Full Paper
lished the accumulation of electron density on the phospho-
rous atom in the transition state for racemization relative to
the ground state h2-allene complex. Together, these data es-
tablish the net transfer of electron density from the allene to
the (L)Au moiety in the conversion of ground state to the tran-
sition state for racemization.
Analysis of solvent effects provides further insight into the
distribution of charge in the transition state for allene racemi-
zation. Because racemization occurs without the creation or
consumption of charge, the strong dependence of the rate of
racemization of 1-aryl-1,2-butadienes on solvent polarity (in
particular the ꢁ50-fold increase in reaction rate for racemiza-
tion of (R)-2b in 1,2-dichloroethane relative to toluene) points
to the localization of positive charge in the transition state rel-
ative to the ground state.[42] More specifically, these observa-
tions suggest that in the ground state h2-allene complex, posi-
tive charge is dispersed across the gold atom and phosphine
and allene ligands and becomes localized on the C1 allene ter-
minus in the transition state for racemization. In support of the
former contention, computational analysis of gold p-com-
plexes {(Ph3P)Au[h2-H2C=C(H)-4-C6H4Me]}+ and [(Ph3P)Au(h2-
MeCꢄCMe)]+ indicates more than 60% of the positive charge
resides on the phosphine and p-ligands.[43] The significantly at-
tenuated effect of solvent polarity on the rate of racemization
of 1-phenyl-1,2-propadienes relative to 1-aryl-1,2-butadienes
points to greater delocalization of positive charge in the transi-
tion state for racemization of 1-phenyl-1,2-propadienes relative
to 1-aryl-1,2-butadienes. This conclusion is in accord with the
attenuated reaction constant for the (Ph3P)AuOTf-catalyzed
racemization of 1-phenyl-1,2-propadienes (1+ =ꢀ1.5) relative
to 1-aryl-1,2-butadienes (1+ =ꢀ2.7), which established more
equal distribution of positive charge between the C1 and C3
allene carbon atoms in the transition state for racemization of
1-phenyl-1,2-propadienes relative to 1-aryl-1,2-butadienes.
The diminished reaction constant for racemization of 1-
phenyl-3-arylpropadienes (1+ =ꢀ1.5) relative to 1-aryl-1,2-bu-
tadienes (1+ =ꢀ2.7–ꢀ2.8) established the lower electron
demand of the aryl-bound allene carbon atom in the transition
state for racemization of 1-phenyl-3-arylpropadienes relative to
1-aryl-1,2-butadienes. This observation suggests that distribu-
tion of positive charge between the terminal allene carbon
atoms in transition state TS1 is varied and depends on the rel-
ative electron-releasing abilities of the allene C1 and C3 sub-
stituents. It is worth noting that the second-order rate con-
stants (krac) for racemization of (R)-3b and (R)-2a, both of
which contain a terminal p-bromphenyl substituent, differed
by only approximately 20% despite the superior electron-
donor properties of the phenyl group of (R)-3b relative to the
methyl group of (R)-2a. This observation points to steric desta-
bilization of transition state TS1 by the terminal phenyl group
of (R)-3b relative to the terminal methyl substituent of (R)-2a.
The reaction constants determined for the gold-catalyzed
racemization of 1-aryl-1,2-butadienes and 1,3-diaryl-1,2-propa-
dienes are less negative than those determined for the hydro-
chlorination of 1-aryl-1,2-propadienes (1+ =ꢀ4.20)[40] and 1-
aryl-1,3-butadienes (1+ =ꢀ2.98)[38] in glacial acetic acid and for
the solvolysis of 1,3-diaryl-3-chloro-1-propenes (1+ =ꢀ3.1)[41] in
60% aqueous acetone, particularly considering the nonpolar
reaction medium employed for gold-catalyzed allene racemiza-
tion. These comparisons suggest that significant positive
charge resides on the allene ligand of the ground state h2-
allene complex for allene racemization. Indeed, the negative
reaction constant (1=ꢀ1.4) determined for the equilibrium
binding affinities of 1-aryl-1,2-butadienes to the (P1)Au+ frag-
ment established depletion of electron density from the C1
atom of bound allene relative to free allene. Furthermore, our
previous determination of a reaction constant of 1=ꢀ3.4 for
binding of p-substituted vinyl arenes to the (P1)Au+ fragment
suggests the reaction constant 1=ꢀ1.4 significantly under es-
timates the extent of electron depletion from the allene upon
binding to gold.[17]
Relationship between the h1-allene species involved in
stereomutation and p-face exchange
Variable temperature NMR analysis of aliphatic gold p-com-
plexes established facile intramolecular p-face exchange of the
allene ligand without stereomutation.[12] To account for this be-
havior, we invoked the involvement of chiral, h1-allene inter-
mediates or transition states III in which gold is bound 458 rel-
ative to the allene axes, presumably interacting with both p-or-
bitals of the central carbon atom (Scheme 4). Similar chiral h1-
allene species have been invoked to account for p-face ex-
change in platinum[44] and iron[33,45] p-allene complexes and
gold h1-allene species have been evaluated computationally.[8,9]
Germane to the present study, analysis of the 4,5-nonadiene
complexes 1a and 1b revealed two discrete p-face exchange
pathways associated with two diastreomeric h1-allene inter-
mediates/transition states that lay well below the transition
state for allene stereomutation (DG° =ꢁ17.4 kcalmolꢀ1):[12]
a lower energy process (DG° =8.8 kcalmolꢀ1) involving either
the trans,trans-III or cis,cis-III that interconverts the allenyl pro-
tons of the more stable h2-allene stereoisomer (cis-5 or trans-5)
and a higher energy process (DG° =9.7 kcalmolꢀ1) involving
cis,trans-III that interconverts the h2-allene diastereomers cis-5
and trans-5 (Scheme 4).[12] Because these two processes
together led to complete interconversion of all four allene
protons of both cis-5 and trans-5, we have no information
regarding the energetics of the third h1-allene diastereomer
(trans,trans-III or cis,cis-III).
The observations noted in the preceding paragraph point to
the central role of p-activation in gold-catalyzed allene racemi-
zation, a contention that is further supported by the large pos-
itive reaction constant (1=2.5) for the racemization of (R)-2b
catalyzed by p-substituted triarylphosphine gold complexes. In
this regard, it is worth noting that we have previously docu-
mented the markedly greater electrophilicity of the twelve-
electron (P1)Au+ fragment relative to cationic AgI,[20] PdII,[21]
and PtII[22] complexes through analysis of vinyl arene relative
binding affinities.[17] Therefore, to the extent that allene racemi-
zation reactivity tracks with the electrophilicity of the metal
complex, these data provide a rationale for the high reactivity
of cationic gold(I) complexes relative to related late transition
metal complexes as allene racemization catalysts.
Chem. Eur. J. 2014, 20, 12245 – 12254
12250
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim