Submillisecond organic synthesis: Outpacing Fries rearrangement through microfluidic rapid mixing

Article abstract of DOI:10.1126/science.aaf1389

In chemical synthesis, rapid intramolecular rearrangements often foil attempts at site-selective bimolecular functionalization.We developed a microfluidic technique that outpaces the very rapid anionic Fries rearrangement to chemoselectively functionalize iodophenyl carbamates at the ortho position. Central to the technique is a chip microreactor of our design, which can deliver a reaction time in the submillisecond range even at cryogenic temperatures.The microreactorwas applied to the synthesis of afesal, a bioactive molecule exhibiting anthelmintic activity, to demonstrate its potential for practical synthesis and production.

Full text of DOI:10.1126/science.aaf1389

RESEARCH
| REPORTS
and tested the structures for mixing efficiency on
the basis of computational fluid dynamics (CFD).
These structures were evaluated by monitoring
the degree of mixing as a function of distance
from the initial mixing point; in general, the mix-
ing efficiency improved with increasing number
of turns (see descriptions of MS-3, -4, and -5 in the
supplementary materials). The turn induced chaot-
ic advection involving rapid distortion and elonga-
tion of the fluid interface (15–17) when the side-on
collision type of serpentine 3D-structured chan-
nel was used. The simulation results showed (fig.
S1, B and C) that the serpentine 3D-structured
channel MS-5, which has a turn before the mix-
ing point and four turns after the mixing point in
total length of 1 mm, provides mixing efficiency
levels of 95% and 82% for total flow rates of 7.5
and 4.5 ml/min, respectively. These flow rates
correspond to a residence time of ~0.3 ms, which
confirmed the feasibility for submillisecond-scale
reaction control for a respectable range of flow
rates.
MICROFLUIDICS
Submillisecond organic synthesis:
Outpacing Fries rearrangement
through microfluidic rapid mixing
Heejin Kim,1* Kyoung-Ik Min,2* Keita Inoue,1 Do Jin Im,3
Dong-Pyo Kim,2† Jun-ichi Yoshida1†
In chemical synthesis, rapid intramolecular rearrangements often foil attempts at site-selective
bimolecular functionalization. We developed a microfluidic technique that outpaces the very
rapid anionic Fries rearrangement to chemoselectively functionalize iodophenyl carbamates
at the ortho position. Central to the technique is a chip microreactor of our design, which can
deliver a reaction time in the submillisecond range even at cryogenic temperatures. The
microreactor was applied to the synthesis of afesal, a bioactive molecule exhibiting anthelmintic
activity, to demonstrate its potential for practical synthesis and production.
hemical synthesis often involves competi-
tion between unimolecular and bimolecu-
lar reactivity of an intermediate. To fully
explore and exploit the synthetic pathways
of a given intermediate, control over its
We initially considered modification of a stain-
less steel microfluidic assembly used earlier (10–12),
which could deliver a residence time of a few
milliseconds. However, the stainless steel was
not suitable for fabricating a minuscule dead vol-
ume, although compatibility with high-pressure
and low-temperature reaction conditions was en-
sured. To further reduce the residence time below
milliseconds (14), it was necessary to devise a
microfluidic device with a smaller internal vol-
ume that would attain a higher level of mixing
efficiency within the temporal frame needed for
the synthesis. Such a device had two major
requirements: The reaction mixture had to be
well-mixed within the lifetime of the reactive in-
termediate, and the device had to endure low
temperature as well as high back pressure ensuing
from the increased viscosity of the reaction fluid
at that temperature.
With the simulation results in hand, we care-
fully considered the choice of material for fab-
rication of the device. Our prior experience with
polymer-based microfluidic chips pointed us to a
fluoroethylene propylene–polyimide film hybrid,
which offers outstanding physical toughness at
high pressure and low temperature as well as
chemical inertness (18–20). Figure 1 showcases
the device we developed, which is a polymer-
based chip microreactor (CMR) with 3D serpen-
tine microchannel design (MS-5).
C
lifetime is needed along with control over its mix-
ing time with external reagents for both its gener-
ation and trapping reactions. One efficient means
of prolonging the lifetime is to lower the reaction
temperature; the range from –78° to –100°C is
slightly above the melting point of many organic
solvents (1). Microfluidic devices (2–9) concomi-
tantly deliver extremely fast mixing times unat-
tainable by batches (10–12). However, minimizing
the mixing time in the microfluidic device is quite
a challenge in the low-temperature range because
of the exponential rise in the viscosity of the sol-
vent with decreasing temperature (13). Here, we
report a microfluidic device that achieves sub-
millisecond mixing time even at cryogenic tem-
peratures, and we demonstrate its utility in
outpacing an intramolecular rearrangement.
We fabricated the device in one step by ther-
mally bonding six stacked polyimide films, each
1Department of Synthetic Chemistry and Biological
Chemistry, Graduate School of Engineering, Kyoto University,
Nishikyo-ku, Kyoto 615-8510, Japan. 2Department of
Chemical Engineering, POSTECH, Pohang 790-784, Korea.
3Department of Chemical Engineering, Pukyong National
University, Busan 608-739, Korea.
*These authors contributed equally to this work. †Corresponding
author. Email: dpkim@postech.ac.kr (D.-P.K.); yoshida@sbchem.
kyoto-u.ac.jp (J.Y.)
On the strength of good mixing reported for a
three-dimensional (3D) serpentine mixer (15, 16),
we designed various microchannel structures (MSs)
Fig. 1. Chip microreactor (CMR)
fabricated with six layers of
polyimide films. (A) Conceptual
scheme for fabricating 3D ser-
pentine microchannel structure
by lamination of the patterned
films and one-step thermal
bonding. (B) Detailed scheme
for a nanoliter reaction space of
rectangular 3D serpentine chan-
nels (200 mm × 125 mm × 1 mm)
with three inlets. (C) Optical
image of the semitransparent
CMR showing the nanoliter
reaction space with three inlets
and one outlet (top view; scale
bar, 500 mm). (D) Optical image
of CMR module (4.5 cm × 3.5 cm,
thickness 0.7 mm). (E) Optical
image of CMR assembly with
metal holders to connect tubes.
SCIENCE sciencemag.org
6 MAY 2016 • VOL 352 ISSUE 6286 691

RESEARCH
| REPORTS
of which was patterned by ultraviolet laser ab-
lation as previously reported (18–20). The internal
volume for the reaction space, 25 nl, is provided
by the rectangular serpentine channels (width
200 mm, height 125 mm, length 1 mm) within the
dotted circle in Fig. 1B, as designed with the help
of the simulation study (Fig. 1 and supplemen-
tary materials). We confirmed that the CMR re-
mained chemically inert toward organolithium
reagents and also robust to several hours of ex-
posure to liquid nitrogen temperature and ap-
plied pressures of 16.2 MPa, as thoroughly verified
in the previous reports (19, 20).
To demonstrate the chemoselectivity attaina-
ble using the CMR, we explored the anionic Fries
rearrangement (21, 22), a well-known reaction
that is often applied to total synthesis of natural
products in organic synthesis (23). For example,
ortho-lithiated O-aryl carbamates undergo facile
intramolecular carbamoyl migration to give a
constitutionally isomeric o-lithiated salicylamide
as the temperature is raised from –78°C to room
temperature. We began our investigation by con-
ducting the reaction using o-iodophenyl diethyl-
carbamate (1) as a precursor (Fig. 2A). The
iodine-lithium exchange of 1 followed by reac-
tion of the resulting aryllithium intermediate 2a′
and/or rearranged 3a′ with methyl chloroform-
ate as an electrophile was conducted at 25°C in
two types of microfluidic devices: (i) a stainless
steel modular microreactor (MMR) consisting of
two T-shaped micromixers and microtube reac-
tors having a circular cross-sectional channel
(inner diameter 250 mm to 1000 mm; see sup-
plementary materials) that is typically used for
flash chemistry (10–12), and (ii) the polyimide-
based CMR. Compound 2a is produced from
aryllithium intermediate 2a′, whereas compound
3a is from rearranged intermediate 3a′.
the starting compound 1 remaining unreacted was
25% because of the mixing efficiency of ~80%,
which is in good agreement with the simulation
result (see supplementary materials). For a better
yield, the number of equivalents of PhLi was in-
creased from 1.05 to 1.2, which raised the yield
of 2a to 91% without any contamination from
isomeric 3a (Fig. 2A, entry 8), achieving >99%
selectivity.
Similar chemoselectivity was attained for Fries
rearrangements with various electrophiles such
as methyl chloroformate, 4-nitrobenzoyl chloride,
chlorotributylstannane, phenyl isocyanate, and
acetyl chloride. The products derived from the
unrearranged intermediate and the rearranged
intermediate were selectively obtained at will
using a suitable device under the optimized con-
ditions (Fig. 2B). With the CMR, 2a to 2d derived
from the unrearranged intermediate 2a′ were
produced in good yields without any detectable
formation of isomeric by-products by gas chro-
matography and nuclear magnetic resonance
(NMR) spectroscopy, whereas 3a to 3d derived
from the rearranged intermediate 3a′ were se-
lectively obtained using MMR-1.
Much more challenging to outpace than the
migration of the carbamoyl groups is the anionic
Fries rearrangement of esters involving acyl
group migration, because of the stronger reac-
tivity of the acyl group relative to the carbamoyl
group and thus the much faster nature of the
reaction. We next explored these reactions using
We sought to achieve high chemoselectivity
between the two isomeric products. The flow
reaction using the MMR-1 microfluidic device
(internal volume of R1, 79 ml) with residence time
in R1 of 628 ms produced only 3a delivering a
91% yield (Fig. 2A, entry 1); 2a derived from
unrearranged 2a′ was not detected at all. Uni-
molecular Fries rearrangement of 2a′ to 3a′ was
evidently complete prior to the bimolecular trap-
ping reaction with methyl chloroformate. When
the internal volume of the microreactors was
decreased from 39 ml (MMR-2) to 0.49 ml (MMR-6),
the yield of 2a increased with the corresponding
decrease in the residence time. However, a small
amount of 3a was still produced even with the
shortest residence time provided by MMR-6 (Fig.
2A, entry 6). In contrast, only 2a was produced,
with no detected 3a, when the reaction was con-
ducted in the CMR, where residence time was
0.33 ms for the reaction volume of 25 nl and a
total flow rate of 4.5 ml/min, delivering a 74%
yield (Fig. 2A, entry 7). This indicates that, by
virtue of the superior mixing and residence time
control of the CMR, the bimolecular reaction to
generate 2a′ and the subsequent bimolecular
trapping reaction with methyl chloroformate were
successfully accomplished before the unimolecular
Fries rearrangement could occur. The portion of
B
Fig. 2. Control of anionic Fries rearrangement reactions of carbamates. (A) Effect of residence time
on the product selectivity. (B) Exceptional chemoselectivity for anionic Fries rearrangements with various
electrophiles.
692 6 MAY 2016 • VOL 352 ISSUE 6286
sciencemag.org SCIENCE

RESEARCH
| REPORTS
various microfluidic devices (MMR-1 to -6 and
CMR) in the range of reaction times from 628 ms
to 0.33 ms (Fig. 3). The strong reactivity of the
acyl group required lowering of the reaction
temperature to –70°C. Even at this low temper-
ature, a total flow rate of 4.5 ml/min could still be
maintained, demonstrating the CMR’s capability
to withstand high pressure up to 160 bar (19, 20).
Figure 3 shows that the yield of 5a derived
from the unrearranged intermediate increased
with decreasing reaction time, but an acceptable
yield was not attained until the reaction time
was reduced to 4 ms, presumably because of both
rearrangement and decomposition of the inter-
mediate. The selectivity diminished for residence
times longer than 4 ms (table S5). However, the
use of CMR gave rise to the formation of the
desired product 5a in 86% yield with no detect-
able isomeric by-products derived from acyl mi-
gration. The extremely short residence (0.33 ms)
achieved by CMR is responsible for the selective
formation of 5a, avoiding the rearrangement
and decomposition. For a comparison, the reac-
tion was conducted in a conventional flask. A
solution of PhLi was added to a solution of 2-
iodophenyl benzoate (4a) at –70°C, and then chlo-
rotributylstannane was added to the resulting
solution within 1 s. The reaction gave the acyl-
migrated product in poor yield (12%) with no
desired 5a. A substantial quantity of by-products
was confirmed by 1H NMR and gas chromatog-
raphy, as expected. Furthermore, using the CMR
at a residence time of 0.33 ms, the aryl compounds
bearing benzoyl, pentanoyl, and acetyl groups
(4) reacted with various electrophiles to give the
desired products in good yields (Fig. 4A). Even in
the case of an aryl iodide bearing a sterically
small acetyl group, the desired products (5i
and 5j) were successfully obtained in moderate
yields without isomeric by-products from the
acyl migration.
The 25-nl reactor volume of the CMR may give
the impression that the throughput must be very
small. To demonstrate the suitability of the CMR
for practical production of specialty chemicals
such as drugs, we applied it to the production
of a biologically active compound having anthel-
mintic activity, 2-(acetyloxy)-5-chloro-N-(2-chloro-
nitrophenyl)-benzamide (afesal) (24), which was
successfully synthesized (Fig. 4B). The sequential
bimolecular reactions, including I-Li exchange
of 4-chloro-2-iodophenyl acetate (6) to generate
the corresponding aryllithium intermediate and
its trapping reaction with 2-chloro-1-isocyanato-4-
nitrobenzene, produced afesal in 67% isolated
yield. Although the O-acetyl functional group has
been reported to be incompatible with organo-
lithium reactions, the O-acetyl function was tol-
erated on the submillisecond time scale of the
CMR device to provide a straightforward syn-
thetic route for afesal. The rapid continuous-
flow nature of the CMR delivered a throughput
of 5.3 g/hour. Thus, stacking multiple small and
portable CMR systems, which individually fit in
one hand, has great potential for scalable indus-
trial chemical synthesis.
Fig. 3. Effect of residence time on the yield of the product derived from the unrearranged
intermediate versus the anionic Fries rearrangement of an ester.
REFERENCES AND NOTES
1. I. R. Sims, Nat. Chem. 5, 734–736 (2013).
2. G. M. Whitesides, Nature 442, 368–373 (2006).
3. K. S. Elvira, X. Casadevall i Solvas, R. C. R. Wootton,
A. J. deMello, Nat. Chem. 5, 905–915 (2013).
4. K. F. Jensen, B. J. Reizman, S. G. Newman, Lab Chip 14,
3206–3212 (2014).
5. S. V. Ley, D. E. Fitzpatrick, R. M. Myers, C. Battilocchio,
R. J. Ingham, Angew. Chem. Int. Ed. 54, 10122–10136 (2015).
6. T. Illg, P. Löb, V. Hessel, Bioorg. Med. Chem. 18, 3707–3719
(2010).
7. K. Geyer, T. Gustafsson, P. H. Seeberger, Synlett 2009,
2382–2391 (2009).
8. T. Tsubogo, H. Oyamada, S. Kobayashi, Nature 520, 329–332
(2015).
9. V. Hessel, E. Shahbazali, T. Noel, S. Zelentsov, ChemBioEng
Rev. 1, 244–261 (2014).
10. J. Yoshida, A. Nagaki, T. Yamada, Chemistry 14, 7450–7459
(2008).
11. H. Kim, A. Nagaki, J. Yoshida, Nat. Commun. 2, 264 (2011).
12. J. Yoshida, H. Kim, A. Nagaki, ChemSusChem 4, 331–340
(2011).
13. C. J. Seeton, Tribol. Lett. 22, 67–78 (2006).
14. D. Bökenkamp, A. Desai, X. Yang, Y.-C. Tai, E. M. Marzluff,
S. L. Mayo, Anal. Chem. 70, 232–236 (1998).
Fig. 4. Control of anionic Fries rearrangement of esters via a submillisecond-controlled CMR
device. (A) Reaction of unrearranged intermediates with various electrophiles. Isolated yields are reported
unless otherwise indicated. (B) Straightforward synthesis of the anthelmintic compound afesal.
SCIENCE sciencemag.org
6 MAY 2016 • VOL 352 ISSUE 6286 693

RESEARCH
| REPORTS
15. R. H. Liu et al., J. Microelectromech. Syst. 9, 190–197
(2000).
16. H. Aref, Phys. Fluids 14, 1315–1325 (2002).
17. Y. Z. Liu, B. J. Kim, H. J. Sung, Int. J. Heat Fluid Flow 25,
986–995 (2004).
18. K.-I. Min et al., Angew. Chem. Int. Ed. 49, 7063–7067 (2010).
19. K.-I. Min, J. Im, H.-J. Lee, D.-P. Kim, Lab Chip 14, 3987–3992
(2014).
23. G. Bringmann et al., J. Org. Chem. 67, 5595–5610 (2002).
24. J. E. Semple, J.-F. Rossignol, Preparation of benzamide
compounds as glutamate receptor modulators and therapeutic
agents, PCT Int. Appl. WO 2012058378 (2012).
simulation and analysis with K.-I.M.; and D.-P.K. and J.Y. directed
the project.
SUPPLEMENTARY MATERIALS
www.sciencemag.org/content/352/6286/691/suppl/DC1
Materials and Methods
Figs. S1 to S3
ACKNOWLEDGMENTS
Supported by National Research Foundation of Korea grants
2008-0061983, NRF-2015R1D1A3A01019112, and NRF-
Tables S1 to S6
20. K.-I. Min, J.-O. Kim, H. Kim, J. Im, D.-P. Kim, Lab Chip 16,
977–983 (2016).
21. K. Fries, G. Finck, Ber. Dtsch. Chem. Ges. 41, 4271–4284
(1908).
2014M1A8A1074940 and by Japan Society for the Promotion of
Science Grant-in-Aid for Scientific Research (S) 26220804. Author
contributions: H.K. and K.-I.M. conceived the concept at POSTECH;
K.-I.M. fabricated the device; H.K. conducted synthetic experiments
at POSETCH and Kyoto University with K.I.; D.J.I. conducted CFD
NMR Spectra
References (25–32)
22 December 2015; accepted 28 March 2016
10.1126/science.aaf1389
22. R. Martin, Org. Prep. Proced. Int. 24, 369–435 (1992).
GEOMORPHOLOGY
shear velocity (U*, meters per second) slightly
exceeds the critical value (U*c) at bankfull con-
ditions. This is called the near-threshold chan-
nel, for which data and theory indicate that
U*/U*c ≈ 1.1 (Fig. 1) (15, 16). Some studies, how-
ever, suggest that this treatment ignores details
of climatic variability that may exert a substan-
tial influence on landscape evolution (1, 17, 18).
Observations reveal that the statistical distribu-
tions of discharge in many rivers possess a power-
law tail (12, 13, 19), whose exponent changes
with climatic setting (17). These observations
have been interpreted to mean that channel
shape may be controlled by climate and, for
rivers with sufficiently heavy-tailed (log-log slope <
–2) discharge distributions, that the rate of sed-
iment transport could be dominated by extreme
events due to climatic variability (17), which
prevents rivers from achieving an equilibrium
geometry over geologic timescales (1, 3). Under-
standing the role of rivers in landscape evolution
requires reconciling the proposed importance of
climatic variability on channel form and dynam-
ics with the apparent equilibrium behavior im-
plied by near-universal hydraulic geometry scaling
relations (15, 20, 21).
Self-organization of river channels as
a critical filter on climate signals
Colin B. Phillips1* and Douglas J. Jerolmack2
Spatial and temporal variations in rainfall are hypothesized to influence landscape evolution
through erosion and sediment transport by rivers. However, determining the relation between
rainfall and river dynamics requires a greater understanding of the feedbacks between flooding
and a river’s capacity to transport sediment. We analyzed channel geometry and stream-flow
records from 186 coarse-grained rivers across the United States.We found that channels adjust
their shape so that floods slightly exceed the critical shear velocity needed to transport bed
sediment, independently of climatic, tectonic, and bedrock controls. The distribution of
fluid shear velocity associated with floods is universal, indicating that self-organization of
near-critical channels filters the climate signal evident in discharge. This effect blunts the
impact of extreme rainfall events on landscape evolution.
nderstanding the control of climate on the
geometry and erosion rate of rivers is es-
sential for reconstructing the geologic his-
tory of landscapes and for predicting the
response of rivers to human-accelerated
generated by climate—defined here as the magni-
tude, frequency, and phase of precipitation—is
represented by a characteristic flood (14). This
“bankfull” flood is the event whose frequency
and magnitude combine to move the most sed-
iment in the long-time limit, and it dictates
channel size (Fig. 1). The second principle applies
to gravel-bed rivers (median bed particle diam-
eter, D ≥ 10 mm), where sediment moves pre-
dominantly as bed load. Gravel-bed rivers adjust
their geometry so that the width-averaged fluid
U
climate change. A natural assumption is to link
river erosion to climate through precipitation
(1–3), yet demonstrating a clear relation is unex-
pectedly challenging (4–7). One reason is that bed-
rock river incision occurs primarily by abrasion
due to the collision of particles with the stream
bed (8) and “plucking” of loose blocks (2), and
therefore it depends on sediment supply as well
as precipitation. Another reason is that bedrock
channel geometry both influences and adjusts to
incision rate (4, 9–11). The effects of climatic varia-
bility (11–13) and bedrock channel geometry (9, 10)
on river incision rates have been explored primar-
ily with numerical models, but empirical observa-
tions remain limited.
Climatic effects on river dynamics are typically
characterized by discharge (Q, cubic meters per
second), which is strongly related to precipitation
(22), and erosion is often modeled using stream
power (the product of discharge and slope, S).
Bed-load motion, however, is driven by applied
Q >> Q bf
h>hbf
U*= ghS
h=hbf
h
S
S
D
tf
ts
U*c
In contrast to the case of bedrock systems, our
understanding of the geometry of alluvial rivers
(channels whose bed and banks are composed of
mobile sediment) is built upon two empirically
vetted theoretical principles. The first is “geomor-
phic work,” in which the wide range of flows
Q = Q bf
h=hbf
time (t)
Fig. 1. Definition sketches. (A) Channel cross section illustrating adjustment to near-threshold bed-load
transport; red regions are above the threshold of motion. The top panel shows flow exceeding bankfull
conditions that induces transport on the banks, resulting in erosion and widening of the channel, which
returns the system to near-threshold conditions (bottom). U*bf was computed from channel surveys of S
and hbf. (B) Definition sketch of a flood, with relevant parameters shown. The gray shaded area (from the
starting time ts to the finishing time tf) represents the part of a flood that is included in the integral for
1St. Anthony Falls Laboratory, University of Minnesota,
Minneapolis, MN 55414, USA. 2Department of Earth and
Environmental Science, University of Pennsylvania,
Philadelphia, PA 19104, USA.
tf
ts
3=2
potential transport, which is calculated as T ¼ ∫ ðU2* − U*2cÞ dt=ðgD520Þ for U* ≥ U*c (26).
*Corresponding author. Email: colinbphillips@gmail.com
694 6 MAY 2016 • VOL 352 ISSUE 6286
sciencemag.org SCIENCE

Submillisecond organic synthesis: Outpacing Fries
rearrangement through microfluidic rapid mixing
Heejin Kim, Kyoung-Ik Min, Keita Inoue, Do Jin Im, Dong-Pyo
Kim and Jun-ichi Yoshida (May 5, 2016)
Science 352 (6286), 691-694. [doi: 10.1126/science.aaf1389]
Editor's Summary
Rapid mixing to race past rearrangement
Chemistry relies on encounters between reactive partners. Sometimes one of the partners changes
shape during the wait, spoiling the desired outcome. Kim et al. designed a microfluidic device to keep
such botched encounters from happening. The device operates at low temperatures to keep individual
reactants from isomerizing. It also achieves fast flow rates to maximize encounters between reactants on
a microsecond time scale. The authors showcase the device by achieving bimolecular carbon-carbon
coupling before one of the reagents can undergo a Fries rearrangement that would shift a neighboring
group to the coupling site.
Science, this issue p. 691
This copy is for your personal, non-commercial use only.
Visit the online version of this article to access the personalization and
article tools:
Article Tools
Permissions
http://science.sciencemag.org/content/352/6286/691
Obtain information about reproducing this article:
http://www.sciencemag.org/about/permissions.dtl
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week
in December, by the American Association for the Advancement of Science, 1200 New York
Avenue NW, Washington, DC 20005. Copyright 2016 by the American Association for the
Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.