Process Safety Evaluation and Scale-up of a Lactam Reduction with NaBH4 and TFA

Article abstract of DOI:10.1021/op500284e

The development of a practical and scalable method for the synthesis of morpholine derivative 2 via reduction of lactam 1 using NaBH₄ and trifluoroacetic acid (TFA) is described. Through the mechanistic studies using ¹¹B NMR spectros

Full text of DOI:10.1021/op500284e

Article
pubs.acs.org/OPRD
Process Safety Evaluation and Scale-up of a Lactam Reduction with
NaBH and TFA
4
Takashi Naka,* Hiroki Ozawa, Ken-ichi Toyama, and Tadashi Shirasaka
Process Development Laboratories, CMC Division, Mitsubishi Tanabe Pharma Corporation, 3-16-89 Kashima, Yodogawa-ku,
Osaka-shi, Osaka 532-8505, Japan
*
S Supporting Information
ABSTRACT: The development of a practical and scalable method for the synthesis of morpholine derivative 2 via reduction of
11
lactam 1 using NaBH and trifluoroacetic acid (TFA) is described. Through the mechanistic studies using B NMR
4
spectroscopy, we observed that active species were generated during TFA addition, which avoids the deactivation of the active
species involved in the reduction process. Process safety assessments were performed on the basis of the reaction mechanism
using reaction calorimetry and gas evolution analyses. Furthermore, safety equipment was prepared to prevent the leakage of
diborane gas during the reaction, and the process was successfully scaled up to the pilot plant at a 35 kg scale.
INTRODUCTION
investigated in an attempt to avoid the disadvantages of using
■
LiAlH (Table 1).
Lactam reduction is employed in a wide range of syntheses by
the pharmaceutical and chemical industries. Lactams are
usually reduced using LiAlH , borane complexes, or NaBH .
4
1
a
2
3
4
Table 1. Screening of lactam-reducing agents
4
4
Generally, these reagents generate hydrogen and high reaction
heat; therefore, safety evaluation of such reduction processes is
essential for plant-scale manufacturing.
Morpholine derivative 2 is a key intermediate in our tau
protein kinase 1 inhibitor program. 2 can be obtained via the
reduction of lactam derivative 1. In the medicinal chemistry
b
c
c
5
temp.
yield
2
3
entry
conditions
(°C)
(%)
(%)
(%)
route, lactam has been reduced using LiAlH , which is difficult
4
1
2
3
4
LiAlH (2.0 equiv)
66
30
15
25
79
86
95
80
97.6
93.5
97.6
99.4
0.6
0.2
4
to handle in plant-scale manufacturing because it is flammable
Vitride (2.0 equiv)
6
and explosive and causes undesirable side reactions. Several
d
BH ·THF (2.3 equiv)
n.d.
n.d.
3
reducing agents having reactivities similar to LiAlH4 are
d
(1) KBH (5 equiv),
1
4
available; however, some reagents cause difficulties when a
LiCl (5 equiv);
(
2) TMSCl (10 equiv)
reaction using them is scaled up. For example, BH ·THF is a
3
d
5
NaBH (3.5 equiv),
25
85
99.6
n.d.
high-cost reagent and presents a storage hazard due to the risk
4
7
,8
TFA (3.5 equiv)
of explosion. Meanwhile, a combination of NaBH and
4
a
Screening were carried out in THF for the solubility of 1 and 2.
Determined by HPLC. Detected by HPLC. Not detected.
additives is equivalent to LiAlH in terms of its reducing
4
b
c
d
6
properties and is easy to handle in a manufacturing plant.
Therefore, we focused on the reduction conditions during the
conversion of 1 to 2 using NaBH and additives. However, the
4
reduction process is highly exothermic, and evaluation of the
Lactam reduction proceeded when several reducing agents
were employed. The amount of byproduct 3 could not be
safety of the process is difficult because the active species
9
,10
involved are not well understood.
We focused on perform-
maintained within the desired range when Vitride (NaAl-
8
ing a safety assessment based on the reaction mechanism
because a deeper understanding of the reaction mechanism can
help us better control the active species involved in the
reduction process.
H (OCH CH OCH ) ) was used (entry 2). Using BH ·THF
2
2
2
3 2
3
caused the lactam moiety to be reduced without forming
byproduct 3 (entry 3). However, BH ·THF is difficult to
3
handle as a storage hazard due to the risk of explosion. KBH
4
11
Here we show that lactam reduction can be achieved using
and LiCl in the presence of TMSCl also reduced 1 without
producing byproduct 3 (entry 4). Furthermore, 1 was
NaBH and TFA, and we report the results of our safety
4
12
evaluation and scale-up of the reduction process.
selectively reduced using a combination of NaBH
4
and TFA
(
entry 5), which is easier to handle than either LiAlH or BH ·
4
3
RESULTS AND DISCUSSION
■
Special Issue: Safety of Chemical Processes 15
Received: September 3, 2014
Screening of Reducing Agents. In the medicinal
chemistry route, lactam reduction has been conducted using
5
LiAlH . Several other methods for reducing lactams were
4
©
XXXX American Chemical Society
A
DOI: 10.1021/op500284e
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
THF. Therefore, NaBH combined with TFA was used in the
subsequent investigations.
trifluoroacetoxy species and the borane species were observed
(Figure 1c and d). We speculated that the reactivity of the
trifluoroacetoxy species depends on the number of trifluor-
oacetoxy groups. The lactam reduction reaction was not
4
The actual reducing agent in the NaBH and TFA mixture is
4
known to be sodium trifluoroacetoxyborohydride
(
NaBH4−n(OCOCF ) (n = 1, 2, or 3)). The effects of using
completed when NaBH and TFA were mixed before 1 was
3 n
4
different additives were also investigated. The reducing mixture
was more reactive with TFA than with acetic acid. However,
the detailed reaction mechanism and safety assessment of the
added. From this result, we speculated that the species
12
containing more TFA moieties had a lower ability to reduce 1.
1
1
Next, the actual reaction system using 1 was investigated. B
NMR spectra of borate derivative 4 are shown in Figure 2a.
When 1.75 equiv for 1 and 3.5 equiv of TFA for 1 was added to
reduction process using NaBH and TFA have not been
4
reported. Therefore, we investigated the reaction mechanism
1
1
using B NMR spectroscopy and performed a safety
assessment for the reaction process on the basis of the reaction
mechanism using reaction calorimetry and gas evolution
1 and NaBH in THF, trifluoroacetoxy species and borane
4
species were both observed (Figure 2d and e). This indicates
that there are two active species involved in the reduction
process, trifluoroacetoxy and borane species.
9
,10
analyses.
1
1
Reaction Mechanism. The detailed reaction mechanism
was investigated so that the safety of the reaction using the
reducing agents could be evaluated. We focused on semibatch
approach by the addition of TFA. Another procedure was not
appropriate for the following reasons: (1) The lactam reduction
From the B NMR results, the following reaction
mechanism is speculated (Scheme 1): NaBH is converted
4
into sodium monotrifluoroacetoxyborohydride 5, along with
hydrogen generation. 5 is then converted into sodium bis-
trifluoroacetoxyborohydride 6, which is a reducing agent
exhibiting less reactivity with 1 than 5. Finally, tetrakis
trifluoroacetoxy borate 7 is generated. However, the reaction
reaction was not completed when NaBH and TFA were mixed
4
before 1 was added. (2) Solid NaBH addition was an unsafe
4
operation.
is not completed when NaBH and TFA are mixed before the
4
addition of 1, indicating that only 5 can react with 1 and
generate borane, which can further react with 1. Therefore, we
conclude that adding TFA is a reasonable sequence that would
result in an appropriate reaction condition. The active species
are generated and 1 is reduced simultaneously when TFA is
added. Adding TFA avoids the deactivation of active species
involved in the reduction process (Scheme 1). We tried to
evaluate the safety of this reduction process involving TFA
addition.
Reaction Calorimetry and Gas Evolution Analyses. As
already discussed, adding TFA is important for controlling the
active species involved in the reduction process. Evaluating the
safety during addition of TFA therefore means evaluating the
safety of the reduction process. Recently Veedhi reported
reaction calorimetry and thermal behavior of the reaction of
Figure 1. 11B NMR spectra of NaBH4 combined with TFA. (a)
NaBH , (b) BH ·THF, (c) NaBH and TFA (0.5 equiv for NaBH ),
4
3
4
4
and (d) NaBH and TFA (1 equiv for NaBH ).
4
4
14
NaBH and TFA. Controlled addition of TFA leads the
4
The active species involved in the reduction reaction were
increasing the thermal stability of the reaction mixture. The
reaction calorimetry of this reduction reaction was conducted
based on this knowledgement using RC1e reaction calorimeter.
The heat released during the reaction was determined to be 521
kJ/mol of 1 (Figure 3). The specific heat capacity was
determined experimentally, and an adiabatic temperature rise of
11
11
analyzed using B NMR spectroscopy. The B NMR chemical
shift values allowed us to estimate the structure of the boron
1
3
11
species. In Figure 1a and b, B NMR spectra of NaBH and
4
BH ·THF each are shown. And when 0.5 and 1.0 equiv of TFA
3
for NaBH was added to NaBH in THF without 1, both the
4
4
Figure 2. 11B NMR spectra of the lactam reduction mixtures. (a) 4, (b) NaBH , (c) BH ·THF, (d) 1, NaBH (3.5 equiv for 1) and TFA (1.75 equiv
4
3
4
for 1), and (e) 1, NaBH (3.5 equiv for 1) and TFA (3.5 equiv for 1).
4
B
DOI: 10.1021/op500284e
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. Proposed reaction mechanism for lactam reduction of 1
The specific heat capacity was determined experimentally, and
an adiabatic temperature rise of 51 K was estimated to occur
during the reaction. Comparing the data with 1 and without 1,
about half of heat reaction of this reduction reaction is caused
by the reaction of NaBH and TFA. This heat output ratio was
4
also in good agreement with the TFA addition ratio. These data
are similar to the results of Veedhi. These results indicate that
TFA addition could control the heat output during the reaction.
Gas evolved when TFA was added; therefore, gas evolution
analysis was conducted. The gas evolution ratio was in good
agreement with the TFA addition ratio (Figure 5). The total
Figure 3. Reaction calorimetry of lactam reduction by NaBH and
4
TFA. Conditions: TFA (3.5 equiv for 1) was added to a mixture of 1
(
0.35 M) and NaBH (3.5 equiv for 1) in THF (10 v/w) at 15 °C
4
(
time for dropwise addition of TFA was 60 min). The conversions
measured by reaction calorimetry and HPLC are compared.
9
3 K was estimated to occur during the reaction. The heat
output ratio was in good agreement with the TFA addition
ratio, and the reaction and heat conversions were almost the
same (78.8% and 73.8%, respectively) at the end of the TFA
addition process. The maximum temperature of synthesis
reaction (MTSR) of this reaction was 34 °C.
The heat released during the reaction between NaBH and
4
Figure 5. Gas evolution during the lactam reduction by NaBH and
TFA. Conditions: TFA (3.5 equiv) was added to a mixture of 1 (0.34
4
TFA without 1 was found to be 263 kJ/mol of 1 (Figure 4).
M) and NaBH (3.5 equiv) in THF (10 v/w) at 15 °C (time for
4
dropwise addition of TFA was 60 min). TFA addition ratio and gas
generation ratio for total amount of off-gas are also compared.
amount of evolved gas corresponded to 3.5 equiv of hydrogen,
which is equivalent to the amount of added TFA. This result
shows that TFA addition could control the amount of gas
evolved.
Pilot Manufacturing. The generation of diborane in the
reaction mixture is described in the reaction mechanism
section. Results of the elemental analysis of the evolved gas
showed that diborane gas was released along with hydrogen.
Diborane gas has to be decomposed to boric acid before it can
be purged into the atmosphere because of its toxicity. A
maximum concentration of diborane gas of 10 ppb is stipulated
1
5
in the appropriate guidelines. To determine the most
appropriate scrubber system, the abilities of different scrubber
solutions were investigated using the laboratory equipment
(Figure 6). When the TFA addition time was set as 1 h and a 1
Figure 4. Reaction calorimetry of the reaction between NaBH and
4
TFA without 1. Conditions: TFA (1.0 equiv for NaBH ) was added to
4
NaBH in THF at 15 °C (time for dropwise addition of TFA was 60
4
min).
C
DOI: 10.1021/op500284e
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
to prevent the release of diborane gas allowed the process to be
successfully scaled up to a 35 kg pilot plant scale.
EXPERIMENTAL SECTION
■
General. All reactions were performed under a nitrogen
atmosphere. Solvents and reagents were used without any
1
13
purification or drying. H and C NMR spectra were acquired
on a Bruker or JEOL spectrometer at frequencies of 400 and
1
00 MHz, respectively. Mass spectra were recorded on Waters
ZQ-2000. HPLC chromatograms were recorded on Shimadzu
LC-20. Reaction calorimetry was performed using Mettler
Toledo MidTemp RC1e vessel. Gas evolution study was
performed using SHINAGAWA wet gas meter WS type.
(
2S)-2-(4-Fluorophenyl)-4-benzylmorpholine (2). (6S)-
Figure 6. Testing of scrubber systems for diborane gas removal.
6
1
-(4-fluorophenyl)-4-benzylmorpholine-3-one (1) (30.0 g,
05.2 mmol) in THF (180 mL) solution was added to a
stirred suspension of NaBH (13.9 g, 368.2 mmol) in THF
4
N aqueous NaOH solution scrubber system was used, the
concentration of the released diborane gas was observed to be
ca. 500 ppb (entry 1). When the TFA addition time was set as 2
h and a 15% aqueous NaOH solution scrubber system was
used, the concentration of the released diborane gas decreased
to ca. 70 ppb (entry 2). Diborane gas was not detected when a
(
120 mL) at 0 °C. Then TFA (42.0 g, 368.2 mmol) in THF
(
120 mL) was added over 2 h maintaining the internal
temperature below 10 °C, followed by stirring for 5 h at 20 °C.
Acetone (42.8 g, 736.4 mmol) and 17% HCl aqueous solution
(
186.12 g, 841.2 mmol) were added to the reaction mixture,
followed by stirring for 3 h at 60 °C. The reaction mixture was
cooled to room temperature and 25% NaOH aqueous solution
1
5% aqueous NaOH solution double scrubber system was used
to quench diborane gas (entry 3). We conclude that a double
scrubber system similar to that employed in the laboratory
experiment would be necessary in the pilot plant.
(
135.0 g, 841.2 mmol) was added, followed by stirring for 30
min. Once settled the upper organic layer was separated and
the aqueous layer was extracted with THF (60 mL). The
combined organic phases were washed with 10% NaCl aqueous
solution (150 mL). The organic layer was evaporated to 90 mL
total volume. Ethanol (600 mL) was added, and the organic
layer was again evaporated to a 90 mL total volume. Ethanol
The addition of TFA was controlled in the pilot plant using a
reducer pipe. A double scrubber system using 15% aqueous
NaOH solution was used. Furthermore, the evolved gas was
diluted with steam before releasing it in order to avoid static
ignition (Figure 7). The scaled up pilot plant allowed 35 kg of 1
to be successfully reduced without releasing diborane gas. The
yield of 2 corresponded with that achieved in the laboratory-
scale experiments. The plant temperature profile data revealed
that the conditions employed were successful in maintaining
the process temperature within the desired range.
(
150 mL) and water (270 mL) were added at 30 °C, and the
resulting slurry was cooled to 10 °C, aged for 1 h. The slurry
was filtered and washed with the mixture of ethanol (15 mL)
and water (75 mL) and dried under reduced pressure to yield
1
24.3 g of 2 (89.4 mmol, isolated yield: 85%). H NMR
(
DMSO-d , 400 MHz) δ 7.34−7.24 (m, 7H), 7.03−6.97 (m,
6
CONCLUSIONS
We have reported the approach for the safety evaluation based
on mechanistic insight into lactam reduction using NaBH and
TFA. The mechanistic study showed that active species were
generated during the addition of TFA, which avoided
deactivation of the active species involved in the reduction
process. Based on this reaction mechanism, the safety
assessments for the process were performed using reaction
calorimetry and gas evolution analyses. Use of safety equipment
2H), 4.54 (dd, J = 10.4, 2.2 Hz, 1 H), 4.00 (ddd, J = 11.5, 3.5,
.5 Hz, 1H), 3.82 (td, J = 11.5, 2.4 Hz, 1 H), 3.53 (s, 2 H), 2.87
(dt, J = 11.5, 2.2 Hz, 1 H), 2.73−2.76 (m, 1 H), 2.27 (td, J =
■
1
4
1
3
11.5, 3.5 Hz, 1 H), 2.07 (dd, J = 11.5, 10.4 Hz, 1 H); C NMR
(100 MHz, DMSO-d ) δ 163.5, 161.0, 137.5, 136.3, 136.3,
129.2, 128.3, 127.9, 127.9, 127.2, 115.2, 115.0, 77.5, 67.1, 63.2,
60.4, 52.8; ESI MS: Calcd for C H FNO [M + H] 272,
Found 272; Anal. Calcd for C H FNO: C, 75.25; H, 6.69; F,
6
+
17
18
17
18
7.00; N, 5.16. Found: C, 75.05; H, 6.68; F, 6.92; N, 5.18.
Figure 7. Schematic of the pilot plant, including the scrubber system. (a) TFA addition tank, (b) reducer pipe to control the TFA addition rate, (c)
reactor tank used for the reduction process, (d) scrubber system using 15% aqueous NaOH solution, and (e) dilution of the evolved gas with steam.
D
DOI: 10.1021/op500284e
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
ASSOCIATED CONTENT
■
*
S
Supporting Information
1
13
Copies of H NMR and C NMR of 1, 2, and 3. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
E-mail: Takashi.Naka@md.mt-pharma.co.jp.
■
*
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Hajime Hiramatsu for support with the analyses,
Dr. Masanori Hatsuda and Dr. Masayuki Utsugi for helpful
discussions, Dr. Hiroshi Iwamura, Mr. Yasunari Takayasu, Mr.
Keita Kawabata, Dr. Kazutoshi Watanabe, Dr. Fumiaki Uehara,
Dr. Toshiyuki Kohara, and Mr. Kenji Fukunaga for supports of
process development, and Mr. Takeshi Tamagawa and Mr.
Akihiro Tsukamoto for the pilot manufacturing.
REFERENCES
■
(
1
1) Magano, J.; Dunetz, J. R. Org. Process Res. Dev. 2012, 16, 1156−
184.
(
2) Brown, H. C.; Weissman, P. M.; Yoon, N. M. J. Am. Chem. Soc.
966, 88, 1458−1463.
3) Kornet, M. J.; Thio, P. A.; Tan, S. I. J. Org. Chem. 1968, 33,
637−3639.
4) Mandal, S. B.; Giri, V. S.; Sabeena, M. S.; Pakrashi, S. C. J. Org.
Chem. 1988, 53, 4236−4241.
5) Okuyama, M.; Uehara M.; Iwamura, H.; Watanabe, K. WO 2007/
11065 A2.
6) Hida, T.; Mitsumori, S.; Honma, T.; Hiramatsu, Y.; Hasizume,
1
(
3
(
(
0
(
H.; Okada, T.; Kakinuma, M.; Kawata, K.; Oda, K.; Hasegawa, A.;
Masui, T.; Nogusa, H. Org. Process Res. Dev. 2009, 13, 1413−1418.
(
7) Nettles, S. M.; Matos, K.; Burkhardt, E. R.; Rouda, D. R.; Corella,
J. A. J. Org. Chem. 2002, 67, 2970−2976.
8) Alimardanov, A. R.; Barrila, M. T.; Busch, F. R.; Carey, J. J.;
Couturier, M. A.; Cui, C. Org. Process Res. Dev. 2004, 8, 834−837.
9) (a) Lobben, P. C.; Leung, S. S-W.; Tummala, S. Org. Process Res.
(
(
Dev. 2004, 8, 1072−1075. (b) Shimizu, S.; Osato, H.; Imamura, Y.;
Satou, Y. Org. Process Res. Dev. 2010, 14, 1518−1520.
(
10) Lam, T. T.; Bagner, C.; Tuma, L. Thermochim. Acta 2005, 426,
1
(
2
(
09−113.
11) Giannis, A.; Sandhoff, K. Angew. Chem., Int. Ed. Engl. 1989, 28,
18−220.
12) (a) Yamada, K.; Takeda, M.; Iwakuma, T. J. Chem. Soc., Perkin
Trans. 1 1983, 265−270. (b) Umino, N.; Iwakuma, T.; Itoh, N.
Tetrahedron Lett. 1976, 10, 763−766. (c) Boros, E. E.; Thompson, J.
B.; Katamreddy, S. R.; Carpenter, A. J. J. Org. Chem. 2009, 74, 3587−
3
(
590.
13) Bregadze, V. I.; Lobanova, I. A.; Timofeev, S. V.; Kudinov, A. R.;
Meshcheryakov, V. I.; Tok, O. L.; Petrovskii, P. V.; Starikova, Z. A.
The substitution at boron atom by the CF COO-group shifts its signal
3
to low field for 15−20 ppm in comparison with the signal of the
unsbstituted boron atom. J. Organomet. Chem. 2002, 657, 171−175.
(
14) Veedhi, S.; Sawant, A. J. Therm. Anal. Calorim. 2013, 111,
1
(
093−1097.
15) The Japan Society for Occupational Health (JSOH)
recommends the Occupational Exposure Limits (OELs) as reference
values for preventing adverse health effects on workers caused by
occupational exposure to chemical substances. J. Occup. Health 2012,
5
4, 387−404.
E
DOI: 10.1021/op500284e
Org. Process Res. Dev. XXXX, XXX, XXX−XXX