Cite this: RSC Advances, 2013, 3, 3194
Received 17th October 2012,
Accepted 4th January 2013
A spring-type piezoelectric energy harvester3
Dongjin Kim,abSeungbum Hong,*abDongjun Li,aHee Seok Roh,cGun Ahn,a
Jiyoon Kim,aMoonkyu Park,aJongin Hong,dTae-hyun Sungeand Kwangsoo No*a
We developed a three-dimensional spring-type piezoelectric
energy harvester using a dip-coating method and multi-direc-
tional electrode deposition. The energy harvester consists of a bi-
layered structure composed of a surface electrode and a ferro-
electric polymer, on a conventional spring which has two roles –
the core electrode and the mechanical substrate for the ferro-
electric polymer. The energy harvester generated an output
voltage of up to 88 mV as a function of cycling compression
stress, which leads to a piezoelectric constant of 28.55 pC N21for
unpoled P(VDF-TrFE) films. Since the spring structure significantly
decreases the resonance frequency of the harvester, the spring-
type energy harvester can effectively generate electricity using
low-frequency vibration energy abundant in the nature.
Vibration-based energy harvesting (VEH) devices have attracted
great interest for use as sustainable and clean electric power
supplies for wireless sensor networks that enable health monitor-
ing of important infrastructures such as power plants, bridges and
remote power grids.1–3Since there are abundant vibration sources
with a low frequency (between 1 and 200 Hz) in nature,4low
frequency vibrations are of high interest and are targeted in VEH
device design for a wide range of potential applications.5–10Several
approaches exist to convert vibrations to electrical power including
electromagnetic, electrostatic and piezoelectric conversion, among
which piezoelectric energy harvesting systems (PEHSs) have
received the most attention. This is because they directly convert
applied mechanical energy into electricity, leading to a simpler
device design in comparison to other mechanisms, which require
complex geometries and numerous additional components.1
However, PEHSs are facing challenges such as low output
power and high resonance frequency.8As the resonant frequency
is usually higher than the vibration frequency with the highest
amplitude in the environment when PEHSs are scaled down to
micron size, they suffer from low output power because energy
harvesters generate the maximum power at the resonance
frequency.11The most commonly adopted ways to reduce the
resonance frequency of the harvester are 1) to add a mass to the
harvester12or 2) to use a spring structure or equivalent that can
decrease the overall system stiffness. We came up with the idea of
a spring-type structure, which can significantly decrease the
resonance frequency toward 1 kHz or less as compared with a
beam-type structure with the same weight. Furthermore, our idea
can be applied to existing spring structures in automobiles,
bridges or even in mattress, which enables us to convert otherwise
wasted volume and energy into useful ones. However, in
fabricating a spring-type piezoelectric energy harvester, there are
processing challenges such as conformal coating of piezoelectric
material onto a substrate with a complex geometry, and uniform
electrode deposition to ensure maximum contact with the
deposited piezoelectric materials.
Here, we report the way we addressed the processing
challenges of spring-type piezoelectric energy harvesters, namely
a combination of dip-coating method and multi-directional
electrode deposition, and measured the output voltage as a
function of cycling compression stress without an external poling
The method and preparation for fabricating the spring-type
energy harvesters are described as follows. Firstly, poly(vinylidene
fluoride trifluoroethylene) (P(VDF-TrFE)) solution, a spring and a
low speed motor were prepared. The spring was commercially
available (Seoul Spring, Inc.), with a wire diameter of 0.97 mm, an
outer spring diameter of 11.55 mm and a length of 40 mm. The
spring had 10 turns. The spring constant was measured by placing
a mass of 500 g at the end of the spring. The precursor solution
was prepared by dissolving 15 wt% P(VDF-TrFE) (VDF:CH2–CF2/
TrFE:CHF–CF2, 75/25) in methyl ethyl ketone solvent (MEK).13The
iron spring was immersed in the solution for 30 s. The spring was
then withdrawn from the solution at a speed of 0.3 mm s21. The
coated P(VDF-TrFE) was left to dry for 1 h in vacuo (less than 1
kPa). To obtain the optimum thickness, we repeated the above
aDepartment of Materials Science and Engineering, KAIST, Daejeon 305-701, Korea.
E-mail: firstname.lastname@example.org; email@example.com
bMaterial Science Division, Argonne National Laboratory, Lemont, IL 60439, USA.
cNuclear Engineering Division, Argonne National Laboratory, Lemont, IL 60439, USA
dDepartment of Chemistry, Chung-Ang University, Seoul 156-756, Korea
eDepartment of Electrical Enginnering, Hanyang University, Seoul, 133-791, Korea
3Electronic supplementary information (ESI) available. See DOI: 10.1039/c2ra22554a
3194 | RSC Adv., 2013, 3, 3194–3198This journal is ? The Royal Society of Chemistry 2013
processes twice more. Then, the coated spring was annealed at 130
uC for 1 h in a chamber at ambient pressure to increase the
crystallinity of the P(VDF-TrFE) films. Afterwards, a Pt electrode
was coated over a 22 mm length (approximately 5.5 turns) of the
P(VDF-TrFE) films covered spring structure by physical vapour
deposition (PVD). We masked a part of the spring (18 mm) using
Teflon and scotch tape to access electrically the core part of the
spring. For the conformal coating on the complex structure of
spring, we coated with Pt four times, rotating the spring about the
long axis with an interval of 90u.
In order to simplify the optimization of the dip-coating process
on the spring structure, we used a model system that has a 99.9%
Fe substrate with a flat geometry on which we deposited P(VDF-
TrFE) by a dip-coating method. We measured the deposition rate
by monitoring the thickness after each coating cycle which was
verified by scanning electron microscopy (SEM, Hitachi S-4800,
operating voltage: 10 kV). Then, we performed X-ray diffraction
(Rigaku D/Max-2500 at a scanning velocity of 1u/min and using the
h/2h scan at 40 kV and 300 mA) and Fourier transform infrared
spectroscopy (FTIR, Bruker Optiks IFS66V S21and Hyperion 3000)
analyses to identify the crystalline phases and the structure of the
polymer chains. Lastly, we conducted P–E hysteresis loop
Ferroelectric Test System with tungsten probe tip (cat-whisker,
T20-7A)) to measure the remnant polarization and the coercive
field, which allowed us to predict the piezoelectric properties of
the P(VDF-TrFE) films deposited on the Fe substrate.
The energy harvester developed here consists of a conventional
spring with a bi-layered structure, composed of a surface electrode
and a ferroelectric polymer, of which a schematic and actual
image along with its circuit are shown in Fig. 1. The spring has
two roles, one as the core electrode and other as the mechanical
substrate for the ferroelectric polymer. The ferroelectric polymer is
then covered by the Pt surface electrode. When the spring is
vertically pressed, the amplitude of the shear stress on the surface
is considerably larger than the vertical stress.14Therefore, it is
expected that the spring-type energy harvester will efficiently
convert the mechanical energy into electricity via the piezoelectric
effect and the amplified conversion of applied vertical stress into
internal shear stress.
Fig. 2a shows the tilted-view SEM image of P(VDF-TrFE) coated
on the Fe plate using the dip-coating method from which we
measured the average thickness of 6.43 mm with a standard
deviation of 0.508 mm. The P(VDF-TrFE) films went through three
iterative dip-coating processes. It should be noted that we were not
able to measure well-defined P–E hysteresis loops from the single
and double coated films. We think it is due to the poor interface
between the electrode and P(VDF-TrFE) based on the shape of P–E
hysteresis loops.15However, as shown in Fig. 2d, we could
measure reliable P–E hysteresis loops from the triple coated film.
In the XRD pattern (see Fig. 2b), the P(VDF-TrFE) film showed a
peak in intensity at 2h = 19.95u, which stems from the reflections
of (110) and (200) crystal planes, and represents the ferroelectric b
phase.16,17Bulk P(VDF-TrFE) has peaks in intensity at 2h = 19.5u
for the (200) reflection and at 2h = 19.8u for the (110) reflection,
and these two Bragg peaks overlap due to the fact that their
natural width (D2h = 0.5u) is larger than the distance (0.3u) between
the (110) and (200) peaks that are associated with the ferroelectric
b phase. No peak was observed near 2h = 18u which is related to
the paraelectric phase.18
Fig. 2c shows a FTIR spectrum of PVDF-TrFE film deposited on
the Fe plate. Absorption bands at 844, 880, 1076, 1119, 1170, 1287,
1400 and 1430 cm21were observed in the IR spectra, and are
related to the crystalline-phase spectra. However, the bands at 844,
880, 1076, 1119, 1170, 1400 and 1430 cm21for the crystalline
phase overlap with the broad disordered-phase bands in this
region.19The 1287 cm21peak does not overlap with any
disordered-phase band and is assigned to the sequences of four
or more trans isomers.19Although 844 and 880 cm21peaks for the
crystal phase overlap with the disordered-phase band, the crystal-
phase bands contribute to the peaks more than the disordered-
Fig. 1 (a) Schematic and (b) actual image of the spring-type energy harvester. The
energy harvester consists of a conventional spring with a bi-layered structure
composed of a surface electrode and a ferroelectric polymer. The spring has two
roles, one as the core electrode and the other as the mechanical substrate for the
This journal is ? The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 3194–3198 | 3195
phase band, and the peaks are assigned to the sequences of three
or more trans units and trans sequences, respectively.19,20Both
XRD and FTIR results confirmed that the P(VDF-TrFE) film
contains a ferroelectric b phase.
Subsequently, P–E hysteresis loops were measured to check the
ferroelectric properties of the P(VDF-TrFE) films deposited on the
Fe substrate (model system), which are related to the piezoelectric
properties via permittivity and an electromechanical coupling
coefficient.21Remnant polarization (Pr) of 16.4 mC cm22and a
coercive field value of 19.4 MV m21were measured (Prvalues
reported in the literature22–24are between 5 and 15 mC cm22). The
results show that the P(VDF-TrFE) films on iron as a model system
using the dip-coating method contain a ferroelectric b phase and
have well-defined ferroelectric properties.
Based on the process optimization results from the model
system, we applied the same processing recipe to the spring
structure, which has a more complicated geometry than the Fe
plate. Fig. 3a shows the schematic of the resulting spring-type
energy harvester under an applied vertical force of up to 17.4 N,
which we measured from the displacement (24 mm) and the
spring constant (725.9 N m21). We were able to collect the output
voltage signals from the energy harvester using a digital
oscilloscope. As shown in Fig. 3b, the spring was cyclically
compressed within the range 16–33 mm. The average displace-
ment of the spring length was 15.92 mm under applied vertical
force while the period of displacement was 0.507 s. The output
signal showed a more complicated shape than the displacement
curve did, which comes from the fact that it consists of signal
induced by piezoelectricity and noise from other sources (see
Fig. 3c). As such, we applied a Fourier transform method to extract
the signal induced by piezoelectricity of which details can be
found in the supplementary information. We obtained an effective
piezoelectric constant of 28.55 pC N21, which is large if we
consider the fact that the spring has not been poled by an
externally applied voltage (poled P(VDF-TrFE) has piezoelectric
coefficients of 231.6 (d33), 10.2 (d31) and 236.0 (d15) pC N21,
respectively).25,26Although the self-poling effect was observed in
the P(VDF-TrFE) thin films deposited on transparent polymer
substrates,27this alone cannot account for such a high piezo-
electric coefficient without external poling. Another possible
reason is that the P(VDF-TrFE) film is exposed to a greater in-
plane stress than the vertical stress through thickness because the
P(VDF-TrFE) film coated on the spring is stretched out when the
spring is deformed.14
In order to calculate the stress distributions on P(VDF-TrFE)
film on the spring, we conducted Finite Element (FE) analysis
using ABAQUS 6.11. Fig. 4 shows stress distributions on P(VDF-
TrFE) film coated on the spring. FE results show that when the
spring is loaded in compression, stress at the outer surface of the
spring is in compression in the 2-direction but in tension in the
1-direction, where the 1- and 2- directions indicate circumferential
and hoop directions respectively, as shown in Fig. 4. With thin
shell theory, it is accepted that electric displacement, q (Coulomb/
m2) for coated P(VDF-TrFE) films is related only to the normal
stresses (s11 and s22) which are the components of in-plane
stresses in shell elements, and it is independent of shear stress
Fig. 2 (a) SEM image of P(VDF-TrFE) coated on the Fe plate using the dip-coating
method. (b) XRD pattern, (c) FTIR spectra, and (d) P–E hysteresis loop of P(VDF-TrFE)
3196 | RSC Adv., 2013, 3, 3194–3198 This journal is ? The Royal Society of Chemistry 2013
s12. Therefore, d31a is main factor on the piezoelectric behaviour
of the spring-type energy harvester. For the case of using a poled
piezoelectric material whose d31is 49 pC N21,26it is expected that
the effective piezoelectric constant would be 383 pC N21for
springs with 10 turns when the external electric field is zero. It
means that the spring structure coated with a P(VDF-TrFE) film
with 10 turns can amplify the effective load by about 7.8 times to
generate piezoelectric charges, leading to a 7.8 times larger
effective piezoelectric coefficient. However, as we only coated 5.5
turns in our experiment, the amplification factor is expected to
reduce to 4.3 times.
Although the effective piezoelectric constant of the energy
harvester was lower than the simulated result for poled P(VDF-
can be developed further due to the following reasons. First, we
can enhance the output power of the spring-type piezoelectric
energy harvesters by tuning the resonance frequency of the spring
to the frequency of larger vibrations in the environment by
increasing the outer spring diameter or number of coils in the
spring. Recent studies reported that energy harvesters generate
maximum power at their resonance frequency,12,28which could be
orders of magnitude greater than the power extracted at other
frequencies. Secondly, from the system’s perspective, it was
recently found that the maximum output power from piezoelectric
energy harvesters is insensitive to the piezoelectric coefficient if
one can tune the load resistance to an optimal value.29Based on
the arguments above, we believe that we can use the spring-type
piezoelectric energy harvester without a poling process, which
could simplify their applications in energy harvesting devices.
In conclusion, we developed a three-dimensional spring-type
piezoelectric energy harvester, which can effectively generate
electricity using low-frequency mechanical energy abundant in
nature. We found that P(VDF-TrFE) films coated on an iron-based
spring (k = 725.9 N m21) using a dip coating method followed by
multi-directional Pt electrode sputter deposition generated a
voltage up to 88 mV under a compressive force of 17.4 N, which
Fig. 3 (a) Schematic of the spring-type energy harvester under vertical stress. Plots
of (b) vertical displacement of the energy harvester under cyclical compression
stress, and (c) output voltage signal generated from the energy harvesters as a
function of elapsed time.
Fig. 4 Stress distributions on P(VDF-TrFE) shell elements (a) s11and (b) s22.
This journal is ? The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 3194–3198 | 3197
RSC Advances Communication
leads to a measured piezoelectric constant of 28.55 pC N21for Download full-text
unpoled P(VDF-TrFE) films. We envision that our simple and
effective recipe toward piezoelectric energy harvesters coated on a
typical spring structure will be widely used not only in energy
harvesting field but also in smart sensors and actuators.
This research was supported by the Mid-career Researcher
Program (No. 2010-0015063) and Conversion Research Center
Program (No. 2011K000674) through the National Research
Foundation of Korea (NRF) funded by the Ministry of
Education, Science and Technology (MEST) and the New &
Renewable Energy of the Korea Institute of Energy Technology
Evaluation and Planning (KETEP) grant (No. 20103020060010)
funded by the Ministry of Knowledge Economy, Korea. Work at
Argonne National Laboratory (S. H. and D. K., data analysis
and writing of manuscript) was supported by UChicago
Argonne, a U.S. DOE Office of Science Laboratory, operated
under Contract No. DE-AC02-06CH11357. J. H. acknowledges
the Chung-Ang University Research Grants in 2011.
Notes and references
1 J. A. Paradiso and T. Starner, IEEE Pervasive Comput., 2005, 4,
2 D. Culler, D. Estrin and M. Srivastava, Computer, 2004, 37,
3 M. Lee, J. Bae, J. Lee, S. Hong and Z. L. Wang, Energy Environ.
Sci., 2011, 4, 3359.
4 S. Roundy, P. K. Wright and J. Rabaey, Computer
Communications, 2003, 26, 1131–1144.
5 L. C. Rome, L. Flynn, E. M. Goldman and T. D. Yoo, Science,
2005, 309, 1725.
6 J. M. Donelan, Q. Li, V. Naing, J. A. Hoffer, D. J. Weber and A.
D. Kuo, Science, 2008, 319, 807.
7 S. Priya, J. Electroceram., 2007, 19, 165–182.
8 S. P. Beeby, M. J. Tudor and N. M. White, Meas. Sci. Technol.,
2006, 17, R175–R195.
9 C. Sun, J. Shi, D. J. Bayerl and X. Wnag, Energy Environ. Sci.,
2011, 4, 4508.
10 Y. Qi and M. C. McAlpine, Energy Environ. Sci., 2010, 3,
11 M. Kim, M. Hoegen, J. Dugundji and B. L. Wardle, Smart Mater.
Struct., 2010, 19, 045023.
12 M. Kim, S. Hong, D. J. Miller, J. Dugundji and B. L. Wardle,
Appl. Phys. Lett., 2011, 99, 243506.
13 Y.-Y. Choi, J. Hong, S. Hong, H. Song, D.-S. Cheong and K. No,
Phys. Status Solidi RRL, 2010, 4, 94–96.
14 A. Fox, Stress Relaxation Testing, ASTM International, Baltimore,
1979, pp. 67.
15 I. K. Yoo, in Nanoscale Phenomena in Ferroelectric Thin Films,
Kluwer Academic Publishers, ed. S. Hong, Boston, 2004, ch. 1.
16 J. F. Legrand, Ferroelectrics, 1989, 91, 303–317.
17 J. Choi, C. N. Borca, P. A. Dowben, A. Bune, M. Poulsen,
S. Pebley, S. Adenwalla and S. Ducharme, Phys. Rev. B: Condens.
Matter, 2000, 61, 5760–5770.
18 A. J. Lovinger, G. T. Davis, T. Furukawa and M. G. Broadhurst,
Macromolecules, 1982, 15, 323–328.
19 K. J. Kim, N. M. Reynolds and S. L. Hsu, Macromolecules, 1989,
20 H. Xu, G. Shanthi and Q. M. Zhang, Macromolecules, 2000, 33,
21 D. Damjanovic, Rep. Prog. Phys., 1998, 61, 1267.
22 S. Ikeda, M. Jimbo, S. Kobayashi and Y. Wada, J. Polym. Sci.,
1985, 23, 1781–1791.
23 F. Xia, H. Xu, F. Fang, B. Razavi, Z.-Y. Cheng, Y. Lu, B. Xu and
Q. M. Zhang, Appl. Phys. Lett., 2001, 78, 1122–1124.
24 S. Fujisaki and H. Ishiwara, Appl. Phys. Lett., 2007, 90, 162902.
25 H. Wang, Q. M. Zhang, L. E. Cross and A. O. Sykes, J. Appl.
Phys., 1993, 74, 3394–3389.
26 T. Yagi, Y. Higashihata, K. Fukuyama and J. Sako, Ferroelectrics,
1984, 57, 327.
27 M. Park, Y.-Y. Choi, J. Kim, J. Hong, H. W. Song, T.-H. Sung and
K. No, Soft Matter, 2012, 8, 1064–1069.
28 F. Lu, H. P. Lee and S. P. Lim, Smart Mater. Struct., 2004, 13,
29 M. Kim, Ph. D. Thesis dissertation, 2011, MIT, ch. 6.
3198 | RSC Adv., 2013, 3, 3194–3198This journal is ? The Royal Society of Chemistry 2013