No stirring necessary: Parallel carbonylation of aryl halides with CO and various alcohols under pressure

Article abstract of DOI:10.1021/jo034112v

The parallel carbonylation of aryl halides with 6-25 bar of CO in 1-mL vials in a standard autoclave was investigated. 4-Bromoacetophenone and 2-chloropyridine were used as model substrates with 102 different O-nucleophiles (primary and secondary alcohols, phenols). No inertization during the loading was necessary. Fifty esters (43 new, yield up to 60%) were isolated and characterized. Ether, ester, ketone, and sometimes even olefin functions were usually tolerated. The new method is suitable for screening and small scale products synthesis.

Full text of DOI:10.1021/jo034112v

SCHEME 1. Ca r bon yla tion of
4-Br om oa cetop h en on e
No Stir r in g Necessa r y: P a r a llel
Ca r bon yla tion of Ar yl Ha lid es w ith CO a n d
Va r iou s Alcoh ols u n d er P r essu r e
Hans-Ulrich Blaser,† Martin Diggelmann,‡ Hans Meier,†
Fre´de´ric Naud,† Elke Scheppach,‡ Anita Schnyder,† and
Martin Studer*,†
Solvias AG, R-1055.6, CH-4002 Basel, Switzerland, and
Syngenta AG, R-1060, CH-4002 Basel, Switzerland
martin.studer@solvias.com
Received J anuary 28, 2003
Abstr a ct: The parallel carbonylation of aryl halides with
6-25 bar of CO in 1-mL vials in a standard autoclave was
investigated. 4-Bromoacetophenone and 2-chloropyridine
were used as model substrates with 102 different O-
nucleophiles (primary and secondary alcohols, phenols). No
inertization during the loading was necessary. Fifty esters
(43 new, yield up to 60%) were isolated and characterized.
Ether, ester, ketone, and sometimes even olefin functions
were usually tolerated. The new method is suitable for
screening and small scale products synthesis.
SCHEME 2. P a r a llel Ca r bon yla tion of
2-Ch lor op yr id in e
Palladium-catalyzed reactions are an important tool
in organic synthesis for the formation of C-C, C-N, and
C-O bonds and, indeed, various technical applications
are found in the fine chemicals industry.1 Due to the
versatility and high functional group tolerance of most
Pd catalysts, there is also growing interest in parallel
synthesis with the goal of applying these reactions in
combinatorial chemistry and to synthesize compound
libraries.2 Carbonylation with CO in the presence of var-
ious nucleophiles is of special importance due to the use-
fulness of the products and the high atom economy of
the method.3 However, using the gaseous and highly toxic
CO under pressure at elevated temperatures requires
high-pressure autoclaves. Since it was always assumed
that stirring of the biphasic reaction mixture is necessary,
this methodology has so far be limited to single experi-
ments. The use of alternative CO sources (no gaseous CO
required) was recently described by Hosoi et al. for the
synthesis of amides with DMF and POCl3,4 and by Kaiser
et al. for microwave assisted reactions with Mo(CO)6, al-
beit at rather high temperature.5 However, both methods
only have a limited value for screenings. To the best of
our knowledge, no simple parallel procedure for carbo-
nylations with CO has been reported up to now. Our goal
was to show that it is possible to perform multiple para-
llel carbonylations in a standard autoclave without the
need to stir, and that the new method allows catalyst
screening as well as the synthesis and isolation of acid
derivatives on a milligram scale.
solubility for CO, carbonylation without stirring might
be possible at elevated pressure. In preliminary tests, we
carried out a few experiments with 4-bromoactophenone,
a substrate that is relatively easy to carbonylate.6,7 For
this purpose we used 1-mL vials closed with a rubber sep-
tum perforated with a small hollow shaft needle. The
loading was carried out in air, and the vials were placed
in a 300-mL autoclave. To ensure a good heat transfer,
the autoclave was always filled with approximately 200
mL of quartz sand.
Working in neat n-butanol we obtained 100% conver-
sion in 14 h without stirring. The isolated yield was 28%
compared to 72% previously reported with stirring (ref
6; see Scheme 1). Even though this yield is significantly
lower, it is nevertheless useful for preparations on small
scale and the synthesis of compound libraries. One
possible explanation for the low yield is a stronger
competition by reduction due to lower CO concentration
in the solution. Next we investigated the effect of solvents
containing 4 equiv of n-butanol. DMF, acetonitrile, dig-
lyme, and DMSO were found to be suitable, as indicated
by a large product peak in the LC trace. Similar results
were obtained with other simple alcohols. However,
4-bromoacetophenone is very reactive and therefore not
a very representative substrate. Furthermore, many of
the esters produced were not suitable for analysis and
isolation on a standard high throughput setup since the
LC-MS typically uses electrospray ionization. Therefore,
most of the products were impossible to detect by MS,
and product analysis and workup were extremely time-
consuming.
The experimental conditions for carbonylation were
chosen as described by Ma¨gerlein et al.6 (See Scheme 1).
We reasoned that if we use solvents with relatively high
† Solvias AG.
‡ Syngenta AG.
(1) Zapf, A.; Beller, M. Top. Catal. 2002, 19, 101.
(2) Schiedel, M.-S.; Briehn, C. A.; Ba¨uerle, R. J . Organomet. Chem.
2002, 653, 200.
(3) Skoda-Foldes, R.; Kollar, L. Curr. Org. Chem. 2002, 6, 1097.
(4) Hosoi, K.; Nozaki, K.; Hiyama, T. Org. Lett. 2002, 4, 2849.
(5) Kaiser, N.-F. K.; Hallberg, A.; Larhed, M. J . Comb. Chem. 2002,
4, 109.
(6) Ma¨gerlein, W.; Beller, M.; Indolese, A. F. J . Mol. Catal. A: Chem.
2000, 156, 213.
(7) (a) Beller, M.; Ma¨gerlein, W.; Indolese, A. F.; Fischer, C.
Synthesis 2001, 1098 and references cited. (b) Beller, M.; Indolese, A.
F. Chimia 2001, 684 and references cited.
10.1021/jo034112v CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/05/2003
J . Org. Chem. 2003, 68, 3725-3728
3725

F IGURE 1. Isolated yields in the carbonylation of 2-chloropyridine in the presence of primary alcohols. A single asterisk indicates
1
isolated esters containing 2-10% impurities according to H NMR. A double asterisk identifies 11-30% impurities found in the
isolated esters.
For these reasons we decided to use 2-chloropyridine,
which is a more relevant substrate and is also established
as a model.7a Similar reaction conditions proved to be
optimal, except for the ligand. Ligand screening experi-
ments indicated that 1,3-bis(diphenylphosphino)propane
(dppp) was better suited than 1,3-bis(diphenylphosphi-
no)ferrocene (dppf) under our conditions (Scheme 2).
Again high conversion and 48% yield were obtained with
n-butanol/DMF in preliminary experiments without stir-
ring.
The next step was to vary the alcohol component to
check the scope of our setup. For this purpose we chose
102 alcohols and phenols with a wide structural diversity:
59 primary alcohols (Figure 1), 31 secondary alcohols
(Figure 2), and 12 phenols (Figure 3). To test functional
group tolerance, some of these also carried reactive
functional groups such as alkenes, alkynes, and others.
In all cases, 12 parallel reactions were carried out in
1-mL vials placed in the 300-mL autoclave. The results
are summarized in Figures 1-3; for clarity, only ranges
of isolated yields are given in the figures. As before, the
conversion of 2-chloropyridine was high in all reactions
investigated. This means that in cases where no product
was isolated either dehalogenation had taken place or
other side products were formed. Since the scope of this
investigation was the development of a preparative
procedure, the side products were not analyzed further.
Of the 59 primary alcohols tested, 30 gave products in
1-60% yield (upper part of Figure 1, for statistical details
see Table 1). A cursory inspection allowed the following
conclusions: (a) ether, ester, and ketones groups were
usually tolerated, and in one case even an unsaturated
alcohol worked; (b) for alcohols containing sulfur groups,
alkynes, aldehydes, or three-membered rings, no ester
could ever be isolated; (c) alcohols with the OH group in
a benzylic position also gave no product except for the
parent benzyl alcohol; and (d) besides these effects, there
was no obvious correlation between the type of functional
group in the alcohol and ester yields. In some cases the
product might be too volatile or too lipophilic to be
detected and isolated with our setup, but there is not
enough evidence to support this hypothesis.
Figure 2 shows the results obtained with secondary
alcohols. As expected these were less reactive nucleo-
philes than the primary alcohols. However, they showed
a very similar pattern concerning functional group toler-
3726 J . Org. Chem., Vol. 68, No. 9, 2003

F IGURE 2. Isolated yields in the carbonylation of 2-chloropyridine in the presence of secondary alcohols. A single asterisk indicates
1
isolated esters containing 2-10% impurities according to H NMR. A double asterisk identifies 11-30% impurities found in the
isolated esters.
F IGURE 3. Isolated ester yields in the carbonylation of 2-chloropyridine in the presence of phenols.
TABLE 1. An a lysis of th e Resu lts Obta in ed in th e
P a r a llel Ca r bon yla tion of 2-Ch lor op yr id in e
products in 7 out of 12 examples. Aromatic aldehydes,
nitro groups, and nitriles were not tolerated, but tertiary
amines, esters, ethers and in one case even an allyl group
were. Furthermore, two sterically hindered phenols with
two ortho substituents also gave the desired product.
primary secondary
nucleophile
alcohol
alcohol
phenol
total
102
49
31/6/13
23
no. of experiments
products isolated
59
30
31
12
12
7
Table 1 shows a statistical analysis of the results. Over
all, about half of the (sometimes highly functionalized)
alcohols tested gave the corresponding esters in useful
purity (>98/>90>70%) 17/5/8
average yield (%) 30
7/1/4
12
7/0/0
14
2
yields. About /3 of the isolated products were analytically
pure (>98% by NMR). Even though the yields were
usually not very high, they were sufficient for analytical
purpose and small-scale (library) synthesis. As shown
above, this method also has some limitations. These can
be due to chemical reasons (functional group tolerance
ance. Here, 12 out of 31 experiments gave the desired
ester and also here a successful example with an unsat-
urated alcohol was found.
Figure 3 shows the results obtained with phenols.
Remarkably, the desired esters could be obtained as clean
J . Org. Chem, Vol. 68, No. 9, 2003 3727

SCHEME 3. Ca r bon yla tion of 2-Ch lor op yr id in e w ith a P h en ol w ith a n d w ith ou t Stir r in g
ridine and 24 µL of Et3N, 600 µL of a stock solution of the alcohol
of carbonylation), analytical (only products containing a
sufficiently basic function can be analyzed by ES+-MS)
or handling problems (loading under air can poison
catalyst), or due to our workup procedure (standard LC
method and isolation procedure might not be suitable for
all products). However, as the system has not been
optimized, it is expected that more products/higher yields
could be obtained if necessary.
To compare the results obtained with 2-chloropyridine
without stirring to those in a stirred autoclave, we
investigated the reaction depicted in Scheme 3 in more
detail. We isolated this sterically demanding ester in 17%
yield with stirring compared to 11%. Thus on small-scale
comparable results can be obtained without stirring even
with relatively unreactive alcohols.
To carry the parallelization one step further, we
performed the same experiments in a standard 96-deep
well plate. Because of its relatively large size, we could
not use the 300-mL autoclave but had to resort to a 16 L
vessel instead. This led to a much longer time span
between the loading of the reaction mixture, the closing
of the autoclave, and the start of the reaction, resulting
most likely in some catalyst decomposition (no inert
conditions). Furthermore, heat transfer was much more
difficult to ensure, handling was more cumbersome, and
last but not least it was not very economical to fill an
almost empty 16-L vessel with the relatively expensive
CO, and using only a very small fraction of it. Neverthe-
less, in many cases we found the desired esters, and even
a few products with alcohols which gave no ester in the
setup described above. Most likely, an optimized ap-
paratus would also allow reliable carbonylation in 96-
deep well plates.
(4 equiv with regard to 2-chloroypridine) in DMF was added.
The vial was closed with a rubber stopper, which was glued to
the vial. To allow gas diffusion, the rubber stopper was perfo-
rated with a small hollow shaft needle. To homogenize the
mixture, the vial was shaken manually and then transferred
into a 300-mL stainless steel autoclave filled with approximately
200 mL of quartz sand. After adding the other 11 vials loaded
with the same procedure, the autoclave was closed and checked
for leaks with He (50 bar). After releasing the helium, the
autoclave was pressurized with CO to 25 bar and heated to 130
°C with a heating rate of 80 °C/h. After 14 h, the autoclave was
cooled to room temperature, depressurized, purged with nitro-
gen, and opened to the atmosphere.
Wor k u p . The reaction mixtures were filtered over cotton to
remove solid Pd-residues. In some cases it was necessary to add
small amounts of DMF to dissolve crystallized Et3N‚HCl. The
clear solutions were analyzed by LC-MS after dissolving 10 µL
of the reaction mixture in 600 µL of CH3CN. Then, the samples
were separated by preparative HPLC-MS. After the separation,
100-µL fractions were drawn from each of the vials containing
the separated products and again analyzed on the analytical LC-
MS machine. The vials containing the products of interest were
evaporated to dryness by blowing warm nitrogen over or in a
vacuum centrifuge. The evaporated vials were weighted to
determine the product yield and the pooled products were
analyzed by NMR.
Sca le-Up Exp er im en t. Syn th esis of 3: Allylphenol (1.0 g),
100 µL of 2-chloropyridine, and 260 µL of Et3N were dissolved
in 25 mL of DMF and placed in a 100-mL autoclave fitted with
a magnet-driven hollow shaft stirrer. After the addition of 23.0
mg of PdCl2(PhCN)2 and 47 mg of dppp, the autoclave was closed
and checked for leaks with He (50 bar). As described for the
parallel experiments, CO (25 bar) was introduced after releasing
the helium and the autoclave was heated to 130 °C with a
heating rate of 80 °C/h. After 14 h, the autoclave was cooled to
room temperature, carefully depressurized, purged with nitro-
gen, opened, and unloaded. The reaction mixture was filtered
through a cartridge filled with silica gel and the filtrate was
purified by flash chromatography (column 3.5 × 35 cm2, EtOAc/
cyclohexane 1:1 as eluent). After removing the solvent, 70 mg
(17%) of the desired ester was isolated as a pale yellow-green
oil. Rf 0.21; 1H NMR (499.9 MHz, CDCl3, 298 K) δ 8.85-8.87
(m, 1 H), 8.29 (d, J ) 7.8 Hz, 1 H), 7.91 (dpt, J ) 7.7 Hz, 1.8 Hz,
1 H), 7.54-7.56 (m, 1 H), 7.19 (pt, J ) 7.9 Hz, 1 H), 6.87-6.90
(m, 2 H), 5.86-5.95 (m, 1 H), 5.00-5.05 (m, 2 H), 3.80 (s, 3H),
3.38 (d, J ) 6.6 Hz, 2 H); 13C{1H} NMR (125.0 MHz, CDCl3, 298
K) δ 162.9, 151.3, 150.2, 147.5, 138.4, 137.1, 135.9, 133.5, 127.2,
126.6, 125.8, 121.8, 116.4, 110.4, 56.0, 34.6; MS (ESP+) 270.1
(M + H), 311.1 (M + CH3CN), 333.1 (M + CH3CN + Na).
In conclusion, we have demonstrated that a simple
reaction setup allows the rapid screening of aryl halide
carbonylation in the presence of alcohols in DMF allowing
the preparation of the corresponding esters on a mil-
ligram scale. No stirring and no inert conditions are
necessary with 4-bromoacetophenone and 2-chloropyri-
dine as substrates. Forty nine different esters (43 new
compounds) were synthesized, isolated, and characterized
1
by MS and H NMR. While in some cases it was not clear
why no product was obtained, we were able to rapidly
synthesize a wide variety of valuable products and to
identify trends and reactivity patterns.
Ack n ow led gm en t. We thank Heinz Steiner for
discussion and technical help.
Exp er im en ta l Section
Su p p or tin g In for m a tion Ava ila ble: 1H NMR and MS
data for all new compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
Typ ica l Exp er im en t for 12 P a r a llel Rea ction s w ith ou t
Stir r in g in a 300-m L Au tocla ve. PdCl2(PhCN)2 (2.0 mg) and
4.0 mg of 1,3-bis(diphenylphosphino)propane (dppp) were placed
in a 1-mL glass vial. After the addition of 12 µL of 2-chloropy-
J O034112V
3728 J . Org. Chem., Vol. 68, No. 9, 2003