ARTICLE

Graphene-Based Conformal Devices

Yong Ju Park,†,^ Seoung-Ki Lee,†,^ Min-Seok Kim,‡ Hyunmin Kim,§ and Jong-Hyun Ahn†,*
†
School of Electrical and Electronic Engineering, Yonsei University, Seoul 120-749, Korea, ‡Center for Mass Related Quantities, Korea Research Institute of Standards
and Science, Daejeon 305-340, Korea, and §Nano & Bio Research Division, Daegu Gyeongbuk Institute of Science and Technology, Daegu 711-873, Korea. ^Yong Ju Park
and Seoung-Ki Lee contributed equally to this work.

ABSTRACT Despite recent progress in bendable and stretchable

thin-film transistors using novel designs and materials, the develop-
ment of conformal devices remains limited by the insufficient flexibility
of devices. Here, we demonstrate the fabrication of graphene-based
conformal and stretchable devices such as transistor and tactile sensor
on a substrate with a convoluted surface by scaling down the device
thickness. The 70 nm thick graphene-based conformal devices dis-
played a much lower bending stiffness than reported previously. The
demonstrated devices provided excellent conformal coverage over an
uneven animal hide surface without the need for an adhesive. In addition, the ultrathin graphene devices formed on the three-dimensionally curved animal hide
exhibited stable electrical characteristics, even under repetitive bending and twisting. The advanced performance and flexibility demonstrated here show
promise for the development and adoption of wearable electronics in a wide range of future applications.

KEYWORDS: ultrathin transistor . tactile sensor . graphene . conformal device . wearable electronics

R
ecently, conformal devices on uneven also been demonstrated as a solution that can
surface have garnered great popular- absorb any strain by bending/flexing of the
ity due to their multifaceted applica- devices.8
10 Although these approaches are
tions in healthcare monitoring systems1
3 potentially useful in three-dimensional elec-
and wearable electronics.4 Such devices tronics systems, its application is restricted to
are hardly achievable using a sophisticated the objects with simple curvatures. Another
planar layout; rather, the fabrication of them approach for conformal devices involves in-
strictly requires a new design of device geom- terconnecting a series of rigid active devices
etry and a relevant adopt of materials that with “shock-absorbing” bridges that can ac-
can compensate structural imperfections. commodate both compressive and tensile
Thus, most of the current research has been stresses.11
15 This method has been exempli-
focusing on how to maximize the structural fied in a course of applications, including elec-
compliance of “implanted” devices with re- tronic eye cameras,16 stretchable transis-
gards to the environment; strain is regarded tors,17 and flexible solar cells.18 Unfortunately,
as one of the major critical challenges to various additional processes must be accom-
overcome, the solution of which is not fully panied to achieve intended results in this
provided yet, especially as for the mechano- case. Prestraining or premolding substrate
compatibility of the devices with respect to can limit the range of available applications.
in vivo circumstances. Indeed, the delicate tolerance to the colli-
Strain, which can seriously degrade de- sion and friction stemming from out-of-plane
vice performance, is locally generated and buckling of the interconnector becomes a
distributed across the device in a physical significant hurdle for this method. Recently,
contact with a rough surface. In an effort to epidermal electronics based on the combina- * Address correspondence to
ahnj@yonsei.ac.kr.
address this issue, various approaches have tion of ultrathin silicon membrane with fila-
been developed. The representative methods mentary serpentine metal mesh have been Received for review June 25, 2014
have been used by employing elastic conduc- extensively exploited; some applications of and accepted July 29, 2014.
tors composed of elastic rubbers and conduc- these systems are medical sensors to monitor
Published online July 29, 2014
tive materials such as micrometer-sized silver body temperature3 and neural/electrophysio- 10.1021/nn503446f
flakes, metal wires and carbon nanotubes.5
7 logical signals.2,4,19 Although there are con-
Net-shaped and fractured metallic film has siderable advantages with established silicon C 2014 American Chemical Society

PARK ET AL. VOL. 8 ’ NO. 8 ’ 7655–7662 ’ 2014 7655

www.acsnano.org
 ARTICLE
technology in the device design, the silicon-
membrane-based devices still have a critical drawback
in bending inflexibility even when assembled with the
soft substrate because the active layer is several hun-
dred nanometers thick. A smaller bending “stiffness”
would be desired to achieve an enough conformal
coverage over the uneven “wrinkle-prone” substrate,
such as the human skin. Also, the silicon membrane
requires an encapsulation layer to obtain conformal
attachment to a substrate for stable operation during
macroscopic movements of the joints in the human
body (i.e., wrist, knee, and elbow), becoming an
additional source for the bending stiffness increase.20
In this light, graphene is an ideal constituent for
conformal devices because it is a one-atomic-thick
platform with extraordinary mechanical flexibility
and optical transmittance.21,22 A few recent studies
have reported the preparation of graphene- or
carbon-based flexible field effect transistors on un-
conventional substrates;23,24 however, it still remains
a significant challenge to integrate the graphene Figure 1. (a) Schematic diagram showing the procedure
devices on rough substrate such as human skin used to fabricate array of ultrathin graphene field effect
or animal hide without degradation of electrical transistors (UT-GFETs). (b) Microscopy image of the as-
fabricated UT-GFETs. The gray regions indicate the gate
characteristics. electrode, and the green regions indicate the channels and
In this work, we report the graphene-based devices source/drain electrodes. (c) Optical image of the UT-GFETs
that can be attached conformably to highly deformed floating on DI water during the transfer process. (d) Optical
image of the transferred UT-GFETs on a glass substrate.
surface such as an animal hide and exhibit stable
electrical and mechanical performances under serious Figure 1(a
d) shows schematic diagrams of the
external deformations, such as flexion and distortion. entire process for fabricating ultrathin graphene field
With direct application of a polymer film used as a gate effect transistors (UT-GFETs) and the optical images of
dielectric and as a supporting layer of device, the device at each step. First, few-layer graphene (FLG) was
thickness of graphene transistor can be reduced less synthesized on a Ni catalyst using chemical vapor
than 70 nm. Consequently, the demonstrated graphene- deposition (CVD). The FLG was then transferred onto
based conformal device produced a bending stiffness a SiO2/Si substrate.27 The gate electrode was defined
(EI ∼1.24 GPa 3 μm4) that was much lower than the using conventional photolithography and oxygen
values in the literature, leading to an excellent adapta- plasma etching processes, as reported previously.22
tional coverage over uneven surfaces without neces- After removing the photoresist, SU-8 epoxy, which
sity of an adhesive layer. Furthermore, we demonstrate functioned as both gate dielectric and supporting layer
a tactile sensor-array composed of monolithically for GFET,28 was spin-coated onto the device, cross-
patterned graphene film, showing good sensitivity linked by UV irradiation, and hard-baked. The thickness
against external force. of the epoxy film was tuned from 70 to 1400 nm by
using different volumetric ratios of the epoxy and
RESULTS AND DISCUSSION epoxy thinner. Because the graphene thickness was
Reducing the device thickness is one strategy for subnanometer in scale, the total thickness of the GFET
realizing conformal devices because the bending stiff- device could be regulated by controlling the thickness
ness is proportional to the cube of the thickness.25 of the SU-8 epoxy layer. The UT-GFETs were completed
Considerable efforts have been applied toward thin- by patterning the source (S)/drain (D) electrodes and
ning the device supporting layer since the supporting semiconducting channel region, after transferring the
layer contributes majority of the device thickness and single layer graphene (SLG) grown on Cu foil onto an
degrades the flexibility of the system.26 To this end, we epoxy coated device.29,30 The detailed fabrication step
designed the fabrication of a stand-alone device pre- was described in Supporting Information. The fabri-
pared without a substrate by substituting a constituent cated UT-GFETs on the SiO2/Si handling substrate had
layer that comprises the transistors for the supporting a channel width of 25 μm and length of 30 μm, as shown
layer. This device structure could be achieved because in Figure 1(b). The UT-GFETs were then floated on
the graphene layer used as both channel and electro- the surface of a dilute HF solution (HF:DI water =1:10)
des was thin, light, and tough, and it did not require an at room temperature to etch away the SiO2 sacrificial
additional supporting layer. layer.31 In this step, the epoxy dielectric film acted as a

PARK ET AL. VOL. 8 ’ NO. 8 ’ 7655–7662 ’ 2014 7656

www.acsnano.org
 ARTICLE
supporting layer to protect the device from destruction
during the transfer process. Thus, we avoided additional
spin-coating steps that would require conventional
transfer steps in an effort to simplify the fabrication
process.32 Complete removal of the sacrificial layer
released the UT-GFETs from the handling substrate,
and the UT-GFETs freely floated on the surface of the
etchant. The UT-GFETs were rinsed with DI water to
remove remaining etchant and were transferred to the
target substrate. The outstanding flexibility of the device
due to the extreme thinness of the device may be
appreciated from the photographic image of the floated
UT-GFET film shown in Figure 1(c). The dry transfer
method is also described in the Supporting Information.
Finally, the devices were transferred to a rough sub-
strate with the help of the capillary forces of water
during drying and adhered to the substrate through van
der Waals forces alone. Figure 1(d) shows a photograph
and magnified image of the UT-GFETs transferred to
transparent substrates. As expected, the optical trans-
mittance exceeded 70% across the visible spectrum
(Supporting Information, Figure S3). The transmittance Figure 2. (a) Calculated stiffness of a UT-GFET prepared
was reduced by ∼3% in the source (S)/drain (D) electro- with a range of device thickness values. Red square indicate
des and channel region, and by ∼26% at 550 nm in the the experimentally measured stiffness values at 70, 220, 430,
and 1400 nm, the thicknesses of the fabricated UT-GFET
gate electrode and epoxy dielectric films. devices. (b) Calculated adhesion energy of the UT-GFET as a
In order to achieve full range of conformal contact function of device thickness. Mechanical modeling revealed
between a GFET film and a rugged surface, an adhesive that the UT-GFET film could achieve conformal contact in
devices <76 nm thick. (c) SEM images of a rugged substrate
energy was investigated using numerical calculation prior to transferring the UT-GFETs. (d
f) SEM images of
of two overlapping cylinder models, and the results the UT-GFETs transferred onto the rugged substrate, for
were compared with experimental measurement as device thickness values of 70, 200, and 430 nm. The inset
image shows the magnified edge part of the GFET film. Scale
shown in Figure 2. Since the adhesive energy depends bar, 100 μm.
strongly on the bending stiffness of a device, a bending
stiffness was calculated based on the thickness of the bending stiffness of the device, R is the radius of the
device as well as the mechanical properties of each cylinders, b is the width of the device, λ is the wave-
layer (i.e., Young's modulus and the Poisson's ratio). length, and γ0 is the arc between the two cylinders
0
Figure 2(a) shows the calculated bending stiffness (Supporting Information, Figure S5). For γ g γc in eq 1,
(EI) variations of fabricated GFETs as a function of a device may make a full conformal contact with the
the thickness. The bending stiffness decreased drama- substrate.1 More detailed information and the calcula-
tically as the device thickness decreased. At a thickness tion process are described in the Supporting Informa-
of 1400 nm, the EI was calculated to be 9192.4 tion. Figure 2(b) shows the calculated and measured
GPa 3 μm4. This value decreases to 266.35, 36.16, and adhesion energy in devices having a thickness of
1.15 GPa 3 μm4 as the GFET becomes thinner to 430, less than 100 nm. According to the calculations, the
220, and 70 nm, respectively. Experimentally measured critical thickness (hc) for perfect conformal contact
EI as marked by red symbols was well matched with the on a rugged substrate was 76 nm. Figure 2(d
f) shows
calculated EI (Supporting Information, Figure S4). The FE-SEM images of devices with different thicknesses
thinnest GFET among the various devices with differ- (tdevice: 70, 220, and 430 nm) transferred onto uneven
ent thickness had a value of 1.24 GPa 3 μm4. It is worth substrates. The surface roughness (Ra) of the rugged
noting that an EI of 1.24 GPa 3 μm4 for the 70 nm thick substrate was 41 μm, measured using a surface pro-
UT-GFETs is the lowest value yet reported for such filer.2 When comparing the surfaces of a bare substrate
devices (Supporting Information, Table S2). Finally, the to the transferred 70 nm thick UT-GFETs, which were
adhesion energy could be calculated according to the thinner than hc, the UT-GFETs were found to make con-
following equation:1 tact deep into the concave features of the surface and
to wrap along the curves of the substrate without
EI
γ g γc0 ¼ f1 þ (1 þ λ)R2 =[(1
λ)γ20 ]g (1) forming cracks or delaminated regions (Figure 2(c,d)).
2R2 b
By contrast, devices thicker than hc made incomplete
where γ (10 mJ/m2) is the adhesion energy in the wet contact with substrates, as shown in Figure 2(e,f).
0
state,1,33 γc is the calculated adhesion energy, EI is the In other words, the GFET film was suspended by the

PARK ET AL. VOL. 8 ’ NO. 8 ’ 7655–7662 ’ 2014 7657

www.acsnano.org
 ARTICLE
surrounding ridges and generated wrinkles as well as
noncontacted region. These differences were scrupu-
lously verified at the edges of the film, as displayed in
the inset of Figure 2(d,f).
The extent of conformal coverage attained by scal-
ing down the device thickness influenced the adhesion
strength (mainly the van der Waals force) between the
GFET film and the substrate.1,34 If the bonding strength
of the GFET film on a substrate is weak, the device can
be detached only by weak external forces. Therefore, it
was necessary to evaluate the bonding (i.e., adhesion)
strength between the device and the substrate.35
37
Figure 3 shows the results of the adhesion strength Figure 3. (a) Schematic illustration of the pull test. The
transferred UT-GFETs contacted the PMDS stamp, and the
tests, as a function of the device thicknesses, measured contact region was monitored using a CCD camera. (b) Calcu-
through a series of pull-off tests using the samples lated contact area between the UT-GFETs and the rugged
shown in Figure 2. A schematic illustration of the pull- substrate (black line), and the measured pull-off force (blue
line) as a function of device thickness.
off test is showed in Figure 3(a). A polydimethylsilox-
ane (PDMS) elastic stamp, which was designed to have the photograph due to their excellent transparency
a contact area of 0.5  0.5 cm2, was mounted on the properties. Rigorous inspection by FE-SEM revealed
end of a cantilever and pressed against the transferred conformal contact between the UT-GFETs and the hide
GFETs with an applied load of 1 N. After a 5 s dwell time, (Figure 4(a)). Interestingly, the UT-GFETs covered the
the PDMS stamp was pulled up from the substrate. bumpy surface well, even the narrow crevices; how-
The pull-off adhesion strength was recorded when the ever, several fine wrinkles were also found due to the
PDMS stamp had been completely released from the presence of extremely small bumps. Compared with a
substrate. The real contact area between GFET film and silicon membrane of the same thickness (tSi: ∼ 70 nm),
PDMS substrate during the whole process was moni- which was studied previously as a potential conformal
tored using a charge-coupled device camera. Figure 3(b) device, the UT-GFETs array film exhibited better cover-
presents plots of the contact area between the GFET film age over uneven surfaces because the devices were
and the substrate, and the pull-off force variations as a composed of an atomic-scale graphene layer and
function of the device thickness. The defined contact a flexible polymer dielectric layer, in contrast with
areas analyzed by the software ImageJ were 4.5, 10.9, the rigid silicon membrane (Supporting Information,
23.9, and 97.1% on 1400, 340, 220, and 70 nm thick Figure S7).
devices, respectively (Supporting Information, Figure S6). Prior to measuring the electrical properties of con-
It should be noted here that the contact area increased formal GFETs, the GFETs were categorized as flat,
dramatically as the devices became thinner. This behav- sloped, and valley depending on the position of the
ior resulted from the relationship between the bending channel on the hide. Namely, the channels that have
stiffness and the device thickness. The bending stiffness, a positive and a negative curvature on the surface were
greatly reduced by the reduced device thickness, dra- classified into sloped and valley, respectively. It is
matically increased the coverage area. Because the desirable to quantify the effects of the external strain
contact strength is proportional to the coverage area, created from the surface curvature on the electrical
the UT-GFETs can adhere strongly to a substrate without characteristics. The representative SEM images of each
the help of an adhesive.35 At a contact load of 1 N, the transistor and the surface profile of the hide was
pull-off force increases by a factor of nearly 30 (from presented in Figure 4(b). Significantly large strain was
2.22 to 68.17 mN) as the thickness is changed from expected to be applied to the transistor at the valley
1400 to 70 nm. positions which corresponded to the regions of great-
The feasibility of using the fabricated device for est deformation from a flat conformation. However,
realistic applications was tested by applying the UT- although the minimum radius of curvature (Rc) in the
GFETs (t: 70 nm) onto animal hide and measuring the valley regions was ∼4 μm, the applied compressive
electrical properties. The inset of Figure 4(a) shows an strain on the device was less than 1% due to extremely
optical image of the UT-GFETs transferred onto hide thin device thickness,38 which falls in the range of
using the dry transfer technique illustrated in Figures 1 endurable strains for the UT-GFETs.39 Moreover, be-
and S2 (Supporting Information). As-fabricated UT- cause other sloped and flat positions spontaneously
GFETs on the SiO2/Si substrate were fully transferred formed conformal contact with the substrate during
onto the target substrate while preserving their origi- the transfer process without introducing any external
nal device array. The T-shaped gray region indicates stresses, the strain at both the sloped and flat positions
a gate electrode made of FLG. The channel and S/D approached zero. To get more reliable verification,
electrodes composed of SLG were difficult to identify in we performed Raman measurement on graphene.

PARK ET AL. VOL. 8 ’ NO. 8 ’ 7655–7662 ’ 2014 7658

www.acsnano.org
 ARTICLE
Figure 4. (a) SEM image of the UT-GFETs mounted on the animal hide substrate. The inset image shows a photograph of the
UT-GFETs. (b) Surface profile of the animal hide. Depending on the position of the channel on the hide, the device was
classified into 3 categories: flat, sloped, and valley. The inset images show representative SEM images of the UT-GFETs on the
flat, sloped, and valley regions. (c) Representative transfer characteristics of the UT-GFETs on flat, sloped, and valley regions.
(d) Distribution of hole mobility of UT-GFETs, which are located on the flat, sloped, and valley regions; inset shows the mobility
histogram of UT-GFETs on hide.

As a result, the 2D Raman band corresponding gra- the flat, sloped, and valley positions, respectively
phene did not reveal any noticeable shifts in each (Supporting Information, Figure S9). The relatively
region, indicating that an undetectable level of strain low mobility of the UT-GFETs may have resulted from
was applied to the graphene (Supporting Information, the poor interfacial properties of the SU-8 epoxy gate
Figure S8).40 The transfer characteristics of the devices insulator and graphene.
were subsequently measured from UT-GFETs posi- Another notable advantage of the conformal UT-
tioned in different regions of the hide. The drain bias GFETs is their good mechanical stretchability across
was fixed at 0.01 V and the gate voltage was swept three-dimensionally curved surfaces. Curved substrates
from
10 to 20 V. Since the device configurations are coated with the UT-GFETs could consume the external
not uniform, statistical analysis of electrical property strain rather than transferring it to the device by trans-
was carried out from 50 devices. Black, red, and blue forming and flattening the surface shape. The stretch-
lines indicate the representative transfer characteris- ability of the device was examined by measuring the
tics of the transistor from flat, sloped, and valley deformations of the UT-GFETs using SEM (Figure 5(a)).
positions, respectively. Interestingly, the transistors As the substrate was stretched up to 4%, the uneven
exhibited similar electrical characteristics on each loca- surface gradually flattened, and only fine vertical wrin-
tion. This indicates that the geometry of hide cannot kles were generated due to the Poisson effects. Con-
greatly influence on the electrical transport property of sequently, the GFETs remained stable under external
UT-GFETs. It is considered that the thickness of the stretching without compromising the physical integrity
device with a few tens of nanometers and wet transfer of the device, as has been observed in wavy structured
method based on surface tension of water can mini- devices.12,13 However, the device located on valley
mize the stress to the GFET when formed on bumpy region was preferentially delaminated from the hide at
substrate.1,38 Undoubtedly, the robust feature of gra- serious stretching because relatively large deformation
phene against external strain plays an indispensable role occurred first in a valley region. This failure can be
to realize conformal device. As shown in Figure 4(c,d), controlled by improving an adhesion between a device
the transistors operated uniformly overall, regardless of and a substrate. The practical feasibility of using con-
the location on hide. The calculated average hole and formal UT-GFETs as wearable electronics was demon-
electron mobilities were 695 ( 41 and 379 ( 13, 692 ( strated by monitoring the electrical characteristic of a
24 and 402 ( 30, and 701 ( 22 and 398 ( 12 cm2/V 3 s on substrate during folding and twisting, which are the

PARK ET AL. VOL. 8 ’ NO. 8 ’ 7655–7662 ’ 2014 7659

www.acsnano.org
 ARTICLE
Figure 5. (a) SEM images of the UT-GFETs released onto the hide (left) or mechanically stretched under a 5% extension (right).
(b) Photographs of the twisted and folded UT-GFETs mounted on the animal hide substrate. (c) Mobility measurements during
repeated deformations of the substrate.

Figure 6. (a) Photographs of a tactile sensor consisting of ultrathin graphene strain gauges on the animal hide (left),
geometric structure of graphene strain gauge (top right), SEM image of graphene strain gauge on animal hide (bottom right).
(b) The pressure map measured by the tactile sensor when a gentle touch of 9 kPa was applied to the sensing area.

basic deformation modes applied by the skin surface displayed stable mechanical durability until 130 re-
as a human body moves (Figure 5(b)). The degrees of peated folding but it showed somewhat vulnerable
bending and twisting in the mechanical tests were behavior at twisting deformation (∼80 cycles). It is sus-
set on the basis of the actual hide deformations that pected that the partial slip or delamination may occur
occur during elbow bending (Supporting Information, between transistor and substrate due to sheer force
Figure S10). Figure 5(c) showed the mobility variations during twisting test, destroying the device. Although
and endurance cycles of a UT-GFET device in a flat the stretchability of the UT-GFETs in conformal contact
configuration and in a bent or twisted configuration with the hide was moderate, it could be increased by
(bending radii: 0.6 cm and twist angle: 25). The mobility transferring the as-fabricated device to a more grooved
variations were statistically analyzed on 25 randomly surface.
selected devices working under identical stresses. As The presented approach to realize mechanically
a result, the mobility of the GFETs did not vary signifi- compliant device is not only useful for transistor appli-
cantly regardless of the bending and twisting deforma- cation but can be applied for various sensing devices.
tion. Unfortunately, we were not able to calculate the To demonstrate the feasibility of using this ultrathin,
exact stress value of device on hide, because the uneven conformable graphene-based device as a tactile sen-
surface can suppress the applied strain by unfolding the sor, we fabricated a 4  4 array with graphene-based
wrinkles; it is carefully estimated that the effective stress resistive strain gauges on an animal hide (Figure 6(a)).
applied to the device is not high. Moreover, the devices The detail fabrication and measurement steps were

PARK ET AL. VOL. 8 ’ NO. 8 ’ 7655–7662 ’ 2014 7660

www.acsnano.org
 ARTICLE
described in the Methods section. All device com- graphene and nanometer-thick polymer dielectric as
ponents including strain gauges and electrical lines the active and supporting layer of the device, respec-
are composed of monolithically patterned graphene tively. The graphene-based conformal devices provided
film. The fabricated graphene-based tactile sensor is excellent conformal coverage over the convoluted sur-
capable of detecting pressure distribution on the sen- faces without using an adhesive, due to the extremely
sing area with a precision better than 9 kPa (Supporting low bending stiffness of the device. Consequently, we
Information, Figure S11) corresponding to the pressure realized conformal and stretchable devices on animal
of human's gentle touch as shown in Figure 6(b). hide, such as UT-GFET and tactile sensor that have
The measured gauge factor of each strain gauge was superior conformability to the previous reported
1.41, which is comparable to the values reported results.4 Moreover, the fabricated devices maintained
previously.41,42 This type of the tactile sensor conform- its original performances well under mechanical bend-
ably attached on an animal hide could be potentially ing and twisting. Although the electrical properties of
applicable to the health care monitoring, nonirritating graphene-based conformal devices was moderate (i.e.,
skin-attachable interactive input/control devices and on/off ratio and gauge factor), it might be possible to be
robotics. improved by combining other two-dimensional semi-
conducting materials. The approach presented here to
CONCLUSION realize the conformal device can offer a key platform for
In this paper, we demonstrated the conformal de- expanding a range of wearable electronic applications
vices on an uneven substrate through scaling down such as fitness tracker, smart medical device and health
the device thickness to 70 nm using atomic-scale monitoring system.

METHODS measuring the weight variation by microelectronic scale. In this

time, value of pressure was calculated by the measured weight
SU-8 Epoxy Preparation. Various thickness of SU-8 layer were
change and the contact area between pressure arm and sensor.
used to demmonstrate wearable electronics. To fabricate 70,
220, 430, and 1400 nm thick SU-8 layer, 7, 20, 30, and 60 wt % of Conflict of Interest: The authors declare no competing
mixture were prepared using SU-8 (SU-8 2005, Micro Chemicals) financial interest.
and cyclopetanone (SU-8 2000 thinner, Micro Chemicals). Then
the mixture was spin-coated on substrate at 2500 rpm for 30 s. Acknowledgment. This work was supported by the Research
After soft bake the sample at at 65 C for 5 min, UV exposure was Program (2012R1A2A1A03006049 and Global Frontier Research
carried out under intensity of 9.3 mJ/cm2 for 2 min. Finally, Center for Advanced Soft Electronics: 2013M3A6A5073170)
robust SU-8 films were completed after hard-baking at 150 C through the National Research Foundation of Korea (NRF),
for 30 min. funded by the Ministry of Science, ICT and Future Planning,
Fabrication and Measurement of Bending Stiffness. To fabricate and the ICT R & D Program (Grant 10041066) funded by the
wavy SU-8 structure, 70, 220, 430, and 1400 nm thick SU-8 layer Institute for Information & Communication Technology Promo-
was prepared on SiO2/Si wafer. Then, we separated the SU-8 film tion, Korea.
from handling wafer by etching the SiO2 sacrificial layer using
diluted HF and rinsed by DI water. At same time, 2 mm thick Supporting Information Available: Transmittance of ultrathin
PDMS slab (Sylgard 184, Dow Corning) was prepared by mixing graphene FET, detail calculation of bending stiffness, Raman
base and curing agent in a 10:1 weight ratio and back at 70 C spectrum of graphene on animal hide, characterization of con-
for 2 h. To geneerate wavy SU-8 structure, the PDMS slabe was formal graphene FET, and comparison table with previous
stretched with mechanical stages before transfer the SU-8 film reported conformal devices. This material is available free of
which was floated on DI-water. When release the strain of PDMS charge via the Internet at http://pubs.acs.org.
after the SU-8/PDMS sample had dried completely, uniform and
regular wavy structure was formed. The wave properties such as
wavelenth and amplitue were measured precisely with atomic REFERENCES AND NOTES
force microscopy and scanning electron microscope. 1. Kim, D.-H.; Viventi, J.; Amsden, J. J.; Xiao, J.; Vigeland, L.;
Conformal Tactile Sensor. The tactile sensor was designed by Kim, Y. S.; Blanco, J. A.; Panilaitis, B.; Frechette, E. S.;
using only two materials, which are graphene and SU-8 epoxy. Contreras, D.; et al. Dissolvable Films of Silk Fibroin for
First, 35 nm thick SU-8 film was formed on 1 μm-thick thermal Ultrathin Conformal Bio-Integrated Electronics. Nat. Mater.
oxide layer (SiO2/Si wafer), and randomly stacked trilayer gra- 2010, 9, 511–517.
phene, which was synthesized using Cu catalyst, was transferred 2. Yeo, W.-H.; Kim, Y. S.; Lee, J.; Ameen, A.; Shi, L.; Li, M.; Wang,
onto SU-8 film. Then the graphene film was monolithically pat- S.; Ma, R.; Jin, S. H.; Kang, Z.; et al. Multifunctional Epidermal
terned to define strain gauges and electrodes. A typical photo- Electronics Printed Directly onto the Skin. Adv. Mater.
lithography process using photoresist (AZ 5214) and reactive ion 2013, 25, 2773–2778.
etching (RIE) with O2 plasma (20 sccm, 100 W, 10 s) was used. 3. Webb, R. C.; Bonifas, A. P.; Behnaz, A.; Zhang, Y.; Yu, K. J.;
Finally, SU-8 epoxy (35 nm) was coated on tactile sensor as a Cheng, H.; Shi, M.; Bian, Z.; Liu, Z.; Kim, Y.-S.; et al. Ultrathin
passivation layer. The transfer method used is same as for UT-GFET. Conformal Devices for Precise and Continuous Thermal
Measurement of Tactile Sensor. The tactile sensor was directly Characterization of Human Skin. Nat. Mater. 2013, 12,
connected to the programmable source-meter unit (Keithley 938–944.
4200-SCS parameter analyzer, Keithley Instruments Inc.) through 4. Jeong, J.-W.; Yeo, W.-H.; Akhtar, A.; Norton, J. J.; Kwack, Y. J.;
flexible printed circuit film (MITFPC Inc.). The sensor was mea- Li, S.; Jung, S. Y.; Su, Y.; Lee, W.; Xia, J.; et al. Materials and
sured by using the passive matrix system which has 1 common Optimized Designs for Human-Machine Interfaces via
and 16 data electrodes. After the connection was completed, Epidermal Electronics. Adv. Mater. 2013, 25, 6839–6846.
the pressure was applied to the sensor using home-built pressure 5. Sekitani, T.; Nakajima, H.; Maeda, H.; Fukushima, T.; Aida, T.;
arm which had hemisphere-shaped contact part (r = 5 mm) while Hata, K.; Someya, T. Stretchable Active-Matrix Organic

PARK ET AL. VOL. 8 ’ NO. 8 ’ 7655–7662 ’ 2014 7661

www.acsnano.org
 ARTICLE
Light-Emitting Diode Display Using Printable Elastic Con- Integrations onto Surfaces of Live Plants and Insects. Nano
ductors. Nat. Mater. 2009, 8, 494–499. Lett. 2014, 14, 2647–2654.
6. Vosgueritchian, M.; Tok, J. B. H.; Bao, Z. N. Stretchable LEDs: 24. Lee, M.-S.; Lee, K.; Kim, S.-Y.; Lee, H.; Park, J.; Choi, K.-H.;
Light-Emitting Electronic Skin. Nat. Photonics 2013, 7, Kim, H.-K.; Kim, D.-G.; Lee, D.-Y.; Nam, S.; et al. High-
769–771. Performance, Transparent, and Stretchable Electrodes
7. Chun, K.-Y.; Oh, Y.; Rho, J.; Ahn, J.-H.; Kim, Y.-J.; Choi, H. R.; Using Graphene-Metal Nanowire Hybrid Structures. Nano
Baik, S. Highly Conductive, Printable and Stretchable Lett. 2013, 13, 2814–2821.
Composite Films of Carbon Nanotubes and Silver. Nat. 25. Jang, H.; Lee, W.; Won, S. M.; Ryu, S. Y.; Lee, D.; Koo, J. B.;
Nanotechnol. 2010, 5, 853–857. Ahn, S.-D.; Yang, C.-W.; Jo, M. H.; Cho, J. H.; et al. Quantum
8. Someya, T.; Kato, Y.; Sekitani, T.; Iba, S.; Noguchi, Y.; Murase, Confinement Effects in Transferrable Silicon Nanomem-
Y.; Kawaguchi, H.; Sakurai, T. Conformable, Flexible, Large- branes and Their Applications on Unusual Substrates.
Area Networks of Pressure and Thermal Sensors with Nano Lett. 2013, 13, 5600–5607.
Organic Transistor Active Matrixes. P. Natl. Acad. Sci. U. S. A. 26. O'Connor, T. F.; Zaretski, A. V.; Shiravi, B. A.; Savagatrup, S.;
2005, 102, 12321–12325. Printz, A. D.; Diaz, M. I.; Lipomi, D. J. Stretching and
9. Takahashi, T.; Takei, K.; Gillies, A. G.; Fearing, R. S.; Javey, A. Conformal Bonding of Organic Solar Cells to Hemisphe-
Carbon Nanotube Active-Matrix Backplanes for Conformal rical Surfaces. Energy Environ. Sci. 2014, 7, 370–378.
Electronics and Sensors. Nano Lett. 2011, 11, 5408–5413. 27. Kim, K. S.; Zhao, Y.; Jang, H.; Lee, S. Y.; Kim, J. M.; Kim, K. S.;
10. Lacour, S. P.; Chan, D.; Wagner, S.; Li, T.; Suo, Z. G. Mechan- Ahn, J.-H.; Kim, P.; Choi, J. Y.; Hong, B. H. Large-Scale
isms of Reversible Stretchability of Thin Metal Films on Pattern Growth of Graphene Films for Stretchable Trans-
Elastomeric Substrates. Appl. Phys. Lett. 2006, 88, 204103. parent Electrodes. Nature 2009, 457, 706–710.
11. Hung, P. J.; Jeong, K. H.; Liu, G. L.; Lee, L. P. Microfabricated 28. Cao, Q.; Xia, M. G.; Shim, M.; Rogers, J. A. Bilayer Organic-
Suspensions for Electrical Connections on the Tunable Inorganic Gate Dielectrics for High-Performance, Low-
Elastomer Membrane. Appl. Phys. Lett. 2004, 85, 6051–6053. Voltage, Single-Walled Carbon Nanotube Thin-Film Tran-
12. Ko, H. C.; Shin, G.; Wang, S. D.; Stoykovich, M. P.; Lee, J. W.; sistors, Complementary Logic Gates, and P-N Diodes on
Kim, D.-H.; Ha, J. S.; Huang, Y. G.; Hwang, K.-C.; Rogers, J. A. Plastic Substrates. Adv. Funct. Mater. 2006, 16, 2355–2362.
Curvilinear Electronics Formed Using Silicon Membrane 29. Li, X. S.; Cai, W. W.; An, J. H.; Kim, S.; Nah, J.; Yang, D. X.; Piner,
Circuits and Elastomeric Transfer Elements. Small 2009, 5, R.; Velamakanni, A.; Jung, I.; Tutuc, E.; et al. Large-Area
2703–2709. Synthesis of High-Quality and Uniform Graphene Films on
13. Kim, D.-H.; Kim, Y. S.; Wu, J.; Liu, Z. J.; Song, J. Z.; Kim, H.-S.; Copper Foils. Science 2009, 324, 1312–1314.
Huang, Y. G. Y.; Hwang, K. C.; Rogers, J. A. Ultrathin Silicon 30. Bae, S.; Kim, H.; Lee, Y.; Xu, X. F.; Park, J.-S.; Zheng, Y.;
Circuits With Strain-Isolation Layers and Mesh Layouts for Balakrishnan, J.; Lei, T.; Kim, H. R.; Song, Y. I.; et al. Roll-
High-Performance Electronics on Fabric, Vinyl, Leather, to-Roll Production of 30-Inch Graphene Films for Transparent
and Paper. Adv. Mater. 2009, 21, 3703–3707. Electrodes. Nat. Nanotechnol. 2010, 5, 574–578.
14. Kim, D.-H.; Choi, W. M.; Ahn, J.-H.; Kim, H.-S.; Song, J. Z.; 31. Thanh, Q. N.; Jeong, H.; Kim, J.; Kevek, J. W.; Ahn, Y. H.; Lee,
Huang, Y. G.; Liu, Z. J.; Lu, C.; Koh, C. G.; Rogers, J. A. S.; Minot, E. D.; Park, J. Y. Transfer-Printing of As-Fabricated
Complementary Metal Oxide Silicon Integrated Circuits Carbon Nanotube Devices onto Various Substrates. Adv.
Incorporating Monolithically Integrated Stretchable Wavy Mater. 2012, 24, 4499–4504.
Interconnects. Appl. Phys. Lett. 2008, 93, 044102. 32. Park, J.-U.; Nam, S.; Lee, M.-S.; Lieber, C. M. Synthesis of
15. Kim, D.-H.; Song, J. Z.; Choi, W. M.; Kim, H.-S.; Kim, R.-H.; Liu, Monolithic Graphene-Graphite Integrated Electronics.
Z. J.; Huang, Y. Y.; Hwang, K.-C.; Zhang, Y. W.; Rogers, J. A. Nat. Mater. 2012, 11, 120–125.
Materials and Noncoplanar Mesh Designs for Integrated 33. Chaudhury, M. K.; Whitesides, G. M. Direct Measurement of
Circuits with Linear Elastic Responses to Extreme Mechan- Interfacial Interactions between Semispherical Lenses and
ical Deformations. P. Natl. Acad. Sci. U. S. A. 2008, 105, Flat Sheets of Poly(dimethylsiloxane) and Their Chemical
18675–18680. Derivatives. Langmuir 1991, 7, 1013–1025.
16. Song, Y. M.; Xie, Y. Z.; Malyarchuk, V.; Xiao, J. L.; Jung, I.; 34. Persson, B. N. J.; Tosatti, E. The Effect of Surface Roughness
Choi, K.-J.; Liu, Z. J.; Park, H.; Lu, C. F.; Kim, R. H.; et al. Digital on the Adhesion of Elastic Solids. J. Chem. Phys. 2001, 115,
Cameras With Designs Inspired by the Arthropod Eye. 5597–5610.
Nature 2013, 497, 95–99. 35. Persson, B. N. J.; Albohr, O.; Creton, C.; Peveri, V. Contact
17. Park, K.; Lee, D.-K.; Kim, B.-S.; Jeon, H.; Lee, N.-E.; Whang, D.; Area between a Viscoelastic Solid and a Hard, Randomly
Lee, H.-J.; Kim, Y. J.; Ahn, J.-H. Stretchable, Transparent Rough, Substrate. J. Chem. Phys. 2004, 120, 8779–8793.
Zinc Oxide Thin Film Transistors. Adv. Funct. Mater. 2010, 36. Palasantzas, G.; De Hosson, J. T. M. Influence of Surface
20, 3577–3582. Roughness on the Adhesion of Elastic Films. Phys. Rev. E:
18. Kaltenbrunner, M.; White, M. S.; Glowacki, E. D.; Sekitani, T.; Stat., Nonlinear, Soft Matter Phys. 2003, 67, 021604.
Someya, T.; Sariciftci, N. S.; Bauer, S. Ultrathin and Light- 37. Varenberg, M.; Peressadko, A.; Gorb, S.; Arzt, E. Effect
weight Organic Solar Cells with High Flexibility. Nat. of Real Contact Geometry on Adhesion. Appl. Phys. Lett.
Commun. 2012, 3, 770. 2006, 89, 121905.
38. Suo, Z.; Ma, E. Y.; Gleskova, H.; Wagner, S. Mechanics of
19. Viventi, J.; Kim, D.-H.; Moss, J. D.; Kim, Y.-S.; Blanco, J. A.;
Rollable and Foldable Film-on-Foil Electronics. Appl. Phys.
Annetta, N.; Hicks, A.; Xiao, J. L.; Huang, Y. G.; Callans, D. J.;
Lett. 1999, 74, 1177–1179.
et al. A Conformal, Bio-Interfaced Class of Silicon Electro-
39. Lee, G. H.; Cooper, R. C.; An, S. J.; Lee, S.; van der Zande, A.;
nics for Mapping Cardiac Electrophysiology. Sci. Transl.
Petrone, N.; Hammerherg, A. G.; Lee, C.; Crawford, B.;
Med. 2010, 2, 22–24.
Oliver, W.; et al. High-Strength Chemical-Vapor Deposited
20. Kim, D.-H.; Lu, N. S.; Ma, R.; Kim, Y. S.; Kim, R.-H.; Wang, S. D.;
Graphene and Grain Boundaries. Science 2013, 340, 1073–
Wu, J.; Won, S. M.; Tao, H.; Islam, A.; et al. Epidermal
1076.
Electronics. Science 2011, 333, 838–843.
40. Ni, G.-X.; Yang, H.-Z.; Ji, W.; Baeck, S.-J.; Toh, C. T.; Ahn, J.-H.;
21. Lee, S.-K.; Kim, B. J.; Jang, H.; Yoon, S. C.; Lee, C.; Hong, B. H.;
Pereira, V. M.; Ozyilmaz, B. Tuning Optical Conductivity of
Rogers, J. A.; Cho, J. H.; Ahn, J.-H. Stretchable Graphene
Large-Scale CVD Graphene by Strain Engineering. Adv.
Transistors with Printed Dielectrics and Gate Electrodes.
Mater. 2014, 26, 1081–1086.
Nano Lett. 2011, 11, 4642–4646.
41. Chang, L. Foundations of MEMS; Pearson Education: Cranbury
22. Lee, S.-K.; Jang, H. Y.; Jang, S.; Choi, E.; Hong, B. H.; Lee, J.; Township, NJ, 2006; pp 207
244.
Park, S.; Ahn, J.-H. All Graphene-Based Thin Film Transis- 42. Huang, M.; Pascal, T. A.; Kim, H.; Goddard, W. A., III; Greer,
tors on Flexible Plastic Substrates. Nano Lett. 2012, 12, J. R. Electronic-Mechanical Coupling in Graphene from
3472–3476. In-Situ Indentation Experiments and Multiscale Simulations.
23. Lee, K.; Park, J.; Lee, M.-S.; Kim, J.; Hyun, B. G.; Kang, D. J.; Nano Lett. 2011, 11, 1241–1246.
Na, K.; Lee, C. Y.; Bien, F.; Park, J.-U. In-Situ Synthesis of
Carbon Nanotube-Graphite Electronic Devices and Their

PARK ET AL. VOL. 8 ’ NO. 8 ’ 7655–7662 ’ 2014 7662

www.acsnano.org