Process optimization of aldol-type reaction by process understanding using in Situ IR

Article abstract of DOI:10.1021/op300186p

A high-yield robust LHMDS-mediated aldol-type reaction of benzyl maltol (2) and benzaldehyde (3) was developed using in situ IR to overcome the problems of low yield and yield fluctuation of the pilot synthesis. In situ IR studies indicated that unexpecte

Full text of DOI:10.1021/op300186p

Article
pubs.acs.org/OPRD
Process Optimization of Aldol-Type Reaction by Process
Understanding Using in Situ IR
Yuki Fukui,* Shinichi Oda, Hiroyuki Suzuki, Toshikazu Hakogi, Daisuke Yamada, Yohei Takagi,
Yasunori Aoyama, Hideyuki Kitamura, Masayoshi Ogawa, and Junko Kikuchi
Chemical R&D Center, CMC Development Laboratories, Shionogi & Co., Ltd., 1-3, Kuise Terajima 2-chome, Amagasaki, Hyogo
660-0813, Japan
S
* Supporting Information
ABSTRACT: A high-yield robust LHMDS-mediated aldol-type reaction of benzyl maltol (2) and benzaldehyde (3) was
developed using in situ IR to overcome the problems of low yield and yield fluctuation of the pilot synthesis. In situ IR studies
indicated that unexpected side reactions of LHMDS and 3 reduced the yield of the aldol-type reaction. On the basis of the
results, the reaction conditions were optimized.
lithiated intermediate was not carbanion on the methyl group
of 2 but a lithium enolate 5 (Figure1). The 1546 cm−1 of 5
increased in proportion to addition of 1.0 equiv of LHMDS
(Figure 2). Although a slight excess amount of LHMDS is
usually used to complete the lithiation, it was found that just 1.0
equiv of LHMDS was sufficient.
Behavior of Remaining LHMDS and 3. We next
investigated the behavior of an excess amount of LHMDS
using in situ IR. 3 was added to the THF solution of LHMDS
at −70 °C. The absorption at 1654 cm−1 increased in
proportion to the addition of 3. After addition of an equivalent
ratio of 3 for LHMDS. The CO stretching absorption (1700
cm−1) of 3 was detected (Figure 3). The absorption of 1654
cm−1 is the CN stretching of benzaldimine 6,11 which is an
unreactive compound for lithium enolate 5. This means that
the excess LHMDS was consumed in the reaction with the
equivalent amount of 3 (Figure 4). (For further information,
see Supporting Information.)
In large-scale production, however, the stirring condition
differs from that of the small lab scale, and the concentration of
the added 3 might be high in some portions of the reaction
mixture. If the reaction does not smoothly proceed for some
unforeseen reason, the reaction of 3 and LHMDS might give a
result different from that of Figure 4. To evaluate the possibility
of this, we changed the addition order of LHMDS and 3 and
observed the behavior of 3. A THF solution of LHMDS was
added to the THF solution of 3 at −70 °C. The CO
stretching absorption (1700 cm−1) of 3 decreased until half the
amount of LHMDS for 3 was added, and then completely
disappeared. To our surprise, no CN stretching absorption
(1654 cm−1) was detected (Figure 5). The reaction mixture of
enolate 5 was added to the reaction mixture, but 4 was hardly
detected and 2 and 3 were recovered, which was confirmed by
HPLC, in spite of no detection of carbonyl absorption of 3 by
in situ IR. These results indicate that one molecule of LHMDS
consumes two molecules of 3. An unreactive complex which
INTRODUCTION
■
The lithiation-mediated aldol type reaction is well-known and
useful for formation of the C−C bond to obtain the
corresponding alcohol derivatives.1 A key intermediate in the
synthesis of active pharmaceutical ingredients (APIs) is 3-
(benzyloxy)-2-(2-hydroxy-2-phenylethyl)-4H-pyran-4-one (4).2
The synthetic process of 4 consists of protection of the
hydroxyl group of maltol (1), lithiation of benzyl maltol (2)3
with lithium hexamethyl disilazide (LHMDS) in THF at −70
°C and a nucleophilic reaction with 3 (Scheme 1). Excess
LHMDS is generally used to complete the lithiation.4 In
accordance with the procedures, we used 1.2 equiv of LHMDS
for preparation of 4 and conducted 100-kg-scale pilot
production on 20 batches. In spite of the simple procedure,
the yield of 4 fluctuated (72−97%), and a corresponding
amount of 2 remained. To stabilize the product yield, we
needed to optimize the process conditions. However, this type
of reaction is often carried out under very low temperatures due
to the generation of unstable intermediates, which cannot be
directly detected by using at-line analytical tools such as TLC,
HPLC, and GC. This prevents general optimization of these
processes. What is required is an understanding of what is
occurring within the reaction mixture in real time, which can
offer useful information on the chemical process. Recently,
process understanding by using process analytical technology
(PAT),5 such as IR,6 NIR,7 Raman,8 and particle counters,9
makes it possible for quality by design (QbD).10 In this work,
we used in situ IR, which can directly detect unstable reaction
intermediates in real time, in order to clarify the chemical
events taking place and develop a robust process.
RESULTS AND DISCUSSION
■
Consideration of the Structure of the Lithiated
Intermediate and Determination of the Required
Amount of LHMDS. We first investigated the lithiation of 2
with 1.2 equiv of LHMDS. The CC stretching absorption of
enolate (A 1546 cm−1) and CC stretching absorption of
enolate conjugated exomethylene (B 1602 cm−1) were detected
as peaks of the intermediate. The IR result showed that the
Received: July 12, 2012
Published: October 2, 2012
© 2012 American Chemical Society
1783
dx.doi.org/10.1021/op300186p | Org. Process Res. Dev. 2012, 16, 1783−1786

Organic Process Research & Development
Article
Scheme 1. Representative procedure for 4
Figure 4. Generation of trimethylsilyl benzaldimine (6).
(Figure 6). (For further information, see Supporting
Information.)
In the case where LHMDS was added into the THF solution
of 3, 3 is constantly sufficiently rich for LHMDS. Therefore,
LHMDS immediately reacts with two molecules of 3 with the
formation of a stable six-membered ring. It seems to be the
reason for the absence of affording benzaldimine 6 in this case.
Process Optimization. According to the verification
experiments for the behavior of 3 and LHMDS, complex 7 is
easy to generate in the case of a high concentration of 3. When
1.2 equiv of LHMDS was used for this reaction, the yield of
product 4 might decrease by generation of complex 7 because
of consumption of up to twice the amount of 3 for 0.2 equiv
excess of LHMDS, and subsequent quenching of the remaining
lithiated intermediate 5 affords 2. We assumed that reduction of
LHMDS is the most important factor for reducing the
consumption of 3 by excess LHMDS in order to complete
and stabilize the main reaction. To prove this hypothesis, the
influence of the addition time of 3 and the amount of LHMDS
were evaluated (Table 1). The original procedure was carried
out as a controlled experiment (entry 1). As expected, fast
dropwise addition of 3 (entry 2) which could generate complex
7 gave a lower yield than entry 1, and a corresponding amount
of 2 was detected by HPLC. Next, the experiment with a
reduced amount of LHMDS was carried out (entries 3 and 4).
It gave quantitative yields without fluctuation of the yield as
Figure 1. Structure of lithiated intermediate 5.
Figure 2. Plots of 1546 cm−1 (enolate 5) and LHMDS addition.
can be reversed to 3 by quenching seems to be generated. We
speculated that the 2:1 complex 7, which is stabilized by six-
membered ring formation, was generated under this condition
Figure 3. Observation of addition of 3 to LHMDS.
1784
dx.doi.org/10.1021/op300186p | Org. Process Res. Dev. 2012, 16, 1783−1786

Organic Process Research & Development
Article
Figure 5. Observation of addition of LHMDS to 3.
MHz Varian FT spectrometer. HRMS were recorded on a
LTQ-Orbitrap mass spectrometer. All in situ infrared spectra
were collected using a Mettler Toledo ReactIR iC10 equipped
with a 6 mm diameter Diamond attenuated total reflection
(ATR) probe (DiComp).
Preparation of Benzyl Maltol (2). Potassium carbonate
(131.5 g, 952 mmol) was portionwise added into the stirring
surry of maltol (1) (100 g, 793 mmol) in acetonitrile (1400
mL) at room temperature. Then benzyl bromide (142.4 g, 833
mmol) was added into the mixturre. The mixture was heated to
70 °C, followed by stirring for 7 h. The reaction mixture was
cooled to room temperature and evaporated to 500 mL.
Toluene (500 mL) was added and washed with water (500
mL), followed by washing with 3%NaCl aq (500 mL). The
organic layer was evaporated to 200 g. The obtained toluene
solution of benzyl maltol (2) was used for the next step. This
compound is already known, and the 1H NMR data
corresponded to the literature data3 1H NMR (500 MHz,
DMSO-d6) δ 8.04 (1H, d, J = 5.6 Hz), 7.42−7.33 (5H, m), 6.37
(1H, d, J = 5.6 Hz) 5.03 (2H, s), 2.12 (3H, s); 13C NMR (500
MHz, DMSO-d6) δ174.9, 160.19, 155.9, 144.1, 137.9, 129.6,
129.3, 129.1, 117.4, 73.6, 15.4 ; HRMS (ESI) m/z calcd for
C13H13O3 [M]+: 217.0859, found 217.0851.
Figure 6. Generation and plausible structure of the 2:1 complex 7.
Table 1. Evaluation of addition time of 3 and amount of
LHMDS
LHMDS
(equiv)
3
addition time of 3
yield of 4
a
entry
(equiv)
(min)
(%)
1
2
3
4
1.20
1.20
1.05
1.05
1.2
1.2
1.2
1.2
60
10
60
10
92
88
>99
>99
a
HPLC yield [4/(2 + 4)] × 100%].
Preparation of 3-(Benzyloxy)-2-(2-hydroxy-2-phenyl-
ethyl)-4H-pyran-4-one (4). The solution of 2 (10 g
containing 8.57 g of 2, 39.6 mmol) was diluted with THF
(68 mL) and cooled to −70 °C. 20% lithium hexamethyl
disilazide (39.81 g, 47.6 mmol) was added dropwise for 1 h,
followed by stirring for 1 h at −70 °C. Benzaldehyde (5.05 g,
47.6 mmol) was added dropwise for 1 h, followed by stirring for
2 h under the same reaction conditions. The reaction mixture
was added to 20% H2SO4 aq, keeping the temperature below 10
°C. The aqueous layer was separated and washed with 3% NaCl
aq. The organic layer was evaporated. Toluene (25 mL) was
added to obtain crystals of 4, and then the slurry was
evaporated to 30 mL. The slurry was cooled to 5 °C and
filtrated. The filtrate was washed with toluene (25 mL) and
dried under reduced pressure to yield 10.74 g of 4 (33.3 mmol,
isolated yield: 84%). 1H NMR (500 MHz, CDCl3) δ 8.08 (1H,
d, J = 5.6 Hz), 7.42−7.22 (10H, m), 6.38 (1H, d, J = 5.6 Hz),
5.62 (1H, d, J = 4.4 Hz), 4.95 (1H, d, J = 11.0 Hz), 4.88 (1H,
m) 4.78 (1H, d, J = 11.0 Hz), 2.94 (1H, dd, J = 14.0, 8.5 Hz),
2.82 (1H, dd, J = 14.0, 8.5 Hz); 13C NMR (500 MHz, DMSO-
d6) δ175.1, 160.5, 156.1, 145.6, 145.4, 138.1, 129.3, 129.2,
with entries 1 and 2 because the consumption of 3 decreased
due to the lower amount of LHMDS. In general, an excess
amount of reagent is used to complete the reaction, but this is
not always optimal, as we have shown here. On the basis of our
findings, we were able to establish a high-yield, robust, and cost-
reduced process.
CONCLUSION
■
In situ IR was effectively applied to the aldol-type reaction
process to understand the chemical events in detail. Depend-
ence of the reaction on relative concentrations of LHMDS and
3 were observed with in situ IR. On the basis of the findings, a
high-yield, robust, and cost-reduced process was established.
EXPERIMENTAL SECTION
■
General. All reactions were run under a nitrogen
atmosphere. Solvents and reagents were obtained from
commercial sources and used without further purification.
NMR spectra were measured on a 500 MHz Inova-500 or 300
1785
dx.doi.org/10.1021/op300186p | Org. Process Res. Dev. 2012, 16, 1783−1786

Organic Process Research & Development
Article
(11) (a) Mauro, P.; Paola, Z. Org. Process. Res. Dev. 1998, 2, 49−59.
(b) Ernest, W. C.; Daniel, M.; Mark, J. N. Tetrahedron 1988, 44,
4157−4172. (c) David, J. H.; Ken-ichi, K.; Dudley, G.; Thomas; Teng-
Kuei, Y. J. Org. Chem. 1983, 48, 289−294. (d) Veeraraghavan, R.;
Thomas, E. B. Chem.Eur. J. 2005, 11, 4387−4395. (e) Yoshio, H.;
Jun, K.; Takeo, H; Akira, T.; Kiyoko, I.; Yoshimi, T.; Michiko, M.;
Hiroshi, M; Tohru, A.; Takeo, O.; Yoshimi, S.; Emiko, Y.; M.; Isao, U.;
Iwao, O. J. Med. Chem. 1998, 41, 2345−2360.
129.1, 129.0, 128.1, 126.5, 117.4, 73.7, 71.0, 39.6 ; HRMS (ESI)
m/z calcd for C20H19O4 [M]+: 323.1278, found 323.1274.
ASSOCIATED CONTENT
■
S
* Supporting Information
1
Copies of H NMR and 13C HMR of 2 and 4. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*yuuki.fukui@shionogi.co.jp.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Yukiko Kan, Ms. Tomoko Sasaki, and Ms. Naomi
Umesako for support with the analyses, and Dr. Takemasa Hida
for helpful discussion.
REFERENCES
■
(1) (a) Timothy, B. D.; Nicolas, B.; Brad, M.; Savall; Noel, A. P.;
William, R. R. J. Am. Chem. Soc. 2004, 126, 9307−9317. (b) William,
R. R.; Thomas, D. B.; Michael, D. W.; Jill, A. J.; Karl, A. S. J. Org. Chem.
2002, 67, 4275−4283. (c) Xuemei, C.; Andrew, J. W.; Raymond, J. H.;
David, F. W. J. Org. Chem. 2002, 67, 9331−9339. (d) Darren, J. D.;
Steven, V. L.; Alessandra, P.; Tom, S. Org. Lett. 2001, 3, 3749−3752.
(e) Paul, M. B.; Jared, T. S.; Woerpel, K. A. J. Org. Chem. 1997, 62,
5674−5675.
(2) (a) Johns, B. A.; Wheatherhead, J. G.. WO/2010/011818 A1.
(b) Aoyama, Y. WO/2010/067176 A1. (c) Yoshida, H. WO/2010/
068253 A1.
(3) (a) Karine, B.; Robert, C. H.; Jan, B.; Johan, N.; Eric, D. C.;
Michel, C. Bioorg. Med. Chem. Lett. 2003, 13, 4371−4374. (b) Willam,
O. N.; Timothy, B. K.; Steven, J. R.; Chris, O. Can. J. Chem. 1988, 60,
123−131.
(4) (a) Brian, H.; Kenneth, L.; Kirk. J. Org. Chem. 2001, 66, 4892−
4897. (b) Paul, M.; Jared, T.; Shaw, K. A.; Woerpel. J. Org. Chem.
1997, 62, 5674−5657. (c) Darren, J. D.; Steven, V. L.; Alessandra,
Polara; Tom, S. Org. Lett. 2001, 3, 3749−3752. (d) Yousuke, Y.;
Hisashi, Y. J. Am. Chem. Soc. 2010, 132, 5354−5356. (e) Fujioka, H.;
Ohba, Y.; Nakahara, K.; Takatsuji, M.; Murai, K.; Ito, M.; Kita, Y. Org.
Lett. 2007, 9, 5605−5608. (f) Ian, L. J.; Felicity, K. M.; Christina, L. L.
C. Org. Lett. 2009, 11, 5526−5529. (g) David, J. M.; Samuel, J. D.
Angew. Chem., Int. Ed. 2007, 46, 7789−7792.
(5) Melissa, B.; Steven, J. F.; Paul, D. H.; Neil, M.; Ivan, M. Org.
Process Res. Dev. 2005, 9, 360−364.
(6) (a) Ana, C. B. S.; Ryan, C.; Thomas, W.; James, P. S.; William, E.;
Thomas, C. Org. Process Res. Dev. 2011, 15, 1458−1463. (b) George,
Z.; Khateeta, E.; Emily, M.; Camille, A.; Osama, S. Org. Process. Res.
Dev. 2012, 16, 204−213.
(7) Michael, P.; Zhihao, L.; Yongkui, S. Org. Process. Res. Dev. 2005, 9,
141−148.
(8) Hongxun, H.; Mark, B.; Yun, H.; Weiyi, S.; Steven, F.; Barbara,
W.; Brian, G. Org. Process Res. Dev. 2012, 16, 35−41.
(9) Christelle, H.; Benoît, H.; Selim, D.; Aurelie, L.; Valerie, V.; Tom,
L. Org. Process Res. Dev. 2012, 16, 49−56.
(10) (a) Vince, M.; Mary, T. E.; Frank, R. B.; Jason, M.; Peter, R.;
Mark, R. B. Pharm. Eng. 2010, 30, 1−16. (b) Kuroda, H. Pharm. Tech.
Jpn. 2009, 25, 17−24. (c) ICH guidelines, Q8(R2) and Q9, http://
www.ich.org/: ICH Q8 Pharmaceutical Development, (R2); U.S.
Department of Health and Human Services, Food and Drug
Administration, Center for Drug Evaluation and Research (CDER):
Rockville, MD, Aug 2009. (d) ICH Q9 Quality Risk Management; U.S.
Department of Health and Human Services, Food and Drug
Administration,Center for Drug Evaluation and Research (CDER):
Rockville, MD, June 2006.
1786
dx.doi.org/10.1021/op300186p | Org. Process Res. Dev. 2012, 16, 1783−1786