Two efficient methods for the preparation of 2-chloro-6-methylbenzoic acid

Article abstract of DOI:10.1021/op0102363

Two efficient methods for the preparation of 2-chloro-6-methylbenzoic acid were developed: one based on nucleophilic aromatic substitution and the other based on carbonylation. In the first approach, 2-chloro-6-fluorobenzaldehyde was converted to its n-butylimine, then treated with 2 equiv of methylmagnesium chloride in THF to give, after hydrolysis, 2-chloro-6-methylbenzaldehyde. Subsequent oxidation of this compound gave the title compound in 85% overall yield. In the second approach, 3-chloro-2-iodotoluene was efficiently carbonylated in methanol to give methyl 2-chloro-6-methylbenzoate, which after hydrolysis afforded the title compound in 94% yield (84% yield after recrystallization). The carbomethoxylation proceeded smoothly even at a high substrate-to-Pd ratio of 10 000. Both methods do not require isolation of intermediates and are suitable for the preparation of kilogram quantities of 2-chloro-6-methylbenzoic acid.

Full text of DOI:10.1021/op0102363

Organic Process Research & Development 2002, 6, 220−224
Two Efficient Methods for the Preparation of 2-Chloro-6-methylbenzoic Acid
Andrzej R. Daniewski,*,† Wen Liu,† Kurt Pu¨ntener,*,‡ and Michelangelo Scalone‡
Chemical Synthesis - Process Research, Non-Clinical DeVelopment, Pre-Clinical Research and DeVelopment,
Hoffmann-La Roche, Inc., Nutley, New Jersey 07110, U.S.A., and F. Hoffmann-La Roche Ltd., Pharmaceuticals DiVision,
Non-Clinical DeVelopment, Process Research and Catalysis, Grenzacherstrasse 124, CH-4070, Basel, Switzerland
Scheme 1. Sandmeyer reaction
Abstract:
Two efficient methods for the preparation of 2-chloro-6-
methylbenzoic acid were developed: one based on nucleophilic
aromatic substitution and the other based on carbonylation.
In the first approach, 2-chloro-6-fluorobenzaldehyde was con-
verted to its n-butylimine, then treated with 2 equiv of
methylmagnesium chloride in THF to give, after hydrolysis,
2-chloro-6-methylbenzaldehyde. Subsequent oxidation of this
compound gave the title compound in 85% overall yield. In
the second approach, 3-chloro-2-iodotoluene was efficiently
carbonylated in methanol to give methyl 2-chloro-6-methyl-
benzoate, which after hydrolysis afforded the title compound
in 94% yield (84% yield after recrystallization). The car-
bomethoxylation proceeded smoothly even at a high substrate-
to-Pd ratio of 10 000. Both methods do not require isolation of
intermediates and are suitable for the preparation of kilogram
quantities of 2-chloro-6-methylbenzoic acid.
Scheme 2. Ortho-metalation via r-amino-alkoxide
Scheme 3. Direct ortho-metalation
Introduction
amine,2 followed by oxidation of 6 to give 1 (Scheme 2),
and (3) direct ortho-metalation of 2-chlorobenzoic acid, 7
(Scheme 3).3
2-Chloro-6-methylbenzoic acid (1) is a rather simple
compound but is difficult to prepare in high purity on a large
scale. A recent program at Roche required multikilogram
The Sandmeyer reaction (Scheme 1) suffers from a low
yield (<50%), high dilution, and tar formationsthus, it is
not suitable for large-scale manufacturing. The second
approach (Scheme 2) requires more than 3 equiv each of
butyllithium and methyl iodide. It afforded 1 with an identical
impurity profile as the commercial material, which was
difficult to purify. The direct ortho-metalation of 7 (Scheme
3) requires a reaction temperature of -90 °C and was not
reproducible in our hands.
To produce high-quality final products in an economical
manner, we then envisioned two potential routes to 1, which
are based on nucleophilic aromatic substitution and carbo-
nylation, respectively.
quantities of this compound as a starting material to meet
the ever increasing supply requirement of our molecules for
toxicological evaluations. We had previously purchased 1
from commercial suppliers, but it was quite expensive
(>$1000/kg) and of insufficient quality (<99% pure). The
impurities contained in the commercial 1 (2-chlorobenzoic
acid, 2-chloro-dimethylbenzoic acid, 2-chloro-6-ethylbenzoic
acid, etc) were difficult to remove from 1 by recrystallization
and, unfortunately, were incorporated into the final product.
There are three known methods for the synthesis of 1,
which are based on: (1) the Sandmeyer reaction of 2-chloro-
6-methylaniline (2) to form the corresponding nitrile, 3,
followed by two more steps to give 1 (Scheme 1),1(2) ortho-
metalation of 2-chlorobenzaldehyde (4) via in situ R-amino-
alkoxide (5) formation with N,N,N′-trimethyl-ethylenedi-
Results and Discussion
Nucleophilic Aromatic Substitution Approach. This
route is based on chemistry described by Cahiez et al.,4 in
which they report a substitution reaction of the fluorine atom
in compound 8 by butylmagnesium bromide to give 9 in
90% yield after hydrolysis (Scheme 4). If a methyl Grignard
reagent would react with 8 as cleanly as butylmagnesium
(2) Comins, D. L.; Brown, J. D. J. Org. Chem. 1984, 49, 1078.
(3) Bennetau, B.; Mortier, J.; Moyroud, J.; Guesnet, J.-L. J. Chem. Soc., Perkin
Trans. 1 1995, 1265.
† Hoffmann-La Roche, Inc., Nutley, NJ.
‡ F. Hoffmann-La Roche, Ltd., Basel, Switzerland.
(1) Ha¨ring, M. HelV. Chim. Acta 1960, 43, 104.
(4) Cahiez, G.; Lepifre, F.; Ramiandrasoa, P. Synthesis 1999, 2138.
220
•
Vol. 6, No. 3, 2002 / Organic Process Research & Development
10.1021/op0102363 CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/22/2002

Scheme 4. Reported nucleophilic aromatic substitution
Scheme 7. Relative reactivities of 14 and 15
with BuMgBr
Scheme 5. Nucleophilic aromatic substitution approach
(although possible, as described later) to remove from 1. If
12 is not removed at this stage, it will be incorporated into
the final product. On the other hand, the formation of a small
amount (<0.2%) of the double-substitution product, 2,6-
dimethylbenzaldehyde (11), which gives 13 upon oxidation,
was of less significance to us. In the preparation of the final
drug substance, 1 is converted into acid chloride 14, then
coupled with an amino acid methyl ester, 16, to give key
intermediate 17 (see Scheme 7). When 1.03 equiv of 1 that
contained ca. 0.2% of 13 was utilized in this coupling
reaction (via the acid chloride), desired 17 was obtained in
high yield, and the dimethylbenzoyl impurity, 18, was
undetected. Indeed, when a 1:1 mixture of 14 and 15 (the
acid chloride of 13) was allowed to react with 1 equiv of
amino acid methyl ester 16 in the presence of sodium
hydroxide (Scheme 7), a 9:1 mixture of 17 and 18 was
formed. Clearly, dimethylbenzoyl chloride, 15, is far less
reactive than 14. Thus, taking advantage of this difference
in reactivity, the substitution reaction giving 6 was allowed
to go to completion, while formation of 11 was minimized
(<0.2%) by running the reaction at ca. 17 °C. At higher
temperatures, the formation of 11 increased.
Scheme 6. Impurities and their fate
bromide, the requisite aldehyde 6 would be obtained in a
good yield. More importantly, the starting material, 2-chloro-
6-fluorobenzaldehyde (10), is quite inexpensive (<$40/kg
for 200-kg quantities).
Oxidation of the aldehyde 6, obtained after acidic workup,
to afford the title compound, 1, was performed with sodium
chlorite-hydrogen peroxide (Scheme 5).5 The reaction was
carried out by the addition of 1.4 equiv of aqueous sodium
chlorite to an aqueous acetonitrile solution of aldehyde 6
and 1.4 equiv of 30% hydrogen peroxide, at 15-19 °C,
buffered with sodium dihydrogen phosphate. Although crude
acid 1 can be obtained directly from the reaction mixture by
removing the acetonitrile, the product thus obtained is colored
and contaminated with inorganic salts. Thus, to remove
inorganic materials, the product was first extracted into
toluene and then into aqueous sodium hydroxide to remove
nonacidic organic materials. Addition of concentrated hy-
drochloric acid to the basic phase precipitated the desired
acid. Product 1, obtained in this fashion, is quite pure,
typically containing less than 0.1 and 0.2% of 12 and 13,
respectively. The overall yield of 1 from 2-chloro-6-
fluorobenzaldehyde (10) is typically 85%.
On addition of n-butylamine (1.13 equiv) to a slurry of
10 in heptane, the solids dissolved rapidly and water formed
(Scheme 5). After washing the organic layer with brine, the
residual water in the organic layer was removed azeotropi-
cally with heptane. Crude 8, thus obtained as an oil, was
used directly in the next step. The starting material, 10, is
sensitive to air oxidation and is usually contaminated with
2-chloro-6-fluorobenzoic acid (12). Although 12 is removed
during the aqueous wash, monitoring by GC to ensure the
complete removal of this impurity is advisable.
The substitution reaction was carried out at ca. 17 °C for
4 h (Scheme 5). Two equivalents of methylmagnesium
chloride were required to achieve complete conversion (i.e.,
less than 0.05% of 10 after hydrolysis). The reaction was
then quenched with acetone at a lower temperature to
minimize over-reaction. Under these conditions, 6 of high
purity is consistently obtained after hydrolysisstypically
containing 0.01-0.05% of 10 and 0.03-0.15% of 11
(Scheme 6). Careful monitoring for complete conversion is
important since 10 can be oxidized in the subsequent step
to give 2-chloro-6-fluorobenzoic acid (12), which is difficult
During the early development work, when reliable
analytical methods had not been established, impure acid 1,
(5) Dalcanale, E.; Montanari, F. J. Org. Chem. 1986, 51, 567.
Vol. 6, No. 3, 2002 / Organic Process Research & Development
•
221

Scheme 8. Carbonylation approach
Table 1. Carbomethoxylation of 19a:a variation of ligands
conversion
(GC area %)
product 20
entry
catalystb
2 h 16 h (GC area %)
1
2
3
4
5
6
7
8
9
Pd(OAc)2[PPh(3,5-di-tBu-Ph)2]2 97 99.8
98
99
97
97
92
95
92
83
31
23
98
98
98
87
Pd(OAc)2[P(3,5-di-tBu-Ph)3]2
Pd(OAc)2[P(o-Tol)3]2
Pd(OAc)2(PPh3)2
96 99.9
88 99.6
87 99.4
48 99.9
87 99.9
64 99.8
52 99.8
containing 0.4% of 12, was produced. This necessitated the
development of an efficient purification method of this
compound. After several fruitless attempts, recrystallization
from dilute aqueous acetic acid, buffered with a small amount
of sodium acetate, was found to be quite effective to give 1
of a high purity with 81% recovery, containing only 0.07%
of the impurity, 12. Without addition of sodium acetate,
recrystallization was less effective. Presumably, sodium
acetate maintains 2-chloro-6-fluorobenzoic acid (12), the
most acidic component in the mixture, ionized at least
partially in the aqueous mixture, preventing crystallization
of this impurity. Once appropriate in-process control methods
had been established, 1 of a high quality has consistently
been produced without resorting to this purification method.
Carbonylation Approach. Benzoic acid derivatives are
generally accessible in good yield by carbonylation of aryl
halides, -triflates, or -diazonium salts.6 Surprisingly, however,
there are only few reported examples employing carbony-
lation for the preparation of 2,6-disubstituted benzoic acids.7
Carbonylation of 2,6-disubstituted substrates, such as 19a-d
(Scheme 8), may be impractical due to steric hindrance,
requiring long reaction times, high catalyst loadings, and
relatively high CO pressures. Moreover, the aryl chloride
function in 19, as well as in the product (e.g., 20), could
undergo carbonylation reaction to generate byproducts,
particularly, if more forcing conditions are applied.
Pd(OAc)2(PPh2Me)2
Pd(OAc)2(DIOP)c
Pd(OAc)2(BINAP)
Pd(OAc)2(dppf)d
Pd(OAc)2(dppp)e
6
5
33
24
10 Pd(OAc)2(dppb)f
11 PdCl2(PPh3)2
97 99.6
92 99.4
86 99.3
65 90
12 PdCl2[PPh(3,5-di-tBu-Ph)2]2
13 PdBr2(PPh3)2
14 PdCl2(NCMe)2/2 PPh3
a S/C 1000, 1.5 equiv of NaHCO3, 40 bar initial CO pressure, 140 °C, 5 wt
% of 19a in methanol. b Isolated or in situ prepared catalysts. c DIOP: O-iso-
propylidene-2,3-dihydroxy-1,4-bis(diphenylphosphino)butane. d dppf: bis(diphe-
nylphosphino)ferrocene. e dppp: bis(diphenyl-phosphino)propane. f dppb: bis-
(diphenylphosphino)butane.
First, the influence of phosphine ligands was investigated
using palladium acetate as the metal source (Table 1, entries
1-10). In the monodentate ligand series (entries 1-5),
catalysts containing bulky ligands (entries 1 and 2) were
found to be more effective than those containing sterically
less hindered phosphines; for example, triphenylphosphine
(entry 4). Catalysts containing bidentate ligands (entries
6-10) were generally less effective; only DIOP was
comparable with triphenylphosphine. Interestingly, dppp,
which was found to be 500 times more effective than
triphenylphosphine in the carbomethoxylation of aryl
triflates,7a,b was one of the least effective ligands examined
(entry 9).
Having identified triphenylphosphine analogues as the
most effective ligands for this reaction, the influence of
anionic ligands (i.e., OAc, Cl, and Br) was examined (Table
1, entries 11-14). PdCl2(PPh3)2 was found to be the most
active catalyst (entry 11), which was comparable with the
more elaborate catalyst, Pd(OAc)2[PPh(3,5-di-tBu-Ph)2]2
(entry 1). Interestingly, the corresponding chloride of the
latter was found to be less active (entry 12). PdBr2(PPh3)2
was similar to Pd(OAc)2(PPh3)2 (entries 4 and 13). Since
PdCl2(PPh3)2 is commercially available and most active of
those examined, this catalyst was chosen for further develop-
ment.
With PdCl2(PPh3)2 as the catalyst, other reaction param-
eters were then examined. Both the substrate, 19a, and the
product, 20, were found to be thermally stablesno major
decomposition products were detected when the reaction was
performed even at 160 °C for 16 h. Lowering CO pressure
from 40 to 10 bar resulted in a slight decrease in selectivity
(99 to 97% GC area%). Of the bases evaluated (NaHCO3,
NaOAc, and Et3N), sodium bicarbonate was found to be
superior to the others in terms of reaction rate. From these
and other experiments, a reaction temperature of 140 °C,
initial CO pressure of 40 bar, and use of 1.5 equiv of sodium
bicarbonate were chosen. Catalyst loading and substrate
concentration were then examined (Table 2). Increasing the
Four chlorotoluene derivatives, 19a-d, were evaluated
as starting materials for the methoxycarbonylation. Iodide
19a smoothly converted to the desired methyl ester 20 under
a variety of conditions. Using bromide 19b, the carbonylation
was quite sluggish even at a very high catalyst loading (S/C
) 10) and elevated temperature (160 °C). Triflate 19c only
gave a mixture of 2-chloro-6-methylphenol and 2-chloro-6-
methylanisole. Diazonium salt 19d gave 3-chlorotoluene as
a major product together with only 14% of 20. As 19a was
found to be a good substrate for methoxycarbonylation,
carbonylation in the presence of water to directly give 1 was
attempted. Unfortunately, under the conditions examined,
3-chlorotoluene was obtained as the major product (50-
80%), and the desired product, 1, was formed only as a minor
product (<30% yield). On the basis of these results,
methoxycarbonylation of iodide 19a was chosen for further
investigation.8
(6) Colquhoun, H. M.; Thompson, D. J.; Twigg, M. V. Carbonylation: Direct
Synthesis of Carbonyl Compounds; Plenum Press: New York, 1991.
(7) (a) Nicolau, K. C.; Ebata, T.; Stylianides, N. A.; Groneberg, R. D.; Carrol,
P. J. Angew. Chem. 1988, 100, 1138. (b) Dolle, R. E.; Schmidt, S. J.; Kruse,
L. I. J. Chem. Soc., Chem. Commun. 1987, 904. (c) Nagira, K.; Kikukawa,
K.; Wada, F.; Matsuda, T. J. Org. Chem. 1980, 45, 2365.
(8) Another plausible approach to prepare 1, metalation of 19a (Grignard or
lithiation) followed by quenching with CO2, was also examined but was
found to be unsuccessful, presumably due to concomitant benzyne formation.
222
•
Vol. 6, No. 3, 2002 / Organic Process Research & Development

Table 2. PdCl2(PPh3)2-catalyzed carbomethoxylation of 19a:a
variation of catalyst loading and substrate concentration
saturated sodium chloride solution (2 × 50 mL) and
concentrated (40 °C/8 mmHg) to give 117 g (overweight)
of crude 8 as a pale-yellow oil.4
conversion
catalyst
(GC area %)
To a solution of 8 (67.3 g, 315 mmol) in THF (100 mL)
at 5 °C was added a 3 M methylmagnesium chloride solution
in THF (210 mL, 630 mmol) over 30 min, while maintaining
the temperature of the reaction mixture at 5-15 °C. The
mixture was then stirred at room temperature (ca. 17 °C)
for 4 h to achieve complete reaction (i.e., less than 0.05%
of 10 after hydrolysis). After cooling to 5 °C with an ice-
water bath, the reaction was quenched by the addition of
acetone (30 mL, 409 mmol), while maintaining the temper-
ature of the mixture at 0-10 °C. Then, 3 N HCl (360 mL)
was added. An exotherm ensued that raised the temperature
of the mixture from 10 to 40 °C. The mixture was allowed
to cool to 25 °C over 1 h and then extracted with hexane (2
× 170 mL). The combined organic layers were washed with
1% sodium chloride solution (2 × 150 mL) and concentrated
to dryness (40 °C/14 mmHg) to give 49.1 g (overweight) of
crude 6 as a light-yellow, low-melting solid.
loading 19a concentration
product 20
entry
S/C
(wt %) in methanol 2 h
16 h (GC area %)
1
2
3
4
5
2000
5000
10000
10000
10000
5
5
93
86
80
91
99.5
99.4
98.7
99.2
98
98
98
98
98
5
10
20
94 >99.9
a 1.5 equiv of NaHCO3, 40 bar initial CO pressure, 140 °C.
substrate concentration (entries 3-5) was found to be
beneficial in terms of reaction rate. Even at a very low
catalyst loading (S/C 10 000) the reaction proceeded smoothly
with high selectivity (entry 5). In a preparative experiment
on a 100-g scale, with a substrate concentration of 20 wt %
and S/C 10 000, the reaction was complete within 8 h (see
Experimental Section). The resulting reaction mixture was
then directly subjected to hydrolysis with aqueous sodium
hydroxide at reflux to give, after acidic workup, 99.5% pure
1 in 94% yield. This material should be suitable for
subsequent transformations. If desired, 99.8% pure 1 with a
palladium content of less than 5 ppm can be obtained after
charcoal treatment and recrystallization from hexane.
GC retention times: 10, 4.50 min; 11, 4.61 min; 6, 4.96
min.
2-Chloro-6-methylbenzoic Acid (1).3 To a mixture of 6
(63.0 g, 408 mmol), sodium dihydrogen phosphate dihydrate
(16 g, 116 mmol), water (100 mL), 30% hydrogen peroxide
(64 mL, 564 mmol), and acetonitrile (190 mL) at 10 °C was
added a solution of 80% sodium chlorite (64 g, 566 mmol)
in water (190 mL) over 1 h, while maintaining the temper-
ature at 15-19 °C with an ice-water bath. After stirring at
room temperature for 1 h, TLC analysis of the organic layer
indicated complete reaction. After cooling with an ice-water
bath, sodium metabisulfite (54 g, 284 mmol) was added in
portions. An exotherm ensued that raised the temperature
of the mixture from 10 to 50 °C. After stirring for 0.5 h, the
mixture was concentrated (40 °C/14 mmHg) to remove most
of the acetonitrile and then acidified by the addition of
concentrated HCl (20 mL) and again concentrated. Toluene
(200 mL) was added to the resulting aqueous slurry, and
the aqueous layer was separated. The organic layer was then
extracted with a solution of sodium hydroxide (31 g, 775
mmol) in water (300 mL) and then with water (30 mL). The
combined aqueous layers were concentrated (40 °C/14
mmHg) to remove residual solvents. The resulting clear
solution was then acidified by the addition of concentrated
HCl (96 mL). After cooling to 10 °C, the resulting solid
was collected by filtration, washed with cold water (2 × 50
mL), and dried by suction overnight to give 58.1 g (85.5%
overall yield from 10) of 1 as a white solid; 99.7% pure by
GC analysis, containing 0.06% of 2-chloro-6-fluorobenzoic
acid (12) and 0.12% of 2,6-dimethylbenzoic acid (13). For
GC analysis, samples (1 mg of 1) were treated with a 1:2
mixture of bis(trimethylsilyl)trifluoroacetamide and pyridine
(350 µL) at 70 °C for 30 min.
Summary
We have developed two efficient methods for the prepa-
ration of 2-chloro-6-methylbenzoic acid (1). The first method
based on nucleophilic aromatic substitution gave the desired
acid in three chemical steps and 85% yield from 2-chloro-
6-fluorobenzaldehyde (10). The second method, which is
based on a catalytic carbonylation, gave 1 in two chemical
steps and 94% yield (84% yield after recrystallization) from
3-chloro-2-iodotoluene (19a). Both methods do not require
isolation of intermediates and are suitable for the preparation
of kilogram quantities of 2-chloro-6-methylbenzoic acid.
Experimental Section
General. Melting points are uncorrected. All reagents and
solvents were obtained from commercial suppliers and used
without further purification. The GC analysis data is reported
in area % and is not adjusted to weight %.
GC conditions.
column
Restek Rtx-35, 15 m × 0.32 mm,
0.25 µm thick coating
He, 65 kPa
250 °C, split ratio of 1:50
50-150 °C at 10 °C/min
300 °C, FID
carrier gas
injection
temperature
detection
Nucleophilic Aromatic Substitution Approach. 2-Chloro-
6-methylbenzaldehyde (6).9 To a slurry of 2-chloro-6-
fluorobenzaldehyde (10, 85.4 g, 539 mmol) in 255 mL of
heptane was added n-butylamine (60 mL, 607 mmol). After
stirring at room temperature for 1 h, the resulting water layer
was separated, and the organic layer was washed with
GC retention times: 13, 6.66 min; 12, 6.79 min; 1, 7.79
min.
RemoVal of 2-Chloro-6-fluorobenzoic Acid (12) from
Impure 1. Ten grams of 1 containing 0.4% of 12 was
dissolved in a mixture of acetic acid (14 mL) and water (100
(9) Bedard, T. C.; Corey, J. Y.; Lange, L. D.; Rath, N. P. J. Organomet. Chem.
1991, 401, 261.
Vol. 6, No. 3, 2002 / Organic Process Research & Development
•
223

mL) by heating to 84 °C. After cooling to 50 °C, sodium
acetate (50 mg) was added, and at 45 °C, seed crystals of 1
were added. The resulting suspension was slowly cooled to
4 °C. The solid was then collected by filtration, washed with
water, and dried to give 8.13 g of 1 (81% recovery),
containing 0.07% of 12.
2-Chloro-6-methylbenzoyl Chloride (14).1 To a mixture
of 1 (20.0 g, 117 mmol), DMF (50 µL), and toluene (70
mL) was added oxalyl chloride (16 mL, 180 mmol). The
resulting solution was stirred at room temperature for 1 h,
and then ca. 50 mL of toluene was removed by distillation
at atmospheric pressure to give a concentrated toluene
solution of 14, which was used in the subsequent acylation
reactions without purification. GC analysis indicated this
material to be 98.4% pure, containing three impurities,
<0.1% of 2-chloro-6-fluorobenzoyl chloride, ca. 0.1-0.2%
of 2,6-dimethylbenzoyl chloride (15), and ca. 1.4% of
2-chloro-6-methylbenzoic anhydride.
Carbonylation Approach. 2-Chloro-6-methylbenzoic Acid
(1).3 3-Chloro-2-iodotoluene (19a, 100 g, 396 mmol), sodium
bicarbonate (49.9 g, 594 mmol), bis(triphenylphosphine)-
palladium(II) chloride (27.8 mg, 0.0396 mmol), and methanol
(500 mL) were charged into an autoclave. After purging with
argon [three cycles of vacuum (0.2 bar)/argon (8 bar)], the
autoclave was pressurized with carbon monoxide (40 bar).
The mixture was heated to 140 °C with stirring (750 rpm)
for 8 h. After cooling, the autoclave was vented and purged
with argon. GC analysis confirmed complete conversion. The
reaction mixture containing methyl ester 20 was transferred
to a flask with the aid of methanol (500 mL), and a solution
of sodium hydroxide (47.53 g, 1.19 mol) in 400 mL of water
was added. After heating to reflux for 21 h, the mixture was
concentrated to a volume of ca. 600 mL. The aqueous
solution was washed with tert-butyl methyl ether (4 × 300
mL), diluted with water (300 mL), and cooled to 5 °C with
an ice-water bath. After dropwise addition of concentrated
HCl (105 mL, 1.24 mol) over 90 min, a white suspension
formed, which was extracted with dichloromethane (3 × 500
mL). The combined organic layers were dried over magne-
sium sulfate, concentrated to dryness, and further dried at
room temperature under vacuum (0.1 mbar) for 16 h to give
64.5 g (94% yield) of 1 as a pale yellow solid; 99.5% pure
by GC analysis. This material was further purified by
treatment with charcoal (3.14 g) in hexane at 50 °C for 1 h.
The charcoal was then removed by filtration, and the product,
1, was crystallized from the hexane by cooling to -23 °C
(not optimized) to give 57.0 g (84% overall yield) of 1 as
white crystals; mp 98-99 °C (lit. mp 100-102 °C),3 99.8%
pure by GC analysis.
Received for review November 7, 2001.
OP0102363
224
•
Vol. 6, No. 3, 2002 / Organic Process Research & Development