Purification of 2,4 dichlorobenzoic acid

Article abstract of DOI:10.1021/op7001547

A practical and efficient method to purify 2,4-dichlorobenzoic acid is described. The formation of an α-methylbenzylamine salt reduces the levels of positional isomer impurities to <0.05%. Although this purification method is not universal for all substituted benzoic acids, it was shown to be applicable to several other benzoic acids.

Full text of DOI:10.1021/op7001547

Organic Process Research & Development 2008, 12, 120–124
Technical Notes
Purification of 2,4 Dichlorobenzoic Acid
Christopher P. Ley and Matthew H. Yates*
Chemical Product Research and DeVelopment, Eli Lilly and Company, Lilly Corporate Center, Indianapolis,
Indiana 46285-4813, U.S.A.
Table 1. ICH guidelines for impurities in new drug
substances
Abstract:
A practical and efficient method to purify 2,4-dichlorobenzoic acid
is described. The formation of an r-methylbenzylamine salt
reduces the levels of positional isomer impurities to <0.05%.
Although this purification method is not universal for all substi-
tuted benzoic acids, it was shown to be applicable to several other
benzoic acids.
maximum
daily
reporting
identification
thresholdc
qualification
thresholdc
dosea
thresholdb,c
e2 g/day
0.05%
0.10% or 1.0 mg 0.15% or 1.0 mg
per day intake
(whichever
is lower)
per day
(whichever
is lower)
0.05%
>2 g/day
0.03%
0.05%
a The amount of drug substance administered per day. b Higher reporting
thresholds should be scientifically justified. c Lower thresholds can be appropriate
if the impurity is unusually toxic.
Introduction
One of the basic goals of process development is to
understand the impurity profile of the drug substance, including
the origin of the impurities, and then demonstrate their control
throughout the process. The ICH guidelines1 for impurities are
described in Table 1. These values have been established to
limit exposure of unqualified impurities to <1 mg/day.
The lower limits, applicable for a high dose compound,
necessitate a very tight control of both the starting materials
and the process. Aromatic, positional isomers are often espe-
cially challenging to control to very low levels. Although some
impurities may be rejected during chemical processing, it is
often more prudent to limit their introduction in the starting
material. This is especially true if the synthesis is short and/or
the starting material is introduced late in the synthetic route.
During the development of a drug candidate we have addressed
the need for isomerically pure 2,4-dichlorobenzoic acid (DCBA).
Because DCBA is a commodity chemical, there is no shortage
of suppliers worldwide and the pricing is competitive. Unfor-
tunately, the major market for DCBA is not pharmaceutical and
most vendors do not test for or control the levels of positional
isomers. DCBA is used extensively in the polymer industry as
a chain modifier for polyamides and organosilicone polymers;
it is also further functionalized in other products as depicted in
Figure 1.
The likely synthetic route for DCBA is outlined in Scheme
1.2,3 The positional isomer impurities are formed during the
dichlorination of toluene. The purity can be controlled at this
stage by fraction distillation or zeolite adsoption.4 However, it
is unclear if the existing technology can achieve the desired
level of purity, which is no impurity >0.05%. Furthermore,
the globalization of commodity chemicals makes it unlikely that
the same sources of raw materials will be used or even available
for future material needs.
Since the quantity of DCBA needed for the pharmaceutical
market is low, we had minimal success influencing vendors to
produce material of exceptionally high quality. Qualification
of the potential impurities is an option, but we prefer to develop
a control strategy that limits the introduction of impurities and
minimizes patient exposure. To this end we sought to develop
a process that could purify commercial grade DCBA of all
positional isomers to <0.05%. Not surprisingly, a screen of
potential recrystallization solvents did not offer any purification.
(2) These are a few patents describing the chlorination of toluene: Dewkar,
G. K.; Thakur, V. Vi.; Pardhy, S. A.; Sudalai, A.; Devotta, S. Catalytic
process for regiospecific chlorination of alkanes, alkenes and arenes.
U.S. Patent 6,825,383, 2004. Catalytic production of 2,4-dichlorotoluene.
Gelfand, S. U.S. Patent 4,006,195, 1977. DiBella, E. P. Production of
2,4-dichlorotoluene. U.S. Patent 3,366,698, 1968.
(3) (a) Nomura, K.; Myawaki, T. Preparation of aromatic carboxylic acids
by using palladium-phosphine-type catalysts. JP 08104661 A 19960423,
1996; Chem. Abstr. 1996, 125, 86311. (b) Kajisori, S.; Kakinami, T.
Preparation of benzoic acid derivatives. JP 02268123 A 19901101, 1990;
Chem. Abstr. 1990, 114, 163745.
Thorough testing of multiple lots of DCBA from the multiple
vendors revealed significant variation in which isomers were
present, and their levels were as seen in Table 2.
(4) (a) Iwayama, K.; Yamakawa, S.; Kitano, Y.; Kitagawa, S. Method for
separation of dichlorotoluene isomer mixture. JP 2000247913 A
20000912, 2000; Chem. Abstr 2000, 133, 239705. (b) Carra, S.;
Paludetto, R.; Storti, G.; Morbidelli, M.; Gurtner, B.; Commandeur, R.
Separation of dichlorotoluene isomers by zeolites. Ger. Offen. DE
3637727. 1987; Chem. Abstr. 1987, 107, 77415
* To whom correspondence should be addressed. Telephone: (317)-651-7729.
E-mail: yatesmh@lilly.com.
(1) Impurities in New Drug Substances Q3A(R2): 2006, ICH Harmonized
Tripartite Guideline.
120
•
Vol. 12, No. 1, 2008 / Organic Process Research & Development
10.1021/op7001547 CCC: $40.75
2008 American Chemical Society
Published on Web 11/15/2007

Figure 1. Use of DCBA in commercial products.
Table 2. Vendor/isomeric impurity comparisons of DCBA
chlorobenzoic acida
dichlorobenzoic acidb
2,6
vendor
lot
year
2
4
2,5
3,4
all other
total impurities
I
A
C
D
E
2002
2002
2002
2002
2002
2005
2006
2006
2006
2006
2006
0.07
2.02
0.06
2.10
0.04
0.03
0.06
0.06
2.09
0.09
2.19
0.04
0.03
0.20
0.26
1.02
0.06
0.22
0.05
I
0.03
0.01
I
0.08
I
I
F
I
G
H
0.08
0.07
0.13
0.06
0.10
0.30
0.06
0.07
0.03
I
0.01
0.52
0.02
II
III
IV
V
0.07
0.1
0.02
0.05
a 3-Chlorobenzoic acid was not present in any samples. b 2,3-Dichlorobenzoic acid and 3,5-dichlorobenzoic acid were not present in any samples.
Scheme 1. A commercial synthesis of DCBA
Table 3. Salt screen of DCBA
A series of experiments were designed to test the rejection
efficiency of the salt for all potential isomers. A solution of
DCBA in isopropanol was spiked with positional isomers to
an approximate level of 0.1%, 0.2%, 0.5%, 0.75%, and 1.0%
of each isomer present in the solution. The mixture was heated
to ensure complete dissolution, and 0.9 equiv of (() R-meth-
ylbenzylamine was added at once. Upon cooling to room
temperature the salt crystallized and formed a thick slurry. As
depicted in Figure 3 below, there was significant purification
across the board, and in most cases the desired limit of <0.05
was met during the initial salt formation.
The salts were isolated and dried to obtain yield information
that was consistently ∼70% when corrected for purity. Condi-
tions were not optimized for this screen, and the yield is likely
to improve with optimization. Reforming the free acid with HCl
in methanol/water provided DCBA in ∼80% yield. Additional
purification was achieved during this step, and the resulting
levels are represented in Figure 4 below.
As demonstrated above, both reprocessing steps afford
purification, and this method can be used to reprocess material
with up to 0.5% of all isomers present. Of course the
commercial supplies do not have all isomers present. As seen
in Table 2, there are usually only a couple present at ∼0.1–0.2%,
and they vary from lot to lot. The purification outlined in
Scheme 2 was demonstrated on commercial bulk material and
shown to be effective as demonstrated in the chromatogram
(Figure 5).
We were interested to see if this purification was specific
for the 2,4 isomer or generally applicable to other chlorobenzoic
acids as well. Because all of the isomers are readily available,
we tested all possible combinations (dichloro 2,6; 2,3; 2,5; 3,4;
3,5; as well as the monosubstituted 2; 3; 4). A series of spiking
experiments at 0.2–0.3% were carried out for each of the above
isomers. Whereas the process rejected all isomers to less than
reduction in
isomers
crystalline salt
sodium
none
potassium
none
ammonium
ethylenediamine
l-phenyl piperazine
(()-R-methylbenzylamine
(S)-R-methylbenzylamine
tert-butylbenzylamine
marginal
most <0.05
most <0.05
all <0.05
all <0.05
all <0.05
A salt screen of common bases identified several crystalline
salts.5 As seen in Table 3, only a few salts provided purification.
Because the metal salts did not offer purification, it was
desirable to have a counterion with a chromophore. This will
simplify its detection and control in the reprocessed material.
DSC analysis of the potential salts, Figure 2, indicated that the
R-methylbenzylamine salt(s) should be the most stable and
would potentially provide the highest level of purification.6
Because our initial studies showed no advantage in using the
enantiomerically pure amine, the racemate was preferred for
price.7
(5) (a) Brzyska, W.; Swita, E. Pol. J. Chem. 1993, 67, 609. (b) Duff, J. G.;
Yung, D. K.; Brenner, R. J.; Wilson, B. J.; Racz, W. J. J. Chem. Educ.
1969, 46, 388. (c) Levitin, E. Y.; Oridoroga, V. O. Pharm. Chem. J.
2003, 37, 653.
(6) Lattice energy is a good but not absolute correlation to impurity
rejection.
(7) The precise nature of the salt, as a racemate or conglomerate, is
unknown.
Vol. 12, No. 1, 2008 / Organic Process Research & Development
•
121

Figure 2. DSC comparison of DCBA amine salts.
Figure 3. Isomer rejection during r-methylbenzylamine·DCBA salt formation.
Figure 4. Isomer rejection during breaking of r-methylbenzylamine·DCBA salt.
0.05% for the 2,5 isomer, all of the others had at least one
impurity that remained g0.05% as depicted in Figure 6 below.8
This procedure may still be useful to purify other chloro isomers
depending on the nature of the impurities present.
122
•
Vol. 12, No. 1, 2008 / Organic Process Research & Development

Figure 5. HPLC overlay of purification process.
Figure 6. Purification of other chlorobenzoic acids.
Beyond the chloro-substituted benzoic acids, we were
interested in the ability of this salt to purify other
substituted benzoic acids. A similar screen of mono- and
disubstituted, methyl and methoxybenzoic acids was
Vol. 12, No. 1, 2008 / Organic Process Research & Development
•
123

Scheme 2. Purification of commercial DCBA
conducted. An analogous process was effective at purify-
ing 2,4-dimethoxybenzoic acid to <0.05%, as well as m-
and p-toluic acids. Information on these screens can be
found in the Supporting Information. Unfortunately, the
high level of purification was not more general across
these series, but again this salt may prove useful depending
on which impurities are present.
(s, 1H), 7.38 (qt, J ) 10 Hz, 3H), 7.32 (t, J ) 7 Hz, 1H), 7.28
(dd, J ) 2, 6 Hz, 1H), 4.33 (qt J ) 6.5 Hz, 1H), 1.49 (d, J )
7 Hz, 3H); 13C NMR (125 MHz, DMSO-d6) δ 168.9, 141.0,
140.8, 132.0, 131.0, 130.6, 129.0, 128.4, 127.1, 127.0, 50.3,
21.7; mp 185.2 °C (DSC).
2,4-Dichlorobenzoic Acid from the (()-r-Methyl Ben-
zylamine Salt. 2,4-Dichlorobenzoate-(()-R-methyl benzyl-
amine (5.0 g, 16.0 mmol) was dissolved in water (50 mL) and
methanol (20 mL). The mixture was heated to 60 °C, and
concentrated HCl was added until the pH was <2.0. The thick
slurry was allowed to air-cool to room temperature. Water (12
mL) was added, the product was isolated by filtration, and the
wetcake was washed with water (30 mL). The product was dried
in vacuo at 40 °C overnight. Yield 94%. 1H NMR (400 MHz,
DMSO-d6) δ 7.81 (d, J ) 8.6 Hz, 1H), 7.71 (d, J ) 2 Hz, 1H),
7.51 (dd, J ) 2, 6 Hz, 1H); 13C NMR (125 MHz, DMSO-d6)
δ 166.2, 136.9, 133.5, 132.8, 130.6, 130.5, 127.9; mp 162.0
°C (DSC).
Conclusion
An efficient purification process has been developed that can
produce ultrapure DCBA through salt formation with (()-R-
methyl benzylamine. This process is operationally simple and
uses inexpensive starting materials and benign solvents. This
salt formation also purifies many other mono- and disubstituted
benzoic acids. This high purity starting material translates into
drug substance of consistently high quality without variation
in the impurity profile.
Experimental Section
(()-r-Methyl Benzylamine Salt of 2,4-Dichlorobenzoic
Acid. 2,4-Dichlorobenzoic acid (10.0 g, 50.2 mmol) was
dissolved in isopropyl alcohol (200 mL). The mixture was
heated to 60 °C and (()-R-methyl benzylamine (5.49 g, 45.3
mmol) was added all at once. The reaction was stirred at 60
°C for 1 h before the slurry was allowed to air-cool to room
temperature. The product was isolated by filtration, and the
wetcake was washed with isopropyl alcohol (25 mL). The
product was dried in vacuo at 40 °C overnight. Yield 79%. 1H
NMR (400 MHz, DMSO-d6) δ 7.49 (d, J ) 7 Hz, 2H), 7.41
Acknowledgment
The authors thank Anne Warner and Glenn McCain for
analysis of samples.
Supporting Information Available
Graphs depicting the rejection efficiency of (()-R-methyl
benzylamine salts of various substituted benzoic acids are
included. This information is availabkboprdle free of charge
via the Internet at http://pubs.acs.org.
Received for review July 5, 2007.
OP7001547
(8) Similar rejection graphs are available in Supporting Information for
the rest of the isomers.
124
•
Vol. 12, No. 1, 2008 / Organic Process Research & Development