Hydrocarbon activation via reversible 1,2-RH-elimination from (tBu3SiNH)3ZrR: Synthetic, structural, and mechanistic investigations

J. Am. Chem. Soc. 1996, 118, 591-611

591

Hydrocarbon Activation via Reversible 1,2-RH-Elimination

from (^tBu₃SiNH)₃ZrR: Synthetic, Structural, and Mechanistic

Investigations

Christopher P. Schaller, Christopher C. Cummins, and Peter T. Wolczanski*

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

Ithaca, New York 14853

ReceiVed August 2, 1995^X

Abstract: Hydrocarbyl complexes, (^tBu₃SiNH)₃ZrR (1-R), were prepared via metatheses of (^tBu₃SiNH)₃ZrCl (1-Cl)

t

with RMgX or RLi (R ) Me, Et, Cy, CH₂Ph, allyl, CHdCH₂, Ph, CH₂Bu, CtCPh, CtC^tBu), through addition of

isobutylene, H₂CdCdCMe₂, and acetylene to 1-H (R ) ⁱBu, dma, or CHdCH₂), and by CH-bond activation; thermal

1,2-RH-elimination from 1-R produced putative (^tBu₃SiNH)₂ZrdNSi^tBu₃(2), which was subsequently trapped by

R′H. Thermolysis of 1-R (∼100 °C, R ) Me or Cy) in the presence of H₂, c-C₃H₆, and CH₄in cyclohexane or neat

C₆H₆, mesitylene, and toluene afforded 1-R (R ) H, ^cPr, Me, Ph, CH₂-3,5-Me₂C₆H₃) and a mixture of 1-CH₂Ph and

1-C₆H₄Me, respectively. Exposure of 1-Cy to C₂H₄or C₆H₆in cyclohexane provided 1-CHdCH₂or 1-Ph, respectively,

but further reaction produced 1₂-(trans-HCdCH) and 1₂-(p-C₆H₄) through double CH-bond activation. Thermolysis

of (^tBu₃SiND)₃ZrCH₃(1-(ND)₃-CH₃) in C₆H₆or C₆D₆yielded CH₃D, and 1C₆H₅or 1-(ND)₃C₆D₅, through reversible

benzene activation. Thermolysis of 1-Cy in neat cyclohexane, and with C₂H₆or CMe₄present, gave cyclometalation

product (^tBu₃SiNH)₂ZrNHSi^tBu₂CMe₂CH₂(3) and 1-NHSi^tBu₃. In THF, thermolysis of 1-CH₃afforded (^tBu₃SiNH)₂-

(THF)ZrdNSi^tBu₃(2-THF); at 25 °C, 1-H lost H₂in the presence of L (L ) THF, Et₂O, NMe₃, PMe₃) generating

2-L; 2-L (L ) Et₂O, py) was also prepared via ligand exchange with 2-THF. Single crystal X-ray diffraction studies

of 2-THF revealed a pseudotetrahedral core, with a long ZrdN bond distance (1.978(8) Å), normal Zr-N(H) bond

lengths (2.028(8), 2.031(8) Å), similar amide (154.7(5), 158.1(5)°) and imide (156.9(5)°) bond angles, and little

O(pπ) f Zr(dπ) bonding. Crystal data: monoclinic, P2₁/n, a ) 13.312(5) Å, b ) 18.268(6) Å, c ) 20.551(7) Å,

â ) 92.30(3)°, Z ) 4, T ) 25 °C. 2-Et₂O thermally eliminated C₂H₄to give 1-OEt through γ-CH activation.

Kinetic isotope effects (KIE) on 1,2-RD-elimination from 1-(ND)₃-R (96.7 °C, R ) CH₃, z_Me) 6.3(1); CH₂Ph, z_Bz

) 7.1(6); Ph, z_Ph) 4.6(4)) and CD₃H loss from 1-CD₃(k(CH₃)/k(CD₃) ) (z′_Me)³) 1.32) revealed a symmetric

H-transfer in a loose transition state. 1,2-RH-elimination rates follow: (96.7 °C, k_R(×10⁴s^-1) ) 22.6(2), Ph; 15.5(2),

t

^cPr; 13.2(4), CHdCH₂; 10.4(2), Cy; 3.21(6), Et; 3.2(1), ⁱBu; 1.3(1), dma; 1.51(6), H; 1.42(4), CH₂Bu; 1.06(2), Me;

0.34(2), CH₂-3,5-Me₂C₆H₃; 0.169(3), CH₂Ph). Competition for (^tBu₃SiNH)₂ZrdNSi^tBu₃(2) by RH/R′H and equilibria

provided information about the stabilities of 1-R relative to 1-^cPr (R ) ^cPr (0.0 kcal/mol) < Ph (0.3) < CH₂Ph (0.7)

t

< Me (1.2) < CH₂Bu (g7.6) < Et (g7.8) < Cy (g10.9)). Transition state energies afforded relative C-H bond

c

activation selectivities (∆∆G^qrelative to Pr-H): ^cPrH ≈ ArH (0.0 kcal/mol) > MeH (3.4) > PhCH₂H (4.0) >

cyclometalation (g8.5) > EtH (g8.9) > ^tBuCH₂H (g9.3) > CyH (g11.2). A correlation of ∆G^q(1,2-RH-elimination)

with D(R-H) indicated generally late transition states but suggested an earlier composition for the alkyls, as rationalized

through a Hammond analysis. Correlation of ∆G^q(1,2-RH-elimination) with RH proton affinity implicated tight

binding of RH in the transition state and possible RH-binding intermediates (2-RH). 1,2-HCtCR-elimination from

1-CtCR was not observed, but second-order exchanges of 1-CtCPh with ^tBuCtCH, and 1-CtC^tBu with HCtCPh

were indicative of an associative pathway. All data can be accommodated by the following mechanism: 1-R + R′H

h 2-RH + R′H h 2-R′H + RH h 1-R′ + RH; a variant where 2 mediates reversible 2-RH + R′H exchange is less

likely.

Introduction

of a C-H bond through divergent pathways, which include (1)

oxidative addition to late metal centers (i.e., L_nM + RH h L_n-

HMR);^3,4(2) late metal assisted heterolytic RH cleavage in polar

media (i.e., L_nM + RH h [L_nMR]^-+ H⁺);^5-73) σ-bond

metathesis by early metal,⁸lanthanide,^9-11and cationic iridium¹²

The transition metal-mediated activation of carbon-hydrogen

bonds has been a forefront area of organometallic research for

over a decade.^1,2In that time, a variety of electronically

unsaturated metal complexes have accomplished the scission

^XAbstract published in AdVance ACS Abstracts, December 1, 1995.

(1) (a) ActiVation and Functionalization of Alkanes; Hill, C., Ed.; John

Wiley & Sons: New York, 1989. (b) SelectiVe Hydrocarbon ActiVation;

Watson, P. L., Ed.; VCH Publishers: Toledo, OH, 1989. (c) Shilov, A. E.

ActiVation of Saturated Hydrocarbons by Transition Metal Complexes;

Reidel: Boston, 1984.

(2) (a) Crabtree, R. H.; Chem. ReV. 1985, 85, 245-269. (b) Bergman,

R. G. Science 1984, 223, 902-908. (c) Graham, W. A. G. J. Organomet.

Chem. 1986, 300, 81-91. (d) Green, M. L. H.; O’Hare, D. Pure Appl.

Chem. 1985, 57, 1897-1910. (e) Halpern, J. Inorg. Chim. Acta 1985, 100,

41-48. (f) Crabtree, R. H.; Hamilton, D. G. AdV. Organomet. Chem. 1988,

28, 299-338. (g) Jones, W. D.; Feher, F. J. Acc. Chem. Res. 1989, 22,

99-106.

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

and references therein.

(4) For an example of a new system, see: Wang, C.; Ziller, J. W.; Flood,

T. C. J. Am. Chem. Soc. 1995, 117, 1647-1648.

(5) (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.

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

(6) (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.

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

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

592 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

systems (i.e., L_nMR + R′H h L_nMR′ + RH); (4) Hg* (³P₁,

5d¹⁰6s¹6p¹) sensitized C-H bond homolysis;¹³(5) attack on RH

by Rh(II) porphyrins (i.e., 2(por)Rh + RH h (por)RhH + (por)-

RhR),¹⁴and related, photochemically generated metalaradicals;¹⁵

(6) H-atom abstractions by metal oxo derivatives,^16,17including

photochemical¹⁸and bioinorganic systems;¹⁹and (7) 1,2-RH-

additions across MdX (X ) O (R ) OH, H,²⁰R′NH),²¹

NR,^21-29CR₂)³⁰multiple bonds.

Scheme 1

These disparate systems can be roughly separated into two

categories: radical C-H bond scissions with no apparent

binding of RH (4, 5, and 6), and concerted processes where

RH binding may play a significant role (1,^4,29-342, 3, and 7).

Within the latter group, the activation of C-H bonds by d⁰metal

imido derivatives^21-27is most intriguing. Isoelectronic oxo

groups exhibit mostly radical-based activations, perhaps because

the reactive functionality is typically bound to a coordinatively

and/or electronically saturated metal center, and H-atom abstrac-

tion must occur at the periphery of the complex. The concerted

nature of the metal imido derivatives suggests that their

reactivity patterns are more complex and that the electrophilicity

of the metal center plays a significant role in the attack of C-H

bonds, just as in late metal systems. Adjacent to the reactive

imido functionalities of transient X_3-nM(dNSi^tBu₃)_n(X )

HNSi^tBu₃, M ) Ti, n ) 1;²³M ) Zr, n ) 1;^22,26M ) V,²⁷

Ta,²⁴n ) 2; X ) OSi^tBu₃, M ) Ti, n ) 1)²⁵and Cp₂ZrdNR^28,29

Current research in homogeneous C-H bond activation is

proceeding along several fronts,³⁸but themes of practical and

fundamental nature stand out. In order to make use of a C-H

bond activation event, catalysis of this transformation must be

coupled with functionalization of the substrate. Radical based

oxidations of hydrocarbons are the most common, but inherent

problems of selectivity persist.¹⁷Electrophilic oxidation pro-

cesses in aqueous and other polar media that provide greater

opportunity for selective hydrocarbon conversion are under

development,^5-7and nonaqueous dehydrogenative methods also

appear promising.³⁹An earlier study featured functionalization

of a C-H bond through an intramolecular isonitrile insertion.⁴⁰

A fundamental understanding of the events that govern the

activation of a C-H bond in each of the above systems is crucial

to the design of future homogeneous, and perhaps heteroge-

neous, catalysts that can selectively activate hydrocarbons.

Typical commercial hydrocarbon activations are autoxidations

that are tantamount to a controlled burn.⁴¹Radical pathways

promise some selectivity based on the relative C-H bond

strengths, but even the radical-based, heterogeneously catalyzed

oxidative coupling of methane is operationally restricted.⁴²

Scheme 1 illustrates the general features of the C-H

2

species are respective d_z/p_zand “d_y”/p_y(z axis along ZrdN)

hybrid, empty orbitals oriented toward the C-H bond of an

approaching substrate. Similar orbitals mediate σ-bond me-

tathesis pathways in related d⁰and d⁰fⁿsystems, but these

second-order exchange processes⁹render possible R-H binding

events difficult to detect by indirect methods.³⁷

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

113, 1833-1835.

(9) 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.

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

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

437.

(12) Burger, P.; Bergman, R. G. J. Am. Chem. Soc. 1993, 115, 10462-

10463.

(13) Muedas, C. A.; Ferguson, R. R.; Brown, S. H.; Crabtree, R. H. J.

Am. Chem. Soc. 1991, 113, 2233-2242.

(14) (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.

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

12, 3705-3723.

(15) Brown, S. N.; Mayer, J. M. J. Am. Chem. Soc. 1994, 116, 2219-

2220.

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

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

1868.

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

Metal Catalyzed Oxidations of Organic Compounds; Academic Press: New

York, 1981.

(18) Jaynes, B. S.; Hill, C. L. J. Am. Chem. Soc. 1993, 115, 12212-

12213.

5878.

(30) (a) McDade, C.; Green, J. C.; Bercaw, J. E. Organometallics 1982,

1, 1629-1634. (b) Chamberlain, L. R.; Rothwell, I. P.; Huffmann, J. C. J.

Am. Chem. Soc. 1986, 108, 1502-1509. (c) Couturier, J.-L.; Paillet, C.;

Leconte, M.; Basset, J. M.; Weiss, K. Angew. Chem., Int. Ed. Engl. 1992,

31, 628-631. (d) van der Heijden, H.; Hessen, B. J. Chem. Soc., Chem.

Commun. 1995, 145-146.

(31) (a) Wasserman, E. P.; Moore, C. B.; Bergman, R. G. Science 1992,

255, 315-318. (b) Weiller, B. H.; Wasserman, E. P.; Bergman, R. G.;

Moore, C. B.; Pimentel, G. C. J. Am. Chem. Soc. 1989, 111, 8288-8290.

(32) Schultz, R. H.; Bengali, A. A.; Tauber, M. J.; Weiller, B. H.;

Wasserman, E. P.; Kyle, K. R.; Moore, C. B.; Bergman, R. G. J. Am. Chem.

Soc. 1994, 116, 7369-7377.

(33) Bengali, A. A.; Schultz, R. H.; Moore, C. B.; Bergman, R. G. J.

Am. Chem. Soc. 1994, 116, 9585-9589.

(34) Periana, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1986, 108, 7332-

7346.

(35) (a) Buchanan, J. M.; Stryker, J. M.; Bergman, R. G. J. Am. Chem.

Soc. 1986, 108, 1537-1550. (b) Bloyce, P. E.; Rest, A. J.; Whitwell, I. J.

Chem. Soc. Dalton Trans. 1990, 813-822.

(36) For methane complexes preceding elimination, see: (a) Parkin, G.;

Bercaw, J.E. Organometallics 1989, 8, 1172-1179. (b) Gould, G. L.;

Heinekey, D. M. J. Am. Chem. Soc. 1989, 111, 5502-5504. (c) Bullock,

R. M.; Headford, C. E. L.; Hennessy, K. M.; Kegley, S. E.; Norton, J. R.

J. Am. Chem. Soc. 1989, 111, 3897-3908.

(37) (a) Piers, W. E.; Shapiro, P. J.; Bunel, E. E.; Bercaw, J. E. Synlett.

1990, 74-84. (b) Piers, W. E.; Bercaw, J. E. J. Am. Chem. Soc. 1990, 112,

9406-9407.

(19) (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.

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

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

115, 4917-4918.

(22) Cummins, C. C.; Baxter, S. M.; Wolczanski, P. T. J. Am. Chem.

Soc. 1988, 110, 8731-8733.

(23) 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.

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

144 and references therein.

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

2179-2180.

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

Soc. 1994, 116, 4133-4134.

(27) 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.

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 593

activation in the zirconium imido system, where 1,2-RH-

elimination occurs from (^tBu₃SiNH)₃ZrR (1-R) to generate the

purported three-coordinate (^tBu₃SiNH)₂ZrdNSi^tBu₃(2). The

imido can then select to activate RH or R′H via 1,2-RH/R′H-

addition across the ZrdN bond.^43,44Herein is described the

culmination of an eight year effort²²to understand the funda-

mental nature of this process through synthetic, kinetics, isotope

effects, thermodynamics, and structural⁴⁵investigations.

observed in either NMR or IR spectra. Infrared stretching

absorptions corresponding to the NH functionality were found

in a 3218-3280 cm^-1range of the spectrum, consistent with a

relatively unperturbed ν(N-H).

Although alkyllithiums generally proved to be less efficient

when directly compared with Grignard reagents (e.g., MeLi vs

MeMgBr), in certain cases alkylations using RLi were simply

more convenient. The addition of 1 equiv of ^tBuCH₂Li to 1-Cl

in hexanes at -78 °C, followed by stirring for 8 h at ambient

t

Results and Discussion

temperature, generated colorless, crystalline 1-CH₂Bu (eq 2).

Synthesis and Characterization. 1. (^tBu₃SiNH)₃ZrR (R

) H, BH₄, Alkyl, Aryl). Three different routes led to the

preparation of hydride and hydrocarbyl tris-amido zirconium

derivatives: metathesis of halides with alkyl anion equivalents,

R-H bond activation (alkane/arene/H₂metathesis), and olefin

insertion. The procedures were usually not optimized, except

for those that provided critical starting materials (1-Me, 1-Cy)

for ensuing mechanistic and reactivity studies. Although many

of the reported isolated yields are modest, reactions that were

monitored by ¹H NMR spectroscopy, typically those involving

C-H bond activation or insertion, manifested conversions of

>90%. Extreme solubilities of the monomeric hydride and

hydrocarbyls have hampered their isolation. Table 1 lists the

+ ^tBuCH Li ^{hexanes, 25 °C, 8 h}8

t

( Bu₃SiNH)₃ZrCl

2

-LiCl

1-Cl

t

(^tBu₃SiNH)₃ZrCH₂Bu

(2)

t

1-CH₂Bu, 51%

Similarly, (^tBu₃SiNH)₃ZrCCPh (1-C₂Ph) was prepared by com-

bining 1-Cl and LiCtCPh in THF at -78 °C then stirring for

8 h at 25 °C; crystallization from hexanes afforded an orange,

+ PhCtCLi ^{THF, 25 °C, 8 h}8

t

( Bu₃SiNH)₃ZrCl

-LiCl

1-Cl

1

¹H and ¹³C{¹H} NMR spectral data for 1-R; H NMR spectra

(^tBu₃SiNH)₃ZrCtCPh

typically exhibit a prominent singlet ascribed to the three

(3)

t

equivalent Bu₃Si fragments, a broad singlet due to the amido

1-C₂Ph, 41%

protons, and resonances attributable to the remaining ligand.

Various hydrocarbyls were synthesized through treatment of

(^tBu₃SiNH)₃ZrCl (1-Cl)⁴⁵with appropriate Grignard reagents

RMgX in Et₂O at -78 °C, followed by stirring at 25 °C for

6-12 h. Recrystallization from hexanes afforded derivatives

(^tBu₃SiNH)₃ZrR (1-R, R ) Me,⁴⁵Et, Cy, CH₂Ph, Ph, CHdCH₂,

+ ^tBuCtCLi ^{C H , 25 °C, 4 h}8

6

(^tBu₃SiNH)₃ZrCl

-LiCl

1

(^tBu₃SiNH)₃ZrCtC^tBu

(4)

t

1-C₂Bu, 81%

^{Et O, 25 °C, 6-12 h}8

2

t

+

RMgX

X ) Cl, R ) Et

Cy

( Bu₃SiNH)₃ZrCl

-MgXCl

1-Cl

microcrystalline solid (eq 3). The related yellow tert-butyl

t

acetylide, (^tBu₃SiNH)₃ZrCC^tBu (1-C₂Bu), was synthesized from

CH₂Ph

t

1-Cl and BuCtCLi in benzene (eq 4).

X ) Br, R ) Me

Treatment of (^tBu₃SiNH)₃ZrCl (1-Cl) with 1 equiv of LiBH₄

at -78 °C in toluene, followed by 7 h at 25 °C, yielded the

colorless borohydride complex (^tBu₃SiNH)₃Zr(η³-BH₄) (1-BH₄,

CHdCH₂

CH₂CHdCH₂

Ph

1

eq 5). In the H NMR spectrum of 1-BH₄, a broad quartet at

(^tBu₃SiNH)₃ZrR

(1)

1-Et, 77%

+ LiBH ^{C H , 25 °C, 4 h}8

7

8

(^tBu₃SiNH)₃ZrCl

1-Cy, 46%

4

-LiCl

1-CH₂Ph, 35%

1-Me, 91%

1-CHdCH₂, 56%

1-Cl

(^tBu₃SiNH)₃Zr(η³-BH₄)

(5)

1-BH₄, 56%

1-allyl, 32%

1-Ph, 32%

allyl ) CH₂CHCH₂) as colorless crystalline solids in 32-91%

yield (eq 1). Evidence of agostic interactions⁴⁶between the

electrophilic zirconium center and any R-protons was not

δ 1.67 signified the presence of the BH₄ligand, which was also

observed as a quintet (J_BH) 81 Hz) in the ¹¹B NMR spectrum

at δ -20.48. In the infrared spectrum of 1-BH₄, a single sharp

A₁band in the terminal B-H stretching region at 2530 cm^-1

,

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

Acc. Chem. Res. 1995, 28, 154-162.

and two absorptions at 2210 and 2150 cm^-1due to the bridging

B-H A₁and E stretching modes, characterized the η³-

conformation.⁴⁷Attempts to form a hydride complex from

1-BH₄through cleavage of the BH₃group upon addition of

Lewis bases (e.g., NH₃, NMe₃) were unsuccessful, as were

additions of alternative hydride sources to chloride, 1-Cl.

Activation of solvent cyclohexane by transient imido (^tBu₃-

SiNH)₂ZrdNSi^tBu₃(2) has not been observed, hence thermoly-

sis of (^tBu₃SiNH)₃ZrR (1-R, R ) Me, Cy) in C₆H₁₂permitted

metathetical syntheses of new hydrocarbyls and convenient

syntheses of several others. In cyclohexane, thermolysis of 1-Cy

at ∼100 °C for 8 h in the presence of ∼20 equiv of cyclopropane

(39) (a) Maguire, J. A.; Petrillo, A.; Goldman, A. S. J. Am. Chem. Soc.

1992, 114, 9492-9498. (b) Maguire, J. A.; Boese, W. T.; Goldman, A. S.

J. Am. Chem. Soc. 1989, 111, 7088-7093. (c) Maguire, J. A.; Boese, W.

T.; Goldman, M. E.; Goldman, A. S. Coord. Chem. ReV. 1990, 97, 179-

192. (d) Sakakura, T.; Sodeyama, T.; Sasaki, K.; Wada, K.; Tanaka, M. J.

Am. Chem. Soc. 1990, 112, 7221-7229.

(40) Jones, W. D.; Hessell, E. T. Organometallics 1990, 9, 718-727.

(41) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis; Wiley-

Interscience: New York, 1992.

(42) Labinger, J. A. Catal. Lett. 1988, 1, 371-376.

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

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

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

P. T. Organometallics 1991, 10, 164-170.

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

36, 1-124.

(47) Marks, T. J; Kolb, J. R. Chem. ReV. 1977, 77, 263-293.

594 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Table 1. ¹H and ¹³C{1H} NMR Spectral Data for (^tBu₃SiNH)₃ZrR (1-R) and (^tBu₃SiNH)₂LZrdNSi^tBu₃(2-L) in C₆D₆Unless Otherwise Noted

¹H NMR (δ, assgmt, mult, J(Hz)) ¹³C NMR (δ, assgmt, J(Hz))

NH

4.89

Schaller et al.

compound

((H₃C)₃C)₃

R

C(CH₃)₃

R

(^tBu₃SiNH)₃ZrCl (1-Cl)

(^tBu₃SiNH)₃ZrH (1-H)

(^tBu₃SiNH)₃ZrMe (1-Me)

(^tBu₃SiNH)₃ZrEt (1-Et)

1.24

1.25

1.24

1.25

30.93

30.84

30.92

23.32

22.91

23.24

23.13

4.87

4.10

3.94

9.60 (H, s)

0.63 (CH₃, s)

28.68 (CH₃)

43.78 (CH₂)

15.49 (CH₃)

65.17 (CH₂)

31.27 (Me₂)

28.90 (CH)

1.77 (CH₃, t, 8)

(CH₂obscured)

(^tBu₃SiNH)₃ZrⁱBu (1-ⁱBu)

1.25

3.89

2.50 (CH, m)

30.93

23.22

1.39 (CH₂, d, 7)

1.27 (Me₂, d, 7)

(^tBu₃SiNH)₃Zr^cPr (1-^cPr)

(in C₆D₁₂)

1.25

1.18

3.83

3.77

0.75-1.00 (m)

31.23

31.21

23.31

23.62

0.87 (CH₂, ddd, 8, 5, 3)

0.74 (CH₂, ddd, 10, 5, 3)

2.68 (CH₂, dm, 14)

1.98 (CH₂, quar d, 13, 3)

1.88 (CH₂, dt, 13, 3)

1.78 (CH, m)

34.95 (CH, 131)

9.72 (CH₂, 161)

68.23 (CH)

36.05 (CH₂)

31.05 (CH₂)

28.02 (CH₂)

(^tBu₃SiNH)₃ZrCy (1-Cy)

1.26

3.89

30.92

23.15

1.43 (CH₂, quin t, 13, 3)

1.35 (CH₂, quin t, 13, 3)

6.61 (CH, quin, 11)

(^tBu₃SiNH)₃Zr(η³-H₂CCHCH₂)

(1-allyl)

1.23

1.25

4.18

4.06

30.89

31.04

23.18

23.23

142.05 (CH, 147)

3.64 ((CH₂)₂, d, 11)

6.08 (CH, t, 9)

81.88 (CH₂, 136)

138.75 (CH)

(^tBu₃SiNH)₃Zr(η³-H₂CCH)CMe₂)

(1-dma)

2.22 (CH₂, d, 9)

1.90 (Me, s)

121.41 (CMe₂)

54.55 (ZrCH₂)

18.24 (Me)

1.83 (Me, s)

(^tBu₃SiNH)₃ZrCH₂Bu

1.27

1.25

1.23

1.25

3.79

4.34

3.94

4.50

1.66 (Me₃C, s)

1.32 (CH₂, s)

31.03

30.98

23.23

74.24 (CH₂)

t

(1-CH₂Bu)

35.59 ((H₃C)₃C)

34.48 (Me₃C)

183.54 (CH)

134.8 (CH₂)

(^tBu₃SiNH)₃ZrCHdCH₂

(1-CHdCH₂)

7.65 (CH, dd, 17, 22)

6.30 (CH_t, dd, 4.5, 22)

6.63 (CH_c, dd, 4.5, 17)

8.9 (CH, s)

[(^tBu₃SiNH)₃Zr]₂

(µ₂:η¹,η¹-trans-C₂H₂)

(1₂-C₂H₂in C₆D₁₂)

(^tBu₃SiNH)₃ZrPh (1-Ph)

8.28 (Ph(o), dm, 7)

7.31 (Ph(m), tm, 7)

7.17 (Ph(p), tm, 7)

31.27

23.67

180.29 (C_ipso

138.77 (Ph)

128.60 (Ph)

127.18 (Ph)

183.01 (C_ipso

136.61 (Ar)

)

(^tBu₃SiNH)₃Zr]₂

(µ₂:η¹,η¹-1,4-C₆H₄)

(1₂-C₆H₄in C₆D₁₂)

(^tBu₃SiNH)₃ZrCH₂Ph

(1-CH₂Ph)

1.16

1.21

4.18

4.14

7.72 (Ar, s)

31.29

30.90

23.65

23.15

)

7.10-7.43 (Ar(2H), m)

6.81-6.92 (Ar(3H), m)

2.83 (CH₂, s)

148.97 (C_ipso

129.05 (Ar)

126.87 (Ar)

121.86 (Ar)

58.65 (CH₂, 119)

148.48 (C_ipso)

(^tBu₃SiNH)₃Zr-CH₂C₆H₃-3,5-Me₂

(1-Mes)

1.22

4.06

6.96 (Ar(2H), s)

30.93

23.16

6.50 (Ar(1H), s)

2.84 (CH₂, s)

2.23 (Me₂, s)

138.16 (Ar)

124.96 (Ar)

124.18 (Ar)

59.89 (CH₂)

21.55 (Me₂)

(^tBu₃SiNH)₃ZrCCPh

(1-CCPh)

1.32

1.29

4.86

4.57

7.56 (Ph(2H), m)

6.95 (Ph(3H), m)

1.21 (Me₃C, s)

(^tBu₃SiNH)₃ZrCC^tBu

(1-CC^tBu)

31.03

23.34

117.92 (ZrC)

104.33 (C^tBu)

31.44 ((H₃C)₃C)

30.93 (Me₃C)

¹¹B NMR δ -20.48

(η³-BH₄, quin,

J_BH) 81 Hz)

67.69 (OCH₂)

19.79 (CH₃)

(^tBu₃SiNH)₃Zr(η³-BH₄)

(1-BH₄in C₆D₁₂)

1.23

1.25

4.68

1.67 (BH₄, br quar)

30.94

31.03

23.33

23.26

(^tBu₃SiNH)₃ZrOCH₂CH₃

(1-OEt)

3.54

3.42

4.12 (OCH₂, q, 7)

1.21 (CH₃, t, 7)

(^tBu₃SiNH)₄Zr (1-NHSi^tBu)

1.28

1.24

31.27

31.04

23.34

23.13

(^tBu₃SiNH)₂ZrNHSi^tBuCMe₂CH₂(3)

3.97 (2H)

3.67 (1H)

1.57 (Me₂, s)

1.42 (CH₂, s)

1.22 (Me₃C, s)

74.89 (CH₂)

35.34 (C(CH₃)₂)

30.17 ((SiC(CH₃)₃)₂)

23.49 ((SiC(CH₃)₃)₂)

23.32 (C(CH₃)₂)

76.77 (OCH₂)

(^tBu₃SiNH)₂(THF)

ZrdNSi^tBu (2-THF)

(^tBu₃SiNH)₂(Et₂O)

ZrdNSi^tBu₃) (2-OEt₂)

(^tBu₃SiNH)₂(Me₃N)

ZrdNSi^tBu₃(2-NMe₃)

(^tBu₃SiNH)₂(Me₃P)

ZrdNSi^tBu₃(2-PMe₃)

(^tBu₃SiNH)₂(py)

1.29

1.44

1.29

1.42

1.27

1.42

1.26

1.42

1.30

1.44

3.85

3.83

3.80

4.36

4.18

4.02 (OCH₂, m)

1.13 (CH₂, m)

3.93 (OCH₂, quar, 7)

0.82 (CH₃, t, 7)

2.36 (Me₃, s)

31.16

31.83

31.14

31.61

23.10

24.09

23.13

24.10

25.22 (CH₂)

69.99 (OCH₂)

13.16 (CH₃)

1.02 (Me₃, d, 7)

8.89 (py(o), m)

6.69 (py(p), m)

6.44 (py(m), m)

ZrdNSi^tBu₃(2-py)

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

provided (^tBu₃SiNH)₃Zr(c-C₃H₅) (1-^cPr, eq 6). Inspection of

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 595

derivative was most conveniently prepared via solid state

1

isolated, colorless 1-^cPr by H NMR spectroscopy revealed

(^tBu₃SiNH)₃ZrR ^C⁶^D¹²^,-RH8 (^tBu₃SiNH)₂ZrNHSi^tBu₂CMe₂CH₂

(3) + (^tBu₃SiNH)₄Zr (1-NHSi^tBu₃) + ^tBu₃SiNH₂+ ... (11)

C₆H₁₂

t

100 °C, 3 h

+ c-C H (excess) _{100 °C, 8 h}8

( Bu₃SiNH)₃ZrCy

3

6

1-Cy

(^tBuSiNH) Zr(c-C H )

+ CyH (6)

+ CyH (7)

3

5

(^tBu₃SiNH)₃ZrMe

^{125 °C}8

1-^cPr, 35%

10 h

1-Me

C₆D₁₂

1-Cy + CH₄(4 atm) _{100 °C, 8 h}8

(^tBu₃SiNH)₂ZrNHSi^tBu₂CMe₂CH₂+ MeH

(12)

t

( Bu₃SiNH)₃ZrMe

3

1-Me

thermolysis of the methyl (1-Me) at 125 °C in dynamic vacuum

(eq 12). In this fashion the colorless “tuck-in” (3) formed with

only ∼5% of the tetraamide (1-NHSi^tBu₃) as a byproduct.

(^tBu₃SiNH)₃ZrPh (1-Ph) also proved to be a surprisingly good

substrate for C-H bond activation when (^tBu₃SiNH)₃ZrCy (1-

Cy) was thermolyzed (∼100 °C, ∼3 h) in cyclohexane (eq 13).

impurities that amounted to ∼5% of the ^tBu region; the principal

byproduct was determined to be (^tBu₃SiNH)₄Zr (1-NHSi^tBu₃).

In a sealed NMR tube, 1-Cy was heated (∼100 °C, 8 h) with

CH₄(4 atm) in C₆D₁₂to produce (^tBu₃SiNH)₃ZrMe (1-Me) in

essentially quantitative yield (eq 7).

1,2-Elimination of MeH (100 °C, 8 h) from 1-Me (eq 8) in

the presence of H₂(4 atm) afforded (^tBu₃SiNH)₃ZrH (1-H), a

consequence of dihydrogen addition across the imido ligand of

C₆D₁₂

+

^{, -CyH}8

t

C₆H₁₂

( Bu₃SiNH)₃ZrCy ( Bu₃SiNH)₃ZrPh

t

100 °C, 3 h

+ H (4 atm)

^{, 100 °C, 8 h}8

( Bu₃SiNH)₃ZrMe

2

1-Cy

1-Ph

1-Me

t

[(^tBu₃SiNH)₃Zr]₂(µ₂:η¹,η¹-1,4-C₆H₄) (1₂-C₆H₄) +

+ MeH (8)

( Bu₃SiNH)₃ZrH

1-H, 62%

At ∼80% conversion, the principal product (∼90%) of the

sealed NMR tube reaction was [(^tBu₃Si-NH)₃Zr]₂(µ₂:η¹,η¹-1,4-

putative intermediate 2. The terminal hydride of 1-H exhibited

a resonance at δ 9.60 in its ¹H NMR spectrum, and an IR stretch

at 1553 cm^-1(ν(ZrD) ) 1117 cm^-1).⁴⁵Hydride 1-H could be

similarly derived from (^tBu₃SiNH)₃ZrPh (1-Ph) and H₂in

cyclohexane-d₁₂.

t

C₆H ) (1₂-C H₄), manifesting a di-p-activation of the benzene

( Bu⁴₃SiNH)₂⁶ZrNHSi Bu₂CMe₂CH₂(3)

(13)

ring.^24,48The colorless para-dizirconabenzene complex (1₂-

C₆H₄) was sparingly soluble, partially crystallizing from the

C₆D₁₂solution (0.65 mL) where the initial [Zr] ∼ 0.05 mM.

Attempts to cleanly prepare (^tBu₃SiNH)₃ZrCHdCH₂(1-

CHdCH₂) afforded similar results. In a sealed NMR tube,

(^tBu₃SiNH)₃ZrMe (1-Me) was exposed to 10 equiv of ethylene

in cyclohexane-d₁₂. After sitting at 25 °C for 2 wk, the

Thermolysis of (^tBu₃SiNH)₃ZrMe (1-Me) at ∼100 °C for 7

h in benzene or mesitylene afforded the respective phenyl, 1-Ph,

and mesityl, (^tBu₃SiNH)₃Zr-CH₂C₆H₃Me₂(1-Mes) derivatives

(eq 9). A similar experiment in toluene led to a mixture of

benzyl (1-CH₂Ph), and para- and meta-aryl products (eq 10)

1

accumulation of a white precipitate was noted, but H NMR

spectroscopy indicated that 1-Me and C₂H₄were still present.

Upon thermolysis at ∼100 °C, 1-CHdCH₂appeared during the

course of 1 h (∼24%), but was accompanied by 6% [(^tBu₃-

SiNH)₃Zr]₂(µ₂:η¹,η¹-trans-C₂H₂) (1₂-C₂H₂), the doubly activated

dizirconaethylene complex. After ∼8 h, the ratio (by metal)

of 1-CHdCH₂:1₂-C₂H₂was 58:42, only 4% of the starting

RH

(^tBu₃SiNH)₃ZrR

(^tBu₃SiNH)₃ZrMe

_{100 °C, 7 h}8

+ MeH

R ) Ph, 1-Ph, 59%

1-Me

CH₂C₆H₃Me₂, 1-Mes, 62%

(9)

C₇H₈

(^tBu SiNH) ZrCH Ph

1-Me _{100 °C, 9 h}8

+

3

2

+ H CdCH (excess) ^{C D}^{, -MeH}8

6

12

1-CH₂Ph

(^tBu₃SiNH)₃ZrMe

2

100 °C, 8 h

(^tBu₃SiNH)₃ZrR′

+ MeH (10)

1-Me

[(^tBu₃SiNH)₃Zr]₂(µ₂:η¹,η¹-trans-C₂H₂) (1₂-C₂H₂) +

(^tBu₃SiNH)₃ZrCHdCH₂(1-CHdCH₂) + ... (14)

R′ ) p-C₆H₄Me, m-C₆H₄Me

1-C₆H₄Me

whose relative composition changed with time (Vide infra). Since

the para:meta ratio of roughly 2:1 was relatively stable over

the 9 h thermolysis, the aryl products were treated together (1-

C₆H₄Me) in ensuing studies.

methyl complex remained, and the remaining ethylene had been

completely converted to polyethylene, assumed to be the white

precipitate (eq 14). Still seeking an alternative preparation of

1-CHdCH₂, (^tBu₃SiNH)₃ZrH (1-H) was treated with excess (4

equiv) acetylene in C₆D₆, and the vinyl complex (1-CHdCH₂)

formed immediately with concomitant polyacetylene (eq 15).

Its subsequent thermolysis (C₆D₁₂, 100 °C, 2 h) in another sealed

NMR tube again produced the dizirconaethylene species 1₂-

When the cyclohexyl or methyl derivatives 1-Cy and 1-Me

were thermolyzed at ∼100 °C in cyclohexane in the absence

of added substrate, alkane extrusion ultimately led to the

t

formation of Bu₃SiNH₂, tetraamide (^tBu₃SiNH)₄Zr (1-NH-

Si^tBu₃), and an intramolecular C-H activation product, (^tBu₃-

C₆D₆

SiNH)₂ZrNHSi^tBu₂CMe₂CH₂(3), among other, uncharacterized

products (eq 11). Attempts to activate ethane in cyclohexane

produced similar results, a clear indication that activation of

C-H bonds by the three-coordinate intermediate, (^tBu₃-

SiNH)₂ZrdNSi^tBu₃(2), was quite selective. The cyclometalated

t

+ HCtCH (excess) _{25 °C, 5 min}8

( Bu₃SiNH)₃ZrH

1-H

(^tBu₃SiNH)₃ZrCHdCH₂

(15)

1-CHdCH₂

596 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

C₆D₁₂

a correspondence between downfield NH shifts and a more

electrophilic metal center in a related titanium system.²³Minor

trends in NH stretching frequencies were provided by infrared

studies of (^tBu₃SiNH)₃ZrR (1-R), along with useful fingerprint

information. The relatively wide range of ν(N-H) (3218-3280

cm^-1) was indicative of a subtle interplay between the N(pπ)

f Zr(dπ) interaction and the R-substituent, whose steric bulk

and electronegativity presumably influence the sp²character of

the amide nitrogen. Partial disruption of N(pπ) f Zr(dπ)

donation should lessen the stretching frequency of the NH by

slightly diminishing its sp²character. Large hydrocarbyl ligands

1-CHdCH_{2 100 °C, 2 h}8

[(^tBu SiNH) Zr] (µ :η¹,η¹-trans-C H )

+ ... (16)

3

2

1₂-C₂H₂

C₂H₂(eq 16). Trace amounts of an olefin polymerization

catalyst apparently coexist with the identifiable zirconium

amides. When ethylene is released from 1-CHdCH₂in a 1,2-

elimination event, transient imido (^tBu₃SiNH)₂ZrdNSi^tBu₃(2)

may recapture it or activate the trans vinylic position in

1-CHdCH₂. Since the polymerization reaction drains away

C₂H₄, the major product becomes 1₂-C₂H₂, which is assumed

to be trans on the basis of molecular models.

Insertion of isobutylene into the zirconium-hydride bond of

in situ generated (^tBu₃SiNH)₃ZrH (1-H) occurred swiftly at 25

°C in hexane solution under 0.5 atm isobutylene to give the

isobutyl complex (^tBu₃SiNH)₃ZrCH₂CHMe₂(1-ⁱBu), which was

isolated in poor yield (eq 17). Although the reaction was clean

t

(e.g., CH₂Bu (3221 cm^-1), ⁱBu (3228), dma (3238), Cy (3230),

Mes (3234), CH₂Ph (3238), Ph (3239)) possess roughly lower

ν(N-H) than corresponding small substituents (e.g., H (3280

cm^-1), Me (3242), ^cPr (3245), CHdCH₂(3245), Et (3248), C₂-

t

Ph (3258), C₂Bu (3265)). Subtle, yet significant steric influ-

ences on the ^tBu periphery of the tris-amide coordination sphere

may slightly distort the ZrNHSi^tBu₃linkages, perhaps lessening

amide π-bonding. More electronegative functionalities, such

as Cl (3255 cm^-1) and the acetylides, may promote a slight

increase in N(pπ) f Zr(dπ) donation relative to the hydrocarbyl

ligands. Tetraamide 1-NHSi^tBu₃represents an anomaly; al-

though more electronegative than the majority of the R groups,

its tremendous size and competing amide π-donation are

reflected in the lowest frequency observed, 3218 cm^-1. The

hydride, 1-H, manifests almost no steric influence and represents

the highest member of the series (3280 cm^-1), despite the lesser

electronegativity accorded H in respect to its hydrocarbyl

cognates.

+ Me CdCH (0.5 atm) ^{25 °C, 20 min}8

t

( Bu₃SiNH)₃ZrH

2

hexane

1-H

(^tBu₃SiNH)₃ZrCH₂CHMe₂

(17)

1-ⁱBu, 15%

25 °C, 3 h

(^tBu SiNH) ZrCH CHdCMe

1-H + H₂CdCdCMe_{2 benzene}8

3

2

1-dma, 24%

(18)

(>90%) when conducted in a sealed NMR tube, the extreme

solubility of 1-ⁱBu hampered crystallization and isolation efforts.

Exposure of 1-H to 3-methyl-1,2-butadiene (1,1-dimethylallene)

yielded the highly soluble, hindered allyl derivative, (^tBu₃-

SiNH)₃ZrCH₂CH)CMe₂(1-dma, dma ) 3,3-dimethylallyl,

24%) (eq 18).

2. (^tBu₃SiNH)₂LZrdNSi^tBu₃. Bis-amidoimido adducts

(^tBu₃SiNH)₂LZrdNSi^tBu₃(2-L) were obtained via 1,2-elimina-

tion reactions of appropriate precursor complexes. Spectral

characteristics of the compounds, some of which were only

prepared in situ on an NMR tube scale, are listed in Table 1.

Thermolysis of (^tBu₃SiNH)₃ZrMe (1-Me) at 100 °C for 10 h

in THF, followed by precipitation from hexanes at -78 °C,

afforded (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃(2-THF) as a colorless

NMR spectra of allyl derivatives (^tBu₃SiNH)₃Zr(CH₂CHCH₂)

(1-allyl) and (^tBu₃SiNH)₃-ZrCH₂CH)CMe₂(1-dma) differed

significantly in appearance. The ¹³C{¹H} NMR spectrum of

1-allyl revealed shifts of δ 142.05 and δ 81.88 for the respective

central and terminal η³-allyl carbons. The central proton

THF

_{100 °C, ∼1 h}8

(^tBu₃SiNH)₃ZrMe

1-Me

1

resonates as a quintet at δ 6.61 (J ) 11 Hz) in the H NMR

(^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃

+ MeH (19)

spectrum, while the syn and anti protons appear as a doublet at

δ 3.64 (J ) 11 Hz), indicative of an η³-allyl undergoing rapid

syn/anti exchange.⁴⁹In contrast, 1-dma is best construed as an

η¹-allyl, since its ¹³C{¹H} spectrum exhibits olefinic resonances

at δ 121.41 and 138.75, accompanied by the ZrCH₂- carbon

at δ 54.55, and methyl constituents at δ 18.24 and 26.39. In

2-THF, 81%

powder in 81% yield (eq 19). The reaction conditions are

reminiscent of those observed in hydrocarbon metatheses

involving 1-R, suggesting that the THF ligand trapped transient

(^tBu₃SiNH)₂ZrdNSi^tBu₃(2) subsequent to 1,2-elimination of

MeH. Adduct 2-THF was also synthesized on a preparative

scale from (^tBu₃SiNH)₃ZrPh (1-Ph) and was routinely produced

and isolated in surprisingly good purity starting from vestiges

of the hydrocarbyl (1-R) syntheses.

Adduct formation appeared to occur in a distinctly different

fashion when (^tBu₃SiNH)₃ZrH (1-H)⁴⁵was employed as the

imide source. Exposure of 1-H to L (L ) THF, Et₂O, NMe₃,

PMe₃) at 25 °C in sealed NMR tubes led to the immediate

evolution of H₂, producing (^tBu₃SiNH)₂LZrdNSi^tBu₃(2-L, L

) THF, Et₂O, NMe₃, PMe₃) in near quantitative yields according

to ¹H NMR (eq 20). The procedure was easily scaled up; using

1

the H NMR spectrum, methyl singlets at δ 1.83 and 1.90 are

accompanied by a ZrCH₂- doublet at δ 2.22 (J ) 9 Hz) and a

vinylic triplet at δ 6.08 (J ) 9 Hz). An asymmetric η³-allyl,

even one dimethylated as in 1-dma, is not expected to display

signals of the dispersity noted in the ¹³C NMR spectrum, and

the resonances of the methylene are clearly more consistent with

an η¹-, rather than η³-allylic ZrCH₂- unit. Moreover, molecular

models strongly suggest that an η³-disposition of the dma ligand

could not be accommodated by the “pocket” formed by the

probable C₃disposition⁴⁵of the amides in the (^tBu₃SiNH)₃Zr

moiety.

Correlations of NH shift vs R-substituent σ- and π-donation

1

to zirconium were not evident in H NMR spectra of (^tBu₃-

C₆D₆

SiNH)₃ZrCl (1-Cl), (^tBu₃SiNH)₃ZrR (1-R, R ) H, BH₄,

hydrocarbyl), and (^tBu₃SiNH)₄Zr (1-NHSi^tBu₃), in contrast to

t

+ L _{25 °C, <5 m}8

( Bu₃SiNH)₃ZrH

1-H

(^tBu₃SiNH)₂LZrdNSi^tBu₃

+ H₂

(20)

(48) (a) Hunter, A. D.; Szigety, A. B. Organometallics 1989, 8, 2670-

2679. (b) Chukwu, R.; Hunter, A. D.; Santarsiero, B. D.; Bott, S. G.;

Atwood, J. L.; Chassaignac, J. Organometallics 1992, 11, 589-597.

(49) 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.

2-L, L ) THF, Et₂O, NMe₃, PMe₃

the dregs from a synthesis of 1-H as starting material, the

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 597

Table 2. Crystallographic Data for (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃

(2-THF)

formula: C₄₀H₉₁N₃OSi₃Zr

formula weight: 805.67

space group: P2₁/n

crystal dimensions: 0.3 × 0.3 × 0.3 mm

µ ) 3.14 cm^-1(no absorption correction)

Nicolet R3m/V: 2θ range: 0-45° ((h, k, l)

reflections: 7135

a ) 13.312(5) Å

b ) 18.268(6) Å

c ) 20.551(7) Å

â ) 92.30(3)°

V ) 4994(3) Å³

Z ) 4

T ) 25 °C

independent reflections: 6557 (R_int) 3.64%)

λ ) 0.71069 Å

(Mo K_R)

reflections (F > 3σ(F)): 4675

R ) 7.70%; R_w) 9.12% (4675 reflections)

w^-1) σ²(F) + 0.0015F²)

F

_calc) 1.072 g/cm³

GOF ) 1.46 (4675

reflections)

addition of Et₂O produced 2-OEt₂in ∼45% isolated yield after

crystallization. Unfortunately, since 1-H remained one of the

more difficult derivatives to prepare, the production of imido

adducts via 1,2-dihydrogen elimination possessed limited utility.

By monitoring a sealed NMR tube of 2-OEt₂and dihydrogen

Figure 1. Molecular view of (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃(2-THF).

₃+ H y^{C D}

6

z

(^tBu₃SiNH)₂(Et₂O)ZrdNSi^tBu

2-OEt₂

Table 3. Selected Interatomic Distances (Å) and Angles (deg) for

(^tBu₃SiNH)₂(THF)ZrdN-Si^tBu₃(2-THF)

2

25 °C

Zr-O

2.229(7)

1.978(8)

2.031(8)

2.028(8)

N1-Si1

N2-Si2

N3-Si3

Si-C (av)

C-C (av)

1.720(9)

1.717(9)

1.728(8)

1.922(14)

1.543(92)

O-C1 1.394(19)

O-C4 1.434(17)

C1-C2 1.435(26)

C2-C3 1.400(28)

C3-C4 1.435(23)

t

+ Et O (21)

( Bu₃SiNH)₃ZrH

Zr-N1

Zr-N2

Zr-N3

2

1-H

(∼3 atm), equilibration with 1-H and Et₂O was noted over the

course of 19 days (eq 21). Through direct integration of all

components, a rough K_eq(25 °C) of 4.7 (∆G° ) -0.9 kcal/mol,

all species 1 M standard states) was obtained.

The displacement of THF from 2-THF by another ligand

occurred with facility provided L was a stronger Lewis base.

Exposure of 2-THF in C₆D₆to 1.0 of equiv pyridine in a sealed

NMR tube (eq 22) caused the immediate formation of (^tBu₃-

SiNH)₂(py)ZrdNSi^tBu₃(2-py) and concomitant free THF, as

O-Zr-N1 102.9(3) Zr-N1-Si1 156.9(5) Zr-O-C1 129.0(9)

O-Zr-N2 105.3(3) Zr-N2-Si2 154.7(5) Zr-O-C4 123.5(7)

O-Zr-N3 107.7(3) Zr-N3-Si3 158.1(5) O-C1-C2 107.0(15)

N1-Zr-N2 112.3(3) N-Si-C (av) 107.9(8) O-C4-C3 108.0(13)

N2-Zr-N3 114.9(3) Si-C-C (av) 111.9 (48) C1-C2-C3 108.1(16)

N1-Zr-N3 112.5(3) C-Si-C (av) 110.9(16) C2-C3-C4 106.4(15)

C-C-C (av) 106.5(42)

Si1 plane are canted such that the N-H/lone pair vectors

(assuming sp²-N-hybridization) are directed slightly toward the

THF apex (dihedral angles: O-Zr-N2-H, 79.9°; O-Zr-N3-

H, 90.4°; O-Zr-N1-(lone pair) ) 83.4°), rendering each

^tBu₃Si fragment approximately equatorial relative to the Zr-O

bond. In this fashion, the distortion from approximate tetra-

hedral geometry is minimal, yet the amides/imide occupy a less

sterically demanding region of space, just as observed for C₃

1-Me, whose H-N-Zr-C dihedral angles are ∼105°.⁴⁵

The plane of the THF ligand is nearly aligned with the Zr-O

vector ( Zr-O-C ) 123.5(7), 129.0(9)°), implicating strong

O(pπ) f Zr(dπ) bonding, but the d(Zr-O) of 2.229(7) Å belies

this argument. In Cp₂(THF)ZrdN^tBu,²⁸the plane of the THF

lies approximately within the Cp₂Zr “wedge” with Zr-O-C

angles of 124.3(4) and 127.3(6)°, and oxygen π-bonding is

negligible because the perpendicularly oriented oxygen lone pair

must π-donate into high energy, CpZr antibonding orbitals.

Despite this factor, the Zr-O bond distance of 2.240(4) Å is

essentially the same as that in 2-THF, hence it is unlikely that

either structure manifests significant oxygen π-donation. Con-

trast this data to the d(Zr-O) of 2.122(14) Å in Cp₂ZrMe-

(THF)⁺,⁵¹where the THF plane is oriented perpendicular to

the Cp₂Zr wedge, thereby ensuring O(pπ) f Zr(dπ) bonding,

despite severe steric constraints.

C₆D₆

(^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃

+ py _{25 °C, <5 m}8

2-THF

(^tBu₃SiNH)₂(py)ZrdNSi^tBu₃

+ THF (22)

2-py

noted by ¹H NMR spectroscopy. In contrast, multiple exposures

(∼10) of 2-THF to diethyl ether solvent were needed to ensure

the formation of (^tBu₃SiNH)₂(Et₂O)ZrdNSi^tBu₃(2-OEt₂) in

>98% purity (¹H NMR) upon isolation from Et₂O (eq 23).

Et₂O, 25 °C

(^tBu SiNH) (Et O)ZrdNSi^tBu

₃+ THF (23)

2-THF y

z

3

2

2-OEt₂

3. Molecular Structure of (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃

(2-THF). A single-crystal X-ray structure determination (mon-

oclinic, P2₁/n, R ) 7.70%, R_w) 9.12%, Table 2) of (^tBu₃-

SiNH)₂(THF)ZrdNSi^tBu₃(2-THF) confirmed its constitution

and geometry. The molecular view of 2-THF is shown in Figure

1, while selected interatomic distances (Å) and angles (deg) are

listed in Table 3. The figures reveal a spiral, C₃“propeller”

arrangement of the sp²-nitrogen-containing imide (Zr-N1-Si1)

and amide (Zr-N2(H)-Si2, Zr-N3(H)-Si3) ligands, a feature

reminiscent of the amide conformations about (^tBu₃SiNH)₃ZrMe

(1-Me)⁴⁵and {(Me₃Si)₂N}₃MCl (M ) Ti, Zr, Hf).⁵⁰The

O-Zr-N angles (105.3(24)° av) are significantly less than the

N-Zr-N angles (113.2(14)° av), in accord with pseudotetra-

hedral symmetry and the greater repulsion between the imide/

amide ligands. Both Zr-N(H)-Si planes and the ZrdN(1)-

The most striking feature of 2-THF concerns the similarity

of the imide and amide linkages. The imide Zr-N1-Si1 angle

(156.9(5)°) is essentially the same as the amide Zr-N-Si angles

(154.7(5)° and 158.1(5)°), while the Zr-N1 bond distance of

1.978(8) Å is ∼0.05 Å shorter than the amide 2.031(8) and

2.028(8) Å bond lengths of Zr-N2 and Zr-N3, respectively.

Minimal lone pair N(pπ) f Zr(dπ) donation is implicated by

(50) Airoldi, C.; Bradley, D. C.; Chudzynska, H.; Hursthouse, M. B.;

Malik, K. M. A.; Raithby, P. R. J. Chem. Soc., Dalton Trans. 1980, 2010-

2015.

(51) Jordan, R. F.; Bajgur, C. S.; Willett, R.; Scott, B. J. Am. Chem.

Soc. 1986, 108, 7410-7411.

598 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

the similarity, proViding a structural basis upon which to predict

its reactiVity as in internal base in C-H bond actiVation

processes. The short amide distances, characteristic of low

coordinate group 4 complexes,^45,50,52,53manifest nitrogen π-do-

nation that is nearly as strong as the formal π-bond of the imide

functionality. The decidedly angular ZrdN-Si fragment

represents a severe departure from typical MdN-R (R )

hydrocarbyl) ligands, whose range of angles is 165-180°.⁵⁴In

complexes containing 18 e^-(without inclusion of the imide lone

pair) or hydrazido ligands, exceptions ( MdN-X < 165°) have

been found, but known zirconium derivatives Cp₂(THF)Zrd

N^tBu,²⁸((2,6-ⁱPr₂C₆H₃)HN)₂(py′)₂ZrdN(2,6-ⁱPr₂C₆H₃) and ((2,6-

ⁱPr₂C₆H₃)O)₂(py′)₂ZrdNPh (py′ ) 4-pyrrolidinopyridine)⁵⁵ex-

hibit nearly linear ZrdN-C angles of 174.4(3), 174.9(3), and

175.5(8)°, respectively.

In view of the recent debunking of bond-stretch isomerism

by Parkin et al.,⁵⁶an imide/amide disorder in the structural

model may be responsible for the unusually small difference in

Zr-N imide and amide bond lengths; a weighted average of

“normal” ZrdN- and Zr-N(H)- bond distances (typically

∼0.2 Å apart) is plausible. However, the amide distances of

2-THF are effectively the same as those in 1-Me (2.039(7) Å

av),⁴⁵rendering this possibility considerably less appealing, since

zirconium imide distances are substantially shorter than corre-

sponding amides. In Cp₂(THF)ZrdN^tBu (d(ZrdN) ) 1.826(4)

Å),²⁸((2,6-ⁱPr₂C₆H₃)HN)₂(py′)₂ZrdN(2,6-ⁱPr₂C₆H₃), and ((2,6-

ⁱPr₂C₆H₃)O)₂(py′)₂ZrdNPh (d(ZrdN) ) 1.868(3), 1.844(9) Å,

respectively),⁵⁵the imide bond distances are significantly shorter

despite ligation to higher-coordinate metal centers where longer

imide bonds are expected. Given a hypothetical 1.86 Å ZrdN

distance for 2-THF, each amide would have to be an unchar-

acteristically long 2.09 Å for the structure to represent a

weighted average of Zr-N bond lengths; furthermore, an

explanation for the statistically different observed amide and

imide distances would still be lacking.

20e^-configurations. When the amides are disposed such that

the N lone pairs are perpendicular to the N₃M plane, all three

ligand orbitals possess the correct symmetry for interaction with

the metal. Even in pseudo-C₃tetrahedral molecules such as

2-THF, these orbital arguments remain pertinent. As a conse-

quence, there is an electronic, as well as steric, component to

the orientation of the NHSi^tBu₃groups about zirconium in 1-Me

and 2-THF. Since the THF oxygen is a poorer donor, the

amides prefer an arrangement that aligns their lone pairs with

the Zr-O vector, and the additional steric constraints cause each

amide to be slightly tipped. Space-filling depictions of 2-THF

indicate that the orientation of the amides provide a strong

impetus for bending of the imide linkage. In essence, the bent

amide bonds permit only a similarly angular imide ligand to

occupy the remaining space in the C₃core. The amide

hydrogens and imide lone pair reside in nearly equivalent

positions. Based on the Zr-N-Si angles (156.6(17)° average),

neither NH nor the lone pair is expected to engage an e set of

metal orbitals in a significant agostic interaction or π-bonding,

although correlations of M-O-R and d(M-OR) have shown

that such inferences can be misleading.⁵⁹

Less well understood is the electronic influence of the ^tBu₃-

Si moiety on the conformation and d(Zr-N) of the imide. If

N(pπ) f Si(dπ) bonding is significant, competition between

Si and Zr for nitrogen π-bonding may partly explain the

lengthened imide bond distance. However, the limited number

of comparative MdN-R (R ) CR′₃vs SiR′₃) structures do

not show a pronounced difference in bond length,⁵⁴and the

existence of significant, related O(pπ) f Si(dπ) interactions is

a matter of some dispute.⁶⁰

4. (^tBu₃SiNH)₂(Et₂O)ZrdNSi^tBu₃Thermolysis. In a sealed

NMR tube, diethyl ether adduct (^tBu₃SiNH)₂(Et₂O)ZrdN-

Si^tBu₃(2-OEt₂) decomposed at 120 °C (C₆D₆, 12 h) to generate

(^tBu₃SiNH)₃ZrOEt (1-OEt, ∼90%) and concomitant ethylene,

consistent with γ-CH activation (eq 24). Scale-up of the

ethoxide derivative, 1-OEt, was accomplished via a direct route

Provided the structural characterization of 2-THF is accurate,

what factors contribute to the unusual imide disposition? In

trigonal (R₂N)₃M complexes (e.g., R₂N ) N(SiMe₃)₂), the

amides adopt a pinwheel dispostion about M, but the N lone

pair is located principally in the N₃M plane. When restricted

to this plane, the N(pπ)-based a₂′ (assuming D_3h) ligand orbital

is not of the appropriate symmetry to interact with any metal

orbital. In d²Os(dNAr)₃and related complexes,^57,58these

symmetry arguments have been used to rationalize apparent

C₆D₆

_{120 °C, 12 h}8

(^tBu₃SiNH)₂LZrdNSi^tBu₃

2-OEt₂

t

+ H CdCH (24)

( Bu₃SiNH)₃ZrOEt

2

1-OEt

heptane

t

+ Et O (excess) _{115 °C, 16 h}8

( Bu₃SiNH)₃ZrCy

2

1-Cy

(52) (a) Planalp, R. P.; Andersen, R. A.; Zalkin, A. Organometallics

1983, 2, 16-20 and references therein. See, also: (b) Andersen, R. A. Inorg.

Chem. 1979, 18, 2928-2932. (c) Andersen, R. A. J. Organomet. Chem.

1980, 192, 189-193.

(^tBu₃SiNH)₃ZrOEt

H₂CdCH₂(25)

1-OEt

(53) (a) Fryzuk, M. D.; Williams, H. D.; Rettig, S. J. Inorg. Chem. 1983,

22, 863-868. (b) Fryzuk, M. D.; Carter, A.; Westerhaus, A. Inorg. Chem.

1985, 24, 642-648. (c) Fryzuk, M. D.; Rettig, S. J.; Westerhaus, A.;

Williams, H. D. Inorg. Chem. 1985, 25, 4316-4325. (d) Fryzuk, M. D.;

Haddad, T. S.; Rettig, S. J. Organometallics 1989, 8, 1723-1732.

(54) (a) Nugent, W. A.; Mayer, J. M. Metal-Ligand Multiple Bonds;

Wiley-Interscience: New York, 1988. (b) Wigley, D. E. Prog. Inorg. Chem.

1994, 42, 239-482.

involving thermolysis of (^tBu₃SiNH)₃ZrCy (1-Cy) in heptane

at 115 °C for 16 h in the presence of ∼10 equiv Et₂O (eq 25).

The formal γ-CH-activation product was isolated in 79% yield

upon crystallization from hexanes.

This apparently internal γ-CH activation is reminiscent of

Parkin’s recent studies of Cp*₂Zr(dO)py (Cp* ) η⁵-C₅Me₅),

which reacts with ^tBuI and RCOMe (R ) Me, ^tBu, Ph) to give

Cp*₂ZrI(OH) and Cp*₂Zr(OH)OCRdCH₂, respectively.²¹

Scheme 2 illustrates plausible mechanisms for these transforma-

tions, highlighting a common, concerted six-atom rearrangement

where the oxo or imido unit functions as an internal base in

attacking the γ-CH bond.

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

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

(56) (a) Parkin, G. Acc. Chem. Res. 1992, 25, 455-460. (b) Parkin, G.

Chem. ReV. 1993, 93, 887-912.

(57) (a) Schofield, M. H.; Kee, T. P.; Anhaus, J. T.; Schrock, R. R.;

Johnson, K. H.; Davis, W. M. Inorg. Chem. 1991, 3595-3604. (b) Williams,

D. S.; Anhaus, J. T.; Schofield, M. H.; Schrock, R. R.; Davis, W. M. J.

Am. Chem. Soc. 1991, 113, 5480-5481. (b) Williams, D. S.; Anhaus, J.

T.; Schofield, M. H.; Schrock, R. R.; Davis, W. M. J. Am. Chem. Soc.

1991, 113, 5480-5481.

1,2-RH-Elimination: Labeling Studies. The prevalence of

σ-bond metatheses^8,9in early transition metal chemistry prompted

a series of labeling studies. Under standard thermolysis

conditions, (^tBu₃SiNH)₃ZrMe (1-Me) in C₆D₆afforded (^tBu₃-

(58) (a) Covert, K. J.; Neithamer, D. R.; Zonnevylle, M. C.; LaPointe,

R. E.; Schaller, C. P.; Wolczanski, P. T. Inorg. Chem. 1991, 30, 2494-

2508. (b) Eppley, D. F.; Wolczanski, P. T.; Van Duyne, G. D. Angew.

Chem., Int. Ed. Engl. 1991, 30, 584-585.

(59) Steffey, B. D.; Fanwick, P. E.; Rothwell, I. P. Polyhedron 1990, 9,

963-968.

(60) Shambayatei, S.; Blake, J. F.; Wierschke, S. G.; Jorgensen, W. L.;

Schreiber, S. L. J. Am. Chem. Soc. 1990, 112, 697-703; 6155.

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 599

Scheme 2

SiND)₃ZrC₆D₅(1-(ND)₃C₆D₅) and CH₄, while (^tBu₃SiND)₃ZrMe

in C₆H₆provided 1-Ph and CH₃D (0.9 equiv, >93% d₁by H

4.91 × 10^-4s^-1in neat C₆D₆(∼11.2 M), hence typical

conditions manifest zero-order in benzene, and cyclohexane

cannot have a pronounced solvent effect. As an additional

check, rough monitoring of the conversion of 1-Cy and methane

(1, 2, 4, and 8 atm) to 1-Me and cyclohexane in C₆D₁₂(eq 7)

1

(^tBu₃SiNX)₃ZrMe

C₆Y₆

_{100 °C, 7 h}8

X ) H, 1-Me; Y ) D

X ) D, 1-(ND)₃-Me; Y ) H

d[1-Me]/dt ) -d[1-Cy]/dt ) k_Cy[1-Cy]¹[MeH]⁰(29)

(^tBu₃SiNY)₃ZrC₆Y₅

X ) H; Y ) D, 1-(ND)₃-C₆D₅

X ) D; Y ) H, 1-Ph

+ X-CH₃(26)

revealed that neither disappearance of 1-Cy (k_Cy(av) ) 1.55

(8) × 10^-3s^-1) nor appearance of 1-Me (k_Me(av) ) 1.5 (1) ×

10^-3s^-1) exhibited any order in methane (eq 29). When

thermolyzed in THF-d₈, the rate (98.7 °C, k_Me) 6.27(9) × 10^-4

s^-1) of MeH-elimination from (^tBu₃SiNH)₃ZrMe (1-Me) re-

flected a 5-fold increase relative to benzene-d₆, despite the

difference in products (2-THF vs 1-(ND)₃-C₆D₅). As a conse-

quence of these experiments, preequilibrium binding of alkane

or solvent is precluded as a significant factor in the 1,2-RH-

elimination events; no further solvent studies were undertaken.

Note that some related reactions appear to be second-order.

In eq 20, (^tBu₃SiNH)₃ZrH (1-H) is seen to be a ready precursor

to (^tBu₃SiNH)₂LZrdNSi^tBu₃(2-L, L ) THF, Et₂O, NMe₃,

PMe₃) upon addition of L (<5 min), a process whose time scale

is suggestive of second-order. Spin saturation transfer experi-

ments revealed that the exchange of free THF with its bound

counterpart in (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃(2-THF) is [THF]-

dependent. The thermoneutrality of this swift exchange may

help enable its observation; assuming a second-order process,

C₆H₁₂

,

t

+ D (4 atm) _{100 °C, 8 h}8

( Bu₃SiNH)₃ZrMe

2

1-Me

t

+ CH (27)

( Bu₃SiND)₃ZrD

4

1-(ND)₃-D

NMR, IR) as shown in eq 26. In a related process, the ultimate

products of deuteration (∼4 atm) of 1-Me in C₆D₁₂were (^tBu₃-

SiND)₃ZrD (1-(ND)₃D) and concomitant CH₄(eq 27). σ-Bond

metathesis pathways are convincingly obviated by these results,

yet the riddle of amide H/D-exchange was introduced and solved

by ensuing kinetics studies. The labeling experiments impli-

cated the mechanism depicted in Scheme 1, where 1,2-RH-

elimination from (^tBu₃SiNH)₃ZrR (1-R) produces transient three-

coordinate imido, (^tBu₃SiNH)₂ZrdNSi^tBu₃(2), and 1,2-R′H-

addition to its reactive ZrdN bond generates a new hydrocarbyl,

1-R′.

k ∼55-85 M^-1s^-1

primary kinetic isotope effect (KIE)⁶⁴for loss of CH₃H/D from

.

Kinetics of 1,2-RH-Elimination. 1. Order and Activation

Parameters. Kinetics investigations portray the 1,2-RH-

elimination as the rate-determining step (Table 4). Thermolysis

of (^tBu₃SiNH)₃ZrMe (1-Me) in C₆D₆(eq 26), monitored by the

2. 1,2-RH-Elimination Kinetic Isotope Effects. The large

1-Me vs 1-(ND)₃/CH₃(eq 30) at 96.7 °C was k_H/k_D) z_Me

)

6.27(8), a value consistent with a relatively linear H-atom

transfer that implicated similar amounts of C-H bond-making

and N-H bond-breaking^61,62within the four-centered transition

state. Kinetic isotope effects for PhCH₃vs PhCH₂D loss from

1-CH₂Ph vs 1-(ND)₃-CH₂Ph (k_H/k_D) z_Bz) 7.1 (6), 96.8 °C),

1

loss of its NH resonances in H NMR spectra, indicated that

the reaction was first-order in 1-Me and zero-order in C₆D₆(k_Me

-

(av) ) 1.02 (5) × 10^-4s^-1, 96.7 °C, eq 28). An Eyring plot

for methane elimination (87.1-127.1 °C) from 1-Me provided

an activation enthalpy (∆H^q) 25.9 (4) kcal/mol) indicative of

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

Interscience: New York, 1984.

-d[1-Me]/dt ) k_Me[1-Me]¹[C₆D₆]⁰

(28)

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

Chemistry, Third Edition; Harper and Row: New York; 1987.

(63) (a) Westaway, K. C. Isotopes in Organic Chemistry, Volume 7:

Secondary and SolVent Isotope Effects; Buncel, E., Lee, C. C., Eds.; Elsevier,

New York, 1987; pp 288-290. (b) Isotope Effects in Chemical Reactions;

Collins, C. J., Bowman, N. S., Eds.; ACS Monograph 167; Van Nostrand

Reinhold: New York, 1970. (c) Wolfe, S.; Kim, C.-K. J. Am. Chem. Soc.

1991, 113, 8056-8061 and references therein.

significant bond-breaking, and a small, negative activation

entropy (∆S^q) -7(1) eu). The latter portrays a rather

constrained, unimolecular transition state^61-63where amide

reorganization is necessary to achieve the coplanar Zr(Me)-

N(H) geometry that permits Me-H bond formation to occur.

With 50.0 equiv of C₆D₆(2.00 M) in C₆D₁₂at 113.0 °C, k_Me

) 4.99(5) × 10^-4s^-1, in comparison to the predicted k_Meof

(64) Methane loss from 1-Me vs 1-(ND)₃CH₃is considered solely a

primary KIE; the assumption that the remaining ND positions do not exert

a significant secondary KIE is implicit.

600 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

Table 4. First-Order Rate Constants,^aFree Energies of Activation and Corresponding D(R-H) for 1,2-RH Elimination of RH from (^tBu₃Si-

NH)₃ZrR (1-R) in C₆D₆, Producing (^tBu₃SiND)₃ZrC₆D₅(1-(ND)₃C₆D₅) (Exceptions Are Noted)

∆G^q

D(R-H)

compound (solvent C₆D₆or as noted)

(^tBu₃SiNH)₃ZrCH₂Ph (1-CH₂Ph)

k (×10⁴s^-1

)

T ((0.3 °C)

(kcal/mol)

[1-R] (M)

(kcal/mol)

0.169(3)

96.7

96.8

96.7

87.1

96.6

96.7

98.7

107.6

113.0

116.9

127.1

96 .7

96.7

29.9

0.036

0.035

0.040

0.0050

0.025

0.050

0.040

0.053

0.040

0.045

0.05

0.037

0.041

0.037

0.038

0.039

0.031

0.039

0.038

0.043

0.040

0.037

88.5(15)

0.130(6)^b

(^tBu₃SiND)₃ZrCH₂Ph (1-(ND)₃-CH₂Ph)

(^tBu₃SiNH)₃ZrCH₂C₆H₃Me₂(1-Mes)

(^tBu₃SiNH)₃ZrCH₃(1-Me)

0.0184(8)^b

0.342(2)

0.387(7)^c

1.096(14)^c

1.02(10)^d

0.980(4)^d

0.968(6)^d

1.06(2)

29.4

88.5(15)

104.9(1)

28.5

(1-Me in THF-d₈; product 2-THF)^e

(1-Me in C₆D₁₂, [C₆D₆] ) 2.00 M)^f

6.27 (9)

3.05(3)^c

4.99(5)

7.00(4)^c

16.3(2)^c

1.104(7)^b

0.977(8)^g

1.02(2)^h

0.176(2)^b

0.74(2)^g

1.00(1)^h

1.42(4)

(^tBu₃SiND)₃ZrCH₃(1-(ND)₃-CH₃)

(^tBu₃SiNH)₃ZrCD₃(1-CD₃)

((d₉-^tBu)₃SiNH)₃ZrCH₃(1-d₈₁-Me)

t

(^tBu₃SiNH)₃ZrCH₂Bu (1-CH₂Bu)

28.3

101.1(4)

104.21(1)

(^tBu₃SiNH)₃ZrH (1-H)

1.51(6)

1.29(9)

1.3(1)

3.2(1)

3.21(6)

10.4(2)

13.2(4)

15.5(2)

(^tBu₃SiNH)₃ZrCH₂CHdCMe₂(1-dma)

(^tBu₃SiNH)₃ZrⁱBu (1-ⁱBu)

28.3

27.7

26.9

26.7

26.6

26.3

88.2(21)

101.1(4)

98.2(5)

(^tBu₃SiNH)₃ZrEt (1-Et)

(^tBu₃SiNH)₃ZrCy (1-Cy)

(^tBu₃SiNH)₃ZrCHdCH2 (1-CHdCH₂)

(^tBu₃SiNH)₃Zr^cPr (1-^cPr)

111.2(8)

(^tBu₃SiNH)₃ZrPh (1-Ph)

22.6(2)

111.2(8)

21.4(10)ⁱ

22.4(4)^b

4.88(5)^b

(^tBu₃SiND)₃ZrPh (1-(ND)₃-Ph)

^aDetermined from nonlinear, least-squares fitting of the differential form of the rate expression; D(RH) values are from ref 72. ^bTandem

measurement for obtaining primary isotope effect: k_H/k_D(PhCH₂H/D loss) ) 7.1(6), from nonweighted fits; k_H/k_D(MeH/D loss) ) 6.27(8); k_H/

k_D(PhH/D loss) ) 4.6(4). ^cValues used in the Eyring plot (87.1-127.1 °C) obtained from triplicate runs. From a weighted, nonlinear, least squares

fit of the data: ∆H^q) 25.9(4) kcal/mol, ∆S^q) -7(1) eu. ^dObtained from 0.05 M stock solution and dilutions. ^eCompare to MeH loss from 1-Me

in C₆D₆(98.7 °C) calcd to be 1.24 × 10^-4s^-1; k(THF)/k(C₆D₆) ) 5.1. ^f50 equiv of C₆D₆. Compare to MeH loss from 1-Me in C₆D₆(113.0 °C)

calcd to be 4.91 × 10^-14s^-1

.

^gTandem measurement for obtaining secondary isotope effect: k(CH₃)/k(CD₃) ) 1.32 (8). ^hTandem measurement for

obtaining peripheral isotope effect: k/k(d₈₁) ) 1.02(2). ⁱDetermined from monitoring the growth and subsequent loss of the NH protons of

(^tBu₃SiNH)₂(^tBu₃SiND)ZrC₆D₅(1-(ND)-C₆D₅) and (^tBu₃SiNH)(^tBu₃SiND)₂ZrC₆D₅(1-(ND)₂C₆D₅) upon thermolysis of 1-Me in C₆D₆as consecutive,

irreversible first-order processes; a k of 1.06 × 10^-4s^-1was used for MeH loss from 1-Me.

C₆D₆

_{100 °C, 7 h}8

secondary KIE observed for 1-Me vs 1-CD₃is consistent with

this depiction.

(^tBu₃SiNX)₃ZrR

X ) H, 1-R

In order for the critical atoms of the Zr-C‚‚‚NH unit (H-

R ) CH₃, CD₃, Ph, CH₂Ph

N-Zr-C dihedral angle ∼105°)⁴⁵in 1-Me to achieve copla-

X ) D, 1-(ND)₃-R

narity in the transition state, rotation about a Zr-N bond must

R ) CH₃, Ph, CH₂Ph

occur, hence ∆G^q_elim(CH₄) can be considered to consist of two

(^tBu₃SiND)₃ZrC₆D₅

X-R

X ) H

X ) D

components: amide rotation (∆G^q_rot) and hydrogen abstraction

(∆G^q_abs). Preliminary molecular mechanics calculations sug-

gested that concomitant rotation of the remaining bulky amides

exerts a steric influence on the Zr-Me, tipping it toward the

N-H. Fully deuterated d₂₇-^tBu₃SiNH₂, prepared from d₉-^tBuLi,

was used to synthesize ((d₉-^tBu)₃SiNH)₃ZrCH₃(1-d₈₁-Me) in

order to investigate the possibility of a steric isotope effect,⁶¹

due to the 81 marginally shorter C-D bonds (by ∼0.009 Å vs

C-H). No evidence of a peripheral KIE was obtained (k_H(1-

Me)/k_D(1-d₈₁-Me) ) 1.02(2), 96.7 °C), and prudence dictated

that other 1-d₈₁-R remain uninvestigated.

(30)

+

1-(ND)₃-C₆D₅

and PhH vs PhD loss from 1-Ph vs 1-(ND)₃Ph (k_H/k_D) z_Ph

)

4.6 (4), 96.7 °C), supported similar contentions, although the

latter number hinted at a significantly less symmetric transition

state for phenyl elimination. Multiple deuteration of the

products was not detected, excluding reversible elimination/

addition prior to RH loss. The R-secondary KIE for CH₄vs

3

CD₃H loss from 1-Me vs 1-CD₃at 96.7 °C was k_H/k_D) z′_Me

3. Amide Deuteration in 1-(ND)_xC₆D₅(x ) 1-3). Ther-

molysis of 1-R in C₆D₆solutions resulted in the elimination of

hydrocarbon RH and successive addition/elimination steps

leading to the formation of (^tBu₃SiND)₃ZrC₆D₅(1-(ND)₃-C₆D₅)

(eqs 31-33). Kinetics analysis of the conversion of (^tBu₃-

SiNH)₃ZrCH₃(1-Me) to 1-(ND)₃-C₆D₅was accomplished via

monitoring the growth and subsequent decay of the NH

resonances corresponding to the phenyl derivative (Figure 2).

) 1.32 (8) or z′_Me) 1.10 per D. A “normal” (i.e., k_H/k_D> 1)

effect is traditionally interpreted as indicating a change from

sp³to sp²character of the methyl group in the transition

state.^61-63Calculations by Cundari on methane elimination from

(H₂N)₃ZrMe, a model of 1-Me, portray the elimination as

occurring with retention at carbon⁴³and with minimal geometric

changes in the transition state CH/D₃fragment. The minor

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 601

Figure 2. Upon thermolysis of 1-Me in C₆D₆, the growth and

subsequent loss of the intensity (¹H NMR spectra) of NH protons

corresponding to (^tBu₃SiNH)₂(^tBu₃SiND)ZrC₆D₅(1-(ND)C₆D₅) and

(^tBu₃SiNH)(^tBu₃SiND)₂ZrC₆D₅(1-(ND)₂C₆D₅) can be modeled as

consecutive, irreversible first-order processes (eqs 31-33); a k_Meof

1.06 × 10^-4s^-1was used for MeH loss from 1-Me.

Figure 3.

(^tBu₃SiNH)₂ZrdNSi^tBu₃(2) to (^tBu₃SiNH)₂ZrNHSi^tBu₂CMe₂CH₂

(3) proved competitive with C-H bond activation.

For some critical cases, these difficulties were circumvented

through competition experiments designed to indirectly provide

ground state information. Imido (^tBu₃SiNH)₂ZrdNSi^tBu₃(2)

was generated in the presence of methane (4 atm) and benzene

(∼2 equiv) via irreversible loss of cyclohexane from 1-Cy in

C₆D₁₂(eqs 34 and 35) at 97 °C. At early conversion (∼17%),

k_Me

(^tBu₃SiNH)₃ZrCH₃

8

C₆D₆

1-CH₃

(^tBu₃SiNH)₂(^tBu₃SiND)ZrC₆D₅

+ CH₄(31)

1-(ND)-C₆D₅

k_Cy

(^tBu₃SiNH)₃ZrCy

( Bu₃SiNH)₂ZrdNSi Bu₃

t

8

+ CyH

(34)

2k_Ph

^/38

∼97 °C

(^tBu₃SiNH)₃(^tBu₃SiND)ZrC₆D₅

1-Cy

2

C₆D₆

1-(ND)-C₆D₅

(^tBu₃SiNH)(^tBu₃SiND)₂ZrC₆D₅

+ C₆D₅H (32)

k_Mevs k_Ph

t

1-(ND)₂-C₆D₅

2 + CH₄+ C₆H₆

8

+

( Bu₃SiNH)₃ZrMe

C₆D₁₂, ∼97 °C

1-Me

k_Ph/3

(^tBu₃SiNH)(^tBu₃SiND)₂ZrC₆D₅

(^tBu₃SiNH)₃ZrPh

8

(35)

(36)

(37)

C₆D₆

1-(ND)₂-C₆D₅

1-Ph

(^tBu₃SiND)₃ZrC₆D₅

+ C₆D₅H (33)

k_Arvs k_Bz

t

2 + C₇H₈

8

+

( Bu₃SiNH)₃ZrC₆H₄Me

1-(ND)₃-C₆D₅

C₇H₈, ∼97 °C

1-Ar

(^tBu₃SiNH)₃ZrCH₂Ph

Using a k_Meof 1.06 × 10^-4s^-1(eq 31) and modeling the

deuteration as two subsequent first-order processes⁶⁵that reflect

the statistics of each event (eq 32, 2k_Ph/3; eq 33, k_Ph/3), a rate

constant of 2.14 × 10^-3s^-1for C₆D₅H elimination was obtained,

consistent with the rate constant of 2.25(3) × 10^-3s^-1generated

from monitoring the loss of C₆H₆from (^tBu₃SiNH)₃ZrPh (1-

Ph). The successful modeling of the deuteration supports the

standard mechanism of Scheme 1.

1-CH₂Ph

k_cPrvs k_Ph

(^tBu SiNH) Zr^cPr

2 + c-C₃H₆+ C₆H₆

_{, ∼97 °C}8

+

3

C₆D₁₂

1-^cPr

(^tBu₃SiNH)₃ZrPh

1-Ph

4. Equilibria and C-H Bond Activation Selectivities.

Inspection of the first-order rate constants for the elimination

of hydrocarbons in Table 4, obtained at 96.7 °C, reveals a

moderate distribution of reaction rates from (^tBu₃SiNH)₃ZrCH₂-

Ph (1-CH₂Ph, k_Bz) 1.69(3) × 10^-5s^-1) to (^tBu₃SiNH)₃ZrPh

(1-Ph, k_Ph) 2.25(3) × 10^-3s^-1). Of critical importance are

the origins of these differences; do they reflect mostly disparate

ground state energies, or do the deviations represent transition

state energetics? Attempts to obtain crucial ground state

information via equilibrium studies were thwarted by unantici-

pated difficulties. For example, thermolysis of 1-Ph in the

presence of MeH in C₆D₁₂resulted in the precipitation of copius

amounts of [(^tBu₃SiNH)₃Zr]₂(µ₂:η¹,η¹-1,4-C₆H₄) (1₂-C₆H₄, eq

13). In other instances, cyclometalation of putative intermediate

where reversibility of R/R′H addition was considered minimal,

the competition for H-Me vs H-Ph bond activation by the

imide revealed that benzene was the favored substrate by ∆∆G^q

) -3.4 kcal/mol. Inclusion of this energy differential on a

standard free energy vs reaction profile diagram enabled

calculation of the free energy difference between 1-Me + C₆H₆,

and 1-Ph + CH₄, which is favored by -1.2 kcal/mol, according

to Figure 3.

This minor ground state free energy disparity is matched in

related experiments involving activation of the aryl (para and

meta) vs benzylic positions of toluene (eqs 34 and 36).

Competition for 2 at 20% conversion revealed that only ∼1%

1-CH₂Ph was observed relative to aryl-activated products,

1-C₆H₄Me, indicating a preference for aryl activation by ∆∆G^q

e -3.4 kcal/mol. The products were then thermally equili-

brated (∼97 °C), and a ground state energy difference of only

(65) Benson, S. D. The Foundations of Chemical Kinetics; McGraw-

Hill: New York, 1960.

602 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

values in parentheses are estimated): ^cPrH [0.0] ≈ ArH [0.0]

> MeH [3.4] > PhCH₂H [4.0] > cyclometalation (>8.5) >

t

EtH (>8.9) > BuCH₂H (>9.3) > CyH (>11.2). These

measurements represent a rare quantitative assessment of

selectivities for C-H bond activation,^{3,9,11,25,34,70}but care must

be taken not to place too much emphasis on the relative positions

of the estimated substrates. Figure 6 illustrates relative transition

state energiessand therefore the ∆∆G^q’s corresponding to CH

bond activation selectivitysfor the appropriate 1,2-RH-elimina-

tions from 1-R, which were calculated by taking the standard

free energy of 1-^cPr as reference (0.0 kcal/mol).

5. Thermochemistry and Relative D(Zr-R)′s. The gen-

eral features of relative elimination rates (i.e., ∆H^q_elim(R) -

∆H^q_elim(R′)) have previously been espoused.²⁴Entropic

factorssadmittedly difficult to assesssare ignored, specifically

because it is likely that significant entropic contributions from

1-R + R′H vs 1-R′ + RH are likely to cancel.⁷¹In addition,

factors such as solvation energies, which can be important, will

also be considered to essentially cancel, thus our initial analysis

will focus on critical bond enthalpies. With these consider-

ations, it can be shown that [∆H^q_elim(R) - ∆H^q_elim(R′)] -

[∆H^q_addn(R) - ∆H^q_addn(R′)] equals ∆H_rxn, which can be defined

in terms of the respective metal-carbon and carbon-hydrogen

bond energies⁷²as shown in eq 38. The ∆H_rxncorresponding

to Figures 3-5 can be estimated from each diagram, thus

Figure 4.

∆H_rxn≈ [D(R-H) - D(R-H)] + [D(M-R) - D(M-R′)]

(38)

comparisons of solution phase metal alkyl bond strengths with

(66) Even when fairly low concentrations of MeH (∼0.025 M) are used

in eq 7, no cyclometalation is evident. Taking the selectivity of (^tBu₃-

SiNH)₂ZrdNSi^tBu₃(2, generated from 1-Cy) for MeH vs cyclometalation

to 3 to be at least 25 (an easily measurable quantity by ¹H NMR), then

r(CH₄)/r(cyclomet) ∼ (k_MeH[2][CH₄]/k_cymet[2]) ∼ (k_MeH[2][0.025]/k_cymet[2])

> 25, and k_MeH/k_cymet> 1000 (1 M standard states). Since k_MeH/k_cymet

)

exp[(∆G^q_cymet- ∆G^q_MeH)/RT] > 1000 M, then ∆G^q_cymet- ∆G^q_MeH> 5.1

Figure 5.

kcal/mol at 370 K.

(67) In each estimate where cyclometalation is observed, other degrada-

tion products rendered NMR spectral measurements difficult. Taking the

selectivity of (^tBu₃SiNH)₂ZrdNSi^tBu₃(2, generated from 1-Cy) for

cyclometalation to 3 vs EtH activation to be at least 5 (assuming a <5:1

∼0.4 kcal/mol favoring 1-C₆H₄Me over 1-CH₂Ph was obtained.

Once the activation energies for elimination of the respective

groups were applied (∆G^qfor elimination of toluene from

1

Ar

3:1-Et ratio could be detected by H NMR spectroscopy; difficulties with

byproduct formation prevented a more definitive estimate by ¹H NMR),

then r(cyclomet)/r(EtH) ∼ (k_cymet[2]/k_EtH[2][C₂H₆]) ∼ (k_cymet[2]/k_EtH[2]‚

1-C₆H₄Me was assumed to be equal to ∆G^q_Phfor 1-Ph), ∆∆G^q

was determined by difference to be ∼ -4.0 kcal/mol (Figure

4), consistent with the results of the competition.

[0.36]) > 5, and k_cymet/k_EtH> 1.8 (1 M standard states). Since k_cymet/k_EtH

)

exp[(∆G^q- ∆G^q_cymet)/RT] > 1.8, then ∆G^q- ∆G^q

> 0.4 kcal/

EtH

cymet

Likewise, competition for (^tBu₃SiNH)₂ZrdNSi^tBu₃(2) be-

tween cyclopropane and benzene revealed little selectivity (97

°C, ∆∆G^q∼0.0 kcal/mol) at early conversion (<15%). Equili-

bration afforded a ground state favoring 1-^cPr + C₆H₆over 1-Ph

+ c-C₃H₆by -0.3 kcal/mol as depicted in Figure 5, a result

fully consistent with the competition.

mol at 370 K. This analysis relies on the fact that 1-Et eliminates about

three times slower than 1-Cy and could be reasonably observed under

reaction conditions.

(68) Taking the selectivity of (^tBu₃SiNH)₂ZrdNSi^tBu₃(2, generated from

t

1-Cy) for cyclometalation to 3 vs BuCH₂-H activation to be at least 25

(assuming a <25:1 3:1-CH₂Bu ratio could be detected by ¹H NMR

t

spectroscopy), then r(cyclomet/r(Me₄C) ∼ (k_cymet[2]/k_NpH[2][Me₄C]) ∼

(k_cymet[2]/k_NpH[2][0.12]) > 25, and k_cymet/k_NpH> 3 (1 M standard states).

Since k_cymet/k_NpH) exp[(∆G^q

- ∆G^q_cymet)/RT] > 3, then ∆G^q

-

NpH

The data may be combined to infer that MeH activation is

favored by -0.6 kcal/mol over PhCH₂H addition or that

cyclopropane activation is favored by -3.4 kcal/mol over that

of methane and by -4.0 kcal/mol over benzylic activation, etc.

Rough estimates⁶⁶based on substrate competitions suggest that

cyclometalation of (^tBu₃SiNH)₂ZrdNSi^tBu₃(2) to (^tBu₃-

NpH

∆G^q

> 0.8 kcal/mol at 370 K. This analysis relies on the fact that

cymet

t

1-CH₂Bu eliminates about seven times slower than 1-Cy and could be

reasonably observed under reaction conditions.

(69) When 1-Me is thermolyzed in cyclohexane-d₁₂, no deuteration of

the amide positions of any products is observed, hence there is no evidence

of C₆D₁₂activation. Assuming the density of C₆D₁₂(0.89 g/mL) is relatively

insensitive to temperature, neat cyclohexane is ∼9.3 M. Assuming a k_H/k_D

for cyclohexane activation to be ∼5, and that >5% deuteration could be

detected by ¹H NMR spectroscopy, the selectivity of (^tBu₃SiNH)₂ZrdN-

Si^tBu₃(2, generated from 1-Me) for cyclometalation to 3 vs CyH activation

would be at least 4; then r(cyclomet)/r(CyH) ∼ (k_cymet[2]/k_CyH[2][C₆H₁₂])

∼ (k_cymet[2]/k_CyH[2][9.3]) > 4, and k_cymet/k_CyH> 37 (1 M standard states).

SiNH)₂ZrNH-Si^tBu₂CMe₂CH₂(3) is disfavored by >5.1 kcal/

mol relative to MeH activation. Likewise, various concentra-

tions of ethane, neopentane, and cyclohexane have been shown

to be uncompetitive substrates when compared with cyclo-

metalation in cyclohexane, hence intramolecular C-H activation

to form 3 is <-0.4, <-0.8, and <-2.7 kcal/mol more favorable

than Et-H,^{67 t}BuCH₂-H,⁶⁸and Cy-H addition.⁶⁹

Since k_cymet/k_CyH) exp[(∆G^q

- ∆G^q_cymet)/RT] > 37, then ∆G^q

-

CyH

∆G^q

> 2.7 kcal/mol at 370 K.

cymet

(70) Periana, R. A.; Bergman, R. G. Organometallics 1984, 3, 508-

510.

(71) (a) Jones, W. D.; Feher, F. J. J. Am. Chem. Soc. 1984, 106, 1650-

1664. (b) Jones, W. D.; Feher, F. J. Acc. Chem. Res. 1989, 22, 91-100.

(72) Berkowitz, J.; Ellison, G. B.; Gutman, D. J. Phys. Chem. 1994, 98,

2744-2765.

Calculation of a rough, relative C-H bond activation

c

selectivity scale, with Pr-H addition taken as the reference,

follows (∆∆G^qin kcal/mol, values in brackets are measured,

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 603

those of gas phase carbon-hydrogen bonds (Table IV) can be

made. Implicit in the discussion is the assumption that

enthalpies and entropies of sublimation associated with the

various metal-alkyl complexes (1-R) can be considered ap-

proximately equal.⁷¹

From Figure 3, ∆H_rxn≈ -0.9 kcal/mol (∼ -0.3 kcal/mol

are attributed to the statistics favoring benzene (6H) vs MeH

(4H) capture) and [D(Ph-H) - D(Me-H)] ≈ 6.3 kcal/mol, thus

[D(Zr-Me) - D(Zr-Ph)] ≈ -7.2 kcal/mol. In essence, the

origin of the minimal ground state difference between 1-Me +

C₆H₆and 1-Ph + MeH stems from the similar differences in

Zr-C^73-76vs C-H bond enthalpies (eq 39).⁷²The diagram

∆H_rxn≈ [D(Zr-Me) - D(Zr-Ph)] + [D(Ph-H) -

D(Me-H)] (39)

(Figure 4) reflecting benzyl vs aryl elimination possesses related

features according to eq 40: ∆H_rxn≈ -0.4 kcal/mol (recall that

the elimination rate for 1-Ph was used), [D(Ar-H) - D(PhCH₂-

H)] ≈ 22.7 kcal/mol, [D(Zr-CH₂Ph) - D(Zr-Ar)] ≈ -22.3

kcal/mol. Toluene is absent in the equilibrium, hence this

analysis assumes that the aryl and benzylic C-H bonds do not

∆H_rxn≈ [D(Zr-CH₂Ph) - D(Zr-C₆H₄Me)] +

[D(Ar-H) - D(ArCH₂-H)] (40)

differ when (^tBu₃SiNH)₃Zr is an aryl substituent. A similar

analysis corresponding to Figure 5 suggests that [D(Zr-Ph) -

D(Zr-^cPr)] ≈ 4.9 kcal/mol, simplified in eq 41. Comparable

free energy pictures are obtained for all of the substrate

competitions (i.e., minor ground state differences that ultimately

Figure 6. Relative ground state (∆G° where ∆G° (1-^cPr) ) 0.0 kcal/

mol) and transition state (∆G°_TS) ∆G° + ∆G^q(1,2-RH-elim)) energies

for 1-R obtained from kinetics studies (colorless), equilibrium (black)

measurements, and estimated ∆G° data (lined) derived from competitive

activations. The light and darked shaded background represents possible

and most likely energetic positions of any intermediate(s). Arrows

indicate that the energies are estimated minima.

∆H_rxn∼ [D(Zr-Ph) - D(Zr-^cPr)] + [D(^cPr-H) -

D(Ph-H)] (41)

relate to large differences in D(Zr-R) - D(Zr-R′)), thus the

assumption that entropic factors and heats of solvation are either

minimal, or essentially cancel in comparison, gains some

credence for these specific cases. The data roughly suggest that

the differences between metal-carbon,^73-76and between cor-

responding carbon-hydrogen bond strengths,⁷²are essentially

the same for the substrates amenable to the equilibrium study;

related conclusions were determined for the previously inves-

tigated tantalum system.²⁴

6. The Nature of the Transition State for 1,2-RH-

Elimination. Figure 7 illustrates a plot of the free energies of

activation for 1,2-RH-elimination (96.7 °C) from 1-R vs D(R-

H), the bond dissociation enthalpies of the corresponding

hydrocarbons.⁷²According to the rudiments of the Hammond

postulate, a basic assessment of the rates is proferred; the

transition state of 1,2-RH-elimination is proposed to reflect

varying degrees of reactant and product character.⁷⁸Choosing

the ordinate scale as D(R-H) leads to two potential interpreta-

tions: (1) a negative slope implicates a correlation with the

strength of the C-H bond formed in the 1,2-RH-elimination

Figure 7. ∆G^q(1,2-RH-elim) for 1-R vs bond dissociation enthalpy,

D(R-H); D(M-R) is predicted to be proportional to the latter. A

negative slope implicates a late transition state (∆G^qdecreases as

D(R-H) increases) and a positive slope is indicative of an early

transition state (∆G^qincreases as D(M-R) increases).

D(Zr-R); assuming D(Zr-R) D(R-H))^72-76san early transi-

tion state. Note that the ∆∆G^q_elimdata span a moderate energy

range (3.6 kcal/mol, ∼13% of ∆G^q) 27.9 kcal/mol), while

av

the scope of bond enthalpies for the hydrocarbons in common

event (i.e., ∆G^q

D(R-H))sa late transition state; (2) a

is substantially greater (23 kcal/mol, ∼23% of D(R-H)_av

)

positive slope indicates a correlation with the strength of Zr-C

bond broken in the 1,2-RH elimination process (i.e., ∆G^q

(77) The parabolic representations of the free energy surface arise from

the usual ill-defined reaction coordinate that comprises Zr-R and N-H

bond-breaking along with C-H and ZrN(π) bond-making and the geometric

changes that accompany these events. In this model, the curvature of all

1-R(R′) are considered equivalent (i.e., same dependence on x). While this

is not rigorously true, only very minor deviations are expected in this energy

region (i.e., the lower portion of each parabola is not consequential). The

model also assumes that the reactant to product transition occurs adiabati-

cally, with a facility that is independent of R. For the origin of such models,

see: Thornton, E. R. J. Am. Chem. Soc. 1967, 89, 2915-2927. See, also:

Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334-338.

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

7715. (b) 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.

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

J. E. J. Am. Chem. Soc. 1987, 109, 1444-1456. (b) Bryndza, H. E.; Bercaw,

J. E. Polyhedron 1988, 7, 1441-1452.

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

(76) For an alternative viewpoint, see: Drago, R. S.; Wong, N. M.; Ferris,

D. C. J. Am. Chem. Soc. 1992, 114, 91-98.

(78) Cundari, T. R.; Matsunaga, N., submitted for publication.

604 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

respective ground states are of little consequence to the disparate

elimination rates. Because of the minimal relative ground state

differences of group representatives 1-CH₂Ph (k_Bz) 1.69(3) ×

10^-5s^-1), 1-Me (k_Me) 1.06(2) × 10^-4s^-1), 1-^cPr (k_c-Pr

)

1.55(2) × 10^-3s^-1) and 1-Ph (k_Ph) 2.26(2) × 10^-3s^-1), these

1-R manifest disparate 1,2-RH-elimination rates primarily as a

consequence of differing positions along the reaction coordinate.

For these complexes, use of transition state energies (e.g., define

∆G^TS) ∆G^q+ ∆G°) naturally leads to the same conclusionsthe

later or more product-like the transition state, the faster the 1,2-

RH-elimination rate. Note that the lowest primary KIE among

the three measured (1-(ND)₃R, R ) Me, CH₂Ph, Ph) is for 1-Ph

(z_Ph) 4.6(4)), where the transition state energy is likely to be

near that of the intermediate state, consistent with a less

symmetric N‚‚‚H‚‚‚Ph conformation that implicates a later

position along the reaction coordinate.

Where ground state energies are significantly dissimilar,

respective transition state energies are additionally effected. Note

that ∆G^q₁< ∆G^q₃despite their identical ground state positions

(x₁) x₃) and despite the lower energy of ∆G^TSrelative to

3

∆G^TS₁. Furthermore, ∆G^TS₁occurs earlier than ∆G^TS₃(x₁(TS)

< x₃(TS)), even though both pathways are generally “late”. The

ground state factor is particularly applicable to the alkyl and

hydride group, whose 1-R roughly correlates with the strength

of the zirconium-carbon (-hydride) bond being broken,

provided D(R-H) is an index of D(R-M). Figure 8 reveals

why 1-Me + EtH (e.g., ∆G°₃) has a slower methane elimination

rate than 1,2-EtH-elimination from 1-Et + MeH (e.g., ∆G°₁),

despite a transition state energy that is lower by <-5.8 kcal/

mol. Since the origins of the reaction coordinate will be similar,

it is the ground state energy of 1-Et + MeH, which is higher

by >6.6 kcal/mol, that translates into an earlier, yet swifter

alkane elimination.

Figure 8. ∆G^q(1,2-RH-elim) for 1-R values may be viewed as arising

from disparate ground state energies (e.g., ∆G°_1,2vs ∆G°₃), and

differing positions (x_1,3vs x₂) along the reaction coordinate.

100.4 kcal/mol). This disparity may portend transition states

for 1,2-RH-elimination that possess a balanced character,

consistent with the kinetic isotope effect data that portray the

hydrogen/deuterium transfer as symmetric.

The data are roughly grouped into three regions. The first

contains the benzylic hydrocarbyls, 1-CH₂Ph and 1-Mes, and

the η¹-dimethylallyl derivative, 1-dma, whose corresponding

hydrocarbons possess similar bond dissociation energies. Data

within a second group containing the alkyls, 1-R (R ) Cy, ⁱBu,

Et, ^tBuCH₂, Me), and the hydride, 1-H, are somewhat scattered,

but a very rough positive correlation with D(R-H) can be

inferred. Species containing Zr-C(sp²) bonds, 1-R (R )

Relative ground and transition state energies for selected 1-R

(Figure 6) show that the former must necessarily be coupled

with positional changes in the reaction coordinate to adequately

understand the correlation between ∆G^q_elimand D(R-H) (Figure

t

7). Ground state energetics pertaining to 1-R (R ) Et, CH₂Bu,

c

CHdCH₂, Ph, Pr) comprise the third group of hydrocarbyls,

Cy) suggest that greater reactant character is infused in the

transition states for alkane (and possibly H₂) elimination, hence

the positive correlation with D(Zr-C) (i.e., D(C-H)) for this

subgroup, even though the process is still generally “late”.

Further inspection of Figure 6 reveals that the intermediate state

(2 + RH’s) must be >11.3 kcal/mol (∆G° for 1-Cy relative to

1-^cPr at 0.0 kcal/mol) and <26.6 kcal/mol (∆G^TSof 1-^cPr). From

the energetics and the parabolic Hammond analysis, the

character of the transition state with respect to Zr-C bond-

breaking and C-H bond-making is likely to be balanced,

although possessing more of the latter. Ab initio calculations

(GAMESS) on the transition states of elimination from several

models of 1-R, (H₂N)₃Zr-R (1′-R), corroborate this postula-

tion.^78,79

assuming this description of the cyclopropyl derivative is apt.⁶²

Figure 8 provides a textbook^62,77parabolic view of 1-R +

R′H and an intermediate (elimination product) state of (^tBu₃-

SiNH)₂ZrdNSi^tBu₃(2) + RH + R′H, enabling interpretation

of the energetics in Figure 6 and Figure 7. Given the caveat of

a common intermediate, the 1,2-RH-elimination data can be

roughly explained with two variables: the disposition of the

1-R + R′H (x_n) relative to 2 + RH + R′H (x_p) along the reaction

coordinate and the relative standard free energies of 1-R + R′H

(∆G°_n). The reaction coordinate is mostly comprised of Zr-C

bond-breaking, C-H bond-making, ZrN(π) bond-making, and

N-H bond-breaking. Figure 8 shows that late transition states

are a general property of endothermic reactions comprised of

similar parabolic surfaces for the reactant and product states,

which are likely when D(Zr-R) approaches that of D(H-R).

It is expected from the work of Schock and Marks et al. that

the zirconium-carbon bonds are strong and that D(Zr-C) are

in a regime where they correlate linearly with related hydro-

carbon BDEs (e.g., D(Zr-R) ∼ RD(R-H) + â).⁷³

The application of a Hammond analysis^62,77to this systemsand

to the previously reported tantalum series,²⁴which can be

interpreted in similar fashionsis satisfying because conventional

logic may be used to assemble a self-consistent depiction

encompassing a variety of substrates. Furthermore, the nature

of the 1,2-RH-elimination reaction is reminiscent of organic

H-CH₂CH₂-X eliminations, where similar interpretive methods

have proven useful. Given this analysis, it is pertinent to ask

why a consistent picture can be obtained, i.e., why the necessary

assumptions work, whether there are alternative explanations,

For 1-R + R′H with similar ground state energies, their

respective ∆G^q

respond to a shift along the reaction

elim

coordinate. Note that ∆G^q₁< ∆G^q₂as a consequence of x₁>

x₂. As a corollary not easily grasped from the figure, {x₁(TS)

- x₁}/{x_p- x₁} > {x₂(TS) - x₂}/{x_p- x₂}, i.e., x₁(TS) is further

along than x₂(TS) on their respective reaction coordinates.

Consequently, path 1 is later than path 2, thereby revealing the

origin of the swifter rate, a situation applicable to 1-Ph (or 1-Ar,

Figure 4) + C₇H₈vs 1-CH₂Ph + PhH (or ArH), where the

t

and specifically why the ground states of 1-Et, 1-CH₂Bu, and

1-Cy are relatively high.

(79) (a) Cundari, T. R.; Gordon, M. S. J. Am. Chem. Soc. 1993, 115,

4210-4217. (b) Cundari, T. R. Organometallics 1993, 12, 4971-4978.

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 605

If similar activation entropies for each 1,2-RH-elimination

can be inferred from the data,⁸⁰entropic contributions from

rotation and alignment of the Bu₃SiNH ligand are probably

of R-Me groups to CH₄leads to an inductive acid-weakening

effect that is compensated by an acid-strengthening polarization

effect.⁸³Methane is therefore a stronger acid than ethane, which

is approximately the same as propane (2-position), while

isobutane (2-position) is stronger than MeH. Solution acidity

data show that increasing substitution at carbon results in a

weakening of acidity. If the acidity data can be considered to

reflect the ionic character of the bond, the data can be interpreted

to mean that increasing substitution at carbon decreases the

ionicity of a bond to that carbon. Electronegativity arguments

support the contention that zirconium-carbon bonds possess a

substantially stronger ionic component than their hydrocarbon

congeners, and thus may be destabilized to a greater degree by

inductive effects relative to RH. An inductive destabilization

t

dominant, since variation of R apparently has minor impact.

Brown and Caffery’s molecular mechanics analysis of the 1,2-

RH-elimination events revealed little to moderate steric influence

from variation of R.^81,82As a corollary, entropic contributions

by R to 1-R and its transition state for 1,2-RH-elimination are

effectively similar. Attempts to interdependently correlate

D(C-H) (R D(Zr-C)) and appropriate steric factors to the

∆G^q_elimdata failed.⁸²Calculational treatments suggest that the

eliminations can be assessed as a replacement of Zr by H at a

carbon center that undergoes minimal distortion, hence the

enthalpic correlations assessing bond-making and -breaking

characteristics were anticipated.^43,78,79Nonetheless, it is cer-

tainly plausible that some of the scatter in the data results from

t

of 1-R (R ) Et, CH₂Bu, Cy) relative to 1-Me would explain

their higher ground states and corresponding greater influence

on the respective transition state species. The drawback to this

argument concerns its specificity; substantial acidity differences

subtle deviations in ∆S^q

.

elim

An earlier, alternative explanation for the 1,2-RH-elimination

vs D(R-H) data (Figure 7) ascribed the relative speed of

elimination from the sp²substrates (1-Ph, 1-^cPr, and 1-CHdCH₂)

to a transition state stabilization²²due to adjacent p-orbital

participation in the H-C bond-forming process. Calculational

support for p-orbital participation has not been uncovered

(except for R ) CtCR, Vide infra),⁷⁸and the possibility of

special ground state stabilization for 1-CH₂Ph etc., is not

supported by experiment. The ¹J_CHof 119 Hz for the Zr-CH₂-

group of the benzyl complex, 1-CH₂Ph, is typical for such

fragments, and not particularly indicative of agostic⁴⁶or allylic

bonding. Low temperature ¹H NMR studies revealed no

peculiar benzyl conformation or agostic effect that would

prevent ready access to the geometry required for 1,2-elimina-

tion, but subtle energetic effects (∼ <5 kcal/mol) would not be

evident in these experiments, hence some ambiguity persists.

The most difficult aspect of the 1,2-RH-elimination rates to

assess concerns the rather dramatic jump in relative ground state

energies for the alkyl derivatives, especially the large difference

between 1-Me and 1-Et. Steric arguments, augmented by the

molecular mechanics study,⁸¹rationalize the destabilization of

c

among the more acidic RH (R ) Pr, Ph, CH₂Ph) substrates

must not generate equivalent inductive influences, perhaps due

to a leveling effect, i.e., the Zr-C bond may have a maximum

ionicity.

Existence of the Three-Coordinate Intermediate, (^tBu₃Si-

NH)₂ZrdNSi^tBu₃(2), and Alkane Binding. 1. Proton

Affinity and 1,2-RH-Elimination. The previous discussion of

1,2-RH-elimination rates assumed that a single intermediate,

(^tBu₃SiNH)₂ZrdNSi^tBu₃(2), mediated the process and its

microcopic reverse, 1,2-RH-addition. Since arene⁸⁶and, more

recently, alkane complexes have been implicated as intermedi-

ates in the activation of C-H bonds by late metal d⁸

complexes,^31-35and as transients indirectly observed in reductive

elimination reactions,^4,36it is prudent to assess the viability of

such species in this d⁰metal imido system.

Consider the binding of an alkane or arene to an electropos-

itive, d⁰zirconium center to be analogous to protonation of RH,

whose enthalpy change is -PA, where PA is the proton affinity

of the hydrocarbon.^87,88Figure 9 illustrates a plot of ∆G^qfor

1,2-RH-elimination versus the proton affinity of the correspond-

ing RH. When available, appropriate experimental PA values

were utilized, but several were obtained from semiempirical

AM1 calculations. In specific cases such as benzylic activation,

calculations were necessary because experimental results reflect

protonation of the arene ring, mimicking an unproductive

reaction coordinate. In such instances, protonation was effected

at the appropriate C-H bond, the structure minimized, and the

proton affinity calculated. The hydride, 1-H, was not considered

because of the obvious orbital differences between H and an R

fragment; its PA of 104.1(11) kcal/mol places it conspicuously

off the curve.

t

1-Cy and 1-CH₂Bu relative to 1-Me, but differentiation of

methyl from ethyl on this basis is untenable in view of the

stability of much bigger substituents such as benzyl, mesityl,

and phenyl.

Attempts to correlate the 1,2-RH-elimination rates and

transition state energies to gas phase^83,84and solution acidities⁸⁵

met with little success but suggested a possible explanation for

the ground state anomalies. For gas phase acidities, the addition

(80) 1,2-XH-elimination activation entropies are -10(3) eu in three

distinct systems: see refs 23, 25 and this work.

(81) Caffery, M. L.; Brown, T. L., personal communication. Employing

a fit to ln k(1,2-RH-elim) ) aE_R′ + bD(C-H) + c, where E_R′ is a ligand

repulsive energy, a correlation coefficient of 0.838 was found for R ) 0.036,

â ) 0.18, and γ ) -27.50. The results were consistent with a late transition

state that manifests some steric acceleration of 1,2-RH-elimination. Mechan-

ics calculations were performed using BIOGRAF, through an academic

collaborators arrangement with Molecular Simulations, Inc. For a related

study, see: Brown, T. L. Inorg. Chem. 1992, 31, 1286-1294.

(82) Attempts to establish a Thornton-More O’Farrell-Jencks type

diagram using D(RH) (R D(Zr-C)) and PA (proton affinity) data as

interdependent indicators of orthogonal reaction coordinates failed to

converge to a solution where both were correlated with the ∆G^qvalues of

1-R. In retrospect, these two data sets are probably not independent--they

reflect similar features of each RH. Other efforts coupling D(RH) (R D(Zr-

C)) and R steric factors (effective A-values) generated by Brown and

Cafferty (ref 81), or PA and A-values also failed to elicit an interdependent

correlation with the ∆G^q_elim(1-R) vlaues, possibly because the steric factors

are of minor consequence in these reactions. (a) More O’Ferrall, R. A. J.

Chem. Soc. B. 1970, 274-277. (b) Jencks, W. P. Chem. ReV. 1972, 72,

705-718.

A relatively smooth correlation is apparent; the greater the

proton affinity of RH, the faster the 1,2-RH-elimination from

(85) Current values were obtained from Andrew Streitweiser, Jr.,

Department of Chemistry, University of California, Berkeley, Berkeley,

California, 94720. See, also: Streitweiser, A., Jr.; Juaristi, E.; Nebenzahl,

L. L. In ComprehensiVe Carbanion Chemistry, Part A; Buncel, E., Durst,

T., Eds.; Elsevier: Amsterdam, 1980.

(86) For an interesting example, see: Jones, W. D.; Feher, F. J. J. Am.

Chem. Soc. 1986, 108, 4814-4819.

(87) (a) McMahon, T. B.; Kebarle, P. J. Am. Chem. Soc. 1985, 107,

2612-2617. (b) Collyer, S. M.; McMahon, T. B. J. Phys. Chem. 1983, 87,

909-912. (c) Rosenstock, H. M.; Buff, R.; Ferreira, M. A. A.; Lias, S. G.;

Parr, A. C.; Stockbauer, R. L.; Holmes, J. L. J. Am. Chem. Soc. 1982, 104,

2337-2345. (d) Cotter, R. J.; Rozett, R. W.; Koski, W. S. J. Chem. Phys.

1972, 57, 4100-4102.(e) Moylan, C. R.; Brauman, J. I. Ann. ReV. Phys.

Chem. 1983, 34, 187-215. (f) Devlin, J. L.; Wolf, J. F.; Taft, R. W.; Hehre,

W. J. J. Am. Chem. Soc. 1976, 98, 1990-1992. (g) Chong, S.-L.; Franklin,

J. L. J. Am. Chem. Soc. 1972, 94, 6347-6351. (h) Walder, R.; Franklin, J.

L. Int. J. M. S. Ion. Phys., 1980, 36, 87.

(83) DePuy, C. H.; Gronert, S.; Barlow, S. E.; Bierbaun, V. M.;

Damrauer, R. J. Am. Chem. Soc. 1989, 111, 1968-1973.

(84) Current values have been compiled by John E. Bartmess, Department

of Chemistry, University of Tennessee, Knoxville, TN 37996.

(88) Lee, C. C.; Hass, E. C.; Obafemi, C. A.; Mezey, P. G. J. Comput.

Chem. 1984, 5, 190-196.

606 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

Figure 9. ∆G^q(1,2-RH-elim) for 1-R vs the proton affinity, PA, from

experiment (b) or from AM1 calculations (O).

Figure 11. Standard free energy diagram illustrating a 1,2-RH-

elimination/addition process mediated by RH-binding transients, 2-RH

and 2-R′H, and dissociative (I) or associative (II) pathways between

them.

Alternatively, RH-bound intermediates (2-RH, B) could play

a role as indicated in Figure 10. The parabolas for 2-RH are

arbitrarily sketched to cross the surface ascribed to (^tBu₃SiNH)₂-

ZrdNSi^tBu₃(2) such that ∆G°_pis no longer at a minimum,

thereby showing that 2 need not be a viable intermediate in

this system; hydrocarbon exchanges could occur between 2-RH

and 2-R′H via dissociative or associative interchange pathways.⁸⁹

For intermediates B, it is clear that stronger binding of RH

results in a lower transition state for elimination (i.e., ∆G^TS

b′

< ∆G^TS_b) that can be considered independent from ground state

or positional influences of 1-R.

Figure 10. The proton affinity correlation may be viewed as reflecting

a purely positional component (x₁vs x₂, A), or indicating complexation

of RH (i.e., 2-RH, B). To highlight the latter possibility, note that 2 is

no longer accorded a minimum in the free energy surface.

In either situation, transition state stabilization (A) or

intermediate state (B) alternative, one common theme derives

from the ∆G^q

vs PA correlation: the more tightly RH is

elim

bound in the transition state, the nearer ∆G^TSis to the reactant

and product states. It is noteworthy that the swiftest 1-R

eliminations are observed for sp²-hybridized hydrocarbyls,

whose Zr-C bond lengths are shorter than corresponding sp³-

hybridized species, and are thereby positionally later in their

respective reaction coordinates. Furthermore, the straightfor-

ward correlation depicted in Figure 9 illustrates that proton

affinity can be used to predict the 1,2-RH-elimination rate of a

new 1-R but cannot assess the true “lateness” of its transition

state because ground state contributions from 1-R cannot be

discerned.

1-R. While the intrinsically enthalpic correlation substantiates

the proposed minimal importance of entropic effects, the data

are also consistent with the previous contention of a generally

late transition state. Attempts to interdependently correlate PA

and appropriate steric factors to the ∆G^q

data failed.⁸²

elim

Consider the parabolic depiction in Figure 10, where the reaction

coordinate again corresponds to the 1,2-elimination of RH to

putative (^tBu₃SiNH)₂ZrdNSi^tBu₃(2), affording (^tBu₃SiNH)₂-

(^tBu₃SiN))Zr(RH) as a transition state (A, 2-RH^q). The

correlation of ∆G^q_elimwith PA implicates a stabilization of the

transition state A. Figure 8 revealed that stabilization of ∆G^TS

could occur via a later reaction coordinate for 1,2-RH-

elimination (e.g., earlier for 1,2-RH-addition) and through

lowering the ground state energy of 1-R; the former results in

a decrease in ∆G^q_elim, but the latter actually results in an

increase, and ∆G^TSreflects both components. Consequently,

it must be inferred that the correlation of ∆G^qwith PA stems

from the change in ∆G^q_elimthat deriVes solely from a positional

Figure 11 presents a conventional depiction of a standard free

energy surface where intermediate (^tBu₃SiNH)₂(^tBu₃SiN))Zr-

(RH) (2-RH) and 2-R′H complexes reside at positions succeed-

ing the respective 1,2-RH-elimination transition states (labeling

experiments discount reversible elimination/addition prior to RH

loss). Transients such as (^tBu₃SiNH)₂ZrdNSi^tBu₃(2) or even

five-coordinate (^tBu₃SiNH)₂(^tBu₃SiN))Zr(RH)(R′H) (2-RH,R′H)

may exist (I) or may simply be representative of transition states

for RH for R′H exchange via dissociative or associative

interchange pathways (II).⁸⁹

2. Associative Exchange in (^tBu₃SiNH)₃ZrCtCR (1-

CtCR). Alkynyl derivatives, (^tBu₃SiNH)₃ZrCtCR (1-CtCPh,

1-CtC^tBu), are conspicuous by their absence in the previous

mechanistic discussions. Thermolysis of 1-CtCPh or 1-Ct

change (e.g., x₂f x₁results in ∆G^TS< ∆G^TS_a′) along the

a

reaction coordinate. An RH substrate more tightly bound in

the transition state would naturally have its R group and H atom

nearer the Zr and N, respectively. As a corollary, note that PA

would not be expected to correlate with ∆G^TSor (x_n(TS) - x_n)/

(x_p- x_n), the fractional position of the transition state along

the reaction coordinate, because these are influenced by both

x_p- x_nand ∆G°_n(1-R).

(89) Atwood, J. D. Inorganic and Organometallic Reaction Mechanisms;

Brooks/Cole: Monterey, CA, 1985.

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 607

C^tBu in benzene-d₆at 96.7 °C for several days did not induce

1,2-elimination of the terminal acetylenes. Phenylacetylene loss

from 1-CtCPh occurred in the presence of excess tert-

butylacetylene, affording 1-CtC^tBu rather than 1-(ND)₃-C₆D₅,

and the reaction rate was roughly first-order in [^tBuCtCH]

according to initial rate studies (<20% conversion, k_CCPh) 1.1-

(4) × 10^-3M^-1s^-1). Likewise, elimination of ^tBuCtCH from

1-CtC^tBu occurred in the presence of excess PhCtCH to

provide 1-CtCPh, and rough initial rate measurements revealed

a first-order dependence on [PhCtCH], with k_CCBu) 1.2 (5)

Calculations of 1,2-HCtCH elimination from 1′-CtCH re-

vealed a large activation energy, and a significant interaction

of the C_Rwith the zirconium center as abstraction of the amide

hydrogen occurs, suggesting that a low energy pathway toward

2 + 2 addition to the resulting imide is plausible.⁷⁸

Thermolysis (C₆D₆, 96.5 °C) of (^tBu₃SiNH)₃ZrCy (1-Cy)

generated 2 in the presence of HCtCPh and HCtC^tBu,

providing a 1.5:1.0 ratio of 1-CtCPh to 1-CtC^tBu at low

conversion (∼6%). The amide resonances of the two products

were again clearly resolved, indicative of no exchange with

C₆D₆. While only limited exploration of the alkynyl derivatives

proved possible, the reactivity can be rationalized via the

associative or dissociative exchange processes involving

2-RCtCH and 2-RH complexes in Figure 11 (II), where the

transition state for HCtCR exchange with benzene is at lower

energy than the transition state for 1,2-PhH-addition to give

1-Ph. In addition, the transition state(s) for HCtCPh exchange

with HCtC^tBu occurs at significantly lower energy than their

respective exchanges with benzene, and formation of 1-CtCR

from 2-RCtCH occurs at lower energy than alkyne exchange.

C₆D₆

(^tBu SiNH) ZrCtC^tBu

+ PhCtCH y

z

3

96.7 °C

1-CtC^tBu

+ ^tBuCtCH (42)

t

( Bu₃SiNH)₃ZrCtCPh

1-CtCPh

× 10^-3M^-1s^-1(eq 42). In support of the initial rate studies,

equilibrium studies revealed that 1-CtCPh is slightly favored

(K(eq 42) ) 0.5, ∆G° ∼ 0.5 kcal/mol). Since the rates of alkyne

exchange appear to be second-order in both directions, an

associative interchange pathway is tentatively proferred.⁸⁹

σ-Bond metathesis⁹represents an alternative pathway, but one

difficult to reconcile with the lack of supporting evidence from

the aforementioned studies of 1-R.

On the basis of the bond strength (D(RCtC-H) ∼ 120-

135 kcal/mol) and proton affinity estimates, 1,2-alkyne elimina-

tion should be swift, provided the conventional mechanistic

scheme is relevant. A pathway consistent with the other 1-R

complexes involves the intermediacy of azametallacyclobutene

Conclusions

Mechanistic Overview. Quantitative aspects of the 1,2-RH-

elimination/addition events have enabled a greater understanding

of the free energy surface, albeit with some ambiguity regarding

the true character of intermediate states. Figure 11 serves to

summarize the mechanistic understanding of this deceptively

simple process. (^tBu₃SiNH)₃ZrR (1-R) complexes possess

strong metal-carbon bonds whose energetic differences (i.e.,

D(Zr-R) - D(Zr-R′)) nearly correspond to their respective

c

hydrocarbon differences for a select group of R (R ) Pr, Ph,

intermediates, (^tBu₃SiNH)₂ZrC(R))C(H)NSi^tBu₃(R ) Ph,

2-HCtCPh; ^tBu, 2-HCtC-^tBu), formed via 1,2-RC₂H-elimina-

tion and subsequent 2+2 addition to the imide. Intermediate

2-HCtCR complexes can be considered the equivalent of the

alkane or arene complexes (2-RH) previously proposed. The

exchange in eq 42 can be accommodated by Figure 11 provided

the transition states for 1,2-elimination of RH (i.e., HCtCPh)

and R′H (i.e., HCtC^tBu) are considered energetically lower

than the transition state pertaining to associative interchange

(II). In vanadium,²⁷titanium,²⁵and zirconium systems,²²

Ar, CH₂Ph, Me). The remaining 1-R derivatives (R ) Et,

t

CH₂Bu, Cy) have ground states that are higher, probably as a

result of the inductive and minor steric effects of additional

R-alkyls. A system with greater versatility will be needed to

answer a broader spectrum of questions concerning ground state

stability and therefore C-H bond activation selectivity.²⁵

In the rate-determining 1,2-RH-elimination event, the large

primary KIEs are consistent with a linear, rather symmetric

transition state (2-RH^q, 2-R′H^q), yet one that is somewhat loose.

The secondary KIE on CD₃H elimination supports calculations

that portray the methyl as relatively undeformed in the transition

state,^43,78,79hence the R group is envisioned as rotating its

azametallacyclobutene (e.g., (^tBu₃SiO)₂TiC(Me))C(Me)N-

Si^tBu₃)²⁵complexes have been isolated and characterized,

suggesting that 2-HCtCR may be energetically near 1-CtCR

(R ) Ph, ^tBu). Related oxametallacyclobutenes of Cp₂Ti have

recently been equilibrated with corresponding hydroxide acetyl-

ide complexes.⁹⁰

Some other comments about the relative free energy of

1-CtCR are germane. When either 1-CtCPh or 1-CtC^tBu

are thermolyzed in C₆D₆for prolonged periods, their amide sites

are not deuterated, precluding reversible addition of benzene-

d₆. Thermolysis (C₆D₆, 100 °C) of (^tBu₃SiNH)₃ZrPh (1-Ph) in

the presence of HCtC^tBu afforded 1-CtC^tBu among other

products, proving that a pathway exists between 1-Ph and

1-CtC^tBu. Estimates place the ground state of 1-CtC^tBu at

least ∼8.6 kcal/mol lower than the ground state 1-Ph,⁹¹hence

the overall barrier of 1,2-RCtCH elimination and alkyne

dissociation from putative 2-RCtCH may be insurmountable.

t

σ-orbital into alignment as the Bu₃SiNH unit functions as a

weak acid, a reasonable consideration in view of strong N(pπ)

f Zr(dπ) bonding by the amide. Geometric considerations and

the calculations depict the hydrogen transfer as linear between

N and R, and within bonding distance of the zirconium via

utilization of an appropriate, empty Zr(dπ)-orbital.⁹²According

to the Hammond analysis, the transition state is generally late,

but possesses a mixed composition, hence the ground state

energy of 1-R and its position along the reaction coordinate

can both dramatically influence the transition state energy and

∆G^q

.

elim

The correlation of ∆G^q

with RH proton affinity suggests

elim

that tight binding of RH in the transition state for 1,2-RH-

elimination expedites the elimination process and evokes the

possibility of intermediate alkane complexes (2-RH, 2-R′H).

Calculations of hypothetical (H₂N)₂(HN)Zr(η²-CH₄) at 373 K

indicate a binding enthalpy of ∼ -6.3 kcal/mol⁴⁴at 373 K

relative to (H₂N)₂ZrdNH + CH₄, but the free energy for binding

may be >0 kcal/mol, hence possible 2-RH species are likely to

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

1995, 117, 5393-5394.

(91) Assume that the prolonged thermolysis (C₆D₆, 24 h, 96.7 °C) of

1-CtC^tBu (0.036 M) yields equilibrium amounts of 1-(ND)_x-Ph-d₅and

HCtC^tBu and that the latter can be confidently detected if present in >5%

(>(0.05)(0.036 M)). No evidence for HCtC^tBu was observed. Assuming

the density of C₆D₆(0.95 g/ml) is relatively insensitive to temperature, neat

benzene-d₆is ∼11.2 M. At equilibrium: K g [C₆D₆][1-CtC^tBu]/[1-(ND)_x-

Ph-d₅][HCtC^tBu] ) [11.2][0.0342]/[0.0018]²; K > 1.2 × 10⁵and ∆G° e

-8.6 kcal/mol.

(92) For related ab initio calculations probing methane activations of a

different type, see: (a) Schreiner, P. R.; von Rague´ Schleyer, Schaefer III,

H. F. J. Am. Chem. Soc. 1993, 115, 9659-9666. (b) Olah, G. A.; Hartz,

N.; Rasul, G.; Surya Prakash, G. K. J. Am. Chem. Soc. 1995, 117, 1336-

1343.

608 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

be energetically similar to (^tBu₃SiNH)₂ZrdNSi^tBu₃(2) + RH.

Provided the acetylide derivatives, 1-CtCR (R ) Ph, ^tBu) can

be described by a related free energy surface, the associative

terminal alkyne exchange can be considered a limiting case of

2-RH + R′H h 2-R′H + RH, providing tentative support for

intermediate alkane and arene complexes as weak solvates.

Similar selectivity trends in systems as disparate as d⁸, 16

t

e^-, Tp′Rh(CNCH₂Bu), and Cp*Rh(PMe₃) in comparison to

putative d⁰solvates of (^tBu₃SiNH)₂ZrdNSi^tBu₃(2) and (^tBu₃-

SiO)₂Ti)NSi^tBu₃suggest that properties of each hydrocarbon

dictate its activation, but the subtleties and varying degrees of

selectivity imply that metal centers can tune this reactivity. In

t

the Tp′Rh(CNCH₂Bu) system, the hydrocarbon selectivities

If 2-RH(R′H) transients exist, exchange of bound hydrocar-

bons may take place with an intermediate of lower (3) or higher

(5) coordination number (Figure 11, I) or without (II). Like

classic interchange substitutions,⁸⁹the latter pathway can have

dissociative (e.g.,the 2‚‚‚RH is virtually broken as the R′H enters

the coordination sphere) or associative (e.g., substantial 2‚‚‚R′H

bonding has incurred as 2‚‚‚RH is starting to break) character

(II). It should be noted that 2 could internally solvate itself

via coordination of a peripheral methyl agostic bond, a possible

parallel the thermodynamic stabilities of the product Tp′Rh-

t

(CNCH₂Bu)(H)R complexes. In the zirconium imido system,

benzylic activation is kinetically less favored than predicted on

the basis of ground state stabilities, but this apparent discrepancy

may be due to an anomalous ground state stabilization of 1-CH₂-

Ph. A perusal of these systems suggest that steric factors can

contribute to substrate discrimination. The bis(amido)imido

ligand arrangement of 2 (or 2-RH) is likely to hinder substrate

approach more than in solvates of (^tBu₃SiO)₂Ti)NSi^tBu₃, where

the linear siloxides provide a more open electrophilic pocket.

intermediate en route to (^tBu₃SiNH)₂ZrNHSi^tBu₂CMe₂CH₂(3),

but thermodynamic parameters associated with this intramo-

lecular solvation are difficult to estimate. Calculations on

(H₂N)₂ZrdNH (2′) and (Me₃SiNH)₂ZrdNSiMe₃(2′′) suggest

that (^tBu₃SiNH)₂ZrdNSi^tBu₃(2) is likely to distort toward a

pyramidal structure; in fact, the zirconium in 2′ and 2′′ lies 0.04

and 0.32 Å above the N₃plane, and ZrdN-Si deviates from

t

Both systems are likely to be less open than Tp′Rh(CNCH₂Bu),

which is in turn more congested than Cp*Rh(PMe₃).

The rate-determining capture of RH by the late metal d⁸

systems leads to an alkane adduct, which subsequently adds to

the metal center. This contrasts with 1,2-RH-addition to the

transient imido, where proposed alkane/arene complexes precede

rate-determining C-H bond activation, an event with intrinsi-

cally greater selectivity. Rate-determining 1,2-RH-addition

across the ZrdN unit of the reactive intermediate clearly requires

greater substrate-specific orientation than rate-determining

formation of alkane/alkene adducts prior to oxidative addition

of RH to d⁸metal centers.⁹³A system in which a greater

number of substrates may be assessed is greatly needed in order

to zero in on the critical factors affecting selectivity.²⁵

linearity (165°).⁷⁸Through distortion, d_z/p_zmixing can occur

2

to enhance the electrophilic character of the orbital oriented

toward the incoming hydrocarbon. Given the potent electro-

positive nature of low-coordinate zirconium, it is unlikely that

2 exists as an unsolvated pseudotrigonal species, hence internal

solvation via an agostic effect or external solvation via substrate

or solvent binding is most reasonable.

In viewing the activation of a bound hydrocarbon, it is

informative to use (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃(2-THF) as

a model. Envision the pair of electrons in the R-H σ-bond

Uncompetitive â-H-Elimination. â-Hydride elimination in

i

1-R (R ) Cy, Bu, Et, ^cPr) was not evidenced, although

insertions of isobutylene, acetylene and 1,1-dimethylallene into

the Zr-H bond of 1-H proved such a pathway exists. The

barrier for CH₂dCMe₂insertion into 1-H is estimated to be

∼15 kcal/mol at 100 °C in view of the swift (<20 min) insertion

under pseudo-first-order conditions at 25 °C. Provided the

transition state for 1,2-RH-elimination (∆G^q) 27.7 ≈ 28 kcal/

mol) from 1-ⁱBu is at least 3 kcal/mol lower than the â-H-

elimination transition state, ∆H° for insertion is e-16 kcal/

mol. Previous arguments ascribe the relative paucity of â-H-

elimination to steric congestion (i.e., Cr(^tBu₄))⁹⁴or the absence

of vacant cis coordination sites in tetrahedral molecules.

Chisholm et al.⁹⁵has proposed that strong X(pπ) f M(dπ)

interactions raise the level of metal-based empty orbitals critical

to accommodating the eliminating alkene. The cycle expressed

above suggests that there is a pronounced endothermicity to a

â-H-elimination event leading to free olefin, hence the stability

of related metal alkyls (i.e., Cp₂MX(^tBu),⁹⁶(Me₂N)₄Ta^tBu,⁹⁵

(^tBu₃SiNH)(THF)^tBuTi)NSi^tBu₃²³) may be predominantly ther-

modynamic in origin. From an enthalpy cycle using ∆H_f°

2

being drawn into the coordination sphere by the empty d_z/p_z

hybrid orbital, eventually binding in the position of the THF.

In 2-THF, the lone pair of the imide nitrogen is not significantly

involved in N(pπ) f Zr(dπ) bonding, thus it may also be free

to act as an internal base in the scission of the C-H bond in

2-RH. Use of the lone pair on nitrogen enables the formal

ZrdN π-bond to essentially remain intact as the R-H bond is

broken, since that specific π-interaction correlates with a N(pπ)

f Zr(dπ) bond of an amide that subsequently rotates to afford

ground state 1-R.

Hydrocarbon Selectivities. The quantitative hydrocarbon

kinetic selectivity data (i.e., ^cPrH ≈ ArH [0.0 kcal/mol] > MeH

t

[3.4] > PhCH₂H [4.0] > RH (R ) Et, CH₂Bu, Cy)), while

limited, exhibits some correspondence to the selectivities for

oxidative addition attributed to [HB(3,5-dimethylpyrazoyl)₃]Rh-

t

(CNCH₂Bu) at -15 °C (i.e., PhH [0.0 kcal/mol] > HCH₂-3,5-

n

c

Me₂C₆H₃[0.15] > MeH [0.4] > PeH [0.80] > PeH [1.7] >

CyH [1.8]),³but the magnitudes of each ∆∆G^q(relative to Ph-

H) are significantly greater. Both groups of data are consistent

with qualitative results of Cp*M(PMe₃) (M ) Rh, Ir) systems,³⁴

where activation of PhH . primary alkyl > cycloalkyl, although

the ∆∆G^q(-60 °C) for benzene vs cyclohexane activation

where M ) Rh is only 1.2 kcal/mol, and the selectivity is

substantially less for M ) Ir.⁷⁰Current data for (^tBu₃-

SiO)₂Ti)NSi^tBu₃(i.e., 25 °C; PhH [0.0 kcal/mol] > MeH [1.0]

> PhCH₂H [2.6]) reveal somewhat less selectivity than the

zirconium system,²⁵but it is still greater than the late metal

derivatives. Selectivity differences for second-order σ-bond

metatheses in the Cp*₂ScR system are difficult to quantify; using

Cp*₂ScMe as the benchmark, benzene activation is ∼0.8 kcal/

mol easier than that of methane.⁹Carbon-hydrogen bond

‚

estimates for (CH₃)₂CHCH₂(14 kcal/mol) and isobutylene

(-3.6 kcal/mol),⁶¹the difference in zirconium-hydride and

-alkyl bond strengths (i.e., D(Zr-H) - D(Zr-C)) in this

system is e18 kcal/mol.

Experimental Section

General Considerations. All manipulations were performed using

either glovebox or high vacuum line techniques. Hydrocarbon solvents

containing 1-2 mL of added tetraglyme were distilled under nitrogen

(93) Cundari, T. R. J. Am. Chem. Soc. 1994, 116, 340-347.

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

(95) Chisholm, M. H.; Tan, L.-S.; Huffman, J. C. J. Am. Chem. Soc.

1982, 104, 4879-4884.

(96) (a) Bu¨chwald, S. L.; Kreutzer, K. A.; Fisher, R. A. J. Am. Chem.

Soc. 1990, 112, 4600-4601. (b) Bu¨chwald, S. L.; Lum, R. T.; Fisher, R.

A.; Davis, W. M. J. Am. Chem. Soc. 1989, 111, 9113-9114.

activations by Cp*₂ThCH₂CMe₂CH₂qualitatively follow a trend

similar to those above.¹¹

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 609

from purple benzophenone ketyl and vacuum transferred from same

prior to use. Benzene-d₆and cyclohexane-d₁₂were dried over activated

4 Å molecular sieves, vacuum transferred, and stored under N₂; THF

and THF-d₈were dried over sodium benzophenone ketyl. All glassware

was base-washed and oven dried, and NMR tubes were additionally

flamed under dynamic vacuum. Methane, ethane, ethylene, allene,

cyclopropane, and isobutylene (Matheson) were typically passed through

a -78 °C trap prior to use. ^tBu₃SiNHLi, (^tBu₃SiNH)₃ZrCl (1), (^tBu₃-

SiNH)₃ZrMe (1-Me), and (^tBu₃SiNH)₃ZrH (1-H) were prepared ac-

cording to published procedures.⁴⁵

IR (Nujol, cm^-1) 3244 (w), 1608 (w), 1565 (w), 1545 (w), 1360 (m),

1069 (s), 1010 (m), 942 (m), 869 (w), 815 (s), 721 (w), 610 (s). Anal.

Calcd for ZrC₃₉H₈₉N₃Si₃: C, 60.39; H, 11.56; N, 5.42. Found: C,

60.23; H, 11.65; N, 5.38.

f. (^tBu₃SiNH)₃ZrC₆H₅(1-Ph). 1-Cl (560 mg, 0.729 mmol) in 20

mL of ether and 0.39 mL of phenyl magnesium bromide in ether (2.0

M, 0.76 mmol) gave 350 mg of 1-Ph (59%). Typical syntheses

involved C-H bond activation. Thermolysis of variable amounts of

1-R at 100 °C in a glass bomb reactor containing C₆H₆(the time was

dependent on R) induced 1,2-RH-elimination to give 2, which was

subsequently trapped by benzene to give 1-Ph in yields >75% when

isolated as above: IR (Nujol, cm^-1) 3239 (w), 1358 (m), 1125 (m),

1065 (s), 1008 (m), 930 (m), 872 (w), 810 (s), 718 (m), 695 (m), 642-

(w), 610 (s). Anal. Calcd for ZrC₄₂H₈₉N₃Si₃: C, 62.15; H, 11.05; N,

5.18. Found: C, 58.08; H, 10.92; N, 4.88.

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 Mattson FT-IR interfaced to an AT&T PC7300

computer or a Perkin Elmer 299B spectrophotometer. Analyses were

performed by Texas Analytical Labs, Stafford, TX or Oneida Research

Services, Whitesboro, NY. We have had difficulty obtaining satisfac-

t

g. (^tBu₃SiNH)₃ZrCH₂Bu (1-CH₂Bu). 1-Cl (510 mg, 0.662 mmol)

t

and ^tBuCH₂Li (57 mg, 0.728 mmol) reacted in 20 mL of ether to afford

tory analyses of the complexes containing the Bu₃SiNH ligand,and

t

274 mg of 1-CH₂Bu (51%): IR (Nujol, cm^-1) 3221 (w), 1360 (m),

they have been difficult to analyze; low carbon percentages (by 1-2%)

are typical, while the hydrogen and nitrogen numbers are sporadic,

despite use of various burning aids. The failures may be due in part

1228 (m), 1210 (w), 1188 (w), 1065 (s), 1006 (m), 929 (m), 870 (w),

810 (s), 610 (s). Anal. Calcd for ZrC₄₁H₉₅N₃Si₃: C, 61.12; H, 11.88;

N, 5.21. Found: C, 61.23; H, 11.95; N, 5.11.

t

from partial decomposition of the Bu₃SiNH ligand to SiC and Si₃N₄

during combustion.⁹⁷

h. (^tBu₃SiNH)₃ZrCCPh (1-CCPh). 1-Cl (478 mg, 0.621 mmol)

and PhCCLi (73 mg, 0.675 mmol) reacted in 20 mL of THF to afford

210 mg of 1-CCPh as an orange powder (41%): IR (Nujol, cm^-1) 3258

(w), 2082 (s), 1595 (w), 1574 (w), 1366 (m), 1208 (m), 1060 (s), 1010

(m), 1005 (m), 942 (m), 876 (w), 815 (s), 758 (w), 686 (m), 620 (s).

Anal. Calcd for ZrC₄₄H₈₉N₃Si₃: C, 63.24; H, 10.73; N, 5.03. Found:

C, 62.36; H, 10.86; N, 4.95.

Procedures. 1. General (^tBu₃SiNH)₃ZrR (1-R) from 1-Cl and

RMgX or RLi. To an appropriately sized flask containing a slurry of

1-Cl in ethereal solvent at -78 °C was added a solution of RMgX or

RLi by syringe under Ar counterflow. In certain cases, solid Grignard

or lithium reagents were mixed with 1-Cl in the flask, and the solvent,

either ethereal or hydrocarbon, was added at -78 °C via distillation.

The mixture was allowed to warm to 25 °C over the course of ∼2 h

and stirred for an additional 8-12 h at 25 °C. Upon removal of the

volatiles, hexane was added to the residue. The slurry was filtered,

and the filter cake was washed repeatedly with hexanes and concen-

trated. Cooling to -78 °C afforded colorless, microcrystalline 1-R in

variable yields upon isolation by filtration; sometimes an additional

crop was taken.

i. (^tBu₃SiNH)₃ZrCC^tBu (1-CC^tBu). 1-Cl (344 mg, 0.446 mmol)

t

and BuCCLi (39 mg, 0.446 mmol) reacted in 15 mL of benzene, but

1

a H NMR spectrum of the mixture revealed 60% completion; 30 mg

t

of BuCCLi 0.341 mmol) was added to ultimately give 285 mg of

1-CC^tBu as a yellow powder (81%): IR (Nujol, cm^-1): 3265 (w), 2090

(m), 1367 (m), 1250 (m), 1207 (w), 1136 (w), 1055 (s), 1012 (m), 943

(m), 876 (w), 815 (s), 932 (w), 620 (s). Anal. Calcd for ZrC₄₂H₉₃N₃-

Si₃: C, 61.84; H, 11.49; N, 5.14. Found: C, 61.62; H, 11.58; N, 5.06.

j. (^tBu₃SiNH)₃Zr(η³-BH₄) (1-BH₄). 1-Cl (318 mg, 0.413 mmol)

and LiBH₄(90 mg, 4.13 mmol) reacted in 6 mL of benzene to afford

274 mg of 1-BH₄(56%): IR (Nujol, cm^-1) 3228 (w), 2530 (m), 2150

(m), 2210 (m), 1360 (m), 1201 (m), 1061 (s), 1006 (m), 929 (m), 876

(w), 805 (s), 610 (s). Anal. Calcd for ZrBC₃₆H₈₈N₃Si₃: C, 57.70; H,

11.84; N, 5.61. Found: C, 57.58; H, 11.89; N, 5.56.

a. (^tBu₃SiNH)₃ZrC₂H₅(1-Et). 1-Cl (1.556 g, 2.02 mmol) in 25

mL of Et₂O and 1.06 mL of EtMgBr in ether (2.0 M, 2.12 mmol) gave

882 mg of 1-Et (57%); a second crop yielded 235 mg (77% total): IR

(Nujol, cm^-1) 3248 (w), 1360 (m), 1060 (s), 1006 (m), 942 (m), 872

(w), 815 (s), 721 (w), 610 (s). Anal. Calcd for ZrC₃₈H₈₉N₃Si₃: C,

59.77; H, 11.75; N, 5.50. Found: C, 59.61; H, 11.86; N, 5.38.

b. (^tBu₃SiNH)₃ZrC₆H₁₁(1-Cy). 1-Cl (500 mg, 0.649 mmol) in

10 mL of Et₂O and 0.36 mL of cyclohexyl magnesium chloride in ether

(2.0 M, 0.715 mmol) gave 250 mg 1-Cy (46%): IR (Nujol, cm^-1) 3230

(w), 1580 (w), 1360 (m), 1069 (s), 1008 (m), 960 (w), 939 (m), 861

(w), 810 (s), 720 (w), 612 (s). Anal. Calcd for ZrC₄₂H₉₅N₃Si₃: C,

61.69; H, 11.71; N, 5.14. Found: C, 65.21; H, 10.86; N, 4.83.

c. (^tBu₃SiNH)₃ZrCH₂Ph (1-CH₂Ph). 1-Cl (950 mg, 1.234 mmol)

in 15 mL of THF and 0.65 mL of benzyl magnesium chloride in ether

2. (^tBu₃SiNH)₃Zr-^cC₃H₅(1-^cPr). To a 200 mL glass bomb

containing 1-Cy (333 mg, 0.407 mmol) was added 40 mL of

cyclohexane via vacuum transfer. Cyclopropane was admitted (∼20

equiv, 1 atm), and the vessel was placed in a 95 °C bath for 4 h. After

cooling to 25 °C, the volatiles were removed. The solid was dissolved

in 5 mL of hexane, and the solution was filtered, concentrated, and

cooled to -78 °C. White crystalline 6-^cPr was collected by filtration

(110 mg, 35%, 5% impurities, chiefly (^tBu₃SiNH)₄Zr (3)): IR (Nujol,

cm^-1) 3245 (w), 1535 (w), 1358 (m), 1060 (s), 1008 (m), 928 (m),

860 (w), 810 (s), 717 (w), 610 (s). Anal. Calcd for ZrC₃₉H₉₀N₃Si₃:

C, 60.31; H, 11.68; N, 5.41. Found: C, 60.22; H, 11.74; N, 5.38.

3. (^tBu₃SiNH)₃Zr-CH₂C₆H₃-3,5-Me₂(1-Mes). To a 50 mL glass

bomb containing 1-CH₃(516 mg, 0.688 mmol) was added 7 mL of

mesitylene via vacuum transfer. The vessel was placed in a 95 °C

bath for 8 h and cooled to 25 °C. The volatiles were removed under

vacuum, and the solid was dissolved in 5 mL of hexane and filtered.

The solvent was removed, and the product was dissolved in ether. The

solution was cooled to -78 °C, giving white, crystalline 1-Mes, which

was collected by filtration (371 mg, 62%): IR (Nujol, cm^-1) 3234 (w),

1600 (m), 1365 (m), 1295 (m), 1160 (m), 1075 (s), 1012 (m), 989 (w),

932 (m), 876 (m), 820 (s), 722 (w), 696 (m), 620 (s). Anal. Calcd for

ZrC₄₅H₉₅N₃Si₃: C, 63.16; H, 11.19; N, 4.91. Found: C, 62.66; H,

11.16; N, 4.66.

(2.0 M, 1.30 mmol) gave 356 mg 1-CH₂Ph (35%): IR (Nujol, cm^-1

)

3238 (w), 1655 (w), 1598 (m), 1365 (m), 1205 (m), 1165 (w), 1152

(w), 1068 (s), 1030 (m), 1011 (m), 990 (m), 932 (m), 872 (w), 810 (s),

742 (m), 721 (w), 695 (m), 610 (s). Anal. Calcd for ZrC₄₃H₉₁N₃Si₃:

C, 62.55; H, 11.11; N, 5.09. Found: C, 61.18; H, 11.29; N, 5.27.

d. (^tBu₃SiNH)₃ZrCHdCH₂(1-CHdCH₂). 1-Cl (510 mg, 0.662

mmol) in 20 mL of ether and 0.35 mL of vinyl magnesium bromide in

ether (2.0 M, 0.695 mmol) gave 282 mg 1-CHdCH₂(56%): IR (Nujol,

cm^-1) 3245 (w), 1550 (w), 1535 (w), 1360 (m), 1050 (s), 1010 (m),

930 (m), 870 (w), 810 (s), 715 (w), 610 (s). Anal. Calcd for

ZrC₃₈H₈₇N₃Si₃: C, 59.93; H, 11.51; N, 5.52. Found: C, 59.77; H,

11.75; N, 5.50.

e. (^tBu₃SiNH)₃Zr(η³-CH₂CHCH₂) (1-allyl). 1-Cl (514 mg, 0.667

mmol) in 10 mL of ether and 0.67 mL of allyl magnesium bromide in

ether (1.0 M, 0.670 mmol) gave 163 mg of off-white 1-allyl (32%):

(97) Sample ¹H and ¹³C{¹H} NMR spectra of selected derivatives are

included in the supporting information for ref 21.

(98) Solubility Data Series; Kertes, A. S., Ed.; Pergamon Press, London;

Vol. 27/28 (Methane).

4. (^tBu₃SiNH)₄Zr (1-NHSi^tBu₃). a. From 1-Me. To a flask

containing 1-Me (610 mg, 0.814 mmol) was added 25 mL of

cyclohexane by vacuum distillation. The solution was brought to reflux

for 10 h and then cooled. The volatiles were removed, and the yellow

residue was taken up in pentane and filtered. The volatiles were again

removed, and the residue was dissolved in 3 mL of THF. Cooling to

-78 °C gave white microcrystals of 3, which were collected by filtration

(150 mg, 0.204 mmol, 25%). b. From 1-Cl. To a glass bomb

(99) Cromer, D. T.; Mann, J. B. Acta Crystallogr., Sect. A 1968, A24,

321-324.

(100) R ) ∑||F_o| - |F_c||/(∑|F_o|); R_w) {∑w(|F_o| - |F_c|)²/∑w(|F_o|)²}^1/2

;

GOF ) {∑[weight(|F_o| - |F_c|)²]}/(M - N) where M ) number of

observations and N ) number of parameters; 3257 (88.9%) reflections with

|F_o| g 3σ(F_o).

610 J. Am. Chem. Soc., Vol. 118, No. 3, 1996

Schaller et al.

containing ^tBu₃SiNHLi (0.175 g, 0.791 mmol) and 1-Cl (0.607 g, 0.788

mmol) was distilled 25 mL of hexanes at -78 °C. The reactor was

then heated at 145 °C for 48 h leading to the precipitation of LiCl.

This white mixture was then filtered and extracted once with 10 mL

of hexanes. The solution was then concentrated and cooled to -78

°C, giving a colorless crystalline 1-NHSi^tBu₃which was collected by

filtration (0.646 g, 86%): IR (Nujol, cm^-1) 3218 (w), 1360 (m), 1066

(s), 1008 (m), 930 (m), 790 (s), 722 (w), 620 (s). Anal. Calcd for

ZrC₄₈H₁₁₂N₄Si₄: C, 58.95; H, 11.40; N, 5.73. Found: C, 58.32; H,

11.63; N, 4.74.

in the dry box, and brought out on needle valve adapters. Deuterated

solvents were distilled into the tubes under vacuum, and the samples

were freeze-pump-thaw degassed (-196 °C). Any gases or volatile

liquids were subsequently admitted via calibrated gas bulbs when exact

amounts were needed (measured via manometer) or via lecture bottles.

The tubes were then sealed using a torch. Example: Equilibration

of 1-Ph and 1-^cPr. An NMR tube sealed to a 14/20 ground glass

joint was charged with 1-Cy (24 mg, 29 mmol). The tube was attached

to a gas bulb and evacuated. C₆D₁₂(0.7 mL) was vacuum transferred

into the tube and freeze-pump-thaw degassed. Benzene was added

(280 Torr in 12.7 mL, 0.194 mmol, 6.6 equiv), followed by cyclopro-

pane (44 Torr in 12.7 mL, 0.030 mmol, 1.0 equiv), and the tube was

sealed with a torch. Thermolysis of the mixture at 96.7 °C for 14 h

gave equilibrium quantities of C₆H₆, 1-Ph, C₃H₆, and 1-^cPr which were

integrated to give K ) ([C₆H₆][1-^cPr]/[C₃H₆][1-Ph] ) 1.57(33).

General Kinetics. 1. 1,2-RH-Elimination from 1-R. Solutions

of 1-R in the appropriate deuterated solvent were prepared in 2-mL

volumetric flasks. Three samples of about 0.6 mL each were transferred

to flame-dried, 5-mm NMR tubes sealed to 14/20 joints and attached

to 180° needle valves. The tubes were freeze-pump-thaw degassed

three cycles (77 K) and flame-sealed under vacuum. The three sample

tubes were simultaneously heated by immersion in a polyethylene glycol

bath with a Tamson immersion circulator. The typical bath temperature

of 96.7 °C was stable to (0.2 °C. Rates of disappearance of amido

NH peaks were monitored in all cases except for the benzene loss from

1-Ph (see text) and the determination of k_H/k_Dpertaining to (^tBu₃-

SiNH)₃ZrMe (1-Me) vs (^tBu₃SiND)₃ZrMe (1-(ND)₃Me). Separate tubes

of these complexes were measured in tandem by the disappearance of

the methyl resonance. Similarly, the isotope effect pertaining to (^tBu₃-

SiNH)₃ZrPh (1-Ph) vs (^tBu₃SiND)₃ZrPh (1-(ND)₃-Ph) was monitored

by following the phenyl ortho hydrogens, and the k_H/k_Dfor (^tBu₃-

SiNH)₃ZrCH₂Ph (1-CH₂Ph) vs (^tBu₃SiND)₃ZrCH₂Ph (1-(ND)₃-CH₂Ph)

was monitored by the disappearance of the methylene resonance. All

runs were monitored for 5-6 half-lives. Single transient spectra were

used to obtain the most reproducible integrals. The data collection,

rates, and uncertainties were obtained by using weighted (1/σ², where

σ was obtained from three simultaneous runs if available) or un-

weighted, nonlinear least-squares fitting to the exponential form of the

rate expression.

5. (^tBu₃SiNH)₂ZrN(H)Si^tBu₂C(Me)₂CH₂(3). A vessel containing

209 mg of 1-Me (0.279 mmol) was attached to a needle valve and

heated at 125 °C under dynamic vacuum for 10 h. After cooling, white,

powdered 3 was collected. ¹H NMR showed 5% 1-NHSi^tBu₃as an

impurity: IR (Nujol, cm^-1) 3250 (w), 3228 (w), 1365 (m), 1195 (w),

1172 (w), 1065 (s), 1010 (m), 942 (m), 810 (s), 725 (w), 620 (s). Anal.

Calcd for ZrC₃₆H₈₄N₃Si₃: C, 58.95; H, 11.40; N, 5.73. Found: C,

55.46; H, 10.79; N, 4.47.

6. (^tBu₃SiNH)₃ZrCH₂CHMe₂(1-ⁱBu). To a thick-walled vessel

containing a frozen solution of 1-Me (734 mg, 0.979 mmol) in 50 mL

of cyclohexane was admitted 1 atm H₂at -195 °C. The solution was

thawed, and the bomb was placed in a 105 °C bath for 8 h to generate

1-H. The volatiles were removed, and the residue was placed in a

small flask and dissolved in 15 mL of hexanes. The flask was opened

to the vacuum manifold and exposed to 0.8 atm isobutylene. The

uptake of isobutylene (the gas pressure in the manifold was monitored

by manometer) was complete after 20 min. After stirring for 3 h, the

solvent volume was reduced to ∼3 mL, and a white powder was

collected by filtration at 20 °C (116 mg, 15% overall yield): IR (Nujol,

cm^-1) 3228 (w), 1360 (m), 1065 (s), 1010 (m), 929 (m), 870 (m), 810

(s), 721 (w), 610 (s). Anal. Calcd for ZrC₄₀H₉₃N₃Si₃: C, 60.76; H,

11.85; N, 5.31. Found: C, 60.30; H, 11.81; N, 4.93.

7. (^tBu₃SiNH)₃Zr(η¹-CH₂CHdCMe₂) (1-dma). To a flask con-

taining a frozen solution of 1-H (270 mg, 0.367 mmol) in 15 mL of

benzene was added 1.1 equiv of dimethylallene (36.7 mL at 204 Torr)

by condensation. The needle valve was closed, and the solution was

stirred at 25 °C for 3 h. Upon removal of the volatiles, the residue

was dissolved in ∼10 mL hexanes, concentrated to ∼2 mL, cooled to

-78 °C, and filtered to afford 1-dma as a white powder (71 mg,

24%): IR (Nujol, cm^-1) 3228 (w), 1690 (w), 1638 (w), 1608 (w), 1360

(m), 1193 (w), 1065 (s), 1010 (m), 932 (m), 810 (s), 726 (w), 692 (w),

610 (s). Anal. Calcd for ZrC₄₁H₉₃N₃Si₃: C, 61.27; H, 11.66; N, 5.23.

Found: C, 61.22; H, 11.61; N, 5.28.

8. (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃(2-THF). To 730 mg of 1-Me

(0.975 mmol) in an 80 mL bomb (dried at 150 °C under vacuum for

8 h) was added 20 mL of benzene by vacuum transfer. The vessel

was placed in a 100 °C bath for 10 h. After cooling, the volatiles

were removed, and a crude ¹H NMR spectrum verified the residue was

clean 1-Ph. THF (20 mL) was added to the bomb by distillation, and

the vessel was heated at 100 °C for 0.75 h. The cooled solution was

transferred to a flask, and the solvent was removed and replaced with

hexanes. Reducing the volume to ∼5 mL and cooling to -78 °C

resulted in the precipitation of 2-THF, collected as a white powder by

filtration: IR (Nujol, cm^-1): 3260 (w), 1540 (w), 1365 (m), 1175 (w),

1045 (s), 1012 (m), 934 (m), 912 (w), 868 (m), 830 (s), 723 (w), 605

(s). Anal. Calcd for ZrOC₄₀H₉₂N₃Si₃: C, 59.63; H, 11.38; N, 5.22.

Found: C, 58.79; H, 11.15; N, 4.93.

2. Equilibrium of 1-CH₂Ph and 1-C₆H₄Me. A 0.033 M solution

of (^tBu₃SiNH)₃ZrMe (1-Me) in toluene was thermolyzed at 96.7 °C

for 1 h (1,2-MeH elimination of 1-Me was ∼20% complete). The

solvent was removed, and a ¹H NMR spectrum revealed ∼1% 1-CH₂-

Ph and ∼99% 1-C₆H₄Me as determined from the ratio of CH₂to CH₃

integrals. The latter was identified as a mixture of para (¹H NMR

(tentative assignments, C₆D₆) δ 2.55 (CH₃, s, 3H), 4.45 (NH, s, 1 H),

8.10, 8.24 (ArH, A₂B₂, J ) 6.3 Hz, 4 H) and meta or ortho (¹H NMR

(tentative assignments, C₆D₆) δ 2.28 (CH₃, s, 3H), 5.05 (NH, s, 1 H),

7.09 (ArH, m, 1H), 8.20 (ArH, m, 3H) isomers and was treated

collectively (see text). Thermolysis for an additional 8 h (∼4 half-

lives) resulted in a dramatic change in the product ratios, which

remained constant with further heating (i.e., ∼38% 1-CH₂Ph and ∼62%

1-C₆H₄Me).

3. Competitive activation of C₆H₆and CH₄. To each of three

tubes was added 20 mg of 1-Cy (0.024 mmol). The tubes were

individually evacuated and C₆D₁₂condensed into each, followed by 2

equiv of C₆H₆(0.048 mmol) as measured by gas bulb. Next, a known

amount of CH₄was condensed into each tube which then contained

the following, according to ¹H NMR spectra: tube 1, 0.69 mL, [1-Cy]

) 0.035 M, [C₆H₆] ) 0.071 M, [CH₄] ) 0.099 M; tube 2, 0.77 mL,

[1-Cy] ) 0.031 M, [C₆H₆] ) 0.064 M, [CH₄] ) 0.120 M; tube 3, 0.66

mL, [1-Cy] ) 0.036 M, [C₆H₆] ) 0.074 M, [CH₄] ) 0.120 M. The

methane concentration was calculated using its standard concentration

in cyclohexane at a partial pressure of 1 atm at 100 °C^9,98and Henry’s

law. Calculated methane concentrations at 25 °C were accurate in

9. (^tBu₃SiNH)₂(Et₂O)ZrdNSi^tBu₃(2-OEt₂). To a flask containing

crude 1-H (1.33 g, 1.80 mmol) was distilled 20 mL of ether at -78

°C; H₂evolution immediately ensued. The reaction mixture continued

to effervesce as it warmed to 25 °C and was maintained at 25 °C for

1 h. Upon removal of the volatiles, the residue was taken up in 12

mL of hexanes and filtered, and the solution volume was reduced to 5

mL. Cooling the solution to -10 °C gave colorless crystals of 2-OEt₂

that were collected by filtration (0.602 g, 41%).

10. (^tBu₃SiNH)₃ZrOCH₂CH₃(1-OEt). To a glass bomb containing

225 mg of 1-Cy (0.275 mmol) was distilled ∼3 mL of Et₂O and ∼12

mL of n-heptane. The reactor was heated at 115 °C for 16 h and cooled

to 25 °C. The solvent was removed in vacuo leaving a crystalline solid

that was dissolved in hexanes, filtered, and crystallized at -78 °C to

afford 165 mg of colorless 1-OEt (79%).

1

comparison to the H NMR spectra. Thermolysis at 97.5 °C ensued,

and the reaction was stopped after 17% conversion. The [1-Me]/[1-

Ph] ratio at low conversion represents k_MeH[2][CH₄]/k_Ph[2][C₆H₆],

permitting calculation of k_MeH/k_PhHand ∆∆G^q) 3.4 kcal/mol as an

average of the three trials.

Single-Crystal X-ray Diffraction Analysis of (^tBu₃SiNH)₂(THF)-

ZrdNSi^tBu₃(2-THF). A colorless needle (0.3 × 0.3 × 0.3 mm) of

(^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃(2-THF), obtained from benzene solu-

11. NMR Tube Reactions. Oven and flame-dried 5 mm NMR

tubes were sealed onto 14/20 ground glass joints, charged with reagents

Hydrocarbon ActiVation Via ReVersible 1,2-RH-Elimination

J. Am. Chem. Soc., Vol. 118, No. 3, 1996 611

tion, was sealed in a capillary. Preliminary X-ray diffraction photo-

graphs revealed monoclinic symmetry. Precise lattice constants,

determined from a least-squares fit of 15 diffractometer-measured 2Θ

values at 25 °C, were a ) 13.312(5) Å, b ) 18.268(6) Å, c ) 20.551(7)

Å, â ) 92.30(3)°. The cell volume was 4994(3) Å³, with a calculated

density of 1.072 g/cm, where Z ) 4. The space group was determined

to be P2₁/n, and the asymmetric unit consisted of C₄₀H₉₁N₃OSi₃Zr. All

unique diffraction maxima (+h,+k,(l) with 2Θ < 45° were measured

on a Nicolet R3m/V automated diffractometer, by a variable-speed,

2Θ-Θ scan (2.00-29.30°/min in ω) with graphite-monochromated

Mo-KR radiation (λ ) 0.71069 Å) at 25 °C. After correction for

Lorentz, polarization, and background effects, 4675 (71.3%) of the

unique data (3967) were judged observed (|F_o| > 3σ |F_o|).⁹⁹All heavy

atoms were located using direct methods (SHELXTL PLUS), and all

non-hydrogen atoms were revealed by successive Fourier syntheses.

Full-matrix, least-squares refinements (minimization of ∑w(F_o- F_c)²,

would be bonded to the metal in the corresponding zirconium

hydrocarbyl was chosen as the site of protonation (e.g., the benzylic

carbon in toluene). The initial symmetry of the protonated carbon was

assigned as C_2Vfor the benzyls, and C_sfor all other cases.

Acknowledgment. Primary support from the National Sci-

ence Foundation (CHE-9218270) is gratefully acknowledged

as are contributions from the Air Force Office of Scientific

Research (AFOSR-87-0103), Cornell University, and the NSF-

REU Program (CCC). We thank Prof. Barry K. Carpenter, Prof.

Alan S. Goldman, Dr. William Tumas, and Dr. Farad Abu-

Hasanyan for helpful discussions. Dr. Jeffrey B. Bonanno, Dr.

Steven M. Baxter, and Mr. David Fuller are thanked for

experimental assistance and Dr. Gregory D. Van Duyne for aid

in the crystallographic study. We thank Prof. Theodore L.

Brown and Prof. Mary Caffery for sharing the results of

molecular mechanics calculations prior to publication. Support

for the Cornell NMR Facility from the NIH and NSF Instru-

mentation Programs is acknowledged.

where w is based on counting statistics modified by an ignorance factor

-1

(w

) σ²(F) + 0.0015F²)), with anisotropic heavy atoms and all

hydrogens included at calculated positions (Riding model, fixed

isotropic U), converged (3967 reflections) to R ) 7.70% and R_w

)

9.12%, with GOF ) 1.46.¹⁰⁰A final difference Fourier map revealed

no peaks greater than 0.74 e^-/Å³.

Calculations. 1. AM1 Calculations of Proton Affinities. AM1

calculations were performed using the AMPAC-IBM program on the

Cornell National Supercomputer Facility. All bond lengths and angles

were left unconstrained. A charge of +1 was used for the protonated

alkanes, and GEO-OK conditions allowed the inclusion of five-

coordinate carbon. Calculations were performed on both the protonated

alkanes and the corresponding unprotonated alkanes (neopentane,

isobutane, toluene, mesitylene; also, calculations were checked against

literature values for methane, cyclopropane and cyclohexane).^87,88The

heats of formation were then used to determine the proton affinity by

Supporting Information Available: X-ray structural infor-

mation pertaining to (^tBu₃SiNH)₂(THF)ZrdNSi^tBu₃(2-THF):

a summary of crystal data encompassing data collection and

solution/refinement, atomic coordinates, isotropic and aniso-

tropic temperature factors, hydrogen atom coordinates, bond

lengths, and bond angles (9 pages); table of observed and

calculated structure factors (24 pages). This material is

contained in many libraries on microfiche, immediately follows

this article in the microfilm version of the journal, can be ordered

from the ACS, and can be downloaded from the Internet; see

any current masthead page for ordering information and Inernet

access instructions.

PA ) -[∆H_f°(RH₂⁺) - ∆H_f°(RH) - ∆H_f°(H⁺)]

(43)

eq 43. In all cases protonation was site-specific; the carbon which

JA950745I

Article Doi

Hydrocarbon activation via reversible 1,2-RH-elimination from (tBu3SiNH)3ZrR: Synthetic, structural, and mechanistic investigations

DOI: 10.1021/ja950745i

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja950745i

Full text of DOI:10.1021/ja950745i

Products guided by the article

R&D Labs maybe for 173314-59-5

Relevant to this article

Hot Product