Reaction safety: A critical parameter in the development of a scaleable synthesis of 2,3-bis-chloromethylpyridine hydrochloride

Article abstract of DOI:10.1021/op025545n

This report describes the development of a scaleable process to prepare 2,3-bis(chloromethyl)pyridine hydrochloride, 2, that is safe, economical, and environmentally acceptable. On a small scale this material is prepared using thionyl chloride as the solv

Full text of DOI:10.1021/op025545n

Organic Process Research & Development 2002, 6, 938−942
Reaction Safety: A Critical Parameter in the Development of a Scaleable
Synthesis of 2,3-Bis-chloromethylpyridine Hydrochloride
J. S. Grimm, C. A. Maryanoff, M. Patel, D. C. Palmer,* K. L. Sorgi, S. Stefanick,* R. R. H. Webster, and X. Zhang
Drug EValuation, Johnson & Johnson Pharmaceutical Research & DeVelopment, L.L.C., 1000 U.S. Route 202,
Raritan, New Jersey 08869, U.S.A.
Abstract:
amounts of excess corrosive and environmentally harmful
reagents or byproducts such as phosphorus oxychloride.
Several alternatives using a slight excess of thionyl chloride
in a reaction solvent were evaluated using reaction calorim-
etry to identify a safe, scaleable alternative to these reaction
conditions.
This report describes the development of a scaleable process
to prepare 2,3-bis(chloromethyl)pyridine hydrochloride, 2, that
is safe, economical, and environmentally acceptable. On a small
scale this material is prepared using thionyl chloride as the
solvent. However, this reaction is not readily scaleable because
of environmental and disposal issues. Reaction calorimetry is
used to evaluate a number of potentially viable chlorination
procedures. In the course of this investigation potentially
hazardous reagent/solvent incompatibilities were identified. The
use of thionyl chloride in MTBE is particularly hazardous owing
to solvent decomposition in the presence of HCl and subsequent
outgassing of large amounts of isobutylene. Reaction conditions
employing thionyl chloride and 1% v/v DMF in toluene were
found to be safe for scale-up. As a result, an environmentally
acceptable, safe process has been developed and scaled to
produce multikilogram amounts of material.
Results and Discussion:
We repeated the literature procedure and treated 1 with
3.0 mol equiv of thionyl chloride as the reactant and solvent.
As the reaction progressed to completion, the mixture became
a thick, difficult-to-stir suspension from which isolation of
2 by filtration was clearly problematic. Addition of precooled
tert-butyl methyl ether (MTBE) to the thick reaction suspen-
sion afforded an easily filtered mixture from which 2 was
isolated in excellent yield. However, the difficulties encoun-
tered on small scale together with our concerns for safety
and disposal of large volumes of excess thionyl chloride
prompted us to investigate alternate approaches. Our goals
were to prepare 2 in high yield and purity while minimizing
the excess of thionyl chloride required and maintaining a
convenient reaction time together with a simple and readily
scaleable workup procedure.
Among the solvents we evaluated initially, most of our
requirements could be met using MTBE as a diluent in the
early research-scale work described above. This solvent was
particularly advantageous for solubility and handling char-
acteristics (e.g., filtration) during the reaction workup. In
addition, we anticipated that the reaction could be run at
ambient conditions. This would avoid the known cleavage
of MTBE under acidic reaction conditions and elevated tem-
perature. For example, a violent acid catalyzed decomposition
of MTBE at elevated temperatures was reported wherein a
fatal explosion and fire resulted when sulfuric acid accident-
ally entered a distillation vessel during removal of MTBE.12
Reaction of 1 with 1.6 equiv of thionyl chloride in MTBE
at ambient temperature afforded 2 in excellent yield and
purity. Product isolation was simple and straightforward,
requiring only filtration and drying to afford 2 as a free
flowing, granular, white solid. These results were confirmed
and reproducible up to 2.0 mol scale. However, at this scale
the reaction became very sluggish and required several hours
Introduction:
Recently, we required a safe, scaleable, and environmen-
tally acceptable synthesis of 2,3-bis-(chloromethyl)pyridine
hydrochloride, 2, to support an ongoing project. Traditionally,
chlorination of 2- or 3-hydroxymethylpyridines has been
accomplished using thionyl chloride,1-4 phosphorus trichlo-
ride,5 phosphorus pentachloride,6 or phosphorus oxychlo-
ride.7,8 Indeed, chlorination of 2,3-bis(hydroxymethyl)-
pyridine, 1, using thionyl chloride as the reactant and solvent
is reported to afford 2 in excellent yield.9-11
However, these classical reaction conditions can be drastic
and often require evaporation and disposal of significant
(1) Bell, I. M.; Erb, J. M.; Freidinger, R. M.; Gallicchio, S. N.; Guare, J. P.;
Guidotti, M. T.; Halpin, R. A.; Hobbs, D. W.; Homnick, C. F.; Kuo, M.
S.; Lis, E. V.; Mathre, D. J.; Michelson, S. R.; Pawluczyk, J. M.; Pettibone,
D. J.; Reiss, D. R.; Vickers, S.; Williams, P. D.; Woyden, C. J. J. Med.
Chem. 1998, 41, 2146.
(2) Ku¨hler, T. C.; Swanson, M.; Shcherbuchin, V.; Larsson, H.; Mellgård, B.;
Sjo¨stro¨m, J.-E. J. Med. Chem. 1998, 41, 177.
(3) Loukiala, S.; Ratilainen, J.; Airola, K.; Valkonen, J.; Rissanen, K. Acta
Chem. Scand. 1998, 58, 593.
(4) Mosher, H. S.; Tessieri, J. E. J. Am. Chem. Soc. 1951, 73, 4925.
(5) Kobayashi, G.; Furukawa, S. Chem. Pharm. Bull. 1953, 1, 347.
(6) Rey-Bellet, G.; Spiegelberg, H. U.S. Patent 2,798,075, 1957; Chem. Abstr.
1957, 52, 2173.
(7) Itai, T.; Ogura, H. J. Pharm. Soc. Jpn. 1955, 75, 296.
(8) Kametani, T.; Fukumoto, K.; Nomura, Y. Chem. Pharm. Bull. 1958, 6,
467.
(9) Kim, D. Y.; Jung, I. C.; Lee, J. W.; Yoon, G. J. J. Korean Chem. Soc.
1998, 42, 102.
(10) Kim, W. J.; Park, T. H.; Kim, B. J.; Kim, M. H.; Pearson, N. Int. Pat.
Appl. WO 94/15938, 1994; Chem. Abstr. 1994, 121, 205332.
(11) Armarego, W. L. F.; Milloy, B. A.; Sharma, S. C. J. Chem. Soc. Perkin
Trans. 1 1972, 2485.
(12) Eur. Chem. News 1995, 64(1678), 5.
938
•
Vol. 6, No. 6, 2002 / Organic Process Research & Development
10.1021/op025545n CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/12/2002

cidentally added, coupled with loss of cooling and stirring.
This exceeds the reflux temperature of MTBE (55-56 °C).
These results, coupled with the uncontrollable outgassing of
isobutylene and sulfur dioxide, represent a serious chemical
reaction safety hazard and therefore were not scaled up. We
were not able to conclusively establish the reason for the
induction period observed during these and other similar
reaction conditions. One possibility is that a change in the
solubility of 1 in the presence of excess thionyl chloride could
contribute to the observed induction period. Alternatively,
the reaction of 1 with thionyl chloride could afford an
intermediate cyclic sulfite 3 that is then converted to 2.13
Sharpless and co-workers have described the facile synthesis
of cyclic sulfites from acyclic vicinal diols (<10 min), using
thionyl chloride in the presence of a base, for example,
triethylamine or pyridine, at 0 °C.14,15 Cyclic sulfites can also
be prepared from acyclic vicinal diols and thionyl chloride
at room temperature in the absence of base.14 In addition,
O’Brien and co-workers16 isolated the cyclic sulfite of
o-xylene-R, R-diol 5 in excellent yield from 4 using thionyl
chloride at 0 °C. However, Berridge and co-workers17
reported that cyclic sulfites in general were unreactive as
substrates in nucleophilic substitution reactions. The hetero-
geneous nature of the reaction mixture rendered attempts to
monitor it by FTIR unsuccessful. We have not further studied
the precise mechanism for conversion of 1 to 2.
Figure 1. RC-1 heat flow curve for experiment in MTBE at
45 °C. (Blue) Reaction temperature Tr. (Red) Heat flow. (Green)
Reactor pressure. (Orange) SOCl2 addition.
for completion. We were concerned about prolonged expo-
sure of MTBE to the acidic reaction conditions and further
investigated this reaction using reaction calorimetry.
All experiments were carried out in a Mettler RC-1
reaction calorimeter. The instrument was equipped with the
1-L MP-10 pressure reactor or the SV01 1-L vessel and a
propeller stirrer, temperature sensor, and calibration heater.
Following the necessary calorimetric measurements, thionyl
chloride was added via a dosing control loop using a
Prominent diaphragm pump for all experiments. Only the
addition time, solvent, and temperature were varied. These
reactions were heterogeneous at all times.
To minimize the reaction time, we evaluated the reaction
in MTBE at 45 °C anticipating that the chlorination would
be complete before MTBE cleavage might occur. The MP10
pressure reactor was charged with MTBE and the starting
diol 1 at 40 °C. Thionyl chloride was added over 30 min at
45 °C. The RC-1 heat flow results clearly indicate that the
addition of thionyl chloride is exothermic under these
conditions (red curve Figure 1).
The heat flow curve (in red) in Figure 1 indicates that
the reaction is clearly not “feed controlled” as evidenced by
the en-masse exotherm after a significant amount of thionyl
chloride had accumulated in the reactor (estimated at 80%
on the basis of the volume added to the reactor). This then
reacted spontaneously and resulted in a serious, uncontrolled
outgassing from the reaction mixture as shown by the
pressure curve (in green) in Figure 1. The outgas was ten-
tatively identified as a mixture of sulfur dioxide, the expected
byproduct of the chlorination reaction and isobutylene
resulting from acid-catalyzed decomposition of MTBE.
Support for this conclusion came from a separate experi-
ment using the Omnical CRC90e calorimeter. Significant
outgassing is observed when concentrated HCl is mixed with
MTBE at 40 °C which strongly supports the presence of
isobutylene as one of the components in the outgas. The
calculated “worst-case temperature rise” (Tad) of 25 °C
indicates that a final reaction temperature of 65 °C can be
attained if the entire charge of thionyl chloride was ac-
We next examined the reaction of 1 with thionyl chloride
in MTBE at ambient temperature in an effort to control the
vigorous outgassing. In this case the SVO1 reactor was
charged with MTBE and the starting diol 1 at 20 °C. Thionyl
chloride was added during 40 min at 25 °C. Once again the
RC-1 heat flow results (green curve, Figure 2) shows an
exothermic addition of thionyl chloride under these reaction
conditions.
The heat flow curve in Figure 2 (green curve) indicates
that the reaction is not “feed-controlled” and that unreacted
thionyl chloride again accumulates in the reactor (red and
orange curves in Figure 2). After approximately 20% of the
thionyl chloride is added, the reaction initiates and exotherms.
Two separate and distinct exothermic events are observed,
with the second and major one initiating only after ca. two-
thirds of the thionyl chloride is added. The calculated worst-
case temperature rise (Tad) of 53 °C indicates that the final
(13) We thank the editor for this suggestion.
(14) Gao, Y.; Sharpless, K. B. J. Am. Chem. Soc. 1988, 110, 7538-7539.
(15) Kim, B. M.; Sharpless, K. B. Tetrahedron Lett. 1989, 30, 655-658.
(16) O’Brien, M. K.; Sledeski, A. W.; Truesdale, L. K. Tetrahedron Lett. 1997,
38, 509-512.
(17) Berridge, M. S.; Franceschini, M. P.; Rosenfeld, E.; Tewson, T. J. J. Org.
Chem. 1990, 55, 1211-1217.
Vol. 6, No. 6, 2002 / Organic Process Research & Development
•
939

Figure 3. RC-1 heat flow curve toluene experiment at 25 °C.
(Blue) Reaction temperature Tr. (Green) Heat flow. (Orange)
SOCl2 addition. (Red) Thermal conversion to product.
Figure 2. RC-1 heat flow curve for experiment in MTBE at
25 °C. (Blue) Reaction temperature Tr. (Red) Thermal conver-
sion to product. (Green) Heat flow. (Orange) SOCl2 addition.
reaction temperature can reach 78 °C if all of the thionyl
chloride is charged with simultaneous loss of reactor cooling
or stirring or both. This again significantly exceeds the reflux
temperature of MTBE (55-56 °C). Additionally, under these
conditions, the HCl generated from the reaction can result
in isobutylene formation together with sulfur dioxide out-
gassing as observed previously. Therefore, a significant
chemical reaction hazard still exists for these reaction
conditions and they were not scaled up.
Substitution of toluene for MTBE would preclude any
safety issues arising from solvent decomposition. In addition,
we expected it to be inert under the reaction conditions and
easily removed. We were gratified to find that the handling
characteristics of 2 in toluene were comparable to those in
MTBE and should permit ready scale-up of the reaction
conditions. Therefore, toluene was selected as the solvent
for the next reactions.
The SV01 reactor was charged with toluene and the
starting diol 1 at 25 °C. Thionyl chloride was added during
40 min at 25 °C. Most surprisingly, there was no reaction at
25 °C as evidenced by the heat flow results from the RC-1
(green curve, Figure 3). Only after the entire charge of thionyl
chloride had been added and the reaction mixture was
warmed to 45 °C did the reaction initiate. Unfortunately,
this batch process again occurs spontaneously and results in
uncontrollable sulfur dioxide outgassing. Thus, these condi-
tions were not acceptable for scale-up.
We then raised the reaction temperature to 45 °C in an
attempt to initiate the reaction immediately upon addition
of SOCl2. In this case the SV01 reactor was charged with
toluene and the starting diol 1 at 45 °C. Thionyl chloride
was then added during 40 min, while maintaining the
temperature at 45 °C. The RC-1 heat flow results (Figure 4)
again indicate that the reaction still did not fully initiate after
addition of 1 equivalent of thionyl chloride as evidenced by
the heat flow exotherm (green curve in Figure 4).
Figure 4. RC-1 heat flow curve toluene experiment at 45 °C.
(Blue) Reaction temperature Tr. (Red) SOCl2 addition. (Green)
Heat flow.
feed-controlled. The calculated worst-case temperature rise
(Tad) was 23 °C, suggesting that the final reaction temper-
ature could reach 68 °C if there is a loss of cooling/stirring
with full accidental charge of the thionyl chloride. This final
Tad temperature was well below the reflux temperature of
toluene (111 °C). However, this was still a batch process
that occurs spontaneously at 40-45 °C. This concern
together with the large volumes of sulfur dioxide that are
generated spontaneously in an uncontrolled manner precluded
scale-up of these reaction conditions as well.
We expected that a feed-controlled reaction profile for
the addition of thionyl chloride would permit control of the
expected exotherm and sulfur dioxide outgassing. In addition,
we wanted to investigate the affects, if any, of a more reactive
chlorinating reagent on the troublesome induction period we
had encountered thus far in both MTBE and toluene.
Therefore, we modified the reaction conditions to incorporate
Again, as the second equivalent of SOCl2 was added, the
reaction occurred spontaneously. The addition of thionyl
chloride was exothermic under these conditions and was not
940
•
Vol. 6, No. 6, 2002 / Organic Process Research & Development

Figure 6. RC-1 heat flow curve toluene-DMF experiment at
45°C. (Blue) Reaction temperature Tr. (Red) Heat flow. (Green)
SOCl2 addition.
Figure 5. RC-1 heat flow curve for MTBE-DMF experiment
at 25 °C. (Blue) Reaction temperature Tr. (Red) Heat flow.
(Green) SOCl2 addition. (Orange) Thermal conversion to
product.
still 50 °C. As with the previous cases, an accidental charge
of all of the thionyl chloride with concomitant loss of cooling/
stirring would still significantly exceed the reflux temperature
of MTBE. This, coupled with the HCl-catalyzed decomposi-
tion of MTBE above 45 °C poses the same significant safety
concerns, and therefore, these conditions were also aban-
doned for scale-up.
On the basis of these results (vide supra) we concluded
that MTBE is not an acceptable solvent for this chlorination
reaction under any conditions we evaluated.
The last reaction conditions investigated employed toluene
as the solvent containing 1% v/v dimethylformamide (DMF).
The SV01 reactor is charged with toluene, 1% v/v DMF,
and the starting diol 1 at 25 °C. Thionyl chloride is added
in three portions at 25 °C during approximately 45 min. The
RC-1 heat flow results indicate that the reaction does initiate
at 25 °C (Figure 6).
Although the addition of the SOCl2 is exothermic under
these conditions, it is still not completely feed-controlled as
the heat flow curve indicates (red curve in Figure 6).
However, the calculated worst-case temperature rise (Tad)
is only 20 °C, indicating that the final reaction temperature
could reach 45 °C if there is a loss of cooling/stirring
combined with an accidental charge of thionyl chloride. This
final Tad temperature is well below the reflux temperature
of toluene. In addition, the reaction is reasonably well
controlled under these conditions, and a moderate evolution
of sulfur dioxide is not observed visually until the second
equivalent of SOCl2 is added. As noted earlier, the use of
toluene as the solvent completely precludes the complication
of isobutylene formation. Therefore, these reaction conditions
were recommended for and successfully scaled up to
multikilogram quantities (see Experimental Section).
a catalytic amount of DMF to effect chlorination under mild
conditions via the well-known Vilsmeier-Haack com-
plex.18,19 For example, N,N-dimethylchloromethaniminium
chloride cleanly converts nucleosides to 5′-chloro-5′-deoxy-
analogues.20,21 Similarly, this same reagent chlorinates
primary, secondary, and benzylic alcohols in very high yield
at room temperature.22 Recently, 4-cyano-2-fluorobenzyl
chloride was prepared from the alcohol via this reagent.23
MTBE was used as the solvent but now containing 1%
v/v dimethylformamide (DMF) to initiate chlorination via
in situ generated N,N-dimethylchloromethaniminium chlo-
ride. The SV01 reactor was charged with MTBE, 1% v/v
DMF, and the starting diol 1 at 25 °C. Thionyl chloride was
added in two portions over 20 min at 25 °C. The RC-1 heat
flow results (Figure 5) indicate that the addition of the first
equivalent of thionyl chloride is exothermic as evidenced
by the heat flow exotherm (red curve in Figure 5). The
exotherm is somewhat feed-controlled and the reaction
initiated once the thionyl chloride feed started. Here, use of
a more reactive chlorinating reagent seemed to be effective
to preclude the induction period we had encountered previ-
ously. Overall, we observed a more classically feed-con-
trolled process for the addition of the of thionyl chloride
which resulted in a more controlled heat generation and
outgassing as the reaction progressed.
Visually, the outgassing (SO2) appears to be much more
controlled under these conditions. However, the calculated
worst-case temperature rise (Tad) for the entire addition is
(18) Vilsmeier, A.; Haack, A. Chem. Ber. 1927, 60, 119.
(19) Giles, P. R.; Marson, C. M. In Encyclopedia of Reagents for Organic
Synthesis; Paquette, L. A., Ed.; John Wiley & Sons: New York, 1995;
Vol. 3, pp 2045-2047.
(20) Dods, R. F.; Roth, J. S. Tetrahedron Lett. 1969, 165.
(21) Dods, R. F.; Roth, J. S. J. Org. Chem. 1969, 34, 1627.
(22) Yoshihara, M.; Eda, T.; Sakaki, K.; Maeshima, T. Synthesis 1980, 746.
(23) Kimura, Y.; Suzuki, H.; Takao, Y. Jpn. Kokai Tokkyo Koho JP 2001172247,
2001; Chem. Abstr. 2001, 135, 61145.
Conclusions
On the basis of the reaction calorimetry results the reaction
conditions employing thionyl chloride and 1% v/v DMF in
Vol. 6, No. 6, 2002 / Organic Process Research & Development
•
941

toluene were found to be safe for scale-up. We also reduced
the amount of thionyl chloride required to a slight excess,
thereby minimizing disposal of excess reagent. As a result,
we have identified an economical, environmentally accept-
able, and safe process that has been scaled up to produce
>3.0 kg of 2 in 89% yield.
Further, this “simple” chlorination is much more complex
than initially believed. We have identified conditions using
thionyl chloride in MTBE as particularly hazardous owing
to solvent decomposition in the presence of HCl and
subsequent outgassing of large amounts of isobutylene.
to afford 2,3-bis(hydroxymethyl)pyridine‚HCl, 1 (352.0 gm,
44.0%) as a white solid of 88.6 wt % purity (HPLC). A
second crop of 1 (205 gm, 25%) was isolated as a white
solid of 99.0 wt % purity by reworking the mother liquors
from above. These crops were combined and used im-
mediately without further purification.
2,3-Bis(chloromethyl)pyridine Hydrochloride, (2): Neat
Thionyl Chloride. A 1-L three-neck Morton flask was
equipped with a reflux condenser, nitrogen inlet, and thermo-
couple. Thionyl chloride (81.55 gm, 0.686 mol) was charged
and cooled to 7.0 °C under nitrogen. 2,3-Bis(hydroxymethyl)-
pyridine hydrochloride, 1 (39.59 gm, 0.227 mol), was added
as a single portion with stirring. The temperature of the
resulting thick, white paste quickly reached 29 °C. An ice
bath was used to cool the suspension to 16 °C as all of the
starting diol dissolved. The reaction mixture was stirred for
an additional 35 min (post addition of the diol), at which
time the reaction was judged complete by HPLC analysis.
During this time the product precipitated as a white solid.
The suspension was diluted with 100 mL of MTBE and was
cooled to 2-3 °C in an ice bath. The suspension was filtered
and the white filter cake was washed with 100-200 mL of
MTBE. The filter cake was dried in a vacuum (28 in. Hg) at
room temperature overnight to afford 2,3-bis(chloromethyl)-
pyridine hydrochloride, 2 (46.01 gm, 96%), as a white solid
of 97.0 wt % purity (HPLC). Mp 150.0-152.0 °C; 1H NMR
(300 MHz, d6-DMSO) δ 10.64 (1H, br s, NH+), 8.68 (1H,
dd, Ar H), 8.19 (1H, dd, Ar H), 7.65 (1H, dd, Ar H), 5.03
(2H, s, ArCH2), 5.01 (2H, s, ArCH2). HRMS: Calcd for
C7H8Cl3N: 176.0034; Found: 176.0034.
2,3-Bis(chloromethyl)pyridine Hydrochloride, (2): 1%
(v/v) DMF in Toluene. A 12-L (4N) flask was equipped
with an air stirrer, nitrogen inlet, reflux condenser, 1-L
addition funnel, and a thermocouple. The flask was blanketed
with nitrogen and charged at room temperature sequentially
with toluene (5.85 L), 2,3-bis(hydroxymethyl)pyridine‚HCl,
1 (972.5 g, 5.5 mol), and DMF (50 mL, 0.65 mol). The
resulting mobile, white suspension was chilled to ca. 15 °C,
and thionyl chloride (940 mL, 6.45 mol) was added via the
addition funnel over 90 min, during which time the temper-
ature rose to 26 °C. At this point the original suspension
was now a white, somewhat tacky, heterogeneous mixture.
As this mixture was heated to 45 °C, most of 1 dissolved to
afford a cloudy solution. Heating was discontinued at this
time, and the reaction mixture was cooled to ambient
temperature during ca. 1.5 h. HPLC analysis of an aliquot
at this time showed no residual 1. The reaction mixture was
chilled to 5 °C, and ethyl acetate (1500 mL) was added to
aid in crystallization of the product. The solids were filtered,
washed with MTBE (1500 mL), and air-dried overnight on
the filter to afford 2,3-bis(chloromethyl)pyridine hydrochlo-
ride, 2 (1161 gm, 92.4%), as a white solid of 93.6 wt %
analysis (HPLC). MS: [M + H]+ ) 176, 178. Anal. Calcd
for C7H8Cl3N: C, 39.56; H, 3.79; N, 6.59; Cl, 50.05,
Found: C, 39.50; H, 3.77; N, 6.50; Cl, 48.72.
Experimental Section:
Melting points were determined on a Thomas-Hoover cap-
illary melting point apparatus and are uncorrected. All re-
agents and solvents were used as received from commercial
sources. 2,3-Pyridinedicarboxylic acid dimethyl ester was
purchased from Polycarbonate Industries. Nonaqueous reac-
tions were performed under an atmosphere of nitrogen. NMR
spectra were recorded on a Bruker 300 MHz spectrometer;
chemical shifts are expressed on the δ scale and are in ppm
downfield of the internal standard tetramethylsilane (TMS).
Spectra were recorded in d6-DMSO or CDCl3. HPLC were
obtained using a Hewlett-Packard HP 1100 equipped with a
photodiode array detector. Mass spectra were recorded on a
Hewlett-Packard series 1100 MSD. A Finnigan Navigator
AQA LC/MS instrument was used to obtain LC/MS data.
Elemental analyses were performed by QTI (Whitehouse,
NJ).
2,3-Bis(hydroxymethyl)pyridine Hydrochloride, (1).24
A 22-L (4N) flask was equipped with an air stirrer, nitrogen
inlet, 2-L addition funnel, and a thermocouple. The flask
was blanketed with nitrogen and sequentially charged at room
temperature with 6.5 L of ethanol, 2,3-pyridinedicarboxylic
acid dimethyl ester (890 gm, 4.56 mol) and sodium boro-
hydride. The resulting thin yellow suspension was stirred
and cooled to -5 to 0 °C. A solution of calcium chloride
(447.5 gm, 4.03 mol) in ethanol (2.75 L) was added slowly,
while maintaining the temperature at <5 °C. After the
addition, the reaction mixture was allowed to slowly warm
to ambient temperature and was judged to be complete after
2 h by HPLC analysis. The reaction mixture was quenched
with 3.6 L of 50:50 aqueous ethanol after which it was
evaporated to recover a yellow solid that was dried in a
vacuum (28 in. Hg) overnight at room temperature. The crude
product was ground to a powder and transferred to a 22-L
(4N) flask. Ethanol (16 L) was added and the suspension
heated to reflux and held for 1.0 h. Agitation was stopped
and the suspension cooled to ambient temperature whereupon
the slightly turbid ethanolic solution of crude diol was
transferred to a 22-L (3N) flask. After this solution cooled
to 10 °C, it was saturated with HCl(g) to precipitate 1 as a
white solid. The product was collected by filtration and
washed, first with a small volume of ethanol and then with
MTBE. After air-drying for several hours the product was
dried in a vacuum (28 in. Hg) at room-temperature overnight
(24) This procedure was modified from that used for reduction of 2,3-pyridine-
dicarboxylic acid diethyl ester. See: Matsumoto, I.; Yoshizawa, J. Jpn.
Kokai Tokkyo Koho JP 49020181, 1974; Chem. Abstr. 1974, 81, 120469.
Received for review April 29, 2002.
OP025545N
942
•
Vol. 6, No. 6, 2002 / Organic Process Research & Development