You are on page 1of 11

Home Search Collections Journals About Contact us My IOPscience

Graphene-nanoplatelet-based photomechanical actuators

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2012 Nanotechnology 23 045501

(http://iopscience.iop.org/0957-4484/23/4/045501)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 138.100.4.44
This content was downloaded on 13/01/2015 at 23:11

Please note that terms and conditions apply.


IOP PUBLISHING NANOTECHNOLOGY
Nanotechnology 23 (2012) 045501 (10pp) doi:10.1088/0957-4484/23/4/045501

Graphene-nanoplatelet-based
photomechanical actuators
James Loomis1 , Ben King1 , Tom Burkhead1 , Peng Xu1 , Nathan Bessler1 ,
Eugene Terentjev2 and Balaji Panchapakesan1
1
Small Systems Laboratory, Department of Mechanical Engineering, University of Louisville,
Louisville, KY 40292, USA
2
Department of Physics, University of Cambridge, Cambridge CB3 0HE, UK

E-mail: b0panc01@louisville.edu

Received 8 November 2011, in final form 27 November 2011


Published 6 January 2012
Online at stacks.iop.org/Nano/23/045501
Abstract
This paper reports large light-induced reversible and elastic responses of graphene nanoplatelet
(GNP) polymer composites. Homogeneous mixtures of GNP/polydimethylsiloxane (PDMS)
composites (0.15 wt%) were prepared and their infrared (IR) mechanical responses studied
with increasing pre-strains. Using IR illumination, a photomechanically induced change in
stress of four orders of magnitude as compared to pristine PDMS polymer was measured. The
actuation responses of the graphene polymer composites depended on the applied pre-strains.
At low levels of pre-strain (39%) the actuators showed reversible expansion while at high
levels (1540%) the actuators exhibited reversible contraction. The GNP/PDMS composites
exhibited higher actuation stresses compared to other forms of nanostructured carbon/PDMS
composites, including carbon nanotubes (CNTs), for the same fabrication method. An
extraordinary optical-to-mechanical energy conversion factor (M ) of 79 MPa W1 for
GNP-based polymer composite actuators is reported.
(Some figures may appear in colour only in the online journal)

1. Introduction actuators already show possible applications encompassing


photonic switches [11], robotics [12], plastic motors [13],
Both in natural and synthetic systems, actuator materials and adaptive micro-mirrors [14]. Recent CNT polymer
change their shape or dimensions upon external stimulus. systems studies have yielded interesting insights into the
The best-known materials for actuation are piezoelectrics [1], nature of photomechanical actuation in these one-dimensional
ferro-electrics [2], shape memory alloys [3], electrostrictive (1D) nanostructures [9, 10, 1518], resulting in renewed
materials [4], liquid crystal elastomers [5], and conducting excitement in photoactive materials and their applications as
polymers [6]. Recent additions include carbon nanotubes micro/nano-optomechanical systems [14, 19, 20].
(CNTs) [7] and porous metallic nanoparticles [8] as Since graphenes 2004 discovery [21], its incredible
they exhibit large stresses and strains from low voltage physical properties have been well documented, including
electromechanical actuation. Similarly, both single and thermal conductivity [22], mechanical strength [23], and
multi-wall CNT (MWCNT) composites have been reported quantum Hall effect at room temperature [24]. While
to undergo photomechanical actuation [9, 10]. Compared numerous articles report on graphene-based composites, most
to electrically/thermally/ionically/phase transition driven of these actually used graphene-like sheets derived from
actuators, light driven actuators have several advantages graphite oxide (GO) or graphite intercalation compounds
including wireless actuation, electromechanical decoupling (GICs). GO and GIC derived fillers in polymer matrices
(and therefore low noise), elimination of electrical circuits can exhibit high electrical conductivities [25], high Youngs
at the point of use, and massive parallel actuation of moduli [26], and easily be functionalized to tailor to
device arrays from a single light source. Photomechanical host polymer properties [27]. While GO and GIC derived

0957-4484/12/045501+10$33.00 1 c 2012 IOP Publishing Ltd Printed in the UK & the USA

Nanotechnology 23 (2012) 045501 J Loomis et al

Figure 1. GNP/PDMS photomechanical actuator characterization and testing: (a) SEM image of GNPs; (b) Raman spectroscopy shift of
PDMS, GNP/PDMS and CNT/PDMS; and (c) layout of the test dynamometer.

filler materials report electrical conductivities better than extraordinary photomechanical effect larger than not only all
those reported for nano-clays [28], these are often lower other carbon-based composites, but also any other material
than the values reported for single-wall CNT fillers [29]. that exhibits a photomechanical effect. These intriguing
Secondly, graphene nanoplatelets (GNPs) and graphene results will allow for expansion of GNP-based micro- and
nanoribbons (GNRs) have properties that are dependent on nano-optomechanical systems.
the number of layers, such as saturable absorption [30],
linear monochromatic optical contrasts [31], and electric field 2. Results and discussion
assisted band gaps [32], which are not exhibited in GO and
GIC derived materials. These interesting properties call for Figure 1(a) shows a scanning electron microscope (SEM)
investigation of graphene-based composite photomechanical image of GNPs deposited on a silicon wafer. During close
actuators. examination, the GNPs were verified to be 3 layers, with
Our work introduces a simple polymer composite system diameters in the range of 12 m. Figure 1(b) shows
whose photomechanical responses to near-infrared (NIR) Raman spectroscopy comparison of plain PDMS polymer,
light are realized solely by incorporation of a homogeneous GNP/PDMS composites, and CNT/PDMS composites. The
dispersion of GNPs within a polydimethylsiloxane (PDMS) Raman peaks of the PDMS are clearly visible in all three
elastomer matrix. Similarly to CNT/PDMS composites, the curves of figure 1(b); the 1582 and 2700 cm1 peaks
actuation responses of the GNP/PDMS composites depended correspond to graphene. The radial breathing mode (RBM),
on the initial applied pre-strain: at low levels (39%), graphite (G), and CNT 2D bands (G0 ) are also clearly
the samples showed reversible expansion; at moderate seen. The SEM and Raman spectroscopy confirmed the high
levels (10%), the samples yielded negligible effect; and purity of these systems. Figure 1(c) presents the experimental
at high levels (1540%), the samples exhibited reversible setup schematic which shows vertical orientation of the
contraction. For GNP/PDMS composites, these correspond test nanocomposite sample within the testing dynamometer
to photomechanical stress changes 2.43.6 times greater assembly.
than for any other carbon form tested. These stress changes Figure 2(a) shows the typical photomechanical actuation
reported are usable and recoverable work done by the and relaxation kinetics during a 3% pre-strain test. For
actuators. An optical-to-mechanical energy conversion factor each pre-strain (340%), the test composites underwent five
(M ) of 79 MPa W1 obtained during testing showed an cycles of NIR illumination on for 60 s, followed by NIR

2
Nanotechnology 23 (2012) 045501 J Loomis et al

Figure 2. Photomechanical responses and actuation kinetics of GNP/PDMS composites: (a) typical photomechanical stress response of the
GNP/PDMS actuator for 3% pre-strain; (b) and (c) actuation and relaxation of the GNP/PDMS actuator, respectively.

illumination off for 30 s. The magnitude of actuation was the nanotubes in the PDMS matrix, the actuation was far
highly repeatable for each test sample. Photomechanical slower with time constants of actuation greater than 20 s
responses started at time t = 0 s and became saturated at for achieving saturation [15]. This shows the dependence
t = 5 s. Similarly, once the light was switched off, the of the actuation kinetics on the number of interconnected
actuators underwent relaxation in t = 5 s. This is significantly nanotubes in the matrix. In marked contrast and regardless
fast actuation compared to past reports on CNT/PDMS of the sample concentrations, GNP/PDMS actuators exhibit
actuators where the actuation achieved saturation at t = fast actuation and relaxation and the quality of the fit
1015 s [15]. The dependences of the actuation and relaxation of the experimental data follows a simple actuation and
responses as a function of time were also examined, with relaxation. This suggests that the energy transfer associated
the experimental data fitted to simple exponential functions with photon absorption leading to subsequent expansive or
by incorporation of a fitting parameter ( ). The experimental contractive actuation and relaxation of the polymeric chains
actuation data was fitted with a simple exponential function is a fast event which is unique to graphene and therefore
of 1 exp[(t/ )], while the relaxation data was fitted a unique photomechanical mechanism. Such fast actuation
with exp[(t/ )]. Both the actuation and relaxation responses and relaxation cannot be explained by simple thermodynamic
followed Debye characteristics [15]. Figure 2(b) shows considerations of steady-state heat flux from irradiated
the experimental actuation response and associated fitting GNPs causing saturation of the photomechanical response.
function when 1.7 s. Figure 2(c) shows the experimental The identical response times for actuation and relaxation
relaxation response and fitting function when 2 s. also suggest constant viscosity and modulus, therefore
When normalized to account for the magnitude of change maintaining the original configuration of the polymer. Future
in stress, the actuation and relaxation responses for all the fast pumpprobe experiments or AFM experiments with
GNP/PDMS test composites were nearly identical. The fast functionalized graphene sheets with single chain polymers
actuation and relaxation of GNP/PDMS composites in both can shed some light on the extraordinary energy transfer
expansive and contractive modes suggests a unique actuation needed to explain this fast actuation and relaxation.
mechanism for GNP/PDMS actuators. Unlike CNT/PDMS In figure 3, starting with a plain PDMS elastomer
actuators which follow a compressed exponential function for (figure 3(a)) and progressing from 0.1 to 5 wt% concentration
actuation [15], GNP/PDMS actuators were found to not only GNP/PDMS composites (figure 3(f)), each plot shows the
follow a simple exponential function during both actuation photomechanical response to 808 nm NIR illumination for
and relaxation, but also exhibit about a three times faster a single 60 s cycle. Since the optical loss of PDMS in the
response. Secondly, the actuation kinetics of the CNT/PDMS NIR region is <0.5 dB cm1 [33], the negligible response in
actuators followed a compressed exponential function only the plain PDMS sample (figure 3(a)) was expected. However,
above the percolation threshold [15]. For values lower than by a concentration of 0.1 wt% GNPs (figure 3(b)), the
the percolation threshold, at 0.02 wt% concentration of photomechanical effect becomes clearly observable through

3
Nanotechnology 23 (2012) 045501 J Loomis et al

Figure 3. Comparison of increasing photomechanically induced stress change in GNP/PDMS composites as a result of NIR illumination
for increasing GNP concentrations: (a) plain PDMS; (b) 0.1 wt% GNPs; (c) 0.5 wt% GNPs; (d) 1 wt% GNPs; (e) 2 wt% GNPs; and
(f) 5 wt% GNPs.

telltale expansion and contraction of the actuator. All the composite sample was used as the benchmark for comparison
composites exhibited that low pre-strain values resulted in throughout this paper. The change from expansion to
film expansion and thus positive-induced stress, while high contraction in all samples is likely a result of orientational
pre-strain values resulted in compression and thus negative effects, similar to those seen in CNT-based actuators [9].
change in stress. More concisely stated, weakly stretched As the applied pre-strains increase the overall test sample
composites show reversible expansion while highly stretched length, the GNPs within the PDMS polymer become more
composites show reversible contraction [9]. aligned/re-arranged with respect to one another. As a result,
From GNP concentrations of 0.12 wt% (figures the macroscopic magnitude of the photomechanical response
3(b)(e)), the magnitude of photoinduced stress was observed is amplified.
to increase, reaching almost four orders of magnitude greater To compare the photomechanical responses of GNP/
than for the pristine PDMS polymer. By a concentration PDMS composites with other forms of carbon/PDMS
of 2 wt%, however, the photomechanical response of the composites, test samples were fabricated for GO/PDMS,
GNP/PDMS composites had become saturated. Comparison CNT/PDMS, and pyrolytic carbon black (CB)/PDMS
with higher GNP concentration of 5 wt% (figure 3(f)) showed using identical fabrication methods. All the samples
only a marginal increase in magnitude of the stress. Therefore, underwent identical testing procedures to the GNP/PDMS
the photomechanical response of the 2 wt% GNP/PDMS composites, Figures 4(a)(d) present these results. All forms

4
Nanotechnology 23 (2012) 045501 J Loomis et al

Figure 4. Comparison of photomechanically induced stress change in 2 wt% composites of various carbon forms: (a) graphite oxide;
(b) carbon nanotubes; (c) carbon black; and (d) graphene nanoplatelets.

of carbon displayed similar photomechanically induced only shear mixed for 5 min before fabricating the actuator
expansion/contraction stress responses to the GNPs, although to obtain this extraordinary M of 79 MPa W1 . It may be
with varying magnitudes. Of these samples, GO/PDMS possible to obtain much larger M values with longer mixing
(figure 4(a)) showed the smallest amount of actuation, from times. In the present study, however, this same short-time
+2 kPa (3% pre-strain) to 9 kPa (40% pre-strain). This shear mixing procedure was followed for all carbon/PDMS
result also shows that studies of graphene composites derived composites, to enable a standardized comparison between
from GO materials are quite different from single, bi-layer, different types of carbon/PDMS actuators. This procedure
or few-layer graphene. The magnitude of induced stress shows the relative ease with which the plate-like GNPs
change in CNT/PDMS (figure 4(b)) was slightly higher, disperse in an elastomeric matrix such as PDMS compared
from +4 kPa (3% pre-strain) to 10 kPa (40% pre-strain). to other carbon forms. The plate-like structure of GNPs may
Finally, CB/PDMS (figure 4(c)) exhibited a stress change enable the individual layers to slide against the polymer,
from +7 kPa (3% pre-strain) to 14 kPa (40% pre-strain). thus achieving rapid and uniform dispersion compared to
The magnitude of the response of GNP/PDMS (figure 4(d)) filament-like CNTs that tend to roll up into particles in the
to stress-induced change ranged from +14 to 36 kPa, an short mixing time. For the GNP/PDMS actuators, when the
average of 4.5 times larger than that of GO/PDMS, 3.6 times actuation is translated into a stroke, this corresponds to a total
that of CNT/PDMS, and 2.4 times that of CB/PDMS for the strain of 25%, significantly larger than lattice expansion
same fabrication methods. The similar magnitude of stress or contraction alone [34]. It is clear that the photomechanical
change seen in CNT/PDMS and CB/PDMS demonstrates actuation of graphene is quite impressive compared to its other
the lack of dispersion of CNTs in the PDMS matrix, so carbon counterparts.
that they act like particles similar to CB rather than tubular Figure 5(a) shows a comparison of the photomechanical
filaments, due to the short mixing times. Past reports have response for three 2 wt% GNP/PDMS samples at pre-strains
shown much larger stress increases for MWCNTs than seen of 3% and 40%. As shown in the figure, actuation and
here [15]. In those studies, the MWCNTs were shear mixed relaxation are relatively consistent throughout the samples
in a high shear laboratory mixer for a minimum of 24 h with marginal dependence on the pre-strains applied, both in
in the PDMS composite before fabricating the actuator. The magnitude of actuation as well as response time. Figure 5(b)
resulting stress of max 50 kPa at 8 mW [9, 15] of absorbed shows the photomechanical stress response versus applied
power (Peffective ) gives rise to a calculated M for CNTs pre-strain, with 1 kPa standard deviation error bars. At
of 6.25 MPa W1 . In the present study, the graphene was low pre-strains (under 15%), the standard deviation is less

5
Nanotechnology 23 (2012) 045501 J Loomis et al

Figure 5. GNP/PDMS actuator characteristics. (a) Comparison of actuation response from three samples at 3% and 40% pre-strains, and
(b) average actuation response versus pre-strain with 1 standard deviation error bars, (c) increase in Youngs modulus of the actuator with
increase in GNP fraction.

than 1.5 kPa; as pre-strain increases, so does the standard well as 40% pre-strain (sample length 70 mm). As expected,
deviation, reaching a maximum of 5.4 kPa at 40% pre-strain. the highest temperature increase also occurs at the highest
This suggests that as pre-strains increase, the orientational GNP concentration (5% GNPs, 0% pre-strain). An 75 C
ordering of GNPs results in larger contraction of the actuator temperature rise quickly diminishes as the distance from the
or better dispersed/aligned composites. Figure 5(c) presents illumination point increases. As pre-strain and thus sample
the change in Youngs modulus (E) as a function of GNP length increase, the amount of graphene exposed to the fixed
fraction dispersed in the polymer. A Youngs modulus size irradiation spot decreases due to orientational ordering,
increase to almost twice that of pristine PDMS elastomer and, therefore, the thermal effects also start to diminish. At
was obtained [35], and thus higher wt% GNP actuators are 40% pre-strain, the maximum temperature increase has fallen
also stiffer. These results suggest that the GNPs in PDMS to 60 C, a 15 C change compared to the unstrained sample.
produced not only a stiffer actuator but also an extraordinary It should be pointed out that at large pre-strain, the overall
optical-to-mechanical energy conversion of 79 MPa W1 . magnitude of the actuation increases, while the temperature
It is common knowledge that any type of energy conversion rise in the actuator decreases. This outcome possibly indicates
is associated with heating effects with only a fraction of the that thermal effects may not be the only effect contributing
energy used to provide useful work. Due to this effect, we to the overall actuation mechanism. We believe that this
studied the macroscopic steady-state temperature increase in response is a result of contributions from electrostatic, elastic,
the actuators. polaronic, and thermal effects that cause this impressive
Laser irradiation on the GNP/PDMS composites was overall photoresponse of GNP/PDMS composites similar to
found to increase actuator temperature at the irradiation spot. CNT/PDMS composites [9, 10, 15].
The temperature decrease as a function of distance from the On exploring GNP/PDMS composites further, questions
illumination source, therefore, was characterized for various arise as to why GNP/PDMS performs better than the other
GNP/PDMS composite concentrations. Figure 6 details this carbon counterparts tested. While it is understandable that
temperature rise for 0% pre-strain (sample length 50 mm) as the GNPs performed better than the CB and GO/PDMS

6
Nanotechnology 23 (2012) 045501 J Loomis et al

Figure 6. Steady-state temperature measurements: temperature decrease along the composite test samples as a result of distance from the
illumination point. 0% pre-strain dimensions and temperature response (left side) are compared with 40% pre-strain dimensions and
temperature response (right side).

mixtures, questions linger as to the reason for the smaller gels [37]. While rapid photomechanical actuation has been
amplitudes of the photomechanical stress for CNT/PDMS. attained in azobenzene containing polymers [38], practical
There are several likely reasons for this which warrant further applications are difficult due to the dependence on the
research. First, the two-dimensional (2D) nature of GNPs precise location of the azobenzene moieties in the dark as
enables sliding of individual platelets on top of one another well as photoplastic effects making them relax slowly, in
and thus results in homogeneous sample dispersions within a 10 min [39]. On the other hand, shape memory (SM)
relatively short time (5 min) shear mixing process. GNPs polymers are quite interesting. When SM polymers are
act like discrete 2D rigid plates. Due to the similar atomic elongated up to 20% followed by irradiation of the actuator
arrangement of carbon and crystal-like behavior, friction at = 260 nm for 60 min, this process releases the stress
and interaction effects between the graphene platelets are and causes relaxation [40]. Such prolonged stress relaxation,
expected to be negligible. However, in the CNT/PDMS however, could lead to a very long actuation time that is
composites, the short time shear mixing was not long not practical for many commercial applications. Recently,
enough to overcome the natural tendency of nanotubes to liquid crystal elastomers (LCEs) have emerged as interesting
form bundles due to their enhanced inter-tube interactions, polymers due to the polymeric network and liquid crystal
thus giving rise to smaller amplitudes of photomechanical anisotropic chain. Thermo-responsive LCEs have shown large
stress. In other words, the nanotubes exhibited particle-like contractions of 3540% throughout the phase transition with
behavior for short mixing times rather than that of a matrix a force of 0.001 MPa C1 , comparable to that of skeletal
of dispersed 1D tubes. Hence the photomechanical stresses muscle [41]. Similarly ferro-electric LCE films have shown
exhibited by CNT/PDMS actuators were quite similar to 4% strains at 1.5 MV cm1 , making them expensive with
CB/PDMS actuators for the short mixing time. Therefore, respect to energy consumption [42]. Compared to these
both the matrix orientation and percolation threshold are previously reported actuators with impressive characteristics,
crucial for enabling high amplitudes of photomechanical the reversible contraction and expansion in GNP/PDMS
stress and faster kinetics for CNT/PDMS actuators [17]. provides fast actuation and relaxation; furthermore, the
Longer mixing times and more sophisticated dispersion actuation is elastic and samples actuate during light on and
methods are needed to separate the nanotubes from the return to the original state within 5 s after light off. While
bundles as well as render them oriented in the PDMS previous nanotube/PDMS composites have exhibited similar
matrix. stresses to the GNP/PDMS composites reported here [15]
The maximum photomechanical stress change reported for different fabrication recipes, nevertheless, GNPs are less
here corresponds to an optical-to-mechanical energy conver- expensive and far easier to disperse compared to CNTs
sion factor, M , of 79 MPa W1 . This conversion factor is in a polymer matrix. With a high energy conversion ratio
significantly higher than for commercial polymers, such as of 79 MPa W1 , GNP-based materials could be one of
polyvinylidene fluoride (PVDF), whose calculated M value is the best material choices for future mechanical actuators,
97 kPa W1 [36], and azobenzene containing polymers and thermoelectric-based devices, and solar energy converters.

7
Nanotechnology 23 (2012) 045501 J Loomis et al

3. Conclusions sample preparation. An 808 nm, 500 mW laser served as


the NIR illumination source. All experiments were conducted
In this study, we have demonstrated the intrinsic photome- in a climate-controlled laboratory. The test equipment was
chanical actuation of graphene-based polymer composites. operated inside a light-isolated enclosure mounted on an
Actuation is witnessed only through the presence of graphene. active air suspension table. Webcams inside the testing
Dispersion of GNPs into a PDMS silicone elastomer matrix enclosure were continuously monitored to ensure that all
resulted in a novel GNP-based photomechanical actuator that control gear operated properly.
showed reversible light-induced elastic expansion and con-
traction. For a 2 wt% sample, a photomechanically induced 4.2. Sample preparation
change in stress of +14 kPa at low pre-strains and 36 kPa
at high pre-strains was successfully demonstrated. With PDMS silicone elastomer obtained from Dow Corning
an optical-to-mechanical energy conversion factor of M (Sylgard 184) was used as the host matrix. PDMS is a two-part
79 MPa W1 , GNP/PDMS composites have outperformed solvent-free flexible silicone organic polymer in the form of
most photoresponsive materials, including all other carbon a base compound with a separate hydrosilane curing agent
forms tested, for the fabrication method described here. that acts as a crosslinker. The term wt% used throughout
Looking forward to a variety of possible future applications of the paper refers to the ratio of carbon additive to PDMS
photomechanical actuators, one can develop surfaces that can base compound. The PDMS composites were fabricated by
undergo simultaneous expansion and contraction. On a small weighing the desired amount of additive (GNPs, CNTs,
scale, this could be used for assembling surfaces to study GO, or CB) and adding to the PDMS base compound. The
mechanical forces on soft matter, such as biological materials additive/base compound combination was then shear mixed
like cells, where the forces are delivered in non-contact for 5 min to facilitate carbon distribution. Taking into account
manner using photomechanical actuation. On a large scale, the additive weight, a crosslinker was added at a ratio of 1:10
a thin film of GNP/PDMS polymer could be utilized as a and further shear mixed for 5 min. To remove trapped air
fundamental smart material in adaptive skins. The temperature pockets, the prepared polymers were degassed for 30 min.
decrease with increase in pre-strain of the actuators could Small amounts of liquid polymer mixtures were deposited on
be useful for developing a new type of strain gauge by the glass slides. A standard spin coating process at 750 rpm
measuring linear temperature change with strain. In the field for 150 s successfully produced nominally 60 m thick films.
of sensing, graphene-based photomechanical actuators could High temperature curing at 125 C for 20 min was employed
be highly useful for sensing adsorbed molecules that result to finish the cross-linking process in the polymer. The PDMS
in a change of the actuators stress state. Finally, the distinct composite sample dimensions were approximately 60 mm
stress response states of the photo-actuators (positive, zero 3 mm (50 mm of active test area, with 5 mm on each side
and negative) could be useful for designing nano-mechanical for test fixture mounting). The level of carbon distribution
memory. Instead of a cantilever gently touching the surface of within the PDMS polymer was qualitatively evaluated in the
the polymer pit to write, it could be stressed simultaneously post-bake composites using optical microscopy. The same
at millions of points using a single light source, enabling shear mixing time of 5 min was used for other carbon forms
nanoscale photomechanical writing. Such photomechanical so as to enable standardized comparison of their performance
writing could be reversible and fast, as well as actuated for such short mixing times.
through control of the number of photons reaching the surface.
We believe that further development of graphene-based 4.3. Stress test experimentation
polymer composites will extend the promising potential of
graphene-based actuation technologies and will serve as a The PDMS composite test samples were mounted vertically
catalyst to inspire continued research into energy conversion between two clamps (figure 1(c) shows this setup). The
devices and systems. bottom clamp was attached to a weighted (50 g) base and
placed on a high accuracy balance (Acculab ALC-80.4). The
4. Methods upper clamp was attached to an automated linear actuator that
was in turn mounted to a high accuracy manual positioning
4.1. General setup stage. The laser diode was placed 75 mm from the middle
of the test strip such that the illumination impacted normal to
The research grade GNPs were purchased from Cheap the PDMS surface. Deformation in the composite strips as a
Tubes, Inc., and had >99% purity (verified using SEM), result of NIR illumination caused a change in weight readings
>700 m2 g1 surface area, <3 nm thickness, <3 layers, and on the balance. Once the light was turned off, both the actuator
0.52 m diameter. The CNTs purchased from Cheap Tubes, and the balance returned to their original length/reading,
Inc., consisted of 60% single-wall CNTs, 35% MWCNTs, respectively. Actuation was quite repeatable from cycle to
and <1.5% ash. The GO purchased from Cheap Tubes, Inc., cycle with nearly the same displacement amplitude. Stress
was >99% purity. The CB was purchased from Polysciences, test standardization was accomplished by finding the zero
Inc. All carbon types were used in original form and not strain value for each sample and zeroing the balance. Stress
surface modified at any time. Fisherbrand 75 mm 50 mm tests on each sample were conducted with pre-strain values
glass slides were used extensively for PDMS composite ranging from 3% to 40%. The timing sequence for each

8
Nanotechnology 23 (2012) 045501 J Loomis et al

pre-strain value was 1 min relaxation wait followed by five experiments. BK conducted the Raman spectroscopy. TB, PX,
cycles of NIR illumination on for 60 s and then off for and NB provided technical support in the imaging. ET and BP
30 s. Engineering stress calculations (referred to as stress worked through e-mail discussions in preparation and editing
throughout the paper) were made by dividing the change in of the paper.
force between illumination on and off by the cross-sectional
area of the test samples. References

4.4. Youngs modulus calculation [1] Powers J M and Sullivan T D 1976 Flexural disk peizoelectric
polymer hydrophones J. Acoust. Soc. Am. 60 S477
[2] Nakamura K and Shimizu H 1989 Hysteresis-free
Youngs modulus values as a function of increasing wt% piezoelectric actuators using linbo3 plates with a
GNP/PDMS were determined by first measuring the stress ferroelectric inversion layer Ferroelectrics 93 2116
induced by pre-straining each test sample from 0 to 5% [3] Kuribayashi K 1989 Millimeter-sized joint actuator using a
in 1% intervals (NIR illumination off ). For a given wt% shape memory alloy Sensors Actuators 20 5764
[4] Uchino K 1986 Electrostrictive actuatorsmaterials and
GNP/PDMS, an average of the resulting change in stress applications Am. Ceram. Soc. Bull. 65 64752
divided by the pre-strain (for pre-strains from 0 to 5%) [5] Huang C, Jakli A and Zhang Q M 2003 Anisotropic liquid
yielded the value for the elastic modulus. The GNP/PDMS crystal gels: self-assembled nanocomposites as solid-state
test samples were observed to have an elastic behavior in this soft actuator materials exhibiting high electromechanical
region. responses Abstracts of Papers of the American Chemical
Society vol 225, p U546
Optical-to-mechanical energy conversion calculation: [6] Smela E and Gadegaard N 1999 Surprising volume change in
PPy(DBS): an atomic force microscopy study Adv. Mater.
Peffective = (Plaser /Aspot )Aincident
11 953
M = 1 total /Peffective . [7] Baughman R H et al 1999 Carbon nanotube actuators Science
284 13404
A laser power meter (Newport 1918-C) was used to [8] Weissmuller J, Viswanath R N, Kramer D, Zimmer P,
calculate the laser power absorbed in the test sample. First, an Wurschum R and Gleiter H 2003 Charge-induced reversible
uninterrupted laser power of 500 mW was measured. Next, strain in a metal Science 300 3125
[9] Ahir S V and Terentjev E M 2005 Photomechanical actuation
a 2 wt% GNP/PDMS test sample was mounted in front of in polymer-nanotube composites Nature Mater. 4 4915
the meter and a value of 160 mW was obtained. Therefore, [10] Lu S X and Panchapakesan B 2005 Optically driven nanotube
the difference (340 mW) is the amount of power absorbed actuators Nanotechnology 16 254854
in the test sample and is referred to as Plaser throughout the [11] Comstock M J, Levy N, Cho J, Berbil-Bautista L,
paper. Aspot is the area of the spot size (3 mm 3 mm, or Crommie M F, Poulsen D A and Frechet J M J 2008
Measuring reversible photomechanical switching rates for a
9 mm2 ), and Aincident is the area for absorption within the film molecule at a surface Appl. Phys. Lett. 92 123107
(3 mm 60 m, or 0.18 mm2 ). The effective laser power [12] Uchino K, Poosanaas P and Tonooka K 2001 Photostrictive
(Peffective ), therefore, is 6.8 mW. The optical-to-mechanical actuatorsnew perspective Ferroelectrics 264 19616
energy conversion factor (M ) is calculated by dividing the [13] Yamada M, Kondo M, Mamiya J I, Yu Y L, Kinoshita M,
total change in engineering stress (1total ) from maximum Barrett C J and Ikeda T 2008 Photomobile polymer
materials: towards light-driven plastic motors Angew.
expansion to maximum contraction (+14 to 36 kPa, or Chem. Int. Edn 47 49868
50 kPa) by the Peffective , resulting in a value of 79 MPa W1 . [14] Lu S and Panchapakesan B 2007 All-optical micromirrors
from nanotube MOMS with wavelength selectivity
J. Microelectromech. Syst. 16 151523
4.5. Thermal response testing [15] Ahir S V and Terentjev E M 2006 Fast relaxation of carbon
nanotubes in polymer composite actuators Phys. Rev. Lett.
Steady-state temperature tests were conducted by placing a 96 133902
thermocouple (Omega, K-type) in contact with the PDMS [16] Lu S X and Panchapakesan B 2007 Photomechanical
composite. The temperature measurements were conducted responses of carbon nanotube/polymer actuators
Nanotechnology 18 305502
in 5 mm intervals from the illumination point to the edge of [17] Lu S X, Ahir S V, Terentjev E M and Panchapakesan B 2007
the testing area (25 mm for 0% pre-strain, 35 mm for 40% Alignment dependent mechanical responses of carbon
pre-strain). Each position was held for 5 min to allow the nanotubes to light Appl. Phys. Lett. 91 103106
temperature to reach its steady-state value (typically reached [18] Ahir S V, Terentjev E M, Lu S X and Panchapakesan B 2007
within 2 min). The plotted temperature data points were Thermal fluctuations, stress relaxation, and actuation in
carbon nanotube networks Phys. Rev. B 76 165437
calculated by averaging the last 60 s of each test period. [19] Lu S X and Panchapakesan B 2006 Appl. Phys. Lett.
88 253107
Acknowledgments [20] Lu S X, Liu Y, Shao N and Panchapakesan B 2007 Nanotube
micro-opto-mechanical systems Nanotechnology 18 065501
[21] Novoselov K S, Geim A K, Morozov S V, Jiang D, Zhang Y,
Funding for this research was partially supported by the NSF Dubonos S V, Grigorieva I V and Firsov A A 2004 Electric
CAREER award ECCS: 0853066 for one of the authors (BP). field effect in atomically thin carbon films Science
NB acknowledges REV support from NSF. 306 6669
[22] Balandin A A, Ghosh S, Bao W Z, Calizo I, Teweldebrhan D,
Author contributions: JL and BP designed the experiments Miao F and Lau C N 2008 Superior thermal conductivity of
and prepared the manuscript. JL performed the fabrication and single-layer graphene Nano Lett. 8 9027

9
Nanotechnology 23 (2012) 045501 J Loomis et al

[23] Lu Q, Arroyo M and Huang R 2009 Elastic bending modulus [34] Rogers G W and Liu J Z 2011 Graphene actuators:
of monolayer graphene J. Phys. D: Appl. Phys. 42 102002 quantum-mechanical and electrostatic double-layer effects
[24] Gusynin V P and Sharapov S G 2005 Unconventional integer J. Am. Chem. Soc. 133 1085863
quantum Hall effect in graphene Phys. Rev. Lett. 95 146801 [35] Armani D, Liu C and Aluru N 1999 Re-configurable fluid
[25] Dreyer D R, Park S, Bielawski C W and Ruoff R S 2010 The circuits by PDMS elastomer micromachining MEMS99:
chemistry of graphene oxide Chem. Soc. Rev. 39 22840 12th IEEE Int. Conf. on Micro Electro Mechanical Systems
[26] Gomez-Navarro C, Burghard M and Kern K 2008 Elastic pp 2227
properties of chemically derived single graphene sheets [36] Sarkisov S S, Curley M J, Fields A, Sarkisov S S and
Nano Lett. 8 20459 Adamovsky G 2004 Photomechanical effect in films of
[27] Bielawski C W, Potts J R, Dreyer D R and Ruoff R S 2011 polyvinylidene fluoride Appl. Phys. Lett. 85 27479
Graphene-based polymer nanocomposites Polymer 52 525 [37] Seki T and Tamaki T 1993 Photomechanical effect in
[28] Alexandre M and Dubois P 2000 Polymer-layered silicate monolayers of azobenzene side-chain polymers Chem. Lett.
nanocomposites: preparation, properties and uses of a new 22 173942
class of materials Mater. Sci. Eng. R 28 163
[38] Tabiryan N, Serak S, Dai X M and Bunning T 2005 Polymer
[29] Thostenson E T, Li C Y and Chou T W 2005 Nanocomposites
film with optically controlled form and actuation Opt.
in context Compos. Sci. Technol. 65 491516
Express 13 74428
[30] Sun Z P, Hasan T, Torrisi F, Popa D, Privitera G, Wang F Q,
[39] Ikeda T, Mamiya J and Yu Y L 2007 Photomechanics of
Bonaccorso F, Basko D M and Ferrari A C 2010 Graphene
mode-locked ultrafast laser ACS Nano 4 80310 liquid-crystalline elastomers and other polymers Angew.
[31] Casiraghi C, Hartschuh A, Lidorikis E, Qian H, Chem. Int. Edn 46 50628
Harutyunyan H, Gokus T, Novoselov K S and [40] Lendlein A, Jiang H Y, Junger O and Langer R 2005
Ferrari A C 2007 Rayleigh imaging of graphene and Light-induced shape-memory polymers Nature 434 87982
graphene layers Nano Lett. 7 27117 [41] Thomsen D L, Keller P, Naciri J, Pink R, Jeon H,
[32] Zhang Y B, Tang T T, Girit C, Hao Z, Martin M C, Zettl A, Shenoy D and Ratna B R 2001 Liquid crystal elastomers
Crommie M F, Shen Y R and Wang F 2009 Direct with mechanical properties of a muscle Macromolecules
observation of a widely tunable bandgap in bilayer 34 586875
graphene Nature 459 8203 [42] Lehmann W, Skupin H, Tolksdorf C, Gebhard E, Zentel R,
[33] Cai D K, Neyer A, Kuckuk R and Heise H M 2008 Optical Kruger P, Losche M and Kremer F 2001 Giant lateral
absorption in transparent PDMS materials applied for electrostriction in ferroelectric liquid-crystalline elastomers
multimode waveguides fabrication Opt. Mater. 30 115761 Nature 410 44750

10