Solution-phase combinatorial synthesis using multicomponent Grignard reagents

Article abstract of DOI:10.1039/b109952n

Mixtures of up to 8 different alkyl Grignard reagents were prepared from the reaction of alkyl halides with magnesium. These multicomponent Grignard reagents were reacted with aldehydes, ketones or esters to produce mixtures of alcohols. Analysis of the p

Full text of DOI:10.1039/b109952n

View Article Online / Journal Homepage / Table of Contents for this issue
Solution-phase combinatorial synthesis using multicomponent
Grignard reagents
Xifu Liang and Mikael Bols*
1
Department of Chemistry, University of Aarhus, Langelandsgade 140 Aarhus C, DK-8000,
Denmark. E-mail: mb@kemi.aau.dk
Received (in Cambridge, UK) 31st October 2001, Accepted 7th January 2002
First published as an Advance Article on the web 22nd January 2002
Mixtures of up to 8 different alkyl Grignard reagents were prepared from the reaction of alkyl halides with
magnesium. These multicomponent Grignard reagents were reacted with aldehydes, ketones or esters to produce
mixtures of alcohols. Analysis of the product mixtures showed that from the reactions of aldehydes and ketones
uniform mixtures of products were formed, while from the reaction with esters uniform libraries were only
obtained after slow addition of the Grignard reagent.
Grignard reactions to the synthetic chemist is its generality as a
Introduction
tool to synthesise compounds with an impressive range of
structures and functional groups. Grignard reagents act as both
prototypical carbon nucleophiles that can undergo addition or
substitution reactions and as strong bases that can deprotonate
acidic substrates, giving conjugate bases or products of
elimination reactions. These reagents react with most organic
functional groups containing polar multiple bonds (e.g.,
carbonyl, nitriles, sulfones, imines), highly strained rings
(e.g., epoxides and some cycloalkenes), acidic hydrogens (e.g.,
alkynes), and some highly polar single bonds (e.g., carbon–
halogen or metal–halogen). Some representative reactions are
outlined in Scheme 2.
Combinatorial chemistry is a range of techniques, some
possibly automated, that permits the rapid synthesis of a large
number of chemical compounds usually through the ‘combin-
ation’ of several different synthetic building blocks (Scheme
1).1–3 Combinatorial chemistry started taking shape in the
Scheme 1 Solution-phase combinatorial synthesis of a mixed library
of n times m products AB from n reagents A and m reagents B.
mid-80’s starting in peptide-chemistry.4,5 However biologically
active peptides do not normally make good drugs, as they are
poorly absorbed orally or are subject to rapid metabolism and
excretion, so in the early 1990’s, several foresighted organic
chemists began to make combinatorial libraries of small drug-
like organic molecules.6,7 Today, combinatorial chemistry has
become an important tool for the discovery and optimisation
processes for novel drugs, affinity ligands, new materials, and
catalysts. The technology has been applied in both academic
and industrial institutions to provide a number of unique
approaches to satisfy the ever-growing need for new chemical
entities with proven or new utility.
While combinatorial chemistry has its origin in solid-phase
synthesis, combinatorial methods based on chemistry in sol-
ution is increasingly popular.8,9 Though solution-phase com-
binatorial chemistry potentially offers some advantages over
solid-phase combinatorial chemistry (such as a wider number
of accessible reactions, easy monitoration by TLC, HPLC etc.,
less limitation on solvents and temperatures, unlimited reaction
scale-up) combinatorial chemistry of mixed libraries is carried
out on solid phase for good reason. The insoluble reactants
ensures that forcing a reaction will produce each library
member in equal amount. The lack of control of the rate of
individual library reactions in solution makes it difficult to
make uniform mixed libraries in solution unless the reactions
employed are fast and high-yielding for each library member.
The Grignard reaction is one of the most frequently used
name reaction in organic synthesis.10 The great value of
Scheme 2 Some of the many reactions that can be carried out with
Grignard reagents.
It would be highly desirable if the Grignard reaction could be
used in solution-phase combinatorial chemistry. By one-pot
reaction between a series of Grignard reagents and an electro-
phile it would be possible to synthesise in a single step a mixture
of analogously C-substituted compounds that could be tested
for a specific application. To be useful it requires however that
one can depend on each product being present in the product
DOI: 10.1039/b109952n
J. Chem. Soc., Perkin Trans. 1, 2002, 503–508
This journal is © The Royal Society of Chemistry 2002
503

View Article Online
mixture, which can only be ensured if each individual Grignard
reaction reacts with roughly equal rate or if each reaction can
be forced to run to the same degree of completion. Based on
what is known about the Grignard reaction the chances for
the first event occurring seemed grim. The rate of reaction
of various alkylmagnesium halides has been shown to vary
widely.11 Thus the ratio of the pseudo-first order rate constants
in reaction of CH3MgBr, C2H5MgBr and n-C4H9MgBr with
acetone are 2 : 4 : 1. The rate of reaction of Grignard reagents
with benzophenone is even more variable.
The purpose of the present work was to investigate whether
uniform compound libraries could be produced through reac-
tion of Grignard reagents with aromatic and aliphatic ketones,
aldehydes or esters. Our strategy involved reaction of a sub-
strate S with multiple Grignard reagents, R1, R2, R3 ؒ ؒ ؒ Rn, to
produce a compound library of n individual products SR1, SR2,
SR3 ؒ ؒ ؒ SRn (Scheme 3), followed by careful investigation of
Fig. 1 HPLC chromatogram of the library obtained from the reaction
of benzaldehyde (1) with a series of Grignard reagents (Scheme 5). The
column was a silica gel column (Nucleosil) and the eluent n-hexane–
propan-2-ol 97 : 3.
Scheme 3 Solution-phase combinatorial synthesis of alcohols as
carried out in this work. S is the substrate (aldehyde, ketone or ester)
and R a Grignard reagent.
whether or to what extent the individual library members were
formed. Mixtures of Grignard reagents were synthesised in
one pot from the halides and magnesium. To the best of our
knowledge, no Grignard reaction has been used in compound
library synthesis, though Grignard reagents have been used in
solid-phase synthesis.12
Results and discussion
Scheme 4 Reaction of benzaldehyde (1) with three Grignard reagents
in equimolar amounts (0.67 equiv. of each with respect to 1).
The most challenging problem in solution phase combinatorial
synthesis is whether a reaction proceeds uniformly and, in other
words, whether a substrate reacts equally with each of the
reagents. For this purpose, a series of simple experiments were
designed. Benzaldehyde (1) was reacted with a mixture of
three different Grignard reagents (CH3MgI, n-C3H7MgBr and
n-C8H17MgBr), which were chosen as typical examples of
straight chain aliphatic Grignard reagents. The mixture of
Grignard reagents in diethyl ether was prepared either in one-
pot or by sequentially adding each halide to excess magnesium.
When using the latter procedure observation of the exothermic
reaction of each halide can be used to ensure that formation of
each Grignard reagent has taken place. The 1 : 1 : 1 mixture of
Grignard reagents, two equivalents in total, was added quickly
to a solution of benzaldehyde in diethyl ether at room temper-
ature giving after stirring for 2–3 h and usual workup a mixture
of secondary alcohols. The obtained mixture was first subjected
Scheme 5 Reaction of benzaldehyde (1) with five Grignard reagents in
equimolar amounts (0.4 equiv. of each with respect to 1).
The complexity was then enhanced to generate an 8-
membered library using the exact same conditions on the reac-
tion of 1 with the series of Grignard reagents from methyl
to octylmagnesium halide (Scheme 6). While NMR analysis
1
to H NMR analysis, which showed that no starting material
was left and was consistent with a roughly equal mixture of the
products 2, 3 and 4. HPLC analysis was then carried out using
individually prepared pure alcohols 2–4 as references, which
showed that the Grignard products had been formed in a ratio
of 4 : 5 : 6 (Scheme 4). This showed clearly that the reactivities
of the three reagents were quite similar, which was somewhat
surprising given the kinetic information available about the
Grignard reaction.10,11 However the fact that only 0.67 equiv-
alents of each Grignard reagent is present ensures a more even
distribution of products that would otherwise occur.
In exactly the same manner, addition of five different
Grignard reagents, CH3MgI, C2H5MgBr, n-C3H7MgBr, n-C4H9-
MgBr, n-C5H11MgBr, to benzaldehyde was carried out (Scheme
5). NMR and HPLC (using comparison with pure reference
compounds) showed that a library of the five compounds 2, 3,
5, 6 and 7 had been obtained, with the ratio between 2, 3, 5, and
6 ϩ 7 being 1 : 1 : 1 : 2. Since no baseline separation was
obtained between compounds 6 and 7 the exact amount of each
of these two could not be quantified (Fig. 1).
Scheme 6 Reaction of benzaldehyde (1) with eight Grignard reagents
in equimolar amounts (0.25 equiv. of each with respect to 1).
showed total conversion and was consistent with the formation
of each of the products 2–9, HPLC-analysis could not separate
all products in this case, and only 2, 3 and 5 were base-line
separated. However, integration of these 3 peaks, and the sum
504
J. Chem. Soc., Perkin Trans. 1, 2002, 503–508

View Article Online
of integration of 4, 6, 7, 8 and 9 was consistent with each
library contributing 12% to the product mixture, and that equal
amounts of each compound had been formed. These experi-
ments convinced us that libraries of simple linear Grignard
reagents can react with aromatic aldehydes to form uniform
libraries of products without the need for special tricks or
precautions.
We also carried out experiments with an aliphatic aldehyde as
the behavior of alkylmagnesium halides towards aliphatic and
aromatic ketones can be widely different.10,11 Therefore the 8
membered mixture of Grignard reagents used in the experiment
above was also added in two fold excess (total) to phenyl-
propanal (10). Also in this experiment NMR and HPLC was
consistent with total conversion of 10 into an 8-membered
library of the alcohols 11–18 (Scheme 7). Again not all
tiary alcohols were formed in 1 : 1 : 1 ratio. In these reactions
significant amounts, 3% to 10%, of secondary alcohols formed
from reduction of the ketones 19 and 20 were observed. Reduc-
tion of the carbonyl compound is a common byproduct in
Grignard reactions especially with hindered ketones.13
Finally the combinatorial reaction of Grignard reagents with
an ester was investigated. As each ester will react with two
equivalents of Grignard reagents this is a considerably more
complex situation. Firstly multiple Grignard reagents can react
with the same ester creating a diversity of tertiary alcohols
through ‘combination’. Secondly the kinetics are complicated
by both the ester and ketones acting as electrophiles in the reac-
tion. In a preliminary experiment, the reaction conditions that
had been used in the reaction of ketones and aldehydes above
were used i.e. 4 equivalents of Grignard mixture was quickly
added to the ester. Thus, a mixture of CH3MgI and n-C8H17-
MgBr was reacted with methyl benzoate (26) at room temper-
ature giving full conversion to the three possible products 21, 23
and 27 (Scheme 9). If the rate of reaction between the two
Grignard reagents and 26 and the intermediate ketones were
identical, the theoretical ratio between the products 21, 23 and
27 would be 1 : 2 : 1 since 23 is racemic and is formed in two
ways. The observed ratio of 21– 23–27 was however 5 : 5 : 1,
which was determined from HPLC analysis using reference
compounds. This result shows that octylmagnesium bromide is
much less reactive than the methylmagnesium iodide towards
the ester functionality. We already know from the ketone
experiments above that CH3MgI and n-C8H17MgBr have the
same reactivity towards 19 or 20 so the excess formation in
methylated products must be caused in the ester reaction. The
result is consistent with an initial conversion of 26 to a 5 : 1
mixture of methyl and octylketone, which then uniformly react
with the methyl and octyl Grignard reagents. The ester 26 was
much less reactive than the other carbonyl compounds investi-
gated in this work, and while most of the above mentioned
reactions were practically instantaneous the reaction between
26 and n-C8H17MgBr took more than 30 min to complete. To
overcome the unequal reactivities of CH3MgI and n-C8H17-
MgBr towards 26, the mixture of these two Grignard reagents
was added slowly using a syringe pump to a solution of 26 over
8 h. In this way added Grignard reagents are allowed to react
before more reagent is added. This afforded a ratio of the three
Scheme 7 Reaction of 10 with eight Grignard reagents in equimolar
amounts (0.25 equiv. of each with respect to 10).
products could be separated on HPLC. The chromatogram
showed seven peaks of which only three were base-line separ-
ated. However the integrals of the individual and mixed peaks
were such that it was consistent with an equal amount of each
product.
The reaction of ketones was now investigated. Two ketones,
acetophenone (19) and 1-phenylbutan-1-one (20), were each
reacted with 2 equivalents of a mixture of CH3MgI, n-C3H7-
MgBr and n-C8H17MgBr. Each reaction gave total conversion
of ketone and the expected products 21–23 and 22, 24, 25,
respectively (Scheme 8). HPLC analysis showed that these ter-
Scheme 8 Reaction of ketones 19 and 20, respectively, with three Grignard reagents in equimolar amounts (0.67 equiv. of each with respect to
ketone).
Scheme 9 Reaction of methyl benzoate (26) with two Grignard reagents in equimolar amounts (2 equiv. of each with respect to 26).
J. Chem. Soc., Perkin Trans. 1, 2002, 503–508
505

View Article Online
Fig. 2 Products from the reaction of methyl benzoate (26) with n-C4H9MgBr, n-C5H11MgBr, n-C6H13MgBr, n-C7H15MgBr and n-C8H17MgBr in
equimolar amounts (0.8 equiv. of each with respect to 26). The compounds have been organised so that compounds on South-west/North-east
diagonals have identical molecular weights.
products 21, 23 and 27 of 3 : 10 : 4 and thus relatively close to
the statistical value of 1 : 2 : 1.
Now the experiment was carried out with five Grignard
reagents. n-C4H9MgBr, n-C5H11MgBr, n-C6H13MgBr, n-C7H15-
MgBr and n-C8H17MgBr were reacted with 26 using the slow
addition procedure giving potentially 25 different products if
counting stereoisomers or 15 different products if not dis-
criminating between enantiomers. These products are listed in
Fig. 2 together with their molecular weights. While the com-
plete conversion of the ester could be ensured by NMR it was
impossible to determine the presence of individual library
members with NMR or HPLC analysis due to the complexity
of the library, but the mixture could be analysed by electro-
spray mass spectroscopy. As can be seen from Fig. 2 a consider-
able number of compounds have identical molecular weights. In
fact between the 25 compounds only 9 different molecular
weights exist. As can be seen from Fig. 2 the number of com-
pounds with each molecular weight is distributed in the ratio
Fig. 3 Electrospray mass spectrum of the product mixture obtained
1 : 2 : 3 : 4 : 5 : 4 : 3 : 2 : 1 from smallest to largest value (Fig. 3).
While the MS spectrum does show the 9 individual peaks,
and even though the intensity of the 9 peaks was quite close to
the statistical value, the peak intensities cannot be relied on
to reflect the concentration of each compound. Therefore the
individual concentration of each substance may be different.
An aliphatic acid ester, methyl 3-phenylpropanoate was also
reacted with the five Grignard reagents using the same pro-
cedure giving a similar result (Fig. 4). Again all nine peaks could
be observed from the MS spectrum.
from reaction of 26 with five Grignard reagents (Fig. 2). The X marked
peaks correspond to expected compounds in the product mixture.
of secondary and tertiary alcohols can be made by solution-
phase combinatorial synthesis using multicomponent Grignard
reagents. This methodology is easy to use. Depending on the
carbonyl compound some adaptation may be needed, but
with slow addition of the Grignard reagent mixture uni-
form libraries must be expected. Further investigations will
focus on using this methodology in the search for useful
compounds.
In summary, we have shown that uniform mixed libraries
506
J. Chem. Soc., Perkin Trans. 1, 2002, 503–508

View Article Online
Fig. 4 Products from the reaction of methyl 3-phenylpropanoate with n-C4H9MgBr, n-C5H11MgBr, n-C6H13MgBr, n-C7H15MgBr and n-C8H17MgBr
in equimolar amounts (0.8 equiv. of each with respect to 3-phenylpropanoate). The compounds have been organised so that compounds on South-
west/North-east diagonals have identical molecular weights.
the solution was first added an aqueous saturated solution of
Experimental
NH4Cl (20 ml) dropwise and then H2O (20 ml). After stirring
All reagents and chemicals were commercially available. Mass
for 10 min, the two phases were separated. The aqueous phase
spectra were carried out on a Micromass LC-TOF instrument.
was extracted with diethyl ether (20 ml × 2). The combined
organic phases were dried over MgSO4 and concentrated in
1H-NMR spectra were obtained on a Varian Gemini 200
instrument.
vacuo to give an alcohol. Products 2, 3, 5, 6, 11, 21 and 22 were
commercially available, while products 7,14 8,14 9,15 4,16 12,17
General procedures for the Grignard reactions
13,18 14,19 15,20 16,21 18,22 23,23 2424 and 2725 were otherwise
known. The identity of these substances was confirmed by
comparison with literature or spectral libraries. Compounds17
and 25 were previously unknown, and their data are thus given
here.
HPLC analysis of the compound libraries: The composition of a
library was determined on a HP Instrument using a Nucleosil
column. (eluent: n-hexane–propan-2-ol 97 : 3; flow rate: 0.5 ml
minϪ1, UV: 210 nm). Within the compound series studied mole-
cules with a higher molecular weight had shorter retention times
than those with lower molecular weight.
1
1-Phenyldecan-3-ol (17). H-NMR (CDCl3): δ 7.2 (m, 5H,
Ph), 3.8 (m, 1H, H3), 2.7 (m, 2H, H1’s), 1.8 (m, 2H, H2’s), 1.3
(m, 12H, H4–H10’s), 0.9 (m, 3H, Me). MS(ES): (obtained on
acetate ester) m/z 299.1982, Calc for C18H28O2 ϩ Na: 299.1987.
Preparation of single component Grignard reagents
Alkyl halide (10 mmol) was added so slowly to a mixture of Mg
(20 mmol) in diethyl ether (20 ml) that the mixture was kept
refluxing. After addition, the mixture was stirred for 2 h. The
solution of Grignard reagent was taken out using a syringe and
employed.
1
4-Phenyldodecan-4-ol (25). Bp: 198 ЊC, H-NMR (CDCl3):
δ 7.3 (m, 5H, Ph), 1.8 (m, 4H, H3’s, H5’s), 1.3 (m, 14H, H2’s,
H6–H11’s), 0.8 (m, 6H, Me’s). IR: 3390 (st, OH), 2926, 2854
(st, CH), 1466 (m, CH2, CH3). MS(ES): m/z 285.2196, Calc. for
C18H30O ϩ Na: 285.2194.
Preparation of multicomponent Grignard reagents
Several different alkyl halides were mole-equivalently mixed
together to give a liquid mixture (30 mmol). The solution was
added so slowly to a mixture of Mg (60 mmol) in diethyl ether
(20 ml) that the mixture was kept refluxing. After addition, the
mixture was stirred for 2 h. The solution of Grignard reagents
was taken out using syringe and employed. Alternatively each
alkyl halide was added dropwise one by one in mole-equivalent
amounts (30 mmol total) to the Mg (60 mmol) in diethyl ether
(20 ml) at such a rate that the solution was kept refluxing. This
ensures that the failure of an alkyl halide to react is detected.
Reaction of a ketone or an aldehyde with multicomponent
Grignard reagent
To a solution of a ketone or an aldehyde (10 mmol) in diethyl
ether (20 ml) was added a solution of multicomponent
Grignard reagent (30 mmol) at 0 ЊC. The reaction solution was
stirred at room temperature for 2–3 h. After reaction, the sol-
ution was cooled to 0 ЊC. To the solution was first added an
aqueous saturated solution of NH4Cl (20 ml) dropwise and
then H2O (20 ml). After stirring for 10 min, the two phases were
separated. The aqueous phase was extracted with diethyl ether
(20 ml × 2). The combined organic phases were dried over
MgSO4 and concentrated in vacuo to give a library of alcohols.
Reaction of a ketone or an aldehyde with a single component
Grignard reagent
To a solution of a ketone or aldehyde (5 mmol) in diethyl ether
(20 ml) was added a solution of Grignard reagent (10 mmol)
at 0 ЊC. The reaction solution was stirred for 2–3 h at room
temperature. After reaction, the solution was cooled to 0 ЊC. To
Reaction of an ester with multicomponents Grignard reagents
To a solution of an ester (5 mmol) in ether (10 ml) was added
a solution of a multicomponent Grignard reagent (20 mmol,
J. Chem. Soc., Perkin Trans. 1, 2002, 503–508
507

View Article Online
10 eds. G. Silverman and P. E. Rakita, Handbook of Grignard Reagents,
Marcel Dekker, Inc., New York, 1996.
40 ml) slowly (8 h) at room temperature. After stirring for 8 h at
the same temperature, the reaction mixture was cooled down to
0 ЊC. To the mixture was first added an aqueous saturated sol-
ution of NH4Cl (20 ml) dropwise and then H2O (20 ml). After
stirring for 10 min, the two phases were separated. The aqueous
phase was extracted with diethyl ether (20 ml × 2). The com-
bined organic phases were dried over MgSO4 and concentrated
in vacuo to give a library of tertiary alcohols.
11 T. Holm, Acta Chem. Scand. Ser. B, 1983, 37, 567–584.
12 R. G. Franzen, Tetrahedron, 2000, 56, 685–691.
13 F. A. Carey and R. J. Sundberg, Organische Chemie: Ein
weiterführendes Lehrbuch, VCH, Weinheim, 1995, p. 1120.
14 J. A. Gautier, M. Miocque and L. Mascrier-Demagny, Bull. Soc.
Chim. Fr., 1967, 1554–1560.
15 M. S. F. Jie, W. L. K. Lam and H. B. Lao, J. Chem. Soc., Perkin
Trans. 1, 1989, 1–11.
16 R. A. O’Brien and R. D. Rieke, J. Org. Chem., 1990, 55, 788–790.
17 M. Julia and D. Uguen, Bull. Soc. Chim. Fr., 1976, 513–518.
18 L. N. Cherkasov, G. I. Pis’mennaya and A. N. Krivosheya, Fiz.,
Khim. Khim. Tekhnol. Sb. Mater. Nauchno-Tekh. Konf. Rab. Nauki
Proizvod., 1969, 217–220.
References
1 N. K. Terrett, Age of the Molecules, ed. N. Hall RSC, London, 1999,
pp. 28–32.
2 ed. H. Fenniri, Combinatorial Chemistry, University Press, Oxford
2000.
19 A. Burger and S. N. Sawhney, J. Med. Chem., 1968, 11, 270–
273.
3 N. K. Terrett, Combinatorial Chemistry, University Press, Oxford,
1998.
20 A. Liguori, G. Sindona and N. Uccella, Gazz. Chim. Ital., 1984, 114,
369–373.
4 R. A. Houghten, C. Pinilla, S. E. Blondelle, J. R. Appel, C. T.
Dooley and J. H. Cuervo, Nature, 1991, 354, 84–86.
5 K. S. Lam, S. E. Salmon, E. M. Hersch, V. J. Hruby, W. M.
Kazmierski and R. J. Knapp, Nature, 1991, 354, 82–84.
6 B. A. Bunin and J. A. Ellman, J. Am. Chem. Soc., 1992, 114,
10997–10998.
21 M. L. Poutsma and P. A. Ibarbia, J. Org. Chem., 1969, 34,
2848–2855.
22 S. Nimgirawath, E. Ritchie and W. C. Taylor, Aust. J. Chem., 1973,
26, 183–193.
23 S. Hata, Y. Yano, H. Matsuda and S. Matsuda, Kogyo Kagaku
Zasshi, 1968, 71, 704–709.
7 B. A. Bunin, M. J. Plunkett and J. A. Ellman, Proc. Natl. Acad. Sci.
U.S.A., 1994, 91, 4708–4712.
8 H. An and P. D. Cook, Chem. Rev., 2000, 100, 3311–3340.
9 C. M. Baldino, J. Comb. Chem., 2000, 2, 89–103.
24 R. Levine, M. J. Karten and W. M. Kadunce, J. Org. Chem., 1975,
40, 1770–1773.
25 P. H. Doe, M. El-Emary, W. H. Wade and R. S. Schechter, J. Am. Oil
Chem. Soc., 1977, 54, 570–577.
508
J. Chem. Soc., Perkin Trans. 1, 2002, 503–508