Continuous production of the diazomethane precursor JV-Methyl-N-nitroso-p- toluenesulfonamide: Batch optimization and transfer into a microreactor setup

Article abstract of DOI:10.1021/op900081g

The goal of this study was to develop a continuous multistep synthesis for the preparation of N-methyl-N-nitroso-p-toluene- sulfonamide (3, MNTS, Diazald) starting from p-toluenesulfonyl chloride (1), making use of microreaction technology (MRT). MNTS is an important precursor for diazomethane, a highly reactive and selective reagent for the production of pharmaceuticals and fine chemicals. Due to the properties of the successive reaction steps (exothermic reactions, use of toxic and highly reactive reagents), it was envisaged that MRT could provide advantages when compared to its batch-wise preparation. The research strategy included preliminary batch investigations, in which the effects of the solvent system, feed concentration, relative molar ratio, temperature, and residence time were established. Starting from these results, the reactions were translated into the MRT setup. As a result, the amidation of 1 to N-methyl-p- toluenesulfonamide (2) as the first reaction step is performed continuously in >90% yield and maximum space-time yields of up to 75 kg L^-1 h^-1. By making use of salting-out effects, the product separates nearly quantitatively in high concentrations in organic solution from the saline-waste stream. It is continuously converted to 3 by addition of NaNO₂ with quantitative conversions: yields of >90% and maximum space-time yields of up to 9 kg L^-1 h^-1. The method presented allows for the connection of the diazomethane precursor preparation to its continuous liberation by addition of a base, and conversion with a substrate, as previously demonstrated using MRT (Struempel, M.; On- druschka, B.; Daute, R.; Stark, A. Green Chem. 2008,10, 41).

Full text of DOI:10.1021/op900081g

Organic Process Research & Development 2009, 13, 1014–1021
Continuous Production of the Diazomethane Precursor
N-Methyl-N-nitroso-p-toluenesulfonamide: Batch Optimization and Transfer into a
Microreactor Setup
Michael Struempel, Bernd Ondruschka, and Annegret Stark*
Friedrich-Schiller UniVersity Jena, Institute for Technical Chemistry and EnVironmental Chemistry, Lessingstrasse 12,
07743 Jena, Germany
3
Abstract:
N-methyl-N-nitroso-p-toluenesulfonamide (MNTS), from which
it is liberated by addition of a base and reacted in situ. Most
The goal of this study was to develop a continuous multistep
synthesis for the preparation of N-methyl-N-nitroso-p-toluene-
sulfonamide (3, MNTS, Diazald) starting from p-toluenesulfonyl
chloride (1), making use of microreaction technology (MRT).
MNTS is an important precursor for diazomethane, a highly
reactive and selective reagent for the production of pharmaceu-
ticals and fine chemicals. Due to the properties of the successive
reaction steps (exothermic reactions, use of toxic and highly
reactive reagents), it was envisaged that MRT could provide
advantages when compared to its batch-wise preparation. The
research strategy included preliminary batch investigations, in
which the effects of the solvent system, feed concentration, relative
molar ratio, temperature, and residence time were established.
Starting from these results, the reactions were translated into the
MRT setup. As a result, the amidation of 1 to N-methyl-p-
toluenesulfonamide (2) as the first reaction step is performed
continuously in >90% yield and maximum space-time yields of
applications are thus restricted to small scale in specialized
4
apparatus, e.g. derivatizations for analytical purposes. For
industrial applications, it has been demonstrated that the
liberation of diazomethane from a precursor and its successive
5,6
conversion is feasible, but requires elaborate safety equip-
7
ment.
The most important precursor is MNTS (3), which in
addition to being a methylating agent can also be used for
8,9
nitrosations under mild conditions. This compound is less
10,11
12
hazardous, albeit still carcinogenic
and sensitizing. Being
a solid, it is more facile to store and to handle in a process than
is gaseous diazomethane.
Although 3 is commercially available, its on-demand pro-
duction for the direct liberation of diazomethane and its
conversion with the desired substrate would make the industrial
production independent from delivery specifications and fluc-
tuations in demand, in particular for small- and medium-sized
businesses (SMBs) producing fine chemicals and specialties.
A further advantage is the reduced cost for the provision of 3,
-1
-1
up to 75 kg L h . By making use of salting-out effects, the
product separates nearly quantitatively in high concentra-
tions in organic solution from the saline-waste stream. It is
since the starting materials 1, methylamine, NaNO
2
, and HCl
continuously converted to 3 by addition of NaNO
2
with
are relatively inexpensive basic chemicals.
Especially with regard to handling dangerous compounds,
microreaction technology (MRT) has shown great potential: the
quantitative conversions: yields of >90% and maximum
space-time yields of up to 9 kg L- h . The method
presented allows for the connection of the diazomethane
precursor preparation to its continuous liberation by ad-
dition of a base, and conversion with a substrate, as
previously demonstrated using MRT (Struempel, M.; On-
druschka, B.; Daute, R.; Stark, A. Green Chem. 2008, 10,
1
-1
high surface/volume ratio allows for efficient heat exchange,
13-16
even for highly exothermal reactions.
The system is
(
3) Diazald in chemical synthesis: http://www.sigmaaldrich.com/chemistry/
chemical-synthesis/technology-spotlights/diazald.html; accessed 10
January 2009.
41).
(
(
(
4) Diazald, MNNG and Diazomethane Generators, Aldrich Technical
Bulletin, AL-180.
5) Archibald, T. G. Large Scale Batch Process for Diazomethane.
(
Aerojet-General Corporation). U.S. Patent 5,817,778, 1997.
6) Archibald, T. G. Continuous Process for Diazomethane from an
N-Methyl-N-nitrosoamine and from Methylurea through N-Methyl-
N-nitrosourea. (Aerojet-General Corporation). U.S. Patent 5,854,405,
Introduction
Diazomethane (4, CH
selectivity under mild conditions in a plethora of reactions
2
2
N ) is valued for its reactivity and
1
997.
1,2
(7) Proctor, L. D.; Warr, A. J. Org. Process Res. DeV. 2002, 6, 884, and
references cited therein.
important for the production of pharmaceuticals and fine
chemicals. However, due to its rather unpleasant properties
(
(
8) Adam, C.; Garc ´ı a-R ´ı o, L.; Leis, J. R.; Norberto, F. J. Org. Chem.
2001, 66, 381.
9) Adam, C.; Garc ´ı a-R ´ı o, L.; Leis, J. R. Org. Biomol. Chem. 2004, 2,
(reactivity, toxicity, explosiveness, gaseous state), it is com-
1
181.
mercially not available as such, but in ‘deactivated’ precursor
forms such as N-methyl-N-nitrosourea and, more importantly,
(
(
10) Searle, C. E. Chemical Carcinogens, 2nd ed.; ACS Monograph 182.
American Chemical Society: Washington, DC, 1985.
11) B o¨ rzs o¨ nyi, M.; Sajg o´ , K.; T o¨ r o¨ k, G.; Pint e´ r, A.; Tam a´ s, J.; Kolar, G.;
Spiegelhalder, B. Neoplasma 1988, 35, 257.
*
Author for correspondence. E-mail: annegret.stark@uni-jena.de. Telephone:
(12) Young, H. S.; Wilson, N. J. E.; Winter, J.; Beck, M. H. Contact
Dermatitis 2004, 50, 313.
0
049 3641 948413. Fax: 0049 3641 948402.
(
(
1) Hopps, H. B. Aldrichimica Acta 1970, 3, 9.
2) Black, T. H. Aldrichimica Acta 1983, 16, 3.
(13) Panke, G.; Schwalbe, T.; Stirner, W. Synthesis 2003, 2827.
(14) Wille, B.; Gabski, H. P.; Haller, T. Chem. Eng. J. 2004, 101, 179.
1
014
•
Vol. 13, No. 5, 2009 / Organic Process Research & Development
10.1021/op900081g CCC: $40.75  2009 American Chemical Society
Published on Web 08/31/2009

Scheme 1. Multistep synthesis of the diazomethane precursor MNTS (3) and the in situ liberation of diazomethane, starting
from p-toluenesulfonyl chloride (1)
Scheme 2. Continuous microreactor setup used for the
equipped with several temperature and pressure sensors, and
amidation or nitrosation (P ) pressure sensor, T )
automatically shuts down if certain threshold values are
temperature sensor, M ) mixer device 1 or 2, S ) residence
overshot. Additionally, due to low spacio-temporal concentra-
tions in such enclosed, continuously operating setups, the risk
of exposure for operating personnel and environment is
dramatically reduced.
section 1 or 2)
With regard to research and development expenditure,
numbering-up instead of scaling-up may prove a valuable tool
to reduce development time. Furthermore, the modular setup
of MRT-based production lines allows for the flexible manu-
facture of chemicals in the same setup with short changeover
times.
toluene (98%), ethanol (98%), carbitol (98%)) or Merck
Chemicals (hydrochloric acid (25 and 32% aq), THF (98%),
acetonitrile (99%), dioxane (98%), diethyl ether (98%), ethyl
acetate (98%)), and were used as received.
All experiments were carried out in a fume-hood. Batch
experiments were carried out on 10 mM scale.
The continuous liberation of diazomethane from its precur-
sors and its conversion in a microreaction setup have been
demonstrated previously on the model compound benzoic
17,18
acid.
Hence, it is the goal of this contribution to provide an
efficient multistep synthesis for the diazomethane precursor
MNTS (3) making use of a continuous microreactor setup.
A literature search was conducted to determine the most
efficient synthetic strategy starting from readily available starting
materials, leading to the design of a two-step synthesis. Starting
from p-toluenesulfonyl chloride (1), N-methyl-p-toluenesulfona-
mide (2) is produced by amidation with methylamine, followed
by nitrosation to N-methyl-N-nitroso-p-toluenesulfonamide (3,
Continuous Setup. The working mechanism of the continu-
ous microreactor setup is shown in Scheme 2.
As mixing devices, either a commercially available Y-shaped
micromixer (mixer A, Little Things Factory GmbH, Ilmenau,
Germany, type ST MI018, 0.12 mL internal volume, channel
dimensions 1.4 × 1.8 × 125 mm disregarding static mixing
elements) was used. For comparison, a PTFE-capillary (Up-
church Scientific, Oak Harbor, USA, inner diameter 1.55 mm,
no mixing elements) connected to simple T-pieces (mixer B,
PTFE) was used. The fluid streams were generated by a
microdosing syringe pump MDSP3f, i-14 (Micro Mechatronic
Technologies GmbH, MMT, Siegen, Germany). The setup was
centrally controlled with a computer, for which a control and
automation program was developed with the visual program-
ming software Agilent Vee Pro 8.5. Additionally, the setup was
equipped with temperature sensors (0-150 °C, Typ K, Ni-CrNi,
Conatex GmbH, St.Wendel, Germany) and pressure sensors
2
MNTS) using NaNO under acidic conditions (Scheme 1).
In order to avoid precipitation and thus clogging in the
microreactor, preliminary data had to be obtained in batch
experiments, with emphasis on solvent choice, concentrations,
and relative ratios of reactants. In addition, the times required
to complete the reaction and heat development observed in batch
experiments were used as an indication for the initial setup (e.g.,
with regard to residence time and cooling unit, respectively).
These conditions were transferred to the microreactor setup,
where further improvements were achieved by alteration of
residence times and flow rates. Finally, the combination of the
single reaction steps to a stable continuous pilot plant was
accomplished.
(0-10 bar, Typ 26PC, Sensortechnics) programmed for shut-
down if the internal pressure reaches 5 bar.
Amidation. Discontinuous Amidation. 1 was dissolved in
the respective solvent (acetonitrile, THF, dioxane, toluene,
diethyl ether, carbitol, ethanol) in the required concentration
Experimental Section
Chemicals. Chemicals and solvents were obtained from
Sigma-Aldrich (p-toluenesulfonyl chloride (1, 97%), N-methyl-
p-toluenesulfonamide (2, 98%), N-methyl-N-nitroso-p-toluene-
sulfonamide (3, MNTS, 99%), sodium hydrogencarbonate
(0.5-7.0 M) and treated with methylamine (40% aq, 1-9 mol
equiv relative to 1). The temperature rise of the exothermic
reaction was monitored. The reaction was quenched by addition
of an excess aqueous HCl (32% aq, 1.1 equiv relative to the
theoretically remaining methylamine). Quantities of 10 µL were
taken as samples, diluted with acetonitrile and water to 0.02
(99.5%), sodium nitrite (40% aq), methylamine (40% aq),
(
(
(
(
15) J a¨ hnisch, K.; Hessel, V.; L o¨ we, H.; Baerns, M. Angew. Chem., Int.
Ed. 2004, 43, 406.
-1
mol L , and filtered (0.2 µm) where necessary. Analysis was
16) Mason, B. P.; Price, K. E.; Steinbacher, J. L.; Bogdan, A. R.; McQuade,
D. T. Chem. ReV. 2007, 107, 2300.
carried out by HPLC (Vide infra).
17) Struempel, M.; Ondruschka, B.; Daute, R.; Stark, A. Green Chem.
Continuous Amidation. Solutions of 1 in acetonitrile or THF
2
008, 10, 41.
18) Ferstl, W. F.; Schwarzer, M. S.; L o¨ bbecke, S. L. Chem.-Ing.-Tech.
004, 76, 1326.
(2-3 M) and methylamine (40% aq, 2.7-5.4 mol equiv relative
2
to 1) were dosed Via PTFE capillaries into the mixing device 1
Vol. 13, No. 5, 2009 / Organic Process Research & Development
•
1015

Scheme 3. Combination of the microreactor setups of the amidation and consecutive nitrosation for the continuous production
of 3 starting from 1 (P ) pressure sensor, T ) temperature sensor, M ) mixer device, S ) residence section)
-
1
(
mixer A or B, total flow rate 100-1080 mL h ), which was
connected to the first residence section S1 (7 cm for mixer A
0.07 mL), 5-90 cm capillary for mixer B (0.09-1.7 mL)).
3) was connected to the phase separator, and M1, M2, S1, and
S2 are hence extended by M3, M4, S3, and S4.
(
Analytics. Yields were determined using calibrations from
commercial standards on a Jasco HPLC equipped with a
Prontosil NC-04 (250 mm × 4 mm), 120-5-C18H 5 µm column
from Bischoff Chromatography, using isocratic 50% water: 50%
acetonitrile as eluent. Detection was achieved with an UV-diode
array detector at 236 nm.
In cases where biphasic solutions resulted, the quantitative
phase separation was often hampered because of the relatively
small scale, and cross-contamination of the respective other
phase occurred; the analytical error was thus estimated to be
10%.
The neutralization (32% aq HCl, 1.0-4.0 mol equiv relative
to 1) took place in the mixing device 2 equipped with a
residence section S2 (5.5 cm for mixer A (0.05 mL) plus 75
cm capillary within mixer A (1.42 mL), 75 cm capillary for
mixer B (1.42 mL)) and entered a phase separator (separating
funnel) after total residence times of approximately 5-60 s.
Temperature sensors were placed in front of S1 and S2, and
behind S1. Pressure sensors were placed in the feed inlets of
methylamine and HCl. When required, cooling was achieved
by placing S1 and S2 in a stirred water bath or thermostat. The
volumes of the phases obtained were determined and 10 µL
samples taken for analysis.
Nitrosation. Discontinuous Nitrosation. 2 was dissolved in
the respective solvent (acetonitrile, THF) at concentrations
between 1 and 4 M, and treated with NaNO (40% aq, 1.3-3.1
mol equiv relative to 2). Acid (HCl, 25% aq, 2.0-3.0 mol equiv,
or acetic acid, 1.3-10.0 equiv relative to 2) was added dropwise.
In all addition steps, the temperature was monitored. Cooling
was not applied to quantify heat development. Quantities of
Results and Discussion
Amidation. In the literature, various reports describe the
19-21
discontinuous synthesis of 2
with yields >70% on laboratory
2
scale. Due to the exothermal behavior of the reaction, cooling
is applied and the amine added slowly. An excess of amine is
required to scavenge HCl formed during the reaction.
However, these batch procedures are not easily transferred
to MRT, since oftentimes liquid-solid heterogeneous reaction
mixtures are used or precipitation of the product occurs.
Therefore, a simple procedure had to be developed for the
amidation of 1 in a completely liquid environment under mild
conditions. It should be mentioned here that the chosen method
is clearly not applicable for large-scale batch synthesis, due to
the fast reaction in conjunction with high heat development.
In the investigation presented herein, a commercially avail-
able aqueous solution of methylamine (40%) was used through-
out for the amidation. Hence, finding a second solvent to
homogenize and avoid solid precipitation was required, since
5
5
0 µL were taken as samples and neutralized by addition of
00 µL of a saturated NaHCO solution. The sample was diluted
3
to 3 mL with acetonitrile, filtered (0.2 µm), and analyzed by
HPLC (Vide infra).
In cases where a solution of 2 was used directly from the
amidation, the addition of slightly overstoichiometric amounts
of HCl relative to the theoretically remaining methylamine was
sufficient prior to the nitrosation.
Continuous Nitrosation. The same setup as for the continu-
ous amidation was used. In the mixing device 1 equipped with
residence section S1 (7 cm for mixer A (0.07 mL), 5 cm
capillary for mixer B (0.09 mL), a solution of 2 (0.5-3.0 M in
2
is a solid and insoluble in water.
The excess of amine was the second aspect: On the one hand,
a 2-fold excess is theoretically required to drive the reaction to
2
THF or acetonitrile) was mixed with NaNO (40% aq, 3 mol
equiv relative to 2). HCl (25% aq, 2.6 mol equiv relative to 2)
was added in the mixing device 2 before the residence section
S2 (5.5 cm within mixer A (0.05 mL) plus either 200 or 260
cm capillary (3.80 or 4.91 mL) after mixer A, and 200 or 260
cm capillary for mixer B (3.80 or 4.91 mL)). The total flow
rate of this reaction was between 90-540 mL h , translating
into a residence time of 33-200 s. Sample preparation was
performed as in the discontinuous experiments.
completion and avoid the formation of N-methyl-N,N-di(p-
22,23
toluenesulfonyl)amide as a side product.
On the other hand,
excessive amounts of amine are disadvantageous from an
environmental and economic point of view.
-1
(19) de Boer, Th. J.; Backer, H. J. Organic Syntheses; Wiley & Sons: New
York, Collect. Vol. 4, 1963; p 943.
(
20) Buchbauer, G.; Pernold, W.; Ittner, M.; Ahmadi, M. T.; Dobner, R.;
Reidinger, R. Monatsh. Chem. 1985, 116, 1209.
(
(
21) Townsend, C. A.; Theis, A. B. J. Org. Chem. 1980, 45, 1697.
22) Hinsberg, O.; Kessler, J. Chem Ber. 1905, 38, 906.
In cases where the solution of 2 from the amidation was
used directly (i.e., the organic phase), a second setup (Scheme
(23) Marvel, C. S.; Gillespie, H. B. J. Am. Chem. Soc. 1926, 48, 2943.
1
016
•
Vol. 13, No. 5, 2009 / Organic Process Research & Development

Table 1. Solvent screening in the batch-wise conducted
amidation of 1 to 2 using methylamine: phase behavior
was obtained. Overall, the reaction outcome appears to be
independent of the solvent used.
max concentration no. of liquid
The influence of the ratio of methyl amine relative to 1 was
tested (Table 2). In any of the solvents used, the yield increased
with increasing ratio, with quantitative yields obtained if an
excess of 2.5-3 mol equiv is used. For stoichiometric composi-
tions (2 mol equiv of methylamine) conversions are quantitative,
but the yield of 2 is <100%. This phenomenon is not fully
understood, but may have its origin in the formation of
N-methyl-N, N-di(p-toluenesulfonyl)amide, which has been
-1 a
b
solvent
carbitol
of 1 [mol L ]
phases
ppt
2.5
1
permanent after
addition of HCl
no
dioxane
acetonitrile
THF
diethyl ether
toluene
3.0
3.0
3.0
0.5
1.0
2
2
2
2
2
no
no
temporary
temporary;
permanent after
addition of HCl
no
22,23
described to occur at substoichiometric compositions
but
EtOH
1.0
1
was not detected with the HPLC method currently employed.
The time dependence was next investigated. It was found
that the amidation reaction takes place immediately and is
completed already after a few seconds, as can be seen from
the first exotherm occurring (Figure 1). The second exothermal
event observed relates to the heat development during the
acidification with HCl. The sensitivity of the thermosensor
detects very precisely the progress of the reaction. This is
reflected in the constant temperature rise (to about 70 °C for 1
M solutions of 1 in carbitol, and to about 90 °C for 3 M
solutions of 1 in dioxane) during the amidation step, irrespective
of the molar equivalents of methylamine used, while the heat
development of the neutralization is proportional to the excess
methylamine. These findings allow for two conclusions for the
transferal of the reaction to the continuous MRT setup. First,
due to the high reaction rates, very short residence times are
a
b
For other conditions and yields, see Table 2, entries 1-7. Number of liquid
phases formed during amidation; no change after acidification with HCl, if not
otherwise indicated.
Table 2. Selected results of the batch-wise amidation
including acidification with HCl) of 1
(
1
in org.
mol equiv
of MeNH
2
rel. to 1
total
yield
[%]
solvent
[mol L ]
-1
entry
solvent
carbitola
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
2.5
3.0
3.0
3.0
0.5
1.0
1.0
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
3.0
3.0
3.0
3.0
3.0
3.0
3.0
1.5
2.2
3.7
1.3
2.2
3.4
0.7
1.5
2.2
3-9
96.4
a,b
dioxane
86.4
97.4
90.8
93.7
98.2
91.7
44.9
65.3
98.5
61.7
95.2
99.5
27.5
57.6
82.4
90-100
a
acetonitrile
a
THF
a
diethyl ether
a
toluene
a
EtOH
diethyl ether
diethyl ether
diethyl ether
toluene
required. Second, the efficient heat exchange properties known
0
1
2
3
4
5
6
7
13-16
for microreactors
will be beneficial to control these fast
and highly exothermic reactions.
toluene
toluene
As pointed out above (Table 1), some solvents which are in
principle water-miscible, such as acetonitrile and THF, showed
liquid-liquid phase separation at certain ratios of methylamine/
1, which was further enhanced by the addition of HCl in the
quench. This salting out effect can be beneficially used in the
process to separate the product-rich organic phase from the sa-
line waste phase and introduce it directly into the next step, the
nitrosation. Several investigations relating to the phase behavior
in dependence of the concentration of 1, methylamine and the
solvent gave a complex picture. For example, the mitigation in
yields for acetonitrile, shown in Figure 2 for >4 mol equiv of
methylamine is due to this phenomenon; the remaining product
traces are found in the aqueous phase.
Other water-soluble solvents, such as ethanol and carbitol,
did not result in biphasic systems. Although these also gave 2
in high yields, they were not further investigated as the
maximum initial concentrations of 1 are rather low (see Table
1). Furthermore, an additional separation step would be required
in these instances to reduce the saline load.
c
acetonitrile
c
acetonitrile
c
acetonitrile
c
acetonitrile
a
Results of experiments displayed in Table 1, T ) autothermal, t ) 10-15
min. b Experiments refer to experiments shown in Figure 1. Experiments refer to
c
experiments shown in Figure 2, yields quoted are for the organic phase only; yield
of entry 17 for molar equivalents of methylamine between 3-9.
Discontinuous Amidation. A solvent screening was con-
ducted to determine solvents which allow for high reactant
concentrations while avoiding the formation of temporary or
permanent precipitates during the amidation and the successive
acidification. Table 1 shows selected results of these investiga-
tions. Up to 3 M solutions of 1 can be processed without the
occurrence of precipitates in dioxane, acetonitrile, and THF,
while lower concentrations were achieved with carbitol and
ethanol. Temporary precipitation occurred in toluene and diethyl
ether (e1.0 M in toluene, e 0.5 M in diethyl ether), while
permanent precipitations were observed at higher concentrations.
In the case of carbitol and toluene, solids formed after the
acidification with HCl. Except for carbitol and ethanol, a
biphasic liquid-liquid system was obtained. In all instances,
boiling of the solvent occurred at concentrations of 1 > 2.5 mol
From these investigations, optimal conditions were derived:
high feed concentrations are achieved in acetonitrile or THF
(<3 M 1), and at molar excesses of 2.5-3 equiv of amine
relative to 1, quantitative yields result. In this instance, the total
volume was reduced by 35-40% aqueous solution, whereas
>90% of 2 remained in the organic phase. Further analysis of
this phase revealed a water content <5% (automated Karl-
Fischer titration), and the concentration of 1 was below the
-1
L
due to the exothermic nature of the reaction.
Table 2 (entries 1-7) shows that high yields (90-98%) of
result in this fully selective reaction in all cases with the
2
-1
exception of dioxane, in which a somewhat lower yield of 86%
detection limit of the HPLC analysis (<0.03 mol L ) in the
Vol. 13, No. 5, 2009 / Organic Process Research & Development
•
1017

Figure 1. Temperature profile of the amidation of 1 with methylamine (40% aq) followed by neutralization with HCl (32% aq).
Left: 1 M 1 in carbitol; right: 3 M 1 in dioxane.
Table 3. Five-hour run of the continuous amidation of 1 in
THF with methylamine (2.7 equiv rel. to 1)
operation time [h]a
>5
9
0 cm (1.7 mL) and
S1, S2
75 cm (1.4 mL)
PTFE-capillaries
mixer A
mixer B
3.0
60
11.6
42
12.0
mixer amidation
mixer acidification
c(1) [mol L ]
flow rate [mL h ]
c(MeNH
flow rate [mL h ]
c(HCl) [mol L ]
flow rate [mL h ]
averaged total yield [%]
average experimental flow rate [mL h ] 107 ( 8
-
1
-
1
-1
2
) [mol L ]
-1
-
1
-1
24
104 ( 10
Figure 2. Effect of the amount of methylamine on the yield of
in the organic phase (0.5 M 1 in acetonitrile, autothermal, 10
-1
2
∆
∆
p [bar]
0.2 ( 0.1
min).
b
T [°C]
0.5 ( 0.2
-1
-1
STY [kg L h ]
11.2 ( 1.0
0.04 ( 0.003
-
1
capacity [kg h ]
a
Values averaged over 17 samples, each drawn after 20 min of continuous
b
operation. Setup temperature controlled with thermostat at 23 °C, temperatures
measured directly after M1 and after residence section 1, outside thermostat fluid.
yield (90-100%) were observed in an autothermal setup,
allowing for extremely low calculated residence times (5-60
s). However, considerable fluctuations resulted due to the boiling
of the solvent (formation of gas bubbles). THF gave results
similar to those of acetonitrile in all cases.
The reaction was found to be independent of the mixer type
(
Y-shaped micromixer, T-piece). However, with mixer B (T-
Figure 3. Dependence of the yield of 2 obtained in the
continuous amidation of 1 with methylamine on the theoretical
flow rate at various excesses of methylamine (2.5 M 1 in
acetonitrile).
piece), the variation of the residence sequence S1 is more
flexible, as the length of the PTFE-capillary is easily varied. In
order to reduce the fluctuations of the flow rate (due to boiling
of the solvent), S1 of the capillary was extended to 90 cm (1.7
mL) and cooled using a water bath. It was found that the
temperature decreases to about 20 °C at the end of S1, with
the flow being steadier. As for the residence section S2, a
capillary length of 75 cm (1.4 mL) was maintained to allow
for the control of the exothermic HCl quench. At the highest
flow rate tested (1080 mL h , Figure 3), a space-time yield
(STY) of approximately 75 kg L
translates into a capacity of 0.23 kg h of this setup with a
reaction volume of 3.1 mL.
samples taken. The chloride content was determined to be 3.0%
of the total chloride derived from 1, i.e. before acidification
3
(automated chloride titration against AgNO ).
Continuous Amidation. During the transfer of the optimized
conditions obtained in batch mode, it occurred that with 3.0 M
solutions of 1 in acetonitrile and THF, slightly turbid solutions
resulted at the outlet of the microreactor, which were, however,
processable without clogging. Hence, the amount of methyl-
amine was varied between 2.7-5.4 mol equiv with respect to
-1
-1
-1
h
is obtained, which
-1
1
under continuous reaction conditions, and the effect of the
The stability of this process step was tested in a 5-h
continuous run in a temperature-controlled setup (Table 3) at a
total flow rate of 126 mL h . From the temperature profile
-1
adjusted (theoretical) flow rate (100-1080 mL h , Figure 3)
was investigated. Within the analytical error, no effects on the
-1
1
018
•
Vol. 13, No. 5, 2009 / Organic Process Research & Development

obtained, a start-up time of 20 min was estimated before the
steady-state of the system would be achieved. It should be noted
that the standard deviation results from the fact that a biphasic
mixture results, which is difficult to quantitatively separate on
a relatively small scale. An averaged experimental flow rate of
-1
1
07 mL h was determined by measuring the volume output
every 20 min. It deviates from its theoretical flow rate due to
nonideal mixing behavior of the solutions. The average yield
obtained during this experiment is quantitative.
In summary, the transfer of the optimum amidation reaction
conditions into the continuous microreactor setup was ac-
complished without problems. Either acetonitrile or THF can
be used as solvents, with 2.5-3.0 M concentrations of 1, 2.7
mol equiv of methylamine, and 1 mol equiv of HCl relative to
2
Figure 4. Nitrosation of 2 to 3 with NaNO : Dependence of
the total yield on the reaction time for different protocols (3 M
2 in THF, 7 mol equiv of acetic acid or 2.6 mol equiv of HCl,
1. As in the discontinuous experiments, a biphasic system
2
and 3 mol equiv of NaNO relative to 2).
resulted, in which >90% yield (at 100% conversion) is isolated
with the organic phase (containing, beside the product, <5%
water and 3% chloride). Both mixer types gave comparable
results. Although the reaction is extremely fast and does not
require long residence times, the addition of a water-cooled
residence section S1 (90 cm) and S2 (75 cm) to allow for
efficient cooling and, hence, reduced fluent fluctuation is
With acetic acid, which is mainly used in literature,24,25 a 7-fold
excess is required to obtain best performance, whereas with
HCl, only a 2.6-fold excess is needed (1 M 2 in acetonitrile,
3
-fold excess NaNO
ations as well as the atom efficiency
CH CO Na-waste) favor HCl. Additionally, with acetic acid, a
2
relative to 2, 4 h). Economic consider-
28,29
(NaCl-waste vs
3
2
-1
advisible. With an optimum flow rate (1080 mL h ) and a
calculated residence time of 13 s, a maximum STY of up to 75
single liquid phase and frequent precipitations also restrict the
transfer into MRT. HCl instead is a driver for liquid-liquid
phase separation, especially in water-miscible solvents. Only
HCl as 25% aqueous solution (or lower) gives precipitation-
free reactions, because otherwise NaNO becomes insoluble.
2
An investigation of the dependency of the reaction time (3
M 2 in THF, 7 mol equiv of acetic acid or 2.6 mol equiv of
HCl, and 3 mol equiv of NaNO relative to 2) showed, that
2
with HCl the product is formed immediately (<1 min), whereas
with acetic acid, the product is formed quantitatively only within
-1
-1
kg L
h
is achieved. A 5-h continuously conducted experi-
ment showed that the system is sufficiently stable for small-
scale production.
Nitrosation. For the nitrosation of 2 to 3, a method was
19,24,25
adapted from literature
using aqueous acidic NaNO
2
.
Conventionally, the acidic environment for NaNO
by addition of acetic acid, and in only a few cases has HCl
been reported.
2
is provided
26
Discontinuous Nitrosation. In batch experiments, the nitro-
sation of 2 was optimized with respect to the type of acid, the
15 min, despite the higher acid concentration (Figure 4). Only
after very long reaction times (several days), does decomposition
of the product 3 occur, thus allowing for the production of the
diazomethane precursor somewhat in advance to its usage.
As in the amidation, the reaction mixture separates into a
biphasic liquid-liquid system in both solvent systems. The
slightly yellow organic phase contains the product 3 (about 90%
of the total yield).
In order to run the process at the highest possible substrate
concentrations of 2 in THF or acetonitrile, concentration studies
were carried out, which showed that up to 4 M of 2 can be
converted, but at >1 M, the solution becomes turbid, and 3
precipitates slowly at >3 M. Hence, the highly concentrated
product solution from the amidation can be used without having
to change solvents. This was tested in batch experiments (one-
pot synthesis); although nitrous gases developed and led to
foaming, consistent yields were obtained. The reaction outcome
reaction time, the type of solvent, the ratio of acid and NaNO
2
with respect to 2, and the concentration of the reactants in
solution.
27
According to literature, the nitrosation is carried out at 0
C. However, in order to obtain information about the heat
°
development during the reaction, experiments were carried out
autothermally while recording the solution temperature. In all
instances, the temperature rise was uncritical with <30 °C.
In the optimization experiments, acetonitrile and THF were
used as these were the most efficient solvents in the preceding
reaction step (Vide supra).
In general, an excess NaNO
nitrosation, due to partial decomposition to nitrous gases such
as NO and NO . Investigations in the range of 1.3 to 3.1 mol
equiv of NaNO relative to 2 (1 M 2 in acetonitrile, 7-fold
excess of acetic acid relative to 2, 4 h) showed that a 2.5-3.0-
fold excess NaNO to 2 gives quantitative conversions. Similar
considerations affect the ratio of acid to 2, because NaNO
2
is needed for complete
2
2
is independent of the sequence of adding acid and NaNO
. However, mixing NaNO with acid before addition of 2 is
disadvantageous, as NaNO decomposes unselectively prior to
2
to
2
2
2
2
+
2
requires acidic conditions to release the active species NO .
reacting with 2. This is evidenced by the formation of larger
amounts of gases (bubbles and brown fumes).
In conclusion, optimal batch conditions for the transfer into
MRT, considering economic and ecological aspects, were
(
(
24) White, E. H. J. Am. Chem. Soc. 1955, 77, 6008.
25) Garcia-Rio, L.; Leis, J. R.; Moreira, J. A.; Norberto, F. J. Chem. Soc.,
Perkin Trans. 2 1998, 1613.
(
26) Iley, J.; Norbert, F.; Rosa, E.; Cardoso, V.; Rocha, C. J. Chem. Soc.,
Perkin Trans. 2 1993, 591.
(28) Trost, B. M. Angew. Chem., Int. Ed. 1995, 34, 259.
(29) Sheldon, R. A. Pure Appl. Chem. 2000, 72, 1233.
(
27) Organikum, 21st ed.; Wiley-VCH: Weinheim, 2001; p 392.
Vol. 13, No. 5, 2009 / Organic Process Research & Development
•
1019

Table 4. Five-hour run of the continuous nitrosation of 2 in
THF with NaNO
Combination of Amidation and Nitrosation. The two
continuous synthetic steps were connected to a single process
step, advantageously exploiting the phase separation occurring
after the amidation to reduce the overall saline load. All
residence sections were cooled in a water bath (20 °C). 1 is
mixed with methylamine and reacted in residence section S1,
HCl is added to quench the amidation (S2). The reaction mixture
separates in a separation funnel. The organic top phase
containing most of the product 2 (2.7-3 M solution) is
continuously fed into M3, where it is mixed with the aqueous
2
operation time [h]a
S1
S2
>5
14.5 cm (0.3 mL)
12.5 cm (0.12 mL)
within mixer A and
2
00 cm (3.8 mL)
PTFE-capillaries
mixer B
mixer A
3.0
30
7.5
30
7.5
30
101 ( 8.5%
90
1.5 ( 0.5
8.0 ( 0.5
4.6 ( 1.0
0.02 ( 0.005
mixer for 2 with NaNO
2
mixer for addition of HCl
-
1
c(2) [mol L ]
-1
flow rate [mL h ]
-1
c(HCl) [mol L ]
2
NaNO solution, followed by the introduction of HCl into M4
-1
flow rate [mL h ]
(Scheme 3). Using this methodology, an overall yield of 3
starting from 1 of 70-80% has been achieved. The experimental
setups described in Tables 3 and 4 were used for this
investigation.
A calculation of the chemical costs at kg-scale, taking into
account net prices of fine chemical catalogues, indicated a price
of approximately 40 Euros/mol MNTS, disregarding personnel
costs and apparatus expenditure.
-
1
c(NaNO
2
) [mol L ]
-1
flow rate [mL h ]
average total yield [%]
theoretical flow rate [mL h ]
-1
∆
p [bar]
T [°C]
b
∆
-1
-1
STY [kg L h ]
capacity [kg h ]
-1
a
Values averaged over 12 samples, each drawn after 20 min of continuous
Combination of Nitrosation with Diazomethane Libera-
tion and Conversion. Previous work has focused on the
liberation of diazomethane from its precursor MNTS by addition
of an alcoholic KOH solution, and successive conversion of a
b
operation. Setup not cooled, temperatures measured directly after M1, M2 and
after residence section 2.
determined to be 3 M solutions of 2 in THF or acetonitrile, 2.6
17
model compound in a continuously operated micoreactor. It
has been demonstrated that yields of 75% and a production rate
of 2.5 mol product (the model compound methyl benzoate) per
day are achievable using a single efficient micromixer, with
mol equiv of HCl and 2.5-3 mol equiv of NaNO
2
relative to
2, at a reaction time <10 min. These conditions lead to
quantitative conversion of 2 and highest yields (>90%) of 3 in
the organic phase.
-1
flow rates between 100 and 800 mL h . In that study, the
Continuous Nitrosation. The optimized conditions of the
batch experiments (2.5 mol equiv of HCl, 2.5 mol equiv of
NaNO relative to 2) were adapted to the continuous mode.
2
diazomethane liberation and conversion was best achieved in
carbitol, since in THF, precipitation occurred. Precipitation was
also a problem in carbitol solution, however, at maximum
All reactions were carried out autothermically.
-1
MNTS concentrations of <0.4 mol L , a homogeneous reaction
Experiments were conducted to investigate the effect of the
residence time (by variation of the flow rate between 90 and
mixture was maintained.
Hence, to prove the concept, the concentrated organic feed
-1
5
7
40 mL h ). It was found that the total yield increased from
0% to 100% for residence times between 30 and 100 s and
-1
from the nitrosation (2.5 mol L ) was washed with a saturated
-1
3
NaHCO solution and diluted with carbitol to 0.4 mol L . This
remained constant thereafter (up to 210 s tested). For a given
solution was then used in the methylation of the model
compound benzoic acid, under the conditions described previ-
-1
preadjusted flow rate of 180 mL h , the required length of the
residence section S2 is hence theoretically >260 cm. Under these
17
ously. Without further optimization, quantitative conversions
-1
-1
-1
conditions, STYs of up to 9 kg L
h
(capacity: 0.04 kg h )
of MNTS and yield of 50-60% of the desired methyl benzoate
were obtained, demonstrating the connectivity of these reaction
steps.
Therefore, future research can focus on the fine-tuning of
the successive reaction steps in conjunction with the transfer
to industrially relevant reactions of diazomethane.
were achieved. Experimental flow rates and residence times
could not be determined, as the formed gases highly accelerated
the flow and led to fluctuation and pressure rise.
Table 4 shows the conditions and results obtained during a
5
-h continuous production of 3 at a theoretical flow rate of 90
-1
mL h , demonstrating the good stability of the system, even
if fluctuations of the flow occur due to the formation of nitrous
gases. Quantitative yields are obtained.
Conclusions
Using a microreactor adds the advantage that liberation of
hazardous gases can be controlled, and with the closed system,
no cooling is required to reduce the amount of nitrous gases to
a minimum.
In conclusion, the transfer of the optimized batch conditions
to MRT leads to a high-performance continuous process with
In conclusion, an efficient two-step synthesis for the diaz-
omethane precursor MNTS has been developed, making use
of a continuous microreactor setup. Starting from readily
available starting materials, p-toluenesulfonyl chloride is con-
verted by amidation to N-methyl-p-toluenesulfonamide, fol-
lowed by nitrosation to N-methyl-N-nitroso-p-toluenesulfona-
-
1
-1
-1
good STYs of 9 kg L
h
and a capacity of 0.04 kg h at a
mide using NaNO under acidic conditions. Deliberate solvent
2
-1
optimum flow rate of 180 mL h with the presently used setup.
Furthermore, these conditions also match the resulting product
phase of the preceding amidation. The combination to a closed
process unit is thus the logical consequence.
choice allows for the conversion of highly concentrated solutions
at yields >90% in both single steps, with the added advantage
of liquid-liquid phase separation due to salting out effects. This
phenomenon reduces the saline load, and hence the risk of
1
020
•
Vol. 13, No. 5, 2009 / Organic Process Research & Development

precipitation to a minimum. Batch-wise extraction of the
aqueous phases may be used to recover the remaining traces
of products from each resulting aqueous phase. The methodol-
ogy presented allows for the connection of the diazomethane
precursor preparation to its liberation by addition of a base, and
conversion with a substrate, as previously demonstrated using
Overall, this methodology now allows for the production
and conversion of the diazomethane precursor MNTS 3 on site
and on demand. This possibility reduces the risk of chemical
storage considerably and achieves independence from the
delivery times for MNTS, especially for SMBs.
17
MRT.
Acknowledgment
Both the amidation and nitrosation are exothermic reactions,
which are advantageously carried out using MRT and its
excellent heat dissipation properties. This aspect, due to the
inherently low spacio-temporal concentrations, renders the
manufacture of the diazomethane precursor operationally safe.
In the light of the recent shortage of acetonitrile on the world
market, flexible solvent choice has become more important with
respect to economic MNTS production. Hence, the similar
performance of both acetonitrile and THF with regard to yields,
maximum concentrations of feeds, and phase separation is
beneficial.
We thank M. Sellin, J. Zimmermann, C. Palik, and W.
F a¨ hndrich for continuous support in the laboratory. For financial
support, M.S. thanks German Federal Environmental Founda-
tion (DBU, especially Dr. Hempel) for a Ph.D. scholarship. A.S.
thanks the German Research Foundation (DFG) for financial
support within the Priority Programme Ionic Liquids (SPP
1191).
Received for review April 3, 2009.
OP900081G
Vol. 13, No. 5, 2009 / Organic Process Research & Development
•
1021