterized by a broad resonance at δ 10.5.4
The compound CpWH(CO)3 behaves identically to its Mo
counterpart, whereas the oxidation of CpWH(CO)2(PMe3) gave
different results depending on conditions: with Fc+ salts (Fc+
) Cp2Fe+, ferrocenium) as the oxidant in MeCN, a 1:1 mixture
of [CpW(CO)2(PMe3)(MeCN)]+ and [CpW(CO)2(PMe3)H2]+
was obtained (cf. eq 5), whereas in the presence of excess 2,6-
lutidine, the MeCN complex was the sole W product with
consumption of 2 mol of oxidant (cf. eq 6). At a Pt anode and
in the absence of added base, however, 2 faradays/mol was
consumed to produce once again the MeCN complex. This
unexpected change of stoichiometry from chemical to electro-
chemical oxidation (both carried out without strong base) was
tentatively attributed to the greater basicity of the “medium” in
the electrochemical experiment; i.e., the tungsten starting
complex is the only available proton acceptor in the dry MeCN
used for the chemical oxidation, whereas traces of another base,
presumably water, unavoidably introduced with the supporting
electrolyte, would serve as the proton scavenger during the
electrochemical experiment. The oxidation of CpMoH(CO)2-
(PPh3), 1, was also investigated chemically and electrochemi-
cally.7 The stoichiometry is that of eq 6 when Fc+ in MeCN is
used as the oxidant in the presence of 2,6-lutidine, to produce
[CpMo(CO)2(PPh3)(MeCN)]+, and the same stoichiometry is
found for the anodic oxidation at a Pt electrode in MeCN (1.94
( 0.08 faraday/mol). A chemical oxidation in a dry medium
and without added strong base has not been reported for this
compound.
There are questions that still do not have an answer from the
above studies: although the basicity of the “medium” has been
invoked to explain a host of different observations, especially
the coulometric results, and it has been hinted that water may
be the responsible player,4 more detailed studies of the
“medium” basicity, specifically concerning the product distribu-
tion and oxidation stoichiometry as a function of the water
concentration, have not been reported. This contribution
addresses just this question on the basis of a comparative study
of the chemical and electrochemical oxidation of the compounds
CpMoH(CO)2L (L: PPh3, 1; PMe3, 2) and their protonation by
HBF4‚Et2O with or without added water. It will be shown that
water is indeed the important player for the above systems and
for many of those previously reported in the literature. The
Preparation of [CpMo(CO)2(PMe3)(MeCN)]+PF6- ([4]+PF6-) by
-
Oxidation of 2 with Fc+PF6 in MeCN. To a yellow solution of 2
(90 mg; 0.31 mmol) in CD3CN (3 mL) was added solid Fc+PF6- (104
mg; 0.306 mmol) in four portions (25, 25, 35, and 19 mg) over a period
of 20 min, causing an effervescence after each addition and a deepening
-
in color (total Fc+PF6 added: 104 mg, 0.306 mmol). The solution
was examined by NMR and IR spectroscopy after each addition of
Fc+PF6-, showing the clean formation of Fc (1H NMR δ 4.14) and a
single product [1H-NMR δ 5.62 (d, 5H, JHP ) 0.6 Hz, Cp), 1.59 (d,
JHP ) 10.4 Hz, 9H, PMe3); 31P{1H}-NMR δ 13.6 (s, PMe3), -140.5
(8) Organometallic Syntheses; King, R. B., Ed.; Academic Press: New
York, 1965; Vol. 1, pp 156-159.
(9) Bainbridge, A.; Craig, P. J.; Green, M. J. Chem. Soc. A 1968, 2715.
(10) Kalck, P.; Pince, R.; Poilblanc, R.; Roussel, J. J. Organomet. Chem.
1970, 24, 445-452.
(6) Klinger, R. J.; Huffman, J. C.; Kochi, J. K. J. Am. Chem. Soc. 1980,
102, 208-216.
(7) Smith, K.-T.; Tilset, M. J. Organomet. Chem. 1992, 431, 55-64.
(11) Becka, A. M.; Miller, C. J. J. Phys. Chem. 1992, 96, 2657-2668.
(12) Allendoerfer, R. D.; Martinchek, G. A.; Bruckenstein, S. Anal. Chem.
1975, 47, 890.

5156 Inorganic Chemistry, Vol. 35, No. 18, 1996
Quadrelli et al.
(septet, JPF ) 707 Hz, PF6-); IR 1987, 1898 cm-1], assigned to
[CpMo(CO)2(PMe3)(CD3CN)]PF6 (cis isomer). The clear yellow-brown
solution was evaporated to dryness, and the ferrocene was extracted
with heptane (4 × 10 mL), leaving an ochre microcrystalline powder.
Yield: 0.07 g (80%). Anal. Calcd for C12H14D3F6MoNO2P2: C, 29.89;
H, 2.93; D, 1.24; N, 2.90. Found: C, 30.0; H, 4.0; N, 2.8.
(d) In the Presence of Water: Fast Titration. Compound 2 (7.5
mg, 0.026 mmol) was dissolved in MeCN (330 µL), and water (5.5
µL, 0.31 mmol) was added. In a separate Schlenk tube, FcPF6 (33.3
mg, 0.101 mmol) was dissolved in MeCN (0.90 mL). Aliquots of the
ferrocenium solution were added to the solution of 2 via microsyringe.
A first addition of 0.43 mL (1.9 equiv) resulted in the complete
bleaching of the ferrocenium blue color within 30 s, without any
noticeable gas evolution, while the yellow color of the initial solution
of 2 was now orange. An additional 40 µL addition (total 2.1 equiv)
did not lead to discoloration of the ferrocenium solution. The overall
addition was complete in less than 5 min.
An analogous reaction was carried out in regular acetonitrile: 2 (155
mg, 0.53 mmol) and Fc+PF6- (178 mg, 0.53 mmol) were added to 10
mL of CH3CN. After 5 min of stirring at room temperature, the solution
was evaporated to dryness and the residue washed with Et2O. 1H-
NMR (CD3CN, δ): 5.62 (d, 5H, JHP ) 0.5 Hz, Cp), 2.38 (d, 3H, JHP
) 2.6 Hz, MeCN), 1.59 (d, 9H, JHP ) 10.3 Hz, PMe3). 1H-NMR
(CD2Cl2, δ): 5.57 (s, 5H, Cp), 2.48 (d, 3H, JHP ) 2.4 Hz, MeCN),
1.62 (d, 9H, JHP ) 10.0 Hz, PMe3). 31P{1H}-NMR (CD2Cl2, δ): 9.7.
In CD3CN, the resonance of coordinated MeCN decreased to ca. 10%
of its intensity after 8 h at room temperature, and the resonance of free
MeCN at δ 1.95 increased in intensity during the same time. Minor
bands in both 1H- and 31P-NMR spectra are assigned to the trans isomer
(< 10% with respect to the main isomer): 1H-NMR (CD2Cl2, δ): 5.39
(d, JHP ) 2.4 Hz, Cp), 2.45 (s, MeCN), 1.65 (d, JPH ) 10.0 Hz, PMe3).
31P-NMR (CD2Cl2, δ): 17.3.
1H-NMR Study of the Reaction between CpMoH(CO)2(PPh3) and
Fc+. Compound 1 (9.2 mg, 0.019 mmol) and n-Bu4NPF6 (78 mg, 0.20
mmol) were introduced into a Schlenk tube and dissolved in CD3CN
(0.4 mL). To this solution was added H2O (1 µL, 0.06 mmol), and the
resulting mixture was transferred to the NMR tube. In a separate
Schlenk tube, a solution of FcPF6 (24 mg, 0.072 mmol) in 0.38 mL of
CD3CN was prepared. Aliquots of the FcPF6 solution (50, 50, and 20
µL; total 0.023 mmol) were added to the NMR tube at 30 min intervals.
1
The reaction was monitored by H-NMR. After each addition, the
characteristic blue color of Fc+ slowly disappeared (t1/2 ca. 5 min) and
the 1H-NMR spectrum showed a broad ferrocene resonance, which
sharpened as the blue color disappeared. After 30 min from each
addition, a sharp Fc resonance at δ 4.1 was again observed. At this
point, 1H-NMR showed the presence of residual resonances of
compound 1 (<20%) and the resonances corresponding to those already
described7 for [3]+ (>80%).
Generation of trans-[4]+ and Its Transformation to the cis
Isomer. To a yellow solution of 2 (12 mg; 0.041 mmol) in CD3CN (3
-
mL), solid Ph3C+BF4 (20 mg; 0.060 mmol) is added. The reaction
immediately showed the formation of Ph3CH [1H NMR δ 7.2] and a
single product [1H-NMR δ 5.59 (s, 5H, Cp), 1.62 (d, JHP ) 10.3 Hz,
9H, PMe3); IR 1992, 1905 cm-1], assigned to trans-[CpMo-
(CO)2(PMe3)(CD3CN)]BF4; a second product accumulated with time
[1H-NMR δ 5.63 (d, JHP ) 0.5 Hz, 5H, Cp), 1.61 (d, JHP ) 10.4 Hz,
Protonation of 2 with HBF4. (a) In Acetonitrile by HBF4/H2O.
2 (35 mg, 0.11 mmol) was dissolved in MeCN (3 mL). To the resulting
pale yellow solution was added HBF4 (48% in H2O, 25 µL, 0.19 mmol).
A smooth change of color to darker orange and gas evolution were
observed. After 1 h of stirring at room temperature, the IR of the
solution in the CO stretching region showed the complete conversion
to cis-[4]+ (1987 and 1897 cm-1).
9H, PMe3); IR 1988, 1900 cm-1
(CO)2(PMe3)(CD3CN)]BF4.
] assigned to cis-[CpMo-
1H-NMR Study of the Reaction between CpMoH(CO)2(PMe3)
and Fc+. (a) In CD3CN with n-Bu4NPF6. Compound 2 (24 mg, 0.082
mmol) and n-Bu4NPF6 (69 mg, 0.18 mmol) were introduced into a
Schlenk tube and dissolved in CD3CN (1 mL), and the resulting mixture
was transferred to the NMR tube. In a separate Schlenk tube, a solution
of FcPF6 (63 mg, 0.19 mmol) in 0.38 mL of CD3CN was prepared.
Aliquots of the FcPF6 solution (50, 50, 20, and 30 µL; total 0.077 mmol)
(b) In Acetonitrile by HBF4/Et2O. This reaction was carried out
as described above in part a, except that HBF4‚Et2O was used in place
of aqueous HBF4. The final result was identical, but the gas evolution
and the conversion to cis-[4]+ were instantaneous. This reaction was
1
1
also monitored by H-NMR in CD3CN under a variety of conditions
were added to the solution of 1. The reaction was monitored by H-
NMR. After each addition, the characteristic blue color of Fc+
immediately disappeared, producing vigorous gas evolution. At this
point, 1H-NMR showed the almost complete disappearance (<5% left)
of the resonance at δ 5.2 (Cp resonance of starting compound 2). The
spectrum of the resulting orange solution showed resonances corre-
sponding to those described above for [4]+ (Vide supra), the ferrocene
resonance at δ 4.1, and an additional resonance at δ 4.5 due to H2.
(b) In the Presence of 2,6-Lutidine. The same experiment as
described above was repeated with the only difference that 2,6-lutidine
(0.072 mmol vs 0.061 mmol of 2 and 0.12 mmol of n-Bu4NPF6) was
also present in the initial solution (1 mL). In this case, no gas evolution
was observed. After the addition of 1.7 oxidizing equiv of Fc+, the
1H-NMR spectrum showed the almost complete replacement of the
resonance at δ 5.2 (due to 2) with a resonance at δ 5.6 (due to [4]+).
(see Results and Discussion).
(c) In Diethyl Ether. Addition of HBF4‚Et2O (48 µL, 0.36 mmol
of HBF4) to a stirring solution of 2 (105 mg, 0.357 mmol) in Et2O (15
mL) that had been precooled to -78 °C resulted in vigorous gas
evolution and formation of a brick-red precipitate. After 1/2 h of stirring
at -78 °C, the solution was decanted and the solid was washed with
ether and dried under vacuum. Yield: 98 mg. The 1H- and 31P-NMR
spectra of this solid indicate the presence of three major products, none
of which correspond to the species generated in situ in acetonitrile. 1H
NMR (δ, CD3CN): 5.45 (s, Cp), 5.31 (d, JHP ) 1.8 Hz, Cp), 5.29 (d,
JHP ) 2.0 Hz, Cp), 1.69 (d, JHP ) 10.2 Hz, PMe3), 1.66 (d, JHP ) 10.2
Hz, PMe3), 1.64 (d, JHP ) 10.2 Hz, PMe3). 31P{1H} NMR (δ,
CD3CN): 42.0, 24.4, 13.6. 19F-NMR (δ, CD3COCD3): 72.8 (w1/2
12 Hz, weak), 73.9 (w1/2 ) 72 Hz, strong). Recrystallization of the
red solid from CH2Cl2/heptane produced X-ray-quality crystals of
[CpMo(CO)3(PMe3)][BF4].
)
1
The amount of residual 2 at this point, by integration of the Cp H-
NMR resonances) is estimated as 10 ( 5% of the initial amount. The
growth of an additional broad resonance was also observed, whose
chemical shift changed as a function of the amount of added Fc+
solution: δ 3.65, 4.45, 5.10, and 5.60 after the addition of 0.40, 0.80,
1.25, and 1.70 mol of Fc+/mol of 2, respectively.
A brick-red precipitate with similar spectroscopic properties was also
obtained by carrying out a similar procedure in heptane solvent.
Protonation of 1 with HBF4‚Et2O in n-Heptane. By a procedure
identical with that described above for compound 2, treatment of 1
with the stoichiometric amount of HBF4‚Et2O in n-heptane at -78 °C
afforded a brick-red precipitate, which was not further investigated.
1H-NMR Study of the Protonation of H2O in CD3CN. A 0.46
mL sample of CD3CN was introduced in the NMR tube, and water (19
µL, 1.05 mmol) was added. To this solution was added HBF4‚OEt2 in
small aliquots via a microsirynge, and the protonation process was
(c) In the Presence of Water: Slow Addition. A stock solution
was prepared in a Schlenk tube by dissolving compound 2 (31 mg,
0.10 mmol) in CD3CN (3.8 mL). A 0.4 mL aliquot (0.011 mol of 2)
was transferred to an NMR tube, and water (1.6 µL, 0.077 mmol) was
added. In a separate NMR tube, a solution of FcPF6 (4.2 mg, 0.013
mmol) in 0.12 mL of CD3CN was prepared. Aliquots of the FcPF6
solution (24, 25, 25, 25, and 6 µL; total 0.0114 mmol) were added to
the solution of 1 at 20 min intervals, and the reaction was monitored
by 1H-NMR on aliquots of the solution. After each addition, the
characteristic blue color of Fc+ immediately disappeared and no gas
evolution was observed. After the addition of 0.0114 mmol of Fc+
(1.7 oxidizing equiv), the resonance at δ 5.2 of 2 had been completely
replaced by the resonance at δ 5.6 of [4]+.
1
monitored by H-NMR.
Results and Discussion
(a) Electrochemical Studies. The cyclic voltammetric
behavior of CpMoH(CO)3 and 1 has already been reported in
the literature, both compounds being oxidized irreversibly in

Protonation of Transition Metal Hydrides
Inorganic Chemistry, Vol. 35, No. 18, 1996 5157
MeCN. The peak potential for the anodic oxidation wave of
CpMoH(CO)3 has been reported by two different groups as
+0.05 V6 and +0.800 V4 vs Fc/Fc+. Our own measurements
of this process seem closer to the pattern reported by Tilset and
complete when the average current during the last data storage
time interval dropped to less than 1% of the first time interval
average current (<1/2 h). Integration of the current vs time curve
gave a consumption of 1.9 ( 0.2 faradays/mol. At the end of
the coulometric experiment, a Karl Fischer titration of the
solution water content revealed [H2O] ) 47 mM, which is a
large excess with respect to the molybdenum hydride complex
([2] ) ca. 4.2 mM), but still a small amount with respect to the
concentration of supporting electrolyte (>0.1 M). The same
experiment carried out on a known amount of ferrocene under
identical experimental conditions gave a consumption of 1.03
( 0.03 faradays/mol. The greater uncertainty on the coulometric
result for the oxidation of 2 is related to the greater difficulty
in weighing the air-sensitive hydride complex. An analogous
coulometric experiment on 1 under identical conditions was
reported by Smith and Tilset to afford a similar result (1.94 (
0.08 faradays/mol of charge consumption).7
co-workers,4 with an irreversible oxidation wave with Ep,a
)
0.54 V and a second irreversible oxidation wave (assigned4 to
the oxidation of the [CpMo(CO)3(MeCN)]+ product), with Ep,a
) 1.11 V. However, we could not obtain reproducible results
from the coulometric experiments. We attribute this behavior
to the known process13 of reductive elimination of CpH from
CpMoH(CO)3 in MeCN to produce Mo(CO)3(MeCN)3 and CpH
and to the subsequent oxidation of these and possibly also other
intermediate products. The half-life of CpMoH(CO)3 in MeCN
at 37.5 °C is reported as less than 30 min.13 Although this
decomposition process is not necessarily the reason for the
discrepancy between the two earlier reports,4,6 it was a sufficient
deterrent for us to continue the investigation of the oxidative
process of CpMoH(CO)3 in acetonitrile and we subsequently
focused uniquely on the phosphine-substituted materials.
Control experiments revealed that both 1 and 2 are stable in
MeCN with no observable decomposition within 3 days at room
temperature. The irreversible oxidation of compound 1 on the
Pt electrode has an anodic peak potential of +0.23 V by direct
calibration against the oxidation of ferrocene, in good agreement
with the potential of +0.26 V reported earlier and calculated
on the basis of the measured potential of +0.35 V with respect
to the Ag/Ag+ reference electrode.7 The decrease in potential
upon going from CpMoH(CO)3 to 1 is as expected, given the
greater donor and/or lower acceptor ability of the phosphine
ligand with respect to CO. The cyclic voltammogram of 2 on
Pt also shows an irreversible oxidation, even at -20 °C, the
anodic peak potential being found at +0.19 V with respect to
ferrocene. The first oxidation wave of 2 is followed, at more
positive potentials in the anodic scan, by a second irreversible
oxidation wave with Ep,a ) 0.62 V, which is assigned to the
oxidation of [CpMo(CO)2(PMe3)(MeCN)]+ ([4]+) produced by
the first oxidation process (Vide infra) and compares with the
analogous process observed at Ep,a ) 0.72 V for the cyclic
voltammogram of 1.7 Therefore, the better donor ligand PMe3
further increases the susceptibility of the system toward oxida-
tion. Similar trends have been observed before for other
systems, for instance CpRuHLL′ (L, L′ ) CO, tertiary
phosphine)14-17 and Cp* analogues18 and for a variety of
octahedral d6 dinitrogen and η2-dihydrogen complexes.19
Constant-current bulk oxidation in THF at -80 °C in an EPR
spectroelectrochemical cell did not lead to the detection of any
transient EPR-active species for either 1 or 2. Given our
instrument sensitivity and our experimental conditions (see
Experimental Section), assuming that the one-electron oxidation
products [1]+ and [2]+ are produced at the electrode, an upper
limit of 10-2 s is estimated for their half-life at -80 °C.
A bulk electrochemical oxidation of 2 in MeCN was carried
out at a Pt anode at a constant potential of +0.11 V with respect
to the anodic peak potential of the hydride complex (+0.30 V
with respect to Fc/Fc+). The electrolysis was considered
(b) Chemical Oxidation Studies of CpMoH(CO)2(PPh3)
in the Presence and Absence of Water and n-Bu4NPF6. As
mentioned in the Introduction, the oxidation of 1 by the
ferrocenium ion had been previously investigated only in the
presence of a strong base (2,6-lutidine), which acts as the proton
acceptor and forces the system to require 2 mol of oxidizing
agent/mol of 1 according to the stoichiometry of eq 6.7 Two
faradays per mole was also required for the electrochemical
oxidation in MeCN at a Pt electrode, and this was attributed to
a proton transfer to the medium. We have repeated the
ferrocenium oxidation experiment without strong base but with
the deliberate addition of n-Bu4NPF6 and water, conditions that
are meant to mimic those of the previously reported7 electro-
chemical oxidation experiment. The reaction was carried out
directly in MeCN-d3 by stepwise additions of substoichiometric
amounts of a solution of FcPF6 in MeCN-d3 and was monitored
1
by H-NMR spectroscopy. A slow reaction took place after
the addition of each aliquot of FcPF6, as visually indicated by
the slow bleaching (t1/2 ≈ 5 min) of the characteristic deep blue
color of the ferrocenium salt. The slow rate of this electron
transfer process is due to the substantially positive potential
necessary for the oxidation of 1 (+0.23 V with respect to the
reversible wave of the Fc/Fc+ couple). However, the start of
the anodic wave upswing occurs at a potential slightly negative
of the ferrocene standard and the oxidation wave is irreversible
because complex [1]+ decomposes immediately after its genera-
tion (Vide supra); therefore an oxidation reaction eventually does
take place. This experiment showed that less than 20% of
starting material was left in solution after addition of 1.2 mol
of Fc+/mol of 1 (thus the complete oxidation of 1 would require
between 1.4 and 1.6 equiv of Fc+, i.e. significantly less than 2
equiv). The only Mo product observed at the end of the reaction
is the acetonitrile adduct [CpMo(CO)2(PPh3)(MeCN-d3)]+,
[3-d3]+. It therefore seems that, under these conditions, the
reaction has a mixed stoichiometry, part of oxidation proceeding
as in eq 6 and part as in eq 8, implying the involvement of H2
evolution. Because of the exceedingly slow oxidation rate (see
above), a noticeable gas evolution was not observed for this
reaction.20
In conclusion, as for the oxidation of CpWH(CO)2(PMe3)
reported earlier,4 we find that the oxidation stoichiometry for 1
changes from the electrochemical experiment (2 faradays/mol)7
to the chemical experiment. However, our result appears
puzzling because our chemical oxidation was carried out under
“wet” and “ionic strength” conditions that simulated the
(13) Kubas, G. J.; Kiss, G.; Hoff, C. D. Organometallics 1991, 10, 2870-
2876.
(14) Ryan, O. B.; Tilset, M.; Parker, V. D. Organometallics 1991, 10, 298-
304.
(15) Ryan, O. B.; Tilset, M. J. Am. Chem. Soc. 1991, 113, 9554-9561.
(16) Ryan, O. B.; Smith, K.-T.; Tilset, M. J. Organomet. Chem. 1991, 421,
315-322.
(17) Smith, K.-T.; Rømming, C.; Tilset, M. J. Am. Chem. Soc. 1993, 115,
8681-8689.
(18) Jia, G.; Lough, A. J.; Morris, R. H. Organometallics 1992, 11, 161-
171.
(20) Professor M. Tilset has informed us that the consumption of 1 mol of
oxidant/mol of 1 in MeCN was also measured in his laboratory, with
1
(19) Morris, R. H. Inorg. Chem. 1992, 31, 1471-1478.
observation of H2 by H-NMR.

5158 Inorganic Chemistry, Vol. 35, No. 18, 1996
Quadrelli et al.
conditions of the electrochemical experiment. Thus, it may seem
that the basicity of the “medium” is comparable with that of
the neutral hydride 1 under our chemical oxidation conditions,
whereas the “medium” appears more basic under the electro-
chemical conditions previously described.7,21 These apparently
contrasting observations are reconciled and fully rationalized
on the basis of the protonation experiments that will be described
in section d.
(c) Chemical Oxidation Studies of CpMoH(CO)2(PMe3)
in the Presence and Absence of Water and n-Bu4NPF6 or
2,6-Lutidine. The oxidation of 2 by FcPF6 proceeds at a faster
rate (essentially instantaneous upon mixing the reagents) with
respect to the oxidation of the corresponding PPh3 complex, in
at δ 4.5 is also observed in the 1H-NMR spectrum and is
assigned to H2, on the basis of a control 1H-NMR spectrum of
a CD3CN solution recorded before and after an H2 purge.
When the oxidation of 2 with Fc+ in CD3CN is carried out
in the presence of a slight excess of 2,6-lutidine, however, there
is no gas evolution, and a stoichiometry of 2 mol of Fc+/mol
of 2 is indicated by the 1H-NMR experiment. A broad
resonance also builds up as the reaction progresses, this being
attributed to the 2,6-lutidinium acidic proton. Its position (see
Experimental Section) changes with amount of lutidinium
formed, presumably because of rapid self-exchange with the
excess of unprotonated 2,6-lutidine.
1
The oxidation-titration with H-NMR monitoring was also
agreement with the measured lower potential for the oxidation
carried out in the presence of a large excess (ca. 10-fold) of
water, the stepwise addition of aliquots of oxidant being carried
out over a time span of 1.5 h. Gas evolution after the addition
of each aliquot of ferrocenium solution was hardly noticeable
in this case, and ca. 1.7 mol of Fc+/mol of 2 was required to
consume 2 completely. Finally, 1.9 mol of the ferrocenium
oxidant was added at one time to 1 mol of 2 and in the presence
of 10 mol of water, resulting in the complete consumption of
the oxidant and no noticeable gas evolution.
-
of this compound. The oxidation product, the salt [4]+PF6
,
was isolated and characterized by H- and 31H-NMR and by
elemental analysis. By analogy with the similar, previously
reported [3]+ and [CpW(CO)2(PMe3)(MeCN)]+ complexes,4,7
complex [4]+ exists in solution as an equilibrium mixture of
cis and trans isomers, with the cis isomer being the prevalent
one. The oxidation of 2 leads directly to the equilibrium mixture
of cis- and trans-[4]+, while generation of this system by hydride
1
-
abstraction from 2 (reaction with Ph3C+BF4 in MeCN) leads
In conclusion, the electrolyte n-Bu4PF6 is shown to have no
effect in the chemical oxidation of 2 in MeCN, whereas water
has a marked effect, but the number of oxidizing equivalents
needed to consume the starting material depends not only on
the presence or absence of water but also on the rate at which
the oxidant is added to the hydride solution. For the stronger
base lutidine, however, 2 mol of oxidant/mol of hydride is
needed even under the same slow-addition conditions where
the presence of water requires a smaller amount of oxidizing
agent. Since the problem appears to lie in the competition
between hydride 1 or 2 and water as bases, we next sought a
deeper insight into these proton transfer processes by studying
the protonation of the 1/H2O/CD3CN and 2/H2O/CD3CN
systems by the strong acid HBF4‚Et2O (see section d).
Before we proceed to illustrate and discuss the results of the
protonation experiments, however, it is interesting to consider
the susceptibility of the oxidized hydride complexes [1]+ and
[2]+ to deprotonation. It has been shown that the pKa of
transition metal hydrides decreases by ca. 20 units upon
oxidation, and the pKa of [CpMoH(CO)3]+ has been estimated
as ca. -6.0 in MeCN.4 The pKa of [2]+ has not been reported
to the best of our knowledge, but it has been shown that a pKa
increase of 8.1 units occurs on going from [CpWH(CO)3]+ to
[CpWH(CO)2(PMe3)]+.4 Assuming no dependence of this
difference on the nature of the metal, a pKa value of ca. +2 is
estimated for [2]+. The pKa of [1]+ is probably lower, given
the lower donor/higher acceptor strength of PPh3 with respect
to PMe3. These values are slightly lower than the pKa reported
for the hydrated proton, H3O+, in MeCN (2.2-2.5),25 i.e., water
has sufficient basic strength to deprotonate [1]+ and [2]+. In
addition, the high formation constants of [H(H2O)n]+ (n ) 2-4;
log Kf ) 3.9-5.3)25 effectively make water an even stronger
base.
selectively to the trans isomer as the kinetically controlled
product. A kinetic study of this hydride abstraction reaction
was recently reported.22 Once formed, the trans complex slowly
isomerizes (t1/2 ≈ 2 days) to the thermodynamically preferred
cis isomer. This behavior is identical to that previously reported
for [3]+ and for [CpW(CO)2(PMe3)(MeCN)]+, and a rational-
ization for such behavior has previously been given.4,7 Moni-
toring of a CD3CN solution of [4]+ at room temperature shows
a smooth MeCN exchange reaction (t1/2 ≈ 2.5 h) forming [4-d3]+
and free MeCN.
The stepwise oxidation of 2 was carried out and monitored
1
by H-NMR spectroscopy in MeCN-d3 as described above for
the PPh3 analogue. In dry CD3CN, each subsequent addition
of the oxidizing agent produced an immediate quenching of the
blue ferrocenium color and caused a Vigorous gas eVolution,
up to the addition of 1 mol of oxidant/mol of 2. Correspond-
ingly, the resonances of the starting material decreased while
the resonances of cis-[4]+ and ferrocene increased. After the
addition of 1 mol/mol of oxidant, the starting material was
completely consumed. This shows that 2 acts as the proton
acceptor for the intermediate, proton donor [2]+, in the dry
CD3CN solvent (eq 8). The consumption of only 1 equiv of
oxidant has precedent in the literature for the ferrocenium
oxidation of CpWH(CO)2(PMe3),4 CpRuH(PPh3)2,16 and
Cp*RuH3(PPh3)23 and for the acetylferrocenium oxidation of
OsH6(PPri3)2.24 The 1:1 stoichiometry does not change when
the experiment is repeated in the presence of n-Bu4NPF6 but
without the deliberate addition of water. In this experiment,
ca. 2 mol of the salt/mol of the hydride was used; thus the
amount of any adventitious water introduced by the salt must
be negligibly small (see section a above). A sharp resonance
(21) A more recent remeasurement of the coulometric oxidation of 1 in
Professor Tilset’s laboratory gives 1.67 ( 0.09 faradays/mol (average
of five experiments). The electrolyte was dried by passing it through
active alumina, i.e. the same treatment described in the previous work
where the consumption of 1.94 ( 0.08 faradays/mol was reported.7
Also, under the same conditions, a coulometric measurement of the
oxidation of 2 gives the consumption of 1.25 ( 0.05 faradays/mol
(average of four experiments). We are grateful to Professor Tilset for
sharing these results with us.
(d) Protonation Studies of CpMoH(CO)2L (L ) PPh3,
PMe3). When either 1 or 2 was protonated in dry MeCN with
the stoichiometric amount of HBF4‚Et2O, immediate and vigor-
ous gas evolution was observed and the color of the solution
changed distinctly from yellow to orange. The only reaction
1
product observed by IR and H-NMR spectroscopies was [3]+
or [4]+, respectively. This result confirms that a proton transfer
from the strong acid [1]+ to 1 (or [2]+ to 2) would result in H2
(22) Cheng, T.-Y.; Bullock, R. M. Organometallics 1995, 14, 4031-4033.
(23) Zlota, A. A.; Tilset, M.; Caulton, K. G. Inorg. Chem. 1993, 32, 3816-
3821.
(24) Smith, K.-T.; Tilset, M.; Kuhlman, R.; Caulton, K. G. J. Am. Chem.
Soc. 1995, 117, 9473-9480.
(25) Izutsu, K. Acid-Base Dissociation Constants in Dipolar Aprotic
SolVents; Blackwell Scientific: Oxford, U.K., 1990.

Protonation of Transition Metal Hydrides
Inorganic Chemistry, Vol. 35, No. 18, 1996 5159
evolution, as observed in the oxidation experiment described
above. Thus, the product of protonation, [CpMoH2(CO)2L]+
(L ) PPh3, PMe3), is unstable in MeCN and decomposes
immediately via replacement of H2 by a solvent molecule. This
behavior prevents us from obtaining quantitative information
on the thermodynamic basicity of 1 and 2 in MeCN relative to
that of water. We have attempted to obtain such information
in other, less coordinating solvents. Jia and Morris have shown
that the thermodynamics of protonation of the compounds
Cp*RuH(L-L) (L-L ) chelating diphosphine) can conveniently
be studied in CH2Cl2 or THF, whereas H2 evolution occurs in
MeCN.26 However, we find that treatment of 1 or 2 with
HBF4‚Et2O in either ether, toluene, or even heptane at -78 °C
always results in vigorous gas evolution and formation of a red
precipitate. For the reaction of 2, the red solid has been
identified as a mixture of [CpMo(CO)3(PMe3)][BF4], CpMo-
(CO)2(PMe3)FBF3, and other unknown species. The tricarbonyl
complex was isolated in the form of single crystals from CH2Cl2/
heptane and was characterized by an X-ray analysis.27 The
presence of CpMo(CO)2(PMe3)FBF3 is suggested by the pres-
ence of two 19F-NMR resonances at δ -72.8 and -73.9 in
acetone-d6 (the resonance of free BF4- is observed at δ -74.5
under the same conditions). Similar compounds, e.g. CpMo-
(CO)2(PR3)FBF3 (R ) Ph, OPh, Et), were described by Beck
et al. as the products of the reaction between CpMoH(CO)2-
(PR3) and Ph3C+BF4-, and their thermal decomposition to
[CpMo(CO)3(PR3)]+ was described.28 It is to be remarked that,
mechanistically, the interactions of H+ and Ph3C+ with the
hydride complex are expected to be identical, involving attack
of the electrophilic reagent on the hydride ligand, formation of
a presumed σ-complex intermediate, and elimination of the
formal product of hydride abstraction, H2 in one case and Ph3CH
in the other. In conclusion, complexes [CpMoH2(CO)2(PR3)]+
are too unstable and lose H2 even upon the action of a
nucleophile as weak as the BF4- counterion in noncoordinating
solvents.
Scheme 1
to analogous complexes with stronger donor ligands, and
cationic complexes have a greater tendency to adopt nonclassical
structure with respect to neutral ones.33 It seems likely,
therefore, that the [CpMoH2(CO)2L]+ (L ) PPh3, PMe3)
complexes adopt a nonclassical structure which could be
unstable toward H2 substitution in the strongly coordinating
MeCN solvent. Molybdenum complexes of this stoichiometry
have not been isolated, to the best of our knowledge, whereas
the tungsten analogue [CpWH2(CO)2(PMe3)]+ has been de-
scribed and shown to have a classical structure.4 The stability
and structure of this complex are consistent with the greater π
basicity of tungsten with respect to molybdenum.
Coming back to the protonation studies in MeCN, when a
large (25-fold) excess of water was introduced, protonation of
the 1/H2O mixture resulted in no gas eVolution and no change
of color, whereas protonation of a 2/H2O mixture did produce
some efferVescence, but at a much less Vigorous rate. Cor-
respondingly, a dampened gas evolution resulted from the
protonation of 2 in MeCN by aqueous HBF4.
For the protonation of 1/H2O by HBF4‚Et2O, continued
addition of small amounts of the acid to the solution of 1/H2O
still failed to induce gas evolution initially, but the subsequent
addition of a large amount of acid finally resulted in vigorous
gas evolution and a color change from yellow to orange.
Repeated experiments with different excess amounts of water
showed that the amount of acid necessary to produce effer-
vescence increased with the amount of water present in solution.
Thus, only when water is fully protonated, does further addition
of acid result in the direct protonation of 1.
These results suggested to us a possible solution to the
puzzling results of the chemical oxidations discussed in sections
b and c. Water indeed competes effectively as a base with either
1 or 2 and will be protonated preferentially under the conditions
utilized in our chemical and electrochemical experiments, but
the basicity of 1 and 2 (especially the latter) is sufficiently high
to allow an equilibrium formation of [CpMoH2(CO)2L]+. Since
formation of the latter species is followed by the irreVersible
loss of H2, the reaction ultimately proceeds, under thermody-
namic control, to the products that would be obtained by direct
protonation of 1 or 2 (see Scheme 1). We can describe the
function of water in this system as a “proton shuttle” between
the strong acid (HBF4‚Et2O in the protonation studies; [1]+ or
[2]+ in the oxidation studies) and the hydride complex 1 or 2.
To seek further evidence in support of this proposal, we
monitored by 1H-NMR the protonation of both 1/H2O and 2/H2O
by HBF4‚Et2O in CD3CN. The results of the latter experiment
are shown in Figure 1. The first spectrum (before addition of
the acid) shows the resonance of water at ca. δ 2.15 and the
relatively broad Cp resonance of 2 at ca. δ 5.2. The weak and
broad resonance at ca. δ 3.6 is due to a THF impurity that was
present in the starting complex 2. Following the addition of
the stoichiometric amount (with respect to 2) of HBF4‚Et2O,
we observe that, after 2 min, the water resonance is shifted to
ca. δ 3.6, whereas the resonance of 2 is essentially unchanged.
This demonstrates that water, not 2, has been protonated. If 2
The instability of [CpMoH2(CO)2L]+ contrasts with the
stability of the corresponding tungsten-PMe3 derivative, ob-
tained as a coproduct from the corresponding oxidation of
CpWH(CO)2(PMe3) according to the stoichiometry of eq 5.4
The stability of dihydride complexes versus reductive elimina-
tion and substitution of H2 by MeCN recently attracted
considerable attention. The susceptibility to this reaction is
greatly increased by protonation; for instance, treatment of
MoH4(PMe2Ph)4, ReH7(PPh3)2, OsH4(PMe2Ph)3, Cp*RuH3-
(PPh3), and OsH6(PPri3)2 with strong acids in MeCN affords
[MoH2(PMe2Ph)4(MeCN)2]2+ 29
,
[ReH(PPh3)2(MeCN)4]2+ 30
,
[Os(PMe2Ph)3(MeCN)3]2+ 31 [Cp*Ru(PPh3)(MeCN)2]+,23 and
,
[OsH3(MeCN)2(PPri3)2]+,24 respectively. It is has been shown
that the process of dihydrogen reductive elimination from a
classical dihydride complex proceeds through an intermediate
nonclassical dihydrogen complex32 and that the relative stability
of classical vs nonclassical dihydrides depends on the metal
electron density. Complexes with stronger π-acidic ligands have
a greater tendency to adopt nonclassical structures with respect
(26) Jia, G.; Morris, R. H. J. Am. Chem. Soc. 1991, 113, 875-883.
(27) Details of the X-ray structure of [CpMo(CO)3(PMe3)][BF4] will be
published elsewhere.
(28) Su¨nkel, K.; Ernst, H.; Beck, W. Z. Naturforsch. 1981, 36B, 474-
481.
(29) Rhodes, L. F.; Zubkowski, J. D.; Folting, K.; Huffman, J. C.; Caulton,
K. G. Inorg. Chem. 1982, 21, 4185-4192.
(30) Allison, J. D.; Walton, R. A. J. Chem. Soc., Chem. Commun. 1983,
401-403.
(31) Bruno, J. W.; Huffman, J. C.; Caulton, K. G. J. Am. Chem. Soc. 1984,
106, 1663-1669.
(32) Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32, 789-805.
(33) Lin, Z. Y.; Hall, M. B. Coord. Chem. ReV. 1994, 135, 845-879.

5160 Inorganic Chemistry, Vol. 35, No. 18, 1996
Quadrelli et al.
Scheme 2
Figure 1. 1H-NMR monitoring of the reaction of HBF4‚Et2O (12 µmol)
with compound 2 (13 µmol) and H2O (111 µmol) in CD3CN (0.48
mL) at room temperature. The first spectrum was taken before the
addition of HBF4‚Et2O. Subsequent spectra were taken after 2, 4, 8,
13, 18, 23, 31, 38, 45, and 84 min.
were protonated, the resonance of 2 would disappear im-
mediately, as demonstrated by the protonation experiment in
dry MeCN discussed above. The shift of the water resonance
is due to the rapid equilibrium, on the NMR time scale, between
the protonated water, [H(H2O)n]+, and residual H2O (proton
exchange reactions can approach the diffusion limit of reaction
rates). The spectrum also shows the quartet resonance due to
the methylene protons of Et2O at δ 3.4. As the reaction
progresses, the following features are observed: (a) the
[H(H2O)n]+/H2O resonance shifts progressively back toward the
position of unprotonated H2O; (b) the Cp resonance of 2 at δ
5.2 slowly decreases in intensity; (c) the Cp resonance of [4]+
at ca. δ 5.6 correspondingly increases in intensity (the sum of
the intensities for the Cp resonances of 2 and [4]+ is constant);
(d) a resonance at δ 4.5 due to dissolved H2 progressively
increases. The escape of H2 into the head space after the initial
phases of the reaction is the probable cause of the reduced rate
of increase of the H2 resonance with respect to the resonances
of the other products after ca. 30 min. The variations of other
resonances not shown in Figure 1 (e.g. the hydride resonance
of 2) are consistent with the above described trends. It is self-
evident that all the above observations are consistent with, and
further support, our proposed Scheme 1.
Given this proposed mechanism, the rate of disappearance
of 2 should follow the rate law rate ) kobs[2][H(H2O)n]+, and
from the initial rate of disappearance of 2 under the conditions
utilized (see Figure 1), we calculate kobs ) 0.02 s-1 M-1. The
same basic features shown in Figure 1 are also observed for
the reaction of HBF4‚Et2O with a 1/H2O mixture in CD3CN.
The main difference between the two reactions is that the decay
of 1, formation of [3]+, formation of H2, and shift of the water
resonance are ca. 15 times slower than those for the corre-
sponding reaction involving 2. This phenomenon can be
attributed to the expected lower basicity of 1 relative to 2, again
consistent with our proposed mechanism of Scheme 1. Inde-
pendent studies on the nature of [H(H2O)n]+ will be illustrated
later in section h.
variables were kept identical in the two experiments, except
the thickness of the monolayer.
Whereas less than 30 min was usually sufficient to complete
the electrolysis on the bare Pt electrode surface, ca. 1 h was
required on the thiol (n ) 2) coated electrode, and the process
was not yet complete after 3 h of operation on the thiol (n ) 4)
coated electrode under otherwise identical experimental condi-
tions. The measured charge consumptions were 1.7 ( 0.2 and
1.3 ( 0.2 farday/mol for the electrode covered with the n ) 2
and n ) 4 thiol, respectively (cf. with 1.9 ( 0.2 faradays/mol
on the bare Pt electrode). When the experiment on the n ) 4
thiol-coated Au electrode was interrupted (after 3 h of elec-
trolysis), the resulting solution still had ca. 20% of unreacted 2
(by cyclic voltammetry). The accumulation of a gas bubble
next to the septum which separates the two electrolytic
compartments was noticed during the slow electrolysis on the
S(CH2)4OH-coated Au electrode, which is believed to arise from
H2 evolved in the bulk. The results observed for these
coulometric experiments are in perfect agreement with the
proposed proton shuttle process of Scheme 1: the stoichiometry
(amount of H2 that evolves and number of faradays per mole
of M-H) depends on the competition between the rates of
consumption of the hydride material by the oxidation at the
electrode and by the acidolysis in the bulk. As a matter of fact,
the terminal OH groups of the ω-hydroxy thiols should in
principle be more basic than water and could capture the protons
of [1]+ and [2]+. If this happens to a significant extent,
however, the results of the experiments show that these basic
centers can function as proton shuttles as water does.
(f) Mechanistic Considerations. The results of our oxidation
studies are summarized and interpreted by the mechanism in
Scheme 2, which is based on that proposed previously by Ryan,
Tilset, and Parker,4 with the addition of the proton shuttle action
of an external base. Following the oxidation of 1 or 2 by
-
Fc+PF6 in dry MeCN, the strongest available bases to
deprotonate [1]+ or [2]+ are the starting materials themselves,
ultimately affording [3]+ or [4]+ with evolution of H2 according
to the stoichiometry of eq 8. The proton transfer step affords
the 17-electron CpMo(CO)2L• and the 18-electron [CpMoH2-
(CO)2L]+ (L ) PPh3, PMe3) intermediates, neither of which is
stable. According to the previously established mechanism,4
the 17-electron species coordinates MeCN and is further
oxidized to the observed product (eqs 3 and 4), while the 18-
electron dihydride complex evolves dihydrogen and coordinates
MeCN to afford more of the observed [3]+ or [4]+ final product
(eq 7).
(e) Coulometric Oxidation of 2 at Thiol-Covered Au
Anodes. Compound 2 was also electrolyzed at thiol-coated Au
electrodes, with two different thicknesses for the self-assembled
monolayer of S(CH2)nOH groups (e.g. n ) 2 and 4). The end-
chain hydroxyl groups help establish a compact and crystalline
monolayer. The effect of the electrode coating is to slow the
kinetics of electron transfer, its rate dramatically decreasing with
the number of methylene groups in the linear alkyl chain.11
Consequently, the time necessary to carry out the electrolysis
becomes longer. The rate of oxidation depends on several
factors besides the thickness of the insulating hydroxy thiol
monolayer, such as the electrode overpotential, the solution
resistance, the rate of stirring, the electrode surface area, the
concentration of residual, unoxidized hydride material, etc. All
For compound 1 or 2, we have clearly demonstrated that
-
neither MeCN nor the PF6 anion of n-Bu4PF6 competes with
2 as a base; thus the basicity of the “medium” in changing the
oxidation stoichiometry can be solely attributed to the presence

Protonation of Transition Metal Hydrides
Inorganic Chemistry, Vol. 35, No. 18, 1996 5161
of water. Water is always introduced with the supporting
electrolyte, and its concentration may be relatively large with
respect to the analyte, given the large excess of electrolyte used
in the typical experiments (see section a above). A reported
method for the dehydration of electrolytic solutions is the
passage through a column of activated neutral alumina,7 but
the results reported indicate that a certain amount of water must
still be present.21 From Scheme 1, the rate of formation of the
product ([3]+ or [4]+) from 1 or 2 and the hydrated proton
should be given by kobs[CpMoH(CO)2L][H(H2O)n]+, with the
second-order observed rate constant kobs equal to k1k2/(k-1[H2O]n
+ k2). This expression simplifies in the following two limiting
cases: (a) rapid proton transfer pre-equilibrium followed by rate-
determining H2 elimination (k-1[H2O]n . k2), kobs ) K1k2/
[H2O]n; (b) rate-determining proton transfer (k-1[H2O]n , k2),
kobs ) k1. A complete analysis of this rate law would require
the determination of kobs at different [H2O], which is beyond
the scope of this study. On the basis of our inability to produce
[CpMoH2(CO)2L]+ (L ) PMe3, PPh3) even in inert solvents at
low temperatures (see Results and Discussions), we know that
the rate of H2 elimination (k2) must be very high, thus suggesting
that the process may be limited by the rate of proton transfer
for compounds 1 and 2. It is pertinent to observe here that any
base can operate as a proton shuttle, provided the basicity of
the transition metal hydride is lower than, but not too distant
from, that of the base and provided the [MH2]+ product
irreversibly decomposes. Thus, 2,6-lutidine does not operate
as a proton shuttle for the oxidation of 1 or 2 but may do so,
for instance, for the oxidation of a more basic hydride.
the bulk. The consumption of 1.94 ( 0.08 faradays/mol for
the oxidation of 1,7 on the other hand, is consistent with our
findings of a very slow proton shuttle process in this case (see
Results and Discussion). A slow proton shuttle would also be
expected for the less basic complexes CpMH(CO)3 (M ) Cr,
Mo, W), consistent with the consumption of 2.1 ( 0.1, 1.9 (
0.1, and 2.2 ( 0.3 faradays/mol, respectively, in those cases.4
CpWH(CO)2(PMe3) should be more basic than 2, as shown by
the higher stability of its protonation product (the stoichiometry
of the ferrocenium oxidation is that of eq 5). The time required
to complete the electrolysis was not specified for this experi-
ment.4 However, the result of 1.9 ( 0.1 faradays/mol indicates
that this compound is still not sufficiently basic to engage in a
significant proton transfer with [H(H2O)n]+ within the time scale
of the electrolytic experiment.
The qualitative order of relative basicity in MeCN for a
selected class of metal hydrides appears to be, on the basis of
the collective results reviewed above, as follows: Cp*RuH3-
(PPh3), CpRuH(PPh3)2 > CpRuH(CO)(PMe3) > CpRuH(CO)-
(PPh3) > CpWH(CO)2(PMe3) > CpMoH(CO)2(PMe3) >
CpMoH(CO)2(PPh3) > CpMH(CO)3 (M ) Cr, Mo, W). We
should emphasize again that the thermodynamic basicity of these
hydrides cannot be determined as easily as their acidity34-39 or
as the basicity of nonhydridic compounds40 because many of
the conjugate acids spontaneously decompose by H2 elimina-
tion,18 even in noncoordinating solvents. A 1H-NMR study of
the kinetics of the proton shuttle process utilizing a suitable
standard base should provide a convenient qualitative assessment
of the relative basicity of hydride compounds.
(h) Protonation of Water in MeCN and Nature of
[H(H2O)n]+. As illustrated in section d, no gas evolution occurs
when a 1/H2O mixture is treated with small amounts (relative
to H2O) of HBF4‚Et2O. The number of moles of acid that is
necessary to start inducing vigorous gas evolution (diagnostic
of the protonation of 1) is much less than the number of moles
of H2O. This is also indicated by an additional protonation
experiment which was carried out on a solution containing
equimolar amounts of 1 and H2O: the introduction of 1 mol of
HBF4‚Et2O/mol of 1 caused an immediate, vigorous, and
copious evolution of gas, qualitatively indistinguishable from
the analogous experiment in dry solvent. It seems, therefore,
that each H+ ion introduced into the solution engages in an
interaction with more than one molecule of water to form a
[H(H2O)n]+ ion with n > 1. Indeed, the values of Kf for
[H(H2O)n]+ for n ) 1-4 in MeCN (1.6 × 102, 8 × 103, 6 ×
104, and 2 × 105, respectively),41 obtained by spectrophotometric
methods, show that the formation of protonated water clusters
is favorable in the acetonitrile solvent.
(g) Survey of Previously Reported Oxidations of M-H
Complexes. The generic scheme of a proton shuttle during the
chemical or electrochemical oxidation of a metal hydride that
we have demonstrated for the oxidation of 1 and 2 (Scheme 1)
brings clarity to a variety of apparently contrasting results
reported in the literature.
For what concerns chemical oxidation processes, the situation
had never been confused because the reported processes either
were carried out in dry MeCN in the absence of an external
base, always requiring 1 equiv of oxidant (ferrocenium oxidation
of CpWH(CO)2(PMe3),4 CpRuH(PPh3)2,16 and Cp*RuH3-
(PPh3)23 and acetylferrocenium oxidation of OsH6(PPri3)224), or
were carried out in the presence of a strong external base with
which the starting hydride complex could not compete, always
requiring 2 oxidizing equiv (ferrocenium/2,6-lutidine oxidation
of CpWH(CO)2(PMe3),4 1,7 and CpRuH(CO)(PPh3),14 and
ferrocenium/pyrrolidine oxidation of CpRuH(CO)(PMe3)15).
More interesting situations derive from previously reported bulk
electrolytic experiments. These are all rationalized on basis of
Scheme 1 and the assumption that adventitious water was
present. The reported14 consumption of 1.70 ( 0.11 faradays/
mol for the oxidation of CpRuH(CO)(PPh3) (whereas 1.93 (
0.15 faradays/mol is consumed in the presence of 2,6-lutidine)
can be attributed to a partial delivery of the proton from
[H(H2O)n]+ back to the hydride complex in the bulk within the
time scale of the bulk electrolytic experiment. In agreement
with the expected trends of metal basicity, the oxidation of
CpRuH(CO)(PMe3) consumes 1.0 ( 0.1 faradays/mol (in the
presence of a 10-fold excess of 2,6-lutidine or a 2-fold excess
of pyrrolidine, this changes to 2.0 ( 0.1 faradays/mol)15 and
the oxidation of CpRuH(PPh3)2 also requires only 1 faraday/
mol.16 In fact, given the partial reversibility of the CpRuH-
(PPh3)2 oxidation,16 it is likely that this oxidation process
involves diffusion of the rather stable 17-electron [CpRuH-
(PPh3)2]+ complex away from the electrode and direct proton
delivery and consequent acidolysis of the starting material in
A number of experimental and theoretical studies to gauge
the stability of [H(H2O)n]+ as a function of n have been
presented,42-45 and several crystal structures of isolated
(34) Jordan, R. F.; Norton, J. R. J. Am. Chem. Soc. 1982, 104, 1255-
1263.
(35) Moore, E. J.; Norton, J. R. J. Am. Chem. Soc. 1986, 108, 2257-2263.
(36) Edidin, R. T.; Norton, J. R. J. Am. Chem. Soc. 1987, 109, 3945-
3953.
(37) Kristja´nsdo´ttir, S. S.; Moody, A. E.; Weberg, R. T.; Norton, J. R.
Organometallics 1988, 7, 1983-1987.
(38) Weberg, R. T.; Norton, J. R. J. Am. Chem. Soc. 1990, 112, 1105-
1108.
(39) Kristja´nsdo´ttir, S. S.; Loendorf, A. J.; Norton, J. R. Inorg. Chem. 1991,
30, 4470-4471.
(40) Wang, D.; Angelici, R. J. J. Am. Chem. Soc. 1996, 118, 935-942.
(41) Chantooni, M. K., Jr.; Kolthoff, I. M. J. Am. Chem. Soc. 1970, 92,
2236-2239.
(42) Peeters, D.; Leroy, G. Bull. Soc. Chim. Belg. 1988, 97, 931-940.
(43) Lee, E. P. F.; Dyke, J. M. Mol. Phys. 1991, 73, 375-405.
(44) Yang, X.; Castleman, A. W., Jr. J. Chem. Phys. 1991, 95, 130-134.

5162 Inorganic Chemistry, Vol. 35, No. 18, 1996
Quadrelli et al.
It should be noted here that nitriles are known to undergo
hydrolysis to amides and carboxylic acids under acid catalysis,51
but this process is slow at room temperature.52 Thus, our data
should not be affected by the hydrolysis process, at least at low
acid concentrations. This conclusion was also reached for the
spectrophotometric study of water protonation in MeCN.49
A
slow drift of chemical shift was observed only for the data at
high acid/water ratios (>1:1) over several hours at room
temperature.
The trend of chemical shifts in Figure 2 can be interpreted
on the basis of a number of equilibria (see eq 9) with all possible
values of n. The nature of the acid in a MeCN solution of
Et2OH+ + nH2O a [H(H2O)n]+ + Et2O
(9)
HBF4‚Et2O is likely to be Et2OH+ or, more rigorously,
[H(Et2O)(MeCN)x]+ (in other words, the strongest base among
those available, e.g. Et2O, BF4-, and MeCN, is probably the
-
ether). The alternative formulation as the ion pair Et2OH+BF4
is also unlikely, given the high dielectric constant of the solvent
MeCN. In the first linear region at small H+/H2O ratios, the
observed resonance is the result of the coalesced resonances of
H2O and [H(H2O)n]+ in rapid exchange.
Figure 2. Position of the water resonance in MeCN-d3 as a function
of the equivalents of HBF4‚Et2O added. [H2O] ) 2.2 M (inset: [H2O]
) 1.4 M).
According to the previous spectrophotometric studies,41,49
a
distribution of the different [H(H2O)n]+ species should always
be present under any given set of experimental conditions. For
our experimental conditions ([H2O] ) 2.2 M) and the Kf values
given in the literature,41 the fractions of the different water
clusters should be [H(H2O)+] ) 6.56 × 10-5, [H(H2O)2+] )
7.22 × 10-3, [H(H2O)3+] ) 0.12, [H(H2O)4+] ) 0.87; hence
the average composition of the protonated water cluster should
be [H(H2O)3.84]+, somewhat in disagreement with the value
{[H(H2O)n]+}mXm- acids for various values of n have been
reported, e.g. [H9O4][FeCl4],46 [H9O4]2[TeBr6],47 and [H5O2]2-
SiFe6,48 but the knowledge of the nature of the hydrated proton
in acetonitrile appears to be limited to the above mentioned
1
spectrophotometric studies.41,49 A H-NMR study of H2O in
MeCN have been described,50 but these have not been extended
1
1
1
determined by the H-NMR technique. The H-NMR method
provides a direct measurement of the average composition of
the protonated water clusters, whereas the previously used
spectrophotometric method is indirect and is based on a number
of assumptions. A thorough reinvestigation of the formation
constants of protonated water clusters is beyond the scope of
to the hydronium ions. Since we have observed that the H-
NMR resonance is rather sensitive to the state of protonation
of water (see Figure 1), we have utilized this technique to probe
this protonation process. Figure 2 shows the position of the
1H-NMR resonance as a function of the amount of added
HBF4‚Et2O. One experiment featured the addition of up to ca.
4 mol of HBF4‚Et2O/mol of H2O, whereas a second one (shown
in the inset) was carried out only up to 0.15 mol of HBF4‚Et2O/
mol of H2O with the collection of a much greater number of
data at low HBF4‚Et2O/H2O ratios. The relative amounts of
water and added acid are measured by volume (microliters of
the H2O and HBF4‚Et2O solutions added by microsyringe to
the NMR tube containing MeCN-d3). Figure 2 can be described,
to a first approximation, in terms of an initial linear increase of
the chemical shift, an intermediate region where the increase
of chemical shift deviates from linearity, and ultimately a new
region of linearity with a much smaller slope with respect to
the first linear region. The second experiment (inset of Figure
2) shows more clearly the linear relationship between the
chemical shift and the amount of added acid in the region of
low Et2OH+/H2O ratios. The extrapolation of the two linear
regions affords a breakpoint corresponding to the addition of
0.34 ( 0.01 mol of acid/mol of water. This indicates that, under
the conditions of water concentration utilized (2.2 M), the
aVerage composition of protonated water in MeCN is [H(H2O)3]+.
1
the present work. However, the H-NMR technique appears
to have a potential greater than previously appreciated for the
investigation of proton transfer equilibria and the nature of acids
in a variety of different solvents.
Conclusions
The studies reported here have afforded, as the principal novel
point of knowledge, the definition of the intimate role of water
(and indeed any base) as a proton shuttle for regulating the
stoichiometry of the oxidation, either chemical or electrochemi-
1
cal, of a transition metal hydride complex. A H-NMR study
of the kinetics of the proton shuttle process utilizing a suitable
standard base may provide a convenient assessment of the
relative basicity of hydride compounds. As a bonus, we have
1
discovered that the H-NMR technique can provide additional
insight into the intimate nature of protonated water in aceto-
nitrile.
Acknowledgment. We gratefully acknowledge support of
this work by DOE (Grant No. DEFG0592ER14230). We also
wish to thank Professors Cary J. Miller and M. Tilset for helpful
discussions.
(45) Honma, K.; Sunderlin, L. S.; Armentrout, P. B. J. Chem. Phys. 1993,
99, 1623-1632.
(46) Gustafsson, T. Acta Crystallogr. 1987, C43, 816-819.
(47) Krebs, B.; Bonmann, S.; Erpenstein, K. Z. Naturforsch. 1991, 46B,
919-930.
IC960394O
(48) Mootz, D.; Oellers, E.-J. Z. Anorg. Allg. Chem. 1988, 559, 27-39.
(49) Kolthoff, I. M.; Chantooni, M. K., Jr. J. Am. Chem. Soc. 1968, 90,
3320-3326.
(51) March, J. AdVanced Organic Chemistry, 4th ed.; Wiley Interscience:
New York, 1992; pp 887-888.
(52) Krieble, V. K.; Alfred, P. L. J. Am. Chem. Soc. 1933, 55, 2326-
2331.
(50) Zelano, V.; Mirti, P. Z. Phys. Chem. (Munich) 1983, 138, 31-43.