A fluorescence-based assay for high-throughput screening of coupling reactions. Application to heck chemistry

Article abstract of DOI:10.1021/ja983419m

A new fluorescence-based assay for high-throughput screening of reactions that couple two organic molecules is described. The assay involves reaction of one substrate containing a tethered fluorophore with a second molecule that is attached to a solid sup

Full text of DOI:10.1021/ja983419m

J. Am. Chem. Soc. 1999, 121, 2123-2132
2123
A Fluorescence-Based Assay for High-Throughput Screening of
Coupling Reactions. Application to Heck Chemistry
Kevin H. Shaughnessy, Paul Kim, and John F. Hartwig*
Contribution from the Department of Chemistry, Yale UniVersity, P.O. Box 208107,
New HaVen, Connecticut 06520-8107
ReceiVed September 25, 1998
Abstract: A new fluorescence-based assay for high-throughput screening of reactions that couple two organic
molecules is described. The assay involves reaction of one substrate containing a tethered fluorophore with a
second molecule that is attached to a solid support. A successful coupling process is signaled by fluorescence
of the solid support, which can be isolated by either simple centrifugation or filtration. To evaluate this assay
in the context of a catalytic process under intense current study, we searched for improved, readily available
phosphines for Heck chemistry. An acrylate containing a tethered coumarin was reacted with an aryl halide
supported on a cross-linked polystyrene resin. Comparison of the results from our fluorescence-based assay
with results from serial GC analysis showed that the fluorescence-based assay accurately selected the most
active ligands for the Heck coupling of aryl bromides and chlorides. Two ligands chosen by the assay, tri-
(tert-butyl)phosphine and di(tert-butylphosphino)ferrocene, were found to be the most active systems to date
for the olefination of unactivated aryl bromides, and di(tert-butylphosphino)ferrocene is the most efficient
ligand for olefination of unactivated aryl chlorides.
Introduction
have been varied over a large number of parallel reactions in
an effort to efficiently optimize reaction conditions.11 In general,
elegant ligand libraries have been prepared for use in homo-
geneous catalysis, but screening of the libraries for activity and
selectivity has typically involved serial analysis by GC or HPLC.
Only recently have high-throughput screening methodologies
been developed for determination of catalytic activity. IR
thermography has been used to analyze catalytic activity of
exothermic reactions in a high-throughput fashion.12-14 The
viability of simple colorimetric assays has been demonstrated
in the analysis of fuel cell cathodes using a fluorescent
indicator,15 and recent studies have demonstrated the ability to
visualize catalytic activity using specific, colored alkene sub-
strates.16
Overall, the synthesis of libraries of potential catalysts has
preceded the ability to conduct high-throughput screening, a
sequence of events that contrasts the initial development of high-
throughput assays for drug discovery. To rectify this problem,
we have initiated studies to develop colorimetric assays for
catalyst activity that could be applied to a wide variety of
reactions, would be amenable to rapid screening processes, could
ultimately be conducted with automated systems, and analyze
for product formation. The success of combinatorial chemistry
The design and selection of organometallic catalysts that are
capable of efficient and selective transformations of organic
substrates has been an important area of chemical research over
the past decade. The search for such catalysts would benefit
from efficient strategies to determine the activity of individual
members of a large set of catalyst systems. High-throughput
screening of small molecules that are potential drugs has driven
the preparation of large libraries of organic molecules by
combinatorial methods.1,2 The rapid screening of potential drugs
has been pursued because the necessary structural characteristics
for the desired biological activity are often poorly understood.
In catalysis, it is also difficult to predict a priori the optimal
metal/ligand combination for a specific transformation, particu-
larly when mechanistic data are unavailable. Thus, combinatorial
chemistry in its various forms has been applied to the preparation
of potential new catalysts.3
Several groups have applied combinatorial chemistry in its
broad definition to the preparation of libraries of ligands for
the enantioselective transformation of organic substrates by
homogeneous catalysis.4-10 In some cases, reaction parameters
(1) Balkenhohl, F.; von dem Bussche-Hu¨nnefeld, C.; Lansky, A.; Zechel,
C. Angew. Chem., Int. Ed. Engl. 1996, 35, 2288-2337.
(2) Thompson, L. A.; Ellman, J. A. Chem. ReV. 1996, 96, 555-600.
(3) Gennari, C.; Nestler, H. P.; Piarulli, U.; Salom, B. Liebigs Ann./
Recl. 1997, 637-647.
(4) Liu, G.; Ellman, J. J. Org. Chem. 1995, 60, 7712-7713.
(5) Gilbertson, S. R.; Wang, X. Tetrahedron Lett. 1996, 37, 6475-6478.
(6) Francis, M. B.; Finney, N. S.; Jacobsen, E. N. J. Am. Chem. Soc.
1996, 118, 8983-8984.
(7) Cole, B. M.; Shimizu, K. D.; Krueger, C. A.; Harrity, J. P. A.;
Snapper, M. L.; Hoveyda, A. H. Angew. Chem., Int. Ed. Engl. 1996, 35,
1668-1671.
(8) Shimizu, K. D.; Cole, B. M.; Krueger, C. A.; Kuntz, K. W.; Snapper,
M. L.; Hoveyda, A. H. Angew. Chem., Int. Ed. Engl. 1997, 36, 1704-
1707.
(10) Gennari, C.; Ceccarelli, S.; Piarulli, U.; Montalbetti, C. A. G. N.;
Jackson, R. F. W. J. Org. Chem. 1998, 63, 5312-5313.
(11) Burgess, K.; Lim, H.-J.; Porte, A. M.; Sulikowski, G. A. Angew.
Chem., Int. Ed. Engl. 1996, 35, 220-222.
(12) Taylor, S. J.; Morken, J. P. Science 1998, 280, 267-270.
(13) Holzwarth, A.; Schmidt, H.-W.; Maier, W. Angew. Chem., Int. Ed.
1998, 37, 2644-2647.
(14) Reetz, M. T.; Becker, M. H.; Ku¨hling, K. M.; Holzwarth, A. Angew.
Chem., Int. Ed. 1998, 37, 2647-2650.
(15) Reddington, E.; Sapienza, A.; Gurau, B.; Viswanathan, R.; Saran-
gapani, S.; Smotkin, E. S.; Mallouk, T. E. Science 1998, 280, 1735-1737.
(16) Cooper, A. C.; McAlexander, L. H.; Lee, D.-H.; Crabtree, R. H. J.
Am. Chem. Soc. 1998, 120, 9971-9972.
(17) Lam, K. S.; Salmon, S. E.; Hersh, E. M.; Hruby, V. J.; Kazmierski,
W. M.; Knapp, R. J. Nature 1991, 354, 82-86.
(9) Porte, A. M.; Reibenspies, J.; Burgess, K. J. Am. Chem. Soc. 1998,
120, 9180-9187.
10.1021/ja983419m CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/27/1999

2124 J. Am. Chem. Soc., Vol. 121, No. 10, 1999
Shaughnessy et al.
Scheme 2. Preparation and Cleavage of Supported Aryl
Scheme 1. Proposal for Visual Assay for Coupling
Reactions
Halides
in drug design rests upon detecting small molecule-macro-
molecule binding events. The most common assays involve
treating a resin-bound library of molecules with the appropriate
receptor containing a dye or radiolabel.17-26 Successful binding
is signaled by a colored or radioactive bead.
Scheme 3. Preparation of Dye-Containing Substrate 6
We felt that an analogous strategy could be applied to the
visual screening of a large set of parallel chemical reactions
that join two molecules by covalent, rather than noncovalent,
interactions. In our study, one substrate (A) would be attached
to a dye molecule, while a second substrate (B) would be
attached to a solid support (Scheme 1). After a successful
coupling reaction, substrate A would be bound to the solid
support via substrate B. Filtering and washing of the beads,
ultimately generated from reactions in microtiter plates with a
fritted bottom, would reveal which wells contained reagents and/
or catalysts that were suitable for the construction of a covalent
bond between A and B. We have initially tethered to substrate
A a fluorescent dye that can be observed with a hand-held UV
lamp typically used to assay TLC plates. In turn, substrates with
a fluorescent tag can be used for a number of additional assays
that allow one to retrofit instrumentation used for macromol-
ecule-small molecule interactions in biological studies. Nev-
ertheless, the simple assay described in this paper should prove
general enough to be applied to any reaction that can be
conducted using a solid-supported substrate and a reagent with
a tethered fluorescent tag.
capable of very high (>105) turnovers with aryl bromides and/
or high activity with aryl chlorides at moderate temperatures.33-36
Improved ligand systems are needed to carry out the reaction
at low temperatures, to conduct chemistry with aryl chlorides
in the absence of bromide additives at mild temperatures,33 and
to conduct reactions with electron-rich aryl bromides. It was
our initial goal to develop an assay for catalytic activity before
construction of ligand libraries. Thus, we elected to use, for
this study, the large pool of phosphine ligands which are either
commercially available or have been prepared previously by
our group. Remarkably, screening of this library with the assay
described here led us to discover two ligand systems that show
marked improvements over previous Heck catalyst systems for
reactions of aryl chlorides and unactivated aryl bromides.
We chose the Heck coupling process as an initial reaction to
demonstrate this simple assay. The Heck reaction can be carried
out using solid-supported aryl halides,27-32 and many groups
have been studying potential Heck catalysts that would be
Results
1. Development of the Assay for Heck Couplings: Selec-
tion of Substrates and Reaction Conditions. For our solid-
supported and soluble tagged substrate, we chose aryl halides
supported on Wang resin and acrylates tethered by an alkyl
group to 4-methyl-7-hydroxycoumarin (4). Aryl halides sup-
ported on Wang resin have been used previously in solid-phase
Heck chemistry.27-32 Coumarin 4 is inexpensive, and alkylation
of the phenolic oxygen does not decrease the fluorescence
intensity.
Treatment of Wang resin with 4-halobenzoyl chloride and
pyridine in methylene chloride resulted in complete conversion
of the OH groups, as determined by FTIR analysis (Scheme
2). Cleavage of the resin-bound aryl bromide (1a) with
trifluoroacetic acid (TFA) followed by treatment with trimeth-
ylsilyldiazomethane gave methyl 4-bromobenzoate (2a). The
loading was found to be 0.57 mmol of ArBr/g of resin
(18) Ohlmeyer, M. H. J.; Swanson, R. N.; Dillard, L. W.; Reader, J. C.;
Asouline, G.; Kobayashi, R.; Wigler, M.; Still, C. W. Proc. Natl. Acad.
Sci. U.S.A. 1993, 90, 10922-10926.
(19) Boyce, R.; Li, G.; Nestler, H. P.; Suenaga, T.; Still, C. W. J. Am.
Chem. Soc. 1994, 116, 7955-7956.
(20) Gennari, C.; Nestler, P.; Salom, B.; Still, C. W. Angew. Chem., Int.
Ed. Engl. 1995, 34, 1765-1768.
(21) Nestler, H. P. Mol. DiVersity 1996, 2, 35-40.
(22) Simon, J. A.; Schreiber, S. L. Chem. Biol. 1995, 2, 53-60.
(23) Chen, J. K.; Schreiber, S. L. Angew. Chem., Int. Ed. Engl. 1995,
34, 953-969.
(24) Wu, J.; Ma, Q. N.; Lam, K. S. Biochemistry 1994, 33, 14825-
14833.
(25) Kassarjian, A.; Schellenberger, V.; Turck, C. W. Pept. Res. 1993,
6, 129-133.
(26) Nestler, H. P.; Wennemers, H.; Sherlock, R.; Dong, D. L.-Y. Bioorg.
Med. Chem. Lett. 1996, 6, 1327-1330.
(27) Yu, K.-L.; Deshpande, M. S.; Vyas, D. M. Tetrahedron Lett. 1994,
35, 8919-8922.
(28) Hiroshige, M.; Hauske, J. R.; Zhou, P. Tetrahedron Lett. 1995, 36,
4567-4570.
(33) Herrmann, W. A.; Brossmer, C.; O¨ fele, K.; Reisinger, C.-P.;
Priermeier, T.; Beller, M.; Fischer, H. Angew. Chem., Int. Ed. Engl. 1995,
34, 1844-1848.
(29) Arumugam, V.; Routledge, A.; Abell, C.; Balasubramanian, S.
Tetrahedron Lett. 1997, 38, 6473-6476.
(30) Collini, M. D.; Ellingboe, J. W. Tetrahedron Lett. 1997, 38, 7963-
7966.
(34) Herrmann, W. A.; Brossmer, C.; Reisinger, C.-P.; Riermeier, T. H.;
O¨ fele, K.; Beller, M. Chem. Eur. J. 1997, 3, 1357-1364.
(35) Reetz, M. T.; Lohmer, G.; Schwickardi, R. Angew. Chem., Int. Ed.
1998, 37, 481-483.
(31) Zhang, H.-C.; Maryanoff, B. E. J. Org. Chem. 1997, 62, 1804-
1809.
(32) Ruhland, B.; Bombrun, A.; Gallop, M. A. J. Org. Chem. 1997, 62,
7820-7826.
(36) Shaw, B. L.; Perera, S. D.; Staley, E. A. J. Chem. Soc., Chem.
Commun. 1998, 1361-1362.

Fluorescence-Based Assay for Heck Coupling Screening
J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2125
Table 1. Comparison of Butyl Acrylate and Fluorescent Acrylate 6
(theoretical maximum ) 0.54 mmol/g) by GC analysis with an
internal standard. The coumarin-containing acrylate (6) was
prepared in two steps in 65% overall yield from 4, as shown in
Scheme 3.
in the Solid-Phase Heck Couplinga
yield
fluorescence
of resind
entry
acrylate
time (h)
3b/7c (%)
BAe
6
BA
6
BA
6
2
2
4
4
6
6
40
55
86
100
97
100
We sought a benchmark system which would show high
activity in both solution- and solid-phase reactions. The standard
Heck coupling system utilizes P(o-tol)3 and Pd(OAc)2 as the
catalyst precursor and NaOAc as base.37 This system has also
been used successfully in solid-phase Heck couplings.27 Cou-
pling of butyl acrylate and 1a at 100 °C for 12 h using the
P(o-tol)3/Pd(OAc)2/NaOAc system resulted in complete con-
sumption of the aryl halide and a quantitative yield of the
cinnamate product 3 (eq 1) after cleavage from the resin. In
1
2
3
4
5
6
weak
strong
strong
a Reactions were run with 55 mg of resin 1a (0.03 mmol of ArBr),
3 equiv of acrylate and sodium acetate, 7.5 mol % Pd(dba)2, and 30
mol % ligand at 100 °C in DMF for the indicated time. Isolation and
cleavage were carried out as described for the preparation of 1a.
b Determined by GC. c Determined by measurement of UV absorbance
at 288 nm. d Observed under UV lamp irradiation. e BA ) butyl
acrylate.
Table 2. Identity of Ligands in the First Screen
comparison, solution-phase coupling of aryl halide 2a and butyl
acrylate by this system at 100 °C for 2 h resulted in 88% isolated
yield of 3. The tagged dye 6 gave nearly identical results in
solution-phase coupling to butyl acrylate: coupling of 2a and
6 using the P(o-tol)3/Pd(dba)2/NaOAc system gave the coupled
product 7 in 92% isolated yield (eq 2).
Pd(dba) , 2.5
2
P(o-tol) , 5 mol %
3
NaOAc
With an acrylate suitably substituted with a dye and an aryl
halide bound to a resin in hand, we carried out experiments
using the solid-supported aryl halide to compare the reactivity
of the dye substrate 6 to that of butyl acrylate. Couplings of
butyl acrylate or 6 with supported aryl halide 1a were carried
out in a series of parallel reactions at 100 °C with the P(o-tol)3/
Pd(OAc)2/NaOAc system over a range of reaction times. Upon
completion of each reaction, the organic products were cleaved
from the resin to give a mixture of unreacted halide 2a, the
simple cinnamate product 3, or the coumarin-tagged cinnamate
7 (Scheme 2, eq 1). The conversion of 2a to 3 was determined
by GC, while the yield of 7 was determined from its UV
absorbance at 288 nm. The results are shown in Table 1. The
rate of formation of coupled products was similar for butyl
acrylate and for coumarin-tagged acrylate 6. After 2 h, both
reactions gave approximately 50% yield, while after 4 h the
reactions were nearly complete.
longer, which were completely labeled with dye, were strongly
fluorescent. In addition, resin-bound aryl halide was treated with
dye substrate 6 in the absence of catalyst, and no fluorescence
was observed. Thus, unreacted dye substrate is not trapped in
the resin, an event that would generate fluorescent beads from
reactions involving inactive catalysts.
2. Use of the Assay for High-Throughput Screening. A.
Selection of Ligands. Two ligand sets were tested for activity
in the Heck reaction. The first ligand set (Table 2, entries 1-20)
was comprised of monophosphines. A series of triarylphosphines
with increasing steric bulk in the ortho position (entries 1-7)
was included, along with phosphines whose electronic properties
varied while the compounds maintained similar steric demands
(entries 8-11). A series of phosphines with groups that could
The resins obtained from the reactions in Table 1 were
observed under UV light to compare the observed fluorescence
with the degree of dye incorporation before the products were
cleaved. The resin obtained after 2 h, which was 55% func-
tionalized with dye, was only faintly fluorescent (Table 1, entry
2). The resin samples obtained from reactions lasting 4 h or
(37) Spencer, A. J. Organomet. Chem. 1983, 258, 101-108.

2126 J. Am. Chem. Soc., Vol. 121, No. 10, 1999
Shaughnessy et al.
magnetically for 4 h. After this time, the vials were centrifuged,
the solvent was decanted, and the resin was washed with DMF,
methylene chloride, methanol, and ether. The resin was then
dried under mild heating and viewed against a black background
simply by using a hand-held UV lamp as an irradiation source.
Figure 1 shows the resin isolated from run 2 of the monophos-
phine ligand assay (entries 1-20, Table 2). The picture was
taken under a short-wave UV radiation source. Samples 4, 5,
6, 14, 15, and 20 appear much brighter than the other resin
samples. The contrast is more significant when viewed in color.
The bright blue of the coumarin fluorophore is easily distin-
guished from the pale blue fluorescence of the polymer beads.
The assays were run twice with each ligand set to determine
the reproducibility of the procedure. The results from both the
monophosphine and diphosphine sets showed that the assay is
highly reproducible (Table 3). Only reactions with P(o-tol)3 gave
different levels of fluorescence during the two runs. Resin
isolated from the first run was dark black, making it impossible
to determine if there was any fluorescence. However, the resin
isolated from the same system in the second run was light tan
and was strongly fluorescent. The cause of this single discrep-
ancy is unknown.
Figure 1. Beads obtained from the first screening study using mono-
phosphine ligands 1-20. Beads are displayed on a black slide after
isolation (top row, ligands 1-5 (left to right); second row, ligands 6-10;
third row, ligands 11-20; bottom row, ligands 16-20).
coordinate weakly (entries 12-16) and a series of alkylphos-
phines with increasing bulk (entries 17-20) were also included
in this first set. Diphosphine ligands comprised the second ligand
set. A range of backbones (entries 21-25, 30, 35-37, and 39)
was explored. The effects of different electronic and steric
properties of diphosphines were explored using a constant
ferrocene backbone (entries 25-29, 31-34). Sterically demand-
ing versions of the different ligands were also included (entries
26, 38, and 40).
B. Initial Round of Screening. Reactions were carried out
in sets of 20, which was the capacity of the available heating
block used in these experiments. Clearly, the number of reactions
can be expanded in future studies. In a drybox, glass vials were
charged with sodium acetate (1.5 equiv based on ArBr) and
the aryl bromide resin (0.54 mmol/g) 1a. To each vial was added
stock solutions in DMF of the dye substrate 6 (1.5 equiv), Pd-
(dba)2 (5 mol % based on ArBr), and the appropriate ligand
solution (3 equiv of P/Pd). The reaction vials were placed in
an aluminum heating block preheated to 100 °C and stirred
The ligands that showed moderate or high activity in the
fluorescence assay appeared to be structurally similar. Di(2,4-
xyl)PPh, P(o-tol)3, P(2,4-xyl)3, P(Np)3, P(o-anis)3, (DPPPh)-
EtOMe, and P(2-MOMPh)3 are all sterically demanding ortho-
substituted arylphosphines. P(t-Bu)3 is also sterically demanding,
but it is more electron-donating than the arylphosphines. The
diphosphines selected by the assay also tended to contain
sterically demanding di(o-tolyl)phosphine groups (DTPDPE and
DTPX). Only two backbones (diphenyl ether and 9,9-dimeth-
ylxanthene) were found in the chelating ligands that gave
positive results.
To determine the accuracy of the assay, each reaction was
repeated using the soluble aryl halide 2a. The reactions were
carried out in a fashion similar to the solid-phase reactions, but
with only 2.5 mol % Pd and for only 2 h. The reactions were
analyzed by GC. The coupling reactions were run multiple times
Table 3. Comparison of Fluorescence Assay with GC Results
fluorescencea
fluorescencea
ligand
run 1
run 2
GC yieldb (%)
ligand
DPPE
DPPBz
DPPP
DPPB
DPPF
DTPF
Dp-MeOPPF
Dp-CF3PPF
Dt-BPF
run 1
run 2
GC yieldb (%)
PPh3
-
-
+
-
12
9
82
99
97
96
6
-
-
-
-
0
0
36
0
0
0
25
0
6
6
0
0
0
1
36
Ph2P(o-tol)
di(2,4-xyl)PPh
P(o-tol)3
P(2,4-xyl)3
P(Np)3
P(2,6-xyl)3
P(o-anis)3
(2-BTF)P(o-anis)2
(2-BTF)2P(o-anis)
P(2-BTF)3
(DPPFc)EtOMe
(DPPFc)EtNMe2
(DPPPh)EtOMe
(2-MOMPh)3P
(2-(1-MOE)Ph)3P
P(n-Pr)3
-
+
-
-
c
-
++
++
++
-
-
-
++
++
-
-
-
-
-
-
-
+
+
7 (20)d
-
-
-
-
64
5
23
10
-
-
-
-
DPPR
-
-
-
-
(DCPFc)DPPE
(DPPFc)DCPE
(DCPFc)DCPE
(Dt-BPFc)DPPE
rac-BINAP
DPPNp
DPPDPE
DTPDPE
DPPX
-
-
-
-
-
-
-
-
0
-
-
+
+
25 (66)d
-
-
++
-
++
-
85
0
2
0
0
-
-
-
-
40
-
-
+
+
35 (42)d
65
P(i-Pr)3
P(cy)3
P(t-Bu)3
-
-
++
-
++
-
-
-
0
100
++
++
95
DTPX
++
++
a Reactions were run for 4 h at 100 °C as described in the Experimental Section. Visual observation of resin fluorescence: -, weak or no
fluorescence; +, moderate fluorescence; and ++, strong fluorescence. b Average yield of 3 from at least two runs. Reactions were run with Pd(dba)2
-4
-3
(3.5 × 10 mmol), ligand (1.0 × 10 mmol), methyl 4-bromobenzoate (0.015 mmol), butyl acrylate, and sodium acetate (0.022 mmol) in 150
µL of DMF for 2 h at 100 °C. c Resin was blackened during the reaction, making observation of fluorescence difficult. d Reactions were run for 4
h.

Fluorescence-Based Assay for Heck Coupling Screening
J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2127
Table 4. Ligands Added for the Second Round of Screens
50 °C for 4 h revealed only four active ligands. Di(2-Me-4-
FPh)PPh and P(2,4-xyl)3, which both contain bulky o-tolyl
substituents, were selected, along with the hindered alkylphos-
phine ligands (t-Bu)2PFc and P(t-Bu)3. The commonly used P(o-
tol)3 was not selected. Reactions using aryl chloride resin 1b at
110 °C for 12 h revealed (t-Bu)2PPh, (t-Bu)2PFc, and P(t-Bu)3
as active ligands for Heck reactions using chloroarenes.
D. More Detailed Solution-Phase Studies Employing the
Ligands Selected by the Fluorescence Assay. A series of
small-scale (0.2 mmol) reactions were conducted in the solution
phase using ligands selected by the assay for Heck reactions
with a variety of aryl halides. P(o-tol)3 was also tested as a
benchmark for comparison because it is a commonly used ligand
that showed poor activity by our assay for reactions of aryl
chlorides and aryl bromides at lower temperatures.38 The ligands
di(2-Me-4-FPh)PPh, P(2,4-xyl)3, (t-Bu)2PFc, and P(t-Bu)3,
which showed activity for reaction of aryl bromides at 50 °C,
were used in the Heck coupling of a series of aryl halides
ranging from electron-deficient 2a to electron-rich 9a (eq 3).
Table 5. Evaluation of the Active Ligand Subset under Various
Conditions
ArBr at
ArBr at
ArCl at
ligand
75 °Ca
50 °Ca
100 °Cb
di(2,4-xyl)PPh
di(2-Me-4-FPh)PPh
P(o-tol)3
P(2,4-xyl)3
P(2-Me-4-FPh)3
P(2,6-xyl)3
P(o-anis)3
(DPPh)EtOMe
P(2-MOMPh)3
(t-Bu)2PPh
(t-Bu)2P(o-tol)
(t-Bu)2PFc
-
-
-
++
++
++
-
++
-
-
-
++
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
++
-
-
++
-
-
++
++
-
++
++
-
++
++
-
P(t-Bu)3
DPPDPE
DTPDPE
DTPX
-
++
-
-
-
-
a Reactions were run at the indicated temperature for 4 h as described
in the Experimental Section. -, weak or no observed fluorescence in
isolated resin; ++, strong or moderate fluorescence in isolated resin.
b Reaction was run with resin 1b for 12 h.
Diphosphine ligand DTPX, which gave a positive result at 75
°C, but not at 50 °C, was also tested to determine if the assay
correctly determined that this ligand produces a catalyst that is
less active than those shown to be active at 50 °C.
for each ligand to generate reliable yield data. Comparison of
the solution-phase reactions analyzed by GC with the reactions
of solid-supported substrate analyzed by fluorescence shows a
high level of agreement (Table 3). All of the ligands that gave
GC yields higher than 80% showed strong fluorescence in the
rapid assay. Only P(o-anis)3, (DPPPh)EtOMe, and DPPDPE
were judged active by the fluorescence assay while giving low
(<35%) yields of 3 by GC analysis. Increasing the reaction time
in the solution-phase reactions of these three ligands showed
that (DPPPh)EtOMe was, indeed, modestly active (66% yield
after 4 h), but the others were not. Thus, the weaker fluorescence
of the beads from reactions using P(o-anis)3 and DPPDPE does
correspond to lower activity in solution. Importantly, all of the
ligands that gave negative results in the fluorescence assay
provided yields of 3 that were less than 65% when analyzed in
solution by GC; in other words, the assay gave no false
negatives.
C. Second Round of Screening. Since we were able to use
the assay to select the best ligands for the coupling reaction,
we used the same procedure to select active ligands under more
demanding reaction conditions. In this round of screening, the
11 active ligands from Table 3 were combined with five new
ligands. The new ligands were chosen because they shared
structural characteristics with those that showed activity in the
initial assay. The new ligands are listed in Table 4. p-Fluorinated
versions of ligands di(2,4-xyl)PPh and P(2,4-xyl)3 were selected,
along with a series of di(tert-butyl)phosphinoarenes ((t-Bu)2PPh,
(t-Bu)2P(o-tol), and (t-Bu)2PFc). This set of 16 ligands was
evaluated using the aryl bromide resin at successively lower
temperatures until only a few ligands showed activity. The
ligand set was also evaluated with an aryl chloride resin, 1b.
Reactions using resin 1a at 75 °C for 4 h resulted in only
half of the ligand set showing activity (Table 5). Reactions at
The Heck couplings were carried out with 2.5 mol % of
catalyst at 75 °C and were analyzed by GC after 2 h. Reactions
involving the electron-deficient aryl halide 2a and P(o-tol)3,
P(2,4-xyl)3, (t-Bu)2PFc, or P(t-Bu)3 as ligand gave >90% yield
of the cinnamate product 3 (Table 6). Reactions involving 2a
and di(2-Me-4-FPh)PPh as ligand, however, gave low yields
(9-27%) of the cinnamate product. As predicted by the assay,
DTPX generated a less efficient catalyst than those, except di-
(2-Me-4-FPh)PPh, chosen by the assay. Reactions using DTPX
produced 3 in only 53% yield. When the reaction was repeated
at 65 °C using only 1 mol % catalyst, P(o-tol)3 and the
structurally similar P(2,4-xyl)3 were the only ligands that led
to significant yields of the coupled product (81 and 85%,
respectively). The tert-butyl ligands (t-Bu)2PFc and P(t-Bu)3
gave little or no conversion at this lower temperature (0 and
22% yield, respectively).
However, reactions using the tert-butyl ligands (t-Bu)2PFc
and P(t-Bu)3 gave nearly quantitative yields for electron-neutral
p-tolyl bromide or electron-rich 4-bromoanisole using 2.5 mol
% catalyst at 75 °C, while use of P(2,4-xyl)3 (51% yield of 10)
and P(o-tol)3 (31% yield of 10) produced less active catalysts.
The low yield observed with P(o-tol)3 is consistent with the
results from the fluorescence assay, which showed that com-
plexes of P(o-tol)3 were less active than those of the two tert-
butyl ligands or P(2,4-xyl)3. Interestingly, reactions involving
di(2-Me-4-FPh)PPh also gave good yields of 10 (67-87%),
while DTPX produced a catalyst with very low activity for the
formation of 10 (4% yield). Results with the electron-rich
bromoanisole were similar to those observed with p-tolylbro-
mide. Use of the tert-butyl ligands (t-Bu)2PFc and P(t-Bu)3 again
(38) de Meijere, A.; Meyer, F. E. Angew. Chem., Int. Ed. Engl. 1994,
33, 2379-2411.

2128 J. Am. Chem. Soc., Vol. 121, No. 10, 1999
Shaughnessy et al.
Table 6. Solution-Phase Heck Couplings with Ligands Selected by Fluorescence Assaya
R ) CO2Me,b
X ) Br,
R ) Me,b
X ) Br,
R ) OMe,b
X ) Br,
R ) CO2Me,a
X ) Cl,
R ) Me,b
X ) Cl,
R ) OMe,b,c
X ) Cl,
ligand
P(o-tol)3
di(2-Me-4-FPh)PPh
P(2,4-xyl)3
(t-Bu)2PPh
(t-Bu)2PFc
P(t-Bu)3
2 h, 75 °C
2 h, 75 °C
2 h, 75 °C
4 h, 110 °C
22 h, 110 °C
24 h, 110 °C
97% (81%)c
31%
31%
0%
0%
0%
5% (30%)d (2%)e
88% (85%)e
67% (87%)d
51%
41% (47%)d
38%
77%
97%
96%
16%
67%
51%
20%
92% (0%)e
94% (22%)e
53%
100%
100%
4%
91%
95%
63% (80%)f
48% (44%)f
DTPX
a GC yields of two runs. ArBr conversion was within 10% of the yield for all runs. Reactions were carried out with 0.2 mmol of aryl halide, 2.5
mol % Pd(dba)2, 5.0 mol % L, 1.5 equiv of butyl acrylate, and 1.5 equiv of NaOAc in 2 mL of DMF. b See eq 3 for substrates’ structures. c 5 mol
% Pd and 10 mol % ligand used. d 5 mol % Pd(dba)2 and 10 mol % ligand used. e Reactions were carried out at 65 °C using 1 mol % Pd(dba)2 and
2 mol % ligand. f Reactions under conditions optimized by Fu et al. (ref 54): 3 mol % Pd(dba)2, 6% ligand, 1 mmol of ArCl, 2.1 mmol of butyl
acrylate, 1.1 mmol of Cs2CO3, 1 mL of dioxane, 110 °C, 24 h.
Table 7. Comparison of Ligand Activity to That of
[Pd(OAc)P(o-tol)2(C6H4CH2)]2
16 h. However, a mixture of Pd(dba)2 and (t-Bu)2PFc showed
total turnover numbers in the same range as those for the dimeric
palladacycle studied by Beller and co-workers.34 Thus, the
activity of this system with reactive substrates at high temper-
atures is similar to those reported previously, and the activity
with unreactive substrates is significantly higher.
Discussion
a Reactions were run with 20 mmol of 2a, 30 mmol of butyl acrylate,
30 mmol of sodium acetate, and 2 × 10-5 mmol of Pd at 100 °C. P/Pd
) 2 except for palladacycle. TON is the average of two runs.
1. Evaluation of the Fluorescence Assay. A. Demonstrated
Strengths. The fluorescence-based assay described here ac-
curately and reproducibly indicates the success of the Heck
coupling reactions >95% of the time. Most importantly, the
procedure selected all of the ligands that produced the most
effective catalysts for Heck reactions conducted in the solution
phase and analyzed by GC. Although two ligands showed
modest activity by the fluorescence assay but low activity in
solution, the inclusion of these ligands in subsequent experi-
ments did not affect selection of the most active catalyst systems.
They did prove to be less active ligands when reactions were
analyzed by the fluorescence assay under more demanding
conditions, such as lower temperature. The assay provides
qualitative yield data, but subtle differences in activities between
ligands were, in fact, detected. For example, DTPX, which
showed activity by the assay at 75 °C but not at 50 °C, was
shown to be less active by standard analytical methods than
the ligand set that was selected by the assay at 50 °C. Similarly,
P(o-tol)3 was not selected by the assay at 50 °C and was shown
to be significantly less active by standard methods for reaction
with electron-neutral and electron-rich aryl bromides 8a and
9a. Thus, di(2-Me-4-FPh)PPh was the only ligand for which
fluorescence analysis showed good activity at 50 °C but GC
analysis of solution-phase reactions showed poor activity.
gave over 90% yield of cinnamate 11. P(o-tol)3 gave a much
lower yield (31%), as did P(2,4-xyl)3 and di(2-Me-4-FPh)PPh
(38 and 41%, respectively).
The ligands shown by the assay to be active for Heck
reactions with aryl chlorides, along with P(o-tol)3, were
examined in these reactions with a series of aryl chlorides at
110 °C. Reactions using P(o-tol)3 as ligand gave no coupling
product with aryl chlorides. Typically, a bromide source and
higher temperatures (>140 °C) are required in order to give
moderate coupling yields with aryl chlorides using the P(o-tol)3
as ligand.33 When using the activated aryl chloride 2b, reactions
involving the ligands (t-Bu)2PFc and P(t-Bu)3 gave high yields
of 3 (93 and 90%, respectively) at 110 °C after 4 h, while (t-
Bu)2PPh gave a lower yield (67%). For reactions of chloro-
toluene, significantly longer times were required in order to
achieve moderate to high levels of conversion. Heating the
reaction for 22 h at 110 °C using (t-Bu)2PFc and P(t-Bu)3 as
ligand gave moderate yields of 10 (67 and 51%, respectively).
Reactions involving (t-Bu)2PPh gave a much lower yield (16%).
Nearly identical results were obtained in the Heck coupling of
the electron-rich chloroanisole. Use of 5 mol % Pd(dba)2 and
10 mol % (t-Bu)2PFc or P(t-Bu)3 again gave moderate yields
of 11 (63 and 48%, respectively) after 24 h. Thus, the activity
of the tert-butyl ligands was superior to that of Pd(OAc)2/P(o-
tol)3 mixtures for the reactions of unactivated aryl bromides
and chlorides.33,34
We also compared the activity of these ligands with that of
the dimeric palladacycle studied by Beller and co-workers34 for
coupling of activated aryl bromides with activated olefins at
low catalyst loads. The coupling of 4-bromobenzonitrile and
butyl acrylate was conducted using 1 × 10-4 mol % Pd at 100
°C in DMF. The reaction was monitored over time, and the
results after 16 and 90 h are shown in Table 7. Although
reactions involving P(t-Bu)3 gave high yields of product for these
substrates under standard catalysts loads (2.5 mol %), this ligand
was significantly less active than either (t-Bu)2PFc or the
palladacycle at low catalyst loads. P(t-Bu)3 appeared to generate
a relatively short-lived catalyst with no activity observed after
In addition to accuracy, the ability to conduct a colorimetric
assay in a high-throughput fashion is crucial to the ultimate
utility. We have not optimized our method for efficiency, but
the procedure is significantly faster than analysis by GC or
HPLC even in its crudest form. Utilizing manual, serial isolation
of the resin, we were able to analyze the resin from 40 reactions
in 2-3 h. By comparison, analysis of these same 40 reactions
by GC required approximately 16 h of GC time. With much
larger libraries, the difference in time would be even more
significant. Clearly, the time required for analysis of reactions
by GC increases linearly with the number of reactions, but the
time required to analyze a library of reactions by fluorescence
does not. Most importantly, workup of these reactions is
analogous to that of standard solid-phase processes that are the
core of robotics-based library synthesis. The use of microtiter

Fluorescence-Based Assay for Heck Coupling Screening
J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2129
plates comprised of wells with fritted bottoms39 should allow
for rapid manual analysis of large sets of reactions, and the
technique should be amenable to rapid automated analysis.
Assets of standard GC analysis are the wide range of reactions
that can be evaluated by this method and the ability to obtain
activity and selectivity data simultaneously. Concerning the issue
of generality, the fluorescence assay should be amenable to
many standard reactions for which appropriate dye-labeled and
solid-supported substrates can be constructed. Solid-phase
organic chemistry has been applied to a wide variety of
structures,40 making it possible to find an appropriate solid-
supported substrate for the study of many different types of
transformations. Similarly, the hydroxyethyl-substituted dye 5
should allow preparation of a wide variety of reactants that
would bear a fluorescent tag. With the exception of some
reactions modifying small molecules such as CO, H2, O2, and
CO2 that cannot be tagged, this assay should be applicable to
many types of reactions.
Concerning the issue of analyzing for activity and selectivity,
any assay for activity will allow one to analyze a subset of
reactions in a more conventional manner for selectivity. Clearly,
one would analyze for selectivity only a subset of the many
systems determined to be highly reactive in an assay for activity
such as our colorimetric one. Should selectivity have been an
issue in our study, only a small set of the initial 40 ligands,
perhaps five, would have been chosen for individual reactions
to be analyzed by GC for selectivity. In addition, there are
certain types of selectivity that would be evaluated along with
activity by the assay for coupling of two molecules. In some
cases, a catalyst may have high activity for a reaction that
consumes one or both substrates but may not provide a coupled
product. For example, a catalytic cycle that involves â-hydrogen
elimination rather than reductive elimination often consumes
two substrates in palladium-catalyzed coupling chemistry but
does not generate a coupled product.41,42 In the case of a catalyst
that is highly active for addition of aryl halide but selects
â-hydrogen elimination rather than reductive elimination, the
solid-supported substrate would not become fluorescent. A
catalyst for aryl halide reduction would, therefore, not be
selected by our assay.
trapped in the resin. Given the strong correlation between the
observed fluorescence and solution-phase yield data, quenching
of the fluorescent dye did not appear to interfere with identifica-
tion of the most active catalyst systems in the Heck reaction.
Again, one should conduct a preliminary set of experiments to
show that increasing conversion does lead to an increase in the
intensity of fluorescence. We are currently exploring solution-
based fluorescence methodologies such as measurements of
fluorescence resonance energy transfer (FRET) and fluorescence
anisotropy that are used to assay binding phenomena in
biological systems and can be adapted to assay for covalent
bond formation. These solution-phase assays may ultimately be
used in cases of distinct reactivity with solid-supported and
dissolved reagent. The solution-phase fluorescence measure-
ments may also increase the speed of the assay by allowing
direct measurement in solution.
2. Improved Ligands for Heck Chemistry. By using the
assay described in this paper, five ligands for Heck couplings
with aryl bromides and three ligands for Heck couplings with
aryl chlorides were selected from a small library of 45
structurally varied phosphines. Of the ligands selected by our
assay for activity in reactions of aryl bromides, four out of five
were also highly active in standard solution-phase reactions with
aryl bromide 2a that were analyzed by GC. Two of these
ligands, (t-Bu)2PFc and P(t-Bu)3, retained high activity with less
reactive substrates such as bromotoluene and bromoanisole,
while P(2,4-xyl)3 gave lower yields when reactions were
conducted with more electron-rich aryl bromides. Both (t-
Bu)2PFc and P(t-Bu)3 produced catalysts generating product in
high turnover numbers (104-105), although (t-Bu)2PFc gave a
longer lived, more active catalyst than did P(t-Bu)3. (t-Bu)2PFc
in combination with Pd(dba)2 showed turnover numbers for
reactions of activated aryl bromides and activated olefins similar
to those for the dimeric palladacycle studied by Beller and co-
workers.34 Moreover, (t-Bu)2PFc and P(t-Bu)3 generated cata-
lysts that produced Heck products in excellent yields using
4-bromotoluene or 4-bromoanisole, substrates that give lower
yields of product when the dimeric palladacycle is used.34
Finally, the use of (t-Bu)2PFc gave good yields of coupled
product after 1 day at 110 °C, even from reactions of butyl
acrylate with chlorotoluene and chloroanisole. The dimeric
palladacycle shows very low activity with these substrates, even
at high temperatures.
B. Potential Limitations. While the fluorescence assay has
been very successful in its application to the Heck reaction,
there are some potential limitations that should be explicitly
stated. First, the reaction relies on parallel reactivity between
reactions of solid-supported substrates and dissolved substrates.
Similar reactivity is usually observed in solution- and solid-
phase reactions and was clearly observed in the Heck reaction.
However, significant differences in reactivity may be observed
for certain catalytic systems, and a preliminary screen to
determine the similarity or difference in reactivity between
soluble and solid-supported substrates is advised before screen-
ing a large number of systems. For example, a reagent that may
lead to stable dimeric versions of the catalyst in solution may
show significantly different reactivity with substrates spacially
isolated on a bead and those in solution.
Thus, (t-Bu)2PFc appears to generate the most active catalyst
for the Heck coupling reaction. tert-Butylphosphines have
recently been reported to be efficient ligands for the Suzuki
coupling44-46 of aryl chlorides as well as the coupling of aryl
halides, including chlorides, with amines.47-51 Previous reports
of the use of tert-butylphosphines in Heck couplings have
involved cyclometalated complexes.52,53 Independent studies in
Fu’s laboratory54 have corroborated our finding that tri(tert-
(43) Scott, R. H.; Balasubramanian, S. Bioorg. Med. Chem. Lett. 1998,
7, 1567-1572.
(44) Shen, W. Tetrahedron Lett. 1997, 39, 5575-5578.
(45) Firooznia, F.; Gude, C.; Chan, K.; Satoh, Y. Tetrahedron Lett. 1998,
39, 3985-3988.
The fluorescence of dyes bound to resins can also be
complicated by fluorescence quenching due to interactions with
the resin, other dye molecules,43 or other species that become
(46) Old, D. W.; Wolfe, J. P.; Buchwald, S. L. J. Am. Chem. Soc. 1998,
120, 9722-9723.
(47) Hamann, B. C.; Hartwig, J. F. J. Am. Chem. Soc. 1998, 120, 7369-
7370.
(39) Filter bottom 96-well plates are sold by Whatman (Clifton, NJ, phone
973-773-5800).
(40) Hermkens, P. H. H.; Ottenheijm, H. C. J.; Rees, D. Tetrahedron
(48) Yamamoto, T.; Nishiyama, M.; Koie, Y. Tetrahedron Lett. 1998,
39, 2367-2370.
(49) Nishiyama, M.; Yamamoto, T.; Koie, Y. Tetrahedron Lett. 1998,
39, 617-620.
1996, 52, 4527-4554.
(41) Driver, M. S.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7217-
(50) Kanbara, T.; Izumi, K.; Narise, T.; Hasegawa, K. Polym. J. (Tokyo)
1998, 30, 66-68.
(51) Kanbara, T.; Izumi, K.; Nakadani, Y.; Narise, T.; Hasegawa, K.
Chem. Lett. 1997, 1185-1186.
7218.
(42) Hayashi, T.; Knoishi, M.; Kumada, M. Tetrahedron Lett. 1979, 21,
1871-1874.

2130 J. Am. Chem. Soc., Vol. 121, No. 10, 1999
Shaughnessy et al.
DTPDPE,62 DPPX,69 DTPX,62 P(2-Me-4-FPh)3,70 (t-Bu)2PPh,71 and (t-
Bu)2P(o-tol).72 Synthesis of new ligands is described below.
butyl)phosphine palladium complexes are unusually active
catalysts for Heck coupling of aryl chlorides, and Littke and
Fu55 have demonstrated high activity in Susuki-Miyaura
reactions56 with aryl chlorides. The high activity of the tert-
butyl-substituted bis-phosphines has been ascribed to their
strongly electron donating character, which accelerates oxidative
addition.47 However, the steric demand of the tert-butyl ligands
should also generate higher concentrations of three-coordinated
palladium complexes that would undergo olefin insertion.
Mechanistic studies to determine the important features of the
(t-Bu)2PFc ligand are in progress.
All Heck reactions were carried out in sealed vials prepared in a
drybox under a nitrogen atmosphere. Reaction vials were heated in an
aluminum block placed on top of a heater/stirrer and surrounded by
glass beads for insulation. Reported reaction temperatures are for the
heating block or oil bath and are (3 °C. GC yields were determined
by comparison of product peaks to an internal standard (naphthalene)
using response factors determined with authentic materials. Spectral
data for Heck coupling products 10 and 11 were consistent with those
reported previously.73
Preparation of [Di(2,4-dimethylphenyl)phosphino]benzene (Di-
(2,4-xyl)PPh). 2,4-Dimethylbromobenzene (10.0 g, 54.1 mmol) was
dissolved in 30 mL of THF and added dropwise to Mg (1.50 g, 61.7
mmol). The reaction was refluxed for 5 h and then cooled to room
temperature. The Grignard solution was added dropwise to a solution
of dichlorophenylphosphine (3.87 g, 27 mmol). The reaction was
refluxed for 4 h and then hydrolyzed with saturated NH4Cl. The mixture
was extracted with ether. After the mixture was dried over MgSO4,
the solvent was removed under reduced pressure. The residue was
recrystallized from ethanol to give white crystals (1.54 g, 18%), mp )
139 °C. 1H NMR (500 MHz, C6D6): 7.02 (dd, J ) 4.37, 8.21 Hz, 3H),
6.89 (d, J ) 5.50 Hz, 3H), 6.79 (d, J ) 7.50 Hz, 3H), 2.48 (s, 9H),
2.07 (s, 9H) ppm. 31P{1H} NMR (202 MHz, C6D6): -21.9 (s) ppm.
Anal. Calcd for C22H23P: C, 82.99; H, 7.28. Found: C, 83.08; H, 7.25.
Preparation of 2-[Di(2-trifluoromethylphenyl)phosphino]anisole
((2-BTF)2P(o-anis)). A solution of 0.667 g (1.67 mmol) of bis[2-
(trifluoromethyl)phenyl]monochlorophosphine74 in dry THF was chilled
to -78 °C under nitrogen in a 50-mL Schlenk flask. A Grignard solution
prepared from 0.411 g (2.20 mmol) of 2-bromoanisole and 58.8 mg
(2.40 mmol) of Mg was added dropwise via cannula. The flask was
allowed to warm to room temperature and then fitted with a reflux
condenser. The reaction was then refluxed overnight, resulting in a
deep reddish-brown solution. The solvent was removed in vacuo, and
the residue was taken up in ether and rinsed with brine. After the organic
phase was dried (MgSO4), the solvent was removed with a rotary
evaporator. The residue was then chromatographed on silica with 5%
EtOAc in hexanes. The crude product was then taken into the drybox
and recrystallized from THF/pentane, resulting in the isolation of 0.255
Conclusion
An assay allowing rapid visual analysis of large numbers of
parallel reactions has been described. Using this assay, we were
able to screen 45 ligands for their ability to form an active
catalyst with palladium for the Heck coupling reaction in a short
period of time. To provide a detailed account of the accuracy
of this assay, each of the reactions conducted using our
fluorescence assay was subsequently analyzed by GC. Although
time-consuming, this set of experiments conclusively showed
the assay to be accurate and reproducible. Using the leads
obtained from our assay, two highly effective ligands for the
Heck coupling of both aryl halides and aryl chlorides were
revealed: di(tert-butylphosphino)ferrocene and tri(tert-butyl)-
phosphine. Di(tert-butylphosphino)ferrocene is the most efficient
ligand yet reported for olefination of unactivated aryl chlorides.
Experimental Section
General Considerations. All reagents were commercially available
and used without further purification unless noted below. DMF for Heck
couplings was anhydrous grade from Aldrich and was stored and
dispensed in a drybox. Wang resin 100-200 mesh with a loading of
57
0.60 mmol/g was purchased from Nova Biochem. Pd(dba)2 and the
palladacycle33 were prepared according to literature procedures. The
following ligands were prepared according to literature methods: P(2,4-
xyl)3,58 P(2-BTF)3,59 P(2-MOMPh)3,60 DPPBz,61 DTPF,62 Dp-MeOP-
PF,63 Dp-CF3PPF,64 Dt-BPF,65 DPPR,66 DPPNp,67,68 DPPDPE,69
1
g (0.596 mmol, 35.7%) of 2 as a white solid. H NMR (500 MHz,
C6D6): 7.47 (m, 2H), 7.15 (br, 2H), 7.05 (br, 1H), 6.86 (m, 4H), 6.68
(m, 2H), 6.40 (m, 1H), 3.05 (s, 3H) ppm. 31P NMR (202 MHz, C6D6)
-28.2 (complex heptet) ppm. Anal. Calcd for C21H15F6OP: C, 58.89;
H, 3.53. Found: C, 58.78; H, 3.54.
(52) Geissler, H.; Gross, P.; Guckes, B.; Klimpel, M. Ger. Offen. DE
19647582, 1998.
(53) Herrmann, W. A.; Brossmer, C.; Reisinger, C.-P.; Riermeier, T. H.;
O¨ fele, K.; Beller, M. Chem. Eur. J. 1997, 3, 1357-1364.
(54) Littke, A. F.; Fu, G. C. J. Org. Chem. 1999, 64, 10-11.
(55) Littke, A. F.; Fu, G. C. Angew. Chem., Int. Ed. 1998, 38, 3387-
3388.
(56) Miyaura, N.; Suzuki, A. Chem. ReV. 1995, 95, 2457-2483.
(57) Rettig, M. F.; Maitlis, P. M. Inorganic Syntheses; Vol. 28, pp 110-
111.
Preparation of 2-[Di(2-methoxyphenyl)phosphino]trifluorometh-
ylbenzene ((2-BTF)P(o-anis)2). A 100-mL Schlenk flask charged with
2.02 g (14.7 mmol) of PCl3, 30 mL of dry THF, and a stir bar was
chilled to -78 °C under nitrogen. A Grignard solution made from 3.30
g (14.7 mmol) of 2-bromobenzotrifluoride and 0.388 g (1.1 equiv) of
Mg was added dropwise via addition funnel. The reaction was warmed
to room temperature and allowed to stir for 5 h. The solvent was
removed in vacuo, and 50 mL of fresh THF was added. The solution
was then chilled to -78 °C, and a second Grignard reagent synthesized
from 5.50 g (29.4 mmol) of 2-bromoanisole and 0.846 g (35.3 mmol)
of Mg was added dropwise via cannula. The reaction was allowed to
warm to room temperature, and the flask was fitted with a reflux
condenser. After the solution was heated at reflux overnight, the reaction
was quenched with 2 mL of methanol, and the solvent was removed
in vacuo. The residue was dissolved in ethyl acetate, rinsed with brine,
and dried over MgSO4. After removal of the solvent, the crude product
was suspended in cold ether and filtered. It was then recrystallized from
degassed ethanol/ethyl acetate until colorless to yield 2.25 g (5.76 mmol,
(58) Culcasi, M.; Berchadsky, Y.; Gronchi, G.; Tordo, P. J. Org. Chem.
1991, 56, 3537-3542.
(59) Neilan, J. P.; Laine, R. M.; Cortese, N.; Heck, R. F. J. Org. Chem.
1976, 41, 3455-3460.
(60) Terfort, A.; Brunner, H. J. Chem. Soc., Perkins Trans. 1 1996,
1467-1479.
(61) Tunney, S. E.; Stille, J. K. J. Org. Chem. 1987, 52, 748-753.
(62) Hamann, B. C.; Hartwig, J. F. J. Am. Chem. Soc. 1998, 120, 3694-
3703.
(63) Yamaguchi, M.; Kido, Y.; Omata, K.; Hirama, M. Synlett 1995,
1181-1182.
(64) Unruh, J. D.; Christensen, J. R. J. Mol. Catal. 1982, 14, 19-34.
(65) Cullen, W. R.; Tae-Jeong, K.; Einstein, F. W. B.; Jones, T.
Organometallics 1983, 2, 714-719.
(66) Kalck, P.; Randrianalimanana, C.; Ridmy, M.; Thorez, A.; tom
Dieck, H.; Ehlers, J. New J. Chem. 1988, 12, 679-686.
(67) Jackson, R. D.; James, S.; Orpen, A. G.; Pringle, P. G. J. Organomet.
Chem. 1993, 458, C3-C4.
(68) van Soolingen, J.; de Lang, R.-J.; den Besten, R.; Klusener, P. A.
A.; Veldman, N.; Spek, A. L.; Brandsma, L. Synth. Commun. 1995, 25,
1741-1744.
(69) Kranenberg, M.; van der Burgt, Y. E. M.; Kamer, P. C. J.; van
Leeuwen, P. W. N. M.; Goubitz, K.; Fraanje, J. Organometallics 1995, 14,
3081-3089.
(70) Paul, F.; Patt, J.; Hartwig, J. F. Organometallics 1995, 14, 3030-
3039.
(71) Corfield, J. R.; De’ath, N. J.; Trippett, S. J. Chem. Soc., C 1971,
1930-1933.
(72) Cheney, A. J.; Shaw, B. L. J. Chem. Soc., Dalton Trans. 1972, 754-
763.
(73) Miura, M.; Hashimoto, H.; Itoh, K.; Nomura, M. J. Chem. Soc.,
Perkins Trans. 1 1990, 2207-2211.
(74) Sihler, R.; Werz, U.; Brune, H.-A. J. Organomet. Chem. 1989, 368,
213-221.

Fluorescence-Based Assay for Heck Coupling Screening
J. Am. Chem. Soc., Vol. 121, No. 10, 1999 2131
1
39.2%) of 3 as a white solid. H NMR (500 MHz, C6D6): 7.54 (m,
166.5, 166.3, 143.0, 138.7, 131.3, 130.0, 127.8, 120.6, 64.5, 52.1 30.7,
19.1, 13.6 ppm. FTIR (KBr pellet): 2949, 1724, 1640, 1280, 1196
cm-1. Anal. Calcd for C15H20O5: C, 68.68; H, 6.92. Found: C, 68.64;
H, 6.75.
1H), 7.34 (m, 1H), 7.09 (t, J ) 7.5 Hz, 2H), 6.94 (br, 2H), 6.86 (m,
2H), 6.74 (t, J ) 7.2 Hz, 2H), 6.46 (dd, Jd ) 7.8 Hz, Jd ) 6.0 Hz, 2H),
3.14 (s, 6H) ppm. 31P NMR (C6D6) -29.1 (q, J ) 95.3 Hz) ppm. Anal.
Calcd for C21H18F3OP: C, 64.62; H, 4.65. Found: C, 64.52; H, 4.60.
Preparation of Tri(2-methoxymethyl)phosphine (P(2-MOMPh)3).
Methyl (2-bromobenzyl) ether (2.21 g, 11.0 mmol) was added to a
suspension of Mg (294 mg, 12.3 mmol) in 10 mL of THF. A solution
of PCl3 (457.7 mg, 3.33 mmol) in 50 mL of THF was prepared under
N2 and cooled to -78 °C. To this solution was added the Grignard
reagent via cannula. The reaction was allowed to warm to room
temperature and was refluxed overnight. The reaction was quenched
by addition of an NH4Cl solution. The aqueous layer was washed with
ether, and the combined organic layers were dried over MgSO4.
Evaporation of the solvent gave a white solid, which was recrystallized
Preparation of Aryl Halide-Functionalized Resin. Wang resin
(3.023 g, 1.814 mmol) was suspended in 50 mL of dry methylene
chloride and stirred for 10 min to allow the beads to swell. Pyridine
(1.46 mL, 18.1 mmol) was added, followed by 4-bromobenzoyl chloride
(1.833 g, 8.960 mmol). The reaction was stirred for 12 h and then
filtered. The resin was then suspended in DMF and stirred for 10 min
before being filtered. This process was repeated once more with DMF
and then with methylene chloride, methanol (2×), and ether. The resin
was then dried in vacuo for several hours until a free-flowing solid
was obtained. FTIR (KBr pellet) showed no residual OH peak and a
strong CdO stretch at 1721 cm-1. To quantify the loading, the aryl
halide was cleaved from the resin by treatment of 100 mg of the resin
with 2 mL of 1:1 TFA/methylene chloride. After being stirred for 1 h,
the mixture was filtered and the resin washed with methylene chloride
and methanol. The solvent was removed from the filtrate, and the
residue was dissolved in 3 mL of 3:1 methylene chloride/methanol and
cooled to 0 °C. A solution of TMS-diazomethane in hexanes was added
until the yellow color persisted. After the solution was stirred for 3 h,
the solvent was removed, the residue was dissolved in methylene
chloride, and acetophenone (8.2 mg) was added. The solution was then
analyzed by GC. Resin 1a was found to have a loading of 0.57 mmol/g
(theoretical maximum ) 0.54 mmol/g), while resin 1b had a loading
of 0.59 mmol/g (theoretical maximum ) 0.55 mmol/g).
1
from degassed ethanol to give 798 mg (61%) of a colorless solid. H
NMR (500 MHz, C6D6): 7.72-7.69 (m, 3 H), 7.22-7.19 (m, 3H),
7.14-7.11 (m, 3H), 6.99-6.96 (m, 3H), 4.80 (s, 6H), 3.09 (s, 9H)
ppm. 31P{1H} NMR (202 MHz, C6D6): -36.0 (s) ppm. Anal. Calcd
for C24H33O3P: C, 73.08; H, 6.90. Found: C, 72.79; H, 6.90.
Preparation of Tri[2-(1-methoxyethyl)phenyl]phosphine (P(2-
MOEPh)3). This ligand was prepared in an identical fashion to P(2-
MOMPh)3 using 1-(2-bromophenyl)ethyl methyl ether (749 mg, 3.50
mmol), magnesium (89.0 mg, 3.70 mmol), and PCl3 (137.3 mg, 1.00
mmol). A greasy, white solid was obtained (240 mg, 55%). This product
is presumably a mixture of four diastereomers. The NMR spectra of
this compound gave very broad resonances. 1H NMR (500 MHz,
C6D6): 7.64-7.61 (m, 3H), 7.14-7.10 (m, 3H), 7.09-7.06 (m, 1H),
7.01-6.98 (m, 2H), 6.87-6.84 (m, 3H), 5.38-5.19 (br m, 3H), 3.11-
2.94 (br m, 9H), 1.53-1.36 (br m, 9H) ppm. 31P{1H} NMR (202 MHz,
C6D6): -40.5 (br m) ppm.
Synthesis of 2-[(4-Methyl-2-oxo-2H-7-chromenyl)oxy]ethyl Acry-
late (6). (i) Preparation of 2[(4-Methyl-2-oxo-2H-7-chromenyl)oxy]-
ethanol (5). Coumarin 4 (5.300 g, 30.08 mmol) and potassium
carbonate (8.376 g, 59.92 mmol) were dissolved in 100 mL of DMF.
To this suspension was added 2-bromoethanol (3.20 mL, 45.1 mmol).
The reaction was heated at 100 °C and stirred overnight. TLC (20%
ethyl acetate/CH2Cl2) showed complete conversion of 4 to a single new
product with lower Rf. The reaction was allowed to cool and was then
poured into 10% HCl and shaken to give a white suspension. The
mixture was extracted with CH2Cl2. The combined organic extracts
were washed with water and brine and then dried over magnesium
sulfate. Removal of solvent gave a pale yellow solid (6.066 g, 92%
yield), which was judged pure by 1H NMR spectroscopy and was used
without further purification. 1H NMR (300 MHz, DMSO-d6): 7.65 (d,
J ) 9.44 Hz, 1H), 6.94 (m, 2H), 6.18 (s, 1H), 4.95 (t, J ) 4.32 Hz,
1H), 4.08 (t, J ) 4.42 Hz), 3.76-3.71 (m, 2H), 2.37 (s, 3H) ppm.
(ii) Acryloylation of 5. Alcohol 5 (6.066 g, 27.55 mmol) was
suspended in 150 mL of CH2Cl2 and cooled to 0 °C. To the suspension
was added triethylamine (5.80 mL, 41.5 mmol), followed by acryloyl
chloride (2.40 mL, 29.5 mmol). The solid was slowly consumed, giving
a yellow solution. After the solution was stirred for 4 h, the reaction
was extracted with water. The resulting aqueous solution was extracted
with methylene chloride, and combined organic extracts were washed
with saturated sodium bicarbonate solution and brine. After the organic
solutions were dried over magnesium sulfate, the solvent was removed
under reduced pressure. The resulting yellow oil was purified by flash
chromatography (SiO2), eluting with a gradient of 0-10% ethyl acetate
in methylene chloride to give 5.385 g (71%) of a pale yellow solid,
Preparation of [Di(4-fluoro-2-methylphenyl)phosphino]benzene
(Di(2-Me-4-FPh)PPh). This ligand was prepared as described for di-
(2,4-xyl)PPh. The Grignard reagent formed from 10 g of 2-bromo-5-
fluorotoluene (90.1 mmol) and Mg (2.22 g, 91.7 mmol) was added to
2.80 mL of dichlorophenylphosphine (90 mmol) in 10 mL of THF.
The crude product was recrystallized from methanol to give 2.1 g (10%)
1
of colorless plates, mp ) 86 °C. H NMR (500 MHz, C6D6): 7.25-
7.21 (m, 2H), 7.07-7.05 (m, 3H), 6.75-6.72 (m, 2H), 6.69 (dt, J )
3.33, 9.41 Hz, 2H), 6.57 (td, J ) 2.64, 8.50 Hz, 2H) ppm. 31P{1H}
NMR (202 MHz, C6D6): -23.1 (s) ppm. Anal. Calcd for C20H17F2P:
C, 73.61; H, 5.25. Found: C, 73.58; H, 5.26.
Preparation of [Di(tert-butyl)phosphino]ferrocene ((t-Bu)2PFc).
Ferrocene (9.311 g, 50.00 mmol) was deprotonated by addition of a
2.6 M solution of t-BuLi (38 mL, 98.8 mmol) by the method of
Guillaneux and Kagan.75 The lithioferrocene was quenched with di-
(tert-butyl)chlorophosphine (25.0 g, 138 mmol). The crude product was
sublimed twice (85 °C, 0.01 mmHg) to generate material that was used
in these studies, but it was contaminated with roughly 5% of ferrocene
(4.46 g, 27% yield). A third sublimation gave analytically pure material.
1H NMR (300 MHz, C6D6): 4.17 (m, 2H), 4.08 (m, 2H), 4.04 (s, 5H),
1.23 (s, 18H). 31P{1H} NMR (202 MHz, C6D6): 27.5 (s) ppm. Anal.
Calcd for C18H27FeP: C, 65.47; H, 8.24. Found: C, 65.51; H, 8.34.
Preparation of Methyl 4-(3-Butoxy-3-oxo-1-propenyl)benzoate
(3). In a drybox, Pd(OAc)2 (12.2 mg, 0.05 mmol), P(o-tol)3 (68.5 mg,
0.23 mmol), sodium acetate (182 mg, 2.22 mmol), and 2a (430 mg,
2.00 mmol) were combined in a small, round-bottom flask and were
dissolved in 10 mL of DMF. To this solution was added butyl acrylate
(0.32 mL, 2.23 mmol). The reaction was placed in an oil bath at 100
°C and stirred for 6 h. The reaction mixture was cooled and poured
into saturated NH4Cl solution. The aqueous layer was extracted with
ether. The ether layers were washed with brine and dried over MgSO4.
Removal of solvent gave a yellow oil that was purified by flash
chromatography, eluting with 5% ethyl acetate in hexanes. The product
was recovered as a yellow solid (462.3 mg, 88%). 1H NMR (300 MHz,
CDCl3): 8.05 (dd, J ) 1.65, 7.17 Hz, 2H), 7.68 (d, J ) 16.04 Hz,
1H), 7.58 (d, J ) 8.33 Hz, 2H), 6.53 (d, J ) 15.89 Hz, 1H), 4.22 (t,
J ) 6.64 Hz, 2H), 0.3.93 (s, 3H), 1.72-1.65 (m, 2H), 1.48-1.41 (m,
2H), 0.97 (t, J ) 7.35 Hz, 3H) ppm. 13C NMR (125 MHz, CDCl3):
1
which was judged to be pure by H NMR spectroscopy. Recrystalli-
zation from ethanol gave an off-white solid that was analytically pure.
1H NMR (500 MHz, CDCl3): 7.50 (d, J ) 8.9 Hz, 1H), 6.88 (dd, J )
2.17, 8.82 Hz, 1H), 6.83 (d, J ) 2.49 Hz, 1H), 6.45 (d, J ) 17.30 Hz,
1H), 6.17 (m, 1H), 6.14 (s, 1H), 5.87 (d, J ) 10.26 Hz, 1H), 4.54 (t,
J ) 4.64 Hz, 2H), 4.28 (t, J ) 4.62 Hz, 2H), 2.39 (s, 3H) ppm. 13C
NMR (125 MHz, CDCl3): 165.9, 161.4, 161.1, 155.2, 152.4, 131.6,
127.9, 125.6, 114.0, 112.5, 112.2, 101.6, 66.3, 62.4, 18.7 ppm. FTIR
(KBr pellet): 2956, 1704, 1613, 1395, 1195 cm-1. Anal. Calcd for
C15H14O5: C, 65.69; H, 5.15. Found: C, 65.40; H, 5.23.
Heck Coupling of Methyl 4-Bromobenzoate (2a) and 6. In a
drybox, Pd(dba)2 (14.4 mg, 0.025 mmol), P(o-tol)3 (15.1 mg, 0.050
mmol), sodium acetate (90.2 mg, 1.10 mmol), 2a (213.8 mg, 0.994
mmol), 6 (301.2 mg, 1.098 mmol), and DMF (10 mL) were placed
into a small, round-bottom flask which was then sealed with a septum.
The flask was heated in an oil bath at 100 °C for 2 h and then allowed
(75) Guillaneux, D.; Kagan, H. B. J. Org. Chem. 1995, 60, 2502-2505.

2132 J. Am. Chem. Soc., Vol. 121, No. 10, 1999
Shaughnessy et al.
to cool to room temperature. The reaction mixture was diluted with
methylene chloride, and the resulting solution was washed with saturated
ammonium chloride. The aqueous layer was extracted with methylene
chloride. The combined organic layers were then washed with water
and dried over magnesium sulfate. Evaporation of solvent under vacuum
gave a yellow solid, which was purified by flash chromatography,
eluting with methylene chloride followed by 5% ethyl acetate in
methylene chloride. The product (7) was recovered as a white solid
transferred to a microcentrifuge tube using DMF. The resin was spun
down and the supernatant poured off. This process was repeated twice
more with DMF, and the beads were suspended in a small amount of
methylene chloride and allowed to stand for 5 min to extract the DMF
from the beads. Methanol was added to sink the beads. The sample
was centrifuged, and the liquid was decanted. The washing/centrifuging
cycle was repeated once more with methanol and then with ether. The
resin was then dried at 40 °C for 1 h. At this point, the fluorescence
was observed by shining a hand-held UV lamp over the resin. To
observe the resin from all reactions simultaneously, the resins were
attached to a blackened microscope slide by using silicone grease. No
change in the fluorescence of the samples was observed over the period
of several months.
1
(375.0 mg, 92%). H NMR (300 MHz, CDCl3): 8.05 (d, J ) 8.3 Hz,
2H), 7.75 (d, J ) 16.06 Hz, 1H), 7.59 (d, J ) 8.31 Hz, 2H), 7.52 (d,
J ) 8.73 Hz, 1H), 6.93-6.86 (m, 2H), 6.57 (d, J ) 15.98 Hz, 1H),
6.16 (s, 1H), 4.62 (m, 2H), 4.32 (m, 2H), 3,93 (s, 3H), 2.40 (d, J )
1.04 Hz, 3H) ppm. 13C NMR (125 MHz, CDCl3): 166.3, 166.2, 161.4,
161.1, 155.2, 152.4, 144.1, 138.4, 131.6, 130.1, 128.0, 125.7, 119.8,
114.0, 112.5, 112.3, 101.7, 66.4, 62.7, 52.3, 18.6 ppm. FTIR (KBr
pellet): 2960, 1715, 1620, 1267, 1176 cm-1. UV (CH2Cl2) λmax ) 288
(ꢀ ) 24 500), 320 (ꢀ ) 16 600) nm. Anal. Calcd for C23H20O7: C,
67.64; H, 4.94. Found: C, 67.61; H, 4.98.
Acknowledgment. We thank Dr. Felix E. Goodson, Dr.
Blake C. Hamann, Grace Mann, and Dr. Fre´de´ric Paul for
preparation of commercially unavailable ligands used in this
paper. Thanks also to William Sacco for taking the photograph
in Figure 1. Finally, we gratefully acknowledge financial support
for this work from the National Institutes of Health, the
Department of Energy, a Camille Dreyfus Teacher-Scholar
Award, and a National Science Foundation Young Faculty
Award. J.F.H. is a grantee of Eli Lilly and is a recipient of a
fellowship from the Sloan Foundation.
General Method for Fluorescence Assay. In a drybox, each of 20
vials was charged with sodium acetate (1.8 mg, 0.02 mmol) and resin
1a or 1b (25 ( 2 mg, 0.014 mmol). To each vial was added 100 µL
of a 0.20 M DMF solution of 6, 25 µL of a 0.027 M DMF suspension
of Pd(dba)2, and 40.5 µL of a toluene solution of ligand that was 0.05
M in monophosphine ligand or 0.025 M in bis-phosphine ligand. The
vials were sealed and removed from the drybox. The vials were placed
in a preheated aluminum block at the appropriate temperature and stirred
for 4 h. After the reactions were allowed to cool, the resin was
JA983419M