The preparation and isolation of chloramphenicol palmitate in toluene

Article abstract of DOI:10.1021/op990200z

The development of a process for the preparation of the antibiotic chloramphenicol palmitate that did not use methylene chloride and that fit into the production plant with improved yields and product quality is described. Addition of DMF as a co-solvent allowed the use of toluene for the reaction and the isolation. The product distribution observed by competitive esterification at the primary and secondary hydroxyl groups of chloramphenicol was well modeled by simple expressions for parallel consecutive reactions and estimated a 1000-fold difference in the rate constants for the two sites. A small increase in the impurity levels found during production plant trials was traced to trans-esterification during the extended processing time required at the 1000-kg scale.

Full text of DOI:10.1021/op990200z

Organic Process Research & Development 2000, 4, 301−304
The Preparation and Isolation of Chloramphenicol Palmitate in Toluene
Edward D. Daugs
Contract Manufacturing SerVices, The Dow Chemical Company, 1710 Building, Midland, Michigan 48674, U.S.A.
Table 1. End of reaction analysis from methylene chloride,
xylene, and toluene
Abstract:
The development of a process for the preparation of the
antibiotic chloramphenicol palmitate that did not use methylene
chloride and that fit into the production plant with improved
yields and product quality is described. Addition of DMF as a
co-solvent allowed the use of toluene for the reaction and the
isolation. The product distribution observed by competitive
esterification at the primary and secondary hydroxyl groups
of chloramphenicol was well modeled by simple expressions for
parallel consecutive reactions and estimated a 1000-fold dif-
ference in the rate constants for the two sites. A small increase
in the impurity levels found during production plant trials was
traced to trans-esterification during the extended processing
time required at the 1000-kg scale.
mol %
yield
(%)
solvent
L/kg
1
2
3
4
methylene
3
2.70
93.36
0.17
3.31
85
chloridea range 1.2-4.7 91.4-94.5 0.12-0.25 3.2-4.2
13% DMF
in xylene 4.5
13% DMF
in toluene 4.5
3
2.4
2.3
1.3
2.0
91.8
94.0
93.7
94.1
0.13
0.17
0.15
0.30
5.7
3.5
4.8
3.7
87
90
89
90
3
a Average of seven production batches.
Table 2. Crystallization from xylene and toluene
mol % isolated
mother liquor
1
2
3
4
4:2
Introduction
Chloramphenicol palmitate (2) is an inert derivative that
is hydrolyzed intestinally to the antibiotic chloramphenicol
(1), and is used to avoid the bitter taste of chloramphenicol.1,2
Selective esterification at the primary hydroxyl group over
the secondary hydroxyl group of 1 is critical for reaction
yield and product quality. The general method of preparation
reported uses palmitoyl chloride and pyridine in carbon
tetrachloride or dichloroethane.3 Production of 2 at the
Aventis pharmaceutical production facility in Garessio, Italy,
used palmitoyl chloride with pyridine in methylene chloride.
A process that avoided the use of chlorinated solvents was
desired for industrial hygiene concerns. Any new process
had to meet or exceed the yield, product quality, and
throughput values demonstrated for the production process.
crude
from xylene
from toluene
0.62
0.36
0.34
85.4
98.2
98.6
0.55
0.15
0.09
13.5
1.3
0.96
0.772:1
0.798:1
further with water before cooling for crystallization and
centrifugation. Although it eliminated the solvent exchange,
the use of xylene or toluene as the reaction solvent was
limited by the low solubility of 1 (0.145 g/L of toluene at
28 °C2).
Results and Discussion
Addition of at least 0.4 g of DMF/g of 1 as a co-solvent
increased the solubility sufficiently to give complete solution
in toluene or xylene at the -10 °C reaction temperature
required for the desired reaction selectivity, allowing the use
of either of these solvents for the esterification. Matching
the production batch size and product crystallization condi-
tions established a maximum solvent loading of 4.5 L/kg of
1. With the 1.15 equiv of pyridine used as the base, a
concentration of 12 to 13 wt % DMF in xylene or toluene
gave complete solution of 1. Minimizing the DMF loading
was important economically, as it would be removed in the
water washes and difficult to recover.
In the production process, chloramphenicol palmitate was
isolated by a series of water washes of the methylene chloride
solution, and then the solvent was removed by distillation
and replaced with xylene. The xylene solution was washed
Both xylene and toluene gave reaction results comparable
to those of the methylene chloride process (Table 1). They
also gave similar crystallization results (Table 2). Product
precipitation near the end of the palmitoyl chloride addition
produced a reaction mixture that, depending on the DMF
loading and reaction temperature, was often difficult to stir
on a laboratory scale. Reaction mixture slurries following
the palmitoyl chloride addition in toluene were thinner and
easier to stir at both reaction concentrations of Table 1, and
(1) Nagabhushan, T.; Miler, G. H.; Varma, K. J. In Kirk-Othmer Encyclopedia
of Chemical Technology, 4th ed.; John Wiley & Sons: New York, 1992;
pp 961-978.
(2) The Merck Index, 8th ed.; Merck; Rahway, NJ, 1968; p 233.
(3) El-Shafie, S. M. M. Chem. Age India 1979, 30, 737. El-Shafie, S. M. M.
Proc. Egypt. Acad. Sci. 1977, 30, 87-8.
10.1021/op990200z CCC: $19.00 © 2000 American Chemical Society and The Royal Society of Chemistry
Published on Web 05/23/2000
Vol. 4, No. 4, 2000 / Organic Process Research & Development
•
301

Figure 1. Comparison of model conversion (lines) with experimental data (points).
thus toluene was chosen as the solvent for further develop-
ment. An initial reaction temperature of approximately -10
°C, near the limit of the production equipment, minimized
the addition time for the exothermic acid chloride addition.
A final reaction temperature of >5 °C was necessary to
ensure agitation.
A simple model can be used to evaluate the product
distribution for consecutive reactions at two competitive sites.
The esterification reactions may be expressed as:
By comparing the product distribution over the course of
the palmitoyl chloride addition with the product profiles
calculated with various values of x, the rate of reaction at
the primary hydroxyl group was estimated to be 1000 times
faster than that of the reaction at the secondary hydroxyl
group (Figure 1) in either methylene chloride or toluene. This
gives a maximum calculated reaction yield of 94 mol %.
With x ) 100 and 500, the maximum calculated yields were
81.9 and 91.4%, respectively. Not shown in the figure is the
level of 3. The model predicts that it reaches a steady state
of 0.02 to 0.03 mol % before decreasing at the end of the
reaction; experimentally, a maximum of 0.5 mol % was
formed, decreasing to 0.2 mol % at the end of the reaction.
The level of 1 remaining at the end of the palmitoyl
chloride addition was slightly overestimated by the simple
model in toluene solvent. The model better matched the level
found using methylene chloride as solvent. The difference
may be due to precipitation of 2 at the end of the palmitoyl
chloride addition in toluene, which lowers the relative
concentration available for the formation of 4 and increases
the relative concentration of 1 for the formation of 2.
Work-up involved a number of water washes to remove
DMF, pyridine hydrochloride, and residual 1 before crystal-
lization. The pH of the wash water was adjusted to 3-3.5
to remove pyridine. GC analysis found that the DMF and
pyridine were effectively removed by the third water wash.
As the DMF was removed from the toluene solution, the
solubility of 2 dropped, such that increasing the wash
temperature to 60 °C was required to maintain a solution.5
Water washes were continued in order to reach <0.25 mol
k1
1 + pyridine + ClCOR
1 + pyridine + ClCOR
2 + pyridine + ClCOR
3 + pyridine + ClCOR
98 2 + pyridine‚HCl
k2
98 3 + pyridine‚HCl
k3
98 4 + pyridine‚HCl
k4
98 4 + pyridine‚HCl
For parallel reactions, the ratio of product concentrations
does not depend on time, but only on the corresponding rate
coefficients.4 Assuming that (a) k1 ) k4 and k2 ) k3, (b) all
species remain in solution, and (c) the reactions are irrevers-
ible and no other side reactions occur, and setting k1 ) xk2,
the rate expressions may be solved by removing the time
dependence on the species concentrations. Simplifying, the
relative concentrations of each species in mole fraction can
be calculated as the concentration of 1 decreases by:
[x/(x + 1)](100 - [1])
[2] )
1 + [1/(x + 1)][(100 - [1])/[1]]
and
[1/(x + 1)](100 - [1])
[3] )
(4) Szabo, Z. G. In ComprehensiVe Chemical Kinetics; Bamford, C. H., Tipper,
C. F. H., Eds.; Elsevier: New York, 1969; Vol. 2, pp 14-24.
(5) The solubility of 2 in toluene at -5, 21, 35, 46, and 53 °C was 0.3, 1.0,
3.4, 6.4, and 23 wt %, respectively.
1 + [x/(x + 1)][(100 - [1])/[1]]
for various values of x.
302
•
Vol. 4, No. 4, 2000 / Organic Process Research & Development

Table 3. Isolated yield and purity of 2
wt %
yield
(%)
process
form
1
3
4
others (%)
methylene chloride
crystalline
amorphous
0.1-0.3%
0.3-1.0%
0.4-1.0%
1.2-2.0%
1.0-1.7%
85
99
90-110 ppm
0.10-0.18
toluene
laba
crystalline
amorphous
crystalline
amorphous
specification
<0.1%
<0.05%
0.23-0.52%
0.24-0.59%
0.70-1.17%
0.65-1.36%
92-93
99
82-92
99
190-245 ppm
505-715 ppm
100-135 ppm
<450 ppm
0.01-0.03%
0.10-0.13%
0.17-0.35%
<2%
planta
0.07-0.08
<0.5
<2%
a Average of three batches.
% of 1, since it was poorly rejected by crystallization of 2
(see Table 2). Following crystallization, the isolated solid
was heated in water at a pH of 4-5 to a melt to convert it
to an amorphous final product (non-polymorph A6). Levels
of 1 decreased slightly during this transformation.
Table 4. Molar balance of palmitates
moles
2
3
4
total
produced
isolated
mother liquor
∆
2956
2817
81.9
-57
3.4
5.4
13.2
115.4
25.0
113.8
+23.4
3075
2847
208.9
-19
The toluene process was demonstrated at the 6000-L scale
in the equipment used for the production of 2. Reaction
selectivity was as observed in the laboratory. As the reactor
loadings were equivalent to the production process, the
palmitoyl chloride addition time required to maintain the
reaction temperature below 5 °C was also equivalent. Six
water washes were necessary to remove the residual 1, a
decrease from the original process. Elimination of the
methylene chloride solvent exchange also decreased the total
processing time, increasing the throughput for the plant.
Isolated yield and purity analysis results for the crystalline
and amorphous forms of 2 are listed in Table 3. The isolated
yield over three 700 to 1000 kg batches averaged 87%, with
a high of 92%. Yield and product quality was improved over
the production process, but was lower than that obtained in
the laboratory. Comparing the laboratory- and plant-scale
results, the isolated yields were higher and the impurity levels
were lower in the laboratory, although the reaction selectivi-
ties, as indicated by analysis at the end of the acid chloride
addition, were similar. This was due to trans-esterification,
especially during the methylene chloride solvent exchange,
which increased the levels of 3 and 4, and decreased the
isolated yield from the process. During the work-up and
solvent exchange of a methylene chloride reaction mixture
in the laboratory, the level of 3 increased from 0.19 to 0.40
mol %, and the level of 4 increased from 6.46 to 7.70 mol
%, decreasing the level of 2 from 92.1 to 91.3 mol %. Some
degree of trans-esterification also occurred during the
extended processing time in the plant-scale trial of the toluene
process, as seen by the molar balance calculations of Table
4. No new impurities were found in the isolated product from
the new process. All of the unidentified minor impurities
detected by HPLC analysis were present in the methylene
chloride production product.
+15.2
increases in isolated yield, product quality, and throughput
over the process in methylene chloride. Reaction selectivity
was well modeled by simple expressions derived from the
rate equations for parallel consecutive reactions, which
predicted the product distribution using a 1000 to 1 difference
in the rate constants for reaction at the primary and secondary
hydroxyl groups of 1. Small increases in the impurity levels
found upon scale-up to 1000 kg were due to trans-
esterification during processing but were less than those
observed during processing using methylene chloride as the
solvent.
Experimental Section
Raw materials were obtained from the Garessio Chloram-
phenicol production plant. HPLC analysis was carried out
at λ ) 254 nm by injecting a 20 µL sample dissolved in
mobile phase onto a 4.6 × 250 mm ZORBAX SIL column
and eluting at 2 mL/min with 4.8 vol % 2-propanol in hexane.
Molar response factors were determined using standard
samples. Final product analyses were carried out by the
Garessio Quality Control Laboratory using established
methods.
Preparation of 2. A 1-L flask was charged with 100 g
(0.3905 mol) of chloramphenicol (1) and 400 mL of toluene,
and 50 mL of solvent was removed by vacuum distillation
to remove residual water. The slurry was chilled to -4 °C,
and 50 mL of toluene, 50 g of DMF, and 28.1 g (0.335 mol)
of pyridine were added. Palmitoyl chloride (90.2 g, 0.328
mol) was added dropwise over 140 min. The final reaction
temperature was 5 °C. After 2.5 h, the slurry was warmed
to 25 °C, and a few drops of 30% hydrogen peroxide in
water were added to remove the pink color. HPLC analysis
found 93.7 mol % of 2, 4.1 mol % of 4, 0.22 mol % of 3,
and 2.0 mol % of 1. The solution was washed with 200 mL
portions of water (1 × 25 °C, 1 × 45 °C, 5 × 60 °C) at a
pH of 3.5 (hydrochloric acid) to give 0.25 mol % of 1
remaining. The solution was cooled to 0 °C, and the solid
Conclusions
The process for the preparation of 2 in toluene and DMF
was developed and demonstrated at the production scale, with
(6) Mitra, A. K.; Ghosh, L. K.; Gupta, B. K. Drug DeV. Ind. Pharm. 1993, 19,
971-80.
Vol. 4, No. 4, 2000 / Organic Process Research & Development
•
303

was isolated, washed with 100 mL of toluene, and dried at
60 °C under vacuum to give 159.7 g (0.2884 mol, 91.9%
yield) of chloramphenicol palmitate (2; 0.38 mol % of 4,
<0.02 mol % of 3, and <0.1 mol % of 1).
Amorphization of 2. A 500-mL flask was charged with
30.41 g of 2 and 272 mL of water at a pH of 4 (hydrochloric
acid). The slurry was heated to 95 °C to give an oil phase,
then rapidly cooled to 30 °C to give a precipitate. The solid
was isolated, washed with water, and dried at 60 °C under
vacuum to give 30.18 g (99.2% recovery) of amorphous 2.
Acknowledgment
The assistance and advice of S. Carrara, A. Coco, and P.
Cifani of the Garessio Pharmaceutical Production Site, P.
Camelia and A. Roberi of the Garessio Chloramphenicol
Plant, and C. Becker and D. Ferrero of the Garessio Quality
Control Laboratory are gratefully acknowledged.
Received for review November 17, 1999.
OP990200Z
304
•
Vol. 4, No. 4, 2000 / Organic Process Research & Development