Directly-observed β-hydrogen elimination of a late transition metal amido complex and unusual fate of imine byproducts

Full text of DOI:10.1021/ja961439n

7010
J. Am. Chem. Soc. 1996, 118, 7010-7011
Scheme 1
Directly-Observed â-Hydrogen Elimination of a
Late Transition Metal Amido Complex and Unusual
Fate of Imine Byproducts
John F. Hartwig
Department of Chemistry
Yale UniVersity, P. O. Box 208107
New HaVen, Connecticut 06520-8107
ReceiVed May 1, 1996
Amido complexes of low-valent platinum group metals
bearing â-hydrogens are rare, particularly for simple terminal
alkylamides.1-5 It is commonly believed that â-hydrogen
elimination reactions are rapid, precluding the isolation of such
compounds.6,7 However, the direct observation of â-hydrogen
elimination from a monomeric amido complex has not been
observed directly,3,8-10 and its microscopic reversesthe insertion
of imine into a metal hydride3sis rare despite its importance
in imine hydrogenations.11 We report here the synthesis and
full characterization of amido compounds with â-hydrogens
based on the classic Vaska’s complex that undergo slow,
directly-observable â-hydrogen elimination chemistry. The
mechanism for this apparently fundamental, but rarely observed,
reaction draws parallels to those for â-hydrogen eliminations
of late metal alkyls.12
and clearly defined their structures, except for the relative
orientation of the sec-Bu groups in dimeric 4. These spectro-
scopic data are included, along with analytical data, as sup-
porting information. As one might expect, the νCO values for
the alkylamides were lower than those for the monomeric
arylamides, reflecting the more electron rich character of
alkylamido complexes. The syn ligand geometry of 4 was
deduced from the presence of a single phosphine resonance in
the 31P{1H} NMR spectrum, but inequivalent i-Bu groups in
1
the H NMR spectrum.
The chemistry we report is summarized in Scheme 1.
Reaction of lithium amides with Vaska’s complex provided a
general route to monomeric terminal alkyl- or arylamido
complexes 1-3 with â-hydrogens.13 For example, reaction of
Vaska’s complex [Ir(PPh3)2(CO)Cl] with an excess of LiNPhMe,
LiN(Ph)CH2Ph, or 1 equiv of LiNHCH2CHMe2 gave the trans
terminal amido complexes 1-3 in isolated yields of 70%, 55%,
and 65%, respectively. In the case of the N-alkylarylamides,
exclusively monomeric materials were obtained. All three
complexes were obtained in analytically pure form by removal
of solvent, extraction of the solid into toluene, and cooling of
the resulting solution to -35 °C after addition of pentane. In
contrast to the formation of monomeric 1-3, reaction of an
excess of LiNHCH2CHMe2 with Vaska’s complex gave the
dimeric, bridging amido complex syn-[Ir(PPh3)(CO)(NHCH2-
CHMe2)]2 (4) as the only phosphine-containing metal product
observed by 31P NMR spectroscopy, along with PPh3. Complex
4 was isolated in 85% yield after crystallization from pentane.
The NMR and IR data for all complexes were straightforward
Thermal reaction of the N-benzylanilide complex 1 occurred
cleanly in toluene solvent at 110 °C and was complete after
1-3 h depending on added phosphine concentration (Vide infra)
to form the stable N-phenyltoluenimine along with the iridium
hydride 5, both in yields exceeding 95% by NMR spectroscopy
employing an internal standard. This reaction is a clear example
of â-hydrogen elimination and constitutes the first direct
observation of this transformation with a monomeric late metal
amido complex.
The N-methylanilide complex 2 was remarkably stable, and
the organic product marked an unusual outcome for the
â-hydrogen elimination chemistry. Complex 2 remained un-
changed for 2 h at 110 °C, but reacted over 1.5 d at 135 °C to
form the hydride 5 in 90% yield. The monomeric imine formed
from â-hydrogen elimination in this case would be highly
reactive and was not observed. Instead, the amidine PhNdC-
(H)NMePh14,15 was formed in 85% yield. This surprising
product was clearly identified by comparison of 1H NMR
spectra, GC retention times, and MS data to an authentic sample
prepared by addition of MeI to Li[PhNCHNPh]. The remainder
of the organic products were N-methylaniline, presumably from
hydrolysis of 2. Thermolysis of 2-d3 containing the amide
N(CD3)Ph led to formation of PhNdC(D)NCD3Ph and to the
formation of 5 containing 60-70% deuterium in the hydride
position,16 with the remaining deuterium incorporated into the
phosphine ligand, by reversible orthometalation of the phos-
phine. These labeling results confirmed that the metal hydride
was generated from the amide â-hydrogen. Methoxide com-
plexes can be dehydrogenated to form CO;17 the formation of
amidine can be envisioned as a combination of N-methylaniline
and phenyl isocyanide that result from formal disproportionation
of unstable methylideneaniline.
(1) Driver, M. S.; Hartwig, J. F. J. Am. Chem. Soc., in press.
(2) An exception is the activation of benzylamine: Glueck, D. S.;
Newman Winslow, L. J.; Bergman, R. G. Organometallics 1991, 10, 1462.
(3) Reversible imine insertion into a dimeric rhodium hydride has been
observed, and therefore partial â-hydrogen elimination of an isolated dimeric
amide has been observed: Fryzuk, M. D.; Piers, W. E. Organometallics
1990, 9, 986.
(4) Bryndza, H. E.; Fultz, W. C.; Tam, W. Organometallics 1985, 4,
939.
(5) Examples of a nickel dimethylamido complex were prepared over
20 years ago: Klein, H. F.; Karsch, H. H. Chem. Ber. 1973, 106, 2438.
(6) Fryzuk, M. D.; Montgomery, C. D. Coord. Chem. ReV. 1989, 95,
1-40.
(7) Bryndza, H.; Tam, W. Chem. ReV. 1988, 88, 1163.
(8) Hartwig, J. F.; Richards, S.; Baran˜ano, D.; Paul, F. J. Am. Chem.
Soc. 1995, 118, 3626.
(9) Bryndza, H. E.; Calabrese, J. C.; Marsi, M.; Roe, D. C.; W., T.;
Bercaw, J. E. J. Am. Chem. Soc. 1986, 108, 4805.
(10) An example of â-elimination from an alkoxide has been observed
directly: Blum, O.; Milstein, D. J. Am. Chem. Soc. 1995, 117, 4582.
(11) For a recent review see: James, B. R. Chem. Ind. 1995, 62, 167.
(12) Cross, R. J. In The Chemistry of the Metal-Carbon Bond; Hartley,
S. P. F. R., Ed.; John Wiley: New York, 1985; Vol. 2, p 559.
(13) Reaction of t-BuNHLi with Vaska’s complex does not give
monomeric amido complexes similar to 1-3, while a mixture of monomeric
and dimeric products was obtained from reaction of t-BuNHLi with the
triethylphosphine version of Vaska’s complex: Rahim, M.; Bushweller,
C. H.; Ahmed, K. J. Organometallics 1994, 13, 4952.
(14) Bredereck, H.; Gompper, R.; Klemm, K.; Rempfer, H. Chem. Ber.
1959, 92, 837.
(15) Oszczapowicz, J.; Kuminska, M. J. Chem. Soc., Dalton Trans. 1994,
103.
(16) These data were determined by 1H and 2H NMR spectrometry. The
amount of deterium in the hydride position was time-dependent, as it
eventually washed completely into the phosphine ligand.
(17) Bryndza, H. E.; Kretchmar, S. A.; Tulip, T. H. J. Chem. Soc., Chem.
Commun. 1985, 977.
S0002-7863(96)01439-4 CCC: $12.00 © 1996 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 118, No. 29, 1996 7011
Scheme 2
displacement of the imine with 2 equiv of PPh3. The most
common pathway for â-hydrogen elimination of late metal alkyl
complexes involves a similar formation of an unsaturated
intermediate that provides a binding site for the resulting
alkene.12,19 The lack of solvent effect suggests that the presence
of the nitrogen heteroatom does not lead to a transition state
with significant charge buildup.
The â-hydrogen elimination chemistry of 2 was inhibited by
added phosphine, although the effect was not measured quan-
titatively as for 1, and may not be a simple inverse first-order
dependence. In contrast, the rate of thermal reaction of 3 was
independent of the concentration of added phosphine. Conver-
sions of 3 to 1 at various reaction times were identical for
reactions conducted in the presence of a 5-fold difference of
added [PPh3]. It is not necessary to invoke a mechanism for
the initial stages of the reaction of 3 that is different from the
one deduced for 1. Rate-determining dissociation of PPh3 or
reversible â-hydrogen elimination followed by associative imine
displacement is consistent with the reaction orders.
It was surprising that such high temperatures were necessary
for the â-hydrogen elimination processes to occur, particularly
the slow reaction of 1 that forms an unusually stable imine.
The analogous alkyl complex [Ir(PPh3)2(CO)(octyl)] generated
in situ from Vaska’s complex and octyllithium produces octene
at room temperature.20,21 Thus, â-hydrogen elimination of late
metal amides can be much slower than elimination of the
corresponding alkyl complexes.
We have shown previously that â-hydrogen elimination
reactions of palladium(II) amides can be more facile than those
directly observed with these Ir(I) compounds and are important
in the catalytic selectivity of late transition metal amides.8
â-Hydrogen elimination can be minimized by the use of bulky
phosphine ligands, as in catalytic amination of aryl halides.22-27
Alternatively, chelating ligands would minimize â-hydrogen
elimination of late metal amides by reducing the equilibrium
for phosphine dissociation. In addition to amination chemistry,
these results are relevant to imine hydrogenation, as noted in
the introduction. Since the â-hydrogen elimination chemistry
of these Vaska-type systems is more rapid from monomers,
imine insertion clearly can involve a single metal center, despite
the fact that some of the best catalysts for asymmetric imine
hydrogenation are dimeric,28 and the only directly observed
imine insertion involves a bridging hydride to form a bridging
amido product.3
The isobutyl complex 3 also formed unusual organic products
for â-hydrogen elimination processes. Compound 3 was less
stable than the arylamido complexes, and thermal chemistry
occurred cleanly at 70 °C. In this case, 0.5 equiv of isobuty-
lamine and 0.5 equiv of isobutyronitrile were formed in a
combined yield that was 90%, along with hydride 5 in 85%
yield, again determined by 1H NMR spectroscopy with an
internal standard. The organic reaction products were identified
1
by comparison of H NMR spectra to those of commercial
materials and reflect a formal disproportionation of the unstable
isobutanimine into amine and nitrile. This transformation
resembles a Cannizarro reaction in which aldehydes undergo
hydroxide-induced formation of carboxylic acid and alcohol.18
Complex 3 must not form dimeric 4 during this â-elimination
chemistry. Complex 4 showed less than 5% conversion during
the time that led to complete conversion of 3, while complete
conversion of 4 after 4 h at 90 °C gave only 50% yield of 5,
35% yield of amine, and less than 5% nitrile.
The mechanism of the simple â-hydrogen elimination from
N-benzylanilide 1 to form N-phenyltoluenimine and hydride 5
was probed by kinetic methods. Reaction rates were measured
at 110 °C by monitoring the decay of the benzylic resonance in
the 1H NMR spectra. Four possible mechanisms for the
â-hydrogen elimination are shown in Scheme 2. Reaction by
pathway A would be first order in 1 and zero order in phosphine
concentration. This mechanism would also be expected to show
a pronounced solvent effect. Pathway B would, again, provide
reactions that would be first order in iridium amide and zero
order in phosphine concentration, but would be unlikely to
provide a significant solvent effect. Reaction by pathway C
would exhibit inverse second-order behavior in phosphine
concentration and second-order behavior in amide 1. Reaction
by pathway D would be first order in 1 and inverse first order
in phosphine concentration.
Acknowledgment. The author gratefully acknowledges support
from an NSF Young Investigator Award, a DuPont Young Faculty
Award, a Union Carbide Innovative Recognition Award, Rhom and
Haas, and Bayer and Johnson Matthey for a loan of iridium trichloride.
The author is a Fellow of the Alfred P. Sloan Foundation.
Observed rate constants were measured with concentrations
of added phosphine (PPh3-d15) that ranged from 2.3 to 9.5 mM.
Linear first-order plots for the decay of N-benzylanilide 1 were
obtained, indicating first-order behavior in 1. Observed rate
constants varied by less than 5% for reactions involving initial
concentrations of 1 that varied by a factor of 3, confirming the
first-order behavior in 1. The kinetic behavior in added
phosphine was clearly inverse first order and the reaction
followed one major pathway. A plot of ln kobs vs ln [PPh3-d15]
gave a slope of -1.1 ( 0.1, and a plot of 1/kobs vs [PPh3-d15]
was linear with a y intercept within experimental error of zero.
Although determined qualitatively by 31P NMR spectroscopy,
reaction rates in C6D6 and THF were essentially identical.
These results are consistent with only pathway D, involving
a 14-electron intermediate that undergoes â-hydrogen elimina-
tion to presumably form an imine complex before final
Supporting Information Available: Spectroscopic and analytical
data for 1-4, along with representative plots of kinetic data (5 pages).
See any current masthead page for ordering and Internet access
instructions.
JA961439N
(19) Whitesides, G. M.; Gaasch, J. F.; Stedronsky, E. R. J. Am. Chem.
Soc. 1972, 94, 5228.
(20) Evans, J.; Schwartz, J.; Urquhart, P. W. J. Organomet. Chem. 1974,
81, C37.
(21) Cannon, J. B.; Schwartz, J. J. Am. Chem. Soc. 1974, 96, 2276.
(22) Louie, J.; Paul, F.; Hartwig, J. F. Organometallics, in press.
(23) A tin free system has also been developed. Louie, J.; Hartwig, J. F.
Tetrahedron Lett. 1995, 36, 3609.
(24) Paul, F.; Patt, J.; Hartwig, J. F. J. Am. Chem. Soc. 1994, 116, 5969.
(25) Driver, M. S.; Hartwig, J. F. J. Am. Chem. Soc. 1995, 117, 4708.
(26) Guram, A. S.; Buchwald, S. L. J. Am. Chem. Soc. 1994, 116, 7901.
(27) Guram, A. S.; Rennels, R. A.; Buchwald, S. L. Angew. Chem., Int.
Ed. Engl. 1995, 34, 1348.
(18) March, J. AdVanced Organic Chemistry, 3rd ed.; John Wiley and
Sons: New York, 1985.
(28) Chan, Y. N. C.; Osborn, J. A. J. Am. Chem. Soc. 1990, 1990, 9400.