Experimental and computational studies of the isomerization reactions of bidentate phosphine ligands in triosmium clusters: Kinetics of the rearrangements from bridged to chelated isomers and X-ray structures of the clusters Os3(CO)10(dppbz), 1,1-Os3(CO) 10(dppbzF4), HOs3(CO)9[μ-1,2- PhP(C6H4

Organometallics 2011, 30, 1253–1268 1253

DOI: 10.1021/om100783f

Experimental and Computational Studies of the Isomerization Reactions

of Bidentate Phosphine Ligands in Triosmium Clusters: Kinetics of the

Rearrangements from Bridged to Chelated Isomers and X-ray Structures

of the Clusters Os₃(CO)₁₀(dppbz), 1,1-Os₃(CO)₁₀(dppbzF₄),

HOs₃(CO)₉[μ-1,2-PhP(C₆H₄-η¹)C₆H₄PPh₂], and HOs₃(CO)₉-

[μ-1,2-PhP(C₆H₄-η¹)C₆F₄PPh₂]

Xue Zhang,^†,‡Srikanth Kandala,^†Li Yang,^†William H. Watson,^§Xiaoping Wang,

David A. Hrovat,^†,‡Weston Thatcher Borden,*^,†,‡and Michael G. Richmond*^,†

^†Department of Chemistry and ^‡Center for Advanced Scientific Computing and Modeling, University of

North Texas, Denton, Texas 76203-5017, United States, ^§Department of Chemistry, Texas Christian

University, Fort Worth, Texas 76129, United States, and Neutron Scattering Science Division, Oak Ridge

National Laboratory, Oak Ridge, Tennessee 37831, United States

Received August 11, 2010

The diphosphine ligand 1,2-bis(diphenylphosphino)benzene (dppbz) reacts with the activated

cluster 1,2-Os₃(CO)₁₀(MeCN)₂(1) at room temperature to furnish a mixture of the triosmium

clusters 1,2-Os₃(CO)₁₀(dppbz) (2) and 1,1-Os₃(CO)₁₀(dppbz) (3), along with a trace amount of the

hydride cluster HOs₃(CO)₉[μ-1,2-PhP(C₆H₄-η¹)C₆H₄PPh₂] (4). The dppbz-bridged cluster 2 forms as

the kinetically controlled product and irreversibly transforms to the corresponding chelated isomer 3

at ambient temperature. The disposition of the dppbz ligand in 2 and 3 has been established by X-ray

crystallography and ³¹P NMR spectroscopy, and the kinetics for the conversion 2 f 3 have been

followed by UV-vis spectroscopy in toluene over the temperature range 318-343 K. The calculated

activation parameters (ΔH^q= 21.6(3) kcal/mol; ΔS^q= -11(1) eu) and lack of CO inhibition support

an intramolecular isomerization mechanism that involves the simultaneous migration of phosphine

and CO groups about the cluster polyhedron. The reaction between 1 and the fluorinated dipho-

sphine ligand 1,2-bis(diphenylphosphino)tetrafluorobenzene (dppbzF₄) was examined under similar

reaction conditions and was found to afford the chelated cluster 1,1-Os₃(CO)₁₀(dppbzF₄) (6) as the

sole observable product. The absence of the expected bridged isomer 1,2-Os₃(CO)₁₀(dppbzF₄) (5)

suggests that the dppbzF₄ligand destabilizes 5, thus accounting for the rapid isomerization of 5 to 6.

Near-UV irradiation of clusters 3 and 6 leads to CO loss and ortho metalation of an ancillary aryl

group. The resulting hydride clusters 4 and HOs₃(CO)₉[μ-1,2-PhP(C₆H₄-η¹)C₆F₄PPh₂] (7) have been

isolated and fully characterized by spectroscopic and X-ray diffraction analyses. Both 4 and 7 react

with added CO under mild conditions to regenerate 3 and 6, respectively, in quantitative yield. The

rearrangements of bridged to chelated diphosphine complexes in this genre of decacarbonyl clusters

have been investigated by DFT calculations. The computational results support a concerted process,

involving the scrambling of equatorial CO and phosphine groups via a classical merry-go-round

exchange scheme. The barriers computed for this mechanism agree well with those that have been

measured, and steric compression within the bridged diphosphine groups of the reactants has been

calculated to reduce the barrier heights for the rearrangement.

Introduction

dissociation/association process (i.e. ligand on, ligand off),²

the permutation of a terminally bound (η¹) ligand(s) between

Ligand fluxionality in polynuclear metal compounds is a

well-established phenomenon. Numerous reports that un-

equivocally demonstrate the nonrigidity of CO, NO, RNC,

and hydride ligands may be found in the chemical literature

dating back to the 1960s.¹In the absence of a ligand

(1) (a) Cotton, F. A.; Kruczynski, L.; Shapiro, B. L.; Johnson, L. F. J.

Am. Chem. Soc. 1972, 94, 6191. (b) 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. (c)

Farrugia, L. J. Dalton Trans. 1997, 1783. (d) Green, M.; Mead, K. A.;

Mills, R. M.; Salter, I. D.; Stone, F. G. A.; Woodward, P. Chem.

Commun. 1982, 51. (e) Musaev, D. G.; Nowroozi-Isfahani, T.; Mor-

okuma, K.; Abedin, J.; Rosenberg, E.; Hardcastle, K. I. Organometallics

2006, 25, 203.

*To whom correspondence should be addressed. W.T.B.: tel, 940-565-

3658; e-mail, borden@unt.edu. M.G.R.: tel, 940-565-3548; e-mail,

cobalt@unt.edu.

r

2011 American Chemical Society

Published on Web 02/22/2011

pubs.acs.org/Organometallics

1254 Organometallics, Vol. 30, No. 5, 2011

Zhang et al.

different metal centers is assumed to proceed via an inter-

mediate or transition structure, involving the simultaneous

coordination of the ligand to two or more metal centers. The

exact degree of ligand bridging depends upon the nuclearity

of the metallic ensemble and the energetics of ligand

exchange.³The bulk of the published studies of intramole-

cular scrambling of an ancillary ligand(s) about a metal

cluster indicate at least one terminal-bridge-terminal

transposition of the migratory ligand(s). These ligands ex-

hibit a high degree of coordinative flexibility, and they

display a gamut of bonding modes from terminal (η¹) to

multisite, face-capping (μ₄) coordination.^3,4

coordinatively flexible, ancillary phosphine ligand in a poly-

nuclear cluster.¹¹A subsequent kinetic investigation pro-

vided compelling evidence for a nondissociative mechanism

for the migration of the dppe ligand in the Ru₄cluster.¹²

Other early examples involving phosphine fluxionality in di-

and trimetallic compounds have been reported, and these

include the reversible migration of phosphine ligands about

the triangular Pt₃face of [Pt₃(μ₃-CO)P(μ-dppm)₃]^2þ(where

P = monodentate phosphine, phosphite) and the “end-over-

end” exchange of one of the bridging diphosphine ligands in

the compounds [Ag₂(dppm)₃]^2þ, [Ag₂(dmpm)₃]^2þ, and

[Ag₂(dppm)_3-x(dmpm)_x]^2þ(where x = 1, 2).^13-16

Tertiary phosphine and phosphite ligands are regarded as

good σ donors and they can also function as π-acceptor

ligands, depending upon the nature of the ancillary

substituents.⁵Historically, such pnictogen ligands have been

assumed to serve as terminal or η¹ligands, and their co-

ordination chemistry has been assumed to be restricted to a

single metal site, irrespective of the nuclearity of the

compound.⁶This provincial view, in which a phosphine acts

as a 2e donor through electron-precise bonding to only a

single metal center, has recently been challenged, and crystal-

lographic and solution NMR data have provided irrefutable

confirmation for the ligation of two or more metal centers by

PF₃and PR₃ligands.

The first structural demonstration of nontraditional bond-

ing between a tertiary phosphine and a polynuclear cluster

may be traced to the study by Balch of the reaction of Pd₂(μ-

dppm)₂Cl₂with PF₃, which affords the trinuclear cluster

[Pd₃(μ₃-PF₃)(μ-Cl)(μ-dppm)₃]^þ.⁷The capping of one of the

triangular faces in the Pd₃cluster by the ancillary PF₃ligand

defies description within a valence-bond perspective. An

electron-deficient or delocalized model of bonding is re-

quired in order to describe the interaction involving the

metallic core and the ancillary PF₃ligand.⁸Other notable

structures, containing dimetallic motifs that are bridged by a

In 2000, Adams presented landmark evidence for the

intramolecular exchange of the PMe₂Ph ligand about plati-

num and ruthenium sites in the heterometallic cluster PtRu₅-

(μ₆-C)(CO)₁₅(PMe₂Ph), as depicted in eq 1.^17,18The

scrambling of the phosphine ligand about the cluster poly-

hedron was investigated by VT NMR measurements, and

¹⁹⁵Pt satellites for the different PMe₂Ph isomers were ob-

served throughout the exchange process. This latter feature is

important, as it eliminates an isomerization process arising

from a ligand dissociation/association scenario.¹⁹A me-

chanism involving a μ₂-PMe₂Ph intermediate, coupled with

a series of CO shifts across Pt-Ru and Ru-Ru bonds, was

proposed for the dynamic isomerization that leads to the

equilibration of the phosphine ligand between the different

metal centers.

tertiary phosphine group(s), have been reported by the

9,10

ꢀ

Werner and Reau groups.

The bridging-to-chelating isomerization of the dipho-

sphine ligand in H₄Ru₄(CO)₁₀(dppe), discovered by Shapley

and Churchill, represents the first demonstrable example of a

(11) Churchill, M. R.; Lashewycz, R. A.; Shapley, J. R.; Richter, S. I.

Inorg. Chem. 1980, 19, 1277.

(12) Kandala, S.; Richmond, M. G. Inorg. Chem. 2006, 45, 5976.

(13) (a) Bradford, A. M.; Jennings, M. C.; Puddephatt, R. J. Orga-

ꢀ

(2) For ligand substitution reactions involving dissociative and asso-

ciative manifolds, see: (a) Vanni, H.; Merbach, A. E. Inorg. Chem. 1979,

18, 2758. (b) Kowaleski, R. M.; Basolo, F.; Trogler, W. C.; Ernst, R. D.

J. Am. Chem. Soc. 1986, 108, 6046. (c) Turaki, N. N.; Huggins, J. M.;

Lebioda, L. Inorg. Chem. 1988, 27, 424.

(3) Band, E.; Muetterties, E. L. Chem. Rev. 1978, 78, 639.

(4) Johnson, B. F. G.; Rodgers, A. In The Chemistry of Metal Cluster

Complexes; Shriver, D. F., Kaesz, H. D., Adams, R. D., Eds.; VCH:

New York, 1990; Chapter 6.

(5) Orpen, A. G.; Connelly, N. G. Chem. Commun. 1985, 1310.

(6) Braunstein, P.; Boag, N. M. Angew. Chem., Int. Ed. 2001, 40,

2427.

(7) Balch, A. L.; Davis, B. J.; Olmstead, M. M. J. Am. Chem. Soc.

1990, 112, 8592. Inorg. Chem. 1993, 32, 3937.

(8) (a) Hoffmann, R. Angew. Chem., Int. Ed. 1982, 21, 711. (b)

Mingos, D. M. P.; Wales, D. J. Introduction to Cluster Chemistry;

Prentice Hall: Englewood Cliffs, NJ, 1990.

(9) (a) Pechmann, T.; Brandt, C. D.; Werner, H. Angew. Chem., Int.

Ed. 2000, 39, 3909. (b) Pechmann, T.; Brandt, C. D.; Roger, C.; Werner,

H. Angew. Chem., Int. Ed. 2002, 41, 2301. (c) Pechmann, T.; Brandt,

C. D.; Werner, H. Dalton Trans. 2004, 959. (d) Pechmann, T.; Brandt,

C. D.; Werner, H. Chem. Eur. J. 2004, 10, 728.

nometallics 1988, 7, 792. (b) Bradford, A. M.; Douglas, G.; Manojlovic-

Muir, L.; Muir, K. W.; Puddephatt, R. J. Organometallics 1990, 9, 409.

(14) Dean, P. A. W.; Vittal, J. J.; Srivastava, R. S. Can. J. Chem. 1987,

65, 2628.

(15) For some additional early reports that address the dynamic

fluxionality of different pnictogen-based ligands at polynuclear systems,

see: (a) Sutin, K. A.; Kolis, J. W.; Mlekuz, M.; Bougeard, P.; Sayer,

B. G.; Quilliam, M. A.; Faggiani, R.; Lock, C. J. L.; McGlinchey, M. J.;

Jaouen, G. Organometallics 1987, 6, 439. (b) Deeming, A. J.; Smith,

M. B. Dalton Trans. 1993, 3383. (c) Yang, K.; Smith, J. M.; Bott, S. G.;

Richmond, M. G. Organometallics 1993, 12, 4779. (d) Yang, K.; Bott,

S. G.; Richmond, M. G. Organometallics 1994, 13, 3788. (e) Laurenczy,

G.; Bondietti, G.; Ros, R.; Roulet, R. Inorg. Chim. Acta 1996, 247, 65.

(16) For a related report describing the chelate f bridge isomeriza-

tion of the dppe ligands in Mo₂Br₄(dppe)₂that proceeds via the internal

rotation of the Mo₂moiety within the cubic cage of the ancillary ligands,

see: Agaskar, P. A.; Cotton, F. A.; Derringer, D. R.; Powell, G. L.; Root,

D. R.; Smith, T. J. Inorg. Chem. 1985, 24, 2786.

(17) (a) Adams, R. D.; Captain, B.; Fu, W.; Pellechia, P. J. Chem.

Commun. 2000, 937. (b) ibid. Inorg. Chem. 2003, 42, 3111.

(18) For other relevant studies dealing with the fluxionality of a

phosphine ligand at di- and polynuclear systems: (a) Ohki, Y.; Suzuki,

H. Angew. Chem., Int. Ed. 2002, 41, 2994. (b) Gallo, V.; Mastrorilli, P.;

Nobile, C. F.; Braunstein, P.; Englert, U. Dalton Trans. 2006, 2343. (c)

Begum, N.; Das, U. K.; Hassan, M.; Hogarth, G.; Kabir, S. E.; Nord-

lander, E.; Rahman, M. A.; Tocher, D. A. Organometallics 2007, 26,

6462. (d) Esswein, A. J.; Veige, A. S.; Piccoli, P. M. B.; Schultz, A. J.;

Nocera, D. G. Organometallics 2008, 27, 1073.

(10) (a) Sauthier, M.; Le Guennic, B.; Deborde, V.; Toupet, L.; Halet,

ꢀ

J.-F.; Reau, R. Angew. Chem., Int. Ed. 2001, 40, 228. (b) Leca, F.;

Sauthier, M.; Deborde, V.; Toupet, L.; Reau, R. Chem. Eur. J. 2003, 9,

ꢀ

3785. (c) Leca, F.; Lescop, C.; Rodriguez-Sanz, E.; Costuas, K.; Halet,

ꢀ

J.-F.; Reau, R. Angew. Chem., Int. Ed. 2005, 44, 4362. (d) Nohra, B.;

Rodriguez-Sanz, E.; Lescop, C.; Reau, R. Chem. Eur. J. 2008, 14, 3391.

ꢀ

Article

One of us (M.G.R.) has held a long-term interest in the

Organometallics, Vol. 30, No. 5, 2011 1255

of bridged to chelated isomers, we investigated the nature of

the preferred pathway that is responsible for nondissociative

ligand isomerization in this genre of clusters.

synthesis and mechanistic study of Ru₃, Os₃, and tetrahe-

drane clusters, containing rigid diphosphine ligands.^15c,20-22

Several of these reports have addressed the dynamic isomer-

ization of the ancillary diphosphine ligand about the cluster

core, as illustrated for the cluster compound Os₃(CO)₁₀-

(dppen) (where dppen = (Z)-1,2-bis(diphenylphosphino)-

ethylene) in eq 2. This particular reaction is reversible, and

the thermodynamically favored isomer contains a chelating

dppen ligand with a K_eqvalue of 6.9 over the temperature

range 358-373K.^21aThe dppen isomerization, from bridging

to chelating, is unaffected by added CO, and the activation

parameters for the forward and reverse reactions indicate an

ordered transition state. The reaction kinetics are consistent

with a nondissociative scenario for this rearrangement, re-

quiring, at a minimum, the formal site exchange of at least

one phosphine and one CO across an Os-Os bond, presum-

ably by way of a transient μ₂-phosphine/μ₂-CO species.

Experimental Section

General Methods. The parent cluster Os₃(CO)₁₂was prepared

from OsO₄and CO,²³and this cluster was used in the prepara-

tion of the bis(acetonitrile) cluster 1,2-Os₃(CO)₁₀(MeCN)₂.²⁴

The dppbzF₄ligand was synthesized from 1,2-dibromotetra-

fluorobenzene and Ph₂PCl according to the procedure of

Meek.²⁵The OsO₄used in our studies was purchased from

Engelhard Chemical Co., and the chemicals Me₃NO xH₂O,

3

dppbz, 1,2-dibromotetrafluorobenzene, chlorodiphenylpho-

sphine, and BuLi (2.5 M in hexanes) were purchased from

Aldrich Chemical Co. The anhydrous Me₃NO employed in

our studies was obtained from Me₃NO xH₂O, after the waters

3

of hydration were azeotropically removed under reflux, using

benzene as a solvent. All reaction solvents were distilled from a

suitable drying agent under argon or obtained from an Innova-

tive Technology (IT) solvent purification system. When not in

use, all purified solvents were stored in Schlenk storage vessels

equipped with high-vacuum Teflon stopcocks.²⁶The NMR

solvents CDCl₃and C₆D₆were purified by bulb-to-bulb dis-

tillation from P₂O₅and sodium/benzophenone ketyl, respec-

tively. Combustion analyses were performed by Atlantic

Microlab, Norcross, GA.

The IR spectra were recorded on a Nicolet 6700 FT-IR

spectrometer in amalgamated NaCl cells having a 0.1 mm path

length, while the quoted ¹H NMR data were recorded at 400 and

500 MHz on Varian VXR-400 and VXR-500 spectrometers,

respectively. The ³¹P NMR spectra were recorded at 202 MHz

on the latter spectrometer. These data were collected in the

proton-decoupled mode and are referenced to external H₃PO₄

(85%), whose chemical shift was set at δ 0.

Preparation of 1,2-Os₃(CO)₁₀(dppbz) from 1,2-Os₃(CO)₁₀-

(MeCN)₂and dppbz. To 0.25 g (0.27 mmol) of 1,2-Os₃(CO)₁₀-

(MeCN)₂in a Schlenk tube was added 0.14 g (0.31 mmol) of

dppbz, followed by 30 mL of CH₂Cl₂via syringe. The solution

was stirred for 3 h at room temperature and then examined by

TLC analysis (1/1 CH₂Cl₂/hexane), which revealed the con-

sumption of the starting cluster, a trace amount of unreacted

dppbz (R_f= 0.53), and a yellow spot (R_f= 0.48) corresponding

primarily to 2 (78%), along with minor amounts of 3 (20%) and

4 (ca. 2%), whose presence was established by NMR spectros-

copy. The ternary cluster mixture, composed of 2-4, could not

be separated into individual components; these products were

isolated as a mixture by flash-column chromatography over

alumina at 0 ꢀC, using the aforementioned eluent. Spectroscopic

data for cluster 2²⁷are as follows. IR (CH₂Cl₂): ν(CO) 2092 (s),

2041 (m), 2013 (s), 2004 (m), 1973 (w) cm^-1. ¹H NMR (CDCl₃):

δ 6.78-7.50 (m, aryl). ³¹P NMR (CDCl₃): δ 6.41 (s).

In order to establish the influence that the differences

between closely related diphosphine ligands have on the

reactivity of the isomeric Os₃(CO)₁₀(P-P) clusters, we have

investigated the reaction of 1,2-bis(diphenylphosphino)-

benzene (dppbz) and 1,2-bis(diphenylphosphino)tetrafluoro-

benzene (dppbzF₄) with the triosmium cluster 1,2-Os₃-

(CO)₁₀(MeCN)₂(1). The dppbz ligand reacts with 1,2-Os₃-

(CO)₁₀(MeCN)₂to furnish bridged 1,2-Os₃(CO)₁₀(dppbz) as

the kinetic product of ligand substitution, and the kinetics

for isomerization to the corresponding chelated isomer, 1,1-

Os₃(CO)₁₀(dppbz), have been followed by UV-vis spectros-

copy. In contrast to the reaction of 1 with dppbz, the reaction

between 1 and dppbzF₄affords chelated 1,1-Os₃(CO)₁₀-

(dppbzF₄) as the sole observable product. Presumably, the

bridged isomer is again formed but rearranges rapidly to the

chelated isomer.

We have performed DFT calculations, in order to explore

the effect that different diphosphine ligands have on the

thermodynamic and kinetic stabilities of the bridged and

chelated isomers in the Os₃(CO)₁₀(P-P) clusters. In the

course of establishing the mechanism for the rearrangement

Isomerization of 1,2-Os₃(CO)₁₀(dppbz) to 1,1-Os₃(CO)₁₀-

(dppbz) under CO. The above cluster mixture (ca. 0.33 g) was

heated in toluene at 60 ꢀC under 1 atm of CO for 24 h, after

which time ³¹P NMR analysis of the solution revealed the

presence of only cluster 3. The solvent was removed under

vacuum and the residue chromatographed over alumina using

a 1/1 CH₂Cl₂/hexane mixture. The crude product was recrys-

tallized from CH₂Cl₂and hexane at 5 ꢀC to afford 3 as a dark

yellow solid. Yield: 0.31 g (94%). IR (CH₂Cl₂): ν(CO) 2094 (s),

(19) Jesson, J. P.; Muetterties, E. L. In Dynamic Nuclear Magnetic

Resonance Spectroscopy; Jackman, L. M., Cotton, F. A., Eds.; Aca-

demic Press: New York, 1975; Chapter 8.

(20) For Ru₃clusters: (a) Shen, H.; Bott, S. G.; Richmond, M. G.

Organometallics 1995, 14, 4625. (b) Bott, S. G.; Shen, H.; Senter, R. A.;

Richmond, M. G. Organometallics 2003, 22, 1953. (c) Bott, S. G.; Shen,

H.; Richmond, M. G. J. Organomet. Chem. 2004, 689, 3426. 2005, 690,

3838. (d) Bott, S. G.; Shen, H.; Huang, S.-H.; Richmond, M. G. J.

Organomet. Chem. 2008, 693, 2327.

(21) For Os₃clusters: (a) Watson, W. H.; Wu, G.; Richmond, M. G.

Organometallics 2005, 24, 5431. (b) Watson, W. H.; Wu, G.; Richmond,

M. G. Organometallics 2006, 25, 930. (c) Watson, W. H.; Poola, B.;

Richmond, M. G. J. Organomet. Chem. 2006, 691, 4676. (d) Watson,

W. H.; Poola, B.; Richmond, M. G. Polyhedron 2007, 26, 3585.

(22) For tetrahedrane clusters: (a) Bott, S. G.; Yang, K.; Talafuse,

K. A.; Richmond, M. G. Organometallics 2003, 22, 1383. (b) Bott, S. G.;

Yang, K.; Richmond, M. G. J. Organomet. Chem. 2005, 690, 3067; 2006,

691, 3771 (c) Watson, W. H.; Bodige, S. G.; Ejsmont, K.; Liu, J.;

Richmond, M. G. J. Organomet. Chem. 2006, 691, 3609.

(23) Drake, S. R.; Loveday, P. A. Inorg. Synth. 1990, 28, 230.

(24) Nicholls, J. N.; Vargas, M. D. Inorg. Synth. 1989, 26, 289.

(25) Eller, P. G.; Meek, D. W. J. Organomet. Chem. 1970, 22, 631.

(26) Shriver, D. F. The Manipulation of Air-Sensitive Compounds;

McGraw-Hill: New York, 1969.

(27) A relatively pure IR spectrum of cluster 2 was obtained after

suitable spectral subtraction of clusters 3 and 4 from the initial ternary

cluster mixture.

1256 Organometallics, Vol. 30, No. 5, 2011

Zhang et al.

Table 1. X-ray Crystallographic Data and Processing Parameters for the dppbz-Substituted Clusters 2 CH₂Cl₂, 3, and 4 CH₂Cl₂

3

4 CH₂Cl₂

2 CH₂Cl₂

3

CCDC entry no.

cryst syst

space group

a, A

b, A

c, A

R, deg

β, deg

770699

triclinic

P1

770697

monoclinic

P2₁/c

10.983(2)

20.218(3)

17.530(3)

770702

orthorhombic

Pbca

21.745(2)

16.774(2)

21.934(2)

11.9406(3)

13.2173(4)

15.0648(4)

86.091(1)

70.006(1)

66.423(1)

2041.2(1)

C₄₁H₂₆Cl₂O₁₀Os₃P₂

1382.06

96.856(3)

γ, deg

V, A³

3865(1)

C₄₀H₂₄O₁₀Os₃P₂

1297.13

8001(1)

C₄₀H₂₆Cl₂O₉Os₃P₂

1354.05

mol formula

fw

formula units per cell (Z)

D_calcd, Mg/m³

λ(Mo KR), A

μ, mm^-1

2

2.249

0.710 73

9.581

4

2.229

0.710 73

9.979

8

2.248

0.710 73

9.773

abs cor

F(000)

semiempirical from equivalents

1288

empirical from equivalents

2408

semiempirical

5040

cryst size, mm

abs cor factor

total no. of rflns

no. of indep rflns

no. of data/residuals/params

R1^a(I g 2σ(I))

wR2^b

0.11 ꢀ 0.08 ꢀ 0.06

0.5971/0.4188

24 301

0.28 ꢀ 0.14 ꢀ 0.03

0.9942/0.3008

32 839

0.24 ꢀ 0.17 ꢀ 0.06

0.5917/0.2026

14 706

8375

9099

14 832

8375/0/523

0.0140

0.0322

9099/0/496

0.0301

0.0492

0.905

0.902, -1.218

14 832/4/516

0.0388

0.0697

1.028

2.434, -2.301

GOF on F²

max, min ΔF, e/A³

1.012

1.114, -0.402

P

^aR1 = ||F_o| - |F_c||/ |F_o|. ^bwR2 = { [w(F_o²- F_c₂)²/ [w(F_o²)²]}^1/2

.

2040 (m), 2011 (s), 2001 (m), 1971 (w) cm^-1. ¹H NMR (C₆D₆): δ

6.78 (unresolved multiplet, 2H, aryl platform), 6.94 (t, 4H, para,

aryl, J_H-H= 7.3 Hz), 7.04 (t, 8H, meta, aryl, J_H-H= 7.3 Hz),

7.30 (unresolved multiplet, 2H, aryl platform), 7.49 (bm, 8H,

ortho, aryl). ³¹P NMR (C₆D₆): δ 26.32 (s). Anal. Calcd (found)

meta, J_H-H= 8.0 Hz, J_P-H= 2.7 Hz), 7.76 (dd, 8H, ortho,

_P-H= 11.5 Hz, J_H-H= 8.0 Hz). ³¹P NMR (CDCl₃): δ 34.00

(s). Anal. Calcd (found) for C₄₀H₂₀F₄O₁₀Os₃P₂: C, 35.09

J

(34.82); H, 1.47 (1.56).

Photolysis of 1,1-Os₃(CO)₁₀(dppbzF₄) To Give HOs₃(CO)₉[μ-

1,2-PhP(C₆H₄-η¹)C₆F₄PPh₂]. A 50 mg (0.037 mmol) portion of

6 was charged to a NMR tube equipped with a J. Young valve,

followed by 0.7 mL of C₆D₆. The solution was evacuated via two

freeze-pump-thaw degas cycles, and the NMR tube was

irradiated at 366 nm at room temperature for a period of 24 h.

The liberated CO was removed by three periodic freeze-

pump-thaw cycles during the course of the reaction. Once the

NMR analysis confirmed the disappearance of 6, the irradiation

was terminated and the solvent removed. The desired product

was subsequently isolated by column chromatography over

silica gel using 1/1 CH₂Cl₂/hexane, after which the product

was recrystallized from hexane/benzene to afford 7 in 90% yield

(44 mg). IR (CH₂Cl₂): ν(CO) 2085 (s), 2025 (s), 2002 (vs), 1996

for C₄₀H₂₄O₁₀Os₃P₂¹/₂CH₂Cl₂: C, 36.31 (36.19); H, 1.88 (1.75).

3

Photolysis of 1,1-Os₃(CO)₁₀(dppbz) To Give HOs₃(CO)₉[μ-

A small Schlenk tube was

1,2-PhP(C₆H₄-η¹)C₆H₄PPh₂].

charged with 0.10 g (0.077 mmol) of 3 and 25 mL of toluene,

after which the vessel was sealed and subjected to three free-

ze-pump-thaw degas cycles. The vessel was then irradiated

(366 nm) at room temperature for a period of 5 days. The CO

that accompanies the formation of 4 was periodically removed

by additional freeze-pump-thaw degas cycles, in order to drive

the reaction to completion. At the end of the photolysis, the

solution was concentrated under vacuum; the residue was

passed over a short pad of alumina, using 1/1 CH₂Cl₂/hexane

as the eluent, to furnish a yellow solid. Recrystallization of 4

from CH₂Cl₂/benzene afforded the desired product in 88% yield

(86 mg). IR (CH₂Cl₂₁): ν(CO) 2075 (s), 2038 (vs), 2013 (s), 1997

(s), 1967 (m) cm^-1. H NMR (C₆D₆): δ -16.43 (dd, hydride,

1

(m), 1951 (m) cm^-1. H NMR (C₆D₆): δ -16.13 (dd, hydride,

J_P-H= 13.8 and 14.2 Hz), 6.88 (bm, 2H, metalated aryl),

6.97-7.16 (m, 10H), 7.56 (m, 3H), 7.68 (m, 2H), 8.05 (bt, 1H,

metalated aryl, J_H-H= 7.5 Hz), 8.37 (m, 1H, metalated aryl).²⁸

³¹P NMR (CDCl₃): δ 41.17 (b, unresolved fluorine coupling),

44.43 (b, unresolved fluorine coupling). Anal. Calcd (found) for

J

_P-H= 12.5 and 16.0 Hz), 6.68-7.52 (m, 22H, aryl), 8.36 (dd,

1H, metalated aryl, J_H-H= 2.5 and 6.5 Hz). ³¹P NMR (C₆D₆):

δ 28.62 (d, J_P-P= 13.3 Hz), 39.54 (d, J_P-P= 13.3 Hz). Anal.

Calcd (found) for C₄₀H₂₄O₁₀Os₃P₂¹/₂C₆H₆: C, 38.53 (38.15);

H, 2.06 (2.50).

C₃₉H₂₀F₄O₉Os₃P₂¹/₂C₆H₆: C, 36.52 (37.01); H, 1.67 (2.18).

3

Isomerization Kinetics. The conversion 2 f 3 was investigated

by UV-vis spectroscopy in quartz UV-visible cells (1.0 cm

width) that were equipped with a high-vacuum Teflon stopcock,

to facilitate handling on the vacuum line. Samples of 2 in toluene

solution were purged with CO at room temperature immediately

before the start of each kinetic experiment, in order to suppress

the formation of the hydride cluster 4. The experiments con-

ducted at 6.8 and 34 atm of CO were carried out in a carbon-steel

300 mL autoclave, where the reaction pressure was controlled by

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

Reaction of 1,2-Os₃(CO)₁₀(MeCN)₂with dppbzF₄To Give 1,1-

Os₃(CO)₁₀(dppbzF₄). To 0.20 g (0.21 mmol) of 1,2-Os₃(CO)₁₀-

(MeCN)₂in 50 mL of benzene was added 0.12 g (0.23 mmol) of

dppbzF₄under argon flush. The solution was stirred at room

temperature and monitored by TLC, which confirmed the

consumption of the starting cluster and the presence of a yellow

spot ascribed to cluster 6 (R_f= 0.79 in 1/1 CH₂Cl₂/hexane) after

ca. 1 h. The solvent was removed under vacuum and the desired

product purified by chromatography over silica gel employing

1/1 CH₂Cl₂/hexane as the mobile phase. The isolated product

was recrystallized from CH₂Cl₂/hexane at 5 ꢀC to furnish 0.25 g

(85%) of analytically pure 6. IR (CH₂Cl₂): ν(CO) 2095 (s), 2046

(s), 2010 (vs), 1992 (m), 1975 (m), 1962 (m), 1927 (w) cm^-1. ¹H

NMR (C₆D₆): δ 7.20 (t, 4H, para, J_H-H= 8.0 Hz), 7.34 (dt, 8H,

(28) The quoted J_P-Hvalues for the bridging hydride were deter-

mined by spectral simulation using the program gNMR.

Article

Organometallics, Vol. 30, No. 5, 2011 1257

Table 2. X-ray Crystallographic Data and Processing Parameters for the dppbzF₄-Substituted Clusters 6 1/2C₆H₆and 7 CHCl₃

3

7 CHCl₃

6

¹/₂C₆H₆

3

CCDC entry no.

cryst syst

space group

a, A

b, A

c, A

R, deg

β, deg

770698

triclinic

P1

770701

monoclinic

P2₁/n

11.4132(6)

19.109(1)

18.870(1)

12.1322(5)

17.5733(8)

21.598(1)

94.371(1)

101.218(1)

106.671(1)

4284.0(3)

C₄₃H₂₃F₄O₁₀Os₃P₂

1408.15

4

2.183

0.710 73

9.025

95.543(1)

4096.0(4)

C₄₀H₂₀Cl₃F₄O₉Os₃P₂

1459.45

4

2.367

γ, deg

V, A³

mol formula

fw

formula units per cell (Z)

D_calcd, Mg/m³

λ(Mo KR), A

μ, mm^-1

0.710 73

9.631

abs cor

F(000)

semiempirical from equivalents

2620

semiempirical from equivalents

2708

cryst size, mm

abs cor factor

total no. of rflns

no. of indep rflns

no. of data/residuals/params

R1^a(I g 2σ(I))

wR2^b

0.27 ꢀ 0.26 ꢀ 0.06

0.6055/0.1927

53 669

0.23 ꢀ 0.11 ꢀ 0.07

0.5096/0.2371

48 738

18 867

8685

18 867/187/1128

0.0240

0.0287

1.026

1.779, -1.696

8685/0/547

0.0238

0.0300

1.041

3.353, -1.123

GOF on F²

max, min ΔF, e/A³

P

^aR1 = ||F_o| - |F_c||/ |F_o|. ^bwR2 = { [w(F_o²- F_c₂)²/ [w(F_o²)²]}^1/2

.

diode array spectrometer employed in our studies was config-

ured with a custom VT cell holder, which was connected to a

Neslab RTE7 constant-temperature circulator. The quoted

reaction temperatures are considered to be accurate to within

(0.5 K. The progress of the isomerization reactions, conducted

at 1 atm of CO, was monitored by following the optical changes

of the 365 nm band 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 Table 3, were determined by nonlinear

regression analysis, using a single-exponential function, with the

initial (A₀) and final (A_¥) absorbance values and the rate

constant (k) treated as free variables in the calculation.²⁹The

activation parameters for the isomerization 2 f 3 were calcu-

The molecular structures were checked using PLATON,³⁴and

the non-hydrogen atoms were refined anisotropically, except

where otherwise noted. All hydrogen atoms were assigned

calculated positions and allowed to ride on the attached carbon

atom. The cluster 4 CH₂Cl₂was found to be twinned in the first

3

domain by 178.1ꢀ about the reciprocal axis -0.010, 1.000, 0.008

(twin ratio 0.625:0.375). The structure was refined using un-

merged data in HKL5 format and contained a disordered

solvent. These drawbacks affected the quality of the collected

X-ray data and prevented the accurate location of the bridging

hydride group that is assumed to span the Os(2)-Os(3) vector in

4 CH₂Cl₂. the cluster 6 ¹/₂C₆H₆exists as two independent 1,1-

Os₃(CO)₁₀(dppbzF₄) molecules in the asymmetric unit of the

triclinic cell (Z⁰=2). Disorder was found for the phenyl groups

on the atoms P2 and P3 and the benzene solvent molecules

located on inversion centers 1f and 1e. The disordered parts were

refined accordingly with distance restraints to simulate the

hexagonal shape of the aromatic rings. Tables 1 and 2 summar-

ize the pertinent X-ray data and processing parameters for these

triosmium clusters.

Computational Methodology. All calculations were performed

with the Gaussian 03 package of programs.³⁵The calculations

were carried out with the B3LYP functional, which is comprised

of Becke’s linear combination of three functionals³⁶and the

correlation functional of Lee, Yang, and Parr.³⁷The Os atoms

were described by Stuttgart-Dresden effective core potentials

and the SDD basis set.³⁸The D95* basis set was used for the P,

O, C, and H atoms.³⁹

3

lated from a plot of ln(k/T) versus T^{-1 30}

with the error limits

representing the deviation of the data points about the least-

squares line of the Eyring plot.

,

X-ray Diffraction Data for Clusters 2-6. Single crystals of the

cluster compounds 2-4, suitable for X-ray diffraction analyses,

were grown from CH₂Cl₂and hexane, while X-ray-quality

crystals of 6 and 7 were grown from benzene/hexane and

CHCl₃/hexane, respectively. With the exception of cluster 3,

whose X-ray data were collected at TCU on a Bruker SMART

1000 CCD-based diffractometer at 213(2) K, all of the other

structures were collected at UNT on a Bruker X8 APEX CCD

diffractometer at 100(2) K. The frames were integrated with the

available SAINT³¹(TCU) and APEX2³²(UNT) software

packages, using a narrow-frame algorithm; the structures were

solved and refined using the SHELXTL program package.³³

(34) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7.

(35) Frisch, M. J. et al. Gaussian 03, Revision E.01; Gaussian, Inc.,

Pittsburgh, PA, 2003.

(29) The rate calculations were performed by using the equation A(t)

= A_¥þ ΔAe^-ktwith the aid of the commercially available program

Origin6.0.

(36) Becke, A. D. Phys. Rev. A 1988, 38, 3098.

(30) Carpenter, B. K. Determination of Organic Reaction Mechan-

isms; Wiley-Interscience: New York, 1984.

(31) Saint Version 6.02; Bruker Advanced Analytical X-ray Systems,

(37) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (b)

Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157,

200.

Inc., 1997-1999.

(38) Dolg, M. In Modern Methods and Algorithms of Quantum

Chemistry; Grotendorst, J., Ed.; John von Neumann Institute for

(32) APEX2 Version 2.14; Bruker Advanced Analytical X-ray Sys-

tems, Inc., Madison, WI, 2007.

(33) SHELXTL Version 5.1; Bruker Advanced Analytical X-ray

€

Computing: Julich, 2000; Vol. 1, p 479.

(39) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry;

Schaefer, H. F., III, Ed.; Plenum: New York, 1976; Vol. 3, pp 1-28.

Systems, Inc., Madison, WI, 1998.

1258 Organometallics, Vol. 30, No. 5, 2011

Zhang et al.

Scheme 1

Geometries were fully optimized, and analytical second deri-

vatives were evaluated at each stationary point, to establish

whether the geometry was an energy minimum (no imaginary

frequencies) or a transition structure (one imaginary frequency).

The computed frequencies were used to make zero-point and

thermal corrections to the electronic energies.

Results and Discussion

I. Diphosphine Bridging and Chelation in the Reaction of

1,2-Os₃(CO)₁₀(MeCN)₂with dppbz. The labile cluster 1,2-

Os₃(CO)₁₀(MeCN)₂reacts rapidly with the diphosphine

ligand dppbz at room temperature, to furnish the corre-

sponding diphosphine-bridged cluster 1,2-Os₃(CO)₁₀-

(dppbz) (2) as the major product. The yield of crude pro-

duct was 78%, as assessed by ³¹P NMR analysis of an aliquot

of the reaction before purification.⁴⁰Accompanying 2 in the

reaction mixture were the isomeric chelated cluster 1,1-Os₃-

(CO)₁₀(dppbz) (3) and the hydride cluster HOs₃(CO)₉[μ-1,2-

PhP(C₆H₄-η¹)C₆H₄PPh₂] (4), which were formed in 20% and

2% yields, respectively. These two minor products both

result from the isomerization of 2 to 3. Decarbonylation of

the latter cluster leads to ortho metalation and the formation

of 4 (vide infra). The relationship between these dppbz-

substituted clusters is summarized in Scheme 1.

TLC analysis of the crude reaction mixture revealed the

presence of only one spot for the three products; therefore,

the ternary mixture was isolated by flash-column chroma-

tography and characterized in solution as an aggregate. The

spectroscopic properties of 2 are consistent with its formu-

lated structure. The bridging dppbz appears as a singlet in the

³¹P NMR spectrum at δ 6.41. The different aromatic hydro-

gens, associated with the four phenyl groups (Ph) and the

benzo group, could not be assigned with certainty due to

extensive overlap.

Figure 1. Thermal ellipsoid plot of 2 CH₂Cl₂at the 50% prob-

3

ability level with the hydrogen atoms shown as small spheres of

arbitrary radii. The CH₂Cl₂solvent has been omitted for clarity.

Selected bond distances (A) and angles (deg): Os(1)-P(1) =

2.3359(6), Os(1)-Os(2) = 2.8507(1), Os(1)-Os(2) = 2.8547(1),

Os(2)-P(2) = 2.3289(6), Os(2)-Os(3) = 2.8821(1), P(1) P-

3 3 3

(2) = 3.6073(9); P(1)-Os(1)-Os(3) = 155.39(2), P(2)-Os(2)-

Os(3) = 156.27(2), C(11)-P(1)-Os(1) = 114.73(8), C(12)-P-

(2)-Os(2) = 120.62(8), C(12)-C(11)-P(1) = 125.3(2), C(11)-

C(12)-P(2) = 126.6(2).

phosphine-substituted polynuclear compounds.^21a,41While

the P(1) and P(2) atoms are equatorially disposed, they are

tipped out of the plane defined by the three osmium atoms by

0.2765(6) and -0.5657(6) A, respectively. The o-phenylene

platform that serves to tether the vicinal Ph₂P moieties is

defined by the atoms C(11)-C(16), and this group is also

tipped out of the metallic plane with a fold angle of 48.89(3)ꢀ.

The ancillary carbonyl ligands are best viewed as linear

groups, and the axial CO groups exhibit the commonly

observed D₃twist with respect to the metallic core.⁴²

The bridging of the Os(1) and Os(2) atoms by the dppbz

ligand in 2 is thermodynamically less favorable than the

chelation in isomer 3. This factor also serves as the driving

force for the facile isomerizations, observed in related,

diphosphine-bridged isomers of Os₃(CO)₁₀(P-P).²¹Three

significantly adverse perturbations result from the stretching

of the dppbz ligand across the Os(1)-Os(2) vector in 2, and

they in turn account for the ground-state destabilization in 2

relative to 3. The bridging dppbz ligand in 2 exhibits a

The thermal ellipsoid plot of the molecular structure of 2,

as the CH₂Cl₂solvate, shown in Figure 1 confirms the

bridging disposition of the dppbz ligand, which spans the

Os(1)-Os(2) vector. The Os-Os bond distances range from

2.8507(1) A (Os(1)-Os(2)) to 2.8821(1) A (Os(2)-Os(3)),

and the Os(1)-P(1) and Os(2)-P(2) bond distances of

2.3359(6) and 2.3289(6) A are slightly longer than those

distances reported by us for 1,2-Os₃(CO)₁₀(dppen) but are

well within the accepted range observed in a variety of

(40) Phosphine ligand substitution in 1,2-Os₃(CO)₁₀(MeCN)₂has

€

been previously shown by Poe to proceed through a dissociative loss

of the MeCN ligands. In the present case, capture of the initial

unsaturated cluster, Os₃(CO)₁₀(MeCN), by the dppbz is predicted to

afford Os₃(CO)₁₀(MeCN)(η¹-dppbz) with a pendant phosphine ligand.

Rapid ring closure to generate 2 is the anticipated result from either

MeCN dissociation or an associative interchange process where MeCN

(41) Orpen, A. G.; Brammer, L.; Allen, F. K.; Kennard, O.; Watson,

D. G.; Taylor, R. Dalton Trans. 1989, S1.

(42) (a) Alex, R. F.; Einstein, F. W. B.; Jones, R. H.; Pomeroy, R. K.

Inorg. Chem. 1987, 26, 3175. (b) Bruce, M. I.; Liddell, M. J.; Skawkataly,

O. B.; Hughes, C. A.; Skelton, B. H.; White, A. H. J. Organomet. Chem.

1988, 347, 207.

€

serves as the leaving group. See: (a) Dahlinger, K.; Poe, A. J.; Sayal,

P. K.; Sekhar, V. C. Dalton Trans. 1986, 2145. (b) Hudson, R. H. E.; Poe,

€

A. J.; Sampson, C. N.; Siegel, A. Dalton Trans. 1989, 2235.

Article

Organometallics, Vol. 30, No. 5, 2011 1259

Table 3. Experimental Rate Constants for the Isomerization of

1,2-Os₃(CO)₁₀(dppbz) to 1,1-Os₃(CO)₁₀(dppbz) under CO^a

entry

temp (K)

10⁴k (s^-1

)

1

2

3

4

5

6

7

8

9

318.0

323.0

328.0

333.0

338.0

343.0

0.29(8)

0.50(7)

0.45(10)^b

0.58(9)^c

0.49(8)^d

1.01(12)

1.41(10)

2.63(16)

3.66(14)

^aThe UV-vis kinetic data were collected in toluene using a ca. 10^-4

M solution of starting cluster by following the increase in the absorbance

of the 365 nm band. Unless otherwise noted, all reactions were per-

formed under 1 atm of CO. ^bReaction carried out in an autoclave under

6.8 atm of CO. ^cReaction carried out in an autoclave under 34 atm of

CO. ^dReaction carried out under argon in the presence of 10 equiv of

PPh₃.

2.9092(5) A (Os(1)-Os(2)) to 2.9263(5) A (Os(1)-Os(3)).

Figure 2. Thermal ellipsoid plot of 3 at the 30% probability

level with the hydrogen atoms shown as small spheres of

arbitrary radii. Selected bond distances (A) and angles (deg):

Os(1)-P(2) = 2.290(1), Os(1)-P(1) = 2.294(1), Os(1)-Os(2) =

2.9092(5), Os(1)-Os(3) = 2.9263(5), Os(2)-Os(3) = 2.9109(4),

The internuclear P(1) P(2) distance of 3.094(2) A has

3 3 3

undergone a significant contraction, relative to that distance

found in 2, and the angles associated with the Os(1)-P-

(1)-C(1) (108.6(2)ꢀ), Os(1)-P(2)-C(2) (108.6(2)ꢀ), C(2)-

C(1)-P(1) (116.9(3)ꢀ), and C(1)-C(2)-P(2) (117.9(3)ꢀ) lin-

kages display values that are in excellent agreement with the

idealized hybridization state of the subtended atom. The

observed angle of 84.90(4)ꢀ for the P(1)-Os(1)-P(2) atoms

is in keeping with the chelating nature of the ancillary dppbz

ligand and the bite angle reported in the three known dppbz-

chelated clusters Ir₄(CO)₁₀(dppbz),⁴⁴H₄Ru₄(CO)₁₀(dppbz),⁴⁵

and HRu₆(μ₅-C)(μ₃-P)(CO)₁₄(dppbz).⁴⁶The remaining

bond distances and angles are unremarkable and require

no comment.

P(1) P(2) = 3.094(2); P(2)-Os(1)-P(1) = 84.90(4), P(2)-Os-

3 3 3

(1)-Os(2) = 168.58(3), P(1)-Os(1)-Os(2) = 164.46(3), C-

(1)-P(1)-Os(1) = 108.6(2), C(2)-P(2)-Os(1) = 108.6(2),

C(2)-C(1)-P(1) = 116.9(3), C(1)-C(2)-P(2) = 117.9(3).

P(1) P(2) internuclear bond distance (3.6073(9) A) that is

3 3 3

>0.4 A longer than the P(1) P(2) internuclear bond

3 3 3

distance reported for the free dppbz ligand.⁴³The angular

deviation from sp²hybridization of the o-phenylene C(11)

and C(12) carbons found in the C(12)-C(11)-P(1)

(125.3(2)ꢀ) and C(11)-C(12)-P(2) (126.6(2)ꢀ) linkages ex-

ceeds the accepted VSEPR angle of 120ꢀ by >5ꢀ. A similar

deformation from the accepted sp³hybridization bond angle

of ca. 109ꢀ is also evidenced in the angles of 114.73(8) and

120.62(8)ꢀ for the C(11)-P(1)-Os(1) and C(12)-P(2)-Os-

(2) linkages, respectively. The observed fold angle of

Our interest in the coordinative flexibility of diphosphine

ligands in polynuclear clusters^21,22led us to measure the

kinetics for the conversion 2 f 3. The isomerization reac-

tions were conveniently monitored by UV-vis spectroscopy,

under CO, over the temperature range 318-343 K; the first-

order rate constants given in Table 3 were obtained. The

progress of each reaction was monitored by following the

absorbance increase in the 365 nm band belonging to 3.

Figure 3 shows the spectral changes in the transformation 2

f 3. The two clearly defined isosbestic points at 397 and 420

nm indicate that the isomerization is well-behaved and is free

of kinetic complications. The absorbance versus time plot,

depicted in the inset of Figure 3, shows the high quality of the

least-squares fit of the 365 nm absorbance data to the growth

of 3 by a first-order reaction. The effect of CO pressure

(entries 2-4 in Table 3) and PPh₃(entry 5) on the reaction

was also explored, and no significant changes in the rate

constant were observed, relative to the standard value re-

ported for entry 2. The insensitivity of the rate constants to

ligand additives and the Eyring activation parameters (ΔH^q

= 21.6(3) kcal/mol; ΔS^q= -11(1) eu) both strongly support

an intramolecular isomerization scenario that involves the

migration of phosphine and CO groups about the cluster

polyhedron. The details of these migrations were

43.95(2)ꢀ along the P(1) P(2) hinge of the six-membered

3 3 3

ring, defined by the Os₂P₂C₂atoms, underscores the non-

planar disposition that exists between the o-phenylene and

Os₃rings in 2.

Heating the ternary mixture of 2-4 under 1 atm of CO at

333 K overnight afforded 3 as the sole observable product, as

1

determined by H and ³¹P NMR spectroscopy. This trans-

formation verifies that 3 is the thermodynamically most

stable of the isomeric Os₃(CO)₁₀(dppbz) clusters. The addi-

tion of CO suppresses the formation of the hydride cluster 4,

but CO addition has been shown not to affect the equilibrium

for the bridged-to-chelated ligand isomerization in other

decacarbonyl clusters Os₃(CO)₁₀(P-P).²¹

Cluster 3 was isolated by column chromatography and

characterized in solution by IR and NMR spectroscopy; the

molecular structure was determined by X-ray diffraction

analysis. The singlet recorded at δ 26.32 in the ³¹P NMR

spectrum of 3 is in full agreement with a triosmium cluster

possessing a chelating dppbz ligand. Figure 2 shows the

thermal ellipsoid plot of the molecular structure of 3 and

confirms the attendant isomerization of the dppbz ligand

upon heating. The Os-Os bond distances range from

(44) Park, B. K.; Miah, M. A.; Kang, H.; Lee, K.; Cho, Y.-J.;

Churchill, D. G.; Park, S.; Choi, M.-G.; Park, J. T. Organometallics

2005, 24, 675.

(45) Nesterov, V. N.; Watson, W. H.; Kandala, S.; Richmond, M. G.

Polyhedron 2007, 26, 3602.

(43) Levason, W.; Reid, G.; Webster, M. Acta Crystallogr. 2006, C62,

440.

(46) Watson, W. H.; Kandala, S.; Richmond, M. G. J. Organomet.

Chem. 2007, 692, 1648.

1260 Organometallics, Vol. 30, No. 5, 2011

Zhang et al.

Figure 3. Representative UV-vis spectral changes for the iso-

merization 2 f 3 in toluene at 323 K under 1 atm of CO. The

inset shows the absorbance versus time plot for the experimental

data (0) and the nonlinear regression fit of the first-order rate

constant k (solid black line).

Figure 4. Thermal ellipsoid plot of 4 CH₂Cl₂at the 50% prob-

3

ability level with the hydrogen atoms shown as small spheres of

arbitrary radii. The CH₂Cl₂solvent has been omitted for clarity.

Selected bond distances (A) and angles (deg): Os(1)-Os(3) =

2.8808(4), Os(1)-Os(2) = 2.9330(3), Os(2)-P(2) = 2.323(1),

Os(2)-P(1) =2.367(2),Os(2)-Os(3) =2.9948(4),Os(3)-C(11) =

investigated by the DFT calculations, the results of which are

described in a subsequent section.

II. Reversible Ortho Metalation in 3 and Formation of

HOs₃(CO)₉[μ-1,2-PhP(C₆H₄-η¹)C₆H₄PPh₂] (4). As part of

our ongoing studies of bond-activation processes, involving

cluster-coordinated bidentate ligands,^{20-22,42,47-49}the ori-

gin and structural composition of the hydride cluster 4 that

was isolated from the reaction between 1 and dppbz were

next examined. While cluster 3 is thermally stable to CO loss

at temperatures below 353 K, it is very sensitive to near-UV

irradiation. Photolysis of a toluene solution containing 3,

using 366 nm light, led to CO loss and formation of 4. The

hydride-bridged cluster 4 could be obtained in quantitative

yield, provided that the liberated CO was periodically re-

moved through freeze-pump-thaw degas cycles or by

active argon purge during irradiation.

In terms of the mechanism of formation of 4 from photo-

lysis of 3, optical excitation within the σ f σ* manifold is

expected to promote the formation of the putative, unsatu-

rated cluster Os₃(CO)₉(dppbz) through loss of one CO. This

latter cluster is extremely sensitive to the presence of the

released CO, which must be removed to avoid the regenera-

tion of cluster 3. Treatment of an isolated sample of 4 with

CO (1 atm) overnight at 333 K in toluene led to the

quantitative formation of 3, thus establishing the reversible

nature of the ortho metalation of the ancillary dppbz ligand.

This reactivity paradigm appears to be general, in that

several other activated diphosphine ligands have been shown

by us to regenerate their respective Os₃(CO)₁₀(P-P) cluster

upon carbonylation.^21,49

2.184(6), P(1) P(2) = 3.076(2); P(2)-Os(2)-P(1) = 81.94(5),

3 3 3

P(2)-Os(2)-Os(1) = 172.92(4), P(1)-Os(2)-Os(1) = 97.55(4),

C(16)-P(1)-Os(2) = 106.7(2), C(17)-P(2)-Os(2) = 109.4(2),

C(17)-C(16)-P(1) = 118.1(4), C(16)-C(17)-P(2) = 116.2(4).

solvate, and confirms the ortho metalation of one of the aryl

rings of the dppbz ligand. 4 contains 48e and is electronically

saturated, assuming that the ancillary PhP(C₆H₄-

η¹)C₆H₄PPh₂ligand functions as a 5e-donor group. The

Os-Os bond lengths range from 2.8808(4) A (Os(1)-

Os(3)) to 2.9948(4) A (Os(2)-Os(3)) and reveal a mean

distance of 2.9362 A that parallels those Os-Os bond

distances reported for other Os₃clusters. Both phosphine

atoms are coordinated to the Os(2) center and display axial

(P(1)) and equatorial (P(2)) dispositions. The bridging hy-

dride was not located during structural refinement, but it

may be confidently assigned to the Os(2)-Os(3) bond, on the

basis of the elongated distance of this bond, relative to the

other two Os-Os bonds, and the stereochemical disposition

of the C(9)O(9) and P(2) groups about the cluster

polyhedron.⁵⁰Here the Os(2)-Os(3)-C(9) and Os(3)-Os-

(2)-P(2) angles of 117.6(2) and 114.87(3)ꢀ, respectively, are

significantly expanded from the idealized value of ca. 90ꢀ

typically found for equatorial substituents in triangular

clusters, due to the presence of the ancillary hydride.⁵¹The

Os(3)-C(11) bond distance of 2.184(6) A is comparable to

the Os-C bond distance found in the triosmium clusters

H₂Os₃(CO)₉(μ-C₆H₄PPh), HOs₃(CO)₉(PPh₃)(μ-SbPh₂)(μ-

C₆H₄), and Os₃(CO)₉[μ-PPh(C₆H₄)CH₂PPh], each of which

possess an activated aryl moiety.⁵²

4 was isolated in 88% yield after purification by column

chromatography over alumina, followed by recrystalliza-

tion. The presence of a bridging hydride at δ -16.43 in the

¹H NMR spectrum and the pair of inequivalent doublets at δ

28.62 and 39.54 in the ³¹P NMR spectrum are in keeping with

the formulated structure of 4. Figure 4 shows the thermal

ellipsoid plot of the molecular structure of 4, as the CH₂Cl₂

(50) (a) Mingos, D. M. P.; Wales, D. J. Introduction to Cluster

Chemistry; Prentice Hall: Englewood Cliffs, NJ, 1990. (b) Teller, R. G.;

Bau, R. Struct. Bonding (Berlin) 1981, 41, 1.

(51) (a) Churchill, M. R.; Hollander, F. J.; Lashewycz, R. A.; Pear-

son, G. A.; Shapley, J. R. J. Am. Chem. Soc. 1981, 103, 2430. (b)

Churchill, M. R.; Hollander, F. J. Inorg. Chem. 1981, 20, 4124.

(52) (a) Colbran, S. B.; Irele, P. T.; Johnson, B. F. G.; Lahoz, F. J.;

Lewis, J.; Raithby, P. R. Dalton Trans. 1989, 2023. (b) Chen, G.; Deng,

M.; Lee, C. K.; Leong, W. K. Organometallics 2002, 21, 1227. (c) Kabir,

S. E.; Miah, M. A.; Sarker, N. C.; Hussain, G. M. G.; Hardcastle, K. I.;

Nordlander, E.; Rosenberg, E. Organometallics 2005, 24, 3315.

(47) Poola, B.; Carrano, C. J.; Richmond, M. G. Organometallics

2008, 27, 3018.

(48) Kandala, S.; Hammons, C.; Watson, W. H.; Wang, X.; Rich-

mond, M. G. Dalton Trans. 2010, 39, 1620.

(49) Huang, S.-H.; Keith, J. M.; Hall, M. B.; Richmond, M. G.

Organometallics 2010, 29, 4041.

Article

Organometallics, Vol. 30, No. 5, 2011 1261

Scheme 2

Decarbonylation of other 1,1-Os₃(CO)₁₀(P-P) clusters,

III. Diphosphine Chelation in the Reaction of 1,2-Os₃-

(CO)₁₀(MeCN)₂with dppbzF₄. Treatment of 1 with a mea-

sured excess of dppbzF₄at room temperature furnished 1,1-

Os₃(CO)₁₀(dppbzF₄) (6) as the sole observable product, as

assessed by ³¹P NMR spectroscopy. Repeating the reaction

at 0 ꢀC in a sealed NMR tube afforded the same product,

albeit more slowly. The inability to observe 1,2-Os₃(CO)₁₀-

(dppbzF₄) (5) spectroscopically indicates that the expected

kinetic substitution product is unstable. Presumably, this

instability is due to unfavorable steric interactions between

the Ph₂P moieties and the ortho fluorines, interactions that

are exacerbated by the coordination of the diphosphine

ligand across the Os-Os bond.⁵⁶The rapid isomerization

of 5 to the thermodynamically more stable, diphosphine-

chelated cluster 6 alleviates the strain in the “stretched”

dppbzF₄ligand. Scheme 3 summarizes the results of the

reaction between 1 and dppbzF₄.

followed by ortho metalation of the diphosphine ligand, is

observed upon heating; however, the composition of the

resulting product is highly dependent upon the nature of the

initial diphosphine, and further activation of the metalated

aryl group is often found under harsh reaction conditions (i.

e., refluxing toluene or heptane). A check of the Cambridge

Structural Database (CSD version 5.31) for Os₃clusters,

possessing a chelating diphosphine that has experienced

ortho metalation, reveals only four structures on file. Scheme

2 shows these four clusters, whose structural diversity is

immediately obvious. Thermolysis of 1,1-Os₃(CO)₁₀(dppp)

in toluene under H₂leads to the elimination of benzene and

formation of the phosphido-bridged cluster H₂Os₃(CO)₈[μ-

Ph₂P(CH₂)₃P(C₆H₄-η¹)].⁵³Heating the corresponding (R)-

BINAP and dppn decacarbonyl derivatives gives rise to the

ortho-metalated products HOs₃(CO)₈[μ-(R)-BINAP(η¹,π)-

H] and HOs₃(CO)₈[μ-Ph₂P(nap)PPh(C₆H₄-η¹,π)], where the

activated aryl ring in each product functions as a 7e face-

capping donor ligand.^54,55Only HOs₃(CO)₉[μ-PhP-

(C₆H₄)CdC(PPh₂)C(O)C=CH(C₅H₄FeCp)C(O)], which

was obtained as a ca. 1:1 mixture of diastereomers

from the photolysis of 1,1-Os₃(CO)₁₀(fbpcd) (where fbpcd

= 2-(ferrocenylidene)-4,5-bis(diphenylphosphino)-4-cyclo-

penten-1,3-dione), exhibits a structure that is similar to that

of 4.^21d

6 was isolated in 85% yield after column chromatography

over silica gel, followed by recrystallization, and it was

characterized by IR and NMR spectroscopy, combustion

analysis, and X-ray crystallography. The ³¹P NMR spectrum

revealed a diagnostic singlet at δ 34.00, while the ¹H NMR

spectrum displayed three aromatic resonances at δ 7.20, 7.34,

and 7.76 in a 4:8:8 integral ratio for the para, meta, and ortho

hydrogens, belonging to the two PPh₂moieties. The struc-

ture of 6 ¹/₂(benzene) is shown in Figure 5, where the

3

(53) Deeming, A. J.; Hardcastle, K. I.; Kabir, S. E. Dalton Trans.

1988, 827.

(54) Deeming, A. J.; Stchedroff, M. Dalton Trans. 1998, 3819.

(55) Bruce, M. I.; Humphrey, P. A.; Schmutzler, R.; Skelton, B. W.;

White, A. J. Organomet. Chem. 2004, 689, 2415.

(56) For manifestations of the buttressing effect in chemical syntheses

and the thermochemical properties of alkyl-substituted phenols, see: (a)

Sammes, P. G.; Weller, D. J. Synthesis 1995, 1205. (b) Choony, N.;

Dadabhoy, A.; Sammes, P. G. J. Chem. Soc., Perkin Trans. 1 1998, 2017.

(c) Verevkin, S. P. J. Chem. Thermodyn. 1999, 31, 1397.

1262 Organometallics, Vol. 30, No. 5, 2011

Zhang et al.

Scheme 3

Figure 6. Thermal ellipsoid plot of 7 CHCl₃at the 50% prob-

3

ability level, with the carbon and hydrogen atoms shown as

small spheres of arbitrary radii. The CHCl₃solvent has been

omitted for clarity. Selected bond distances (A) and angles (deg):

Os(1)-C(11) = 2.186(4), Os(1)-Os(2) = 2.8714(3), Os(1)-Os-

(3) = 2.9812(3), Os(2)-Os(3) = 2.9399(3), Os(3)-P(1) =

2.307(1), Os(3)-P(2) = 2.366(1), P(1) P(2) = 3.139(2); P-

Figure 5. Thermal ellipsoid plot of one of the two independent

molecules of 6 ¹/₂C₆H₆at the 50% probability level. The

3 3 3

(1)-Os(3)-P(2) = 84.41(4), P(1)-Os(3)-Os(2) = 165.45(3),

P(2)-Os(3)-Os(2) = 98.38(3), C(34)-P(1)-Os(3) = 108.6(1),

C(35)-P(2)-Os(3) = 106.4(1), C(35)-C(34)-P(1) = 117.1(3),

C(34)-C(35)-P(2) = 118.8(3).

3

hydrogen atoms are shown as small spheres of arbitrary radii,

and the C₆H₆solvent has been omitted for clarity. Selected bond

distances (A) and angles (deg): Os(1)-P(2) = 2.276(2), Os-

(1)-P(1) = 2.290(1), Os(1)-Os(2) = 2.8871(3), Os(1)-Os(3) =

photochemical properties of 6 were explored, in order to

check whether there are any reactivity differences between it

and 3. 6 is stable to CO loss at 333 K but reacts in a manner

identical with that of 3 under photolysis conditions (366 nm),

to afford the corresponding hydride-bridged cluster 7. Treat-

ment of 7 with excess CO at 333 K regenerates 6 in quanti-

tative yield. The ortho metalation of the ancillary

diphosphine ligand occurs without complications, and the

spectroscopic properties for 7 are consistent with the pro-

posed structure

2.9093(3), Os(2)-Os(3) = 2.8948(4), P(1) P(2) = 3.106(2);

3 3 3

P(2)-Os(1)-P(1) = 85.74(5), P(2)-Os(1)-Os(2) = 165.99(4),

P(1)-Os(1)-Os(3) = 167.75(3), C(11)-P(1)-Os(1) = 107(1),

C(12)-P(2)-Os(1) = 107.2(2), C(12)-C(11)-P(1) = 116.9(4),

C(11)-C(12)-P(2) = 117.6(4).

presence of a chelating dppbzF₄ligand is verified. The two

independent molecules of 6 found in the unit cell do not

exhibit any significant differences in geometries, and to our

knowledge, the present example represents the first structu-

rally characterized organometallic compound that contains

the dppbzF₄ligand.⁵⁷The mean Os-Os bond distance in 6 is

2.8971 A and is slightly shorter than the mean distance found

for the Os-Os bonds in 3. The ancillary aryl groups asso-

ciated with the P(2) and P(3) atoms in the molecules of 6 are

disordered over two positions. The observed bite angles for

the chelating dppzF₄ligands in the two independent mole-

cules found in the crystal structure are 85.74(5) and 86.81(6)ꢀ

for the P(1)-Os(1)-P(2) and P(3)-Os(4)-P(4) linkages and

are similar to the 84.90(4)ꢀ bite angle observed in 3 for the

dppbz ligand.

Figure 6 shows the thermal ellipsoid plot of 7 CHCl₃,

3

whose overall molecular architecture is compositionally

similar to that of 4. The Os-Os bond distances vary from

2.8714(3) A (Os(1)-Os(2)) to 2.9812(3) A (Os(1)-Os(3)),

with the latter bond presumably serving as the locus for the

bridging hydride that was not found during data reduction

and refinement. The Os(1)-C(11) bond distance of 2.186(4)

IV. Synthesis and X-ray Diffraction Structure of HOs₃-

(CO)₉[μ-1,2-PhP(C₆H₄-η¹)C₆F₄PPh₂]. The thermal and

(57) For a report on an iron complex containing the related ligand

1,2-bis(dimethylphosphino)tetrafluorobenzene, see: Higgins, S. J.; Jew-

iss, H. C.; Levason, W.; Webster, M. Acta Crystallogr. 1985, C41, 695.

Article

Organometallics, Vol. 30, No. 5, 2011 1263

Scheme 4

Table 4. Eyring Activation Data for the Isomerization of 1,2-

Os₃(CO)₁₀(P-P) to 1,1-Os₃(CO)₁₀(P-P)

one of the phosphorus atoms across one of the Os-Os

vectors. The second mechanism consists of a merry-go-

round migration of two CO groups and one of the phos-

phorus atoms about the equatorial plane of the cluster.⁵⁸

Both mechanistic possibilities draw on established para-

digms for CO fluxionality in polynuclear clusters,⁵⁹and each

isomerization process requires the formation of a nonclassi-

cal, bridging phosphorus atom along the reaction coordi-

nate, during the transit of phosphorus across an Os-Os

bond.⁶⁰

Entry into the pairwise exchange process mandates the

presence of an axially disposed phosphine moiety, prior to

the simultaneous migration of this phosphine and a trans

axial CO group across a connecting Os-Os bond. The

isomerization of a phosphine moiety from its initial equator-

ial site (e) in the bridged isomer (B) to an axial position (a)

could be achieved by exchange between Os₁and Os₃of two

axial CO groups, situated on opposite sides of the three-

membered Os ring.^61,62The exchange mechanism requires

conrotation of the other four substituents that lie

ligand

ΔH^q(kcal/mol)

ΔS^q(eu)

ref

dppen^a

bpcd

bmi

fbpcd

dppbz

bmf^a

26.5(6)

25.0(7)

24.0(1)

23.1(3)

21.6(3)

26.9(7)

-5(2)

-2(2)

-5(1)

-7(1)

-11(1)

6(1)

21a

21b

21c

21d

this work

unpublished

^aThe isomerization is reversible, and only the activation parameters

for the forward portion of the reaction are reported.

A compares favorably to the corresponding Os-C bond

distance found in 4. The remaining bond distances and

angles are unremarkable with respect to 4 and related

osmium clusters characterized by our groups.

V. Possible Mechanisms for Diphosphine Isomerization in

Os₃(CO)₁₀(P-P). The kinetics and activation parameters for

the conversion 2 f 3 are in keeping with those found in

related Os₃(CO)₁₀(P-P) clusters. Scheme 4 shows the differ-

ent bidentate diphosphine ligands that have been examined

and that undergo bridging to chelating isomerization in

Os₃(CO)₁₀(P-P) clusters. Table 4 summarizes the Eyring data

for the different diphosphine-substituted clusters. The data

reveal a high degree of uniformity and strongly support a

common mechanism for the isomerization of these dipho-

sphine ligands about the Os₃polyhedron, namely a nondis-

sociative, first-order reaction, whose rate depends only on

the Os₃(CO)₁₀(P-P) cluster.

As already noted, the dppbzF₄ligand is atypical, inas-

much as the expected ligand-bridged cluster 5 was not

observed when 1 was allowed to react with dppbzF₄. Instead,

only the chelated isomer 6 was observed. This finding

suggests that the enthalpy of activation for the rearrange-

ment of 5 to 6 is lower than the enthalpies of activation, given

in Table 4, for the rearrangements of all of the other Os₃-

(CO)₁₀(P-P) clusters in the table. The data in Table 4 allow

one to estimate that ΔH^q< 20 kcal/mol for the rearrange-

ment of 5 to 6.

(59) For reports that deal with the elucidation of reaction mechan-

isms in polynuclear clusters using DFT calculations, see: (a) Persson, R.;

Monari, M.; Gobetto, R.; Russo, A.; Aime, S.; Calhorda, M. J.; Nord-

lander, 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. Organometallics 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) Li, Q.-S.; Xu, B.; Xie, Y.; King, R. B.; Schaefer, H. F. Dalton

ꢀ

~

Trans. 2007, 4312. (f) Cabeza, J. A.; Perez-Carreno, E. Organometallics

ꢀ

2008, 27, 4697. (g) Cabeza, J. A.; del Rıo, I.; Fernandez-Colinas, J. M.;

´

ꢀ

Perez-Carreno, E.; Sanchez-Vega, M. G.; Vazquez-Garcı

metallics 2009, 28, 1832. (h) Cabeza, J. A.; Fernandez-Colinas, J. F.;

~

ꢀ

´

a, D. Organo-

ꢀ

~

Perez-Carreno, E. Organometallics 2009, 28, 4217. (i) Cabeza, J. A.; Van

der Maelen, J. F.; Garcıa-Granda, S. Organometallics 2009, 28, 3666. (j)

Cabeza, J. A.; del Rıo, I.; Goite, M. C.; Perez-Carreno, E.; Pruneda, V.

´

ꢀ

~

´

Chem. Eur. J. 2009, 15, 7339.

(60) A kinetically indistinguishable mechanism, involving the unsa-

turated cluster Os₃(CO)₁₀(η¹-P-P), formed through the dissociation of

one arm of the diphosphine ligand from 1,2-Os₃(CO)₁₀(P-P), cannot be

completely eliminated from consideration. However, given the rigid

nature and syn disposition of the diphosphine ligands depicted in

Scheme 4 and the uniformly negative entropies of activation in Table 4,

this alternative does not seem likely.

(61) The isomerization of a phosphine moiety from an equatorial site

to an axial site in the bridged isomer could, in principle, also be achieved

by a 90ꢀ trigonal twist of the four substiuents at one osmium center.

However. in Os₃(CO)₁₂, the activation barrier for the 90ꢀ trigonal-twist

pathway is computed to be 5.4 kcal/mol higher in energy than the type of

isomerization process depicted in eq 3 involving exchange of trans

diaxial CO molecules between adjacent Os atoms. The 5.4 kcal/mol

lower energy of activation that is computed for the isomerization

mechanism in eq 3 is consistent with the VT NMR data reported for

CO exchange in Os₃(CO)_12-xP_x(where x = 1-4) and related clusters. In

these clusters carbonyl scrambling preferentially occurs by exchanges of

pairs of CO molecules between adjacent Os atoms, rather than by

localized, 90ꢀ, tripodal, ligand rotations at a single Os metal center.⁶²

The migration of P in the bridging-to-chelating transfor-

mation of the ancillary diphosphine ligand in Os₃(CO)₁₀(P-

P) must be accompanied, at a minimum, by the permutation

of at least one CO group between adjacent osmium centers.

Two scenarios can account for the phenomenologically

observed diphosphine isomerization in these molecules.

The first scenario involves a pairwise exchange of a CO with

(58) In theory, these two isomerization mechanisms may be differ-

entiated on the basis of the temperature-dependent behavior of the

ancillary CO ligands in 1,2-Os₃(CO)₁₀(P-P). Unfortunately, global CO

scrambling has been found to be competitive with the phosphine

migration process, negating a mechanistic distinction by VT ¹³C

NMR spectroscopy.

1264 Organometallics, Vol. 30, No. 5, 2011

Zhang et al.

Figure 7. Optimized B3LYP structures for the intermediates and transition structures involved in the two-center and merry-go-round

mechanisms for bridged to chelated isomerization reactions of Os₃(CO)₁₀(dppen). ΔE and ΔH are given in kcal/mol, relative to B-ee.

approximately in the plane containing the two migrating CO

groups and Os₁and Os₃. Equation 3 illustrates the ligand

movements starting from bridged 1,2-Os₃(CO)₁₀[(Z)-

Ph₂PCHdCHPPh₂] (B-ee). (The labeling scheme for the

three osmium atoms is that employed in the ensuing discus-

sion of the results of our DFT calculations, and the axial CO

groups in eq 3 have been color-coded for clarity).

resulting permutations of the groups attached to these two

osmium atoms are shown schematically in eq 4.

Following the trans CO and P migrations in eq 4, the C-ea

product is a chelated isomer; however, it has one phosphorus

axial and the other equatorial, whereas, 1,1-Os₃(CO)₁₀[(Z)-

Ph₂PCHdCHPPh₂] and all the other chelated diphosphines

that have been isolated have both phosphorus atoms equa-

torial (C-ee). However, as shown in eq 5, a mechanism,

similar to those in eqs 3 and 4 and involving migrations

between Os₁and Os₂of a pair of trans diaxial CO groups,

would put both of the phosphine groups that are attached to

Os₂in equatorial positions.

After the B-ea isomer has been formed, the axial phos-

phorus on Os₁can migrate to Os₂, concomitant with the

migration of the axial CO, trans to the migrating phos-

phorus, from Os₂to Os₁. This process is fundamentally the

same as that in eq 3, wherein the double migration again

requires conrotation of axial and equatorial CO groups, but

in this case at Os₁and Os₂. The migrations, rotations, and

(62) (a) Deeming, A. J.; Donovan-Mtunzi, S.; Kabir, S. E. J. Orga-

nomet. Chem. 1985, 281, C43. (b) Alex, R. F.; Pomeroy, R. K. Organo-

metallics 1987, 6, 2437.

Article

Organometallics, Vol. 30, No. 5, 2011 1265

Figure 8. Potential energy surface for the conversion of B-ee to C-ee via in-plane (pairwise) migration and merry-go-round processes

(where P = PPh₂). ΔE and ΔH are given in kcal/mol, relative to B-ee.

The mechanism that involves a merry-go-round migration

of two CO groups and one of the phosphorus atoms in the

equatorial plane of the B-ee cluster is shown in eq 6. This

mechanism achieves in one step, involving the simultaneous

migration of three equatorial groups, what takes three

separate steps in the mechanism that involves the pairwise

exchange of a CO and one of the phosphorus atoms (eq 4).

However, the simplicity and economy of the merry-go-round

mechanism in eq 6 does not, of course, necessarily make it the

energetically preferred mechanism.

barrier to rearrangement of 2 to 3 by 5 kcal/mol and why

fluorination of the benzo group apparently effects an even

larger reduction in the enthalpy of activation, as evidenced

by our failure to detect 5 as an intermediate in the formation

of 6.

Our computational studies were initiated by performing

DFT calculations on the mechanism of the rearrangements

of bridged to chelated Os₃(CO)₁₀(P-P) clusters, using dppen

as the diphosphine ligand. The mechanism involving the

pairwise exchange of a P and a CO ligand was studied first.

Optimized geometries of the two steroisomers of the bridged

cluster 1,2-Os₃(CO)₁₀[(Z)-Ph₂PCHdCHPPh₂] were ob-

tained: one stereoisomer, B-ee, with both phosphorus atoms

equatorial, and the other stereoisomer, B-ea, with one phos-

phorus equatorial and the other axial. B-ea was computed to

be higher in enthalpy than B-ee by 9.1 kcal/mol (the energies

and all of the geometrical parameters of the minima and

transition structures are given in the Supporting

Information).

The mechanism for isomerization of B-ee to B-ea in eq 3

involves the pairwise exchange of the trans diaxial CO

groups across the Os₁-Os₃bond, accompanied by conrota-

tion of four other substituents that are attached to these two

osmium atoms.⁶¹The transition structure (TS1) is the dou-

bly CO bridged species shown in Figure 7. The barrier height

for the mechanism in eq 3 was computed to be 10.3 kcal/mol,

only 1.2 kcal/mol higher than the enthalpy of the B-ea

isomer.

VI. Calculations of the Mechanisms and Barrier Heights

for Diphosphine Isomerization in Os₃(CO)₁₀(P-P). We

decided to use the results of electronic structure calculations

to determine which of the above mechanisms is favored in the

rearrangements of Os₃(CO)₁₀(P-P) clusters. As a test of our

computational prediction of the preferred mechanism, we

could assess how well the barrier heights, computed for the

preferred mechanism, would reproduce the experimental

activation enthalpies in Table 4. If the results of our calcula-

tions were thus validated, we could then use them to attempt

to understand the observed variations in barrier heights with

the structures of the phosphine ligands. For example, we

could address the question of why replacement of the etheno

group in dppen by the benzo group in dppbz lowers the

As shown in eq 4, the bridged intermediate B-ea can be

transformed into the chelated structure C-ea by the pairwise

exchange of axial CO and phosphine groups across the

Os₂-Os₁bond, concomitant with conrotation of the other

1266 Organometallics, Vol. 30, No. 5, 2011

Zhang et al.

four groups that lie approximately in the same plane as the

two migrating groups and the two osmium atoms to which

they are attached. The transition structure (TS-ea) for this

process was located, and its structure is given in Figure 7. As

showninFigure8, TS-ea is calculated to be 29.0 kcal/mol higher

in enthalpy than B-ee and 18.7 kcal/mol higher than TS1. TS-ea

is the highest transition structure in the isomerization of B-ee to

C-ee by the three-step mechanism in eqs 3-5.

TS1 and TS-ea are both doubly bridged. Why is the latter

transition structure computed to be so much higher in

enthalpy than the former? Part of the reason is that passage

from B-ea back to B-ee via TS1 is more exothermic than

passage from B-ea to C-ea by 6.6 kcal/mol, but this difference

can be only a small contributor to the 18.7 kcal/mol differ-

ence between the two transition structures.

TS-ee is, indeed, a bona fide transition structure and not a

short-lived intermediate on the PES for rearrangement of B-

ee to C-ee was verified by normal-mode analysis, which

found an imaginary frequency of 77i for TS-ee. An IRC

calculation confirmed that TS-ee does, in fact, connect B-ee

directly to C-ee.

It should be noted, however, that in Os₃(CO)₁₂, a D_3h

structure with three bridging equatorial carbonyls is calcu-

lated to be ca. 5 kcal/mol higher in energy than a C_2vdoubly

bridged transition structure for exchanging a pair of trans

diaxial carbonyl groups.⁶⁴Therefore, at first it seems surpris-

ing that the preferred mechanism for the rearrangement of B-

ee to C-ee is predicted to occur via the triply bridged, merry-

go-round transition structure (TS-ee), rather than by the

doubly bridged transition structure for exchange of an axial

phosphine and a trans axial CO between just two Os atoms

We believe that most of the difference is caused by the fact

that in TS1 both of the bridging groups are carbonyls; the

presence of an empty π* MO in an η¹terminal CO group

makes it easy for the carbonyl to accept a pair of electrons

from an adjacent metal atom, to become an η²bridging

carbonyl group. In contrast, a phosphine does not have such

a low-energy, empty MO. This difference between carbonyl

groups and phosphines is the reason bridging CO groups are

common in transition-metal chemistry but bridging phos-

phines are not. It is the bonding of the phosphine group

simultaneously to the Os₁and Os₂atoms in TS-ea that makes

this transition structure much higher in energy than either

TS1 or TS2, in which both of the bridging ligands are CO.

Evidence in support of this explanation for the high energy

of TS-ea, relative to TS1 and TS2, comes from bond lengths.

For example, the CO that bridges in TS-ea has an average

Os-C bond length that is only 0.204 A longer in this

transition structure than the average length of the Os-C

bonds to this carbonyl group in B-ea and C-ea. This is typical

of the modest difference between the lengths of the Os-C

bonds to the bridging carbonyl groups in TS1, TS2, and TS-

ea and to the terminal carbonyl in B-ee, B-ea, and C-ea. In

contrast, the phosphine that bridges in TS-ea has an average

Os-P bond length that is fully 0.355 A longer in this

transition structure than the average length of the Os-P

bonds in B-ea and C-ea.

As shown in Figure 8, C-ea is computed to have a 2.5 kcal/

mol lower enthalpy than B-ea. A barrier of only 0.4 kcal/mol is

computed to separate C-ea from C-ee by doubly CO bridged

TS2, which is below B-ea by 2.1 kcal/mol. The computed

enthalpy difference of -1.7 kcal/mol between B-ee and C-ee,

in favor of the latter species, compares well with the experimen-

tally determined ΔGꢀ value of -1.4 kcal/mol that was obtained

in our thermal equilibration studies on Os₃(CO)₁₀(dppen).^21a

There is a plethora of NMR reports on CO scrambling

about triangular faces in polynuclear clusters.⁶³Indeed, our

calculations find TS-ee (Figure 8), the triply bridged transi-

tion structure for the one-step conversion of B-ee to C-ee in

eq 6, to be 4.3 kcal/mol lower in enthalpy than TS-ea, the

highest energy transition structure for the three-step conver-

sion of B-ee to C-ee via the mechanism in eqs 3-5. That

(TS-ea). Nevertheless, the results of the calculations by Li et

64

al. on Os₃(CO)₁₂

and ours on Os₃(CO)₁₀[(Z)-

Ph₂PCHdCHPPh₂] are, in fact, wholly consistent with each

other. As shown in Figure 8, although TS-ee is 4.3 kcal/mol

lower in enthalpy than TS-ea, 9.1 kcal/mol of the enthalpy of

TS-ea comes from the enthalpy that is required to isomerize

B-ee to B-ea, so that the ligand exchange reaction can occur.

Starting from B-ea, the barrier to exchange of an axial

phosphine and a trans axial CO via TS-ea is only ΔH^q=

19.9 kcal/mol, which is 4.8 kcal/mol lower than ΔH^q= 24.7

kcal/mol for exchange of three equatorial ligands via TS-ee.

Unfortunately, the B3LYP/SDD calculations on Os₃-

(CO)₁₂by Li et al. were performed without polarization

functions in the D95 basis set.⁶⁴Therefore, the barrier

heights to rearrangement, computed by Li et al., are not

directly comparable to the barrier heights for rearrangement

of Os₃(CO)₁₀[(Z)-Ph₂PCHdCHPPh₂], computed by us with

the D95* basis set on phosphorus, oxygen, and carbon. In

order to make meaningful comparisons between the barrier

heights for rearrangement of these two molecules, we re-

peated the B3LYP/SDD calculations on Os₃(CO)₁₂with the

D95* basis set on phosphorus, oxygen, and carbon.

Our calculations on Os₃(CO)₁₂found the energies of the

doubly bridged C_2vand triply bridged D_3htransition struc-

tures to be respectively 13.0 and 17.7 kcal/mol higher than

that of the D₃equilibrium geometry. Both of these transition

structure energies are 4.7 kcal/mol higher than the values

published by Li et al. Clearly, inclusion of polarization

functions in the basis sets for carbon and oxygen stabilizes

the equilibrium geometry of Os₃(CO)₁₂more than the transi-

tion structures for CO migrations in this molecule.

The energy ΔΕ = 13.0 kcal/mol for the doubly bridged

Os₃(CO)₁₂transition structure, relative to the equilibrium

geometry, is 8.1 kcal/mol smaller than ΔE = 21.1 kcal/mol

between TS-ea and B-ea; the relative energy ΔΕ = 17.7 kcal/

mol for the triply bridged Os₃(CO)₁₂transition structure is

8.5 kcal/mol smaller than ΔE = 26.2 kcal/mol between TS-

ee and B-ee. Therefore, we believe that ca. 8 kcal/mol

represents a reasonable estimate of the energetic penalty

paid by Os₃(CO)₁₀[(Z)-Ph₂PCHdCHPPh₂], relative to Os₃-

(CO)₁₂, for having a bridging phosphine in place of a

bridging carbonyl group in the transition structures for

ligand exchanges between both two and three Os atoms.⁶⁵

The calculated activation parameters of ΔH^q= 24.7 kcal/

mol and ΔS^q= -2.7 cal/(mol K) for isomerization of B-ee to

(63) (a) Cotton, F. A. Inorg. Chem. 1966, 5, 1083. (b) Orlandi, A.;

Frey, U.; Suardi, G.; Merbach, A. E.; Roulet, R. Inorg. Chem. 1992, 31,

ꢁ

1304. (c) Aime, S.; Dastru, W.; Gobetto, R.; Krause, J.; Violano, L.

Inorg. Chim. Acta 1995, 235, 357. (d) Besancon, K.; Laurenczy, G.;

Lumini, T.; Roulet, R.; Bruyndonckx, R.; Daul, C. Inorg. Chem. 1998,

37, 5634. (e) Waterman, S. M.; Humphrey, M. G. Organometallics 1999,

18, 3116. (f) Canal, J. P.; Jennings, M.; Yap, G. P. A.; Pomeroy, R. K.

Dalton Trans. 2008, 1375.

(64) Li, Q.-S.; Xu, B.; Xie, Y.; King, R. B.; Schaefer, H. F. Dalton

Trans. 2007, 4312.

Article

Organometallics, Vol. 30, No. 5, 2011 1267

Table 5. Computed Energy and Enthalpy Differences between the Bridged and Chelated Os₃(CO)₁₀(diphosphine) Isomers and Computed

Activation Energies and Enthalpies for the Bridged to Chelated Isomerization Reaction via the Merry-Go-Round Mechanism^d

^aDue to the size of and numbers of degrees of freedom in this ligand, no attempt was made to locate the transition structure for the rearrangement of

the bridged to the chelated isomer. ^bThe corresponding experimental value for entry a is ΔH^q= 26.5(6) kcal/mol. ^cThe corresponding experimental

value for entry b is ΔH^q= 21.6(3) kcal/mol. ^dAll energies are in kcal/mol.

C-ee via triply bridged TS-ee are in good agreement with

the experimental values (Table 4) of ΔH^q= 26.5(6) and

ΔS^q= -5(2) cal/(mol K).^21aThis agreement provides sup-

port for the computational finding that the one-step, merry-

go-round mechanism in eq 6 really is preferred to the three-

step mechanism in eqs 3-5 for the rearrangement of B-ee to

C-ee. As discussed in section VII, further evidence comes

from the good agreement of the activation parameters,

calculated for the rearrangements via merry-go-round tran-

sition structures of Os₃(CO)₁₀(P-P) complexes, containing

other diphosphine ligands (Table 5), with the values obtained

experimentally (Table 4).

Figure 9. Schematic depiction of how, on going from the

bridged to the chelated isomer, the decrease in P-C-C bond

angle and the concomitant increase in the R-C-P angle reduce

the steric interactions between each phosphorus atom and the

phenyl groups attached to it with the R group that is bonded to

the same carbon atom as the phosphorus.

VII. Computed Reactivity Trends as a Function of the

Ancillary Diphosphine Ligand in Isomeric Os₃(CO)₁₀(P-P)

Clusters. Having established computationally that the mer-

ry-go-round scenario is the operative mechanism for the

isomerization, from bridging to chelating, of the dipho-

sphine ligand in 1,2- and 1,1-Os₃(CO)₁₀[(Z)-Ph₂PCHd

CHPPh₂], we computed the energetics associated with this

mechanism for the isomerization of other rigid diphosphine

ligands about the triosmium cluster. The results of these

calculations are summarized in Table 5. All of the isomeriza-

tion reactions are computed to be exothermic. The calculated

ΔH values for the isomerizations display more than an order

of magnitude difference, ranging from ΔH = -1.7 kcal/mol

for the dppen ligand (entry a) to ΔH = -21.5 kcal/mol for

the highly substituted 1,2-bis(diphenylphosphino)-2,3,4,5-

tetramethylbenzene ligand (entry f).

The size of this decrease is consistent with our conjecture that

the reason only chelated isomer 6 was observed in the

reaction between 1 and dppbzF₄is that 5 rearranges to 6

too quickly to be detected in our experiments.

The computational results in Table 5 underscore the influ-

ence that steric effects within the diphosphine ligands have on

both the thermodynamics and kinetics of these isomerization

reactions. As already discussed in section I, in connection with

the X-ray structures of 2 and 3, the rearrangement of the

former to the latter decreases the C-C-P bond angle from ca.

126ꢀ in the six-membered ring of 2 to ca. 117ꢀ in the five-

membered ring of 3. Similar sized reductions in these internal

angles are calculated to occur in the rearrangements of the

bridged to the chelated isomers for all of the ligands in Table 5

(see the Supporting Information for the optimized structures).

As shown schematically in Figure 9, decreasing the C-C-P

bond angle and, concomitantly, increasing the R-C-P angle

reduce the strain that is engendered by steric interactions

between each phosphorus atom and the phenyl groups at-

tached to it with the group, R, that is bonded to the same

carbon as the phosphorus. Consequently, as the size of the R

group increases, so does the amount of strain relieved by the

rearrangement of the bridged to the chelated isomer, and as

might be anticipated, the increased exothermicity of this

reaction lowers the height of the barrier to it.

The calculated activation enthalpies range from ΔH^q=

24.7 kcal/mol for the dppen ligand to ΔH^q= 18.2 kcal/mol

for the dppbzF₄ligand (entry e), the most highly substituted

diphosphine that has been studied experimentally. The com-

puted value of ΔH^q= 22.5 kcal/mol for the isomerization of

2 to 3 (entry b in Table 5) agrees well with the value of ΔH^q=

21.6(3) kcal/mol, measured in the current study (Table 4).

Upon perfluorination of the benzo group, the enthalpy of

activation for rearrangement is computed to decrease by 4.3

kcal/mol for the isomerization of 5 to 6 (entry e in Table 5).

The effective size of the R group increases from left to right

across Table 5, from R = H in entry a to R = C-CH₃for the

o-methyl groups in entry f, and the exothermicity of the

conversion of the bridged to the chelated isomer also

(65) We find that a similar energetic penalty of ca. 8 kcal/mol is

incurred when the PH₃group in Os₃(CO)₁₁(PH₃) replaces a bridging CO

in Os₃(CO)₁₂in the two-center and three-center transition structures for

rearrangement.

1268 Organometallics, Vol. 30, No. 5, 2011

Zhang et al.

increases in magnitude across Table 5. An exception to this

trend is provided by entries d and e, where the two additional

fluorines in the meta positions of entry e have little effect on

the exothermicity of the rearrangement. This fact provides

evidence that the exothermicity of the rearrangements of

both of these fluorinated ligands is due to steric effects from

the o-fluorines, rather than to electronic factors.

Conclusions

The synthesis and characterization of new dppbz- and dppbzF₄-

substituted triosmium decacarbonyl clusters are reported, and the

rate of conversion of bridged 1,2-Os₃(CO)₁₀(dppbz) (2) tochelated

1,1-Os₃(CO)₁₀(dppbz) (3) has been measured over a 25 ꢀC temp-

erature range. The slightly negative entropy of activation supports

a nondissociative mechanism for the isomerization process.

The results of DFT calculations indicate that a merry-go-

round mechanism, involving the concerted exchange of CO and

diphosphine ligands about the equator of the cluster polyhedron

is the lowest energy pathway. The calculations confirm the

ability of a tertiary phosphine to function in a nonclassical

coordination mode and to simultaneously bind to adjacent

metal centers in a bridging fashion. However, calculations of

the barrier heights for pairwise exchange of axial ligands indicate

the bridging by phosphorus provides considerably less stabiliza-

tion than bridging by a carbonyl carbon. Comparison of the

energies computed for P-bridged structures in Os₃(CO)₁₀[(Z)-

Ph₂PCHdCHPPh₂] and the corresponding doubly and triply

CO bridged structures in Os(CO)₁₂indicates an energy penalty

of about 8 kcal/mol for P versus CO bridging.

Increasing the effective size of the substituent, R, that is

attached to the same carbon as phosphorus in the diphosphine

ligands is shown, by the results of both our experiments and our

calculations, to increase the reactivity of the bridging dipho-

sphine ligand in this genre of cluster compounds. DFT calcula-

tions also find that increasing the geminal substituent size has an

even greater effect on increasing the exothermicity of the re-

arrangement of the bridged to the chelated isomer. These

findings are attributed to the ca. 10ꢀ decrease in the C-C-P

bond angle that occurs during the transformation of the six-

membered ring in the bridged isomer to the five-membered in the

chelated isomer and the concomitant increase in the R-C-P

angle. Increasing the latter angle decreases the magnitude of the

nonbonded interactions of phosphorus and the phenyl groups

attached to it with the substituent that is attached to the same

carbon. The larger the effective size of the substituent group, R,

the larger is the amount of steric strain removed on the rearran-

gement of the bridged to the chelated isomer of Os₃(CO)₁₀(P-P).

To the extent that the steric interactions between the

geminal R and PPh₂groups in the bridged isomer decrease,

not only in the chelated isomer but also in the transition

structure leading to it, one might expect the barrier heights to

decrease with the size of R and, hence, with the increase in the

magnitude of the exothermicity of the rearrangement.⁶⁶An

exception to the correlation between increased exothermicity

and decreased barrier height is R = CH₃(entry c). Although

the exothermicity of the rearrangement of entry c is com-

puted to be 1.5 kcal/mol greater than that of entry b (R =

C-H), the barrier to the rearrangement of entry c is actually

computed to be 1.1 kcal/mol higher than that of entry b.

Apparently the methyl groups in entry c undergo some small

destabilizing interactions in the transition structure for

rearrangement that are absent in the chelated product.

If the increase in exothermicity and decrease in barrier

height in Table 5 both have their origin in relief of strain in

the bridged reactant, the effect of the size of R should have a

larger influence on the exothermicity than on the barrier

height. For example, the 9.1 kcal/mol calculated increase in

exothermicity from entry a to entry e is accompanied by only

a 6.9 kcal/mol decrease in barrier height. As would be

expected, relief of the strain that is present in the bridged

reactant is less complete in the transition structure leading to

the chelated product than in the chelated product itself.

(66) A reviewer has suggested that charge effects, as opposed to steric

effects, may be important in the transition state of our isomerization

reactions, especially for the dppz and dppzF₄ligands. With the

o-(Ph₂P)₂C₆F₄ligand, the migrating phosphine (P₁) shows evidence

for enhanced rehybridization in the TS, in order to help relieve the steric

interaction between the P₁Ph₂group and the adjacent fluorine (the

Ph-P₁-Ph angle is 113.8ꢀ in the dppbzF₄TS vs 108.6ꢀ in the dppbz TS,

as illustrated below). This angular change leads to a flattening of P₁, and

this allows it to bond to Os₂more strongly in the dppbzF₄TS than in the

dppbz TS, as shown by Os₂-P₁bond lengths of 2.75 A and 3.05 A,

respectively (see the Supporting Information). The additional electron

donation from the P₁Ph₂group in the dppbzF₄TS makes Os₂more

negative than in the dppbz TS (-1.546 vs -1.487). Although P₁is

slightly more positive in the former TS than in the latter (1.119 vs 1.100),

most of the positive charge that one might have expected to find on P₁

seems to be neutralized by the transfer of electrons from Os₁back to P₁.

The charges on Os₁are respectively -1.324 and -1.403 in the dppbzF₄

TS and dppbz TS. Since the NBO charges for the bridged and chelated

isomers of the dppz- and dppzF₄-substituted clusters reveal no signifi-

cant differences, NBO differences in the TS charges are best viewed as

arising from the rehybridization that accompanies the P₁atom in the

dppzF₄TS, relative to the dppz TS, as influenced by the neighboring

fluorine atom.

Acknowledgment. Financial support from the Robert A.

Welch Foundation (Grant B-0027 to W.T.B. and Grant

B-1093 to M.G.R.) and from the National Science Founda-

tion (to W.T.B.) is greatly appreciated. X.W. acknowledges

support by the U.S. Department of Energy, Office of

Science, under Contract No. DE-AC05-00OR22725 mana-

ged by UT Battelle, LLC. The NSF-MRI program grant

CHE-0840518 is gratefully acknowledged for support of the

NMR facilities at UNT. NSF support of the UNT computa-

tional facilities through grant CHE-0741936 is acknowl-

edged. Preliminary experimental work conducted by Mr.

Sean Hunt and preliminary calculations by Dr. Haiyan Wei

are acknowledged.

Supporting Information Available: Text, figures, tables, and

CIF files giving an Eyring plot of the dppbz isomerization data,

atomic coordinates and energies of all optimized stationary points

and transition states, NBO charges and selected bond distances for

the merry-go-round TS structures containing dppz and dppzF₄

ligands, and crystallographic data. This material is available free of

charge via the Internet at http://pubs.acs.org.

Article Doi

Experimental and computational studies of the isomerization reactions of bidentate phosphine ligands in triosmium clusters: Kinetics of the rearrangements from bridged to chelated isomers and X-ray structures of the clusters Os3(CO)10(dppbz), 1,1-Os3(CO) 10(dppbzF4), HOs3(CO)9[μ-1,2- PhP(C6H4

DOI: 10.1021/om100783f

Source and publish data:

Authors:

Article abstract of DOI:10.1021/om100783f

Full text of DOI:10.1021/om100783f

Products guided by the article

R&D Labs maybe for 1273331-81-9

Relevant to this article

Hot Product