47-electron organometallic clusters derived by chemical and electrochemical oxidation of trihydrido(alkylidyne)triruthenium and -triosmium clusters. Ligand additivity in metal clusters

Article abstract of DOI:10.1021/om970769t

Cyclic voltammograms for H₃Ru₃(μ₃-CX)(CO)_9-nL_n (X = OMe, SEt, Me, Et, Ph, NMeBz, Br; L = PR₃, AsPh₃, SbPh₃; n = 2, 3), H₃Ru₃(μ₃-COMe)(CO)₆(PPh ₃)₂(L) (L = P(OMe)₃, CNCH₂-Ph), and H₃Os₃(μ₃-COMe)(CO)₇(PPh ₃)₂ each display a quasi-reversible to reversible, one-electron oxidation followed by an irreversible to quasi-reversible, one-electron oxidation. The ligand and substituent effects upon the oxidation potential are analyzed as an example of ligand additivity in a cluster system; the oxidation potential of the alkylidyne cluster core is as sensitive to π-donor substituent effects as are aromatic π complexes such as CpFe-(C₅H₄X) to substituents on the rings, and the dependence of the oxidation potential of the cluster upon ligand substitution on any given Ru atom is 37% of that expected for the oxidation of a monometallic complex. Reactions with 1 equiv of Ag⁺ or ferricenium give the 47-electron cations [H₃Ru₃(CX)(CO)₆L₃]¹⁺ (X = OMe; L₃ = (PPh₃)₂;L? L? = CO, PPh₃, P(OMe)₃, CNCH₂Ph; X = Ph, NMeBz, SEt, L = PPh₃), characterized by EPR spectroscopy; the equilibrium between isomers H₃Ru₃(COMe)(CO)₆(ax-PPh₃) ₂(ax- and eq-PPh₃)^0/1+ (ax = axial coordination, eq = equatorial coordination) favors the axial coordination for the 48-electron cluster (eq/ax = 0.15) and equatorial coordination for the 47-electron species (eq/ax = 6.4). The rate constant for ax-eq isomerization for the 47-electron species (6.5 s^-1) is 4 orders of magnitude greater than that for the 48-electron species (2 × 10^-4 s^-1). The 47-electron clusters have lifetimes which are correlated with the oxidation potentials of the precursors, ranging from seconds to many hours at room temperature. Slow decomposition of [H₃Ru₃-(COMe)(CO)₆L₃]¹⁺ in the absence of added CO forms the new 46-electron clusters [H₃Ru₃-(CO)₇L₃]¹⁺ (L = PPh₃, AsPh₃). [H₃Ru₃(COMe)(CO)₆(PPh₃) ₃]¹⁺ does not react with CCl₄ but does react with Lewis bases such as CO and acetonitrile. Disproportionation occurs with acetonitrile, chloride, and pyridine; the CO products are [HRu₃(CO)₉(PPh₃)₃]¹⁺, [MePPh₃]¹⁺, and H₃Ru₃(COMe)(CO)_9-n(PPH₃)_n (n = 1, 2).

Full text of DOI:10.1021/om970769t

872
Organometallics 1998, 17, 872-886
47-Electr on Or ga n om eta llic Clu ster s Der ived by
Ch em ica l a n d Electr och em ica l Oxid a tion of
Tr ih yd r id o(a lk ylid yn e)tr ir u th en iu m a n d -tr iosm iu m
Clu ster s. Liga n d Ad d itivity in Meta l Clu ster s
William G. Feighery,† Huirong Yao, Anthony F. Hollenkamp,‡
Robert D. Allendoerfer, and J erome B. Keister*
Department of Chemistry, University at Buffalo, State University of New York,
Buffalo, New York 14260-3000
Received August 29, 1997
Cyclic voltammograms for H3Ru3(µ3-CX)(CO)9-nLn (X ) OMe, SEt, Me, Et, Ph, NMeBz,
Br; L ) PR3, AsPh3, SbPh3; n ) 2, 3), H3Ru3(µ3-COMe)(CO)6(PPh3)2(L) (L ) P(OMe)3, CNCH2-
Ph), and H3Os3(µ3-COMe)(CO)7(PPh3)2 each display a quasi-reversible to reversible, one-
electron oxidation followed by an irreversible to quasi-reversible, one-electron oxidation. The
ligand and substituent effects upon the oxidation potential are analyzed as an example of
ligand additivity in a cluster system; the oxidation potential of the alkylidyne cluster core
is as sensitive to π-donor substituent effects as are aromatic π complexes such as CpFe-
(C5H4X) to substituents on the rings, and the dependence of the oxidation potential of the
cluster upon ligand substitution on any given Ru atom is 37% of that expected for the
oxidation of a monometallic complex. Reactions with 1 equiv of Ag+ or ferricenium give the
47-electron cations [H3Ru3(CX)(CO)6L3]1+ (X ) OMe; L3 ) (PPh3)2L′; L′ ) CO, PPh3, P(OMe)3,
CNCH2Ph; X ) Ph, NMeBz, SEt, L ) PPh3), characterized by EPR spectroscopy; the
equilibrium between isomers H3Ru3(COMe)(CO)6(ax-PPh3)2(ax- and eq-PPh3)0/1+ (ax ) axial
coordination, eq ) equatorial coordination) favors the axial coordination for the 48-electron
cluster (eq/ax ) 0.15) and equatorial coordination for the 47-electron species (eq/ax ) 6.4).
The rate constant for ax-eq isomerization for the 47-electron species (6.5 s-1) is 4 orders of
magnitude greater than that for the 48-electron species (2 × 10-4 s-1). The 47-electron
clusters have lifetimes which are correlated with the oxidation potentials of the precursors,
ranging from seconds to many hours at room temperature. Slow decomposition of [H3Ru3-
(COMe)(CO)6L3]1+ in the absence of added CO forms the new 46-electron clusters [H3Ru3-
(CO)7L3]1+ (L ) PPh3, AsPh3). [H3Ru3(COMe)(CO)6(PPh3)3]1+ does not react with CCl4 but
does react with Lewis bases such as CO and acetonitrile. Disproportionation occurs with
acetonitrile, chloride, and pyridine; the CO products are [HRu3(CO)9(PPh3)3]1+, [MePPh3]1+,
and H3Ru3(COMe)(CO)9-n(PPh3)n (n ) 1, 2).
In tr od u ction
changes of the metal cluster framework induced by
electrochemical oxidations or reductions.3 Radical cat-
ions formed by oxidation of saturated dinuclear com-
plexes are usually very unstable with respect to frag-
mentation. Odd-electron clusters of three or more metal
atoms, especially clusters capped by main-group atoms,
seem to be more stable than dinuclear odd-electron
species.
The electrochemistry of alkylidynetrimetallic clusters
has been the subject of a number of studies.4 The most
extensive studies have concerned Co3(CX)(CO)9 and
substituted derivatives. The HOMO for this series has
Co-Co bonding character.5 These clusters undergo one-
electron reductions to form 49-electron radical anions,
and phosphine-substituted derivatives can be oxidized;
The redox properties of molecular metal clusters are
of interest with respect to the chemical reactivity of
radicals,1 novel electrical properties of organometallics,2
and fundamental chemistry of metal-metal bonds.3
Electrochemical studies of organometallic clusters have
received increasing attention because of the importance
of odd-electron intermediates in many reaction mech-
anisms and also because of examples of structural
† Present address: Chemistry Department, Indiana University at
South Bend, South Bend, IN 46615.
‡ On leave from the Commonwealth Scientific and Industrial
Research Organization (CSIRO), Division of Minerals, Port Melbourne,
Victoria, 3207, Australia.
(1) (a) Brown, T. L. In Organometallic Radical Processes; Trogler,
W. C., Ed.; Elsevier: Amsterdam, 1990; p 67. (b) Kochi, J . K. J .
Organomet. Chem. 1986, 300, 139. (c) Tyler, D. R. Prog. Inorg. Chem.
1988, 36, 125. (e) Connelly, N. G. Chem. Soc. Rev. 1989, 18, 153. (f)
Geiger, W. E. Prog. Inorg. Chem. 1985, 33, 275. (g) Trogler, W. C. In
Organometallic Radical Processes; Trogler, W. C., Ed.; Elsevier:
Amsterdam, 1990; p 306.
(3) (a) Lemoine, P. Coord. Chem. Rev. 1982, 47, 55. (b) Lemoine, P.
Coord. Chem. Rev. 1988, 83, 169. Other leading references: (c) Etienne,
M.; Mathieu, R.; Lugan, N. Organometallics 1993, 12, 140. (d) Pugh,
J . R.; Meyer, T. J . J . Am. Chem. Soc. 1992, 114, 3784. (e) Bullock, J .
P.; Palazotto, M. C.; Mann, K. R. Inorg. Chem. 1991, 30, 1284. (f)
Connelly, N. G.; Geiger, W. E. Adv. Organomet. Chem. 1985, 24, 87.
(2) Astruc, D. Electron Transfer and Radical Processes in Transition-
Metal Chemistry; VCH Publishers, Inc.: New York, 1995; Chapter 4.
S0276-7333(97)00769-3 CCC: $15.00 © 1998 American Chemical Society
Publication on Web 02/11/1998

Ligand Additivity in Metal Clusters
Organometallics, Vol. 17, No. 5, 1998 873
Ru3(CSEt)(CO)9,14 H3Ru3(CX)(CO)9-nLn (X ) MeO, Me, Ph; n
) 2, 3; L ) AsPh3),15 HRu3(COMe)(CO)10-nLn,16 H2Os3(CO)10,17
H2Os3(CO)9(PPh3),18 and H3Ru3(CNMeBz)(CO)6(SbPh3)311 were
prepared as previously described. Synthetic procedures for the
however, the products are not very stable. The redox
series [FcCCo3(CO)6 L3]2+/1+/0/1- (Fc ) ferrocenyl) was
examined by electrochemical methods.4i,j The first oxi-
dation corresponds to the ferrocenyl moiety, whereas the
second involves the Co3C unit. The [1+,1+] species
could be prepared by chemical oxidation. A linear
relationship between E1/2[1,0] and the number of PR3
ligands was noted. Mixed-metal analogues have also
been studied.4f,g,h
The 49/48/47-electron species [Co3(µ3-CPh)2Cp3]1-/0/1+
have been generated electrochemically. The 48/47-
electron oxidation is at 0.34 V vs SCE and is electro-
chemically reversible. The 47-electron cation displays
a broad EPR signal at g ) 2.2. The HOMO of the 48-
electron cluster is a degenerate pair, implying a J ahn-
Teller distorted 47-electron species.6
15
preparation of H3Ru3(COMe)(CO)6(PPh3)3 and H3Ru3(CPh)-
(CO)6(PPh3)315 have been previously published; however, these
procedures have been modified, and the new preparations are
described in the Supporting Information. Dichloromethane
was distilled under nitrogen from calcium hydride before use.
Other chemicals were of reagent grade purity and were used
as received. All manipulations were carried out under a
blanket of nitrogen and with nitrogen-saturated solvents using
standard Schlenk procedures. However, workup and spectral
analyses were conducted without exclusion of air.
P h ysica l Meth od s of Ch a r a cter iza tion s. Infrared spec-
tra were recorded on Mattson Instruments Alpha Centauri
FTIR, Nicolet Magna 550 FTIR, or Beckman 4250 spectro-
photometers. 1H NMR spectra were obtained on J EOL FX-
90, Varian Associates Gemini 300, or Varian Associates VXR-
400S instruments, using deuteriochloroform or dichloromethane-
d2 as the solvent and TMS as the reference. 13C NMR spectra
were recorded either on the Gemini 300 or the VXR-400S
The electrochemistry of Ru3(µ3-CMe)(µ-CO)3(η5-C5-
Me5)3 displays two electrochemically reversible one-
electron oxidations at 0.17 and 0.86 V (reference not
specified).7 The strong donor Cp* ligands stabilize the
47-electron product, so much so that the radical cation
could be isolated and crystallographically characterized.
The structure has approximate C3v symmetry. While
no EPR signal was detectable, a magnetic moment of
2.0 µB was measured. The 46-electron species was also
isolated; deprotonation generated the vinylidene [Ru3-
instrument in deuteriochloroform and referenced to TMS. 31
P
NMR spectra were recorded on the VXR-400S instrument in
deuteriochloroform, and chemical shifts are reported relative
to o-phosphoric acid. EPR spectra were recorded on an IBM/
Bruker X-band ER-200 SRC spectrometer, with a microwave
power of 20 mW, in dichloromethane solution.20
H3Ru 3(CNMeBz)(CO)6(P P h 3)3. To a solution of HRu3-
(CNMeBz)(CO)10 (105 mg) in 50 mL of cyclohexane was added
150 mg (4 equiv) of PPh3. Hydrogen gas was bubbled through
the solution, and it was heated at reflux for 8 h, during which
time the color changes from yellow to red. The solvent was
removed, and the product recrystallized from a dichloromethane/
methanol mixture. Yield: 157 mg, 77%. Anal. Calcd for
Ru3C69H58O6P3N: C, 59.48; H, 4.19. Found: C, 59.11; H, 3.98.
H3Ru 3(CSEt)(CO)6(P P h 3)3. To a solution of H3Ru3(CSEt)-
(CO)9 (105.1 mg) in 15 mL of dichloromethane was added 176.6
mg (4 equiv) of PPh3. This solution was stirred under nitrogen
at room temperature overnight. The solvent was removed, and
(µ3-η2-CCH2)(µ-CO)3(η5-C5Me5)3]1+
.
The class of clusters H3Ru3(µ3-CX)(CO)9-n(PPh3)n is
closely related to the tricobalt series, but the presence
of hydride ligands and differences in the nature of the
HOMO as a result make the reaction chemistry very
different.8 The HOMO is Ru-CX bonding in nature and
involves π conjugation with a filled p orbital of the X
substituent, so that a one-electron oxidation should
generate a radical cation in which the SOMO has
metal-carbon bonding character.9 We report here,
concerning the electrochemical properties of these com-
pounds, the characterization of 47-electron radical
cation oxidation products and also some chemical reac-
tions of these 47-electron clusters. A preliminary ac-
count of some of this work has appeared previously.10
the product was recrystallized from
a dichloromethane/
methanol mixture. Yield: 203.6 mg, 92%. Anal. Calcd for
Ru3C63H53O6P3S: C, 56.71; H, 4.00. Found: C, 57.02; H, 3.99.
H3Ru 3(COMe)(CO)6(P P h 3)2(CNCH2P h ). The reaction of
H3Ru3(COMe)(CO)7(PPh3)2 (2.3 mM) and CNBz (2.4 mM) in
CDCl3 (0.7 mL) was monitored by 1H NMR spectroscopy. The
substitution initially produced H3Ru3(COMe)(CO)6(PPh3)2(ax-
CNBz), which isomerized to an equilibrium mixture (eq/ax )
88/12) with H3Ru3(COMe)(CO)6(PPh3)2(eq-CNBz). TLC of the
residue on silica eluting with 2:1 dichloromethane:hexanes
yielded 18 mg, 51%. Anal. Calcd for Ru3C52H43NO7P3: C,
53.89; H, 3.74. Found: C, 53.72; H, 3.71.
Rate constants were determined from the integrals of the
hydride resonances. The rate constant for the initial substitu-
tion was determined from the slope of a plot of ln(mol fraction
of H3Ru3(COMe)(CO)7(PPh3)2) vs time (ksub ) 9.8 × 10-5 sec-1
Exp er im en ta l Section
St a r t in g Ma t er ia ls. HRu3(µ-CNMeBz)(CO)10,11 H3Ru3-
(COMe)(CO)9,12 H3Os3(COMe)(CO)9,12 H3Ru3(CPh)(CO)9,13 H3-
(4) (a) Peake, B. M.; Robinson, B. H.; Simpson, J .; Watson D. J .
Inorg. Chem. 1977, 16, 405. (b) Bond, A. M.; Peake, B. M.; Robinson,
B. H.; Simpson, J .; Watson, D. J . Inorg. Chem. 1977, 16, 410. (c) Bond,
A. M.; Dawson, P. A.; Peake, B. M.; Reiger, P. H.; Robinson, B. H.;
Simpson, J . Inorg. Chem. 1979, 18, 1413. (d) Downard, A. J .; Robinson,
B. H.; Simpson, J . Organometallics 1986, 5, 1122. (e) Kotz, J . C.;
Petersen, J . V.; Reed, R. C. J . Organomet. Chem. 1976, 120, 433. (f)
Lindsay, P. N.; Peake, B. M.; Robinson, B. H.; Simpson, J .; Honrath,
U.; Vahrenkamp, H.; Bond, A. M. Organometallics 1984, 3, 413. (g)
Honrath, U.; Vahrenkamp, H. Z. Naturforsch. 1984, 39b, 545. (h)
Honrath, U.; Vahrenkamp, H. Z. Naturforsch. 1984, 39b, 555. (i)
Colbran, S. B.; Robinson, B. H.; Simpson, J . Organometallics 1983, 2,
952. (j) Colbran, S. B.; Robinson, B. H.; Simpson, J . Organometallics
1983, 2, 943. (k) Downard, A. J .; Robinson, B. H.; Simpson, J .
Organometallics 1986, 5, 1132. (l) Downard, A. J .; Robinson, B. H.;
Simpson, J . Organometallics 1986 5, 1140. (m) Robinson, B. H.;
Simpson, J .; Trounson, M. E. Aust. J . Chem. 1986, 39, 1435. (n) Worth,
G. H.; Robinson, B. H.; Simpson, J . Organometallics 1992, 11, 3863.
(5) Chesky, P. T.; Hall, M. B. Inorg. Chem. 1981, 20, 4419.
(6) Enoki, S.; Kawamura, T.; Yonezawa, T. Inorg. Chem. 1983, 22,
3821.
(10) (a) Feighery, W. G.; Allendoerfer, R. D.; Keister, J . B. Organo-
metallics 1990, 9, 2424. (b) Churchill, M. R.; Lake, C. H.; Feighery,
W. G.; Keister, J . B. Ibid. 1991, 10, 2384.
(11) Keister, J . B.; Payne, M. W.; Muscatella, M. J . Organometallics
1983, 2, 219.
(12) Keister, J . B.; Shapley, J . R.; Strickland, D. A. Inorg. Synth.
1990, 27, 196.
(13) Keister, J . B.; Horling, T. L. Inorg. Chem. 1980, 19, 2304.
(14) Churchill, M. R.; Ziller, J . W.; Dalton, D. M.; Keister, J . B.
Organometallics 1987, 6, 806.
(15) Rahman, Z. A.; Beanan, L. R.; Bavaro, L. M.; Modi, S. P.;
Keister, J . B.; Churchill, M. R. J . Organomet. Chem. 1984, 263, 75.
(16) Dalton, D. M.; Barnett, D. J .; Duggan, T. P.; Keister, J . B.;
Malik, P. T.; Modi, S. P.; Shaffer, M. R.; Smesko, S. A. Organometallics
1985, 4, 1854
(7) Connelly, N. G.; Forrow, N. J .; Knox, S. A. R.; Macpherson, K.
A.; Orpen, A. G. J . Chem. Soc., Chem. Commun. 1985, 16.
(8) Keister, J . B. Polyhedron 1988, 7, 847.
(17) Kaesz, H. D. Inorg. Synth. 1990, 28, 238.
(9) Sherwood, D. E., J r.; Hall, M. B. Organometallics 1982, 1, 1519.
(18) Deeming, A. J .; Hasso, S. J . Organomet. Chem. 1976, 114, 313.

874 Organometallics, Vol. 17, No. 5, 1998
Feighery et al.
at 19 °C). Rate constants for isomerization were determined
from a nonlinear least-squares fit of the equation [ax] - [ax]eq
) (ka[S]0/(kb - 9.8 × 10-5 s-1))exp(-9.8 × 10-5 s-1 t) + ([ax]0
- ((ka[S]0)/(kb - 9.8 × 10-5 s-1))exp(-kbt), where [ax] and [ax]eq
are the mole fractions of the axial isomer at time t and at
equilibrium, respectively, [S]0 and [ax]0 are the mole fractions
added. After the mixture was allowed to stand overnight, the
NMR spectrum contained resonances due to [H3Ru3(CO)7-
(PPh3)3]+ (42%) and H3Ru3(COMe)(CO)6(PPh3)3 (30%), the
yields determined by integration versus the toluene standard.
No other hydride resonances were observed.
A solution of [H3Ru3(COMe)(CO)6(PPh3)3]1+, prepared by
reaction of H3Ru3(COMe)(CO)6(PPh3)3 (27 mg, 20 µmol) with
AgSO3CF3 (5 mg, 23 µmol) in dichloromethane (20 mL), was
allowed to stand overnight under nitrogen. Following evapo-
of the starting material and the axial isomer at t ) 0, ka
ksub - kr, and kb ) kf + kr. The forward and reverse rate
constants for isomerization of H3Ru3(COMe)(CO)6-
)
1
(PPh3)2(ax-CNBz) to H3Ru3(COMe)(CO)6(PPh3)2(eq-CNBz) are
ration of the solution to dryness, the H NMR spectum of the
kf ) 8.8 × 10-5 and kr ) 1.2 × 10-5 s-1, respectively.
residue in deuteriochloroform showed the presence of [H3Ru3-
(CO)7(PPh3)3]1+, a doublet at 2.54 ppm attributed to MePPh3
,
1+
The product could also be obtained by reaction of CNBz with
H3Ru3(COMe)(CO)6(PPh3)3. Over a several hour period, the
disappearance of H3Ru3(COMe)(CO)6(PPh3)3 and formation of
a mixture of axially and equatorially substituted H3Ru3-
(COMe)(CO)6(PPh3)2(CNBz) was observed. The initial product
mixture was richer in the axially substituted isomer, but
overnight the equilibrium mixture was obtained.
and very broad resonances at -15.2 and 4.2 ppm assigned to
[H3Ru3(COMe)(CO)6(PPh3)3]1+/0. Carbon monoxide was bubbled
through the solution to decompose the remaining radical cation
and to convert [H3Ru3(CO)7(PPh3)3]1+ to the red species
believed to be [HRu3(CO)9(PPh3)3]1+
.
The spectrum now
showed the presence of H3Ru3(COMe)(CO)6(PPh3)3 (-15.1
ppm), H3Ru3(COMe)(CO)7(PPh3)2 (-15.85, -16.35 ppm), H3-
Ru3(COMe)(CO)8(PPh3) (-16.9, -17.5 ppm), [HRu3(CO)9-
H3Ru 3(COMe)(CO)6(P P h 3)2(P (OMe)3). A solution of H3-
Ru3(COMe)(CO)7(PPh3)2 (32 mg, 0.030 mmol) and P(OMe)3 (3.9
µL, 0.033 mmol) in dichloromethane (30 mL) was stirred under
nitrogen for 9 h. TLC on silica eluting with 2:1 dichlo-
romethane:hexanes yielded two yellow bands characterized as
H3Ru3(COMe)(CO)6(PPh3)2(P(OMe)3) (17.5 mg, 50%) and H3Ru3-
(COMe)(CO)6(PPh3)(P(OMe)3)2 (9.4 mg, 30%) in order of elu-
tion. Anal. Calcd for Ru3C47H45O10P3: C, 48.415; H, 3.89.
Found: C, 48.22; H, 3.99.
H3Os3(COMe)(CO)9-n (P P h 3)n (n ) 1, 2). To a stirred
solution of H3Os3(COMe)(CO)9 (127 mg) in cyclohexane (20 mL)
was added PPh3 (39 mg, 1 equiv). The solution was heated to
reflux under a nitrogen atmosphere for 24 h. The products
were separated by thin-layer chromatography on silica gel,
eluting with a 5:1 cyclohexane:dichloromethane mixture, af-
fording three yellow bands. The top band was unreacted
starting material (30 mg, 24%) followed by the monosubsti-
tuted product (51 mg, 32%) and then by the di-substituted
product (60 mg, 31%). Anal. Calcd for Os3C45H36O8P2: C,
40.42; H, 2.71. Found: C, 40.50; H, 2.53.
Ch em ica l Oxid a tion of Su bstitu ted Alk ylid yn e Clu s-
ter s. For each of the clusters H3Ru3(CX)(CO)9-nLn (X ) OMe,
L ) AsPh3 and PPh3, n ) 0-3; X ) NMeBz, L ) PPh3, n ) 3,
X ) SEt, L ) PPh3, n ) 3; X ) Ph, L ) PPh3, n ) 3), H3-
Ru3(COMe)(CO)6(PPh3)2L (L ) CNCH2Ph, P(OMe)3), and
H3Os3(COMe)(CO)9-n(PPh3)n (n ) 0-2), the procedure for the
oxidation was as follows. To a dichloromethane solution of
the cluster was added 1 equiv of the oxidant ([Fe(o-phen)3]-
(ClO4)3, AgCF3SO3, AgBF4, or FeCp2PF6) under nitrogen. The
solution immediately turned from yellow/orange to green. The
solvent was removed under vacuum. For ESR analysis, the
mixture was dissolved in a small amount of dichloromethane
(approximately 10-3 M in cluster), filtered into an ESR tube,
and vacuum sealed. For NMR analysis, the mixture was
dissolved in CDCl3 and filtered into an NMR tube.
Red u ction of [H3Ru 3(COMe)(CO)6(P P h 3)3]1+. To a solu-
tion of 19.6 mg of H3Ru3(COMe)(CO)6(PPh3)3 in dichlo-
romethane was added 4.8 mg (1 equiv) of AgCF3SO3. The
solution turned green. A freshly prepared solution of Na/
benzophenone in THF was then added dropwise until the
cluster solution turned from green to orange. After the solvent
was removed and redissolution in CDCl3, the 1H NMR spec-
trum showed that only H3Ru3(COMe)(CO)6(PPh3)3 was present.
After TLC, the starting material was recovered (18.1 mg, 93%
yield).
(PPh3)3]1+ ( -17.05 ppm) and MePPh3 (2.54 ppm), in molar
1+
ratios of 0.18:0.86:0.16:1:1.
Rea ction s of H3Ru 3(COMe)(CO)6(P P h 3)31+ w ith Ca r bon
Mon oxid e. Carbon monoxide was bubbled through a solution
of [H3Ru3(COMe)(CO)6(PPh3)3]1+, prepared by reaction of H3-
Ru3(COMe)(CO)6(PPh3)3 (50.3 mg, 38.6 µmol) with FeCp2PF6
(14.1 mg, 42.5 µmol) in dichloromethane (20 mL). The color
of the solution rapidly changed from yellow-green to red. IR
spectra were taken periodically until no further change was
noted after 2 h. The solution was evaporated to dryness, and
the residue was applied to a silica gel TLC plate, which was
eluted with dichloromethane/hexanes (1:1 v/v) to remove
ferrocene (isolated 7.0 mg) and H3Ru3(COMe)(CO)9-n(PPh3)n
(n ) 3, 11.6 mg (23%); n ) 2, 2.4 mg, impure). The bottom
red band was extracted with ethyl acetate/dichloromethane;
evaporation yielded 40.5 mg of a red solid. The red material
was dissolved in a minimum amount of methanol, and a
methanol solution of sodium tetraphenylborate (16.1 mg, 46.8
µmol) was added. A red-purple precipitate was collected by
filtration and washed with methanol. This material is likely
a mixture of products. The red solid was dissolved in dichlo-
romethane, filtered, and then recrystallized from dichlo-
romethane-methanol at -20 °C. Red-purple crystals were
isolated from the near-colorless solution. (18.9 mg, 11.4 µmol,
29.5% based upon starting material). IR (CH2Cl2): 2096 vw,
2064 m, 2026 vs, 2011 sh, 1992 vw sh, 1967 vw cm-.1 1H NMR
(CDCl3): -17.05 (dt, J PH 3, 6 Hz) ppm. 31P NMR (CDCl3), PF6
salt: 26.1 (1 P, d, J 12 Hz), 29.2 (1 P, s), 32.6 (1 P, d, J 12 Hz),
-144.5 (1 P, sept, J PF 714 Hz), 29.2 (0.2 P, s) ppm. MS
(positive-ion FAB), PF6 salt: m/e M+ 1345, (M - CO)+ 1317
(
102Ru3). Anal. Calcd for BRu3C87H66O9P3: C, 62.86; H, 4.00.
Found: C, 62.09; H, 4.00.
Ca r bon Tetr a ch lor id e. Oxidation of H3Ru3(COMe)(CO)6-
(PPh3)3 (0.86 mM) was achieved by addition of ferricinium
hexafluorophosphate. IR analysis indicated complete conver-
sion to [H3Ru3(COMe)(CO)6(PPh3)3]1+
. A 5 mL portion was
transferred to a second flask, and carbon tetrachloride (7 µL,
17 equiv) was added. No differences between the IR spectra
of the two solutions were noted after 25 min.
Aceton itr ile. To a 5 mL portion of the solution prepared
as described above was added acetonitrile (7.0 µL, 30 equiv).
The color of the solution immediately began to change from
yellow-green to dark red. After 30 min, the IR spectrum of
the solution contained absorptions at 2065 w, 2036 s, 2009 s,
1973 m, and 1960 m cm-1. Chromatographic workup allowed
the recovery of H3Ru3(COMe)(CO)6(PPh3)3 in 27% yield.
P yr id in e. The addition of pyridine (5 µL, 7.7 equiv) to a
solution of [H3Ru3(COMe)(CO)6(PPh3)3]+ (formed from 10.5 mg
H3Ru3(COMe)(CO)6(PPh3)3 and FeCp2PF6 (2.7 mg) in 5 mL
CH2Cl2) caused the color of the solution to change from green-
yellow to red-orange within 15 min. The IR spectrum con-
Decom p osit ion of [H3R u 3(COMe)(CO)6(P P h 3)3]1+ t o
[H3Ru 3(CO)7(P P h 3)3]1+. To a solution of H3Ru3(COMe)(CO)6-
(PPh3)3 (30 mg) in CDCl3 (0.7 mL) was added 2 µL of toluene.
The methyl proton resonance of toluene was used as a
reference peak to measure the relative amounts of clusters
present during various stages of the reaction. Oxidation of
the cluster with 1 equiv of Ag+ turned the color of the solution
from orange to green. A small amount of water (ca. 5 µL) was

Ligand Additivity in Metal Clusters
Organometallics, Vol. 17, No. 5, 1998 875
tained absorptions due to H3Ru3(COMe)(CO)6(PPh3)3 in addi-
ucts. Then the solution was shaken vigorously after addition
of an excess of CF3COOH (5-8 µL) or CF3SO3H. The hydride
signal due to the trisubstituted cluster disappeared im-
mediately, giving rise to a very small signal at -17.36 ppm
(d, J 3 Hz). After the solution was allowed to stand overnight,
the 1H NMR spectrum indicated that the protonation of the
disubstituted cluster was complete, while the monosubstituted
cluster remained unreacted. The protonation of HRu3(COMe)-
tion to a broad absorption at 2060 cm-1
.
Ch lor id e. The addition of PPNCl (6.2 mg, 1.2 equiv) to a
solution of [H3Ru3(COMe)(CO)6(PPh3)3]+ (formed from 11.7 mg
H3Ru3(COMe)(CO)6(PPh3)3 and FeCp2PF6 (3.2 mg) in 6 mL
CH2Cl2) caused the color to change from green-yellow to yellow
upon mixing. The IR spectrum contained absorptions due the
H3Ru3(COMe)(CO)6(PPh3)3 (ca. 67% recovery).
1
(CO)8(PPh3)2 was indicated clearly by H NMR spectroscopy.
Tr ip h en ylp h osp h in e. Addition of PPh3 (11.2 mg) to a
solution of [H3Ru3(COMe)(CO)6(PPh3)3]+ (formed from 25.2 mg
H3Ru3(COMe)(CO)6(PPh3)3 and AgSO3CF3 (ca. 8 mg)) resulted
in slow changes over a 40 min period. After 24 h, the solution
was evaporated to dryness. The 1H NMR spectrum (CDCl3
solution) displayed signals at 3.90 (br) and -15.2 (br) ppm
assigned to [H3Ru3(COMe)(CO)6(PPh3)3]0/+, 2.92 (d, J 13 Hz)
ppm, assigned to [MePPh3]1+, and -11.1 (br), -11.55 (br d),
and -12.3 (br) assigned to [H3Ru3(CO)7(PPh3)3]+. A ratio of
MePPh3]1+/[H3Ru3(CO)7(PPh3)3]1+ of 1.1 was determined. Next,
carbon monoxide was bubbled through solution for a few
minutes, causing an immediate color change to red. The NMR
spectrum displayed hydride resonances due to H3Ru3(COMe)-
(CO)6(PPh3)3, H3Ru3(COMe)(CO)7(PPh3)2, and [HRu3(CO)9-
(PPh3)3]1+ in molar ratios of 1.2:0.2:0.9.
Reaction of [H3Ru 3(COMe)(CO)6(P P h 3)3]1+ with [H3Ru 3-
(COMe)(CO)7(P P h 3)2]1+ a n d P P h 3. In one 50 mL Schlenk
flask was placed H3Ru3(COMe)(CO)6(PPh3)3 (16.6 mg, 12.7
µmol), PPh3 (3.7 mg, 14.4 µmol), and dichloromethane (10 mL),
and in a second flask H3Ru3(COMe)(CO)7(PPh3)2 (14.9 mg, 13.9
µmol) and dichloromethane (10 mL) were placed. To each flask
was added FeCp2PF6 (4.7 mg (14.2 µmol) and 5.2 mg (15.7
µmol), respectively). Then the contents of the second flask
were transferred via pipet to the first flask. The IR spectrum
of the resulting solution displayed absorptions at 2080 m, 2072
m br, 2027 s, 2017 vs, and 1964 m cm-.1 After 15 min, the
solution was evaporated to dryness. The 1H NMR spectrum
displayed resonances at -11.1, -11.5, and -12.3 ppm assigned
to [H3Ru3(CO)7(PPh3)3]1+ (relative intensity 0.54), -15.75 and
-16.35 ppm assigned to H3Ru3(COMe)(CO)7(PPh3)2 (relative
intensity 0.15), -12.9 (br, 1 H) and -16.65 (br, 1 H), unidenti-
fied (relative intensity 0.22), and -17.01 assigned to [HRu3-
(CO)9(PPh3)3]1+ (relative intensity 0.09).
The growth of new signals (-17.24 ppm, dd, 1H, J PH ) 20.8,
8.4 Hz; -13.19 ppm, d, 1H, J PH ) 10.4 Hz) due to the
protonation product [H2Ru3(COMe)(CO)8(PPh3)2]+ were ac-
companied by diminution of the hydride signal due to the
starting material HRu3(COMe)(CO)8(PPh3)2.
Oxid a tion of H3Os3(COMe)(CO)8(P P h 3). To a solution
of 30 mg of H3Os3(COMe)(CO)8(PPh3) in dichloromethane was
added 7 mg (1 equiv) of AgO3SCF3. The solution turned green.
After the mixture was stirred overnight, the color had changed
to yellow-orange. The 1H NMR spectrum of the products
showed the presence of [H3Os3(CO)9(PPh3)]+, starting material,
and other unidentified products. The mixture was dissolved
in MeOH (which deprotonates [H3Os3(CO)9(PPh3)]1+). Thin-
layer chromatography (silica gel, hexane elution) yielded H2-
Os3(CO)9(PPh3) (10 mg, 34%).
P r oton a tion of H2Os3(CO)10 a n d H2Os3(CO)9(P P h 3).
NMR samples of [H3Os3(CO)10]1+ and [H3Os3(CO)9(PPh3)]1+
were prepared as previously reported.18 NMR data for [H3-
Os3(CO)10]1+ (CDCl3). 13C NMR: 151.8 (2 C(a), d, J CH 14.6 Hz),
164.4 (1 C(b), d, J CH 4.0 Hz), 166.8 (1 C(c)), 170.1 (1 C(d)), 170.2
1
(1 C(e)), 170.5 (2 C(f)), 172.0 (2 C(g), d, J CH 11.0 Hz) ppm. H
NMR: -12.33 (2 H(a), d, J ab 1.4 Hz), -14.49 (1 H(b), t, J ab 1.4
Hz) ppm. NMR data for [H3Os3(CO)9(PPh3)]1+ (CDCl3). 13C
NMR: 165.4 (1 C(a), d, J CH 4 Hz), 165.7 (1 C(b), dd, J CH 10
Hz, J CP 9 Hz), 166.1 (1 C(c)), 168.0 (1 C(d), dd, J CH 13 Hz, J CP
10 Hz) 170.6 (1 C(e), d, J CH 9 Hz), 170.7 (1 C(f)), 171.2 (1 C(g)),
173.6 (1 C(h), d, J CH 13 Hz), 175.0 (1 C(i), dd, J CH 10 Hz, J CP
10 Hz) ppm. 1H NMR: -11.89 (1 H (a), d, J aP 7.3 Hz), -12.03
(1 H (b), J bP 30.5 Hz), and -13.78 ppm (1 H(c), d, J cP 9.3 Hz)
ppm.
Electr och em istr y. Cyclic voltammetric experiments were
performed with Bioanalytical Systems, Inc. (West Lafayette,
IN) instrumentation (BAS-100 and BAS-100W electrochemical
analyzers). Measurements were made in accordance with
standard techniques, which included purging of solutions with
nitrogen to exclude oxygen. Dichloromethane was freshly
distilled from calcium hydride and stored under nitrogen. The
supporting electrolyte was 0.1 M tetrabutylammonium tet-
rafluoroborate, TBATFB (Kodak, also prepared by ion ex-
change of tetrabutylammonium bromide with sodium tetraflu-
oroborate), which had been recrystallized 3 times from ethanol/
water and vacuum-dried. The working electrode for the
experiments at relatively low scan rates was a platinum disk
electrode (diameter ) 3 mm). For work at higher scan rates
and for some of the steady-state investigations, smaller
platinum disks (BAS, diameter ) 100 and 10 µm) were
employed. A platinum wire served as the auxiliary (counter)
electrode. The concentration of the analyte was 10-3 M unless
stated otherwise. The reference electrode used was either the
saturated calomel electrode (SCE) or Ag|AgCl|LiCl(satd) in
C2H5OH. For voltammetric studies that involved relatively
high scan rates, a quasi-reference electrode (silver wire) was
employed. Compensation for resistive losses (“iR drop”) was
employed for all measurements. Under typical experimental
conditions, the reversible oxidation for ferrocene was centered
on E ) 0.53 V vs SCE. The potential peak-to-peak separation,
∆Ep ) Ep,a - Ep,c, was within a range of 65 ( 5 mV. All
potentials are reported relative to the ferrocene/ferricinium
couple as 0 V. Simulated cyclic voltammograms were com-
puted by means of DigiSim 2.1 software, obtained from
Bioanalytical Systems, Inc. Details of the procedures followed
are provided under Results and Discussion.
Tw o-E lect r on Oxid a t ion of H3R u 3(COMe)(CO)9-n
-
(P P h 3)n . When 2 equiv of AgBF4 was added to the dichlo-
romethane solution of H3Ru3(COMe)(CO)7(PPh3)2, the solution
turned into dark green and then dark red within minutes. Bulk
electrolysis at the potential of the second oxidation also gave
the same product, based on IR spectroscopy. The CO stretch-
ing frequencies shifted to higher values. The 1H NMR
spectrum displays hydride resonances at -17.24 (dd) ppm,
with phosphorus-hydride coupling constants of 20.8 and 8.4
Hz, and -13.19 (d) ppm, with a phosphorus-hydride coupling
constant of 10.4 Hz. The resonance assigned to a methoxy
group appears at 4.53 ppm. In the 31P NMR spectrum, two
singlets at 38.1 and 28.1 ppm were observed. The product of
the two-electron oxidation is proposed to be [H2Ru3(COMe)-
(CO)8(PPh3)2]+. This cation can be prepared by the alternative
route of protonation of HRu3(COMe)(CO)8(PPh3)2.
Treatment of H3Ru3(COMe)(CO)6(PPh3)3 with 2 equiv of
AgBF4 caused decomposition, and no products were identifi-
able. However, oxidation with 2 equiv of FeCp2PF6 in the
presence of 1 equiv of PPh3 generates [H3Ru3(CO)7(PPh3)3]1+
.
P r ot on a t ion of H R u 3(µ-COMe)(CO)10-n (P P h 3)n (n
)
1-3). A mixture of clusters HRu3(COMe)(CO)10-n(PPh3)n (n
) 1-3) was prepared in cyclohexane.16 The ratio of the
products was controlled by the amount of PPh3 for the ligand
substitution reaction. Typically, three phosphine-substituted
clusters, e.g., n ) 1-3, can be obtained with a ratio of 2:2:1
by adding 2.5 equiv of PPh3 and stirring overnight. Cyclo-
hexane was removed by rotary evaporation. The residue was
1
dried in vacuo and dissolved in CDCl3 in a NMR tube. A H
NMR spectrum was recorded to determine the ratio of prod-

876 Organometallics, Vol. 17, No. 5, 1998
Feighery et al.
The diffusion coefficient for H3Ru3(CNMeBz)(CO)6(PPh3)3
was determined to be 1.61(38) × 10-6 cm2 s-1 by using normal
pulse voltammetry and linear regression of (id2π)/(nFACR*)2
vs 1/tp for experiments at a number of pulse widths (5, 50, 100
ms).26 Assuming that the diffusion coefficients of the other
clusters are similar to the value for the diffusion coefficient of
H3Ru3(CNMeBz)(CO)6(PPh3)3, the number of electrons, n,
involved in the first oxidation process can be calculated from
the cathodic peak current. For example, from the cyclic
voltammetric data for H3Ru3(COMe)(CO)6(PPh3)3, the cathodic
peak current and the Randles-Sevcik equation gives n )
1.04.21
Con tr olled P oten tia l Cou lom etr y of H3Ru 3(CNMeBz)-
(CO)6(P P h 3)3. The coulometric cell was assembled as fol-
lows: To a jacketed cell, kept at constant temperature (22 °C),
was added 25 mL of a 0.1 M TBATFB solution of dichlo-
romethane and 11.0 mg (7.89 × 10-6 moles) of H3Ru3(CNMeBz)-
(CO)6(PPh3)3. A silver wire was used as a quasi-reference
electrode, and a platinum wire served as the auxiliary elec-
trode. Both electrodes made contact with the solution through
fritted disks. The working electrode was a platinum gauze
electrode (Aesar Unimesh, gauze type 39/1, 25 mm diameter
× 50 mm height cylinder), placed directly into the solution. A
cyclic voltammogram was performed prior to the coulometry
to determine the correct potential. The solution was stirred
vigorously and purged with nitrogen while the experiment was
performed. After approximately 145 min, the current had
decayed to background levels and the experiment was halted.
The amount of charge passed was 0.73 C, and the value
calculated for n, the number of electrons, was 0.96.
F igu r e 1. Structures of the isomers H3Ru3(µ3-COMe)-
(CO)6(ax-L)2(ax-L′) and H3Ru3(µ3-COMe)(CO)6(ax-L)2(eq-
L′).
cally irreversible in acetonitrile, most likely because of
the nucleophilic attack by the solvent on the oxidized
species. For the monosubstituted clusters, the cyclic
voltammetry for the oxidation is also chemically ir-
reversible in dichloromethane. Anodic peak potentials,
from cyclic voltammograms recorded at a scan rate of
100 mV/s, are given in Table 3. By contrast, cyclic
voltammograms for dichloromethane solutions of disub-
stituted and trisubstituted clusters exhibit two oxidation
responses, the first of which is either electrochemically
reversible or quasi-reversible at scan rates of ca. 100
mV/s. The second process, at more positive potentials,
is either quasi-reversible or irreversible and has not
been examined in detail. Figure 2 shows a cyclic
voltammogram for H3Ru3(µ3-CSEt)(CO)6(AsPh3)3, which
displays an electrochemically reversible first oxidation
and a near-reversible second oxidation. Reversibility
was assessed by comparing the values of the peak
current ratio (ip,c/ip,a) and peak-to-peak separation (∆Ep)
for a scan rate of 100 mV/s with those for the ferrocene/
ferricenium couple under the same conditions (ip,c/ip,a
) 1; ∆Ep ) 65 ( 5 mV). Electrochemical data for the
cluster series H3Ru3(µ3-CX)(CO)9-n(L)n (n ) 1-3) are
collected in Table 3. For H3Ru3(µ3-CNMeBz)(CO)6-
(PPh3)3, the first one-electron oxidation process is es-
sentially electrochemically reversible at a scan rate of
100 mV/s (Figure 3a). Controlled potential coulometry
performed on H3Ru3(µ3-CNMeBz)(CO)6(PPh3)3 yielded
a value of 0.96 F/mol for the first oxidation process.
For most of the other disubstituted and trisubstituted
cluster compounds, however, the oxidation responses are
best described as electrochemically quasi-reversible.
This was confirmed by recording cyclic voltammetric
responses over a range of scan rates. For most of the
compounds, ∆Ep was found to increase with increasing
Resu lts a n d Discu ssion
H3Ru 3(µ3-CX)(CO)9-n Ln . The syntheses and char-
acterizations of a number of substituted clusters H3M3-
(µ3-CX)(CO)9-nLn (M ) Ru, X ) OMe, L ) AsPh3, PPh3,
n ) 1-3; M ) Ru, X ) SEt, Ph, NMeBz, L ) PPh3, n )
3; M ) Os, X ) OMe, L ) PPh3, n ) 1, 2) have been
described previously. The EPh3 ligands occupy the axial
coordination sites, trans to the Ru-CX bonds, e.g., H3-
Ru3(CPh)(CO)7(ax-AsPh3)2.15 For small ligands, pyri-
dine and NCMe, equatorial coordination, cis to the Ru-
CX bonds, is preferred.22 We have now prepared the
mixed substitution products H3Ru3(COMe)(CO)6(PPh3)2L
(L ) P(OMe)3 and CNCH2Ph) which exist in solution
as equilibrium mixtures of H3Ru3(COMe)(CO)6(ax-
PPh3)2(ax-L) and H3Ru3(COMe)(CO)6(ax-PPh3)2(eq-L),
eq/ax ) 2.8 and 7.3, respectively (Figure 1). We have
also determined that this equilibrium exists for H3Ru3-
(COMe)(CO)6(PPh3)3 (eq/ax ) 0.15 in dichloromethane).
Because of ring strain, H3Ru3(COMe)(CO)7(µ-(PPh2)2-
CHPPh2) is equatorially substituted but H3Ru3(COMe)-
(CO)6{µ3-(PPh2CH2)3CMe} is axially substituted.10b The
coordination geometry is easily established by 1H NMR
spectroscopy. Representative NMR and IR data are
given in Tables 1 and 2, respectively; data for all
compounds are reported in the Supporting Information
(Tables 1S and 2S).
scan rate while ip,c/ip,a increased. The increase in ip,c
/
ip,a indicates that a follow-up chemical process causes
decomposition of the radical cation. In the present case,
such a decomposition may involve reaction with the
supporting electrolyte (TBATFB), as several of the
oxidation products have long lifetimes in pure dichlo-
romethane (the half-life is 1 h for decomposition of [H3-
Ru3(µ3-COMe)(CO)6(PPh3)3]1+). Other examples of elec-
trolyte-induced decomposition have been reported.23
Chemical evidence (vide infra) indicates that for some
of the clusters ligand isomerization, where a phosphine
ligand migrates from an axial site to an equatorial site,
accompanies the first oxidation process. In particular,
H3Ru3(µ3-CX)(CO)6(PPh3)3 (X ) OMe and SEt) undergo
this isomerization. Figure 3b shows a typical cyclic
voltammogram for H3Ru3(µ3-COMe)(CO)6(PPh3)3. Com-
Electr och em istr y. The electrochemical oxidations
of the clusters (µ-H)3Ru3(µ3-CX)(CO)9-nLn are all chemi-
(19) Bryan, E. G.; J ackson, W. G.; J ohnson, B. F. G.; Kelland, J .
W.; Lewis, J .; Schorpp, K. T. J . Organomet. Chem. 1976, 108, 385.
(20) For a more detailed account of the EPR experiment, see: Male,
R.; Samotowka, M. A.; Allendoerfer, R. D. Electroanalysis 1989, 1, 333.
(21) (a) Bard, A. J .; Faulkner, L. R. Electrochemical Methods; J ohn
Wiley and Sons: New York, 1980. (b) Nicholson, R. S.; Shain, I. Anal.
Chem. 1964, 36, 706.
(22) Beanan, L. R. Ph.D. Thesis, State University of New York at
Buffalo, Buffalo, NY, 1986.

Ligand Additivity in Metal Clusters
Organometallics, Vol. 17, No. 5, 1998 877
Ta ble 1. 1H a n d 31P NMR Da ta in CDCl3 Solu tion a
1H NMR (ppm)
cluster
31P NMR (ppm)
22.0
H3Ru3(COMe)(CO)6(ax-PPh3)3
4.00 (s, 3H)
-15.14 (t, 3H, J PH ) 7.7 Hz)
-15.54 (br, 1 H)
-15.94 (br d, 1 H, J PH ca. 40Hz)
-16.24 (br, 1 H)
3.69 (s, 3 H)
H3Ru3(COMe)(CO)6(ax-PPh3)2(eq-PPh3)
H3Ru3(CNMeBz)(CO)6(PPh3)3
H3Ru3(CSEt)(CO)6(PPh3)3
2.93 (s, 3H)
4.51 (s, 2H)
-15.11 (t, 3H, J PH ) 6.3 Hz)
4.08 (t, 3H, J HH ) 7.1 Hz)
3.32 (q, 2H)
22.1
23.0
-15.33 (t, 3H, J PH ) 7.6 Hz)
[H3Ru3(COMe)(CO)6(PPh3)3]1+
[H3Ru3(COMe)(CO)6(AsPh3)3]1+
[H3Ru3(COMe)(CO)7(PPh3)2]1+
[H3Ru3(CSEt)(CO)6(PPh3)3]1+
[H3Ru3(CNMeBz)(CO)6(PPh3)3]1+
4.1 (v br)
-15.1 (v br)
4.0 (v br)
-15.2 (v br)
3.9 (v br)
-15.8, -16.4 (v br)
3.5, 1.4 (v br)
-15.6 (v br)
9.2-8.8 (v br)
6.0, 5.2 (v br)
hydride not observed
H3Ru3(COMe)(CO)6(PPh3)2(CNBz)
isomer I, axial CNBz
-16.06 (dd, 2H, J PH ) 12.2 Hz, J HH ) 2 Hz),
-15.72 (tt, 1H, J PH ) 11.3 Hz, J HH ) 2 Hz)
4.53 (s, 2H)
22.0 (s)
3.95 (s, 3H)
H3Ru3(COMe)(CO)6(PPh3)2(CNBz)
isomer II, equatorial CNBz
-16.92 (d, 1H, J PH ) 11.7 Hz)
-16.53 (d, 1H, J PH ) 11.1 Hz)
-15.72 (t, 1H, J PH ) 11.3 Hz)
4.70 (d, 1H, J HH ) 17.0 Hz)
4.78 (d, 1H, J HH ) 17.0 Hz)
3.95 (s, 3H)
22.5 (s, 1P)
21.3 (s, 1P)
H3Ru3(COMe)(CO)6(PPh3)2(P(OMe)3)
isomer I, equatorial P(OMe)3
-16.71 (ddt, 1 H, J HH 2 Hz, J PH 10.6, 15.9 Hz)
-16.06 (br t, 1 H, J PH ca. 8.5 Hz)
-15.81 (br t, 1 H, J PH ca. 9 Hz)
3.44 (d, 9 H, J PH 12 Hz)
3.97 (s, 3 H)
H3Ru3(COMe)(CO)6(PPh3)2(P(OMe)3)
isomer II, axial P(OMe)3
ca. -15.9 (2 H, obscured)
-15.65 (br t, 1 H, J PH 9.5 Hz)
3.27 (d, 9 H, J PH 12 Hz)
4.02 (s, 3 H)
[H3Ru3(CO)7(PPh3)3]1+
[H3Ru3(CO)6(CS)(PPh3)3]1+
[H3Ru3(CO)7(AsPh3)3]1+
-11.13 (br s, 1H)
29.4 (1P)
33.7 (1P)
40.4 (d, 1P,
J PH ) 48 Hz)
-11.56 (d, 1H, J PH ) 48 Hz)
-12.35 (br s, 1H)
-11.07 (br s, 1H)
-11.50 (d, 1H, J PH ) 45 Hz)
-12.29 (br s, 1H)
-10.96 (m, 1H)
-12.49 (m, 1H)
-12.59 (m, 1H)
a
Phenyl resonances are not reported.
Ta ble 2. IR Da ta for Selected [(µ-H)3Ru 3(µ3-CX)(CO)9-n Ln ]0/1+ in Dich lor om eth a n e
cluster
ν(CO), cm-1
H3Ru3(COMe)(CO)6(PPh3)3
H3Ru3(CNMeBz)(CO)6(PPh3)3
H3Ru3(CSEt)(CO)6(PPh3)3
2035 s, 2010 vs, 1959 m
2031 s, 2005 s br, 1953 m
2036 s, 2013 vs, 1963 m
2069 w sh, 2062 w, 2045 vs, 2037 sh, 2007 s, 1961 w
2057 s, 2036 vs, 1993 s
2085 vw, 2065 m, 2044 vs, 2036 vs, 2028 m, 2014 vs, 2005 s, 1992 w, 1964 w
2178 w, 2030 m, 2011 vs, 2001 sh, 1967 m, 1954 m, 1939 sh
2036 m, 2032 m, 2011 vs, 2002 m sh, 1964 m, 1954 m
[H3Ru3(COMe)(CO)6(PPh3)3]1+
[H3Ru3(CNMeBz)(CO)6(PPh3)3]1+
[H3Ru3(CSEt)(CO)6(PPh3)3]1+
H3Ru3(COMe)(CO)6(CNBz)(PPh3)2
H3Ru3(COMe)(CO)6(P(OMe)3)(PPh3)2
parison of this response with that of Figure 3a (for the
µ3-CNMeBz analog) shows a significantly larger value
of ∆Ep for the former. While such an increase in this
parameter may be due simply to slower electron-
transfer kinetics, it is also consistent with the ECEC
mechanism. Here, the electrochemical behavior is
described by the so-called “square” reaction scheme
(Scheme 1). The behavior that follows this scheme has
been noted for a number of organometallic redox
processes.1f
Further support for the presence of a mechanism that
involves both chemical and electrochemical steps was
obtained by a more extensive examination of the cyclic
voltammetric responses over a range of potential scan
rates. To minimize the distortion of responses due to
the effects of uncompensated solution resistance, vol-

878 Organometallics, Vol. 17, No. 5, 1998
Feighery et al.
Ta ble 3. Cyclic Volta m m etr ic Da ta for H3Ru 3(µ3-CX)(CO)9-n Ln (n ) 1-3) in Dich lor om eth a n ea
cluster
H3Ru3(COMe)(CO)9
(Ep,a + Ep,c)/2 [Ep,a1] (V)
∆Ep (mV)
ip,c/ip,a
Ep,a2 (V)
[0.53]
[0.29]
-0.02
-0.22
-0.24
-0.19
[0.37]
0.10
-0.08
-0.59
-0.37
-0.18
-0.09
0
H3Ru3(COMe)(CO)8(PPh3)
H3Ru3(COMe)(CO)7(PPh3)2
H3Ru3(COMe)(CO)6(PPh3)3
H3Ru3(COMe)(CO)6(PPh3)2(P(OMe)3)
H3Ru3(COMe)(CO)6(PPh3)2(CNCH2Ph)
H3Ru3(COMe)(CO)8(AsPh3)
H3Ru3(COMe)(CO)7(AsPh3)2
H3Ru3(COMe)(CO)6(AsPh3)3
H3Ru3(CNMeBz)(CO)6(PPh3)3
H3Ru3(CNMeBz)(CO)6(SbPh3)3
H3Ru3(CSEt)(CO)6(PPh3)3
H3Ru3(CSEt)(CO)6(AsPh3)3
H3Ru3(CPh)(CO)6(PPh3)3
H3Ru3(CMe)(CO)6(PPh3)3
H3Ru3(CEt)(CO)6(PPh3)3
H3Ru3(CBr)(CO)6(PPh3)3
70
95
72
72
0.88
0.77
0.97
0.93
0.24
0.07
0.08
0.13
74
86
61
68
77
72
74
73
74
120
81
0.78
0.84
0.97
0.91
0.78
0.83
0.93
0.91
0.93
0.76
0.60
0.34
0.15
-0.18
0.08
0.28
0.29
0.50
0.42
0.45
0.68
0.10
0
-0.01
0.22
H3Ru3(COMe)(CO)6{(PPh2CH2)3CEt}
H3Ru3(COMe)(CO)7{(PPh2)2CHPPh2}
H3Os3(COMe)(CO)9
H3Os3(COMe)(CO)8(PPh3)
H3Os3(COMe)(CO)7(PPh3)2
-0.11
[0.12]
[0.54]
[0.31]
-0.01
85
0.66
0.27
a
Cluster (10-3 M) in 0.1 M TBATFB in dichloromethane. The working electrode was a 5 mm diameter platinum disk, and the auxiliary
electrode was a platinum wire. The reference electrode was either a silver wire or SCE. The scan rate was 100 mV/s. The potentials are
referenced to the ferrocene/ferricenium couple (0 V, ∆Ep ) 72 mV) under the same conditions.
F igu r e 2. Cyclic voltammogram for H3Ru3(µ3-CSEt)-
(CO)6(AsPh3)3.
tammograms were recorded at a working electrode of
relatively small dimensions (platinum disk, diameter )
100 µm). Use of this electrode, together with a platinum-
wire quasi-reference electrode, allowed essentially full
compensation for resistive losses. From this basis, it
was then possible to attribute any changes in peak
position and/or peak-to-peak separation to the influence
of kinetic (chemical and electrochemical) parameters.
Figure 4 shows typical cyclic voltammetric responses
for the first oxidation process of H3Ru3(µ3-COMe)(CO)6-
(PPh3)3 in dichloromethane solution, for scan rates from
1 to 100 V/s. At 1 V/s (Figure 4a), it is clear that the
oxidation step for this compound is not a simple process.
F igu r e 3. Cyclic voltammogram for the first oxidation of
(a) H3Ru3(µ3-CNMeBz)(CO)6(PPh3)3 and (b) H3Ru3(µ3-
COMe)(CO)6(PPh3)3 at 100 mV/s.
The response for oxidation on scanning toward more
positive potentials is followed on the reverse scan by
chemical reaction (C) follows an electrochemical step (E)
and yields a product that is also electroactive (second
two reduction processes. These are separated by ∼100
mV. Comparison with Figure 3b, recorded at a scan
rate of 100 mV/s, suggests that the single reduction
process observed at the lower scan rate coincides with
the process at the more negative potential in Figure 4a.
The general shape of the latter response is consistent
with one of the well-known ECE systems, where a
E).
In fact, the oxidation response for H3Ru3(µ3-COMe)-
(CO)6(PPh3)3 involves a second oxidizable species. This
becomes apparent when two cycles through the potential
range are completed (Figure 4b). Now the response is
consistent with two redox processes, separated by ∼100

Ligand Additivity in Metal Clusters
Organometallics, Vol. 17, No. 5, 1998 879
Sch em e 1
grams feature charging currents that are of the same
order as those for the experimental counterparts. From
the H NMR spectra, recorded for dichloromethane-d2
1
solutions, integration of the hydride resonances provided
K1 ) 0.15. An initial value for the equilibrium constant
(K2) for the interconversion of Ax+ and Eq+ was obtained
from voltammetric studies on a solution of H3Ru3(µ3-
COMe)(CO)6(PPh3)3 that had been treated with 1 equiv
of AgBF4 (vide infra). Oxidation with Ag+ produces
solutions in which the predominant form of the cationic
species is the equatorial isomer (Eq+). An analysis of
the voltammetric responses suggested a value for [Eq+]/
[Ax+] of around 6. Ultimately, the lack of accuracy in
the estimation of K2 had no effect on the quality of the
simulation of Scheme 1 for H3Ru3(µ3-COMe)(CO)6-
(PPh3)3. Indeed, close fitting of simulated and experi-
mental data provided an accurate estimate for ∆E° (the
difference in the potentials for the Ax and Eq couples)
which, in combination with K1, fixes the value of K2, on
thermodynamic grounds.
The general applicability of the ECEC mechanism
was tested by making a series of relatively coarse
adjustments to the least-well-defined of the input
parameters. Having obtained a good qualitative cor-
respondence between the experimental and computed
voltammograms, the fitting routine of DigiSim was
utilized to refine the parameter set. The results are
shown in Figure 5, parts of which may be compared with
the corresponding voltammograms in Figure 4. An
important feature of the simulation, that is not evident
in Figures 4 and 5, is the dependence of the cyclic
voltammetric responses on the concentration of the
cluster. Comparison of the responses obtained for 1 and
10 mM solutions showed slightly less current for the
reduction of Eq+ at the higher concentration. This
observation is consistent with the effects of the “cross-
reaction” defined in Scheme 1. Second-order rate
constants for this reaction were obtained by conducting
a further series of simulations for analyte concentrations
set at 1 and 10 mM. As a result, the final parameter
set provides an accurate description of the electrochemi-
cal behavior over at least one order of magnitude in
concentration. Table 4 contains a summary of rate and
equilibrium constants derived from the refined simula-
tion.
Liga n d a n d Su bstitu en t Effects Up on Oxid a tion
P oten tia l. The relationship between the structure of
a molecule and its oxidation potential is one of the most
fundamental of all chemical concepts. Correlations of
oxidation potentials with substituent effects and with
HOMO energies have been noted by a number of
workers.24,25 Ligand additivity, the postulate that the
effect of a particular ligand upon the oxidation potential
is a function of that ligand only and is independent of
the other ligands of the complex, is a well-established
concept for substituted mononuclear complexes, for
which the HOMO usually has dπ character.24 To date,
ligand additivity in metal cluster systems has not been
mV, for which all four participating species are stable
on the time scale of the potential sweep. This situation
fits well with the spectroscopic data that show these
compounds are prone to undergo isomerization between
so-called “axial” and “equatorial” forms. Given that the
dominant form of the neutral compound in bulk dichlo-
romethane solution is the axial (Ax) isomer, the oxida-
tion-reduction pair at the more positive potential is
assigned to the couple Ax-Ax+, where Ax+ represents
the corresponding cationic species. The second redox
pair is due to the equatorial form (neutral species and
cation, Eq-Eq+).
According to the general representation for the ECEC
system (Scheme 1), the influence of the kinetic param-
eters for the two isomerization reactions (k1, k-1, k2, and
k-2) is revealed by the dependence of the cyclic volta-
mmetric responses on scan rate. As shown in Figure
4c, increasing the scan rate (to 5 V/s) causes a change
in the response as the current due to the reduction
process for Eq+ becomes smaller. At 100 V/s (Figure
4d), the cyclic voltammogram consists almost exclusively
of responses for the Ax-Ax+ couple. (In fact, a response
for the Eq couple would be expected at this scan rate,
due to the small amount of this isomer present in the
bulk solution. As shown below, however, broadening
of the Eq response (due to the encroachment of electro-
chemical kinetic effects) causes this response to disap-
pear into the large background (charging) current.)
Further characterization of the kinetic parameters for
Scheme 1 was undertaken through simulation of the
voltammetric responses with DigiSim (v. 2.1). Data for
H3Ru3(µ3-COMe)(CO)6(PPh3)3 were used as the basis for
the simulation.
Estimates of the values for the input parameters were
derived from the electrochemical and spectroscopic
studies. As noted above, careful choice of the electrode
size and instrument settings allowed effectively com-
plete compensation for the effects due to solution
resistance. Consideration of the latter, along with the
cell time constant, also provided a good estimate for the
double layer capacitance. Hence, simulated voltammo-
(23) Paw, W.; Lake, C. H.; Churchill, M. R.; Keister, J . B. Organo-
metallics 1995, 14, 3768.
(24) Bursten, B. E.; Green, M. R. Prog. Inorg. Chem. 1988, 36, 393
and references therein.
(25) Lever, A. B. P. Inorg. Chem. 1990, 29, 1271.
(26) Alexiou, A. D. P.; Toma, H. E. J . Chem. Res., Synop. 1993, 464.
(27) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. (Washington, D.C.)
1991, 91, 165.

880 Organometallics, Vol. 17, No. 5, 1998
Feighery et al.
F igu r e 4. Cyclic voltammograms for 10-3 M H3Ru3(µ3-COMe)(CO)6(PPh3)3 at scan rates of (a) 1 V/s; (b) 1 V/s, s first
cycle, - - - second cycle; (c) 5 V/s; (d) 100 V/s.
investigated in a systematic manner. The electrochem-
istry of the series (µ-H)3Ru3(µ3-CX)(CO)9-n(L)n provides
valuable insight into ligand additivity and substituent
effects in cluster systems. Here, we adopt the terminol-
ogy that “ligand effects” refer to effects due to the steric
and electronic properties of the ligand directly bonded
to a metal center (e.g., CO vs PPh3), whereas “substitu-
ent effects” are effects due to the steric and electronic
properties of substituents of organic moieties remote
from the metal center (e.g., substituents on the hydro-
carbyl fragment).
A variety of parameter sets have been developed to
correlate the various properties of the metal complexes
with the steric and electronic properties of ligands such
as tertiary phosphines. Lever has established a set of
ligand parameters EL based upon an extensive compila-
tion of oxidation potentials for octahedral Ru(II)/Ru(III)
mononuclear complexes.25 This ligand parameter set
allows for the prediction of E1/2 for any given octahedral
mononuclear Ru(II) complex. The most general expres-
sion for the oxidation potentials for octahedral com-
plexes of other metals is given in eq 1, where IM and
SM are characteristic for the metal couple.
Linear free-energy relationships involving substituent
effects are used extensively in organic and inorganic
chemistry.27 Many correlations involving σp or σp+ have
been reported. The former parameter is based upon the
pKa of para-substituted benzoic acids and does not
involve conjugative interactions. The latter set is
derived from rates of solvolysis of substituted cumyl
chlorides and, therefore, reflects direct conjugation
between the substituent and an electron-deficient cen-
ter. One recent example is the correlation of the
oxidation potentials of (η6-C6H5X)Cr(CO)3 with the
substituent parameters.28 The oxidation potentials of
Fe(η5-C5H4X)Cp have also been subjected to analysis;29
although the best correlation involves σP for most
substituents, strong π donors appear to correlate better
with the σp+ values. Lever has correlated ligand-based
redox potentials with σp values.30
It is enlightening to correlate the oxidation potentials
for (µ-H)3Ru3(µ3-CX)(CO)9-n(L)n with the ligand param-
eters EL and the substituent parameters for the hydro-
carbyl ligand. Electrochemical one-electron oxidation
of (µ-H)3Ru3(µ3-CX)(CO)9-n(L)n (n ) 2 or 3) is a kineti-
cally facile process. As the oxidations of the monosub-
stituted clusters are irreversible, we have used for
comparison the anodic peak potentials as determined
by cyclic voltammetry at 100 mV/s. It should be noted
that (Ep,a - E1/2) will increase as the couple becomes
more irreversible, and therefore, comparisons of data
for the quasi-reversible and irreversible examples should
be viewed with caution. Also, we have a very limited
variety of ligand and substituent types. Nonetheless,
i
Eobsd(V) ) S [ a E (L )] + I
M
(1)
∑
M
i
L
i
0
One report of the application of the Lever parameters
to a cluster has appeared. The oxidation potential of
[Ru3(µ3-O)(OAc)6L3]n (n ) -1 to +2) was studied by
Toma.26 This compound, which is a 51-e cluster in the
+1 state, differs from the clusters described below
because the HOMO is not associated with the organo-
metallic cluster bonding framework (indeed there are
no Ru-Ru bonds).
(28) Hunter, A. D.; Mozol, V.; Tsai, S. D. Organometallics 1992, 11,
2251.
(29) Britton, W. E.; Kashyap, R.; El-Hashash, M.; El-Kady, M.;
Herberhold, M. Organometallics 1986, 5, 1029 and references therein.

Ligand Additivity in Metal Clusters
Organometallics, Vol. 17, No. 5, 1998 881
case the single methylidyne substituent. The constant
IM3 serves to scale the potential to a hypothetical
complex H3Ru3(µ3-CH)L9 where EL(L) ) 0 and corrects
for differences in reference potentials (Lever’s EL pa-
rameters are based upon the NHE reference, whereas
we are using ferrocene/ferricenium ) 0 V).
The coefficient F represents the sensitivity of the
oxidation potential to the alkylidyne substituent. The
π interaction is manifested in a very high value. The
28
oxidation potentials of Cr(η6-C6H5X)(CO)3 and Fe(η5-
C5H4X)Cp29 have been previously correlated with σI and
σp, respectively. The nonlinear least-squares fits of the
published oxidation potentials for Cr(η6-C6H5X)(CO)3
and Fe(η5-C5H4X)Cp to eq 2 (with SM3 ) 0) with ring
substituent constants σp+ yield F ) 3.8(0.3) (correlation
0.981) and 5.2(0.4) (correlation 0.975), respectively.
Thus, the oxidation potential of the alkylidyne cluster
core is even more sensitive to π-donor substituent effects
than are the aromatic π complexes to substituents on
the rings.
The bonding of (µ-H)3Ru3(µ3-CX)(CO)9 (X ) H, Cl, and
Br) has been the subject of a previous study. Fenske-
Hall MO calculations performed by Sherwood and Hall9
indicate that the HOMO for H3Ru3(CX)(CO)9 (X ) H,
Cl, Br) is involved primarily in the Ru-C bonding
interaction. The methylidyne carbon is best described
as sp-hybridized with delocalized metal-carbon bonds,
consisting of a weak bond involving the carbon “lone
pair”, and a stronger interaction with the carbon 2p
orbitals. The HOMO for X ) H is described as Ru-C
“lone pair” antibonding. A pair of degenerate bonding
orbitals which lie somewhat lower in energy is described
as being formed from the Ru3 {eg*} and C 2pπ orbitals.
However, for the π-donor substituents Cl and Br, the
delocalization of the halogen lone pairs into the degen-
erate set makes these orbitals the HOMOs. For X )
Ph, OMe, SEt, and NR2, the lower symmetry will split
the degeneracy but the description of the HOMO will
be as for the halogen derivatives. When the methyli-
dyne substituent is a strong π-donor group such as MeO,
the electron pairs are strongly delocalized onto the
cluster. Thus, it is not surprising that the oxidation
potential of the cluster is strongly dependent upon the
π-donor properties of X, in addition to the degree of
phosphine substitution on the metal atoms.
F igu r e 5. Computed cyclic voltammograms (DigiSim v.
2.1) for 10-3 M H3Ru3(µ3-COMe)(CO)6(PPh3)3 at scan rates
of (a) 1 V/s; (b) 5 V/s; (c) 100 V/s.
The first term in eq 2 represents the effect of ligand
substitution for CO upon the oxidation potential of the
cluster. The coefficient S for a mononuclear Ru(II)/
Ru(III) couple is ideally 1. The linearity of the fit, which
includes clusters differing in the number of metal atoms
which are substituted, indicates that each metal atom
contributes equally to the HOMO. As for monometallic
complexes, ligand additivity is conserved.24 The value
of SM3 ) 0.37 indicates that delocalization over three
Ru atoms reduces the dependence of the oxidation
potential upon a ligand on any given Ru atom.
a good fit (SM3 ) 0.37(0.03), F ) 6.0(0.7), and IM3
)
-2.5(0.2) V (correlation 0.962, Figure 6) of the observed
anodic peak potentials (at 100 mV/s) can be made to eq
2.
9
Eobsd(V) ) SM E (L ) + 2.303(RT/nF)Fσ+p(X) + IM
∑
3i)1
L
i
3
(2)
Equation 2 is a modified version of Lever’s25 with
inclusion of a Hammet term.27 The correlation with the
Substitution of each phosphine or arsine ligand for a
carbonyl decreases the oxidation potential by approxi-
mately 150-200 mV per ligand. Surprisingly, there is
a very significant difference in the donor properties of
EPh3, E ) P > As g Sb. Lever’s EL parameters are
essentially the same for these (E ) P (0.39), As (0.38),
and Sb(0.38)). Other parameter sets also suggest
similar electronic properties for these ligands.31
+
substituent parameters σp is much poorer than with σp
,
indicating that π conjugation is important in stabilizing
the radical cation. The factor 2.303(RT/nF) is included
to provide units of volts. Here, the ligand parameters
EL are summed over the CO or ER3 ligands of the
triruthenium core, and the substituent parameters are
summed over the n hydrocarbyl substituents Xi, in this

882 Organometallics, Vol. 17, No. 5, 1998
Ta ble 4. Su m m a r y of P a r a m eter s for Sim u la tion of Volta m m ogr a m s
Feighery et al.
kinetics
thermodynamics
electron transfer
E°′
R
ks
Ax+ + e- h Ax
Eq+ + e- h Eq
+0.172 V (E1°)
+0.267 V (E2°)
0.5
0.5
0.15 cm s-1
0.10 cm s-1
kinetics
reaction
thermodynamics
isomerization
K
kf
kb
Ax h Eq
0.15 (K1)
6.44 (K2)
43.0 (K3)
2.0 × 10-4 s-1
2 × 105 M-1 s-1
1.3 × 10-3 s-1
4.7 × 103 M-1 s-1
Ax+ h Eq+
6.5 s-1
1.0 s-1
Ax+ + Eq h Ax + Eq+
The first was an electrochemical experiment. A cyclic
voltammogram of H3Ru3(COMe)(CO)6(PPh3)3 was re-
corded, then Ag+ was added to the solution in the
amounts of 0.25, 0.5, 0.75, and 1 equiv, each time
recording the voltammogram. In each case, the volta-
mmogram was unchanged, despite the fact that the
solution had turned from orange to green. The second
experiment involved oxidation of a solution of the
cluster, followed by chemical reduction and a determi-
nation of the starting material recovered. A solution
of H3Ru3(COMe)(CO)6(PPh3)3 was oxidized with Ag+.
Dropwise addition of a THF solution of Na/benzophe-
none ketal caused the solution to turn from green to
orange. TLC of this mixture and work up produced 93%
of the starting material used. Thus, the integrity of the
cluster is retained after oxidation.
F igu r e 6. Correlation of experimental anodic peak poten-
tials (at 100 mV/s), experimental Ep,a, vs potentials calcu-
lated, calculate Ep,a, from eq 2, with SM3 ) 0.37, F ) 6.0,
and IM3 ) -2.5 V.
1
The H NMR spectrum for the green solution formed
upon addition of 1 equiv of AgSO3CF3 or FeCp2PF6 to a
solution of H3Ru3(µ3-COMe)(CO)6(PPh3)3 displays broad
resonances in the methyl (4 ppm) and hydride (-15
ppm) regions of the spectrum, which are indicative of
the formation of a paramagnetic species, proposed to be
the 47-electron radical cation [H3Ru3(µ3-COMe)(CO)6-
(PPh3)3]1+. A broad resonance is also observed in the
31P NMR spectrum of this compound, centered at 20.9
ppm. Even when 0.5 equiv of oxidant is added, only
these broad resonances are observed, suggesting rapid
electron transfer between 47- and 48-electron clusters.
In general, the same observations were made for all of
the di- and trisubstituted clusters studied. However,
because the stabilities of the oxidized clusters vary
greatly, data could not be obtained for all of the
47-electron clusters. For each, the IR spectrum (Table
2) displays higher CO stretching frequencies than its
48-electron precursor, as expected for positively charged
clusters. The IR spectrum for the oxidation product of
H3Ru3(CNMeBz)(CO)6(PPh3)3 displays the same pattern
as the neutral precursor, suggesting a similar symmetry
with three axial PPh3 ligands. However, spectra for the
oxidation products of H3Ru3(COMe)(CO)6(EPh3)3 (E )
As, P) display a higher number of bands, consistent with
lower symmetry due to the presence of one equatorial
There appears to be only a small dependence of the
oxidation potential upon the geometry of the substitu-
tion product. The oxidation potential for H3Ru3(COMe)-
(CO)6(ax-PPh3)2(eq-PPh3) is 95 mV less positive than
that of H3Ru3(COMe)(CO)6(ax-PPh3)3, and only one
oxidation wave is found for H3Ru3(COMe)(CO)6(ax-
PPh3)2(P(OMe)3), which exists as a 3:1 mixture of
equatorially and axially substituted isomers.
On e-Electr on Oxid a tion s Usin g Ch em ica l Re-
a gen ts. As was the case with electrochemical oxidation,
chemical oxidations in acetonitrile did not produce
characterizable cluster products. However, in dichlo-
romethane solution, a variety of oxidants, including
AgSO3CF3, AgBF4, [FeCp2]PF6, and [Fe(o-phen)3]3+
,
caused the color of the solution of the di- or trisubsti-
tuted cluster to change immediately from red-orange to
dark green. The addition of Ag+ to solutions of the
cluster also leads to the formation of a silver mirror;
exceptions are H3Ru3(µ3-CR)(CO)6(PPh3)3 (R ) Me, Ph)
which undergo a parallel one-electron oxidation and
formation of adducts with Ag+. The nature of these
adducts will be presented in a later paper. The oxida-
tion products are proposed to be the 47-electron clusters
[H3Ru3(CX)(CO)6(PPh3)3]1+; on the basis of the IR and
EPR data (vide infra), these clusters are proposed to be
isostructural with the 48-electron precursor, with all-
axial PPh3 ligands for X ) Ph and NMeBz (C3v sym-
metry) but one equatorial PPh3 ligand and two axial
PPh3 ligands for X ) OMe or SEt (C1 symmetry).
To verify that the clusters remain intact upon chemi-
cal oxidation, two separate experiments were performed.
1
and two axial PPh3 ligands. IR and H NMR data for
the oxidized clusters are given in Tables 1 and 2.
The green solutions of the oxidized clusters were
found to be EPR-active. Figure 7 shows the EPR
spectrum for the radical cation [H3Ru3(CNMeBz)(CO)6-
(PPh3)3]1+, and Figure 8 shows the spectrum of [H3Ru3-
(COMe)(CO)6(PPh3)3]1+. The EPR data for each of the
clusters are summarized in Table 5. A qualitative
assessment of the stability of the radical cations can be
made from the observation of the persistence of the
green solution and the decay of the EPR signal with
(30) Masui, H.; Lever, A. B. P. Inorg. Chem. 1993, 32, 2199.
(31) (a) Brown, T. L.; Lee, K. J . Coord. Chem. Rev. 1993, 128, 89.
(b) Tolman, C. A. Chem. Rev. (Washington, D.C.) 1977, 77, 313.

Ligand Additivity in Metal Clusters
Organometallics, Vol. 17, No. 5, 1998 883
with large hyperfine couplings for trans phosphines and
unobservable couplings for phosphines cis to the Ru-C
bonds (Scheme 1).
The lack of resolvable hyperfine coupling for the
equatorial phosphine ligands is confirmed by the EPR
spectra of the oxidation products from treatment of H3-
Ru3(µ3-COMe)(CO)6{µ-dppm}(PPh3) (dppm ) (PPh2)2-
CH2) with ferricenium. The dppm ligand bridges equa-
torial sites on two different Ru atoms, whereas the PPh3
ligand is axially coordinated. The EPR spectrum for the
species proposed to be [H3Ru3(µ3-COMe)(CO)6{µ-dppm}-
(PPh3)]1+ is a doublet (a(31P) 32 mT); no hyperfine
coupling to the equatorial 31P nuclei is observed.32
Another result which is consistent with axial to
equatorial ligand isomerization as the explanation for
the low symmetry of some of the 47-electron clusters
comes from the oxidation of H3Ru3(COMe)(CO)6[(ax-
PPh2CH2)3CCH3].10b This cluster should have a similar
oxidation potential to that of H3Ru3(COMe)(CO)6(PPh3)3,
with the exception that ligand isomerization is not
possible. Oxidation with Ag+ in this case was shown
to produce a very unstable radical cation, which displays
a quartet EPR signal. Thus, ligand isomerization from
an axial to equatorial site is the most likely explanation
for the low symmetry of the radical cations [H3Ru3-
(COMe)(CO)6(PPh3)3]+ and [H3Ru3(CSEt)(CO)6(PPh3)3]+.
F igu r e 7. EPR spectrum for [H3Ru3(µ3-CNMeBz)(CO)6-
(PPh3)3]1+ in dichloromethane at 20 °C.
Ra d ica l Ca tion Str u ctu r e a n d Isom er iza tion .
The spectroscopic data indicate that the structures of
radical cations [H3Ru3(CX)(CO)6(PPh3)3]1+ (X ) OMe,
SEt) differ from those of the 48-electron precursors,
whereas [H3Ru3(CX)(CO)6(PPh3)3]1+ (X ) NMeBz, Ph)
retain the all-axially substituted structure. The all-
axial substitution adopted by H3Ru3(CX)(CO)6(PPh3)3 is
apparently due to steric factors. Smaller ligands such
as pyridine, acetonitrile, and isocyanides preferentially
occupy equatorial positions, and in fact, the kinetic
product H3Ru3(COMe)(CO)6(ax-PPh3)2(ax-CNBz) was
shown to rearrange to the more stable H3Ru3(COMe)-
(CO)6(ax-PPh3)2(eq-CNBz) within several hours (k ) 8.8
× 10-5 s-1 at 19 °C). The isomerization of the radical
cations indicates that the electronic preference for
equatorial substitution is of greater importance relative
to the steric preference for axial substitution. The all-
axial structure is retained by [H3Ru3(CX)(CO)6(PPh3)3]1+
for the larger methylidyne substituents X ) NMeBz and
Ph. The rate of the isomerization is 4-5 orders of
magnitude faster for the 47-electron species than for the
48-electron cluster. The rate constant for isomerization
of [H3Ru3(COMe)(CO)6(ax-PPh3)3]1+ to the more stable
[H3Ru3(COMe)(CO)6(ax-PPh3)2(eq-PPh3)]1+ is 6.5 s-1 at
19 °C.
F igu r e 8. EPR spectrum for [H3Ru3(µ3-COMe)(CO)6-
(PPh3)3]1+ in dichloromethane at 20 °C.
time. The stability at room temperature follows the
order NMeBz (hours) > OMe > SEt > Ph (seconds). The
stability appears to be related to the oxidation potential,
the lower the oxidation potential, the more stable the
radical cation.
The EPR spectra of [H3Ru3(µ3-CPh)(CO)6(PPh3)3]1+
,
[H3Ru3(µ3-CNMeBz)(CO)6(PPh3)3]1+, and [H3Ru3(µ3-
COMe)(CO)7(PPh3)2]1+ each show that the unpaired
electron is coupled equivalently to each of the phospho-
rus atoms in the cluster. This suggests that no signifi-
cant changes in the cluster geometry have occurred
upon oxidation. In contrast, the EPR spectra of [H3-
Ru3(µ3-CX)(CO)6(PPh3)3]1+ (X ) OMe or SEt) each show
that the unpaired electron is coupled inequivalently to
only two of the three phosphorus atoms in the molecule.
EPR signals for [H3Ru3(µ3-COMe)(CO)6{P(OMe)3}-
(PPh3)2]1+ and [H3Ru3(µ3-COMe)(CO)6(CNBz)(PPh3)2]1+
appear as triplets, most likely because of coincidentally
equal hyperfine coupling constants to the nonequivalent
axial PPh3 ligands; there is no evidence for the presence
of two isomers in either spectrum, and the center line
of the triplet for [H3Ru3(µ3-COMe)(CO)6{P(OMe)3}-
(PPh3)2]1+ is only slightly broader than that for [H3Ru3-
(µ3-COMe)(CO)6(CNBz)(PPh3)2]1+. These results can be
explained in terms of a ligand isomerization from an
axial site to an equatorial site following oxidation,
reducing the symmetry of the cluster from C3v to C1,
Rapid isomerizations for 17-electron complexes are
well-known, but rarely can quantitative comparisons of
18- and 17-electron species be determined. Very re-
cently, the relative rates for interconversion of isomers
of Cp*IrCp2Co2(CO)3, involving bridging and terminal
CO ligands, were determined to be 48-electron (1) , 49-
electron (104) , 47-electron (108) species.33
Decom p osition of th e Ra d ica l Ca tion s. Slow
decomposition of [H3Ru3(COMe)(CO)6(PPh3)3]1+ occurs
in solution under nitrogen. Overnight, the green solu-
tion gradually turns red in color. The major cluster
products are starting material and a diamagnetic mate-
rial which we propose to be [H3Ru3(CO)7(PPh3)3]1+, a

884 Organometallics, Vol. 17, No. 5, 1998
Ta ble 5. EP R Sp ectr a l Da ta fr om Dich lor om eth a n e Solu tion
Feighery et al.
cluster
T (K)
g
a(31P), mT
[H3Ru3(COMe)(CO)6(PPh3)3]+
[H3Ru3(CSEt)(CO)6(PPh3)3]+
295
295
295
225
295
295
295
295
2.08 (dd)
2.06 (dd)
2.06 (q)
2.09 (q)
2.09 (t)
2.09 (t)
2.08 (t)
2.19 (br)
4.0, 6.0
3.5, 4.8
3.6
2.6
4.8
[H3Ru3(CNMeBz)(CO)6(PPh3)3]+
[ H3Ru3(CPh)(CO)6(PPh3)3]+
[H3Ru3(COMe)(CO)6(P(OMe)3)(PPh3)2]+
[H3Ru3(COMe)(CO)6(CNBz)(PPh3)2]+
[H3Ru3(COMe)(CO)7(PPh3)2]+
[H3Os3(COMe)(CO)7(PPh3)2]+
4.9
5.1
[H3Ru3(COMe)(CO)6(PPh3)3]1+ + PPh3 f
H3Ru3(CO)7(PPh3)3 + MePPh3 (3)
1+
H3Ru3(CO)7(PPh3)3 +
[H3Ru3(COMe)(CO)6(PPh3)3]1+
[H3Ru3(CO)7(PPh3)3]1+
H3Ru3(COMe)(CO)6(PPh3)3 (4)
f
+
F igu r e 9. Possible structures for [H3Os3(CO)10]1+; struc-
ture b is the proposed structure.
of product solutions from decomposition of [H3Ru3-
(COMe)(CO)6(PPh3)3]1+ contain a doublet at 2.80 ppm
(J 13 Hz), attributed to [MePPh3]1+ in a 1:1 molar ratio
to [H3Ru3(CO)7(PPh3)3]1+ and substantial amounts of
H3Ru3(COMe)(CO)6(PPh3)3 are also present when the
decomposition proceeds in the absence of added CO.
No cluster products were identified from the decom-
position of the trisubstituted NMeBz or Ph analogues.
However, for H3Ru3(COMe)(CO)8(PPh3) and H3Ru3-
(COMe)(CO)7(PPh3)2, decomposition regenerates start-
ing material with low recovery. We were unable to
isolate pure materials due to decomposition during
purification by TLC or recrystallization.
F igu r e 10. Proposed structure for [H3Os3(CO)9(PPh3)]1+
.
Oxidations of the osmium alkylidyne clusters H3Os3-
(COMe)(CO)9-n(PPh3)n (n ) 0-2) were examined to
confirm the nature of the decomposition reaction of the
radical cations [H3Ru3(COMe)(CO)6L3]1+. The Os clus-
ters show similar electrochemical behavior to their Ru
analogues (Table 3), with the parent carbonyl and
monosubstituted clusters showing an irreversible oxida-
tion wave and the disubstituted cluster displays a quasi-
reversible wave followed by an irreversible wave. Oxi-
dation of H3Os3(COMe)(CO)7(PPh3)2 with Ag+ produces
a green solution, which displays a singlet EPR signal
(Table 5). No hyperfine coupling to phosphorus is
observed, but the line width of the signal is quite broad.
Addition of 1 equiv of Ag+ to a dichloromethane solution
of H3Os3(COMe)(CO)8(PPh3) caused the yellow solution
to turn fleetingly green before turning yellow again. 1H
NMR analysis of this solution indicated the presence of
[H3Os3(CO)9(PPh3)]1+, as well as starting material and
other unidentified products. Treatment of the solution
with methanol allowed for the recovery of H2Os3(CO)9-
(PPh3) in 34% yield after TLC. This indicates that the
net loss of the methyl radical, leading to the formation
of [H3M3(CO)10-nLn]1+ (M ) Ru, L ) PPh3, AsPh3, n )
3; M ) Os, L ) PPh3, n ) 1), is indeed a pathway for
radical cation decomposition.
F igu r e 11. Proposed structure for [H3Ru3(CO)7(PPh3)3]1+
.
new 46-electron cluster. Although no Ru clusters of this
or related compositions had been previously reported,
the Os analogues [H3Os3(CO)10-nLn]1+ (L ) PPh3; n )
0, 1) have been prepared by protonation of unsaturated
H2Os3(CO)10-nLn.19 This formulation is based upon the
1H and 31P NMR data and comparison to the analogous
cation [H3Os3(CO)9(PR3)]1+
. Decompositions of [H3-
Ru3(CSEt)(CO)6(PPh3)3]1+ and [H3Ru3(COMe)(CO)6-
(AsPh3)3]1+ form analogous products with similar
hydride NMR resonances. The chemistries of [H3Ru3-
(CSEt)(CO)6(PPh3)3]1+ and its decomposition product are
under investigation and will be the subject of a later
paper. The NMR data are in Table 1. On the basis of
the spectroscopic characterization of [H3Os3(CO)9-
(PPh3)]1+ (vide infra), the proposed structure of [H3Ru3-
(CO)7L3]1+ (L ) PPh3, AsPh3) is shown in Figure 11.
Attempts to purify these clusters, or their deprotonated
forms, proved unsuccessful, due to their instability.
These clusters are formally unsaturated, having 46-
valence electrons, a factor which may account for their
instability. Carbon monoxide reacts with [H3Ru3(CO)7-
(PPh3)3]1+ within minutes. The red product displays an
IR spectrum identical to the product of CO with the 47-
electron [H3Ru3(COMe)(CO)6(PPh3)3]1+, which is char-
The likely mechanism by which the radical cations
decompose is attack by PPh3, released by decomposition,
at the methyl group (eqs 3 and 4). The 1H NMR spectra

Ligand Additivity in Metal Clusters
Organometallics, Vol. 17, No. 5, 1998 885
acterized as [HRu3(CO)9(PPh3)3]1+ (vide infra), a 48-
electron cluster.
P r ot on a t ion of H2Os3(CO)10 a n d H2Os3(CO)9-
(P P h 3). Because we were unable to purify [H3Ru3-
(CO)7L3]1+, we could obtain limited information con-
cerning the structure. The alternative synthetic pathway
available for the osmium analogues provided a means
of obtaining complete spectroscopic characterization.
We, therefore, reinvestigated the structures of [H3-
F igu r e 12. Proposed structure for [HRu3(CO)9(PPh3)3]1+
.
1
Os3(CO)9(PPh3)]1+ and [H3Os3(CO)10]1+ by H, 13C, and
in dichloromethane at 100 mV/s was unaffected by the
addition of 1-10 equiv of methyl iodide, benzyl bromide,
acetonitrile, alkynes, or pyridine to the electrochemical
cell. This suggested that the radical cation cluster was
too stable to undergo ligand substitution, atom transfer,
or other reactions on the time scale of the cyclic
voltammetric experiment.
31P NMR spectroscopy. J ohnson, Lewis, and co-workers
had originally proposed two possible structures for
[H3Os3(CO)10-nLn]1+ (Figure 9, structures a and b) which
1
were consistent with the available H NMR data (Ex-
perimental Section).19 These workers preferred the
structure shown in Figure 9 a. However, the 13C NMR
data (Experimental Section) which we obtained are
consistent with the structure shown in Figure 9b, not
Figure 9a. For structure a, no more than four carbonyl
resonances are expected from the Cs symmetry. How-
ever, the observed 13C NMR spectrum displays seven
carbonyl resonances, as expected for the C1 symmetry
of structure b. Tentative assignments of the NMR
resonances in the Experimental Section are according
to Figure 9b.
Since [H3Ru3(COMe)(CO)6(PPh3)3]1+ has a reasonable
lifetime, we were able to examine its reactions with
various reagents. Addition of chloride, acetonitrile, or
pyridine to solutions of the 47-electron species, gener-
ated by oxidation with ferricenium, caused regeneration
of the 48-electron precursor in yields of up to 60%.
Disproportionation induced by donor ligands is a com-
mon reaction of 17-electron species. Slow decomposition
in the absence of added Lewis bases gives the 46-
electron species [H3Ru3(CO)7(PPh3)3]1+, in addition to
MePPh31+ and varying amounts of H3Ru3(COMe)(CO)6-
(PPh3)3. Addition of CO at 1 atm causes rapid (<15
min) decomposition to [HRu3(CO)9(PPh3)3]1+, and vary-
ing amounts of H3Ru3(COMe)(CO)7(PPh3)2, H3Ru3-
(COMe)(CO)8(PPh3), and MePPh31+, were identified in
the 1H NMR spectrum of the product mixture. After
workup, red-brown crystals of [HRu3(CO)9(PPh3)3]-
[BPh4] were isolated in 30% yield (based upon the initial
amount of H3Ru3(COMe)(CO)6(PPh3)3), along with 23%
recovery of the latter. The disappearance of the radical
cation under CO is faster than its decomposition in the
absence of CO or in the presence of 1 equiv of PPh3.
Characterization of the product [HRu3(CO)9(PPh3)3]-
Further evidence for this structure is obtained from
the spectra of [H3Os3(CO)9(PPh3)]1+. The structure of
H2Os3(CO)9(PPh3) has Cs symmetry, with the phosphine
ligand occupying an equatorial site on a hydride-bridged
metal atom; the 13C NMR spectrum contains the ex-
pected 2:2:2:1:1:1 pattern.34 Protonation of this cluster
to form the analogue to Figure 9, structure a or b, would
not change the symmetry, and a similar pattern would
be expected. However, the low symmetry evident in the
13C NMR spectra is indicative of the structure in Figure
10. This structure is identical to the one adopted by
the isoelectronic cluster H3Os3 (CO)9(SiPh3), for which
the structure has been determined crystallographi-
cally.35 The carbonyl assignments in the Experimental
Section and Figure 10 are made using both the coupled
and decoupled spectra, assuming that trans couplings
to both phosphorus and hydrogen are large and cis
1
[BPh4] is based upon its IR spectrum, H and 31P NMR
spectra, mass spectrum, and elemental analysis. The
mass spectrum displays a low intensity ion at m/e 1344,
with a higher intensity ion at m/e 1316; computer fitting
of the mass envelope of the latter is consistent with the
composition [HRu3(CO)8(PPh3)3]1+ and suggests [HRu3-
(CO)9(PPh3)3]1+ as the composition of the molecular ion.
A structure which is consistent with the spectroscopic
data is shown in Figure 12, analogous to structures of
previously characterized [HOs3(CO)9(PR3)3]1+, which can
be prepared by protonation of Os3(CO)9(PR3)3.36 The
solution displays three 31P signals, with one large 31P-
31P coupling constant, which is consistent with the
presence of two PPh3 ligands trans to one another across
a Ru-Ru bond. The 1H NMR spectrum contains a
single hydride resonance at -17.05 (dt, J PH 3, 6 Hz)
ppm, consistent with two PR3 ligands cis to a bridging
hydride and a third further removed; the 1H NMR signal
is similar to that reported for [HOs3(CO)9(PR3)3]1+, e.g.
PMe2Ph -18.93 (t, J PH 9.7 Hz).36 The IR spectrum for
[HRu3(CO)9(PPh3)3]1+ (IR (CH2Cl2) 2096 vw, 2064 m,
1
couplings are small. The H NMR spectrum of [H3Os3-
(CO)9(PPh3)]1+ displays three hydride resonances, two
of which have phosphorus coupling constants charac-
teristic of a cis P-H relationship (7.3 and 9.3 Hz) and
the third which has an unusually large coupling con-
stant (30.5 Hz), indicative of a trans phosphorus-
hydride relationship required by the structure in Figure
10.
Ch em ica l Rea ctivity of th e Ra d ica l Ca tion Clu s-
ter [H3Ru 3(COMe)(CO)6(P P h 3)3]+. An examination
of the reactivities for the radical cluster [H3Ru3(COMe)-
(CO)6(PPh3)3]+ with various reagents on the electro-
chemical time scale did not afford much information.
The cyclic voltammogram of H3Ru3(COMe)(CO)6(PPh3)3
(32) Bierdeman, D. J .; Kourouklis, D.; Keister, J . B. Unpublished
results.
(33) Geiger, W. E.; Shaw, M. J .; Wu¨nsch, M.; Barnes, C. E.;
Foersterling, F. H. J . Am. Chem. Soc. 1997, 119, 2804.
(34) 1H and 13C NMR data for H2Os3(CO)9(PPh3) in deuteriochloro-
form solution: 13C NMR 176.8 (2 C, d, J CP 13.5 Hz), 177.3 (1 C), 181.6
(2 C, d, J CP 11.0 Hz), 183.2 (1 C), 184.1 (1 C), 185.4 (2 C) ppm; 1H
NMR -10.3 ppm (d, J PH 7.2 Hz).
(36) (a) Bradford, C. W.; Nyholm, R. S. J . Chem. Soc., Dalton Trans.
1973, 529. (b) Deeming, A. J .; J ohnson, B. F. G.; Lewis, J . J . Chem.
Soc. A 1970, 2967. (c) Deeming, A. J .; Donovan-Mtunzi, S.; Kabir, S.
E.; Hursthouse, M. B.; Abdul Malik, K. M.; Walker, N. P. C. J . Chem.
Soc., Dalton Trans. 1987, 1869.
(35) Willis, A. C.; Einstein, F. W. B.; Ramadan, R. M.; Pomeroy, R.
K. Organometallics 1983, 2, 935.

886 Organometallics, Vol. 17, No. 5, 1998
Feighery et al.
2026 vs, 2011 sh, 1992 vw, 1967 w cm-1) is almost
identical with that reported for [HOs3(CO)9(PPh3)3]Br
(IR(CDCl3) 2097 w, 2064 m, 2021 vs, 2007 sh, 1985 w,
1965 w cm-1).36a One caveat: we have been unable to
prepare the equivalent material by protonation of
Ru3(CO)9(PPh3)3 with CF3CO2H, CH3SO3H, or CF3SO3H
in dichloromethane or chloroform solution, rather de-
composition to a yellow solution, containing no hydride
resonance, occurs upon mixing. Dissolution in sulfuric
acid, followed by addition of aqueous ammonium hex-
fluorophosphate, produced a very small amount of a red
solid having an IR spectrum similar to that attributed
to [HRu3(CO)9(PPh3)3]1+, but the amount was too small
to allow purification and complete characterization.
X ) OMe rather than at the metal centers. These 47-
electron alkylidyne clusters are also less reactive than
other 47-electron clusters in which the unpaired electron
is in a metal-metal bonding orbital, such as HRu3(µ3-
η3-XCCRCR′)(CO)9-nLn37 or H2Ru3(µ3-η2-XCCR)(CO)9-n
-
Ln.23 The possible explanations for the low reactivity
are that (1) the unpaired electron is delocalized over the
entire Ru3(CX) framework, stabilizing the radical, cf. a
17-electron monometallic complex, (2) the unpaired
electron is in a bonding orbital which is inaccessible to
external reagents, and (3) the Ru-C bond strength is
significantly greater than the Ru-Ru bond strength,
stabilizing the radical, cf. 47-electron clusters in which
the SOMO has metal-metal bond character.
Con clu sion
Ack n ow led gm en t. This work was funded by a
grant from the National Science Foundation (Grant No.
CHE9213695) to J .B.K.
The radical cations [H3Ru3(CX)(CO)6(PPh3)3]1+ have
been prepared by chemical or electrochemical oxidations
of the appropriate 48-electron, neutral precursor. These
47-electron clusters are surprisingly unreactive. Mono-
metallic 17-electron complexes undergo extremely rapid
halogen-atom abstraction and associative ligand sub-
stitutions.1 These clusters do not react with the halogen-
atom donors chloroform and carbon tetrachloride. Hard
donors such as chloride ion, pyridine, and acetonitrile
induce disproportionation, regenerating the 48-electron
starting material in fair yield. Reactions with triph-
enylphosphine occur by attack at the methyl group for
Su p p or tin g In for m a tion Ava ila ble: Text giving the
syntheses of H3Ru3(COMe)(CO)9-n(PPh3)n (n ) 1-3) and H3-
Ru3(CPh)(CO)6(PPh3)3 and tables of 1H and 31P NMR data
(Table 1S), IR data (Table 2S), and data used for ligand
additivity fitting (Table 3S) (5 pages). Ordering information
is given on any current masthead page.
OM970769T
(37) Yao, H.; McCargar, R. D.; Allendoerfer, R. D.; Keister, J . B.
Organometallics 1993, 12, 4283.