From alcohol to 1,2,3-triazole: Via a multi-step continuous-flow synthesis of a rufinamide precursor

Article abstract of DOI:10.1039/c6gc01133k

Rufinamide is an antiepileptic drug used to treat the Lennox-Gastaut syndrome. It comprises a relatively simple molecular structure. Rufinamide can be synthesized from an organohalide in three steps. Recently we have shown that microreactor flow networks have better sustainability profiles in terms of life-cycle assessment than the respective consecutive processing in a batch. The analysis was based on the results of a single step conversion from batch to continuous mode. An uninterrupted continuous process towards rufinamide was developed, starting from an alcohol precursor, which is converted to the corresponding chloride with hydrogen chloride gas. The chloride is then converted to the corresponding organoazide that yields the rufinamide precursor via cycloaddition to the greenest and cheapest dipolarophile available on the market. The current process demonstrates chemical and process-design intensification aspects encompassed by novel process windows. Single reaction steps are chemically intensified via a wide range of conditions available in a microreactor environment. Meanwhile, the connection of reaction steps and separations results in process-design intensification. With two in-line separations the process consists of five stages resulting in a total yield of 82% and productivity of 9 g h^-1 (11.5 mol h^-1 L^-1). The process minimizes the isolation and handling of strong alkylating or energetic intermediates, while minimizing water and organic solvent consumption.

Full text of DOI:10.1039/c6gc01133k

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: S. Borukhova, T.
Noel, B. Metten, E. De Vos and V. Hessel, Green Chem., 2016, DOI: 10.1039/C6GC01133K.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/greenchem

Please do not adjust margins
Green Chemistry
Page 1 of 20
View Article Online
DOI: 10.1039/C6GC01133K
Green Chemistry
ARTICLE
From alcohol to 1,2,3-triazole via multi-step continuous-flow
synthesis of rufinamide precursor
Svetlana Borukhova,a Timothy Noёl,a,b Bert Metten,c Eric de Vosc and Volker Hessel*a
Received 00th January 20xx,
Accepted 00th January 20xx
Rufinamide is an antiepileptic drug used to treat the Lennox-Gastaut syndrome. It comprises a relatively simple
molecular structure. Rufinamide can be synthesized from an organohalide in three steps. Recently we have
shown that microreactor flow networks have better sustainability profiles in terms of life-cycle assessment than
the respective consecutive processing in batch. The analysis was based on results of a single step conversion from
batch to continuous mode. An uninterrupted continuous process towards rufinamide was developed, starting
from an alcohol precursor, which is converted to the corresponding chloride with hydrogen chloride gas. The
chloride is then converted to the corresponding organoazide that yields the rufinamide precursor via
cycloaddition to the greenest and cheapest dipolarophile available on the market. The current process
demonstrates chemical and process-design intensification aspects encompassed by Novel Process Windows.
Single reaction steps are chemically intensified via a wide range of conditions available in a microreactor
environment. Meanwhile, the connection of reaction steps and separations results in process-design
intensification. With two in-line separations the process consists of five stages resulting in a total yield of 82% and
productivity of 9g h-1 (11.5 mol h-1 L-1). The process minimizes the isolation and handling of strong alkylating or
energetic intermediates, while minimizing water and organic solvent consumption.
DOI: 10.1039/x0xx00000x
www.rsc.org/
Our aim was to design and realize such a flow microreactor
Introduction
network consisting of
a
multi-step synthesis in
compartmentalized continuous manner in micro flow reactors.
The biggest advantage behind the use of micro flow reactors is
the access to a larger playground in terms of process
conditions without compromising safety4. Chemical reactions
can be kinetically accelerated via high temperature,
concentration and pressure5–8. In addition, a combination of
those operating parameters may lead to new chemical
transformations and operating regimes that are not accessible
In nature, complex molecules are constructed into one entity
via series of reactions such as bio-assembly lines1. According to
Fujita,
a
bio-assembly line represents an energy-driven
product undergoes series of
conveyer-belt, where
a
a
modifications on its way to the end1. Within a cell, single
reactions are carried out continuously within the bio-assembly
line under simple operating conditions, such as a single set of
solvent type, temperature and pressure values. The activation
of reactions is realised via use of enzymes, while
compartmentalization takes place in supramolecular structures
or organelles2. On the contrary, in fine chemical and active
pharmaceutical ingredient (API) synthesis it is common to build
molecules from intermediates that are gathered from various
sites in the world or produced on a site in quantities governed
by the capacity of the equipment available. We believe that by
mimicking nature we can increase the sustainability and
in standard batch technology9–12
. The aforementioned
operating conditions accompanied with a fresh attitude on
process design with the aim of intensification, comprise Novel
Process Windows13–15,6. Substantial numbers of reactions and
processes have been shown to benefit from the concept16–19
.
In collaboration with Ajinomoto OmniChem N.V., we aimed at
developing a continuous process to synthesize a rufinamide
precursor, while leaving amidation, the last step, to be handled
according to a conventional batch procedure. Rufinamide is a
sodium-channel blocker that is used in treatment of a type of
epilepsy called Lennox-Gastaut syndrome. Herein, we describe
a five-stage multi-step process with minimized use of organic
solvent and catalyst. The process consists of gas-liquid,
biphasic liquid-liquid and solid forming reactions and two in-
line separation steps.
maximize the efficiency in chemical processes.
A first
technological equivalent to a bio-assembly line is a micro flow
reactor network based on reactor cascades operated under
non-classical conditions with integrated separation units3.
PLEASE INSERT SCHEME 1.TIFF HERE
Scheme 1. Retrosynthetic sequence of rufinamide.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 1
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 2 of 20
View Article Online
DOI: 10.1039/C6GC01133K
ARTICLE
Journal Name
Background
difluorobenzyl azide in cycloaddition to yield triazole (Scheme
2), minimal amounts are available at any point in time,
minimizing the risk associated with detonation of organic
azides in general.
Rufinamide was developed by Novartis Pharma, AG in 2004
and currently is marketed by Eisai. The initial synthetic path
developed by Novartis relied on constructing the 1,2,3-triazole
precursor via 1,3-dipolar cycloaddition of 2,6-difluorobenzyl Step 1- Chlorodehydroxylation
azide to 2-chloroacrylonitrile (1) as shown in Scheme 120.
The first step of the 5-stage multi-step synthesis is
Alkynes (2)-(4) have been shown to be readily reactive,
however, regioselectivity needs to be directed by a copper
catalyst, otherwise a mixture of 1,4- and 1,5-cycloadduct
products is formed21,22. Mudds et al. proposed a solventless
batch sequence using (E)-methyl 3-methoxyacrylate (5) as an
alternative23. The alkene is a significantly safer and cheaper
alternative to all previously mentioned dipolarophiles. In
addition, due to the presence of the methoxy leaving group no
catalyst is needed to direct regioselectivity. However, the
biggest drawback is its relatively low reactivity requiring high
concentrations. Furthermore solventless reactions with such
potentially energetic compounds are not scalable in batch.
Following batch procedure developments, continuous-flow
processes have been developed. First, Borukhova et al.
showed how the reactivity of (5) can be enhanced in a micro
flow capillary reactor under neat and superheated
conditions24. The reaction time was reduced from >24 hrs to
chlorodehydroxylation of 2,6-difluorobenzyl alcohol to afford
2,6-difluorobenzyl chloride. Recently, we realized the
continuous synthesis of chloroalkanes from alcohols using
either concentrated hydrochloric acid27 (HCl (aq)) or hydrogen
chloride28 (HCl) gas as greener alternatives to more wasteful
chlorinating agents. Valuable intermediates were formed with
water as a sole by-product. The intermediates were further
continuously utilized in the construction of antiemetic drugs
and antihistamines. In order to prepare the chlorides, HCl (aq)
was introduced into a continuous flow reactor via a reagent
loop, which was refilled when needed. Rapid conversion of
neat alcohols was afforded with 3 equivalents of HCl (aq) in 15
min residence time at 120 ̊C. Despite the great practicality of
the process, the use of HCl gas proved to be an even greener
alternative to the use of aqueous HCl due to the possibility of
truly continuous operation (without any need for reagent
loop) and of less equivalents needed for a full conversion of
investigated alcohols. Benzyl alcohol served as a model
compound, which was fully converted to a corresponding
10 min when the temperature was increased from 135
̊C in
batch to 210 ̊C in flow. Moreover, continuous synthesis was
combined with crystallization of the rufinamide ester
precursor upon collection and cooling of the product. 86% of
isolated yield with a productivity of 9 g/h (16.6 mol h-1 L-1) was
obtained. Subsequently, Jamison et al. demonstrated a total
synthesis of rufinamide using 2,6-difluorobenzyl bromide and
methyl propiolate (3) as starting reagents25. 2,6-difluorobenzyl
bromide was converted to 2,6-difluorobenzyl azide, while (3)
was converted to the corresponding amide (4) after reacting
with aqueous ammonia. Without any intermediate separation,
chloride in 20 min residence time at 100 ̊C with 1.2 equivalents
of HCl gas. When the same conditions were applied to 2,6-
difluorobenzyl alcohol 87% yield was obtained. The lower yield
was attributed to the presence of fluorine atoms, that slow
down the dissociation of the protonated hydroxyl group to
yield the benzylic cation. Figure 1.a shows a schematic
representation of the experimental platform, where further
optimization was performed. When the temperature was
increased from 100 ̊C to 110 ̊C, a temporary decrease in
flowrate at the outlet was observed. The decrease in flow rate
can be explained by the higher consumption of HCl gas, which
results in a larger volume available for the liquid phase to
occupy. The constant flow of liquid phase was once again
observed after 40 min reaction time. The resulting product
contained >99wt% of 2,6-difluorobenzyl chloride and no
dibenzyl ether by-product was observed.
PLEASE INSERT SCHEME 2.PNG HERE
Scheme 2. Proposed alternative reagents to prepare rufinamide precursor.
the intermediates reacted to form rufinamide in the copper
tubing, where the reaction was catalysed by the leached ionic
copper species within the reaction stream. The overall process
concentration was 0.25 M in DMSO and resulted in 92% yield
in 11 min with a productivity of 217 mg/h (2.1 mol h-1 L-1) .
Novel Process Windows, apart from concentrating on chemical
PLEASE INSERT FIGURE 1.PNG HERE
and process-design intensification, advocate
a holistic
Figure 1. a. Schematic representation of 2,6-difluorobenzyl chloride synthesis
and optimization results. b. Schematic representation of 2,6-difluorobenzyl
azide synthesis and optimization results. c. Schematic representation of 5-stage
multi-step process to deliver rufinamide precursor.
approach to the process design to yield green processes and
sustainable manufacturing. Recently, we performed a life-cycle
assessment26 and saw a potential within the synthetic route
based on 2,6-difluorobenzyl alcohol and (5) as key reagents, as
shown in Scheme 2. It is proposed that in order to minimize
waste, instead of high-weight halides, like bromides, chlorides
should be used in the synthesis of 2,6-difluorobenzyl azide.
Another green alternative is the route to the chloride itself, by
using pure hydrogen chloride gas and generates water as the
sole by-product. The chloride is converted into the azide.
Finally, due to the immediate consumption of synthesized 2,6-
Step 2- Synthesis of azide
Organoazides are often high energetic substances that are
prone to detonate upon the release of nitrogen under minimal
energy input, such as friction, pressure, heat or light29. Thus,
the production and storage of large quantities of organoazides
present high safety risks. Previously performed safety
assessment of the 2,6-difluorobenzyl azide indicated that the
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 3 of 20
Journal Name
View Article Online
DOI: 10.1039/C6GC01133K
ARTICLE
reagent was relatively stable and not shock sensitive24. reaction. However, crystallization of the product occurred at
Borukhova et al. demonstrated safe cycloaddition of the outlet of the inline liquid-liquid separator. Moreover, 2,6-
organoazides with a variety of dipolarophiles under high difluorobenzyl chloride is a lachrymator, which is less stable
temperature and pressure in micro flow reactors30. The most and more toxic than the corresponding azide. Therefore, in
straightforward synthesis of azide is the use of organohalide order to minimize the handling of the chloride it was decided
and azide source, such as sodium azide. Upon solvation of to connect two steps by the neutralization of the excess of HCl
sodium azide (NaN3) in water the highly toxic and explosive (aq), without any intermediate separation, as shown in Figure
hydrazoic acid (HN3) is formed as shown in (1).
1.c. A significant difference in results was observed between
the optimization results for azide synthesis only and the azide
synthesis in 2-in-1 flow. The results of the optimization are
tabulated in table 1. The same concentration of 0.4 M NaOH,
(1)
Table 1. Intermediate results for optimization of 2-step 2,6-difluorobenzyl azide after the neutralization of excess HCl from reaction 1, was
maintained in reactor 2, which led to partial hydrolysis of the
chloride. Decreasing the concentration of base improved the
results, again increasing the selectivity towards azide. Higher
temperatures and NaN3 equivalents were required to reach
full conversion of 2,6-difluorobenzyl chloride, indicating higher
mass transfer barrier due to the stronger phase separation at
higher concentration of sodium chloride formed during the
neutralization.
Res. time
(min)
NaN3
(eq)
NaOH
(M)
Conv.
(%)
Yield
(%)
Entry
T (̊C)
1
2
3
4
5
6
7
8
140
140
140
150
150
150
160
160
30
30
30
30
30
40
30
40
-
0.4
0.4
0.4
0.4
0.2
0.2
0.2
0.2
28
73
77
83
91
98
98
>99
0
1.1
1.5
1.5
1.5
1.5
1.5
1.6
60
66
72
86
92
91
95
2,6-difluorobenzyl azide is liquid under atmospheric conditions
and immiscible with water. Previously, Borukhova et al.
described the design and fabrication of an inline Teflon
membrane-based liquid-liquid separator27. In order to afford a
perfect separation, the ‘right’ pressure difference should be
applied over the membrane to allow permeate, organic phase,
synthesis from 2,6-difluorobenzyl alcohol.
to pass through the membrane without causing
a
breakthrough of the retained, aqueous, phase or alternatively,
without carrying the permeate within the retentate. An ultra-
low volume adjustable BPR was attached to the aqueous
outlet of the separator and a pressure of 2 psi was set.
Meanwhile, 2,6-difluorobenzyl azide was collected at the
organic side via a tubing of 20 cm with 250 µm ID. After
reaching steady-state, 2,6-difluorobenzyl azide was collected
for 4 hr resulting in 93% isolated yield. Meanwhile, 1.2 % of
total organic phase was lost based on the collected volume
measurements. Karl Fischer titration indicated the presence
0.8 % of water within the separated organic phase.
In order to force an equilibrium to the reactant side and
prevent the formation of the acid, sodium hydroxide (NaOH)
can be added. However, to see if there was any effect of the
base on conversion and reaction, the reaction was first ran
without the base. Figure 1.b shows the experimental platform
used along with the conditions applied and corresponding
results. It should be noted that 2,6-difluorobenzyl chloride is a
solid under atmospheric conditions. Since the aim was to work
with as concentrated conditions as possible, 2,6-difluorobenzyl
chloride was melted and 10 wt% of toluene was added.
However, this lead to crystallization upon cooling. Finally, 25
wt% of toluene was needed to afford homogeneous reactant
solution.2,6-difluorobenzyl chloride and sodium azide
solutions were mixed within an ETFE T-mixer to yield a two-
By connecting two steps, the separation of the hazardous
chloride intermediate and the complication due to its
precipitation that would require the addition of a solvent or a
phase transfer catalyst were circumvented. Thus, a decrease in
the number of unit operations and solvent use was achieved.
Total residence time of 80 min was needed for the two-step
synthesis of 2,6-difluorobenzyl azide via two nucleophilic
substitutions. If desired, the rate of the reactions could be
increased by the addition of a polar solvent, which would
accelerate the reactions. However, due to the fact that our
goal was to deliver a solvent-free process, we avoided the
addition of any solvent.
phase slug flow and heated 140 ̊C. The starting operating
conditions were the ones previously investigated in the
optimization of benzyl azide synthesis. As expected, the
addition of base increased the hydrolysis of 2,6-difluorobenzyl
chloride and decreased the conversion, as tabulated in Figure
1.b. Decreasing equivalents of base increased the selectivity
towards 2,6-difluorobenzyl azide. However, more equivalents
of NaN3 were needed to obtain full conversion in 30 min
residence time.
Step 1+2+separation+3+crystallization
Step 1+2+separation
The intensified cycloaddition of 2,6-difluorobenzyl azide to (5)
greatly benefited from being carried out in a continuous
manner under superheated conditions24. The most important
benefit is the increased safety due to the circumvention of any
In order to connect the synthesis of 2,6-difluorobenzyl chloride
to the synthesis of azide, the chloride had to be separated
from the acidified water phase, formed in the course of the
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 4 of 20
View Article Online
DOI: 10.1039/C6GC01133K
ARTICLE
Journal Name
significant decomposition of organic azide. Similarly to a factors contribute to a better sustainability profile of the
previously published process, 2,6-difluorobenzyl azide was individual steps and the flow microreactor network as a whole.
mixed with (5), heated to 210 ̊C and allowed 10 min reaction
Finally, if compared to conventional batch process:
time. In order to carry out continuous synthesis and
purification, a stream of methanol was introduced at the
outlet of the reactor. The streams were allowed to mix within
a 1 ml mixing zone. Upon collection needle-like crystals
crashed out of the solution. However, the analysis of the
mother liquor indicated the presence of unconverted azide.
We attribute the difference in the results to the difference in
the starting composition of the azide feed. Therefore, we
increased the residence time to 15 min, which gave full
conversion of the azide. GC-MS showed that the main side-
products were benzyl alcohol, and decomposition products of
2,6-difluorobenzyl azide. Further cooling, filtration and drying
under vacuum resulted in 88% isolated yield based on 2,6-
difluorobenzyl azide, leading to an overall 82% yield based on
2,6-difluorobenzyl alcohol used. As a result, a productivity of
the rufinamide precursor of 9 g/h was afforded (11.5 mol h-1 L-
1).
ꢀ
ꢀ
Operating time is drastically decreased from days to hours;
Solvent use is circumvented without compromising the
safety usually afforded by dilution;
ꢀ
ꢀ
ꢀ
Overall isolated yield is higher;
Production capacity is variable based on demand;
Inline separation minimizes manual handling.
Meanwhile, if the process is compared to the existing
continuous total synthesis:
ꢀ
Synthetic route starts off with a cheaper, greener and more
accessible starting compound;
ꢀ
ꢀ
No solvent is used, as opposed to DMSO use;
9 times longer residence times are needed to make a
precursor;
ꢀ
No catalyst is used, thus no leached metal in the final
product;
Conclusion
ꢀ
ꢀ
Inline separation affords more concentrated intermediate;
6 times higher space time yield is obtained.
Continuous flow and solvent-free processing comprise two
eminent issues of Green Chemistry and Green Engineering.
Yet, these two processing issues are difficult to combine due to
the high susceptibility of micro flow reactors to clog.
Nevertheless, we believe that the combination should be the
next step in the synthesis of fine chemicals to minimize solvent
waste and energy used within a given process. Moreover,
incorporation of Novel Process Windows as a concept to assist
in chemical and process-design intensification proves to be an
important stepping stone on the way to sustainable and
efficient processes.
Acknowledgements
We would like to thank Erik van Herk for his work on
mechanical aspect of the process. Funding by the Advanced
European Research Council Grant “Novel Process Windows –
Boosted Micro Process Technology” (Grant number: 267443) is
kindly acknowledged.
Experimental Section
Herein, we demonstrated
a 5-stage 3-step continuous
synthesis with integrated separation steps. A total yield of 82% General Information. HPLC pumps were purchased from
of rufinamide precursor with productivity of 9 g/h (11.5 mol h-1 Knauer, while ETFE micro capillary reactors along with
L-1) was delivered. The total residence time needed for the connections were purchased from IDEX Health & Science
whole process is 95 min, which is relatively long when Technologies. Reagents were pumped via pumps into the ETFE
compared to other continuous-flow 5-stage processes31. This T- or Y-mixers. Hastelloy tubing and stainless steel connections
stems from the absence of solvent. In this case nucleophilic were ordered from Valco instruments.
substitutions could proceed at much higher rates if diluted
Hydrogen chloride gas was purchased in 10 L bottle from Linde
with polar solvents. However, when considering the separation
Gas Benelux. 2,6-difluorobenzyl alcohol was provided by
of the solvent and its recycling or disposal, and the purity of
Ajinomoto OmniChem N.V. The rest of the chemicals were
the final product, prevention of the waste generation
purchased from Sigma-Aldrich and used as received without
surpasses the benefits of higher reaction speed. The aim of
any additional purifications.
minimizing solvent use and, thus, energy needed for its later
Continuous synthesis of 2,6-difluorobenzyl chloride. The HPLC
separation was realized. Similarly, water use was minimized by
pump was first purged with isopropanol and then with the 2,6-
combining introduction of a quenching reagent (NaOH) and a
difluorobenzyl alcohol. The reactor tubing was connected
reactant for the subsequent step (NaN3) at once. Moreover,
directly to the pump and the reactor was filled with alcohol.
isolation of toxic and unstable lachrymator, alkylating chloride
Meanwhile, the hydrogen chloride bottle was opened,
intermediate, was circumvented by connecting its synthesis to
followed by the mass flow controller being set to 0.05 g/min
the synthesis of the corresponding azide. Finally, in the
(1.4 mmol/min). A Y-mixer was attached to the gas, alcohol
isolation of 2,6-difluorobenzyl azide no organic extractant was
and reactor outlets. 2,6-difluorobenzyl alcohol was pumped at
added, instead the azide was separated in its neat form and
0.128 ml/min (1.1 mmol/min). The pressure was set to 12 bar
used as it is in the subsequent step. All of the mentioned
at the Equilibar back pressure regulator (BPR), the heating
4 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 5 of 20
Journal Name
View Article Online
DOI: 10.1039/C6GC01133K
ARTICLE
plate was turned on and reaction temperature was set. After Third step- continuous synthesis of methyl 1-(2,6-
three times a residence time of 40 min, the product was difluorobenzyl)-1H-1,2,3-triazole-4-carboxylate. All HPLC
collected into a vial containing water and ethyl acetate. The pumps were purged with isopropanol and later with either 2,6-
product was extracted and organic phase was analysed further difluorobenzyl azide, EMMA or methanol. The collected 2,6-
with GC. 1H NMR (400 MHz, CDCl3) δ 7.30 (m, J= 4 Hz, 1H), 6.92 difluorobenzyl azide was pumped at 100 µl/min (0.7
(t, J= 8 Hz, 2H), 4.67 (s, 2H); 13C (100 MHz, CDCl3) δ 162.45 (d, mmol/min) into a T-mixer to mix with EMMA, pumped at 112
J= 7 Hz), 159.94 (d, J= 7 Hz), 132.63 (t, J= 10 Hz), 114.08 (t, J= 9 µl/min (1.05 mmol/min), in order to proceed into 3.2 ml
Hz), 111.61 (q, J= 5 Hz), 32.40 (t, J=5 Hz).
Hastelloy micro capillary reactor. MeOH was pumped at 3.2
ml/min (15 v/v) to dilute the product stream. A closed heating
Continuous synthesis of 2,6-difluorobenzyl azide. 25 gr of 2,6-
difluorobenzyl chloride was melted and mixed with 10 ml of
toluene. The solution was allowed to cool down to room
temperature. HPLC pumps were first purged with isopropanol
and the first pump was purged with the solution, while a
second with aqueous sodium azide solution (4.0 M). 2,6-
difluorobenzyl chloride solution was pumped at 21 µl/min,
while sodium azide solution was pumped at 46 µl/min. Two
streams were mixed within an ETFE T-mixer and proceeded
into a 2 ml ETFE reactor (0.03” ID). The inline 7 bar (100 psi)
BPR was attached to the other end of the reactor and the
bath (Lauda Proline 8) was heated to 210 ̊C and 15 min
residence time was allowed for the reaction to take place.
After passing through a mixing zone of 1 ml volume Hastelloy
micro capillary (0.04” ID) and BPR of 1000 psi, the product was
collected. The collected solution was cooled to room
temperature and later to 5
precipitated yellow product was then filtered and dried under
vacuum overnight. Yield 88%; mp 136-137
C; 1H NMR (400
̊C in the fridge overnight. The
̊
MHz, DMSO-d6) δ 8.84 (s, 1H), 7.49 (m, J= 4 Hz, 1H), 7.15 (t, J=
8 Hz, 2H), 5.71 (s, 2H), 3.80 (s, 3H); 13C (100 MHz, DMSO-d6) δ
162.43 (d, J= 8 Hz), 160.98, 159.95 (d, J= 7 Hz), 138.90, 132.30
(t, J= 10 Hz), 129.91, 112.38 (q, J= 5 Hz), 111.25 (t, J= 19 Hz),
51.25, 41.77 (t, J=4 Hz).
reactor was heated to 140 ̊C. After three times residence time
of 30 min, the product was collected into a vial containing
water and ethyl acetate. The product was extracted and
organic phase was analysed further with GC. Toluene served as
an internal standard. 1H NMR (400 MHz, CDCl3) δ 7.32 (m, J= 4
Hz, 1H), 6.95 (t, J= 8 Hz, 2H), 4.43 (s, 2H); 13C (100 MHz, CDCl3)
δ 162.80 (d, J= 7 Hz), 160.31 (d, J= 7 Hz), 132.65 (t, J= 10 Hz),
111.74 (d, J= 6 Hz), 111.41 (d, J= 6 Hz), 41.79 (t, J=3 Hz).
References
1.
2.
M. Fujita, J. Theor. Biol., 1982, 99, 9–13.
L. F. Tietze and U. Beifuss, Angew. Chemie Int. Ed. English,
1993, 32, 131–163.
Two-step continuous synthesis of 2,6-difluorobenzyl azide.
Two HPLC pumps were first purged with isopropanol. 2.6 gr
(65 mmol) of NaOH and 26 gr (400 mmol) of NaN3 were
dissolved in water to yield 100 ml of solution. The first pump
was purged with 2,6-difluorobenzyl alcohol, while the second
with demi water. Meanwhile, the hydrogen chloride bottle was
opened, followed by the mass flow controller being set to 0.05
g/min (1.4 mmol/min). All the reactors were connected and
immersed into heating baths. A Y-mixer was attached to the
gas, alcohol and reactor outlets. 2,6-difluorobenzyl alcohol was
pumped at 128 µl/min (1.1 mmol/min). Another pump was set
to 128 µl/min to pump pure water, and mixed streams flowed
into a second reactor of 24 ml (1.6 mm ID). For this experiment
two inline cartridge BPRs of 7 bar pressure were used. One
BPR was attached first, after a significant pressure was
attained within the reactor, the second BPR was attached. The
3.
4.
D. Webb and T. F. Jamison, Chem. Sci., 2010, 1, 675–680.
P. Poechlauer, J. Colberg, E. Fisher, M. Jansen, M. D.
Johnson, S. G. Koenig, M. Lawler, T. Laporte, J. Manley, B.
Martin, and A. O’Kearney-McMullan, Org. Process Res.
Dev., 2013, 17, 1472–1478.
5.
6.
V. Hessel, Chem. Eng. Technol., 2009, 32, 1655–1681.
V. Hessel, D. Kralisch, N. Kockmann, T. Noël, and Q. Wang,
ChemSusChem, 2013, 6, 746–89.
7.
T. Razzaq, T. N. Glasnov, and C. O. Kappe, Chem. Eng.
Technol., 2009, 32, 1702–1716.
8.
B. Gutmann, D. Cantillo, and C. O. Kappe, Angew. Chemie
Int. Ed., 2015, 54, 6688–6728.
9.
M. Baumann, I. R. Baxendale, L. J. Martin, and S. V. Ley,
Tetrahedron, 2009, 65, 6611–6625.
10.
11.
L. Ducry and D. M. Roberge, Angew. Chemie, 2005, 117
8186–8189.
,
first heating bath was set to 110 ̊C, while second to 70 ̊C. After
D. M. Roberge, L. Ducry, N. Bieler, P. Cretton, and B.
one residence time (40 min), the pure water solution was
replaced by the NaOH and NaN3 solution and the flowrate was
gradually increased to 460 µl/min. The second heating bath
Zimmermann, Chem. Eng. Technol., 2005, 28, 318–323.
D. M. Roberge and N. Bieler, 2008, 1155–1161.
T. Illg, V. Hessel, P. Löb, and J. C. Schouten, Chem. Eng. J.,
2011, 167, 504–509.
12.
13.
was set to 160 ̊C. After the period equivalent to two residence
times, the BPR outlet was connected to the inline liquid-liquid
separator. An ultra-low volume adjustable BPR was attached
to the aqueous outlet and a pressure of 2 psi was set.
Meanwhile, 2,6-difluorobenzyl azide was collected at the
organic side via a tubing of 20 cm with 250 µm ID. The first
millilitre of the product was discarded, which was followed by
a collection.
14.
15.
16.
S. Borukhova and V. Hessel, in Process Intensification for
Green Chemistry, John Wiley & Sons, Ltd, 2013, pp. 91–
156.
V. Hessel, D. Kralisch, N. Kockmann,, Novel Process
Windows: Innovative Gates to Intensified and Sustainable
Chemical Processes, 2015.
M. Damm, T. N. Glasnov, and C. O. Kappe, 2010, 14, 215–
224.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 5
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 6 of 20
View Article Online
DOI: 10.1039/C6GC01133K
ARTICLE
Journal Name
17.
F. Benito-Lopez, R. M. Tiggelaar, K. Salbut, J. Huskens, R. J.
M. Egberink, D. N. Reinhoudt, H. J. G. E. Gardeniers, and W.
Verboom, Lab Chip, 2007, , 1345–51.
7
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
T. Razzaq, T. N. Glasnov, and C. O. Kappe, European J. Org.
Chem., 2009, 2009, 1321–1325.
D. Cantillo, H. Sheibani, and C. O. Kappe, J. Org. Chem.,
2012, 77, 2463–73.
W. S. and M. S. R. Portmann, U. C. Hofmeier, A. Burkhard,
2004.
H. C. Kolb and K. B. Sharpless, Drug Discov. Today, 2003, 8,
1128–37.
C. D. Smith, I. R. Baxendale, S. Lanners, J. J. Hayward, S. C.
Smith, and S. V Ley, Org. Biomol. Chem., 2007, , 1559–61.
5
W. H. Mudd and E. P. Stevens, Tetrahedron Lett., 2010, 51
3229–3231.
,
S. Borukhova, T. Noël, B. Metten, E. de Vos, and V. Hessel,
ChemSusChem, 2013, 6, 2220–5.
P. Zhang, M. G. Russell, and T. F. Jamison, Org. Process Res.
Dev., 2014.
D. Ott, S. Borukhova, and V. Hessel, Green Chem., 2016, 18
,
1096–1116.
S. Borukhova, T. Noël, and V. Hessel, ChemSusChem, 2016,
9, 67–74.
28.
29.
30.
31.
S. Borukhova, T. Noël, and V. Hessel, Org. Process Res.
Dev., 2016, 20, 568–573.
R. A. Abramovitch and E. P. Kyba, in The Azido Group
(1971), John Wiley & Sons, Ltd., 1971, pp. 221–329.
S. Borukhova, A. D. Seeger, T. Noël, Q. Wang, M. Busch,
and V. Hessel, ChemSusChem, 2015, 8, 504–512.
D. R. Snead and T. F. Jamison, Angew. Chemie Int. Ed.,
2015, 54, 983–987.
6 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 7 of 20
Green Chemistry
View Article Online
DOI: 10.1039/C6GC01133K
Scheme 1. Retrosynthetic sequence of rufinamide.
148x132mm (300 x 300 DPI)

Green Chemistry
Page 8 of 20
View Article Online
DOI: 10.1039/C6GC01133K
Scheme 2. Proposed alternative reagents to prepare rufinamide precursor.

Page 9 of 20
Green Chemistry
View Article Online
DOI: 10.1039/C6GC01133K
Figure 1. a. Schematic representation of 2,6-difluorobenzyl chloride synthesis
and optimization results. b. Schematic representation of 2,6-difluorobenzyl azide synthesis and optimization
results. c. Schematic representation of 5-stage
multi-step process to deliver rufinamide precursor.

Green Chemistry
Page 10 of 20
View Article Online
DOI: 10.1039/C6GC01133K

Page 11 of 20
Green Chemistry
View Article Online
DOI: 10.1039/C6GC01133K
Supporting information
From alcohol to 1,2,3-triazole via multi-step continuous-flow
synthesis of rufinamide precursor
Svetlana Borukhova,a Timothy Noёl,a,b Bert Metten,c Eric de Vosc and Volker Hessel*a
Corresponding author’s contact information: v.hessel@tue.nl
a.
Department of Chemical Engineering and Chemistry, Technische Universiteit Eindhoven, Den
Dolech 2, 5600 MB Eindhoven, Netherlands.
b.
c.
Department of Organic Chemistry, Ghent University, Krijgslaan 281 (S4), 9000 Ghent, Belgium.
S.A. Ajinomoto OmniChem N.V., Cooppallaan 91, 9230 Wetteren, Belgium.

Green Chemistry
Page 12 of 20
View Article Online
DOI: 10.1039/C6GC01133K
1. General Reagent Information
Hydrogen chloride gas was purchased in 10 L bottle from Linde Gas Benelux. 2,6ꢀdifluorobenzyl alcohol was
provided by S.A. Ajinomoto OmniChem N.V. The rest of the chemicals were purchased from SigmaꢀAldrich
and used as received without any additional purifications.
HPLC pumps were purchased from Knauer, while ETFE micro capillary reactors along with connections were
purchased from IDEX Health & Science Technologies. Reagents were pumped via pumps into the ETFE Tꢀ or
Yꢀmixers. Hastelloy tubing and stainless steel connections were ordered from Valco instruments.
2. General Analytical Information
NMR (Brukerꢀ Avance 400 (400 MHz)) and GCꢀMS (Shimadzu GCꢀ2010 Plus coupled to a Mass Spectrometer,
1
Shimadzu GCMSꢀQP 2010 Ultra) were used for characterization of isolated products. Hꢀ NMR and 13Cꢀ NMR
1
spectra were used to characterize and determine the purity of isolated products. H chemical shifts are reported
in parts per million (ppm) downfield from tetramethylsilane (TMS) (0.00 ppm), while 13C chemical shifts are
reported relative to CDCl3 (77.2 ppm). Multiplicity of the peaks is abbreviated as s ꢀ singlet, d ꢀ doublet, t ꢀ
triplet, qꢀ quartet, mꢀ multiplet and brꢀbroad. GCꢀFID was used to quantify product distribution. GCꢀMS was
used to identify the decomposition products.
3. GC method
GC method for followꢀup of 2,6ꢀdifluorobenzyl alcohol and chloride conversion
Parameter
Column
Length
Specification
CPꢀSil 5 CB
30 m
Inner Diameter
Film thickness
Internal Pressure
Injection volume
Split ratio
250 µm
1.0 µm
1.8 psi (He)
0.2 µl
1/5
Injection temperature 220 ˚C
Detector temperature 280 ˚C
Air
300 ml/min
Steam
30 ml/min
Make up
Initial temperature
Ramp 1
25 ml/min
80 ˚C (2 min)
15 ˚C/min (till 200
20 ˚C/min (till 240
̊
̊
Ramp 2

Page 13 of 20
Green Chemistry
View Article Online
DOI: 10.1039/C6GC01133K
HPLC method for follow up of 1,2,3ꢀtriazole formation
Parameter
Column
Specification
C18
Length
Inner Diameter
Pore size
150 mm
4.6 mm
5 µm
Column temperature 35 ˚C
Injection volume
Eluent
1 µl
775 ml of DW, 1 ml of
Formic acid (98%), 150 ml
MeOH, 75 ml THF
UV, 230 nm
Detection
Flow rate
1 ml/min
Running time
35 min

Green Chemistry
Page 14 of 20
View Article Online
DOI: 10.1039/C6GC01133K
4. General Information Regarding experimental setups and experimental procedure
4.1 Continuous chlorodehydroxylation
4.1.1
Reaction setup for continuous chlorodehydroxylation
Scheme S1. Schematic representation of setup assembled for chlorodehydroxylation including hydrogen
chloride gas delivery unit (with a permission from [1] Copyright © 2016, American Chemical Society ). More
detailed information on assembly and operation is available in ref. 1.

Page 15 of 20
Green Chemistry
View Article Online
DOI: 10.1039/C6GC01133K
Scheme S2. Schematic representation of setup assembled for continuous chlorodehydroxylation
Component
Aꢀ1
Description
Scheme
Knauer HPLC pumps of
Smartline 1050 series
Bꢀ1
Cꢀ1
Yꢀmixer (Pꢀ512 from IDEXꢀHS)
ETFE tubing of 1/16” external
diameter and 0.03” internal
diameter (1530L from IDEXꢀHS)
Dꢀ1
Equilibar back pressure regulator
4.1.2 Experimental procedure for continuous chlorodehydroxylation of 2,6-
dilfuorobenzyl alcohol
The HPLC pump was first purged with isopropanol and then with the 2,6ꢀdifluorobenzyl alcohol. The
reactor tubing was connected directly to the pump and the reactor was filled with alcohol. Meanwhile,
the hydrogen chloride bottle was opened, followed by the mass flow controller being set to 0.05 g/min
(1.4 mmol/min). A Yꢀmixer was attached to the gas, alcohol and reactor outlets. 2,6ꢀdifluorobenzyl
alcohol was pumped at 0.128 ml/min (1.1 mmol/min). The pressure was set to 12 bar at the Equilibar
back pressure regulator (BPR), the heating plate was turned on and reaction temperature was set. After
three times a residence time of 40 min, the product was collected into a vial containing water and ethyl
acetate. The product was extracted and organic phase was analysed further with GC. 1H NMR (400
MHz, CDCl3) δ 7.30 (m,
162.45 (d, = 7 Hz), 159.94 (d,
32.40 (t, =5 Hz).
J
= 4 Hz, 1H), 6.92 (t,
J
= 8 Hz, 2H), 4.67 (s, 2H); 13C (100 MHz, CDCl3) δ
= 10 Hz), 114.08 (t, = 9 Hz), 111.61 (q, = 5 Hz),
J
J
= 7 Hz), 132.63 (t,
J
J
J
J

Green Chemistry
Page 16 of 20
View Article Online
DOI: 10.1039/C6GC01133K
4.2 Continuous 2-step synthesis
4.2.1 Reaction setup for 2-step continuous synthesis of 2,6-dilfuorobenzyl azide
Reaction setup for 2ꢀstep synthesis consisted of the chlorination setup, shown in scheme S1, connected to a
liquidꢀ liquid separator, shown in scheme S3. Two inserts were modelled and fabricated from ETFE. Inserts
were used to compress hydrophobic Zefluor membrane (P5PQ047), available from Pall Corporation. Two
stainless steel holders were used to create a shell for the separator.
Scheme S3. Liquidꢀliquid membrane based separator (with a permission from [1]. Copyright © 2016 WILEYꢀ
VCH Verlag GmbH & Co. KGaA, Weinheim ).
a. Design of the top insert with two drillings for inlet of slugꢀflow and outlet of aqueous phase.

Page 17 of 20
Green Chemistry
View Article Online
DOI: 10.1039/C6GC01133K
b. Design of the bottom insert with single drilling for outlet of organic phase.
c. Photograph of a liquidꢀliquid separator with white membrane in between two inserts.

Green Chemistry
Page 18 of 20
View Article Online
DOI: 10.1039/C6GC01133K
4.1.1
Experimental procedure for 2-step continuous synthesis
Two HPLC pumps were first purged with isopropanol. 2.6 gr (65 mmol) of NaOH and 26 gr (400 mmol) of
NaN3 were dissolved in water to yield 100 ml of solution. The first pump was purged with 2,6ꢀdifluorobenzyl
alcohol, while the second with demi water. Meanwhile, the hydrogen chloride bottle was opened, followed by
the mass flow controller being set to 0.05 g/min (1.4 mmol/min). All the reactors were connected and immersed
into heating baths. A Yꢀmixer was attached to the gas, alcohol and reactor outlets. 2,6ꢀdifluorobenzyl alcohol
was pumped at 128 µl/min (1.1 mmol/min). Another pump was set to 128 µl/min to pump pure water, and
mixed streams flowed into a second reactor of 24 ml (1.6” ID). For this experiment two inline cartridge BPRs of
7 bar pressure were used. One BPR was attached first, after a significant pressure was attained within the
reactor, the second BPR was attached. The first heating bath was set to 110
residence time (40 min), the pure water solution was replaced by the NaOH and NaN3 solution and the flowrate
was gradually increased to 460 µl/min. The second heating bath was set to 160 C. After the period equivalent to
̊C, while second to 70 ̊C. After one
̊
two residence times, the BPR outlet was connected to the inline liquidꢀliquid separator. An ultraꢀlow volume
adjustable BPR was attached to the aqueous outlet and a pressure of 2 psi was set. Meanwhile, 2,6ꢀ
difluorobenzyl azide was collected at the organic side via a tubing of 20 cm with 250 µm ID. The first millilitre
of the product was discarded, which was followed by a collection.
Scheme S4. Schematic representation of setup assembled for 2ꢀstep continuous synthesis

Page 19 of 20
Green Chemistry
View Article Online
DOI: 10.1039/C6GC01133K
Scheme S5. Photograph of a setup for 2ꢀstep continuous synthesis
4.3 Synthesis of methyl 1-(2,6-difluorobenzyl)-1H-1,2,3-triazole-4-carboxylate

Green Chemistry
Page 20 of 20
View Article Online
DOI: 10.1039/C6GC01133K
Scheme S6. Schematic representation of setup assembled for 3ꢀstep continuous synthesis of rufinamide
precursor.
Experimental procedure for 3ꢀstep continuous synthesis of methyl 1ꢀ(2,6ꢀdifluorobenzyl)ꢀ1Hꢀ1,2,3ꢀtriazoleꢀ4ꢀ
carboxylate:
All HPLC pumps were purged with isopropanol and later with either 2,6ꢀdifluorobenzyl azide, EMMA
or methanol. The collected 2,6ꢀdifluorobenzyl azide was pumped at 100 µl/min (0.7 mmol/min) into a
Tꢀmixer to mix with EMMA, pumped at 112 µl/min (1.05 mmol/min), in order to proceed into 3.2 ml
Hastelloy micro capillary reactor. MeOH was pumped at 3.2 ml/min (15 v/v) to dilute the product
stream. A closed heating bath (Lauda Proline 8) was heated to 210 ̊C and 15 min residence time was
allowed for the reaction to take place. After passing through a mixing zone of 1 ml volume Hastelloy
micro capillary (0.04” ID) and BPR of 1000 psi, the product was collected. The collected solution was
cooled to room temperature and later to 5
̊
filtered and dried under vacuum overnight. Yield 88%; mp 136ꢀ137
̊
δ 8.84 (s, 1H), 7.49 (m,
DMSOꢀd6) δ 162.43 (d,
J= 4 Hz, 1H), 7.15 (t, J
J
J
J
112.38 (q,
J= 5 Hz), 111.25 (t,
J
J