Ligand elaboration mediated by a Cp*W(NO) template: Stepwise incorporation of small molecules into a tungsten vinyl fragment

854

Organometallics 1998, 17, 854-871

Liga n d Ela bor a tion Med ia ted by a Cp *W(NO) Tem p la te:

Step w ise In cor p or a tion of Sm a ll Molecu les in to a

Tu n gsten Vin yl F r a gm en t

Peter Legzdins* and Sean A. Lumb

Department of Chemistry, The University of British Columbia,

Vancouver, British Columbia, Canada V6T 1Z1

Victor G. Young, J r.^†

Department of Chemistry, The University of Minnesota, Minneapolis, Minnesota 55455

Received December 9, 1997

Thermolysis of the alkyl vinyl complex Cp*W(NO)(CH₂SiMe₃)(CPhdCH₂) (1) in the

presence of unsaturated, heteroatom-containing compounds such as esters and nitriles

quantitatively affords metallacyclic products of reductive coupling. These are trapped as

18e complexes via either intramolecular rearrangement or intermolecular reaction with added

trapping reagents. The nature of these metallacycles is consistent with the intermediacy of

the acetylene complex Cp*W(NO)(η²-CPhtCH) (A) derived in situ from the reductive

elimination of SiMe₄from 1. With esters ROAc (R ) Me, Et), reductive coupling and C-O

bond cleavage yields the alkoxide-containing oxametallacyclopentadiene complexes Cp*W-

(NO)(η²-OdC(Me)CHdCPh)(OR) (2, R ) Me; 3, R ) Et). Thermolysis of 1 in RCN (R ) Me,

Et, ⁱPr) containing small excess amounts of R′OH yields the respective hydroxide or alkoxide

compounds Cp*W(NO)(η²-NHdC(R)CHdCPh)(OR′) (4, R ) Me, R′ ) H; 5, R ) Et, R′ ) H;

i

6, R ) Pr, R′ ) H; 7, R ) Me, R′ ) C₃H₅). Utilization of cyclopentadiene (CpH) as the

trapping agent in MeCN affords the aminopentafulvene complex Cp*W(NO)(HNC(dC(C₄H₄))-

1

(Me))(η²- NHdC(Me)CHdCPh) (8). VT H NMR spectroscopy reveals the fluxional solution

behavior of the fulvene ligand in 8. Thermolysis of 1 in RCN (R ) Me, ⁱPr) containing trace

amounts of acetone gives the bicyclic species Cp*W(NO)(η³-OC(Me)₂NdC(Me)CHdCPh) (9,

R ) Me; 10, R ) ⁱPr). In the absence of added trapping reagent, thermolysis of 1 in RCN (R

) Me, Et) yields the vinyl amidinate complexes Cp*W(NO)(η³-NHC(R)dNC(dC(R¹)(R²))-

CHdCPh) (11, R ) Me, R¹) R²) H; 12, R ) Et, R¹) H, R²) Me). The molecular structures

proposed for 2, 4, 8, and 10 are confirmed by single-crystal X-ray analyses. Mechanistic

proposals to account for the observed chemistry are corroborated by the results of labeling

studies, while a kinetic study of the transformations yielding 3, 4, and 11 implicates the

rate-limiting generation of acetylene intermediate A. A qualitative orbital overlap rationale

is proposed to account for the observed chemistry.

Sch em e 1

In tr od u ction

Of fundamental importance is the bond-forming pro-

cess involving the coupling of two unsaturated sub-

strates. That a transition metal (TM) environment

possesses the ability to act as a template for such bond

formation represents one of the most significant ap-

plications of modern organotransition metal chemistry.¹

Since the first report of Ni-induced alkyne cyclotrimer-

ization,²TM-induced reductive coupling of unsaturated

organic substrates has provided access to a wealth of

polyfunctional cyclic hydrocarbons and heterocycles.³

Reductive coupling is typically mediated by TM

complexes containing a coordinatively unsaturated, low-

valent (electron-rich) metal center, the former charac-

* To whom correspondence should be addressed.

^†To whom correspondence regarding X-ray crystallographic analyses

should be addressed.

(1) 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. (b) Crabtree, R. H.

The Organometallic Chemistry of the Transition Metals, 2nd ed.; Wiley-

Interscience: Toronto, 1994.

(2) Reppe, W.; Schlichting, O.; Klager, K.; Toepel, T. Liebigs Ann.

Chem. 1948, 560, 1.

(3) This field has been reviewed extensively, see, for example: (a)

Vollhardt, K. P. C. Acc. Chem. Res. 1977, 10, 1. (b) Buchwald, S. L.;

Nielsen, R. B. Chem. Rev. 1988, 88, 1047. (c) Carnahan, E. M.;

Protasiewicz, J . D.; Lippard, S. J . Acc. Chem. Res. 1993, 26, 90. (d)

Broene, R. D.; Buchwald, S. L. Science 1993, 261, 1696. (e) Negishi,

E.-I.; Takahashi, T. Acc. Chem. Res. 1994, 27, 124.

Publication on Web 01/31/1998

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 855

Sch em e 2

teristic being prerequisite for the initial coordination of

unsaturated substrate while the latter is required for

the activation of the coordinated substrate toward

coupling. While many late transition metal complexes

exist in this state and can be manipulated as such,

analogous low-valent, “naked” complexes of the early

transition metals tend to be unstable and must be

nation of alkane dramatically expanded the range of

cyclic and heterocyclic compounds made accessible to

the synthetic chemist.

In an earlier communication describing the double

C-H activation of n-pentane and n-hexane,⁷we pro-

posed the formation of Cp*W(NO)(η²-PhCtCH) (A) by

a â-H reductive elimination of SiMe₄from Cp*W(NO)-

(CH₂SiMe₃)(CPhdCH₂) (1) under thermolysis conditions

in the appropriate hydrocarbon solvent. The structural

similarity of A to the benzyne complex of Erker and

Buchwald prompted us to investigate the reductive

coupling chemistry of A in the presence of such unsat-

urated, heteroatom-containing substrates as esters and

nitriles, with the expectation that the products of

reductive coupling would mirror any orbital differences

between the zerovalent Cp*W(NO) template and the

isolobal divalent zirconocene analogue. A recent com-

munication reported our preliminary results in this

regard.⁸In this paper, we present the full details of

this study, including a kinetic analysis of representative

reactions, and provide mechanistic evidence and a

molecular orbital proposal to account for the chemistry

observed. Unlike the group 4 analogues, the group 6

metallacycles initially formed in the reductive coupling

event are not isolable as stable, 18-valence-electron (18e)

complexes and continue to react, either intramolecularly

or with trapping reagents present in the reaction

mixture, until electronic saturation is attained.

generated in situ from higher-valent precursors.

A

particularly versatile method toward this end involves

the in situ reductive elimination of hydrocarbon from a

parent hydrocarbyl complex (Scheme 1). Because one

of the two unsaturated ligands is generated intra-

molecularly in the reductive elimination step, the pres-

ence of a different unsaturated substrate in the reaction

mixture results in products arising from cross-coupling.

Further variation can be introduced into the overall

reaction by modifying the substituents R′, R′′, and R′′′

in the parent hydrocarbyl complex.

While current interest in this area spans a number

of groups and ligand sets, its foundation is set in the

chemistry developed for the group 4 metallocenes.^3b,d,e

Initial reports by Erker described the thermal elimina-

tion of benzene from Cp₂ZrPh₂and proposed the forma-

tion of a transient benzyne complex which is capable of

coupling to olefins and acetylenes, yielding a plethora

of functionalized aliphatic zirconacycles.⁴Buchwald

subsequently structurally characterized the phosphine-

trapped benzyne complex by X-ray diffraction and

extended the reactivity to heteroatom-containing sub-

strates.⁵The method was then generalized so that a

variety of alkyne complexes could be obtained through

â-H reductive elimination of methane from the ap-

propriate methyl alkenyl precursor.⁶The development

of this particular methodology involving reductive elimi-

Resu lts a n d Discu ssion

The thermolysis of Cp*W(NO)(CH₂SiMe₃)(CPhdCH₂)

(1) in the presence of several substituted esters and

(6) (a) Buchwald, S. L.; Lum, R. T.; Dewan, J . C. J . Am. Chem. Soc.

1986, 108, 7441. (b) Buchwald, S. L.; Watson, B. T.; Huffman, J . C. J .

Am. Chem. Soc. 1987, 109, 2544.

(7) Debad, J . D.; Legzdins, P.; Lumb, S. A.; Batchelor, R. J .; Einstein,

F. W. B. J . Am. Chem. Soc. 1995, 117, 3288.

(4) (a) Erker, G. J . Organomet. Chem. 1977, 134, 189. (b) Erker,

G.; Kropp, K. J . Am. Chem. Soc. 1979, 101, 3659.

(5) Buchwald, S. L.; Watson, B. T.; Huffman, J . C. J . Am. Chem.

Soc. 1986, 108, 7411.

(8) Legzdins, P.; Lumb, S. A. Organometallics 1997, 16, 1825.

856 Organometallics, Vol. 17, No. 5, 1998

Legzdins et al.

Ta ble 1. Nu m ber in g Sch em e, Yield , a n d An a lytica l Da ta for Com p lexes 2-12

anal. found (calcd)

complex

compd no.

color (yield, %)

C

H

N

Cp*W(NO)(OMe)(η²-OdC(Me)CHdCPh)

Cp*W(NO)(OEt)(η²-OdC(Me)CHdCPh)

Cp*W(NO)(OH)(η²-HNdC(Me)CHdCPh)‚0.5MeCN

Cp*W(NO)(OD)(η²-DNdC(Me)CHdCPh)

Cp*W(NO)(OH)(η²-HNdC(Et)CHdCPh)

Cp*W(NO)(OH)(η²-HOdC(ⁱPr)CHdCPh)

Cp*W(NO)(OCh₂CHdCH₂)(η²-NHdC(Me)CHdCPh)

Cp*W(NO)(HNC(dC(C₄H₄))(Me))(η²-HNdC(Me)-

N(Me)CHdCPh)‚MeCN

2

3

purple (89)

purple (83)

orange (84)

47.96 (47.99)

48.77 (49.00)

47.52 (47.15)

b

44.64 (48.99)

48.39 (48.11)

47.52 (50.20)

54.63 (54.39)

5.09 (5.18)

5.35 (5.42)

5.27 (5.22)

2.50 (2.67)

2.47 (2.60)

7.04 (6.60)

4‚0.5MeCN

4-d₂

5

6

orange (>95)^a

red-orange (92)

red-orange (83)

red-orange (81)

deep red (75)

5.15 (5.79)

5.57 (5.38)

5.00 (5.49)

5.75 (5.66)

5.12 (5.19)

5.49 (5.34)

4.65 (5.09)

8.82 (8.75)

7

8‚MeCN

Cp*W(NO)(η³-OC(Me)₂NdC(Me)CHdCPh)

Cp*W(NO)(η³-OC(Me)₂NdC(ⁱPr)CHdCPh)

Cp*W(NO)(η³-HNC(Me)dNC(dCH₂)CHdCPh)

Cp*W(NO)(η³-DNC(CD₃)dNC(dCD₂)CHdCPh)

Cp*W(NO)(η³-HNC(Et)dNC(dCHMe)CHdCPh)

9

10

11

11-d₆

12

red-brown (72)

red-brown (71)

dark red (77)

dark red (95)^a

red-brown (68)

b

52.56 (52.92)

48.90 (49.52)

b

5.94 (5.92)

5.09 (5.10)

5.01 (4.84)

6.69 (7.88)

51.59 (51.32)

5.53 (5.57)

7.55 (7.49)

a

b

Yield as indicated in the ¹H NMR spectrum of the reaction mixture. Not determined.

Ta ble 2. Ma ss Sp ectr oscop ic Da ta a n d IR Sp ectr a l

Da ta for Com p lexes 2-12

compd no.

MS (m/ z)^a

probe temp^b(°C) IR (Nujol, cm^-1

)

2

3

4

539

525

510

200

150

1532 (ν_NO

1538 (ν_NO

3562 (ν_OH

1532 (ν_NO

2630 (ν_OD

2282 (ν_ND

1531 (ν_NO

3580 (ν_OH

)

4-d₂

512

120

150

5

494 [P⁺- NO]

periods up to 1 week. The Nujol-mull IR bands at-

tributable to the NO stretch in 2 and 3 are located at

1532 and 1538 cm^-1, typical of 18e alkoxide-containing

1587 (ν_CdN

1535 (ν_NO

1733 (ν_CdN

1588 (ν_NO

1730 (ν_CdN

1589 (ν_NO

1730 (ν_CdN

1589 (ν_NO

)

6

7

8

508 [P⁺- NO]

150

1

complexes.⁹The H and ¹³C NMR signals displayed by

)

complexes 2 and 3 are readily assigned. For example,

the resonances of the acyl methyl substituents in the

¹H NMR spectra are singlets that integrate for 3 protons

at about δ 2.32 and 2.44 for 2 and 3, respectively, and

the signals for the analogous carbon nuclei are identifi-

able in the ¹³C NMR spectra at about 26 ppm for both

complexes. Likewise, the vinyl proton resonance is

evident as a singlet at 7.18 ppm for complex 2 and 7.37

ppm for complex 3. The assignment of the carbon

backbone containing the PhCdCH unit in these com-

plexes is deduced from the combination of the gate-

decoupled ¹³C NMR results and from HMBC experi-

ments. On the other hand, NOE and HMBC data do

not distinguish between the structure depicted in eq 1

for 2 and 3 and the expected oxametallacyclopentene

ring structure erroneously reported⁸to result from the

simple reductive coupling of carbonyl and acetylene

moieties in the metal’s coordination sphere.

The four-legged piano-stool arrangement of ligands

in the solid-state molecular structure of 2 has been

definitively established by an X-ray crystallographic

analysis (Figure 1). Of particular note is the existence

of the tungsten-bound methoxy group (W-O bond

length of 2.001(7) Å) that confirms a C-O bond-cleavage

process, in addition to reductive coupling, affords the

observed 18e products. The σ-bound methoxy and alkyl

ligands are trans-disposed in the metal’s coordination

sphere, and the acyl function is coordinated in a cis

fashion to both σ-ligands with a W-O dative bond

length of 2.203(6) Å. Interestingly, this distance is

shorter than the 2.535(7) Å Ni-O dative bond length

in the late transition metal complex Ni(PMe₃)₂(Cl)(η²-

550

599

)

9

10

550^c

578

150

1598 (ν_CdN

)

1542 (ν_NO

3268 (ν_NH

)

e

11

533^d

80

1590 (ν_CdC

)

1517 (ν_NO

)

11-d₆

12

539

561

150

n/a^f

1598 (ν_CdN

1529 (ν_NO

)

a

Values for the highest intensity peak of the calculated isotopic

cluster (¹⁸⁴W). Probe temperatures. ^cHigh-resolution EI mass

b

spectrum (150 °C), found (calcd): 550.18242 (550.18170), [C₂₃H₃₀

-

N₂O₂¹⁸⁴W, P⁺]. High-resolution EI mass spectrum (150 °C), found

(calcd): 553.16608 (533.16638), [C₂₂H₂₇N₃O¹⁸⁴W, P⁺]; 503.16662

(503.16837), [P⁺- NO]. ^eKBr pellet. ^fNMR-tube reaction.

d

nitriles under differing experimental conditions affords

the product complexes depicted in Scheme 2. Each

reaction portrayed is considered in turn in the following

sections. The numbering scheme and analytical data

for complexes 2-12 are collected in Table 1. Mass and

IR spectroscopic data are presented in Table 2. ¹H and

¹³C NMR data for selected compounds are summarized

in Table 3, while X-ray crystallographic data for com-

plexes 2, 4, 8, and 10 are given in Table 4.

Red u ctive Cou p lin g of Ester s: Acyl C-O Bon d

Clea va ge. Thermolysis of the red tungsten vinyl

complex 1 at 45 °C in the presence of esters such as

methyl or ethyl acetate affords the oxametallacyclopen-

tadienyl complexes Cp*W(NO)(OMe)(η²-OdC(Me)CHdC-

Ph) (2) and Cp*W(NO)(OEt)(η²-OdC(Me)CHdCPh) (3),

respectively, in high yield and purity (eq 1). These four-

legged, piano-stool complexes are burgundy crystalline

solids that are air-stable at ambient temperatures for

(9) Legzdins, P.; Lundmark, P. J .; Rettig, S. J . Organometallics 1993,

12, 3545.

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 857

Ta ble 3. ¹H a n d ¹³C NMR Sp ectr oscop ic Da ta for Com p lexes 2, 4, 4-d ₂, 8, 9, 11, a n d 11-d ₆

compd no.

¹H NMR^aδ (ppm)

¹³C NMR^aδ (ppm)

251.3 (OdCMe)

204.8 (CPhdCH)

150.1 (Ph i-C)

136.2 (CPhdCH)

128.5 (Ph)

3

2

7.60 (d, 2H, J _HH) 7.2 Hz, Ph o-H)

7.37 (s, 1H, CPhdCH)

3

7.31 (vt, 2H, J _HH) 7.2 Hz, Ph m-H)

3

7.31 (vt, 1H, J _HH) 7.2 Hz, Ph p-H)

4.42 (s, 3H, MeO)

2.44 (s, 3H, MeCdO)

1.72 (s, 15H, C₅Me₅)

127.9 (Ph)

127.0 (Ph)

112.7 (C₅Me₅)

62.6 (OMe)

26.3 (MeCdO)

9.2 (C₅Me₅)

4

8.69 (br s, 1H, NH)

229.5 (s, CPhdCH)

182.6 (s, HNdCMe)

150.9 (s, Ph i-C)

3

7.55 (d, 2H, J _HH) 6.9 Hz, Ph o-H)

3

7.27 (vt, 2H, J _HH) 7.2 Hz, Ph m-H)

3

1

7.18 (vt, 1H, J _HH) 6.9 Hz, Ph p-H)

135.2 (d, J _CH) 155.8 Hz, CPhdCH)

4

1

7.06 (d, 1H, J _HH) 3.5 Hz, CPhdCH)

127.8 (d, J _CH) 156.7 Hz, Ph)

1

2.25 (s, 3H, HNdCMe)

1.67 (s, 15H, C₅Me₅)

0.86 (br s, 1H, OH)

127.7 (d, J _CH) 156.7 Hz, Ph)

1

127.3 (d, J _CH) 158.8 Hz, Ph)

112.0 (s, C₅Me₅)

1

23.6 (q, J _CH) 132.0 Hz, HNdC(Me))

1

9.7 (q, J _CH) 127.4 Hz, C₅Me₅)

3

4-d₂

7.55 (d, 2H, J _HH) 6.9 Hz, Ph o-H)

b

3

7.27 (vt, 2H, J _HH) 7.2 Hz, Ph m-H)

7.18 (vt, 1H, J _HH) 6.9 Hz, Ph p-H)

3

7.06 (s, 1H, CPhdCH)

2.25 (s, 3H, NDdCMe)

1.67 (s, 15H, C₅Me₅)

8

8.11 (br s, 1H, HNdCMe)

228.5 (s, CPhdCH)

185.2 (s, HNdCMe)

175.9 (s, HN-C(d(C₄H₄))Me)

150.1 (s, Ph i-C)

3

7.53 (d, 2H, J _HH) 8.1 Hz, Ph o-H)

3

7.24 (vt, 2H, J _HH) 8.1 Hz, Ph m-H)

3

7.24 (vt, 1H, J _HH) 8.1 Hz, Ph p-H)

4

1

7.06 (d, 1H, J _HH) 3.0 Hz, CPhdCH)

135.0 (d, J _CH) 158.5 Hz, CPhdCH)

1

6.52 (br s, 1H, HNC(dC(C₄H₄))Me)

6.50 (br s, 1H, HNC(dC(C₂H_aH_bC₂H₂))Me)

6.44 (br s, 1H, HNC(dC(C₂H_aH_bC₂H₂))Me)

6.23 (br s, 1H, HNC(dC(C₂H₂C₂H_aH_b))Me)

6.10 (br s, 1H, HNC(dC(C₂H₂C₂H_aH_b))Me)

2.34 (s, 3H, HNdCMe)

127.9 (d, J _CH) 159.2 Hz, Ph)

1

127.2 (d, J _CH) 159.0 Hz, Ph)

1

127.1 (d, J _CH) 159.0 Hz, Ph)

121.5 (br s, HNC(dC(C₄H₄))Me)

117.8 (br s, HNC(dC(C₄H₄))Me)

115.6 (br s, HNC(dC(C₄H₄))Me)

115.0 (br, s, HNC(dC(C₄H₄))Me)

111.7 (s, C₅Me₅)

2.25 (s, 3H, HNC(dC(C₄H₄))Me)

1.70 (s, 15H, C₅Me₅)

107.5 (br s, HNC(dC(C₄H₄))Me)

1

23.7 (q, J _CH) 127.9 Hz, NdCMe)

1

21.2 (q, J _CH) 128.3 Hz, HN-C(fulv)Me)

1

9.7 (q, J _CH) 128.0 Hz, C₅Me₅)

3

9

7.26 (d, 2H, J _HH) 6.9 Hz, Ph o-H)

226.2 (s, CPhdCH)

170.7 (s, NdCMe)

150.3 (s, Ph i-C)

3

7.14 (t, 2H, J _HH) 6.9 Hz, Ph m-H)

3

7.08 (t, 1H, J _HH) 6.9 Hz, Ph p-H)

3

1

6.59 (s, 1H, J _WH) 14 Hz, CPhdCH)

134.5 (d, J _CH) 157.5 Hz, CPhdCH)

1

1.95 (s, 3H, HNdCMe)

1.51 (s, 15H, C₅Me₅)

1.35 (s, 3H, OCMe₂)

1.20 (s, 3H, OCMe₂)

127.5 (d, J _CH) 158.0 Hz, Ph)

1

126.8 (d, J _CH) 158.0 Hz, Ph)

1

126.6 (d, J _CH) 158.0 Hz, Ph)

113.0 (s, C₅Me₅)

99.2 (s, (OCMe₂)

1

30.1 (q, J _CH) 127.5 Hz, OCMe₂)

1

29.0 (q, J _CH) 127.5 Hz, OCMe₂)

1

18.3 (q, J _CH) 127.5 Hz, HNdCMe)

1

9.9 (q, J _CH) 127.5 Hz, C₅Me₅)

3

11

7.43 (d, 2H, J _HH) 8.1 Hz, Ph o-H)

191.8 (CPhdCH)

171.1 (NHC(Me)dN)

156.1 (NCdCH₂)

150.9(Ph i-C)

3

7.26 (vt, 2H, J _HH) 8.4 Hz, Ph m-H)

3

7.11 (vt, 1H, J _HH) 7.5 Hz, Ph p-H)

6.83 (s, 1H, CPhdCH)

4.73 (br s, 1H, HNC(Me)d)

4.45 (s, 1H, NCdCH_aH_b)

4.28 (s, 1H, NCdCH_aH_b)

2.11 (s, 3H, HNC(Me)d)

1.73 (s, 15H, C₅Me₅)

1

138.8 (d, J _CH) 127.7 Hz, CPhdCH)

127.7 (Ph)

125.4 (Ph)

111.8 (C₅Me₅)

90.6 (NCdCH₂, J _CH) 152 Hz)

22.8 (HNC(Me)d)

1

9.9 (C₅Me₅)

b

3

11-d₆

7.43 (d, 2H, J _HH) 8.1 Hz, Ph o-H)

3

7.26 (vt, 2H, J _HH) 8.4 Hz, Ph m-H)

7.11 (vt, 1H, J _HH) 7.5 Hz, Ph p-H)

3

6.83 (s, 1H, CPhdCH)

2.11 (s, 3H, HNC(Me)d)

1.73 (s, 15H, C₅Me₅)

a

b

Sample spectra recorded in CDCl₃unless otherwise noted. Spectrum not recorded.

Ta ble 4. X-r a y Cr ysta llogr a p h ic Da ta for Com p lexes 2, a n d 4, 8, a n d 10

2

4

8

10

Crystal Data

empirical formula

cryst habit, color

cryst size, mm

cryst syst

C

₂₁H₂₇NO₃W

C₂₂H₂₉N₃O₂W

C₃₁H₃₉N₅OW

C₂₅H₃₄N₂O₂W

block, red

0.28 × 0.24 × 0.19

monoclinic

P2₁/c

2341.90(4)

10.5087(1)

15.4352(2)

15.1367(1)

94.787(1)

4

long plate, red

prism, orange

0.50 × 0.18 × 0.18

monoclinic

P2₁/n

2159.21(9)

8.4729(2)

21.2313(5)

12.1060(3)

97.483(1)

4

irregular block, red

0.28 × 0.19 × 0.08

monoclinic

P2₁

1485.68(3)

7.2125(1)

18.4419(1)

11.3542(2)

100.348(1)

2

0.32 × 0.095 × 0.012

monoclinic

P2₁/c

1964.2(2)

7.2421(3)

15.3723(7)

17.8615(8)

98.962(1)

4

space group

vol, Å³

a, Å

b, Å

c, Å

â, deg

z

fw

525.29

1.776

5.901

1032

551.33

1.696

5.371

1088

681.52

1.523

3.919

684

578.39

1.640

4.955

1152

density (calcd, mg/m³)

abs coeff, mm^-1

F(000)

Data Collection

Siemens SMART

Platform CCD

0.710 73

diffractometer

Siemens SMART

Platform CCD

0.710 73

Siemens SMART

Platform CCD

0.710 73

Siemens SMART

Platform CCD

0.710 73

wavelength, Å

temp, K

173(2)

θ range for data collection, deg

1.76-25.10

1.92-25.03

1.82-25.01

1.89-25.03

index ranges

-8 < h < 8, -18 < k < 18,

-20 < l < 21

9664

-10 < h < 9, -25 < k < 25,

-14 < l < 10

-8 < h < 8, -21 < k < 21,

0 < l < 13

-11 < h < 11, 0 < k < 18,

0 < l < 17

no. of reflns collected

no. of indep reflns

11 123

3800 (R_int) 0.0378)

7150

11 298

4064 (R_int) 0.0288)

3457 (R_int) 0.0522)

4957 (R_int) 0.0199)

Solution and Refinement

SHELXTL-V5.0

system used

SHELXTL-V5.0

solution

refinement method

weighting scheme

direct methods

full-matrix least-squares on F²

w ) 2(F_o²) + (AP)²+ (BP)]^-1

,

w ) 2(F_o²) + (AP)²+ (BP)]^-1

,

w ) 2(F_o²) + (AP)²+ (BP)]^-1

,

w ) 2(F_o²) + (AP)²+ (BP)]^-1

,

where P ) (F_o+ 2F_c²))/3,

2

A ) 0.0546, and B ) 0.0

SADABS (Sheldrick, 1996)

1.000 and 0.572

A ) 0.0139, and B ) 3.9846

A ) 0.0384, and B ) 0.789

SADABS (Sheldrick, 1996)

1 and 0.835

A ) 0.0136, and B ) 1.1696

SADABS (Sheldrick, 1996)

1 and 0.784

abs corr

semiempirical

max and min transmission

data/restraints/params

final R indices [I > 2σ(I)]

R indices (all data)

0.66286 and 0.39415

3800/0/288

R₁) 0.0223, wR₂) 0.0489

R₁) 0.0271, wR₂) 0.0512

1.125

3455/0/243

4955/1/343

4064/0/280

R₁) 0.0516, wR₂) 0.1031

R₁) 0.0791, wR₂) 0.1117

1.059

R₁) 0.0284, wR₂) 0.0675

R₁) 0.0326, wR₂) 0.0697

1.04

R₁) 0.0277, wR₂) 0.0479

R₁) 0.0401, wR₂) 0.0506

1.072

goodness-of-fit on F²

largest diff. peak and hole, e Å^-3

1.794 and -2.329

0.576 and -0.730

0.616 and -0.729

0.452 and -0.542

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 859

Sch em e 4

F igu r e 1. Thermal ellipsoid plot of compound 2. Thermal

ellipsoids depict the 50% probability level.

C by the reductive coupling of the two coordinated

ligands. While the analogous Cp₂Zr oxametallacyclo-

pentene complexes are readily isolable and have been

studied extensively,^3b,dthe Cp*W(NO)-containing in-

termediate C cannot be isolated as such. Indeed,

though the metal-bound acetal O has the capacity to

function as a 3e donor to the metal center in the correct

orientation¹⁴and as such would render the complex

electronically saturated, space-filling models reveal that

the restrictive geometry of the five-membered ring

enforces a conformation that leaves the filled O-centered

orbitals pointed away from the metal center. The O

atom, thus, functions predominantly as a 1e donor,

resulting in a build-up of electron density at the het-

eroatom and leaving the metal center in C electronically

unsaturated. Subsequent acetal C-O-bond cleavage

yields the observed 18e complexes.

Sch em e 3

Only one of two possible routes to the observed

products is presented in Scheme 3. An alternate path

involving direct acyl C-O oxidative addition to the

metal center in A followed by acyl/acetylene coupling

is conceivable, yet such examples involving alkyl car-

boxylate esters are rare and reports of this mode of

reactivity seem limited to low-valent late transition

metal systems.¹⁵On the other hand, coordination of an

ester is a prerequisite to reductive coupling, and our

recent observation of η²-ester coordination in the related

Cp*W(NO)(PPh₃) fragment¹⁶supports a pathway in-

volving the intermediacy of B.

Acetal â-C-O bond cleavage by a TM complex is an

area of considerable current interest, probably the most

well-known example being the Kulinkovich reaction

which generates cyclopropanols following reductive

coupling of the coordinated olefin and ester at Ti(II)

centers (Scheme 4).¹⁷While the coupling leading to

complexes 2 and 3 is reminiscent of the Kulinkovich

reaction, the products of the two reactions are quite

different. Displacement of RO^-in the putative acetal

intermediate C by nucleophilic attack of the quaternary

vinyl C at the acetal carbon center to give the O-bound

OdC(CH₂SiMe₃)CdCPh)¹⁰as well as the 2.361(3) Å

Zr-O bond length previously reported by Rosenthal for

related zirconacyclic complexes¹¹yet longer than the

Re-O bond length of 2.099 Å in the isoelectronic [CpRe-

(NO)(PPh₃)(η¹-OdCMePh)]⁺cation.¹²The CdO bond

length of 1.277(12) Å is longer than a typical C-O

double bond and is characteristic of a coordinated acyl

function.¹³The oxametallacyclopentadiene ring is clearly

evident from the short-long-short pattern of bond

lengths for the O(1)dC(12)-C(13)dC(14) fragment

(1.277(12), 1.413(14), and 1.337(12) Å, respectively),

comparable to those found in the structure of the above-

mentioned Ni (II) complex.

The mechanism that we propose for the overall

process leading to 2 and 3 is depicted in Scheme 3.

Intermediate B, formed by coordination of an ester to

W in A, is transformed into the oxametallacyclopentene

(14) (a) Ashby, M. T.; Enemark, J . H. J . Am. Chem. Soc. 1986, 108,

730. (b) Hubbard, J . L.; McVicar, W. K. Inorg. Chem. 1992, 31, 910.

(15) For

a review of this chemistry, see: Yamamoto, A. Adv.

Organomet. Chem. 1992, 34, 111.

(16) Burkey, D. J .; Debad, J . D.; Legzdins, P. J . Am. Chem. Soc.

1997, 119, 1139.

(10) Carmona, E.; Gutie´rrez-Puebla, E.; Monge, A.; Mar´ın, J . M.;

Paneque, M.; Poveda, M. Organometallics 1989, 8, 967.

(11) Peulecke, N.; Ohff, A.; Tillack, A.; Baumann, W.; Kempe, R.;

Burlakov, V. V.; Rosenthal, U. Organometallics 1996, 15, 1340.

(12) Dalton, D. M.; Ferna´ndez, J . M.; Emerson, K.; Larsen, R. D.;

Arif, A. M.; Gladysz, J . A. J . Am. Chem. Soc. 1990, 112, 9198.

(13) Debad, J . D.; Legzdins, P. Organometallics 1993, 12, 2094.

(17) (a) Lee, J .; Kim, H.; Cha, J . K. J . Am. Chem. Soc. 1996, 118,

4198. (b) Lee, J .; Kim, H.; Cha, J . K. J . Am. Chem. Soc. 1996, 118,

291. (c) Corey, E. J .; Rao, A. R.; Noe, M. C. J . Am. Chem. Soc. 1994,

116, 9345. (d) Kulinkovich, O. G.; Sviridov, S. V.; Vasilevskii, D. A.

Synthesis 1991, 3, 234.

860 Organometallics, Vol. 17, No. 5, 1998

Legzdins et al.

cyclopropenyl ligand (as proposed to account for the

products in the Kulinkovich reaction) is not observed,

presumably a result of the reduced oxophilicity of W(II)

relative to Ti(IV) and the greater ring strain in cyclo-

propene relative to cyclopropane.¹⁸Other examples of

acetal â-C-O bond cleavage have been described in the

literature. Yamamoto has reported the generation of

acetal complexes by the addition of a M-H bond (M d

Rh,^19aCo^19b) across the carbonyl function of the coor-

dinated ester and subsequent â-C-O bond cleavage to

produce the analogous aldehydes. Likewise, Grotjahn

has recently invoked acetal C-O bond cleavage to

account for the thermal generation of an alkoxycarbene

complex from a CpRu acetal complex.²⁰Given these

literature precedents, the pathway outlined in Scheme

3 appears to be the more reasonable one. While num-

erous examples of an ester coupling to an olefin exist,

the reductive coupling of an ester to an acetylene in the

coordination sphere of a metal appears to be unprec-

edented. Because the overall transformation involves

the formal exchange of vinyl for alkoxide at the acyl

carbon, this reaction can be viewed as the conversion

of a carboxylate ester to an R,â-unsaturated ketone via

C-O bond cleavage.²¹Variation of the ester substitu-

ents and the precursor vinyl substituents, thus, offers

potential access to a range of substituted vinyl ketones.

Red u ctive Cou p lin g w ith Nitr iles. Simple reduc-

tive coupling of acetylene and coordinated nitrile in the

tungsten’s coordination sphere does not yield the ex-

pected azametallacycle, Cp*W(NO)(η²-NdC(R)CHdCPh),

following the generation of the acetylene intermediate

A in nitrile solvent. Instead, analogous to the ther-

molysis of 1 in esters, the products obtained result from

the subsequent reactivity of the intermediate azamet-

allacyclopentadiene complex obtained through initial

reductive coupling. Unlike the acetal intermediate C,

however, this aza complex does not achieve electronic

saturation by rearrangement. Rather, trapping of this

intermediate by a variety of electrophilic sources occurs.

A. Tr a p p in g by P r otic Sou r ces. Thermolysis of

1 in nitrile-containing 0.1-0.01 M H₂O or allyl alcohol

affords complexes 4-7 quantitatively (eq 2). These

F igu r e 2. Thermal ellipsoid plot of complex 4. Thermal

ellipsoids depict the 50% probability level.

stable for periods of up to 2 weeks at ambient temper-

atures, thereby attesting to the existence of an 18e,

closed-shell configuration at the metal center. The

solid-state molecular structure of prototypal complex 4

is shown in Figure 2. Analogous to the structure of

complex 2, imine N coordination to W diagonal to the

NO ligand in these four-legged piano-stool complexes

confers electronic saturation at the metal center. Al-

though the hydroxyl H was not located during the

structure refinement, the W-O bond length of 2.064(2)

Å is characteristic of a tungsten-oxygen single bond.²²

Hence, protonation at the coordinated imine must be

invoked to support the hybridization at N and to

preserve the electroneutrality of the complex.

The solid-state structural features highlighted in the

preceding paragraph are also evident in the spectro-

scopic data for complex 4 (Table 3). For example,

signals for both the N-H and O-H protons are visible

1

in the H NMR spectra (CDCl₃) of complexes 4-6, the

former appearing at approximately 8.65 ( 1.0 ppm and

the latter in the range 0.84 ( 0.2 ppm. In addition, the

singlet attributable to the vinyl H in all cases appears

at about 7.1 ppm (¹J _CH≈ 157 Hz). Signals in the ¹³C-

{¹H} NMR spectrum at about 228 and 189 ( 7 ppm

divulge the presence of the carbenoid C_Ratom of the

vinyl moiety and the quaternary imine carbon, respec-

tively. The downfield shift of the resonance due to the

carbenoid C_Ratom suggests a degree of electronic

delocalization in the azametallacyclopentadienyl ring.²³

Interestingly, only the O-H stretch in the Nujol-mull

IR spectrum of 4 is readily observable, the N-H stretch

being masked by the broad unsaturated C-H absorp-

tions in the range 3000-3200 cm^-1(Table 2). However,

the isotopic shift of these absorptions in the IR spectrum

can be examined for the complex in which a deuterium

label is incorporated at the hydroxyl and imido sites (eq

3). The Nujol-mull IR spectrum of 4-d₂reveals a band

at 2282 cm^-1which can be attributed to the N-D

stretch. Calculation of the expected N-H stretch for

complex 4 based on the Redlich-Teller rule²⁴gives a

value of ν_N-Hfor 4 of 3195 cm^-1, in a region obscured

products are orange-red crystalline solids that are air-

(18) A recent theoretical analysis describes ring strain and bond

enthalpies as the major contributors to the stability of methylcyclo-

propane relative to methylcyclopropene, see: J ohnson, W. T. G.;

Borden, W. T. J . Am. Chem. Soc. 1997, 119, 5930.

(19) (a) Yamamoto, T.; Miyashita, S.; Naito, Y.; Komiya, S.; Ito, T.;

Yamamoto, A. Organometallics 1982, 1, 808. (b) Hayashi, Y.; Yama-

moto, T.; Yamamoto, A.; Komiya, S.; Ito T.; Kushi, Y. J . Am. Chem.

Soc. 1986, 108, 385.

(20) Grotjahn, D. B.; Lo, H. C. Organometallics 1996, 15, 2860.

(21) For other examples of ketone-stabilized vinyl complexes, see:

(a) Alt, H. G.; Eichner, M. E.; J ansen, B. M. Angew. Chem., Int. Ed.

Engl. 1982, 21, 861. (b) Alt, H. G.; Hayen, H. I. J . Organomet. Chem.

1986, 103, 1501. (c) Bianchini, C.; Innocenti, P.; Melli, A.; Sabat, M.

Organometallics 1986, 5, 72. (d) Reference 10.

(22) The complex [Mo(O)(OH)(CN)₄]^3-has ModO ) 1.697(7) Å and

Mo-OH ) 2.077(7) Å; see: Robinson, P. R.; Schlemper, E. O.;

Murmann, R. K. Inorg. Chem. 1975, 14, 2035. Typical terminal W-oxo

bond distances lie in the range 1.68-1.72 Å, see: Nugent, W. A.; Mayer,

J . M. Metal Ligand Multiple Bonds; Wiley: New York, 1988.

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 861

Sch em e 5

by the bands resulting from the vinylic and aromatic

C-H stretches. Prediction of the O-D stretch expected

for 4-d₂employing the same calculation based on the

observed ν_O-Hfor 4 yields a value of 2544 cm^-1, which

is in reasonable agreement with the experimentally

observed value of 2630 cm^-1

.

Other protic sources with pK_avalues comparable to

that of water can also be employed as trapping reagents

in this reaction. For example, thermolysis of 1 in

acetonitrile containing small amounts of allyl alcohol

results in the formation of the alkoxide complex 7, which

can be isolated in good yield as air-stable red crystals.

positions occurs during the thermolysis of hydroxide 4

in THF-d₈containing D₂O (vide supra), thereby imply-

ing that these two protons are labile. In addition, the

thermolysis of allyl 7 in wet THF-d₈results in its

conversion to complex 4, at a rate significantly slower

than that observed for label exchange. Taken together,

these results indicate that the two processes are inde-

pendent. In other words, while the hydroxyl and amine

protons are labile and readily exchange in the presence

of D₂O, the formation of alkoxide complexes 4-7 by the

addition of R′OH is slow by comparison, with inter-

conversion probably being governed by the pK_aof the

two protic sources.²⁷Because E is a high-energy

intermediate, it is uncertain whether the hydration step

can be regarded as reversible. On the other hand, the

formation of a strong C-C single bond and the conjuga-

tion within the ligand backbone strongly suggests that

the D f E ring closure is irreversible. It is also

1

The N-H proton resonance in the H NMR spectrum

of allyl 7 is a singlet at 8.62 ppm, analogous to that

exhibited by the hydroxyl complex 4. Signals attribut-

able to the allylic moiety of the alkoxide ligand are

distinguishable as a series of multiplets in the charac-

teristic range of δ 4.5-6.2.²⁵The remaining spectro-

scopic data are consistent with the formulated structure

and are readily assigned by comparison to the data for

complexes 4-6.

The proposed mechanism to account for this reactivity

is shown in Scheme 5. Following the initial generation

i

of A, coordination of RCtN (R ) Me, Et, Pr) and

reductive coupling of acetylene and nitrile in the metal’s

coordination sphere give the azametallacyclopentadienyl

intermediate E. The geometry of the resulting five-

membered ring enforces an imide conformation that

severely restricts the donation of the nitrogen atom’s

sp²lone-pair electron density to the metal. Hence, the

imide N only functions as a 1e donor, and the result is

a 16e, Lewis acidic complex. Analogous to intermediate

C, the build-up of lone pair electron density on the imido

N in E renders it susceptible to protonation by the

external protic source, R′OH. The hydration of E and

coordination of the amido N in a manner analogous to

the acyl O in C thus affords the observed 18e products.

Similar protonated or alkylated azametallacyclopenta-

dienyl complexes of Ti,^26aNb,^26bTa,^26cW^26d,eand Ir^26f

have been synthesized via analogous or other prepara-

tive routes.

t

pertinent that the analogous thermolysis in wet BuCN

does not afford any tractable products. While an η¹

t

interaction of the bulkier BuCN with the metal center

can presumably occur, unfavorable steric interactions

probably slow either the formation of intermediate D

or the D f E ring-closing step to the extent that

decomposition pathways, thus, become competitive and

no product is observed. Interestingly, the coupling of

t

two BuCtCH units on the (RO)₃Ta fragment (RO )

2,6-diisopropylphenoxide) reported by Wigley is facile

at room temperature.²⁸Even more striking is the fact

that under thermolysis conditions, ^tBuCN couples readily

to coordinated benzyne in the Buchwald Cp₂Zr system.²⁹

It would, thus, appear that the substantial size of the

Cp* and phenyl acetylene ligands in our complexes is

playing a significant role in the chemo- and regioselec-

tivity leading to complexes 4-7.

Mechanistic studies have afforded evidence in support

of the reaction pathway proposed in Scheme 4. Incor-

poration of deuterium at the amine and hydroxyl

(23) Aromaticity has been invoked for similar aza- and thiapenta-

dienyl iridium systems, see: (a) Alvarado, Y.; Daff, P. J .; Pe´rez, P. J .;

Poveda, M. L.; Sa´nchez-Delgado, R.; Carmona, E. Organometallics

1996, 15, 2192. (b) Bleeke, J . R.; Ontwerth, M. F.; Rohde, A. M.

Organometallics 1995, 14, 2813.

(24) Ebsworth, E. A. V.; Rankin, D. W. H.; Cradock, S. Structural

Methods in Inorganic Chemistry; Blackwell Scientific Publications:

London, 1987; pp 217-218.

(25) Debad, J . D.; Legzdins, P.; Lumb, S. A. Organometallics 1995,

14, 2543.

In light of the mechanism invoked to account for the

formation of complexes 4-7 (vide supra), next we chose

cyclopentadiene (CpH) as an atypical protic source to

probe the extent of this reactivity. In doing so we

discovered the unexpected result that under these

(26) (a) Cohen, S. A.; Bercaw, J . E. Organometallics 1985, 4, 1006.

(b) Lorente, P.; Carfagna, C.; Etienne, M.; Donnadieu, B. Organome-

tallics 1996, 15, 1090. (c) Strickler, J . R.; Wigley, D. E. Organometallics

1990, 9, 1665. (d) Filippou, A. C.; Vo¨lkl, C.; Kiprof, P. J . Organomet.

Chem. 1991, 415, 375. (e) Filippou, A. C.; Vo¨lkl, C.; Rogers, R. D. J .

Organomet. Chem. 1993, 463, 135. (f) Reference 23a.

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

Organic Chemistry, 2nd ed.; Harper & Row: New York, 1981.

(28) Smith, D. P.; Strickler, J . R.; Gray, S. D.; Bruck, M. A.; Holmes,

R. S.; Wigley, D. W. Organometallics 1992, 11, 1275.

(29) Buchwald, S. L.; Watson, B. T.; Lum, R. T.; Nugent, W. A. J .

Am. Chem. Soc. 1987, 109, 7137.

862 Organometallics, Vol. 17, No. 5, 1998

Legzdins et al.

F igu r e 4. Resonance contributors to the electronic struc-

ture of dimethylaminopentafulvene.

Examination of the solution behavior of 8 by ¹H NMR

spectroscopy reveals the intriguing fluxional nature of

the aminofulvene ligand, further evidence for the elec-

tronic influence of the amine group (and resonance form

b, Figure 4) on the structure of the fulvene ring (vide

F igu r e 3. Thermal ellipsoid plot of compound 8. Thermal

ellipsoids depict the 50% probability level.

1

supra). The four fulvene H resonances in the H NMR

experimental conditions a second molecule of aceto-

nitrile is incorporated into the complex and a fulvene

substituent is obtained from the cyclopentadienyl frag-

ment (eq 4). Solution ¹H NMR data (CDCl₃) for this

spectrum (CDCl₃) appear in the region δ 6.5-6.1. The

broadness of these signals and the lack of observed

interproton coupling in the RT (room temperature)

spectrum are both symptomatic of a fluxional process.

Consistently, changing the NMR solvent to toluene-d₈

results in a shift of the fulvene signals into the aromatic

region and their coalescence into two broad singlets,

implying a fluxional process that is faster in the less-

polar solvent. Indeed, a VT-¹H NMR experiment (MeCN-

d₃) reveals that at 60 °C the R- and R′-H signals coalesce

to a broad singlet, as do the â- and â′-H signals. No

further change occurs with an increase in temperature.

Because the four sites are inequivalent in the static

fulvene unit, coalescence implies that the R-site is

equilibrating with the R′-site, as is the â- with the â′-

site, through a rotation about the fulvene C₂axis

(Scheme 6, A). The approximate rotation rate of

red-brown crystalline product are consistent with the

existence of the protonated azametallacyclopentadienyl

ring found in the molecular structures of complexes 4-7.

The results of a single-crystal X-ray diffraction study

confirm the structure shown for fulvene complex 8 in

eq 4.

272(5) s^-1at coalescence corresponds³²to a ∆G^q

)

15.3(2) kcal mol^-1. The source of this hindered rotation

presumably lies in the partial reduction of the unique

CdC bond order, as highlighted in the structural

discussion above. Although a double bond is drawn

between the two unique carbon nuclei of the fulvene unit

in the formal bonding depiction, delocalization of amine

lone-pair electron density to the N-C bond imparts

considerable single-bond character to the fulvene link-

age, permitting hindered rotation about this axis.

In addition, the ¹H NMR spectrum of 8 in CDCl₃

containing added D₂O shows the surprising result that

²H label is incorporated into the R- and R′-H positions

The solid-state molecular structure depicted in Figure

3 reveals that the second molecule of MeCN is incorpo-

rated into the complex as a component of an aminopen-

tafulvene ligand. In addition, the angles around both

C(11) and C(13) sum to 360.0°, in accord with the

proposed sp²hybridization at these centers, revealing

the fulvene character in this five-membered ring. On

the other hand, the C(11)-C(13) bond length and the

fulvene ring bond lengths in 8 deviate substantially from

those in the analogous 6,6-dimethylpentafulvene,³⁰the

C(sp²)dC(sp²) bonds containing significant single-bond

character and the single bonds being contracted relative

to the accepted C(sp²)-C(sp²) bond length. A structural

study of the influence of dipolar resonance forms in

aminopentafulvenes conducted by Ammon has revealed

that the presence of the electron-releasing amine sub-

stituent results in significant charge separation in the

ground state of the molecule; i.e., resonance structures

a and b depicted in Figure 4 make a significant

contribution to the ground state.³¹

1

(δ 6.50 and 6.44 in the H NMR spectrum) but not into

either of the â-H positions, in addition to the expected

exchange at the two amine sites. Consistently, selective

irradiation of the fulvene amine H signal results in a

dramatic reduction in intensity of both R-H signals by

equivalent amounts via magnetization transfer between

the amine-H and two R-H sites. An exchange process

is, thus, implicated (Scheme 6, B). A 1,4-H shift

transfers H_nto the R-position, yielding two equivalent

R-H bound to the sp³R-C. In the microscopic reverse,

either one of these two protons may be transferred back

to the amine N. Since the amine N readily exchanges

with D₂O, label is incorporated into the fulvene ring via

mechanism B. Because the label appears in both the R

and R′ sites in the slow-exchange RT spectrum, the

exchange mechanism (B) must be operating indepen-

The aminopentafulvene N in 8 functions as a 1e donor

as a result of the electronic saturation of the metal

center. Thus, it reasonable to assume that a substantial

contribution by resonance form b to the ground state of

the aminopentafulvene ligand is being reflected in the

structural parameters determined for this ligand.

(31) Ammon, H. L. Acta Crystallogr., Sect. B 1974, 30, 1731.

(32) Gu¨nther, H. NMR Spectroscopy; Wiley: New York, 1980.

(30) Chiang, J . F.; Bauer, S. H. J . Am. Chem. Soc. 1970, 92, 261.

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 863

Sch em e 6

Sch em e 7

dent of the rotational mechanism (A) to corroborate the

experimental observations.³³That this process is sig-

nificantly slower than fulvene rotation is established by

the simultaneous reduction of the H_Rand H_R′signal

intensity upon irradiation of H_n. Unfortunately, the

separate measurement of the H_nexchange rate at

coalescence by magnetization transfer NMR spectros-

copy is hampered by the near-coincidence of the coales-

cent H_Rsignals and that of H_n.

The mechanism to account for the formation of 8 is

shown in Scheme 7 and is most easily expressed as a

continuation of that presented in Scheme 4 for com-

plexes 4-7. Protonation by CpH at the basic imido site

in E affords F , an intermediate analogous to compounds

4-7. Acetonitrile insertion may then proceed via one

of two pathways. Dissociation of either Cp^-or the

pendent imine followed by MeCN coordination yields the

adducts G or H, respectively. Attack at the quaternary

nitrile carbon by Cp^-in an inter- or intramolecular

fashion then affords the azomethine intermediate I, and

a 1,3-H tautomerization yields the observed enamine

product.

Either pathway leading to I is plausible. The pendent

imine fragment certainly could possess a degree of

lability under the thermolysis conditions employed.

Likewise, outer-sphere “η⁰” Cp^-anions have been

reported, but these seem to predominate in conjunction

with late transition metal coordination compounds.³⁴

Considerable literature precedent exists for the η¹-

(33) An intermolecular exchange cannot be absolutely discounted,

but the kinetic inaccessibility of the fulvene R-H suggests that such a

process is unlikely.

(34) Casey, C. P.; O’Connor, J . M.; Haller, K. J . J . Am. Chem. Soc.

1985, 107, 1241.

864 Organometallics, Vol. 17, No. 5, 1998

Legzdins et al.

bonding mode of the Cp ligand,³⁵although examples of

insertion chemistry into such metal-carbon linkages

are scarce. Casey has described the formation of an

oxafulvene of rhenium by the insertion of CO into a Re-

η¹-Cp bond in the presence of PMe₃.³⁴More recently,

Carmona has reported the formation of the C-bound

aminofulvene ligand [C(dC(C₄H₄))N(H)^tBu] by invoking

a 1,3-H shift in the imido ligand derived from the

t

insertion of BuNC into the Pd-C bond of an η¹-Cp

ligand.³⁶

B. Ad d ition of th e Ca r bon yl F u n ction a cr oss th e

W-N Bon d . We have suggested that the chemistry

occurring beyond the initial reductive coupling event

leading to compounds 2-8 results because the hetero-

atom lone pair does not stabilize the 16e W center in

intermediates C and E (vide supra). The nature of

products 4-8 derived from the thermolyses in nitrile

solvents containing protic acids and the proposed path-

way by which they form reflect both the nucleophilic

character of the imido nitrogen atom conferred by a

build-up of electron density at N and the resultant

electronic unsaturation of the metal center in E.

Further evidence supporting this rationale is provided

by the very interesting and unusual products derived

from the thermolysis of 1 in 0.1 M solutions of acetone

in MeCN or ⁱPrCN. After 24 h at 45 °C in these

solutions, the alkyl vinyl complex 1 is cleanly converted

to 9 or 10, respectively (eq 5). These product complexes

F igu r e 5. Thermal ellipsoid plot of complex 10. Thermal

ellipsoids depict the 50% probability level.

Sch em e 8

are isolable in relatively high yields from their respec-

tive reaction mixtures as red-brown crystalline blocks

and are air-stable in the solid state for up to 2 weeks.

Analogous to complexes 4-8 described previously, the

phenyl ring, vinyl proton, and methyl substituent are

readily identified from the characteristic signals in the

¹H and ¹³C NMR spectra of 9 in CDCl₃(Table 3). While

these assignments are straightforward, extra signals in

the proton and carbon spectra exist which are in accord

with the presence of two additional methyl substituents

and an extra quaternary carbon nucleus. The masses

of the parent ions in the mass spectra of 9 and 10 are

consistent with the inclusion of one molecule of acetone

in the molecular formulas for these complexes, a feature

also corroborated by their elemental analyses.

That acetone is incorporated into these extended ring

systems is unequivocably confirmed by the solid-state

molecular structure detemined for complex 10 by X-ray

crystallography (Figure 5). The carbonyl unit has

clearly added across the W-N bond yielding a C(11)-

N(2) single bond distance of 1.478(5) Å and a W(1)-

O(1) single bond length of 2.075(3) Å. Likewise, the

C(11)-O(1) distance of 1.425(5) Å reflects the reduction

of the C-O bond order to that of a single bond. The

sum of the bond angles about N(2) (360.0°) reveals the

sp²nature of the imine moiety, and the W(1)-N(2)

distance of 2.095(3) Å is characteristic of coordination

of N to W.²⁵

The mechanism we propose to yield complexes 9 and

10 is depicted in Scheme 8 and, like that proposed for

8, stems from the generation of the 16e intermediate E

containing a nucleophilic imido N center. Following the

formation of E, acetone coordination (J ) facilitates the

addition of the W-N bond across the carbonyl fragment.

The regioselectivity of this addition reflects both the

affinity of W for O^1aas well as the polarity of the W-N

linkage; the nucleophilic imido N attacks at the elec-

trophilic quaternary carbonyl carbon with concomitant

O attack at the electrophilic, electronically unsaturated

metal center. That 9 and 10 are formed exclusively

implies that the J f product step is irreversible.

Consistently, 9 does not convert to 4 under thermolysis

conditions in wet THF-d₈. At present, this ring expan-

sion seems to be limited to sterically unencumbered

ketones and aldehydes. For example, an NMR-scale

thermolysis of 1 in MeCN containing small amounts of

acetaldehyde affords the analogous addition products,

as evinced in the ¹H NMR spectrum (CDCl₃) of the final

reaction mixture. Two signals of equal intensity are

observed for each environment, thereby implying the

formation of two regioisomers in which the methyl

substituent of the alkoxy fragment is presumably

(35) O’Connor, J . M.; Casey, C. P. Chem. Rev. 1987, 87, 307.

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 865

directed either toward or away from the Cp* ring. In

contrast, the analogous thermolyses in acetonitrile

solvent containing small amounts of larger substrates

such as ethyl acetate, dimethyl acetamide, or benzal-

dehyde afford only complex 11 (vide infra) and small

amounts of unidentified decomposition products. An

interesting observation is the selectivity of intermedi-

ates A and E for the reagents in solution. The higher-

energy intermediate A is observed to couple with nitrile

exclusively, implying that the reductive coupling step

is under kinetic control and the resultant product

distribution is governed only by the relative concentra-

tions of nitrile and ketone. On the other hand, the

reaction of the more stable intermediate E is governed

by the more thermodynamically favorable W-O inter-

action in J and forms the oxa/aza bicyclic product

exclusively (vide infra), despite the difference in con-

centration of the two donor solvents.

The ring-expanding insertions of unsaturated hydro-

carbons,³⁷ketones,³⁸and carbon monoxide³⁹into met-

allacycle M-C σ bonds are well-known. Conversely, the

addition of unsaturated molecules across metal-nitro-

gen bonds of amido complexes has been observed only

in the case of CO and other compounds such as CO₂or

PhNCO that contain sites of electrophilicity.⁴⁰To the

best of our knowledge, the addition of a ketone across a

metal-imide bond to form bicyclic complexes such as 9

and 10 is unprecedented. This transformation consti-

tutes a unique sequential [2 + 2] regioselective cyclo-

addition of phenyl acetylene, nitrile, and acetone within

the metal’s coordination sphere.

F igu r e 6. NOE results for the indicated environments in

complex 11. [W] represents Cp*W(NO). T indicates an

observed NOE.

imine N atom is correctly disposed relative to the NO

ligand for dative bonding to W, thus rendering 11 and

12 electronically and coordinatively saturated at the

metal center.²⁵

The close similarity of the bicyclic ring in these

compounds to that established for the structurally

characterized 10 aids in the formulation of these

compounds on the basis of their spectroscopic data.

Utilizing compound 11 as an example, the characteristic

signals of the azametallacyclopentenyl ring components

are easily identified in the NMR spectral data (Table

3). A broad singlet at 4.73 ppm in the ¹H NMR

spectrum and a ν_NHat 3268 cm^-1in the KBr-pellet IR

spectrum reveal the presence of the amine H. Likewise,

the signal attributable to the resonance of the ring vinyl

proton is clearly discernible at 6.83 ppm. A short-range

¹H-¹³C correlation experiment identifies the carbon

signals to which the proton environments, particularly

the vinyl signals, are coupled. For example, the en-

docyclic vinyl proton resonance at δ 6.83 couples to the

carbon signal at 138.8 ppm with a coupling constant of

158 Hz, a value obtained from a gate-decoupled ¹³C

NMR experiment. Likewise, the exocyclic vinyl reso-

C. Nitr ile Ad d ition a cr oss th e W-N Bon d . In the

absence of an added protic source or electrophile, the

blood-red products resulting from the thermolysis of 1

in MeCN or EtCN are generated in virtually quantita-

tive yield (eq 6). These bicyclic amidinate complexes

1

nances at 4.73 and 4.45 ppm in the H NMR spectrum

couple to the carbon-13 signal at 90.6 ppm (¹J _CH) 152

Hz). Hence, the ring substituents (the vinyl H, CH₂

vinyl group, methyl substituent, and phenyl group) are

identifiable from these data, although their connectivity

is not. The bicyclic structure of this ring system is

implied by analogy to the solid-state molecular structure

determined for complex 10, but the location of the NH,

methyl, and vinyl CH₂units can only be assigned

1

through the combination of NOE and long-range H-

¹³C correlation data. The spatial arrangement of these

substituents is clearly revealed by the results of an NOE

experiment (Figure 6).

are isolable as microcrystalline solids and are air stable

in solution or in the solid state for periods up to 1 month.

As described for compounds 9 and 10 (vide supra), the

It, thus, remains to assign the quaternary nuclei to

which these substituents are bound, utilizing the results

of a HMBC H-¹³C correlation experiment. Thus, both

(36) Al´ıas, F. M.; Belderra´ın, T. R.; Paneque, M.; Poveda, M. L.;

Carmona, E. Organometallics 1997, 16, 301.

1

the endocyclic vinyl H signal and that of the phenyl

ortho-H couple via a two-bond interaction to the qua-

ternary signal at 191.9 ppm, which can be assigned as

the C R to W. Likewise, the endocyclic vinyl H signal

shows a two-bond coupling to the signal at 156.1 ppm,

as does one of the signals attributable to the geminal

vinyl protons. This signal is, therefore, attributed to

the quaternary imine carbon. Both geminal vinyl

proton signals couple to the endocyclic vinyl carbon

signal at 138.8 ppm. Finally, the quaternary carbon

signal at 171.1 ppm shows a coupling to the singlet

assigned to the methyl substituent at 2.11 ppm, which

permits its assignment as the quaternary amidinate

carbon. Complex 12 has been characterized in an

analogous manner.

(37) (a) Reference 3a. (b) Tebbe, F. N.; Parshall, G. W.; Reddy, G.

S. J . Am. Chem. Soc. 1978, 100, 3611. For more recent examples, see:

(c) Bruck, M. A.; Copenhaver, A. S.; Wigley, D. E. J . Am. Chem. Soc.

1987, 109, 6525. (d) Christensen, N. J .; Legzdins, P. Organometallics

1991, 10, 3070. (e) Yeh, W.-Y.; Ho, C.-L.; Chiang, M. Y.; Chen, I.-T.

Organometallics 1997, 16, 2698 and references therein. (f) Yang, S.-

M.; Chan, M. C.-W.; Cheung, K.-K.; Che, C.-M.; Peng, S.-M. Organo-

metallics 1997, 16, 2819.

(38) (a) Reference 26a. (b) Meinhart, J . D.; Grubbs, R. H. Bull. Chem.

Soc. J pn. 1988, 61, 171. (c) Christensen, N. J .; Legzdins, P.; Trotter,

J .; Yee, V. C. Organometallics 1991, 10, 3070. (d) Buchwald, S. L.;

Grubbs, R. H. J . Am. Chem. Soc. 1983, 105, 5490. (e) Yasuda, H.;

Okamoto, T.; Mashima, K.; Nakamura, A. L. Organomet. Chem. 1989,

363, 61. (f) Beckhaus, R. J . Chem. Soc., Dalton Trans. 1997, 1991.

(39) For a leading reference, see: Kablaoui, N. M.; Hicks, F. A.;

Buchwald, S. L. J . Am. Chem. Soc. 1997, 119, 4424.

(40) See, for example: (a) Cabeza, J . A.; del R´ıo, I.; Franco, J .;

Grepioni, F.; Riera, V. Organometallics 1997, 16, 2763. (b) Cowan, R.

L.; Trogler, W. C. J . Am. Chem. Soc. 1989, 111, 4750. (c) Bryndza, H.

E.; Fultz, W. C.; Tam, W. Organometallics 1985, 4, 939.

866 Organometallics, Vol. 17, No. 5, 1998

Legzdins et al.

Sch em e 9

For example, no ligand substitution chemistry occurs

for 11 under thermolysis conditions in the presence of

MeCN-d₃and PMe₃.⁴¹This fact coupled to the absence

of characteristic IR⁴²and ¹³C NMR⁴³evidence suggests

that the second equivalent of acetonitrile has added

across the W-N linkage. A labeling study indicates

that the tautomerization step is a reversible one. Thus,

thermolysis of authentic 11 in THF-d₈containing D₂O

results in label incorporation at the amine position as

well as at the exocyclic methyl position and the exocyclic

vinyl position, thereby affording 11-d₆but not facilitat-

ing the conversion of 11 to 4-d₂. Since the intramolecu-

lar tautomerization is reversible, the ring expansion

affording L must be irreversible to account for the fact

that 11 is not converted to 4 in the presence of water.

(This deuterated complex, 11-d₆, can also be prepared

independently by thermolysis of 1 in MeCN-d₃.) The

extensive deuteration in the ring substituents of 11-d₆

leads to the conclusion that both a formal 1,3- and a

formal 1,5-tautomerization process⁴⁴are in operation,

with the thermodynamic product resulting from a 1,5-H

shift. To probe this tautomerization mechanism further,

a crossover experiment involving thermolysis of an

equimolar mixture of 11-d₆and 12 in THF-d₈was

performed. Monitoring of this mixture by ¹H NMR

spectroscopy reveals the slow incorporation of the

deuterium label into the amine, ethyl, and methylvinyl

susbtituents in 12 and proton incorporation into the

analogous sites in 11-d₆, a clear indication of the

intermolecular nature of this tautomerization process.⁴⁵

A striking feature of this experiment is the slow rate of

conversion. After 24 h at 65 °C, the deuterium scram-

bling is approximately 50% complete, as judged by

integration of the methyl signals at about 2.1 and 1.3

The mechanism leading to these complexes (Scheme

9) can be viewed as being analogous to that proposed

for the formation of 9 and 10 (Scheme 6). The azamet-

allacyclopentadiene E is first formed by reductive

coupling of the acetylene and RCN at W.

A second molecule of acetonitrile coordinates (K) and

adds across the polar W-imine link (L), a process

facilitated by attack at the electrophilic quaternary

nitrile carbon nucleus by the nucleophilic imido N in

K. A coordinatively saturated intermediate species (L)

is thus generated. This species, an 18e analogue of E,

contains a basic imido N center that is quenched upon

tautomerization via a formal 1,5-H shift to afford the

observed products.

1

ppm in the H NMR spectrum. Since the thermolyses

to form complexes 2-12 require approximately 24 h at

45 °C, both the dilute concentration of the NMR sample

and an isotope effect presumably retard the rate of

tautomerism in this experiment, a conclusion fully in

accord with the proposed bimolecular mechanism.⁴⁶

The competition by MeCN and water for reaction with

intermediate E has been tested by conducting the

thermolysis of 1 in MeCN containing 2, 5, and 10 equiv

of water ([H₂O] ) 19, 46, and 93 mM, respectively).

Formation of 11 (∼5%) was observed only in the case

where [H₂O] ) 19 mM, as detected by ¹H NMR

spectroscopy of the three mixtures after their thermoly-

sis for 24 h. Quantitative conversion to 4 was indicated

in the ¹H NMR spectra of the final mixtures in the latter

two cases. Thus, the ring expansion afforded by the

insertion of a second acetonitrile molecule must be a

more energetically demanding process; only at very low

water concentrations does this process become competi-

tive with protonation of E by water.

t

i

Interestingly, when either BuCN, PhCN, or PrCN

is employed in this reaction, no tractable products are

t

obtained. This is not surprising in the case of BuCN

since no product was isolated in the presence of added

water either, a result that we attribute to unfavorable

steric interactions between the metal complex and

incoming solvent molecule, inhibiting the reductive

coupling event (vide supra). In the absence of steric

effects, the lack of accessible protons on the quaternary

â-carbon prohibits the tautomerization step which af-

fords the final product in these transformations. This

is also the case for PhCN. In contrast, it is remarkable

that no product is obtained with ⁱPrCN since the

isolation of 6 confirms that the reductive coupling event

does indeed occur. On the other hand, the steric

interactions limiting the scope of the ring expansion that

affords complexes 9 and 10 might also play a key role

in the analogous expansion involving nitriles. If so, it

is of no consequence that thermolysis of 1 in anhydrous

ⁱPrCN yields no tractable products since during nitrile

thermolysis in the presence of large organic carbonyls

only complex 11 is obtained. Like compounds 9 and 10,

the insertion of a nitrile into a metal-N bond appears

to be unprecedented, constituting a unique, regioselec-

tive coupling of 2 equiv of nitrile with phenyl acetylene

to give the vinyl amidinate complexes 11 and 12.

Several other experiments have been performed to

corroborate our structural and mechanistic proposals.

(41) Legzdins, P.; Sayers, S. F. J . Am. Chem. Soc. 1994, 116, 12105.

(42) Nakamoto, K. Infrared and Raman Spectra of Inorganic and

Coordination Compounds, 4th ed.; Wiley-Interscience: New York,

1986.

(43) Yeh, W.-Y.; Ting, C.-S.; Chih, C.-F. J . Organomet. Chem. 1991,

427, 257.

(44) The formal 1,5-tautomerization can also be viewed as two

sequential 1,3-H shifts; the first transfer occurs from the methyl group

to the internal imine N and the second from the internal imine N to

the terminal amide N. That the tautomerization is shown to be

intermolecular renders these labels formal descriptors only.

(45) For an elegant kinetic and mechanistic study of an analogous

intermolecular tautomerization process in (DIPP)₃Ta(η²-HNC(dCH₂)C-

(Ph)dCPh) , see ref 26c.

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 867

Ta ble 5. Kin etic Da ta for th e Con ver sion of 1 to

11, 4, a n d 3

temp [H₂O] product

solvent (K)

k_obs

av k_obs

(M) complex (×10^-5s^-1

)

R²

(×10^-5s^-1

)

MeCN 318

11

4

4.9

4.7

4.5

4.7

5.0

4.4

4.5

4.6

4.8

4.4

4.7

4.9

2.8

3.0

9.7

0.9994

0.9997

0.9996

0.9997

0.9998

0.9999

0.9997

0.9998

0.9997

0.9983

0.9998

0.99996

0.9998

0.9997

0.9998

4.7 ( 0.2

MeCN 318 0.12

MeCN 318 0.06

MeCN 318 0.03

MeCN 318 0.006

EtOAc 318

4.6 ( 0.1

4.6 ( 0.4

4.6 ( 0.1

4.7 ( 0.3

2.9 ( 0.1

4

3

EtOAc 318

3

EtOAc 318

3

F igu r e 7. Possible tautomeric pathways affording 18e

complexes from intermediate L.

EtOAc 327

3

EtOAc 327

EtOAc 336

EtOAc 341

EtOAc 348

3

11

51

49

90

93

103

265

268

295

0.99992 10 ( 1

0.99996

0.9997

0.9996

0.9994

0.9996

50 ( 1

95 ( 6

Analogous to fulvene formation in complex 8, the

extensive conjugation upon tautomerization undoubt-

edly provides the stabilization necessary for the forma-

tion of the less-stable enamine tautomer 11 under

thermodynamic conditions. That an intermolecular

proton shift from the internal Me substituent to the

terminal imido N results in the exclusive formation of

11 as the thermodynamic product can be rationalized

(Figure 7) by considering conjugation effects and orbital

overlap arguments analogous to those presented for E.

As noted previously (Scheme 9), the internal imido N

is properly oriented for dative bonding to W in both L

and 11, thereby conferring a closed-shell electronic

configuration on both of these complexes. That L

proceeds to abstract a proton intermolecularly is an

indication of the basic nature of the terminal imido N.

On the other hand, the internal imido N is reduced in

its Lewis basicity as a result of coordination to W; a

proton transfer affording the internal amine (X) is

therefore improbable. Such a shift would also lead to a

disruption of the ring conjugation, again yielding the

thermodynamically least-stable tautomer of the three.

On the other hand, the formation of both Y and 11 result

in substantial ring conjugation, and yet 11 is the only

thermodynamic product observed. The relative stability

of 11 over Y is probably a result of the formation of the

more stable amidinate unit in the conjugated backbone

in 11.⁴⁷

0.9996 276 ( 16

0.9998

EtOAc 348

3

produce the η²-acetylene intermediate A.⁴⁸We propose

here a rate-limiting generation of A through an E1-type

dissociation pivotal to the mechanisms giving rise to

both modes of reactivity. The results outlined in Table

5 support the proposed rate-limiting generation of η²-

acetylene intermediate A in all of these reactions and

mitigate against an alternative bimolecular S_N2-type

associative mechanism; in such a case, a significant rate

dependence on the nature of the incoming substrate

would be expected.⁴⁹Hence, the rates of conversion of

1 in acetonitrile doped with water afford observed rate

constants from 4.6 × 10^-5s^-1to 4.7 × 10^-5s^-1for water

concentrations ranging from 0.006 to 0.12 M, eliminat-

ing the possibility of a bimolecular saturation mecha-

nism involving hydration of azametallacycle E by water.

More importantly, the rates of conversion of 1 in the

presence of different substrates are remarkably similar,

ranging from 2.98 × 10^-5s^-1to 5.04 × 10^-5s^-1. The

slight variance in the observed rate constants for the

formation of 11, 4, and 3 can be ascribed to the

differences in the bulk properties of the solvents and

not to any definable mechanistic differences.⁵⁰A study

of the temperature dependence of the reaction rate

(Figure 8) yields an enthalpy and entropy of activation

of 32(3) kcal mol^-1and 21(2) cal (mol‚K)^-1, respectively.

The magnitude of the enthalpy reflects considerable

Kin etic Stu d ies. The fact that the conversions

described in this paper are very clean prompted us to

consider furthering our understanding of this chemistry

through kinetic studies. UV-vis spectroscopy was

employed as the method of choice for collecting well-

defined kinetic data. Quantification of these data is

effected by monitoring the growth of the shoulder at 336

or 350 nm, the results of which are summarized in Table

5. The conversions of complex 1 to 11, 4, and 3 under

thermolysis conditions at 45 °C were monitored in this

way, and the observed rate constants were determined

from a first-order analysis of the data.

(46) The results of this crossover experiment only confirm the

intermolecular component to this tautomerization. An intramolecular

component cannot be ruled out in the absence of the appropriate kinetic

studies.

(47) A very rapid isomerization (on the NMR time scale) between Y

and 11 cannot be discounted from the experimental data.

(48) A manuscript describing the scope of the thermal C-H activa-

tion chemistry observed for this system in the presence of a variety of

hydrocarbons, including a detailed kinetic and mechanistic analysis,

is currently in preparation.

The mechanism originally proposed⁷to account for the

observed C-H activation of n-pentane and n-hexane

emphasized an initial elimination of SiMe₄from 13 to

(49) Carmona proposes such a mechanism for the formation of an

analogous iridapyrrole complex: see ref 23a.

(50) Connors, K. A. Chemical Kinetics: The Study of Reaction Rates

in Solution; VCH Publishers: New York, 1990; Chapter 8.

868 Organometallics, Vol. 17, No. 5, 1998

Legzdins et al.

F igu r e 8. Eyring plot for the thermolysis of 1 in EtOAc,

308 < T < 348 K.

bond scission at the transition state, while the small

and positive value of the activation entropy implies a

degree of dissociative character in the rate-limiting step.

Comparable findings have been reported for a similar

rate-limiting reductive elimination process.⁵¹

F igu r e 9. Frontier orbitals in isolobal CpW(NO) (I-III)

and Cp₂Zr (IV) complexes.

from the N p-type filled orbital into this orbital, yielding

an 18e, saturated metal center. This phenomenon is

reflected structurally in the coplanarity of the W-NO

and the N-R or O-R vectors in the molecular struc-

tures determined for such compounds⁵⁵and is supported

by computational modeling.¹⁴In the case of intermedi-

ate E, however, the filled orbital on N is forced to point

away from the metal center by the conformation of the

ring (case III). Reduced orbital overlap occurs, which,

in combination with the diminished oxophilicity of W

relative to Zr, leads to an increase in the Lewis basicity

of the imido N and Lewis acidity of the metal center. In

comparison, the analogous LUMO in the zirconocene

system contains two disproportionately larger lobes.⁵⁶

These larger, more accessible lobes are exterior to the

azametallacycle, one being adjacent to the imide N,

making donation from the filled N orbital into this

LUMO facile (case IV). Thus, complexes of this type

are stabilized for the group 4 case and are readily

isolable. That a kinetic opportunity exists for the

addition of ROH, ketones, or nitriles across the W-N

bond in intermediate E we propose to be a consequence

of the greater electron density at the imide N in addition

to the reduced steric requirements of the NO ligand. The

analogous chemistry is not observed in the isolobal

zirconocene system, despite the fact that small mol-

ecules such as H₂O and CO react with the zirconocene

analogues to release the ring from the metal center. The

smaller metallocene wedge in the latter system probably

offers little opportunity for reaction with larger, unsat-

urated heteroatom-containing molecules at the metal

center once the products of reductive coupling have

formed.

Rin g Exp a n sion a n d Liga n d Ela bor a tion . In

Schemes 5, 8, and 9 we propose that the chemistry

leading to complexes 4-12 arises from the specific

incapability of the N atom in intermediate E to function

as a 3e donor to W, an effect caused by the conforma-

tional restrictions of the metallacycle. Two salient

consequences of this are (i) the electronic unsaturation

of the metal center in these intermediate complexes and

(ii) the resultant basicity or nucleophilicity conferred on

the imido N which affords a reactivity beyond the initial

reductive coupling event. The origin of these chemical

characteristics in the azametallacycle E (and also in the

O-containing analogue C) deserves further discussion

in light of the fact that the analogous complexes for the

Cp₂Zr system are readily isolated and have been exten-

sively studied. In addition to the greater oxophilicity

of the group 4 metals being a stabilizing influence in

these metallacycles, an orbital-overlap argument can

also be made for the differences in reactivity. Such a

rationale requires a brief review of the Lewis acidity

inherent to the 16e model complexes CpM(NO)Me₂(M

) Mo, W). It has been demonstrated computationally⁵²

that there exists a metal-centered d_xy-type LUMO in

these complexes oriented perpendicular to the M-NO

vector, with the most sterically accessible lobe bisecting

the R-W-R angle (Figure 9, case I).

Furthermore, it has been shown experimentally⁵³that

16e bis(alkyl) complexes of type I are Lewis acidic and

that the coordination of 2e Lewis bases occurs cis to both

alkyl ligands at the most kinetically accessible lobe of

the LUMO.⁵⁴Thus, it is a reasonable assumption that

a similar metal-centered LUMO exists in the related

complexes containing an NR₂or OR ligand (case II, NR₂

ligand depicted). Here, however, the heteroatom-

containing ligand can orient to stabilize the metal-

centered LUMO through donation of electron density

Ep ilogu e

In this paper, the products of the reductive coupling

of esters and nitriles with coordinated phenyl acetylene

in the coordination sphere of the Cp*W(NO) template

are described. Driving the chemistry that occurs beyond

the reductive coupling event is the inability to attain

an 18e configuration at the W center, a feature in

contrast to previously published work involving the

isolobal zirconocene system. A number of pathways are

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

Soc. 1986, 108, 1537.

(52) (a) Legzdins, P.; Rettig, S. J .; Sa´nchez, L. J .; Bursten, B. E.;

Gatter, M. G. J . Am. Chem. Soc. 1985, 107, 1411. (b) Bursten, B. E.;

Cayton, R. H. Organometallics 1985, 6, 2004.

(53) (a) Legzdins, P.; Rettig, S. J .; Sa´nchez, L. J . Organometallics

1988, 7, 2394. (b) Reference 25. (c) Legzdins, P.; Veltheer, J . E. Acc.

Chem. Res. 1993, 26, 41.

(54) The other three lobes of the LUMO are kinetically inaccessible.

One of the three extends upwards toward the Cp* ring, whereas access

to the other two lobes is presumably blocked by the Cp* methyl

substituents (which have been omitted for clarity in Figure 9).

(55) (a) Kuzelka, J .; Legzdins, P.; Rettig, S. J .; Smith, K. M.

Organometallics 1997, 16, 3569.

(56) Lauer, J . W.; Hoffmann, R. J . Am. Chem. Soc. 1976, 98, 1729.

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 869

employed by the Cp*W(NO)(η²-PhCtCH) fragment in

accomplishing this end. In the case involving reductive

coupling to esters, â-C-O bond cleavage in the putative

16e intermediate oxametallacyclopentene complex af-

fords alkoxide-containing, 18e oxametallacyclopentadi-

ene complexes. In nitrile solvent, the proposed 16e

azametallacyclopentadiene complex is trapped by pro-

tonation of the basic imine N by water, allyl alcohol, or

CpH. Coordination of the conjugate base, thus, fulfills

the electronic requirements of the metal, yielding 18e

compounds. In the latter case, the tautomeric amino-

pentafulvene ligand is formed which exhibits interesting

fluxional behavior at room temperature in solution. In

the presence of acetone, the intermediate azametalla-

cyclopentadiene complex is trapped by a chemo- and

regioselective ring-expanding addition of ketone. In the

absence of any trapping agent, a second equivalent of

RCN inserts in a manner analogous to acetone, an

intermolecular proton shift affording the observed vinyl

amidinate complex. A kinetic study of these transfor-

mations implicates a mechanism involving rate-limiting

elimination of SiMe₄from the parent bis(hydrocarbyl)

complex. In all cases studied, reaction ensues beyond

the initial reductive coupling event until electron satu-

ration of the metal center is achieved. This phenom-

enon has been rationalized through a qualitative dis-

cussion of frontier orbitals; a reduced compatibility of

the heteroatom lone-pair orbital(s) with the metal-

centered LUMO in the proposed 16e transient W species

results in a Lewis-acidic metal center and nucleophilic

heteroatom. This incompatibility is in direct contrast

to the 18e configuration achievable by the analogous

zirconocene system and reflects the difference in the

electronic structure and oxophilicity of the two systems.

The scope of the reactivity that results is an extension

of, and provides an interesting contrast to, the body of

work surrounding the analogous zirconocene system.

Kin etic Stu d ies. Kinetic studies were performed using a

HP8452 UV-vis spectrometer equipped with a thermostated

cell holder connected to a VWR 1150 constant-temperature

bath which was accurate to (0.25 °C. Typical kinetic runs

monitored the product band at 336 or 360 nm arising from

the thermolysis of 0.1-1 mg of 1 dissolved in 3 mL of solvent

over not less than 3.5 half-lives. Absorbance values for t_∞were

obtained by computer optimization of the squared residual,

R², for the regression line fitted to the data through a first-

order analysis.

X-r a y Diffr a ction Stu d ies. Data collection and structure

solution for complexes 2, 4, 8, and 10 were conducted at the

X-ray Crystallographic Laboratory at the University of Min-

nesota. All calculations were performed using SGI Indy

RR4400-SC or Pentium computers using the SHELXTL V5.0

suite of programs.

A crystal of each compound was attached to a glass fiber

and mounted on a Siemens SMART system for data collection

at 173(2) K. An initial set of cell constants was calculated from

reflections harvested from three sets of 20 frames. Final cell

constants were calculated from sets of strong reflections

obtained during the actual data collection. Crystallographic

data are presented in Table 4. In the hemisphere collection

technique employed during this work, a randomly oriented

region of reciprocal space was surveyed to the extent of 1.3

hemispheres to a resolution of 0.84 Å. Three major swaths of

frames were collected with 0.30° steps in ω. This collection

strategy provided a high degree of redundancy in the data as

well as providing good ψ input in the event that an empirical

absorption correction had to be applied. The space groups

were determined based on systematic absences and intensity

statistics. In each case, a successful direct-methods solution

was calculated which provided most non-hydrogen atoms from

the E-map. Several full-matrix least-squares difference-Fou-

rier cycles were performed which located the remainder of the

non-hydrogen atoms. All non-hydrogen atoms were refined

with anisotropic displacement parameters unless stated oth-

erwise. All hydrogen atoms were placed in ideal positions and

were refined as riding atoms with individual or relative

isotropic displacement parameters.

The refinement of the data collected for complex 2 was

hampered by twinning in the crystal, as evinced by an R1 of

7.2% and a list of 50 worst reflections containing largely h )

4 reflections. A successful attempt was made to locate enough

reflections of the twin to index it; the twin law thus determined

is (by rows) 100, 0,-1,0, -³/₄,0,1, which explains why all h )

Exp er im en ta l Section

Gen er a l Meth od s. All reactions and subsequent manipu-

lations were performed under anaerobic and anhydrous condi-

tions using an atmosphere of dinitrogen or argon. General

procedures routinely employed in these laboratories have been

described in detail previously.⁵⁷The organometallic reagent

Cp*W(NO)(CH₂SiMe₃)(CPhdCH₂) (1) was prepared according

to the reported procedure.⁷Ethyl acetate, methyl acetate,

acetonitrile, propionitrile, isobutyronitrile, tert-butyl nitrile,

benzonitrile, and allyl alcohol (Aldrich) were distilled or

vacuum-transferred from CaH₂. Acetone was refluxed over

Linde 4A molecular sieves, distilled onto vigorously anhydrous

Linde 4A molecular sieves, and vacuum-transferred as needed.

CpH was cracked from cyclopentadiene dimer and stored cold

in the dark under an inert atmosphere. Wet acetonitrile was

prepared by deaerating HPLC-grade solvent obtained directly

from Aldrich or by adding microliter amounts of distilled,

deionized water to rigorously dried acetonitrile via a micro-

syringe. Acetonitrile-d₃(CIL) was dried over CaH₂and

vacuum-transferred. Deuterated water (CIL) was deaerated

under a flow of argon immediately prior to use. Spectroscopic

data, yields, and elemental analyses of all new complexes

synthesized during this work are collected in Tables 1-3.

Where appropriate, NOE, HMQC, HMBC, and gate-decoupled

¹³C experiments were performed on a Bruker AMX 500 MHz

NMR spectrometer to facilitate ¹H and ¹³C spectral assign-

ments.

2

4 reflections have F_o. F_c

.

The program UNTWIN⁵⁸was

used to create a HKLF 5-type reflection file from the original

HKLF 4 file. No data from the minor twin were used to create

the new reflection file. The residual improved 2%, and all

atoms became positive definite. The twin is ∼19% of the mass

of the main specimen.

P r epar ation of Cp*W(NO)(OMe)(η²-OdC(Me)CHdCP h )

(2). Cp*W(NO)(CH₂SiMe₃)(CPhdCH₂) (1, 135 mg, 0.25 mmol)

was dissolved in MeOAc (10 mL) in a thick-walled bomb. The

resulting deep red solution was heated at 45 °C for 24 h, during

which time its color changed to purple. Solvent was removed

from the final solution in vacuo, and the purple residue was

extracted with a minimum of a CH₂Cl₂/hexanes mixture

(1:1). The extract was filtered through a plug of Celite (1.5 ×

0.5 cm) supported on a frit, and the volume of the filtrate was

reduced in vacuo. Cooling the concentrated solution to -30

°C overnight induced the deposition of dark purple needles of

complex 2 (89 mg), which were isolated by removal of the

supernatant solution. A second crop of crystals (28 mg) was

obtained by reducing the volume of the supernatant solution

and cooling for a further 24 h.

(58) Young, V. G., J r., UNTWIN; X-ray Crystallographic Laboratory,

Chemistry Department, The University of Minnesota: Minneapolis,

MN, 1997.

(57) Legzdins, P.; Rettig, S. J .; Ross, K. J .; Batchelor, R. J .; Einstein,

F. W. B. Organometallics 1995, 14, 5579.

870 Organometallics, Vol. 17, No. 5, 1998

Legzdins et al.

P r ep a r a tion of Cp *W(NO)(OEt)(η²-OdC(Me)CHdCP h )

(3). Complex 3 was prepared in a manner analogous to that

described in the preceding paragraph for 2. After having been

heated for 24 h, the reaction mixture was reduced in volume

and the burgundy solution was cooled to -30 °C overnight to

induce the deposition of 3 as analytically pure needles (112

mg), which were isolated by removal of the supernatant

solution and subsequent drying under vacuum.

P r epar ation of Cp*W(NO)(OH)(η²-HNdC(Me)CHdCP h )

(4). Reactant 1 (135 mg, 0.25 mmol) was dissolved in wet

MeCN (10 mL) in a thick-walled bomb. The solution was

heated for 24 h at 45 °C, whereupon it became light orange in

color and a red crystalline solid deposited in the bottom of the

bomb. The orange supernatant solution was removed by pipet,

and the crystals were dried under vacuum. The supernatant

solution was filtered through a Celite plug (1.5 × 0.5 cm), and

the volume of the filtrate was reduced in vacuo. The resulting

dark orange solution was then cooled to -32 °C overnight to

induce the deposition of a second fraction of 4 as orange

microcrystals (107 mg), which were isolated as described

above.

P r epar ation of Cp*W(NO)(OD)(η²-DNdC(Me)CHdCP h )

(4-d ₂). Complex 4-d₂was prepared in a manner similar to that

employed for 4 (vide supra), except that reactant 1 (135 mg,

0.25 mmol) was first dissolved in dry MeCN (10 mL) and the

red solution was then doped with D₂O (25 µL, 1.4 mmol). After

thermolysis, the dark orange solution was concentrated and

cooled to induce crystallization. Red crystalline 4-d₂was

isolated by removal of the solvent and washing with Et₂O. This

complex was characterized by EI-MS and by comparison of its

¹H NMR spectroscopic properties to those displayed by 4.

P r ep a r a tion of Cp *W(NO)(OH)(η²-HNdC(Et)CHdCP h )

(5) a n d Cp *W(NO)(OH)(η²-HNdC(ⁱP r )CHdCP h ) (6). Both

complexes 5 and 6 were prepared in a manner similar to that

employed for 4. Following the thermolyses, the final reaction

mixtures were concentrated and cooled to -32 °C overnight

to induce the deposition of 5 as red needles (82 mg) and 6 as

red blocks (87 mg). Further concentration and cooling of the

supernatant solutions afforded second crops of the products

(18 mg of 5 and 25 mg of 6), which were isolated by removal

of the mother liquor and drying under vacuum.

taken to dryness under reduced pressure, and the residue was

triturated with pentane (3 × 5 mL), washed with Et₂O (3 × 5

mL) and dried under high vacuum. The remaining solid was

dissolved in a minimum volume of CH₂Cl₂/hexanes (1:1), and

the solution was filtered through Celite (1 × 2 cm plug). The

filtrate was reduced in volume and cooled to -32 °C overnight.

Complex 11 was isolated the next day as a dark red micro-

crystalline powder (207 mg) after removal of the supernatant

solution and drying under vacuum.

P r ep a r a tion of Cp *W(NO)(η³-DNC(CD₃)dNC(dCD₂)-

CHdCP h ) (11-d ₆). Complex 1 (30 mg, 0.056 mmol) was

dissolved in MeCN-d₃(0.5 mL) in an NMR tube and was

heated at 45 °C for 24 h. The solvent was then removed in

vacuo, and the red solid was washed twice with Et₂O and dried

under high vacuum for 4 h. Characterization of 11-d₆was

1

effected by EI-MS and by comparison of its H NMR spectrum

(CDCl₃) to that exhibited by 11.

P r ep a r a tion of Cp *W(NO)(η³-HNC(Et)dNC(dCHMe)-

CHdCP h ) (12). Complex 12 was prepared in a manner

identical with that for 11, except that the dark red-brown

solution produced by thermolysis was reduced in volume and

cooled overnight twice to induce the deposition of two crops of

12 (96 mg total) as brown, analytically pure microcrystals.

Rea ction of 4 w ith D₂O. Complex 4 (20 mg, 0.040 mmol)

was dissolved in MeCN-d₃(0.5 mL) in an NMR tube, and D₂O

(10 µL, 0.56 mmol) was added. The resulting solution was

1

heated at 45 °C overnight, and its H NMR spectrum was then

recorded. Quantitative formation of 4-d₂was indicated by a

comparison of this spectrum to that exhibited by an authentic

sample of 4-d₂.

Th er m olysis of 4 in EtOAc. Complex 4 (20 mg, 0.040

mmol) was dissolved in EtOAc (3 mL) in a glass bomb and

heated for 24 h at 45 °C. The volatiles were removed under

vacuum after cooling the orange solution. Comparison of the

¹H NMR spectrum (CDCl₃solution) of the resultant orange

solid to that of authentic 4 indicated no reaction during

thermolysis in EtOAc.

Th er m olysis of 7 in CD₃NO₂con ta in in g D₂O. Complex

7 (15 mg) was dissolved in CD₃NO₂, and D₂O (2 µL) was added.

1

After thermolysis at 50 °C for 16 h, the H NMR spectrum of

the mixture was recorded. Signals attributable to 4-d₂and

free allyl alcohol indicated partial conversion. Further heating

for an additional 72 h resulted in ∼35% conversion to 4, as

evinced in the ¹H NMR spectrum of the solution.

P r ep a r a t ion of Cp *W(NO)(OCH ₂CH dCH ₂)(η²-H NdC-

(Me)CHdCP h ) (7). Complex 7 was prepared in a manner

identical with that for 5, except that allyl alcohol (280 µL, 16

equiv) was added via microsyringe in place of the water. The

final dark orange solution was concentrated and cooled twice

to obtain red needles of 7 in two crops (78 and 33 mg). The

crystals were isolated and dried in the usual fashion.

P r ep a r a t ion of Cp *W(NO)(H NC(dC(C₄H ₄))(Me))(η²-

NHdC(Me)CHdCP h ) (8). Complex 8 was prepared in a

manner identical with that for 5, except that excess CpH was

added via syringe in the place of water. The thermolyzed dark

red solution was concentrated and cooled to obtain 8 as dark

red needles (112 mg) after isolation and drying under high

vacuum.

Th er m olysis of 9 in THF -d ₈con ta in in g D₂O. Complex

9 (15 mg) was dissolved in THF-d₈, and D₂O (2 µL) was added

to the solution. After thermolysis at 50 °C for 16 h, the ¹H

NMR spectrum of the mixture was recorded. Only signals

attributable to 9 were evident, indicating no conversion to 4-d₂.

Th er m olysis of 1 in MeCN con ta in in g 2, 5, a n d 10

equ iv of H₂0. Reactant 1 (30.0 mg, 55.6 µmol) was dissolved

in dry MeCN (6.0 mL), and the resulting solution was divided

into three 2-mL aliquots which were placed into three glass

bombs. Two, five, and ten equivalents of water (0.7, 1.7, and

3.5 µL) were added to the first, second, and third bombs,

thereby affording solutions with [H₂O] ) 19, 47, and 95 mM,

respectively. Following thermolyses at 45 °C for 24 h, the

volatiles were removed from the final mixtures under high

vacuum and the red-orange residues were taken up in CDCl₃

(0.5 mL). The ¹H NMR spectra of the three resulting solutions

were recorded and compared to the ¹H NMR spectrum (CDCl₃)

of an equimolar mixture of authentic 11 and 4, revealing that

in only the case where [H₂O] ) 19 mM was 11 observed to

form, in approximately 5%. The formation of 4 was quantita-

tive for [H₂O] g 47 mM.

P r ep a r a t ion

of

Cp *W(NO)(η³-OC(Me)₂NdC(Me)-

CHdCP h ) (9). Complex 9 was prepared in a manner identical

with that for 8, except that excess acetone was added by

vacuum transfer onto a frozen MeCN solution of 1 (135 mg,

0.25 mmol). After thermolysis, the resultant red-brown solu-

tion was concentrated and cooled to obtain 9 as brown irregular

blocks (99 mg) after isolation and drying under high vacuum.

P r ep a r a t ion

of

Cp *W(NO)(η³-OC(Me)₂NdC(ⁱP r )-

CHdCP h ) (10). Complex 10 was prepared in a manner

analogous to that for 9. and was isolated in two crops as

brown-red blocks (102 mg) by concentration and cooling of the

final reaction mixture.

P r ep a r a t ion of Cp *W(NO)(η³-H NC(Me)dNC(dCH ₂)-

CHdCP h ) (11). Reactant 1 (270 mg, 0.5 mmol) was dissolved

in MeCN (20 mL) in a thick-walled bomb, and the solution

was heated for 24 h at 50 °C. The final blood-red solution was

Th er m olysis of 11 in THF -d ₈con ta in in g D₂O. An excess

of D₂O was added via microsyringe to a solution of pure 11

dissolved in THF-d₈. The solution was placed in a thermo-

stated water bath at 55 °C, and its ¹H NMR spectra were

recorded at regular intervals over a 48 h period. The resulting

Incorporation into a Tungsten Vinyl Fragment

Organometallics, Vol. 17, No. 5, 1998 871

spectra revealed quantitative conversion to 11-d₆by compari-

son to the proton spectra of authentic 11 and 11-d₆.

sions. We are grateful to the Natural Sciences and

Engineering Research Council of Canada for support of

this work in the form of grants to P.L. and a postgradu-

ate scholarship to S.A.L.

Th er m olysis of 11-d ₆a n d 12 in THF -d ₈. The crossover

experiment involving equimolar amounts of 11-d₆and 12

dissolved in THF-d₈was performed. An initial ¹H NMR

spectrum was recorded, and the solution was then placed in a

1

65 °C oil bath for 72 h, during which time its H NMR spectra

Su p p or tin g In for m a tion Ava ila ble: Thermal ellipsoid

plots for compounds 2, 4, 8, and 10 (Figures S1-S4) and tables

listing full crystallographic information, atomic coordinates

and U_eq, anisotropic thermal parameters, intramolecular bond

distances and angles (Tables S1-S24), and full NMR data

(Table S25) (39 pages). Ordering information is given on any

current masthead page.

were recorded at regular intervals. The progress of the

experiment was followed by comparison of the resultant

spectra to those for authentic 11 and 12 and revealed an

intermolecular exchange process between the acidic sites in

the two compounds.

Ack n ow led gm en t. We thank Prof. E. Carmona, Dr.

P. Daff, and Prof. R. Waymouth for stimulating discus-

OM971080R

Article Doi

Ligand elaboration mediated by a Cp*W(NO) template: Stepwise incorporation of small molecules into a tungsten vinyl fragment

DOI: 10.1021/om971080r

Source and publish data:

Authors:

Article abstract of DOI:10.1021/om971080r

Full text of DOI:10.1021/om971080r

Products guided by the article

R&D Labs maybe for 202531-06-4

Relevant to this article

Hot Product