series of scans with different heating rates and modeling the
adiabatic case using the AKTS software. The relevant value
with respect to production is always the temperature at which
the TMRad is 24 h (TD24) which was calculated to be 88 °C for
the hydrochloride of 10. To judge the criticality the relative
positions of four different temperature levels have to be
considered: process temperature (65 °C), boiling point of the
mixture (73 °C), TD24 (88 °C), and MTSR (161 °C). Accord-
ingly, the reaction can be classified as criticality class 4. This
means that a runaway is prevented only by the boiling of the
solvent, which (and its eventual reflux) must be possible also
in the case of a cooling failure.
Due to the high exothermal decomposition potential of 9
the question arose whether the leaving group might form an
explosive gas mixture during the reaction. By monitoring a
laboratory reaction we found that the maximum gas evolution
occurred after 2 h with a gas production of 0.84 L/(min·kg).
Conclusion
Due to the high interest in the PPI razeprazole (4), a high-
yielding telescoped procedure for the conversion of the nitro-
compound 9 to the rabeprazole synthon 12 (Scheme 3) was
developed which turned out to be highly reproducible in the
laboratory as well as on 30-kg scale. Due to reduction of
crystallizations and optimized reaction conditions, cycle time
could be significantly reduced and the throughput increased in
5-8
comparison to those from the standard laboratory procedures.
As the starting material 9 is known to possess a strong
exothermal decomposition potential, extensive safety investiga-
tions were made, and the whole process was adapted in a safe
and reliable manner. With these precautions, no safety issues
were observed either in the laboratory or in the pilot plant.
Furthermore, the chloro-compound 10 would also serve as a
starting material for the synthesis of the lansoprazole synthon
1
1 which presents a cost-saving and safe alternative to the
11
single-step procedure described recently.
2
The resulting gas contained mainly N O (19%), NO (31%), and
HCl (14%), thus resembling a mixture of HCl and HNO
known for its high corrosive potential. In addition, the formation
of an explosive mixture of organic compounds such as EtOH
x
,
Experimental Section
Differential scanning calorimetry was performed using a
Mettler DSC-821e differential scanning calorimeter. RC1
experiments were performed in a Mettler RC1 Classic reaction
calorimeter. Calculations of TMRad were done by using either
the Netzsch Thermokinetics or the AKTS software. Monitoring
of the reaction progress and assay of the final compounds were
done by HPLC and comparison to reference substances. As
potentially unstable compounds are involved, the noted tem-
peratures and distillation times must not be exceeded.
and EtOAc with NO and N
be achieved by heavily purging the reaction vessel with N
2
O had to be avoided which could
all
2
the time. No formation of less volatile nitrogen-containing
byproduct was observed.
During the synthesis of the chloro-N-oxide 10 several
distillations and solvent exchanges are necessary; therefore, the
thermal decomposition behaviors as well as the stabilities of
the intermediate hydrochloride and the final free base 10 had
to be investigated. As described above, the hydrochloride of
5
,10
4-Chloro-2,3-dimethylpyridine-N-oxide (10). Acetyl chlo-
ride (30 kg, 382 mol) was slowly (2 h) added to a cooled (4 °C)
solution of 9 (25.8 kg, 153.6 mol) in ethanol (65 L) keeping the
reaction temperature below 30 °C. After full addition, the mixture
was heated to 65 °C and stirred for 5 h. After distillation of the
volatiles, water (22 L) and toluene (33 L) were added followed by
pH adjustment by adding aqueous NaOH (50%) (12 L) targeting
a pH of 7.5-8.5. After phase separation at 70 °C, the aqueous
layer was re-extracted with toluene (16 L), and the combined
organic layers were evaporated to a residual volume of 50 L.
10 showed a strong exothermal decomposition of 906 J/g
starting at 124 °C, and the TD24 was calculated to be 88 °C. To
ensure safe distillation conditions, the following maximum
temperatures were defined for production: the internal temper-
ature was limited to 75 °C, and the jacket temperature was
limited to 100 °C. In addition, the distillation must not take
longer than 24 h under these conditions.
The free base 10 showed a strong autocatalytic decomposi-
tion of 1378 J/g starting at 202 °C, resulting in a TD24 ) 98 °C.
Accordingly, the severity for a runaway reaction has to be
classified as high, but the probability of occurrence at process
temperature is low, resulting again in a low overall risk for a
thermally initiated decomposition. Furthermore, our strategy to
isolate 10 as a solution even lowers the risk for a runaway
reaction as the heat of decomposition is lowered significantly
and the shape of the exothermal peak (DSC) indicates a change
in the kinetic parameters towards a less pronounced autocatalytic
decomposition.
(Caution: Distillation must not take longer than 24 h due to safety
reasons!) Water (1.6 L) and toluene (7 L) were added to the
distillation sump, and again the volatiles were distilled off followed
by the addition of 3-methoxy-1-propanol (32.4 L) which gave a
solution of 23.2 kg of 10 (147.5 mol, 96% yield) in a mixture of
3-methoxy-1-propanol and toluene (61 L, 60 kg).
To obtain an analytically pure sample, this solution was
evaporated to dryness and crystallized from cyclohexane, giving
1
10 in 75% as slightly yellowish crystals, mp 103-106 °C. H
NMR (300 MHz, CDCl
J ) 6.9 Hz, 1H), 2.57 (s, 3H), 2.40 (s, 3H).
-(3-Methoxypropoxy)-2,3-dimethylpyridine-N-oxide (12).
MTBACl (4.86 kg, 75% in water) and aqueous NaOH (50%)
16 L) were added to a solution obtained in step 1 (61 L
containing 23.2 kg of 10 (147.5 mol) and 32.4 L of 3-methoxy-
-propanol (338 mol)). The mixture was heated to reflux (100
3
): δ 8.10 (d, J ) 6.9 Hz, 1H), 7.18 (d,
Further on in the sequence, no serious safety issues requiring
extraordinary high precautions were observed. The conversion
of 10 to 12 was found to be slightly exothermal with a heat of
4
(
reaction (Q′ ) ) 43.3 kJ/kgreaction mass (determined by RC1
r
measurements). Carrying out the same systematic approach as
described above, the criticality class for this reaction is 3. The
isolated product 12 showed an exothermal decomposition of
1
°C) and stirred for 8 h. H O (34 L) was added, and the phases
2
were separated at 80 °C. After distilling off the volatiles, a
585 J/g with a TD24 ) 116 °C.
residue (37.4 kg) containing 12 (29.3 kg, 138.7 mol, 94% yield)
5
66
•
Vol. 14, No. 3, 2010 / Organic Process Research & Development