Reactions in elastomeric nanoreactors reveal the role of force on the kinetics of the Huisgen reaction on surfaces

Article abstract of DOI:10.1021/ja504137u

The force dependence of the copper-free Huisgen cycloaddition between an alkyne and a surface-bound azide was examined in elastomeric nanoreactors. These studies revealed that pressure and chain length are critical factors that determine the reaction rate. These experiments demonstrate the central role of pressure and surface structure on interfacial processes that are increasingly important in biology, materials science, and nanotechnology.

Full text of DOI:10.1021/ja504137u

Communication
pubs.acs.org/JACS
Reactions in Elastomeric Nanoreactors Reveal the Role of Force on
the Kinetics of the Huisgen Reaction on Surfaces
Xu Han,† Shudan Bian,† Yong Liang,‡ K. N. Houk,‡ and Adam B. Braunschweig*,†
†Department of Chemistry, University of Miami, Coral Gables, Florida 33146, United States
‡Department of Chemistry and Biochemistry, University of California, Los Angeles, Los Angeles, California 90095, United States
S
* Supporting Information
absence of CuI have alternately concluded that the reaction
does23,25 and does not24 proceed under applied pressure. Before
ABSTRACT: The force dependence of the copper-free
Huisgen cycloaddition between an alkyne and a surface-
bound azide was examined in elastomeric nanoreactors.
These studies revealed that pressure and chain length are
critical factors that determine the reaction rate. These
experiments demonstrate the central role of pressure and
surface structure on interfacial processes that are
increasingly important in biology, materials science, and
nanotechnology.
the potential of force-induced reactions on surfaces can be
realized, there is a need for new experimental methods that can
deliver soft organic and biologically active materials onto a
surface, apply a controlled localized force, and print over large
areas with micrometer scale resolution.
We report how arrays composed of approximately 10,000
elastomeric tips26−28 mounted on the piezoactuators of an
atomic force microscope (AFM) can be used to apply a localized
force that accelerates the Huisgen 1,3-dipolar cycloaddition
reaction between azide-terminated monolayers and alkyne inks
in the absence of the CuI catalyst. Upon bringing the tips into
contact with the surfaces, a nanoreactor is formed between the
tips and the surfaces that confines the reactants under an applied
force (Figure 1). Several aspects of this approach are ideal for
patterning applications and for understanding how force and
interface structure contribute to the rate constant of reactions,
namely (1) these tip arrays uniformly deposit delicate organic
and biologically active materials onto surfaces,29,30 (2) the
piezoelectric tip actuation provides precise control over reaction
time and the applied force between the tips and the substrates,
he velocities of reactions in solution k are accelerated at a
T
rate proportional to their activation volume V⧧ and the
applied pressure P according to a relation first described by Van’t
Hoff (eq 1).1
⎛
⎜
⎞
⎟
d lnk
dP
−V⧧
RT
=
⎝
⎠
(1)
T
The effect of isotropic pressure on reactions in solution is now
well established,2−4 and the ability of anisotropic shearing forces
to induce predictable bond cleavage in molecules with
mechanically responsive functional groups has become the
focus of considerable interest because of their utility in stimuli
responsive materials.4−10 Only recently have studies on the role
of isotropic force been extended to include mechanically induced
bond formation,11−17 which could be used to covalently pattern
molecules onto surfaces, leading to breakthroughs in biosensing,
molecular electronics, and the understanding of reactivity at
interfaces. To this end, we have lately demonstrated the first
method to covalently pattern graphene by employing a force-
accelerated Diels−Alder reaction that does not proceed in the
absence of applied pressure.18 However, the mechanisms by
which force accelerates surface reactions remains poorly
understood, because there are few experimental methods capable
of controlling precisely critical reaction parameters such as
reaction time t and applied P. Forces can be controlled accurately
using a single AFM tip,19−22 but the low throughput of this
method necessitates complex single-molecule analyses. Alter-
natively soft lithography has been used to investigate force-
accelerated reactions over large areas,23−25 which simplifies
analysis, but the applied P is difficult to control. As a result, these
AFM and soft lithography experiments can lead to inconsistent
and contradictory results. For example, attempts to use force to
induce the Huisgen cycloaddition reaction between alkyne-
terminated surfaces and azide substrates, or vice versa, in the
Figure 1. (a) Elastomeric tip array coated with an ink mixture consisting
of alkyne and PEG (red) printing onto an azide-terminated Au surface.
(b) Magnified view of the inked tip array approaching the surface. (c)
Tip forms a nanoreactor with the surface by applying a force that
accelerates the Huisgen cycloaddition reaction. (d) After washing, only
covalently bound molecules remain on the surface.
Received: April 25, 2014
Published: July 16, 2014
© 2014 American Chemical Society
10553
dx.doi.org/10.1021/ja504137u | J. Am. Chem. Soc. 2014, 136, 10553−10556

Journal of the American Chemical Society
Communication
and (3) the large patterns (cm2) created by the massively parallel
tip arrays facilitate fluorescence microscopy and electrochemical
analysis of the surface reactions. In these studies we show that the
rate constant k of the Huisgen reaction on surfaces is sensitively
dependent on applied force and, surprisingly, the alkyl chain
length of the azide-terminated monolayer.
The ability of compressive force to induce the Huisgen
reaction was investigated by printing fluorescent 1 and redox-
active 2 alkyne inks31 onto azide-terminated SiO2 and Au
surfaces, respectively. Previously, we have used these same inks
to study the ability of polymer pen lithography26,27 to induce the
CuI catalyzed azide−alkyne click reaction in the absence of
applied force and short (<20 s) print times, and we found that
under those conditions the reaction does not proceed without
the CuI catalyst.32 To carry out the printing, molded elastomeric
tips composed of polydimethylsiloxane and azide-terminated
monolayers on SiO2 and Au were prepared following reported
literature protocols.31,32 In a typical experiment, a mixture
composed of alkyne (3 mM) and polyethylene glycol (PEG)
(2000 g/mol, 5 mg) was dissolved in 1 mL 4:1 EtOH:H2O, and
this mixture was sonicated for 3 min to guarantee solution
homogeneity. The pen array was exposed to O2 plasma to render
the surface of the arrays hydrophilic, and the ink mixture was spin
coated onto the tip arrays (2000 rpm, 2 min). The PEG is added
to the ink mixture to encapsulate the reactive species and ensure
uniform transport through the aqueous meniscus that forms
between the tips and the surface.30 The tip arrays were mounted
onto an AFM specially equipped with a tilting stage to level the
pen array, an environmental chamber to control humidity, and a
customized lithography software to control the force and contact
time between the tips and the surfaces. The forces were varied by
changing the extension of the z-piezoactuators holding the
tips.33,34 Following printing, the surfaces were washed with
copious amounts EtOH and H2O to remove molecules that were
not covalently immobilized from the surface.
Initially, bond formation was confirmed by printing
fluorescent ink 1 and analyzing the resulting patterns by
fluorescence microscopy. In these experiments six different
spots were printed by each tip, where t was held constant, the
force was varied from 0.29 to 0.42 mN, and printings were
repeated at t of 0, 60, 180, 300, 420, and 600 s (Figure 2a). The
force was determined by measuring the tip edge length during
printing, which can be converted to the applied force following a
previously described relation.33 The fluorescence images (λex=
532−587 nm, λem= 608−687 nm) demonstrate that uniform
fluorescent patterns are formed over large areas,31 delivering
evidence of covalent immobilization. Although the fluorescence
data provide only qualitative information on the extent of the
reaction, fluorescence signal-to-background increases with both
force and t (Figure 2b). As a control experiment, fluorescent ink
1 was printed onto an amine-terminated glass surface at 0.42 mN
for 600 s, and no fluorescent pattern was visible after rinsing with
EtOH and H2O, which confirms that both azide and alkyne are
required for ink immobilization.
Figure 2. (a) Fluorescent images of 2 × 3 dot arrays of 1 printed at
different t (0, 60, 180, 300, 420, 600 s) and P (0.29, 0.32, 0.34, 0.37, 0.39,
0.42 MPa). (b) Signal-to-background of the printed features of 2
increases with P and t. Error values were determined from three
independent measurements.
presence of 2 easy to identify, and CV provides quantitative data
on the surface density within the printed features and can also be
used to confirm covalent immobilization on conductive
surfaces.35 For all printing t and P above 0 s and 0.1 MPa, a
redox wave indicative of the fc/fc+ redox couple was observed.
The linear plots of ln(scan rate) vs ln(current) further confirmed
that the molecules were surface confined as a result of triazole
formation.35 A quantitative analysis of the surface density of fc as
a function of t and P (Figure 3a) revealed that conversion
increases with both increasing P and increasing t. As long as some
force is applied, the reaction proceeds to near completion given
long enough time to react. However, in the absence of force or
very short t (<150 s), no triazole formation was observed. The
linear fits of ln([azide]) for a given P provide the different k,
which represents the rate of loss of azide on the surface, because
we assume that the reaction follows pseudo-first order kinetics,
where 2 is present in large excess. The initial concentration of
azide, N3, is based on the number of possible fc units that can be
immobilized in a well-organized monolayer, which is 2.7 × 1014
molecules per cm2.35 It should be noted that when printing under
high force, fc surface densities as high as 4.3 × 1014 molecules per
cm2 have been observed,35−37 and this discrepancy has been
explained by suggesting that subsurface fc is embedded within
the monolayer,37 but in our experiments no surface densities
above 2.7 × 1014 molecules per cm2 were detected. The value of k
ranges from 2.89 0.39 × 10−3 s−1 for a P of 2.72 0.21 MPa to
4.52 0.19 × 10−3 s−1 for a P of 7.95 0.75 MPa (Figure 3b).
This sensitive dependence of k on applied P can explain
inconsistencies in previous studies on the force dependence of
the Huisgen reaction.23,25 The applied P must be controlled
precisely to achieve full conversion; although seemingly
Redox-active ink 2 was subsequently deposited onto Au
surfaces coated with a SAM composed of 11-azido-undecane-1-
thiol, where each tip created a pattern composed of 16 spots that
were printed with identical force and t. The patterning was
repeated at t ranging from 0 to 600 s and P ranging from 0.1−7.95
MPa. The force was measured as described previously33 and then
converted to P by the method of Mullen and Basche.20,31 Cyclic
voltammetry (CV) was used to characterize reactions between 2
and the azide SAMs because the characteristic fc peaks make the
10554
dx.doi.org/10.1021/ja504137u | J. Am. Chem. Soc. 2014, 136, 10553−10556

Journal of the American Chemical Society
Communication
Figure 4. (a) Plot of ln k vs P at different alkyl chain lengths, whose slope
provides V⧧obs. Error values were determined from three independent
measurements. (b) Rotation of the C11 azide molecule sweeps out a
large cylinder. (c) Monolayer is compressed when tips are brought into
contact with surface, decreasing the cylinder volume in the reaction.
Figure 3. (a) The variation of fc surface density on Au surfaces as a
function of P and t. (b) Plots of ln([azide]) to t at different P, whose
slope provides k. Error values were determined from three independent
measurements.
the sum of the activation volume of the cycloaddition reaction
V⧧c and the change in the volume of the monolayer ΔVm (eq 2).
V⧧ = V⧧ + ΔVm
(2)
obs
c
contradictory, all previous studies on the ability of force to
catalyze the Huisgen reaction are likely correct once the role of P
on k is accounted for.
To test this hypothesis, we reasoned that the height of the
monolayer would alter V⧧ because of the differences in the
obs
A further series of control experiments was performed to
confirm the covalent immobilization. Ferrocenecarboxylic acid,
which does not react with N3, was dissolved in a PEG matrix and
printed onto an azido-functionalized Au surface. After printing at
7.95 MPa for 600 s and washing with EtOH and H2O, the
absence of any remaining fc on the surface was confirmed by CV.
To show that the rate acceleration under applied force is not
simply a reflection of increased reactant concentration, the
reaction configuration was inverted: rather than coat the tips with
ink and bring them into contact with the surface, the surfaces
were coated with ink, and the tips were used to locally apply
pressure. The concentrations of 2 spin coated onto the surfaces
were varied over two decades (0.3, 0.6, 3, 15, 30 mM), and in
each case the tips were used to apply a P of 4.78 MPa for 300 s to
print a 4 × 4 array. The surfaces were characterized by CV, and
no significant differences in the value of k were observed.31 We
concluded that the reaction was zero order in alkyne when
present in great excess.
change of ΔVm during compression. Therefore, 2 was printed on
SAMs containing alkane chains of different length between the
thiol and the azide with a printing time of 5 min and at pressures
of 2.72−7.95 MPa (Figure 4a). Based on a maximum packing
density of 2.7 × 1014 molecules per cm2, the measured V⧧
obs
decreases gradually from −189 cm3 mol−1 for an undecyl chain to
−39 cm3 mol−1 for a propyl chain (Table 1).31 This indicates
that, assuming that V⧧ remains constant (−13 cm3/mol), the
c
volume change of the monolayer ΔVm becomes smaller and
smaller with the decrease of the height of the monolayer. A
simple model can semiquantitatively account for the change in
ΔVm (Figure 4b). The free rotation of the C11 azide molecule on
Au in a standing up configuration36 sweeps out a cylinder. From
Table 1. Observed Activation Volumes from Experiments and
Calculations for the Huisgen Reaction on Monolayers with
Different Alkane Chain Lengths
alkane chain length of
N3(CH2)nSH
experimental activation
computed activation
b
The CV studies were used to determine the V⧧ (observed
ab
,
volume (cm3/mol)
volume (cm3/mol)
obs
dependence of ln k on pressure) by plotting ln k versus P (Figure
C11
C8
C6
C5
C3
−189 41
−179 45
−144 45
−137 21
−191
−169
−147
−124
−91
4a).38 An activation volume of −189
41 cm3 mol−1 was
determined for the Huisgen cycloaddition on the C11 11-azido-
undecane-1-thiol SAM. Calculations at the M06-2X/6-31G(d)
level in the gas phase show that the (3 + 2) cycloaddition of alkyl
azide and terminal alkyne has a V⧧ of about −13 cm3 mol−1.31
This value is significantly smaller than that observed in the
experiment.
−39
9
a
Error values were determined from three independent measurements.
Calculated values are a combination of the activation volume of the
b
cycloaddition reaction V⧧ (−13 cm3/mol) and the change in the
We attribute the difference between these two values to the
monolayer compression that occurs during printing. We propose
that the observed activation volume V⧧obs in the nanoreactor is
c
volume of the monolayer ΔVm. Maximum possible surface density was
calculated as 2.7 × 1014 molecules per cm2.31
10555
dx.doi.org/10.1021/ja504137u | J. Am. Chem. Soc. 2014, 136, 10553−10556

Journal of the American Chemical Society
Communication
the density of 2.7 × 1014 molecules/cm2 of the C11 chain on the
surface,35 the occupied area of the single molecule is calculated to
be 0.37 nm2. The height of the C11 monolayer is about 2.0 nm,31
so the volume of the cylinder is 0.74 nm3 or 445 cm3 mol−1.
Calculations indicate that, when two anti conformers in the
middle of the alkane chain are changed to gauche in the reaction,
the height of the C11 monolayer decreases to 1.2 nm.31 This will
result in a Δh (Figure 4c) of −0.8 nm, leading to an ΔVm of
−0.30 nm3 or −178 cm3 mol−1. According to this model, the
computed ΔVm values for the C8, C6, C5, and C3 monolayers
are −156, −134, −111, and −78 cm3 mol−1, respectively,31 and
these data reproduce the trend observed in the experiments
(Table 1). The largest discrepancy between experimental and
computational values occurs with the C3 monolayers, which are
more likely to lie flat on the substrate and are therefore less
compressible along the force vector than the closely packed
monolayers used for computations.
(9) Hickenboth, C. R.; Moore, J. S.; White, S. R.; Sottos, N. R.; Baudry,
J.; Wilson, S. R. Nature 2007, 446, 423.
(10) Larsen, M. B.; Boydston, A. J. J. Am. Chem. Soc. 2013, 135, 8189.
(11) Kersey, F. R.; Yount, W. C.; Craig, S. L. J. Am. Chem. Soc. 2006,
128, 3886.
(12) Black, A. L.; Orlicki, J. A.; Craig, S. L. J. Mater. Chem. 2011, 21,
8460.
(13) Todres, Z. V. Organic Mechanochemistry and Its Practical
Applications; CRC Press: Boca Raton, FL, 2006.
(14) Fanselow, D. L.; Drickamer, H. G. J. Chem. Phys. 1974, 61, 4567.
(15) Wilson, D. G.; Drickamer, H. G. J. Chem. Phys. 1975, 63, 3649.
(16) Kawamura, I.; Degawa, Y.; Yamaguchi, S.; Nishimura, K.; Tuzi, S.;
Saito, H.; Naito, A. Photochem. Photobiol. 2007, 83, 346.
(17) Tabata, M.; Tanaka, Y.; Sadahiro, Y.; Sone, T.; Yokota, K.; Miura,
I. Macromolecules 1997, 30, 5200.
(18) Bian, S. D.; Scott, A. M.; Cao, Y.; Liang, Y.; Osuna, S.; Houk, K.
N.; Braunschweig, A. B. J. Am. Chem. Soc. 2013, 135, 9240.
(19) Paxton, W. F.; Spruell, J. M.; Stoddart, J. F. J. Am. Chem. Soc. 2009,
131, 6692.
In conclusion, we have used massively parallel tip arrays to
form nanoreactors that can apply force to accelerate bond-
forming reactions with negative V⧧ on surfaces. This approach
was used to show that the velocity of the Huisgen reaction on
surfaces is sensitively dependent on force and monolayer chain
length, and a semiquantitative model was developed that explains
the previous contradictory results in the literature. Force may
have an important role in many interfacial processes, and the
approach described in this work can be used to address
challenges in materials science, biology, and nanotechnology.
(20) Stottinger, S.; Hinze, G.; Diezemann, G.; Oesterling, I.; Mullen,
K.; Basche, T. Nat. Nanotechnol. 2014, 9, 182.
(21) Brantley, J. N.; Bailey, C. B.; Wiggins, K. M.; Keatinge-Clayth, A.
T.; Bielawski, C. W. Polym. Chem. 2013, 4, 3916.
(22) Lantz, M. A.; Hug, H. J.; Hoffmann, R.; van Schendel, P. J. A.;
Kappenberger, P.; Martin, S.; Baratoff, A.; Guntherodt, H. J. Science
2001, 291, 2580.
(23) Mehlich, J.; Ravoo, B. J. Org. Biomol. Chem. 2011, 9, 4108.
(24) Spruell, J. M.; Sheriff, B. A.; Rozkiewicz, D. I.; Dichtel, W. R.;
Rohde, R. D.; Reinhoudt, D. N.; Stoddart, J. F.; Heath, J. R. Angew.
Chem., Int. Ed. 2008, 47, 9927.
(25) Rozkiewicz, D. I.; Janczewski, D.; Verboom, W.; Ravoo, B. J.;
Reinhoudt, D. N. Angew. Chem., Int. Ed. 2006, 45, 5292.
(26) Huo, F. W.; Zheng, Z. J.; Zheng, G. F.; Giam, L. R.; Zhang, H.;
Mirkin, C. A. Science 2008, 321, 1658.
(27) Eichelsdoerfer, D. J.; Liao, X.; Cabezas, M. D.; Morris, W.; Radha,
B.; Brown, K. A.; Giam, L. R.; Braunschweig, A. B.; Mirkin, C. A. Nat.
Protoc. 2013, 8, 2548.
(28) Giam, L. R.; Mirkin, C. A. Angew. Chem., Int. Ed. 2011, 50, 7482.
(29) Braunschweig, A. B.; Huo, F. W.; Mirkin, C. A. Nat. Chem. 2009,
1, 353.
ASSOCIATED CONTENT
* Supporting Information
Experimental details and computational data. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
a.braunschweig@miami.edu
■
(30) Huang, L.; Braunschweig, A. B.; Shim, W.; Qin, L. D.; Lim, J. K.;
Hurst, S. J.; Huo, F. W.; Xue, C.; Jong, J. W.; Mirkin, C. A. Small 2010, 6,
1077.
Notes
The authors declare no competing financial interest.
(31) See Supporting Information for details.
(32) Bian, S. D.; He, J. J.; Schesing, K. B.; Braunschweig, A. B. Small
2012, 8, 2000.
(33) Liao, X.; Braunschweig, A. B.; Zheng, Z. J.; Mirkin, C. A. Small
2010, 6, 1082.
(34) Liao, X.; Braunschweig, A. B.; Mirkin, C. A. Nano Lett. 2010, 10,
1335.
(35) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.; Mujsce, A. M.
J. Am. Chem. Soc. 1990, 112, 4301.
(36) Vericat, C.; Vela, M. E.; Benitez, G.; Carro, P.; Salvarezza, R. C.
Chem. Soc. Rev. 2010, 39, 1805.
(37) Pellow, M. A.; Stack, T. D. P.; Chidsey, E. D. C. Langmuir 2013,
ACKNOWLEDGMENTS
■
A.B.B. is grateful to the Air Force Office of Scientific Research
(Young Investigator Award FA9550-13-1-0188) and the Na-
tional Science Foundation (DBI-1340038), and K.N.H. is
grateful to the National Science Foundation (CHE-1059084)
for financial support. Calculations were performed on the
Extreme Science and Engineering Discovery Environment
(XSEDE), which is supported by the NSF (OCI-1053575).
REFERENCES
■
(1) Kuster, C. J. T.; Scheeren, H. W. In High Pressure Chemistry; van
Eldik, R., Klaerner, F.-G., Eds.; Wiley-VCH: Weinheim, 2007; pp 284−
304.
29, 5383.
(38) Jezowski, S. R.; Zhu, L. Y.; Wang, Y. B.; Rice, A. P.; Scott, G. W.;
Bardeen, C. J.; Chronister, E. L. J. Am. Chem. Soc. 2012, 134, 7459.
(2) Blandamer, M. J.; Burgess, J.; Robertson, R. E.; Scott, J. M. W.
Chem. Rev. 1982, 82, 259.
(3) Ribas-Arino, J.; Marx, D. Chem. Rev. 2012, 112, 5412.
(4) Beyer, M. K.; Clausen-Schaumann, H. Chem. Rev. 2005, 105, 2921.
(5) Wiggins, K. M.; Brantley, J. N.; Bielawski, C. W. Chem. Soc. Rev.
2013, 42, 7130.
(6) Caruso, M. M.; Davis, D. A.; Shen, Q.; Odom, S. A.; Sottos, N. R.;
White, S. R.; Moore, J. S. Chem. Rev. 2009, 109, 5755.
(7) Brantley, J. N.; Wiggins, K. M.; Bielawski, C. W. Science 2011, 333,
1606.
(8) Lenhardt, J. M.; Ong, M. T.; Choe, R.; Evenhuis, C. R.; Martinez, T.
J.; Craig, S. L. Science 2010, 329, 1057.
10556
dx.doi.org/10.1021/ja504137u | J. Am. Chem. Soc. 2014, 136, 10553−10556