Risk assessment and safety evaluation study for ozonolysis of i-pinene: Raw material of a novel prostaglandin D2 receptor antagonist S-5751

Article abstract of DOI:10.1021/op100065d

Safety evaluation for ozonolysis of β-pinene was conducted using DSC, ARC, and RC1e. ARC measurement showed that the concentration of the ozonized mixture affects its thermal stability. The sample at a high concentration of 1.83 mol/L was very unstable even at room temperature and decomposed rapidly. ADT₂₄ of the diluted sample of 0.73 mol/L was 28.2°C. Although reductive decomposition of the ozonized mixture at extremely low temperature is very dangerous because of accumulation of the heat of reaction, its process temperature was validated to not exceed the ADT₂₄. As a result, pilot manufacturing on a 300-L reactor scale could be achieved.

Full text of DOI:10.1021/op100065d

Organic Process Research & Development 2010, 14, 1485–1489
Risk Assessment and Safety Evaluation Study for Ozonolysis of ꢀ-Pinene: Raw
Material of a Novel Prostaglandin D2 Receptor Antagonist S-5751
Takemasa Hida,*,† Junko Kikuchi,‡ Makoto Kakinuma,§ and Hideo Nogusa§
Shanghai Office, Shionogi & Co., Ltd., 306A, 3F, Building A, Far East International Plaza, No. 319, Xianxia Road,
Shanghai 200051, China, DiscoVery Research Laboratories, Shionogi & Co., Ltd., 12-4, Sagisu 5-chome, Fukushima-ku,
Osaka 553-0002, Japan, and CMC DeVelopment Laboratories, Shionogi & Co., Ltd., 1-3, Kuise Terajima 2-chome, Amagasaki,
Hyogo 660-0813, Japan
Scheme 1. New synthetic route to S-5751 1 from ꢀ-pinene 2
Abstract:
Safety evaluation for ozonolysis of ꢀ-pinene was conducted using
DSC, ARC, and RC1e. ARC measurement showed that the
concentration of the ozonized mixture affects its thermal stability.
The sample at a high concentration of 1.83 mol/L was very
unstable even at room temperature and decomposed rapidly.
ADT24 of the diluted sample of 0.73 mol/L was 28.2 °C.
Although reductive decomposition of the ozonized mixture
at extremely low temperature is very dangerous because of
accumulation of the heat of reaction, its process temperature
was validated to not exceed the ADT24. As a result, pilot
manufacturing on a 300-L reactor scale could be achieved.
have been reported for obtaining its key intermediate.8-10
new synthetic route from 3 to (+)-2-((1R,2R,3R,5S)-2-amino-
6,6-dimethylbicyclo[3.1.1]hept-3-yl)-ethanol (4), which is one
of the key intermediates of S-5751, was reported by the
Shionogi group.10 To obtain 3, ozonolysis of ꢀ-pinene 2 is
necessary. Although a procedure for obtaining 3 from 2 using
ozonolysis has been reported by Grimshaw et al.,11 a detailed
safety evaluation study is necessary for scale-up manufacturing
of 3. Here we describe a risk assessment and detailed safety
evaluation study of ozonolysis and present a case of practical
application for scale-up ozonolysis.
A
Introduction
Since Scho¨nbein first studied ozonolysis in 1855,1 this
reaction has been applied to many types of industrial produc-
tion,2 for example, for the preparation of antibiotics such as
ceftibuten and cefaclor.3,4 Ozonolysis is a useful reaction for
obtaining carbonyl compounds directly from olefins, and ozone
is a very “green” oxidant because it only produces oxygen as
a coproduct after the reaction. However, ozonolysis is known
to be a very dangerous reaction which may cause explosions.
Recently, microreactor systems have been used for kilogram-
scale synthesis,5 but detailed safety evaluation study and risk
assessment of ozonolysis are needed for practical industrial
manufacturing.
Results & Discussion
Identification of Products from Ozonolysis and Their
Evaluation Using CHETAH. Reaction products were deter-
mined by NMR of the reaction mixture. The reaction mecha-
nism by Criegee12 is outlined as follows (Scheme 2). In the
first step, ozone and 2 made the 1,2,3-trioxolane structure
(primary ozonide 5). In the second step, 5 underwent heterolysis
to give a zwitterion (6 or 7) and a carbonyl compound (3 or
formaldehyde 8). In the final step, the zwitterion combined with
the carbonyl compound to give normal ozonide 9. However
this ozonolysis was carried out in methanol. In a protonic
solvent, here methanol, 7 can be easily diverted to 10 instead
(+)-Nopinone 3 is a raw material of the prostaglandin D2
(PGD2) receptor antagonist S-5751 1, which was discovered at
Shionogi Research Laboratories.6,7 Several synthetic methods
* To whom correspondence should be addressed. Fax: 86-21-6235-1311.
Telephone: 86-21-6235-1311. E-mail: takemasa.hida@shionogi.co.jp.
† Shanghai Office, Shionogi & Co., Ltd.
‡ Discovery Research Laboratories, Shionogi & Co., Ltd.
§ CMC Development Laboratories, Shionogi & Co., Ltd.
(1) Scho¨nbein, C. F. J. Pract. Chem. 1855, 66, 282.
(2) Van Ornum, S. G.; Champeau, R. M.; Pariza, R. Chem. ReV. 2006,
106, 2990.
(8) (a) Campos, K. R.; Lee, S.; Journet, M.; Kowal, J. J.; Cai, D.; Larsen,
R. D.; Reider, P. J. Tetrahedron Lett. 2002, 43, 6957. (b) Campos,
K. R.; Journet, M.; Lee, S.; Cai, D.; Kowal, J. J.; Lee, S.; Larsen,
R. D.; Reider, P. J. J. Org. Chem. 2003, 68, 2338.
(3) Yoshioka, M. Pure Appl. Chem. 1987, 59, 1041.
(4) Bernasconi, E.; Lee, J.; Sogli, L.; Walker, D. Org. Process Res. DeV.
2002, 6, 169.
(5) Hu¨bner, S.; Bentrup, U.; Budde, U.; Lovis, K.; Dietrich, T.; Freitag,
A.; Ku¨pper, L.; Ja¨hnisch, K. Org. Process Res. DeV. 2009, 13, 952.
(6) Arimura, A.; Yasui, K.; Kishino, J.; Asanuma, F.; Hasegawa, H.;
Kakudo, S.; Ohtani, M.; Arita, H. J. Pharmacol. Exp. Ther. 2001,
298, 411.
(9) Jennequin, T.; Labat, S.; Toupet, L.; Caille, J. C.; Mauduit, M. Synlett
2008, 11, 1669.
(10) Hida, T.; Mitsumori, S.; Honma, T.; Hiramatsu, Y.; Hashizume, H.;
Okada, T.; Kakinuma, M.; Kawata, K.; Oda, K.; Hasegawa, A.; Masui,
T.; Nogusa, H. Org. Process Res. DeV. 2009, 13, 1413.
(11) Grimshaw, J.; Grimshaw, J. T.; Junejia, H. R. J. Chem. Soc. Perkin 1
1972, 50.
(7) Mitsumori, S.; Tsuri, T.; Honma, T.; Hiramatsu, Y.; Okada, T.;
Hashizume, H.; Kida, S.; Inagaki, M.; Arimura, A.; Yasui, K.;
Asanuma, F.; Kishino, J.; Ohtani, M. J. Med. Chem. 2003, 46, 2446.
(12) Criegee, R.; Scro¨der, G. Chem. Ber. 1960, 93, 689.
10.1021/op100065d  2010 American Chemical Society
Published on Web 06/15/2010
Vol. 14, No. 6, 2010 / Organic Process Research & Development
•
1485

Scheme 2. Reaction mechanism and products on the ozonolysis of ꢀ-pinene 2
of 9.13 According to this mechanism, zwitterion 7 might
dimerize to 13 or polymerize to 14 (Figure 1), both of which
seem to be difficult to reduce, but these compounds were not
detected in the NMR chart. The major products were methoxy
hydroperoxide 10 and 11, and (+)-nopinone 3 and ozonide 9
were detected as minor products. During the NMR measure-
ment, 10 was found to be very unstable to heat and was
decomposed to 3 even at room temperature overnight (Figure
2). On the other hand, 9 was difficult to decompose to 3. These
results mean that the reaction is fraught with the danger of
turning into an explosive decomposition by 10 under adiabatic
conditions.
the results, although |∆Hc - ∆Hd| of both 9 and 10 are not so
high, both compounds are explosive. The ∆Hd’s of both
compounds are comparatively large,14 and they should be treated
carefully.
Thermal Screening with DSC and ARC. Using DSC, the
exothermal energy value and onset temperature of the ozonized
mixture in methanol, which is charged on 1.83 mol/L of 2 (∼29
w/w% of 10), were measured using both gold and SUS cells
(Table 2 and Figure 3). The exothermal energy and peak shape
are similar, regardless of the material of cell. Thus, the SUS
reactor can also be used for actual manufacturing. Although
DSC shows the onset point at over 70 °C, the actual temperature
rise of the exothermal peak starts from 50 °C. The exothermic
curve is sharp, and the decomposition seems to proceed
explosively. These data indicated that the ozonized mixture
should be treated at very low temperature. We concluded that
such an extremely unstable compound should be investigated
in more detail; therefore, we carried out ARC measurements.
The potential explosive risks of 9 and 10 were assessed by
theoretical calculation using CHETAH software (Table 1). From
Figure 1. Putative dimer and polymer.
Figure 2. NMR of ozonized mixture and its transformation at room temperature for overnight.
1486
•
Vol. 14, No. 6, 2010 / Organic Process Research & Development

Table 1. Results of CHETAH
net
plosive
density
oxygen balance
gO2/100 g
overall energy
release potential
b
compound
∆Hca kJ/g
∆Hd kJ/g
|∆Hc - ∆Hd| kJ/g
9
–29.472
–2.608
27.725
LOW
26.917
LOW
–223.346
MEDIUM
–217.106
MEDIUM
–0.337
HIGH
–0.348
HIGH
0.277
MEDIUM
–2.566
PLOSIVE
0.275
10
–30.322
MEDIUM
PLOSIVE
a Fuel value. b Maximum heat of decomposition.
Table 2. DSC data of ozonized mixturea
chart of 1.83 mol/L (Figure 4), in which the pressure curve is
a single one while the exothermic curve has two more slopes,
the decomposition seems to proceed by interaction with two
molecules of 10 but not with methanol. In 0.73 mol/L of the
ozonized mixture, the decomposition seems to proceed with
one molecule of 10 itself because the exothermic curve is only
a slope (Figure 5). Thus, dilution of the product is necessary to
stabilize the product.
material
of cell
peak
top (°C)
energy
∆H (J/g)
onset (°C)
gold
SUS
71
81
107
105
519
523
a Using ozonized mixture (concentration substrate/methanol: 1.83 mol/L of
10). Scanning ratio is 10 K/min. Range of scanning temperature dis from -25
to 350 °C.
Heat of Reaction (RC1e). The heat of reaction for ozo-
nolysis of 2 was measured by RC1e (Table 3 and Figure 6). It
is a very large value of 498.95 kJ/mol but can be controlled by
the flow rate of O3/O2. In addition to controlling the heat of
ozonolysis, the reaction temperature must be kept low from the
viewpoint of assuring safety. The ozonolysis is carried out under
an oxygen atmosphere, and the concentration of methanol in
vapor should be controlled to below the lower explosive limit
in oxygen.
Reductive decomposition of the ozonized mixture using tri-
methyl phosphite was carried out to obtain 3, and the heat of
reaction was measured at -25 °C (Table 4 and Figure 7). It was
found that the accumulation of heat on dropwise addition of 1.0
equiv of trimethyl phosphite is over 33%. Then the heat of reaction
was measured at -10 °C; however, the heat of reaction at dropwise
addition of 1.0 equiv of trimethyl phosphite still accumulated 25%
(Table 4 and Figure 6). Although the ozonized mixture is very
unstable to heat and the upper process temperature should be
controlled at below ADT24, reductive treatment should not be
carried out at extremely low temperatures. The control of reductive
decomposition seems to be the most important point for safety
The ozonized mixture, the concentration of which is the same
as that of the DSC sample, ∼29 w/w% of 10, was measured
by ARC (Figure 4). It was found to be too unstable even at
room temperature to detect the onset temperature of self-heating.
The rate of self-decomposition was fast and its adiabatic
temperature was over 108.70 °C (after ꢁ correction).15 There-
fore, a diluted ozonized mixture, 0.73 mol/L (∼14 w/w% of
10), was prepared and subjected to measurement (Figure 5). It
showed an onset temperature of self-heating of 42.09 °C, and
thus the ADT24 was calculated to be 28.2 °C after ꢁ correction
and extrapolation.16 Self-heating was raised to 65.42 °C, and
thus the adiabatic temperature rise was calculated to be 44.39
°C from the ꢁ correction.
The process temperature of ozonolysis and reductive de-
composition should be controlled at below ADT24. Next, we
measured the heat of reaction and set the upper limits for the
process temperature.
The ARC measurements yielded experimental results show-
ing that the stability of the ozonized mixture is closely related
to the reaction concentration. The exothermic curve of 1.83
mol/L differs from that of 0.73 mol/L. In view of the ARC
Figure 3. DSC chart of ozonized mixture (1.83 mol/L).
Vol. 14, No. 6, 2010 / Organic Process Research & Development
•
1487

Figure 4. ARC chart of ozonized mixture (1.83 mol/L).
Figure 5. ARC chart of ozonized mixture (0.73 mol/L).
Table 3. Result of RC1e experimenta
specific heat
of reaction mass
process
temperature (°C)
∆Hreac
(kJ/mol)
b
(J/(g ·K))
∆Tad (K)
-40
498.95
2.15
185.9
a Ozonolysis was carried out with 0.36 mol/L of 2. b Corrected value of 0.73
mol/L.
assurance. The maximum temperature of the synthetic reaction
(MTSR) was calculated at each temperature (Table 4). MTSR by
dropwise addition of trimethyl phosphite at -25 to -10 °C for 90
min was below ADT24, which means that the risk of a runaway
reaction by self-heating decomposition is not very large. The time
maximum rate (TMR) at -10 °C is 7.6 × 104 h; thus, there is no
Figure 6. Heat flow of ozonolysis of 2.
problem such as exceeding the upper limit temperature of reductive
decomposition.
1488
•
Vol. 14, No. 6, 2010 / Organic Process Research & Development

Table 4. RC1e experimental results of reduced
decomposition was investigated by RC1e. In conclusion, the
hazardous points of ozonolysis are: the thermal instability of
the products due to a high concentration of the reaction mixture,
the process temperature of ozonolysis, the process temperature
of reductive decomposition, dropwise addition time of the
reducing reagent and complete decomposition of ozonide.
Particularly hazardous is the accumulation of heat at the
reductive decomposition, requiring careful operation and control
of the process temperature. Twenty-seven batches of pilot
manufacturing on a 17 kg/lot of 2 basis were achieved
successfully according to the safety evaluation study.
decomposition of ozonized mixture and MTSRa
process
specific heat of
heat
temperature ∆Hreac
reaction mass accumulation
(°C)
(kJ/mol)
(J/(g · K))
(%)b
∆Tad (K)c MTSRd
-25
-10
429.25
442.39
2.52
2.52
33.3
25.0
140.24
144.61
25.36
27.80
a Experiment was carried out using 0.37 mol/L of ozonized solution of 2. b Heat
accumulation at dropwise addition of 1.0 equiv. of trimethyl phosphite. c This
value is converted to the solution of 0.73 mol/L of 2. d MTSR ) TP + (1 - XST
/
100) × ∆Tad × MF/MST (TP: Process temperature, XST: Conversion at dropwise
addition of stoichiometric reagent, MST: Mass of reaction mixture at dropwise
addition of stoichiometric reagent, MF: Mass of reaction mixture at finish of
dropwise addition, ∆Tad adiabatic temperature rise of reaction heat.). Dropwise
addition time of trimethyl phosphite.
Experimental Section
NMR spectra were measured on a VarianUnity Inova-500 or
Inova-600. DSC was measured on a Perkin-Elmer Pyris 1. ARC
was measured by Asahi Kasei Chemicals Corporation Explo-
sives R&D Center.
Procedure for Manufacturing of (+)-nopinone 3 in a
Pilot Plant. In a SUS reactor (300 L), ꢀ-pinene 2 (17.0 kg,
124.8 mol) was mixed with methanol (150 kg, 187 L), and the
mixture was cooled to -57 °C and stirred with a stream of
ozonized oxygen gas for 8 h. The reaction was checked for
completion by TLC, and then the mixture was heated to -27
°C. Trimethyl phosphite (31 kg, 249.8 mol) was added dropwise
under the same conditions for over 3 h. The reaction mixture
was heated to 0 °C for 1 h, and aqueous 10% H2SO4 (50 kg)
was added. After stirring at 25 °C for 4 h, the disappearance of
ozonide 9 was checked by TLC. After the checking for
disappearance of peroxide using KI starch paper, the reaction
mixture was transferred to glass-lined reactor (1000 L). Toluene
(88 kg, 102 L) and water (340 kg) was added and stirred. The
organic layer was separated and washed with aqueous 5%
NaHCO3 (85 kg), followed by aqueous 10% Na2SO3 (42.5 kg).
The aqueous layer was extracted by toluene and then separ-
ated. The organic layers were combined. After the removal of
toluene, the residue was distilled under 5 mmHg at 90-100
°C of jacket temperature and gave (+)-nopinone 311 (11.25 kg,
65.2%, boiling point at 5 mmHg: 78-80 °C).
Figure 7. Heat flow of reductive decomposition of ozonized
mixture.
Other Risk Assessments. Ozonolysis of 2 in methanol gave
ozonide 9 as well as 10. Reduction of 9 proceeded with
difficulty, which compares with that of 10. The remaining 9
can cause an explosion on distillation of 3. Reductive decom-
position of 9 was accelerated by addition of sulfuric acid at
room temperature. The subsequent operation was carried out
only after the complete termination of reductive decomposition
of 9 was confirmed.
The 3 obtained was extracted with toluene and condensed
under reduced pressure. The thermal stability of the residue was
evaluated with ARC. Its ADT24 is calculated as 113.5 °C and
TMR at 100 °C is 1.32 × 103 h. A high-performance pump,
which can reduce the pressure to at least below 20 mmHg, is
necessary to distill 3 at below 100 °C of jacket temperature.
Acknowledgment
We thank Ms. Mao Kuriyama and Mr. Shuji Kawabe in
Asahi Kasei Chemicals Corporation Explosives R&D Center
for measurement of ARC and helpful discussions on safety
evaluation study.
Conclusion
A risk assessment and safety evaluation studies for ozo-
nolysis of ꢀ-pinene were conducted in detail. The thermal
stability of the ozonized mixture was clarified by DSC and ARC
measurement and the accumulation of heat of the reductive
Supporting Information Available
Copies of two-dimensional NMR spectra of 3, 9, and 10;
copies of TLC for checking reaction termination; ARC chart
of residue of (+)-nopinone 3, and summarized table of ARC
measurements. This material is available free of charge via the
Internet at http://pubs.acs.org.
(13) Murray, R. W. Acc. Chem. Res. 1968, 1, 313.
(14) CHETAH has been developed by ASTM (American Society for
Testing and Materials), which can calculate the released energy of
organic compound under ideal gas state.
(15) ARC use sample container of anti pressure type, and some of
exothermal heat of decomposition can be consumed for heating of
sample container. Therefore experimental value must be corrected by
correction factor ꢁ. ꢁ ) 1 + [MW × CPW/(MS × CPS)] (MW: mass of
container, CPW: specific heat of container, MS: mass of sample, CPS
:
specific heat of sample).
(16) ADT24 (adiabatic decomposition temperature for 24 h) is the temper-
ature at which time to maximum rate (TMR) is 24 h.
Received for review March 8, 2010.
OP100065D
Vol. 14, No. 6, 2010 / Organic Process Research & Development
•
1489