Selectivities in hydrocarbon activation: Kinetic and thermodynamic investigations of reversible 1,2-RH-elimination from (silox)2((t)Bu3SiNH)TiR (silox = (t)Bu3SiO)

10696

J. Am. Chem. Soc. 1997, 119, 10696-10719

Selectivities in Hydrocarbon Activation: Kinetic and

Thermodynamic Investigations of Reversible

1,2-RH-Elimination from (silox)₂(^tBu₃SiNH)TiR (silox )

^tBu₃SiO)

Jordan L. Bennett and Peter T. Wolczanski*

Contribution from Baker Laboratory, Department of Chemistry, Cornell UniVersity,

Ithaca, New York 14853

ReceiVed March 7, 1997^X

Abstract: Addition of 2.0 equiv of Na(silox) to TiCl₄(THF)₂afforded (silox)₂TiCl₂(1), which yielded (silox)₂-

(^tBu₃SiNH)TiCl (2-Cl) upon treatment with ^tBu₃SiNLi. Grignard or alkyllithium additions to 2-Cl or 1,2-RH-addition

to transient (silox)₂TidNSi^tBu₃(3) produced (silox)₂(^tBu₃SiNH)TiR (2-R; R ) Me, Et, CH₂Ph ) Bz, CHdCH₂)

Vy, ^cBu, ⁿBu, Ph, H, ^cPr, ^cPe, CH₂-3,5-Me₂C₆H₃) Mes, ^neoHex, ^cHex, η³-H₂CHCH₂, η³-H₂CCHCHMe). Insertions

t

of C₂H₄, butadiene, HC₂H, and HC₂Bu into the titanium-hydride bond of 2-H generated (silox)₂(^tBu₃SiNH)TiR

(2-R; R ) Et, η³-H₂CCHCHMe, Vy, E-CHdCH^tBu). Trapping of 3 by donors L afforded (silox)₂LTidNSi^tBu₃

(3-L; L ) OEt₂, THF (X-ray, two independent molecules: d(TidN) ) 1.772(3), 1.783(3) Å), py, PMe₃, NMe₃,

t

NEt₃) and metallacycles (silox)₂(^tBu₃SiN)TiCRdCR′ (3-RC₂R′; RC₂R′ ) HC₂H, MeC₂Me, EtC₂Et, HC₂Bu) and

(silox)₂(^tBu₃SiN)TiCH₂CH₂(3-C₂H₄). Kinetics of 1,2-RH-elimination from 2-R revealed a first-order process (24.8

°C): R ) Bz < Mes < H < Me (1.54(10) × 10^-5s^-1) < ^neoHex < Et < ⁿBu < ^cBu < ^cPe < ^cHex < ^cPr < Vy <

Ph. Kinetics data, large 1,2-RH/D-elimination KIE’s (e.g., MeH/D, 13.7(9), 24.8 °C), and Eyring parameters (e.g.,

2-Me, ∆H^q) 20.2(12) kcal/mol, ∆S^q) -12(4) eu) portray a four-center, concerted transition state where the N‚‚‚H‚‚‚R

linkage is nearly linear. Equilibrium measurements led to the following relative standard free energy scale: 2-^cHex

> 2-^cPe > 2-ⁿPr ∼ 2-ⁿBu > 2-^neoHex > 2-Et > 2-^cBu > 2-CH₂SiMe₃> 2-Ph > 2-Me > 2-Bz > 2-^cPr ∼ 2-Mes >

2-Vy > 3-C₂H₄> 3-NEt₃> 2-H > 3-OEt₂> 3-EtC₂Et >3-MeC₂Me > 3-THF > 3-NMe₃> 3-PMe₃> 3-py. A

correlation of D(TiR)_relto D(RH) revealed greater differences in titanium-carbon bond energies. THF loss from

3-THF allowed a rough estimate of ∆G°(3). Using thermochemical cycles, relative activation energies for 1,2-RH-

c

n

c

addition were assessed: ^cHexH > PeH > BuH > ^neoHexH > EtH > BzH > BuH > MesH > MeH > PhH >

^cPrH > VyH > 3-C₂H₄formation > H₂. On the basis of a parabolic model, C-H bond activation selectivities are

influenced by the relative ground state energies of 2-R and a parameter representing the reaction coordinate. A

more compressed reaction coordinate for sp²- vs sp³-substrates eases their activation.

Introduction

tive additions of RH to coordinatively unsaturated d⁸and d⁶

metal centers (i.e., L_nM + RH h L_nHMR)^1-3,7-9and related

late-metal, solvent-assisted heterolytic RH cleavages (i.e., L_nM

+ RH h [L_nMR]^-+ H⁺)^10-12have dominated investigations

in this area, and surprising parallels to these electrophilic attacks

on RH have been discovered in d⁰early metal systems. σ-Bond

metatheses of hydrocarbons with coordinatively unsaturated

(4) (a) Wayland, B. B.; Ba, S.; Sherry, A. E. J. Am. Chem. Soc. 1991,

113, 5305-5311. (b) Zhang, X.-X.; Wayland, B. B. J. Am. Chem. Soc.

1994, 116, 7897-7898.

(5) Cook, G. K.; Mayer, J. M. J. Am. Chem. Soc. 1994, 116, 1855-

1868.

Despite greater than 15 years of research on carbon-hydrogen

bond activation,^1,2explanations of selectivities in systems

exhibiting concerted C-H bond-breaking events are still

somewhat speculative, because the energetics of critical ground

states are uncertain.³Understanding the factors that permit L_nM

or a metal-ligand functionality to react with R¹H vs R²H is

very important, because the ultimate utilization of C-H bond

reactivity depends crucially on whether the event can be made

selective.

Systems based upon H atom abstractions manifest selectivities

that essentially correlate inversely with the bond dissociation

energies of the substrate C-H bonds,^4-6but concerted systems

contrast markedly, exhibiting almost the opposite trend. Oxida-

(6) For related oxidation processes, see: Sheldon, R. A.; Kochi, J. K.

Metal Catalyzed Oxidations of Organic Compounds; Academic Press: New

York, 1981.

(7) For a classic study, see: (a) Jones, W. D.; Feher, F. J. Acc. Chem.

Res. 1989, 22, 99-106. (b) Jones, W. D.; Feher, F. J. J. Am. Chem. Soc.

1984, 106, 1650-1664.

^XAbstract published in AdVance ACS Abstracts, October 15, 1997.

(1) (a) Crabtree, R. H. In The Chemistry of Alkanes and Cycloalkanes;

Patai, S., Rappoport, Z., Eds.; John Wiley & Sons: New York, 1992; p

653. (b) Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T. H.

Acc. Chem. Res. 1995, 28, 154-162. (c) Davies, J. A., Ed. SelectiVe

Hydrocarbon ActiVation; VCH: Weinheim, Germany, 1990. (d) Crabtree,

R. H.; Hamilton, D. G. AdV. Organomet. Chem. 1988, 28, 299-338. (e)

Graham, W. A. G. J. Organomet. Chem. 1986, 300, 81-91. (f) Green, M.

L. H.; O’Hare, D. Pure Appl. Chem. 1985, 57, 1897-1910.

(8) For recent d⁶activations, see: (a) Burger, P.; Bergman, R. G. J. Am.

Chem. Soc. 1993, 115, 10462-10463. (b) Arndtsen, B. A.; Bergman, R.

G. Science 1995, 270, 1970-1973. (c) Luecke, H. F.; Arndtsen, B. A.;

Burger, P.; Bergman, R. G. J. Am. Chem. Soc. 1996, 118, 2517-2518. (d)

Holtcamp, M. W.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 1997,

119, 848-849. Calculational support for the oxidative addition pathway is

given in: (e) Strout, D. L.; Zaric, S.; Niu, S.; Hall, M. B. J. Am. Chem.

Soc. 1996, 118, 6068-6069.

(2) Crabtree, R. H. Chem. ReV. 1995, 95, 987-1007.

(9) For recent advances in catalytic dehydrogenation, see: (a) Maguire,

J. A.; Petrillo, A.; Goldman, A. S. J. Am. Chem. Soc. 1992, 114, 9492-

9498. (b) Gupta, M.; Hagen, C.; Kaska, W. C.; Cramer, R. E.; Jensen, C.

M. J. Am. Chem. Soc. 1997, 119, 840-841.

(3) Leading references to d⁸oxidative additions and a set of ground state

data crucial to understanding RH oxidative additions are included in: Jones,

W. D.; Hessell, E. T. J. Am. Chem. Soc. 1993, 115, 554-562.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10697

metal alkyl complexes (i.e., L_nMR + R′H h L_nMR′ + RH) of

the early metals,^13,14lanthanides,^15,16and actinides¹⁷show trends

in reactivity that crudely mirror their late-metal counterparts.

Concerted 1,2-RH-additions across MX (X ) O (R ) OH,

H,¹⁸R′′NH),¹⁹NR′ (R ) hydrocarbyl,^20-27R′′NH),^28,29CR′₂)^30,31

multiple bonds comprise a relatively recent class of C-H bond

cleavage reactions of which d⁰metal imido complexes³²

constitute the most prevalent subgroup. Hydrocarbon reactivity

is typified by 1,2-additions of RH to imido functionalities of

electrophilic transients Cp₂ZrdN^tBu (R ) arene,²⁶sp²-

substrates),²⁷(^tBu₃SiNH)₂ZrdNSi^tBu₃(R ) hydrocarbyl),^20,21

(^tBu₃SiNH)₂TidNSi^tBu₃(R ) benzene),²²(^tBu₃SiNH)Ta-

Scheme 1

(10) (a) Lin, M.; Sen, A. Nature 1994, 368, 613-614. (b) Kao, L.-C.;

Hutson, A. C.; Sen, A. J. Am. Chem. Soc. 1991, 113, 700-701. (c) Sen,

A.; Gretz, E.; Oliver, T. F.; Jiang, Z. New J. Chem. 1989, 13, 755-760.

(d) Sen, A. Acc. Chem. Res. 1988, 21, 421-428.

(11) (a) Luinstra, G. A.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem.

Soc. 1993, 115, 3004-3005. (b) Labinger, J. A.; Herring, A. M.; Lyon, D.

K.; Luinstra, G. A.; Bercaw, J. E.; Horva´th, I. T.; Eller, K. Organometallics

1993, 12, 895-905.

(12) Periana, R. A.; Taube, D. J.; Evitt, E. R.; Loffler, D. G.; Wentrcek,

P. R.; Voss, G.; Masuda, T. Science 1993, 259, 340-343.

(13) Guram, A. S.; Jordan, R. F.; Taylor, D. F. J. Am. Chem. Soc. 1991,

113, 1833-1835 and references therein

(14) (a) Quignard, F.; Le´cuyer, C.; Choplin, A.; Olivier, D.; Bassett, J.-

M. J. Mol. Catal. 1992, 74, 353-363. (b) Le´cuyer, C.; Quignard, F.;

Choplin, A.; Olivier, D.; Bassett, J.-M. Angew. Chem., Int. Ed. Engl. 1991,

30, 1660-1661. (c) Quignard, F.; Choplin, A.; Bassett, J.-M. J. Chem. Soc.,

Chem. Commun. 1991, 1589-1590.

(15) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.; Nolan,

M. C.; Santarsiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am. Chem.

Soc. 1987, 109, 203-219.

(16) Watson, P. L.; Parshall, G. W. Acc. Chem. Res. 1985, 18, 51-56.

(17) Fendrick, C. M.; Marks, T. J. J. Am. Chem. Soc. 1986, 108, 425-

437.

(dNSi^tBu₃)₂(R

)

methane, benzene),²³(^tBu₃SiNH)V-

(dNSi^tBu₃)₂(R ) hydrocarbyl),²⁵and (silox)₂TidNSi^tBu₃(3,

(18) Parkin, G.; Bercaw, J. E. J. Am. Chem. Soc. 1989, 111, 391-393.

(19) Howard, W. A.; Waters, M.; Parkin, G. J. Am. Chem. Soc. 1993,

115, 4917-4918.

t

silox ) Bu₃SiO, R ) hydrocarbyl).²⁴During the course of

seeking a stable 3-coordinate imido derivative, investigations

of the last were initiated. While transient 3 remains elusive,

the corresponding alkyl species, (silox)₂(^tBu₃SiNH)TiR (2-R),

undergo facile 1,2-RH-eliminations at 24.8 °C and are amenable

to equilibrium studies that establish relative ground state

energies.

(20) (a) Schaller, C. P.; Cummins, C. C.; Wolczanski, P. T. J. Am. Chem.

Soc. 1996, 118, 591-611. (b) Cummins, C. C.; Baxter, S. M.; Wolczanski,

P. T. J. Am. Chem. Soc. 1988, 110, 8731-8733.

(21) Schaller, C. P.; Bonanno, J. B.; Wolczanski, P. T. J. Am. Chem.

Soc. 1994, 116, 4133-4134.

(22) Cummins, C. C.; Schaller, C. P.; Van Duyne, G. D.; Wolczanski,

P. T.; Chan, E. A.-W.; Hoffmann, R. J. Am. Chem. Soc. 1991, 113, 2985-

2994.

Scheme 1 illustrates the general scope of reactivity reported

herein, showing 1,2-RH-elimination and -addition pathways and

the trapping of (silox)₂TidNSi^tBu₃(3) by Lewis bases and

alkynes. Equilibrium studies are reported that link all 2-Rⁿ,

3-L, and 3-RC₂R′, and kinetic investigations of 1,2-RH-

elimination from 2-R enable selectivities of 1,2-R¹H- vs 1,2-

(23) Schaller, C. P.; Wolczanski, P. T. Inorg. Chem. 1993, 32, 131-

144.

(24) Bennett, J. L.; Wolczanski, P. T. J. Am. Chem. Soc. 1994, 116,

2179-2180. Included herein is an estimate of the enthalpy for binding

ethylene (∆H_σ-bind∼ 11 kcal/mol, ∆G_σ-bind∼9 kcal/mol) to d⁰(silox)₂-

TidNSi^tBu₃(3).

(25) de With, J.; Horton, A. D. Angew. Chem., Int. Ed. Engl. 1993, 32,

903-905. Upon reassessment of the data in this paper, ∆S^q_elim∼ 8(23) eu,

a value lacking the precision necessary for interpretation: Bennett, J. L.;

Wolczanski, P. T. Unpublished work.

R²H-addition to be determined as ∆∆G^q

) ∆G^q_addn(R¹H)

addn

- ∆G^q_addn(R²H). The origin of these selectivities is presented

in terms of a model that describes reactant (i.e., 2-R) and product

(i.e., 3) free energy surfaces as intersecting parabolas, such that

ground state and positional dependencies can be delineated.

While certain trends are surprisingly similar to late-metal

oxidative additions and other concerted activations, the greater

selectivity exhibited by (silox)₂TidNSi^tBu₃(3) may be rational-

ized on the basis of a different rate-determining event.

(26) Walsh, P. J.; Hollander, F. J.; Bergman, R. G. Organometallics 1993,

12, 3705-3723.

(27) Lee, S. Y.; Bergman, R. G. J. Am. Chem. Soc. 1995, 117, 5877-

5878.

(28) Zambrano, C. H.; Profilet, R. D.; Hill, J. E.; Fanwick, P. E.;

Rothwell, I. P. Polyhedron 1993, 12, 689-708.

(29) Jolly, M.; Mitchell, J. P.; Gibson, V. C. J. Chem. Soc., Dalton Trans.

1992, 1329-1330.

(30) (a) van der Heijden, H.; Hessen, B. J. Chem. Soc., Chem. Commun.

1995, 145-146. (b) Coles, M. P.; Gibson, V. C.; Clegg, W.; Elsegood, M.

R. J.; Porrelli, P. A. J. Chem. Soc., Chem. Commun. 1996, 1963-1964. (c)

Tran, E.; Legzdins, P. J. Am. Chem. Soc. 1997, 119, 5071-5072.

(31) (a) Lockwood, M. A.; Clark, J. R.; Parkin, B. C.; Rothwell, I. P. J.

Chem. Soc., Chem. Commun. 1996, 1973-1974. (b) Chamberlain, L. R.;

Rothwell, I. P.; Huffmann, J. C. J. Am. Chem. Soc. 1986, 108, 1502-1509.

(c) McDade, C.; Green, J. C.; Bercaw, J. E. Organometallics 1982, 1, 1629-

1634. (d) Couturier, J.-L.; Paillet, C.; Leconte, M.; Basset, J. M.; Weiss,

K. Angew. Chem., Int. Ed. Engl. 1992, 31, 628-631. (e) Polse, J. L.;

Andersen, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1996, 118, 8737-

8738.

(32) (a) Wigley, D. E. Prog. Inorg. Chem. 1994, 42, 239-482. (b)

Wigley, D. E.; Gray, S. D. In ComprehensiVe Organometallic Chemistry

II; Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.; Pergamon Press:

Oxford, U.K., 1995, Vol. 5, pp 57-153. (c) Nugent, W. A.; Mayer, J. M.

Metal-Ligand Multiple Bonds; Wiley-Interscience: New York, 1988.

Results and Discussion

Syntheses and Spectral Characterizations. 1. (silox)₂-

(^tBu₃SiNH)TiR (R ) Cl (2-Cl), H (2-H), hydrocarbyl (2-

R)). Treatment of TiCl₄(THF)₂with 2 equiv of Na(silox) in

diethyl ether for 4 h at 25 °C afforded colorless, crystalline

(silox)₂TiCl₂(1,³³silox ) ^tBu₃SiO) in 90% yield upon recrys-

tallization from hexanes (eq 1). The addition of LiNHSi^tBu₃

to 1 in ether at 25 °C for 30 min yielded colorless (silox)₂-

(^tBu₃SiNH)TiCl (2-Cl) in 87% yield (eq 2), also after crystal-

lization from hexanes.

(33) Covert, K. J. Ph.D. Thesis, Cornell University, 1991.

10698 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

Et₂O

at δ 189.01 (Ti-CHd) and δ 129.44 (dCH₂), while a ν(CdC)

(silox) TiCl

TiCl₄(THF)₂+ 2Na(silox) _{25 °C, 4 h}8

₂+ 2NaCl

(1)

was observed in the IR spectrum at 1554 cm^-1

.

2

1

Chloride 2-Cl reacted within 5 min with cyclobutyllithium

in diethyl ether at -30 °C, generating off-white (silox)₂(^tBu₃-

SiNH)Ti^cBu (2-^cBu) upon crystallization from hexanes. Simi-

larly, (silox)₂(^tBu₃SiNH)TiⁿBu (2-ⁿBu) was synthesized by

Et₂O

1 + ^tBu₃SiNHLi

8

+ LiCl

t

(silox)₂( Bu₃SiNH)TiCl

25 °C, 30 min

2-Cl

n

addition of BuLi to 2-Cl in diethyl ether at 0 °C (eq 5). The

(2)

Et₂O

Three distinct routes led to the preparation of (silox)₂(^tBu₃-

SiNH)TiR (2-R, R ) H, hydrocarbyl): metathesis of (silox)₂-

(^tBu₃SiNH)TiCl (2-Cl) with appropriate alkyl anion equivalents,

R-H bond activation by the thermally generated transient

(silox)₂TidNSi^tBu₃(3), and olefin insertion into the Ti-H bond

of (silox)₂(^tBu₃SiNH)TiH (2-H). The extreme solubilities of

certain hydrocarbyls hampered isolation and purification, and

the yields of most derivatives were not optimized unless copious

quantities of the complex were necessary for ensuing studies.

¹H NMR spectra of 2-R revealed a 27 H singlet for the ^tBu₃SiNH

ligand that was typically slightly upfield of the 54 H silox

resonance, a broader 1 H amide signal that varied with

substituent, and the spectral signature of the hydrocarbyl;

¹³C{¹H} NMR spectra possessed related features (Table 1). The

NH chemical shifts (C₆D₆) of 2-Cl and 2-R essentially parallel

those of (^tBu₃SiNH)₃ZrX (X ) Cl, hydrocarbyl),²⁰roughly

correlating with the electronegativity of the substituent.²²

The reaction of (silox)₂(^tBu₃SiNH)TiCl (2-Cl) with Grignard

reagents generally proceeded smoothly, provided the subsequent

elimination of RH from 2-R was slow (k_R< 1 × 10^-4s^-1),

permitting isolation of a reasonable amount of 2-R. In some

cases, alkyllithium reagents were required, although these

reactions were often not as clean as the reactions of 2-Cl with

RMgX. Crude alkyl derivatives were typically >80% pure, and

these often sufficed for kinetics studies.

2-Cl + RLi _{-30 °C f 0 °C, 5 min}8

(silox)₂(^tBu₃SiNH)TiR

+ LiCl (5)

2-R

c

R: Bu, 35% (90% purity);

ⁿBu, 55% (90% purity)

alkyl 2-ⁿBu could not be induced to crystallize and was isolated

as a precipitate upon filtration and removal of the volatiles.

Metathesis was also used to obtain the deuterated amide phenyl

complex (silox)₂(^tBu₃SiND)TiPh (2-(ND)-Ph), which was con-

veniently isolated from hexanes in 39% yield after treatment

of 2-(ND)-Cl with phenyllithium in ether at 0 °C (eq 6). Neither

Et₂O

(silox) (^tBu SiND)TiCl

+ C₆H₅Li _{0 °C, 5 min}8

2

3

2-(ND)-Cl

(silox)₂(^tBu₃SiND)TiC₆H

₅+ LiCl (6)

2-(ND)-Ph

^tBuLi nor BuCH₂Li successfully alkylated 2-Cl, and while

metathesis was observed with Me₃SiCH₂Li, byproducts seriously

complicated subsequent isolation efforts.

t

Direct 1,2-RH-addition to transient (silox)₂TidNSi^tBu₃(3)

was an attractive synthetic alternative for 2-R species whose

subsequent 1,2-RH-elimination was swift and in cases where a

suitable RMgX or RLi was not readily available. The methyl

complex (silox)₂(^tBu₃SiNH)TiMe (2-Me) was typically used in

the thermal generation of 3 because its synthesis was optimized

(reproducibly >75%) and byproduct CH₄could easily be

removed as a competitive substrate via degassing. In a typical

experiment, 2-Me was dissolved in the neat hydrocarbon

substrate or in a cyclohexane solution containing the desired

hydrocarbon. The reaction solution was sealed in a glass bomb

and heated at 55-80 °C for 3-24 h (eqs 7 and 8). At intervals,

Treatment of 2-Cl in diethyl ether with excess (1.5 equiv)

MeMgBr, EtMgCl, or PhCH₂MgCl for ∼1 h at 25 °C produced

the corresponding alkyl amides, (silox)₂(^tBu₃SiNH)TiR (2-R,

R ) Me, Et, Bz ) PhCH₂) after swift workup from hexanes

(eq 3). Alkyls 2-Me and 2-Et were colorless crystalline solids

Et₂O

_{25 °C, 1 h}8

t

RMgX

+

(silox)₂( Bu₃SiNH)TiCl

X ) Br, R ) Me

2-Cl

X ) Cl, R ) Et, Bz

(silox) (^tBu SiNH)TiR

+ MgXCl (3)

2

3

2-R

RH/C₆H₁₂

_{55-80 °C, 3-24 h}8

(silox)₂(^tBu₃SiNH)TiMe

R: Me, 76%; Et, 51%;

Bz, 72%

2-Me

t

+ CH (7)

[(silox)₂TidNSi Bu₃]

4

while 2-Bz was bright yellow. Initial attempts to synthesize

2-Vy (Vy ) CHdCH₂) from 2-Cl and VyMgBr in Et₂O gave

products derived from ethylene loss, prompting a switch to

vinyllithium and lower temperatures. Freshly prepared, recrys-

tallized vinyllithium was combined with 2-Cl in diethyl ether,

and the mixture was stirred at 0 °C for 15 min, providing

(silox)₂(^tBu₃SiNH)TiCHdCH₂(2-Vy) as an off-white powder

upon workup in hexanes (eq 4). 2-Vy exhibited a classic ABX

3

(silox) (^tBu SiNH)TiR

3 f

(8)

2

3

2-R

R: H, 67%; ^cPr, 33%;

^cPe, 43%; Mes, 42%;

^neoHex, 20%; Ph, 69%

Et₂O

the reaction mixture was cooled and then frozen at 77 K and

methane was removed by vacuum. Upon completion of the

reaction, the volatiles were removed and the product was isolated

by crystallization from hexanes. A synthesis of the hy-

dride,^24,34,35(silox)₂(^tBu₃SiNH)TiH (2-H, δ(TiH) ) 7.34, ν(TiH/

D) ) 1645/1180 cm^-1), was effected through this procedure,

as were preparations of several hydrocarbyls, (silox)₂(^tBu₃-

2-Cl + CH₂dCHLi _{0 °C, 15 min}8

(silox)₂(^tBu₃SiNH)TiCHdCH₂

+ LiCl (4)

2-Vy, 30%

pattern with coupling constants consistent with a vinyl deriva-

tive: J_cis) 14.2 Hz, J_trans) 19.2 Hz, J_gem) 2.8 Hz. ¹³C NMR

(C₆D₁₂) spectroscopy corroborated the assignment, showing

resonances

c

t

SiNH)TiR (2-R, R ) Pr, Pe, ^neoHex (^neoHex ) CH₂CH₂Bu),

Mes (Mes ) CH₂-3,5-Me₂C₆H₃), Ph). The hydrocarbyls were

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10699

Table 1. ¹H and ¹³C{¹H} NMR Spectral Data for (^tBu₃SiO)₂(^tBu₃SiNH)TiR (2-R), (^tBu₃SiO)₂LTidNSi^tBu₃(3-L),

(^tBu₃SiO)₂TiC(R)dC(R)NSi^tBu₃(3-RCCR), and Related Complexes in C₆D₆Unless Otherwise Noted^a,b

¹H NMR, δ(assgnt, mult, J (Hz))

¹³C{1H} NMR, δ (assgnt, J (Hz))

c

compound

((H₃C)₃C)₃

NH

R/L

C(CH₃)₃

R/L

(silox)₂TiCl₂(1)

1.22

1.30

1.28

1.29

1.27

1.29

1.27

1.30

1.28

1.36

1.34

29.86

30.81

31.12

30.70

24.14

23.99

23.65

23.35

30.99

23.75

23.41

23.71

23.31

24.20

23.84

(silox)(^tBu₃SiNH)TiCl (2-Cl)

(silox)₂(^tBu₃SiNH)TiH (2-H)

(silox)(^tBu₃SiNH)TiMe (2-Me)

(silox)(^tBu₃SiNH)TiEt (2-Et)

(silox)₂(^tBu₃SiNH)TiCHdCH₂(2-Vy)^b

9.01

7.34

7.39

7.21

7.96

8.64 (TiH, s)

1.35 (CH₃, s)

22.87

47.30 (CH₃)

30.78

31.12

30.79

31.14

31.11

31.43

1.74 (CH₃, t, 7.4)

67.02 (CH₂)

18.81 (CH₃)

129.44 (CH₂)

189.01 (TiC)

1.98 (CH₂, q, 7.4)

6.10 (CH_c, dd, 19, 3)

6.18 (CH_t, dd, 14, 3)

7.73 (CH, dd, 19, 14)

0.65-0.75 (CH_â, m)

1.4-1.6 (TiCH, m)

0.8-1.0 (CH₃, m)

2.0-2.3 (CH₂, m)

(silox)₂(^tBu₃SiNH)Ti-c-C₃H₅(2-^cPr)^d

(silox)₂(^tBu₃SiNH)TiC₄H₉(2-ⁿBu)^e

1.30

1.28

1.31

1.29

7.20

7.24

30.82

31.15

30.78

31.14

23.73

23.38

23.74

23.34

13.39 (C_â)

57.87 (C_R)

13.63 (CH₃)

27.94 (CH₂)

36.10 (CH₂)

74.46 (TiC)

22.20 (C_γ)

36.42 (C_â)

85.80 (C_R)

(silox)₂(^tBu₃SiNH)Ti-c-C₄H₇(2-^cBu)

1.31

1.28

7.10

1.97 (“quin”, 9.2)

2.19 (m)

2.55 (m)

30.80

31.14

23.64

23.23

3.02 (“quin”, 8.9)

3.27 (“quin”, 9.2)

0.29 (CH₃, s)

2.04 (CH₂, s)

0.75-1.05 (m, 2H)

1.55-2.60 (m, 7H)

(silox)₂(^tBu₃SiNH)TiCH₂SiMe₃(2-CH₂SiMe₃)

(silox)₂(^tBu₃SiNH)Ti-c-C₅H₉(2-^cPe)

1.31

1.29

1.32

1.30

7.06

7.10

30.87

31.22

23.70

23.31

27.46 (C_γ)

36.78 (C_â)

89.47 (C_R)

(silox)₂(^tBu₃SiNH)Ti(CH₂)₂Bu (2-^neoHex)

1.32

1.30

7.24

0.91 (CH₃, s)

2.11 (Ti(CH₂)₂, m)

30.79

31.14

23.77

23.37

28.98 ((CH₃)₃)

33.12 (CMe₃)

47.74 (CH₂)

70.64 (TiC)

t

(silox)₂(^tBu₃SiNH)Ti-c-C₆H₁₁(2-^cHex)^a,f

1.23

1.20

6.93

7.63

1.6-2.0 (CH, CH₂, m)

2.60 (CH₂, dm, 11)

1.01 (CH₃, s)

(silox)₂(^tBu₃SiNH)TiCHdCH^tBu (2-E-CHdCH^tBu)^g

6.69 (CH, d, 17.8)

7.42 (TiCH, d, 17.8)

7.0-7.2 (Ph_m,p, m)

8.2 (Ph_o, “dt”, 7.8, 1.6)

(silox)₂(^tBu₃SiNH)TiPh (2-Ph)^b

1.30

8.27

31.11

31.45

24.22

23.82

126.84 (Ph)

134.74 (Ph)

136.83 (Ph)

187.74 (C_ipso

128.68 (Ar)

191.70 (C_ipso

73.81 (CH₂)

123.34 (Ph)

127.92 (Ph)

128.49 (Ph)

148.76 (C_ipso

21.43 (CH₃)

74.59 (TiC)

125.25 (C_p)

125.81 (C_o)

137.48 (C_m)

148.65 (C_ipso

)

[(silox)₂(^tBu₃SiNH)Ti]₂(µ-η¹,η¹-1,4-C₆H₄) (2₂-C₆H₄)^b,h

(silox)(^tBu₃SiNH)TiCH₂Ph (2-Bz)

31.10

31.40

30.79

31.14

24.22

23.82

23.71

23.31

)

1.26

7.75

7.65

3.30 (CH₂, s)

6.8-7.4 (Ph, m)

)

(silox)₂(^tBu₃SiNH)TiCH₂C₆H₃-3,5-Me₂(2-Mes)

1.27

2.23 (CH₃, s)

3.31 (CH₂, s)

6.55 (Ar H_p, s)

6.96 (Ar H_o, s)

30.76

31.17

23.73

23.35

[silox)₂(^tBu₃SiN)Ti]₂(µ₂:η¹,η¹-1,3-(CH₂)₂C₆H₃-5-Me)

(2₂-C₆H₃(CH₂)₂Me)^a,i

2.05 (CH₃, s)

2.96 (CH₂, s)

(silox)₂(^tBu₃SiNH)Ti(η³-H₂CCHCH₂)

(2-η³-H₂CCHCH₂)^a,b,j

1.22

1.19

7.53

7.36

1.57 (H_syn, dd, 6, 1.5)

4.90 (CH, “tm”, 18)

7.25 (H_anti, dd, 18, 1.5)

2.60 (CH₃, d, 8)

5.03 (CH, “sext”, 10)

5.77 (CH, br m)

115.93 (CH₂)

133.43 (CH)

(silox)₂(^tBu₃SiNH)Ti(η³-H₂CCHCHMe)

(2-η³-H₂CCHCHMe)^a,b

1.23

1.20

6.05 (CH, “sext”, 10)

[(silox)₂(^tBu₃SiN)Ti]₂(3₂)

1.19

1.18

1.35

1.36

1.35

1.33

1.22

1.16

31.26

31.58

23.51

23.86

23.52

23.89

24.03

24.35

24.01

(silox)₂(THF)TidNSi^tBu₃(3-THF)

(silox)₂(Et₂O)TidNSi^tBu₃(3-OEt₂)

(silox)₂(Me₃N)TidNSi^tBu₃(3-NMe₃)^a,b

1.08 (R-CH₂, t, 6.6)

4.24 (â-CH₂, t, 6.6)

0.88 (CH₃, t, 6.9)

4.16 (CH₂, q, 6.9)

2.96 (CH₃, s)

31.09

31.68

31.12

31.65

31.53

31.98

25.25 (â-CH₂)

78.04 (R-CH₂)

12.85 (CH₃)

70.93 (CH₂)

52.90 (CH₃)

10700 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Table 1 (Continued)

Bennett and Wolczanski

¹H NMR, δ(assgnt, mult, J (Hz))

¹³C{1H} NMR, δ (assgnt, J (Hz))

c

compound

((H₃C)₃C)₃

NH

R/L

C(CH₃)₃

R/L

(silox)₂(Et₃N)TidNSi^tBu₃(3-NEt₃)

1.32

1.29

1.34

1.38

0.95 (CH₃, t, 7.0)

4.14 (CH₂, q, 7.0)

6.48 (py_m, m)

31.35

31.85

31.08

31.70

23.74

24.11

23.64

24.08

13.64 (CH₃)

49.97 (CH₂)

124.58 (py)

140.49 (py)

(silox)₂(py)TidNSi^tBu₃(3-py)

6.70 (py_p, m)

9.15 (py_o, m)

1.16 (CH₃, d, J_PH) 7.8)

151.15 (py)

14.76 (CH₃, J_PC) 19.8)

(silox)₂(Me₃P)TidNSi^tBu₃(3-PMe₃)

(silox)₂(^tBu₃SiN)TiCH₂CH₂(3-C₂H₄)^k

(silox)₂(^tBu₃SiN)TiCHdCH (3-HC₂H)^b

(silox)₂(^tBu₃SiN)TiCMedCMe (3-MeC₂Me)^l

1.32

1.35

31.39

31.99

23.85

24.22

1.26

1.23

3.08 (CH₂, br s)

30.99

31.88

24.02

24.58

50.04 (CH₂, J_CH) 146.5)

1.25

1.20

7.80 (TiCH, d, 11.1)

10.65 (NCH, d, 11.1)

30.93

31.49

23.94

24.13

153.65 (NC)

198.31 (TiC)

1.24

2.24 (CH₃, s)

2.48 (CH₃, s)

30.79

31.65

23.48

24.5

21.67 (CH₃)

22.30 (CH₃)

162.05 (NC)

216.99 (TiC)

(silox)₂(^tBu₃SiN)TiCEtdCEt (3-EtC₂Et)^a,b,l,m

(silox)₂(^tBu₃SiN)TiC^tBudCH (3-^tBuC₂H)^b,m

1.23

1.26

2.33 (CH₂, q, 7.5)

2.89 (CH₂, q, 7.5)

31.09

32.0

23.92

25.3

166.53 (NC)

223.69 (TiC)

1.26

1.18 (CH₃, s)

10.72 (NCH, s)

31.03

31.59

23.78

24.21

152.27 (NC)

225.50 (TiC)

a

b

t

¹H NMR in C₆D₁₂solvent. ¹³C{¹H} NMR in C₆D₁₂solvent. ^cFirst resonance due to (silox)₂; second due to BuSiNH or Bu₃SiN. ^dOther

â-CH resonance obscured. ^eOther CH₂resonances obscured. ^fRemaining H’s obscured by Bu resonances; low solubility and swift 1,2-^cHexH-

elimination rate prevented observation by ¹³C{¹H} NMR. ^{g t}Bu resonance obscured by those of 2-H and 3-^tBuC₂H. ^hObserved with 2-Ph; solubility

too low to enable confident ¹H NMR ^tBu, etc. resonance assignments. ⁱObserved with 2-Mes; solubility too low to enable confident assignment of

^tBu, etc. resonances, which are assumed to be obscured by 2-Mes. ^jRemaining ¹³C{¹H} resonances could not be differentiated from those of more

t

soluble trace impurities. ^kMethylene resonance broad due to fluxionality; for description of -130 to 20 °C data in C₇D₁₄, see text. ^lImide Bu

t

carbons very broad. ^mOther carbon assignments ambiguous with regard to trace impurities.

colorless, crystalline complexes, except for 2-Mes and 2-Ph,

which were yellow.

In the case of the phenyl derivative, extended manipulation

in cyclohexane led to the formation of a yellow insoluble

material tentatively formulated as [(silox)₂(^tBu₃SiNH)Ti]₂(µ-

η¹,η¹-1,4-C₆H₄) (2₂-C₆H₄), the product of double C-H bond

complexes. In these cases, steric hindrance to double activation

activation of benzene (eq 9). A dinuclear species was antici-

is minimal, because the metals are either oppositely disposed

about the substrate (e.g., 1,4-C₆H₄, 1,3-c-C₄H₆) or linked by a

spacer of g3 atoms.

Considerable effort was expended toward the synthesis of

the cyclohexyl complex, (silox)₂(^tBu₃SiNH)TiC₆H₁₁(2-^cHex),

considered the least thermodynamically stable yet observable

alkyl. Thermolysis of (silox)₂(^tBu₃SiNH)TiMe (2-Me) in neat

pated from precedents set in previous zirconium and tantalum

cyclohexane for 7 d at 55 °C afforded a yellow, microcrystalline

(i.e., [(^tBu₃SiNH)₃Zr]₂(µ-η¹,η¹-1,4-C₆H₄),²⁰[(^tBu₃SiNH)₂-

material that proved to be a mixture of (silox)₂(^tBu₃SiNH)-

(^tBu₃SiN)Ta]₂(µ-η¹,η¹-1,4-C₆H₄))²³systems. As a testament to

the more sterically accessible pocket of transient (silox)₂-

TiC₆H₁₁(2-^cHex) and [(silox)₂Ti]₂(µ-NSi^tBu₃)₂(3₂), formulated

as a dimer on steric grounds (eq 11). The composition was

TidNSi^tBu₃(3), evidence of additional double activations was

obtained. When 2-Mes was allowed to stand in C₆D₁₂, ¹H NMR

spectra gave evidence for [(silox)₂(^tBu₃SiN)Ti]₂(µ-η¹,η¹-1,3-

(CH₂)₂C₆H₃-5-Me) (2₂-C₆H₃(CH₂)₂Me) and free mesitylene (eq

10). In the synthesis of 2-^cBu (eq 5), part of the isolation and

purification problems can be attributed to formation of a 1,3-

cyclobutanediyl derivative. ¹H NMR spectra of crude product

mixtures in attempted syntheses of 2-CH₂SiMe₃and 2-^cBu

provided hints of the formation of sparingly soluble dinuclear

C₆H₁₂

(silox)₂(^tBu₃SiNH)TiMe (2-Me) _{55 °C, 7 d}8

(silox)₂(^tBu₃SiNH)TiC₆H₁₁(2-^cHex) +

[(silox)₂Ti]₂(µ-NSi^tBu₃)₂(3₂) + CH₄(11)

difficult to assay because of the low solubilities of the species,

especially 3₂, but was estimated to be approximately 1:1 by

NMR spectroscopy. The low solubility of the mixture limited

its utility as a starting material, but solutions enriched in 2-^cHex

were generated and used to determine the 1,2-^cHexH-elimination

rate constant.

(34) Discrete, terminal titanium hydrides are relatively uncommon.

See: (a) No¨th, H.; Schmidt, M. Organometallics 1995, 14, 4601-4610.

(b) You, Y.; Wilson, S. R.; Girolami, G. S. Organometallics 1994, 13,

4655-4657. (c) Cummins, C. C.; Schrock, R. R.; Davies, W. M.

Organometallics 1992, 11, 1452-1545. (d) Frerichs, S. R.; Kelsey Stein,

B.; Ellis, J. E. J. Am. Chem. Soc. 1987, 109, 5558-5560. (e) Pattiasina, J.

W.; van Bolhuis, F.; Teuben, J. H. Angew. Chem., Int. Ed. Engl. 1987, 26,

330-331. (f) Aitken, C. T.; Harrod, J. F.; Samuel, E. J. Am. Chem. Soc.

1984, 106, 1859-1860. (g) Spaltenstein, E.; Palma, P.; Kreutzer, K. A.;

Willoughby, C. A.; Davis, W. M.; Buchwald, S. L. J. Am. Chem. Soc. 1994,

116, 10308-10309.

Derivatization of (silox)₂(^tBu₃SiNH)TiH (2-H) was accom-

plished via insertion of ethylene into its titanium-hydride bond

(eq 12), but this reaction was not amenable to a preparative

scale because it required several hours, allowing 1,2-H₂-

elimination to compete. For larger substrates, H₂loss from 2-H

was observed, leading to allyl formation via C-H bond

(35) Bercaw, J. E.; Marvich, R. H.; Bell, L. G.; Brintzinger, H. H. J.

Am. Chem. Soc. 1972, 94, 1219-1238.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10701

insert into 2-H to form (silox)₂(^tBu₃SiN)Ti(E-CHdCH^tBu) (2-

C₆H₁₂

t

+ CH dCH

8

(silox)₂( Bu₃SiNH)TiH

t

E-CHdCH^tBu) but subsequent elimination of BuCHdCH₂

2

25 °C

2-H

t

followed by [2+2] addition of BuCtCH to the imide of 3

(silox)₂(^tBu₃SiNH)TiCH₂CH₃

(12)

rendered this an unattractive synthetic pathway (eqs 15-17).

2-Et

C₆D₁₂

t

+ RCtCH

8

(silox)₂( Bu₃SiNH)TiH

activation by (silox)₂TidNSi^tBu₃(3). The reactions of 2-H with

propene and either cis- or trans-2-butene afforded respective

yellow η³-allyl and red-orange η³-1-methylallyl complexes

according to eq 13. Despite being precipitated in nearly

25 °C

2-H

(silox)₂(^tBu₃SiNH)TiCHdCHR

R ) H (2-Vy), ^tBu (2-E-CHdCH^tBu)

(15)

C₆H₁₂

C₆D₁₂

_{25 °C}8

(silox)₂(^tBu₃SiNH)TiCHdCHR

2-H + H₃CCHdCHR

8

25 °C

R ) H (2-Vy), ^tBu (2-E-CHdCH^tBu)

(silox) (^tBu SiNH)Ti(η³-H CCHCHR)

+ H₂(13)

2

3

2

t

R ) H (2-η³-H₂CCHCH₂)

+ H CdCHR (16)

[(silox)₂TidNSi Bu₃]

2

3

R ) Me (2-η³-H₂CCHCHMe)

quantitative yield, these derivatives were extremely difficult to

characterize due to their limited solubility. In 2-η³-H₂CCHCH₂,

separate anti (δ 7.25, J_anti)18, J_gem) 1.5 Hz) and syn (δ 1.57,

J_syn) 6, J_gem) 1.5 Hz) resonances distinct from the central

proton at δ 4.90 (m, J_anti) 18 Hz) are assigned on the basis of

coupling magnitude, but must be considered tentative, since the

chemical shifts are opposite that observed in typical static, dⁿ

(n g 0) systems.³⁶This oddity may be a consequence of the

d⁰titanium center. Its ¹³C{¹H} NMR spectrum revealed two

resonances at δ 133.43 (CH) and δ 115.93 (CH₂), consistent

with the expected C_ssymmetry. The extremely low solubility

A crude estimate of the rate constant, k_{Vy- Bu}, for H₂CdCH^tBu

loss from 2-E-CHdCH^tBu is 7(4) × 10^-4s^-1, roughly 3 times

faster than ethylene loss from 2-Vy.

t

2. Adducts (silox)₂LTidNSi^tBu₃(3-L). Adducts (silox)₂-

LTidNSi^tBu₃(3-L, L ) OEt₂, THF, NMe₃, NEt₃, py, PMe₃)

were synthesized by allowing any 2-R species to stand in the

appropriate donor solvent or solvent mixture for longer than 5

half-lives for RH elimination. In the case of ethereal adducts,

stirring 2-R, typically 2-Me, in neat solvent afforded crude 3-L

(L ) OEt₂(3-OEt₂), THF (3-THF)), which could be purified

by recrystallization from hexanes (eq 18). Adducts of pyridine,

1

of 2-η³-H₂CCHCHMe prevented corroboration based on H

NMR assignments of the η³-H₂CCHCHMe ligand, and its

stereochemistry could not be unambiguously determined. A

related η³-2-methylallyl species formed from 2-H and iso-

butylene could not be spectrally (¹H and ¹³C{¹H} NMR)

discerned from trace impurites and was not pursued. Insertion

of butadiene into the Ti-H bond of 2-H provided an inde-

pendent synthesis of 2-η³-H₂CCHCHMe (eq 14). Formation

L

(silox)₂(^tBu₃SiNH)TiR

_{25 °C}8

2-R

(silox)₂LTidNSi^tBu₃

+ RH (18)

C₆H₁₂

2-H + H₂CdCHCHdCH₂

8

3-L

25 °C

L: OEt₂, 69%; THF, 40%

(silox)₂(^tBu₃SiNH)Ti(η³-H₂CCHCHMe)

2-η³-H₂CCHCHMe

(14)

tertiary amines, and PMe₃were synthesized in the same manner,

except that a hexanes/L mixture was used as solvent (eq 19).

of the allyls occurred at a rate qualitatively similar to that for

elimination of H₂from 2-H without olefin present, thereby

rendering assistance of H₂loss by olefin precomplexation

unlikely.

In an NMR-tube reaction, acetylene inserted into the Ti-H

bond of 2-H to provide 2-Vy, but this route was not viable

because subsequent ethylene elimination proved to be swift, and

L, hexanes or C₆D₁₂

(silox)₂(^tBu₃SiNH)TiR

8

25 °C

2-R

(silox)₂LTidNSi^tBu₃

+ RH (19)

3-L

L: py, 66%; PMe₃, 48%;

NMe₃; NEt₃

azametallacyclobutene^26,37-42(silox)₂(^tBu₃SiN)TiCHdCH (3-

t

HC₂H) was produced. Likewise, BuCtCH was observed to

Donor adducts 3-L were various shades of yellow and were

extremely soluble in hydrocarbon solvents, which hampered

their isolation. Trialkylamine adducts 3-NR₃(R ) Me, Et) were

only prepared (>90%) on an NMR (Table 1) tube scale in

C₆D₁₂, and 3-NEt₃was markedly less soluble than the other

3-L compounds, perhaps because the ethyls of NEt₃interlock

(36) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles

and Applications of Organotransition Metal Chemistry; University Science

Books: Mill Valley, CA, 1987.

(37) Walsh, P. J.; Hollander F. J.; Bergman, R. G. J. Am. Chem. Soc.

1988, 110, 8729-8731.

(38) Walsh, P. J.; Carney, M. J.; Bergman, R. G. J. Am. Chem. Soc.

1991, 113, 6343-6345.

t

with the Bu groups.

(39) (a) Walsh, P. J.; Baranger, A. M.; Bergman, R. G. J. Am. Chem.

Soc. 1992, 114, 1708-1719. (b) Baranger, A. M.; Walsh, P. J.; Bergman,

R. G. Ibid. 1993, 115, 2753-2763. (c) Meyer, K. E.; Walsh, P. J.; Bergman,

R. G. J. Am. Chem. Soc. 1995, 117, 974-985.

3. [2+2] Adducts (silox)₂(^tBu₃SiN)TiCRdCR′ (3-RC₂R′).

Azametallacyclobutenes^26,37-42were formed by allowing 2-R

to eliminate RH in the presence of alkynes, similar to the

synthesis of 3-L. Typically, 2-Me was allowed to stir in the

presence of 1-3 equiv of alkyne (HCtCH, MeCtCMe,

(40) de With, J.; Horton, A. D.; Orpen, A. G. Organometallics 1993,

12, 1493-1496.

(41) (a) McGrane, P. L.; Jensen, M.; Livinghouse, T. J. Am. Chem. Soc.

1992, 114, 5459-5460. (b) Bryan, J. C.; Burrell, A. K.; Miller, M. M.;

Smith, W. H.; Burns, C. J.; Sattelberger, A. P. Polyhedron 1993, 12, 1769-

1777.

(42) Doxsee, K. M.; Farahi, J. B.; Hope, H. J. Am. Chem. Soc. 1991,

113, 8889-8898.

10702 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

t

EtCtCEt, BuCtCH) in hexanes (eq 20). Only the 2-butyne

-130 °C, and line shape analysis of this fluxional process (-130

to +20 °C, ∆G^q(25 °C) ) 8.9(10) kcal/mol) revealed a

significant activation enthalpy of ∆H^q) 7.9(4) kcal/mol and

small activation entropy of ∆S^q) -3(2) eu. The activation

parameters were unchanged with 10 equiv of ethylene present,

and no involvement of the free olefin was spectroscopically

detected below 20 °C. The simplest dynamic process that

renders the methylenes equivalent is ethylene rotation, and

disruption of TidN(dπ/pπ)fCdC(pπ*) bonding, depicted in

the azametallacycle as formal Ti-C and C-N bonds, is incurred

regardless of pathway. Some TiO(dπ/pπ)fCdC(pπ*) back-

bonding could help mitigate olefin rotation, but the TiO

π-bonding orbitals are so low in energy with respect to the π*-

orbital of ethylene that significant stabilization is unlikely. Total

loss of π-interaction leaves the ethylene bound solely by the

corresponding CdC(π^b)fTi(dσ) σ-bond to the d⁰metal center,

and previous estimates place its binding enthalpy at ∼11 kcal/

mol and free energy at ∼9 kcal/mol.²⁴

[2+2] adduct was isolated in 55% yield upon crystallization

from hexanes, but all the adducts were generated numerous

times on an NMR tube scale and characterized by ¹H and

¹³C{¹H} NMR spectroscopy. These complexes appear to have

more rigidity and/or steric congestion than the other species as

evidenced by broad Bu resonances in their H and ¹³C{¹H}

NMR spectra; however, when 3-MeC₂Me was heated to 100

°C in the NMR spectrometer probe, no tendency toward

coalescence of the alkyne methyls was noted. The azametalla-

cyclobutenes are dark burgundy in color and were previously

noted as byproducts in the attempted insertion of alkynes into

2-H (eqs 15-17). Unlike related oxametallacyclobutenes

reported by Polse et al.,⁴³which were observed in equilibrium

with hydroxide acetylides, no evidence for the acetylide amide

t

1

Examples of azametallacyclobutanes such as 3-C₂H₄have

surfaced in recent years. Horton et al. reported that treatment

of (^tBu₃SiNH)(Et₂O)V(dNSi^tBu₃)₂with ethylene resulted in

displacement of ether and formation of (^tBu₃SiNH)(^tBu₃SiN)-

t

isomers of 3-HC₂H and 3-HC₂Bu (e.g., (silox)₂(^tBu₃SiNH)-

VCH₂CH₂NSi^tBu₃,⁴⁰which exhibited ethylene fluxionality and

slowly converted to the vinyl complex (^tBu₃SiNH)₂(^tBu₃SiN)-

VCHdCH₂. Exposure of Cp₂ZrdN^tBu(THF) to excess C₂H₄

and norbornene led to displacement of THF and formation of

azazirconacyclobutanes, according to Bergman et al.²⁶Rotation

t

TiCtCR, R ) H, Bu) was obtained, and they are considered

to be g2.7 kcal/mol⁴⁴endoergic with respect to the aza-

metallacycles.

4. (silox)₂TidNSi^tBu₃(3) and C₂H₄. Trapping of thermally

generated (silox)₂TidNSi^tBu₃(3, eq 8) by ∼2 equiv of ethylene

(eqs 21 and 22) afforded the azametallacyclobutane (^tBu₃SiO)₂-

t

of the ethylene fragment of Cp₂BuNZrCH₂CH₂was proposed

to account for a fluxional process that equilibrated the meth-

ylenes (∼12-13 kcal/mol at 250 K) without exchange with free

ethylene. An X-ray crystal structure of the norbornene adduct

revealed a d(Zr-N) of 2.013(3) Å, slightly shorter than a typical

zirconium-amide bond distance (e.g., Cp₂Zr(NH^tBu)(3,5-

Cl₂C₆H₃), d(Zr-N) ) 2.060(3) Å),²⁷and a normal d(Zr-C) of

2.241(3) Å⁴⁶that corroborates the azametallacyclobutane depic-

tion.

C₆H₁₂

t

+ H CdCH (excess)

8

[(silox)₂TidNSi Bu₃]

2

25 °C

3

(silox)₂(^tBu₃SiNH)TiCHdCH₂(2-Vy) +

(^tBu₃SiO)₂(^tBu₃SiN)TiCH₂CH₂(3-C₂H₄)

(21)

Olefin complexes of d⁰metals are expected to have low

stability because of the lack of M(dπ)fCdC(π*) bonding, but

some examples of olefins bound to formally d⁰metals have been

reported. Kress and Osborn⁴⁷found spectroscopic evidence for

a W(VI) alkylidene-cycloheptene complex, alternatively de-

scribed as a disrupted metallacyclobutane derivative, that

displays fluxionality related to the azametallacycles above (e.g.,

∆G^q_rot(240 K) ) 10.6(1) kcal/mol). More recently, Jordan et

al.⁴⁸and Casey and Hallenback⁴⁹found structural and spectro-

scopic evidence, respectively, of pendant olefins binding purely

via σ-interactions to d⁰metal centers (e.g., [Cp₂Zr(OCMe₂(CH₂)₂-

CHdCH₂)][MeB(C₆F₅)₃],⁴⁸Cp*₂Y(CH₂CH₂CMe₂CHdCH₂))⁴⁹

directly related to Ziegler-Natta systems.^50,51

(^tBu₃SiN)TiCH₂CH₂(3-C₂H₄) and (silox)₂(^tBu₃SiNH)TiCHdCH₂

(2-Vy), which was independently prepared via eq 4.⁴⁵At early

conversion of 2-R to 3, the product ratio for 3-C₂H₄:2-Vy of

92:8 translated to a kinetic preference for [2+2] addition of

ethylene to 3 versus C-H activation of ∆∆G^q_addn) -1.4 kcal/

mol. The ratio 3-C₂H₄:2-Vy was redetermined at equilibrium

to be 87:13 at 25 °C, corresponding to ∆G° ) -1.2 kcal/mol

for eq 23. A van’t Hoff analysis of the equilibrium constants

determined over a 24.8-87.7 °C range for this reaction gave

∆H° ) -5.72(11) kcal/mol and ∆S° ) -15.2(3) eu.

Molecular Structures. 1. (^tBu₃SiO)₂(^tBu₃SiNH)Ti-

t

CH₂CH₂Bu (2-^neoHex). After a struggle to obtain X-ray-

quality crystals of 2-Me, 2-Bz, and several other alkyl deriva-

NMR spectra of 3-C₂H₄exhibited the expected ^tBu resonances

tives, a single-crystal X-ray structural study (Table 2) of

and broad singlets at δ 3.08 (¹H) and δ 50.04 (¹³C{¹H}, J_CH

)

t

(^tBu₃SiO)₂(^tBu₃SiNH)TiCH₂CH₂Bu (2-^neoHex) confirmed its

146.5 Hz) corresponding to the methylenes of the azametalla-

cycle. In methylcyclohexane-d₁₄, the δ 3.09 resonance at 20

°C split into two multiplets at δ 2.57 and 3.60 when cooled to

constitution and geometry, as the molecular view in Figure 1

(46) Cardin, D. J.; Lappert, M. F.; Raston, C. L. Chemistry of Organo-

Zirconium and -Hafnium Compounds; Ellis Horwood Limited: New York,

1986.

(43) Polse, J. L.; Andersen, R. A.; Bergman, R. G. J. Am. Chem. Soc.

1995, 117, 5393-5394.

(47) Kress, J.; Osborn, J. A. Angew. Chem., Int. Ed. Engl. 1992, 31,

1585-1587.

(44) Assuming g1% of (silox)₂(^tBu₃SiNH)TiCtCR (2-CtCR) could

be observed by NMR spectroscopy relative to 3-HC₂R, at 24.8 °C, ∆G° g

RT ln{0.01/1-0.01} g 2.7 kcal/mol.

(48) Wu, Z.; Jordan, R. F.; Petersen, J. L. J. Am. Chem. Soc. 1995, 117,

5867-5868.

(45) Note the parallels to late-metal systems in: (a) Stoutland, P. O.;

Bergman, R. G. J. Am. Chem. Soc. 1988, 110, 5732-5744. (b) Bell, T.

W.; Brough, S. A.; Partridge, M. G.; Perutz, R. N.; Rooney, A. D.

Organometallics 1993, 12, 2933-2941.

(49) Casey, C. P.; Hallenbeck, S. L.; Pollock, D. W.; Landis, C. R. J.

Am. Chem. Soc. 1995, 117, 9770-9771.

(50) Jordan, R. F. AdV. Organomet. Chem. 1991, 32, 325-387.

(51) Marks, T. J. Acc. Chem. Res. 1992, 25, 57-65.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10703

Table 2. Crystallographic Data for Orthorhombic

sets failed to resolve the issue. Diffraction to better than 0.8 Å

was observed, and data to 0.85 Å were used in a successful

solution, resulting in a statistical distinction between d(Ti-N)

and d(Ti-O).

t

(^tBu₃SiO)₂(^tBu₃SiNH)TiCH₂CH₂Bu (2-^neoHex) and

(silox)₂TidNSi^tBu₃(THF) (3-THF)

2-^neoHex

3-THF

One molecule (A) has a disordered THF and the other a

formula

fw

space group:

λ, Å

a, Å

C₄₂H₉₄NO₂Si₃Ti

777.35

Pnma

C₄₀H₈₉NO₃Si₃Ti

764.29

Pbca

t

disordered Bu group (B), as the molecular views in Figure 2

illustrate, but common bond distances and angles are within

error. Greater repulsion among the silox and imide ligands again

accounts for the slightly distorted tetrahedral geometry of

3-THF: O_THF-Ti-(N/O) ) 101.9(19)° average, O-Ti-

(N/O) ) 115.9° average. The Ti-N-Si (173.0(27)° average)

and Ti-O-Si (172.1(36)° average) angles are essentially

identical, while the titanium-imide bond lengths (d(Ti-N) )

1.772(3), 1.783(3) Å) are significantly shorter than the silox

Ti-O bonds of 1.824(4) Å (average), which are comparable to

those of (silox)₃TiNH₂(d(Ti-O) ) 1.815 Å average).⁵³Nearly

linear M-O-Si linkages are normally observed,⁵⁴and the

similar imide linkage reflects the cylindrical symmetry imposed

by pseudotetrahedral coordination of four potential π-donors,

even though N(pπ) f Ti(dπ) donation is moderate since the

d(Ti-N) is long relative to those of other complexes, which

exhibit titanium-imide distances from 1.672 to 1.723 Å.³²

The Lewis acidity possessed by (silox)₂TidNSi^tBu₃(3),

0.7107

0.908

18.445(10)

19.375(12)

14.200(14)

5075(7)

293(2)

17.319(3)

25.447(5)

41.861(8)

18449(6)

165(5)

b, Å

c, Å

V, Å³

T, K

Z

4

16

1.101

0.297

0.0637

0.1749

0.972

F

_calc, g/cm³

1.017

0.269

0.1388

0.2838

µ, mm^-1

R [I > 2σ(I)]

R_w²[I > 2σ(I)]

GOF

1.397

2

localized in its vacant d_z/p_z(z-axis to the ligand plane) hybrid

orbital, plays the crucial role in capturing L or the pair of

electrons in a C-H bond. Provided 3-THF can be considered

a rough model of the transient responsible for C-H bond

activation, either 3 or the related hydrocarbon adduct (silox)₂-

(RH)TidNSi^tBu₃(3-RH), its long TidN bond length may

signify the availability of a nitrogen lone pair to function as an

internal base in the 1,2-RH-addition process. In the related

zirconium-based C-H bond activation system, a similar conclu-

sion was applied to (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃,²⁰whose

d(ZrdN) of 1.978(8) Å was ∼0.1 Å longer than expected. The

zirconium derivative also contains a d(Zr-O_THF) of 2.229(7)

Å, which is slightly longer than the d(Ti-O_THF) of 2.037(1) Å

corresponding to 3-THF, once a 0.13 Å correction for differing

covalent radii is applied. Although THF adducts of other

titanium imides are not available for comparison, the bond

lengths of the zirconium derivative and related congeners

suggest that the O_THF(pπ)fTi(dπ) interaction is minimal in

3-THF, despite a favorable π-bonding orientation featuring a

relatively flat THF ( Ti-O-C ) 124.8(24)° average) roughly

aligned with the TidN vector.

t

Figure 1. Molecular view of (silox)₂(^tBu₃SiNH)TiCH₂CH₂Bu (2-CH₂-

t

CH₂Bu) showing a nonagostic neohexyl group (d(Ti-C) ) 2.09(3)

Å, Ti-C-C ) 111(2)°) and Bu₃Si disorder. The silox and Bu₃-

t

SiNH groups cannot be differentiated.

indicates. Although crystallographic C_ssymmetry is enforced

on the molecule, the structure solution exhibits severe disorder

in the periphery of the Bu₃SiX (X ) O, NH) ligand on the

t

mirror plane. Further unresolved disorder of the ^tBu₃SiX ligands

is likely, on the basis of the unusual discrepancy between the

U_isofor O (U_iso) 69(5) Å²) and N (U_iso) 10(6) Å²);

consequently, the amide and silox groups cannot be confidently

distinguished. The pseudotetrahedral geometry possessed by

2-^neoHex ( O-Ti-O/N ) 115.7(24)° average, C-Ti-O/N

) 102.1(6)° average) reflects greater steric interactions among

Kinetics of 1,2-RH-Elimination. 1. Rate Expression.

Monitoring by ¹H NMR spectroscopy revealed that thermolysis

of (silox)₂(^tBu₃SiNH)TiR (2-R) in benzene-d₆produced RH and

(silox)₂(^tBu₃SiND)TiC₆D₅(2-(ND)-C₆D₅), consistent with the

following mechanism:

t

^{k (elim)}8

+ RH

the Bu₃Si-containing ligands. No evidence of an agostic

R

t

(silox)₂(^tBu₃SiNH)TiR

(silox)₂TidNSi Bu₃

interaction is evident, as judged by the normal Ti-C-C angle

of 111(2)° and standard Ti-C bond length of 2.09(3) Å, which

is comparable to the Ti-CH₃bond distance in [(^tBu₃CO)TiMe₂]-

(µ-OMe)₂of 2.063(10) Å.⁵²

C₆D₆

2-R

3

(23)

k_PhD(addn)

(silox) (^tBu SiND)TiC D

5

2. (^tBu₃SiO)₂(THF)TidNSi^tBu₃(3-THF). A single-crystal

X-ray structural determination of (^tBu₃SiO)₂(THF)TidNSi^tBu₃

(3-THF) revealed an asymmetric unit (Table 2, Figure 2)

composed of two molecules that have essentially identical,

slightly distorted tetrahedral geometries. To conclusively

establish whether a silox/imide disorder, similar to the silox/

amide disorder of 2-^neoHex, was problematic to 3-THF, a high-

quality data set was collected at 165(5) K using synchrotron

radiation after conventional Mo KR and Cu rotating-anode data

3

8

(24)

(25)

2

3

6

C₆D₆

2-(ND)-C₆D₅

L

(silox)₂LTidNSi^tBu₃

3-L

3

^{k (trap)}8

C₆D₆

Rate-determining loss of RH generated transient (silox)₂-

TidNSi^tBu₃(3, eq 23), which was scavenged via C-D bond

(53) Mass, J. L.; Wolczanski, P. T. Unpublished results.

(54) Wolczanski, P. T. Polyhedron 1995, 14, 3335-3362 and references

therein.

(52) Lubben, T. V.; Wolczanski, P. T. J. Am. Chem. Soc. 1987, 109,

424-435.

10704 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

Figure 2. Molecular views of two independent molecules of (silox)₂(THF)TidNSi^tBu₃(3-THF), one exhibiting THF backbone disorder and the

t

other (primed) revealing Bu group disorder. Pertinent interatomic distances (Å) and angles (deg): Ti1-N1, 1.772(3); Ti1′-N1′, 1.783(3); Ti-

OSi_av, 1.824(4); Ti-O(THF)_av, 2.037(3); N-Si_av, 1.684(3); O-Si_av, 1.644(3); N1-Ti1-O1, 114.38(12); N1′-Ti1′-O1′, 115.43(11); N1-Ti1-

O3, 118.52(13); N1′-Ti1′-O3′, 116.26(11); N1-Ti1-O4, 103.17(10); N1′-Ti1′-O4′, 104.44(11); O1-Ti1-O3, 115.14(11); O1′-Ti1′-O3′,

115.54(11); O1-Ti1-O4, 103.17(10); O1′-Ti1′-O4′, 104.44(11); O3-Ti1-O4, 100.51(10); O3′-Ti1′-O4′, 99.17(10); Ti1-N1-Si2, 174.9(2);

Ti1′-N1′-Si2′, 171.2(2); Ti1-O1-Si1, 175.6(2); Ti1′-O1′-Si1′, 172.9(2); Ti1-O3-Si3, 172.6(2); Ti1′-O3′-Si3′, 167.1(2); Ti1-O4-C37,

123.6(2); Ti1′-O4′-C37′, 121.5(3); Ti1-O4-C40, 126.2(2); Ti1′-O4′-C40′, 128.0(3).

activation to give 2-(ND)-C₆D₅(eq 24). Under these conditions,

loss of RH proceeds to completion and follows first-order

kinetics. The isotopically labeled product is inconsistent with

a σ-bond metathesis pathway,¹⁵and precedent in related

zirconium²⁰and tantalum²³systems suggested that the 1,2-RH-

elimination/addition mechanism was operative.

molecular transition state, thereby supporting the proposed 1,2-

RH-elimination mechanism where amide reorganization to

achieve R-Ti-N-H planarity is a critical feature.^55,56Similar

activation parameters have been observed for MeH loss from

(^tBu₃SiNH)₃ZrMe (∆H^q) 25.9(4) kcal/mol, ∆S^q) -7(1) eu)²⁰

t

and Bu₃SiNH₂elimination from (^tBu₃SiNH)₃TiCl (∆H^q

)

Thermolysis of (silox)₂(^tBu₃SiNH)TiMe (2-Me) at different

concentrations of 2-Me established that the reaction is first order

in 2-Me (k_Me(24.8 °C) ∼ 1.4 × 10^-5s^-1, Table 3). The reaction

was zero order in [C₆D₆] according to experiments conducted

23.3(8) kcal/mol, ∆S^q) -11(2) eu),²²which are processes

considered 1,2-eliminations.

In order to interpret rates of 1,2-RH-elimination, it is assumed

that enthalpies of activation are primarily responsible for the

3.5 kcal/mol range in observed ∆G^q

values (22.2 (2-Ph) to

in C₆D₆/C₆D₁₂mixtures (24.8 °C, [C₆D₆] ) 11.2 M (k_Me

)

elim

1.5(2) × 10^-5s^-1), 4.52 M (3(1) × 10^-5s^-1), 1.12 M (2.6(6)

× 10^-5s^-1), but problems in integrating overlapping resonances

rendered the rates imprecise. When the elimination was carried

out in neat THF-d₈or in C₆D₆with 20 equiv of THF present,

thereby producing (silox)₂(THF-d₈)TidNSi^tBu₃(3-THF-d₈, k_Me

) 1.46(5) × 10^-5) or 3-THF (k_Me) 1.54(7) × 10^-5s^-1, eqs

23 and 25), respectively, no variation in the rate constant was

noted relative to neat benzene-d₆. The rate law for disappear-

ance of 2-Me, and by inference all remaining 2-R conducted

under the aforementioned conditions, is given in eq 26.

c

25.7 kcal/mol (2-Mes, 2-Bz)). Disparate 2-R (R ) Me, Pe,

Bz) were chosen for the Eyring analyses in order to dispel the

notion that variations in ∆S^qcould significantly affect trends

in free energy of activation. It is noteworthy that three

extremely different 2-R derivatives gave essentially equal

entropies of activation, yet the typical ((2-4) eu error in ∆S^q

at 24.8 °C amounts to 0.6-1.2 kcal/mol, a lack of precision

that again necessitates the assumption that activation entropies

play an insignificant role in ∆G^q_elimcomparisons. In correlations

below, it is argued that enthalpies of activation are critical to

understanding the 1,2-RH-elimination rates of various 2-R, a

supposition buttressed by accompanying equilbrium studies.

3. 1,2-RH-Elimination Kinetic Isotope Effects. Loss of

CH₃D from 2-(ND)-Me proceeded with a large KIE at 24.8 °C

(k_H/k_D) z_Me) 13.7(9)). A similar loss of toluene-d₁from

2-(ND)-Bz also exhibited a large primary KIE, which was

measured at higher temperatures because of the sluggishness

of the reaction. The values at 52.4 °C (z_Bz) 10.5(7)), 70.2 °C

(z_Bz) 9.6(8)), and 90 °C (z_Bz) 5.6(2)) crudely extrapolate to

a z_Bznear that of z_Meat 24.8 °C. The value obtained via loss

of C₆H₅D from 2-(ND)-Ph, z_Ph) 7.4(3) at 24.8 °C, is somewhat

smaller in magnitude.

-d[2-R]/dt ) k_obs[2-R]¹[trap]⁰

(26)

Furthermore, it is clear that solvent effects in this system are

unremarkable, in corroboration of the proposed rate-determining

1,2-RH-elimination and constrained four-center transition state

(Scheme 1).

c

2. Activation Parameters for 2-R (R ) Me, Pe, Bz).

Eyring analyses were conducted for clean, first-order loss of

RH from a primary alkyl (2-Me) and secondary alkyl (2-^cPe)

over a 24.8-71.3 °C range and for a benzylic alkyl (2-Bz) over

a 24.8-90.2 °C range (Table 3). Similar analyses for 2-Ph or

other sp²(e.g., 2-Vy) derivatives were precluded by the swiftness

of their eliminations at elevated temperature. For all three

species, a relatively large enthalpy of activation was observed,

accompanied by a moderate, negative entropy of activation (2-

Bz, ∆H^q) 22.2(5) kcal/mol, ∆S^q) -12(2) eu; 2-Me, ∆H^q)

20.2(12) kcal/mol, ∆S^q) -12(4) eu; 2-^cPe, ∆H^q) 19.6(6)

kcal/mol, ∆S^q) -12(2) eu). The activation parameters indicate

significant bond-breaking occurring within a constrained, uni-

Large primary kinetic isotope effects are consistent with a

four-center transition state in which the transfer of H from N

to R is relatively linear and there are similar amounts of N-H

bond-breaking and C-H bond-making character.^55,56Support

(55) Carpenter, B. K. Determination of Reaction Mechanisms; Wiley-

Interscience: New York, 1984.

(56) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic

Chemistry, 3rd ed.; Harper and Row: New York, 1987.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10705

for this transition state is evident from calculational studies of

model systems such as (H₂N)₃MMe (M ) Ti, Zr, Hf)^57-60and

(HO)₂(H₂N)TiMe.⁶¹The smaller isotope effect observed for

2-Ph vs 2-(ND)-Ph portends a less symmetric transition state

for H transfer in 1,2-PhH/D-elimination (i.e., force constants

that describe NH/D vs CH/D in the transition state are less

similar in magnitude than those in the Me and Bz cases).

Equilibria Relating 2-R, 3-L, and 3-RCCR. ¹H NMR

spectroscopy was used to monitor the approach to equilibrium

of four types of reactions, indicated by the generic equations

(27)-(30). In a typical experiment, an NMR tube was charged

kinetic studies. The 19.1 kcal/mol ladder in Figure 3 is the

result of a least-squares fit of the experimental data indicated

by the arrows (see Supporting Information).

In relative terms, the least thermodynamically stable organo-

titanium species are the secondary alkyls 2-^cHex and 2-^cPe,

n

c

followed by a host of sp³-alkyls (ⁿPr, Bu, ^neoHex, Et, Bu,

CH₂SiMe₃, Me) and finally the sp²-alkyls (2-Ph, 2-Vy, 2-^cPr).

Two adducts of 3 (3-C₂H₄and 3-NEt₃) are less stable than the

hydride 2-H, which is in turn less stable than the remaining

adducts and azametallacyclobutenes. The spread among the

alkyl derivatives is only 5.5 kcal/mol, while the adducts range

from azametallacyclobutane 3-C₂H₄at -6.7 kcal/mol to 3-py

at -19.1 kcal/mol, a range of 12.4 kcal/mol. Within the adducts,

steric factors are important to relative stability. Triethylamine

adduct 3-NEt₃is 9.0 kcal/mol less stable than the corresponding

trimethyl derivative, 3-NMe₃, and the azametallacycle derived

from 2-butyne (3-MeC₂Me) is 4.0 kcal/mol more stable than

the 3-hexyne adduct (3-EtC₂Et). Within the nitrogen bases, the

least basic but planar and effectively smaller py ligand makes

an adduct (3-py) that is 3.6 and 12.6 kcal/mol more stable than

3-NMe₃and 3-NEt₃, respectively.

t

+

(silox)₂( Bu₃SiNH)TiR

2-R

t

R′H h

+ RH (27)

+ RH (28)

+ L (29)

( Bu₃SiO)₂( Bu₃SiNH)TiR′

2-R′

t

+ L h

(silox)₂( Bu₃SiNH)TiR

2-R

(^tBu₃SiO)₂LTidNSi^tBu₃

3-L

The temperature dependencies of selected equilibria were

roughly determined in order to assess contributions from ∆S°,

which contains a factor based solely on the relative numbers of

symmetry-equivalent hydrogens available for activation on the

competing substrates. The statistical factor pertaining to 2-R

+ R′H h 2-R′ + RH (eq 28) may be trivially separated from

the standard free energy change, yielding a corrected free energy

according to eq 31. ∆G°_corrmay be considered a “per H”

(^tBu₃SiO)₂LTidNSi^tBu₃

3-L

+ L′ h

(^tBu₃SiO)₂L′TidNSi^tBu₃

3-L′

(^tBu₃SiO)₂LTidNSi^tBu₃

3-L

+ R′CtCR′′ h

(silox) (^tBu SiN)TiCR′dCR′′

∆G°_corr) ∆G° - RT ln(W/W′)

(31)

+ L (30)

2

3

3-R′C₂R′′

W (W′) ) number of equivalent H’s on RH (R′H) that can

be activated

with 2-R followed by addition of C₆D₁₂solvent and an excess

of R′H, whose concentration was chosen to obtain a quantifiable

range of K_eq. Convenient limits for observing equilibria were

-4 < ∆G° < 4 kcal/mol, although in some cases ∆G° values

as high as 7.8 kcal/mol were determined.

standard free energy change for any given equilibrium. Van’t

Hoff analyses (24.8, 50.2, 70.2 °C) revealed standard entropy

changes that are slightly larger in magnitude than the respective

statistical contributions. For example, measurements of the 2-Et

+ MeH h 2-Me + EtH equilibrium yielded a ∆S° of -2.8(6)

eu that includes -0.8 eu based on statistics. Similarly,

2-CH₂SiMe₃+ MeH h 2-Me + SiMe₄was shown to have

∆S° ) -6.8(24) eu, a value that incorporates -2.2 eu purely

due to the statistics of available C-H bonds that can be

activated. This equilibrium was chosen because the (trimeth-

ylsilyl)methyl group is extremely bulky and representative of

2-R species whose entropic factors reflect this factor. The

moderate ∆S° is still within 2σ of all the others, even though it

is somewhat greater in magnitude. Other cases manifested even

In lieu of monitoring the approach to equilibrium from both

sides of a reaction, which in some cases was unfeasible,

verification of an established equilibrium was often attained via

cross-checking through measurement of related, independent

equilibria. As a direct consequence, each complex was linked

to another via a ladder of relative standard free energies and

arbitrarily set relative to 2-^cHex at 0.0 kcal/mol (Figure 3). In

addition, the propagation of error was minimized by the cross-

checks, which helped ensure that no single experiment having

a large (but potentially unknown) systematic error could unduly

skew all of the data. With all complexes in the ladder sharing

a common free energy surface, each 2-R position in the ladder

marks the relative standard free energy of the complex and all

other substrates (e.g., 2-R¹+ R²H + ... + L¹+ L²+ ... +

RC₂R′ + ...). Likewise, each 3-L position corresponds to the

relative standard free energy of 3-L¹+ R¹H + R²H + ... + L²

+ ... + RC₂R′ + .... Implicit in the energy scale is the

assumption that various secondary effects, such as those derived

from the medium of each individual experiment, are negligible.

This assumption appears reasonable, given the covalent character

of the complexes involved and the lack of evidence pertaining

to significant solvent interactions, even in the aforementioned

less deviation (e.g., 2-Et + PrH h 2-^cPr + EtH, ∆S° )

c

-1.1(13)). By inference from the selected temperature de-

pendence studies, ground state entropic factors pertaining to each

2-R state are also considered moderate and are deemed

negligible in much of the remaining discussion. As a conse-

quence, the standard enthalpy differences between states are

considered to parallel the standard free energy differences

exhibited in Figure 3.

Note that the equilibrium 3-THF + C₂H₄h 3-C₂H₄+ THF

is endoergic by 9.0 kcal/mol, whereas the related

t

Cp₂ZrdN^tBu(THF) + C₂H₄h Cp₂BuNZrCH₂CH₂+ THF

equilibrium was reported to be endoergic by only 1.6 kcal/mol

at 25 °C.²⁶Likewise, the displacement of ether by ethylene in

(^tBu₃SiNH)(Et₂O)V(dNSi^tBu₃)₂must also have a modest ∆G°.⁴⁰

The dramatic difference exhibited by the titanium system is

attributed to the greater electrophilicity of the group 4 bis(silox)-

(57) Cundari, T. R. J. Am. Chem. Soc. 1992, 114, 10557-10563.

(58) Cundari, T. R. Organometallics 1993, 12, 1998-2000.

(59) Cundari, T. R.; Gordon, M. S. J. Am. Chem. Soc. 1993, 115, 4210-

4217.

(60) Cundari, T. R. Organometallics 1994, 13, 2987-2994.

(61) Cundari, T. R. Organometallics 1993, 12, 4971-4978.

10706 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

Table 3. First-Order Rate Constants,^aFree Energies of Activation, and Corresponding D(RH)^bfor 1,2-RH-Elimination of RH from

(^tBu₃SiO)₂(^tBu₃SiNH)TiR (2-R) in C₆D₆, Producing (^tBu₃SiO)₂(^tBu₃SiND)TiC₆D₅(2-(ND)-C₆D₅) (Exceptions Noted)

compound (solvent C₆D₆or as noted)

10^-5k, s^-1

T ((0.3), °C

∆G^‡, kcal/mol

[2-R], M

D(RH) kcal/mol

(silox)₂(^tBu₃SiNH)TiCH₂C₆H₃Me₂(2-Mes)

(silox)₂(^tBu₃SiNH)TiCH₂Ph (2-Bz)

0.0872(7)

0.086(6)^c

2.26(5)^c,d

17(1)^c,d

92(2)^c,d

0.21(1)^d

1.80(5)^d

16.3(5)^d

0.58(5)^e

1.54(10)^f

1.42(7)

1.3(2)^g

24.8

52.4

70.2

90.2

52.4

70.2

90.2

24.8

35.1

50.2

63.6

71.3

24.8

40.4

50.5

60.0

70.9

24(1)

24.8

25(1)

25.1

24.8

25.7

0.037

0.038

0.037

0.032

0.041

0.032

0.045

0.041

0.023

0.005

0.042

0.041

0.040

0.05

0.049

0.045

0.02

0.042

0.04

88.5(15)

(silox)₂(^tBu₃SiND)TiCH₂Ph (2-(ND)-Bz)

(silox)₂(^tBu₃SiNH)TiH (2-H)

24.6

24.0

104.2(1)

104.9(1)

(silox)₂(^tBu₃SiNH)TiCH₃(2-Me)

4.9(7)^f

30(2)^f

79(6)^f

170(20)^f

1.46(5)^h

1.5(2)ⁱ

2-Me in THF-d₈; product 3-THF-d₈

(2-Me in C₆D₁₂, [C₆D₆] ) 4.52 M)

(2-Me in C₆D₁₂, [C₆D₆] ) 1.12 M)

3(1)^i,j

2.6(6)^i,k

1.54(7)^d,l

0.112(6)^d,l

1.71(8)

1.86(10)

2.05(3)

2.13(8)

4.6(2)^m

25(2)^m

(silox)₂(^tBu₃SiND)TiCH₃(2-(ND)-Me)

t

(silox)₂(^tBu₃SiNH)TiCH₂CH₂Bu (2-^neoHex)

23.9

23.8

23.4

99.9(15)

101.1(4)

99.9(19)

96.5(10)

94.4(10)

(silox)₂(^tBu₃SiNH)TiCH₂CH₃(2-Et)

(silox)₂(^tBu₃SiNH)TiCH₂(CH₂)₂CH₃(2-ⁿBu)

(silox)₂(^tBu₃SiNH)Ti-c-C₄H₇(2-^cBu)

(silox)₂(^tBu₃SiNH)Ti-c-C₅H₉(2-^cPe)

0.04

0.040

0.019

0.020

0.018

0.019

0.04

0.042

0.043

0.02

70(1)^m

174(4)^m

434(12)^m

8.8(6)

(silox)₂(^tBu₃SiNH)Ti-c-C₆H₁₁(2-^cHex)

(silox)₂(^tBu₃SiNH)Ti-c-C₃H₅(2-^cPr)

(silox)₂(^tBu₃SiNH)TiCHdCH₂(2-Vy)

(silox)₂(^tBu₃SiNH)TiC₆H₅(2-Ph)

23.0

22.8

22.4

22.2

95.6(10)

106.3(3)

111.2(8)

113.5(5)

11.6(5)

23.9(4)

33.3(3)ⁿ

6.7(8)ⁿ

(silox)₂(^tBu₃SiND)TiC₆H₅(2-(ND)-Ph)

0.02

^aDetermined from nonlinear least-squares fitting of the differential form of the rate expression. For details regarding the individual experiments,

consult the Experimental Section. ^bD(RH) values are from ref 62. ^cValues used in the Eyring plot (24.8-90.2 °C) obtained from triplicate runs.

From a weighted nonlinear least-squares fit of the data: ∆H^q) 22.2(5) kcal/mol, ∆S^q) -12(2) eu. ^dTandem measurements for obtaining the

primary isotope effect: k_H/k_D(PhCH₂H/D loss) ) 10.5(7) (52.4 °C), 9.6(8) (70.2 °C), 5.6(2) (90.2 °C); k_H/k_D(MeH/D loss) ) 13.7(9). ^eObtained

from the instantaneous rate determined at early conversion. ^fValues used in the Eyring plot (24.8-71.3 °C) obtained from triplicate runs. From a

weighted nonlinear least-squares fit of the data: ∆H^q) 20.2(12) kcal/mol, ∆S^q) -12(4) eu. ^gObtained from dilution of a 0.017 M stock solution.

^hCompare to MeH loss from 2-Me in C₆D₆; k(THF-d₈)/k(C₆D₆) ) 0.95(7). ⁱC₆D₆and C₆D₆/C₆D₁₂runs conducted in parallel. ^j90 equiv of C₆D₆.

^k22 equiv of C₆D₆. ^lConducted with 20 equiv of THF (product 3-THF) for the purposes of obtaining k_H/k_D; note that k_Me(20 equiv THF)/k_Me(C₆D₆)

) 1.00(8). ^mValues used in the Eyring plot (24.8-70.9 °C) obtained from triplicate runs. From a weighted nonlinear least-squares fit of the data:

∆H^q) 19.6(6) kcal/mol, ∆S^q) -13(2) eu. ⁿMeasurements used in determination of the primary isotope effect: k_H/k_D(PhH/D loss) ) 7.4(3).

(amide)metal center in comparison to the later, less electro-

D(TiR)_rel) [D(TiR) - D(TiBz)] )

∆H°_reacn+ [D(RH) - D(BzH)] (33)

positive vanadium and metal centers ligated by better e^--donors

(e.g., Cp, imide). Steric factors also mitigate O(pπ)fZr(dπ)

bonding in Cp₂ZrdN^tBu(THF).

Figure 4 displays a generally strong correlation of D(TiR)_rel

with D(RH)⁶²(slope ) 1.1, r ) 0.95) that improves measurably

when the R ) Bz, Mes, H, and Ph points are removed (slope

) 1.36, r ) 0.995). Other researchers have noted a roughly

linear correlation between D(RH) and D(L_nM-R) for disparate

organometallic systems.^63-70Its existence typically implies

strong metal-carbon bonding and minimal involvement of

Thermochemistry. 1. Relative D(TiR). The equilibria

describing the interconversion of the 2-R derivatives (eq 28)

were used to determine relative Ti-R bond strengths. With

some assumptions, the ∆H_reacnfor eq 27 is dependent on the

differences of the relative titanium-carbon and carbon-

hydrogen bond strengths according to eq 32. The aforemen-

∆H°_reacn≈ [D(R′H) - D(RH)] + [D(TiR) - D(TiR′)] (32)

(62) (a) Berkowitz, J.; Ellison, G. B.; Gutman, D. J. Phys. Chem. 1994,

98, 2744-2765 and references therein. (b) Davico, G. E.; Bierbaum, V.

M.; DePuy, C.H.; Ellison, G. B.; Squires, R. R. J. Am. Chem. Soc. 1995,

117, 2590-2599. (c) Martinho Simo˜es, J. A.; Beauchamp, J. L. Chem. ReV.

1990, 90, 629-688. (d) D(H^cPr): Baghal-Vayjooee, M. H.; Benson, S. W.

J. Am. Chem. Soc. 1979, 101, 2838-2840. (e) D(H^cBu): CRC Handbook

of Chemistry and Physics, 75th ed.; CRC Press: Boca Raton, FL, 1994. (f)

D(H^neoHex) assumed to equal D(HBu). (g) D(HMes) assumed to equal

D(HBz).

tioned statistical correction (eq 31) is included, but other entropic

factors contributing to the free energies are assumed to be

negligible or cancel, as implied by the temperature dependence

studies. Likewise, enthalpic contributions such as heats of

solvation of the various 2-R and R′H are assumed to be

negligible or cancel, as in previous cases.^20,23With D(TiBz)

arbitrarily chosen as a reference, the various relative titanium-

carbon bond strengths corresponding to (silox)₂(^tBu₃SiNH)TiR

(2-R, eq 33) were compiled and plotted relative to the absolute

D(RH) of the corresponding hydrocarbons, as shown in Figure

4.

(63) Bryndza, H. E.; Fong, L. K.; Paciello, R. A.; Tam, W.; Bercaw, J.

E. J. Am. Chem. Soc. 1987, 109, 1444-1456.

(64) Bryndza, H. E.; Domaille, P. J.; Tam, W.; Fong, L. K.; Paciello, R.

A.; Bercaw, J. E. Polyhedron 1988, 7, 1441-1452.

(65) Bulls, A. R.; Bercaw, J. E.; Manriquez, J. M.; Thompson, M. E.

Polyhedron 1988, 7, 1409-1428.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10707

Figure 4. Relative Ti-C bond strengths (kcal/mol) in (silox)₂(^tBu₃-

SiNH)TiR (2-R) versus the C-H bond strength of the corresponding

hydrocarbon.⁶²D(TiR)_rel) D(TiR) - D(TiBz) as in eq 33, with ∆H°_reacn

estimated from ∆G°_corr(eq 31) as explained in the text. The line (slope

) 1.36, r ) 0.9953) is a least-squares fit to the points except Bz, Mes,

H, and Ph. A least-squares line to all points had a slope of 1.1 (r )

0.95).

ligand aberrant when grouped with hydrocarbyls. Unlike L_nM

and the hydrocarbyls, hydride also has virtually no ability to

accommodate charge⁷³and is intrinsically less polarizable. The

absence of any steric perturbation on the hydride ligand, here

probably not a factor, and the lack of a significant ionic

contribution to the correlated homonuclear H-H bond^70,74have

also been proposed to explain its unique character.

Figure 3. Ladder of standard free energies accorded hydrocarbyl (and

hydride) states 2-R, adduct states 3-L, and metallacyclic states 3-RC₂R′

and 3-C₂H₄, as defined in the text. (silox)₂(^tBu₃SiNH)Ti^cHex (2-^cHex)

was chosen as the reference state, hence each ∆G° in parentheses refers

to the value for 2-^cHex + RH/L/RC₂R′ + ... h 2-R/3-RC₂R′/3-L +

^cHexH + .... These optimized values are derived from a least-squares

fit of individual equilibria indicated by the arrows and listed in

Supporting Information.

The remaining outlying hydrocarbyls may exhibit secondary

ground state influences in addition to their intrinsic bond

strengths, as defined by the correlation with D(RH). The benzyl

and mesityl substituents have ∼6-7 kcal/mol of additional

stabilization, whereas the metal-carbon bond of (silox)₂-

(^tBu₃SiNH)TiPh (2-Ph) is weaker than expected by ∼4 kcal/

mol. These substantial deviations in magnitude imply enthalpic

rather than entropic origins. Agostic⁷²or η³-binding⁷⁵of the

benzyl and mesityl substituents could account for the apparent

extra strength of their bonds. The J_CHof the benzyl methylene

group remained at 124 Hz upon cooling from +25 to -70 °C,

secondary effects; consequently, deviations from linearity may

suggest that lesser effects (e.g., steric,⁷¹π-bonding, agostic,⁷²

etc.) may be operative. Note that the sp³-centers, excluding

benzylic bonds and adding the sp²-like cyclopropyl and sp²-

vinyl, comprise a well-behaved group that reflects little devia-

tion. Three of the larger substituents, cyclopentyl, neohexyl,

and CH₂SiMe₃, show no deviation, while cyclohexyl is slightly

(∼1-2 kcal/mol) destabilized relative to the line, but no more

so than ⁿPr, and both are within reasonable error of the

correlation and its assumptions. It is inferred that ground state

steric effects are minimal, as corroborated by the X-ray

crystallographic characterization of 2-^neoHex, and the assumption

that entropic differences are minimal also gains credence.

The titanium-hydride bond is about 7 kcal/mol stronger than

the line implies, but the magnitude of the difference is relatively

small and typical of systems that manifest strong D(L_nM-

R(H)),^63-70where secondary influences are minimal. Presum-

ably the unique s-orbital overlap of the hydride renders this

1

and the disposition of the aromatic signals in the H NMR

spectrum did not change substantially with temperature (-85

°C f 25 °C). Stretches in the infrared spectrum in the range

2350-2700 cm^-1are considered diagnostic for agostic C-H

bonds,⁷²but the IR spectrum of 2-Bz was essentially featureless

from 1600 to 2800 cm^-1. Coupled with the regular disposition

of the neohexyl group in the X-ray crystal structure of (silox)₂-

t

(^tBu₃SiNH)TiCH₂CH₂Bu (2-^neoHex), the absence of diagnostic

spectral features in 2-Bz suggests that an agostic interaction is

not operative, but a minor, spectroscopically (NMR) undetect-

able stabilization (<7 kcal/mol) by an η³-interaction is probable.

In support, allyllic binding in (silox)₂(^tBu₃SiNH)Ti(η³-

H₂CCHCHR) (R ) H (2-η³-H₂CCHCH₂), Me (2-η³-H₂-

CCHCHMe)) is strong, and X-ray crystal structures of related

d⁰benzyl derivatives exhibit allylic stabilization.⁷⁵Finally,

while 2-R are generally colorless, 2-Bz is yellow, a color

common to early-metal benzylic species that exhibit allylic

(66) Stoutland, P. O.; Bergman, R. G.; Nolan, S. P.; Hoff, C. D.

Polyhedron 1988, 7, 1429-1440.

(67) Schock, L. E.; Marks, T. J. J. Am. Chem. Soc. 1988, 110, 7701-

7715.

(68) Dias, A. R.; Martinho Simo˜es, J. A. Polyhedron 1988, 7, 1531-

1544.

(69) Diogo, H. P.; de Alencar Simoni, J.; Minas da Piedade, M. E.; Dias,

A. R.; Martinho Simo˜es, J. A. J. Am. Chem. Soc. 1993, 115, 2764-2774.

(70) Labinger, J. A.; Bercaw, J. E. Organometallics 1988, 7, 926-928.

(71) Halpern, J. Inorg. Chem. Acta 1985, 100, 41-48.

(72) Brookhart, M.; Green, M. L. H.; Wong, L. Prog. Inorg. Chem. 1988,

36, 1-124.

(73) Siegbahn, P. E. M. J. Phys. Chem. 1995, 99, 12723-12729.

(74) Pearson, R. G. J. Chem. Soc., Chem. Commun. 1968, 65-67.

(75) (a) Davies, G. R.; Jarvis, J. A. J.; Kilbourn, B. T.; Pioli, A. J. P. J.

Chem. Soc., Chem. Commun. 1971, 677. (b) Davies, G. R.; Jarvis, J. A. J.;

Kilbourn, B. T. J. Chem. Soc., Chem. Commun. 1971, 1511-1512. (c) Bassi,

I. W.; Allegra, G.; Scordamaglia, R.; Chioccola, G. J. Am. Chem. Soc. 1971,

93, 3787-3788.

10708 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

bonding. Alternatively, the electron-withdrawing capability of

a phenyl group may impart greater ionic character⁷³to each

metal-benzyl bond, thereby increasing D(MBz) relative to those

of the remaining hydrocarbyls, or the benzyl point signifies the

origin of nonlinearity in this correlation (Vide infra).

The most difficult outlying substituent to rationalize is phenyl,

whose modest ∼4 kcal/mol destabilization relative to the line

is greater than a steric effect and counter to an expected increase

in stability on the basis of the increased electron-withdrawing

ability of an sp²-center. (silox)₂(^tBu₃SiNH)TiPh (2-Ph) contains

three π-donors of electronegativity greater than that of the

phenyl: two siloxides and the amide. The inclusion of even a

very weak π-donor such as phenyl may provide unwanted

competition for empty titanium d-orbitals used in X(pπ)fTi(dπ)

(X ) N, O) bonding, resulting in a net destabilization relative

to a substituent whose bonding is solely σ. Repeated attempts

to gain structural information on 2-Bz and 2-Ph have failed.

kcal/mol,⁸¹a value reasonable in view of thermochemical

estimates on related cyclopentadienyl compounds.⁶⁸In the

Matcha analysis, best fits were determined with ø_Ti∼ 2.7,⁸¹

a

value near those of the hydrocarbyls (ø ∼ 2.55-2.58),^82,83while

the ECT treatment revealed a covalency for (silox)₂(^tBu₃SiNH)-

Ti approaching that of a methyl group. However, it is likely

that determinations based on the Pauling equation and variants

give erroneous ø(L_nM) values.^67,70,81

While the methods above are subject to some interpretation,

the Ti-R bonds of 2-R are best described as covalent, with a

lesser ionic component, yet such factors are responsible for the

slope greater than 1. For example, if ø_Mcan be realistically

assumed to be <ø_H) 2.20,⁸¹eq 35 predicts that ∂D(MX)/

∂D(HX) will be >1; i.e., ionic contributions to M-R bonding

are greater than in R-H bonding and cause the deviation from

unity slope. Siegbahn has also concluded that ionic contribu-

tions are critical in related metal-carbon bonds.⁷³

In (dppe)(Me)PtX and Cp*(Me₃P)₂RuX systems investigated

by Bryndza et al.,^63,64with hydrocarbyl and other substituents

that span ∆D(HX) ) 48 kcal/mol, the D(MX)_relvs D(HX)

correlation was also basically linear with a slope of ∼1.0.

Furthermore, deviations are apparently attenuated by changes

in the ancillary ligand bond strengths, as evidenced by variations

in D(RuP) as X was changed.⁶⁴In essence, the data comply

with Pauling’s electroneutrality principle,⁷⁶where the more

polarizable L_nM fragment electronically adjusts in response to

changes in X, thereby maintaining ∂D(MX)/∂D(HX) near unity

while minimizing the total energy of L_nMX. In a similar vein,

variation of both X(σ)fTi(dσ) and X(pπ)fTi(dπ) (X ) O, N)

The most important feature of the D(TiR)_relvs D(RH)

correlation concerns the slope of the line, which is 1.36 with

the outlying substituents removed, and still greater than one with

those included. For comparison, Jones’ D(RhR)_relvs D(RH)

plot pertaining to Tp′(^tBuCH₂NC)Rh(H)R (Tp′ ) HB(3,5-

diethylpyrazoyl)₃) with the outlying R ) CH₂Ph point removed

has a slope of ∼1.5.³D(MR) typically do not approach D(RH)

in magnitude; consequently, it is curious that differences in

D(MR) are greater than differences in D(RH).

Covalent bonds are typically described by expressions that

contain electronegativity differences, such as the Pauling

equation⁷⁶and Matcha’s variant,⁷⁷or components that describe

electronic differences, such as the electrostatic, covalent, and

transfer factors of the ECT approach developed by Drago.^78-80

Consider the simplest of these, the arithmetic mean formulation

of the Pauling equation that describes D(MR) as a function of

D(RH) and electronegativity parameters (eq 34; M ) L_nM′; ê

t

components of silox and Bu₃SiNH bonding may render the

(silox)₂(^tBu₃SiNH)Ti polarizable as R is varied in (silox)₂(^tBu₃-

SiNH)TiR (2-R), resulting in the generally linear D(MR)_relvs

D(RH) correlation.

2. Thermodynamics of â-H-Elimination. Evidence of â-H-

c

elimination from (silox)₂(^tBu₃SiNH)TiR (2-R, R ) Hex, Pe,

ⁿPr, ⁿBu, ^neoHex, Et, ^cBu) has not been obtained. Construction

of a thermodynamic cycle (eqs 36-38, 25 °C) allows assessment

of the free energy of the reaction 2-Et f 2-H + C₂H₄(eq 39),

D(MR) ) D(RH) + 0.5[D(MM) - D(HH)] +

ê[(ø_M- ø_R)²- (ø_R- ø_H)²] (34)

2-Et + C₂H₄f 2-Vy + C₂H₆

C₂H₆(g) f H₂(g) + C₂H₄(g)

2-Vy + H₂f 2-H + C₂H₄

∆G° ) -2.9 kcal/mol

(36)

is typically in kcal‚mol^-1/EN unit²). Its derivative shows that

deviations from a slope of unity can be ascribed to how

electronegativity differences vary with changes in D(RH) (eq

35). Because of the intrinsic interdependence of ø_R, ø_H, and

∆G° ) +24.2 kcal/mol

(37)

∆G° ) -5.0 kcal/mol (38)

∆G° ) +16.3 kcal/mol (39)

∂D(MR)/∂D(RH) )

1 + ê∂[(ø_M- ø_R)²- (ø_R- ø_H)²]/∂D(RH) (35)

2-Et f 2-H + C₂H₄

D(RH), it is difficult to predict the magnitude or sign of the

deviation.

with the assumption that the ∆G° for ethane dehydrogenation

is reasonable for cyclohexane solution as well as the gas phase

(eq 37).⁸⁴On the basis of these calculations, â-hydride

elimination is expected to be endoergic by ∼16 kcal/mol.

The absence of vacant cis-coordination sites in pseudotetra-

hedral 2-R, the nonlability of the ancillary ligands, and steric

hindrance⁸⁵are all kinetic factors that hamper â-hydride

elimination. Additionally, Chisholm has pointed out that

X(pπ)fM(dπ) interactions raise the energy of the empty orbitals

The lack of curvature and relatively small deviation from a

slope of unity in the D(TiR)_relvs D(RH) correlation bear closer

scrutiny. A few systems with related characteristics have been

investigated thermochemically, and these exhibit strong, covalent

metal-carbon bonds.^63-70By fitting of the D(TiR)_reldata to

the Matcha variant of the Pauling equation and through related

ECT-based procedures, crude estimates of the absolute bond

strengths in (silox)₂(^tBu₃SiNH)TiR (2-R) place D(TiMe) at ∼65

(81) Bennett, J. L.; Vaid, T. P.; Wolczanski, P. T. Inorg. Chim. Acta, in

press. Although ø_Ti∼ 2.7 gave the best fit in the Matcha analysis, the ø_Ti

value is misleading because the Pauling equation and variants are inherently

limited (e.g., D(TiH) - D(TiR) is poorly assessed).

(76) Pauling, L. The Nature of the Chemical Bond, 3rd ed.; Cornell

University Press: Ithaca, NY, 1960.

(77) Matcha, R. L. J. Am. Chem. Soc. 1983, 105, 4859-4862.

(78) Drago, R. S.; Wong, N. Inorg. Chem. 1995, 34, 4004-4007.

(79) Drago, R. S. Applications of Electrostatic-CoValent Models in

Chemistry; Surfside Scientific Publishers: Gainesville, FL, 1994.

(80) Drago, R. S.; Wong, N. M.; Ferris, D. C. J. Am. Chem. Soc. 1992,

114, 91-98.

(82) Boyd, R. J.; Boyd, S. L. J. Am. Chem. Soc. 1992, 114, 1652-1655.

Generally, sp²-carbon centers are considered more electronegative that sp³.

(83) Allen, L. C. J. Am. Chem. Soc. 1989, 111, 9003-9014.

(84) CRC Handbook of Chemistry and Physics; 75th ed.; CRC Press:

Boca Raton, FL, 1994.

(85) Kruse, W. J. Organomet. Chem. 1972, 42, C39-C42.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10709

Figure 5. Standard free energy vs reaction coordinate diagram

corresponding to the 1,2-RH-elimination/addition pathway of Scheme

1.

needed for a 5-coordinate d⁰olefin intermediate.⁸⁶Assuming

an ∼18 kcal/mol barrier for the observed slow insertion of

ethylene into 2-H at 25 °C, the reverse â-H-elimination would

have a ∆G^q∼ 36 kcal/mol, with much of the activation having

a thermodynamic origin. The stabilities of metal alkyls such

as Cp₂MX(^tBu),⁸⁷(Me₂N)₄Ta^tBu,⁸⁶and (^tBu₃SiNH)-

(THF)^tBuTidNSi^tBu₃²²and those of directly related systems^20,23

are probably best understood on the basis of thermodynamic

considerations.⁶⁷

Ground and Transition States Energies. 1. General

Considerations. Figure 5 illustrates how the combination of

kinetic and thermodynamic data in Table 3 and Figure 4,

respectively, may be used to establish the transition state energy

for each 1,2-RH-elimination and -addition event. From the

standard free energy surface, differences in transition state

energies (i.e., ∆G^q_addn(R²H) - ∆G^q_addn(R¹H)), obtained from

∆G^q_elim(2-R¹), ∆G^q_elim(2-R²), and ∆G° (or ∆G°_corr, eq 31)

according to eq 40, permit evaluation of the most critical goal

in this investigationskinetic selectivities for the activation of

an RH vs R′H bond.

Figure 6. Ground and transition state free energies (kcal/mol, 1 M

standard states) of 2-Rⁿ(i.e., 2-R¹+ R²H + ... + L¹+ L²+ ... +

RC₂R′ + ...), 3-Lⁿ(i.e., 3-L¹+ R¹H + R²H + ... + L²+ ... + RC₂R′

+ ...) and 3-RC₂R′ (i.e., 3-RC₂R′ + R¹H + R²H + ... + L¹+ L²+ ...

+ R′′C₂R′′′ + ...) relative to the reference ground state of (^tBu₃-

SiO)₂pyTidNSi^tBu₃(3-py) at 0.0 kcal/mol. Ground states are repre-

sented by shaded columns, ∆G^q_elim(RH)’s with unshaded columns, and

transition state energies for 1,2-RH-elimination/addition by the sum

of the two. Cyclohexane and benzene solvent effects are assumed to

be negligble or equivalent.

composed of the organometallic species plus all of the other

hydrocarbons, dative ligands, and alkynes in a 1 M standard

state, with one caveat; cyclohexane and benzene, the solvents

used for the respective equilibria and rate studies, are considered

to be equivalent in terms of influence on the relative state

energies. For simplicity, these energies are referred to as the

energy of the organometallic species, i.e., 2-R or 3-L.

∆∆G^q_addn) ∆G^q_addn(R²H) - ∆G^q_addn(R¹H) )

∆G^q_elim(2-R²) + ∆G° - ∆G^q_elim(2-R¹) (40)

2. Relative Energy of (silox)₂TidNSi^tBu₃(3). The isolation

of (silox)₂TidNSi^tBu₃(3) was a paramount goal of this project,

since absolute second-order rate constants for the activation of

RH by 3 could then be measured. Although this has not been

achieved, estimates of the free energy pertaining to 3 + R¹H +

The compilation of ground and transition states for each

(silox)₂(^tBu₃SiNH)TiR (2-R) and ground states accorded each

(^tBu₃SiO)₂LTidNSi^tBu₃(3-L) are not readily accommodated

by a standard free energy vs reaction coordinate diagram; hence

the data are presented in a stacked columnar graph in Figure 6.

The ground and transition state energies are presented relative

to the ground state of (^tBu₃SiO)₂pyTidNSi^tBu₃(3-py) at 0.0

kcal/mol. Ground states are graphed in order of decreasing

energy and are represented by shaded columns, the ∆G^q_elim(2-

R)’s from Table 3 are indicated with unshaded columns, and

the sum of the two for a specific 2-R provides the requisite

transition state energy for 1,2-RH-elimination/addition relative

to 3-py. Each relative standard free energy refers to a state

+

R²H + ... + L¹+ L²+ ... RCtCR′ + ... (referred to as the

ground state of 3) can be made. Because addition of RH to 3

should be exoergic for all RH, the relative energy of 3 will be

higher than the energy of 2-^cHex, 19.1 kcal/mol relative to 3-py.

An experiment that yields an upper bound on the energy of

3 is loss of THF-d₈from 3-THF-d₈(3.7 kcal/mol relative to

3-py), which was measured in C₆D₆with 15 or 43 equiv of

THF present ([THF] ) 0.81 M, k ) 5.2 × 10^-3s^-1, ∆G^q)

20.5 kcal/mol; [THF] ) 1.96 M, k ) 4.0 × 10^-3s^-1) and in

neat THF ([THF] ) 12.3 M, k ) 3.4 × 10^-3s^-1) as shown in

eq 41.⁸⁸The data support a dissociative or dissociative

(86) Chisholm, M. H.; Tan, L.-S.; Huffman, J. C. J. Am. Chem. Soc.

1982, 104, 4879-4884.

(87) (a) Buchwald, S. L.; Kreutzer, K. A.; Fisher, R. A. J. Am. Chem.

Soc. 1990, 112, 4600-4601. (b) Buchwald, S. L.; Lum, R. T.; Fisher, R.

A.; Davis, W. M. J. Am. Chem. Soc. 1989, 111, 9113-9114.

(88) Loss of L from 3-L is not always straightforward. Exchange of py-

d₅with (silox)₂(^tBu₃SiNd)Tipy (3-py) is faster than that of 3-THF, an

observation suggestive of an associative substitution. Further studies are

underway by L. M. Slaughter and P. T. Wolczanski.

10710 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

^{C D /THF}8

6

(silox)₂(THF-d₈)TidNSi^tBu₃

3-THF-d₈

24.8 °C

(silox)₂(THF)TidNSi^tBu₃

3-THF

+ THF-d₈(41)

interchange process whose transition state must have an energy

equal to or greater than that of 3, whose maximum can therefore

considered to be 24.2 kcal/mol. The bracketed energy region

where 3 resides (i.e., 19.1 kcal/mol < ∆G°(3) e 24.2 kcal/

mol) can be narrowed further since ∆G° for 2-^cHex h 3 +

^cHexH is likely to be >3 kcal/mol (i.e., 3 is not observed in

solutions containing 2-^cHex and 3₂). Capture of (silox)₂-

TidNSi^tBu₃(3) by THF was considered to be almost barrierless

(i.e., dissociative interchange), leading to ∆G°(3) ) 24(1) kcal/

mol for the purposes of further discussion.

With this estimate of the free energy of (silox)₂TidNSi^tBu₃

(3), ∆G^q_addn(RH) for 3 + RH f 2-R can be roughly assessed

according to eq 42. The ∆G^q_addn(RH) values calculated range

∆G^q_addn∼ ∆G°(2-R, relative to 3-py) +

∆G^q_elim(2-R) - 24(1) kcal/mol (42)

from ∆G^q_addn(H₂) ∼ 9(1) kcal/mol to ∆G^q_addn(^cHexH) ∼ 18(1)

kcal/mol and represent rare, experimentally supported, quantita-

tive estimates of the barrier for activation of a C-H (or H-H)

bond, in this case a 1,2-RH-addition to transient (silox)₂-

TidNSi^tBu₃(3). Calculations by Cundari result in a barrier of

∆Hq_addn∼ 14.5 kcal/mol for CH₄addition to (H₂N)₂TidNH, a

value consistent with the experimental ∆G^q_addn∼ 15 kcal/mol.⁶⁰

The value of ∆G°(3) also allows for a crude estimate of its

imide π-bond strength as 34-40 kcal/mol.⁸⁹

3. Intramolecular Competitions. While the majority of

the selectivity data obtained was determined from the relative

ground state and ∆G^q’s for 1,2-RH-elimination according to

Figures 5 and 6, some intramolecular competitions were used

to check the data. Transient (silox)₂TidNSi^tBu₃(3) was

generated by thermolysis (25 °C) of (silox)₂(^tBu₃SiNH)Ti^cPr

(2-^cPr) in the presence of 20 equiv of toluene in C₆D₁₂solution

(eqs 43 and 44). Resonances attributable to 2-Bz and 2-Ar,

Figure 7. 1,2-RH-Elimination/addition reactant and product standard

free energy surfaces (kcal/mol, 1 M standard states) of 2-Rⁿ(i.e., 2-R¹

+ R²H + ... + L¹+ L²+ ... + RC₂R′ + ...) and 3 (i.e., 3 + R¹H +

R²H + ... + L¹+ L²+ ... + RC₂R′ + ...) represented by parabolas

defined by the respective standard free energies (∆G°(2-Rⁿ), ∆G°(3))

and corresponding positional coordinates, x(2-Rⁿ) and x(3). Transition

states (∆G^TS) are located at x(TS_n) such that the extent of 1,2-RH-

n

elimination can be assessed as {x(TS_n) - x(2-Rⁿ)}/{x(3) - x(2-Rⁿ)},

and the extent of 1,2-RH-addition, as {x(3) - x(TS_n)}/{x(3) - x(2-

Rⁿ)} (see Appendix).

considered as para- and meta-activation products due to the

presence to two diagnostic methyl signals,²⁰were observed by

¹H NMR spectroscopy. At 34% conversion of 2-^cPr, a kinetic

ratio of 2-Bz:2-Ar of 1:24 was observed. Since detection of

2-Bz was not accomplished under true early conversion condi-

tions (usually <10%), activation of an aryl C-H bond over a

benzylic one is preferred by <-1.9 kcal/mol. By assuming

∆G^q_elim(2-Ph) ∼ ∆G^q_elim(2-Ar), eq 40 predicts an aryl vs benzyl

activation preference of -2.5 kcal/mol, a value in decent

agreement. In a related competition for 3 by ethylene, formation

C₆D₁₂

_{25 °C}8

(silox)₂(^tBu₃SiNH)Ti^cPr

2-^cPr

+ ^cPrH (43)

t

[(silox)₂TidNSi Bu₃]

of the metallacycle (^tBu₃SiO)₂(^tBu₃SiN)TiCH₂CH₂(3-C₂H₄) was

favored over (silox)₂(^tBu₃SiNH)TiCHdCH₂(2-Vy) by -1.4

kcal/mol, while ∆G° ) -1.2 kcal/mol. In an initial rate study,

3

t

3 +

f

+

(silox)₂( Bu₃SiNH)TiBz

C₇H₈

loss of ethylene from 3-C₂H₄occurred with ∆G^q) 22.1(2)

diss

2-Bz

(20 equiv)

kcal/mol, in excellent agreement with -∆G° + ∆G^q_elim(2-Vy)

(silox)₂(^tBu₃SiNH)TiC₆H₄Me

+ ∆∆G^q

) ∆G^q_diss(C₂H₄) ) 22.2 kcal/mol (eq 40).²⁴

(44)

addn

2-Ar

Parabolic Model of the Reaction Coordinate. 1. General

Considerations.²⁰The standard free energy vs reaction coor-

dinate diagram of Figure 5 is depicted as overlapping parabo-

las^56,93,94corresponding to the ground state 2-R¹+ R²H + ...

+ L¹+ L²+ ... + RC₂R′ + ... (solid line, state 2-R¹) and the

intermediate state 3 + R¹H + R²H + ... + L¹+ L²+ ... +

RC₂R′ + ... (state 3) in Figure 7. The diagram assumes similar

parabolic surfaces for 2-R¹, 2-R², etc., a single surface for 3,

(89) Consider Figure 6 and the addition of MeH to (silox)₂TidNSi^tBu₃

(3): 3 + MeH h 2-Me. From the assessment of ∆G°(3), ∆G°_reacn∼ -9

kcal/mol, and assuming D(TiMe) ∼ 65-71 kcal/mol^68,81and ∆S°_reacn

∼

-30 eu (T∆S° ∼ -9 eu), ∆H°_reacncan be estimated as -18 kcal/mol and

applied to the following (D(HMe) ) 105 kcal/mol, D(NH) ) 92 kcal/mol):

60,62

∆H°_reacn∼ D(TidN) + D(HMe)-[D(TiMe) + D(NH) + D(TiN)];

∆H°_reacn∼ -18 ∼ D(π(TiN)) + 13 - D(TiMe). The difference between

the titanium-nitrogen interaction in 3 vs 2-Me (i.e., D(TidN) - D(Ti-

N))⁶⁰is taken to be the imide π-bond enthalpy, D(π(TiN)) ∼ 34-40 kcal/

mol. It is greater than gas phase estimates of D(TidNH⁺) ∼ 25 kcal/mol,⁹⁰

(91) (a) Schrock, R. R.; Sharp, P. R. J. Am. Chem. Soc. 1978, 100, 2389-

2399. (b) Schrock, R. R.; Messerle, L. W.; Wood, C. D.; Guggenberger, L.

J. J. Am. Chem. Soc. 1978, 100, 3793-3800.

(92) Gable, K. P.; Phan T. N. J. Am. Chem. Soc. 1993, 115, 3036-

3037.

(93) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334-338.

(94) Thornton, E. R. J. Am. Chem. Soc. 1967, 89, 2915-2927.

consistent with ∆G^q∼ D(π(TadC)) g 20 kcal/mol for (MeC₅H₄)-

rot

CpMeTadCH₂,⁹¹and somewhat less than Gable’s estimate of ∼50 kcal/

mol for D(π(RedO)) in Cp*ReO₃.⁹²

(90) (a) Clemmer, D. E.; Sunderlin, L. S.; Armentrout, P. B. J. Phys.

Chem. 1990, 94, 3008-3015. (b) Clemmer, D. E.; Sunderlin, L. S.;

Armentrout, P. B. J. Phys. Chem. 1990, 94, 208-217.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10711

2. 1,2-RH-Addition to (silox)₂TidNSi^tBu₃. Figure 6

highlights the general correspondence between ground state and

transition state energies for 1,2-RH-elimination/addition. The

ease of RH addition to (silox)₂TidNSi^tBu₃(3) generally parallels

the respective ground state energies of 2-R, indicating the

predominant factor in R-H bond activation selectivity is the

strength of the titanium-carbon bond that will eventually be

formed. Therefore, the slope of the D(TiR)_relvs D(RH) line

(Figure 4) is the critical indicator of selectiVity. If all versions

of eq 27 were effectively thermoneutral (i.e., ∂D(TiR)_rel/∂D(RH)

∼ 1), there would be no thermodynamic impetus behind the

kinetic selectivity in hydrocarbon activation. Lacking any

pronounced secondary effects (i.e., steric etc.), such a system

would exhibit stochastic selectivity.

Figure 8 graphs ∆G^q_addn(RH, calculated with ∆G°(3) ) 24

kcal/mol) as a function of ∆G°(2-R) and reveals the effect of

the free energy on ∆∆G^q_addn. Generally, the greater the stability

of 2-R, the easier the activation of the corresponding hydro-

carbon. The relationship is fairly linear (eq 45) for the sp³-

Figure 8. Linear free energy relationship as defined by ∆∆G^q

)

addn

∆G^q_addn(R²H) - ∆G^q_addn(R¹H) ∼ â{∆G°(2-R²) - ∆G°(2-R¹)} (eq 45,

referenced to ∆G°(3) ) 24 kcal/mol, â(sp³) ) 0.77, R²) 0.98).

∆∆G^q_addn) ∆G^q_addn(R²H) -

∆G^q_addn(R¹H) ∼ â{∆G°(2-R²) - ∆G°(2-R¹)} (45)

and two variables to interconnect them: (1) the relative standard

free energies accorded the ground states of 2-Rⁿand 3, as

determined via the measured equilibria and above estimates of

∆G°(3); (2) the reaction coordinate defined by the disposition

of 2-Rⁿ(x(2-Rⁿ)) relative to 3 (x(3)). The use of surfaces with

like curvature for all 2-Rⁿis necessary to employ this simple

model, and there exists reasonable justification for this assump-

tion. Since the reaction coordinate for each 2-Rⁿcontains

elements of Ti-C and N-H bond-breaking accompanied by

TiN(π) and C-H bond-making, the corresponding parabolic

surfaces are expected to be quite alike. This is especially the

case in the relatively low energy region relevant to these

reactions, where ∆G^q_elim(2-Rⁿ) ∼ 20 kcal/mol << D(TiRⁿ) ∼

65 kcal/mol^68,81or D(NH) ∼ 92 kcal/mol^60,62and ∆G^q_addn(RⁿH)

<< D(RⁿH) etc. Under these circumstances the relevant bond

stretches, bends, etc. are well described as harmonic oscillators,

leading to a surface that is parabolic as a sum of such

components. Furthermore, if the curvatures were different, the

surface would be expected to be the steepest for 2-Ph and 2-Vy,

which have the strongest D(Ti-R); a greater steepness would

lead to greater transition state energies and higher respective

substituents, and the line of slope â ) 0.77 (R²) 0.98) provides

a numerical indication of the thermodynamic influence on

transition state energies. In the parabolic model of Figure 7, a

substantial slope occurs where differences in 1,2-RH-addition

activation energies (e.g., ∆G^q- ∆G^q_3a) approach the differ-

1a

ences in the ground states of the products (e.g., ∆G°(2-R¹) -

∆G°(2-R³)), provided the ground states of 2-R¹and 2-R³are

positionally similar (i.e., x(2-R¹) ∼ x(2-R³)), and the dispositions

of the respective transition states are distinct from 3 (i.e., x(TS₁)

and x(TS₃) are not near x(3), and ∂∆G°(3)/∂x(3) << 0). When

RH addition to 3 is not very “early” along the reaction

coordinate, ∂∆G^q_addn/∂∆G° will be relatively linear and approach

unity, as in this situation.

1,2-R(sp³)H-addition to transient (silox)₂TidNSi^tBu₃occurs

with a transition state of balanced character, consistent with the

relatively low energy estimate of 3 and correspondingly high

estimates of ∆G^q_addn(R(sp³)H) ∼ 14.7-18.1 kcal/mol. Kinetic

isotope effect data that portray the transition state as

symmetricscarbon-hydrogen bond-breaking and metal-carbon

bond-making are both importantscorroborate this depiction. No

significant deviations occur within the sp³-group, suggesting that

steric factors are similar in both ground and transition states,

entropic influences are relatively constant, and variations in x(2-

Rⁿ) must correlate with ∆G°(2-Rⁿ).

∆G^q

and ∆G^q_elim, yet the opposite is obserVed for these

addn

hydrocarbyls. The model also assumes that transfer between

the surfaces is adiabatic and that coupling between the reactant

(2-R) and product (3) surfaces occurs to the same degree for

each 2-Rⁿand can be neglected. While the reaction coordinate

is not measured directly, by assuming similar surfaces for 2-R¹,

2-R², etc. relative to a common intermediate surface (3), one

can calculate x(2-Rⁿ) relative to x(3), because knowledge of

∆G^q_elim(RⁿH) affords a unique disposition of each 2-Rⁿparabola

relative to that of 3 (see Appendix).

In Figure 7, the left surfaces are centered at x(2-Rⁿ) and

correspond to each 2-Rⁿstate at lower energy than the upper

right surface that describes state 3, whose position is denoted

as x(3). The respective transition states for 1,2-RH-elimination

are positioned at x(TS_n), and the extent of reaction⁹⁵at that point

can be assessed as {x(TS_n) - x(2-Rⁿ)}/{x(3) - x(2-Rⁿ)},

providing a numerical estimate of the “early” or “late” character

of the 1,2-RH-elimination. Likewise, the extent of 1,2-RH-

addition can be assessed as {x(3) - x(TS_n)}/{x(3) - x(2-Rⁿ)}.

Two other groups of hydrocarbyls are evident from the plot

in Figure 8: sp²-hybridized derivatives, including 2-^cPr, and

benzyl and mesityl. It is more difficult than expected (∼1-2

kcal/mol) for 3 to activate the benzylic bonds of toluene and

mesitylene relative to sp³C-H bonds, yet clearly much of the

aforementioned 6-7 kcal/mol of ground state stabilization

accorded 2-Bz, and by inference 2-Mes, translates to their

respective transition states. Noting that the ground states of

2-Me and 2-Bz are similar in energy, consider the surfaces of

2-R¹(R¹) Me) and 2-R²(R²) Bz) in Figure 9 to be

representative. ∆G^q_addn(MeH) is lower than ∆G^q_addn(BzH), i.e.,

∆G^q_1a< ∆G^q_2a, because a higher transition state for the benzyl,

∆G(TS₂), results from a lesser disposition along the reaction

coordinate, i.e., x(2-R²) < x(2-R¹). As another example,

transition states for 1,2-RH-addition of mesitylene and ethane

(95) For views of related reaction coordinates, see: (a) Crabtree, R. H.;

Holt, E. M.; Lavin, M. E.; Morehouse, S. M. Inorg. Chem. 1985, 24, 1986-

1992. (b) Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32, 789-

805. (c) Bu¨rgi, H. B.; Dunitz, J. Acc. Chem. Res. 1983, 16, 153-161.

to 3 are approximately equal (i.e., ∆G^TS∼ ∆G^TS), despite

ground state differences that clearly favor mesityl activation (i.e.,

∆G°(2-Mes) < ∆G°(2-Et)). Given the constraints of the model,

Mes

Et

10712 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

Figure 9. Depiction of transition states for 1,2-R(sp³)H- and 1,2-

(R(sp²)H-addition/elimination that reveal a more compressed reaction

coordinate (and correspondingly less reorganizational energy) for

H-transfer to an sp²-hydrocarbyl.

the similar ∆G^q_addnvalues necessarily derive from x(2-Mes) <

x(2-Et), the relative dispositions of the hydrocarbyl ground

states.

Figure 10. ∆G^q_addn(RH) vs {x(3) - x(TS_n)}/{x(3) - x(2-Rⁿ)}, a

measure of the reaction coordinate derived from the parabolic model

(see Appendix). The line is a least-squares fit to the sp³data (m )

110, R²) 0.99).

The origin of the displacement in reaction coordinate cannot

be discerned from the model, but the probable weak η³-

coordination of the benzyl and mesityl substituents could

dramatically change x(2-Bz) and x(2-Mes) from an sp³-hydro-

carbyl position. η³-Coordination may result in an R-carbon that

is significantly further from the amide hydrogen, one that

requires greater bond length and angular changes than those of

a simple hydrocarbyl in achieving the transition state geometry.

In essence, the additional olefinic interaction of the η³-

coordination mode provides a greater thermodynamic impetus

for benzylic activation that is partially offset by a greater reaction

coordinate. It is also conceivable that inductive influences of

the phenyl and 3,5-dimethyphenyl substituents on 2-CH₂R′ may

provide differing stabilizations to the ground and transition states

of the repective benzyl and mesityl species, thereby changing

the nature of reactant and product surface coupling, but such

effects are beyond the predictability of the parabolic model and

are exceedingly difficult to experimentally address.⁹⁶

(silox)₂(^tBu₃SiNH)TiPh (2-Ph) could be interpreted as having

either a ground state destabilization or transition state stabiliza-

tion of ∼1.5 kcal/mol, factors that would also have to be applied

to 2-^cPr and 2-Vy. While comprising a small data set, the two

sp²-substituents and cyclopropyl appear to parallel the line in

Figure 8, indicating that ground state influences on the transition

states of the sp²- and sp³-substituents are the same (i.e., â(sp³’s)

∼ â(sp²’s)), albeit with different intercepts. Transition state

stabilization via p-orbital participation is often invoked to

explain speedier rates of sp²-substituents in various reactions.

Greater coupling between reactant and product free energy

surfaces would accommodate this rationale, but a simpler

explanationsone that is less subject to interpretationsis pro-

vided by the reaction coordinate. The sp²-hybridized substit-

uents (C(sp²) r_cov∼ 0.67 Å) possess inherently shorter bond

lengths than corresponding sp³-derivatives (C(sp³) r_cov∼ 0.77

Å),⁷⁶and a correspondingly compressed reaction coordinate for

1,2-RH-elimination/addition will translate into swifter additions

(i.e., {x(3) - x(TS(sp²))}/{x(3) - x(2-R(sp²))} < {x(3) - x(TS-

(sp³))}/{x(3) - x(2-R(sp³))}. Figure 9 illustrates this principle

by depicting symmetric transition states for R(sp³)H and R(sp²)H

1,2-RH-addition/elimination that indicate H-transfer occurs over

a shorter distance for the latter, provided the R‚‚‚H‚‚‚N distance

is greater than d(Ti-N). As a consequence of the more

compressed reaction coordinate, less reorganizational energy is

needed for a 1,2-R(sp²)H-addition or -elimination event, which

translates into generally swifter rates.

To emphasize the interdependence of 1,2-RH-addition se-

lectivities on ground state energies and the reaction coordinate,

Figure 10 plots ∆G^q_addn(RH, calculated with ∆G°(3) ) 24 kcal/

mol) vs extent of reaction, {x(3) - x(TS_n)}/{x(3) - x(2-Rⁿ)}

(see Appendix). The latter is <0.5 for all cases,⁹⁷implicating

somewhat early transition states, and an earlier ∆G^TSgenerally

n

corresponds to a swifter addition of RH. The data again break

roughly into the same three groups: sp³-substituents (0.44-

0.47), benzylic substituents (∼0.44), and sp²-substituents,

including 2-^cPr (0.42-0.44). The denominators of the positional

parameters differentiate the groups, but the slope within each

should be similar in this particular regime (i.e., x(TS_n) are not

near x(3), and ∂∆G(3)/∂x(3) << 0). Assuming a relatively

constant {x(3) - x(2-Rⁿ)} within each group, it may be inferred

that {x(3) - x(TS_n)} changes are primarily due to ground state

energy differences of 2-R. Within the sp³-substituents, the

dependence of ∆∆G^q_addnon ∆{x(3)-x(TS_n)}/{x(3) - x(2-Rⁿ)}

is quite striking, with R²) 0.99. While the sample sizes are

small, it may be inferred from the graph that similar depend-

encies affect the sp²-groups and benzylic groups; differing

ground states influence the sp³, sp², and benzylic sets of {x(3)

- x(TS_n)}/{x(3) - x(2-Rⁿ)} to the same degree.

In summary, the 1,2-RH-addition/elimination (R ) hydro-

carbyl) preferences may be rationalized solely on the basis of

the relative free energies of 2-Rⁿand an understanding that the

putative four-center transition state has geometric parameters

influenced by the ground state structure of 2-R (e.g., d(Ti-

R)). Additional or alternative effects, such as those affecting

coupling of reactant and product free energy surfaces (i.e.,

p-orbital participation of the sp²-substituents), may be present

but are unnecessary for interpretation of the data; computational

support of such influences has not been discovered.^57-61

Dihydrogen addition has been excluded in part on the basis

that H-binding via an s-orbital is intrinsically different from

that of spⁿ-hybridized substituents. Normally, the nondirec-

tionality of the s-orbital is held responsible for faster rates in

(96) Many aryl substituents (X) change the electron-donating or -with-

drawing capacity of CH₂Ph or Ph, but appropriate D(H-C₆H₄X) and D(H-

CH₂C₆H₄X) are unknown and assumed to be equivalent to D(H-Ph) and

D(H-CH₂Ph) from Benson additivity relationships (Benson, S. D. Ther-

mochemical Kinetics; John Wiley & Sons: New York, 1968). The origin

and interpretation of such substituent effects are thereby rendered moot.

(97) Since the parabolic surfaces of 2-R and 3 are equivalent (see

Appendix), the transition state of a thermoneutral reaction is at x(TS_n) )

0.5, and the transition states of endothermic and exothermic reactions are

necessarily positioned at <0.5 and >0.5, respectively.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10713

Figure 11. Linear free energy relationship as defined by ∆∆G^q

)

Figure 12. ∆G^q_elim(2-R) vs {x(TS_n) - x(2-Rⁿ)}/{x(3) - x(2-Rⁿ)}, a

measure of the reaction coordinate derived from the parabolic model

(see Appendix). The line is a least-squares fit to the sp³data (m ) 32,

R²) 0.78).

elim

∆G^q_elim(2-R²) - ∆G^q_elim(2-R¹) ∼ â′{∆G°(2-R²) - ∆G°(2-R¹)} (eq 46,

referenced to ∆G°(3) ) 24 kcal/mol, â(sp³) ) -0.23, R²) 0.83).

various insertions (e.g., L_nM(olefin)H f L_nMR), and the

parabolic model could assimilate this logic through greater

coupling of the 2-R and 3 surfaces. The 3 + H₂h 2-H reaction

coordinate is expected to be greatly compressed relative to the

hydrocarbyls, again supporting swifter H₂addition. In contrast,

the difficulty in polarizing the H-H bond, a critical factor in

the assessment of ground state energetics, may slow dihydrogen

addition/elimination. In conclusion, parabolas used to model

2-R (R ) hydrocarbyl) may poorly describe 2-H.

3. 1,2-RH-Elimination from 2-R. By the principle of

microscopic reversibility, the preceding model provides comple-

mentary trends in 1,2-RH-elimination from 2-R. The relation-

ship between ∆G^q_elim(2-R²) and ∆G°(2-R) in Figure 11 can be

interpreted within the same three groups: sp³- and sp²-

substituents and 2-CH₂Ar. A moderate correlation within the

sp³-group (eq 46, R²) 0.83) shows that ground state influences

the sp³-data that {x(3) - x(2-Rⁿ)} is fairly constant within the

group, higher ground state energies for 2-R lead to faster

eliminations, and the slope (R²) 0.78) is a consequence of

{x(TS_n) - x(2-Rⁿ)} changes that reflect ∆G°(2-Rⁿ) differences.

As a corollary, 2-Bz, 2-Me, and 2-Ph have similar extents of

reaction values, but the denominators {x(3) - x(2-Rⁿ)} for each

are different, with x(2-Bz) < (2-Me) < (2-Ph) leading to

∆G^q_elim(2-Bz) > ∆G^q_elim(2-Me) > ∆G^q_elim(2-Ph) disparities that

are greater than predicted on the basis of a linear free energy

relationship alone! Again the ∼0.1 Å difference in covalent

radii account for the more compressed reaction coordinate for

the sp²-substituents and 2-^cPr, resulting in generally swifter

eliminations. Weak η³-coordination by aryl groups in 2-Bz and

2-Mes should elongate the reaction coordinate, rendering slower

1,2-RH-elimination rates.

4. Other Correlations. Correlations of proton affinity with

∆G^q_elim(2-R) were previously used to suggest the presence of

alkane intermediates,^20,21and related studies in this system did

little but support the contention of sp²vs sp³substrate classes.

Similarly, attempts to correlate 1,2-RH-addition and -elimination

processes with gas- and solution-phase pK_a’s manifested the

three standard groups (i.e., sp³, sp², and 2-^cPr, 2-CH₂Ar).

Significant scatter was also evident, and no obvious conclusions

could be reached, in concert with experimental and calcula-

tional^57,60,61examinations that have failed to elicit evidence of

charge build-up on crucial intermediates; these reactions are best

considered as concerted four-center processes.

∆∆G^q_elim) ∆G^q_elim(2-R²) -

∆G^q_elim(2-R¹) ∼ â′{∆G°(2-R²) - ∆G°(2-R¹)} (46)

on the 1,2-RH-elimination are small (â′ ) -0.23) yet clearly

indicative of speedier eliminations from 2-R of higher ground

states. This is expected from the parabolic model, since x(TS_n)

are not near x(3) and ∂∆G(3)/∂x(3) << 0. Under these

circumstances, the 1,2-RH-eliminations are not particularly late,

and ∆G^TS- ∆G^TSwill be less than, but approach, the

2

1

respective ground state differences, resulting in a small ∆∆G^q

elim

dependence on ground state differences. The remaining groups

have too few members to manifest a trend yet are presumably

differentiated on the basis of disparate reaction coordinates, with

the 2-CH₂Ar eliminating slower than predicted by an all-

encompassing linear free energy relationship and the sp²-

substituents undergoing RH loss faster than expected on that

basis. The greater scatter in the data as viewed from the

elimination standpoint is consistent with a lesser dependence

on ground state in a moderately late reaction coordinate, where

subtler influences will be magnified.

Conclusions

Mechanistic Overview. The mechanism of 1,2-RH-elimina-

tion from (silox)₂(^tBu₃SiNH)TiR (2-R) and of 1,2-RH-addition

to (silox)₂TidNSi^tBu₃(3) is portrayed in Scheme 1 and Figure

5 and summarized as follows: (1) 1,2-RH-elimination occurs

via a four-center transition state in which the amide/imide

nitrogen, the transferring hydrogen, and the R-carbon of R are

relatively linear, as large KIE’s indicate and calculations

corroborate; (2) activation parameters, KIE evidence, and

thermodynamic estimates accorded 2-R h 3 + RH suggest a

concerted process, with balanced amounts of N-H and Ti-R

bond-breaking and C-H bond-making with little charge build-

up; (3) intermediate 3 or solvated 3 is a potent electrophile and

As illustrated by the ∆G^q_elim(2-Rⁿ) vs extent of reaction plot

in Figure 12, the benzylic substituents and sp²-substituents are

shown to be distinct within the somewhat late (i.e., {x(TS_n) -

x(2-Rⁿ)}/{x(3) - x(2-Rⁿ)} > 0.5)⁹⁷character of the 1,2-RH-

elimination transition states. The extent of reaction is near 0.5,

indicative of transition states possessing the balanced character

implied by the KIE experiments. Again, one can infer from

2

uses a d_z/p_zhybrid orbital to attack the pair of electrons in a

substrate C-H bond; (4) 1,2-RH-addition occurs concomitant

10714 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

Table 4. Pairwise Kinetic Selectivity of (silox)₂TidNSi^tBu₃(3) for R¹H vs R²H Expressed as ∆∆G^q

Calculated via Eq 40 (kcal/mol at 24.8 °C)^a-c

) ∆G^q_addn(R²H) - ∆G^q_addn(R¹H) As

addn

R¹H

R²H

^cHexH

^cPeH

ⁿBuH

^neoHexH

EtH

BzH

^cBuH

MesH

MeH

PhH

^cPrH

VyH

C₂H₄

H₂

^cHexH

^cPeH

ⁿBuH

^neoHexH

EtH

0.0

-0.5

-1.4

-1.7

-1.8

-1.9

-2.6

-3.4

-4.3

-5.5

-6.1

-7.5

-8.9

0.5

0.0

1.4

0.9

0.0

1.4

0.9

0.0

1.7

1.2

0.3

0.0

-0.1

-0.2

-0.9

-1.7

-2.6

-3.8

-4.4

-5.8

-7.2

1.8

1.3

0.4

0.1

1.9

1.4

0.5

0.2

0.1

0.0

-0.7

-1.5

-2.4

-3.6

-4.2

-5.6

-7.0

2.6

2.1

1.2

0.9

0.8

0.7

0.0

-0.8

-1.7

-2.9

-3.5

-4.9

-6.3

3.4

2.9

2.0

1.7

1.6

1.5

0.8

0.0

-0.9

-2.1

-2.7

-4.1

-5.5

4.3

3.8

2.9

2.6

2.5

2.4

1.7

0.9

5.5

5.0

4.1

3.8

3.7

3.6

2.9

2.1

1.2

0.0

-0.6

-2.0

-3.4

6.1

5.6

4.7

4.4

4.3

4.2

3.5

2.7

1.8

0.6

0.0

-1.4

-2.8

7.5

7.0

6.1

5.8

5.7

5.6

4.9

4.1

3.2

2.0

1.4

0.0

-1.4

8.9

8.4

7.5

7.2

7.1

7.0

6.3

5.5

4.6

3.4

2.8

1.4

0.0

-0.9

-1.2

-1.3

-1.4

-2.1

-2.9

-3.8

-5.0

-5.6

-7.0

-8.4

0.0

-0.3

-0.4

-0.5

-1.2

-2.0

-2.9

-4.1

-4.7

-6.1

-7.5

-0.3

-0.4

-0.5

-1.2

-2.0

-2.9

-4.1

-4.7

-6.1

-7.5

BzH

0.0

^cBuH

MesH

MeH

PhH

-0.1

-0.8

-1.6

-2.5

-3.7

-4.3

-5.7

-7.1

0.0

^cPrH

VyH

C₂H₄

H₂

-1.2

-1.8

-3.2

-4.6

^aAzametallacyclobutane (silox)₂(^tBu₃SiN)TiCH₂CH₂formation according to eq 22 is included and listed as R¹H or R²H ) C₂H₄. ^bThe selectivity

is given on a per molecule basis. ^cFor kinetic selectivity on a per hydrogen basis, ∆G^q

may be generated by using ∆G°_corr(eq 31) in eq 40.

addn

Table 5. Measured and Estimated Relative Rates and ∆G^qfor 1,2-RH-Elimination from Group 4 and 5 XY(^tBu₃SiNH)MR Complexes

elim

group 4 compound

k_rel,^as^-1

∆G^q_elim, kcal/mol

group 5 compound

k_rel,^as^-1

∆G^q_elim, kcal/mol

(silox)₂(^tBu₃SiNH)TiMe

460

5500

24.9

23.0

(^tBu₃SiNH)₂(^tBu₃SiN)VMe^b

(^tBu₃SiNH)₂(^tBu₃SiN)VPh^b,c

12

140

27.6

25.8

(silox)₂(^tBu₃SiNH)TiPh

(^tBu₃SiNH)₃ZrMe

(^tBu₃SiNH)₃ZrPh

1

22

29.4

27.1

(^tBu₃SiNH)₃HfMe^b,d

2.0 × 10^-3

34.0

31.3

(^tBu₃SiNH)₂(^tBu₃SiN)TaMe^b

(^tBu₃SiNH)₂(^tBu₃SiN)TaPh^b

1.8 × 10^-5

7.7 × 10^-4

37.4

34.7

(^tBu₃SiNH)₃HfPh^b

7.9 × 10^-2

^aRate constants were recalculated at 97 °C and corrected for the number of NH units per molecule relative to (^tBu₃SiNH)₃ZrMe. ^b∆S^qassumed

c

d

to be -10 eu. k_Ph/k_Meassumed to be similar to that of the titanium case. k_Ph/k_Meassumed to be similar to that of the tantalum case.

with or subsequent to RH-binding via the four-center transition

state; (5) selectivities of C-H bond activation parallel the

strengths of the Ti-R bonds formed, with additional perturba-

tions due to a more compressed reaction coordinate for sp²-

substrates and a slightly elongated one for benzylic and

presumably allylic activations; (6) the steric “pocket” in 3 must

be quite open to accommodate the breadth of substrates

observed, including secondary C-H bond activation of cyclo-

alkanes, which had not previously been noted for 1,2-RH-

abstractions that intrinsically favor activation of the saturated

products, because their carbon-hydrogen bonds are weaker than

those of the methane feedstock.^2,99Unlike these heterogeneous

catalysts, the low-temperature, stoichiometric system herein

exhibits greater selectivity for unsaturated hydrocarbons;

H-CHdCH₂addition is preferred over H-CH₃addition by 2.7

kcal/mol, while benzene activation is favored over methane

activation by 0.9 kcal/mol. Should catalysts be developed that

operate via 1,2-RH-addition to either homogeneous or hetero-

geneous MdX functionalities, or by related concerted additions,

it is the latter selectivity that will ultimately present the most

difficult problems in methane conversion.^2,100Selectivity often

originates in the steric features of homogeneous⁴and biologi-

cal¹⁰¹systems, yet the trends in C-H activation by (silox)₂-

TidNSi^tBu₃svirtually opposite those in heterogeneous Rideal-

type processessare inherent to the concerted character of the

reaction and its exothermicity, which derives from the strength

of the titanium-carbon bonds formed.

c

additions aside from those of PrH; (7) by inference, similar

characteristics apply to related Ti, Zr, V, and Ta systems. 1,2-

RH-Elimination studies of the allylic derivatives were not

available for study because of substantial ground state stab-

lilizations derived from η³-coordination. Alkynyl derivatives

were also not amenable to investigation, because (silox)₂(^tBu₃-

SiNH)TiCtCR (2-CtCR) derivatives are thermodynamically

unstable with respect to corresponding [2+2] addition products,

azametallacycles (silox)₂(^tBu₃SiN)Ti(RCdCH) (3-HC₂R). Po-

tentially interfering cyclometalation^20,98reactions were not

observed for (silox)₂TidNSi^tBu₃(3), presumably because the

geometric requirements for an electrophilic attack on the C-H

2. Comparisons. In a comparison of the selectitivites

exhibited by this system, two observations are particularly

striking. The general activation trends evident in Table 5 are

roughly present in all concerted activations: sp²-substrates ∼

^cPrH > sp³-primary alkanes > sp³-cycloalkanes. Some variation

in the position of benzylic and allylic activations with respect

to these groups has been noted. Quantitative and qualitative

selectivities in oxidative additions of RH to transient [HB(3,5-

t

bonds of silox or Bu₃SiN were prohibitive.

Selectivities for R¹H vs R²H Activation by (silox)₂-

TidNSi^tBu₃(3). 1. Ramifications. While the parabolic model

is adequate to detail the C-H bond activation process, what

are the ramifications of the selectivities on a per molecule basis?

As Table 4 indicates, methane activation is preferred by 1.7

kcal/mol over ethane activation, by 2.0 kcal/mol over terminal

activation of butane (and presumably propane), and by 2.9 and

3.4 kcal/mol over activation by ^cPeH and ^cHexH, respectively.

These significant selectivities obviate the most critical difficulty

encountered by heterogeneous metal oxide catalysts that promote

the oxidative coupling of methane. Rideal-type mechanisms

operative in [MO_n]_xcatalysts are predicated on H atom

t

dimethylpyrazoyl)₃]Rh(CNCH₂Bu),³Cp*MPMe₃(M ) Rh,

(98) (a) Rothwell, I. P. Acc. Chem. Res. 1988, 21, 153-159. (b) Rothwell,

I. P. Polyhedron 1985, 4, 177-200.

(99) Labinger, J. A.; Ott, K. C. J. Phys. Chem. 1987, 91, 2682-2684.

(100) For comments on the limitations of MeH activation in Rideal

processes, see: Labinger, J. A. Catal. Lett. 1988, 1, 371-376.

(101) (a) Watanabe, Y.; Groves, J. T. In The Enzymes, 3rd ed.; Academic

Press: New York, 1992. (b) Stewart, L. C.; Klinman, J. P. Annu. ReV.

Biochem. 1988, 57, 551-592.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10715

Ir),¹⁰²σ-bond metatheses of RH with Cp*₂ScR,¹⁵Cp*₂LuR,¹⁶

and Cp*₂ThCH₂CMe₂CH₂,¹⁷and transient metal imido systems

that also display 1,2-RH-additions all conform to this trend.¹⁰³

Since exoergic C-H bond activation events necessarily

involve the formation of strong metal-carbon bonds, similar

activation trends in seemingly disparate systems probably stem

from the same factors. The strength of the metal-carbon bonds

formed is the primary factor that dictates selectivities in all

mononuclear systems exhibiting concerted processes, with some

moderation by other influences such as the reaction coordinate

and steric effects.

Weaker metal-carbon bonds in the first-row elementsrelative

to Zr and Hfspermit the transition state, and its partially

rendered Ti‚‚‚R bond, to be more easily achieved. The ∼4-5

kcal/mol difference in 1,2-RH-elimination rates between Zr and

Hf is less easily reasoned but is not unusual. Such disparities

are usually attributed to slightly stronger Hf-R bonds,⁶⁷

augmented by relativistic effects.¹⁰⁴In addition to ground state

influences, the shorter covalent radius of titanium (Ti r_cov

∼

1.32 Å) derivatives leads to a compressed reaction coordinate

relative to its Zr (Zr r_cov∼ 1.45 Å) and Hf (Hf r_cov∼ 1.44 Å)

congeners and correspondingly swifter rates.

Values of ∆∆G^q(R¹H vs R²H) for 1,2-RH-addition to (silox)₂-

TidNSi^tBu₃(3) span a greater range than those for oxidative

The slower 1,2-RH-elimination rates upon shifting to group

5 appear to derive primarily from the thermodynamics of the

event. Consider silox and ^tBu₃SiNH to be essentially equivalent

3e^-donors and the related imido ligand ^tBu₃SiN as a 4e^-donor.

As a consequence, (^tBu₃SiNH)₂(^tBu₃SiNd)MR and (^tBu₃SiNH)-

(^tBu₃SiNd)₂M (M ) V, Ta) can be considered 16e^-species,

in contrast to their related 14e^-group 4 derivatives. Wigley

interpreted this disparity in electron count as “π-loading”,¹⁰⁵

because the essential difference between groups 4 and 5 is the

additional N(pπ)fM(dπ) bond of the imido ligand common to

intermediate and ground states. From the standpoint of a

Hammond analysis, it was suggested that “π-loading”, or,

generally, a more electron rich metal center, should lead to

earlier and swifter C-H bond activation events in the case of

(R′NH)(R′Nd)₂M (M ) V, Ta) + RH vs group 4 derivatives.

In essence, “π-loading” or the greater electron density

stabilizes product (R′NH)₂(R′SiNd)MR states relative to in-

termediate (R′NH)(R′Nd)₂M + RH states to a greater degree

than in group 4. As a corollary, the reverse 1,2-RH-elimination

event will be later and slower in group 5 than in group 4.

Unfortunately, the lack of ground state data for the V and Ta

systems leaves this analysis moot, but calculations by Cundari

support this contention, despite commentary to the contrary.⁶⁰

Clearly, decreases in 1,2-RH-elimination rates upon shifting

from group 4 to 5 need to be assessed by a much more detailed

model than is possible from the available kinetic and thermo-

dynamic data and may only be reconciled through careful

calculations.^57-62

Generality of the Parabolic Model. Lacking major steric

influences or critical reactant/product surface coupling, concerted

reactions should generally follow the parabolic model, respond-

ing to thermodynamic and positional effects. While most

recognizable in its application to electron transfer reactions, i.e.,

Marcus theory,¹⁰⁶the application of parabolas to model free

energy surfaces is an old concept in physical organic chemistry

with its origins in Hammond analyses,⁹³related models,⁹⁴and

their predecessors.¹⁰⁷Given this historical precedent, it is

somewhat surprising that the reaction coordinate is often

underappreciated when compared to influences of orbital

character and π-effects, constructs of modern valence bond and

molecular orbital theories. For example, in concerted systems,

hydrocarbon selectivities are recognized as crudely paralleling

the s-character of the C-H bonds activated. Greater s-character

leads to lower activation barriers, perhaps as a consequence of

lesser orbital directionality and greater acidity, although slow

t

addition events in [HB(3,5-dimethylpyrazoyl)₃]Rh(CNCH₂Bu)

(i.e., ∆∆G^qrelative to C₆H₆, -15 °C: PhH [0.0 kcal/mol] >

n

H-CH₂-3,5-Me₂C₆H₃[0.15] > MeH [0.4] > PeH [0.80] >

^cPeH [1.7] > CyH [1.8])³and Cp*MPMe₃(M ) Rh,

∆∆G^q(-60 °C, C₆H₆vs C₆H₁₂) ) 1.2 kcal/mol; M ) Ir,

∆∆G^q(-60 °C) < 0.5 kcal/mol) systems.¹⁰²Relative σ-bond

metathesis C-H bond activation events are more difficult to

analyze due to the lack of quantitative data and the second-

order nature of these reactions, but C₆H₆activation by Cp*₂ScMe

is about 0.8 kcal/mol easier than that by MeH.¹⁵

In unsaturated d⁸L_nM fragments such as [HB(3,5-

t

dimethylpyrazoyl)₃]Rh(CNCH₂Bu) or Cp*MPMe₃(M ) Rh,

Ir), R¹H vs R²H binding is the discriminating event. In contrast,

transition states for C-H activation by d⁰X_nMdN- complexes

exhibit substantial M-R¹(R²) and N-H bond-making, and

R¹(R²)-H bond-breaking in the transition state. Since the C-H

bond is broken to a greater degree than it is during a simple

binding event, the activations evident in the metal imido systems

are intrinsically more selective.

Periodicity and 1,2-RH-Elimination/Addition. Upon

completion of this study, examples of 1,2-RH-eliminations are

known for all group 4^20,24and 5^23,25metals except niobium;

specific cases are listed in Table 5. Two assumptions were made

in converting measured rates to a common temperature (97 °C)

and in estimating missing data: (1) where ∆S^qwas unknown,

a value of -10 eu was assigned, and (2) when either the Me or

Ph derivative was known but not both, the k_Ph/k_Meratio of the

nearest system was used.

Progressing down group 4, 1,2-RH-elimination slows by

increments of ∆∆G^q

∼ 4-5 kcal/mol, whereas a shift to

elim

group 5 within a row constitutes a ∼3 kcal/mol increase in

∆G^q_elim. Overall, the table spans a ∼10⁷decrease in 1,2-RH-

elimination rates upon traversing from titanium (2-R) to

corresponding tantalum complexes. Substitution of silox for

^tBu₃SiNH has already been shown to induce a modest rate

reduction by a factor of 4 in the tantalum system;²³hence this

ligand variation is not considered important.

The trends in Table 5 can be rationalized within the current

model. Recall that lower ground states of (silox)₂(^tBu₃SiN)-

Ti-R (2-R, R ) sp³) led to only slightly higher ∆G^q_elimvalues,

because the 1,2-RH-elimination is not particularly “late”. In

the parabolic model, a greater influence of ∆G°(M-R) on the

1,2-elimination reaction is predicted for later reactions. The

data in the table reflect the increasingly “late” character of the

elimination event as one proceeds down and to the right on the

periodic table. In general terms, 1,2-RH-addition to X₂MdN-

will become earlier and more exoergic descending group 4, the

(104) (a) Kaltsoyannis, N. J. Chem. Soc., Dalton Trans. 1997, 1-11.

(b) Pyykko¨, P. Chem. ReV. 1988, 88, 563-594.

(105) (a) Chao, Y. W.; Rodgers, P. M.; Wigley, D. E.; Alexander, S. J.;

Rheingold, A. L. J. Am. Chem. Soc. 1991, 113, 6326-6327. (b) Smith, D.

P.; Allen, K. A.; Carducci, M. D.; Wigley, D. E. Inorg. Chem. 1992, 31,

1319-1320. (c) Bryan, J. C.; Burrell, A. K.; Miller, M. L.; Smith, W. H.;

Burns, C. J.; Sattelberger, A. P. Polyhedron 1993, 12, 1769-1777.

(106) Marcus, R. A. Angew. Chem., Int. Ed. Engl. 1993, 32, 1111-

1121.

∆G^q

will decrease accordingly, i.e., Ti > Zr > Hf, and the

addn

corresponding 1,2-RH-eliminations will be later descending the

group, with ∆G^q

increasing as Ti < Zr < Hf.

elim

(102) Periana, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1986, 108,

7332-7346.

(103) Cundari, T. R. J. Am. Chem. Soc. 1994, 116, 340-347.

(107) Leffler, J. E.; Grunwald, E. Rates and Equilibria of Organic

Reactions; John Wiley & Sons: New York, 1963.

10716 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

then dissolved in hexanes (40 mL), and the mixture was filtered. The

precipitate was washed with hexanes (7 × 40 mL), and the extracts

were combined and concentrated (20 mL). Cooling to -78 °C and

filtering yielded 7.35 g of microcrystalline 2-Cl . A second crop gave

1.17 g (8.52 g, 90%). IR (Nujol, cm^-1): 1380 (m), 1370 (m), 1360

(m), 1085 (m), 1015 (w), 940 (m), 858 (s), 822 (s), 805 (s), 720 (w),

627 (s). Anal. Calcd for TiSi₃ClNO₂C₃₆H₈₂: C, 59.34; H, 11.34; N,

1.92. Found: C, 59.28; H, 11.40; N, 1.86.

3. (silox)₂(^tBu₃SiNH)TiMe (2-Me). Into a solution of 2-Cl (1.367

g, 1.877 mmol) in 15 mL of ether at 25 °C was syringed 1.25 mL of

MeMgBr in ether (3.0 M, 2 equiv). The solution was stirred for 1 h,

and the volatiles were removed. The solid was triturated with hexanes

(3 × 15 mL) and then dissolved in hexanes (10 mL), and the mixture

was filtered. The precipitate was washed with hexanes (5 × 15 mL),

and the extracts were combined and concentrated to 5 mL. Cooling to

-78 °C and filtering gave 0.852 g (64%) of microcrystalline 2-Me. A

second crop afforded 0.157 g (1.009 g total, 76%). IR (Nujol, cm^-1):

3248 (w), 1390 (m), 1380 (w), 1368 (m), 1089 (m), 1016 (w), 1009

(w), 950 (s), 940 (s), 875 (s), 828 (s), 775 (w), 725 (w), 625 (s). Anal.

Calcd for TiSi₃NO₂C₃₇H₈₅: C, 62.75; H, 12.10; N, 1.98. Found: C,

62.23; H, 12.11; N, 1.63.

4. (silox)₂(^tBu₃SiNH)TiCH₂CH₃(2-Et). To a slurry of 2-Cl (462

mg, 0.634 mmol) in ether was added 0.35 mL of EtMgCl (2.0 M in

ether) at 25 °C. The mixture was stirred for 4 h, and the volatiles

were removed. The solid was triturated with hexanes (3 × 5 mL) and

then dissolved in hexanes, and the mixture was filtered. The residue

was washed with hexanes (7 × 3 mL), and the extracts were combined

and concentrated to 2 mL. Cooling to -78 °C and filtering afforded

microcrystalline 2-Et (187 mg, 41%). A second crop yielded 45 mg

(232 mg, 51%). IR (Nujol, cm^-1): 3228 (w), 1382 (s), 1372 (m), 1360

(s), 1085 (s), 1009 (m), 1001 (m), 945 (s), 941 (s), 914 (s), 865 (s),

815 (s), 720 (w), 620 (s). Anal. Calcd for TiSi₃NO₂C₃₈H₈₇: C, 63.19;

H, 12.14; N, 1.94. Found: C, 63.08; H, 12.21; N, 1.90.

5. (silox)₂(^tBu₃SiNH)TiCH₂Ph (2-Bz). To a solution of 2-Cl (620

mg, 0.790 mmol) in 15 mL of ether was added 0.64 mL of PhCH₂-

MgCl (2.0 M in ether) at 25 °C. The mixture was stirred for 3 h, and

the solution turned yellow-orange. The volatiles were removed, the

residue was triturated with hexanes (3 × 5 mL) and then dissolved in

hexanes (5 mL), and the mixture was filtered. The precipitate was

washed with hexanes (5 × 5 mL), and the extracts were combined and

concentrated to 3 mL. Cooling to -78 °C and filtering gave

microcrystalline 2-Bz (480 mg, 72%). IR (Nujol, cm^-1): 3222 (w),

1604 (w), 1390 (m), 1379 (w), 1368 (m), 1216 (w), 1105 (s), 1032

(w), 1025 (w), 1015 (w), 1000 (w), 933 (s), 895 (m), 870 (s), 820 (s),

744 (m), 696 (m), 625 (s). Anal. Calcd for TiSi₃NO₂C₄₃H₈₉: C, 65.85;

H, 11.44; N, 1.79. Found: C, 65.58; H, 11.57; N, 1.73.

benzylic and allylic activations in these systems seem to

contradict the latter. Since d(R-H) and d(M-R) vary inversely

with s-character, the primary effect of hybridization may be to

influence the reaction coordinate! Rationalizations of reactivity

on the basis of participation or nonparticipation of π-orbitals

are often popular and yet are difficult to assess.⁹⁶Again, bond-

making and -breaking events that involve sp- and sp²-carbon

centers of unsaturation are likely to reflect the consequences of

a compressed reaction coordinate relative to an sp³-carbon,

leading to generally speedier reactions.¹⁰⁸In summary, ther-

modynamic influences on reaction rate, usually described via

linear or nonlinear free energy relationships, are readily ac-

cepted, yet positional dependencies should receive greater

attention than less transparent, orbital-based rationalizations

(e.g., s-character and π-effects).

Experimental Section

General Considerations. All manipulations were performed using

either glovebox or high-vacuum techniques. Hydrocarbon and ethereal

solvents were dried over and vacuum transferred from sodium benzo-

phenone ketyl (with 3-4 mL of tetraglyme/L added to hydrocarbons).

Benzene-d₆was sequentially dried over sodium and 4 Å molecular

sieves and then stored over and vacuum-transferred from sodium

benzophenone ketyl. Cyclohexane-d₁₂was dried over sodium and then

stored over and vacuum-transferred from Na/K alloy. All glassware

was base-washed and oven-dried. NMR tubes for sealed tube experi-

ments were flame-dried under dynamic vacuum immediately prior to

the experiment. TiCl₄(THF)₂was prepared according to the literature

110,111

procedure,¹⁰⁹using TiCl₄(Aldrich) as received. NaOSi^tBu₃

and

112

LiNHSi^tBu₃were prepared according to the literature procedures.

Methane, ethane, propane, butane, ethylene, propene, cis-2-butene,

trans-2-butene, cyclopropane, and isobutylene (Matheson) were passed

through a -78 °C trap before use. 2-Butyne (Farach Chemical Co.)

was dried over Na and stored over 4 Å molecular sieves in a glass

bomb. Neohexane, tetramethylsilane, PMe₃, and NMe₃(Aldrich) were

dried over Na and stored in glass bombs over Na prior to use.

¹H and ¹³C{¹H} NMR spectra were obtained using Varian XL-200,

XL-400, VXR-400S, and Unity-500 spectrometers. Infrared spectra

were recorded on a Perkin-Elmer 299B spectrophotometer. Combustion

analyses were performed by Oneida Research Services, Whitesboro,

NY, Texas Analytical Labs, Stafford, TX, or Robertson Microlit

Laboratories, Madison, NJ. The Center for High Energy Synchrotron

Studies (CHESS) at Cornell University was used for the X-ray

crystallographic study of (silox)₂(THF)TidNSi^tBu₃(3-THF).

Procedures. 1. (silox)₂TiCl₂(1). To a flask containing TiCl₄-

(THF)₂(1.69 g, 5.1 mmole) and Na(silox) (2.45 g, 10.2 mmol) was

added 60 mL of ether at -78 °C. On warming to 25 °C, the bright

yellow solution of TiCl₄(THF)₂bleached to a pale yellow solution. After

overnight stirring under argon, the ether was removed and replaced

with 40 mL of hexanes. Filtration in hexanes, concentration, and

cooling yielded 1 as a white crystalline solid (2.20 g, 78%). IR (Nujol,

cm^-1): 1018 (w), 1009 (w), 980 (s), 900 (s), 815 (s), 630 (s), 610 (s).

Anal. Calcd for C₂₄H₅₄Si₂O₂Cl₂Ti: C, 52.44; H, 9.90. Found: C,

52.48; H, 9.78.

6. (silox)₂(^tBu₃SiNH)TiCHdCH₂(2-Vy). A 25 mL round-bottom

flask was charged with 2-Cl (333 mg, 0.457 mmol) and vinyllithium

n

(prepared from BuLi and tetravinyltin, 37 mg, 1.09 mmol).¹¹³Ether

was added via vacuum transfer at -78 °C, and the mixture was stirred

for 1 h. Volatiles were removed under vacuum at 0 °C. ¹H NMR

spectroscopy showed the mixture to be ∼90% 2-Vy, ∼ 5% 2-Cl, and

the remainder 3-C₂H₄and 3-OEt₂. The solid was triturated with hexanes

(2 × 5 mL) and then dissolved in hexanes, and the mixture was filtered.

The filtrate was concentrated (0.5 mL) to yield an off-white powder,

which was collected by filtration (98 mg, 30%). IR (Nujol, cm^-1):

3225 (w), 1554 (w), 1394 (w), 1382 (s), 1371 (w), 1360 (m), 1080

(m), 1009 (m), 1000 (m), 990 (m), 952 (m), 930 (s), 900 (s), 865 (s),

810 (s), 711 (w), 615 (s).

7. (silox)₂(^tBu₃SiNH)TiPh (2-Ph). A solution of 2-Me (145 mg,

0.205 mmol) in benzene (5 mL) was heated (60 °C) for 4 h in a 30 mL

glass bomb. The bomb was immersed in liquid nitrogen, and volatiles

(methane) were removed under vacuum. The vessel was then heated

for 4 h at 60 °C and cooled to 25 °C. The volatiles were removed to

yield a yellow powder (109 mg, 69%). A portion was set aside for

characterization while the rest was immediately taken up in C₆D₆for

use in kinetics studies. Attempts to concentrate the solution and

recrystallize this material from benzene led to the precipitation of a

yellow insoluble material, probably 2-C₆H₄-2 (see text). IR (Nujol,

cm^-1) 3215 (w), 1578 (w), 1386 (m), 1373 (w), 1363 (m), 1090 (m),

2. (silox)₂(^tBu₃SiNH)TiCl (2-Cl). A 100 mL round-bottom flask

was charged with ^tBu₃SiNHLi (2.876 g, 12.99 mmol) and (silox)₂TiCl₂

(7.142 g, 13.18 mmol). Ether (80 mL) was added via vacuum transfer

at -78 °C. The mixture was allowed to warm to 25 °C over the course

of 20 min, and the solids dissolved. The pale yellow solution was

stirred 30 min at 25 °C, and the volatiles were removed to yield an

off-white solid, which was triturated with hexanes (3 × 30 mL) and

(108) For example, note that the reductive elimination rates of (Ph₃P)₂-

Pt(H)R (R ) Ph > Et > Me > allyl) conform to the parabolic model. See:

Abis, L.; Sen, A.; Halpern, J. J. Am. Chem. Soc. 1978, 100, 2915-2916.

(109) Manzer, L. Inorg. Synth. 1982, 21, 135-140.

(110) Covert, K. J.; Wolczanski, P. T.; Hill, S. A.; Krusic, P. J. Inorg.

Chem. 1992, 31, 66-78.

(111) Weidenbruch, M.; Pierrard, C.; Pesel, H. Z. Z. Naturforsch., B:

Anorg. Chem. Org. Chem. 1978, 33B, 1468-1471.

(112) Nowakowski, P. M.; Sommer, L. H. J. Organomet. Chem. 1979,

178, 95-103.

(113) Seyferth, D.; Weiner, M. A. J. Org. Chem. 1961, 26, 3583-3586.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10717

1010 (m), 925 (s), 880 (s), 815 (s), 803 (s), 775 (m), 720 (w), 620 (s).

Anal. Calcd for TiSi₃NO₂C₄₂H₈₇: C, 65.47; H, 11.38; N, 1.82.

Found: C, 65.38; H, 11.47; N, 1.84.

Anal. Calcd for TiSi₃NO₂C₄₂H₉₅: C, 64.81; H, 12.30; N, 1.80.

Found: C, 64.73; H, 12.38; N, 1.75.

14. (silox)₂(^tBu₃SiNH)TiCH₂C₆H₃(Me)₂(2-Mes). In a small glass

bomb, 2-Me (445 mg, 0.628 mmol) was dissolved in 10 mL of

mesitylene. The solution was heated at 55 °C for 14 h. The resulting

orange solution was cooled to 25 °C, the volatiles were removed to

yield an orange solid, which was taken up in hexanes (6 mL), and the

mixture was filtered. Yellow crystalline 2-Mes was collected by

recrystallization from 2 mL of hexanes at 0 °C (207 mg, 41%). IR

(Nujol, cm^-1): 3585 (w), 3222 (w), 1600 (w), 1385 (m), 1380 (m),

1370 (m), 1160 (w), 1090 (m), 1020 (w), 930 (s), 850 (s), 820 (s), 810

(s), 670 (w), 650 (w), 620 (s). Anal. Calcd for TiSi₃NO₂C₄₅H₉₃: C,

66.54; H, 11.54; N, 1.72. Found: C, 66.43; H, 11.81; N, 1.66.

8. (silox)₂(^tBu₃SiNH)TiH (2-H). To a 30 mL glass bomb contain-

ing 2-Me (1.756 g, 2.479 mmol) at -196 °C was added 12 mL of

cyclohexane. Hydrogen (500 Torr) was admitted at -196 °C; then

the vessel was warmed to 25 °C, heated at 90 °C for 2 h, and cooled

to 25 °C. The volatiles were removed, the solid was dissolved in

hexanes (10 mL), and the mixture filtered. The residue was washed

with hexanes (5 × 5 mL), and the extracts were concentrated to 3 mL.

Cooling to -78 °C and filtration gave a thick slurry of microcrystalline

2-H (885 mg). A second crop yielded 279 mg (1.164 g, 67%). IR

(Nujol, cm^-1): 3225 (w), 1645 (m, TiH), 1348 (m), 1370 (w), 1360

(m), 1080 (m), 1010 (w), 950 (m), 930 (w), 875 (s), 830 (s), 815 (s),

765 (m), 715 (w), 620 (s). Anal. Calcd for TiSi₃NO₂C₃₆H₈₃: C, 62.29;

H, 12.05; N, 2.02. Found: C, 62.01; H, 11.98; N, 1.44.

9. (silox)₂(^tBu₃SiNH)Ti(CH₂)₃CH₃(2-ⁿBu). To a slurry of 2-Cl

(395 mg, 0.542 mmol) in 10 mL of ether was added 0.35 mL of ⁿBuLi

(1.6 M in hexanes, 1.03 equiv) at 0 °C. Upon addition of the

alkyllithium, the solution turned yellow and a fine white precipitate

began to appear. After 5 min, the volatiles were removed to give a

yellow foam, which was triturated with hexanes (3 × 5 mL). The

product would not crystallize from cold hexanes, so the crude solid

(>90% 2-ⁿBu by ¹H NMR, principal impurity 2-Cl) was removed (225

mg, 55%) and used for kinetics studies. IR (Nujol, cm^-1): 3225 (w),

1390 (s), 1380 (s), 1370 (s), 1190 (w), 1100 (m), 1015 (m), 930 (s),

880(3), 820 (s), 610 (m).

15. (silox)₂(^tBu₃SiNH)Ti-c-C₄H₇(2-^cBu). To a solid mixture of

2-Cl (517 mg, 0.710 mmol) and ^cBuLi (46 mg, 0.742 mmol) was added

15 mL of ether at -78 °C. The resulting slurry was warmed to -30

°C, and the solids were dissolved to generate a brown solution. After

5 min, the volatiles were removed while the reaction mixture was kept

at -30 °C. The solids were triturated with hexanes (3 × 10 mL) at 0

°C and then dissolved in hexanes, and the mixture was filtered. The

residual was washed with hexanes (2 × 10 mL), and the extracts were

combined and concentrated to 3 mL. Cooling to -78 °C and filtration

gave microcrystalline 2-^cBu (186 mg, 35%). IR (Nujol, cm^-1): 3233

(w), 1385 (s), 1380 (m), 1240 (w), 1195 (w), 1090 (s), 1035 (w), 1020

(m), 940 (s), 870 (s), 820 (s), 665 (w), 620 (s). 2-^cBu was not of

1

sufficient purity (>90% by H NMR) for combustion analysis.

16. (silox)₂(^tBu₃SiN)TiCH₂CH₂(3-C₂H₄). A glass bomb was

charged with 2-Et (317 mg, 0.439 mmol) and 12 mL of cyclohexane.

The bomb was immersed in liquid nitrogen and 2 equiv of C₂H₄(125.2

mL at 130 Torr) was added. The mixture was warmed to 25 °C and

stirred for 20 h and then warmed to 45 °C and stirred for 4 h. The

solution was cooled to 25 °C, concentrated to 4 mL, and filtered to

remove the polyethylene that had formed. The residue was washed

with cyclohexane (2 × 3 mL), and the extracts were combined. The

solvent was removed and replaced with pentane (1 mL). Concentrating

to 0.6 mL, cooling to -78 °C, and filtration afforded yellow

microcrystalline 3-C₂H₄(131 mg, 42%). IR (Nujol, cm^-1): 1390 (m),

1380 (m), 1365 (m), 1190 (w), 1005 (w), 945 (m), 910 (s), 880 (s),

820 (s), 720 (w), 625 (s). Because of the thermal instability of 3-C₂H₄,

combustion analysis was not attempted.

17. (silox)₂(THF)TidNSi^tBu₃(3-THF). A solution of 2-Me (503

mg, 0.710 mmol) in THF (8 mL) was stirred in a bomb reactor for 3.5

d. The volatiles were removed, the resulting yellow solid was triturated

with hexanes (3 × 10 mL) and then dissolved in hexanes, and the

mixture was filtered. Concentrating to 1.5 mL, cooling to -78 °C,

and filtration yielded microcrystalline 3-THF (220 mg, 40%). IR

(Nujol, cm^-1): 1395 (w), 1385 (s), 1365 (m), 1180 (w), 1074 (s), 1040

(w), 1013 (m), 930 (s), 890 (s), 821 (s), 755 (w), 615 (s). Anal. Calcd

for TiSi₃NO₃C₄₀H₈₉: C, 62.86; H, 11.74; N, 1.83. Found: C, 62.78;

H, 11.86; N, 1.80.

18. (silox)₂(py)TidNSi^tBu₃(3-py). To a solution of 2-Me (480

mg, 0.678 mmol) dissolved in hexanes (25 mL) was added 4 mL of

pyridine. The resulting solution was sealed in a glass bomb and stirred

at 65 °C for 7 h. The volatiles were removed to give a yellow powder,

which was subsequently triturated with hexanes (3 × 10 mL) and then

dissolved in hexanes (5 mL). Yellow crystalline 3-py was filtered (350

mg, 66%) from the hexanes (3 mL) at -78 °C. IR (Nujol, cm^-1) 1615

(m), 1380 (s), 1365 (m), 1220 (w), 1135 (w), 1080 (s), 1075 (s), 1045

(w), 1020 (m), 920 (s), 890 (s), 735 (m), 700 (m), 625 (w), 610 (s).

Anal. Calcd for TiSi₃N₂O₂C₄₁H₈₆: C, 63.85; H, 11.24; N, 3.63.

Found: C, 64.08; H, 11.50; N, 3.51.

19. (silox)₂(OEt₂)TidNSi^tBu₃(3-OEt₂). A solution of 2-Me (1.196

g, 1.69 mmol) in Et₂O (25 mL) was stirred for 4 d at 25 °C. Upon

removal of the volatiles, the solid was triturated with hexanes (3 × 15

mL) and then dissolved in hexanes (15 mL) and filtered. Faint yellow

crystalline 3-OEt₂was filtered (895 mg, 69%) from the hexanes (3

mL) at 0 °C. IR (Nujol, cm^-1): 3680 (w), 1391 (m), 1380 (m), 1369

(m), 1189 (w), 1150 (w), 1092 (m), 1063 (m), 1036 (w), 1015 (w),

1008 (w), 999 (m), 920 (s), 890 (s), 822 (s), 770 (m), 721 (w), 625 (s).

Anal. Calcd for TiSi₃NO₂C₄₀H₉₁: C, 62.69; H, 11.97; N, 1.84.

Found: C, 62.55; H, 12.80; N, 1.46.

10. (silox)₂(^tBu₃SiNH)Ti-c-C₃H₅(2-^cPr). A small bomb was

charged with 2-Me (460 mg, 0.649 mmol) and hexanes (10 mL).

Cyclopropane (10 equiv) was admitted via a calibrated volume gas

bulb. The bomb was heated at 60 °C for 14 h and then cooled to 25

°C, whereupon colorless crystals precipitated from the pale yellow

solution. The crystals were collected by decantation and washed with

cold hexanes (3 mL) to afford 156 mg of 2-^cPr (33%). IR (Nujol,

cm^-1): 3235 (w), 1390 (m), 1379 (m), 1367 (m), 1088 (m), 1015 (w),

945 (m), 875 (s), 840 (s), 820 (s), 725 (w), 627 (s). Anal. Calcd for

TiSi₃NO₂C₃₉H₈₇: C, 63.79; H, 11.94; N, 1.91. Found: C, 63.61; H,

12.03; N, 1.86.

11. (silox)₂(^tBu₃SiNH)Ti-c-C₅H₉(2-^cPe). A small bomb was

charged with 2-Me (484 mg, 0.683 mmol) and cyclopentane (4 mL).

The solution was thrice subjected to heating (50 °C) followed by a

freeze-pump-thaw degas cycle to remove methane, thereby driving

the reaction to completeness. Product 2-c-C₅H₉was was precipitated

from a 1:1 mixture of cyclopentane/Me₃SiOSiMe₃at 0 °C and collected

via filtration (223 mg, 42%). IR (Nujol, cm^-1): 3210 (w), 1395 (m),

1383 (s), 1372 (m), 1360 (m), 1296 (w), 1093 (s), 1009 (w), 1004 (w),

930 (s), 860 (s), 815 (s), 730 (w), 620 (s). Anal. Calcd for TiSi₃-

NO₂C₄₁H₉₁: C, 64.60; H, 12.03; N, 1.84. Found: C, 64.41; H, 12.14;

N, 1.80.

12. a. (silox)₂(^tBu₃SiNH)Ti-c-C₆H₁₁(2-^cHex) and [(silox)₂-

TidNSi^tBu₃]₂(3₂). A glass bomb was charged with 2-Me (566 mg,

0.799 mmol) and 15 mL cyclohexane. The solution was stirred for 7

days at 55 °C, concentrated (10 mL), and allowed to cool slowly to 25

°C, affording yellow microcrystalline material (122 mg). ¹H NMR

analysis indicated a mixture of sparingly soluble 2-Cy and 3₂. IR

(Nujol, cm^-1): 1395 (w), 1384 (m), 1360 (w), 1280 (m), 1200 (s),

1035 (m), 962 (s), 898 (s), 865 (s), 812 (m), 720 (m), 618 (m).

b. 2-^cHex for Kinetic Study. A small glass bomb was charged

with 2-Me (577 mg, 815 mmol) and 10 mL of cyclohexane. The bomb

was sealed and heated at 65 °C for 3 h. The solution was thrice

subjected to 77 K freeze-pump-thaw degas cycles to remove methane.

Removal of the volatiles afforded a yellow solid, which was extracted

with 2.2 mL of C₆D₆. This solution was filtered to remove residual

solids and used immediately for kinetics studies.

t

13. (silox)₂(^tBu₃SiNH)TiCH₂CH₂Bu (2-^neoHex). A glass bomb

was charged with 2-Me (373 mg, 0.527 mmol) and 7 mL of CH₃CH₂C-

(CH₃)₃. The solution was heated for 6 h at 50 °C, cooled to 25 °C,

and concentrated to 2 mL, which induced crystallization. Colorless

2-^neoHex was obtained via filtration (82 mg, 20%). IR (Nujol, cm^-1):

3238 (w), 1390 (m), 1380 (w), 1365 (m), 1238 (w), 1092 (m), 1061

(w), 1014 (m), 1005 (m), 940 (s), 870 (s), 820 (s), 735 (w), 625 (s).

10718 J. Am. Chem. Soc., Vol. 119, No. 44, 1997

Bennett and Wolczanski

20. (silox)₂(PMe₃)TidNSi^tBu₃(3-PMe₃). To a 10 mL round-

bottom flask charged with 2-Me (221 mg, 0.312 mmol) and 5 mL of

hexanes at -78 °C was added 2.3 equiv of PMe₃via a gas bulb. The

reaction mixture was stirred for 5 d at 25 °C, and the volatiles were

removed to afford a yellow powder (115 mg, 48%). IR (Nujol, cm^-1):

1380 (m), 1363 (m), 1308 (w), 1288 (w), 1064 (s), 1010 (w), 960

(m), 945 (w), 930 (w), 910 (s), 883 (s), 842 (w), 818 (s), 730 (w), 618

(m). Anal. Calcd for TiSi₃NO₂PC₃₉H₉₀: C, 60.97; H, 11.81; N, 1.82.

Found: C, 60.76; H, 12.05; N, 1.84.

Occasionally only two of the three tubes would survive the entire kinetic

run, but in all cases at least two tubes survived.

General Equilibria Procedures. 1. Sample Preparation.

A

flame-dried 5 mm NMR tube joined to a 14/20 joint and attached to a

needle valve was charged with ca. 20 mg of an organotitanium species.

The apparatus was attached to the vacuum line and evacuated.

Cyclohexane-d₁₂(0.4-0.7 mL) was admitted to the tube via vacuum

transfer followed by a known amount of a volatile reagent admitted

through a calibrated gas bulb. In some cases, a second reactant was

added (H₂for example). The mixture was cooled to 77 K, and the

tube was sealed with a torch.

21. (silox)₂(^tBu₃SiN)TiC(Me)dCMe (3-MeCCMe). Into a solution

of 2-H (463 mg, 0.667 mmol) in cyclohexane (5 mL) was condensed

1.1 equiv of 2-butyne (287.8 mL at 47 Torr). The solution was stirred

for 1 h at 25 °C, whereupon it turned orange-brown. The volatiles

were removed, the residue was taken up in hexanes (3 mL), and the

mixture was filtered. The residual was washed with hexanes (3 × 5

mL), and the extracts were collected and concentrated to 1 mL. Cooling

to -78 °C, stirring for 1 h, and filtration afforded dark orange

microcrystalline 3-MeCCMe (227 mg). A second crop yielded 45 mg

(272 mg total, 55%). IR (Nujol, cm^-1): 1382 (m), 1372 (w), 1360

(w), 1232 (w), 1147 (s), 1040 (w), 1000 (w), 908 (s), 867 (s), 814 (s),

713 (m), 619 (m). Anal. Calcd for TiSi₃NO₂C₄₀H₈₇: C, 64.38; H,

11.75; N, 1.88. Found: C, 64.19; H, 11.70; N, 1.71.

2. Measurement of Equilibria. Equilibrium concentrations of

reactants and products were determined by 400 MHz (or 500 MHz in

1

certain cases of difficult signal overlap) H NMR spectroscopy. All

tubes were measured at least twice after it was certain the reactions

had come to equilibrium. Because of the need to accurately measure

very low concentrations of species in many cases, the following protocol

for data collection was observed: (1) a delay time (D₁) of 120 s was

used between acquisitions, (2) the filter bandwidth was set to ensure

no attenuation of signal within the spectral window, and (3) the

acquisition time was set to the largest value allowed by the computer

software (ca. 8-12 s; sweep width dependent). These precautions

ensured complete relaxation of nuclei under investigation, accuracy of

integrals of peaks near the edge of the spectrum, and elimination of

ringing in intense peaks. Depending on the intensity of the weakest

resonance to be measured, spectra were acquired with 4-32 transients.

In a few cases, deconvolution of overlapping peaks was required.

NMR Tube Reactions. a. (silox)₂(^tBu₃SiN)TiC(Et)dCEt (3-

EtCCEt). An NMR tube attached to a ground glass joint was charged

with 16 mg (0.023 mmol) of 2-Me and attached to a calibrated gas

bulb. The apparatus was attached to a vacuum line, evacuated, and

cooled to -78 °C. Cyclohexane-d₁₂(0.4 mL) was admitted via vacuum

transfer. The tube was cooled to 77 K and reevacuated to admit

3-hexyne (1 equiv) via the gas bulb. The tube was sealed with a torch

and allowed to stand at 25 °C for 4 d. 3-EtCCEt formed in >95%

Single-Crystal X-ray Diffraction Studies. 1. (^tBu₃SiO)₂(^tBu₃SiNH)-

t

TiCH₂CH₂Bu (2-^neoHex). Large block-shaped crystals (approximately

0.5 × 0.5 × 0.5) formed upon concentration of a solution of 2-^neoHex

at room temperature. These crystals were mounted in thin walled glass

capillaries which were subsequently flame-sealed. Preliminary rotation

photographs indicated orthorhombic Laue symmetry. Precise lattice

constants (Table 2) were determined from a least-squares fit of 15

diffractometer-measured 2θ values at 25 °C. The structure was solved

by direct methods and refined by least-squares methods (SHELXL-

1

yield as judged by H NMR spectroscopy.

b. (silox)₂(^tBu₃SiN)TiC(H)dCH (3-HCCH). An NMR tube at-

tached to a ground glass joint was charged with 30 mg (0.042 mmol)

of 2-Me and attached to a calibrated gas bulb. The apparatus was

attached to a vacuum line, evacuated, and cooled to -78 °C.

Cyclohexane-d₁₂(0.4 mL) was admitted via vacuum transfer. The tube

was cooled to 77 K and reevacuated to admit acetylene (1.2 equiv) via

the gas bulb. The tube was sealed with a torch and allowed to stand

t

93). Severe disorder (silox cannot be distinguished from Bu₃SiNH)

in the periphery of the molecule limited interpretation of the structural

model to a description of the overall geometry and the alkyl fragment.

2. (silox)₂(THF)TidNSi^tBu₃(3-THF). A supersaturated solution

of 3-THF in hexanes was prepared in a 10 mL flask by heating 800

mg of 3-THF in 3 mL of hexanes. The hot flask was wrapped with

glass wool and aluminum foil and allowed to stand undisturbed for 14

h. Upon standing, many small crystals of 3-THF formed along with a

few large (ca. 3 mm on an edge), well-formed, crystalline blocks.

Preliminary unit cell determination using a fragment (ca. 0.8 × 0.8 ×

0.8 mm) of a larger crystal revealed orthorhombic symmetry but even

this large crystal exhibited a sharp drop in observed reflections with

increasing 2θ on a sealed-tube instrument.

A similar fragment (0.8 × 0.8 × 0.4 mm) was coated in Paratone-N

oil and cooled in a nitrogen cold stream at the CHESS A-1 station.

X-ray data were collected using 13.65 keV (0.908 Å) X-rays generated

by the A-line 1.3 T, 24-pole wiggler magnet. A total of 1.66 mrad of

the available 4.22 mrad synchrotron radiation was doubly focused into

the A-1 hutch by a horizontally focusing Si(111), cylindrically bent,

triangular monochromator followed by a vertically focusing, cylindri-

cally bent, rhodium-coated silicon mirror with final collimation and

beam shaping by a 0.5 mm double-pinhole collimator. The uncolli-

mated beam is 0.15 × 2.6 mm (fwhm) in dimension at the focus with

a flux of 1.3 × 10¹³photons/s and an energy resolution of 52 eV at 13

keV. A flux of 4 × 10¹¹photons/s with an energy resolution of 15 eV

has been measured through a 0.3 mm collimator in this configuration.

A Molecular Structure Corp. cryostat operating around -165 °C was

used to flash cool the crystal in Paratone-N oil and maintain it at

cryogenic temperatures throughout the data collection (Table 2).

Diffraction patterns were recorded with a 2K × 2K CCD operating in

binned mode (effective 1K × 1K). The data were merged and

processed using DENZO and SCALEPACK. The reflections were

corrected for background effects, integrated using DENZO, and scaled

together using SCALEPACK, and the structure was solved by direct

methods and refined by least-squares methods (SHELXL-93). For the

1

at 25 °C for 4 d. 3-HCCH formed in >70% yield as judged by H

NMR spectroscopy. A black flocculant solid, assumed to be poly-

acetylene, also formed.

c. (silox)₂(R₃N)TidNSi^tBu₃(3-NR₃, R ) Me, Et). An NMR tube

attached to a ground glass joint was charged with 30 mg (0.042 mmol)

of 2-Me and attached to a calibrated gas bulb. The apparatus was

attached to a vacuum line, evacuated, and cooled to -78 °C.

Cyclohexane-d₁₂(0.4 mL) was admitted via vacuum transfer. The tube

was cooled to 77 K and reevacuated to admit NR₃(1.2 equiv) via the

gas bulb. The tube was sealed with a torch and allowed to stand at 25

1

°C for 4 d. 3-NR₃formed in >90% yield as judged by H NMR.

General Kinetics Procedures. Solutions of 2-R were prepared in

n

c

C₆D₆in 2 mL volumetric flasks. In the cases of R ) Bu, Bu, and

^cHex, the solutions were filtered to remove insoluble material. Either

Me₃SiSiMe₃or Me₃SiOSiMe₃(∼1 µL) was added as an internal

standard. Three samples of 0.6 mL each were transferred to flame-

dried 5 mm NMR tubes joined to 14/20 joints and attached to needle

valves. The tubes were subjected to three freeze-pump-thaw degas

cycles and flame-sealed under vacuum. Temperatures above 25 °C

were regulated in a polyethylene glycol bath with a Tamson immersion

circulator. The temperature of 24.8 °C was obtained by placing the

tubes in an Hewlett Packard model 5890A gas chromatograph oven

kept in a room at 20-22 °C with the nominal temperature set to 25

°C. In both cases, the temperature was stable to (0.3 °C. Rates of

disappearance of amido NH peaks were monitored in most cases. For

some 2-R derivatives a better signal intensity was obtained by

monitoring proton resonances from the R group (R ) Et, Bz, ^cPr, ^cBu,

^cPe, ^cHex, ^neoHex). All runs were monitored for 5-6 half-lives. Single

transient spectra were used to obtain the most reproducible integrals.

Rates and uncertainties were obtained from nonlinear, nonweighted

least-squares fitting to the exponential form of the rate expression.

SelectiVities in Hydrocarbon ActiVation

J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10719

∆G°(2-Rⁿ) + ∆G^q_elim(2-Rⁿ) ) ∆G°(3) + (x(TS_n))²(A4)

11 536 unique integrated observations of ca. 18 500 observed reflections

an R_sym) 3.1% was obtained to 0.8 Å resolution.

Solving for x(TS_n), which is <0, since x(3) was defined as 0,

Acknowledgment. Support from the National Science

Foundation (Grant CHE-9218270 (past), Grant CHE-9528914

(current)) is gratefully acknowledged. We thank Barry K.

Carpenter, Thomas R. Cundari, and Thomas P. Vaid for helpful

discussions, Steve Ealick, Dan Thiel, Marian Szebenyi, Rick

Walter, Emil Lobkovsky, Jorge Rios Steiner, Dave Fuller, and

Aidan Harrison for assistance in X-ray crystallography and

NMR measurements, and Lee M. Slaughter for experimental

aid. Funding for the Cornell NMR Facility from the NIH and

NSF Instrumentation Programs is acknowledged.

we have

x(TS_n) ) -(∆G°(2-Rⁿ) + ∆G^q_elim(2-Rⁿ) - ∆G°(3))^1/2(A5)

At the intersection, we rewrite eq A2 using eq A3 and rearrange

to a polynomial in x(2-R) using A5:

∆G°(2-Rⁿ) + ∆G^q_elim(2-Rⁿ) )

∆G°(2-Rⁿ) + (x(TS_n) - x(2-Rⁿ))²(A6)

(x(2-Rⁿ))²- 2(x(TS_n))(x(2-Rⁿ)) + (∆G°(2-Rⁿ) -

∆G°(3)) ) 0 (A7)

Appendix: Parabolas 2-Rⁿ, 3, and ∆G^TSat x(TS_n) in

n

Figure 7

Solving for x(2-R), we obtain eqs A8 and A9.

The parabolas corresponding to 2-Rⁿand 3 in Figure 7 are

indicated in eqs A1 and A2, where ∆G°(2-Rⁿ), ∆G°(3), and

x(2-Rⁿ) ) x(TS_n) - (∆G^q_elim(2-Rⁿ))^1/2

(A8)

x(2-Rⁿ) ) -(∆G°(2-Rⁿ) + ∆G^q_elim(2-Rⁿ) - ∆G°(3))^1/2

-

∆G(3) ) ∆G°(3) + K(x - x(3))²

(A1)

(A2)

(∆G^q_elim(2-Rⁿ))^1/2(A9)

∆G(2-Rⁿ) ) ∆G°(2-Rⁿ) + K(x - x(2-Rⁿ))²

Supporting Information Available: A table listing indi-

vidual equilibrium measurements, a discussion of NH IR

stretching frequencies for 2-R, comments on azametalla-

cyclobutene vs alkylidene-imine structures, a perspective on

cyclometalation, and X-ray structural information pertaining to

their free energies are in kcal/mol relative to ∆G°(3-py) ) 0

kcal/mol. x(2-Rⁿ) and x(3) are the positions of the corresponding

states along the reaction coordinate, x. The steepness of

parabolas 2-Rⁿand 3 is set arbitrarily (K ) 1). The following

are defined or known: x(3) ) 0, ∆G°(3) ) 24 kcal/mol, ∆G°(2-

Rⁿ), and ∆G^q_elim(2-Rⁿ). Values of x(TS_n) and x(2-Rⁿ) are sought.

t

(^tBu₃SiO)₂(^tBu₃SiNH)TiCH₂CH₂Bu (2-^neoHex) and (silox)₂-

(THF)TidNSi^tBu₃(3-THF), including tables of crystal data

encompassing data collection and solution/refinement, atomic

coordinates, isotropic and anisotropic temperature factors,

hydrogen atom coordinates, and bond lengths, and bond angles

(28 pages). See any current masthead page for ordering and

Internet access instructions.

The 2-R and 3 parabolas intersect at x(TS_n) and ∆G^TS, and

n

q

∆G^TS) ∆G°(2-R ) + ∆G _elim(2-Rⁿ)

(A3)

n

At the intersection (i.e., ∆G(2-Rⁿ) ) ∆G^TS), eq A1 is rewritten

n

as

JA9707419

Article Doi

Selectivities in hydrocarbon activation: Kinetic and thermodynamic investigations of reversible 1,2-RH-elimination from (silox)2((t)Bu3SiNH)TiR (silox = (t)Bu3SiO)

DOI: 10.1021/ja9707419

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja9707419

Full text of DOI:10.1021/ja9707419

Products guided by the article

R&D Labs maybe for 154030-91-8

Relevant to this article

Hot Product