Reactive dyes as a method for rapid screening of homogeneous catalysts [16]

Full text of DOI:10.1021/ja9818607

J. Am. Chem. Soc. 1998, 120, 9971-9972
9971
Reactive Dyes as a Method for Rapid Screening of
Homogeneous Catalysts
Alan C. Cooper, Lenore H. McAlexander, Dong-Heon Lee,
Matthew T. Torres, and Robert H. Crabtree
Yale Chemistry Department, Box 208107
New HaVen, Connecticut 06520-8107
ReceiVed May 26, 1998
The advent of combinatorial chemistry1 has raised the question
of how to apply these methods to homogeneous catalyst
discovery1b,c where a wide variety of rapid parallel assays will
be needed. Recent reports have proposed IR thermography1b or
laser photoionization1d to select active catalysts. The formation
of colored products has long been used2 for monitoring enzyme
reactions, and here we discuss reactive dyes that change color
upon undergoing a catalytic reaction. Visual selection of the most
active candidate catalysts should then be possible by noting which
catalysts cause the most rapid color change. Further studies will
always be required to confirm that the “hits” indeed correspond
to active homogeneous catalysts and to determine the other
properties of each catalyst. In this paper we describe the screening
of conventional catalyst candidates for alkene and imine hydrosi-
lation with reactive dye-substrates. We have chosen a well-
known reaction3 with well-established characteristics to test the
method but will extend this approach to other reactions in the
future.
The new dyes contain an electron donor (D) and an acceptor
(A) group linked by the appropriate reactive functionality (RF),
a CdC or a NdC bond in this case: D-CHdCH-A and D-Nd
CH-A. When the reactive functionality is saturated upon under-
going the catalytic reaction of interest, the electronic connection
between the D and the A groups is severed, and the intensity of
color is strongly diminished; the result is an apparent bleaching
of the dye color.
absorption maxima (Table 1); 3 was deep purple and 4 was dark
blue in THF or EtOAc. The dyes are somewhat solvatochromic.6c
The 3J(H,H′) 1H NMR coupling constant7 for the CHdCH group
of 3 was 15.9 Hz, consistent with a trans geometry. N-substituted
imines such as 4 are not expected8 to be rigid but tend to adopt
a trans geometry. Each product showed a useful IR signature
from the RF group (3, ν(CC) ) 1675 cm-1{m}; 4, ν(NC) ) 1636
cm-1{ms}).
With the dyes 3 and 4 in hand, we were able to implement the
rapid screening protocol. Twelve hydrosilation catalysts9a-c were
assayed (Table 2). Some were known to be active (e.g., entries
4 and 6), some were not previously studied for alkene hydrosi-
lation (entries 8 and 109d), and one was an entirely new compound
(entry 11). Assays were run in a homemade 60-well plate
machined from 3/4-in Teflon (well capacity, 1 mL) in a glovebox
under N2 and Ar.10 Stock solutions of Ph2SiH2 (5.0 × 10-2M),9e
dye (alkene, 3.3 × 10-4M; imine, 6.6 × 10-4M),9f and catalysts
(3.3 × 10-4M) were prepared in degassed anhydrous THF. Fixed
volumes (silane, 0.100 mL; dye, catalyst 0.200 mL each) of the
(6) (a) A mixture of ferrocenylamine (1.41 g, 7.0 mmol) and 4-pyridine-
carboxaldehyde (0.75 g, 7.0 mmol) in benzene (50 mL) was refluxed with
MgSO4 (0.5 g) for 2 h. The reaction mixture was cooled to room temperature
and the MgSO4 filtered off. The filtrate, concentrated under reduced pressure,
gave a dark red solid, which was purified by chromatography on alumina
with 10/90 (v/v) ethyl acetate/hexane. The dye was obtained as a red solid
We have avoided the commercially preferred4 A and D groups
which have undesired functionality, such as -NR2, -NO2 or -OR,
that could bind to, oxidize, or otherwise inhibit the catalysts under
study. Instead, we have used a ferrocenyl (Fc) D group and a
pyridinium A group, so that the RF group is the only part of the
dye expected to show significant affinity for the catalyst.
We have synthesized the known compound trans-15 and the
new compound 26 in acceptable yields by conventional routes.
Pro-dyes 1 and 2 are strongly colored, being violet-red (1) or
dark red (2) in THF or EtOAc; the UV-vis spectral data are
shown in Table 1.
Quaternization proceeds readily (DMF, 25-80 °C) with a
variety of benzyl halides, but 4-(tert-butyl)benzyl bromide gave
the most soluble derivatives. Anion exchange (excess NaBPh4,
acetone) to give6b the BPh4- salts 3 and 4 avoids the presence of
potentially ligating halide counterions. Conversion of pyridines
1 and 2 to the more strongly electron-accepting pyridinium salts
3 and 4 led to a significant intensification and red shift of the
1
(1.9 g, 94% yield). H NMR (CDCl3, 25 °C) 8.66 (d, J ) 7.5 Hz, 2H), 8.56
(s, 1H), 7.63 (d, J ) 7.5 Hz, 2H), 4.62 (t, J ) 2.1 Hz, 2H), 4.36 (t, J ) 2.1
Hz, 2H), 4.21 (s, 5H). Anal. Calcd for C16H14FeN2: C, 66.20; H, 4.86; N,
9.65. Found: C, 66.62; H, 4.88; N, 9.38. (b) The pro-dyes 1 and 2 (3.46
mmol) were treated with 4-(tert-butyl)benzyl bromide (0.96 mL, 5.2 mmol)
in DMF (50 mL) at 80 °C for 15 min. The cooled solution was poured into
1:1 (v/v) Et2O/hexanes (100 mL) to precipitate a dark-colored solid, which
was filtered, washed with Et2O (3 × 30 mL), and dissolved in acetone (50
mL). Addition of NaBPh4 (1.2 g, 3.5 mmol), stirring for 12 h, and removal of
the acetone in vacuo gave a solid which was extracted with CH2Cl2 (2 × 20
mL). Evaporation of the combined extracts and recrystallization from 5:1 (v/
v) MeOH/THF gave crystals of the products 3 (78%) and 4 (86%). (c) For an
example, in DMSO, λmax (ꢀ) for 3 is 554 nm (62 000). This means that control
wells need to be run in the same solvent during any catalyst assays.
(7) 1H NMR (CDCl3, 25 °C) of 3: 7.55 (br s, 8H), 7.32 (d, J ) 7.5 Hz,
2H), 7.12 (d, J ) 15.9 Hz, 1H), 6.95 (m, 8H), 6.78 (d, J ) 7.5 Hz, 2H), 6.77
(m, 4H), 6.25 (d, J ) 5.4 Hz, 2H), 6.18 (d, J ) 15.9 Hz, 1H), 6.08 (d, J )
5.4 Hz, 2H), 4.54 (br s, 4H), 4.18 (br s, 5H), 4.10 (s, 2H), 1.29 (s, 9H). Anal.
Calcd. for C52H50BFeN: C, 82.69; H, 6.69; N, 1.85. Found: C, 82.40; H,
1
6.92; N, 1.78. H NMR (CDCl3, 25 °C) of 4: 8.42 (s, 1H), 7.41 (m, 13H),
(1) (a) Lam, K. S.; Lebl, M.; Krchnak, V. Chem. ReV. 1997, 97, 411.
Combinatorial Chemistry, Wilson, S. R., Czarnik, A. W., Eds.; Wiley: New
York, 1997. Combinatorial Peptide and Nonpeptide Libraries, G. Jung, Ed.;
VCH: New York, 1996. (b) Taylor, S. J.; Morken, J. P. Science 1998, 280,
267; (c) Hoveyda, Angew. Chem., Int. Ed. Engl. 1996, 35, 1668. Burgess, K.;
Lim, H.-J.; Porte, A. M.; Sulikowski, G. A. Angew. Chem., Int. Ed. 1996, 35,
220. (d) Senkan, S. M. Nature 1998, 394, 350.
(2) Walsh, C. Enzymatic Reaction Mechanisms; W. H. Freeman: San
Francisco, 1979. Jencks, W. P. Catalysis in Chemistry and Enzymology;
Dover: New York, 1969.
(3) (a) Applied Homogeneous Catalysis; Cornils, B., Herrmann, W. A.,
Eds.; VCH: New York, 1996. (b) Speier, J. F. AdV. Organomet. Chem. 1979,
17, 407. (c) Ojima, I.; Kogure, T.; Nagai, Y. Tetrahedron Lett. 1973, 2475.
(4) Handbook of US Colorants, Marmion, D. M. Wiley: New York, 1984.
(5) Bhadbhade, M. M.; Das, A.; Jeffery, J. C.; McCleverty, J. A.; Navas
Badiola, J. A.; Ward, M. D. J. Chem. Soc., Dalton Trans. 1995, 2769.
6.87 (m, 9H), 6.84 (m, 7H), 4.65 (t, J ) 2.1 Hz, 2H), 4.57 (t, J ) 1.8 Hz,
2H), 4.43 (s, 2H), 4.18 (s, 5H), 1.29 (s, 9H). Anal. Calcd for C51H49BFeN2:
C, 81.0; H, 6.53; N, 3.70. Found: C, 80.68; H, 6.39; N, 3.70.
(8) The Chemistry of Double Bonded Functional Groups, S. Patai, Ed.;
Wiley: New York, 1977-1989.
(9) (a) Some of these compounds had shown homogeneous catalytic activity
in prior work.3,9b,c (b) Chaloner, P. A. Handbook of Coordination Catalysis
in Organic Chemistry; Butterworths: London, 1986. (c) Barber, D. E.; Lu,
Z,; Richardson, T.; Crabtree, R. H. Inorg. Chem. 1992, 31, 4709. (d) Herrmann,
W. A.; Brossmer, C.; Reisinger, C.-P.; Riermeier, T. H.; Ofele, K.; Beller,
M., Chem. Eur. J. 1997, 3, 1357. (e) Et2SiH2 and Ph2SiH2 were found to be
the most reactive silanes tried in a preliminary assay, and Ph2SiH2 was therefore
used subsequently. (f) More of the less intensely colored imine dye was
required per well than of the more intense alkene. (g) Digital images for a
sample screen are available on the Web at http://ursula.chem.yale.edu/
∼crabtree/ under “Research”.
S0002-7863(98)01860-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/11/1998

9972 J. Am. Chem. Soc., Vol. 120, No. 38, 1998
Table 1. The New Dyes
Communications to the Editor
dye
formulaa
color (native)b
λ
max (ꢀ)c
color (reduced)d
1
3
2
4
FcCHdCH(4-py)
[FcCHdCH(4-pyBz)]BPh4
FcNdCH(4-py)
violet-red
deep purple
dark red
460 (1900)
607 (12600)
496 (2000)
686 (5200)
lt yellow
lt yellow
lt yellow
lt yellow
[FcNdCH(4-pyBz)]BPh4
dark blue
a Fc ) ferrocenyl; py ) pyridyl; Bz ) p-(tert-butyl)benzyl. b In EtOAc or THF. c In ethyl acetate solution. λmax in nm; ꢀ in M-1 cm-1 d In THF.
.
Table 2. Results of Hydrosilation Catalyst Screens Using Dyes
3-4
prove that hydrosilation is the cause. In the case of one of the
best catalysts, RhCl(PPh3)3, we therefore confirmed that the color
change was indeed a result of catalytic hydrosilation of the RF
group of the dye by running a conventional reaction with Et2SiH2
on dyes 3 and 4 in an NMR tube. The identity of the hydrosilated
dyes was verified by NMR spectroscopy.11
time for dye bleachinga (min)
dye 3 (alkene)
ti tf tf - ti
0:05 >45c >45c 0:05 >45c >45c
0:03 0:35 0:32 0:02
dye 4 (imine)
compd
ti tf tf - ti
1
2
3
4
5
6
7
8
9
[Ir(cod)(PPh3)2]BF4
[Rh(cod)(PPh3)2]PF6
[Rh(nbd)(PPh3)2]PF6
RhCl(PPh3)3
[Rh(octanoate)2]2
RuCl2(PPh3)3
Because of the substantially different ꢀ values of alkene and
imine dyes, a higher substrate/catalyst ratio was required to mask
the catalyst color for the imine versus alkene dye, making the
comparison of rates for alkene and imine unreliable. In future
work, we plan to make a more intense imine dye to allow direct
comparisons. Catalyst solubility can also be a limitation (entry
12).
9
9
0:04
0:03
1
2
0:56 0:15 1:45 1:30
1:57 0:03 2:57
NRb NR
3
NRb NR
0:10 >45c >45c 0:10 >45c >45c
NiCl2(PPh3)2
NRb NR
NRb NR
NR NR
NRb NR
NRb NR
NRb NR
[Ni(tss)]2
Cp2ZrClH
10 [Pd(Ar2PC6H4CH2)OAc]2 0:01 1:15 1:14 0:01 1:15 1.14
To see if the dye substrates have reactivity comparable to that
of a conventional substrate, we looked at a conventional reaction
of the best catalysts for Ph2SiH2 hydrosilation of cyclooctene.
Because of the high dilution in the assay, good NMR data could
only be obtained in more concentrated solutions. This and the
presence of initial lag times for some catalysts made direct
quantitative comparison of NMR and assay data difficult, but the
qualitative rate patterns were maintained between assay and NMR
data; for example, in the case of dye 3, the catalysts of entries 4
and 10 were clearly the fastest catalysts, followed by entries 1-3.
The dyes are activated substrates, however, and react much faster
than standard alkenes such as cyclooctene, a feature which is
helpful for rapid screening because many hydrosilation reactions
normally require elevated temperature and long reaction times.
11 [(nbd)Rh(triphos)]SbF6
12 PtCl2(NH3)2
NR NR
d
NR NR
d
a cod ) 1,5-cyclooctadiene; nbd ) norbornadiene; tss ) salicylal-
dehyde thiosemicarbazonate (2-) (o-O-C6H4CHdN-N-C(S)NH);
triphos ) Ph2PC2H4PPhC2H4PPh2; Ar ) o-tolyl. ti, initial bleaching
time; tf, final bleaching time; NR, no discernible initial loss of color
over 45 min. Times in min or min:s. b Very weak activity evident for
these catalysts only at much higher (0.2-1.0 µmol/well) catalyst
loadings; these were considered too slow to be useful. c Cases in which
catalyst color tends to mask dye color and somewhat obscures final
bleaching. d Compound insufficiently soluble in THF (or any common
organic solvent) for meaningful assay.
stock solutions, added to the appropriate wells with an adjustable
microliter pipet, delivered the following molar amounts to each
well: catalyst, 66 nmol; alkene dye, 66 nmol; imine dye, 130
nmol; silane, 5.0 µmol. The amounts were chosen so that the
dye masks the catalyst color. The fastest catalysts (entries 1-4,
6 and 10) were also successfully run at a dye/catalyst mole ratio
of 5 to show that a stoichiometric reaction of catalyst with dye is
not responsible for bleaching. The silane is involatile, and the
solvent is sufficiently involatile for room temperature use. Manual
recording of color changes was supplemented by recording images
on a digital camera.9g Catalyst + dye controls were always run,
and dye + silane and catalyst + silane controls were run once.
Two color change events, initial and final, were recorded,
corresponding to the first sign of bleaching and full bleaching,
respectively. The latter is somewhat less reliable when the
catalyst color interferes, but the times were reproducible (initial
(5%; final (15%). The order of initial times does not always
correlate with the order of final times for different catalysts. The
initial time is likely to be most influenced by any lag time due to
slow catalyst activation and the difference between initial and
final times to the rate of the activated catalyst. For the
concentrations used in the screens, the earliest visually detectable
bleaching corresponds to ca. 40% conversion and complete visual
bleaching to ca. 95% dye conversion.
We conclude that the use of dyes containing suitable reactive
functionalities is a useful, inexpensive method for the rapid visual
screening of potential homogeneous catalysts and that the data
obtained are relevant to the reactions of standard substrates
containing the same functionality. In future work, we will report
on the application of related soluble dyes to assay polymer-bound
catalysts obtained from a combinatorial phosphine library.12
Acknowledgment. We thank the U.S. Department of Energy for
funding, Roy Periana, John Hartwig, and Istvan Horvath for helpful
discussions, Hye Kwon and James Kuo for experimental assistance, and
Jennifer Loch and Christopher Beck for providing catalysts.
Supporting Information Available: General procedures for the assay
and synthesis of reactive dyes (1 page, print/PDF). See any current
masthead page for ordering information and Web access instructions.
JA9818607
(10) A glovebag, used in initial runs, was satisfactory for the more air-
stable catalysts.
(11) The hydrosilation of the dyes was verified by 1H NMR and IR
spectroscopy. Hydrosilation of 3: characteristic product resonances grow in
at δ 4.58 (m, 1H, C-SiHEt2), δ 4.08 (br s, 1H, CH-SiHEt2), and δ 2.38 (s,
2H, CH2) while the olefinic resonances at δ 7.12 and δ 6.18 disappear. The
ν(CdC) band of 3 at 1675 cm-1 disappears and the ν(Si-H) band of the
hydrosilated product appears at 2110 cm-1. Hydrosilation of 4: characteristic3c
product resonances grow in at δ 4.47 (m, 1H, N-SiHEt2) and δ 4.58 (br s,
2H, N-CH2-py), while the imine proton resonance at δ 8.42 disappears. The
ν(CdN) band of 4 at 1636 cm-1 disappears and the ν(C-N) and ν(Si-H)
bands of the hydrosilated product at 1312 and 2112 cm-1, respectively, both
appear.
The bleaching times for the catalysts (Table 2) show that
Wilkinson’s catalyst is, as expected, among the most highly active.
The most active for initial bleaching, however, is the new
palladacyclic Heck reaction catalyst (entry 10) of Herrmann9d et
al., a compound not previously considered for hydrosilation. Even
in so small a group of catalysts, we therefore have a significant
new hit.
(12) (a) Initial data12b suggest that other related soluble dyes can be useful
for screening combinatorial libraries on polymer beads. (b) Cooper, A.; Loch,
J.; Crabtree, R. H. J. Am. Chem. Soc., in preparation.
The bleaching process indicates a change has taken place, such
as loss of conjugation between A and D groups, but does not