Nitration chemistry in continuous flow using fuming nitric acid in a commercially available flow reactor

Article abstract of DOI:10.1021/op200055r

The paper will describe the use of flow chemistry for scaling up exothermic or hazardous nitration reactions. Such reactions often cause time delays to the delivery of larger batches of intermediates or final compounds for medicinal chemistry projects, be

Full text of DOI:10.1021/op200055r

TECHNICAL NOTE
pubs.acs.org/OPRD
Nitration Chemistry in Continuous Flow using Fuming Nitric Acid
in a Commercially Available Flow Reactor
Cara E. Brocklehurst,* Hansj€org Lehmann, and Luigi La Vecchia
Global Discovery Chemistry, Novartis Institutes for Biomedical Research, Preparations Laboratories, Klybeckstrasse 141,
4057 Basel, Switzerland
ABSTRACT: The paper will describe the use of flow chemistry for scaling up exothermic or hazardous nitration reactions. Such
reactions often cause time delays to the delivery of larger batches of intermediates or final compounds for medicinal chemistry
projects, because considerable time is required for safety evaluation and, if necessary, modification of the procedure so that it can be
scaled-up and run in a safe manner. A commercially available continuous flow reactor was used in the scale up of three challenging
nitrations including a reaction involving a potentially explosive mixture of acetic acid and fuming nitric acid, with a productivity of
97 g/h.
’ INTRODUCTION
the speed at which the heat is released from a strongly exothermic
reaction.
In recent years, continuous flow chemistry has emerged as a
useful technology for research and development chemists within
the pharmaceutical sector.1ꢀ4 Highly efficient mixing is com-
bined with superior heat exchange and narrowly distributed, well-
defined, reaction times. One major advantage of flow reactors, in
comparison to reactions in batch mode, is the small reaction
volume, which allows the running of highly exothermic or
hazardous reactions in a safer manner to provide >100 g
quantities. Flow chemistry automation enables chemists to
rapidly find optimized conditions for a reaction on trial scale.
The refined conditions can then be directly translated to the
production of larger amounts by increasing the volume of the
reactor (and hence flow rate), using multiple reactors in parallel
(numbering-up), and/or prolonging the running time of the
system. The fast transfer of processes is in stark contrast to the
typical scale-up of batch chemistry, which often requires addi-
tional time for the optimization of parameters.
Differential scanning calorimetry (DSC) can provide valuable
information about the onset temperature of a reaction or
decomposition, and how much energy is released per gram of
product. According to our internal safety guidelines, any reaction,
starting material, or product showing an exotherm over 300 J gꢀ1
is considered to require further safety assessment as it could
result in an uncontrollable temperature rise within a reactor
vessel. Performing such reactions in a microreactor or a con-
tinuous flow reactor is advantageous because the reaction takes
place in narrow tubes, and any heat evolved is produced in a small
reaction volume and can be more rapidly dissipated. Indeed, the
heat-transfer rate between the reactor and the surrounding
medium can be magnitudes of orders faster compared to that
for a batch reactor.6 In contrast to this, the scale-up of exothermic
reactions in batch can be problematic due to low mixing
efficiency and poor heat transfer, leading to the formation of
hot spots and, as a further consequence, to the formation of
undesired by-product.
Such technology is attractive for use in our Preparation
Laboratories to aid the often time-critical supply of intermedi-
ates, building blocks, and final compounds in multihundred gram
amounts to the Medicinal Chemistry laboratories of Novartis
Research.
’ LITERATURE EXAMPLES OF NITRATIONS IN CON-
TINUOUS FLOW
Nitration of aromatic compounds is not only fast but is often
accompanied by a strong exotherm upon addition of the nitrating
acid mixture, making these reactions interesting targets for
adaptation to continuous flow. It is known that continuous
processing can be used to control the reaction temperature and
any selectivity of a given reaction with more precision. Nitration
reactions are often performed using mixtures of nitric and acetic
acids (sometimes with acetic anhydride) which are known to be
potentially explosive.7,8 Use of excess nitrating agent often results
in polynitration, meaning that the accurate reaction times and
’ CONTROLLING ENERGY RELEASE USING FLOW
CHEMISTRY
The Novartis Preparation Laboratories became interested in
continuous flow chemistry in order to scale up exothermic or
hazardous reactions. Hazardous chemistry often causes us delays
and requires extensive safety assessments in order to scale up the
batch reaction in a safe manner or to find alternative reaction
conditions so that the hazard can be minimized when the
reaction is performed on a larger scale in a batch reactor.5 In
order to fully understand the advantages of adapting batch
chemistry to continuous flow, it is insightful to consider the
kinetics of the reaction and the potential energy released in the
case of a runaway reaction. Another important consideration is
Special Issue: Safety of Chemical Processes 11
Received: March 4, 2011
Published: April 29, 2011
r
2011 American Chemical Society
1447
dx.doi.org/10.1021/op200055r Org. Process Res. Dev. 2011, 15, 1447–1453
|

Organic Process Research & Development
TECHNICAL NOTE
Scheme 1. Nitration of 8-bromo-1H-quinolin-2-one 1 in
batch and under continuous flow
scaled to 25 tonnes by DSM and subsequently performed under
cGMP using a Corning microreactor system.17 The sensitive step
in this reaction was not the nitration itself but the risk of
exothermic decomposition of the nitrated product. The nitration
utilized 65% nitric acid and involved the quenching of the outlet
by stepwise addition of water and sodium hydroxide streams.
’ RESULTS AND DISCUSSION
The following section will outline the adaptation of three
nitration examples to continuous flow and demonstrate the high
productivity we have observed using relatively simple, commer-
cially available equipment.18 The first nitration we investigated is
outlined in Scheme 1, showing the original batch conditions for
the nitration of 8-bromo-1H-quinolin-2-one 1, with the subse-
quent optimized flow conditions beneath. DSC of a mixture of
fuming nitric acid and acetic acid showed a strong decomposition
releasing 1374 J gꢀ1 with an exotherm starting at 130 °C,
presumably from the decomposition of acetyl nitrate. DSC of
the reaction mixture after 50% addition of nitric acid showed a
strongly exothermic decomposition starting at 105 °C, releasing
633 J gꢀ1, and after 100% addition, a second strongly exothermic
decomposition releasing 1620 J gꢀ1 with an onset at 225 °C.
Under adiabatic conditions, the total energy released corre-
sponds to a temperature rise of 840 °C. Reaction calorimetry
also indicated that the reaction demonstrated the potential for
accumulation.
The original nitration reaction had been performed on small
scale by the Medicinal Chemistry team with the acidic solution at
reflux (approximately 115 °C). The proximity of the reaction
temperature to the onset of decomposition was unacceptable for
scale up. For safety reasons, the scale of this reaction was limited
to <0.1 mol for batch reactions, and we sought alternatives to be
able to run this reaction on the required 250-g scale.
In adapting this nitration chemistry to continuous flow, not
only would a relatively small quantity of high-temperature
reaction mixture be present in the reactor at any one time, but
also, should a temperature rise occur, the machine must be able
to be manually stopped or will automatically stop if a rapid
change in pressure is detected. Once product is formed in the
reactor, it is continuously removed and pumped into the quench
solution. There is therefore relatively little time for the nitrated
product to undergo decomposition compared to the correspond-
ing batch reaction.
As is often the case for adapting batch chemistry to flow, we
started our trials with small-scale batch experiments. Not only
does this provide us with information about the reaction rate
(which was monitored by HPLC), but it also provides us with an
opportunity to check whether the reaction is homogeneous.
Figure 1 shows the reaction conversion at three different con-
centrations and the fourth experiment using diluted nitric acid.
The rate of reaction using 65% nitric acid was deemed too slow
for adaptation to flow: the reaction showing only approximately
50% conversion after one hour. However when the concentra-
tion was increased from 0.2 to 1 M, the reaction was complete in a
matter of minutes in small-scale batch experiments.
specific stoichiometry achievable in flow is particularly attractive.
The precision of the reaction times in flow means that the
nitrating mixture can be quenched before overreaction takes
place. There are relatively few literature examples of nitration
reactions which have been performed under continuous condi-
tions. One reason is because of the lack of commercially available
lab-scale equipment able to pump the corrosive acids that are
required.
Schwalbe et al. used a stainless steel CYTOS microreaction
system from Cellular Process Chemistry, for the nitration of a key
intermediate in the synthesis of sildenafil, the active ingredient of
Viagra.9,10 By adapting this chemistry to flow, not only was the
handling of the highly energetic reaction made easier, but also the
selectivity of the reaction was improved. Exothermic decomposi-
tion via decarboxylation was avoided by utilizing accurate
temperature control within the reactor. Further examples of nitra-
tions in continuous flow include the nitration of toluene,10ꢀ13
phenol,14 benzaldehyde,15 and salicylic acid16 using a range of
equipment, mainly including syringe pumps and self-made
apparatus. In many of these literature examples, the aromatic
component to be nitrated was used as the solvent, and the
reaction was run under two-phase plug flow with the aqueous
nitrating acid. A recent literature example (Pelleter et al. from
AstraZeneca) details the nitration of 3-alkylpyrazoles on 100-
gram scale with a productivity of 0.82 g/h.6 A very impressive
example for a nitration on production scale in flow is the nitration
of a key intermediate in the synthesis of naproxcinod which was
A set of screening experiments using the Vapourtec R2C/R4
equipment was performed, using 1 M substrate in acetic acid and
a second inlet stream of fuming nitric acid. The outlet stream was
quenched by dripping directly into water and the resultant
precipitate removed by filtration. The screen quickly revealed
that only 3-min residency time (Rt) were necessary for near
1448
dx.doi.org/10.1021/op200055r |Org. Process Res. Dev. 2011, 15, 1447–1453

Organic Process Research & Development
TECHNICAL NOTE
complete conversion (Table 1, entry 4). When the reactor
temperature was varied and the 3-min residency time main-
tained, complete conversion was achieved at just 90 °C (Table 1,
entry 5), a reduced temperature compared to that of batch. If the
mixture was heated above this temperature, or reacted for longer
than three minutes, a number of undefined baseline by-products
were observed in HPLC. However a residency time of at least
three minutes was required for >99% conversion (Table 1,
entries 7 and 8), and when the equivalents of nitric acid were
reduced, the reaction rate quickly dropped (Table 1, entries 9
and 10). Lower temperatures led to incomplete conversion.
The conditions which gave complete conversion (Table 1,
entry 5) were chosen for scale-up trials with 3-min residency time
at 90 °C using a 20-fold excess of fuming nitric acid. The reaction
was stepwise scaled up by first using a 2-mL reactor followed by
increasing reactor length until the maximum reactor size of
40 mL for the Vapourtec equipment was reached. At each stage
the power usage of the R4 unit was observed to check that the
unit was able to cope with the heat evolved from the reaction.
The pumps were monitored over longer reaction times and were
found to reliably pump the required flow rates over 2 h or more.
For the largest-scale run, an additional 10-mL reactor coil at
room temperature was attached behind the reaction coils at
90 °C so that the outlet stream had time to cool before reaching
the back-pressure regulator. The maximum scale up is outlined in
Scheme 1 with an outlet flow rate of 13.3 mL/min. Filtration of
the quenched reaction mixture gave 86% yield of high-purity
product in comparable yield to the original batch procedure. In
total, 201 g of material was produced over 124 min, giving a
satisfying production rate of 97 g/h. Not only were the screening
experiments complete within a matter of hours, but we were also
able to produce the required 250 g of material in just one day’s
work (including workup) using the Vapourtec setup. The flow
reaction shows a considerable time saving compared to the batch
experiment which required two months of safety evaluation and a
subsequent week to safely perform the chemistry in a 2.5-L
reactor using a trifluoroacetic/nitric acid alternative as the
nitrating mixture.
The second nitration reaction which was performed is out-
lined in Scheme 2. Initial scale-up trials were performed under
batch conditions using an analogue procedure described in
literature.20 The method gave varying results and considerable
by-product formation, leading to a gummy product which was
difficult to purify and which could not be used directly for further
reactions. When the reaction was scaled up in batch, to >100-g
scale, it became difficult to keep the relevant reaction parameters
(e.g., temperature, dosing time, stoichiometry) under precise
control, and it was not possible to obtain a constantly high quality
of product. Either dosing the nitric acid slowly at low tempera-
ture in order to control the heat evolution, or adding the acetic
anhydride too early in the reaction led to the formation of
acylated by-product 5 in significant quantities. The formation of
by-product 5 made isolation of the product by precipitation or
the workup extremely challenging. The by-product was removed
for the next reaction step by additional crystallization, which in
turn led to a significant loss in yield. We once again measured the
DSC of the neat reaction product and observed two large
exotherms, the first with an onset at 140 °C and a potential
energy release of 641 J gꢀ1, the second with an onset at 220 °C
and an energy release of 559 J gꢀ1. The first exotherm was
considered risky for scale up in batch.
Following the success of the previous nitration in continuous
flow, and knowing that the reaction was homogeneous, we
directly ran a screening experiment using the Vapourtec equip-
ment in which we varied temperature, stoichiometry, and
residency time. Suitable conditions for a two-step reaction using
three inlet streams, including our additional R1 pumping unit,
were found after only a few hours of screening experiments, the
time bottleneck often being the running of HPLC analytics. For
the scale-up reaction, 2-amino-4-bromo-benzoic acid methyl
ester 3, as a 0.8 M solution in acetic acid, was mixed with a
6-fold excess of fuming nitric acid and then subjected to a 1.3-min
Figure 1. Batch conversion with time for varying substrate concentra-
tions and 65% nitric acid for the nitration of 8-bromo-1H-quinolin-2-
one 1.
Table 1. Screening experiments for the nitration of 8-bromo-1H-quinolin-2-one 1 using a 2-mL PFA coil and collecting the outlet
in water19
entry
Rt (mins)
temp. (°C)
stoichio-metric ratio (1:HNO3)
1 (%)a
2 (%)a
by-product (%)a,b
1
5
5
5
3
3
3
2
1
5
5
120
105
90
1:20
1:20
1:20
1:20
1:20
1:20
1:20
1:20
1:15
1:11
1
0
96
98
3
2
2
2
0
0
0
0
3
1
2
3
1
97
4
120
90
1
97
5
0
100
92
6
70
8
7
90
2
98
8
90
2
98
9
90
11
86
10
90
52
47
a Area percentage determined from HPLC (254 nm). b Baseline impurities in HPLC of which the structures were not confirmed.
1449
dx.doi.org/10.1021/op200055r |Org. Process Res. Dev. 2011, 15, 1447–1453

Organic Process Research & Development
TECHNICAL NOTE
Scheme 2. Nitration of 2-amino-4-bromo-benzoic acid
methyl ester 3 in batch and under continuous flow
Scheme 3. Nitration of 1-benzosuberone 6 in batch and
under continuous flow
from batch reactions. It is also noteworthy that the total lab time
taken for screening, analytics, and subsequent scale-up was only
one week. Since our group is concerned with the time-critical
supply of intermediates, building blocks and final compounds in
multihundred gram amount, time savings of this type can be
directly translated into cost benefits.
The third nitration, outlined in Scheme 3, uses neat 1-benzo-
suberone and fuming nitric acid. The reaction in batch showed a
strong exotherm and formation of approximately 20% of an
ortho-substituted by-product. DSC of the dried crude material
from a batch procedure showed a large exothermic decomposi-
tion of 983 J gꢀ1 with a slow onset, beginning at approximately
150 °C. Although the reaction was run at low temperature in
batch, addition of 1-benzosuberone 6 to a stirred reactor contain-
ing fuming nitric acid was very exothermic, and we did not want
to work in a temperature regime anywhere close to the exother-
mic decomposition observed in the DSC. By using flow chem-
istry, we sought to adequately control this reaction and in doing
so to possibly reduce by-product formation. Neat 1-benzosuber-
one (at 6.68 M) was mixed with fuming nitric acid at a range of
residency times and stoichiometric ratios using a 5-mL PFA coil
in a chilled reactor according to Table 2.
residency time in the first reactor. After this time, the outlet from
the first reactor was mixed with the third inlet of a 4 M solution of
acetic anhydride in acetic acid in 2-fold excess relative to the
substrate. The reaction was subjected to a second residency time
of one minute to give an outlet stream of 10 mL/min which was
quenched directly by flowing into stirred iceꢀwater. Formation
of a crystalline solid was greatly influenced by the temperature of
the quench water. Gum formation was observed when room
temperature water was used for the quench, but a fine crystalline
solid was obtained when ice was used. Filtration of the quenched
reaction mixture gave 84% yield of a crystalline solid containing
no by-product and pure enough to be used directly in the next
step. In total, 123 g of material was produced over 105 min giving
a production rate of 70 g/h. The flow experiment provided pure
material, in contrast to the varying quality of product obtained
The outlet stream was dosed into stirred iceꢀwater and the
product removed by filtration for analysis by NMR. As previously
observed for the other nitrations, a large excess of nitric acid
1450
dx.doi.org/10.1021/op200055r |Org. Process Res. Dev. 2011, 15, 1447–1453

Organic Process Research & Development
TECHNICAL NOTE
Table 2. Screening experiments for the nitration of 1-benzosuberone 6 using a 5-mL PFA coil and collecting the outlet in
iceꢀwater
entry
Rt (mins)
temp. (°C)
stoichio-metric ratio (6:HNO3)
6 (%)a
7 (%)a
8 (%)a
batch
15
2.5
2.5
2.5
2.5
5
10
10
10
10
0
1:10
1:2.5
1:5
0
23
9
83
65
71
78
79
79
80
17
12
20
22
16
21
20
1
2
3
4
5
6
1:10
1:10
1:10
1:10
0
5
10
10
0
7.5
0
a Percentage from NMR integration.
(in this case 10 equiv) was required for high conversion, and we
observed similar quantities of the by-product compared to that
from batch (Table 2, entries 1ꢀ3). No significant change was
observed when the reaction was run at lower temperatures or for
longer residency time (Table 2, entries 4ꢀ6). Appropriate
conditions were found to be 2.5-min residency time at 10 °C,
and so we scaled up to the equipment maximum of 20 mL in two
10-mL PFA coils. When the flow rates were increased propor-
tionally by a factor of 4 relative to the screening experiments (to a
total flow rate of 8 mL/min), the dry ice-chilled air cooling unit
was not able to maintain the reaction temperature of 10 °C, and
the internal temperature rose to 30 °C. The reaction was
immediately manually stopped. The reaction was performed
with 5-min residency time (with a total flow rate of 4 mL/min),
but again the temperature of 10 °C could not be maintained by
the cooling unit, the internal temperature rose to 15 °C, and the
reaction was again stopped. The fastest we were able to perform this
reaction in a safe manner was at a total flow rate of 2.67 mL/min,
equating to a residency time of 7.5 min. It would be possible to
dilute this reaction so that the energy released from the reaction
per centimeter of tubing would be reduced; however, this would
result in a proportionally lower productivity. Experiments
in which the 1-benzosuberone 1 was diluted with a solvent
(e.g., acetic acid), or using a stainless steel reactor to aid heat
transfer, were not attempted.
high and thus limits the maximum flow rates that can be achieved
in continuous mode.
In all three examples detailed above, comparable yields to that
of batch were observed using continuous processing. In the
second example, an additional advantage was that by-product
formation was completely avoided under flow conditions. The
impurity profile was improved, and this may have been due to the
exact stoichiometry at the mixing point and the accurately
defined reaction times before the quench. Selectivity was not
altered or improved in the other two examples.
In summary, scaling up nitrations to produce some hundreds
of grams was shown to be possible in a rapid manner (within a 1-
to 2-week time frame, including screening) using the Vapourtec
equipment. Three real examples of challenging chemistry en-
countered by our research group, which showed problems on
scale-up, have been transferred to continuous flow with good
results. Nitrations in continuous flow provide a safe alternative to
running dangerous exothermic reactions in batch and show
significant time savings, especially in the screening of reaction
parameters to find the optimal conditions.
’ EXPERIMENTAL PROCEDURES
All chemicals were purchased from Fluka or Sigma-Aldrich
unless otherwise mentioned. Fuming nitric acid was anhydrous
and 100% concentration. 8-Bromo-1H-quinolin-2-one 1 was
prepared according to a procedure by Schlosser et al.21 Reverse
phase HPLC analyses were performed on an Agilent-1100
machine using a Zorbax Eclipse XDB-C18 column, 4.6 mm ꢁ
50 mm, 1.8 μm, using acetonitrile and water as eluent (both
containing 0.05% TFA), a column temperature of 35 °C, a flow
rate of 1.0 mL/min, and measuring at 216 nm. The standard
gradient used was 5 to 100% MeCN over 6 min, 100% MeCN for
1.5 min, followed by 100 to 5% MeCN over 0.5 min. GC was
performed on an Agilent 7890A GC system with a BGB Silaren
column (30 m ꢁ 0.32 mm ID, 0.12 μm film). The standard 12-
min run started at 40 °C which was held for 0.3 min, followed by a
temperature ramp at 25 °C/min up to 220 °C and a second ramp
of 40 °C/min up to 280 °C, and this temperature was held for 3
min. The hydrogen flow was 2 mL/min, the front inlet tempera-
ture was 220 °C, and the front detection temperature was 300 °C.
NMR was performed using a 400 MHz Varian machine, AS 400
Oxford. 1H shifts were referenced to d6-DMSO at 2.49 ppm and
CDCl3 at 7.26 ppm. 13C shifts were referenced to d6-DMSO at
39.52 ppm and CDCl3 at 77.16 ppm. MS was measured using VG
Platform (Fisons Instruments), Spectraflow 783 detector, HP
1100 series HPLC. Melting points were measured using a B€uchi
B-545 machine.
Using a residency time of 7.5 min, crude product was isolated
in quantitative yield at 78% purity. The by-product was removed
by crystallization of the crude material to give a final isolated yield
of 70%; a comparable yield to that of the batch procedure. As a
proof of concept, 12 g of >97% purity material (after crystal-
lization) was prepared over 19 min at a production rate of 39 g/h.
’ CONCLUSIONS
The Novartis Preparation Laboratories have demonstrated the
scale up of three nitration reactions using commercially available
equipment at high production rates. All nitration solutions were
homogeneous which is in contrast to some literature examples
which utilize biphasic, organic, and nitrating acid mixtures in two-
phase plug flow. In our experience, it is not necessary to employ a
complex micromixer or chip, and in the three examples chosen, a
simple Y-piece was sufficient in to effect efficient mixing and high
conversion. It should also be noted that acetic acid or acetic
anhydride with fuming nitric acid mixtures provides an alter-
native to traditional nitrating mixtures, such as a mixture of nitric
and sulfuric acids. Despite the low corrosive nature of the latter
(making it possible to use stainless steel reactors) the viscosity is
1451
dx.doi.org/10.1021/op200055r |Org. Process Res. Dev. 2011, 15, 1447–1453

Organic Process Research & Development
TECHNICAL NOTE
Differential scanning calorimetry was measured on a Mettler
Toledo DSC823 machine. A sample of the starting material,
product, or reaction mixture was heated in a steel capsule, plated
with 5 μm of gold, starting from 35 °C and increasing to 400 °C
at a rate of 5 °C/min.
8.55 (1H, d, J = 2.53), 8.73 (1H, s), 11.21 (1H, br s). HPLC
(Zorbax XDB-C18) rt = 3.73 min; >99% purity. 13C NMR (101
MHz, d6-DMSO) δ 124.3, 124.8, 125.0, 128.5, 140.6, 140.9.
HRMS (FAB) calcd for C9H5BrN2O3 þ H: 268.95563, found:
268.95548/270.95344; calcd for C9H5BrN2O3 þ Na: 290.93758,
found: 290.93741/292.93533.
Specifications of the Vapourtec Equipment. A Vapourtec
R2C/R4/R1C setup was used for all three flow experiments. The
R2C unit is the pumping unit containing two adapted Knauer
pumps which are able to pump highly concentrated and corrosive
acids. The tubing and machine parts are all made from perfluor-
oalkoxy (PFA) or polytetrafluoroethylene (PTFE or Teflon)
plastics, and the pump heads are ceramic. The liquid stream does
not come into contact with metal parts, and corrosion can be
avoided. The R4 is the heater unit with four heating positions and
two cooling positions. The R1C pump is an additional pump
(again adapted for corrosive reagents) which was used when
three inlet streams were required. The tubing used was PFA with
an internal diameter of 1 mm and kept to a minimum between
reactors to avoid unwanted heat transfer. The Y-piece sits
exterior to the heated reactors and inside the cooled reactors,
the latter allowing the chilling of inlet streams prior to mixing.
The Y-pieces are made of PTFE and have an inner diameter of 1
mm and no narrow point where crystallization can occur. Each
reactor unit contains PFA tubing wrapped around into a coil
(for example, a 10-mL reactor contains 13 m of tubing). The
temperature sensor sits on the wall of the PFA tubing, approxi-
mately in the middle of the reactor. The reactor manifold,
surrounding the reactor coil, is approximately 400 cm3.
In order to fill the R2C pumping unit with the strongly
corrosive acids, the following pumping protocol was followed.
The system was flushed with isopropyl alcohol followed by water
and finally glacial acetic acid with all pumps flowing at 2 mL/min
and for 5 min for each inlet. The back of the pump heads were
flushed with water followed by acetic acid. After the reaction the
pumps were flushed in the reverse order. The system was flushed
with glacial acetic acid followed by water, 0.1 M NaOH, and
water again, and finally isopropyl alcohol with all pumps flowing
at 2 mL/min and for 5 min for each inlet. The back of the pump
heads were flushed with water followed by isopropyl alcohol. The
machine was left to stand in isopropyl alcohol.
8-Bromo-6-nitro-1H-quinolin-2-one 2. The flow reactor was
configured using a combination of the R2C pump module and
R4 heater/chiller module. Four 10-mL PFA tubing reactors were
installed in the R4 module, plus an additional 10-mL coil at room
temperature to allow the outlet to cool, along with an 8-bar
ceramic back-pressure regulator fitted in-line between the reactor
outflow and the collection valve. The solvent bottle was filled
with glacial acetic acid, and the reagent stock bottles were filled
with 8-bromo-1H-quinolin-2-one 1 in acetic acid (1 M) and
fuming nitric acid (22.4 M), respectively. The equipment was set
to flow with a substrate:nitric acid ratio of 1:20. Pump 1 delivered
7.04 mL/min substrate solution and 6.29 mL/min fuming nitric
acid to give a residency time of 3 min at 90 °C and a final outlet
flow rate of 13.3 mL/min.
Methyl 4-bromo-2-(nitroamino)benzoate 4. The flow re-
actor was configured using a combination of the R2C pump
module, the R1C pump module, and R4 heater/chiller module.
Two 10-mL PFA tubing reactors were installed in the R4 module
along with an 8-bar ceramic back-pressure regulator fitted in-line
between the reactor outflow and the collection valve. The solvent
bottle was filled with glacial acetic acid, and the reagent stock
bottles were filled with 3 in acetic acid (0.8 M), fuming nitric acid
(22.4 M), and acetic anhydride in acetic acid (4 M), respectively.
The equipment was set to flow with a substrate/HNO3/Ac2O
ratio of 1:6:2. Pump 1 delivered 6.24 mL/min substrate solution
and 1.26 mL/min fuming nitric acid to give a residency time of
1.3 min at 30 °C in the first reactor. The outlet of the first reactor
had a flow rate of 7.5 mL/min. The third pump delivered
2.50 mL/min of the acetic anhydride solution to give a residency
time of 1 min at 30 °C in the second reactor.
The outlet stream was collected for 105 min and poured
directly into stirred ice water (8 L). When collection stopped, the
cream-colored suspension was stirred for 10 min before filtering
and washing with water (2 ꢁ 500 L). The solid was dried under a
filter paper in the vacuum oven at 40 °C overnight to give a
cream-colored crystalline solid (123 g, 84%). Rate of production =
70 g/h. 1H NMR (400 MHz, CDCl3) δ 3.97 (3H, s), 7.41 (1H,
dt, J = 8.60, 1.95), 7.94 (1H, dd, J = 8.59, 1.56), 8.42 (1H, t, J =
1.95), 13.37 (1H, br s). HPLC (Zorbax XDB-C18) rt = 4.72 min;
>99% purity. 13C NMR (101 MHz, d6-DMSO) δ 53.1, 126.4,
126.4, 130.7, 132.2, 132.8, 135.7, 165.3. HRMS (FAB) calcd for
C8H7BrN2O4 ꢀ H: 272.95165, found: 272.95169/274.94938.
8-Nitro-1-benzosuberone 7. The flow reactor was config-
ured using a combination of the R2C pump module and R4
heater/chiller module. Two 10 mL PFA tubing reactors were
installed in the R4 module in chilled reactors along with an 8-bar
ceramic back-pressure regulator fitted in-line between the reactor
outflow and the collection valve. The chilled reactors were
cooled with nitrogen which had been passed over dry ice. The
solvent bottle was filled with glacial acetic acid, and the reagent
stock bottles were filled with neat 1-benzosuberone (6.68 M)
and fuming nitric acid (22.4 M), respectively. The equipment
was set to flow with a substrate/nitric acid ratio of 1:10. Pump 1
delivered 0.67 mL/min substrate solution and 2.00 mL/min
fuming nitric acid to give a residency time of 7.5 min at 10 °C and
a final outlet flow rate of 2.67 mL/min.
The outlet stream was collected for 18.7 min and poured
directly into stirred ice water (500 mL). When collection
stopped, the yellow suspension was stirred for 10 min before
filtering and washing with water (2 ꢁ 50 mL). The solid was
dried under a filter paper in the vacuum oven at 40 °C overnight
to give a pale-yellow solid (17.2 g, >99%). The crude material
(containing approximately 20% of the ortho-nitrated isomer)
was crystallized twice with 10% TBME in heptane (2 ꢁ 80 mL)
and the resultant solid dried in the vacuum oven at 40 °C
overnight to give a pale-yellow solid (12.0 g, 70%). Rate
The outlet stream was collected for 124 min and poured
directly into stirred water (8 L). When collection stopped, the
yellow suspension was stirred for 30 min before filtering and
washing with water (2 ꢁ 1 L). The solid was dried under a filter
paper in the vacuum oven at 50 °C for 2 h then under a filter
paper in the fume cupboard overnight to give a pale-yellow solid
1
of production = 39 g/h. H NMR (400 MHz, d6-DMSO)
δ 1.67ꢀ1.74 (4H, m), 1.78ꢀ1.85 (4H, m), 2.67ꢀ2.77 (2H, m),
3.03ꢀ3.07 (2H, m), 7.60 (1H, d, J = 8.53), 8.27 (1H, s), 8.29
(1H, d, J = 2.51). HPLC (Zorbax XDB-C18) rt = 4.55 min; >99%
1
(201 g, 86%). Rate of production = 97.2 g/h. H NMR (400
MHz, d6-DMSO) δ 6.75 (1H, d, J = 9.09), 8.16 (1H, d, J = 9.35),
1452
dx.doi.org/10.1021/op200055r |Org. Process Res. Dev. 2011, 15, 1447–1453

Organic Process Research & Development
TECHNICAL NOTE
purity. GC (BGB Silaren) rt = 9.20 min; 97.3% purity. 13C NMR
(101 MHz, d6-DMSO) δ 20.5, 24.8, 31.8, 40.4, 123.2, 126.7,
132.2, 139.7, 146.8, 149.4, 203.6. HRMS (FAB) calcd for
C11H11NO3 þ H: 206.08117, found: 206.08110.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: cara.brocklehurst@novartis.com. Telephone: þ41 61
6962970. Fax: þ41 61 6968016.
’ ACKNOWLEDGMENT
We thank Duncan Guthrie, Chris Butters, and David Griffin
(from Vapourtec UK) and Jutta Hollmann and Benjamin Martin
(from Novartis Chemical and Analytical Development) for
continued advice. We also thank Pablo Kuhn, Martin Roggwiller,
Doris Brandenberger, and Dominik Wiss for technical support.
’ REFERENCES
(1) Roberge, D. M.; Zimmermann, B.; Rainone, F.; Gottsponer, M.;
Eyholzer, M.; Kockmann, N. Org. Process Res. Dev. 2008, 12, 905–910.
(2) Ley, S. V.; Baxendale, I. R. Chimia 2008, 62, 162–168.
(3) LaPorte, T. L.; Wang, C. Curr. Opin. Drug Discovery Dev. 2007,
10, 738–745.
(4) Geyer, K.; Codee, J. D. C.; Seeberger, P. H. Chem.—Eur. J. 2006,
12, 8434–8442.
(5) All batch experiments were carried out in glass, round-bottomed
flasks with magnetic stirring (for reactions <1 L) or glass reactors with
overhead stirring (for reactions >1 L). Heating was achieved using an oil bath.
(6) Pelleter, J.; Renaud, F. Org. Process Res. Dev. 2009, 13, 698–705.
(7) Bretherick, L. Handbook of Reactive Chemical Hazards, 7th ed.;
Elsevier Ltd.: Dordrecht, the Netherlands, 2007; Vol 1, p 1634.
(8) Raikova, V. M. Chem. Abstr. 1983, 98, 5963.
(9) Schwalbe, T.; Autze, V.; Wille, G. Chimia 2002, 56, 636–646.
(10) Panke, G.; Schwalbe, T.; Stirner, W.; Taghavi-Moghadam, S.;
Wille, G. Synthesis 2003, 2827–2830.
(11) Antes, J.; Boskovic, D.; Krause, H.; Loebbecke, S.; Lutz, N.;
Tuercke, T.; Schweikert, W. Chem. Eng. Res. Des. 2003, 81, 760–765.
(12) Burns, J. R.; Ramshaw, C. Chem. Eng. Commun. 2002,
189, 1611–1628.
(13) Dummann, G.; Quittmann, U.; Groschel, L.; Agar, D. W.;
Worz, O.; Morgenschweis, K. Catal. Today 2003, 79, 433–439.
(14) Ducry, L.; Roberge, D. M. Angew. Chem., Int. Ed. 2005,
44, 7972–7975.
(15) Kulkarni, A. A.; Kalyani, V. S.; Joshi, R. A.; Joshi, R. R. Org.
Process Res. Dev. 2009, 13, 999–1002.
(16) Kulkarni, A. A.; Nivangune, N. T.; Kalyani, V. S.; Joshi, R. A.;
Joshi, R. R. Org. Process Res. Dev. 2008, 12, 995–1000.
(17) Braune, S.; Pochlauer, P.; Reintjens, R.; Steinhofer, S.; Winter, M.;
Lobet, O.; Guidat, R.; Woehl, P.; Guermeur, C. Chim. Oggi 2009, 27, 26–29.
(18) http://www.vapourtec.co.uk/.
(19) Area percentages from HPLC were not adjusted for the
extinction coefficient of starting material compared to product. The
values are sufficient to give an indication of reaction conversion and
purity in order to choose the correct conditions for scale-up. It should
also be noted that a 1:1 molar solution of 1:2 showed an area % ratio
of 40:60.
(20) Porcs-Makkay, M.; Mezei, T.; Simig, G. Org. Process Res. Dev.
2007, 11, 490–493.
(21) Cottet, F.; Marull, M.; Lefebvre, O.; Schlosser, M. Eur. J. Org.
Chem. 2003, 1559–1568.
1453
dx.doi.org/10.1021/op200055r |Org. Process Res. Dev. 2011, 15, 1447–1453