Retention of configuration and regiochemistry in allylic alkylations via the memory effect

Article abstract of DOI:10.1021/om0306224

The results of a systematic variation of achiral monodentate ligands used for the Pdcatalyzed alkylation of crotyl and 3-buten-2-yl acetates and carbonates are presented. A significant degree of retention of regiochemistry and stereochemistry was observed for these substrates, especially for nonracemic 3-buten-2-yl ethyl carbonate with PCy_xPh_s-x and carbene ligands.

Full text of DOI:10.1021/om0306224

Organometallics 2004, 23, 2179-2185
2179
Reten tion of Con figu r a tion a n d Regioch em istr y in
Allylic Alk yla tion s via th e Mem or y Effect
J . W. Faller* and Nikos Sarantopoulos
Department of Chemistry, Yale University, New Haven, Connecticut 06520-8107
Received October 1, 2003
The results of a systematic variation of achiral monodentate ligands used for the Pd-
catalyzed alkylation of crotyl and 3-buten-2-yl acetates and carbonates are presented. A
significant degree of retention of regiochemistry and stereochemistry was observed for these
substrates, especially for nonracemic 3-buten-2-yl ethyl carbonate with PCyxPh3-x and
carbene ligands.
Sch em e 1. Mo-Ca ta lyzed En a n tioselective
Alk yla tion of 2a
In tr od u ction
A number of different metals have been successfully
employed as catalysts for the alkylation of allylic
substrates1 and have been shown to promote the reac-
tion with varying regioselectivity. Iridium and rhodium
complexes have recently been used when substitution
on the more substituted allyl terminus was desired.2,3
Furthermore, molybdenum-based catalysts have given
excellent results with diaminocyclohexane-based lig-
ands,4,5 such as 1, allowing the regio- and enantiose-
a
The branched isomer (4) is produced selectively: 4:3 ) 49:
1.
lective alkylation of aryl allyl carbonates 2 (Scheme 1)
with the production of a preponderance of the branched
isomer.
Sch em e 2. Mem or y Effect Obser ved Wh en Usin g
P Cy3 a s a Liga n d for P a lla d iu m , P r od u cin g a
P r ep on d er a n ce of 7 (7:6 ) 92:8)
On the other hand, palladium-based catalysts are
usually chosen when alkylation at the less substituted
terminus of an allyl is desired.6 There are cases,
however, where this selectivity can be changed, par-
ticularly by ligand modification, such that the opposite
regioselectivity is observed.7-9 More specifically, the use
(1) Pfaltz, A.; Lautens, M. Comprehensive Asymmetric Catalysis II;
J acobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer: Berlin, 1999;
Chapter 24, pp 833-884.
(2) Takeuchi, R.; Kashio, M. J . Am. Chem. Soc. 1998, 120, 8647-
8655.
(3) (a) Evans, P. A.; Nelson, J . D. Tetrahedron Lett. 1998, 39, 1725-
1728. (b) Evans, P. A.; Nelson, J . A. J . Am. Chem. Soc. 1998, 120,
5581-5582. (c) Evans, P. A.; Kennedy, L. J . Org. Lett. 2000, 2, 2213-
2215.
of PCy3 as a ligand for the Pd-catalyzed alkylation of
racemic 3-buten-2-yl acetate (5) has been shown to
promote the formation of the branched product 7
(Scheme 2).10 This is an example of a “memory effect”,
wherein the original site of attachment of the leaving
group plays a significant role in determining the regio-
chemistry of the product. This contrasts with a pattern
that is frequently seen for palladium catalysts with
other phosphorus donor ligands, particularly chelates,
(4) Trost, B. M.; Hachiya, I. J . Am. Chem. Soc. 1998, 120, 1104-
1105.
(5) Glorius, F.; Pfaltz, A. Org. Lett. 1999, 1, 141-144.
(6) (a) Trost, B. M.; Van Vranken, D. L. Chem. Rev. 1996, 96, 395-
422. (b) Trost, B. M.; Crawley, M. L. Chem. Rev. 2003, 103, 2921-
2943.
(7) Hayashi, T.; Kawatsura, M.; Uozumi, Y. Chem. Commun. 1997,
561-562.
(8) (a) Lloyd-J ones, G. C.; Stephen, S. C. Chem. Eur. J . 1998, 4,
2539-2549. (b) Lloyd-J ones, G. C.; Stephen, S. C.; Murray, M.; Butts,
C. P.; Vyskocil, S.; Kocovsky, P. Chem. Eur. J . 2000, 6, 4348-4357.
(c) Note that anionic intermediates have also been implicated by:
Cantat, T.; Genin, E.; Giroud, C.; Meyer, G.; J utand, A. J . Organomet.
Chem. 2003, 687, 365-376.
(9) (a) Hilgraf, R.; Pfaltz, A. Synlett 1999, 1814-1816. (b) Pretot,
R.; Pfaltz, A. Angew. Chem., Int. Ed. 1998, 37, 323-325. (c) You, S.
L.; Zhu, X. Z.; Luo, Y. M.; Hou, X. L.; Dai, L. X. J . Am. Chem. Soc.
2001, 123, 7471-7472.
(10) Blacker, A. J .; Clarke, M. L.; Loft, M. S.; Williams, J . M. J .
Org. Lett. 1999, 1, 1969-1971.
10.1021/om0306224 CCC: $27.50 © 2004 American Chemical Society
Publication on Web 03/30/2004

2180 Organometallics, Vol. 23, No. 9, 2004
Faller and Sarantopoulos
Sch em e 3. Mem or y Effect Obser ved Wh en Usin g
(t-Bu )2NHC a s a Liga n d for P d , P r od u cin g a n
Excess of 7
Sch em e 5. Ster eoch em ica l Mem or y Effect Usin g
P Cy3 a s a Liga n d for P d a
a
The branched isomer is formed in excess (87:13) with
predominant retention of configuration.
Ta ble 1. Ster eoch em ica l Mem or y Effect Usin g a
P Cy3/P d Ca ta lyst
Sch em e 4. P r od u cts Obser ved Wh en Usin g
(t-Bu )2NHC a s a Liga n d for P d , P r od u cin g a n
Excess of th e Lin ea r P r od u ct
ee of 8 (R), %
PCy3:Pd
nucleophile:8
ee of 7 (R). %
89
91
91
91
2
2
3
2
1.5:1
1.5:1
1.5:1
3.0:1
79
82
72
82
reaction shown in Scheme 2;11 however, no comment
about its significance was made at that time. We sought
to investigate this phenomenon in analogous reactions
such as that shown in Scheme 3, and thus, we required
a nonracemic branched carbonate as a substrate for
these studies.
Compound (R)-8 can be prepared in high enantiomeric
purity from the commercially available, though expen-
sive, (R)-3-buten-2-ol. We also found that nonracemic 8
could be prepared via kinetic resolution (see Experi-
mental Section). With nonracemic 8 in hand, we sought
to investigate the stereochemistry of the memory effect
in the palladium-catalyzed reaction.
where the original point of attachment of the leaving
group is irrelevant and regioselectivity is controlled by
the ligand set. The absence of a memory effect is an
important feature in the operation of asymmetric ally-
lation catalysts, as stereocontrol based on the position
of attachment of the leaving group, rather than the
influence of the chiral ligands, would reduce the enan-
tioselectivity of the catalyst.
We examined the palladium-catalyzed alkylation of
the allylic carbonate 8 using PCy3 as the ligand. When
(R)-(+)-8 was utilized, we found that the reaction yielded
(R)-(+)-7 (Scheme 5). Table 1 summarizes the effects of
changing the nucleophile or the PCy3 loading on the ee
of the alkylated product.
We also examined the effect of using less bulky
phosphines containing both phenyl and cyclohexyl groups.
The results are summarized in Table 2. Generally, Ph3P-
containing catalysts are presumed to not have a large
memory effect;6 however, we have observed that if the
reactions are run at room temperature or below there
can be a significant regiochemical memory effect that
is more pronounced with carbonates relative to acetates.
One should note that the crotyl carbonate still yields a
significant amount of branched product (41%), whereas
we observed that crotyl acetate shows only 17% branched
product.
Resu lts
We have observed a memory effect similar to that
shown in Scheme 3 for the carbonate analogue of 5 with
Pd/PCy3 (which yields 7:6 ) 87:13). We have also found
that a modest memory effect can be observed using the
carbonate analogue of 5 as a substrate and N-hetero-
cyclic carbenes (NHC’s) as ligands and for which a 7:6
ratio of 56:44 was observed for (t-Bu)2-NHC and 57:43
was observed for (i-Pr2C6H3)2-NHC (Schemes 3 and 4).
This contrasts with the result using the linear carbonate
and (t-Bu)2-NHC, which gave a 7:6 ratio of 34:66.
It is reasonable to expect that when a regiochemical
memory effect exists, it could also involve a memory
effect in terms of stereochemistry. Thus, there are two
facets to this phenomenon: (1) a regiochemical memory
effect and (2) a stereochemical memory effect. One
isolated instance of partial chirality retention was
previously observed with a nonracemic acetate in the
(11) Acemoglu, L.; Williams, J . M. J . Adv. Synth. Catal. 2001, 343,
75-77.

Allylic Alkylations
Organometallics, Vol. 23, No. 9, 2004 2181
Sch em e 6. Ra cem iza tion via η3-η1-η3 Allyl
In ter con ver sion a n d Cis-Tr a n s Isom er iza tion of
a n η3-Allyl In ter m ed ia te
Ta ble 2. Com p a r ison of Mem or y Effects Obser ved
w ith Su bstr a te 8 a n d Sh ow n by Differ en t
P h osp h in es
ee (R) of
substrate, % T, °C
phosphine ligand
(10%)
ee (R) of
product, %
B:La
0
0
0
0
0
0
0
0
-78b
-24
23
65
23
-15
23
23
23
PPh3
PPh3
PPh3
PPh3
PCyPh2
PCyPh2
PCy2Ph
PCy2biphenyl
PCyPh2
PCy2Ph
PCy3
62:38
48:52
45:55
39:61
72:28
70:30
81:19
61:39
69:31
81:19
87:13
0
0
0
0
0
0
0
0
neither retention of regiochemistry nor retention of
stereochemistry was observed.
91
91
91
25
79
82
23
25
The Na[CH3C(CO2Me)2] reaction, however, is decep-
tive, since the bulkiness of the nucleophile biases the
attack on the less hindered terminus and tends to yield
linear product. We found that, on carrying out the
reaction with Na[HC(CO2Me)2] in thf in the presence
of PPh3 (2:1 P:Pd), crotyl acetate (9) and 3-buten-2-yl
acetate (5) yielded linear-to-branched ratios of 83:17 and
58:42. respectively, which indicate a definite regio-
chemical memory effect for PPh3-containing catalysts.
In the presence of the bulky monodentate ligand
2-(diphenylphosphino)-2′-methoxy-1,1′-binaphthyl (MeO-
MOP), Hayashi found that a greater retention of regio-
chemistry could be observed in alkylations. Thus, crotyl
acetate and 3-buten-2-yl acetate yielded linear-to-
branched ratios of 95:5 and 38:62, respectively, in the
presence of 2:1 MeO-MOP:Pd.13 This was attributed to
the large cone angle15 (>210°) of MeO-MOP, which
would favor intermediate complexes containing one
phosphine. Even though an excess of phosphine was
present in this case, only the neutral (allyl)Pd(L)X was
formed. The excess MeO-MOP was not effective in
promoting cis-trans isomerization. The cone angle of
Cy3P is still large (170°)15 but not as large as that of
MeO-MOP, and so a similar mechanism presumably
holds. Generally one would not expect a carbene to
dissociate easily, so that a bulky carbene should behave
similarly.
a
The reactions were performed in thf with 1.5 equiv of
b
NaHC(CO2Me)2. Held at -78 °C for 4 days and then warmed to
23 °C.
Ta ble 3. Ster eoch em ica l Mem or y Effect Usin g
NHC/P d Ca ta lysta
ee of 8 (R), % NHC:Pd
NHC
ee of 7 (R), %
6:7
91
91
1
1
(t-Bu)2NHC
(i-Pr2Ph)2NHC
68
66
61:39
66:34
a
100% conversion was observed after 20 h of heating under
reflux in thf.
Using the in situ generated carbene following the
procedure of Sato et al.,12 we observed significant
retention of configuration indicating a high stereochem-
ical but only a modest regiochemical memory effect
(Table 3).
When oxidative addition occurs, Hayashi suggests
that the leaving group (OAc in his case) remains cis to
the carbon to which it was originally attached. If cis-
trans isomerism is slow, then attack occurring trans to
the phosphine16-18 should result in the regiochemistry
being retained.
This also could provide a rationale for the retention
of regiochemistry and stereochemistry with 8 and PCy3
and other bulky phosphines, if η3-η1-η3 allyl intercon-
version at the unsubstituted terminus were slow. This
is unexpected because, if the phosphine were trans to
the methyl, it would promote racemization via an η1-
crotyl group, as shown in Scheme 6. A phosphine
generally weakens the bond trans to it and promotes
formation of an η1-crotyl group cis to the phosphine
ligand. On the other hand, if the phosphine were cis to
the methyl, a σ-bond would be formed at the more
Discu ssion
For palladium-catalyzed reactions, the usual pre-
sumption is that the regiochemistry of the product is
independent of the isomeric nature of the substrate. For
example, in the presence of PPh3 (2:1 P:Pd), crotyl
acetate (9) and 3-buten-2-yl acetate (5) both yield a
similar linear-to-branched ratio of 81:19 and 82:18,
respectively, when the nucleophile was [CH3C(CO2-
Me)2]-.13 The position of attachment of the leaving group
in this case appears to be generally irrelevant, owing
to the rapid formation of the same distribution of
reactive η3-allylic intermediates after oxidative addition
prior to nucleophilic attack. An η3-η1-η3 allyl inter-
conversion often provides a pathway for racemization
of an allylic palladium intermediate (Scheme 6).14 Thus,
(14) Faller, J . W.; Thomsen, M. E.; Mattina, M. J . J . Am. Chem.
Soc. 1971, 93, 2642.
(15) Tolman, C. A. Chem. Rev. 1977, 77, 313.
(16) (a) Pfaltz, A. Acta Chem. Scand. 1996, 50, 189. (b) Brown, J .
M.; Hulmes, D. I.; Guiry, P. J . Tetrahedron 1994, 50, 4493.
(17) J onasson, C.; Kritikos, M.; Backvall, J .-E.; Szabo, K. J . Chem.
Eur. J . 2000, 6, 432.
(18) Faller, J . W.; Sarantopoulos, N. Unpublished X-ray crystal-
lographic results, manuscript in preparation.
(12) Sato, Y.; Yoshino, T.; Mori, M. Org. Lett. 2003, 5, 31-33.
(13) Hayashi, T.; Kawatsura, M. Uozumi, Y. J . Am. Chem. Soc. 1998,
120, 1681-1687.

2182 Organometallics, Vol. 23, No. 9, 2004
Faller and Sarantopoulos
Sch em e 7. Syn -An ti Isom er iza tion via η3-η1-η3
Cl (11) which have Pd-C bond lengths of 2.21 and 2.16
Å trans to NHC and 2.10 and 2.10 Å trans to Cl,
respectively. We feel that our values for these distances
in the t-Bu2NHC complex may be more reliable, because
the thermal parameters for the central carbon in the
(i-Pr2Ph)2NHC complex suggest some disorder. (crotyl)-
Pd(Cy2P-2-biphenyl)Cl (12) has Pd-C bond lengths of
2.21 Å trans to P and 2.11 Å trans to Cl.
Needless to say, the reaction pathway is complicated.
One might also note that it appears that the distribution
of the regioisomers of the products is affected somewhat
by the optical purity of the starting material (Table 2
and Scheme 3). For example, note the results for
CyPPh2. These results were reproducible when the
reactions were repeated, so we do not believe this is an
artifact, although there is the possibility that impuri-
ties are the basis of the difference. We are continuing
to investigate this phenomenon, which could possibly
be accounted for by still having product bound in an
intermediate when substrate attacks. This observation
needs further study but does not significantly change
our overall conclusions.
Allyl In ter con ver sion
substituted terminus. This does not provide a path for
racemization but only isomerization. Since the syn
isomer is thermodynamically more stable,13 this path
would only provide for a small concentration of the anti
isomer and the configuration at carbon remains the
same (Scheme 7). The most important feature, however,
is that an excess of a bulky phosphine does not promote
a kinetically important alternative route for racemiza-
tion. There is significant interaction between the phos-
phine and the methyl so that, given the possibility of
reaching cis-trans equilibrium, a single isomer (>90%)
would predominate. We found by NMR that (crotyl)Pd-
(Cy3P)Cl is a single isomer with the P trans to the
methyl at equilibrium. In contrast to the results with
other less bulky phosphines, a 1.3:1 ratio of 7:6 is
observed with crotyl acetate;10 hence, the Cy3P promotes
formation of the branched isomer, even though a
memory effect would suggest a predominance of the
linear isomer. This may be partially attributed to the
directing influence of the trans phosphine in competition
with a steric preference for adding to the least substi-
tuted terminus.
When structural features do not allow isomerization
or racemization of the allyl, retention of the stereochem-
istry at the alkylated carbon is typical for palladium-
catalyzed allylic alkylations in medium-polarity solvents
such as THF. Thus, the results are consistent with
retention of configuration via either double inversion
or double retention and a lack of significant racemiza-
tion in the intermediate. This suggests that capture of
the allyl intermediate occurs before racemization can
occur. Although Hayashi stresses the importance of the
neutral (allyl)Pd(L)X intermediate,14 others have sug-
gested that a more reactive ion-paired intermediate is
more important.8 With the exception of bulky aliphatic
phosphines such as Cy3P, memory effect studies have
focused on phosphines with aromatic substitution. It
would appear that the Cy3P has an ideal combination
of (1) bulk that prevents rapid isomerization via a
second-order attack on an allyl intermediate by a second
phosphine, (2) electronic properties that provide for slow
isomerization and racemization, and (3) basicity to
promote rapid oxidative addition of the carbonate.
Although it has been suggested that a bis MeO-MOP
Pd complex is unlikely, owing to steric effects,14 we have
isolated (Cy3P)2PdCl(CH2Cl) from mixtures of [(allyl)-
PdCl]2 and excess PCy3 in CH2Cl2.18 This attests to the
ease of oxidative addition of (Cy3P)nPd0 and the poten-
tial stability of a bis Cy3P complex that could be involved
in the catalytic cycle. The Pd-C bond lengths we have
observed in (allyl)Pd(PCy3)Cl (10) of 2.19 Å trans to P
and 2.11 Å trans to Cl suggests that attack should occur
trans to P in analogous compounds based on studies
indicating that attack occurs at the more weakly bound
carbon atom.17,18 Similar conclusions can be drawn from
(allyl)Pd[(i-Pr2Ph)2NHC]Cl19 and (allyl)Pd(t-Bu2NHC)-
Mem or y Effects in P er sp ective. Bosnich20,21 has
provided one of the most comprehensible explanations
for the lack of memory effects in many situations. He
has elegantly outlined the factors that determine the
behavior of a Pd/(S,S)-chiraphos based allylic alkylation
catalyst, and the two key features of his discussion are
(1) the fast addition of the allylic substrate on Pd and
(2) the rapid equilibration of intermediate Pd complexes
that are formed by oxidative addition of the regioiso-
meric substrates to Pd. The equilibration of a common
π-allyl intermediate must occur rapidly relative to the
rate at which the nucleophile attacks. The first argu-
ment concerns the reactivity of the Pd0/(S,S)-chiraphos
species toward oxidative addition, while the second
argument is the one that requires a single π-allyl
intermediate, which subsequently undergoes alkylation.
Provided that the rates of the η3-η1-η3 rearrangements
of a Pd allyl complex which provide the equilibration of
the intermediates are dictated by the nature of the
ligands, one could argue that ligands such as PCy3 have
the necessary steric and electronic characteristics that
produce relative rates that allow deviations from the
demands of the Bosnich model.
Fiaud and Malleron22 reported the first example of a
substrate-dependent distribution of products in the
catalytic alkylation of allylic substrates in 1981. Their
results showed that (1) the enantiomeric purity of the
substrate played a role in determining the enantiomeric
purity of the product and (2) enantiomeric catalysts gave
products of different enantiopurity when a partially
optically resolved substrate was used. Since then, a
limited number of reports have appeared concerning
similar observations, although they were not always
explicitly described as a memory effect. Most of them
concern Pd-based catalytic systems,7,10-12,23-27 but this
seems to be a more widespread behavior than one might
(20) Auburn, P. R.; Mackenzie, P. B.; Bosnich, B. J . Am. Chem. Soc.
1985, 107, 2033-2046.
(21) Mackenzie, P. B.; Whelan, J .; Bosnich, B. J . Am. Chem. Soc.
1985, 107, 2046-2054.
(22) Fiaud, J . C.; Malleron, J . L. Tetrahedron Lett. 1981, 22, 1399-
1402.
(19) Viciu, M. S.; Germaneau, R. F.; Navarro-Fernandez, O.; Stevens,
E. D.; Nolan, S. P. Organometallics 2002, 21, 5470-5472.
(23) Lloyd-J ones, G. C.; Stephen, S. C. Chem. Eur. J . 1998, 4, 2539-
2549.

Allylic Alkylations
Organometallics, Vol. 23, No. 9, 2004 2183
Sch em e 8. Kin etic Resolu tion of th e Ra cem ic
Ca r bon a te 8
originally have anticipated, and it extends to other
metals such as Ru,28 W,29 Rh,3,30,31 Ir,32,33 and Pt.34
There are also more complicated rationalizations of the
memory effect that involve the active participation of
the sodium ion.25,26
In conclusion, bulky phosphines are not the only class
of ligands that can promote a memory effect in the Pd-
catalyzed alkylation of open allylic substrates. N-het-
erocyclic carbenes are also capable of demonstrating a
regioselective memory effect, favoring the formation of
the branched product when the branched substrate is
employed. Although N-heterocyclic carbenes also show
a stereochemical memory effect, this is less pronounced
than in the case of Cy3P and other cyclohexyl-substi-
tuted phosphines. For Cy3P, the use of the ethyl carbon-
ate leaving group improved the observed stereoselec-
tivity of the alkylation, increasing the ee of the product
from 64%11 to 82%. In addition, we have noted that even
relatively smaller ligands, such as Ph3P, can show a
regiochemical memory effect, especially when the leav-
ing group is a carbonate and reactions are carried out
at lower temperatures. Considering the recent interest
in chiral monodentate ligands35,36 and their potential
use in allylic alkylations,37 there is a great likelihood
that memory effects will play a more important role in
the observed enantioselectivity than has been observed
with bidentate ligands.
optical rotation were performed on a Perkin-Elmer Model 341
polarimeter.
3-Bu t -1-en yl E t h yl Ca r bon a t e (8). The procedure that
follows was used for making both the racemic 8 and (R)-8. Into
a flame-dried Schlenk flask that was filled with nitrogen were
introduced 10 mL of Et2O, 3-buten-2-ol (200 mg, 2.77 mmol),
and pyridine (2-3-fold excess). The resulting mixture was
cooled externally by the use of an ice bath. Dropwise addition
of ethyl chloroformate (2-3-fold excess) caused the precipita-
tion of pyHCl. After all of the ethyl chloroformate was added,
the ice bath was removed and the mixture was stirred at room
temperature overnight. Water was added to the reaction
mixture, and extraction with ether followed. The combined
organic extracts were dried over anhydrous Na2SO4, and the
ether was evaporated on a rotary evaporator, yielding a pale
1
yellow oil (200 mg, 50% yield). H NMR (400 MHz, CDCl3 δ):
5.80 (m, 1H), 5.22 (d, J ) 17 Hz, 1H), 5.12 (m, 2H), 4.12 (q, J
) 7.2 Hz, 2H), 1.30 (d, J ) 6.4 Hz, 3H), 1.24 (t, J ) 7.6 Hz,
3H). GC (HP 5890A, Cyclodex-B, 30 °C, He (15 psi)): 72.2 min,
(R)-8; 73.4 min, (S)-8.
Exp er im en ta l Section
Na HC(CO2Me)2. NaH (200 mg, 8.3 mmol) was suspended
in 10 mL of THF contained in a flame-dried Schlenk flask.
The flask was externally cooled by the use of an ice bath, and
dimethyl malonate (1 g, 7.54 mmol) was added dropwise. After
the cessation of hydrogen evolution the mixture was stirred
until it became clear. The solvent was evaporated, and the
resulting white solid was kept under nitrogen (88% yield).
The following spectra were used to identify products by
NMR (see ref 5).
All manipulations were carried out under an inert atmo-
sphere of dinitrogen. THF and acetonitrile or propionitrile were
distilled prior to their use over sodium/benzophenone and
calcium hydride, respectively. The ligand (+)-1 (Strem), PCy3
(Strem), PCy2Ph (Aldrich), PCyPh2 (Strem), ethyl chlorofor-
mate (Aldrich), dimethyl malonate (Aldrich), dimethyl meth-
ylmalonate (Aldrich), Pd2(dba)3 (Strem), Cs2CO3 (Strem),
Mo(CO)6 (Strem), rac-3-buten-2-ol (Fluka), and (R)-(-)-3-
buten-2-ol (Fluka) were used as received. The palladium
dimers [(C3H5)PdCl]2 and [(C4H7)PdCl]2 are prepared according
to a previously published procedure.38 Spectra were recorded
on a Bruker 400 spectrometer. Shift experiments using (+)-
Eu(hfc)3 were recorded on a Bruker 500 spectrometer. The
enantiomeric excess of 8 in the kinetic resolution experiments
was determined by GC using a Hewlett-Packard 5890A gas
chromatograph with a Cyclodex-B column. Determinations of
((E)-1-Bu t-2-en yl)m a lon ic Acid Dim eth yl Ester (6). 1H
NMR (300 MHz, CDCl3, δ): 5.61-5.49 (m, 1H), 5.42-5.31 (m,
1H), 3.73 (s, 6H), 3.41 (t, J ) 7.5 Hz, 1H), 2.57 (m, 2H), 1.64
(dd, J ) 6.1, 1.0 Hz, 3H).
1
2-(1-Meth yla llyl)m a lon ic Acid Dim eth yl Ester (7). H
NMR (400 MHz, 25 °C, CDCl3): δ 5.70 (ddd, J ) 17.2, 10.4,
8.0, 1H), 5.03 (d, J ) 17.2 Hz, 1H), 4.95 (d, J ) 10.4 Hz, 1H),
3.67 (s, 3H), 3.63 (s, 3H), 3.25 (d, J ) 8.8 Hz, 1H), 2.89 (m,
1H), 1.03 (d, J ) 6.8 Hz, 3H).
(24) Lloyd-J ones, G. C.; Stephen, S. C.; Murray, M.; Butts, C. P.;
Vyskocil, S.; Kocovsky, P. Chem. Eur. J . 2000, 6, 4348-4357.
(25) Lloyd-J ones G. C. Synlett 2001, 161-183.
(26) Fairlamb, I. J . S.; Lloyd-J ones, G. C.; Vyskocil, S.; Kocovsky,
P. Chem. Eur. J . 2002, 8, 4443-4453.
3-Bu ten -2-ol via Kin etic Resolu tion . We investigated the
use of Mo(CO)3(CH3CN)3 and Mo(CO)3(C2H5CN)3 with the
ligand (+)-1 as a catalyst for the kinetic resolution of the
racemic ethyl carbonate 8 (Scheme 8). Particularly with the
acetonitrile adduct we found the percent conversion after 15
h to be erratic, but one result (with 0.5 equiv of malonate)
showed an ee for 8 of 64% S after 47% conversion. In another
run for 36 h (with 0.7 equiv of malonate) showed an ee for 8
of 78% S. Thus, we anticipated that refinement of the
procedure might well provide a route to reasonable quantities
of 8 in high enantiomeric purity.
P r oced u r e A. Molybdenum hexacarbonyl (13.12 mg, 0.050
mmol) in 5 mL of acetonitrile was heated under vigorous reflux
for 5 h. The solvent was evaporated, and THF (5 mL) that had
been degassed by two freeze-pump-thaw cycles was intro-
duced to the reaction flask that contained the fresh Mo(CO)3-
(MeCN)3. Addition of (+)-1 (24.3 mg, 0.075 mmol) to the
resulting THF solution afforded a deep violet color. The
solution was heated under reflux for 1 h, and then the
nucleophile and 8 were added to the reaction mixture, which
had been cooled to room temperature. Stirring was continued
(27) Trost, B. M.; Bunt, R. C. J . Am. Chem. Soc. 1996, 118, 235-
236.
(28) Trost, B. M.; Fraisse, P. L.; Ball, Z. T. Angew. Chem., Int. Ed.
2002, 41, 1059-1061.
(29) Lehmann, J .; Lloyd-J ones, G. C. Tetrahedron 200x, 51, 8863-
8874.
(30) Tsuji, J .; Minami, I.; Shimizu, I. Tetrahedron Lett. 1984, 25,
5157-5160.
(31) Takeuchi, R.; Kitamura, N. New J . Chem. 1998, 659-660.
(32) Bartels, B.; Helmchen, G. Chem. Commun. 1999, 741-742.
(33) Bartels, B.; Garcia-Yebra, C.; Rominger, F.; Helmchen, G. Eur.
J . Inorg. Chem. 2002, 2569-2586.
(34) Blacker, A. J .; Clarke, M. L.; Loft, M. S.; Mahon, M. F.;
Humphries, M. E.; Williams, J . M. J . Chem. Eur. J . 2000, 6, 353-
360.
(35) Lagasse, F.; Kagan, H. B. Chem. Pharm. Bull. 2000, 48, 315-
324.
(36) Tang, W. J .; Zhang, X. M. Chem. Rev. 2003, 103, 3029-3069.
(37) Hayashi, T. J . Organomet. Chem. 1999, 576, 195-202.
(38) Tatsuno, Y.; Yoshida, T.; Otsuka, S. Inorg. Synth. 1990, 28,
342-345.

2184 Organometallics, Vol. 23, No. 9, 2004
Faller and Sarantopoulos
for 15-36 h. Conversions were determined by NMR spectra,
integrating the vinylic methyl groups of substrate 8 and
products 6 and 7. The determination of the ee of 8 by GC was
performed on an aliquot of the reaction mixture that was added
to pentane, filtered through Celite, and dissolved in dichlo-
romethane after evaporation of pentane on a rotary evapor-
ator. The retention times for the two enantiomers of 8 at 30
°C using a Hewlett-Packard 5890A with a Cyclodex-B column
and a pressure of 15 psi for He are 72.2 min (R) and 73.4 min
(S).
P r oced u r e B. A sample of Mo(CO)3(EtCN)3 was prepared
by heating 2 g of molybdenum hexacarbonyl in 20 mL of EtCN
under reflux for 24 h. The tan powder that was obtained after
evaporation of the solvent was stored under N2, and samples
of it were used in the kinetic resolution experiments in place
of the MeCN analogue. The reaction workup was exactly the
same as in procedure A described above.
Typ ica l P r oced u r e for th e P d -Ca ta lyzed Alk yla tion s
of (R)-8 u sin g P Cy3 a s Liga n d . A flame-dried Schlenk flask
was charged with THF (10 mL), [(C3H5)PdCl]2 (2.2 mg, 0.006
mmol), PCy3 (6.8 mg, 0.024 mmol), (R)-(+)-8 (35 mg, 0.243
mmol), and NaHC(CO2Me)2 (56 mg, 0.363 mmol) and then
stirred for 1 h at room temperature. Conversions were
determined by NMR spectra, integrating the vinylic methyl
groups of substrate 8 and product 6 and the allylic methyl
group of 7. The determination of the ee of the alkylated product
was performed on an aliquot where the solvent was evaporated
on a rotary evaporator and the oily residue was dissolved in
pentane and filtered through Celite. The shift reagent (+)-
Eu(hfc)3 was used to split the allylic methyl group doublet of
the alkylated product 7 (initially at δ 1.03) into two unequal
doublets (minor δ 1.59, major δ 1.52), the integrals of which
were used to calculate the ee. The sign of rotation of the
reaction mixture, +, indicated5 the presence of (R)-(+)-7 in
excess. Thus, the minor doublet at δ 1.59 corresponds to the
allylic methyl group on the complex of Eu with (S)-(-)-7, and
the major doublet at δ 1.52 corresponds to the allylic methyl
group on the complex of Eu with (R)-(+)-7.
resulting pale yellow residue with pentane afforded the
product as an off-white powder in high yield (57 mg, 94%
yield). H NMR (400 MHz, CDCl3, δ): 5.12 (1H, ddd, Hc, J )
1
12.0 Hz, J ) 11.6 Hz, J ) 6.8 Hz), 4.26 (1H, ddq, Ha′, J ) 6.8
Hz, J (Me-H) ) 6.8 Hz, J (P-H) ≈ 7 Hz), 3.16 (1H, dd, Hs, J )
6.8, 2.0 Hz); 2.37 (1H, d, Ha, J ) 11.6 Hz), 2.15-1.21 (36H, m,
Cy and crotyl Me resonances). The spectrum showed less than
3% of other isomers. 31P{1H} NMR (202 MHz, CDCl3, δ): 43.5
(s) for the syn isomer. No other isomers were observed. Anal.
Calcd for C22H40ClPPd: C, 55:35; H, 8.45. Found: C, 54.26;
H,8.32. Multiple analysis attempts deviated from the calcu-
lated values. We were unable to obtain satisfactory analyses.
1
(η3-a llyl)P d Cl(P Cy3) (10). H NMR (400 MHz, CDCl3, δ):
5.32 (1H, m, Hc), 4.54 (1H, ddd, Hs, J (PH) ) 8.0, 8.0, 2.0 Hz),
3.50 (1H, dd, Ha, J ) 11.7 Hz, J (PH) ) 8.8 Hz); 3.33 (1H, d,
Hs’, J ) 6.4 Hz, J ≈ 2 Hz), 2.52 (1H, d, Ha′, J ) 12.0 Hz), 2.14-
1.12 (33H, m, Cy resonances). 31P{1H} NMR (162 MHz, CDCl3,
δ): 41.9 (s). Anal. Calcd for C21H38ClPPd: C, 54.43; H, 8.27.
Found: C, 54.28; H, 8.24.
(η3 -syn -c r o t y l)P d C l(d ic y c lo h e x y lp h o s p h in o -2-b i-
p h en yl) (12). 1H NMR (500 MHz, CDCl3, δ): 4.45 (1H, br,
Hc), 3.87 (1H, br, Ha), 2.88 (1H, d, Hs', J ) 6.5 Hz), 2.34-1.10
(26H, m, Cy, Ha′ and crotyl Me resonances). 31P{1H} NMR (122
MHz, CDCl3, δ): 31.4 (s). Anal. Calcd for C28H38ClPPd: C,
61.43; H, 7.00. Found: C, 61.44; H, 7.05.
(η3-C3H5)P dCl(1,3-di-ter t-bu tyl-im idazol-2-yliden e) (11).
1H NMR (400 MHz, CDCl3, δ): 7.13 (1H, d, NCHCHN, J )
2.2 Hz), 7.11 (1H, d, NCHCHN, J ) 2.2 Hz), 5.25 (1H, m, Hc),
4.11 (1H, dd, Hs, J ) 7.2, 2.0 Hz), 3.37 (1H, d, Hs, J ) 6.4 Hz),
3.28 (1H, d, Ha, J ) 12.8 Hz), 2.26 (1H, d, Ha, J ) 11.6 Hz),
1.86 (9H, s, C(CH3)3), 1.69 (9H, s, C(CH3)3). Anal. Calcd for
C
14H25N2ClPd: C, 46.29; H, 6.94; N, 7.71. Found: C, 46.58;
H, 7.05; N, 7.54.
X-r a y Cr ysta llogr a p h y of (a llyl)P d Cl(P Cy3), (a llyl)-
P d Cl(NHC(t-Bu )2), a n d (cr otyl)P d Cl(P Cy2bip h en yl) (10-
12). Data were collected on a Nonius KappaCCD (Mo KR
radiation). The structures were solved by direct methods
(SIR92)39 and refined on
F for all reflections using the
Typ ica l P r oced u r e for th e P d -Ca ta lyzed Alk yla tion s
of (R)-8 u sin g NHC(t-Bu )2 a s Liga n d . A flame-dried three-
necked flask equipped with a condenser was charged with THF
(5 mL), Pd2(dba)3 (5.6 mg, 0.006 mmol), NHC-t-Bu (2.2 mg,
0.012 mmol), (R)-(+)-8 (35 mg, 0.243 mmol), H2C(CO2Me)2
(64.2 mg, 0.485 mmol), and Cs2CO3 (158 mg, 0.48 mmol), and
the mixture was heated under reflux for 20 h. Conversions
were determined by NMR spectra, integrating the olefinic
protons of substrate 8 and products 6 and 7. The determination
of the ee of the alkylated product was performed on an aliquot
where the solvent was evaporated on a rotary evaporator and
the oily residue was dissolved in pentane and filtered through
Celite. The shift reagent (+)-Eu(hfc)3 was used to split the anti
olefinic proton doublet of the alkylated product 7 to two
unequal doublets, whose integrals were used to calculate the
ee. The sign of rotation of the reaction mixture, +, indicated5
the presence of excess (R)-(+)-7. When the major doublet was
shifted to δ 5.27, it corresponded to the olefinic anti proton of
(R)-(+)-7, and the less shifted minor doublet at δ 5.19 cor-
responded to the olefinic anti proton of (S)-(-)-7.
P r ep a r a t ion of (a llyl)P d Cl(L) a n d (cr ot yl)P d Cl(L)
Com p lexes. The following procedure was used for the prepa-
ration of (η3-syn-crotyl)PdCl(PCy3). All other allyl and crotyl
complexes were prepared in over 90% yield by analogous
manipulations. Crystals for use in X-ray crystal studies were
obtained by slow diffusion of pentane into a toluene solution
of each complex.
(η3-syn -cr otyl)P d Cl(P Cy3). A Schlenk flask was flame-
dried and charged with [(crotyl)PdCl]2 (25.0 mg, 0.063 mmol)
and PCy3 (35.6 mg, 0.127 mmol). THF (10 mL) was introduced
using a syringe, and the very pale yellow solution was stirred
at room temperature for approximately 1 h. Evaporation of
THF on a rotary evaporator and repeated washing of the
TEXSAN40 package. Non-hydrogen atoms were refined with
anisotropic displacement parameters. Hydrogen atoms were
included at calculated positions. One should note that the
hydrogen atoms attached to allyl carbons were treated by
assuming them to be methylene hydrogen atoms on a Pd-
C-C fragment. This provides a more realistic placement of
the anti protons than would be obtained by assuming the
terminal allyl hydrogen atoms are in the plane of the C-C-C
fragment of the allyl. This choice can affect the Pd-C distance
by ∼0.01 Å and is an important consideration if comparing
distances with other published structures where the choice for
calculating positions is not known. Relevant crystal and data
parameters are presented below. ORTEP diagrams are shown
in Figures 1-3. One should note that many allyl-Pd com-
plexes cocrystallize with small percentages of other isomers
in the lattice and this can influence the accuracy of the bond
lengths. Both 10 and 12 crystallize with small percentages of
other isomers. The nature of the superpositions of atoms
suggests that bond length errors would be small. We also noted
more complicated disorders in (crotyl)PdCl(PCy3) and (2-
methylallyl)PdCl(NHC(t-Bu)2), which made the metrical pa-
rameters less reliable.
The colorless compound 10 crystallized in
a primitive
orthorhombic cell with the dimensions a ) 19.1477(5) Å, b )
11.3394(4) Å, c ) 9.9827(4) Å, and V ) 2167.48(11) Å3 for Z )
4 and FW ) 463.36, with a calculated density of 1.420 g/cm3
and an absorption coefficient µ(Mo KR) of 10.55 cm-1. A large
(39) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A. J .
Appl. Crystallogr. 1993, 26, 343.
(40) TEXSAN for Windows version 1.06: Crystal Structure Analysis
Package; Molecular Structure Corp., The Woodlands, TX, 1997-1999.

Allylic Alkylations
Organometallics, Vol. 23, No. 9, 2004 2185
F igu r e 3. ORTEP view of 12 with 50% probability
ellipsoids. Selected distances are as follows: Pd1-Cl1,
2.375(1); Pd1-C1, 2.214(4); Pd1-C2, 2.118(7); Pd1-C3,
2.132(3). Note that there is a 9% population of the anti
isomer present, which is not shown.
F igu r e 1. ORTEP view of 10 with 50% probability
ellipsoids. Selected distances are as follows: Pd1-Cl1,
2.369(2); Pd1-P1, 2.304(1); Pd1-C1, 2.191(6); Pd1-C2,
2.115(8); Pd1-C3, 2.112(6).
Å, c ) 14.3349(6) Å, and V ) 1576.59(10) Å3 for Z ) 4 and FW
) 363.22, with a calculated density of 1.530 g/cm3 and an
absorption coefficient, µ(Mo KR), of 13.34 cm-1. The systematic
absences of 0kl (k + l ) (2n + 1) and h0l (h ) (2n + 1)
determined the space group to be Pna21 (No. 33) or centrosym-
metric Pnam (No. 62). Ultimately, solution in Pna21 (No. 33)
was successful. The polarity was determined (marginally,
suggesting a possible racemic twin) by inverting the structure,
which gave higher R factors (R ) 0.0335 and Rw ) 0.0327
compared to R ) 0.0331 and Rw ) 0.0326). The final cycle of
full-matrix least-squares refinement on F was based on 1479
observed reflections (I > 2.00σ(I)) and 162 variable parameters
and converged with unweighted and weighted agreement
factors of R ) 0.033 and Rw ) 0.033.
The pale yellow compound 12 crystallized in a primitive
monoclinic cell with dimensions a ) 10.3007(3) Å, b )
13.3964(5) Å, c ) 19.1082(8) Å, and V ) 2587.5(2) Å3 for Z )
4 and FW ) 363.22, with a calculated density of 1.405 g/cm3
and an absorption coefficient, µ(Mo KR), of 8.96 cm-1. The
systematic absences of h0l (l ) (2n + 1) and 0k0 (k ) (2n +
1) uniquely determined the space group to be P21/c (No. 14).
The solution was straightforward, but the final difference
Fourier map showed two peaks near C1. These suggested
cocrystallization of an anti crotyl isomer bound to the opposite
face of the allyl with the methyl trans to the phosphorus.
Carbon atoms at these peak positions were refined isotropically
including the occupancy, with the constraint that their oc-
cupancies were equal and their sum with the occupancies of
their counterparts, C2 and C4, totaled 1.000. This led to a
refined occupancy of 9.1% for the minor isomer. The final cycle
of full-matrix least-squares refinement on F was based on 2456
observed reflections (I > 3.00σ(I)) and 280 variable parameters
and converged with unweighted and weighted agreement
factors of R ) 0.032 and Rw ) 0.032.
F igu r e 2. ORTEP view of 11 with 50% probability
ellipsoids. Selected distances are as follows: Pd1-Cl1,
2.379(2); Pd1-C1, 2.162(7); Pd1-C2, 2.118(7); Pd1-C3,
2.101(6).
crystal was used, and the data were corrected (SORTAV).41
The systematic absences of 0kl (k + l ) (2n + 1) and h0l (h
) (2n + 1) determined the space group to be Pna21 (No. 33)
or centrosymmetric Pnam (No. 62). Ultimately, the solution
in Pna21 (No. 33) was successful. The polarity was determined
by inverting the structure, which gave higher R factors (R )
0.0284 and Rw ) 0.0330 compared to R ) 0.0273 and Rw
)
0.0314). The final cycle of full-matrix least-squares refinement
on F was based on 1771 observed reflections (I > 3.00σ(I)) and
220 variable parameters and converged with unweighted and
weighted agreement factors of R ) 0.027 and Rw ) 0.031.
The yellow compound 11 crystallized in a primitive ortho-
rhombic cell with dimensions a ) 11.0410(5) Å, b ) 9.9613(4)
Ack n ow led gm en t. We thank the National Science
Foundation for support for this work.
Su p p or tin g In for m a tion Ava ila ble: Tables of bond
distances and angles, anisotropic thermal parameters, and
atomic coordinates for 10-12. This material is available free
of charge via the Internet at http://pubs.acs.org.
(41) (a) Blessing, R. H. Acta Crystallogr. 1995, A51, 33. (b) Blessing,
R. H. J . Appl. Crystallogr. 1997, 30, 421.
OM0306224