Design and Optimization of a Continuous Stirred Tank Reactor Cascade for Membrane-Based Diazomethane Production: Synthesis of α-Chloroketones

Article abstract of DOI:10.1021/acs.oprd.9b00115

The development of a continuous diazomethane generator comprising a continuous stirred tank reactor (CSTR) cascade and membrane separation technology is reported. This reactor concept was applied for the telescoped three-step synthesis of a chiral α-chloroketone, a key building block for many HIV protease inhibitors, via a modified Arndt-Eistert reaction starting from N-protected l-phenylalanine. The initial mixed anhydride was generated in a coil reactor and directly introduced into the CSTR diazomethane cascade. The use of a semipermeable Teflon membrane (AF-2400) allowed the generation of anhydrous diazomethane, which diffuses through the membrane into the CSTR where it is immediately consumed by the anhydride to furnish the corresponding diazoketone. The subsequent halogenation with concentrated HCl was performed downstream in batch and allowed production of the α-chloroketone on a multigram scale, with a productivity of 1.54 g/h (5.2 mmol/h).

Full text of DOI:10.1021/acs.oprd.9b00115

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC
Full Paper
Design and Optimization of a CSTR Cascade for Membrane-
based Diazomethane Production - Synthesis of #-Chloroketones
Michaela Wernik, Peter Poechlauer, Christoph Schmoelzer, Doris Dallinger, and C. Oliver Kappe
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00115 • Publication Date (Web): 21 Jun 2019
Downloaded from http://pubs.acs.org on June 23, 2019
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 25
Organic Process Research & Development
1
1
2
3
4
5
Design and Optimization of a CSTR Cascade for Membrane-
based Diazomethane Production - Synthesis of -Chloroketones
6
7
8
9
Michaela Wernik, †,‡ Peter Poechlauer,§ Christoph Schmoelzer,§ Doris
Dallinger,*†,‡ and C. Oliver Kappe*,†,‡
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
† Center for Continuous Synthesis and Processing (CCFLOW), Research Center
Pharmaceutical Engineering GmbH (RCPE), Inffeldgasse 13, 8010 Graz, Austria
‡ Institute of Chemistry, University of Graz, NAWI Graz, Heinrichstrasse 28, 8010 Graz,
Austria
§ Patheon Austria GmbH & Co KG, Sankt-Peter-Straße 25, 4020 Linz, Austria
____________________
* Corresponding authors. E-mail: do.dallinger@uni-graz.at, E-mail: oliver.kappe@uni-graz.at
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 25
2
1
2
3
4
TABLE OF CONTENTS ENTRY
5
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
ACS Paragon Plus Environment

Page 3 of 25
Organic Process Research & Development
3
1
2
3
4
Abstract
5
6
7
8
The development of a continuous diazomethane generator comprising a continuous stirred tank
reactor (CSTR) cascade and membrane separation technology is reported. This reactor concept
was applied for the telescoped three step synthesis of a chiral α-chloroketone, a key building
block for many HIV protease inhibitors, via a modified Arndt-Eistert reaction starting from N-
protected L-phenylalanine. The initial mixed anhydride was generated in a coil reactor and
directly introduced into the CSTR diazomethane cascade. The use of a semi permeable Teflon
membrane (AF-2400) allowed to produce anhydrous diazomethane, which diffuses through the
membrane into the CSTR where it is immediately consumed by the anhydride to furnish the
corresponding diazoketone. The subsequent halogenation with concentrated HCl was
performed downstream in batch and allowed production of the α-chloroketone on a multigram
scale, with a productivity of 1.54 g/h (5.2 mmol/h).
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
Keywords: anhydrous diazomethane, membrane separation, CSTR, α-chloroketone synthesis
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 25
4
1
2
3
4
Introduction
5
6
7
8
Since the late 1990s, the FDA has approved a series of highly effective, orally bioavailable
protease inhibitors for the treatment of HIV, such as Atazanavir, Palinavir or Lopinavir, that
contain a chiral amino alcohol as key building block (Scheme 1).1,2 This central moiety is
generally introduced by a nucleophilic ring-opening of the respective N-protected
aminoepoxide via the amino functionality of the C-terminal building block. The chiral
aminoepoxide is usually obtained by selective reduction of the corresponding halomethyl
ketone, which is readily available via halomethylation of an N-protected amino acid.2 This most
direct and cost effective route involves an Arndt-Eistert type reaction of an activated amino
acid with anhydrous diazomethane (CH2N2) to the α-diazoketone, followed by an α,α-
substitution with a hydrogen halide to furnish the corresponding α-halo ketone (Scheme 1).
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
PG
HN
O
PG
HN
O
O
PG
HN
O
PG
HN
O
Base
CH2N2
HX
N2
X
OH
O
OEt
ClCO2Et
R1
R1
R1
R1
O
Ph
HN
OMe
O
CONHt-Bu
O
H
Ph
O
N
Ph
N
O
O
O
N
N
N
H
H
N
H
MeO
N
NH
O
O
OH
N
N
N
NH
H
H
O
OH
OH
O
Ph
N
Atazanavir
Palinavir
Lopinavir
N
Scheme 1. Synthesis of -haloketones from CH2N2 and N-protected amino acids.
CH2N2 is a versatile C1 building block for numerous reactions,3,4 with the advantage that those
transformations are generally fast, clean (often nitrogen is the sole byproduct) and can be
carried out under mild conditions. On the other hand, as with many alkylating reagents, CH2N2
is extremely toxic and a potent carcinogen.5 Due to high volatility and extreme sensitivity to
explosive decomposition it is also challenging to handle or to store CH2N2.6 Diazomethane is
usually prepared by base-mediated decomposition of an N-methyl-N-nitroso amine precursor
such as Diazald.6-8 To generate anhydrous CH2N2, classical distillation techniques have
historically been employed and special distillation kits were developed to co-distil an ethereal
CH2N2 solution.6,9 As an alternative to the distillation procedures a nitrogen stream can be
employed to transport the CH2N2 into the gas phase and further into the substrate-containing
receiver flask.10
Since the interest in diazo-compounds has increased in recent years, the development of
safer techniques for the preparation of CH2N2 is in constant progress. Over the last years,
ACS Paragon Plus Environment

Page 5 of 25
Organic Process Research & Development
5
1
2
3
4
5
6
7
8
9
different methods to generate CH2N2 in continuous flow systems have been developed.3,11-16
This allows the in situ on-demand generation of CH2N2 without the hazards of storage, isolation
or accumulation of large amounts of material. For instance, a recent approach by Lehmann
applied a micro reactor and a liquid–liquid phase separator for continuous CH2N2 generation
and purification on a 95-117 mmol/h scale.17
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 safe method to generate pure, anhydrous CH2N2 is the use of membrane technology,
which allows continuous separation/isolation of gaseous CH2N2 from the basic aqueous stream.
A number of different reactor types have been developed using microporous, semipermeable
membranes, which are chemically and mechanically stable and selectively allow the transport
of low-molecular hydrophobic molecules.18-24 Two examples of these reactor types were
introduced by our group: the tube-in-tube20,21 and the tube-in-flask reactor,22-24 both of them
relying on a commercially available Teflon AF-2400 membrane as separation and purification
tool. The fully continuous tube-in-tube reactor, which was initially developed by Ley to saturate
a liquid phase with gas,25 was used to generate CH2N2 in an aqueous inner stream. Subsequently,
the gaseous CH2N2 diffused through the membrane into an outer substrate stream which was
enclosed within a thick walled impermeable outer tubing. In an attempt to further simplify the
set-up with respect to the handling of solids, concentrations, and scale-up, a tube-in-flask
reactor was developed where the membrane is coiled inside a glass flask, which allows the
gaseous CH2N2 to diffuse into the flask where it is immediately consumed by the substrate
solution.22,23
Both set-ups have been employed by our group for the multistep synthesis of α-
chloroketones starting from N-protected amino acids.21,24 In the tube-in-tube reactor, CH2N2
was generated from Diazald as precursor,21 whereas the tube-in-flask protocol started from
cheap and benign N-methyl urea and sodium nitrite.24
Although the tube-in-tube reactor operates fully continuously, its limitation is a very
low productivity (1.25 mmol/h) especially when slow reactions such as the intended
diazoketone synthesis are considered. A larger throughput can be obtained with the semibatch
tube-in-flask reactor, in particular by simple parallelization of the membranes.22 Nevertheless,
this approach still has its limits with regard to scale-up as the size of the flask, for safety reasons,
cannot be increased to a commercially relevant volume. In consideration of a higher throughput,
we report here the design of a tube-in-CSTR, which is a fully continuous version of the tube-
in-flask model. In a proof-of-concept study, the three-step synthesis of an α-chloroketone from
N-protected L-phenylalanine and anhydrous CH2N2 was performed continuously employing the
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 25
6
1
2
3
4
5
6
tube-in-CSTR set-up. For reaction control and monitoring of the CH2N2 generation,
respectively, in-line FTIR was implemented as PAT tool.
7
8
9
Results and Discussion
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
Before implementing the tube-in-CSTR concept, the tube-in-flask set-up (Figure 1) was used
for simplicity reasons to execute a thorough evaluation of the membrane performance with
respect to the anhydrous nature of CH2N2,the mass balance and the throughput.A commercially
available AF-2400 membrane tubing (0.8 mm ID, 1.0 mm OD) was coiled inside a standard
250 mL GLS 80 wide necked glass flask (Ø neck: 80 mm, Ø: 95 mm, h: 105 mm), as previously
reported.22,23 A custom-made safety screw cap was attached to the glass flask (see Figure S1,
Supporting Information). The semipermeable membrane was connected to standard gas-tight
PFA (perfluoroalkoxy alkane) tubing (1/16 inch OD) on both sides. The inlet tubing was fed
with the CH2N2 solution that was generated in situ from Diazald and KOH solutions. The two
streams were combined in a Y-mixer. Precipitation and blockages at the mixing point were
avoided by immersing the Y-mixer into a water bath at 30 °C. The outlet tubing was connected
to a back pressure regulator (BPR) and further led into a quench solution (e.g., acetic acid). The
glass flask was filled with 50 mL of solution (solvent/substrate) which was stirred by a Teflon
coated stirring bar and cooled in an ice bath to 0 °C.
Although CH2N2 is best soluble in diethyl ether,22 its use as solvent is strongly
discouraged, if not forbidden, in industrial laboratories.26 Therefore, methyl tert-butyl ether
(MTBE) was considered a more appropriate solvent despite the lower CH2N2 solubility.22 All
reactions were monitored with in situ FTIR spectroscopy (Mettler Toledo ReactIR 15 with a
silicon probe). The probe was inserted into the flask via a screw-cap adapter. The described
reactor (Figure 1) was the standard set-up for the subsequent experiments.
ACS Paragon Plus Environment

Page 7 of 25
Organic Process Research & Development
7
1
2
3
4
5
6
7
8
9
O
S
O
NO
CH3
H3C
FTIR probe
30 °C
Diazald in DMF
(1 M)
quench
BPR
5 bar
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
aq KOH
(2 M)
CH2N2
Figure 1. Schematic overview of the tube-in-flask reactor for initial optimization studies (see
also Figure S1, Supporting Information).
Generation of Anhydrous Diazomethane. Access to anhydrous CH2N2 is crucial for the
modified Arndt-Eistert reaction and therefore for the synthesis of the amino alcohol building
block of HIV protease inhibitors. Previous reports describe the reaction of acyl chlorides (e.g.,
benzoyl chloride) with CH2N2 as an indicator of anhydrous reaction conditions.20,22
Consequently, the water sensitivity of benzoyl chloride was examined. As it turned out, the
expected benzoic acid was only detectable via GC-FID upon addition of at least 50 vol% of
water, showing that this method was not sufficiently sensitive to prove the generation of
anhydrous CH2N2. Hence, other methods to determine the water content in the organic CH2N2
solution inside the flask were explored. FTIR spectroscopy could in principle be used to monitor
the characteristic absorption spectrum of water at ca. 3400 cm–1,27 but unfortunately, the
absorption of water is in the blind spot area of the silicon probe used in this study. Ultimately,
the water content was determined by Karl-Fischer titration. A 0.1 M solution of CH2N2 in
anhydrous MTBE (0.0003% H2O) was generated inside the flask and subsequently quenched
with benzoic acid. A reference Karl-Fischer titration was performed for a benzoic acid solution
in MTBE, and the value subtracted from the one of the quenched CH2N2 solution. The Karl-
Fischer titration proved our concept of anhydrous CH2N2 generation by revealing 347 ± 7 ppm
of water in the quenched solution.
FTIR-Monitoring. As recently reported by our group, FTIR spectroscopy allows to accurately
monitor the generation and consumption of CH2N2 based on the characteristic C=N2 stretch at
2100 cm–1.22 By employing the Mettler Toledo iCQuant software, the factual concentration of
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 25
8
1
2
3
4
5
6
7
8
9
CH2N2 in the flask could be determined in real time. Therefore, CH2N2 was generated by
pumping 10 mL of a 1 M solution of Diazald in DMF and a 2 M KOH solution in MeOH/H2O
(1:2) with a flow rate of 200 µL/min each through the AF-2400 membrane tubing (3.5 m). The
flask was filled with 50 mL of anhydrous MTBE and samples were taken in definite intervals.
The conversion of a stoichiometric amount of benzoic acid to methyl benzoate was determined
via HPLC−UV/vis peak integration at 215 nm and used for an external calibration. This allowed
to directly convert the peak area of the C=N2 stretch at 2100 cm-1 into the CH2N2 concentration
in the flask over time (Figure 2).
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
As can be seen in Figure 2b, the CH2N2 concentration increases linearly until it reaches
a plateau at a maximum concentration of 0.10 M, which correlates to 50% CH2N2 diffusion into
the flask, after 48 min. After ca. 55 min a quench with acetic acid was performed, upon which
the concentration of CH2N2 drops to zero. The calculated values from the iCQuant software
(0.104 M) were in good agreement with the experimental values from the external calibration
(0.110 M). The collected reference data were directly converted into a quantitative model,
which enabled the direct measurement of the CH2N2 concentration without an external control
reaction with benzoic acid. This CH2N2 quantification concept was implemented in the
membrane performance studies described below for experiments where only pure solvent was
contained inside the flask.
ACS Paragon Plus Environment

Page 9 of 25
Organic Process Research & Development
9
1
2
3
4
5
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
Figure 2. Correlation of CH2N2 concentration (HPLC analysis) with a) FTIR peak area of the
C=N2 stretch at 2100 cm-1, and b) time.
Membrane Performance/Mass Balance Evaluation. In order to achieve a high diffusion of
CH2N2 through the membrane tubing into the solution inside the flask, in-depth optimization
studies were performed with respect to substrates/type of CH2N2 consumption, residence
time/flow rates and throughput. Significant emphasis was put on the careful determination of
the mass balance. Standard conditions for the membrane performance studies were based on
those we have reported previously for the tube-in-flask reactor.22,23 10 mL of a 1 M Diazald
solution (10 mmol) in DMF and a 2 M KOH solution in MeOH/H2O (1:2) were pumped through
the AF-2400 membrane (3.5 m, 1.75 mL residence volume) at equal flow rates. The mass
balance is based on complete conversion of Diazald to CH2N2.
The diffusion of CH2N2 into pure MTBE was compared to the diffusion into a reaction
mixture containing an equimolar amount of benzoic acid, with respect to Diazald and thus
CH2N2, inside the flask. The final CH2N2 yield was determined after the complete consumption
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 25
10
1
2
3
4
5
6
7
8
9
of the Diazald solution. For no-substrate containing studies, CH2N2 was quantified via FTIR
(vide supra). In the other case, the conversion of benzoic acid to methyl benzoate inside the
flask (HPLC analysis) provided information on the CH2N2 yield. In addition, CH2N2, which did
not diffuse through the membrane tubing, was analyzed in the waste solution. The aqueous
outlet stream was directed into a solution of benzoic acid in MTBE that was equimolar to the
employed Diazald. The corresponding methyl ester that was extracted into the organic phase
was determined by HPLC.
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
In the process of CH2N2 generation from Diazald, potassium p-toluenesulfonate (PPTS)
is formed as by-product,15 which in turn might be further converted to methyl p-
toluenesulfonate (MPTS) by the reaction with CH2N2 in the MeOH/H2O/DMF environment
inside the membrane tubing (Scheme 2). This pathway was corroborated by independent
experiments treating PPTS with CH2N2 (see Supporting Information). MPTS was therefore
expected in the waste solution, and indeed its presence was confirmed by GC-MS and HPLC
versus reference material (Figures S2-S8, Supporting Information). To take MPTS into account
for the mass balance of CH2N2 it was quantified via HPLC.
inside
CH3
O
ON
O
membrane tubing
CH2N2
+
OK
OMe
N
S
O S O
O S O
+
KOH
- H2O
- KOH
- N2
Me
Me
Me
Diazald
PPTS
MPTS
diffusion through
membrane tubing
substrate
product
Scheme 2. CH2N2 generation from Diazald and KOH and its potential consumption pathways.
For more details, see the Supporting Information.
To investigate the impact of the residence time in the membrane on CH2N2 yield and mass balance,
solutions of Diazald and KOH were pumped through the reactor at varying flow rates. As can be seen
in Figure 3, by applying longer residence times, the amount of both CH2N2 and MPTS in the waste
stream decreased. Conversely, unrecovered CH2N2 (), most likely due to decomposition under aqueous
conditions, increased with prolonged residence times: This outcome is not surprising since previous
ACS Paragon Plus Environment

Page 11 of 25
Organic Process Research & Development
11
1
2
3
4
5
6
7
8
9
studies report a half-life of 750 s for CH2N2 in water.14 No-substrate (Figure 3a) and substrate containing
(Figure 3b) reactions in the tube-in-flask system show comparable trends for the mass balance in the
waste stream. However, higher CH2N2 yields inside the flask could be achieved when benzoic
acid was present directly in the flask. Due to the concentration gradient that is maintained, the
diffusion of CH2N2 from the membrane into the flask is favored, according to Le Chatelier’s
principle. Whereas the results for a residence time of 4.4 min are fairly comparable for the two
approaches (52% vs 64% CH2N2 in the flask and 29% vs 28% CH2N2 in the waste), by doubling
the residence time to 8.8 min a 23% higher CH2N2 yield could be achieved when the substrate
was contained inside the flask. This corresponds exactly to the difference of CH2N2 that remains
unrecovered (33% vs 10%, Figure 3) which leads to the conclusion that at some point
decomposition will prevail over diffusion if the CH2N2 is not immediately consumed by a
substrate (see 17.6 min residence time, Figure 3a). Approximately 3-11% of the CH2N2 inside
the membrane react with PPTS to furnish MPTS.
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 3. CH2N2 mass balance with respect to residence times/flow rates. a) CH2N2 was
generated with pure MTBE inside the flask. b) CH2N2 was generated with benzoic acid
dissolved in MTBE inside the flask. DM: diazomethane. c) Flow chart of CH2N2 pathway and
mass balance based on the experimental data shown in Figure 3b (4.4 min residence time).
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 25
12
1
2
3
4
5
6
7
8
9
Since the reaction of CH2N2 and benzoic acid is considered to be essentially instantaneous, also
the mass balance for the reaction of N-Boc-L-phenylalanine (Boc-L-Phe) anhydride 2 and
CH2N2 (Scheme 3) was determined, which is known to proceed at a slower reaction rate.21,24
Compound 2 was generated inside the flask by adding 1 equiv Bu3N and 1.5 equiv ClCO2Et to
a solution of 5 mmol Boc-L-Phe in 50 mL MTBE at 0 °C. Excess CH2N2 is necessary for this
modified Arndt−Eistert reaction, therefore 3 equiv of Diazald were employed. Similar results
for the mass balance as described above (Figure 3) were obtained: With a residence time of 4
min inside the membrane, 15-20% of CH2N2 did not diffuse through the membrane and thus
could be detected in the waste and ca. 5% of CH2N2 reacted with PPTS to MPTS. These results
show that, even if the reaction in the flask is slower, the concentration gradient is not affected
by the rate of CH2N2 consumption, and hence a similar membrane performance can be attained.
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
Boc O
HN
Boc O
HN
O
Boc O
HN
Boc O
HN
HCl
ClCO2Et
Bu3N
CH2N2
N2
Cl
OH
O
OEt
1
2
3
4
Scheme 3. Synthesis of α-chloroketone 4 used for the proof-of-concept study of the tube-in-
CSTR reactor.
In view of the envisaged scale-up in a CSTR, a throughput screening was additionally
performed aiming for maximum diffusion of CH2N2 in correlation with time and expenditure
of starting material. Again, different residence times in the membrane were tested, but now,
rather the correlation with different filling volumes inside the flask was examined (Table 1),
which should provide information about the impact of dilution and headspace. The same
conditions with respect to Diazald and KOH were applied as described above. We have shown
that a higher concentration gradient of CH2N2 can be maintained with the substrate-in-flask
approach (Figure 3b) which in turn leads to higher CH2N2 yields, hence it was also applied for
this study. While - as anticipated - the CH2N2 yield increases from 51 to 93% by increasing the
residence time from 2 to 8 min (Table 1, entries 1-3), the throughput is only affected marginally,
as it decreases by only 15%. The solvent volume in the flask and thus the concentration of
CH2N2, on the other hand, has a stronger effect on the throughput (Table 1, entries 2-5). Similar
yields were obtained independently of the filling volume, resulting in a decrease of the
throughput due to higher dilution. In addition, this observation implies that the headspace does
ACS Paragon Plus Environment

Page 13 of 25
Organic Process Research & Development
13
1
2
3
4
5
6
not have an effect on the reaction outcome, indicating that CH2N2 does not accumulate in the
headspace under these conditions.
Since the conditions shown in entry 2 (Table 1) represent a good compromise between
material consumption, total reaction time (40 min), CH2N2 yield (71%) and thus throughput
(3.80 mmol CH2N2/h), we decided on a 4 min residence time in the membrane and 50 mL
MTBE in the flask for the CSTR experiments.
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
Table 1. CH2N2 Throughput Screeninga
residence
time (min)
MTBE in flask CH2N2 yield CH2N2 conc
CH2N2 throughput
entry
(mL)
50
(%)
51
71
93
80
92
(mol/L)
0.10
(mmol/h)
3.9
1
2
3
4
5
2
4
8
4
8.4
50
0.14
3.8
3.3
1.2
1.0
50
0.19
100
150
0.08
0.06
a Conditions: 10 mL (1 M solution in DMF) of Diazald and 20 mL (2 M solution in MeOH/H2O
1:2) of KOH were pumped at equal flow rates, 5 m AF-2400, 10 mmol benzoic acid dissolved
in MTBE at 0 °C, yield of CH2N2 determined by conversion of benzoic acid to methyl benzoate
and monitored by HPLC peak area% integration.
Temperature Screening. Initially, the tube-in-flask reactor was operated at 0 °C to minimize
the risks associated with CH2N2. At this temperature, no CH2N2 was detectable in the headspace
as measured by FTIR. To simplify the set-up for the CSTR approach we aimed to operate the
flask at slightly higher, potentially even ambient, temperatures. As mentioned above, CH2N2 is
extremely sensitive to explosion, therefore, generating a concentrated solution of CH2N2 in
MTBE at 25 °C should be avoided. Instead, a stoichiometric amount of benzoic acid was added
to the flask in order to directly consume the CH2N2. While increasing the temperature in a
stepwise manner from 0 °C to 25 °C, no CH2N2 was detected in the headspace at any
temperature, not even at 24 °C where MTBE was observed (Figure S10, Supporting
Information). In addition, a headspace experiment at 20 °C for the slower reaction of CH2N2
with Boc-L-Phe mixed anhydride 2 was performed, also here, no CH2N2 was detected in the
headspace (Figure S11, Supporting Information). Presumably, CH2N2 is sufficiently soluble in
MTBE to remain in the liquid phase. Therefore, experiments could be performed at 15-20 °C,
without the risk of having gaseous CH2N2 in the headspace.
CSTR Cascade. As described above, the tube-in-tube and tube-in-flask reactors both have their
limitations considering scale-up. In continuation with our development of larger scale
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 25
14
1
2
3
4
5
6
7
8
9
diazomethane generators, we decided to employ a CSTR concept, which can be considered as
a continuous version of the tube-in-flask reactor. The three-step synthesis of α-chloroketone 4
starting from Boc-L-Phe 1 was used to proof the concept of the tube-in-CSTR (see, Scheme 3).
The first step (activation of amino acid 1) was conducted in flow employing a coil reactor, as
previously reported by Pinho et al.21 The synthesis of the diazoketone 3 was performed in a
CSTR cascade in a continuous flow approach and finally the α-chloroketone synthesis was done
in batch.
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 first CSTR unit (Scheme 4) consisted of a 5 m AF-2400 Teflon membrane inside a
double walled 250 mL CSTR, which was cooled to 15-20 °C and equipped with a Teflon-bladed
mechanical stirrer. A nylon net was employed to hold the fragile membrane tubing in place
inside the CSTR, in addition, this design also provided protection from the stirrer blades.
Similar to the tube-in-flask set-up, the membrane was connected to standard PFA tubing (1/16”
OD) on both sides. The custom-made Teflon lid (Figure S13, Supporting Information) was
equipped with appropriate ports to introduce the inlet and outlet fittings. The inlet tubing was
connected to two syringe pumps via a Y-mixer, which was immersed into a 30 °C water bath.
For these experiments, the same conditions as described above were applied: solutions of
Diazald (1 M in DMF) and KOH (2 M in MeOH/H2O 1:2) were pumped through the membrane
tubing at equal flow rates. The outlet tubing was directed to a quench solution (either benzoic
acid for mass balance or acetic acid). In contrast to the tube-in-flask reactor, the BPR was
omitted in the tube-in-CSTR set-up, since the BPR blocked at ambient temperatures when being
used for a longer period of time (> 3 h). Membrane performance tests revealed that the applied
4-5 bar back pressure had no effect on the performance of the membrane. Therefore, to avoid a
potential bottleneck, the BPR was removed from the set-up. It should be noted, however, that
without the BPR the actual residence time is shorter due the formation of gaseous CH2N2.
The substrate feed entered through the Teflon lid of the CSTR and was directed to the
bottom near the stirrer blades. To maintain a constant solvent level in the CSTR and a constant
flow rate by simply using gravity, a goose neck was installed. A Teflon tubing (2.4 mm ID, 3.2
mm OD) was connected to the bottom outlet of the CSTR and then to a Y-piece (13.6 µL inner
volume, thru-hole 2 mm) which was kept at the same height as the desired solvent height in the
CSTR. On one side of the Y-piece another Teflon tubing was attached as outlet. The third
opening of the Y-piece was left open and enlarged by simply drilling a larger hole (3 mm, see
Figure S13b, Supporting Information).
ACS Paragon Plus Environment

Page 15 of 25
Organic Process Research & Development
15
1
2
3
4
5
6
7
8
9
9.7 m PFA
tubing
20 °C
7 min
1
Boc-L-Phe (0.3 M)
Bu3N (0.3 M)
in MTBE
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
ClCO2Et (0.45 M)
in MTBE
quench/waste
Diazald (1 M)
in DMF
CH2N2
gas release
aq KOH (2 M)
30 °C
Boc O
HN
N2
5 m AF-2400
15-20 °C, 4 min
3
Scheme 4. Tube-in-CSTR unit with continuous substrate feed and goose neck outlet (see also
Figures S12 and S13, Supporting Information).
To test the set-up, the mass balance for the CSTR unit was determined using a stoichiometric
amount of benzoic acid dissolved in 50 mL MTBE inside the CSTR. Similar results as in the
tube-in-flask reactor (see, Figure 3b) were obtained, with a 68-77% CH2N2 yield, 15-20%
CH2N2 in the waste and ca. 5 % MPTS generation. As mentioned before, the unrecovered 3-
8 % CH2N2 probablydecomposein the aqueous environment due to the short half-life of CH2N2.
This shows that the transfer of the tube-in flask set-up to the CSTR does not affect the
performance of the AF-2400 membrane.
In accordance with the previously reported conditions for the continuous flow approach
towards activation of the amino acid,21 a solution of ClCO2Et (0.45 M in MTBE, 1.50 equiv.)
was combined in a Y-mixer with a premixed equimolar solution of Boc-L-Phe (1) and Bu3N
(0.3 M in MTBE, 1.00 equiv.). The solutions were pumped through a PFA coil reactor (9.7 m,
0.8 mm ID, 4.85 mL residence volume) at equal flow rates (0.348 mL/min, 7 min residence
time) at 20 °C to reach a conversion of 97% to the desired mixed anhydride 2. As previously
reported,21 the remaining 3% of 1 could not be converted to the anhydride.
Starting with one CSTR unit, several optimization experiments were performed for the
diazoketone synthesis. First, the amount of Diazald required to convert the mixed anhydride 2
to the corresponding diazoketone 3 was screened. Similar to the tube-in-tube reactor, all other
parameters depend on the CH2N2 production efficiency. The residence time in the AF-2400
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 25
16
1
2
3
4
5
6
7
8
9
membrane therefore also determines the flow rate of the substrate feed - thus the Diazald/CH2N2
to substrate ratio - and the residence time inside the CSTR. As described above, the best
conditions with respect to throughput and total reaction time involve a residence time of 4 min
inside the membrane which relates to a total flow rate of 0.626 mL/min. Our studies started
with the optimum conditions that were determined in the tube-in-flask set-up which necessitated
3 equiv of Diazald for full conversion of the mixed anhydride 2 to diazoketone 3. The residence
time in the CSTR therefore was 72 min, and 3 was detected in only 90% conversion (entry 1,
Table 2). Applying more than 3 equiv of Diazald (entries 2-3, Table 2) had no significant effect
on the conversion. The residual 3% of the protected amino acid 1 from the previous step were
converted to the corresponding methyl ester.
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 2. Optimization Studies for the Generation of Diazoketone 3 in One or Two CSTRsa
entry
tres (min)
eq Diazald
DK 3 (%)b
AH 2 (%)b
CSTR 1
CSTR 2
CSTR 1
CSTR 2
1
2
3
4
5
72
77
84
48
72
-
-
3
3.2
3.5
2
-
-
90
91
90
85
98
10
9
-
-
10
15
2
48
72
1.2
0.5
3
a CSTR 1: 1 M solution of Diazald in DMF and 2 M solution of KOH in MeOH/H2O 1:2, total
flow rate: 0.626 mL/min, tres: 4 min, 50 mL MTBE at 20 °C. CSTR 2: 0.5 M solution of Diazald
in DMF and 1 M solution of KOH in MeOH/H2O 1:2, total flow rate: 0.2 mL/min, tres: 4 min,
50 mL MTBE at 20 °C. Anhydride Synthesis: 0.3 M Boc-L-Phe and Bu3N in MTBE and 0.45
M ClCO2Et in MTBE, total flow rate: 0.696 mL/min, tres: 7 min. Conversions measured after 4
CSTR volumes of the first and second CSTR, respectively
b
Conversions of anhydride 2 (AH) to diazoketone 3 (DK) determined by HPLC peak area%
integration, 3 % Boc-L-Phe-OMe from the first step is not included
The remaining 9-10% of anhydride 2 can be explained by the broad residence time distribution,
which is typical for a CSTR. Hence, achieving full conversion within a single CSTR is
challenging. To overcome this issue a second CSTR was installed in series as additional
residence time unit. The outlet of the goose neck attached to the first CSTR was therefore
directed into the second reactor. At the outlet of the second CSTR another goose neck was
installed, to allow a residence time of 14 min and a constant flow rate of 0.696 mL/min over
both units. Unfortunately, the reaction 2  3 did not proceed further, even with a total residence
time of 86 min. Therefore, we concluded that a second residence time unit was not required.
ACS Paragon Plus Environment

Page 17 of 25
Organic Process Research & Development
17
1
2
3
4
5
6
7
8
9
Ultimately, to reach higher conversion a second CH2N2 generator was installed in the
second CSTR (Scheme 5). The CH2N2 generator in this unit was constructed similarly to the
first, with minor adjustments. Since a glass lid was used, the inlet and outlet of the 3.2 m
AF-2400 membrane were connected via PFA tubing through a septum (see Figure S14,
Supporting Information). A solution of Diazald (0.5 M in DMF) and KOH (1 M in MeOH/H2O
1:2) were used in this CH2N2 generator. Since considerably less CH2N2 would be needed to
complete the reaction in the second CSTR, the concentration of the Diazald solution was
minimized to 0.5 M to increase the flow rate to 0.2 mL/min and thus prevent blockage of the
membrane. By applying lower flow rates, occasional precipitation of Diazald and/or PPTS was
noticed. Each of the two CSTRs was filled with 50 mL MTBE. We first considered that lower
amounts of Diazald could be used in the CSTR cascade, which would entail the advantage of a
dramatically decreased residence time of 48 min in each CSTR by applying a total flow rate of
1.04 mL/min (entry 4, Table 2). Unfortunately, only 85% of the anhydride could be converted
to the diazoketone, leaving 15% unreacted anhydride. Best results were obtained with 3 equiv
of Diazald in the first reactor and an additional 0.5 equiv in the second (this corresponds to
3 equiv based on the 10% remaining anhydride 2), which gave a 98% conversion to the desired
diazoketone 3 (entry 5, Table 2) within a residence time of 72 min in each CSTR. Therefore,
these conditions were employed for the steady state experiment. It should be noted, that
compared to the tube-in-tube approach,21 a 1.4 times lower amount of Diazald was required for
the synthesis of diazoketone 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
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 25
18
1
2
3
4
5
6
7
8
9
9.7 m PFA
tubing
aq KOH
(1 M)
Diazald (0.5 M)
in DMF
Feed E
25 °C
7 min
1
Boc L-Phe (0.3 M)
Bu3N (0.3 M)
in MTBE
Feed D
Feed C
30 °C
Feed F
Feed A
ClCO2Et (0.45 M)
in MTBE
2
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
CH2N2
FTIR
quench/waste
gas release
Diazald (1 M)
in DMF
quench/waste
CH2N2
aq KOH (2 M)
30 °C
Feed B
gas release
5 m AF-2400
15-20 °C, 4 min
Boc O
HN
N2
3.2 m AF-2400
15-20 °C, 4 min
3
Scheme 5. Synthesis of diazoketone 3 in the CSTR cascade with two separate CH2N2 generators
(see also Figure S12, Supporting Information).
By conducting a residence time distribution (RTD) experiment (tracer experiment monitored
by FTIR), it was determined that 5 reactor volumes of CSTR 2 are required to reach steady state
(see Figure S15, Supporting Information). This corresponds to a start-up time of 6 h. The FTIR
was placed in the second CSTR and used for monitoring the formation of diazoketone 3 at 2107
cm–1. Unfortunately, the concentration of CH2N2 could not be determined via FTIR, since the
characteristic stretches of CH2N2 (2097 cm-1) and diazoketone 3 overlap. The FTIR spectrum
shown in Figure 4 indicates that, as predicted by the RTD calculations, after ca. 6 h a steady
state is reached. Product was collected in fractions for 4 h whilst the system was at steady state
and HPLC analysis revealed 96-98% diazoketone.
ACS Paragon Plus Environment

Page 19 of 25
Organic Process Research & Development
19
1
2
3
4
5
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
Figure 4. FTIR spectrum of the steady state run for the synthesis of diazoketone 3 employing
the CSTR cascade. Diazoketone 3 was monitored at 2107 cm-1.
The last step in the synthesis toward α-chloroketone 4 requires the addition of HCl to the
diazoketone 3 solution. We opted for aqueous (concentrated) HCl as cost efficient reagent, as
has already been described for the tube-in-flask approach.24 However, attempts for a fully
continuous synthesis of α-chloroketone 4 using the CSTR cascade were unsuccessful. The
goose neck was not reliable, since the surface tension of water is too high, which resulted in a
rising solvent level in the CSTR. Additionally, aqueous HCl accumulated in the dead volume
(15 mL) of the CSTR, resulting in a phase separation even with vigorous stirring, which was
detrimental to reaction performance. Due to these issues, the α-chloroketone synthesis was
conducted in batch by quenching the collected fractions with 3 equiv of concentrated HCl at
room temperature (10 min stirring). Thus, diazoketone 3 is fully converted to α-chloroketone
4, and additionally any excess CH2N2 is destroyed. Applying more than 3 equiv of concentrated
HCl led to deprotection of the α-chloroketone.
For the activation of the amino acid generally an amine base (Et3N or Bu3N) is used as
HCl scavenger which could be challenging for both, a continuous process and the isolation of
the α-chloroketone, respectively. The insolubility of the Et3N•HCl salt in MTBE disqualified
Et3N as suitable base for a continuous flow process. As previously reported in the fully
continuous chloroketone synthesis, Bu3N•HCl is soluble in THF, but separation by column
chromatography was necessary.21 Switching the solvent from THF to MTBE allowed us to
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 25
20
1
2
3
4
5
6
7
8
9
remove the Bu3N•HCl by a simple extraction procedure. An acidic (1 M HCl) and basic
(saturated Na2CO3) washing step and evaporation of the solvent afforded α-chloroketone 4 in
80-85% yield. In the 1H-NMR the only detectable side product is the Boc-L-Phe methyl ester
(2%), which is a result of unconverted amino acid 1 reacting with CH2N2. In total, 7.4 g of
α-chloroketone 4 could be isolated from the steady state run, which corresponds to a
productivity of 1.54 g/h (5.2 mmol/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
Conclusion
In conclusion, we have developed a laboratory-scale multistep process to synthesize α-
chloroketone 4, an important key element of HIV protease inhibitors, from Boc-L-Phe with
multigram-per-day productivity. The synthetic pathway using anhydrous CH2N2 is a cost-
effective and straightforward approach to this building block. A modified Arndt-Eistert reaction
was performed in a continuous set-up with two consecutive CH2N2 generatorshoused in aCSTR
for the synthesis of diazoketone 3. The continuous on-site-on-demand generation and
immediate consumption of CH2N2 eliminates the need to store and transport this hazardous
reagent. Subsequent treatment of the diazo compound with HCl allowed synthesis of the α-
chloroketone 4 in good yields (80-85%) with a purity of 98%. With the set-up described in this
work, an α-chloroketone 4 throughput of 1.54 g/h could be reached. Compared to the tube-in-
tube reactor the productivity could be increased by a factor of 4.2. The tube-in-CSTR cascade
represents a productive continuous alternative to the tube-in-flask reactor previously developed.
Initial studies in the tube-in-flask reactor allowed the development of a robust set-up with the
two CSTRs in series, which ran for several hours without intervention. The same membrane
was used for all the optimization studies in the CSTR and the steady state run, which totals in
70 h of reliable use. Moreover, by optimizing the reaction conditions in the anhydride activation
step, we were able to simplify the purification process of the final product from chromatography
to extraction.
Potentially, the reactor could be scaled up to reach a higher productivity by
parallelization of the membrane in the CSTRs. As demonstrated by our group previously in the
tube-in-flask set-up, the productivity could be increased 4-fold with four membranes – each 4
m in length (2 mL residence volume) – wrapped in parallel.22 Theoretically, with more
membranes housed inside the CSTR, a higher substrate concentration could be used since more
CH2N2 is generated at the same time. This could be achieved by introducing the substrate at a
higher flow rate, which would in turn also reduce the residence time inside the CSTR cascade
and thus the time until steady state is reached. By adjusting the number of membrane tubings
ACS Paragon Plus Environment

Page 21 of 25
Organic Process Research & Development
21
1
2
3
4
5
6
7
8
9
used per CSTR, the throughput could be adjusted as needed: e.g., four membranes with the
same characteristics employed in parallel for CH2N2 generation, would potentially lead to ca.
50 g of α-chloroketone 4 within one working day (8 h, excluding the time needed for reaching
steady state) which corresponds to a throughput of 6.2 g/h. Another, yet more favorable, option
to increase productivity is the application of membranes with a higher CH2N2 diffusion rate,
which allows the processing of higher substrate concentrations. A membrane screening with
respect to increased CH2N2 diffusion capabilities and, importantly, concomitant generation of
anhydrous CH2N2 is under investigation in our laboratories.
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
While various ways to provide the required membrane area for scale-up are known and
have been realized on larger scale,12,17 there is still a need to provide sufficient contact time for
substrates reacting slower with diazomethane. For such substrates, the use of stirred tank
cascades is among the cheapest and best-understood ways to provide the required contact time.
Experimental Section
General. 1H NMR spectra were recorded on a Bruker 300 MHz instrument. 13C NMR spectra
were recorded on the 300 MHz instrument at 75 MHz. Chemical shifts (δ) are expressed in ppm
downfield from TMS as internal standard. The letters s, d, dd, t, q, and m are used to indicate
singlet, doublet, doublet of doublets, triplet, quadruplet, and multiplet. Analytical HPLC
(Shimadzu LC20) analysis was carried out on a C18 reversed-phase (RP) analytical column
(150 mm × 4.6 mm, particle size 5 μm) at 37 °C using mobile phases A (H2O/MeCN (90:10
v/v) + 0.1% TFA) and B (MeCN + 0.1% TFA) at a flow rate of 1.5 mL/min. The following
gradient was applied: linear increase from 30% B to 100% B within 10 min. GC-MS spectra
were recorded using a ThermoFisher Focus GC coupled with a DSQ II (EI, 70 eV). A TR-5MS
column (30 m×0.25 mm×0.25μm) was used, with helium as the carrier gas (1 mL min−1constant
flow). The injector temperature was set to 280 °C. After 1 min at 50 °C, the temperature was
increased by 25 °C min−1 to 300 °C and kept at 300 °C for 3 min.In-line FTIR analysis (software
package iC IR 7.0) was performed using a Mettler Toledo ReactIR 15 with a AgX Fiber Conduit
with an integrated attenuated total reflectance (ATR) gold-sealed silicon sensor and a mercury
cadmium telluride (MCT) detector. Melting points were obtained in a Stuart SMP3 melting
point apparatus in open capillary tubes. The water content was analyzed by an automatic
Metrohm Titrando 831 KF coulometric Karl Fisher titration method (EN ISO 12937:2000) in
triplicate. All solvents and chemicals were obtained from standard commercial vendors and
were used without any further purification. Diazald was synthesized following a literature
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 25
22
1
2
3
4
5
procedure.28 All compounds synthesized herein are known in the literature. Proof of purity was
obtained by 1H NMR and HPLC−UV/Vis spectroscopy.
6
7
8
9
CAUTION: CH2N2 is mutagenic, carcinogenic, toxic and explosive. All reactions should be
carried out in an efficient fumehood with the sash closed, and laboratory personnel working
with CH2N2 MUST familiarize themselves with the potential hazards and prevention measures.
Excess CH2N2 should be destroyed by the addition of acetic acid.
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 α-Chloroketone 4 from N-Boc-L-Phenylalanine (1) in a CSTR Cascade:
The set-up shown in Scheme 5 was used. A 1 M Diazald (225 mmol) solution in DMF (feed A)
and a 2 M KOH (450 mmol) solution in MeOH/H2O 1:2 (feed B) were prepared. For the second
CH2N2 generator a 0.5 M Diazald (37.5 mmol) solution in DMF (feed C) and a 1 M KOH (75
mmol) solution in MeOH/H2O 1:2 (feed D) were prepared. CSTR 1 and CSTR 2 were each
filled with 50 mL of anhydrous MTBE. Both CSTRs were water cooled to 15-20 °C. For feed
E, Boc-L-Phe (75 mmol) and Bu3N (75 mmol) were dissolved in 250 mL of MTBE (0.3 M).
For feed F, ethyl chloroformate (112.5 mmol) was dissolved in 250 mL of MTBE (0.45 M).
Feeds E and F were pumped into a Y-mixer (1.7 µL, 0.5 mm thru-hole) at 20 °C and further
through a 9.7 m PFA tubing (1.6 mm OD, 0.8 mm ID. 4.85 mL internal volume) by a syringe
pump (Syrris Asia) at flow rates of 0.348 mL/min each. The resulting anhydride solution was
directed into CSTR 1. Seven minutes after starting the anhydride synthesis, the CH2N2
generation in both CSTRs was started by switching from pure DMF to Diazald. Feeds A and B
were pumped into a Y-mixer (1.7 µL, 0.5 mm thru-hole), that was immersed in a water bath at
30 °C, by two syringe pumps (Syrris Asia) at flow rates of 0.313 mL/min each. The mixture
was directed through a short PFA tubing (1.6 mm OD, 0.8 mm ID, 110 µL) and then further
through a 5 m AF-2400 Teflon membrane (1.0 mm OD, 0.8 mm ID, 2.5 mL). Feeds C and D
were also pumped into a Y-mixer, that was immersed in a water bath at 30 °C, by two syringe
pumps (Syrris Asia) at flow rates of 0.100 mL/min each. The stream then was directed through
a 3.2 m AF-2400 Teflon membrane (1.0 mm OD, 0.8 mm ID, 1.6 mL).The aqueous streams
leaving the membranes were quenched with AcOH. The product stream was collected for 4 h
(50 mL fractions) once steady state was reached (after 6 h). The 4 x 50 mL fractions were
quenched by dropwise adding concentrated HCl (22.5 mmol). After 10 min under vigorous
magnetic stirring at room temperature, the reaction to the desired α-chloroketone was complete.
Each fraction was extracted with 1 M HCl (50 mL) and saturated Na2CO3 (50 mL). The organic
phase was dried over Na2SO4 and the solvent was removed under reduced pressure. 1.80 - 1.90
ACS Paragon Plus Environment

Page 23 of 25
Organic Process Research & Development
23
1
2
3
4
5
6
7
8
9
1
g (80-85%) of a white solid was isolated. mp. 100-102 °C (lit.29 102-103 °C); H NMR (300
MHz, CDCl3) δ 7.31-7.20 (m, 3H), 7.14-7.11(m, 2H), 5.02-4.99 (d, J = 5.01 Hz 1H), 4.65-4.58
(m, 1H), 4.13 (d, J = 4.13 Hz, 1H), 3.94 (d, J = 3.94 Hz, 1H) 3.08-2.97 (m, 2H), 1.36 (s, 9H);
13C NMR (75 MHz, CDCl3) δ 201.5, 155.3, 135.7, 129.3, 129.1, 127.5, 80.7, 58.6, 47.7, 37.8,
28.4
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
Supporting Information Available. Additional experimental information, supplementary
figures and 1H and 13C NMR spectra. This material is available free of charge via the Internet
at http://pubs.acs.
Acknowledgement
The CC Flow Project (Austrian Research Promotion Agency FFG No. 862766) is funded
through the Austrian COMET Program by the Austrian Federal Ministry of Transport,
Innovation and Technology (BMVIT), the Austrian Federal Ministry of Science, Research and
Economy (BMWFW), and by the State of Styria (Styrian Funding Agency SFG).
References
(1)
Antiretroviral
drugs
used
in
the
treatment
of
HIV
infection
(accessed
https://www.fda.gov/forpatients/illness/hivaids/treatment/ucm118915.htm
March 13, 2019).
(2)
(3)
(4)
(5)
Izawa, K.; Onishi, T. Industrial Syntheses of the Central Core Molecules of HIV Protease
Inhibitors. Chem. Rev. 2006, 106, 2811−2827
a) Dallinger, D.; Kappe, C. O. Recent Enabling Technologies for Diazomethane
Generation and Reactions. Aldrichim. Acta 2016, 49, 57–66.
Black, T. H. The Preparation and Reactions of Diazomethane. Aldrichim. Acta 1983, 16,
3–10.
National Institute for Occupational Safety and Health. Diazomethane.
https://www.cdc.gov/niosh/docs/81-123/ (accessed March 13, 2019).
de Boer, T. J.; Backer, H. J. Diazomethane. Org. Synth. 1963, 4, 250.
Arndt, F. Diazomethane. Org. Synth. 1943, 2, 165.
(6)
(7)
(8)
(9)
Moore, J. A.; Reed, D. E. Diazomethane. Org. Synth. 1973, 5, 351.
Hudlicky, M. An Improved Apparatus for the Laboratory Preparation of Diazomethane.
J. Org. Chem. 1980, 45, 5377–5378.
(10) Carlson, E.; Duret, G.; Blanchard, N.; Tam, W. Synthesis of Cyclopropanated [2.2.1]
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 25
24
1
2
3
4
Heterobicycloalkenes: An Improved Procedure. Synth. Commun. 2016, 46, 55–62.
5
6
7
8
(11) For recent reviews, see: (a) Deadman, B. J.; Collins, S. G.; Maguire, A. R. Taming
Hazardous Chemistry in Flow: The Continuous Processing of Diazo and Diazonium
Compounds Chem. - Eur. J. 2015, 21, 2298–2308; (b) Ford, A.; Miel, H.; Ring, A.;
Slattery, C. N.; Maguire, A. R.; McKervey, M. A. Modern Organic Synthesis with α-
Diazocarbonyl Compounds Chem. Rev. 2015, 115, 9981–10080; (c) Müller, S. T. R.;
Wirth, T. Diazo Compounds in Continuous-Flow Technology ChemSusChem 2015, 8,
245–250; (d) Gutmann, B.; Cantillo, D.; Kappe, C. O. Continuous-Flow Technology-A
Tool for the Safe Manufacturing of Active Pharmaceutical Ingredients. Angew. Chemie
Int. Ed. 2015, 54, 6688–6728; (e) Movsisyan, M.; Delbeke, E. I. P.; Berton, J. K. E. T.;
Battilocchio, C.; Ley, S. V.; Stevens, C. V. Taming Hazardous Chemistry by Continuous
Flow Technology. Chem. Soc. Rev. 2016, 45, 4892–4928.
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
(12) Poechlauer, P.; Braune, S.; Dielemans, B.; Kaptein, B.; Obermüller, R.; Thathagar, M.
On-Site-on Demand Production of Hazardous Chemicals by Continuous Flow Processes.
Chim. Oggi/Chemistry Today 2012, 30, 51–54.
(13) Archibald, T. G.; Barnard, J. C.; Reese, H. F. Continuous Process for Diazomethane from
an N-Methyl-N-Nitrosoamine and from Methylurea through N-Methyl-N-Nitrosourea.
US Patent 5854405, 1998.
(14) Proctor, L. D.; Warr, A. J. Development of a Continuous Process for the Industrial
Generation of Diazomethane. Org. Process Res. Dev. 2002, 6, 884–892.
(15) Struempel, M.; Ondruschka, B.; Daute, R.; Stark, A. Making Diazomethane Accessible
for R&D and Industry: Generation and Direct Conversion in a Continuous Micro-
Reactor Set-Up. Green Chem. 2008, 10, 41–43.
(16) Rossi, E.; Woehl, P.; Maggini, M. Scalable in Situ Diazomethane Generation in
Continuous-Flow Reactors. Org. Process Res. Dev. 2012, 16, 1146–1149.
(17) (a) Lehmann, H. A Scalable and Safe Continuous Flow Procedure for In-Line Generation
of Diazomethane and Its Precursor MNU. Green Chem. 2017, 19, 1449–1453; (b)
Lehmann, H. Diazomethane and flow chemistry: filling a gap in pharma R&D. Chim.
Oggi/Chemistry Today 2018, 36, 22–24.
(18) (a) Mallia, C. J.; Baxendale, I. R. The Use of Gases in Flow Synthesis. Org. Process Res.
Dev. 2016, 20, 327–360; (b) Brzozowski, M.; O’Brien, M.; Ley, S. V.; Polyzos, A. Flow
Chemistry: Intelligent Processing of Gas-Liquid Transformations Using a Tube-in-Tube
Reactor. Acc. Chem. Res. 2015, 48, 349–362.
(19) Maurya, R. A.; Park, C. P.; Lee, J. H.; Kim, D. P. Continuous in Situ Generation,
ACS Paragon Plus Environment

Page 25 of 25
Organic Process Research & Development
25
1
2
3
4
5
6
7
8
9
Separation, and Reaction of Diazomethane in a Dual-Channel Microreactor. Angew.
Chemie - Int. Ed. 2011, 50, 5952–5955.
(20) Mastronardi, F.; Gutmann, B.; Kappe, C. O. Continuous Flow Generation and Reactions
of Anhydrous Diazomethane Using a Teflon AF-2400 Tube-in-Tube Reactor. Org. Lett.
2013, 15, 5590–5593.
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
(21) Pinho, V. D.; Gutmann, B.; Miranda, L. S. M.; De Souza, R. O. M. A.; Kappe, C. O.
Continuous Flow Synthesis of Alpha-Halo Ketones: Essential Building Blocks of
Antiretroviral Agents. J. Org. Chem. 2014, 79, 1555–1562.
(22) Dallinger, D.; Pinho, V. D.; Gutmann, B.; Kappe, C. O. Laboratory-Scale Membrane
Reactor for the Generation of Anhydrous Diazomethane. J. Org. Chem. 2016, 81, 5814–
5823.
(23) Dallinger, D.; Kappe, C. O. Lab-Scale Production of Anhydrous Diazomethane Using
Membrane Separation Technology. Nat. Protoc. 2017, 12, 2138–2147.
(24) Garbarino, S.; Guerra, J.; Poechlauer, P.; Gutmann, B.; Kappe, C. O. One-Pot Synthesis
of α-Haloketones Employing a Membrane-Based Semibatch Diazomethane Generator.
J. Flow Chem. 2016, 6, 211–217.
(25) O’Brien, M.; Baxendale, I. R.; Ley, S. V. Flow Ozonolysis Using a Semipermeable
Teflon AF-2400 Membrane To Effect Gas−Liquid Contact. Org. Lett. 2010, 12, 1596–
1598.
(26) Prat, D.; Wells, A.; Hayler, J.; Sneddon, H.; McElroy, C. R.; Abou-Shehada, S.; Dunn,
P. J. Selection Guide of Classical- and Less Classical-Solvents. Green Chem. 2015, 18,
288–296.
(27) Pershin, S. M.; Bunkin, A. F.; Lukyanchenko, V. A.; Nigmatullin, R. R. Detection of the
OH Band Fine Structure in Liquid Water by Means of New Treatment Procedure Based
on the Statistics of the Fractional Moments. Laser Phys. Lett. 2007, 4, 809–813.
(28) de Boer, T. J.; Backer, H. J. p-Tolylsulfonylmethylnitrosamide. Org. Synth. 1954, 34,
96.
(29) Wang, X.; Thottathil, J. K.; Poliaszek, R. P. Preparation of α-Chloroketones by the
Chloroacetate Claisen Reaction. Synlett 2000, 6, 902–904.
ACS Paragon Plus Environment