Experimental and computational study of the transformation of terminal alkynes to vinylidene ligands on trans-(chloro)bis(phosphine)Rh fragments and effects of phosphine substituents

Organometallics 2007, 26, 3385-3402

3385

Experimental and Computational Study of the Transformation of

Terminal Alkynes to Vinylidene Ligands on

trans-(Chloro)bis(phosphine)Rh Fragments and Effects of Phosphine

Substituents

Douglas B. Grotjahn,*^,†Xi Zeng,^†Andrew L. Cooksy,^†W. Scott Kassel,^‡

Antonio G. DiPasquale,^‡Lev. N. Zakharov,^‡and Arnold L. Rheingold^‡

Department of Chemistry and Biochemistry, 5500 Campanile DriVe, San Diego State UniVersity,

San Diego, California 92182-1030, and Department of Chemistry and Biochemistry,

UniVersity of California, San Diego, La Jolla, California 92093-0385

ReceiVed April 11, 2007

Experimental and computational evidence points to unimolecular transformation of terminal alkynes

on the title Rh(I) metal fragments. Lack of isotopic scrambling in double-crossover experiments is

inconsistent with a previously proposed bimolecular pathway. Focusing on a unimolecular manifold,

alkyne binding to the metal forms the Rh(I) alkyne π-complex 2, which isomerizes to the Rh(III) hydrido-

(alkynyl) species 4, ultimately leading to Rh(I) vinylidene product 5. In making alkyne-free precursors,

use of heterocyclic ligand (i-Pr)₂PIm′ (1b, Im′ ) 1-methyl-4-tert-butylimidazol-2-yl) led to species 8

with a labile P,N chelate, whereas a geometrically similar o-tolyl ligand suffered metalation at the methyl

group and was unsuitable for alkyne transformation studies. Kinetic studies comparing 1b and (i-Pr)₂PPh

(1c) allowed determination of rate constants for the alkyne binding event and conversion of 2 to 5 (the

latter, k_2-5, being 9.6 times faster for 1b). Based on a scan of the two-dimensional reaction surface,

combined density functional/molecular mechanics calculations predict that η²-(C,H) alkyne complex 3 is

in a fast equilibrium with the lower energy hydrido(alkynyl) complex 4, and neither species is expected

to be present at observable concentrations. Eyring model estimates of the rate constants from these

computational data predict the available experimental values in this work to within a factor of 2 and the

ratio of the rate constants k_2-5for 1b and 1c to within 10%. The calculations also agree with the qualitative

observation that reaction rates are faster for both ligands 1b and 1c than for (i-Pr)₃P and predict that

reactions using triphenylphosphine will be faster than those with (i-Pr)₃P. NMR coupling constants,

1

particularly J_CCvalues, were used to evaluate bonding and back-bonding in isotopomers of 2a-c and

5a-c derived from H¹³C¹³CH.

enables dramatic changes both in the stability of coordinated

hydrocarbon ligands and in the specificity of reactions on the

ligand undergoing transformation. For example, in the gas

phase, vinylidene (:CdCH₂) is at least 40 kcal mol^-1less

stable than its isomer acetylene (HCtCH),^10,11whereas this

relationship can be reversed on coordination of both species

to a metal.¹²Moreover, depending on metal and ligands,

the changes in reactivity at carbons on going from a free ter-

minal alkyne to a coordinated vinylidene allow a wide variety

of functionalizations at C1 of the former alkyne that are

impossible to achieve on the alkyne itself. Furthermore, vi-

nylidene complexes are emerging as useful catalysts for alkene

and alkyne metathesis chemistry. For all of these reasons, the

increasing importance of vinylidenes in organic synthesis,¹³

organometallic catalysis, and materials chemistry is highlighted

Introduction

Many organic transformations of importance and utility in

both industrial-scale^1,2and fine-chemical synthesis involve

activation of hydrocarbon C-H bonds, a topic of intense and

continuing interest.^3-9

Such reactions offer the ability to change the size of carbon

frameworks or add new and higher-value functional groups at

carbon. In addition, the use of transition metal compounds

* Corresponding author. E-mail: grotjahn@chemistry.sdsu.edu.

^†San Diego State University.

^‡University of California, San Diego. X-ray crystal structures.

(1) Satterfield, C. N. Heterogeneous Catalysis in Industrial Practice, 2nd

ed.; McGraw-Hill: New York, 1991.

(2) Weissermel, K.; Arpe, H.-J. Industrial Organic Chemistry, 2nd

Revised and Extended Edition; VCH: New York, 1993.

(3) Crabtree, R. H. J. Chem. Soc., Dalton Trans. 2001, 2437-2450.

(4) Ritleng, V.; Sirlin, C.; Pfeffer, M. Chem. ReV. 2002, 102, 1731-

1770.

(5) Klei, S. R.; Golden, J. T.; Burger, P.; Bergman, R. G. J. Mol. Catal.

A: Chem. 2002, 189, 79-94.

(10) Ervin, K. M.; Ho, J.; Lineberger, W. C. J. Chem. Phys. 1989, 91,

5974-5992.

(11) Chen, Y.; Jonas, D. M.; Kinsey, J. L.; Field, R. W. J. Chem. Phys.

1989, 91, 3976-3987.

(6) Labinger, J. A.; Bercaw, J. E. Nature 2002, 417, 507-514.

(7) ActiVation and Functionalization of C-H Bonds; Goldberg, K. I.,

Goldman, A. S., Eds.; American Chemical Society: Washington, D.C.,

2004; Vol. 885.

(8) Handbook of C-H Transformations; Dyker, G., Ed.; Wiley-VCH:

Weinheim, Germany, 2005.

(12) Wakatsuki, Y. J. Organomet. Chem. 2004, 689, 4092-4109.

(13) In particular, for recent uses of Rh vinylidenes in synthesis: (a)

Trost, B. M.; Rhee, Y. H. J. Am. Chem. Soc. 2003, 125, 7482-7483. (b)

Kim, H.; Lee, C. J. Am. Chem. Soc. 2005, 127, 10180-10181. (c) Kim,

H.; Lee, C. J. Am. Chem. Soc. 2006, 128, 6336-6337. (d) Chen, Y.; Lee,

C. J. Am. Chem. Soc. 2006, 128, 15598-15599. (e) Trost, B. M.; McClory,

A. Angew. Chem., Int. Ed. 2007, 46, 2074-2077.

(9) Lersch, M.; Tilset, M. Chem. ReV. 2005, 105, 2471-2526.

Publication on Web 06/01/2007

3386 Organometallics, Vol. 26, No. 14, 2007

Scheme 1. Mechanistic Alternatives for Alkyne-to-Vinylidene Transformations

Grotjahn et al.

by the number and variety of recent reviews,^14-19along with

earlier ones.^20-24

of the vinyl C-H toward the metal to give E.²⁸One variant of

this process starts with a preformed M-H species rather than

requiring protonation.^29,30

The utilization of novel alkyne and vinylidene reactivity in

synthesis can be supported by studying the mechanism of

alkyne-to-vinylidene transformation, a subject recently re-

viewed.¹²The first proposal of alkyne-to-vinylidene conversion

on a metal was made in 1972.²⁵Basic mechanistic alternatives

shown in Scheme 1 all proceed from an initial alkyne-metal π

complex (A). The first route (path a) involves oxidative addition

of the alkyne C-H bond to give B, followed by a 1,3-shift of

the hydride atom over the M-C1-C2 framework leading to

C. One variant of this mechanism invokes deprotonation of the

M-H unit and reprotonation at C2.^25,26The second route (path

b) features a direct 1,2-shift of the terminal C-H of the

coordinated alkyne to C2, perhaps by way of an η²-(C,H)

complex (D).²⁷Finally, a third possibility (path c) features initial

protonation of coordinated alkyne in A, followed by 1,2-shift

In our ongoing studies of catalytic alkyne transformations

on CpRu fragments facilitated by an internal base within a

phosphine ligand,^31-33thus far we have found alkyne π-com-

plexes to be elusive. This led us to study other metal fragments.

A thorough series of pioneering synthetic papers from Werner

and co-workers^34-39showed that alkyne π-complexes of trans-

(chloro)bis(phosphine)M (M ) Rh, Ir) (e.g., 2, Scheme 2) could

be readily prepared using bulky, electron-rich phosphines such

as (i-Pr)₃P (1a). More importantly, these species isomerized to

vinylidene complexes 5 under moderate conditions (e.g., hours

at 50 °C for Rh derivatives). Intriguingly, there were indications

that added pyridine could either accelerate the conversion³⁷or

bind to the metal and allow trapping of hydrido(alkynyl) isomers

of type B.⁴⁰Finally, a major 1997 computational study⁴¹and

some circumstantial evidence³⁷suggested that the conversions

of the latter compounds to vinylidenes followed an unusual

bimolecular pathway (B to C, illustrated at the bottom of

Scheme 1).

(14) Selegue, J. P. Coord. Chem. ReV. 2004, 248, 1543-1563.

(15) Katayama, H.; Ozawa, F. Coord. Chem. ReV. 2004, 248, 1703-

1715.

(16) Valyaev, D. A.; Semeikin, O. V.; Ustynyuk, N. A. Coord. Chem.

ReV. 2004, 248, 1679-1692.

(17) Werner, H. Coord. Chem. ReV. 2004, 248, 1693-1702.

(18) Bruneau, C.; Dixneuf, P. H. Angew. Chem., Int. Ed. 2006, 45, 2176-

2203.

(28) Tokunaga, M.; Suzuki, T.; Koga, N.; Fukushima, T.; Horiuchi, A.;

Wakatsuki, Y. J. Am. Chem. Soc. 2001, 123, 11917-11924.

(29) Olivan, M.; Eisenstein, O.; Caulton, K. G. Organometallics 1997,

16, 2227-2229.

(19) Varela, J. A.; Saa, C. Chem.-Eur. J. 2006, 12, 6450-6456.

(20) Bruce, M. I.; Swincer, A. G. AdV. Organomet. Chem. 1983, 22,

59-128.

(30) Olivan, M.; Clot, E.; Eisenstein, O.; Caulton, K. G. Organometallics

1998, 17, 3091-3100.

(21) Bruce, M. I. Chem. ReV. 1991, 91, 197-257.

(22) Dixneuf, P. H.; Bruneau, C.; Derien, S. Pure Appl. Chem. 1998,

70, 1065-1070.

(31) Grotjahn, D. B.; Incarvito, C. D.; Rheingold, A. L. Angew. Chem.,

Int. Ed. 2001, 40, 3884-3887.

(32) Grotjahn, D. B.; Lev, D. A. J. Am. Chem. Soc. 2004, 126, 12232-

(23) Bruneau, C.; Dixneuf, P. H. Acc. Chem. Res. 1999, 32, 311-323.

(24) Puerta, M. C.; Valerga, P. Coord. Chem. ReV. 1999, 193-195, 977-

1025.

12233.

(33) Grotjahn, D. B. Chem.-Eur. J. 2005, 11, 7146-7153.

(34) Alonso, F. J. G.; Ho¨hn, A.; Wolf, J.; Otto, H.; Werner, H. Angew.

Chem., Int. Ed. Engl. 1985, 24, 406-408.

(35) Werner, H.; Garcia Alonso, F. J.; Otto, H.; Wolf, J. Z. Natursforsch.,

B 1988, 43b, 722-726.

(36) Werner, H.; Brekau, U. Z. Naturforsch., B 1989, 44b, 1438-1446.

(37) Ho¨hn, A.; Werner, H. J. Organomet. Chem. 1990, 382, 255-272.

(38) Werner, H.; Hampp, A.; Peters, K.; Peters, E. M.; Walz, L.; von

Schnering, H. G. Z. Naturforsch., B 1990, 45b, 1548-1558.

(39) Canepa, G.; Brandt, C. D.; Werner, H. Organometallics 2001, 20,

604-606.

(25) Chisholm, M. H.; Clark, H. C. J. Am. Chem. Soc. 1972, 94, 1532-

1539.

(26) Bianchini, C.; Peruzzini, M.; Vacca, A.; Zanobini, F. Organome-

tallics 1991, 10, 3697-3707.

(27) Reviews on alkane complexes: Hall, C.; Perutz, R. N. Chem. ReV.

1996, 96, 3125-3146. Jones, W. D.; Vetter, A. J.; Wick, D. D.; Northcutt,

T. O. In ActiVation and Functionalization of C-H Bonds; American Chemical

Society: Washington DC, 2004; Vol. 885, pp 56-69. Recent leading

references: Castro-Rodriguez, I.; Nakai, H.; Gantzel, P.; Zakharov, L. N.;

Rheingold, A. L.; Meyer, K. J. Am. Chem. Soc. 2003, 125, 15734-15735.

Lawes, D. J.; Geftakis, S.; Ball, G. E. J. Am. Chem. Soc. 2005, 127, 4134-

4135. Lawes, D. J.; Darwish, T. A.; Clark, T.; Harper, J. B.; Ball, G. E.

Angew. Chem., Int. Ed. 2006, 45, 4486-4490.

(40) Wolf, J.; Werner, H.; Serhadli, O.; Ziegler, M. Angew. Chem., Int.

Ed. Engl. 1983, 22, 414-415.

(41) Wakatsuki, Y.; Koga, N.; Werner, H.; Morokuma, K. J. Am. Chem.

Soc. 1997, 119, 360-366.

Transformation of Terminal Alkynes to Vinylidene Ligands

Organometallics, Vol. 26, No. 14, 2007 3387

Scheme 2. Specific Ligands and Metal Complexes Considered in This Study

Therefore, we entered a study of Rh-alkyne and -vinylidene

chemistry by preparing imidazolylphosphine 1b as an analogue

of 1a. Remarkably, immediately on adding colorless 1b and

then 1-hexyne to a pale orange solution of Rh dimer [(µ-Cl)-

Rh(cyclooctene)₂]₂, a deep permanganate-purple color appeared,

and the resulting solution contained not the expected alkyne

complex 2b but rather its vinylidene isomer 5b. This surprisingly

rapid reaction (relative to that using 1a) prompted the studies

reported here. We recently made a preliminary report of

experimental evidence from double-crossover experiments: the

most significant finding was retention of isotopic composition,

which was inconsistent with a bimolecular pathway from B to

C.⁴²Here we detail and expand on this experimental work and

report a full computational study of the alkyne-to-vinylidene

transformation.⁴³In focusing on the unimolecular reaction

pathway consistent with our experimental results, an η²-(C,H)

alkyne complex (D) emerged as a potential intermediate. In

addition, we report a series of experiments and calculations

elucidating the effects of the heterocycle, which in the case of

chlorobis(phosphine)Rh(I) are minimal on the alkyne-to-vi-

nylidene pathway but are profound on the structure and reactivity

of alkyne-free precursors. Of more general significance, this

paper also examines the steric and electronic effects of the

phosphine substituents, a matter of increasing interest as

synthetic applications continue to appear.¹³

Therefore, in addition to (i-Pr)₂PIm′ (1b, Im′ ) 1-methyl-4-

tert-butylimidazol-2-yl) we made two model ligands, (i-Pr)₂PPh

(1c) and (i-Pr)₂P(o-Tol) (1d),⁴⁴for comparison along with (i-

Pr)₃P. The Ph ligand 1c was chosen to be a closer electronic

analogue to (i-Pr)₂PIm′ (1b) than 1a, and the o-Tol analogue

1d was chosen as an analogue electronically similar to 1c with

the additional steric contribution of the ortho-methyl group,

designed to model the steric effect of the N-methyl group in

1b.

With a total of four ligands (1a-1d) in hand, our major

objectives were (1) to make trans-(chloro)(CO)[(i-Pr)₂PR]₂Rh

complexes 6a-6d and use the infrared stretching frequency of

the CO ligand and solid-state bond lengths and angles as

reporters of the steric and electronic effects of phosphine

substituent R, (2) to make pure complexes of the form [(chloro)-

((i-Pr)₂PR)₂Rh]_n(n ) 1 or 2) for alkyne binding studies, (3) to

make authentic samples of alkyne and vinylidene complexes

from these species, and (4) to study kinetics of alkyne

transformation on Rh. Interesting effects of the heterocycle on

precursor structure are described below. As our experimental

study evolved, a fifth major objective became the synthesis and

characterization of isotopically labeled alkyne and vinylidene

complexes and crossover experiments involving the alkyne

species. Not only did isotopic labeling prove decisive in

determining the molecularity of alkyne-to-vinylidene transfor-

Experimental Results and Discussion

(43) While this work was in progress, a computational study of alkyne-

to-vinylidene transformation on the simple fragment trans-(chloro)Rh(PH₃)₂

appeared: Suresh, C. H.; Koga, N. J. Theor. Comput. Chem. 2005, 4, 59-

73. No other phosphines were considered in this study.

Experimental Design. Our initial goal in this study was to

elucidate the role of the phosphine heterocyclic substituent in

the surprisingly rapid formation of a Rh vinylidene complex.

(44) Riihimaki, H.; Kangas, T.; Suomalainen, P.; Reinius, H. K.;

Jaaskelainen, S.; Haukka, M.; Krause, A. O. I.; Pakkanen, T. A.; Pursiainen,

J. T. J. Mol. Catal. A: Chem. 2003, 200, 81-94.

(42) Grotjahn, D. B.; Zeng, X.; Cooksy, A. L. J. Am. Chem. Soc. 2006,

128, 2798-2799.

3388 Organometallics, Vol. 26, No. 14, 2007

Table 1. Comparison of Selected Data for trans-(chloro)(CO)[(i-Pr)₂PR¹]₂Rh (6)^a

Grotjahn et al.

infrared

NMR^b

¹³C{¹H} for CO (CDCl₃)

X-ray diffraction

complex

R¹

ν

_CO(cm^-1, CH₂Cl₂)

³¹P{¹H} (CDCl₃)

Rh-P (Å)

6a

6b

6c

6d

i-Pr

Im′

Ph

1946^c

1966

1965

1960

189.0 (dt, J_C-Rh) 75.0, J_C-P) 15.0)

186.4 (dt, J_C-Rh) 73.6, J_C-P) 15.5)

187.3 (dt, J_C-Rh) 73.6, J_C-P) 15.3)

186.8 (dt, J_C-Rh) 74.6, J_C-P) 15.4)

49.9 (J ) 119.5)^d,48

25.3 (J ) 118.5)

42.6 (J ) 124.4)

31.1 (J ) 121.0)

2.3488(3)⁴⁷

2.3229(16)

2.3355(11)

2.3336(19)

o-Tol

^{a 31}P NMR and X-ray data for 6a are from the literature cited, whereas all other values in Table 1 are from this work. ^bChemical shifts in ppm, coupling

c

constants in Hz. In CDCl₃unless otherwise stated. Literature values: 1950,⁴⁹1947,⁴⁷1946⁵⁰(all in CH₂Cl₂), 1948 (CsI pellet),⁵¹1943.2 (C₆D₆),⁵²1940

(matrix unspecified),⁵³1938 (Nujol).^{48,54 d}In C₆D₆.

Figure 2. ORTEP view of 6d, with ellipsoids shown at 30%

probability. Only one position of the disordered Cl atom and CO

ligand is shown for clarity.

Figure 1. ORTEP view of 6b, with ellipsoids shown at 30%

probability. Only one position of the disordered Cl atom and CO

ligand is shown for clarity.

are identical within experimental uncertainty (ca. 0.003 Å). For

6b, this value is only 0.01 Å shorter than for 6c and 6d, whereas

for 6a, it is ca. 0.014 Å longer. These differences are only

slightly greater than experimental uncertainties. In summary,

the three aryl(di-isopropyl)phosphine ligands resemble each

other as a group, distinct from (i-Pr)₃P.

mation, but it also allowed further characterization of metal-

ligand back-bonding and hence effects of phosphine substituent

R.

CO Complexes. Bubbling CO through a solution of the dimer

[(µ-Cl)Rh(cyclooctene)₂]₂, followed by addition of (i-Pr₂)PR,

evaporation of solvent, and recrystallization of the residue

(which was essentially pure product containing some residual

cyclooctene and solvent) afforded trans-(chloro)(CO)[(i-

Pr)₂PR]₂Rh complexes 6b-6d as yellow crystalline solids in

89-91% yields. The structures of the complexes and purities

were verified by a combination of spectral data (key values of

which are shown in Table 1), combustion analysis, and X-ray

crystallography (Table 1 and Figures 1, 2, and S1).⁴⁵

Inspection of IR data acquired in CH₂Cl₂would suggest that

the three ligands bearing one aromatic substituent have very

similar properties,⁴⁶close to being identical within experimental

uncertainty. All three ligands are less electron-rich than the

trialkylphosphine (i-Pr)₃P.⁴⁶

In the crystal structures of 6b-6d, because of the disorder

of the Cl and CO ligands, a well-known phenomenon in

structures of this general type,⁴⁷determination of Rh-Cl, Rh-

C, and C-O distances with precision needed to discuss subtle

bonding changes is impossible. However, data not subject to

this limitation are the Rh-P bond lengths for 6c and 6d, which

Precursors for Study of Alkyne Binding and Transforma-

tion. From the work of Werner et al.,⁵⁵the symmetrical dimer

7c derived from (i-Pr)₂PPh is a known compound that is isolated

as a violet solid in near-quantitative yield from the phosphine

and [(µ-Cl)Rh(cyclooctene)₂]₂in 4 to 1 molar ratio. In accord

with literature data, its ³¹P{¹H} NMR spectrum showed a sharp

doublet at δ 58.9 ppm (¹J_RhP) 197.7 Hz) in C₆D₆and at 55.6

ppm in CD₂Cl₂.

In contrast, when (i-Pr)₂PIm′ and [(µ-Cl)Rh(cyclooctene)₂]₂

were combined in a molar ratio of 4 to 1 in CD₂Cl₂or CH₂Cl₂,

in addition to free cyclooctene, a single unsymmetrical species

was seen, with two sharp ³¹P{¹H} signals at 47.0 ppm (dd, ¹J_RhP

2

1

) 181.1 Hz and J_PP) 42.2 Hz) and 30.0 ppm (dd, J_RhP

)

2

162.6 Hz and J_PP) 42.2 Hz). Structure 8 is consistent with

the data. The upfield ³¹P NMR signal with the smaller Rh-P

coupling is assigned to the ³¹P nucleus involved in the four-

(47) Wilson, M. R.; Prock, A.; Giering, W. P.; Fernandez, A. L.; Haar,

C. M.; Nolan, S. P.; Foxman, B. M. Organometallics 2002, 21, 2758-

2763.

(48) Moigno, D.; Callejas-Gaspar, B.; Gil-Rubio, J.; Werner, H.; Kiefer,

W. J. Organomet. Chem. 2002, 661, 181-190.

(49) Otto, S.; Roodt, A. Inorg. Chim. Acta 2004, 357, 1-10.

(50) Vastag, S.; Heil, B.; Marko, L. J. Mol. Catal. 1979, 5, 189-195.

(51) Intille, G. M. Inorg. Chem. 1972, 11, 695-702.

(52) Wang, K.; Rosini, G. P.; Nolan, S.; Goldman, A. S. J. Am. Chem.

Soc. 1995, 117, 5082-5088.

(45) See Supporting Information for full details.

(46) For uses of Rh-CO stretching frequencies see: (a) Wang, K.;

Goldman, M. E.; Emge, T. J.; Goldman, A. S. J. Organomet. Chem. 1996,

518, 55-68. (b) Huang, A.; Marcone, J. E.; Mason, K. L.; Marshall, W. J.;

Moloy, K. G. Organometallics 1997, 16, 3377-3380. (c) Clarke, M. L.;

Cole-Hamilton, D. J.; Slawin, A. M. Z.; Woollins, J. D. Chem. Commun.

2000, 2065-2066. (d) Cooney, K. D.; Cundari, T. R.; Hoffman, N. W.;

Pittard, K. A.; Temple, M. D.; Zhao, Y. J. Am. Chem. Soc. 2003, 125,

4318-4324. (e) Reference 47.

(53) Busetto, C.; D’Alfonso, A.; Maspero, F.; Perego, G.; Zazzetta, A.

J. Chem. Soc., Dalton Trans. 1977, 1828-1834.

(54) Moigno, D.; Kiefer, W.; Callejas-Gaspar, B.; Gil-Rubio, J.; Werner,

H. New J. Chem. 2001, 25, 1389-1397.

(55) Werner, H.; Kukla, F.; Steinert, P. Eur. J. Inorg. Chem. 2002, 1377-

1389.

Transformation of Terminal Alkynes to Vinylidene Ligands

Organometallics, Vol. 26, No. 14, 2007 3389

Figure 3. Molecular structure of 8, with ellipsoids shown at 30% probability.

membered chelate.^56,57Crystallization afforded pure 8 as a

yellow solid in 95% yield.

All NMR spectra for 8 acquired in CD₂Cl₂at temperatures

ranging from 30 to -80 °C showed a single species whose NMR

data were consistent with the solid-state structure shown in

Figure 3, which is of relevance to the kinetics experiments

described below, conducted at -20 °C in CD₂Cl₂. However,

NMR data on solutions in less-polar solvents such as THF-d₈,

benzene-d₆, and toluene-d₇suggested equilibration of 8 with

the minor, symmetrical species 7b with spectral data and

structure like that of dimer 7c. Evaporation of the less-polar

solvent containing 8 and 7b and dissolution of the residue in

CD₂Cl₂led to complete conversion to 8. Full details and

discussion are in the Supporting Information.⁴⁵

The solid-state structure of 8 shown in Figure 3 confirms

that there is a cis arrangement of phosphines in a distorted

square-planar environment. The greatest deviation among the

five atoms Rh, Cl, P, P, and N used to define a least-squares

plane was seen by the Rh atom (0.0028 Å). Obvious distortions

include the angle subtended by the chelate [P-Rh-N angle )

69.83(5)°] and the P-Rh-P angle of 101.694(18)°. Because

almost identical distortions are seen even in a related (imidazole-

2-thiolato)M complex,⁵⁸in which the bulky (i-Pr)₂P- group is

replaced by a sterically undemanding S atom, these distortions

are considered to be a consequence of the four-membered

chelate rather than the size of the phosphine substituents.

The solid-state structure suggests hindrance to rotation about

the Rh-P(2) bond caused by the pairs of i-Pr groups on each

P as well as by the heterocyclic substituents. Indeed, cooling a

CD₂Cl₂solution of 8 in the NMR probe to -60 °C did not

alter the already-sharp ³¹P NMR resonances but did result in

Whereas one of the heterocyclic nitrogens in 8 could occupy

a coordination site on Rh(I), interaction of a ligand C-H bond

was seen using (i-Pr)₂P(o-Tol). When 1d was combined with

[(µ-Cl)Rh(cyclooctene)₂]₂in a molar ratio of 4 to 1 in CD₂Cl₂,

in addition to free cyclooctene, a major (ca. 80%) unsymmetrical

species was seen, tentatively assigned structure 9. As in the case

of 8, spectral data were consistent with the presence of two

different phosphines; however, a large P-P coupling constant

indicated trans orientation, as seen at 30 °C from two broadened

³¹P{¹H} signals that sharpened at 0 °C to resonances at 72.3

1

the appearance of four featureless H NMR multiplets for the

now-unique i-Pr methine protons, at δ 2.84, 2.48, 1.98, and 1.64

ppm. The signals near 1.2 ppm for the i-Pr methyls also changed

in appearance but were overlapping and not well-resolved. Using

line-shape analysis of ¹H NMR spectra observed between -80

and +30 °C, the barrier for rotation is calculated to be

approximately 13 kcal mol^-1. A second dynamic process was

indicated by observation of spectra at higher temperatures

(between 30 and 105 °C, requiring THF-d₈or toluene-d₈). Under

these conditions, in ¹H and ³¹P{¹H} NMR spectra of 8 changes

indicative of coalescence of the two sets of resonances for the

two imidazolylphosphines suggested the involvement of a

fluxional exchange of the bound and free basic imidazole

nitrogens, ∆H^q) 20.7 kcal mol^-1and ∆S^q) +13 eu. The

positive ∆S^qvalue suggests a dissociative process.

1

2

ppm (dd, J_P-Rh) 119.2 Hz and J_P-P-trans) 367.5 Hz) and

1

2

39.1 ppm (dd, J_P-Rh) 111.5 Hz and J_P-P-trans) 367.5 Hz).

The H NMR spectrum at 0 °C showed a singlet at 1.95 ppm

1

as expected for a methyl group on an aromatic ring, but the

integral corresponded to three protons, not six. Moreover, at

0 °C there was a broad, featureless peak centered on -9.4 ppm

integrating to approximately 2 H. Cooling the sample below

0 °C sharpened many of the ¹H NMR resonances, for example

in the aromatic region, and importantly, upfield, revealing two

broad, featureless one-proton peaks centered at -7.4 and -11.4

ppm. Spectra acquired on the same material dissolved in toluene-

d₈at various temperatures were more complex. Assignment as

the oxidative addition product 9 was supported by the upfield

¹H NMR resonances and by observation in the infrared spectrum

(56) Garrou, P. E. Chem. ReV. 1981, 81, 229-266.

(57) Grotjahn, D. B.; Gong, Y.; DiPasquale, A. G.; Zakharov, L. N.;

Rheingold, A. L. Organometallics 2006, 25, 5693-5695.

(58) Miranda-Soto, V.; Perez-Torrente, J. J.; Oro, L. A.; Lahoz, F. J.;

Martin, M. L.; Parra-Hake, M.; Grotjahn, D. B. Organometallics 2006, 25,

4374-4390.

of a weak and slightly broad absorption centered at 2110 cm^-1

,

which was absent in an infrared spectrum of the ligand (i-Pr)₂P-

(o-Tol). This infrared stretch is more consistent with the presence

3390 Organometallics, Vol. 26, No. 14, 2007

Grotjahn et al.

of a Rh-H unit in 9 than of an agostic interaction in 10.⁵⁹

Finally, in a simple study of reactivity, bubbling CO through a

solution containing 9 did lead to a mixture containing 6d as

the major component, with spectral data identical to those

presented above. However, overall, it was decided that because

of uncertainty over the precise structure of the product from

(i-Pr)₂P(o-Tol) and our inability to free it completely from

impurities, or to get this mixture to react cleanly, the compound

as obtained was unsuitable for kinetic studies described below.

Nonetheless, it is interesting to contrast the activation of the

methyl group on the aromatic ring of (i-Pr)₂P(o-Tol) with lack

of such activation on (i-Pr)₂PIm′, in which the N-methyl group

could in principle be metalated as was N-benzyl-2-diphe-

nylphosphinoimidazole by Ir(I).⁶⁰Interaction of a benzylic C-H

bond ortho to an aromatic -PR₂substituent through cyclom-

etalation or agostic interaction is known for a variety of

metals.^61-64In contrast, the heterocyclic nitrogen in imida-

zolylphosphine 1b offers an alternative way for a low-coordinate

metal to achieve a ligand environment that is stable, but not

too stable.

Table 2. Selected NMR Coupling Constants J (Hz) for

H¹³C¹³CH and Complexes Derived from It, Relevant to

Bonding in These Species^a

²J_H-Rh

¹J_H-C

²J_H-C³J_H-H

¹J_C-C

¹J_C-Rh

H¹³C¹³CH

na

2.5

2.6^c

2.5

na

248.8

229.7

228.5

49.6

28.0

27.5

9.6

1.7

1.7^c

1.8

1.7

na

169.9

114.6

114.5

115.8^c

116.2

115.8

57.0

na

2a-H¹³C¹³CH

2a-H¹³C¹³CH^b

2b-H¹³C¹³CH

2c-H¹³C¹³CH

2c-H¹³C¹³CH^d

16.0

15.7

14.7

nd

15.3

57.0

55.4

55.6

232.9 ^c29.3^c

232.8

233.2

161.7

163.0

29.0

29.1

na

b

5a-¹³C¹³CH₂

5b-¹³C¹³CH₂

5c-¹³C¹³CH₂

na

58.0

58.7

^aData acquired on samples in CD₂Cl₂at 30 °C unless otherwise indicated.

“na” means not applicable. “nd” means not determined. ^bIn C₆D₆. At

c

d

20 °C. At -20 °C.

doubly labeled acetylene, H-¹³C¹³C-H, to make 2b- and 2c-

H¹³C¹³CH. In CD₂Cl₂at -20 °C both H-¹³C¹³C-H complexes

gave rise to ¹³C NMR signals near 68 ppm showing coupling

to both Rh and the two P [for (i-Pr)₂PPh complex δ 68.7 (dt,

¹J_C-Rh) 15.3 Hz, ²J_C-P) 2.0 Hz) and for (i-Pr)₂PIm′ analogue

Alkyne Complexation and Vinylidene Formation. Syn-

thetic Studies. At ambient temperatures, adding 1-hexyne to

solutions of complexes 7c and 8 led to formation of vinylidene

complexes 5c-CCHBu and 5b-CCHBu, isolated as impressively

deep violet solids. These species were also prepared in one pot,

by first adding the appropriate phosphine to [(µ-Cl)Rh(cy-

clooctene)₂]₂, followed by 1-hexyne. Satisfactory combustion

analyses were consistent with identity and purity of the samples.

NMR spectral data diagnostic for the vinylidene ligand include

downfield ¹³C NMR resonances near δ 298 ppm with coupling

1

2

δ 67.5 (dt, J_C-Rh) 13.1 Hz, J_C-P) 2.0 Hz)]. Notably, for

the alkyne protons a 10-line pattern for the AA′XX′ system

could be observed. Analysis of the pattern⁴⁵in the case of the

free alkyne as well as its complexes allowed determination of

several coupling constants (Table 2), most significantly ¹J_C-C

,

which was identical within experimental uncertainty (115.8-

116.2 Hz) in the two complexes 2b- and 2c-H¹³C¹³CH.⁶⁶The

strong back-bonding of the metal to the alkyne π-system is

1

obvious by noting that for the free alkyne, J_C-C) 169.9 Hz,

consistent with literature values of 165.8-173.6 Hz in various

media.^67,68There are extensive ¹³C-¹³C coupling data for

organic molecules,^69-71but fewer examples for metal com-

plexes.⁷²In this paper we are considering very closely related

derivatives of the same alkyne, so it is reasonable to relate the

magnitude of ¹³C-¹³C coupling to C-C distance^73,74and hence

the degree of metal-ligand back-bonding.⁷⁴From data in Table

2 it is clear that the imidazol-2-yl and phenyl ligands are

electronically very similar in this system. This conclusion was

additionally supported (Table 2) by our data for a third system,

2

to both Rh and P [for 5b-CCHBu, δ 298.0 (td, J_C-P) 17.4

Hz, ¹J_C-Rh) 55.2 Hz), for 5c-CCHBu, 297.7 (td, ²J_C-P) 17.4

1

Hz, J_C-Rh) 55.2 Hz)] and upfield H NMR resonances near

0 ppm for the RhdCdCHCH₂CH₂CH₂CH₃proton, which

couples to Rh, the two P, and to neighboring methylene protons

[for 5b-CCHBu, δ 0.37 ppm (dtt, ³J_H-Rh) 1 Hz, ³J_H-P) 3.0

3

Hz, J_H-CH2) 8 Hz, 1 H), for 5c-CCHBu δ 0.00 ppm (dtt,

³J_H-Rh) 1.8 Hz, ³J_H-P) 3.5 Hz, ³J_H-CH2) 8 Hz, 1 H)]. These

data are in complete accord with those observed by Werner’s

group on many Rh(I) vinylidene complexes.^34-38,65

1

(i-Pr)₃P complex 2a-H¹³C¹³CH. In this complex, J_C-Cis

Reaction of (i-Pr)₂PPh was slow enough at ambient temper-

atures to observe the intermediate π-alkyne complex 2c-HCCBu

by NMR, but to do so with the heterocyclic analogue required

lowering the reaction temperature to -20 °C. In both cases the

reactivity of 1-hexyne π-complexes 2b- and 2c-HCCBu pre-

cluded isolation, but for these species in the reaction mixture,

diagnostic resonances for the terminal alkyne proton were found

at 2.92 ppm for the (i-Pr)₂Ph case (2c-HCCBu) and 3.25 ppm

for the (i-Pr)₂PIm′ case (2b-HCCBu) (dt ) q, J ) 2.5 Hz),

along with in each case the appearance of a new doublet in the

³¹P{¹H} NMR spectrum.

definitely less (by ca. 1.5 Hz) than that seen in complexes

(66) The values given in this paper differ slightly from those appearing

in our original communication,⁴²because an empirical fitting of the data

has now been supplanted by a rigorous analysis. See Experimental Section

and Supporting Information for full details.

(67) Wigglesworth, R. D.; Raynes, W. T.; Kirpekar, S.; Oddershede, J.;

Sauer, S. P. A. J. Chem. Phys. 2000, 112, 3735-3746.

(68) Jackowski, K.; Wilczek, M.; Pecul, M.; Sadlej, J. J. Phys. Chem. A

2000, 104, 5955-5958.

(69) Krivdin, L. B.; Kalabin, G. A. Prog. Nucl. Magn. Reson. Spectrosc.

1989, 21, 293-448.

(70) Kamienska-Trela, K. Annu. Rep. NMR Spectrosc. 1995, 30, 131-

230.

The identity of alkyne π-complexes and characterization of

the bonding in these species were further secured by using

1

(71) For experimental and computational studies of J_C-Cin acetylene

derivatives, see refs 71a-f. (a) Sebald, A.; Wrackmeyer, B. Spectrochim.

Acta, Part A 1981, 37A, 365-368. (b) Wrackmeyer, B. Spectroscopy

(Amsterdam, Netherlands) 1982, 1, 201-208. (c) Lambert, J.; Klessinger,

M. Magn. Reson. Chem. 1987, 25, 456-461. (d) Zinchenko, S. V.; Kalabin,

G. A.; Krivdin, L. B.; Proidakov, A. G.; Bazhenov, B. N. Zh. Org. Khim.

1988, 24, 1595-1605. (e) Biedrzycka, Z.; Kamienska-trela, K. Pol. J. Chem.

2003, 77, 1637-1648. (f) Pecul, M.; Ruud, K. Magn. Reson. Chem. 2004,

42 (Special Issue), S128-137.

(59) Brookhart, M.; Green, M. L. H. J. Organomet. Chem. 1983, 250,

395-408.

(60) Tejel, C.; Ciriano, M. A.; Jimenez, S.; Oro, L. A.; Graiff, C.;

Tiripicchio, A. Organometallics 2005, 24, 1105-1111.

(61) Rheingold, A. L.; Fultz, W. C. Organometallics 1984, 3, 1414-

1417.

(62) Herrmann, W. A.; Brossmer, C.; Reisinger, C.-P.; Riermeier, T. H.;

O¨ fele, K.; Beller, M. Chem.-Eur. J. 1997, 3, 1357-1364.

(63) Baratta, W.; Del Zotto, A.; Herdtweck, E.; Vuano, S.; Rigo, P. J.

Organomet. Chem. 2001, 617-618, 511-519.

(64) Sjoevall, S.; Andersson, C.; Wendt, O. F. Organometallics 2001,

20, 4919-4926.

(72) For studies of ¹J_C-Con metal complexes in general, see ref 69, pp

186-196. For dramatic differences between mono- and dinuclear alkyne-

metal complexes, see in particular: Aime, S.; Osella, D.; Giamello, E.;

Granozzi, G. J. Organomet. Chem. 1984, 262, C1-C4.

(73) Benn, R.; Rufinska, A. Organometallics 1985, 4, 209-214.

(74) Grotjahn, D. B.; Collins, L. S. B.; Wolpert, M.; Lo, H. C.;

Bikzhanova, G. A.; Combs, D.; Hubbard, J. L. J. Am. Chem. Soc. 2001,

123, 8260-8270.

(65) Werner, H.; Wolf, J.; Ho¨hn, A. J. Organomet. Chem. 1985, 287,

395-407.

Transformation of Terminal Alkynes to Vinylidene Ligands

Organometallics, Vol. 26, No. 14, 2007 3391

Scheme 3. Results of Crossover Experiments Forming 5^a

We note that a double-crossover experiment was attempted

on hydrido(alkynyl)iridium species like 4a, but unfortunately,

rapid H-D exchange between the hydridic starting materials

prevented a conclusion regarding molecularity.³⁷In our work,

adding an equimolar mixture of terminal alkynes DCC(CH₂)₃-

CH₃and HCCCHMe₂(0.6 equiv of each relative to the amount

of Rh) to either 7c or 8 at -20 °C produced mixtures of alkyne

π-complexes 2, which evolved to vinylidene complexes (5) as

the mixtures were warmed. Significantly, with an estimated

detection limit of 5-10%, no crossoVer was detected, regardless

of ligand type, inconsistent with significant bimolecular 1,3-H

shift. We note that Li and Crabtree performed similar crossover

experiments on a (hydrido)Ir(III) system and observed a lack

of crossover in products derived from alkyne-to-vinylidene

transformation at some stage.^76,77

However, because the propensity for bimolecular reaction

could be a function of hindrance on the alkyne or the phosphine,

and because the previous studies^37,41focused on acetylene itself

and the (i-Pr)₃P ligand, we allowed a 1:1 mixture of 2a-HCCH

and 2a-D¹³C¹³CD to convert to vinylidene complexes 3 at 25-

30 °C in C₆D₆. Remarkably, formation of 5a-CCH₂and 5a-

¹³C¹³CD₂is consistent with an absence of crossover even in

this system. Notably, our estimated limit of detection of D in

5a-CCH₂is 1%. The absence of isotopomer 5a-CCHD from

2a-HCCH in the presence of the D¹³C¹³CD isotopomer is the

most compelling eVidence against crossoVer.

Surprisingly, the rearrangement of pure 2a-D¹³C¹³CD in

silylated and dried NMR tubes gave 5a-¹³C¹³CD₂along with

10-20% of the ¹³C¹³CHD isotopomer 5a-¹³C¹³CHD, which

somewhat complicated the crossover analysis. In these experi-

ments, the source of the H could not be determined, but it would

not appear to be extraneous water, because the amount of H

incorporation was unaffected by pretreating the NMR tube with

D₂O before drying or by the presence or absence of 2a-HCCH.

Moreover, the vinylidene protons of 5-CCH₂did not exchange

with D₂O at room temperature in C₆D₆. Finally, we ruled out

the role of a kinetic isotope effect by observing that separate

solutions of 2a-D¹³C¹³CD and 2a-HCCH rearrange to the

respective vinylidenes at 30 °C in the NMR probe. Comparing

^aUpper bounds given are estimated detection limits.

derived from the other two ligands, following the trends in the

infrared stretching frequency in the corresponding CO com-

plexes (Table 1). In fact, although the focus of attention here

1

has been on J_C-C, by looking at the data in Table 2, one can

see from the other coupling constants of sufficient magnitude

(¹J_H-C, ²J_H-C, and even ¹J_C-Rh) that the values for the imidazol-

2-yl and phenyl derivatives are identical within experimental

uncertainty or nearly so, and distinct from those for 2a-

H¹³C¹³CH.

The doubly labeled acetylene complexes evolved to the

corresponding vinylidenes 5a-, 5b-, and 5c-¹³C¹³CH₂, which

were characterized in solution by several key NMR resonances.

The ¹³C NMR spectrum of the Ph analogue 5c-¹³C¹³CH₂

showed two strong mutually coupled (¹J_C-C) 58.7 Hz) peaks

at δ 297.9 (²J_C-P) 17.1 Hz, ¹J_C-Rh) 55.6 Hz) and 90.7 ppm

(³J_C-P) 6.5 Hz, ²J_C-Rh) 15.7 Hz) for C1 and C2 of the Rhd

¹³Cd¹³CH₂unit. Similar data were found for the Im′ analogue.

1

Most significantly, the value of J_C-Cin both systems (58.7

and 58.0 Hz) was virtually identical, again suggesting virtually

identical C-C bond distances and hence metal-vinylidene

back-bonding (Table 2),⁷⁵especially considering that the same

parameter had a value of 57.0 Hz for the (i-Pr)₃P analogue.

In our work, all attempts to detect hydrido(alkynyl) isomers

4 or other intermediates in conversions of alkyne complexes to

vinylidenes were unsuccessful. The conversions of 2b- and 2c-

H¹³C¹³CH to the corresponding vinylidene isomers were not

completely clean, particularly if less than 1 equiv of labeled

alkyne was used and some unreacted 8 or 7c was present. More

conclusive results regarding the absence of detectable amounts

of 4 came from careful examination of reactions of 8 and 7c

with 1-hexyne, which showed only alkyne-free precursor 8 or

7c, corresponding alkyne π-complex 2b-HCCBu or 2c-HC-

CBu, and final vinylidene 5b-CCHBu or 5c-CCHBu. It could

be estimated that the lower detection limit in these latter

experiments was from 1% to 5%. Computational and kinetics

results (Vide infra) are consistent with our inability to detect

hydride complexes 4 or other reaction intermediates at these

levels.

rates of disappearance of starting materials showed that k_HCCH

/

k_D13C13CD) 1.67, a relatively small effect compared with the

low or undetectable D incorporation in crossover experiments

above. Katayama and co-workers found k_PhCCH/k_PhCCD) 1.69-

(5) for similar transformation on a Ru(II) center.¹⁵

Kinetics. A series of kinetics experiments was undertaken

with the goals of determining microscopic rate constants and

evaluating the effects of the heterocyclic ligand on each step of

alkyne-to-vinylidene transformation. To a cooled solution of

precursor 8 or 7c in CD₂Cl₂was added 1-hexyne, and the

resulting cold NMR tube was inserted into an NMR probe

precooled to -20 °C. The evolution of the mixture from alkyne-

free precursor and 1-hexyne to alkyne complex 2-HCCBu to

vinylidene complex 5-CCHBu was followed by ¹H NMR. The

rate constants for three elementary steps were then fit to the

observed concentration versus time profiles: step 1 ) alkyne

binding to precursor, forming 2; step 2 ) transformation of 2

to the hydrido(alkynyl) species 4; step 3 ) transformation of 4

to 5. For a given mechanism and set of rate constants, the

concentrations were predicted as a function of time using a

fourth-order Runge-Kutta numerical integration, and the rate

Crossover Experiments. In order to determine the molecu-

larity of the conversion of hydrido(alkynyl) intermediate 4 to

vinylidene product 5, we turned to a series of double-crossover

experiments on isotopically labeled species (Scheme 3).

(76) Li, X.; Incarvito, C. D.; Crabtree, R. H. J. Am. Chem. Soc. 2003,

125, 3698-3699.

(77) Li, X.; Vogel, T.; Incarvito, C. D.; Crabtree, R. H. Organometallics

2005, 24, 62-76.

(75) The identity of the trans-ligand in (halo)(¹³C¹³CH₂)Rh[(i-Pr)₃P]₂

affects ¹J_C-C; see ref 48. Apparently, in ref 48 the reported values of ¹J_C-C

are off by a factor of 2.

3392 Organometallics, Vol. 26, No. 14, 2007

Grotjahn et al.

Figure 4. B3LYP/LANLDZ lowest energy conformers of 5g and 5h. In each case the upper view is from a point perpendicular to the P-P

axis, whereas the lower one is from a point on the P-P axis, showing the anti orientation of aromatic groups.

constants were then iteratively optimized in a least-squares fit

to the data. (These data do not differentiate between the

unimolecular and bimolecular mechanisms for step 3, because

if the unobserved intermediate 4 is in steady state, then the

product formation rate is approximately proportional to the

concentration of 2 for either mechanism.) The numerical

integration predicts that the concentration of 4 never exceeds

0.002 mM in any of these experiments, consistent with its failure

to be detected by NMR. As a result of this, however, the rate

constants for steps 2 and 3 cannot be independently determined

from the available data, and these were therefore combined into

a single effective rate constant, k_2-5. For 8, the rate constant

for formation of 2 is k ) 9.52(23) × 10^-5mM^-1min^-1and

to-vinylidene conversion rate k_2-5was seen. Finally, the inability

to detect hydride intermediates and follow their concentrations

with time precluded determination of the rate law for the

vinylidene-forming step or the microscopic rate constants for

individual steps, so effects of the phosphine substituents had to

be evaluated further using computational means.

Kang et al. computed the relative energies of the PH₃species

2f, 4f, and 5f, optimizing geometries at the MP2 level and

calculating energies at the MP4 level.⁸³Although they calculated

the barrier for alkyne rotation in 2f, they did not probe the

kinetics of the reaction from 2 to 5. The following year,

Wakatsuki et al. ’s detailed computational investigation of the

reaction of 2f and 2a⁴¹employed a combination of the MP2 ab

initio and MM3 molecular mechanics techniques, making the

treatment of the i-Pr groups in 2a, 4a, and 5a tractable. That

work found barriers of over 30 kcal/mol to both steps (Scheme

1, A f B and B f C) of the unimolecular reaction, inconsistent

with the observed rate of reaction and inspiring the proposal of

a bimolecular 1,3-H shift mechanism shown at the bottom of

Scheme 1. With the advantage of experiments inconsistent with

a bimolecular mechanism, we therefore carried out a new

computational investigation in an effort to describe the reaction

mechanism in greater detail and shed light on the discrepancy

between the experiments and previous calculations.

the effectiVe rate constant for formation of 5 from 2 is k_2-5

)

3.32(22) × 10^-2min^-1. For 7c, k ) 2.01(65) × 10^-3mM^-1

min^-1and k_2-5) 3.46(19) × 10^-3min^-1. Thus, the rate

constant k_2-5for the overall H-shift is 9.6 times greater with

the heterocyclic system than with the nonheterocyclic one.

Conclusions from Experiments and Rationale for Calcula-

tions. Profound changes in complex structure as a function of

ligand in this study are seen before addition of alkyne: the

hybrid P,N-ligand 1b leads to the fluxional chelate complex 8

with intact N-methyl substituents, whereas in 1d lack of a

potentially coordinating nitrogen leads to methyl metalation (9).

In contrast, all available X-ray diffraction and IR and NMR

spectral data on trans-chlorobis(phosphine)Rh(I)-L species (L

) CO, H¹³C¹³CH, or vinylidene ¹³C¹³CH₂) highlight the

similarity of ligands with either imidazol-2-yl (1b) or phenyl-

derived (1c, 1d) substituents, distinct from the (i-Pr)₃P case 1a.

As for alkyne-to-vinylidene transformation, the crossover

experiments on the isomerization of alkyne π-complexes on

trans-(chloro)bis(phosphine)Rh are inconsistent with a bimo-

lecular pathway leading to vinylidene complexes. Hydride

intermediates were not detectable. For a given alkyne, the

rearrangement appears to be fastest for (i-Pr)₂PR¹where R¹)

aryl. The imidazolyl unit leads to a somewhat faster reaction,

perhaps because of the hemilability^78-82of the chelate in 8, a

process for which we have determined activation parameters.

From the kinetics experiments, a 9.6-fold increase in alkyne-

Computational Results and Discussion

The computational study consisted of (1) conformational

analysis at the B3LYP/LANL2DZ level of the R¹group

orientations, (2) mapping of the B3LYP/LANL2DZ reaction

surface that connects structures 2f-5f in order to identify the

relevant intermediates and transition states, (3) B3LYP/

LANL2DZ geometry optimizations and frequency calculations

for structures 2g-5i, based on the previous conformational and

reaction surface analyses, (4) geometry optimizations of struc-

tures 2a-5a, 2c-5c, 2d-5d, and 2e-5e by BLYP/DQZ/UFF

ONIOM calculations (henceforth “BDMM”). Details surround-

ing these steps and the choice of methods are given in the

Experimental Section. The most stable conformations identified

in all cases are characteristic of a stabilizing interaction between

the R¹group and Cl atom, as exemplified in Figure 4.

Calculations on conformers of 5g as well as on prior intermedi-

(78) Braunstein, P. J. Organomet. Chem. 2004, 689, 3953-3967.

(79) Werner, H. Dalton Trans. 2003, 3829-3837.

(80) Braunstein, P.; Naud, F. Angew. Chem., Int. Ed. 2001, 40, 680-

699.

(81) Bo¨rner, A. Eur. J. Inorg. Chem. 2001, 327-337.

(82) Schneider, J. J. Nachr. Chem. 2000, 48, 614-616, 618-620.

(83) Kang, S. K.; Song, J. S.; Moon, J. H.; Yun, S. S. Bull. Kor. Chem.

Soc. 1996, 17, 1149-1153.

Transformation of Terminal Alkynes to Vinylidene Ligands

Organometallics, Vol. 26, No. 14, 2007 3393

to be most stable in a pseudo-C₂conformation, as shown for

compound 4e-HCCH in Figure 5.

Potential Energy Surface. The relaxed potential energy

surface (PES) of the unsubstituted [RhCl(PH₃)₂](C₂H₂) system

was determined at the B3LYP/LANL2DZ level in order to

investigate the possibility that unanticipated intermediates might

contribute to the observed reaction kinetics. To our knowledge,

the previous computational studies^41,43,83did not make a scan

of the entire PES. The surface was parametrized as a function

of the C2-C1-H and Rh-C1-C2 bond angles, which unam-

biguously distinguish between 2f, 4f, and 5f, as shown in Figure

6. However, the values of these two bond angles are similar

for the hydrido(alkynyl) 4f and for the η²-(C,H) alkyne complex

3f, and this two-dimensional PES is supplemented by an

additional scan of the C1-Rh-H bond angle near the 4f

geometry in Figure 6. One species not previously identified to

our knowledge is the η²-(C,C) vinylidene complex seen at the

left of Figure 6. Because the geometries of this central

framework are not greatly changed upon substitution by the (i-

Pr)₂PR¹ligands, the characteristic geometries of the stationary

Figure 5. BDMM-optimized geometry of 4e-HCCH, viewed

roughly along the RhCC axis, showing the approximate C₂

symmetry.

ates in which the imidazole nitrogen(s) could be acting as

internal base revealed that none of these conformers were more

stable than the ones just described and shown in Figure 4. In

the subsequent BDMM calculations, the i-Pr groups were found

Figure 6. B3LYP/LANL2DZ relaxed PES of the [RhCl(PH₃)₂](HCCH) reaction system, with energies relative to that of 5f: (a) the smoothed

surface as a function of Rh-C1-C2 and C2-C2-H angles; (b) graph of the potential energy as a function of the C1-Rh-H angle near

the geometries of 3f and 4f, resolving the barrier between them; (c) contour plot of the surface in (a) with contours every 10 kcal mol^-1

.

3394 Organometallics, Vol. 26, No. 14, 2007

Grotjahn et al.

Figure 7. B3LYP/LANL2DZ-optimized critical point geometries for species on the surface shown in Figure 6.

point geometries are shown for the simpler PH₃ligands in this

series of calculations in Figure 7.

barrier of only 0.6 kcal/mol in the case of B3LYP/LANL2DZ

calculations for [RhCl(PH₃)₂](C₂H₂). Under these conditions,

3 and 4 are subject to a fast equilibrium. Although 4 is not

directly coupled to 5 by the minimum energy path, the fast

equilibrium approximation allows the integrated rate law for

production of vinylidene to be written roughly as [vinylidene]

The rather surprising stability of the η²-(C,C) vinylidene

complex suggested the possibility of a kinetic contribution from

a triplet-spin surface, a situation that to our knowledge has not

been considered previously. To test this, the geometries of the

alkyne π-complex 2f, hydrido(alkynyl) species 4f, the “normal”

RhdCdCH₂complex 5f, and its η²-(C,C) vinylidene isomer

were all optimized in their triplet-spin states at the B3LYP/

LANL2DZ level, using the unsubstituted model species. The

triplet alkyne complex and vinylidene were both found to lie at

energies only 33-35 kcal/mol above the singlet vinylidene and

10 kcal mol^-1more stable than TS45, whereupon a search was

made for low-lying triplet geometries with the C2-C1-H bond

angle fixed to 69°, its value in the singlet transition states. No

points were found that suggested a lower energy path existed

between reactant and product along the triplet surface, so these

possible structures were not considered further.

) [hydride]₀{(1 - e^-at) - [a(e^{-k t}- e^-at)/(a - k_2-4)]}, where

2-4

a ) k_4-5/(1 + K_3-4). In the limit that the fast equilibrium

constant K_3-4. 1, this equation predicts kinetics consistent

k_2-4

k_4-5/K_3-4

with the reaction sequence π-complex

8 hydride

8

vinylidene. The 1,3-shift required for one-step conversion of 4

to 5 therefore remains, effectively, the mechanism for the

reaction. The relaxed PES establishes that a 1,2-shift does not

take place (along a minimum energy path) directly from 2,

because decreasing the C2-C1-H angle, to place the H atom

above the π-bond, causes the energy barrier along the Rh-

C1-C2 angle to disappear. Once one H atom moves signifi-

cantly from one end of the acetylene toward the other, the carbon

atom to which it was bound falls toward the Rh atom.

The PES does reveal that the η²-(C,C) vinylidene species (at

Rh-C1-C2 ) 79.6°, C2-C1-H ) 25.6°) is also a stable

geometry in the model species, with a barrier of only about 4

kcal/mol to formation of the product vinylidene. However, no

realistic mechanism to reach the η²-(C,C) vinylidene was

identified, and therefore this species was not studied further.

Kinetics Analysis. The η²-(C,H) complex 3 was not identified

in the previous study by Wakatsuki et al.,⁴¹who instead obtained

the hydrido(alkynyl) structure 4 with a C-Rh-H bond angle

of 69.5° and Rh-H bond length of 1.50 Å. In marked contrast,

the recent DFT study by Suresh et al. found the η²-(C,H) alkyne

complex 3 to be an intermediate in the hydrogen shifts to and

from the metal center.⁴³The present calculations similarly find

both 4 (C-Rh-H angle of 86-91°, Rh-H bond length 1.52

Å) and the η²-(C,H) alkyne complex 3 (C-Rh-H ) 38-39°,

Rh-H ) 1.67-1.71 Å) to be distinct local minima on the PES

of each compound studied [H₂PR¹and (i-Pr)₂PR¹for R¹) i-Pr,

Ph, Im, o-Tol]. Complex 3 is positioned geometrically between

the three other species (2, 4, 5) in the reaction system such that

it appears along each minimum energy pathway connecting the

other three species on the PES. However, its impact on the

kinetics appears to be negligible, because the hydrido(alkynyl)

species 4 is lower in energy and separated from 3 by an energy

Given that the experimental observations on 2b show

reactions near completion after 1 h at -20 °C, a maximum value

of 3600 s may safely be assumed for the half-life of the hydrido-

(alkynyl) complex at this temperature, corresponding to a

minimum rate constant k_4-5of 2 × 10^-4s^-1. Assuming a typical

Arrhenius pre-exponential factor of 10¹²s^-1, which can be

estimated from published studies of C-H activation via alkane

Transformation of Terminal Alkynes to Vinylidene Ligands

Organometallics, Vol. 26, No. 14, 2007 3395

Figure 8. Reaction diagram for [RhCl(Pi-Pr₂Im)₂]C₂H₂determined at the BDMM level.

Table 3. BDMM-Computed Relative Energies (kcal mol^-1

along the Reaction Connecting 2 and 5 in Scheme 2

)

k

_2-5is measured, the computations predict the k_4-5value to be

limiting for the formation of vinylidene. The predicted k_4-5

values each lie within a factor of 2 of the observed rate constants,

and the 9.6 factor increase for Im over Ph is predicted to within

10%. The excellent agreement between the observed rate

constants and the predicted rate constants for the rate-limiting

4 f 5 step is partly fortuitous, as discussed below. Nonetheless,

the successful prediction of the factor of 10 increase in reaction

rate from Ph to Im is hoped to be an indication of the value of

calculations on systems of this type.

R¹

∆E_2-4

E_a,2-3

∆E_3-4

∆E_4-5

E_a4-5

∆E_2-5

i-Pr (a)

Im (e)

Ph (c)

3.4

7.8

5.9

7.8

16.1

18.5

18.0

19.0

3.8

2.2

3.7

1.9

-20.9

-22.2

-21.1

-22.7

21.9

18.3

20.3

18.5

-17.5

-14.4

-15.1

-14.9

o-Tol (d)

^aReference 41.

Table 4. Computed Free Energies of Activation and Rate

Constants at -20 °C^a

Empirically, the computations indicate that the observed

increase in reaction rate with the monoaryl-substituted ligands

may be ascribed to an accompanying stabilization of the reaction

surface relative to the hydrido(alkynyl) species 4. The values

of ∆E_2-4, ∆E_3-4, E_a4-5, and ∆E_4-5in Table 3 show that

replacement of one i-Pr group by Ph in each ligand raises the

energy of 4c relative to species 2c and TS45c by about 2 kcal

mol^-1, although this effect is absent when considering 3c and

5c. More striking is the effect of substitution by Im or o-Tol,

which raises the energy of 4e and 4d relative to all the other

species (2, 3, 5, and the two transition states) by 1.3 to 4.4 kcal

mol^-1. The transition state for the 2 f 4 step lies closer in

geometry to 4 than to 2, and its relative energy follows the same

trends as 4. As a consequence, the barrier to the 2 f 4 reaction

step increases as the ligands progress in the series i-Pr, Ph, Im,

o-Tol, while the barrier to the 4 f 5 reaction step decreases.

Because the 4 f 5 step is the rate-limiting step in all but the

o-Tol case, the reaction rate increases along this series.

∆G_a,24

k_2-4

∆G_a,45

k_4-5

k_2-5,obs

(min^-1

)

R¹

(kcal mol^-1

)

(min^-1

)

(kcal mol^-1

)

(min^-1

)

i-Pr (a)

Im (e)

Ph (c)

13.9

15.5

15.8

16.5

4.8

19.6

16.7

17.9

15.1

5.7 × 10^-5nd

0.21

0.11

0.027

0.019

0.0019

0.45

0.033

0.0035

na

o-Tol (d)

^aEnergies from BDDM calculations, with thermal and zero-point

corrections estimated from B3LYP/LANL2DZ calculations on the

[RhCl(H₂PR¹)₂](C₂H₂)-derived species.

complexes,^84,85the maximum activation energy should be 21

kcal/mol. The predicted kinetics are highly sensitive to both

computational method and basis set among those investigated

in this work. However, qualitatively accurate results are obtained

using the BDMM-predicted relative energies, presented in Table

3. The reaction diagram based on these values is illustrated for

[RhCl(Pi-Pr₂Im)₂]C₂H₂in Figure 8.

These computational results are consistent with the inability

of the NMR studies to detect the hydrido(alkynyl) intermediate

4. The net reaction is irreversible, because the barrier from

vinylidene 5 to 4 is in all cases predicted to be greater than 39

kcal mol^-1. In contrast, the first step of the reaction has a reverse

barrier about 7 kcal mol^-1lower than the TS45 energy, allowing

4 and π-complex 2 to remain in partial equilibrium during the

reaction, with an equilibrium constant less than 10^-4. At any

time during the reaction, the concentration of the hydrido-

(alkynyl) complex is therefore expected to be insubstantial

compared to that of the π-complex.

The physical basis for the relative stability of TS45 with the

monoaryl ligands is more elusive. The hydrido(alkynyl) complex

is distinct from the other stationary points on the reaction surface

in representing the Rh(III) oxidation state with square-pyramidal

coordination rather than Rh(I) with square-planar coordination.

The shift observed in the CO stretching frequency from R¹)

i-Pr to o-Tol (Table 1) is consistent with the (i-Pr)₃P ligand

contributing a slight destabilization to the CO bond, and the

NMR coupling constant data available in Table 2 for 2a-2c

and 5a-5c suggest that this effect is also felt in the C₂H₂and

CCH₂analogues. This interpretation of the spectroscopic

evidence is supported by small differences observable in the

Rh-C and C-C bond lengths in the optimized BDMM

geometries, as listed in Table 5. The Rh-C distance in the

π-complex 2a is 0.008 to 0.012 Å shorter than that in 2b, 2c,

and 2d, and the Rh-C1 distance in TS45a is 0.009 to 0.011 Å

shorter than with the other ligands studied. In contrast, the C-C

bond lengths are longer by 0.001 to 0.005 Å in 2a, TS45a, and

Table 4 gives the rate constants for steps 2 and 3 of the

reaction (2 f 4 and 4 f 5), estimated from elementary

transition state theory for the unimolecular case, k ) (k_BT/h)

exp[-∆G^q/(RT)]. For the two ligands for which the rate constant

(84) Lian, T.; Bromberg, S. E.; Yang, H.; Proulx, G.; Bergman, R. G.;

Harris, C. B. J. Am. Chem. Soc. 1996, 118, 3769-3770.

(85) Northcutt, T. O.; Wick, D. D.; Vetter, A. J.; Jones, W. D. J. Am.

Chem. Soc. 2001, 123, 7257-7270.

3396 Organometallics, Vol. 26, No. 14, 2007

Grotjahn et al.

Table 5. Comparison of Selected BDMM-Optimized Bond

Lengths (Å) or Angles (deg) and Relevant Experimental

NMR Data (J values, Hz)

reaction system using triphenylphosphine ligands, derivatives

of which have been used recently in catalytic transformations

for organic synthesis.¹³In order to adequately represent the

electronic structure of the phenyl groups, which was neglected

for the i-Pr groups in our BDMM calculations, optimizations

of 2-5 and partial optimizations near the anticipated TS23 and

TS45 geometries were carried out at the BLYP level using the

DQZ,6-31+G(d,p) basis, without molecular mechanics. The

limiting step is again predicted to be the hydride-to-vinylidene

H-shift. The estimated activation barrier for the Ph₃P system

4j converting to 5j lies roughly 4 kcal mol^-1lower than that

for the (i-Pr)₃P analogue 4a converting to 5a (calculated at the

same level of theory). This suggests that, like the heterocyclic

ligand in the present work, triphenylphosphine ligands such as

those used recently for catalytic reactions¹³will contribute a

substantial enhancement of the reaction rate over that using (i-

Pr)₃P.

a

b/e

c

d

j^a

2

Rh-C

C-C

C-C-H

¹J_C-Rh

¹J_C-C

2.136

1.272

151.8

16.0

114.6

2.148

1.270

152.7

14.7

115.8

2.146

1.271

152.8

15.3

116.2

2.144

1.271

152.7

nd

2.134

1.270

153.9

nd

3

4

Rh-C

C-C

2.014

2.023

2.019

2.020

1.244

1.252

1.699

1.990

1.241

1.332

1.664

1.244

1.276

1.673

1.243

1.246

1.709

1.244

1.251

1.704

C-H

Rh-H

Rh-C

C-C

1.977

1.239

1.980

1.238

1.985

1.239

1.982

1.239

1.969

1.236

TS45

1.940

1.273

Rh-C

C-C

1.929

1.278

1.939

1.277

1.938

1.277

1.933

1.272

Conclusions

5

Whereas alkyne transformation on CpRu-bis(phosphine)

systems is very strongly influenced and accelerated by hetero-

cyclic groups on the phosphines,^31-33in the present case the

most profound effect of the basic imidazole nitrogen in ligand

1b appears to be in forming alkyne-free Rh(I) precursor 8 with

a labile P,N chelate. In contrast, (i-Pr)₂PPh (1c) appears to

form a normal dimer with [Rh₂Cl₂] core, whereas its o-Tol

analogue (i-Pr)₂P(o-Tol) (1d) suffers metalation of the methyl

group on the aromatic ring, unlike the methyl group of 1b. The

effects of phosphine substituent R in (i-Pr)₂PR on bonding in

complexes of CO, HCCH, and its isomer CCH₂could be

evaluated using a combination of conventional (e.g., Rh-CO

infrared absorption data for 6) and nontraditional means

(particularly ¹³C-¹³C coupling data in 2a-c and 5a-c derived

from H¹³C¹³CH), leading to a consistent picture: the electronic

effects of (i-Pr)₂R (R ) Ph, Im′) were similar and different from

those of (i-Pr)₃P.

Qualitatively, ligand 1b led to much faster formation of

vinylidene complex 5 than did 1a, whereas model ligand 1c

was intermediate in its effects. Alkyne complexes 2 could be

formed from 1b and 1c below room temperature, but kinetics

experiments could not shed light on the molecularity of the

conversion of 2 to 5, nor on the microscopic rate constant for

conversion of hydrido(alkynyl) complex 4 to final product.

However, lack of isotopic scrambling in double-crossover

experiments is significant experimental evidence against alkyne

activation on Rh(I) and transformation to vinylidene by way of

binuclear species.

These results motivated a computational search of the

potential energy surface for alkyne-to-vinylidene transformation

on trans-(Cl)(R₂PR¹)Rh, which revealed an η²-(C,H) complex,

3, not far above hydrido(alkynyl) isomer 4. The magnitudes

and trends in the observed rate constants are predicted well by

DFT computations using the RCEP core potential and basis of

Stevens et al. for the rhodium atom and incorporating the i-Pr

groups, using a combination of molecular mechanics and DFT.

These methods in conjunction with careful surveying of the

reaction surface appear to offer tractable and accurate means

for modeling reaction dynamics in these and similar studies of

C-H activation, catalysis, and transformations useful for organic

synthesis.

Rh-C

C-C

1.833

1.330

57.0

1.837

1.327

55.4

58.0

1.837

1.328

55.6

58.7

1.836

1.328

nd

1.842

1.326

nd

¹J_C-Rh

¹J_C-C

^aResults of full basis-set calculations.

nd

5a than when the monoaryl ligands are used. Both effects are

consistent with the (i-Pr)₃P ligand encouraging back-bonding

from the metal into a π-antibonding orbital of the C-C bond,

strengthening the Rh-C bond while weakening the C-C bond.

The weakened C-C bond raises the energy of the complex at

these geometries with (i-Pr)₃P. However, this difference between

the ligands is reduced at the hydrido(alkynyl) geometry 4, where

the C-C bond lengths lie within 0.001 Å of each other for all

four ligands, and 4a therefore does not experience the same

destabilization as TS45, relative to the other ligands. Conse-

quently, the energy difference between TS45 and 4, the barrier

for the rate-determining step, is lower for the monoaryl ligands

as a group, and this accounts for the observed rate difference.

The importance of back-bonding in determining the comparative

energies is consistent with the conclusions of Suresh et al. in

their recent analysis of the alkyne-dependence of this reaction.⁴³

Contributions of steric effects may also be considered. Looking

at calculated geometrical parameters in Table 5, among monoar-

yl ligands, the Im and o-Tol ligands seem indistinguishable from

the Ph analogue. In contrast, looking at computed energies and

considering the faster observed rate of reaction using 1b

compared with 1c, it appears that the overall conversion of 2

to 5 is more favorable using the bulkier ligands 1e and 1d and

that both ∆E_4-5and E_a4-5are reduced when the bulkier ligands

are employed. Here it may be useful to consider the relief of

interligand steric interactions on going from five-coordinate

intermediate 4 to four-coordinate product 5.

Because the BDMM calculations treat the two i-Pr groups in

(i-Pr)₂PR¹by molecular mechanics, the effect of these substit-

uents on back-bonding will be underestimated. Partial optimiza-

tions (with constrained i-Pr dihedral angles) of 4a and TS45a

using a full 6-31+G(d,p) basis in place of the crude MM model

predict the activation barrier to be nearly 6 kcal mol^-1higher

than the value in Table 4. These calculations remain too

resource-intensive for routine use, however, and the present

BDMM calculations appear to provide a useful balance of

accuracy and efficiency.

Experimental Section

In view of the qualitative success of the present computational

work, exploratory calculations were carried out on the same

General Procedures. All manipulations were carried out in a

nitrogen-filled glovebox or using Schlenk techniques. The starting

Transformation of Terminal Alkynes to Vinylidene Ligands

Organometallics, Vol. 26, No. 14, 2007 3397

Table 6. Collection Data for Crystal Structures of the Complexes

6b

6c

6d

8

formula

MW

cryst system

space group

color, habit

a [Å]

C₂₉H₅₄ClN₄OP₂Rh

675.06

monoclinic

P2₁/n

yellow block

7.9000(4)

24.4193(12)

8.9974(4)

96.793(1)

1723.53(14)

2

C₂₅H₃₈ClOP₂Rh

554.85

monoclinic

P2₁/n

yellow blade

8.5410(8)

7.7860(7)

19.8080(17)

96.300(2)

1309.3(2)

2

C₂₇H₄₂ClOP₂Rh

582.91

monoclinic

P2₁/c

yellow blade

10.232(11)

15.409(17)

9.569(10)

113.509(15)

1383(3)

2

C₄₁H₄₆ClN₄OP₂Rh

811.12

monoclinic

P2₁/n

colorless needle

9.6038(6)

10.5140(6)

19.1961(12)

103.768(1)

1882.6(2)

2

b [Å]

c [Å]

â [deg]

V [Å³]

Z

D

_calcd[g cm^-3

]

1.301

1.407

1.399

1.431

T [K]

100

208

100

2θ_max[deg]

55.0

10 730

3928

195

3.62

8.72

1.137

50.0

7073

2314

146

4.96

5.31

1.210

52.0

8269

2714

151

4.75

11.19

1.089

55.0

11 589

4284

241

3.10

8.56

1.061

no. of measd reflns

no. of indep reflns

no. of params

R(F) (I > 2σ(I)), %^a

R(wF²) (I > 2σ(I)), %^b

GOF

res electron density

0.488

0.928

0.712

0.536

a

b

2

R ) ∑||F_o| - |F_c||/∑|F_o|. R(wF²) ) {∑[w(F_o- F_c)²]/∑[w(F_o)²]}^1/2; w ) 1/[σ²(F_o) + (aP)²+ bP], P ) [2F_c+ max(F_o,0)]/3.

materials [(µ-Cl)Rh(cyclooctene)₂]₂⁸⁶and [RhCl((i-Pr)₂PPh)₂]₂(7c)⁵⁵

were prepared as described in the literature. Phosphine ligand 1b

was made as recently described.⁸⁷Acetylene gas (HCCH) from a

cylinder containing acetone was passed through concentrated

sulfuric acid, then KOH pellets, then powdered K₂CO₃. Labeled

derivatives H-¹³C¹³C-H (99% ¹³C, >98% chemical purity) and

D-¹³C¹³C-D (99% ¹³C and 98% D, >98% chemical purity) were

used as received from Cambridge Isotope Labs (CIL). Care was

taken not to pressurize acetylene because of its potentially explosive

nature. Benzene, toluene, dichloromethane, hexane, and pentane

were distilled from calcium hydride under nitrogen, whereas THF

and diethyl ether were distilled from sodium and benzophenone.

Acetone, ethyl alcohol, and H₂O were deoxygenated prior to use

by nitrogen for 30 min. C₆D₆and CDCl₃were purchased from CIL,

dried over calcium hydride, and vacuum transferred prior to use.

Nitrogen was bubbled through samples of CD₂Cl₂, THF-d₈, and

toluene-d₈from CIL before use. NMR tubes were silanized by

boiling hexamethyldisilazane in the tube for about a minute,

followed by rinsing the cooled tube with pentane and heating the

tube under vacuum before filling it with nitrogen. Unless noted

otherwise, ¹H NMR spectra were obtained on either a Varian

Gemini 200 MHz spectrometer or Varian Inova 500 MHz spec-

trometer. ¹³C NMR spectra were obtained at 125.7 MHz on a Varian

Inova 500 MHz instrument. Both ¹H NMR and ¹³C NMR chemical

shifts were reported in parts per million downfield from tetram-

ethylsilane and referenced to the solvent resonances (¹H NMR: 7.16

ppm for C₆HD₅, 7.27 ppm for CHCl₃, 5.28 ppm for CHDCl₂; ¹³C

NMR: 128.39 ppm for C₆D₆, 77.23 ppm for CDCl₃, and 54.00

ppm for CD₂Cl₂), where ¹H NMR chemical shifts are followed by

multiplicity, coupling constants J in hertz, and integration in

parentheses. The values for coupling constants in the AA′XX′

systems of H¹³C¹³CH complexes were determined using guidance

from the literature^88,89and simulation of spectra using the program

WINDNMR. ³¹P NMR spectra were obtained at 80.95 MHz on

the Gemini or at 202.374 MHz on the Inova, and chemical shifts

were referenced to external 85% H₃PO₄(aq). Deuterium NMR

experiments were run at JEOL USA by Dr. Ashok Krishnaswami

using an ECA-500 spectrometer. Elemental analyses were per-

formed at NuMega Resonance Labs, San Diego, CA.

Data relating to the X-ray structural determinations are collected

in Table 6. All data were collected on Bruker diffractometers

equipped with APEX CCD detectors. All structures were solved

by direct methods and refined with anisotropic thermal parameters

and idealized hydrogen atoms. All software is contained in the

SMART, SAINT, and SHEXTL libraries distributed by Bruker-

AXS (Madison, WI). Complete disclosures about the crystal-

lographic work may be found in the deposited CIF files.

Preparation of 6b. In the glovebox, a mixture was made in a

test tube using [Rh(µ-Cl)(cyclooctene)₂]₂(116.0 mg, 0.16 mmol)

and dry C₆H₆(3 mL). Carbon monoxide was bubbled through the

mixture for 1 min, until the color of the solution lightened to yellow.

A solution of 1b (163.2 mg, 0.64 mmol) in dry C₆H₆(3 mL) was

added. After 10 min the yellow solution was concentrated in Vacuo

to 0.5 mL and pentane (3 mL) was added. The complex precipitated

directly. Removing the supernatant with a pipet and washing with

pentane (2 mL) followed by drying in Vacuo resulted in a pure

pale yellow powder (193.8 mg, 90%). ¹H NMR (CDCl₃, 500

MHz): δ 6.71 (s, 2H, imidazole H-4), 4.12 (s, 6H, N-CH₃), 3.00-

3.10 (m, 4H, Me₂CH-), 1.28 (dvt, ³J_H-H) 7.0 Hz, N ) 14.5 Hz,

12H, (CH₃)₂CH-), 1.26 (s, t-Bu, 18H), 1.17 (dvt, ³J_H-H) 7.0 Hz,

N ) 17.5 Hz, 12H, (CH₃)₂CH-). ³¹P NMR (CDCl₃, 80.95 MHz):

δ 25.8 (d, ¹J_P-Rh) 118.5 Hz). ¹³C NMR (CDCl₃, 125.7 MHz): δ

186.39 (d, ¹J_C-Rh) 73.6 Hz, ²J_C-P) 15.5 Hz), 153.3 (vt, N ) 8.3

Hz), 137.7 (vt, N ) 62.1 Hz), 118.5 (s), 36.3 (s), 32.0 (s), 30.3 (s),

24.9 (vt, N ) 28.3 Hz), 19.5 (vt, N ) 5.3 Hz), 18.7 (s). IR (CDCl₃,

NaCl): ν_CO1968.2 cm^-1. Anal. Calcd for C₂₉H₅₄ClN₄OP₂Rh

(675.07): C, 51.60; H, 8.06; N, 8.30. Found: C, 51.20; H, 7.83;

N, 8.28.

Diffusion of a small fraction from toluene-petroleum ether

resulted in yellow crystals suitable for crystal structure determi-

nation.

Synthesis of 6c. In the glovebox, a vial was charged with a stir

bar and [Rh(µ-Cl)(cyclooctene)₂]₂(33.4 mg, 0.0465 mmol) and CH₂-

Cl₂(2 mL). Outside the glovebox, CO was bubbled through the

mixture for ca. 30 s, and the cap was closed quickly. The solution

lightened in color. The vial was brought into the glovebox and (i-

Pr)₂PPh (37.9 mg, 0.195 mmol) was added using CH₂Cl₂(total 5

mL, in portions). After 1 h, the solution was concentrated and the

residue crystallized from THF (ca. 3 mL) initially warmed, into

which was allowed to diffuse petroleum ether for 1 day. After this

time some crystals had formed but significant color remained in

(86) van der Ent, A.; Onderdelinden, A. L. Inorg. Synth. 1973, 14, 92-

95.

(87) Grotjahn, D. B.; Gong, Y.; Zakharov, L. N.; Golen, J. A.; Rheingold,

A. L. J. Am. Chem. Soc. 2006, 128, 438-453.

(88) Pople, J. A.; Schneider, W. G.; Bernstein, H. J. High-resolution

Nuclear Magnetic Resonance; McGraw-Hill: New York, 1959.

(89) Gu¨nther, H. Angew. Chem., Int. Ed. Engl. 1972, 11, 861-874.

3398 Organometallics, Vol. 26, No. 14, 2007

Grotjahn et al.

the liquid, so petroleum ether was added to the mixture and the

vial stored at -40 °C for 2 days. The cold supernatant was removed

by pipet and the yellowish crystals stored under vacuum to yield

m, 4H), 1.7-2.2 (br featureless m, 2 or 3H), 1.31 (dd, J ) 7.0,

15.6 Hz, 12H), 1.16 (dd, J ) 7.0, 13.5 Hz, 12H), -6 to -13 ppm

(br featureless peak, 1 or 2H). In addition, minor broad, featureless

peaks were seen at 3.90-4.02 (0.15H), 3.56-3.64 (0.12H), and

-20 to -22 ppm (ca. 0.2H). Partial spectrum at -30 °C: 7.55 (t,

J ) 7.0 Hz, 1H), 7.42 (t, J ) 7.0 Hz, 1H), 7.32-7.37 (m, 1H),

7.26-7.32 (m, 2H), 7.18-7.26 (m, 2H), 7.14 (t, J ) 7.0 Hz, 1H),

1.90 (s, 3H), 1.0-1.3 (m, 24H), -6.0 to -8.5 (br featureless m,

1H), -10 to -13 (br featureless m, 1H). Because the sample was

not pure, and some resonances were overlapping, and there were

other, minor peaks, it is very likely that even for the major species

not all peaks were assigned. ³¹P{¹H} NMR (CD₂Cl₂, 202.374 MHz,

0 °C): δ 72.3 ppm (dd, ¹J_P-Rh) 119.2 Hz and ²J_P-P) 367.5 Hz)

1

6c (46.0 mg, 89%). H NMR (CDCl₃, 500 MHz): δ 7.76-7.82

(m, 4H), 7.42-7.46 (m, 6H), 2.95 (septet of vt, ³J_H-H) 7.0 Hz, N

3

) 6.0 Hz, 4H), 1.29 (dvt, J_H-H) 7.0 Hz, N ) 17.0 Hz, 12H),

3

1.13 (dvt, J_H-H) 7.0 Hz, N ) 14.0 Hz, 12H). ¹³C{¹H} NMR

(CDCl₃, 125.7 MHz): δ 187.3 (dt, J_C-Rh) 73.6 Hz, J_C-P) 15.3

Hz), 134.9 (vt, N ) 10.6 Hz), 130.2 (s), 129.3 (dvt, J_C-Rh) 1.8

Hz, N ) 35.4 Hz), 127.7 (vt, N ) 9.7 Hz), 29.9 (s), 22.2 (vt, N )

23.9 Hz), 19.2 (vt, N ) 5.5 Hz), 18.2 ppm (s). ³¹P{¹H} NMR

1

(CDCl₃, 80.95 MHz): δ 42.6 ppm (d, J_P-Rh) 124.4 Hz). Anal.

Calcd for C₂₅H₃₈ClOP₂Rh (554.87): C, 54.11; H, 6.90. Found: C,

54.38; H, 6.78.

1

2

and 39.1 ppm (dd, J_Rh-P) 111.5 Hz and J_P-P) 367.5 Hz). IR

(CH₂Cl₂, NaCl): 2960, 2929, 2872 (m), 2110 (w, sl br).

Crystals suitable for X-ray diffraction were found in the bulk

sample crystallized from THF-petroleum ether.

Preparation of 5c-CCHBu. A solution of [(µ-Cl)Rh(cy-

clooctene)₂]₂(144.5 mg, 0.201 mmol) in C₆H₆(3 mL) was treated

with a solution of (i-Pr)₂PPh (184.4 mg, 0.949 mmol) in C₆H₆(3

mL). After stirring for 10 min, 1-hexyne (0.10 mL, 0.872 mmol)

was added to the solution via syringe. The reaction was stirred

overnight. The solvent was removed in Vacuo, and the remaining

solid was washed three times with portions (3 mL) of pentane. The

product was obtained after recrystallizing from pentane at -53 °C

Synthesis of 6d. By a procedure similar to that used to make

6c, [Rh(µ-Cl)(cyclooctene)₂]₂(35.7 mg, 0.0497 mmol) and (i-Pr)₂P-

(o-tol) (43.5 mg, 0.209 mmol) gave 6d (52.7 mg, 91%). ¹H NMR

3

(CDCl₃, 500 MHz): δ 7.42-7.47 (m, 2H), 7.33 (t, J_H-H) 7.5

3

Hz, 2H), 7.27-7.31 (m, 2H), 7.23 (t, J_H-H) 7.5 Hz, 2H), 3.12

3

(s, 6H), 2.80-2.95 (br featureless m, 4H), 1.32 (dvt, J_H-H) 7.5

Hz, N ) 15.5 Hz, 12H), 1.15 (dvt, ³J_H-H) 7.0 Hz, N ) 14.0 Hz,

12H). ¹³C{¹H} NMR (CDCl₃, 125.7 MHz): δ 186.8 (dt, J_C-Rh

)

1

as violet crystals, yield 176.6 mg (72%). H NMR (500 MHz,

74.6 Hz, J_C-P) 15.4 Hz), 144.4 (vt, N ) 13.1 Hz), 132.7 (s),

132.3 (vt, N ) 7.4 Hz), 129.6 (s), 126.6 (vt, N ) 33.6 Hz), 124.8

(vt, N ) 6.0 Hz), 25.3 (vt, N ) 9.4 Hz), 23.8 (br vt, N ) ca. 20

Hz), 19.9 (vt, N ) 5.3 Hz), 18.2 ppm (sl br s). ³¹P{¹H} NMR

CDCl₃): δ 7.74-7.69 (m, 4H, Ph), 7.40-7.37 (m, 6H, Ph), 3.03-

5

3

2.95 (m, 4H, PCHMe₂), 1.58 (tq, J_H-P) 2 Hz, J_{H-vinylidene H}

≈

3

³J_H-CH2) 7.7 Hz, 2H, CdCHCH₂CH₂CH₂CH₃), 1.27 (dvt, J_H-H

3

) 7 Hz, N ) 14 Hz, PCHCH₃), 1.09 (dvt, J_H-H) 7 Hz, N ) 14

1

3

(CDCl₃, 80.95 MHz): δ 31.1 ppm (d, J_P-Rh) 121.0 Hz). Anal.

Hz, PCHCH₃), 0.73 (∼sextet, J_H-H≈ 7.5 Hz, 2H, CdCHCH₂-

Calcd for C₂₇H₄₂ClOP₂Rh (582.93): C, 55.63; H, 7.26. Found: C,

55.51; H, 6.88.

Crystals suitable for X-ray diffraction were found in the bulk

CH₂CH₂CH₃), 0.56 (t, J ) 7.0 Hz, 3H, CdCHCH₂CH₂CH₂CH₃),

0.21 (∼quintet, ³J_H-H) 7.5 Hz, 2H, CdCHCH₂CH₂CH₂CH₃), 0.00

3

(dtt, J_H-Rh) 1.8 Hz, J_H-P) 3.5 Hz, J_H-CH2) 8 Hz, 1H, Rhd

CdCH). Assignments of the proton resonances in the vinylidene

side chain were made by means of a COSY spectrum. ³¹P NMR

sample crystallized from THF-petroleum ether.

Preparation of RhCl[(i-Pr)₂PIm′]₂(8). To a solution of [(µ-

Cl)Rh(cyclooctene)₂]₂(87.2 mg, 0.122 mmol) in CH₂Cl₂(10 mL)

was added 1b (123.6 mg, 0.486 mmol). After stirring for 10 min

at room temperature, the solvent was evaporated in Vacuo. The

remaining yellow or brownish-yellow solid was washed with three

(80.95 MHz, CDCl₃): δ 40.1 (d, J_P-Rh) 142.3 Hz). ¹³C (125.7

1

2

1

MHz, CDCl₃): δ 297.7 (td, J_C-P) 17.4 Hz, J_C-Rh) 55.2 Hz,

RhdC), 135.0 (vt, N ) 10.4 Hz, PPh-meta or ortho), 129.4 (dvt,

²J_C-Rh) 2.1 Hz, N ) 35.6 Hz, PPh-ipso), 129.3 (s, PPh-para),

127.0 (vt, N ) 8.7 Hz, PPh-ortho or meta), 106.1 (td, ³J_C-P) 6.9

1

3 mL portions of hexanes and dried, yield 148.4 mg (95%). H

2

Hz, J_C-Rh) 15.1 Hz, RhdCdC), 33.3 (s, CdCHCH₂CH₂CH₂-

NMR (200 MHz, CD₂Cl₂): δ 6.68, 6.49 (both s, 2H, Im′-H), 4.53,

3.60 (both s, 6H, N-CH₃), 2.55, 2.17 (both br, 4H, PCHMe₂), 1.37,

1.26 (both s, 18H, tBu), 1.43-1.14 (m, 24H, PCHCH₃). ³¹P NMR

CH₃), 21.94 (vt, N ) 22.8 Hz, PCHCH₃), 21.93 (s, CdCHCH₂-

CH₂CH₂CH₃), 19.4 (vt, N ) 5.9 Hz, PCHCH₃), 18.2 (s, PCHCH₃),

15.7 (s, CdCHCH₂CH₂CH₂CH₃), 13.7 (s, CdCHCH₂CH₂CH₂CH₃).

Carbon assignments were made by analogy with assignments made

for 5b-CCHBu using gCOSY, gHMQC, and gHMBC. Anal. Calcd

for C₃₀H₄₈ClP₂Rh (609.01): C, 59.17; H 7.94. Found: C, 59.56;

H, 7.96.

1

2

(80.95 MHz, CD₂Cl₂): δ 47.0 (dd, J_P-Rh) 181.1 Hz, J_P-P

)

42.2 Hz), 30.0 (dd, ¹J_P-Rh) 162.6 Hz, ²J_P-P) 42.7 Hz). ¹³C NMR

(125.7 MHz, CD₂Cl₂): δ 155.0 (dd, ³J_C-P) 10.1 Hz, ⁵J_C-P) 6.3

Hz, P-Im′-C4), 151.6 (d, ³J_PC) 7.0 Hz, Im′-C4), 149.6 (dd, ¹J_C-P

) 27.8 Hz, ³J_C-P) 6.7 Hz, P-Im′-C2), 143.4 (d, ¹J_C-P) 60.0 Hz,

Im′-C2), 117.5, 117.1 (both s, Im′-C5), 36.8, 34.9 (both s, Im′-N-

CH₃), 32.1, 32.0 (both s, Im′-CMe₃), 30.6, 30.1 (both s, Im′-

C(CH₃)₃), 23.7, 23.5 (both s, PCHMe₂), 20.7, 20.5 (both s,

PCHMe₂), 19.2 (s, PCH(CH₃)₂). Anal. Calcd for C₂₈H₅₄ClN₄P₂Rh

(647.06): C, 51.97; H, 8.41; N, 8.66. Found: C, 48.01; H, 7.47;

N, 7.73. For C₂₈H₅₄ClN₄P₂Rh‚CH₂Cl₂(731.99): calcd C, 47.58;

H, 7.71; N. 7.65.

Preparation of 5b-CCHBu. A solution of [(µ-Cl)Rh(cy-

clooctene)₂]₂(188.3 mg, 0.262 mmol) in C₆H₆(3 mL) was treated

with a solution of 1b (276.3 mg, 1.09 mmol) in C₆H₆(3 mL). After

stirring for 10 min, 1-hexyne (0.10 mL, 0.872 mmol) was added

to the solution via syringe and the reaction was stirred overnight.

The solvent was removed in Vacuo, and the remaining solid was

washed three times with portions (3 mL) of pentane. The product

was obtained after recrystallizing from pentane at -53 °C as violet

Synthesis of 9. The dimer [Rh(µ-Cl)(cyclooctene)₂]₂(44.8 mg,

0.0624 mmol) was suspended in pentane (1 mL), and (i-Pr)₂P(o-

tol) (53.5 mg, 0.257 mmol) was dissolved in pentane (1 mL). The

solution was added to the stirred suspension, and pentane (total 4

mL) was used to rinse the weighing vial. Within 5 min, the solids

had dissolved to give an orange solution. After 30 min, the mixture

was concentrated. Pentane was added to the red, oily residue and

the mixture concentrated. This was repeated. Attempts to crystallize

the product did not succeed, and further analysis was performed

on the residue (68.8 mg). Anal. Calcd for C₂₆H₄₂ClP₂Rh (554.92):

C, 56.27; H, 7.63. Found: C, 54.25; H, 7.74.

1

crystals, yield 245.4 mg (64%). H NMR (500 MHz, CDCl₃): δ

6.67 (s, 2H, Im′-H), 3.97 (s, 6H, Im′-NCH₃), 3.17-3.08 (m, 4H,

5

3

PCHCH₃), 1.94 (tq, J_H-P) 2 Hz, J_{H-vinylidene H}≈ J_H-CH2) 7.7

Hz, 2H, CdCHCH₂CH₂CH₂CH₃), 1.26 (s, 18H, Im′-C(CH₃)₃), 1.23,

1.15 (both dvt, J_H-H) 0.5 N ) 7.0 Hz, 12H, PCHCH₃), 1.01

3

(∼sextet, J_H-H) 7.5 Hz, 2H, CdCHCH₂CH₂CH₂CH₃), 0.74 (t,

3

3H, J ) 7.5 Hz, CdCHCH₂CH₂CH₂CH₃), 0.70 (∼quintet, J_H-H

) 7.5 Hz, 2H, CdCHCH₂CH₂CH₂CH₃), 0.37 (dtt, ³J_H-Rh) 1 Hz,

³J_H-P) 3.0 Hz, J_H-CH2) 8 Hz, 1H, RhdCdCHCH₂CH₂CH₂-

3

CH₃). ³¹P NMR (80.95 MHz, CDCl₃): δ 23.2 (d, J_P-Rh) 135.3

1

2

¹H NMR (CD₂Cl₂, 500 MHz, 30 °C): δ 7.4-7.6 (br featureless

m, 2H), 7.1-7.4 (br featureless m, 6H), 2.68-2.82 (br featureless

Hz). ¹³C NMR (125.7 MHz, CDCl₃): δ 298.0 (td, J_C-P) 17.4

1

Hz, J_C-Rh) 55.2 Hz, RhdC), 152.8 (vt, N ) 7.7 Hz, Im′-C4),

Transformation of Terminal Alkynes to Vinylidene Ligands

Organometallics, Vol. 26, No. 14, 2007 3399

138.0 (vt, N ) 60.4 Hz, Im′-C2), 117.7 (s, Im′-C5), 104.9 (td, ³J_C-P

) 6.5 Hz, ²J_C-Rh) 15.5 Hz, RhdCdC), 36.1 (s, Im′-N-CH₃), 33.6

(s, CdCHCH₂CH₂CH₂CH₃), 31.8 (s, Im′-C(CH₃)₃), 30.1 (s, Im′-

C(CH₃)₃), 24.0 (vt, N ) 27.3 Hz, PCHCH₃), 21.9 (s, CdCHCH₂-

CH₂CH₂CH₃), 19.6 and 18.7 (two sl br s, PCHCH₃), 14.7 (s, Cd

CHCH₂CH₂CH₂CH₃), 13.7 (s, CdCHCH₂CH₂CH₂CH₃). Proton and

carbon assignments were made using gCOSY, gHMQC, and

gHMBC. Anal. Calcd for C₃₄H₆₄ClN₄P₂Rh (729.20): C, 56.00; H,

8.85; N, 7.68. Found: C, 56.12; H, 8.73; N, 7.86.

spectra were acquired at low temperature, the sample was allowed

to warm to room temperature. NMR showed conversion of

π-complex to the corresponding vinylidene RhCl(PiPr₂Im′)₂(d

¹³Cd¹³CH₂) (5b-¹³C¹³CH₂), although the mixture contained minor

unidentified species, which made complete spectral assignments

difficult. ¹H NMR (500 MHz, CD₂Cl₂, 30 °C): δ 6.78 (s, 2H, Im′-

H), 3.96 (s, 6H, Im′-NCH₃), 3.10-3.20 (m, 4H, PCHCH₃), 1.30

(s, 18H, Im′-C(CH₃)₃), 1.25 (dvt, signal partially hidden by other

3

resonances, 12H, PCHCH₃), 1.17 (dvt, J_H-H) 0.5 N ) 7.0 Hz,

12H, PCHCH₃), 0.06 (dtd, ²J_H-C) 3.0 Hz, ²J_H-P) 3.5 Hz, ¹J_H-C

) 163.0 Hz, 2H, Rhd¹³Cd¹³CH₂). ³¹P NMR (202.374 MHz, CD₂-

Cl₂, 30 °C): δ 25.7 (ddd, ³J_P-C) 5.9 Hz, ²J_P-C) 17.0 Hz, ¹J_P-Rh

) 135.3 Hz). Partial ¹³C NMR (125.7 MHz, CD₂Cl₂, 30 °C): δ

296.7 (tdd, ²J_C-P) 17.1 Hz, ¹J_C-Rh) 55.4 Hz, ¹J_C-C) 58.0 Hz,

Reaction of [RhCl((i-Pr)₂PPh)₂]₂(7c) with ¹³C-Labeled Acety-

lene. A J-Young tube containing a solution of 7c (10.1 mg, 0.010

mmol) in CD₂Cl₂(1.0 mL) was stored in the glovebox freezer

(-40 °C) for a few minutes before it was removed, and ¹³C-labeled

acetylene H¹³Ct¹³CH (2 mL in a syringe) was bubbled slowly for

1 min. The J-Young tube was then sealed and immediately taken

out of the glovebox and placed in an acetone-dry ice bath for

transport to a precooled (-20 °C) NMR probe. The NMR data

were acquired immediately after brief experiment setup. ¹H NMR

(500 MHz, CD₂Cl₂, -20 °C) for 2c-H¹³C¹³CH: δ 7.58-7.42 (m,

10H, Ph), 3.30 (HH′CC′ 10 peaks with additional coupling ²J_H-Rh

3

2

Rhd¹³Cd¹³CH₂), 89.5 (tdd, J_C-P) 5.9 Hz, J_C-Rh) 16.5 Hz,

¹J_C-C) 58.0 Hz, Rhd¹³Cd¹³CH₂).

Kinetics Experiments. All reagents, solvents, NMR tube, and

syringes were precooled to -35 °C before any manipulations.

Complex 7c or 8 was dissolved in CD₂Cl₂(1.00 mL) in a J-Young

NMR tube, 1-hexyne was added using a microsyringe, and the NMR

tube was then sealed with the screw cap. Quantities used were the

following: run 1, complex 8 (13.7 mg, 0.0212 mmol) and 1-hexyne

(50 µL, 0.436 mmol, 21.8 equiv); run 2, complex 8 (14.0 mg, 0.0216

mmol) and 1-hexyne (50 µL, 0.436 mmol, 20.2 equiv); run 3,

complex 8 (25.7 mg, 0.0397 mmol) and 1-hexyne (50 µL, 0.436

mmol, 11.0 equiv); run 4, complex 8 (37.0 mg, 0.057 mmol) and

1-hexyne (7.0 µL, 0.061 mmol, 1.07 equiv); and finally, run 5,

complex 7c (26.0 mg, 0.0493 mmol) and 1-hexyne (6.0 µL, 0.05

mmol, 1.0 equiv). The tube was immediately removed from the

glovebox and placed in an acetone-dry ice bath to prevent the

starting of the reaction, and shaken again to obtain a well-mixed

solution just before insertion into the NMR probe precooled to

-20 °C. The experiment was started immediately afterward, after

allowing about 1 min for thermal equilibration and experiment

setup. Acquisition parameters included 1 transient of 90° pulse

width, followed by acquisition, and then in intervals of between 4

and 10 min between pulses. The data were collected at regular

interval times, each one with identical acquisition parameters. The

concentration of each component was calculated from the peak

intensities of the resonance of each complex against the peak of

internal standard (Me₃Si)₄C, which was set arbitrarily equal to 100.

1

2

3

) 2.5 Hz: J_H-C) 233.2 Hz, J_H-C) 29.1 Hz, J_H-H) 1.7 Hz,

¹J_C-C) 115.8 Hz, 2H, Rh-η²-H¹³Ct¹³CH′), 2.54-2.65 (m, 4H,

PCHCH₃), 1.25 (dvt, J_H-H) 0.5 N ) 7.5 Hz, PCHCH₃), 1.07

3

(dvt, ³J_H-H) 0.5 N ) 7.5 Hz, PCHCH₃). ³¹P NMR (202.374 MHz,

CD₂Cl₂, -20 °C): δ 35.8 (d, ¹J_P-Rh) 120.5 Hz). Partial ¹³C NMR

(125.7 MHz, CD₂Cl₂, -20 °C): δ 68.7 (dt, ¹J_C-Rh) 15.3 Hz, ²J_C-P

) 2.0 Hz, Rh-(η²-H¹³C¹³CH)). [In a separate experiment using less

alkyne, a sample observed quickly at 30 °C showed peaks centered

on 3.25 ppm for the HH′CC′ spin system, ²J_H-Rh) 2.5 Hz, ¹J_H-C

2

3

1

) 232.8 Hz, J_H-C) 29.0 Hz, J_H-H) 1.8 Hz, J_C-C) 116.2

Hz.] The sample was allowed to warm to room temperature.

NMR showed conversion of π-complex 2c-H¹³C¹³CH to vinyli-

dene RhCl(PiPr₂Ph)₂(d¹³Cd¹³CH₂) (5c-¹³C¹³CH₂). ¹H NMR (500

MHz, CD₂Cl₂, 30 °C): δ 7.73-7.77 (m, 10H, Ph), 2.94-3.06 (m,

4H, PCHCH₃), 1.25 (dvt, ³J_H-H) 0.5 N ) 7.8 Hz, PCHCH₃), 1.13

3

2

(dvt, J_H-H) 0.5 N ) 7 Hz, PCHCH₃), -0.26 (dtd, J_H-C) 3.0

Hz, ⁴J_H-P) 3.5 Hz, ¹J_H-C) 163.0 Hz, 2H, Rhd¹³Cd¹³CH₂). ³¹

P

1

NMR (80.95 MHz, CD₂Cl₂, 30 °C): δ 41.6 (ddd, J_P-Rh) 142.3

Hz, J_P-C) 17.3 Hz, J_P-C) 6.7 Hz). Partial ¹³C NMR (125.7

2

3

2

1

MHz, CD₂Cl₂, 30 °C): δ 297.9 (tdd, J_C-P) 17.1 Hz, J_C-Rh

)

55.6 Hz, J_C-C) 58.7 Hz, Rhd¹³Cd¹³CH₂), 90.7 (tdd, J_C-P

1

3

6.5 Hz, J_C-Rh) 15.7 Hz, J_C-C) 58.7 Hz, Rhd¹³Cd¹³CH₂).

2

1

Crossover Reaction of [RhCl((i-Pr)₂PPh)₂]₂(7c) with 1-Hex-

yne-1-d₁and 3-Methyl-1-butyne. To a NMR tube containing

[RhCl((i-Pr)₂Ph)₂]₂(7c) (13.2 mg, 0.025 mmol) was added CD₂-

Cl₂(0.5 mL). The tube was shaken and sonicated for 2 min to make

sure all complex dissolved. The sample was stored at -78 °C in a

dry ice-acetone bath. A solution of 1-hexyne-1-d₁(2.5 µL, 0.021

mmol) and 3-methyl-1-butyne (2.5 µL, 0.024 mmol) in CD₂Cl₂(0.5

mL) was added to the NMR tube at -78 °C. The NMR tube was

immediately inserted into the precooled NMR probe (-20 °C) after

allowing about 2 min for thermal equilibration and experiment

setup. The NMR spectra were recorded at three different temper-

atures: -20, 0, and 30 °C. Spectra at low temperatures (-20 and

Reaction of RhCl((i-Pr)₂PIm′)₂(8) with ¹³C-Labeled Acety-

lene. A J-Young tube containing a solution of 8 (12.8 mg, 0.020

mmol) in CD₂Cl₂(1.0 mL) was stored in the glovebox freezer

(-40 °C) for a few minutes before it was removed and ¹³C-labeled

acetylene H¹³Ct¹³CH (2 mL in a syringe) was bubbled slowly for

1 min. The J-Young tube was then sealed and immediately taken

out of the glovebox and placed in an acetone-dry ice bath for

transport to a precooled (-20 °C) NMR probe. The NMR data

were acquired immediately after brief experiment setup. ¹H NMR

(500 MHz, CD₂Cl₂, -20 °C) for 2b-H¹³C¹³CH: δ 6.79 (s, 2 H,

Im′-H), 4.23 (s, 6 H, Im′ N-CH₃), 3.78 (HH′CC′ 10 peaks with

2

additional coupling, J_H-Rh) 2.4 Hz; broadness of peaks under

0 °C) showed doublets in ³¹P NMR spectra at δ 32.8 (d, ¹J_P-Rh

)

1

these conditions precluded determining the various coupling

constants), 2.1-2.8 (featureless m, 6H, PCHCH₃), 1.38 (s, 18H,

Im′-C(CH₃)₃), 0.8-1.4 (featureless m, integral hard to determine

accurately, PCHCH₃). ³¹P NMR (202.374 MHz, CD₂Cl₂,

123.7 Hz) and 35.0 (d, J_P-Rh) 123.7 Hz), corresponding to

π-complex [RhCl((i-Pr)₂PPh)₂(η²-CDtCCH₂CH₂CH₂CH₃)] (2c-

DCCBu) and [RhCl((i-Pr)₂PPh)₂(η²-CHtCCH(CH₃)₂)] (2c-HC-

CiPr), respectively. After isomerization of π-complex to vinylidene

was completed at 30 °C, there were only two doublets, at δ 42.4

1

-20 °C): δ 19.4 (d, J_P-Rh) 117.4 Hz). ¹³C NMR (125.7 MHz,

1

CD₂Cl₂, -20 °C): δ 152.6 (vt, N ) 7.2 Hz, Im′ C4), 137.6 (vt, N

) 57.8 Hz, Im′ C2), 118.1 (s, Im′ C5), 67.5 (dt, ¹J_C-Rh) 13.1 Hz,

²J_C-P) 2.0 Hz, Rh-(η²-H¹³C¹³CH)), 37.2 (s, Im′ N-CH₃), 31.9 (s,

C(CH₃)₃), 30.1 (s, C(CH₃)₃), 20-23 (br, CH(CH₃)₂), 17.5-18.8

(br, CH(CH₃)₂)). [In a separate experiment using less alkyne, a

sample observed quickly at 20 °C showed peaks centered on 3.69

ppm for the HH′CC′ spin system, ²J_H-Rh) 2.6 Hz, ¹J_H-C) 232.9

(d, J_P-Rh) 141.5 Hz) and 42.1 (d, J_P-Rh) 142.7 Hz),

corresponding to the vinylidene [RhCl((i-Pr)₂PPh)₂(dCdCDCH₂-

CH₂CH₂CH₃)] (5c-CCDBu) and [RhCl((i-Pr)₂PPh)₂(dCdCHCH-

(CH₃)₂)] (5c-CCHiPr), respectively. ¹H NMR spectra indicated no

crossover reaction, for 5c-CCDBu: δ 1.67 (m, RhdCdCDCH₂,

2H), 0.78 (m, RhdCdCDCH₂CH₂CH₂CH₃, 2H), 0.60 (t, J ) 7.5

Hz, RhdCdCDCH₂CH₂CH₂CH₃, 3H), 0.30 (m, RhdCdCDCH₂-

CH₂CH₂CH₃, 2H), 0.12 (none, RhdCdCD); for 5c-CCHiPr: δ

2

3

1

Hz, J_H-C) 29.3 Hz, J_H-H) 1.7 Hz, J_C-C) 115.8 Hz.] After

3400 Organometallics, Vol. 26, No. 14, 2007

Grotjahn et al.

3

0.05 (d, J_H-H) 6.5 Hz, RhdCdCHCH(CH₃)₂, 6H), -0.12 (dt,

Over 2 days at room temperature the complex formed 5a-

¹³C¹³CH₂: ¹H NMR (500 MHz, C₆D₆, 30 °C): δ 2.78-2.90 (m,

³J_H-H) 10.0 Hz, J_H-P) 3.5 Hz, RhdCdCH, 1H).

4

3

PCH, 6H), 1.32 (dvt, J_H-H) 6.0 Hz, N ) 13.5 Hz, PCHCH₃),

The results from this crossover experiment were compared with

those from a similar experiment using an equimolar mixture of

undeuterated 1-hexyne and 3-methyl-1-butyne in order to estimate

detection limits.

-0.13 (qd, ³J_H-Rh) ⁴J_H-P) 3.2 Hz, ¹J_H-C) 161.7 Hz). ³¹P NMR

1

(80.95 MHz, C₆D₆, 30 °C): δ 42.3 (d, J_P-Rh) 135.3 Hz). ¹³C

NMR (125.7 MHz, C₆D₆, 30 °C): δ 291.0 (tdd, ²J_C-P) 16.0 Hz,

1

3

¹J_C-Rh) J_C-C) 57.0 Hz, Rhd¹³Cd¹³CH₂), 89.4 (tdd, J_C-P

)

Crossover Reaction of RhCl((i-Pr)₂PIm′)₂(8) with 1-Hexyne-

1-d₁and 3-Methyl-1-butyne. To a NMR tube containing RhCl-

((i-Pr)₂PIm′)₂(8) (13.4 mg, 0.021 mmol) was added CD₂Cl₂(0.5

mL). The tube was shaken and sonicated for 2 min to make sure

all complex dissolved. The sample was then stored at -78 °C in a

dry ice-acetone bath. A solution of 1-hexyne-1-d₁(1.5 µL, 0.013

mmol) and 3-methyl-1-butyne (1.5 µL, 0.015 mmol) in CD₂Cl₂(0.5

mL) was added to the NMR tube at -78 °C. The NMR tube was

immediately inserted into the precooled NMR probe (-20 °C) after

allowing about 2 min for thermal equilibration and experiment

setup. The NMR spectra were recorded at three different temper-

atures: -20, 0, and 30 °C. Interestingly, in ³¹P NMR spectra at

lower temperatures, -20 and 0 °C, for the alkyne π-complexes a

broadened peak near 18 ppm was observed, whereas peaks for 8

and final vinylidene complexes were sharp. At 30 °C, the two

5.9 Hz, ²J_C-Rh) 17.0 Hz, ¹J_C-C) 56.7 Hz, Rhd¹³Cd¹³CH₂). The

assignment of J_C-Creported previously⁴⁸must be incorrect. One

1

interesting comparison regarding this isotopomer and the natural-

abundance one is that the magnitude of ³J_H-Rhfor the ¹H resonance

at -0.13 ppm is different (3.2 Hz here and 1.0 Hz in the case of

5a-CCH₂).

Preparation of π-Complexes [RhCl((i-Pr)₃P)₂(η²-HCtCH)]

(2a-HCCH) and [RhCl((i-Pr)₃P)₂(η²-D¹³Ct¹³CD)] (2a-D¹³C¹³CD)

and Their Isomerization to Vinylidenes. The π-alkyne complex

from HCCH was prepared as described by Werner et al.,⁹⁰whereas

the labeled one was made using D¹³C¹³CD. A stirred solution of

[(µ-Cl)Rh(cyclooctene)₂]₂(0.232 g, 0.323 mmol) in pentane (30

mL) was treated with (i-Pr)₃P (1a) (0.503 g, 3.14 mmol) and stirred

for 10 min at room temperature. The solution was cooled to 0 °C

in an ice bath. The acetylene gas (20 mL in the case of D¹³C¹³CD)

was bubbled through the solution for 10 s at 1 atm, forming instantly

a yellow precipitate. The solution was cooled further to -78 °C

for 2 h and filtered, and the solid washed with cold (-35 °C)

pentane and dried. Yield for 2a-HCCH: 165.2 mg, 53%; for 2a-

D¹³C¹³CD: 166.5 mg, 52%.

1

doublets at δ 25.6 (d, J_P-Rh) 135.8 Hz) and 25.4 (d, J_P-Rh

)

136.4 Hz) confirmed the formation of vinylidenes [RhCl((i-

Pr)₂PIm′)₂(dCdCDCH₂CH₂CH₂CH₃)] (5b-CCDBu) and [RhCl-

((i-Pr)₂PIm′)₂(dCdCHCH(CH₃)₂)] (5b-CCHiPr). ¹H NMR spectra

indicated no crossover reaction, for 5b-CCDBu: δ 2.04 (m, Rhd

CdCDCH₂, 2H), 1.06 (m, RhdCdCDCH₂CH₂CH₂CH₃, 2H), 0.91

(m, RhdCdCDCH₂CH₂CH₂CH₃, 3H), 0.77 (m, RhdCdCDCH₂CH₂-

CH₂CH₃, 2H), 0.53 (none, RhdCdCD); for 5b-CCHiPr: δ 0.48

NMR data for 2a-HCCH matched those in the literature:^{90 1}

H

3

NMR (200 MHz, C₆D₆, 30 °C): δ 3.30 (d, J_H-Rh) 2.6 Hz, 2H,

(d, ³J_H-H) 7.0 Hz, RhdCdCHCH(CH₃)₂, 6H), 0.23 (dt, ³J_H-H

)

3

HCtCH), 2.33 (m, PCH, 6H), 1.25 (dvt, J_H-H) 6.4 Hz, N )

4

13.2 Hz, PCHCH₃). ³¹P NMR (81.0 MHz, C₆D₆, 30 °C): δ 34.3

(d, ¹J_P-Rh) 116.6 Hz). Isomerization takes place in C₆D₆solution

at 30 °C. NMR data for 5a-CCH₂matched those in the literature:

^{34 1}H NMR (500 MHz, C₆D₆, 30 °C): δ 2.85 (m, PCH, 6H), 1.32

10.0 Hz, J_H-P) 3.0 Hz, RhdCdCH, 1H).

The results from this crossover experiment were compared with

those from a similar experiment using an equimolar mixture of

undeuterated 1-hexyne and 3-methyl-1-butyne in order to estimate

detection limits.

3

(dvt, J_H-H) 6.0 Hz, N ) 13.5 Hz, PCHCH₃), -0.13 (dt, J_H-Rh

) 1.0 Hz, J_H-P) 3.5 Hz). ³¹P NMR (80.95 MHz, C₆D₆, 30 °C):

δ 42.3 (d, J_P-Rh) 135.3 Hz).

4

Preparation of π-Complex [RhCl((i-Pr)₃P)₂(η²-H¹³Ct¹³CH)]

(2a-H¹³C¹³CH) and Isomerization to the Vinylidene. The π-alkyne

complex from H¹³C¹³CH was prepared in analogy with the method

described by Werner et al.⁹⁰A stirred solution of [(µ-Cl)Rh-

(cyclooctene)₂]₂(0.1218 g, 0.170 mmol) in pentane (15 mL) was

treated with (i-Pr)₃P (1a) (0.2595 g, 1.62 mmol) and stirred for 20

min at room temperature. The solution was cooled to 0 °C in an

ice bath. The acetylene gas (12.5 mL) was bubbled through the

solution over 30 s at 1 atm, lightening the reaction mixture color

to brownish-yellow. A yellow precipitate began to form. After 10

min the solution was cooled further to -78 °C for 2 h and filtered,

and the solid washed with cold (-35 °C) pentane and dried. Yield

of 2a-H¹³C¹³CH: 113 mg, 68%.

1

NMR data for 2a-D¹³C¹³CD: ¹H NMR (200 MHz, C₆D₆,

30 °C): δ 2.21-2.44 (m, PCH, 6H), 1.25 (dvt, ³J_H-H) 6.2 Hz, N

) 13.0 Hz, PCHCH₃). ³¹P NMR (81.0 MHz, C₆D₆, 30 °C): δ 34.4

1

(d, J_P-Rh) 116.5 Hz). ¹³C NMR (50.3 MHz, C₆D₆, 30 °C): δ

67.2-69.4 (m, D¹³Ct¹³CD), 22.2 (vt, N ) 17.6, PCH), 20.2 (s,

PCHCH₃). Isomerization takes place in C₆D₆solution at 30 °C.

NMR data for 5a-¹³C¹³CD₂: ¹H NMR (500 MHz, C₆D₆, 30 °C):

3

δ 2.70-2.95 (m, PCH, 6H), 1.32 (dvt, J_H-H) 6.0 Hz, N ) 13.0

Hz, PCHCH₃). Although there should be no high-field signal,

centered on -0.13 ppm was a wide doublet of featureless multiplets,

¹J_H-C) 160 Hz, for 5a-¹³C¹³CHD. The total integration of the

two peaks corresponded to 0.28 H, meaning that the vinylidene

terminal carbon bears 14% H. In the crossover experiment described

below the amount of this material was cut to around half, reflecting

the fact that in the crossover experiment only half of the starting

material was 2a-D¹³C¹³CD. ³¹P NMR (80.95 MHz, C₆D₆, 30 °C):

NMR data for 2a-H¹³C¹³CH: ¹H NMR (500 MHz, C₆D₆,

30 °C): δ 3.30 (HH′CC′ 10 peaks with additional coupling, ²J_H-Rh

1

2

3

) 2.5 Hz, J_H-C) 228.5 Hz, J_H-C) 27.5 Hz, J_H-H) 1.7 Hz,

¹J_C-C) 114.5 Hz, 2H, Rh-η²-H¹³Ct¹³CH′), 2.25-2.40 (m, PCH,

6H), 1.25 (dvt, ³J_H-H) 6.2 Hz, N ) 13.0 Hz, 36 H, PCHCH₃). ¹H

NMR (500 MHz, CD₂Cl₂, 30 °C): δ 3.43 (HH′CC′ 10 peaks with

1

2

3

δ 42.4 (ddd, J_P-Rh) 135.3 Hz, J_P-C) 16.2 Hz, J_P-C) 5.7

Hz). ¹³C NMR (50.3 MHz, C₆D₆, 30 °C): δ 291.3 (tt, J_C-Rh

)

1

additional coupling, ²J_H-Rh) 2.5 Hz, ¹J_H-C) 229.7 Hz, ²J_H-C

)

¹J_C-C) 57.0 Hz, J_C-P) 16.0 Hz, RhdC), 87.5-90.5 (m, Rhd

2

28.0 Hz, J_H-H) 1.7 Hz, J_C-C) 114.6 Hz, 2H, Rh-η²-H¹³Ct

3

1

¹³CH′), 2.34 (septet of vt, N ) J_HH) 7.2 Hz, PCH, 6H), 1.27

3

CdCD₂), 23.8 (vt, N ) 19.9 Hz, PCH), 20.6 (s, PCHCH₃).

(dvt, J_H-H) 6.2 Hz, N ) 13.0 Hz, 36 H, PCHCH₃). ³¹P NMR

3

Crossover Reaction of [RhCl((i-Pr)₃P)₂(η²-HCtCH)] (2a-

HCCH) and [RhCl((i-Pr)₃P)₂(η²-D¹³Ct¹³CD)] (2a-D¹³C¹³CD).

A J-Young NMR tube was charged with 2a-HCCH (11.8 mg,

0.0243 mmol) and 2a-D¹³C¹³CD (12.2 mg, 0.0250 mmol), and the

two solids were dissolved in C₆D₆. The reaction was monitored by

¹H NMR spectroscopy. After 18 h, about 60% of the alkyne

π-complexes had been converted to vinylidenes, a process com-

pleted after 3 days. The amount of 5a-¹³C¹³CHD formed appeared

to be about half that obtained from pure 2a-D¹³C¹³CD, as measured

by integration of the small, wide doublet of featureless multiplets

2

1

(81.0 MHz, C₆D₆, 30 °C): δ 34.4 (td, J_P-C) 2.4 Hz, J_P-Rh

)

116.5 Hz). ¹³C NMR (125.7 MHz, C₆D₆, 30 °C): δ 68.9 (td, ²J_C-P

) 2.4 Hz, J_C-Rh) 15.7 Hz, H¹³Ct¹³CH), 22.2 (vt, N ) 17.3,

1

PCH), 20.0 (s, PCHCH₃). ¹³C NMR (125.7 MHz, CD₂Cl₂, 30 °C):

δ 68.4 (td, J_C-P) 2.4 Hz, J_C-Rh) 16.0 Hz, H¹³Ct¹³CH), 22.0

2

1

(vt, N ) 17.2, PCH), 19.8 (s, PCHCH₃).

(90) Werner, H.; Wolf, J.; Schubert, U.; Ackermann, K. J. Organomet.

Chem. 1986, 317, 327-356.

Transformation of Terminal Alkynes to Vinylidene Ligands

Organometallics, Vol. 26, No. 14, 2007 3401

for 5a-¹³C¹³CHD. Thus, the amount of H at the terminal carbon of

5a-¹³C¹³CD₂was no greater than when pure 2a-D¹³C¹³CD was

used, and therefore H incorporation in Scheme 3 is shown as “ca.

0% H”.

remains tractable for the present systems, provided that the i-Pr

groups are replaced by hydrogens in the model compounds. The

kinetics and ligand-dependence of the chemistry were therefore

initially investigated by B3LYP/LANL2DZ calculations on the

[RhCl(H₂PR¹)₂](C₂H₂)-derived species.

Deuterium NMR was performed on a solution made by evapora-

tion of the reaction mixture and redissolution of the residue in C₆H₆

(1 mL) and C₆D₆(ca. 0.05 mL). A large doublet (¹J_D-C) 24.7

Hz) for the d¹³Cd¹³CD₂moiety of 5a-¹³C¹³CD₂was seen,

The final calculations presented in this paper take advantage of

the recent benchmark study by Truhlar and co-workers of DFT

techniques applied to organometallics.⁹⁵They found that the

nonhybrid BLYP method was superior to B3LYP for a broad range

of predicted properties. That study relied on the relativistic, compact

effective potentials (RCEP) of Stevens et al.⁹⁶and either the

accompanying double-ú basis set or a modified Hay-Wadt triple-ú

basis set, supplemented by Pople’s 6-31+G(d,p) basis for the main

group atoms. Their findings found little improvement upon exten-

sion from the double-ú (DQZ) to triple-ú basis. The final optimiza-

tions reported here were therefore carried out at the BLYP level

with the DQZ basis set. Selected calculations were also carried out

at the MP2 and CISD levels (see Supporting Information) to test

the sensitivity of our initial results to computational technique.

To simplify computations on the i-Pr-substituted species, the

experimentally studied molecules were modeled by ONIOM

calculations⁹⁷in which the i-Pr groups were treated exclusively by

UFF molecular mechanics and coupled to the simplified [RhCl-

(H₂PR¹)₂](C₂H₂) system, which in turn was optimized at the BLYP/

DQZ level. These ONIOM calculations, designated BDMM in this

work, are designed after the IMOMM calculations described by

Wakatsuki et al.⁴¹A series of optimizations were also carried out

with a modified version of their IMOMM MP2/MM3 calculations,

using a slightly larger basis set (replacing 4-31G by 6-31G* and

using the n+1 basis of Hay and Wadt^93,94and UFF instead of MM3

molecular mechanics); these calculations support the previously

reported⁴¹reaction barrier of 30 kcal/mol. However, Grimme’s

recently proposed spin-component scaling of the MP2 energy

contributions⁹⁸reduces this barrier to 17 kcal mol^-1, in rough

agreement with DFT calculations that use the same basis set. All

calculations were carried out using the Gaussian 03 suite of

programs⁹⁹running on a Pentium Xeon cluster under Linux.

Conformational Analysis. Even in the absence of the i-Pr

groups, there is considerable conformational flexibility in these

calculations, particularly for the Im- and o-Tol-substituted species.

Therefore, 10 distinct initial geometries of the [HRhCl(PH₂Im)₂]-

CCH hydrido(alkynyl) species 4g were chosen to survey potential

effects of different intermolecular interactions between locations

on the imidazole and on the central [HRhCl]CCH framework. The

hydrido species was chosen in case its lower symmetry, relative to

that of the π-complex and vinylidene isomers, influenced the

preferred conformation. Similar but less rigorous conformational

tests were carried out for the Ph- and o-Tol-substituted compounds

(4h and 4i) and for species 2 and 5, and frequencies were calculated

in all cases to ensure that these were minimum energy geometries.

In all cases with R¹) Im, Ph, or o-Tol, the lowest energy

conformations of the [RhCl(H₂PR¹)₂](HCCH)-derived species 2-5

placed both ligand rings roughly anti with respect to each other,

each with one CH bond oriented toward the Cl atom for slight

electrostatic stabilization. In the case of the Im and Tol ligands,

the Cl-oriented hydrogen atom is located on the methyl group.

Geometries were subsequently optimized from initial conformations

that exploited this stabilizing characteristic. Although the relative

energies of these conformers (e.g., E(gauche) - E(anti) for 2g)

4

2

presumably slightly broadened by J_D-Pand J_D-C. In the center

of this doublet was a very small peak, which could be from 5a-

CCHD. Deconvolution of the spectrum assuming Lorentzian line

shape led to the fit shown in the Supporting Information. The ratio

of integral areas for the large peaks of the doublet for 5a-¹³C¹³CD₂

and what may be the singlet for 5a-CCHD was 270 to 1, which

would correspond to a molar ratio of 135 to 1. Thus, a 1% detection

limit is given for D in 5a-CCH₂.

Ruling Out Isotope Effect as Responsible for Lack of

Crossover. In separate silanized and dried J. Young NMR tubes,

solutions of 2a-D¹³C¹³CD and 2a-HCCH in C₆D₆were allowed

to rearrange to the respective vinylidenes at 30 °C in the NMR

probe. Comparing rates of disappearance of starting materials

¹³C¹³CD

showed that k_HCCH/k_D

) 1.67, a relatively small effect

compared with the low or undetectable D incorporation in crossover

experiments above.

Computational Methods. Two methodologies are employed for

the principal computational results reported in this study:

1. B3LYP/LANL2DZ calculations were used to carry out (a)

the initial conformational analysis on structures 4g, 4h, and 4i; (b)

mapping of the reaction surface for structures 2f-5f (Figure 6);

and (c) frequency calculations for zero-point energy and thermal

contributions to the free energies of 2a-5a, 2c-5c, 2d-5d, and

2e-5e.

2. All tabulated numerical results from the computational work

(Tables 3, 4, and 5) are obtained by BLYP/DQZ/UFF ONIOM

calculations (labeled BDMM), carried out on structures 2a-5a, 2c-

5c, 2d-5d, and 2e-5e, with free energy corrections from the

B3LYP/LANL2DZ results above.

The experimental work on imidazolylphosphine complexes used

ligand 1b, which featured a tert-butyl group adjacent to the basic

nitrogen (Im′). All calculations described in this paper were made

tractable by using complexes of imidazolylphosphines 1e and 1g,

ligands lacking the tert-butyl group, where the 1-methylimidazol-

2-yl unit is abbreviated as Im; the evidence indicates that this is a

reasonable simplification for the systems considered here.⁹¹The

initial calculations were carried out on simplified systems, the

[RhCl(H₂PR¹)₂](C₂H₂)-derived complexes, with R¹) H, Im, Ph,

o-Tol (bearing phosphine ligands 1f-1i, Scheme 2) and with the

C₂H₂group arranged to correspond to each of four minimum-energy

species A-D in Scheme 1, specifically analogues f-i of 2-5 in

Scheme 2. The final series of calculations replaced the H atoms in

the H₂PR¹ligands with the i-Pr groups present in the experimentally

studied compounds and were carried out for each of the species

2-5 derived from acetylene complexes [RhCl((i-Pr)₂PR¹)₂](C₂H₂)

(2a-, 2c-, 2d-, and 2e-HCCH, with R¹) i-Pr, Im, Ph, o-Tol).

Previous experimental studies on similar systems in this labora-

tory have been successfully modeled⁹²using the B3LYP hybrid

density functional and the LANL2DZ basis set, based on Hay and

Wadt’s effective core potentials.^93,94This computational approach

(91) In other applications of heterocyclic phosphines the steric hindrance

at the basic nitrogen is quite important,³²but as seen in calculated structures

here, there is no evidence of the involvement of basic nitrogen in alkyne-

to-vinylidene transformation on Rh(I). Therefore, we feel that lack of a

ring tert-butyl substituent in calculated species should be of little conse-

quence.

(92) Lev, D. A.; Grotjahn, D. B.; Amouri, H. Organometallics 2005,

24, 4232-4240.

(93) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270-283.

(94) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299-310.

(95) Schultz, N. E.; Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2005,

109, 11127-11143.

(96) Stevens, W. J.; Krauss, M.; Basch, H.; Jasien, P. G. Can. J. Chem.

1992, 70, 612-630.

(97) Maseras, F.; Morokuma, K. J. Comput. Chem. 1995, 16, 1170-

1179.

(98) Grimme, S. J. Chem. Phys. 2003, 118, 9095-9102.

(99) Frisch, M. J.; et al. Gaussian 03; Gaussian, Inc.: Pittsburgh, PA,

2003. For complete reference, see Supporting Information.

3402 Organometallics, Vol. 26, No. 14, 2007

Grotjahn et al.

change by as much as 3 kcal mol^-1, and more when the i-Pr groups

are added, the energy difference between any specific pair of

conformers along the reaction path (e.g., ∆E_2-4(gauche) -

∆E_2-4(anti) for 2g) was not found to exceed 1.5 kcal mol^-1. The

ONIOM optimizations of structures featuring (i-Pr)₂PR¹ligands

yielded no geometries that differed markedly in the central

framework from those featuring the PH₂R¹ligands. (A series of

BDMM calculations were carried out on structures 2g-5g explicitly

to test the possibility of the imidazolyl ligand acting as an internal

base to stabilize the HCCH system, but none of these conformers

were more stable than the conformers described above.) Minimum

energy geometries were then identified by analytical optimizations

from initial geometries consistent with the most stable conformers

identified in the survey.

Transition-State Optimizations. For the R ) H model systems,

analytical optimizations of the transition states TS23 and TS45 were

carried out from initial geometries suggested by the relaxed PES.

Frequency analysis confirmed that these were saddle points. This

analysis was repeated for TS34 only in the simplest model system,

[RhCl(PH₃)₂](C₂H₂), because the minimum associated with 3 is

apparently too shallow to contribute substantially to the observed

kinetics. (It is for this reason that the transition state that technically

connects geometries 3 and 5 is referred to as TS45 rather than

TS35.) The transition states in the ONIOM calculations were

numerically optimized. The local coordinate most associated with

the reaction coordinate near each transition state was identified from

the Hessian matrix of the corresponding, analytically optimized

[RhCl(H₂PR¹)₂](C₂H₂) transition state. These coordinates, the Rh-

C2 bond length for TS23 and the C2-C1-H bond angle for TS45,

were initially fixed to the value optimized at the level B3LYP/

LANL2DZ for the simplified [RhCl(PH₂R)₂](C₂H₂) model system.

The transition states were then identified in the ONIOM calculations

by stepping this parameter in both directions and repeating the

partial optimization to determine the location of the peak of the

relaxed PES along this coordinate. It was found that the reaction

coordinate near TS23 was too heavily coupled to the Rh-C2-C1

bond angle for this method, and therefore the Rh-C2-C1 angle

was also coarsely scanned in roughly 9° increments to improve

the predicted TS23 energy to an estimated precision of 1 kcal/

mol.

and TS45 when using DFT methods rather than MP2 calculations,

and the lowest predicted barrier height in the preliminary B3LYP/

LANL2DZ calculations is 26 kcal mol^-1. Results by Hirao¹⁰⁰and

our own vibrational and geometrical benchmarks using RhC and

RuO₄(see Supporting Information) suggest that the DFT methods

should be more accurate than the MP2. However, it is only upon

adoption of the DQZ basis that the reaction barriers predicted under

these conditions become consistent with the experimental results,

with similar relative energies obtained using either the B3LYP or

BLYP methods. The two major differences between this basis and

the LANL2DZ are the implementation of all-electron functions for

P and Cl and use of the RCEPs of Stevens et al. rather than the

more demanding ECP of Hay and Wadt. The RCEPs were devised

using an energy-based optimization scheme, distinct from the

orbital-optimization scheme used to develop the Hay-Wadt ECPs,

and comprise a relatively compact and efficient expansion of

Gaussians to model the pseudopotential.⁹⁶Suresh et al. also predict

energy barriers substantially lower than those of Wakatsuki et al.⁴¹

by means of B3LYP calculations, using the Hay-Wadt pseudo-

potential and basis functions for Rh, and 6-31++G(d,p) basis

functions for the main group atoms.⁴³The simpler PH₃ligands in

their study permit use of a larger basis set than is tractable for the

experimentally relevant ligands investigated in the present work.

Acknowledgment. D.B.G. and X.Z. thank the U.S. National

Science Foundation for partial support under Grant No. 0415783,

Dr. LeRoy Lafferty for help with advanced NMR experiments

at SDSU, and Dr. Ashok Krishnaswami of JEOL USA for

deuterium NMR experiments. A.L.C. thanks the NSF for partial

support under Grant No. CHE-0216563 and Prof. William

Richardson for helpful discussions.

Supporting Information Available: Molecular structure of

6c, CIF files for all structures, details of analysis of spectra of

H¹³C¹³CH isotopomers, tables of BDMM absolute energies and

optimized geometries for 2a-5d and transition states, tables for

computational methods comparison and benchmarks, and complete

ref 99. This material is available free of charge via the Internet at

http://pubs.acs.org.

Wakatsuki et al.’s previous analysis had indicated that the

reaction barriers for the unimolecular reaction were too high to

account for the observed reaction rates.⁴¹The present work found

a significant decrease in those predicted barrier heights at TS23

OM700355R

(100) Yanagisawa, S.; Tsuneda, T.; Hirao, K. J. Chem. Phys. 2000, 112,

545-553.

Article Doi

Experimental and computational study of the transformation of terminal alkynes to vinylidene ligands on trans-(chloro)bis(phosphine)Rh fragments and effects of phosphine substituents

DOI: 10.1021/om700355r

Source and publish data:

Authors:

Article abstract of DOI:10.1021/om700355r

Full text of DOI:10.1021/om700355r

Products guided by the article

R&D Labs maybe for 944834-68-8

Relevant to this article

Hot Product