Accelerated Chemical Reactions and Organic Synthesis in Leidenfrost Droplets

Article abstract of DOI:10.1002/anie.201605899

Leidenfrost levitated droplets can be used to accelerate chemical reactions in processes that appear similar to reaction acceleration in charged microdroplets produced by electrospray ionization. Reaction acceleration in Leidenfrost droplets is demonstrated for a base-catalyzed Claisen–Schmidt condensation, hydrazone formation from precharged and neutral ketones, and for the Katritzky pyrylium into pyridinium conversion under various reaction conditions. Comparisons with bulk reactions gave intermediate acceleration factors (2–50). By keeping the volume of the Leidenfrost droplets constant, it was shown that interfacial effects contribute to acceleration; this was confirmed by decreased reaction rates in the presence of a surfactant. The ability to multiplex Leidenfrost microreactors, to extract product into an immiscible solvent during reaction, and to use Leidenfrost droplets as reaction vessels to synthesize milligram quantities of product is also demonstrated.

Full text of DOI:10.1002/anie.201605899

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201605899
German Edition: DOI: 10.1002/ange.201605899
Reaction Kinetics
Accelerated Chemical Reactions and Organic Synthesis in Leidenfrost
Droplets
Ryan M. Bain, Christopher J. Pulliam, Fabien Thery, and R. Graham Cooks*
In memory of Klaus Biemann
Abstract: Leidenfrost levitated droplets can be used to
accelerate chemical reactions in processes that appear similar
to reaction acceleration in charged microdroplets produced by
electrospray ionization. Reaction acceleration in Leidenfrost
droplets is demonstrated for a base-catalyzed Claisen–Schmidt
condensation, hydrazone formation from precharged and
neutral ketones, and for the Katritzky pyrylium into pyridi-
nium conversion under various reaction conditions. Compar-
isons with bulk reactions gave intermediate acceleration factors
milligram scale synthesis using reactions in electrosprayed
droplets have also been reported using multiplexed ESI
emitters.
The Leidenfrost effect, discovered in 1756, occurs when
a liquid is dropped on a surface that is heated to a temperature
significantly higher than the boiling point of the liquid
(Figure 1). As the droplet approaches the heated surface, an
[6]
(
2–50). By keeping the volume of the Leidenfrost droplets
constant, it was shown that interfacial effects contribute to
acceleration; this was confirmed by decreased reaction rates in
the presence of a surfactant. The ability to multiplex Leiden-
frost microreactors, to extract product into an immiscible
solvent during reaction, and to use Leidenfrost droplets as
reaction vessels to synthesize milligram quantities of product is
also demonstrated.
E
lectrospray ionization (ESI) is a commonly used ionization
technique for mass spectrometry (MS) in which analyte ions
are generated and transferred from solution into the gas
phase. In ESI, analytes in a charged solution are sprayed
pneumatically to produce a plume of charged microdroplets.
These droplets undergo desolvation and Coulombic fission
events to produce numerous smaller and more concentrated
Figure 1. The Leidenfrost reaction system used in these experiments.
Additional reaction mixture was added with a pipette to keep the
volume of the Leidenfrost droplets constant. The added droplets
impact the surface and fuse with the main levitated droplet, which sits
atop a cushion of solvent vapor.
[
1]
droplets. These microdroplets are the nexus for accelerated
[
2]
chemical reactions. On-line monitoring experiments have
demonstrated that increasing the distance between the
sprayer and the MS increases the product to starting material
ratio, indicating the occurrence of rapid reactions in the
droplets. Zare and co-workers demonstrated that the
reaction rate acceleration in the spray (compared to the
bulk) for the Pomeranz–Fritsch synthesis of isoquinoline was
more than six orders of magnitude. In a study of the base-
catalyzed Claisen–Schmidt condensation, we showed that
acceleration is not simply a result of increased concentration
insulating vapor layer is formed that keeps the solution from
[7]
boiling rapidly. With this vapor cushion in place, the droplet
levitates, and solvent gradually evaporates. The Leidenfrost
phenomenon has been used as a method of thermal desorp-
tion employing dielectric barrier discharge ionization to
[
3]
[
8]
record mass spectra. Leidenfrost droplets have also been
used as microreactors for the preparation of metal nano-
[
4]
[9]
particles; for example, Park and co-workers used this
technique to prepare nanoporous Pd by placing a Pd(NO3)2
solution onto a hotplate and allowing the droplet to com-
[
5]
[9b]
but is influenced by interfacial effects. Successful efforts at
pletely evaporate, leaving Pd black.
Inverse Leidenfrost
experiments have also been performed by levitating droplets
[
10]
on liquid nitrogen. For traditional Leidenfrost droplets, the
vapor pressure and droplet evaporation processes have been
well characterized. Furthermore, Leidenfrost droplets have
[
*] R. M. Bain, C. J. Pulliam, F. Thery, Prof. R. G. Cooks
Department of Chemistry, Purdue University
[11]
560 Oval Drive, West Lafayette, IN (USA)
been shown to have localized pH extremes within the
E-mail: cooks@purdue.edu
[
8]
droplets. Herein, we demonstrate the use of the Leidenfrost
effect to perform accelerated chemical synthesis in droplets.
We show that interfacial effects are involved in acceleration
and prepared macroscopic amounts of synthetic products
within minutes.
Homepage: http://aston.chem.purdue.edu
F. Thery
On leave from: Laboratory of Neuroscience
University of Lille 1 (France)
Supporting information for this article can be found under:
http://dx.doi.org/10.1002/anie.201605899.
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

Angewandte
Communications
Chemie
Droplet levitation and manipulation techniques have
been characterized using high electric fields in field-induced
droplet ionization (FIDI) experiments. This technique dates
back nearly a century and was initially applied to neutral
measure of reaction rate acceleration as it does not take into
account ionization efficiency differences, amongst other
factors. Acceleration factors were measured from the prod-
uct/reagent ratios for the Leidenfrost reaction with the 5 mm
solution vs. 20 mm bulk solution (to account for the concen-
tration increase in the Leidenfrost droplets that is due to
evaporation). The result is a product (3)/reagent (1) ion ratio
that is notably higher (by a factor of 4) than that for the bulk
solution (see the Supporting Information, Section S4). The
addition of acid increases product formation in both experi-
ments. The acceleration is also evident from the fact that the
product/reactant ratio in the Leidenfrost droplets after 2 min
(0.4) approximately matched that (0.6) reached after 1 h in
the corresponding bulk experiment (Table 1).
[
12]
droplets. Later, Grimm and Beauchamp reported the use of
strong electric fields to extract positive and negative ions from
[
13]
neutral droplets. Further work was performed with on-line
FIDI-MS to study the chemistry at the air–droplet inter-
[
14]
face. Smith and co-workers have used similar contactless
droplet levitation techniques making use of acoustic waves to
study the fundamentals of ionization and the properties of the
[
15]
air–water interface.
Hydrazone formation between cortisone (1) and Girardꢀs
Reagent T (2) to form the precharged hydrazone product 3
(
Scheme 1) was selected as a model reaction because a similar
The reaction of neutral isatin (4) with phenylhydrazine (5)
to form hydrazone 7 via an observable precondensation
intermediate 6 (Scheme 2) was also investigated. This reac-
tion is known to be accelerated in the charged droplets
generated by nESI when the distance between the spray
source and the MS inlet is increased. This is ascribed to
solvent evaporation and increased analyte concentrations as
[17]
well as to increased surface/volume ratios. The Leidenfrost
version of this reaction also showed a substantial acceleration
Scheme 1. Cortisone (1) and Girard’s reagent T (2) react to form
hydrazone 3.
system has been characterized in the bulk phase and also by
using acceleration in desorption electrospray ionization
[
16]
(
DESI) droplets.
The results (Table 1) show significant
reaction acceleration, expressed as an acceleration factor,
which was defined as the ratio of the MS product ion signal to
the reagent ion signal in the Leidenfrost droplet versus the
ratio in the bulk experiment. This ratio of ratios is a crude
Scheme 2. Reaction of 4 and 5 to form hydrazone 7 via intermediate 6.
Table 1: Bulk-phase and Leidenfrost droplet reactions and acceleration factors (acc. factors).
2
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!

Angewandte
Communications
Chemie
ment, eight ceramic spacers were used to isolate eight
Leidenfrost droplets and keep them from fusing. The droplets
were levitated for approximately 2 min while their individual
volumes (0.5 mL) were maintained by pipetting in fresh
reaction mixture. When the spacers were removed to allow
the droplets to fuse, a single large droplet formed, and it was
immediately extracted from the surface with a pipette. The
resulting sample (18 mg) was analyzed by MS and found to be
Figure 2. The bulk-phase reaction of 4 and 5 showed mainly reagent 5
and a small amount of intermediate 6 after 2 min (a) whereas the
parallel Leidenfrost experiment led to significant product (7) formation
9
0% pure by standard addition of reagent 8 (10% uncon-
verted 8). This simple experiment illustrates the ease of
manipulating these droplets, multiplexing this mesoscale
reactor system, and synthesizing organic compounds on the
milligram scale. The Supporting Information (Section S5)
includes an image of the multiplexed experimental setup.
One feature of Leidenfrost droplets is reaction acceler-
ation. Another is the ability to use immiscible solvents and so
to perform liquid/liquid extractions in the reacting solution,
removing product as it is formed and so driving the reaction.
This capability was demonstrated by the base-catalyzed
aqueous hydrolysis of adipic acid monoethyl ester (Sec-
tion S1). The ester was placed in a chloroform droplet that
was levitated in contact with a methanol/water droplet
containing base. The data in the Supporting Information
indicate extensive hydrolysis with product transfer into the
aqueous phase during the reaction (Section S6).
(
b).
factor (ꢁ 6 after correction for concentration effects) over the
bulk phase reaction (Table 1 and Figure 2).
The Katritzky reaction (Scheme 3) of 2,4,6-triphenylpyry-
lium (8) with 4-methoxyaniline (9) to yield the corresponding
Scheme 3. The pyrylium cation 8 reacts with 4-methoxyaniline (9) to
produce pyridinium cation 10.
Lastly, a base-catalyzed Claisen–Schmidt condensation
was investigated because the acceleration of this reaction has
[
5,19]
been well-characterized by paper spray and ESI.
pyridinium cation 10 was also investigated. This base-
catalyzed reaction has previously been investigated by
paper spray ionization (thin-film experiment), and showed
6-Hydroxy-1-indanone (11) was reacted with benzaldehyde
(12) in the presence of base to form the condensation product
13 (Scheme 4). Both the starting material (11) and the
[18]
a significant reaction acceleration.
The reaction (2 min,
constant volume by addition of reaction solution) in Leiden-
frost droplets showed acceleration factors of 50 and 8 for the
reactions performed with and without the base catalyst,
respectively (Figure 3, Table 1). Note that the time taken for
the bulk reaction to reach the same product/reagent ion
intensities (an alternative measure of the acceleration factor)
is > 10 h vs. the 2 min Leidenfrost time.
Scheme 4. Condensation of 11 and 12 to form product 13, examined
in the negative-ion mode.
The potential for scaling up of Leidenfrost reactions was
explored using the Katritzky reaction. In a prototype experi-
product (13) are presumably deprotonated at the aromatic
hydroxy group and have similar ionization efficiencies in the
negative-ion mode. The peak intensities in the mass spectra
(Figure 4) reflect the extent of the reaction, which is
significantly greater for the Leidenfrost experiment than in
the bulk phase (acceleration factor: 17).
To explore the role of the surface in Leidenfrost droplet
acceleration, surfactants were added to the reaction mixture.
The addition of triton X-100 (Section S7) had a significant
effect on reaction acceleration. The addition of 0.01% (v/v)
triton had no effect, but a 1% (v/v) triton-containing reaction
mixture, when subjected to the same Leidenfrost conditions,
showed significant suppression of acceleration (Figure 4),
whereas 5% (v/v) triton suppressed the reaction to bulk
values. Higher concentrations of triton did not allow the
droplet to be maintained at the Leidenfrost temperature. A
blank bulk-phase experiment with 1% triton showed no
change in the extent of acceleration. These experiments
Figure 3. Bulk-phase Katritzky reactions with and without base ((a)
and (b)) gave smaller product (10) versus starting material (8) ratios
than the corresponding Leidenfrost experiments ((c) and (d)), which
displayed acceleration factors of 50 and 8, respectively.
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

Angewandte
Communications
Chemie
mixture in aliquots over the 2 min course of droplet levitation
Figure 1). The volume was kept constant by monitoring the diameter
(
of the droplet over the time course of the experiment using a scale
bar. As the droplets concentrate the reaction mixture by a factor of
approximately four over the initial concentration during the constant-
volume experiment, bulk-phase reactions were also performed at
fourfold higher concentrations to correct for the concentration effect.
Additional experiments were performed in which a constant volume
was maintained by addition of pure solvent. Comparisons to bulk
reactions were performed at room temperature at the initial concen-
trations used for the Leidenfrost experiments. These additional
experiments helped separate effects of concentration changes from
those associated with other features of the levitated droplets, notably
interfacial effects.
Leidenfrost experiments were conducted for 2 min in a Petri dish
atop a heater/stirrer with a surface temperature between 400 and
Figure 4. Surface effects in Claisen-Schmidt reactions. a) The 13/11
ratio in the bulk-phase reaction after 2 min. b) The 13/11 ratio in the
Leidenfrost reaction, showing an acceleration factor of 17. The addition
of triton X-100 at 1% and 5% v/v suppressed the acceleration effect to
factors of just 2 and 1 as shown in (c) and (d), respectively. The m/z
peaks with a spacing of 44 are PEG-spaced peaks of triton.
5008C. All analyses were performed by nanoESI using an LTQ XL-
Orbitrap hybrid mass spectrometer (Thermo Fisher Scientific, San
Jose, CA). Additional experimental details on each reaction system
can be found in the Supporting Information, Section S1. All reagents
and solvents were purchased from Sigma–Aldrich (St. Louis, MO)
unless otherwise specified. The determination of the acceleration
factors and the nESI-MS method used for chemical analysis are
discussed in the Supporting Information, Sections S2 and S3.
demonstrate the role of the surface in the acceleration of
these reactions as small concentrations of triton block the
surface and diminish accelerated product formation. Further
experiments on the role of the concentration and details of Acknowledgements
the Leidenfrost methods can be found in the Supporting
Information, Section S8.
We acknowledge the intellectual and artistic contributions of
Heat is also a factor in promoting reactions in Leidenfrost
droplets and has been investigated as a potential source of
acceleration. The temperatures in Leidenfrost droplets are
not well characterized, but there is significant mixing and
a temperature approximately 10–208C below the boiling
point of the solvent is considered to be an acceptable
Stephen T. Ayrton and support from the National Science
Foundation (CHE-1307264). This material is also based upon
work supported by the U.S. Department of Energy, Office of
Science, Office of Basic Energy Sciences, Separations and
Analysis Program (DE-FG02-06ER15807).
[9]
approximation. The effect of temperature was examined
by performing the Claisen–Schmidt condensation in bulk at
Keywords: Leidenfrost effect · mass spectrometry ·
reaction acceleration · reaction kinetics · synthesis
4
58C in methanol (methanol boils at 658C). The increased
temperature provided an acceleration factor of just 2. Even
when the same reaction was performed under reflux, that is, at
temperatures well above those in Leidenfrost droplets, the
acceleration factor was only 5 (significantly less than the
acceleration factor of 17 seen for the Leidenfrost droplet).
In conclusion, the use of Leidenfrost droplets for small-
scale synthesis has been demonstrated through simple exam-
ples, and a variety of chemical reactions have been accel-
erated. The observed acceleration is modest by comparison
with that seen in the much smaller droplets generated by ESI,
which is consistent with the smaller surface/volume ratios of
the larger droplets and with the role played by the interface in
reaction acceleration. Direct evidence for surface reactions
was obtained by experiments with added surfactant. Concen-
tration and temperature effects also play a role in reaction
acceleration. This study suggests that Leidenfrost droplets can
be used to screen reactions and create milligram quantities of
material in minutes.
[1] a) L. Konermann, E. Ahadi, A. D. Rodriguez, S. Vahidi, Anal.
Chem. 2013, 85, 2 – 9; b) N. B. Cech, C. G. Enke, Mass Spectrom.
Rev. 2001, 20, 362 – 387.
[
2] X. Yan, R. M. Bain, R. G. Cooks, Angew. Chem. Int. Ed. 2016,
5, DOI: 10.1002/anie.201602270; Angew. Chem. 2016, 128,
DOI: 10.1002/ange.201602270.
5
[3] a) J. K. Lee, S. Kim, H. G. Nam, R. N. Zare, Proc. Natl. Acad.
Sci. USA 2015, 112, 3898—3903; b) R. M. Bain, C. J. Pulliam,
R. G. Cooks, Chem. Sci. 2015, 6, 397 – 401; c) R. M. Bain, C. J.
Pulliam, S. A. Raab, R. G. Cooks, J. Chem. Educ. 2015, 92, 2146 –
2
151.
4] S. Banerjee, R. N. Zare, Angew. Chem. Int. Ed. 2015, 54, 14795 –
4799; Angew. Chem. 2015, 127, 15008 – 15012.
5] Y. Li, X. Yan, R. G. Cooks, Angew. Chem. Int. Ed. 2016, 55,
3433 – 3437; Angew. Chem. 2016, 128, 3494 – 3498.
[
[
1
[6] T. Mꢂller, A. Badu-Tawiah, R. G. Cooks, Angew. Chem. Int. Ed.
012, 51, 11832 – 11835; Angew. Chem. 2012, 124, 12002 – 12005.
2
[
[
7] D. Quꢃrꢃ, Annu. Rev. Fluid Mech. 2013, 45, 197 – 215.
8] S. Saha, M. K. Mandal, H. Nonami, K. Hiraoka, Anal. Chim.
Acta 2014, 839, 1 – 7.
Experimental Section
The Leidenfrost droplets were maintained at a constant volume of ca.
[9] a) R. Abdelaziz, D. Disci-Zayed, M. K. Hedayati, J.-H. Pçhls,
A. U. Zillohu, B. Erkartal, V. S. K. Chakravadhanula, V. Duppel,
L. Kienle, M. Elbahri, Nat. Commun. 2013, 4, 2400; b) D. W. Lee,
0
.5 mL (orders of magnitude larger than the submicrometer diameter
droplets of ESI) by adding approximately 2 mL of the reaction
4
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!

Angewandte
Communications
Chemie
M. H. Jin, C. B. Lee, S. W. Lee, J. W. Park, D. Oh, J. C. Park, J. S.
Park, Chem. Mater. 2015, 27, 5151 – 5160.
[15] M. S. Westphall, K. Jorabchi, L. M. Smith, Anal. Chem. 2008, 80,
5847 – 5853.
[
[
[
10] M. Adda-Bedia, S. Kumar, F. Lechenault, S. Moulinet, M.
Schillaci, D. Vella, Langmuir 2016, 32, 4179 – 4188.
11] I. V. Kazachkov, WSEAS Trans. Heat Mass Transfer 2012, 7, 47 –
[16] M. Girod, E. Moyano, D. I. Campbell, R. G. Cooks, Chem. Sci.
2011, 2, 501.
[17] R. M. Bain, C. J. Pulliam, S. T. Ayrton, K. R. Bain, Rapid
Commun. Mass Spectrom. 2016, accepted.
[18] X. Yan, R. Augusti, X. Li, R. G. Cooks, ChemPlusChem 2013, 78,
1142 – 1148.
[19] R. M. Bain, C. J. Pulliam, X. Yan, K. F. Moore, T. Mꢂller, R. G.
Cooks, J. Chem. Educ. 2014, 91, 1985 – 1989.
5
7.
12] a) C. T. R. Wilson, Philos. Trans. R. Soc. London Ser. A 1921,
21, 73; b) G. I. Wilson, C. T. R. Taylor, Proc. Cambridge Philos.
Soc. 1925, 22, 728; c) W. A. Macky, Proc. R. Soc. London Ser. A
931, 133, 565 – 587.
13] R. L. Grimm, J. L. Beauchamp, J. Phys. Chem. B 2005, 109,
244 – 8250.
14] R. L. Grimm, R. Hodyss, J. L. Beauchamp, Anal. Chem. 2006, 78,
800 – 3806.
2
1
[
[
8
Received: June 17, 2016
Published online: && &&, &&&&
3
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

Angewandte
Communications
Chemie
Communications
Reaction Kinetics
Chemical reactions can be accelerated in
Leidenfrost levitated droplets in process-
es that are similar to reaction acceleration
in charged microdroplets produced by
electrospray ionization. Acceleration fac-
tors of 2 to 50 compared to the corre-
sponding bulk-phase reactions were ach-
ieved for a range of transformations.
R. M. Bain, C. J. Pulliam, F. Thery,
R. G. Cooks*
&&&& — &&&&
Accelerated Chemical Reactions and
Organic Synthesis in Leidenfrost Droplets
6
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!