alization of furans, dihydrofurans,13 thiophenes,13 azoles,
1
2
catalysts, and three solvents were selected for additional
study. Solvents were selected in part by the diversity of their
physical properties.
14
and oxazoles, among others. Additionally, functionalization
of 2-furaldehyde has been achieved through oxidative
coupling using stoichiometric amounts of Pd(OAc)
ever, homocoupling and lack of regioselectivity render these
conditions problematic. Indeed, palladium-mediated dimer-
ization of furfural has been used preparatively.15
1
8
15
2
; how-
A complete factorial design involved 45 combinations;
however, the number of experiments was reduced to 30 in
such a way that information could still be statistically
19
interpreted. This permitted a broad exploration of the
In an effort to evaluate the overall feasibility of the
reaction, we evaluated a number of parameters in preliminary
experimentation including bases, solvents, additives, cata-
lysts, and phosphines at two levels using a half-factorial
parameter space and highlighted synergies that a traditional
vary-one-factor-at-a-time approach might have overlooked.
Figure 1 displays results from the 30 reactions. For this study
16
design (16/32 experiments, Table 1). Generally, conversions
Table 1. Screening Factors
additive
catalyst
phosphine
base (2 equiv)
KOAc
solvent (1 equiv) (5 mol %) (10 mol %)
DMA
Bu4NBr
CuI
PdCl2
none
N,N-diisopropyl- CH3CN
Pd(OAc)2
(o-tolyl)3P
ethylamine
were low and ranged from 0 to 41%. Analysis of the data
clearly showed selection of base as the major factor influenc-
ing percent conversion. Remarkably, KOAc gave a 9-fold
increase vs N,N-diisopropylethylamine as an average over
Figure 1. Conversion (%). Data points shown (30/45) reflect a
statistically balanced design. All experiments were preformed with
all 16 reactions. Similarly, PdCl
increase in conversion with respect to Pd(OAc)
2
produced almost a 2-fold
. Homocou-
1
equiv of Bu
Solvents: DMF, 4-methyl-2-pentanone (MIBK), NMP. Catalysts:
PdCl , 5% Pd/C (wet and dry). Phopshine ligands: P ) none, P
(2-furyl) P, P ) Cy P, P ) (o-tol) P, and P ) (t-bu) P.
Conversion is defined from HPLC areas at the 6 h time-point: 100-
product/product + starting material + homocoupling; 220 nm).
4
NBr, 2 equiv of KOAc, and 3 equiv of 2-furaldehyde.
2
pling was a deleterious side reaction and thus reduced
significantly the amount of arylfurfural. Further analysis of
the data revealed an 8-fold increase on average in the amount
of homocoupling when using N,N-diisopropylethylamine
over KOAc as a base. Additionally, the absence of phosphine
ligand resulted in a 1.5-fold increase in the amount of
homocoupling.
2
0
1
)
3
2
3
3
3
4
3
(
the amount of 2-furaldehyde was increased to 3 equiv in an
effort to further minimize homocoupling.
Phosphines have been used in cross-coupling reactions to
promote transmetalation and thereby reduce homocoupling.
We initially attempted reactions at 80 °C and found this
temperature to be completely ineffective, therefore neces-
sitating an increase to 110 °C in order to obtain reasonable
17
In Heck-type reactions, elimination of the hydridopalladium
8
10
halide intermediate is reversible. Phosphines have been used
conversion rates. Consistent with Jeffery’s observations,
to prevent or slow the re-addition of the hydridopalladium
complex to the double bond. The use of phosphines to
the palladium(0) catalyst preformed well in both DMF and
9
NMP in the absence of phosphine ligand when Bu NBr was
4
promote oxidative addition of aryl halides to palladium is
well-known; however, in this particular reaction, alkene
insertion is arguably the step with the most significant energy
barrier. To suggest the role phosphine might play in alkene
insertion for this reaction would be purely speculative;
however, examination of a range of sterically and electroni-
cally diverse phosphines was warranted.
Using a statistically driven approach, an array of solvents,
catalysts, and ligands was explored with iodochlorobenzene
again as a common substrate. Five phosphines, three
used as the additive. Similarly, Jeffery observed pronounced
effects from water, either beneficial or detrimental. A marked
effect is also seen here when comparing wet Pd/C vs dry.
Clearly PdCl with DMF produced promising results with
2
and without added ligand. Ultimately, Cy P was carried
3
forward because it worked well for other substrates under
investigation. In a number of cases, low conversions were
still the direct result of extensive homocoupling.16 We later
found that this could be avoided altogether or minimized
(<10%) with the slow addition of the aryl halide via syringe
1678
Org. Lett., Vol. 3, No. 11, 2001
McClure
