FULL PAPER
Results and Discussion
slower (at least 24 h for complete conversion). Indeed, in
order to shorten the reaction time, reactions with potassium
aryltrifluoroborates were best conducted at 508C; in all re-
actions examined, complete conversion was observed within
hours.
We have recently shown that, in the presence of commer-
cially available [{Rh
ACHTUNGTRENNUNG
acids and potassium trifluoroAHCNUTGTRENNUNG
Baylis–Hillman substrates derived from a,b-unsaturated
esters via 1,4-addition/b-hydroxyelimination process gave
trisubstituted alkenes with high yields and stereoselectivities
above 96:4 in favor of the E isomer.[12,13] However, when we
wanted to extend these conditions to other Baylis–Hillman
adducts, the results were disappointing. For example, reac-
tion of phenylboronic acid (1a) with 3a under previously
described conditions[12] afforded only 50% yield of the ex-
pected alkene 5aa, even if the conversion was quantitative
(Table 1, entry 1); lower reaction temperatures did not in-
A comparison of the reactivity of the two catalyst precur-
sors, [{RhACHTUNRTGNNEUG(cod)OH}2] and [{RhACHTUTGNREN(NUNG cod)Cl}2], was evaluated in
the reaction of Baylis–Hillman adduct 3a with phenylboron-
ic acid (1a) under optimized conditions (Figure 1). In the
case of the hydroxy–rhodium precursor, the reaction was
completed within a few minutes at room temperature,
whereas more than 5 h were necessary with the chloride–
rhodium precursor. More surprisingly, fast reaction was also
observed using [{RhACTHNURGTNEUNG(cod)OH}2] when the reaction was con-
ducted at 08C (Figure 1). In view of the postulated reaction
mechanism, where a hydroxy–rhodium species is generated
after b-hydroxy elimination,[12,13] the result suggests that the
chloride anion is not a spectator ligand in this reaction, be-
cause, on the contrary, similar reaction rates should be ob-
served with this two catalyst precursors after an inductive
period. Further studies are currently under way in our labo-
ratory in order to investigate the exact role played by the
rhodium counter-anion.
Table 1. Optimization of the reaction conditions for the addition of 1a to
3a.[a]
Entry
[Rh]
Solvent
Conversion[b]
t
1
2
3
4
5
6
7
8
9
[{Rh
[{Rh
N
MeOH
MeOH
MeOH
MeOH
MeOH
isopropanol
toluene
dioxane
DMF
100[c] (50)
100 (60)
100
100 (89)
100[d] (88)
83
72
77
< 10
0
30 min
6 h
[Rh
A
30 min
5 min
5 min
30 min
30 min
30 min
1 h
[{Rh
[{Rh
[{Rh
[{Rh
[{Rh
[{Rh
[{Rh
10
H2O
4 h
[a] Reactions conducted by using 0.5 mmol Baylis–Hillman adduct 3a,
1 mmol of 1a with 2 mol% [Rh] at RT in 2 mL of solvent. [b] Conversion
determined by GC. Isolated yields in brackets. The isomeric ratio was
99:1 in favor of the E isomer. [c] Reaction conducted at 508C. [d] Re-
action conducted with 1 mol% [Rh].
Figure 1. Compared reactivity of [{RhACHTNUTRGENN(GU cod)OH}2] and [{RhACHTUNGTERN(NUGN cod)Cl}2] in
&
the addition of 1a to 3a (0.5 mol% [{RhACHTNURGTNEUNG(cod)OH}2] at 20 ( ) and at 08C
&
*
(
); 0.5 mol% [{RhACTHGUNTER(NNUG cod)Cl}2] at 208C ( )).
crease the yields that much (entry 2). In order to achieve ac-
ceptable yields on the substrates, other reaction conditions
were evaluated and particularly other rhodium catalyst pre-
cursors. We were pleased to find that a cationic rhodium
(entry 3), and, even better, a hydroxy–rhodium (entry 4)
were particularly suitable for this reaction, which was now
fast even at room temperature. Most impressive, the reac-
tion conducted with 1 mol% [{RhACHTUNRGTNEUNG(cod)OH}2] was finished
in less than five minutes and the expected alkene 5aa was
isolated in 88% yield (entry 5).
Other solvents were also evaluated using the hydroxy–
rhodium catalyst precursor (entries 5–10) but no beneficial
effect was observed in nonprotic solvents. Indeed, as shown
before,[12] alcoholic solvents are also most suitable for the re-
action of Baylis–Hillman adducts with boronic acids using
The generality of the reaction conditions was evaluated
on various Baylis–Hillman adduct with either boronic acids
1 or trifluoroborate species 1’ (Scheme 1). The new opti-
mized experimental conditions were first re-evaluated on
the reaction of Baylis–Hillman adducts 2 derived from ester
(EWG = CO2Me). In reactions with arylboronic acids,
Baylis–Hillman adducts 3 afforded the expected trisubstitut-
ed alkenes in good yields and high stereoselectivity
(Table 2). With all substrates examined, reactions were com-
pleted within a few minutes at room temperature using only
1 mol% of rhodium catalyst precursor. Moreover, faster re-
actions and generally higher yields of alkenes were observed
compared with the reaction using previously published reac-
[{Rh
previously optimized solvent system (toluene/MeOH 1:1),[13]
the addition of potassium trifluoro(phenyl)borate (1a’) to
3a also occurred at room temperature, but the reaction was
ACHTUNGTRENNUNG(cod)OH}2]. Using the same catalyst precursor, and the
tion conditions: [{RhACHTUNTGRENUNG(cod)Cl}2] at 508C (entries 2, 4, 9,
G
14).[12] Once again, it appeared that either aromatic or ali-
phatic Baylis–Hillman adducts show similar reactivity.[17]
Chem. Eur. J. 2009, 15, 4710 – 4715
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4711