Process Research and Impurity Control Strategy of Esketamine

Article abstract of DOI:10.1021/acs.oprd.9b00553

An improved synthesis of (S)-ketamine (esketamine) has been developed, which was cost-effective, and the undesired isomer could be recovered by racemization. Critical process parameters of each step were identified as well as the process-related impurities. The formation mechanisms and control strategies of most impurities were first discussed. Moreover, the (S)-ketamine tartrate is a dihydrate, which was disclosed for the first time. The practicable racemization catalyzed by aluminum chloride was carried out in quantitative yield with 99% purity. The ICH-grade quality (S)-ketamine hydrochloride was obtained in 51.1% overall yield (14.0% without racemization) by chiral resolution with three times recycling of the mother liquors. The robust process of esketamine could be industrially scalable.

Full text of DOI:10.1021/acs.oprd.9b00553

Subscriber access provided by NATIONAL UNIV OF SINGAPORE
Full Paper
Process Research and Impurity Control Strategy of Esketamine
Shenghua Gao, Xuezhi Gao, Zhezhou Yang, and Fu-Li Zhang
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00553 • Publication Date (Web): 18 Mar 2020
Downloaded from pubs.acs.org on March 18, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Process Research and Impurity Control
Strategy of Esketamine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Shenghua Gao, † Xuezhi Gao, † Zhezhou Yang,*, † and Fuli Zhang*, †
†Shanghai Institute of Pharmaceutical Industry, China State Institute of Pharmaceutical
Industry, 285 Gebaini Road, Pudong District, Shanghai 201203, China
Abstract
An improved synthesis of (S)-ketamine (Esketamine) has been developed, which was
cost-effective and the undesired isomer could be recovered by racemization. Critical
process parameters of each step were identified as well as the process-related impurities.
The formation mechanisms and control strategies of most impurities were first
discussed. Moreover, the (S)-ketamine tartrate is a dihydrate, which was disclosed for
the first time. The practicable racemization catalyzed by aluminum chloride was carried
out in quantitative yield with 99% purity. The ICH-grade quality (S)-ketamine
hydrochloride was obtained in 51.1% overall yield (14.0% without racemization) by
chiral resolution with three times recycling of the mother liquors. The robust process
of Esketamine could be industrially scalable.
Keywords
Esketamine; Major Depressive Disorder; Process; Racemization; Impurity
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 45
1
2
3
4
5
Introduction
6
7
8
9
Major depressive disorder has become a common and major public health
challenge, and approximately 16% of the world population suffer from depression at
some point in their lives1, 2. However, the current antidepressants, usually several weeks
to months, are necessary to effectively alleviate the clinical symptoms2. This significant
time lag and unsatisfactory therapeutic efficacy limit the application prospect of current
antidepressants. As a result, it is urgent to develop antidepressants with better
pharmacodynamics and pharmacokinetics3, 4.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(2S)-(2-chlorophenyl)-2-(methylamino)-cyclohexanone ((S)-1, (S)-ketamine,
Esketamine), typically assumed to N-methyl-D-aspartate (NMDA) receptor antagonist,
has been approved as a nasal spray formulation for the treatment of major depressive
disorder on March 5th, 2019 by FDA. (S)-ketamine has higher potency or affinity for
the NMDA reception and thus potentially allowing for lower dosages than its
enantiomer (R)-ketamine.5
(S)-Ketamine is a 3-4 fold more potent inhibitor of the NMDA receptor, which
promoted the development of its asymmetric synthesis strategies. For instance, Kiyooka
6
and colleagues reported a procedure for the asymmetric synthesis of (S)-Ketamine
with an overall yield of 21% (99% e.e.) in ten steps; however, the synthetic route has
low atom economy and high energy consumption. Another catalytic asymmetric route
with similar idea was reported recently by Chen and colleagues7. The procedure, based
2
ACS Paragon Plus Environment

Page 3 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
on catalytic enantioselective transfer hydrogenation of a cyclic enone and [3, 3]-
sigmatropic rearrangement of an allylic cyanate to isocyanate, afforded Esketamine (99%
8
9
8
e.e) in 50% overall yield. A concise route designed by Toste , starting from o-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
chlorophenyl cyclohexanone, using direct asymmetric electrophilic amination with di-
tert-butyl azodicarboxylates to install the desired quaternary stereocenter, was
conducted in three steps with 30% yield (99% e.e.). The key step was carried out at
40 °C for 50-60 h and relied on the complex chiral organophosphorus ligands which
have not yet been industrialized. More recently, Gohari9 utilized Ellman’s (S)-tert-
butanesulfinamide as a chiral auxiliary to form the stereocenter following Grignard
reagent as nucleophile at -78 °C, and (S)-1 was synthesized under ozonolysis condition
and gave the unsatisfactory 75% ee.
Although several asymmetric syntheses of Esketamine have been reported, an
efficient and scalable synthesis to enantiopure Esketamine was desired. Classical
resolution10 of the rac-1 could be a better choice for the industrialization of (S)-1 if the
unwanted (R)-1 isomer lost in the mother liquor could be racemized. Recently, another
economical and environmental process for the preparation of racemic ketamine under
sustainable continuous flow conditions was reported by Monbaliu. However, it should
be emphasized that the final step was carried out on Montmorillonite K10 with a high
pressure (35 bar) and high temperature (180 °C), and the conversion of was the
unsatisfactory 57-75%, leading to a low yield 62%11b. Considering the productivity and
cost of three procedures 11-12, the general route reported by Stevens and colleagues12
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 45
1
2
3
4
5
6
7
8
9
was more suitable for further investigation of scale-up (Scheme 1).The process
involved a bromination-imination-thermal rearrangement sequence and was adopted
for the preparation of ketamine analogs by the Janssen Pharmaceutical Company in the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13
laboratory . However, the bromination step was typically poor atom economy with
the using of N-bromobutanimide14, bromine15 or copper bromide13. Subsequently, the
corresponding intermediate 5 was next subjected to further imination which usually
carried out with excessive liquid methylamine at low temperature -40 °C. During the
thermal rearrangement step, α-hydroxyimine 6 hydrochloride was heated to 170-180 o
C in decalin. Herein, a modification of the general route for ketamine, followed by a
chiral resolution, was developed with the aim of industrially scalable and raising the
product quality.
Scheme 1. The General Route of (rac)-1
Results and Discussion
We herein report an improved synthetic route for the preparation of (S)-1 with the
racemization of unwanted (R)-1, as shown in Scheme 2.
4
ACS Paragon Plus Environment

Page 5 of 45
Organic Process Research & Development
1
2
3
4
5
Scheme 2. Improved Synthetic Route to Esketamine (S)-1a
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a Reagents and conditions: (a) (1) AlCl3, CH2Cl2，−10°C, (2) AlCl3, cyclopentane, 40 °C,
3 h, vacuum distillation,70.3%; (b) 48% HBr, 30% H2O2, 70-80 °C, 3 h, 96.8%; (c) (1)
15-20% CH3NH2 aqueous solution, 15-20 °C, 48 h, 90%; (2) 2 N EA-HCl; (d)
PhCOOEt, AlCl3, 125 °C, 5 h, 93.6%; (e) L-(+)-tartaric acid, acetone/H2O, 41%; (f) (1)
NaOH,; 2 N EA-HCl; (2) crystallized with acetone /H2O, 80%; (g) 1N NaOH (aq); (h)
(1) 2 N EA-HCl in EA; (2) 10 mol % AlCl3; PhCOOEt, 150 °C, 24 h, 99%.
Preparation of (2-chlorophenyl) cyclopentyl ketone (4)
It was reported that 4 was synthesized via Grignard reaction using o-
chlorobenzenitrile and cyclopentyl magnesium bromide as raw materials. However, the
anhydrous and anaerobic conditions are always required for Grignard reaction, which
necessitate rigorous equipment and operation controls. Moreover, the relatively high
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 45
1
2
3
4
5
6
7
8
9
price of bromocyclopentane (14.3 K U.S. dollars per ton in China) and large dosage
(>1.5 equiv) in the reaction increase the cost. Furthermore, the yield of Grignard
reaction was not high (68%) 14. All above these prompted us to explore other processes
to synthesize 4. The Nenitzescu Reductive Acylation is one of the important methods
to prepare ketones with mild reaction conditions, and the raw materials used are cheaper
than bromocyclopentane. Therefore, the cheaper o-chlorobenzoyl chloride and
cyclopentene were used to replace Grignard reagent as raw materials to prepare 4 via
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nenitzescu Reductive Acylation reaction 16-17
.
In the general procedure for the reductive acylation reaction, two moles of
aluminum chloride (AlCl3) were added to a mixture of the olefin and acyl chloride in
cyclohexane solvent at about -l0 °C, and upon warming slowly to 70 °C, HCl was
evolved and the saturated ketone was obtained 16-17. The probable mechanism for the
preparation of 4 was shown in Scheme 3. The reaction process was divided into two
stages, and the intermediates 14, 15, 17 were obtained via Friedel Crafts acylation in
the first stage. When acyl chloride 2 was completely converted to the intermediate states,
another batch of AlCl3 was added, and 14, 15, 17 returned to carbocation 13 or 16
through Friedel crafts alkylation, and then the product 4 was obtained after the
reduction of saturated alkane. The intermediates 14, 15, 17 were separated from the
reaction mixture and structures characterized by HRMS and NMR spectra, which are
shown in the Supporting Information.
6
ACS Paragon Plus Environment

Page 7 of 45
Organic Process Research & Development
1
2
3
4
5
Scheme 3 the Probable Mechanism for the Preparation of 4
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Under the above conditions, a larger impurity 19 was detected at approximately
21.2% on LC-MS, respectively. According to the corresponding molecular weight
222.1, the impurity was generated as a result of the competitive reaction between
cyclopentene and the byproduct cyclohexene produced from the dehydrogenation of
cyclohexane in the reaction mixture. The formation mechanism of 19 was proposed as
show in Scheme 4. Therefore, the cyclopentane was a good substitute for cyclohexane
to avoid the generation of impurity 19. Moreover, dichloromethane was used to replace
cyclopentane as solvent to improve the reaction and reduce the amount of AlCl3 due to
the better solubility for AlCl3.
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 45
1
2
3
4
5
Scheme 4. The Proposed Formation Mechanism of Impurity 19
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Another two impurities were also detected, of which molecular weights detected
by LC-MS were both 276.1, just one more cyclopentane than that of 4. The two
impurities were presumed to be 20 and 21, and the formation mechanisms were
proposed as show in Scheme 5. The two compounds with indefinite structures were
obtained by the preparative chromatography and then characterized by 1D and 2D
NMR spectra. However, the structures were confirmed as compounds 20 and 22 rather
than 21. According to the results, we speculate that the carbocation 16 was more stable
due to the γ position being less affected by the electronegativity of carbonyl group. The
stability of reactive intermediates 13 and 16 was supported by Density Functional
Theory (DFT) calculations and the Gibbs free energy was 0.0 kcal/mol and - 7.4
kcal/mol, respectively (Scheme 6). 16 was easily obtained from the rearrangement of
13 through the transition state of TS migration. The free energy profiles of both
carbocation were calculated by the M06 method18 in dichloromethane (DCM). Even if
8
ACS Paragon Plus Environment

Page 9 of 45
Organic Process Research & Development
1
2
3
4
5
two electron withdrawing substituents are attached to the benzene ring, the impurity 22
6
7
was still generated via Friedel-Crafts alkylation.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 5. The Proposed Formation Mechanism of Impurities 20 and 22
Scheme 6. Free-Energy for 13 and 16
9
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 45
1
2
3
4
5
6
7
8
9
In view of the above analysis, the carbocation 16, competing with acetyl cation
12, would react with cyclopentene in the reaction mixture, and 20 was obtained after
further hydrogenation. Undoubtedly, the electrophilicity of 12 is stronger than 16. In
view of the reaction mechanism, the equivalent of cyclopentene was reduced, which
will cause part 12 not to be involved in the acylation, so that the newly generated
cyclopentene would react with unreacted 12 first rather than 16. Moreover, the reaction
time was strictly controlled to 3 h, because of the tendency of impurities 20 and 22 to
increase gradually with the reaction time prolonging (Table 1, entry 4). The results
were summarized in Table 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Nenitzescu Reductive Acylation under Various Conditions a
equivalence ratio
2 : 3
HPLC (%)b
20
Entry
4
22
6.4
2.1
2.3
5.5
1
2
1:1.4
78.3
92.1
63.8
81.9
15.3
1:1.1
5.8
3c
4d
1:0.9
11.2
1:1.1
12.6
10
ACS Paragon Plus Environment

Page 11 of 45
Organic Process Research & Development
1
2
3
4
a
Reaction conditions: AlCl3 (1.1+1.2 equiv) were added in two batches, solvent:
5
6
7
8
9
b
CH2Cl2, reaction time: 3 h, reductant: cyclopentane. Only peak areas (%) of 20, 22
and 4 were integrated; c Peak areas (%) of 2, 4, 20 and 22were integrated, 22.7% of 2
remaining, reaction time was prolonged to 6 h; d the reaction time was prolong to 6 h.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The workup procedure involved adding the reaction mixture slowly to cold water
followed by 10-15% sodium hydroxide aqueous solution to remove AlCl3, and the
brown oily liquid was obtained in the purity 87.2% after the dichloromethane was
concentrated under reduced pressure. In order to improve the purity of the product and
remove the residual AlCl3, 4 was purified to a pale yellow liquid with high purity of
98.7% on HPLC by vacuum rectification (2 mmHg, 116-120 °C fraction). Intermediates
15 and 17 were removed during the vacuum distillation process and the impurities 14,
20 and 22 were all reduced to be less than 0.8% by HPLC analysis.
Preparation of (1-bromocyclopentyl) (2-chlorophenyl) ketone (5).
The bromination step was typically poor atom economy with the using of N-
bromobutanimide (NBS) 14, bromine (Br2) or copper bromide (CuBr2) . The raw
material costs, conversion rate, waste streams and risk factors for various industrial
bromination reagents, are presented in Table 2. After comprehensive comparison, we
finally chose hydrogen bromide (HBr) in combination with hydrogen peroxide (H2O2)
as the bromination reagent with high atomic economy, less environmental pollution and
low cost for bromination reaction. The bromine anion was oxidized by H2O2 and
15
13
11
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 45
1
2
3
4
5
completely transferred to 5. The by-product was only environment-friendly H2O, which
6
7
saved the cost of waste disposal comparing to the other bromination reagents.
8
9
Table 2. The Comparison of Various Bromination Reagents for the preparation of 5 a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Br
Stoichiometric
Waste
stream
Hazard
factor c
Br utilization
rate
Conver-
sion rated
Entry
Cost b
source
reagents required
1
2
Br2(l)
50%
50%
Br2
middle
high
HBr
high
low
15%
99%
CuBr2(s)
CuBr2
CuBr
3
4
NBS(s)
HBr(l)
100%
100%
NBS+p-TsOH
high
low
succinimide
4H2O
low
88%
95%
2HBr+H2O2
middle
2KBr+
5
KBr(s)
100%
low
4H2O+K2SO4
low
89%
H2O2+H2SO4
a Based on the reaction “4” + XBr
“5”; Prices from the latest quotation in
b
China (¥ ton-1): bromine (l) 75,000; CuBr2 125,000; NBS 58000; KBr 36,500; HBr (48
c
d
wt % aq) 20,000; hydrogen peroxide (30 wt % aq) 7,500; Reaction /storage risks;
HPLC purity of the reaction mixture.
In view of the similarity between the oxidative bromination on activated aromatics
reported in the literature19 and the bromination of 4, both 1.0 equivalent HBr and H2O2
was explored. However, the reaction did not reach completion with the H2O2 dripping
slowly, approximately 14.2% ketone remained by HPLC, as shown in Table 3 (entry
12
ACS Paragon Plus Environment

Page 13 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
1). The bromine, which formed inevitably during the reaction process, guided us to
increase the amount of hydrogen bromide. Subsequently, when the equivalent of HBr
and H2O2 were increased to 1.5, the conversion rate was still the unsatisfactory 95%
(Table 3 entry 2). The addition of excessive HBr and prolonging the reaction time will
lead to an increase of impurity 24, whose proposed formation mechanism was shown
in the Scheme 7. Compounds 25 and 14 were detected in the reaction solution by HPLC
with the content of 0.2% and 1.0%. HRMS and NMR spectra of 24 are displayed in the
Supporting Information. The mixing speed and temperature were two major factors
responsible for the incomplete reaction, the faster the stirring speed, the greater the
reaction rate. The best result was obtained under the conditions that maintain the
reaction temperature at 70-80 °C by adjusting the dropping speed of 1.5 equiv hydrogen
bromide at the stirring speed of 400 rpm (Table 3, entry 3).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 7. The Proposed Formation Mechanism of Impurity 24
The workup process involved the addition of saturated sodium sulfite solution to
quench the unreacted hydrogen peroxide, followed by the addition of n-hexane for
13
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 45
1
2
3
4
5
6
7
8
9
extraction. The residual oxidant and hydrogen bromide were removed from n-heptane
after washing three times with saturated sodium sulfite solution followed by
concentration of the organic phase. The brown liquid 5 was obtained in 98.6% purity
with the yield of 96.8%. The minor quantities of 14, 24 and the residual 4 were removed
in the next step. The hydrolytic impurity of 25 was detected by HPLC to be less than
0.4%, and it could be convert to the next intermediate α-hydroxyimine 6.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. Screening of Reaction Stoichiometry for the Bromination Reaction a
equivalence ratio
4:HBr:H2O2
1:1.0:1.0
HPLC b
Entry
t(h)
4
5
14
24
1
2
20
20
3
14.2%
84.2%
92.9%
98.7%
1.1%
0.5%
1:1.5:1.5
1:1.5:1.5
5.2%
0.5%
1.1%
0.6%
0.8%
0.2%
3 c
a Reaction temperature: 10 °C ∼ 60 °C, solvent-free, 200 rpm, H2O2 dripping slowly. b
Only peak areas (%) of 4, 5, 14 and 24 were integrated on HPLC. c 400 rpm, temperature:
70 °C ∼ 80 °C
14
ACS Paragon Plus Environment

Page 15 of 45
Organic Process Research & Development
1
2
3
4
5
Preparation of α-Hydroxyimine (6)
6
7
8
9
The α-bromo ketone 5 in the treatment with excessive liquid methylamine at low
temperature -40 °C afforded the α-hydroxyimine 6 12f, 13. Following the above
conditions, a more abundant impurities with the molecular weight of 237.7 was detected
at a level of approximate 30% on LC-MS (Table 4, entry 1). The impurity structure
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
was confirmed as amino ketone 27 by HRMS, 1D NMR and 2D NMR. The reaction
12f
mechanism reported by Stevens
gave us the direction of the optimization. The
methylamination conducted in liquid methylamine containing a molar equivalent of O18
enriched water, and the α-hydroxyimine 6 demonstrated no measurable incorporation
of the O18 label. The evidence supported the intramolecular nature of the oxygen
migration through the epoxyamine 28 as an intermediate (Scheme 8, path A). In other
words, water did not play any role in the reaction. Considering that 40% methylamine
is more convenient to store and commercially available, liquid methylamine was
replaced and the experimental results were listed in Table 4. Initially, the α-bromo
ketone 5 reacted with 40% methylamine aqueous solution in THF at -20 °C for 24 h,
and the conversion was less than 40% presumably due to the low temperature (Table
4, entry 2). Subsequently, the reaction was completed when temperature increased to
25 °C (Table 4, entry 3). However, two impurities were generated, one of which was
the substituted impurity 27 at a level of approximate 25.3% on HPLC, and the other
one was 25 with the content of 10.8%, caused by the hydrolysis of bromide in alkaline
aqueous solution. Moreover, the hydrolytic impurity 25 showed a trend of increasing at
15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 45
1
2
3
4
5
6
7
8
9
early timepoints and then decreasing, due to the nucleophilic addition of the
methylamine to the carbonyl group despite the presence of water. This is another path
to get the target product 6 (Scheme 8 path B). From the above study, the reaction
mechanism of methylamination with methylamine aqueous solution was put forward in
Scheme 8.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 8. Two Paths of Methylamination in Methylamine Aqueous Solution (Curved
arrow only for Path A)
16
ACS Paragon Plus Environment

Page 17 of 45
Organic Process Research & Development
1
2
3
4
5
Table 4. Preparation of 6 under Various Temperature and Solvent a
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HPLC(%) b
Entry
T (°C)
Solvent
5
6
25
0
27
1c
2d
3
-40
-20
25
-
THF
THF
CH2Cl2
-
0.2
60.8
1.1
0.8
1.6
67.2
29.9
62.8
41.7
77.2
32.6
8.4
0.9
10.8
8.6
4.3
25.3
48.9
16.9
4
25
5
25
5 react with excess methylamine, 24 h; b HPLC purity of the reaction mixture, only
a
c
d
peak areas (%) of 5, 6, 25 and 27 were integrated. liquid methylamine; 40%
methylamine aqueous solution for entries 2-5.
According to the above analysis, the next optimization direction was to avoid the
generation of 27, because it could not be converted into 6. The reaction was then carried
out in different solvents. Interestingly, the impurity 27 content was formed at the lowest
levels in solvent-free condition, followed by THF, but formed at the highest levels in
water-insoluble CH2Cl2 (Table 4, entries 3-5). The lipid-soluble 5 and most
17
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 45
1
2
3
4
5
6
7
8
9
methylamine will be dissolved in dichloromethane when the solvent is dichloromethane.
In this case, the reaction mixture becomes oil-water two-phase and methylamine will
react with 5 in the same phase. Compared with water as solvent, more methylamine
will react with 5 in dichloromethane by nucleophilic substitution. It is speculated that
the content of 27 was mainly affected by the concentration of methylamine interacting
with 5. Therefore, we reduced the concentration of methylamine aqueous solution, and
the results showed that the amounts of 27 decreased with the decrease of the
concentration of methylamine (Table 5, entries 1-3). As the temperature affects the
selectivity of nucleophilic substitution, reaction temperature was further investigated.
When the temperature reduced to 15 °C, 27 dropped to approximately 3.8% on HPLC
(Table 5, entry 4). However, if the temperature was adjusted to 5 °C, the reaction time
would be needed to be prolonged to more than 3 days for complete conversion, which
was not conducive to large-scale production (Table 5, entry 5). Moreover, the low
reaction temperature was favorable for blocking the rearrangement of the hydrolytic
impurity 25 to 29. It could be found that the content of 29 decreased as the temperature
was lowered (Table 5, entries 3-5). The generation mechanism of 29 and the summary
of all side reactions involved in methylamination were proposed as show in Scheme 9.
According to the above experimental data, the substituted impurity 27 and the
hydrolytic impurity 25 could be controlled to less than 3.0% under the synergistic action
of the concentration of methylamine and the reaction temperature (Table 5, entry 6).
The optimum reaction conditions for the preparation of α-hydroxyimine 6 were
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
ACS Paragon Plus Environment

Page 19 of 45
Organic Process Research & Development
1
2
3
4
5
determined as follows: 15% methylamine aqueous solution reacted with 5 at 15-20 °C
6
7
for 48 h without solvent.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 9. All the Side Reactions Involved in Methylamination
Table 5. Preparation of 6 with Various Concentration of Methylamine Aqueous
Solution under Various Temperature a.
Concentration
HPLC(%) b
27
Entry
T/t (°C/h)
25/24
of CH3NH2
(%)
6
25
29
14
1
40
76.3
3.1
16.5
3.3
0.8
19
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 45
1
2
3
4
5
6
7
8
9
2
3
4
5
6
30
20
20
20
15
25/24
25/48
15/48
5/80
79.6
86.1
91.2
90.0
91.8
4.4
2.3
2.4
5.8
3.0
12.3
7.2
3.8
3.1
2.9
3.0
3.7
2.1
0.8
1.8
0.7
0.7
0.5
0.3
0.5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
20/48
a Reaction condition: 15 V methylamine aqueous solution with various concentration;
400 rpm, the residue of 5 was below 1.5% ; b HPLC purity of the reaction mixture.
Due to the insolubility of 6 in water, the precipitates would generate during the
reaction processing, and desired precipitates was easily filtered from the reaction
mixture. Little remaining product and most of the impurities were detected in the
mother liquors by HPLC. The workup procedure was simpler and more efficient than
the extraction operation as reported by the literature. After drying under vacuum at
35 °C for 8 h, the obtained solid was an HCl salt in ethyl acetate to give 6 hydrochloride.
The impurities 14, 25, 29 and most of the impurity 27 remained in ethyl acetate. In
order to obtain a higher purity of 6 and reduce the purification pressure of the target
product, 6 hydrochloride was then reslurried by THF under reflux temperature for 2 h.
The off-white solid with 99.0% purity was obtained. The hydrolytic impurity 25 was
0.08% and the substituted impurity 27 was less than 1.0%.
Rearrangement of α-Hydroxyimine Hydrochloride (rac-1)
20
ACS Paragon Plus Environment

Page 21 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
The thermal rearrangement, also known as ring-expansion reaction, was carried
out in ethyl benzoate or decalin at 170-180 °C for 0.5-1.0 h 12-13. Under the above
conditions, the obtained product was a black gray precipitate, which indicated the
ketamine was carbonized at such high temperature. What was worse, the hydrolytic
impurity 25 from α-hydroxyimine 6 was detected at approximately 25% on HPLC
(Table 6, entry 1), resulting from a small amount of moisture from 6 and solvent in the
reaction mixture. Because trace amount of water is enough to cause the hydrolysis of
imine hydrochloride at high temperature, leading to a low yield of rearrangement (70%).
Therefore, it was pressing and necessary to seek a solution to the carbonization and
hydrolysis of imine hydrochloride salt. When the temperature reduced to 140 °C, the
rearrangement was conducted smoothly with the time extension to 18 h (Table 6, entry
2). Although the carbonization was improved and obtained the grey solid, the hydrolytic
impurity 25 did not decreased. Subsequently, a rearrangement with free base of 6 was
attempted to avoid hydrolysis. However, the conversion and yield were much lower
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
than that of hydrochloride, which was consistent with what was reported 12d
.
Based on the acid catalytic mechanism of the reaction (Scheme 10), the
frequently-used Lewis acid aluminum chloride (AlCl3), was selected to catalyze the
thermal rearrangement and absorb the trace moisture in the reaction mixture. The
experimental results were roughly the same as expected and summarized in Table 6.
When the hydrochloride salt of 6 was subjected to the thermal rearrangement at 140 °C
with 0.1 equivalent AlCl3, 25 was reduced to 2.1% and the reaction time was shortened
21
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 45
1
2
3
4
5
6
7
8
9
to 2 h from 18 h. However, a new impurity with the same molecular weight as 25 was
detected at approximately 8.6% on HPLC (Table 6, entry 3). It was presumed that 25
was rearranged to the impurity 29 at the presence of AlCl3. When the amount of AlCl3
was 1.5 equivalent of the moisture which was determined through Karl Fischer
Moisture Titrator in 6 hydrochloride and solvent, the content of 29 reduced to 3.3%
under the nitrogen atmosphere. The lower reaction temperature was still expected to
reduce risk in manufacturing scale-up. To our delight, the thermal rearrangement still
could be successfully completed in 5 h at 125 °C, but when the temperature decreased
to 110 °C, there was still 5.6 % residual when the reaction time prolonged to 24 h.
Ultimately, the optimized process was as follows: under a nitrogen atmosphere, 6
hydrochloride and AlCl3 (1.5 equiv of the moisture) reacted in ethyl benzoate at 125 °C
for 5-6 h. The critical process parameter, the moisture content in the solvent and 6, was
lower than 150 ppm and 2000 ppm respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 10. The Proposed Mechanism of the Thermal Rearrangement
22
ACS Paragon Plus Environment

Page 23 of 45
Organic Process Research & Development
1
2
3
4
5
Table 6. The Thermal Rearrangement with AlCl3 under Various Temperature a
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
HPLC (%) b
Entry
Catalyst
T/t (°C/h)
6
-
1
25
25.3
22.7
2.1
29
1
-
180/2
140/18
140/2
140/2
125/5
110/24
74.4
77.1
89.3
95.7
95.5
91.2
0.3
0.2
8.6
3.3
3.8
2.8
2
-
0.1
0.1
-
3c
4d
5d
6d
AlCl3
AlCl3
AlCl3
AlCl3
1.0
0.1
5.6
0.7
0.4
a
Reaction conditions: 0.1 g/mL α-hydroxylimine hydrochloride 6 in ethyl benzoate
under a nitrogen atmosphere; b Only peak areas (%) of 5, 25, 29 and 1 were integrated
on HPLC; c 0.1 equivalent AlCl3; d 1.5 equivalent AlCl3 (the moisture in 6 and solvent).
Afterwards, the solution was cooled to room temperature and then cyclohexane
was slowly added. The resulting precipitates were filtered from the mixture and washed
with cyclohexane, achieving the 99% chemical purity on HPLC. Most of the impurities
23
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 45
1
2
3
4
5
6
7
25, 27 and 29 remained in the mother liquors detected by HPLC (25<0.1%, 27<0.3%,
29<0.2% in crude ketamine).
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Preparation of Esketamine Hydrochloride (S-1).
According to the relevant research, the commercially available and inexpensive
L-(+)-tartaric acid was finally chosen as the chiral resolution agent and performed in a
mixture of acetone and H2O 10. Subsequently, different proportions of acetone and H2O
were screened, and the yield of resolution was up to 46% (lit 10a 45%) when the chiral
resolution was performed in the mixture acetone and H2O at 15:1, and the chiral purity
of tartrate 11 was 99.5% with gradient cooling process.
Acetone was removed from the mother liquors after the resolution and then the
solid enriched with (R)-isomer was obtained after basification with NaOH aqueous
solution. What puzzled us was that the ratio of the (R)-1 and (S)-1 displayed in chiral
HPLC was R: S=85:15. However, the content of (S)-isomer in mother liquor should be
at most 8% according to the 46% yield. In order to make the contradiction clear, the
resulting tartrate was subsequently alkalized by sodium hydroxide aqueous solution and
extracted by dichloromethane, yielding only 41%, which had a big difference with the
resolution yield (46%). Many factors were investigated, and eventually SCXRD gave
us the final answer that the crystalline unit cell of (S)-ketamine tartrate contains two
water molecules (cultured from acetone/ H2O, Figure 1). At present, there is no
information about the crystal water of (S)-ketamine tartrate through an extensive
24
ACS Paragon Plus Environment

Page 25 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
literature review. This is the first report that the (S)-ketamine tartrate is a dihydrate,
which makes the molecular weight of dihydrate is corrected to 423 but not 387 reported
in the literature 10c. According to the definite molecular weight 423, the resolution yield
was 41%, which is consistent with the yield of alkalization of the resulting (S)-ketamine
tartrate with sodium hydroxide.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. The SCXRD of Tartrate of Esketamine
Initially, the direct conversion of tartrate to hydrochloride in 5V isopropanol was
considered, but the experimental result showed that the yield was only 85% which
might be caused by the free water in isopropanol and the crystalline water in tartrate.
Finally, the conventional process was adopted, where the tartrate was alkalized with
diluted sodium hydroxide aqueous solution to get the free base and further formed a
salt with hydrochloric acid in ethyl acetate to obtain the desired crude drug substance.
The purification of (S)-1 hydrochloride via a recrystallization process was performed
in the mixture of acetone and water at the ratio of 7:1 (v/v) and a good yield (∼80%) of
drug substance with ICH-grade quality was obtained (99.8% purity and 99.9% e.e.).
25
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 45
1
2
3
4
5
6
7
8
9
The physical properties and the structure of purified (S)-1 hydrochloride was
characterized by DSC, TGA, EA, HRMS, NMR and SCXRD, and all the date were
available in the Supporting Information.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Recycling of the Unwanted (R)-1 Enantiomer
The unwanted (R)-1 isomer lost in the resolution mother liquors not only lowers
the total yield of the route, but also produces a large amount of waste. However, the
defects can be overcome adequately supposing that the conversion of (R)-1 isomer back
to the racemic 1. Herein, the recycling of the unwanted (R)-1 enantiomer from the
resolution mother liquors was studied. According to our previous study (the
unpublished work is under revision at present, Manuscript ID: op-2019-00436u.R1),
the catalytic efficiency of several frequently-used Lewis acids are arranged as follow:
AlCl3 ≈ MgCl2 > FeCl3 > ZnCl2 > BF3 > CaCl2. Considering the high catalytic efficiency
and removability, AlCl3 was finally selected to catalyze the racemization of (R)-1.
The equivalent of AlCl3 was further investigated and the results were summarized
in Table 7. 20 mol % of AlCl3 was required for the racemization of (R)-1 in previous
studies (entry 1). The possibility of further racemization of the AlCl3 was attempted
and it was observed that 10 mol % of AlCl3 was not enough to afford a racemic mixture
of (R)-1/(S)-1 in 10 h (entry 2). To our delight, when the time was extended to 24 h, a
racemic mixture of (R)-1/(S)-1 = 49.5 / 50.5 was obtained. However, when reaction
temperature was dropped to 130 °C with 20 mol % of AlCl3, the racemization rate was
26
ACS Paragon Plus Environment

Page 27 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
much slower compared to that at 150 °C (entry 4). The updated racemization procedure
was as follows: The (R)-1 hydrochloride was treated with 10 mol % of AlCl3 in ethyl
benzoate at 150 °C for 24 h with 99% yield. The racemized ketamine was subsequently
resolved by L-(+)-tartaric acid to obtain (S)-1 in 41% yield and with 99.5% e.e. With
three times recycling of the (R)-1, the yield of resolution could reach to 87%, and the
overall yield was up to 51.1% (14.0% without racemization). Such a concise and cost-
efficient approach of the racemization could be industrially useful to recycle the waste
(R)-ketamine into the resolution process to obtain the (S)-ketamine.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 7. Optimization of aluminum chlorideand temperature for the (R)-1 racemization.
a
equiv
Entry
T/t (°C/h)
(R)-1/(S)-1 (%)b
6 (%)c
(AlCl3)
0.2
1
2
3
4
150/10
150/10
150/24
130/10
49.8/50.2
65.2/34.8
49.5/50.5
81.2/18.8
0.4
10.9
0.2
0.1
0.1
0.2
30.5
27
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 45
1
2
3
4
5
a
b
Reaction conditions: 0.1 g/mL (R)-1 hydrochloride (98.5% e.e) in ethyl benzoate;
6
7
c
Measured by HPLC on a Chiralpak OD-3 column; Measured by HPLC on a Waters
8
9
X-Bridge C18 column, only peak areas (%) of 6 and 1 were integrated.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Control Strategy and Impurity Limit of Process Related Impurity in Esketamine
The optimized process could not only be industrially scalable but also be quality
controllable. The process parameters were established and improved well as above
described. The control strategy and impurity limit of process related impurities were
proposed, as shown in Table 8.
Table 8 Control Strategy and Impurity Limit of Process Related Impurity in
Esketamine
Control strategy
Compounds
Impurity content a
Impurity limit
and results
14 (1.5%)
15 (0.3%)
17 (0.4%)
20 (5.8%)
22 (2.3 %)
4 (0.1%)
14 ≤ 0.5%;
15 ≤ 0.3%;
17 ≤ 0.3%;
20 ≤ 1.0%;
22 ≤ 0.5%;
4 ≤ 0.1%;
5 ≤ 0.1%;
Purification by vacuum
distillation
14 ≤ 0.4%; 15 ≤ 0.1%;
17 ≤ 0.1%; 20 ≤ 0.8%;
22 ≤ 0.5%;
4
6
(1) Salt-forming process
5 (0.5%)
28
ACS Paragon Plus Environment

Page 29 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
14 (0.3%)
4 (ND b); 5 (ND); 14
14 ≤ 0.1%;
25 ≤ 0.5%;
27 ≤ 1.0%;
(ND);
25 (0.8%)
27 (1.4%)
29 (0.1 %)
25 ≤ 0.3%; 27 ≤ 1.0%;
29 (ND)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(2) THF reslurry
25 ≤ 0.3%, 27 ≤ 1.0%
6 (0.08%)
25 (0.02%)
6 ≤ 0.1%;
25 ≤ 0.1%;
Recrystallized with
acetone/H2O:
6 (ND); 25 (ND);
27 ≤ 0.03%; 29 (ND);
Ethyl benzoate (ND)
Ketamine
27 (0.24%)
27 ≤ 0.1%;
29 (0.11 %)
29 ≤ 0.1%;
Ethyl benzoate (0.9 %)
Ethyl benzoate ≤ 0.1%
Recrystallized with
acetone/H2O
Esketamine
(R)-1 (0.45 %)
(R)-1 ≤ 0.1%
a Content assay of impurities using peak area normalization method; Not detected on
b
HPLC.
CONCLUSION
In conclusion, a cost-effective, recoverable, and industrially scalable synthetic
process for Esketamine, has been developed and significantly optimized. Throughout
the synthetic route, common and inexpensive reagents were used. The preparation of
29
ACS Paragon Plus Environment