Enantioselective palladium-catalyzed addition of 1,3-dicarbonyl compounds to an allene derivative

Article abstract of DOI:10.1002/chem.200500826

Enhancing atom economy of the metal-catalyzed asymmetric allylic alkylation (AAA) shifts from the usual nucleophilic displacement of a leaving group to an addition of a pronucleophile to a double bond. Using 1-alkoxyallenes as proelectrophiles, the palladium-catalyzed AAA proceeds with 1,3-dicarbonyl compounds as pronucleophiles with excellent regioselectivity and enantiomeric excess under optimized conditions. The pH of the medium proved crucial for reactivity/selectivity. By using the more acidic Meldrum's acids, the reactions required a co-catalytic amount of Bronsted acid, such as trifluoroacetic acid. Single regioisomeric products of 82-99% ee were obtained. On the other hand, the less acidic 1,3-diketones failed to react under such conditions. The fact that a less acidic acid like benzoic acid sufficed, suggested the need for general base catalysis as well. Thus, a mixture of triethylamine and benzoic acid proved optimal (ee's 93-99). Employment of the (R,R)-phenyl Trost ligand gave a product with S configuration. A model to rationalize the results has been developed.

Full text of DOI:10.1002/chem.200500826

FULL PAPER
DOI: 10.1002/chem.200500826
Enantioselective Palladium-Catalyzed Addition of 1,3-Dicarbonyl
Compounds to an Allene Derivative
Barry M. Trost,* Alessandro B. C. Simas, Bernd Plietker, Christoph Jäkel, and Jia Xie[a]
Abstract: Enhancing atom economy of
the metal-catalyzed asymmetric allylic
alkylation (AAA) shifts from the usual
nucleophilic displacement of a leaving
group to an addition of a pronucleo-
phile to a double bond. Using 1-alkoxy-
allenes as proelectrophiles, the palla-
dium-catalyzed AAA proceeds with
1,3-dicarbonyl compounds as pronu-
cleophiles with excellent regioselectivi-
ty and enantiomeric excess under opti-
mized conditions. The pH of the
medium proved crucial for reactivity/
selectivity. By using the more acidic
Meldrumꢀs acids, the reactions required
a co-catalytic amount of Brønsted acid,
such as trifluoroacetic acid. Single re-
gioisomeric products of 82–99% ee
were obtained. On the other hand, the
less acidic 1,3-diketones failed to react
under such conditions. The fact that a
less acidic acid like benzoic acid suf-
ficed, suggested the need for general
base catalysis as well. Thus, a mixture
of triethylamine and benzoic acid
proved optimal (eeꢀs 93–99). Employ-
ment of the (R,R)-phenyl Trost ligand
gave a product with S configuration. A
model to rationalize the results has
been developed.
Keywords:
compounds · asymmetric catalysis ·
palladium proelectrophiles
pronucleophiles
alkylation
·
allylic
·
·
With the current development status of synthetic method-
ology, there is a widespread perception that chemists are
able to build “from scratch” any material at will. However,
the discovery of novel natural products featuring high struc-
tural complexity and outstanding biological properties[1] as
well as the demand for efficient design of large libraries of
micromolecules, as required by the emerging field of chemi-
cal genetics,[2] pose new challenges to the field. Chemical se-
lectivity[3] and atom economy,[4] as displayed by reactions
featuring asymmetric induction,[5–7] are major goals in the
modern chemical enterprise in its search for creativity-
driven syntheses and environment-friendly processes. Inven-
tion of new reactions or improvement of known processes
along these lines hold promise to remove present design
constraints in the synthesis of complex molecules.[8] In this
context, transition-metal catalysis[7,9] has been particularly
fruitful in delivering new synthetic technology targeting
such high expectations.
can be introduced to generate chiral compounds, which can
be further elaborated in the synthesis of more complex mol-
ecules. However, this reaction needs stoichiometric amounts
of base and the starting material possesses a leaving group
that is not incorporated in the product. The final goal would
be the development of an asymmetric allylic alkylation with-
out the use of stoichiometric amounts of base or of leaving
groups.
Some years ago, our group[10] and others[11] disclosed the
palladium-catalyzed addition of pronucleophiles to allenes,
also referred to as hydrocarbonation of allenes [Eq. (1)].
These investigations showed that compounds possessing
acidic hydrogen atoms (such as 2a) add to allenes (such as
1) under neutral conditions or in the presence of minimal
amounts of base to form adducts (such as 3) regioselectively.
Based upon experimental evidence,[10] our group proposed
a mechanism involving a hydropalladation step instead of
the carbopalladation alternative, as outlined in Scheme 1.
Allylic alkylation represents a very useful and broadly ex-
plored transformation among the so-far developed catalytic
methodologies. A variety of synthetically useful nucleophiles
[a]Prof. B. M. Trost, A. B. C. Simas, B. Plietker, C. Jäkel, J. Xie
Department of Chemistry, Stanford University
Stanford, CA 94305-5080 (USA)
Fax : (+1)650-725-0002
E-mail: bmtrost@stanford.edu
Chem. Eur. J. 2005, 11, 7075 – 7082
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7075

1,3-dicarbonyl compounds. Herein, we report a full account
of our studies on the enantioselective addition of Meldrumꢀs
acid derivatives to allene 11 under the chiral catalytic
system developed by our group[6] and the recent findings re-
garding its application to reactions in which 1,3-diketones
are pronucleophiles.
Results
Methyl Meldrumꢀs acid (2a) was successfully added to ben-
zyloxyallene 11 to afford adduct 13a, albeit in low selectivi-
ties [Eq. (3)], in the presence of the catalyst based on the
chiral ligand (R,R)-12 and under basic conditions establish-
ed in our previous work with allenes.[10] The choice of alkox-
yl substituent in the allenic substrate 11 followed the prece-
dent in the literature[15] and aimed at selecting the branched
adduct (13a) over the linear one (14). Decreasing the cata-
lyst load[16] improved the ee, but left the regioselectivity un-
altered (Table 1, entries 1,2). Moreover, addition of tetrabu-
Scheme 1. Allylic alkylation: substitution versus hydrocarbonation.
The pronucleophile 4 releases its acidic proton to a palladi-
um(0) complex 5 producing palladium hydride 6, which in
turn interacts with the allene substrate 7 affording, after hy-
dropalladation, p-allylpalladium intermediate 8, the
common intermediate in Pd-catalyzed allylic alkylations.
The nucleophile generated in the catalytic cycle (9) attacks
this intermediate liberating the palladium complex to ini-
tiate another cycle and the product 10. As a simple addition
reaction with a high level of chemoselectivity, this methodol-
ogy meets the principles of atom economy. Further studies
demonstrated its usefulness in the construction of macrocy-
clic structures.[12]
More recently, a preliminary report of the first enantiose-
lective version of this process from our laboratory ap-
peared.[13] Through the use of two distinct conditions, both
Meldrumꢀs acid derivatives [Eq. (2)]and azalactones were
regioselectively added to benzyloxyallene providing
branched products with high regio- and enantioselectivity.
Thus, compared to the highly successful palladium-catalyzed
asymmetric allylic alkylations,[6,14] the hydrocarbonation of
allenes might be considered a step ahead, as a highly effi-
cient asymmetric induction is attained without the necessity
to use equivalent amounts of base or electrophile activation
through leaving groups.
Table 1. Addition of methyl Meldrumꢀs acid 2a to allene 11 under basic
conditions.[a]
Entry
Conditions[a]
THF, 808C, 12 h[b]
Ratio
3.0:1.0
2.8:1.0
7.3:1.0
–
>99:1
>99:1
>99:1
>99:1
Yield [%]
ee [%]
1
2
3
4
5
6
7
8
69
62
67
–
86
90
89
89
38
55
72
–
83
84
84
83
THF, 808C, 12 h
THF, 4% TBAB, 808C, 12 h
THF, 4% TBAB, RT, 48 h
DMSO, 4% TBAB, RT, 12 h
DMSO, 6% TBAB, RT, 12 h
DMSO, 8% TBAB, RT, 12 h
DMSO, RT, 12 h
[a]Catalyst: 1.0% [ h3-(C3H5PdCl)2], 2.5% (R,R)-12, unless otherwise
stated. [b]2.0% [ h3-(C3H5PdCl)2], 5.0% (R,R)-12.
tyl ammonium bromide (TBAB)[17] improved both the
branched/linear product ratio and enantioselectivity
(entry 3). Substituting THF by dimethyl sulfoxide (DMSO)
as solvent allowed the reaction to be conducted at room
temperature (entries 4,5). Under such conditions, a dramatic
enhancement of reaction efficiency was observed. Further
experimentation showed that tetraalkylammonium halides
did not have a significant effect on the ee for reactions run
in DMSO (entries 6–8). Unfortunately, when different
batches of 2a were employed, the best yield and ee obtained
before could not be reproduced. Purification of these mate-
rials was not able to restore the stereoselectivity, which
then ranged well below the previous level (45–56%). As a
matter of fact, we observed that higher stereoselectivities
were obtained with impure samples of 2a. Variation of the
base as well as employment of additives were tried with no
success.
To learn more about how the reaction conditions depend
upon the nature of the pronucleophiles for good enantiose-
lectivity and about its scope, as well as gaining some mecha-
nistic insight, we focused our attention on the reactions of
As methylmalonic acid is an important contaminant in
commercial samples of 2a, the effect of acid on this reaction
was investigated. Experiments using malonic acid as additive
demonstrated that, up to a certain level, an increase in acid
7076
www.chemeurj.org
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7075 – 7082

Asymmetric Catalysis
FULL PAPER
Table 3. Reaction reversibility determined by cross-experiments involv-
ing 13a and 2b.
content resulted in an improvement of ee (Table 2, en-
tries 1–3) although, an excess of the additive led to a lower
yield (entry 4). While use of tBuOK at low concentration
Entry
Additive (%)
Ratio 13a:13b
Yield [%]
1
2
3
KOtBu (2)
–
malonic acid (100)
1.0:1.1
1:0
1:0
99
80
93
Table 2. Addition of methyl Meldrumꢀs acid 2a to allene 11 under acidic
conditions.[a]
Entry
Acid (%)
Yield [%]
ee [%]
occur when the same experiment was run under neutral con-
ditions or in the presence of malonic acid (entries 2,3).
Therefore, at least part of the selectivity erosion in the reac-
tion run in basic medium could result from oxidative addi-
tion to the formed adduct reforming the p-allyl–Pd inter-
mediate. In the presence of acid, such an unproductive path-
way was possibly shut off by protonation of palladium. The
protonation of palladium(0) complexes to afford the corre-
sponding hydridopalladium(ii) species requires relatively
strong acids[20] with respect to their nickel or platinum coun-
terparts. Amatore and co-workers demonstrated that com-
plete protonation of the palladium species in the Pd/PPh3/
DMF system only occurred when an excess of acetic acid
was employed.[21] Thus, considering the lower basicity of
ligand 12, an equilibrium involving the palladium hydride in-
termediate probably occurs. Unfortunately, under the most
effective protocol identified so far, the addition of different
Meldrumꢀs acid derivatives to allene 11 displayed high varia-
bility with regard to enantioselectivity. These results indicat-
ed the need for further development, that is, the definition
of a more consistent catalytic system. Moreover, it was clear
to us that by using stoichiometric amounts of acid, the full
potential of this reaction concerning atom economy had
been compromised. We reasoned that the ee inconsistency
could be aided by using less coordinating solvents (i.e., non-
polar solvents). Furthermore, by using less basic solvents
than DMSO, less buffering by the solvent would allow great-
er effectiveness of the acid at lower concentrations.
1
2
3
4
5
6
7
8
9
malonic (10)
malonic (40)
malonic (100)
malonic (400)
malonic (100)[b]
malonic (100)[c]
AcOH (100)
TFA (100)
75
90
96
66
87
80
81
34
18
57
77
87
85
86
71
72
–
TsOH (100)
–
[a]2.0% [Pd(OTFA) 2], 2.5% (R,R)-12, 2% tBuOK, additive, DMSO,
RT, 16 h. [b]No base was added. [c]20% tBuOK instead.
(with 1.0 equiv of malonic acid) had no effect on reaction ef-
ficiency (entry 5), a tenfold increase of this additive caused
a substantial drop in stereoselectivity (entry 6). Use of
AcOH brought about a lower ee (entry 7), whereas use of
stronger acids had a deleterious effect on both yield and re-
gioselectivity, such that substantial amounts of the linear re-
gioisomer were formed (entries 8,9). The beneficial effect of
acid in this case may be related to the nature of the pronu-
cleophile. Under such conditions, the enol form of methyl
Meldrumꢀs acid would probably act as nucleophile. Its lower
reactivity relative to that of the conjugate base of 2a would
slow down the nucleophilic addition step, enabling a more
effective p–s–p equilibration of the p-allyl–Pd intermedi-
ate.[18]
Nonetheless, the following crossover experiments indicat-
ed an additional role of acid additives [Eq. (4)]. When
adduct 13a was treated with Meldrumꢀs acid derivative 2b
under basic conditions, group exchange was observed
(Table 3, entry 1). It had previously been shown that
carbon-based substituents in allylic positions can serve as
leaving groups in palladium-catalyzed reactions if the gener-
ated anion is sufficiently stabilized.[19] The crossover did not
As anticipated, changing the solvent to CH2Cl2 or THF
led to high eeꢀs even when smaller quantities of malonic
acid were employed (Table 4, entries 1,2). Experiments with
the more sensitive substrate 2b showed that, with 1% of
malonic acid as additive, CH2Cl2 is a better solvent for this
reaction (entries 3,4). Furthermore, we learned that the
Chem. Eur. J. 2005, 11, 7075 – 7082
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7077

B. M. Trost et al.
Table 4. Addition of methyl Meldrumꢀs acid 2a or 2b to allene 11 in
Table 6]. Interestingly, the successful conditions designed for
Meldrumꢀs acid derivatives completely failed when applied
to diketone 15a (entry 1). Acetal 17 was the only observed
product. Byproduct 17 is supposedly formed after partial
degradation of allene 11 under the reaction conditions fol-
lowed by catalyzed addition of released BnOH to 11.[22] In
fact compound 17 was found to be a common side product
in sluggish additions to 11. Fortunately, we found that in the
absence of TFA, compound 15a reacts to afford adduct 16a
in very high ee, albeit rather slowly (entry 2). A higher cata-
lyst load substantially increased chemical yield, but the for-
mation of 16a was still sluggish (entry 3). Moreover, this
modification in the reaction conditions did not lead to a de-
crease in ee, as previously observed with other systems.[16]
nonpolar solvents under acidic conditions.[a]
Entry Meldrumꢀs acid Acid (mol%) Solvent Yield [%] ee [%]
1
2
3
4
5
6
7
8
2a
2a
2b
2b
2b
2b
2b
2b
malonic (10)
malonic (10)
malonic (1)
malonic (1)
–
TFA (1)
TFA (2)
TFA (4)
THF
CH2Cl2 46[b]
THF 81
CH2Cl2 78
CH2Cl2 78
CH2Cl2 81
CH2Cl2 78
CH2Cl2 40
43[b]
97
98
66
85
86
94
91
88
[a]2.0 mol% [Pd(OTFA) 2], 2.5 mol% (R,R)-12, RT, overnight. [b]Re-
action halted after 5 h.
same level of efficiency could be achieved in the absence of
malonic acid (entry 5). Gratifyingly, higher eeꢀs could be at-
tained by employment of 1% trifluoroacetic acid (TFA),
confirming the decisive role of acid in the reaction under
study (entry 6). Raising the TFA concentration, however,
decreased both the ee and yield as other acid-catalyzed reac-
tions become significant (entries 7,8).
Taking the conditions of entry 6, Table 4, as optimal, the
reaction was extended to other Meldrumꢀs acid derivatives
as summarized in Equation (5) and Table 5.
It is worth noting that in contrast to the reactions with
Meldrumꢀs acid derivatives, reaction of diketone 15a pro-
ceeded under neutral conditions in high ee (Table 6, en-
tries 2 and 3). The reactivity problem should nonetheless be
tackled. It might stem from this compoundꢀs lower acidity
relative to that of 2a (pKaDMSO: Meldrumꢀs acid: 7.3; 2,4-
pentanedione: 13.3).[23] Addition of acid would facilitate
protonation of palladium, but make the conjugate base of
the pronucleophile unavailable. Thus, it appeared to us that
a buffered system should be applied here. We chose 2:1
PhCO2H/Et3N mixtures to probe the acid–base combination
effect on the reaction rate. The following results showed
that the buffer system indeed made the addition of 15a to
11 more efficient (2% Pd). Whereas stereoselectivity was
maintained, the reaction rate was dramatically improved
(Table 6, entry 4). Decrease of additive concentration
helped to uniformly increase chemical yield (entries 5,6).
Further decrease made its effect irrelevant. Additional ex-
periments were carried out to clarify partial contributions of
Table 5. Addition of methyl Meldrumꢀs acid derivatives 2 to allene 11
under optimized conditions.
Entry
Meldrumꢀs acid
Yield [%]
ee [%]
1
2
3
4
5
2a
2c
2d
2e
2 f
75
63
61
90
82
99
82
88
91
96
À
acid (PhCO2H and Et3NH+) and base (PhCO2 ) compo-
nents of the buffer. They disclosed that both PhCO2H and
À
PhCO2 had a positive effect on reaction rate (entries 7–9).
After the encouraging results
obtained for palladium-cata-
lyzed addition of Meldrumꢀs
acid derivatives to allene 11, we
set out to investigate the use of
1,3-diketones as pronucleo-
philes. Their addition to allene
11 would reveal any differences
in reactivity and selectivity be-
tween cyclic and noncyclic sub-
strates. We started with reac-
tions of 3-methylpentane-2,4-
dione (15a) using the same
chiral catalyst system [Eq. (6),
Table 6. Effect of additives on the addition of diketone 15a to allene 11.
Entry
Conditions[a]
Additive (%)
T [h]Yield of
16a [%]
ee [%]
1
2
3
4
5
6
7
8
9
10
A
B
C
B
B
B
C
C
C
C
TFA (2)
–
–
PhCO2H (20); Et3N (10)
PhCO2H (10); Et3N (5)
PhCO2H (5.0); Et3N (2.5)
Bu4NOBz (3)
PhCO2H (3)
PhCO2H (3); Et3N (3)
TFA (3); Et3N (3)
15
36–40
36
12–15
12–15
12–15
24
24
24
24
–
–
48[b]
63
98
98
96
96
96
96
97
97
97
59
65
73
65
71
68
36[b]
[a]A: 2% [Pd(OTFA) 2], 4% ligand (R,R)-12. B: 1% [h3-(C3H5PdCl)2], 2.5% ligand (R,R)-12. C: 1.5% [h3-
(C3H5PdCl)2], 3.75% ligand (R,R)-12. [b]Not complete.
7078
www.chemeurj.org
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7075 – 7082

Asymmetric Catalysis
FULL PAPER
À
The same did not apply to Et3NH+/CF3CO2 (entry 10). We
applied the optimum conditions to the additions of a series
of pentane-2,4-dione derivatives to allene 11 [Eq. (7),
Table 7]. To obtain higher yields, 3% Pd and a small excess
Cyclopentanedione 18a reacted with 11 very quickly under
neutral conditions to give adduct 19a in high ee. Conversely,
analogous diketone 18b added to allene 11 in moderate ee.
The reactions of both pronucleophiles resulted in very dif-
ferent behavior towards the buffered system successfully
used before. A significant drop in ee resulted when the
buffer system was employed in the reaction of five-mem-
bered diketone 18a, while no effect on the ee was observed
with cyclohexanedione 18b. However, in the latter case, the
yield was higher when the reaction was additive-free. Indane-
diones 18c and 18d also reacted under neutral conditions.
Although both compounds underwent fast additions, their
performance regarding stereoselectivity was very different,
with benzyl-substituted derivative 18d showing a far better
enantioselectivity. As a matter-of-fact, substance 18c was ex-
pected to react more selectively than 18d as its conjugate
base is more stable. A slower rate in the nucleophilic addi-
tion should lead to higher ee. A co-addition experiment
showed that indanedione 18c is nonetheless more reactive
than the counterpart 18d. Such a result indicates that either
palladium atom protonation or hydropalladation of allene
11 is the rate-limiting step. In other words, enantioselection
occurs in the fast step of these reactions. The low selectivity
in the reaction of pronucleophile 18c may be accounted for
by the nucleophileꢀs steric hindrance, which would disrupt
molecular recognition by the catalyst. No effort was made
to optimize the results obtained with cyclic 1,3-diketones by
means of controlling the rate of the nucleophilic addi-
tion.[16,17c,d,18]
Table 7. Generality of the addition of pentane-2,4-dione derivatives 15 to
allene 11.
Entry
Pronucleophile
Yield [%][a]
ee [%]
1
2
3
4
5
15b
15c
15d
15e
15 f
86
99
98
99
98
93
83[b]
91
97
95[c]
[a]Optimized conditions: 1.5% [ h3-(C3H5PdCl)2], 3.75% ligand (R,R)-
12, allene 11 (1.3 mol equiv), 5% PhCO2H, 2.5% Et3N, RT, 15 h. [b]The
same as [a], except for: 1.0% Pd dimer, 2.5% ligand (R,R)-12, allene 11
(1.0 mol equiv). [c]Reaction time: 1 h.
of allene (0.3 mol equiv) was used. Excellent yields and ster-
eoselectivities in the formation of adducts 16 resulted. Inter-
estingly, derivative 15 f reacted very quickly and resulted in
a slightly lower enantioselectivity (entry 5). It is possible
that in the absence of buffer, a higher efficiency in the for-
mation of 16 f could be accomplished.
Performance of cyclic 1,3-diketones as pronucleophiles
was investigated [Eq. (8), Table 8]. A marked difference in
reactivity compared to acyclic diketones 15 was observed.
To determine the enantiodirection of the process under
study, we chose to use adduct 13a, the absolute configura-
tion of which was disclosed by Mosherꢀs NMR method.[24]
Thus, this compound was transformed into the saturated sec-
ondary alcohol 21, which upon reaction with the two enan-
tiomeric Mosherꢀs acid chlorides, led to diastereomers 22a
and 22b (Scheme 2).
According to the accepted model, shielding effects by the
phenyl substituent (some of which are shown in Figure 1) in-
dicated the relative configuration shown by diastereomers
22a and 22b to be as depicted. Thus, in reactions of pronu-
cleophiles and allene 11 in the presence of the catalyst
based on the (R,R)-12 ligand, adducts possessing the S con-
figuration are formed.
Table 8. Addition of cyclic 1,3-diketones 18 to allene 11.
Entry
Pronucleophile
Conditions[a]
Yield [%]
ee [%]
1
2
3
4
5
6
18a
18a
18b
18b
18c
18d
A
B
A
B
A
A
67
66
90
70
88
77
93
78
77
77
<8
81
Scheme 2. Derivatization of adduct 13a to form distereomeric esters
22a,b.
[a]A: no additive. B: 20% PhCO 2H, 10% Et3N.
Chem. Eur. J. 2005, 11, 7075 – 7082
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7079

B. M. Trost et al.
served additive effect may be mainly related to the nucleo-
philic addition step, possibly by general base catalysis. As an
indication of that, we point out that both PhCO2H and
À
PhCO2 species (in the form of Bu4N+ or Et3NH+ salts)
À
were equally effective additives. With the former, PhCO2 is
made available after formation of the palladium hydride in-
termediate. Thus, co-catalysis by the PhCO2H/PhCO2 addi-
Figure 1. Difference in chemical shifts (1H and 13C NMR spectra) shown
by diastereomers 22a,b.
À
tive system afforded the best results (even with lower cata-
lyst loading) by simultaneously providing activation of the
pre-catalyst by protonating palladium and generation of the
requisite nucleophile, presumably the 1,3-diketone enolate.
The reactions of pentane-2,4-dione derivatives 15 were
surprisingly robust with regard to stereoselectivity. The ee
remained virtually constant over very different conditions.
Under neutral conditions, after the hydropalladation step,
the nucleophilic species in the nucleophilic addition would
naturally be the pronucleophileꢀs conjugate base. It is worth
noting that in such a medium, the same degree of enantiose-
lection is obtained as that under acidic or buffered condi-
tions; the enol form of the pentane-2,4-dione derivative
likely takes part in this step. It should be noted that acetyla-
cetonate (acac) functions as a ligand to PdII. At a minimum,
a pH-dependent equilibrium wherein the acac is bound to
PdII exists under the conditions of the reaction. Since such
coordination inhibits alkylation, the more favorable this
equilibrium the poorer the reaction rate.
The cyclic 1,3-diketones showed dramatically different ef-
fects. First, the reactions are considerably faster which may
relate to the fact that they cannot serve as bidentate ligands
to Pd in contrast to the acyclic 1,3-diketones. Furthermore,
their reactions may be more sensitive to additives resulting
in lower enantioselectivities. The source of the effect may
derive from an increase in rate of nucleophilic addition in
the presence of the additive. On the other hand, by running
under neutral conditions, the slower rate of alkylation
allows the intermediate p-allyl complexes to equilibrate
faster than they can react and thus give higher ee. Similar ef-
fects have previously been noted.[25]
Discussion
Our work has shown that hydrocarbonation of allene 11
with Meldrumꢀs acid derivatives 2 becomes more efficient
when acid additives are employed in the reaction. As briefly
discussed earlier, this finding suggested that the enol form
of the pronucleophile, instead of its conjugate base, must
take part for higher stereoselectivities to occur. When base
was employed, a lower ee resulted. Such observations point-
ed out the need of efficient p–s–p equilibration in the p-
allyl–Pd intermediate (Curtin–Hammett conditions). A less
reactive nucleophile would allow a more effective equilibra-
tion and, thus, the “matched” p-allyl–Pd intermediate could
be more efficiently selected. The observed effect of TBAB
on the addition of Meldrumꢀs acid derivatives in THF in the
presence of base furnished additional evidence for such an
assertion. It has been proposed that this halide effect origi-
nates from halide binding to palladium that favors a change
of hapticity.[17a,b] Conversely, a higher concentration of the
nucleophilic species, which may be the case for reactions
run in DMSO and in the presence of acid additives, acceler-
ates the rate of nucleophilic addition. As pointed out previ-
ously, insuring that equilibration of the chiral p-allyl–Pd
complexes is fast relative to nucleophilic attack is necessary
for high ee—a requirement that may not be met when nucle-
ophilic addition is very fast.
The experiments involving pentane-2,4-dione derivatives
15 indicated a different role for additives in the reaction
under study. We reasoned that with less acidic pronucleo-
philes, the use of acidic additives would be necessary to
attain satisfactory rates. However, under such conditions,
the pronucleophileꢀs conjugate base would be unavailable
and, thus, an effective concentration of its enol form is nec-
essary. We assumed that use of a buffer, such as the
PhCO2H/Et3N system, might ensure both palladium proto-
nation and efficient formation of the enol tautomer. Natu-
rally, the p-allyl–Pd species involved must be reactive
enough towards such a nucleophile. Our results showed that
this is, in fact, the best additive system for reactions of pen-
tane-2,4-diones 15. Conversely, TFA completely inhibited
addition of 15a to allene 11, whereas its conjugate base ap-
parently had no effect on this reaction. Thus, the formation
of a palladium hydride species does not solely guarantee
adduct formation. It is clear that the reactivity of the 1,3-di-
ketoneꢀs enol form plays an important role here. Pentane-
2,4-dione derivatives are known to form enols, as easily de-
In general, our results herein affirm the need for nucleo-
philic addition to be slow in order to obtain good stereose-
lectivity. In fact, we have shown that additives and steric
hindrance at the nucleophile have such an effect on the
mechanism that allows an efficient p–s–p equilibration of
the p-allyl–Pd intermediate. Allenes, such as 11, give rise to
this species through hydropalladation according to the ac-
cepted mechanism. Allene 11 can undergo hydropalladation
through pathways a and b, which would be initiated by a
À
palladium complexation to either the C C double bond in
the allenic system generating either species 23 or 24, respec-
tively (Scheme 3). Whereas complexes 23a, 23b, and 24b
preferentially form the syn-p-allyl–Pd intermediates 27a
and 27b (via 25 and 26b) after hydropalladation, complex
24a necessarily affords the anti-p-allyl–Pd intermediates 28
(via 26a). As complex 24b would be disfavored by steric in-
teraction between the catalyst and the syn benzyloxy group,
pathway b is likely to produce 24a preferentially. Although
hydropalladation of 11 via complexes 23a,b is electronically
1
tected by H NMR spectroscopy, and are able to add to 11
under neutral conditions, albeit slowly. Therefore, the ob-
7080
www.chemeurj.org
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7075 – 7082

Asymmetric Catalysis
FULL PAPER
1 mLminÀ1
, 254 nm): t1 =11.47, t2 =
12.78). [a]D (c=0.92 in CHCl3):
+22.98; 1H NMR (300 MHz, CDCl3):
d=7.37–7.20 (m, 5H), 6.11 (ddd, J=
19.8, 10.5, 9.3 Hz, 1H), 5.53 (dd, J=
9.5, 1.5 Hz, 1H), 5.38 (dd, J=18.7,
1.2 Hz, 1H), 4.55 (d, J=10.8 Hz, 1H),
4.29 (m, 2H), 1.70 (s, 3H), 1.67 (s,
3H), 1.46 ppm (s, 3H); 13C NMR
(75 MHz, CDCl3): d=132.0, 128.2,
123.5, 123.3, 123.1, 118.3, 101.1, 81.8,
66.1, 48.1, 24.8, 23.9, 17.0 ppm; IR
(film): n˜ =1741.5, 1379, 1290, 1204,
1069 cmÀ1
; elemental analysis calcd
(%) for C17H20O5 (304.34): C 67.09, H
6.62; found: C 66.91, H 6.47.
Representative procedure for addition
of 1,3-diketones to allene 11
3-(1-Benzyloxy-allyl)-3-furan-2-yl-
methyl-pentane-2,4-dione (16e): De-
gassed CH2Cl2 (0.7 mL) was added to
a mixture of ligand (R,R)-12 (8.85 mg,
Scheme 3. Mechanistic pathways in the hydropalladation of allene 11.
3.75 mol%), (allyl)2Pd2Cl2 (1.87 mg,
1.5 mol%) and PhCO2H (2.09 mg,
favored over complexes 24a,b due to p-basicity, sterically
the opposite is true. Thus, pathways a and b are expected to
compete leading to formation of both complexes 27 and 28.
Reductive elimination converts intermediates 27a and 28b
into product 29 and 27b and 28a into ent-29. Previous work
showed that, under the catalytic system employed herein, ef-
ficient p-facial discrimination occurs in some processes.[26,27]
Selection of either 25a or 25b could involve diastereoselec-
tion in the formation of either complexes 23 or lowering of
the energy of one of the involved transition states. The fact
that the product obtained corresponds to ent-29 when using
the R,R ligand requires that 27b and/or 28a be the reactive
intermediates in this ligand-controlled reaction. Neverthe-
less, considering that our results suggests that high stereose-
lectivities depend on efficient equilibration of the p-allyl–Pd
intermediate, such initial p-facial diastereoselectivity could
explain the substantial formation of minor enantiomer 29 in
some reactions. Furthermore, structural studies[6a,18] do not
support the possibility of syn and anti intermediates 27b and
28a, respectively, converging to the same product ent-29.
The allyl systems thereof would not fit equally well in the
chiral catalyst pocket.
5.0 mol%), and the resulting mixture was allowed to stir at RT for
ꢀ30 min under inert atmosphere. Then, neat 3-furylmethyl-2,4-pentane-
dione, (15e; 61.6 mg, 57.0 mL, 0.342 mmol), Et3N (0.86 mg, 2.5 mol%;
added as a freshly prepared solution in degassed CH2Cl2), and neat
allene 11 (65 mg, 66.7 mL, 0.445 mmol), in this order, were added by sy-
ringe, and the resulting mixture was stirred for 15 h at RT. The reaction
mixture was then filtered through a short plug of silica gel and eluted
with EtOAc/petroleum ether (1:1). After evaporation of the volatiles, the
crude product was purified by flash chromatography to give compound
16e as a colorless oil (108 mg, 97%) with an ee=98% (HPLC: Daicel
OD-H, heptane/2-propanol (90:10), 254 nm; 1 mLminÀ1
; 13.76 min
(minor) 17.85 min (major)). [a]2D2:2 =+8.75 (c=1.12 in CHCl3); 1H NMR
(300 MHz, CDCl3): d=7.36–7.28 (m, 5H), 6.25 (m, 1H), 6.01 (d, J=
3.0 Hz, 1H), 5.59 (m, 1H), 5.47–5.41 (m, 3H), 4.62 (d, J=11.1 Hz, 1H),
4.45 (d, J=7.5 Hz, 1H), 4.33 (d, J=11.1 Hz, 1H), 3.28 (d, J=15.4 Hz,
1H), 3.08 (d, J=15.4 Hz; 1H), 2.21 (s, 3H), 2.09 ppm (s, 3H); 13C NMR
(75 MHz, CDCl3): d=207.5, 205.4, 150.3, 141.6, 137.5, 132.6, 128.3, 127.6,
127.4, 121.4, 110.5, 108.8, 81.7, 72.4, 70.9, 30.5, 30.3, 28.0 ppm; IR (neat):
n˜ =1699.1, 1421.5, 1355.4, 1193.4, 1146.1, 1068.4, 1010.9, 939.3, 735.9,
698.7 cmÀ1; elemental analysis calcd (%) for C20H22O4: C 73.60, H 6.79;
found:
326.1517.
C 73.72, H 6.70; HRMS (EI): m/z calcd: 326.1518; found:
Mosherꢀs ester 22a: A mixture of alcohol 21 (18.5 mg, 98.3 mmol) (S)-
MTPA chloride (15.0 mL, 80.3 mmol), and DMAP (12.0 mg, 98.2 mmol)
in CH2Cl2 (0.5 mL) was stirred at room temperature for 2 h ((S)-
MPTA=a-methoxy-a-(trifluoromethyl)phenylacetic acid; DMAP=4-di-
methylamino pyridine). Then, it was concentrated and purified by flash
chromatography to give ester 22a (24.9 mg, 77%). [a]D =+27.96 (c=1.0
in CH2Cl2); 1H NMR (500 MHz, CDCl3): d=7.59–7.58 (m, 2H), 7.42–
7.40 (m, 3H), 5.30 (dd, 1J=10 Hz, 2J=2.5 Hz, 1H), 3.69 (d, J=11.5 Hz,
1H), 3.60 (d, J=12 Hz, 1H), 3.56 (s, 3H), 3.45 (d, J=11.8 Hz, 1H), 3.37
(d, J=12 Hz, 1H), 1.73–1.69 (m, 1H), 1.64–1.59 (m, 1H), 1.38 (s, 3H),
1.35 (s, 3H), 0.93 (t, J=7.5 Hz, 3H), 0.92 ppm (s, 3H); 13C NMR
(125 MHz, CDCl3): d=166.2, 132.0, 129.7, 128.5, 127.5, 98.2, 79.9, 66.5,
66.4, 55.5, 37.3, 23.7, 22.6, 15.8, 11.1 ppm; IR (neat): n˜ =2977, 2940, 2877,
1746, 1451, 1393, 1374, 1257, 1207, 1189, 1166, 1121, 1085, 1017, 908, 826,
768, 713 cmÀ1; elemental analysis calcd (%) for C20H27O5F3: C 59.40, H
6.75; found: C 59.18, H 6.86.
Experimental Section
Representative procedure for addition of Meldrumꢀs acid derivatives 2 to
allene 11
5-(1-Benzyloxy-allyl)-2,2,5-trimethyl-[1,3]-dioxa-4,6-dione (13a):[13] In a
4 mL Minivert pressure vial, a mixture of [Pd(OTFA)2](6.7 mg), ligand
(R,R)-12 (17 mg) and methyl Meldrumꢀs acid 2a (316.3 mg, 2.0 mmol)
was evacuated and flushed with Ar three times and then diluted with
CH2Cl2 (4.5 mL). Then, a freshly prepared 1m solution of TFA in CH2Cl2
(20 mL) was added by syringe. After 30 min of stirring at RT, allene 11
(2.0 mmol) was added by syringe. After stirring overnight, the product
was purified by direct flash chromatography on silica gel (3:2 petroleum
ether/Et2O) to yield 13a[13] (527.7 mg, 87%) as a colorless oil, which sol-
idified upon standing, with an ee of 97%. Enantiomeric excess was deter-
mined by chiral HPLC (Chiralcel AD, heptane/2-propanol (99:1), flow:
Mosherꢀs ester 22b: The procedure was the same as the one employed in
the preparation of compound 22a, except for the use of (R)-MTPA chlo-
ride. The yield was 56% (83% based on recovered starting material).
[a]D =À10.96 (c=1.0 in CH2Cl2); 1H NMR (500 MHz, CDCl3): d=7.57–
7.55 (m, 2H), 7.42–7.41 (m, 3H), 5.26 (dd, 1J=10 Hz, 2J=2.5 Hz, 1H),
3.70 (d, J=12 Hz, 1H), 3.64 (d, J=12 Hz, 1H), 3.50 (s, 3H), 3.45 (d, J=
Chem. Eur. J. 2005, 11, 7075 – 7082
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7081

B. M. Trost et al.
11.8 Hz, 1H), 3.38 (d, J=12 Hz, 1H), 1.68–1.64 (m, 1H), 1.61–1.56 (m,
1H), 1.39 (s, 3H), 1.37 (s, 3H), 0.93 (s, 3H), 0.89 ppm (t, J=7.5 Hz, 3H);
13C NMR (125 MHz, CDCl3): d=166.4, 131.7, 129.8, 128.6, 127.9, 98.2,
80.0, 66.6, 66.5, 55.2, 37.3, 23.9, 23.6, 22.4, 15.8, 10.8 ppm; IR (film): n=
2977, 2940, 2877, 1746, 1451, 1393, 1374, 1257, 1207, 1189, 1166, 1121,
1085, 1017, 908, 826, 768, 713 cmÀ1; elemental analysis calcd (%) for
C20H27O5F3: C 59.40, H 6.75; found: C 59.18, H 6.86.
[9]a) B. M. Trost, Angew. Chem. 1995, 107, 285; Angew. Chem. Int. Ed.
Engl. 1995, 34, 259–281; b) B. M. Trost, Acc. Chem. Res. 2002, 35,
695–705.
[10]B. M. Trost, V. J. Gerusz, J. Am. Chem. Soc. 1995, 117, 5156–5157.
[11]Y. Yamamoto, M. Al-Masum, N. Asao, J. Am. Chem. Soc. 1994, 116,
6019–6020.
[12]B. M. Trost, P.-Y. Michellys, V. J. Gerusz, Angew. Chem. 1997, 109,
1837–1839; Angew. Chem. Int. Ed. Engl. 1997, 36, 1750–1753.
[13]B. M. Trost, C. Jäkel, B. Plietker, J. Am. Chem. Soc. 2003, 125,
4438–4439.
Acknowledgements
[14]a) B. M. Trost, J. Org. Chem. 2004, 69, 5813–5837; b) B. M. Trost,
M. L. Crawley, Chem. Rev. 2003, 103, 2921–2943.
We thank the National Science Foundation and National Institutes of
Health (GM-13598) for generous support of our programs. A.S. thanks
CAPES and Universidade Federal do Rio de Janeiro/Nfflcleo de Pesqui-
sas de Produtos Naturais (Brazil), C.J. thanks the Alexander von Hum-
boldt Foundation, and B.P. thanks the Deutscher Akademischer Aus-
tauschdienst for postdoctoral fellowships. We also thank Dr. Kelvin Yong
for GC-MS analyses of some adducts. Mass spectra were provided by the
Mass Spectrometry Facility, University of California-San Francisco sup-
ported by the NIH Division of Research Resources.
[15]Y. Yamamoto, M. Al-Masum, Synlett 1995, 969–970.
[16]a) I. J. S. Fairlamb, G. C. Lloyd-Jones, Chem. Commun. 2000, 2447–
2448; b) B. M. Trost, J.-P. Surivet, J. Am. Chem. Soc. 2000, 122,
6291–6292.
[17]a) M. P. T. Sjogren, S. Hansson, B. Škermark, A. Vitagliano, Orga-
nometallics 1994, 13, 1963; b) U. Burckhardt, M. Baumann, A.
Togni, Tetrahedron: Asymmetry 1997, 8, 155; c) B. M. Trost, R. C.
Bunt, J. Am. Chem. Soc. 1998, 120, 70–79; d) B. M. Trost, G. M.
Schroeder, J. Org. Chem. 2000, 65, 1569–1573.
[18]B. M. Trost, F. D. Toste, J. Am. Chem. Soc. 1999, 121, 4545–4554.
[19]Y. I. M. Nilsson, P. G. Andersson, J. E. Bäckvall, J. Am. Chem. Soc.
1993, 115, 6609–6613; N. Vicart, J. Gore, B. Cazes, Synlett 1996,
850–852.
[20]V. V. Grushin, Chem. Rev. 1996, 96, 2011–2033.
[21]a) C. Amatore, A. Jutand, G. Meyer, I. Carelli, I. Chiarotto, Eur. J.
Inorg. Chem. 2000, 1855–1859; for recent findings on the impor-
tance of palladium hydride species, see: b) I. D. Hills, G. C. Fu, J.
Am. Chem. Soc. 2004, 126, 13178–13179.
[22]a) F. P. J. T. Rutjes, T. M. Kooistra, H. Hiemstra, H. E. Schoemaker,
Synlett 1998, 192–194; b) H. Ovaa, M. A. Leeuwenburgh, G. A.
van de Marcel, J. H. van Boom, Tetrahedron Lett. 1998, 39, 3025–
3028.
[23]F. Bordwell, Acc. Chem. Res. 1988, 21, 456–463.
[24]S. K. Latypov, J. M. Seco, E. QuiÇoµ, R. Riguera, J. Org. Chem.
1996, 61, 8569–8577.
[1]a) D. A. Spiegel, J. T. Njardarson, I. M. McDonald, J. L. Wood,
Chem. Rev. 2003, 103, 2691–2727; b) S. J. Mickel, D. Niederer, R.
Daeffler, A. Osmani, E. Kuesters, E. Schmid, K. Schaer, R. Gambo-
ni, W. Chen, E. Loeser, F. R. Kinder Jr., K. Konigsberger, K.
Prasad, T. M. Ramsey, O. Repic, R.-M. Wang, Org. Process Res.
Dev. 2004, 8, 122–130 and preceding papers; c) C. H. Heathcock, M.
McLaughlin, J. Medina, J. L. Hubbs, G. A. Wallace, R. Scott, M. M.
Claffey, C. J. Hayes, G. R. Ott, J. Am. Chem. Soc. 2003, 125, 12844–
12849; d) J. L. Hubbs, C. H. Heathcock, J. Am. Chem. Soc. 2003,
125, 12836–12843.
[2]a) S. L. Schreiber, Chem. Eng. News 2003, 81(9), 51–61; b) See also:
S. Borman, Chem. Eng. News 2004, 82(9), 32–40.
[3]B. M. Trost, Pure Appl. Chem. 1979, 51, 787–800.
[4]B. M. Trost, Science 1991, 254, 1471–1477.
[25]G. Giambastiani, G. Poli, J. Org. Chem. 1998, 63, 9608–9609.
[26]a) B. M. Trost, F. D. Toste, J. Am. Chem. Soc. 1998, 120, 9074–9075;
b) B. M. Trost, M. J. Krische, R Radinov, G. Zanoni, J. Am. Chem.
Soc. 1996, 118, 6297–6298.
[5]R. Noyori, M. Kitamura, Angew. Chem. 1991, 103, 90; Angew.
Chem. Int. Ed. Engl. 1991, 30, 49–69.
[6]a) B. M. Trost, Acc. Chem. Res. 1996, 29, 355–364; b) B. M. Trost,
D. L. Van Vranken, Chem. Rev. 1996, 96, 395–422.
[27]For cases in which p-facial discrimination does not seem to operate
efficiently, see: a) reference [18]; b) B. M. Trost, H. C. Shen, L.
Dong, J. -P. Surivet, C. Sylvain, J. Am. Chem. Soc. 2004, 126, 11966–
11983.
[7]a) R. Noyori, M. Kitamura, T. Ohkuma, Proc. Natl. Acad. Sci. USA
2004, 101, 5356–5362; b) B. M. Trost, Proc. Natl. Acad. Sci. USA
2004, 101, 5348–5355; c) Comprehensive Asymmetric Catalysis,
Vols. I–III (Eds.: E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Spring,
Berlin, 2000.
Received: July 15, 2005
[8]T. Hudlicky, Chem. Rev. 1996, 96, 3–30.
Published online: September 30, 2005
7082
www.chemeurj.org
ꢁ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7075 – 7082