Ortho-metalation dynamics and ligand fluxionality in the conversion of Os3(CO)10(dppm) to HOs3(CO)8[μ- PhP(C6H4-μ2,ν1)CH 2PPh2]: Experimental and DFT evidence for the participation of agostic C-H and π-aryl intermediates at an intact triosmium cluster

Article abstract of DOI:10.1021/om100475v

The mechanism for the stepwise conversion of Os₃(CO) ₁₀(dppm) to HOs₃(CO)₉[μ-PhP(C ₆H₄)CH₂PPh₂] and HOs ₃(CO)₈[μ-PhP(C₆H₄- μ₂,ν¹)CH₂PPh₂] has been investigated. The octacarbonyl cluster HOs₃(CO)₈[μ- PhP(C₆H₄-μ₂,ν¹)CH ₂PPh₂] is in rapid equilibrium with the isomer containing a metalated phenyl ring that is bound to a single osmium, HOs ₃(CO)₈[μ-PhP(C₆H₄- ν¹)CH₂PPh₂], which in turn captures CO to give HOs₃(CO)₉[μ-PhP(C₆H₄)CH ₂PPh₂] in a reaction that is first-order in cluster and CO with a rate constant of 23.9(3) × 10^-3 M^-1 s ^-1 at 288 K. The kinetics for the transformation of HOs ₃(CO)₉[μ-PhP(C₆H₄)CH ₂PPh₂] to Os₃(CO)₁₀(dppm) have been studied in toluene over the temperature range 317-340 K and found to be first-order in starting cluster and independent of CO. Important insight into the reductive coupling process was obtained from the carbonylation kinetics employing DOs₃(CO)₉[μ-(Ph-d₂)P(C ₆H₃D)CH₂PPh₂-d₄], which was prepared from Os₃(CO)₁₀(dppm-d₈) and where all of the ortho sites on the aryl groups contained deuterium. Here, a significant inverse isotope effect (k_H/k_D = 0.50) was found, whose origin actually derives from an inverse equilibrium isotope effect, and this supports a preequilibrium phase of the reaction involving a hydride (deuteride) cluster and an intact Os₃ cluster containing a coordinated aryl moiety, prior to the rate-limiting formation of the unsaturated cluster Os₃(CO)₉(dppm-P_a,P_e). Experimental proof for the intermediacy of an extremely labile cluster-aryl complex(es) in the proposed preequilibrium step has been demonstrated by photochemical experiments using the isotopically labeled cluster Os ₃(CO)₁₀(dppm-d₄), where each aryl group contains one ortho hydrogen and deuterium atom. Photolysis of Os ₃(CO)₁₀(dppm-d₄) in toluene-d₈ gives a 70:30 mixture of the octa- and nonacarbonyl HOs₃(CO) _8,9[μ-(Ph-d₁)P(C₆H₃D)CH ₂PPh₂-d₄] and DOs₃(CO) _8,9[μ-(Ph-d₁)P(C₆H₄)CH ₂PPh₂-d₄], respectively, over the temperature range 243-298 K. These data indicate that the intermediates formed after CO loss are kinetically labile and that the ortho metalation leads to a thermodynamically equilibrated mixture of hydride and deuteride clusters. DFT calculations corroborate the experimental findings and provide crucial details on the nature and lability of those cluster species that do not lend themselves to direct spectroscopic observation.

Full text of DOI:10.1021/om100475v

Organometallics 2010, 29, 4041–4057 4041
DOI: 10.1021/om100475v
Ortho-Metalation Dynamics and Ligand Fluxionality in the Conversion of
Os3(CO)10(dppm) to HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2]:
Experimental and DFT Evidence for the Participation of Agostic
C-H and π-Aryl Intermediates at an Intact Triosmium Cluster
Shih-Huang Huang,† Jason M. Keith,‡ Michael B. Hall,*,‡ and Michael G. Richmond*,†
†Department of Chemistry, University of North Texas, Denton, Texas 76203, and ‡Department of Chemistry,
Texas A&M University, College Station, Texas 77843
Received May 16, 2010
The mechanism for the stepwise conversion of Os3(CO)10(dppm) to HOs3(CO)9[ μ-PhP-
(C6H4)CH2PPh2] and HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] has been investigated. The octa-
carbonyl cluster HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] is in rapid equilibrium with the isomer
containing a metalated phenyl ring that is bound to a single osmium, HOs3(CO)8[ μ-PhP(C6H4-η1)-
CH2PPh2], which in turn captures CO to give HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] in a reaction that
is first-order in cluster and CO with a rate constant of 23.9(3) ꢀ 10-3 M-1 s-1 at 288 K. The kinetics
for the transformation of HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] to Os3(CO)10(dppm) have been
studied in toluene over the temperature range 317-340 K and found to be first-order in starting
cluster and independent of CO. Important insight into the reductive coupling process was obtained
from the carbonylation kinetics employing DOs3(CO)9[ μ-(Ph-d2)P(C6H3D)CH2PPh2-d4], which was
prepared from Os3(CO)10(dppm-d8) and where all of the ortho sites on the aryl groups contained
deuterium. Here, a significant inverse isotope effect (kH/kD = 0.50) was found, whose origin actually
derives from an inverse equilibrium isotope effect, and this supports a preequilibrium phase of the
reaction involving a hydride (deuteride) cluster and an intact Os3 cluster containing a coordinated
aryl moiety, prior to the rate-limiting formation of the unsaturated cluster Os3(CO)9(dppm-Pa,Pe).
Experimental proof for the intermediacy of an extremely labile cluster-aryl complex(es) in the
proposed preequilibrium step has been demonstrated by photochemical experiments using the
isotopically labeled cluster Os3(CO)10(dppm-d4), where each aryl group contains one ortho hydrogen
and deuterium atom. Photolysis of Os3(CO)10(dppm-d4) in toluene-d8 gives a 70:30 mixture of the
octa- and nonacarbonyl HOs3(CO)8,9[ μ-(Ph-d1)P(C6H3D)CH2PPh2-d4] and DOs3(CO)8,9[ μ-(Ph-d1)P-
(C6H4)CH2PPh2-d4], respectively, over the temperature range 243-298 K. These data indicate that the
intermediates formed after CO loss are kinetically labile and that the ortho metalation leads to a
thermodynamically equilibrated mixture of hydride and deuteride clusters. DFT calculations
corroborate the experimental findings and provide crucial details on the nature and lability of those
cluster species that do not lend themselves to direct spectroscopic observation.
Introduction
C-H bond, has been forever removed.2 The reaction dy-
namics for the metal-mediated activation of common and
exotic feedstocks have been elucidated through the use of
transient spectroscopic and computational methodologies.3,4
The use of these techniques, especially when utilized in con-
cert, has provided heretofore unprecedented and intimate
details into catalytic processes involving the activation of
The directed functionalization of relatively unreactive
C-H bonds remains a top priority in the field of catalysis.1
Only recently have the intricacies associated with the inter-
molecular C-H bond activation of alkane and alkene sub-
strates at mononuclear compounds been unraveled, and the
shroud of ambiguity concerning the nature and reactivity of
agostic metal-alkane and π metal-arene compounds, the
species that typically precede the formal breaking of a
(2) For a few representative examples, see: (a) Janowicz, A. H.;
Bergman, R. G. J. Am. Chem. Soc. 1982, 104, 352. (b) Hoyano, J. K.;
Graham, W. A. G. J. Am. Chem. Soc. 1982, 104, 3723. (c) Jones, W. D.;
Feher, F. J. J. Am. Chem. Soc. 1984, 106, 1650. (i) 1986, 108, 4814.
(d) Buchanan, J. M.; Stryker, J. M.; Bergman, R. G. J. Am. Chem. Soc.
1986, 108, 1537. (e) Wick, D. D.; Reynolds, K. A.; Jones, W. D. J. Am.
Chem. Soc. 1999, 121, 3974. (f ) Northcutt, T. O.; Wick, D. D.; Vetter, A. J.;
Jones, W. D. J. Am. Chem. Soc. 2001, 123, 7257. (g) Churchill, D. G.; Janak,
K. E.; Wittenberg, J. S.; Parkin, G. J. Am. Chem. Soc. 2003, 125, 1403.
(h) 1983, 105, 3929.
*Corresponding authors. E-mail: (M.B.H.) mbhall@tamu.edu; (M.G.R.)
cobalt@unt.edu.
(1) (a) Bergman, R. G. Science 1984, 223, 902. (b) Arndtsen, B. A.;
Bergman, R. G.; Mobley, T. A.; Peterson, T. H. Acc. Chem. Res. 1995, 28,
154. (c) Crabtree, R. H. Chem. Rev. 1995, 95, 987. (d) Shilov, A. E.; Shul'pin,
G. B. Chem. Rev. 1997, 97, 2879. (e) Labinger, J. A.; Bercaw, J. E. Nature
2002, 417, 507.
r
2010 American Chemical Society
Published on Web 09/01/2010
pubs.acs.org/Organometallics

4042 Organometallics, Vol. 29, No. 18, 2010
Huang et al.
C-H bonds and the subsequent conversion of the resulting
intermediates to commodity chemicals.
of the diphosphine-bridged cluster Os3(CO)10(dppm) leads
to the formation of the ortho-metalated cluster HOs3(CO)8-
[ μ-PhP(C6H4-μ2,η1)CH2PPh2], as depicted in eq 1.9 Two points
are noteworthy as it pertains to this octacarbonyl cluster: (1)
C-H bond formation is readily reversible under CO, and (2)
it remains an active platform for the construction of a plet-
hora of triosmium clusters, whose synthesis and isolation
would otherwise be hampered through traditional thermal
activation of Os3(CO)10(dppm).10,11
The high intrinsic reactivity displayed by those unsatu-
rated species able to activate hydrocarbon substrates can
also lead to the deleterious activation of an auxiliary ligand-
(s) in the coordination sphere of the metal. In the case of
metal-bound phosphine ligands, C-P and C-H bond clea-
vage manifolds are common, and these reactions are unde-
sirable, as they are believed to retard the catalytic activity of
certain reactions.5 The activation of alkyl and aryl C-H
bonds in coordinated phosphine ligands gives rise to cyclo-
metalation and ortho-metalation products, respectively. In-
terestingly, the exact mechanism associated with the above
phosphine activation processes is far from clear,6 and while
one would assume that these ligand-based C-H activations
would follow reactivity paths analogous to those systems
whose intermolecular activation of hydrocarbons is now well
understood, the situation remains, for the most part, largely
unaddressed. In an early landmark study, Jones investigated
the selectivity of different unsaturated Cp*Rh(I) species for
their intramolecular metalation reactivity (ortho and cyclo)
at the ancillary phosphine ligands versus intermolecular
activation of benzene.7 Here, the kinetic and thermodynamic
aspects related to intra- and intermolecular C-H bond
activation were dissected, and the data unequivocally revea-
led a pronounced thermodynamic preference for the intra-
molecular activation of the ancillary P-ligand and a kinetic
preference for intermolecular C-H bond activation of the
benzene solvent.
Our groups maintain an interest in the mechanistic and
computational study of metal cluster compounds, especially
those systems that exhibit an enhanced activation of small
molecules and bidentate ligands.12,13 Recently one of us (M.
G.R.) has published a study concerning the ortho metalation
of the chelating diphosphine in the cluster Os3(CO)10(bpcd)
[where bpcd = 4,5-bis(diphenylphosphino)-4-cyclopentene-
1,3-dione].14 Thermolysis or photolysis of Os3(CO)10(bpcd)
leads to CO loss and ortho metalation of one of the aryl
substituents to give the hydride cluster HOs3(CO)9[ μ-PhP-
(C6H4)CdC(PPh2)C(O)CH2C(O)]. Treatment of the resulting
hydride with CO quantitatively regenerates Os3(CO)10(bpcd)
The ligand activation described above is not limited to
mononuclear entities and, in fact, has been observed in
numerous polynuclear metal clusters. Unlike their mono-
nuclear counterparts, polynuclear entities display an added
dimension of complexity, as contiguous metal centers often
promote the multisite activation of a cluster-bound ligand,
leading to reactivity patterns that have no parallel in mono-
nuclear coordination chemistry.8 For example, thermolysis
(9) (a) Clucas, J. A.; Foster, D. F.; Harding, M. M.; Smith, A. K.
Chem. Commun. 1984, 949. (b) Brown, M. P.; Dolby, P. A.; Harding, M. M.;
Mathews, A. J.; Smith, A. K. Dalton Trans. 1993, 1671.
(10) For selected reports employing HOs3(CO)8[ μ-PhP(C6H4-
μ2,η1)CH2PPh2] as a starting material in the synthesis of other poly-
nuclear osmium clusters, see: (a) Clucas, J. A.; Harding, M. M.; Smith,
A. K. Chem. Commun. 1985, 1280. (b) Bartlett, R. A.; Cardin, C. J.; Cardin,
D. J.; Lawless, G. A.; Power, J. M.; Power, P. P. Chem. Commun. 1988, 312.
(c) Brown, M. P.; Dolby, P. A.; Harding, M. M.; Mathews, A. J.; Smith, A. K.;
Osella, D.; Arbrun, M.; Gobetto, R.; Raithby, P. R.; Zanello, P. Dalton Trans.
1993, 827. (d) Abedin, S. M. T.; Hardcastle, K. I.; Kabir, S. E.; Malik,
K. M. A.; Mottalib, M. A.; Rosenberg, E.; Abedin, M. J. Organometallics
2000, 19, 5623. (e) Deeming, A. J.; Hassan, M. M.; Kabir, S. E.; Nordlander,
E.; Tocher, D. A. Dalton Trans. 2004, 3709. (f ) Kabir, S. E.; Miah, M. A.;
Sarker, N. C.; Golzar-Hussain, G. M.; Hardcastle, K. I.; Nordlander, E.;
Rosenberg, E. Organometallics 2005, 24, 3315. (g) Hassan, M. R.; Hogarth,
G.; Golzar-Hossain, G. M.; Kabir, S. E.; Raha, A. K.; Saha, M. S.; Tocher,
D. A. Organometallics 2007, 26, 6473.
(11) For a recent review that outlines the stabilizing role played by the
dppm ligand in M3(CO)10(dppm) clusters (where M = Ru, Os), see:
Kabir, S. E.; Hogarth, G. Coord. Chem. Rev. 2009, 253, 1285.
(12) (a) Adams, R. D.; Captain, B.; Fu, W.; Hall, M. B.; Smith, M. D.;
Webster, C. E. Inorg. Chem. 2004, 43, 3921. (b) Adams, R. D.; Captain, B.;
Hall, M. B.; Smith, J. L., Jr.; Webster, C. E. J. Am. Chem. Soc. 2005, 127,
1007. (c) Adams, R. D.; Captain, B.; Beddie, C.; Hall, M. B. J. Am. Chem.
Soc. 2007, 129, 986. (d) Adams, R. D.; Captain, B.; Smith, M. D.; Beddie, C.;
Hall, M. B. J. Am. Chem. Soc. 2007, 129, 5981. (e) Adams, R. D.; Boswell,
E. M.; Hall, M. B.; Yang, X. Organometallics 2008, 27, 4938. (f ) Adams,
R. D.; Hall, M. B.; Pearl, W. C., Jr.; Yang, X. Inorg. Chem. 2009, 48, 652.
(13) (a) Kandala, S.; Hammons, C.; Watson, W. H.; Wang, X.;
Richmond, M. G. Dalton Trans. 2010, 39, 1620. (b) Poola, B.; Carrano,
C. J.; Richmond, M. G. Organometallics 2008, 27, 3018. (c) Watson, W. H.;
Kandala, S.; Richmond, M. G. J. Organomet. Chem. 2007, 692, 1648; 968.
(d) Kandala, S.; Richmond, M. G. Inorg. Chem. 2006, 45, 5976.
(e) Watson, W. H.; Wu, G.; Richmond, M. G. Organometallics 2005, 24,
5431. (f ) Bott, S. G.; Shen, H.; Richmond, M. G. J. Organomet. Chem.
2004, 689, 3426. (g) Bott, S. G.; Shen, H.; Senter, R. A.; Richmond, M. G.
Organometallics 2003, 22, 1953. (h) Bott, S. G.; Yang, K.; Talafuse, K. A.;
Richmond, M. G. Organometallics 2003, 22, 1383.
(3) (a) Whittlesey, M. K.; Mawby, R. J.; Osman, R.; Perutz, R. N.;
Field, L. D.; Wilkinson, M. P.; George, M. W. J. Am. Chem. Soc. 1993,
115, 8627. (b) Schultz, R. H.; Bengali, A. A.; Tauber, M. J.; Weiller, B. H.;
Wasserman, E. P.; Kyle, K. R.; Moore, C. B.; Bergman, R. G. J. Am. Chem.
Soc. 1994, 116, 7369. (c) Bengali, A. A.; Schultz, R. H.; Moore, C. B.;
Bergman, R. G. J. Am. Chem. Soc. 1994, 116, 9585. (d) Bromberg, S. E.;
Lian, T.; Bergman, R. G.; Harris, C. B. J. Am. Chem. Soc. 1996, 118, 2069.
(e) Blake, A. J.; George, M. W.; Hall, M. B.; McMaster, J.; Portius, P.; Sun,
ꢀ
X. Z.; Towrie, M.; Webster, C. E.; Wilson, C.; Zaric, S. D. Organometallics
2008, 27, 189. (f ) Calladine, J. A.; Vuong, K. Q.; Sun, X. Z.; George, M. W.
Pure Appl. Chem. 2009, 81, 1667.
ꢀ
(4) (a) Niu, S.; Zaric, S.; Bayse, C. A.; Strout, D. L.; Hall, M. B.
Organometallics 1998, 17, 5139. (b) Bergman, R. G.; Cundari, T. R.;
Gillespie, A. M.; Gunnoe, T. B.; Harman, W. D.; Klinckman, T. R.; Temple,
M. D.; White, D. P. Organometallics 2003, 22, 2331. (c) Hartwig, J. F.; Cook,
K. S.; Hapke, M.; Incarvito, C. D.; Fan, Y.; Webster, C. E.; Hall, M. B. J. Am.
Chem. Soc. 2005, 127, 2538. (d) Evans, M. E.; Li, T.; Vetter, A. J.; Rieth,
R. D.; Jones, W. D. J. Org. Chem. 2009, 74, 6907. (e) Zhu, H.; Ziegler, T.
Organometallics 2009, 28, 2773. (f ) Yang, Z.; Hall, M. B. J. Phys. Chem. A
2009, 113, 2152. (g) Flener-Lovitt, C.; Woon, D. E.; Dunning, T. H., Jr.;
Girolami, G. S. J. Phys. Chem. A 2010, 114, 1843. (h) Balcells, D.; Clot, E.;
Eisenstein, O. Chem. Rev. 2010, 110, 749.
(5) (a) Lavigne, G.; de Bonneval, B. In Catalysis by Di- and Poly-
nuclear Metal Cluster Complexes; Adams, R. D.; Cotton, F. A., Eds.; Wiley-
VCH: New York, 1998. (b) Garrou, P. E.; Dubois, R. A.; Jung, C. W.
CHEMTECH 1985, 15, 123. (c) Moser, W. R.; Papile, C. J.; Weininger, S. J.
J. Mol. Catal. 1987, 41, 293. (d) Herrmann, W. A.; Brossmer, C.; Oefele, K.;
Beller, M.; Fischer, H. J. Organomet. Chem. 1995, 491, C1.
(6) Albrecht, M. Chem. Rev. 2010, 110, 576, and references therein.
(7) Jones, W. D.; Feher, F. J. J. Am. Chem. Soc. 1985, 107, 620.
(8) (a) Muetterties, E. L.; Rhodin, T. N.; Band, E.; Brucker, C.;
Pretzer, H. Chem. Rev. 1979, 79, 91. (b) Dyson, P. J. Coord. Chem. Rev.
2004, 248, 2443.
(14) Watson, W. H.; Wu, G.; Richmond, M. G. Organometallics
2006, 25, 930.

Article
Organometallics, Vol. 29, No. 18, 2010 4043
Scheme 1
and establishes the reversible nature associated with the
reactivity of the C-H bond in terms of oxidative cleavage
and reductive coupling steps. On the basis of kinetic and isotopic
substitution studies on the reductive coupling step, we have
confirmed that C-H bond formation exhibits an inverse
equilibrium isotope effect (EIE). These data strongly support
the intermediacy of the transient π-complex Os3(CO)9[ μ-
PhP(μ-C6H5)CdC(PPh2)C(O)CH2C(O)], whose formation
precedes the generation of the coordinatively unsaturated
cluster Os3(CO)9(bpcd). Scheme 1 illustrates this particular
reaction.
Experimental Section
General Methods. The Os3(CO)10(dppm) cluster was prepared
from the reaction of dppm with either Os3(CO)12 or Os3(CO)10-
(MeCN)2, the latter which was synthesized from Os3(CO)12,
Me3NO, and MeCN.17 The parent cluster Os3(CO)12 was pre-
pared from OsO4 and CO.18 The OsO4 was purchased from
Engelhard Chemical Co., and the chemicals dppm, 2-bromoani-
line, 2,6-dibromoaniline, diphenylacetic acid, BuLi (2.5 M in
hexanes), and Me3NO xH2O were purchased from Aldrich
Chemical Co. The anhydrous Me3NO employed in our studies
was obtained from Me3NO xH2O, after the waters of hydration
3
3
were azeotropically removed under reflux using benzene as a
solvent. The deuterated compounds D2O (99.5% D), NaOD
(40% in D2O; 99.5% D), benzene-d6 (99.6% D), toluene-d8
(99.6% D), CD2Cl2 (99.8% D), CDCl3 (99.8% D), and MeOD
(99.8% D) were purchased from Cambridge Isotope Labora-
tories. All reaction solvents were distilled from a suitable drying
agent under argon or obtained from an Innovative Technology
solvent purification system; when not in use, the purified
solvents were stored in Schlenk storage vessels equipped with
high-vacuum Teflon stopcocks.19 The NMR solvents benzene-
d6 and toluene-d8 were purified by bulb-to-bulb distillation from
Na/benzophenone, with the CD2Cl2 and CDCl3 solvents dis-
tilled from P2O5. The molarity of the BuLi used in our reactions
was periodically checked by titration against diphenylacetic acid.20
The photochemical studies were conducted with an Oriel universal
power supply equipped with a 200 W high-pressure Hg lamp. The
quartz NMR tubes used in these studies were equipped with a high-
vacuum stopcock, and the low-temperature UV photolysis experi-
ments were carried out in a homemade Dewar constructed from
Suprasil quartz tubing.
Notwithstanding the fact that the activation of the dppm
ligand in Os3(CO)10(dppm) is facile and reversible and
that HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] serves as a
cluster synthon for “Os3(CO)8(dppm)”, no mechanistic data
exist for any of the steps associated with this transforma-
tion aside from empirical product distributions based on
ligand substitution studies that have employed these two
clusters.10c-e,15,16 The unanswered questions concerning the
oxidative cleavage and reductive coupling steps attendant in
the activation of the dppm ligand in Os3(CO)10(dppm) and
HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] require resolu-
tion before generalities involving the ortho metalation in
different Os3(CO)10(diphosphine) clusters can be made with
certainty. Moreover, knowledge of such ligand activation
is required if these and related cluster systems are to be
exploited in any catalysis scenario. In the present work, we
report our kinetic and isotope studies on the ortho metala-
tion of the bridging diphosphine ligand in Os3(CO)10(dppm),
and the results from this study are contrasted with the ortho-
metalation data obtained from Os3(CO)10(bpcd). DFT cal-
culations have also been performed in order to shed insight
on the intermediates in this reaction and to facilitate the
construction of a working mechanism that is consistent with
the experimental results.
IR spectra were recorded on Nicolet 20SXB or 6700 FT-IR
spectrometers in sealed 0.1 mm NaCl cells. The quoted 1H, 13C,
and 31P NMR data were recorded on a Varian VXR-500
1
spectrometer at 500, 125, and 202 MHz, respectively. The H
and 13C spectral data have been referenced against the residue
protiated and carbon resonance(s) of the NMR solvent. The
reported 31P chemical shift data were recorded in the proton-
decoupled mode and are referenced to external H3PO4 (85%),
whose chemical shift was set at δ = 0. The GC-MS mass spectral
data for the samples 2-deuterio-iodobenzene, 2,6-dideuterio-
iodobenzene, PPh3-d6ortho and PPh3-d3ortho were recorded on a
Finnigan/Thermo-Electron GC-MS system, while the ESI-MS
spectra of the dppm isotopomers and their corresponding
triosmium clusters were recorded on a Thermo Finnigan Deca
XP ion trap mass spectrometer using 1% AcOH in MeCN or
MeCN/KI as the sample matrix.
(15) To our knowledge, only one kinetic study on Os3(CO)10(dppm)
exists, and this involves the kinetic measurement for the ring closure of
1
€
the dppm ligand in Os3(CO)11(η -dppm): Poe, A.; Sekhar, V. C. J. Am.
Chem. Soc. 1984, 106, 5034.
(16) (a) Raha, A. K.; Ghosh, S.; Manzurul Karim, M.; Tocher, D. A.;
Begum, N.; Sharmin, A.; Rosenberg, E.; Kabir, S. E. J. Organomet.
Chem. 2008, 693, 3613. (b) Azad, S. M.; Azam, K. A.; Kabir, S. E.; Saha,
M. S.; Golzar Hossain, G. M. J. Organomet. Chem. 2005, 690, 4206.
(c) Abedin, S. M. T.; Azam, K. A.; Hursthouse, M. B.; Kabir, S. E.; Malik,
K. M. A.; Mottalib, M. A.; Rosenberg, E. J. Cluster Sci. 2001, 12, 5.
(d) Kabir, S. E.; Abdul Malik, K. M.; Molla, E.; Abdul Mottalib, M.
J. Organomet. Chem. 2000, 616, 157. (e) Azam, K. A.; Hursthouse,
M. B.; Islam, M. R.; Kabir, S. E.; Malik, K. M. A.; Miah, R.; Sudbrake, C.;
Vahrenkamp, H. Dalton Trans. 1998, 1097. (f ) Bruce, M. I.; Horn, E.; Bin
Shawkataly, O.; Snow, M. R.; Tiekink, E. R. T.; Williams, M. L.
J. Organomet. Chem. 1986, 316, 187. (g) Cartwright, S.; Clucas, J. A.;
Dawson, R. H.; Foster, D. F.; Harding, M. M.; Smith, A. K. J. Organomet.
Chem. 1986, 302, 403.
Preparation of dppm-d8ortho. To 50 mL of liquid ammonia in a
100 mL Schlenk flask under argon at -78 ꢀC was added 0.69 g
(17) Nicholls, J. N.; Vargas, M. D. Inorg. Synth. 1989, 26, 289.
(18) Drake, S. R.; Loveday, P. A. Inorg. Synth. 1990, 28, 230.
(19) Shriver, D. F. The Manipulation of Air-Sensitive Compounds;
McGraw-Hill: New York, 1969.
(20) Kofron, W. G.; Baclawski, L. M. J. Org. Chem. 1976, 41, 1879.

4044 Organometallics, Vol. 29, No. 18, 2010
Huang et al.
(30 mmol) of sodium. The solution was stirred for several
minutes prior to the addition of 4.0 g (15 mmol) of PPh3-d6ortho,
(68%) of DOs3(CO)8[ μ-(Ph-d2ortho)P(C6H3D-μ2,η1)CH2PPh2-
d
4ortho]. 1H NMR (C6D6): δ 3.96 (ddd, 1H, methylene, 2JHH
14.1 Hz, JPH = 11.0, 6.3 Hz), 4.65 (ddd, 1H, methylene,
=
3
at which time the initially blue solution slowly turned dark
orange. After stirring for 0.5 h at -78 ꢀC, 0.80 g (15 mmol)
of NH4Cl was added in one portion to quench the generated
phenyl sodium from the reduction reaction. Stirring was con-
tinued for 1 h at -78 ꢀC, and the orange solution was then
treated with 0.48 mL (7.5 mmol) of CH2Cl2 in 2 mL of Et2O to
give a pale yellow solution that signaled the consumption of the
sodium diphenylphosphide. The ammonia was allowed to eva-
porate, and 50 mL of degassed water was added to the remaining
residue, followed by extraction of the aqueous layer with Et2O
(3 ꢀ 50 mL). The combined organic layers were dried over
MgSO4, after which the solution was filtered and then concen-
trated to dryness to afford the crude dppm-d8ortho. The product
was purified by flash-column chromatography over silica gel
using CH2Cl2/hexane (3:7) to give dppm-d8ortho as a colorless
solid in 70% yield (2.1 g). 1H NMR (C6D6): δ 2.78 (t, 2H,
methylene, JPH = 2.0 Hz), 7.04 (s, 12H, meta and para). 31P
NMR (C6D6): δ -23.21.21 ESI-MS: m/z 393.47 [M þ H]þ.
Synthesis of Os3(CO)10(dppm-d8ortho). To a large Schlenk tube
containing 0.46 g (0.49 mmol) of Os3(CO)10(MeCN)2 was added
100 mL of CH2Cl2 via cannula, followed by 0.20 g (0.51 mmol)
of dppm-d8ortho. The solution was stirred for 12 h at room
temperature and then examined by TLC, which confirmed the
presence of the desired product. Os3(CO)10(dppm-d8ortho) was
subsequently purified by column chromatography over silica gel
using a 3:7 mixture of CH2Cl2/hexane as the mobile phase.
Recrystallization of the isolated product using benzene furn-
ished 0.42 g (68%) of orange Os3(CO)10(dppm-d8ortho). 1H
NMR (C6D6): δ 4.85 (t, 2H, methylene, JPH = 10.5 Hz), 6.89
(s, 12H, meta and para). 31P NMR (C6D6): δ -28.14. ESI-MS:
m/z 1286.73 [M þ K]þ and 1258.92 [M - CO þ K]þ.
3
3
2JHH = 14.1 Hz, JPH = 11.1, 8.4 Hz), 6.24 (dt, Hb, JHH
=
5
7.2 Hz from coupling with Ha and Hc, JPH = 2.2 Hz),
6.73-6.81 (m, 2H), 6.95-7.06 (m, 2H), 7.08-7.15 (m, Hc and
5H), 9.02 (ddd, Ha, 3JHH = 7.2 Hz from coupling with Hb, 4JHH
= 1.3 Hz from coupling with Hc, 4JPH = 2.0 Hz).22 31P NMR
(C6D6): δ -19.82 (d, JPP = 69 Hz), -18.67 (d, JPP = 69 Hz).
ESI-MS: m/z 1190.9 [M]þ and m/z peaks for the consecutive loss
of 1-7 CO groups.
Synthesis of DOs3(CO)9[ μ-(Ph-d2ortho)P(C6H3D)CH2PPh2-
d4ortho
]
from DOs3(CO)8[ μ-(Ph-d2ortho)P(C6H3D-μ2,η1)CH2-
PPh2-d4ortho] and CO. To 0.10 g (0.084 mmol) of DOs3(CO)8-
[ μ-(Ph-d2ortho)P(C6H3D-μ2,η1)CH2PPh2-d4ortho] in a medium
Schlenk tube was added 40 mL of toluene via syringe, after
which the solution was cooled to ca. 5 ꢀC. CO was slowly bubb-
led into the toluene solution for a period of 1 h, at which time
TLC analysis revealed the complete consumption of the starting
cluster and the presence of the nonacarbonyl product DOs3-
(CO)9[ μ-(Ph-d2ortho)P(C6H3D)CH2PPh2-d4ortho]. The solvent
was removed under vacuum and the residue purified by flash-
column chromatography over silica gel using a 7:3 mixture of
1
CH2Cl2/hexane as the eluent. Yield: 97 mg (95%). H NMR
2
3
(C6D6): δ 3.39 (ddd, 1H, methylene, JHH = 12.4 Hz, JPH
=
10.2, 9.2 Hz), 4.91 (ddd, 1H, methylene, 2JHH = 12.4 Hz, 3JPH
= 12.7, 9.2 Hz), 6.02 (ddd, Hc, 3JHH = 7.3 Hz from coupling
with Hb, 3JHH = 2.9 Hz from coupling with Ha, 4JPH = 1.0 Hz),
3
6.42 (dt, Hb, JHH = 7.3 Hz from coupling with Ha and Hc,
5JPH = 2.5 Hz), 6.65 (m, 2H), 6.74 (m, 1H), 6.83-7.09 (m, 6H),
8.20 (ddd, Ha, 3JHH = 7.3 Hz from coupling with Hb, 4JHH
=
2.9 Hz from coupling with Hc, JPH = 1.2 Hz). 31P NMR
(C6D6): δ -23.23 (d, JPP = 71 Hz), -14.21 (d, JPP = 71 Hz).
ESI-MS: m/z 1223.9 [M - CO þ 2H2O]þ.
4
Synthesis of DOs3(CO)8[ μ-(Ph-d2ortho)P(C6H3D-μ2,η1)CH2-
PPh2-d4ortho] from Os3(CO)10(dppm-d8ortho). To a 100 mL
Schlenk tube under argon flush was charged 0.20 g (0.16 mmol)
of Os3(CO)10(dppm-d8ortho) and 50 mL of toluene. The solution
was heated at reflux for 8 h, during which time the solution color
gradually changed from orange to green as the octacarbonyl
product formed. The solution was allowed to cool and the
product isolated by column chromatography to give 0.13 g
Preparation of dppm-d4ortho. To 50 mL of liquid ammonia in a
100 mL Schlenk flask under argon at -78 ꢀC was added 0.60 g
(26 mmol) of sodium, and stirring continued for 0.5 h before the
addition of 3.5 g (13 mmol) of PPh3-d3ortho. The solution was
stirred for an additional 0.5 h at -78 ꢀC and then treated with
0.69 g (13 mmol) of NH4Cl. At this point, 0.42 mL (6.5 mmol) of
CH2Cl2 in 3 mL of Et2O was added to the ammonia solution
dropwise, and the solution was allowed to warm to room
temperature with slow warming. The ammonia was allowed to
evaporate and the crude product isolated upon a preliminary
extractive workup. The dppm-d4ortho was purified by flash-
column chromatography over silica gel using a CH2Cl2/hexane
(3:7) as a colorless solid in 75% yield (1.9 g). 1H NMR (C6D6):
δ 2.79 (t, 2H, methylene, JPH = 2.0 Hz), 7.05 (bs, 12H, meta and
para), 7.44 (b, 4H, ortho). 31P NMR (C6D6): δ -23.13. ESI-MS:
m/z 389.40 [M þ H]þ.
(21) Cf.: the 31P chemical shift of dppm-d0 at δ -22.91 recorded under
identical conditions.
(22) The labeling scheme employed in the assignment of the hydro-
gens on the metalated aryl ring in the d8 isotopomers of DOs3(CO)8-
[ μ-(Ph-d2ortho)P(C6H3D-μ2,η1)CH2PPh2-d4ortho] and DOs3(CO)9[ μ-(Ph-
d2ortho)P(C6H3D)CH2PPh2-d4ortho] is shown below. The reported cou-
pling constants for the hydrogens on the metalated aryl ring were
determined from spectral simulations using gNMR.
Synthesis of Os3(CO)10(dppm-d4ortho). To 0.25 g (0.27 mmol)
of Os3(CO)10(MeCN)2 in 60 mL of CH2Cl2 in a Schlenk flask
was added 0.10 g (0.26 mmol) of dppm-d4ortho. The solution was
allowed to stir overnight, after which time the solvent was
removed and the residue purified by flash-column chromatog-
raphy. Recrystallization of the crude product using benzene
1
afforded 0.23 g (72%) of orange Os3(CO)10(dppm-d4ortho). H
NMR (C6D6): δ 4.86 (t, 2H, methylene, JPH = 10.8 Hz), 6.90
(bs, 12H, meta and para), 7.26 (bs, 4H, ortho). 31P NMR
(C6D6): δ -27.10. ESI-MS: m/z 1282.93 [M þ K]þ and
1255.97 [M - CO þ K]þ.
Carbonylation Kinetics. The UV-vis kinetic studies employ-
ing HOs3(CO)8,9[ μ-PhP(C6H4)CH2PPh2] and DOs3(CO)8,9
-
[ μ-(Ph-d2ortho)P(C6H3D)CH2PPh2-d4ortho] were analyzed in Sup-
rasil quartz UV-visible cells (1.0 cm width) that were equipped
with a high-vacuum Teflon stopcock to facilitate handling on
the vacuum line. The clusters HOs3(CO)8[ μ-PhP(C6H4-
μ2,η1)CH2PPh2] and DOs3(CO)8[ μ-(Ph-d2ortho)P(C6H3D-μ2,η1)-
CH2PPh2-d4ortho] are extremely sensitive to added CO, and the
samples were thermally equilibrated in the UV-vis cell holder

Article
Organometallics, Vol. 29, No. 18, 2010 4045
calculated from a plot of ln(k/T ) versus T-1 24
with the error
limits representing the deviation of the data points about the
least-squares line of the Eyring plot.
Table 1. Experimental Rate Constants for the Carbonylation of
HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] to HOs3(CO)9[ μ-
PhP(C6H4)CH2PPh2] in Toluenea
,
Computational Methodology and Modeling Details. All initial
calculations were performed with the hybrid DFT functional
B3LYP as implemented in the Gaussian 03 programming
package.25 This functional utilizes the Becke three-parameter
exchange functional (B3)26 combined with the correlation func-
tional of Lee, Yang, and Parr (LYP)27 and is known to produce
good descriptions of reaction profiles for transition metal con-
taining compounds.28,29 The Os atoms were described using the
LANL2DZ effective core potential (ecp) and basis set of Hay
and Wadt30 with the 6p orbitals replaced with the split valence
functions from Couty and Hall.31 The P atoms were described
using the LANL2DZ(d) ecp basis set.32,33 All other atoms
(C, H, O) were modeled using a Pople style34 double-ζ 6-31G-
(d0,p0) basis set with polarization functions optimized for
heavy atoms.35
entry
temp (K)
CO pressure (atm)
104kobsd (s-1
)
1
2
3
4
5
6
7
8
288.1
288.1
288.1
288.1
293.2
301.5
307.9
316.2
1.0
1.0
17.0
34.0
1.0
1.0
1.0
1.0
5.30(3)
5.23(2)b
39.3(1)
64.1(3)
7.26(2)
13.2(5)
23.2(5)
40.3(8)
a The UV-vis kinetic data were collected using a ca. 10-4 M solution
of HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] by following the decrease
in the absorbance of the 388 nm band, unless otherwise noted. b Reaction
carried out using a ca. 10-3 M solution of HOs3(CO)8[ μ-PhP(C6H4-
μ2,η1)CH2PPh2] by following the decrease in the absorbance of the 600
nm band.
All geometries were fully optimized and evaluated for the
correct number of imaginary frequencies through calculation
of the vibrational frequencies using the analytical Hessian.
Zero imaginary frequencies (positive eigenvalues) correspond
to an intermediate or local minimum, whereas one imaginary
frequency designates a transition state. From the analytical
Hessian zero-point energies as well as enthalpy and entropy
corrections for 298.15 K were also calculated and added to the
total electronic energy to obtain a total free energy, ΔG
[298.15].
Implicit solvent effects were incorporated by using the polar-
izable continuum model36 with toluene as the solvent using the
parameter ε = 2.379. For this method we chose to employ the
radii optimized for COSMO-RS by Klamt et al.37 These radii
were found to provide good results for systems involving
hydrogen atom transfer. The solvent effects were calculated at
geometries calculated in the gas phase. The resulting solvation
free energy correction was added to the total free energy from
above to obtain the total free energy.
Standard-state corrections were added to all species to con-
vert concentrations from 1 atm to 1 M. This was accomplished
via the equation ΔGꢀ0 = ΔGꢀ þ RT ln(Qꢀ0/Qꢀ), where the initial
concentration Qꢀ = 1 atm (1/24.5 M for an ideal gas) and the
final concentration Qꢀ0 = 1 M.38 Thus, a value of ca. -1.894
kcal/mol was added to the total free energy for each species. The
overall effect of this correction is the addition of (M - N) ꢀ
1.894 kcal/mol for the conversion of M molecules to N mole-
cules. This approximation is added in order to better model the
Table 2. Experimental Rate Constants for the Carbonylation
of HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] to Os3(CO)10(dppm)
in Toluenea
entry
temp (K)
CO pressure (atm)
105k (s-1
)
1
2
3
4
5
6
7
317.2
323.0
328.5
328.5
328.5
334.0
339.8
1.0
1.0
1.0
17.0
34.0
1.0
4.38(7)
9.59(9)
20.5(3)
16.8(3)
23.2(6)
31.3(4)
69(6)
1.0
a The UV-vis kinetic data were collected using a 8.73 ꢀ 10-3
M
solution of HOs3(CO)8[ μ-PhP(C6H4)CH2PPh2] by following the in-
crease in the absorbance of either the 358 or 425 nm bands.
prior to the admission of CO. For those reactions involving the
octacarbonyl cluster conducted under 1 atm of CO, the CO was
administered directly into the reaction solution via a gastight
syringe. In the case of HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] and
DOs3(CO)9[ μ-(Ph-d2ortho)P(C6H3D)CH2PPh2-d4ortho], the reac-
tion solutions were purged with CO at room temperature
immediately before the start of each kinetic experiment. The
experiments performed under high CO pressure were carried out
in a carbon-steel 300 mL autoclave, with the internal CO pres-
sure regulated with the aid of a Tescom pressure regulator. The
autoclave was equipped with a tip tube that enabled sample
removal for UV-vis analysis while maintaining constant CO
pressure. The Hewlett-Packard 8452A diode array spectrometer
employed in our studies was configured with a custom VT cell
holder, which was connected to a VWR constant-temperature
circulator. The quoted reaction temperatures are considered to
be accurate to within (0.5 K. The UV-vis kinetics performed
under 1 atm of CO were monitored by following the optical
changes of the 388 or 600 nm band of the octacarbonyl cluster
(decay) and the 358 or 425 nm band of the nonacarbonyl cluster
(growth) as a function of time for at least 4-6 half-lives, while
those reactions carried out under higher CO pressure were
examined out to 3 half-lives. The UV-vis-derived rate constants
quoted in Tables 1 and 2 were determined by nonlinear regres-
sion analysis using a single exponential function, with the initial
(Ao) and final (A¥) absorbance values and the rate constant (k)
treated as free variables in the calculation.23
(25) Frisch, M. J.; et al. Gaussian 03, Revision E.01; Gaussian, Inc.:
Pittsburgh, PA, 2003.
(26) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(27) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(28) Baker, J.; Muir, M.; Andzelm, J.; Scheiner, A. In Chemical
Applications of Density-Functional Theory; Laird, B. B.; Ross, R. B.;
Ziegler, T., Eds.; ACS Symposium Series 629; American Chemical Society:
Washington, DC, 1996.
(29) Niu, S.; Hall, M. B. Chem. Rev. 2000, 100, 353.
(30) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(31) Couty, M.; Hall, M. B. J. Comput. Chem. 1996, 17, 1359.
€
€
(32) Hollwarth, A.; Bohme, M.; Dapprich, S.; Ehlers, A. W.; Gobbi,
€
A.; Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 237.
(33) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284.
(34) (a) Hariharan, P. C.; Pople, J. A. Chem. Phys. Lett. 1972, 16, 217.
(b) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon, M. S.;
DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77, 3654.
(35) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem.
Phys. 1980, 72, 650.
The quoted activation parameters for the conversion of
HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] to Os3(CO)10(dppm) were
(36) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999.
(37) Klamt, A.; Jonas, V.; Buerger, T.; Lohrenz, J. C. W. J. Chem.
(23) The rate calculations were performed by using the equation
A(t) = A¥ þ ΔAe(-kt) with the aid of the commercially available
program Excel or Origin6.0.
Phys. 1998, 102, 5074.
(24) Carpenter, B. K. Determination of Organic Reaction Mechan-
isms; Wiley-Interscience: New York, 1984.
(38) Cramer, C. J. Essentials of Computational Chemistry, 2nd ed.;
Wiley: Chichester, UK, 2004; pp 378-379.

4046 Organometallics, Vol. 29, No. 18, 2010
Huang et al.
changes in solvent phase translational entropy when there are
changes in the number of molecules present.
Kinetic isotope effects (KIE) were examined through calcula-
tion of the analytic Hessian upon substitution of deuterium for
hydrogen. The new zero-point energies and enthalpy and en-
tropy terms for the deuterated species were thus determined and
used to obtain the corresponding free energies. The approxi-
mated KIEs were then calculated from kH/kD = e(ΔG
D - ΔGH/RT )
.
Results and Discussion
I. Carbonylation Kinetics for the Conversion of HOs3-
(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] to HOs3(CO)9[ μ-PhP-
(C6H4)CH2PPh2]. Thermolysis of Os3(CO)10(dppm) in re-
fluxing toluene leads to CO loss and the formation of HOs3-
(CO)9[ μ-PhP(C6H4)CH2PPh2], the latter presumably from
the putative unsaturated cluster Os3(CO)9(dppm), which
serves as the platform for the observed ortho-metalated
product. HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] is not stable
under the reaction conditions and readily loses CO to
afford the dimetalated cluster HOs3(CO)8[ μ-PhP(C6H4-
μ2,η1)CH2PPh2]. Our initial attempts to measure the rate
of CO loss from Os3(CO)10(dppm) and HOs3(CO)9[ μ-PhP-
(C6H4)CH2PPh2] were complicated by the fact that the
cluster products from decarbonylation are extremely sensi-
tive to the liberated CO in solution, leading us to question
the validity of our computed rate constants. While we
could measure the rate of disappearance of Os3(CO)10-
(dppm) and HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] in solu-
tion by IR or UV-vis spectroscopy, we could not guarantee
the absence of a back reaction between the liberated CO and
the resulting decarbonylation product. Under these condi-
tions the determined rate constants serve as a lower limit for
the true loss of CO from the starting cluster. Control experi-
ments conducted under CO (1 atm) confirmed our suspicions
regardingthe sensitivityof the decarbonylationproductsfrom
Os3(CO)10(dppm) and HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2]
to extraneous CO. Heating Os3(CO)10(dppm) under CO led
to no perceptible spectral changes when monitored by IR
spectroscopy, while HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] re-
acted with CO to afford Os3(CO)10(dppm) in essentially
quantitative yield.
The above limitations were circumvented by carrying out
the carbonylation studies starting from the clusters HOs3-
(CO)8,9[ μ-PhP(C6H4)CH2PPh2]. The reaction between HOs3-
(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] and excess CO proceeds
rapidly in toluene to give HOs3(CO)9[ μ-PhP(C6H4)CH2-
PPh2].39 We would point out that no conversion of HOs3-
(CO)9[ μ-PhP(C6H4)CH2PPh2] to Os3(CO)10(dppm) was ob-
served under the time scale of these experiments. The pro-
gress of these reactions was conveniently monitored by
following the decrease in the absorbance of the 388 nm band
of the starting cluster. The pseudo-first-order rate constants
for the first carbonylation step are listed in Table 1. As an
independent check of the value of kobsd measured at 388.1 K
(entry 1) and the use of the 388 nm band of HOs3(CO)8[ μ-
PhP(C6H4-μ2,η1)CH2PPh2] to accurately assess the progress
of the reaction over the examined temperature range of
288-316 K, we also conducted one experiment using the
weak visible band at 600 nm displayed by HOs3(CO)8[ μ-
PhP(C6H4-μ2,η1)CH2PPh2] (entry 2).40 The carbonylation
Figure 1. UV-vis spectral changes for HOs3(CO)8[ μ-PhP-
(C6H4-μ2,η1)CH2PPh2] in the presence of CO (1 atm) in toluene
at 288.1 K. The inset shows the absorbance data versus time for
the experimental data (9) and the nonlinear regression fit of the
pseudo-first-order rate constant kobsd (solid red line).
product is nonabsorbing at this wavelength, and this ensures
that the measured rate of the reaction truly reflects the con-
sumption of the HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2]
cluster. Figure 1 shows the UV-vis spectral changes for the
carbonylation reaction under 1 atm of CO and 288 K, with
the inset depicting the least-squares fit of the rate constant
(kobsd) and the absorbance data.
The effect of CO on the reaction was investigated in more
detail at 288 K. Entries 3 and 4, which represent reactions
carried out at CO pressures of 17 and 34 atm, respectively,
confirm that the rate is dependent on the [CO]. Graphical
analysis of the 288 K data in terms of a plot of kobsd versus
[CO] (see Supporting Information) afforded a rate constant
of 23.9(3) ꢀ 10-3 M-1 s-1. As mentioned earlier, the first
carbonylation step proceeds rapidly, and attempts to mea-
sure the rate of the reaction at higher temperatures as a
function of [CO] were deemed unfeasible using the conven-
tional apparatus at our disposal.41 The kinetic data are
consistent with a reaction that is first-order in cluster and
CO, and the rate-limiting step was initially assumed to
involve the direct attack of CO on the cluster, coupled with
the concerted conversion of the dimetalated aryl moiety to an
η1-aryl group. Equation 2 shows this transformation, where
the rate law obeys the form
rate ¼ k1½clusterꢁ½COꢁ
The simple nucleophilic addition of CO to the osmium
center containing only two carbonyl groups depicted in
eq 2 is kinetically valid provided the dimetalated phenyl
moiety remains strongly tethered to both metals. However,
if the aryl moiety is fluxional in terms of a reversible
(39) The concentration of CO in toluene was determined by using the
published solubility data. See: Basickos, L.; Bunn, A. G.; Wayland,
B. B. Can. J. Chem. 2001, 79, 854.

Article
Organometallics, Vol. 29, No. 18, 2010 4047
Scheme 2
dissociation/coordination process with respect to the
Os(CO)2 center, this would lead to an enhanced exposure
of the Os(CO)2 center to CO attack in the “opened” cluster
HOs3(CO)8[ μ-PhP(C6H4-η1)CH2PPh2]. Such coordinative
flexibility on the part of the metalated aryl moiety would
be similar to the windshield wiper motion exhibited by
nonclassical norbornyl cations and the migratory behavior
of a wide variety of ligands in di- and polynuclear com-
pounds.42,43 Scheme 2 illustrates this scenario and shows the
expected rate law for the capture of CO by HOs3(CO)8-
[ μ-PhP(C6H4-η1)CH2PPh2] (cluster B),44 which is kineti-
cally indistinguishable from the elementary second-order
process depicted in eq 2. A distinction between the two
mechanisms is possible based on the effect of [CO], provided
saturation kinetics can be observed. Whereas the mechan-
ism in eq 2 will always show a first-order dependence on
[CO], steady-state considerations predict that the carbon-
ylation will actually become independent of the attacking
ligand at high concentrations of CO (k2[CO] . k-1). In the
saturation regime, the reaction will reduce to a simple first-
order reaction (rate = k1[cluster A]). Unfortunately, this
concept cannot be experimentally tested in the present case
because the carbonylation is too rapid.
Increases in computing speed over the past decade have
greatly facilitated the examination of complex metal cluster
reactions by density functional theory (DFT). Several im-
portant studies have recently demonstrated the ability of
DFT calculations to address multistep reaction mechanisms
involving ligand fluxionality and substrate activation at
different polynuclear ruthenium and osmium clusters.45
Accordingly, we sought to resolve the above carbonylation
conundrum in HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] by
means of DFT calculations, and here we initially optimized
the structures of HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2]
(A1) and HOs3(CO)9[ μ-PhP(C6H4-η1)CH2PPh2] (B1). The
DFT-calculated structures for A1 and B1 are shown in
Figure 2, with the computed structure of A1 in good agree-
ment with the X-ray diffraction structure.9a The bridging
hydride in A1, which was not located during structural
refinement, spans the Os-Os vector that also binds the
metalated aryl moiety, as expected when the disposition of
the ancillary CO groups are taken into account.46 The com-
puted versus experimentally measured Os-C bond distances
˚
˚
˚
˚
of 2.38 A [2.297(15) A] and 2.33 A [2.283(13) A] associated
with the aryl moiety that is bound to the Os(CO)3 and
Os(CO)2 centers, respectively, confirm the metalation of this
Os-Os vector. The computed structure of B1 is in accord
with the formulated structure for the nonacarbonyl cluster,
especially with respect to the bond distances and angles exhi-
bited by the metalated η1-aryl moiety and the capping of the
triangular Os3 face by the 5e PhP(C6H4)CH2PPh2 donor
ligand. Finally, B1 compares well with the overall architec-
ture exhibited by the structurally characterized derivative
(40) (a) The concentration of CO at 288 K may be estimated at
ca. 8.0 ꢀ 10-3 M, leading to a CO/cluster ratio of ca. 8. While this ratio is
slightly lower than that technically required for a reaction run under
strict pseudo-first-order conditions (i.e., a 10-fold excess of ligand), it
nonetheless gave a kobsd value in excellent agreement with that reaction
conducted under true pseudo-first-order conditions (entry 1). (b) For
the kinetic treatment of different pseudo-first-order scenarios, see:
Espenson, J. H. Chemical Kinetics and Reaction Mechanisms; McGraw-
Hill: New York, 1995.
(41) The computed half-life of 4.3 min (17 atm CO) and 1.8 min (34
atm CO) for the two high-pressure carbonylations strongly support the
need for stopped-flow techniques if one wishes to measure accurately the
rate of the reaction at higher temperatures and CO pressure.
(42) For a general discussion on the properties and reactivity of
norbornyl cations, see: March, J. Advanced Organic Chemistry, 4th
ed.; Wiley: New York, 1992.
(43) For some representative organometallic examples involving the
windshield wiper motion of ligands, see: (a) Huang, S.-H.; Watson,
W. H.; Carrano, C. J.; Wang, X.; Richmond, M. G. Organometallics
2010, 29, 61. (b) Bergamo, M.; Beringhelli, T.; D'Alfonso, G.; Mercandelli,
P.; Sironi, A. J. Am. Chem. Soc. 2002, 124, 5117. (c) Rosenberg, E.; Milone,
L.; Gobetto, R.; Osella, D.; Hardcastle, K.; Hajela, S.; Moizeau, K.; Day, M.;
Wolf, E.; Espitia, D. Organometallics 1997, 16, 2665. (d) Curtis, M. D.; Han,
K. R.; Butler, W. M. Inorg. Chem. 1980, 19, 2096.
(45) (a) Persson, R.; Monari, M.; Gobetto, R.; Russo, A.; Aime, S.;
Calhorda, M. J.; Nordlander, E. Organometallics 2001, 20, 4150.
(b) Abdul Mottalib, M.; Begum, N.; Tareque Abedin, S. M.; Akter, T.; Kabir,
S. E.; Arzu Miah, M.; Rokhsana, D.; Rosenberg, E.; Golzar Hossain, G. M.;
Hardcastle, K. I. Organometallic 2005, 24, 4747. (c) Musaev, D. G.;
Nowroozi-Isfahami, T.; Morokuma, K.; Rosenberg, E. Organometallics
2005, 24, 5973. (d) Musaev, D. G.; Nowroozi-Isfahami, T.; Morokuma, K.;
Abedin, J.; Rosenberg, E.; Hardcastle, K. I. Organometallics 2006, 25, 203.
ꢀ
~
(e) Cabeza, J. A.; Perez-Carreno, E. Organometallics 2008, 27, 4697. (f )
ꢀ
ꢀ
~
Cabeza, J. A.; del Río, I.; Fernandez-Colinas, J. M.; Perez-Carreno, E.;
ꢀ ꢀ
Sanchez-Vega, M. G.; Vazquez-García, D. Organometallics 2009, 28, 1832.
(44) The conversion of the dimetalated aryl group in HOs3(CO)8[ μ-
PhP(C6H4-μ2,η1)CH2PPh2] to its η1 isomer can theoretically occur at
either osmium center that binds the metalated aryl group. This act will
furnish two different osmium hinge or anchor points from which the η1
isomer may transverse the original Os-Os vector. Given the steric
congestion imposed about the face of the cluster by the axially disposed
diphosphine ligand, only the η1 isomer shown in Scheme 2 is regarded as
feasible.
ꢀ
ꢀ
~
(g) Cabeza, J. A.; Fernandez-Colinas, J. F.; Perez-Carreno, E. Organome-
tallics 2009, 28, 4217. (h) Cabeza, J. A.; Van der Maelen, J. F.; García-
Granda, S. Organometallics 2009, 28, 3666. (i) Cabeza, J. A.; del Río, I.;
ꢀ
~
Goite, M. C.; Perez-Carreno, E.; Pruneda, V. Chem.;Eur. J. 2009, 15, 7339.
(46) (a) Churchill, M. R.; Hollander, F. J.; Lashewycz, R. A.; Pearson,
G. A.; Shapley, J. R. J. Am. Chem. Soc. 1981, 103, 2430. (b) Churchill, M. R.;
Hollander, F. J. Inorg. Chem. 1981, 20, 4124.

4048 Organometallics, Vol. 29, No. 18, 2010
Huang et al.
Figure 2. Optimized B3LYP structures for intermediates A1, A2, and B1 and transition states TSA1A2 and TSA2B1.
HOs3(CO)8(PPri3)[ μ-PhP(C6H4)CH2PPh2] that is obtained
from the reaction of PPri3 with HOs3(CO)8[ μ-PhP(C6H4-
μ2,η1)CH2PPh2].9b Here the addition of the PPri3 ligand to
the cluster is accompanied by a formal opening of the
metalated aryl moiety analogous to that expected for the
carbonylation product HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2].
The direct addition of CO to the Os(CO)2 center in A1 that
is opposite the face-capping diphosphine ligand seemed a
promising route to B1. The incoming CO was allowed to
approach the metal center with an axial-like trajectory, and
a two-dimensional scan was performed with constraints on
the cleavage of the Os-C(aryl) and the formation of the
Os-CO bond distances. Results of the scan revealed a
monotonic increase in the overall energy of the associative
region (long Os-C(aryl) and short Os-CO distances) of the
potential energy surface (PES) with no apparent saddle point
and energies > 50 kcal/mol for the ΔEgas. All attempts to
find a concerted transition state in this process collapsed to
product, reagents, or a third species resembling B1 but with
an unrealistically long Os-COnew bond distance. The dis-
sociation or elongation of the Os-C(aryl) bond leading to
the species HOs3(CO)8[ μ-PhP(C6H4-η1)CH2PPh2] was next
examined, and the dissociative region of the PES (long
Figure 3. Free-energy profile for conversion of A1 to B1 in the
presence of CO. Energy values are ΔG’s in kcal/mol with respect
to A1.
Os-C(aryl) and Os-COnew distances) was found to be
ꢁ
significantly lower in energy vis-a-vis the initial CO addition
route examined. The geometry-optimized structure of HOs3-
(CO)8[ μ-PhP(C6H4-η1)CH2PPh2] is represented by species
A2, and this closely resembles the structure of B1 except for
the missing CO group. The binding of the triangular metal
face by the PhP(C6H4)CH2PPh2 ligand in A2 is similar to the
diphosphine ligand in HOs3(CO)8(PPri3)[ μ-PhP(C6H4)-
CH2PPh2].9b Moreover, elongation of the Os-C(aryl) vector
TSA1A2 and TSA2B1, whose optimized geometries are
depicted in Figure 2. TSA1A2 corresponds to the breaking
of the Os-C(aryl) bond as the dimetalated aryl cluster
transforms to the η1 isomer. This transition state is fairly
late and resembles species A2 to a high degree. The non-
˚
bonding Os-C(aryl) bond distance of 3.43 A in TSA1A2 is
˚
˚
from 2.33 A (A1) to 3.99 A (A2) exceeds the sum of the van
der Waals radii of the carbon and osmium atoms, precluding
any direct bonding between these centers.47,48
significantly weakened relative to that distance found in A1
and fully consistent with this process. The addition of CO to
A2 leads to TSA2B1, which corresponds to the high-energy
point on the PES with a ΔGq value of 21.5 kcal/mol relative
to A1. TSA2B1 is an early transition state and resembles A2
with the distant CO moiety approaching the Os(CO)2 moiety
from the exposed polyhedral face of the cluster. That
TSA2B1 amounts to the rate-determining step for the overall
process is in full harmony with the reaction kinetics and the
sequence of events depicted in Scheme 2.49
Figure 3 shows the free-energy profile for the conversion
of A1 to B1, along with the pertinent transition states
(47) The opening of the dimetalated aryl group in HOs3(CO)8[ μ-
PhP(C6H4-μ2,η1)CH2PPh2] that accompanies the addition of PPri3 leads
to a concomitant increase of the initial Os-C(aryl) bond distance of
˚
˚
2.28(1) A to over 4.08 A. See: ref 9b.
(48) Bondi, A. J. Phys. Chem. 1964, 68, 441.

Article
Organometallics, Vol. 29, No. 18, 2010 4049
(η1,η3-C3H4) prior to the formation of the 32e species
Cp*2Ir2(η1,η3-C3H4) in the rate-limiting step.51
III. Synthesis of dppm-d8ortho and Deuterium Isotope Ef-
fects in the Carbonylation of DOs3(CO)8,9[ μ-(Ph-d2ortho)P-
(C6H3D)CH2PPh2-d4ortho]. The effect of deuterium on the
reductive coupling step that accompanies the formation of
Os3(CO)10(dppm) was investigated by using an isotopically
substituted dppm ligand where all of the ortho hydrogens
were replaced by deuterium atoms. The needed dppm-d8ortho
ligand was prepared according to the protocol outlined in
Scheme 3, and the introduction of deuterium was made
through a modification of the functional-group-directed
reductive debromination described by Schlosser.52,53 The
undesirable incorporation of the original amine hydrogens
into the ortho sites of the desired 2,6-dideuteroaniline during
the reductive debromination was minimized by using 2,6-
dibromoaniline-d2.54 The deuterium content at the ortho
sites in 2,6-dideuteroaniline was found to be >98% isotopic
purity by 1H NMR, and this material was immediately dia-
zotized and then converted to 2,6-dideutero-iodobenzene.
The mass spectrum of the latter compound revealed a mole-
cular ion two mass units greater than iodobenzene, and the
1H and 13C NMR properties were fully consistent with the
formulated structure. Lithiation of 2,6-dideutero-iodoben-
zene, followed by treatment with PCl3, gave PPh3-d6ortho in
an overall yield of 30%. While PPh3-d6ortho may also be
prepared through the metal-catalyzed deuteration of PPh3
using D2 gas,55 our procedure lends itself to the large-scale
synthesis of PPh3-d6ortho, eliminates the incomplete deutera-
tion of the ortho sites, and avoids any degradation of the
phosphine ligand through P-C bond cleavage. Finally,
reductive cleavage of one of the P-Ph bonds in PPh3-d6ortho
was best executed by using Na in liquid NH3, followed by
neutralization of the NaPh-d2ortho with NH4Cl and then
Figure 4. UV-vis spectral changes for HOs3(CO)9[ μ-PhP-
(C6H4)CH2PPh2] in the presence of CO (1 atm) in toluene at
334.0 K. The inset shows the absorbance data versus time for the
experimental data (9) and the nonlinear regression fit of the
first-order rate constant k (solid red line).
II. Carbonylation Kinetics for the Conversion of HOs3-
(CO)9[ μ-PhP(C6H4)CH2PPh2] to Os3(CO)10(dppm). The
carbonylation reaction leading to Os3(CO)10(dppm), and
with it the important reductive coupling step that regenerates
the aryl C-H bond, proceeds cleanly and without the pre-
sence of spectroscopically observable intermediates in toluene
over the temperature range 317-340 K. The kinetics were
investigated by UV-vis spectroscopy under CO (excess) by
following the increase in absorbance of the 358 or 425 nm
band belonging to Os3(CO)10(dppm). Figure 4 shows the
UV-vis spectral changes for the reaction of HOs3(CO)9[ μ-
PhP(C6H4)CH2PPh2] with CO at 334 K, while the inset
reveals the least-squares fit of the first-order rate constant
(k) as a function of the absorbance data. The first-order rate
constants for these reactions are summarized in Table 2.
Inspection of entries 3-5 confirms that the rate of the
reaction is independent of added CO, and this supports a
process that is first-order in starting cluster and consistent
with the rate law
1
addition of CH2Cl2, which furnished dppm-d8ortho. The H
NMR spectrum of dppm-d8ortho recorded in C6D6 showed no
detectable ortho hydrogens, and the ESI mass spectrum
showed a m/z peak at 393.47 for the species [M þ H]þ.
The starting cluster Os3(CO)10(dppm-d8ortho) was next
prepared from Os3(CO)10(MeCN)2 and dppm-d8ortho in
68% yield. The isolated product was characterized by
NMR and mass spectrometry, with the data in accord with
the formulated structure. Controlled thermolysis of Os3-
(CO)10(dppm-d8ortho) in toluene afforded the deuteride cluster
DOs3(CO)8[ μ-(Ph-d2ortho)P(C6H3D-μ2,η1)CH2PPh2-d4ortho], and
this material was judged to be very pure by 1H NMR
spectroscopy, showing no evidence for the presence of ortho
hydrogens or a bridging hydride ligand.56 The absence of the
latter ligand is important, as it allows us to eliminate a
rate ¼ k1½clusterꢁ
The Eyring activation parameters computed for the pre-
sent carbonylation [ΔHq = 24.5(6) kcal/mol and ΔSq =
-1(2) eu] are comparable to those data found by us in the
conversion of HOs3(CO)9[ μ-PhP(C6H4)CdC(PPh2)C(O)-
CH2C(O)] to Os3(CO)10(bpcd),14,50 where compelling kin-
etic evidence was presented for the formation of an inter-
mediate π-complex, prior to the rate-limiting release of the
unsaturated cluster Os3(CO)9(bpcd) (see Scheme 1). We
would also note that the activation parameters computed for
HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] are not unlike those val-
ues reported by Bergman for the reductive coupling of ben-
zene from dinuclear compound Cp*2Ir2(Ph)(μ-H)(η1,η3-
C3H4), which presumably affords Cp*2Ir2(η2-benzene)-
(51) McGhee, W. D.; Hollander, F. J.; Bergman, R. G. J. Am. Chem.
Soc. 1988, 110, 8428.
(52) Heiss, C.; Marzi, E.; Schlosser, M. Eur. J. Org. Chem. 2003, 4625.
(53) For a comprehensive treatise on the use of organozinc and
related organometallics in organic synthesis, see: Organometallics in
Synthesis: A Manual, 2nd ed.; Schlosser, M., Ed.; Wiley: Chichester,
England, 2002.
(54) Here the exchange of the original amino hydrogens by deuterium
was achieved by treating 2,6-dibromoaniline with three separate 50 mL
portions of MeOD. Each individual exchange reaction was stirred
overnight, after which the solvent was removed and fresh MeOD added
to the deuterium-enriched aniline. The ratio of deuterium:hydrogen in
the first exchange cycle was ca. 15:1, and this ratio increases with each
successive exchange cycle.
(49) From the calculated energy differences in Figure 3, we estimate
k-1(A2fTSA1A2) and k2(A2fTSA1A2) as ca. 1.8 ꢀ 10-4 s-1 and 1.2ꢀ
10-6 s-1 M-1, respectively. The rate differences reinforce our assump-
tion concerning Scheme 2 and the fact that k-1. k2. Under the present
conditions, the ligand-independent path for carbonylation would be-
come competitive with the second-order process at [CO] > 150 M.
(50) Since the transformation of HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2]
to Os3(CO)10(dppm) does not occur via a single elementary reaction, any
mechanistic inference based solely on the activation parameters is best
acknowledged tenuously.
(55) Parshall, G. W.; Knoth, W. H.; Schunn, R. A. J. Am. Chem. Soc.
1969, 91, 4990.
(56) Cf.: the 1H NMR spectrum of HOs3(CO)8[ μ-PhP(C6H4-
μ2,η1)CH2PPh2] recorded in C6D6 under identical conditions.

4050 Organometallics, Vol. 29, No. 18, 2010
Huang et al.
Scheme 3
scenario involving the proton/deuteron exchange between
the chromatographic support and the isotopically labeled
cluster.57,58 The carbonylation of DOs3(CO)8[ μ-(Ph-d2ortho)P-
(C6H3D-μ2,η1)CH2PPh2-d4ortho] to DOs3(CO)9[ μ-(Ph-d2ortho)-
P(C6H3D)CH2PPh2-d4ortho] was investigated at 288.1 K un-
der 1 atm of CO in a manner identical to that employed for
the protiated isotopomer. Two independent carbonlyations
were conducted, and each displayed well-behaved pseudo-
first-order decay curves for the starting cluster, from which the
average kobsd value of 5.16(2) ꢀ 10-4 s-1 was computed.59 This
latter value compares well with the kobsd value of 5.23(2) ꢀ
10-4 s-1 from the d0-isotopomer (Table 1) and confirms the
absence of any significant isotope effect in the first carbony-
lation step. The insensitivity of the rate constants to isotopic
substitution reinforces the carbonylation mechanism that is
depicted in Scheme 2.
display an inverse EIE when the vibrational frequencies of
the participant atoms are taken into account.63 The transfer
of the deuteride ligand from the osmium metal to the ortho
carbon atom leads to a stronger aryl C-D bond relative to
the osmium-deuteride bond in DOs3(CO)9[ μ-(Ph-d2ortho)P-
(C6H3D)CH2PPh2-d4ortho]. The faster reductive coupling
experimentally found for the deuterated isotopomer stems
from its greater preequilibrium concentration of Os3(CO)9-
[ μ-(Ph-d2ortho)P(C6H3D2)CH2PPh2-d4ortho] relative to Os3-
(CO)9[ μ-PhP(C6H5)CH2PPh2].64 Scheme 4 illustrates the
reductive coupling reaction leading to Os3(CO)9(dppm-Pa,
Pe) (where Pa = axial phosphine; Pe = equatorial phos-
phine). Not depicted in the scheme is the rapid isomerization
of the dppm ligand and the trapping of CO by Os3(CO)9-
(dppm-Pe,Pe) to give Os3(CO)10(dppm-Pe,Pe), which occur
after the rate-limiting step involving the generation of Os3-
(CO)9(dppm-Pa,Pe). Admittedly simplistic, the fundamen-
tals at play in this scenario have been shown to be operative
in the carbonylation of HOs3(CO)9[ μ-PhP(C6H4)CdC-
(PPh2)C(O)CH2C(O)] to Os3(CO)10(bpcd),14 in addition to
numerous mononuclear compounds that display a dissocia-
tive loss of an alkane and arene from hydrido(alkyl) and
hydrido(aryl) precursors.60,65 In the latter examples, transi-
ent metal-bound σ-alkane and π-arene species have been
unequivocally demonstrated as active participants in the
rate-limiting step.
The carbonylation of DOs3(CO)9[ μ-(Ph-d2ortho)P(C6H3D)-
CH2PPh2-d4ortho] to Os3(CO)10(dppm-d8ortho) was examined
by UV-vis spectroscopy at 323.0 and 334.0 K under 1 atm of
CO. The first-order rate constants of 19.1(2) e-5 s-1 (323 K)
and 62.4(5) ꢀ 10-5 s-1 (334 K) measured for the consump-
tion of DOs3(CO)9[ μ-(Ph-d2ortho)P(C6H3D)CH2PPh2-d4ortho
]
are faster than those reactions employing the d0 isotopomer
(Table 2), giving rise to an inverse isotope effect of 0.50 at
both temperatures. The observed inverse isotope effect is
readily understood in terms of a reductive coupling sequence
that contains a preequilibrium that exhibits an inverse equili-
brium isotope effect (EIE), followed by a slower rate-limiting
step that is isotope insensitive (i.e., Kk).2d,g,60,61 Using
the deuterated isotopomer for illustrative purposes, the
formation of the aryl C-D bond through reductive coupling
would yield the preequilibrium species Os3(CO)9[ μ-(Ph-
d2ortho)P(C6H3D2-π)CH2PPh2-d4ortho], whose aryl ring must
remain bound to the cluster at the locus responsible for
ortho metalation.62 Such a multistep scenario is predicted to
IV. NMR Evidence for Low-Energy H/D Scrambling in
Ortho Metalation. Additional experiments were conducted in
order to establish the existence of a labile π-arene (or agostic
C-H) compound as an intermediate that precedes ortho
metalation on the reaction coordinate. Ideally, the testing of
a putative π/agostic species may be achieved by studying the
ortho-metalation step (i.e., oxidative coupling) under kineti-
cally and thermodynamically controlled reaction conditions.
Here valuable information may be gleaned concerning the
Scheme 4

Article
Organometallics, Vol. 29, No. 18, 2010 4051
discriminatory C-H versus C-D bond activation at the
cluster. To this end, we have prepared the isotopically sub-
stituted ligand dppm-d4ortho, whose ortho sites on each aryl
ring contain one hydrogen and one deuterium relative to the
phosphorus atom. Starting from 2-bromoaniline and using the
methodology employed in the synthesis of dppm-d8ortho, we
successfully prepared the desired diphosphine in an overall
yield of 42%. The synthetic details for dppm-d4ortho are shown
in eq 3, and the NMR and mass spectral data are in accord with
the formulated structure.
(dppm-d4ortho) at temperatures in excess of 100 ꢀC is expected
to proceed with loss of CO and the formation of Os3(CO)9-
(dppm-d4ortho), the harsh conditions will afford a thermo-
dynamic mixture of the corresponding hydride and deuteride
clusters HOs3(CO)9[ μ-(Ph-d1ortho)P(C6H3D)CH2PPh-d2ortho
]
and DOs3(CO)9[ μ-(Ph-d1ortho)P(C6H4)CH2PPh-d2ortho]. No
useful information concerning the kinetic aspects of the
ortho metalation can be obtained under these conditions.
The photolysis of Os3(CO)10(dppm) was next explored as a
potential route to HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] be-
fore we conducted our experiments with the isotopi-
cally substituted cluster Os3(CO)10(dppm-d4ortho). Near-
UV photolysis of Os3(CO)10(dppm) in benzene-d6 at room
temperature in sealed NMR tubes produced liminal evidence
for ortho metalation. Purging the liberated CO from solution
with argon during photolysis facilitated the formation of the
ortho-metalated product when the sample was irradiated for
24 h, but the yield was low (<5%) and the time scale for the
reaction was deemed impractical for our studies. The argon
purge prevents the unsaturated Os3(CO)9(dppm) cluster from
scavenging the released CO and regenerating Os3(CO)10-
(dppm). Accompanying the formation of the expected
ortho-metalation product HOs3(CO)9[ μ-PhP(C6H4)CH2-
PPh2] is the dimetalated cluster HOs3(CO)8[ μ-PhP(C6H4-
μ2,η1)CH2PPh2], which presumably derives from the photo-
chemically promoted loss of CO in the former cluster. This
supposition was subsequently confirmed by a control experi-
ment where HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] was irra-
diated under similar conditions and found to give the
octacarbonyl cluster. Although no quantum limits have been
measured for these reactions, qualitatively HOs3(CO)9[ μ-
PhP(C6H4)CH2PPh2] appears to be more photosensitive
than Os3(CO)10(dppm) with respect to CO loss. Os3(CO)10-
(dppm) is extremely photosensitive to 254 nm light (UV), and
the cluster compounds HOs3(CO)8,9[ μ-PhP(C6H4)CH2PPh2]
were observed as the major products under controlled photolysis
(i.e., short irradiation times and active CO purge).
Treatment of Os3(CO)10(MeCN)2 with dppm-d4ortho furn-
ished the diphosphine-substituted cluster Os3(CO)10(dppm-
d4ortho) in 72% yield after recrystallization. This cluster
serves as a suitable platform for the investigation of the
kinetic selectivity attendant in the ortho metalation, pro-
vided the unsaturated cluster Os3(CO)9(dppm-d4ortho), the
presumed precursor to the π/agostic complex, can be gener-
ated under mild conditions. While thermolysis of Os3(CO)10-
(57) We have purified DOs3(CO)8[ μ-(Ph-d2ortho)P(C6H3D-μ2,η1)-
CH2PPh2-d4ortho] by recrystallization and column chromatography,
albeit with some material loss using the former method. Both purifica-
1
tion methods afforded indistinguishable NMR spectra. The H NMR
spectrum of the chromatographed cluster was critically scrutinized due
to reports of isotope loss from facile H/D exchange between the
stationary phase and the metal-deuteride species. See ref 58.
(58) (a) Keister, J. B.; Shapley, J. R. J. Am. Chem. Soc. 1976, 98, 1056.
(b) Andrews, M. A.; Kirtley, S. W.; Kaesz, H. D. Inorg. Chem. 1977, 16,
1556.
One more experimental aspect had to be established before
the ortho-metalation study using Os3(CO)10(dppm-d4ortho)
could be carried out. We had to be sure that the hydride and
deuteride clusters HOs3(CO)8,9[ μ-(Ph-d1ortho)P(C6H3D)CH2-
PPh-d2ortho] and DOs3(CO)8,9[ μ-(Ph-d1ortho)P(C6H4)CH2-
PPh-d2ortho] could be accurately and easily quantified by 1H
NMR spectroscopy. In principle, NMR analysis using the
hydrogens on the metalated aryl ring would permit a direct
assessment of the isotopic composition associated with the
ortho-metalation products HOs3(CO)9[ μ-(Ph-d1ortho)P(C6H3D)-
CH2PPh-d2ortho] and DOs3(CO)9[ μ-(Ph-d1ortho)P(C6H4)CH2-
PPh-d2ortho] and the corresponding octacarbonyl hydride/
deuteride clusters. The formation of the latter clusters is not
expected to involve any change in the isotope ratio from the
nonacarbonyl hydride/deuteride clusters, as the conversion
only involves a loss of CO. Unfortunately, the chemical shift
data for the aryl hydrogens on the metalated ring have not
been reported prior to this study. This problem was alle-
viated through a 1H COSY analysis of HOs3(CO)9[ μ-PhP-
(C6H4)CH2PPh2] and HOs3(CO)8[ μ-PhP(C6H4)CH2PPh2]
recorded in toluene-d8, with the spectral assignments receiv-
ing additional corroboration by the NMR data in hand from
the dppm-d8 isotopomers of these clusters (vide supra). The
structures shown below report the 1H chemical shift data for
the hydrogens on the aryl ring of interest. Both clusters exhi-
bit an ABMX spin system for the hydrogens on the metalated
ring, where the lowest field resonance may be confidently
(59) Here the kobsd values of 5.33(1) and 4.98(2) ꢀ 10-4 s-1 were
obtained by nonlinear regression analysis of the optical changes in the
600 and 386 nm band, respectively, from two different experiments.
These data were averaged to give the quoted kobsd value.
(60) (a) Parkin, G.; Bercaw, J. E. Organometallics 1989, 8, 1172.
(b) Bullock, R. M.; Headford, C. E. L.; Hennessy, K. M.; Kegley, S. E.;
Norton, J. R. J. Am. Chem. Soc. 1989, 111, 3897. (c) Janak, K. E.; Churchill,
D. G.; Parkin, G. Chem. Commun. 2003, 22. (d) Jones, W. D. Acc. Chem.
Res. 2003, 36, 140.
(61) While a single elementary process involving the reductive cou-
pling reaction cannot be definitively excluded, we consider this scenario
unlikely given the totality of the available data. See refs 1, 2, 7, and 60.
(62) We acknowledge the possibility of a cluster intermediate having
an agostic C-H(D) interaction with the aryl moiety as a viable pre-
equilibrium species, but we can make no kinetic distinction between
these two aryl-tethered species based on the data at hand.
(63) (a) Wolfsberg, M. Acc. Chem. Res. 1972, 7, 225. (b) Kubas, G. J.
Metal Dihydrogen and σ-Bond Complexes; Kluwer Academic/Plenum
Publishers: New York, 2001.
(64) For a few examples involving structurally and spectroscopically
characterized π-bound metal compounds, see: (a) Jones, W. D.; Dong,
L. J. Am. Chem. Soc. 1989, 111, 8722. (b) Cheng, T.-Y.; Szalda, D. J.;
Bullock, R. M. Chem. Chem. 1999, 1629. (c) Johansson, L.; Tilset, M.;
Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 2000, 122, 10846.
(d) Norris, C. M.; Templeton, J. L. Organometallics 2004, 23, 3101.
(65) For an example of bimetallic cooperativity in an ortho-metala-
tion process that exhibits an inverse isotope effect, see: Esswein, A. J.;
Veige, A. S.; Piccoli, P. M. B.; Schultz, A. J.; Nocera, D. G. Organome-
tallics 2008, 27, 1073.

4052 Organometallics, Vol. 29, No. 18, 2010
Huang et al.
assigned to Ha. With the exception of Hc in HOs3(CO)8[ μ-
PhP(C6H4)CH2PPh2], whose resonance is obscured by other
aryl hydrogens, all of the metalated hydrogens appear as
distinct resonances, and this allows these hydrogens to be
used as internal markers in the determination of the ortho-
metalation products generated during photolysis.
of 70:30 for the hydride:deuteride clusters. The absence of
nonacarbonyl products is attributed to the greater photo-
chemical sensitivity of the nonacarbonyl clusters, which in
turn lose CO rapidly to afford the observed octacarbonyl
clusters. Finally, we also explored the photolysis at 201 K
(dry ice/EtOH), but no ortho metalation was observed when
the sample was irradiated for an extended period of time. The
absence of ortho metalation at the lowest temperature may
reflect an enhanced geminate recombination of the released
CO and Os3(CO)9(dppm-d4ortho).68
The ineludible conclusion that must be drawn from the
above NMR experiments is the fact that the ortho-metala-
tion products have undergone thermal equilibration during
the time scale of these reactions. The initially formed pro-
ducts from C-H(D) bond activation are not stable and must
undergo a rapid and reversible equilibration via a labile
cluster intermediate(s). Scheme 5 outlines the gross aspects
of a scenario that agree with the photolysis experiments.
Optical excitation of Os3(CO)10(dppm-d4ortho) leads to loss
of CO and formation of the unsaturated cluster Os3(CO)9-
(dppm-d4ortho), followed by ortho metalation. Since the
ortho metalation mandates the capping of one of the triangular
Os3 faces by the diphosphine ligand, the ancillary diphosphine in
the Os3(CO)10(dppm-d4ortho), which is equatorially disposed,
must undergo a conformational change prior to ortho metala-
tion, a feature that is addressed by the DFT calculations
presented in the next section.
V. Calculation of the Mechanism for Ortho Metalation.
We now turn our attention to the reaction of HOs3(CO)9[ μ-
PhP(C6H4)CH2PPh2] (species B1) with CO to give Os3-
(CO)10(dppm) (species C1). The conversion of B1 to C1 is
complex and occurs in three discrete stages: (1) reductive
coupling, (2) ligand rearrangement, and (3) CO capture. The
order of these steps is critical, and the fine points of each
portion of the reaction warrant a thorough discussion. The
labeling scheme for B1 shown below will serve to aid the
forthcoming DFT discussions.
Photolysis of Os3(CO)10(dppm-d4ortho) at 254 nm in
toluene-d8 at 275 K gave HOs3(CO)9[ μ-(Ph-d1ortho)P-
(C6H3D)CH2PPh-d2ortho] and DOs3(CO)9[ μ-(Ph-d1ortho)P-
(C6H4)CH2PPh-d2ortho] in 8.8% combined yield and HOs3-
(CO)8[ μ-(Ph-d1ortho)P(C6H3D-μ2,η1)CH2PPh-d2ortho] and
DOs3(CO)8[ μ-(Ph-d1ortho)P(C6H4-μ2,η1)CH2PPh-d2ortho] in 12%
combined yield after 1 h.66,67 The hydride:deuteride isomer
ratio for both sets of products was found to be 67:33 in favor
of the hydride cluster. This ratio is identical to that value
determined earlier by us in the thermally equilibrated ortho-
metalated clusters obtained from Os3(CO)10(bpcd-d4ortho).14
Continued irradiation for an additional 1 h furnished the
above products in ca. 15% and 38% yield, respectively, with
an isomeric ratio of 70:30. No significant change in the
hydride:deuteride ratio in the cluster products was observed
when the sample was next briefly heated at 323 K. The only
noticeable reaction involved the conversion of some of the
nonacarbonyl to the octacarbonyl cluster. After heating, the
NMR tube was charged with CO and allowed to sit overnight
at room temperature in the dark. NMR analysis the following
day revealed the presence of only unreacted Os3(CO)10-
(dppm-d4ortho) and HOs3(CO)9[ μ-(Ph-d1ortho)P(C6H3D)CH2-
PPh-d2ortho] and DOs3(CO)9[ μ-(Ph-d1ortho)P(C6H4)CH2PPh-
d2ortho], as the octacarbonyl products were completely con-
sumed. The hydride:deuteride isotope ratio for the nonacar-
bonyl clusters was determined to be 72:28. We also
investigated the photolysis reaction at 243 K with the hopes
of kinetically stabilizing the initially formed ortho-metalation
products. Irradiation of Os3(CO)10(dppm-d4ortho) for 3 h,
followed by NMR analysis at the same temperature, revealed
the presence of HOs3(CO)8[ μ-(Ph-d1ortho)P(C6H3D-μ2,η1)-
That B1 functions as the nomological platform for C-H
reductive coupling, and formation of species B2 was quickly
established through an appropriate step-scan analysis. The
transition state TSB1B2 differs from B1 considerably, with
CH2PPh-d2ortho] and DOs3(CO)8[ μ-(Ph-d1ortho)P(C6H4-
μ2,η1)CH2PPh-d2ortho] in ca. 5% yield and an isotopic ratio
˚
the Os2-H1 bond distance increasing from 1.80 to 2.27 A
˚
and the Os1-H1 distance decreasing from 1.87 to 1.82 A as
(66) This experiment was also conducted with 1,4-di-tert-butylben-
zene as an internal standard, and no significant material loss was noted.
The uncertainty in the reported percent yields and the hydride:deuteride
ratios is conservatively estimated as (5%.
˚
the C1-H1 bond forms, whose distance of 1.39 A indicates
that this bond has not completely formed. A counterclock-
wise tilting of the Ph1 ring relative to the metallic plane
facilitates the reductive coupling and allows the incipient
agostic bond to adopt a favorable orientation relative to the
Os1 center.69 Figure 5 shows the free-energy profile for this
(67) Integration of the upfield hydride resonance belonging to the
clusters HOs3(CO)8,9[ μ-(Ph-d1ortho)P(C6H3D)CH2PPh-d2ortho] relative
to any of the hydrogens on the metalated aryl ring also provides an
independent check of the hydride:deuteride ratios in these two clusters.
Accurate integration information is obtained only after waiting the
customary delay period of five T1 between acquisitions. The T1 values
for the different hydrogens were determined by performing a standard
inversion-recovery experiment on HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2]
and HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] in toluene-d8. The T1
value measured from the bridging hydride group in these clusters is ca.
1.5 s and ca. 4.7 s, respectively.
(68) For a report on the temperature-dependent photochemistry in a
tetraruthenium cluster, see: Yamamoto, S.; Asakura, K.; Mochida, K.;
Nitta, A.; Kuroda, H. J. Phys. Chem. 1993, 97, 565.
(69) Here the tilt angle is defined by the atoms Os1-C1-C4 and
decreases from 176ꢀ (B1) to 161ꢀ (TSB1B2).

Article
Organometallics, Vol. 29, No. 18, 2010 4053
Scheme 5
and the subsequent steps leading to C1, with Figure 6
showing the DFT-optimized geometric structures. An IRC
calculation performed on TSB1B2 confirmed that B1 leads
to the agostic compound B2, whose Os1-H1 distance of
nearly orthogonal to the metallic plane and leads to the first
of two distinct η2-π complexes, B3 and B4. The main
difference between B3 and B4 involves the turnstile-type
rotation of the aryl moiety and the three CO groups at Os1
that leads to the exposure of the Os1 center opposite the
dppm ligand in the latter isomer. All species from B1 through
B4 and their transition states are coordinatively saturated
and contain 48e.
The second stage in the mechanistic pathway involves the
rearrangement of the dppm ligand and proceeds from B4 to
B5 via TSB4B5. This transition state represents the highest
barrier in the overall process en route to C1. TSB4B5 lies
21.6 kcal/mol above B1 in energy, and its computed value is
in good agreement with the ΔGq298 value of 25 kcal/mol extra-
polated from the experiments involving the carbonylation of
HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2] to Os3(CO)10(dppm)
(vide supra). The transformation from B4 to TSB4B5 is
accompanied by significant ligand redistribution about the
cluster polyhedron. Here, an in-plane migration of two CO
groups about the Os1-Os3 vector, coupled with the release of
the π-bound arene moiety and an axial to equatorial permu-
tation of the phosphine ligand at Os3, is reflected in the
˚
2.093 A and an Os1-H1-C1 bond angle of 109.0ꢀ are in good
agreement with the published criteria for an agostic M-
C-H linkage.70 While agostically bound substrates at mono-
nuclear compounds are well established, their observation in
metal clusters is rare. Deeming and Kabir have presented
compelling NMR data for the participation of the agostic
silane species Os3(HSiR3)(CO)9(dppm) in the reversible iso-
merization of HOs3(SiR3)(CO)9(dppm).10e B2 resembles
TSB1B2 with further tilting of the Ph1 ring to 142ꢀ and an
˚
elongation of the Os1-C1 bond to 2.67 A. The C1-H1 bond
˚
distance of 1.12 A is only slightly longer than the C-H bond
distance of 1.084(6) A found in benzene.71 Continued tilting
˚
of the Ph1 ring in B2 to 104ꢀ furnishes an aryl group that is
(70) (a) Brookhart, M.; Green, M. L.; Parkin, G. Proc. Natl. Acad.
Sci. U.S.A. 2007, 104, 6908. (b) Lein, M. Coord. Chem. Rev. 2009, 253,
625.
(71) Handbook of Chemistry and Physics, 56th ed.; Weast, R. C., Ed.;
CRC Press: Cleveland, OH, 1975.

4054 Organometallics, Vol. 29, No. 18, 2010
Huang et al.
Figure 5. Potential energy surface for conversion of B1 to C1 in the presence of CO. Energy values are ΔG’s in kcal/mol with respect to A1 (B1).
Figure 6. Optimized B3LYP structures for the intermediates B2-B6 and C1 and the corresponding transition states.
optimized structure of TSB4B5. The pairwise exchange of
CO groups about the Os1-Os3 vector is not unexpected and,
in fact, has a phenomenological foundation that is docu-
mented by 13C NMR studies on the fluxional behavior of

Article
Organometallics, Vol. 29, No. 18, 2010 4055
CO ligands in numerous polynuclear clusters.72 Os3(CO)9-
(dppm-Pa,Pe) (B5) is coordinatively unsaturated by virtue of
its 46e count, and its reaction with CO will be considered
shortly in detail. The remaining axial phosphine moiety at
Os2 relaxes to an equatorial disposition by an analogous CO
exchange process about the Os1-Os2 vector to afford Os3-
(CO)9(dppm-Pe,Pe) (B6), which lies 5.2 kcal/mol above B1 in
energy. The orientation of the three CO groups at Os1 in B6 is
similar to that found in species B4 and B5.
The formation of B5 from B4 may also be envisioned
through a process involving the localized scrambling of three
of the four ligands at Os3, and this possibility was investi-
gated accordingly.73 A tripodal exchange involving the axial
phosphine and the two equatorial CO groups furnishes B5
via the transition state TSB4B5_alt, as shown in Figure 7.
Concurrent with the 3-fold scrambling of ligands at Os3 is the
release of the π-bound aryl ring that must accompany the
phosphine moiety in its transit to the equatorial site ortho-
gonal to the Os3-Os2 vector in B5. A similar exchange of
phosphine and CO ligands at Os2 completes the reaction
sequence and gives B6. The computed energies of
TSB4B5_alt and TSB5B6_alt are nearly identical, but more
importantly they are considerably higher in energy in com-
parison with their in-plane migration counterparts already
discussed. These computational data agree with the available
VT NMR data on Os3(CO)12-nPn clusters where the tripodal
exchange of phosphine and CO ligands reveals a higher free
energy of activation compared with the in-plane migration
sequence for ligand permutation.72b,73
Figure 7. Potential energy surface for conversion of B4 to B6
depicting the lower energy in-plane migration and alternative
tripodal rotation exchange schemes. Energy values are ΔG’s in
kcal/mol with respect to A1 (B1).
Figure 8. Potential energy surface for the alternative pathway involving the attack of CO on B4 to give C1, including the lower energy
in-plane migration and alternative tripodal rotation isomerization sequences. Energy values are ΔG’s in kcal/mol with respect to A1 (B1).

4056 Organometallics, Vol. 29, No. 18, 2010
Huang et al.
associative pathway where the rate-limiting step involves the
attack of CO on B4.76 The Os1-π interaction has decreased
slightly on the basis of the Os1-C2 and Os1-C3 bond
˚
distances of 3.97 and 4.20 A, respectively, and TSB4C2 lies
25.9 kcal/mol above B1 in energy. Figures 8 and 9 illustrate
these data. The dppm ligand exhibits axial coordination in
C2, and this addition product is an isomer of C1. Tandem in-
plane migrations of axial CO groups across the Os1-Os2 and
Os1-Os3 vectors afford C3 and C1, respectively, and this
process serves as the lowest energy manifold for the stereo-
chemical permutation of the two Ph2P moieties. The alter-
native process involving sequential tripodal rotations proceeds
through the higher energy transition states TSC2C3_alt and
TSC3C1_alt and mimics those trends found in the conversion
of B4 f B5 f B6. Energetically, the computed ΔGq value of
25.9 kcal/mol (TSB4C2) is close to that computed for the
dissociative manifold involving B4 f TSB4B5 (ΔGq = 21.6),
making a mechanistic distinction problematic in the absence
of kinetic data.77 Fortunately, the carbonylation kinetics
confirmed the zero-order dependence of CO on the reaction,
which allows us to eliminate the associative process from
consideration.
Figure 9. Fully optimized B3LYP structures for intermediates
C2 and C3 and the corresponding transition states along the
alternative pathway from B4 to C1.
The scavenging of CO by B6 completes the carbonylation
reaction and affords the known cluster Os3(CO)10(dppm-Pe,Pe)
(C1). The transition state associated with this substitution,
TSB6C1, occurs early and is endergonic by ca. 10 kcal/mol
relative to B6. The optimized structure computed for C1
shows an excellent correspondence with the solid-state struc-
ture for Os3(CO)10(dppm).74 As mentioned earlier, the rate-
limiting step for the conversion of B1 to C1 is TSB4B5, with
an overall ΔGq = 21.64 kcal/mol. The mechanistic pathway
reveals a first-order dependence on B1 and is independent of
CO, features in accord with the experimental data. More-
over, isotopic substitution of H1 with deuterium gave a
ΔΔGq of 0.71 kcal/mol and an inverse isotope effect of 0.30
consistent with the experimentally measured isotope effect of
0.50.75 The nature of the species in the preequilibrium, which
was initially assumed to involve only a single π intermediate,
may now be thoroughly addressed. The equilibria between
species B1-B4 give rise to the observed and computed EIE,
and the events depicted in Figure 5 nicely account for the
thermodynamic hydride/deuterium equilibration found in
the photolysis of Os3(CO)10(dppm-d4ortho). The transforma-
tion of B4 to B5 shows no significant isotope effect (ΔΔGq =
0.01 kcal/mol), as expected for a process that does not
involve the breaking of any isotopically sensitive bonds.
In this final section we address the reaction of CO with B4
and probe the dynamics for the CO-promoted displacement
of the π-bound aryl moiety from the Os1 center. CO addition
to the Os1 center in B4 affords the transition state TSB4C2
and resembles those transition states computed for the
addition of CO to A2 and B6. This scenario is akin to an
Conclusions
The carbonylation kinetics for HOs3(CO)8[ μ-PhP(C6H4-
μ2,η1)CH2PPh2] to HOs3(CO)9[ μ-PhP(C6H4)CH2PPh2]
have been investigated, and the reaction was found to be
first-order in cluster and CO and does not exhibit any
noticeable isotope effect. Of the two kinetically indistin-
guishable processes supported by the carbonylation data,
DFT calculations have proven invaluable in terms of un-
raveling the kinetic enigma. It is shown that the dimetalated
aryl moiety in HOs3(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] is
in equilibrium with the isomeric cluster HOs3(CO)8[ μ-PhP-
(C6H4-η1)CH2PPh2], and it is the latter species that is attacked
by CO. The carbonylation of HOs3(CO)9[ μ-PhP(C6H4)CH2-
PPh2] to Os3(CO)10(dppm) is independent of CO and dis-
plays an equilibrium isotope effect of 0.50. These data
support a reductive coupling scenario involving, at a mini-
mum, a preequilibrium species that precedes the rate-limiting
formation of Os3(CO)9(dppm). DFT calculations indicate
that C-H bond formation gives rise to an agostic inter-
mediate, followed by two discrete π-aryl cluster species
before the generation of the isomeric clusters Os3(CO)9-
(dppm). The energy barrier between the different agostic and
π intermediates is low (<5 kcal/mol), and this accounts for
the kinetic lability displayed by these species. The agreement
between the experimental and computational data is im-
pressive, and the utility of DFT calculations to provide
unprecedented insight into complex bond-forming and
bond-breaking reactions occurring at polynuclear metal
clusters cannot be overemphasized.
(72) (a) Adams, R. D.; Cotton, F. A. In Dynamic Nuclear Magnetic
Resonance Spectroscopy; Jackman, L. M., Cotton, F. A., Eds.; Academic
Press: New York, 1975; Chapter 12. (b) Alex, R. F.; Pomeroy, R. K.
Organometallics 1987, 6, 2437. (c) Cooke, J.; Takats, J. Organometallics
1995, 14, 698.
Acknowledgment. Financial support from the Robert
A. Welch Foundation (Grants A-0648-MBH and B-1093-
MGR) and the National Science Foundation (CHE-
0518074, CHE-0541587, and DMS-0216275) is greatly
appreciated. Profs. Shariff E. Kabir and Manfred Schlosser
(73) This process has also been referred to as a restricted trigonal twist
and has been observed in different Os3(CO)12-nPn clusters (where n =
1-4) containing monodentate phosphine/phosphite ligands. See: (a)
Deeming, A. J.; Donovan-Mtunzi, S.; Kabir, S. E. J. Organomet. Chem.
1985, 281, C43. (b) Ref 72b.
(74) Azam, K. A.; Hursthouse, M. B.; Kabir, S. E.; Abdul Malik,
K. M.; Abdul Mottalib, M. J. Chem. Crystallogr. 1999, 29, 813.
(75) A rate-limiting step involving the conversion of B1 to B2 may be
eliminated from consideration on the basis of the computed isotope
effect. A normal KIE (ΔΔGq = -0.60) has been computed for the single-
step reductive coupling process, clearly incongruent with the isotope
ratio experimentally determined for the carbonylation.
(76) Darensbourg, D. J. Adv. Organomet. Chem. 1982, 21, 113.
(77) The isotope effect for the associative process has been computed
as 0.29, and this reflects the preequilibria involving species B1-B4. A
mechanistic distinction based solely on this criterion is not possible.

Article
Organometallics, Vol. 29, No. 18, 2010 4057
are thanked for providing an initial sample of HOs3-
(CO)8[ μ-PhP(C6H4-μ2,η1)CH2PPh2] and helpful comments
concerning the functional-group-directed deuteration of halo
arenes, respectively. We also acknowledge the reviewers
for helpful comments. Ms. Nicole Ledbetter (UNT) and
Dr. Yongxuan Su (UCSD) are thanked for their assistance
in recording the reported mass spectra of our isotopically
substituted ligands and cluster compounds.
Supporting Information Available: Experimental details re-
lated to the synthesis of the deuterium-substituted intermediates
employed in the preparation of the dppm-d8ortho and dppm-
d4ortho ligands, spectroscopic data for selected clusters, plots
of selected kinetic data, complete ref 25, 2-D scan for the A1 to
B1 reaction surface, and atomic coordinates of all optimized
stationary points and transition states. These materials are
available free of charge via the Internet at http://pubs.acs.org.