Lead-free Mn-doped (K0.5, Na0.5)NbO3 (KNN) thin films were fabricated by the chemical solution deposition method. The addition of small concentration of Mn dopant effectively reduced the leakage current density and enhanced the piezoelectric properties of the films. The leakage current density of 0.5 mol. % Mn-doped KNN film showed the lowest value of ∼10-7 A/cm2 at 10 V compared to the films with other doping concentrations and the piezoelectric d33 and e31 coefficients of this film were ∼90 pm/V and −8.5 C/m2, respectively. The maximum power and power density of the lead-free thin film-based vibrational energy harvesting device were 3.62 μW and 1800 μW/cm3 at the resonance frequency of 132 Hz and the acceleration of 1.0 G. The results prove that the 0.5 mol. % Mn-doped KNN film is an attractive candidate transducer layer for the piezoelectric MEMS energy harvesting device applications with a small volume and a long-lasting power source.

Piezoelectric materials are of importance due to their high energy-conversion efficiency, in particular, for converting from mechanical energy to electrical energy, and vice versa.1–4 In recent years, there has been a growing interest in microelectromechanical system (MEMS) piezoelectric vibration energy harvesting with a resonant frequency between 50 Hz and 200 Hz, which is the range of common environmental vibration sources, for their potential to small-scale self-powered microelectronic devices, especially for wireless sensor applications.2–4 

The objective of this research is to develop bio-compatible and size-scaled MEMS vibrational energy scavenging devices using lead-free perovskite piezoelectric thin films, with emphasis on self-powering. While many research groups have demonstrated bulk-scale prototypes of piezoelectric energy harvesters5,6 or relatively large-scale bending mode power generators,7,8 only a few groups have fabricated small-scale thin film-type vibrational energy harvesters with a low resonance frequency capable of generating useful power—by far, the vast majority using Pb(Zr,Ti)O3 (PZT)-based thin films in Pb-based material systems9,10 and AlN thin films in non-toxic lead-free material systems.11,12 The toxicity of lead (Pb) in PZT, however, has led to global efforts to identify a replacement system, and this search is particularly critical for environmental-friendly electronic device applications. Comparing lead-free perovskite thin film-based and AlN-based vibrational MEMS energy harvesters, the optimal resistance of AlN thin film-based devices is much larger than that of perovskite piezoelectric thin film-based ones due to a low electromechanical coupling coefficient of AlN films.10–12 In addition, the fabrication of AlN thin films is not suitable for cost effective chemical solution deposition method even if the figure of merit (FOM) of AlN-based energy harvesters is comparable with other perovskite thin film-based devices due to lower dielectric constant.

In this letter, we demonstrate practical power generation by small-scale MEMS energy harvesting devices using bio-compatible lead-free perovskite (K0.5, Na0.5)NbO3 (KNN)-based piezoelectric films with optimal Mn dopant concentration by cost effective chemical solution deposition method.

The 1 μm-thick lead-free Mn-doped KNN thin films were prepared by the chemical solution deposition process. Approximately 15% excess sodium and potassium were added to the precursor solutions to compensate for the loss of these components during the annealing process of the films. For MEMS energy harvesting devices, the device structure was based on Pt (150 nm) top electrode/Mn-doped KNN film (1 μm)/Pt (150 nm) bottom electrode/Ti (10 nm) adhesion layer/500 nm-thick electrical passivation SiO2 layer/20 μm-thick Si device layer/500 nm-thick buried SiO2 layer/500 μm-thick handling layer. The whole MEMS process used to fabricate this kind of device is described in our previous paper.9 The device had dimensions of 1 mm (W) × 5 mm (L) × 26 μm (H) with Si poof mass dimensions of 1 mm (W) × 3 mm (L) × 526 μm (H). The effective volume (beam + proof mass) of the device was around 0.002 cm3. The fabricated devices were encapsulated conformally with parlyene-C films using a polymer CVD system (PDS2010 Labcoater, Specialty Coating Systems). Parylene-C polymer is an excellent moisture barrier and bio-compatible protector which has been widely used for active electronics and medical devices with its pin-hole free deposition.13 The performance of the device was measured using a vibration system (LW-132-4.5, Labworks, Inc.).

The appropriate choice of materials is the most important factor in enhancing the performance and reliability of lead-free piezoelectric thin film-based energy harvesters. Ever since Saito et al. reported comparable piezoelectric properties of KNN-based ceramics with PZT-based systems,14 and St. Jude Medical AB demonstrated their bio-compatibility for medical implants,15 KNN-based systems have been highlighted as the leading candidates for lead-free piezoelectric device applications. Therefore, intense research has been devoted to improve the piezoelectric properties of KNN-based perovskite thin films for small-scale electronic devices. However, unlike bulk ceramics, a serious leakage current problem has to be overcome to apply these materials to small scale piezoelectric devices due to low breakdown strength by low leakage tolerance of thin films, which causes insufficient poling, easy breakdown, and poor piezoelectric properties. Leakage current density of KNN-based thin films is mainly affected by non-uniform and porous microstructures, interfacial reaction with substrates, and defect structure.4,16,17 Among various leakage current sources, many studies have demonstrated that the control of both oxygen vacancies and free carrier holes is the most important parameter in reducing the severe leakage current density of this material system,16–18 but so far, very few successful lead-free KNN thin film-based piezoelectric devices have been reported19 due to low breakdown strength and high leakage current.

Figure 1(a) shows the leakage current density of 1 μm-thick KNN films as a function of Mn content at the applied voltage of 10 V. The leakage current density of the film is decreased significantly with increasing the Mn concentration to 0.5 mol. % where it shows the lowest value of ∼10−7 A/cm2. Thereafter, the leakage current density is slightly increased to 1.0 mol. % Mn addition and then rapidly increased with further addition, approaching the similar value with un-doped KNN film. The significant increase in the leakage current density beyond 1.0 mol. % Mn addition might be attributed to the accumulation of Mn at grain boundaries due to solubility limit of Mn in our KNN films,16 where it acts as a secondary phase. Based on these results, we selected the Mn content to 0.5 mol. % in the KNN films for MEMS energy harvesting device applications. Figure 1(b) shows the leakage current density of the un-doped and the 0.5 mol. % Mn-doped KNN films as a function of applied voltage. It is well known that main charged defects and the source of high leakage current density of KNN-based thin films are initially oxygen vacancies (VO) due to high volatilization of alkali ions such as K+ and Na+ during pyrolysis and annealing process.16,18 Generally, to compensate for the loss of alkali ions and to reduce the oxygen vacancies, excessive sodium and potassium elements (15% excess in this study) are added in the initial precursor solution. However, this is not the perfect solution to keep the leakage current density low. During a relatively long final anneal, the leakage current density is increased again by the continuous incorporation of oxygen into the film lattice which induces free carrier holes (2h). The leakage current in this stage is governed by p-type electron-hole conduction. This reaction subsequently proceeds by the following equation

VO+12O2OO×+2h,
(1)

where VO is oxygen vacancy and OO× is O2− at the oxide-ion site.

FIG. 1.

Leakage current density of (a) KNN films as a function of Mn dopant concentration at the applied voltage of 10 V and (b) pure KNN and 0.5 mol. % Mn-doped KNN films as a function of applied voltage.

FIG. 1.

Leakage current density of (a) KNN films as a function of Mn dopant concentration at the applied voltage of 10 V and (b) pure KNN and 0.5 mol. % Mn-doped KNN films as a function of applied voltage.

Close modal

The increase in the leakage current density has also been observed for KNN ceramics annealed in O2 atmosphere, which indicates that the leakage current is dominated by the hole conduction after high temperature annealing. To reduce the concentration of the free carrier holes in the films, Mn is incorporated as a dopant element into KNN films. The valence state of Mn can be changed from +2 to +3 or +4 by absorbing the free carrier holes at room temperature,18 which effectively suppresses the electrical conductivity of KNN films. With 0.5 mol. % Mn dopant, the leakage current density of the films is dramatically decreased due to the reduced concentration of free carrier holes, as shown in Fig. 1. The leakage current density for the 0.5 mol. % Mn-doped films is ∼10−7 A/cm2 at 10 V, which is 104–105 times lower than that of pure KNN thin films.

Figures 2(a) and 2(b) show the typical polarization hysteresis loops of the un-doped and the 0.5 mol. % Mn-doped KNN films as a function of the applied voltage, respectively. The hysteresis loops of the pure KNN film show huge polarization values and an abnormal hysteretic behavior without saturation, which are due to its high leakage current density. The abnormal hysteresis shape and the huge non-saturated polarization values usually indicate large leakage currents in the film because the ferroelectric tester also measures the leakage charges collected on the integrated capacitor and converts them to polarization values,20 which means that the polarization values in Fig. 2(a) are not real. In contrast, the 0.5 mol. % Mn-doped KNN film shows stable and well-saturated ferroelectric hysteresis loops even at high applied voltage of 20 V, as shown in Fig. 2(b). The measured remanent polarization and the coercive voltage of the films are 10 μC/cm2 and 1.7 V at 10 V. These values do not change much with increasing the applied voltage to 20 V.

FIG. 2.

Polarization-voltage hysteresis loops of (a) pure KNN thin film and (b) 0.5 mol. % Mn-doped KNN thin film.

FIG. 2.

Polarization-voltage hysteresis loops of (a) pure KNN thin film and (b) 0.5 mol. % Mn-doped KNN thin film.

Close modal

The reliable characterization with high resolution of piezoelectric properties is very important for accurate evaluation of material properties as well as for the proper design of piezoelectric devices. A resolution of approximately 10−3–10−2 nm is required to investigate the low-signal piezoelectric response due to the small film thickness if the displacements are non-linear with respect to the applied voltage. Such a resolution can be accomplished with the sensitive interferometric technique.21 In addition, if a double-beam optical scheme is used with this technique, the artifacts from substrate bending during actuation can be eliminated. Before the piezoelectric measurement of the films, the calibration of the double-beam laser interferometer with the maximum resolution of 0.2 pm/V (aixDBLI, aixACCT system, Germany) is performed by measuring the displacement of X-cut quartz which has negligible nonlinearities in d11 even from low driving voltage to an extremely high excitation level. Figure 3(a) shows the displacement signals measured by the lock-in amplifier as a function of driving voltage applied to the quartz sample at 1 kHz. The d11 value determined by the proportionality in the graph is approximately 2.25 pm/V, which is within 3% of the table value for quartz (d11 = 2.3 pm/V). Figure 3(b) shows the piezoelectric hysteresis loop for the 0.5 mol. % Mn-doped KNN thin film using a double-beam laser interferometer. The piezoelectric d33 coefficient of the film is measured as a small signal response to a small ac field as a function of a much larger dc field. The measured d33 value of the film is ∼90 pm/V. Piezoelectric e31 coefficient is also measured using 4-point bending method22 and the average value is around −8.5 C/m2. These piezoelectric coefficients are much higher than other lead-free perovskite piezoelectric thin films.23,24 The result shows that the KNN thin film with a small concentration of Mn-dopant is an attractive candidate lead-free piezoelectric material system for replacing the PZT-based piezoelectric films.22 Unfortunately, we could not measure the piezoelectric d33 or e31 value of the pure KNN film due to the low breakdown strength induced by the high leakage current density.

FIG. 3.

(a) Linear response of the double-beam laser interferometric displacement signal of X-cut quartz and (b) piezoelectric coefficients of 0.5 mol. % Mn-doped KNN thin film as a function of driving voltage.

FIG. 3.

(a) Linear response of the double-beam laser interferometric displacement signal of X-cut quartz and (b) piezoelectric coefficients of 0.5 mol. % Mn-doped KNN thin film as a function of driving voltage.

Close modal

For MEMS-based vibrational energy harvesting devices, the design of the device structure should be beneficial in creating a high strain on the piezoelectric film layer. In this study, the cantilever beam with a Si proof mass, which is the freestanding structure with one-side clamping, is designed to produce high strain and low resonance frequency under a seismic force. The bending or warping of the cantilever beam due to residual stress between the film and structure layers is prevented by adding a Si support layer. Figures 4(a) and 4(b) show the optical images of various types of patterned d31 mode energy harvesting cantilevers and the fabricated energy harvester using the 0.5 mol. % Mn-doped KNN thin film, respectively. The insets in Fig. 4(b) show the SEM plane-view and the cross-sectional images of the film. It is well known that pure KNN films have non-uniform microstructure with high porosity due to high volatilization of sodium and potassium elements.17,25 However, it is clearly indicated that the 0.5 mol. % Mn-doped KNN film with 15% excessive sodium and potassium addition is very effective in developing dense and uniform morphology without any pores. It has also been proven in many other lead-free perovskite material systems that Mn is an effective dopant to improve the density and electromechanical properties.26,27

FIG. 4.

Optical images of (a) the mask-patterned d31 mode energy harvesting cantilevers, (b) the fabricated MEMS energy harvester using the 0.5 mol. % Mn-doped KNN thin film, (c) the measured resonance frequency and (d) the measured peak voltage and the average power output of the energy harvesting device at the resonance frequency of 132 Hz with and without poling process as a function of resistive loads. The insets in Fig. 4(b) show the surface and the cross-sectional microstructures of the 0.5 mol. % Mn-doped KNN thin film.

FIG. 4.

Optical images of (a) the mask-patterned d31 mode energy harvesting cantilevers, (b) the fabricated MEMS energy harvester using the 0.5 mol. % Mn-doped KNN thin film, (c) the measured resonance frequency and (d) the measured peak voltage and the average power output of the energy harvesting device at the resonance frequency of 132 Hz with and without poling process as a function of resistive loads. The insets in Fig. 4(b) show the surface and the cross-sectional microstructures of the 0.5 mol. % Mn-doped KNN thin film.

Close modal

For the given device stacked-structure, the dimensions of the cantilever length and the proof mass are very important because they determine the resonance frequency of the vibrational energy harvesting device. Since the power output signal is severely reduced even by a small deviation from the resonance frequency due to the narrow bandwidth of small-scale vibrational energy harvesting devices,9–12 the precise dimension control is required for matching the target frequency. We designed the cantilever beam length and the Si proof mass to have a target resonance frequency under 200 Hz because general environmental vibration frequency is below this value. Figure 4(c) demonstrates that our design effectively restricts the resonance frequency to 132 Hz. Figure 4(d) shows the measured peak voltage and the average power output of this device at the resonance frequency of 132 Hz with and without poling process. For achieving maximum power output of the piezoelectric device, the poling process is essential prior to the power measurement. The poling process for dipole alignment is performed by applying 12 V at 80 °C for 20 min. Since 0.5 mol. % Mn-doped KNN film shows low leakage current density and high breakdown strength, the sufficient dipole alignment process is possible using this poling condition without any electrical damage to the film. However, the pure KNN film-based device cannot be poled using the same condition and shows no recordable power generation due to high leakage current density and easy breakdown. When excited at 1.0 G (G = 9.81 m/s2) acceleration, the peak voltage increases with increasing resistive load. The average power reaches the maximum value at 9.6 kΩ resistive load, which is the optimal load value in our system. The output voltage of the poled device shows three times higher than that of the un-poled device. The maximum power, the power density, the corresponding peak voltage, and the calculated figure of merit (FOM = e312/ε0ε33) of the poled device are 3.62 μW, 1800 μW/cm3, 520 mV, and 10.2 GPa, respectively. These are comparable to those in the conventional PZT-based vibrational MEMS energy harvesting device system.9,10

In summary, Mn-doped KNN thin film was systematically investigated for small-scale piezoelectric MEMS energy harvesting applications. The small concentration of Mn addition in KNN film was very effective in reducing the leakage current density. The 0.5 mol. % Mn-doped KNN thin film with dense and uniform microstructure also showed good piezoelectric d33 and e31 values of 90 pm/V and −8.5 C/m2, respectively. Based on material property characterization, we optimally fabricated the thin film-based MEMS piezoelectric cantilever with a Si proof mass for energy harvesting device applications. The maximum power output, the power density, and the corresponding peak voltage of the poled device showed 3.62 μW, 1800 μW/cm3, and 520 mV at the resonance frequency of 132 Hz, which is comparable device performance with PZT-based MEMS devices. These results suggest that the bio-compatible 0.5 mol. % Mn-doped KNN thin film-based piezoelectric MEMS energy harvesting devices shows a strong feasibility for small-scale self-powered electronic device applications as a long-lasting green power source.

This research was financially supported by National Science Foundation (NSF), under Award No. 1408344. The authors at University of Ulsan would like to thank Basic Science Research Program (2014R1A1A4A01004404) through the National Research Foundation of Korea (NRF) funded by the Ministry of Education for the support.

1.
S.
Roundy
,
P. K.
Wright
, and
J.
Rabaey
,
Comput. Commun.
26
,
1131
(
2003
).
2.
L. M.
Miller
,
E.
Halvorsen
,
T.
Dong
, and
P. K.
Wright
,
J. Micromech. Microeng.
21
,
045029
(
2011
).
3.
S.-B.
Kim
,
H.
Park
,
S.-H.
Kim
,
H. C.
Wikle
,
J.-H.
Park
, and
D.-J.
Kim
,
J. Microelectromech. Syst.
22
,
26
(
2012
).
4.
S.-H.
Kim.
,
A.
Leung
,
C. Y.
Koo
,
L.
Kuhn
,
W.
Jiang
,
D.-J.
Kim
, and
A. I.
Kingon
,
Mater. Lett.
69
,
24
(
2012
).
5.
S. R.
Anton
and
H. A.
Sodano
,
Smart. Mater. Struct.
16
,
R1
(
2007
).
6.
H. W.
Kim
,
S.
Priya
,
K.
Uchino
, and
R.
Nwenham
,
J. Electroceram.
15
,
27
(
2005
).
7.
W.
Wu
,
S.
Bai
,
M.
Yuan
,
Y.
Qin
,
J. L.
Wang
, and
T.
Jing
,
ACS Nano
6
,
6231
(
2012
).
8.
S. Y.
Chung
,
S.
Kim
,
J.
Lee
,
K.
Kim
,
S.
Kim
,
C.
Kang
,
S.
Yoon
, and
Y. S.
Kim
,
Adv. Mater.
24
,
6022
(
2012
).
9.
D.
Shen
,
J.-H.
Park
,
J. H.
Noh
,
S.-Y.
Choe
,
S.-H.
Kim
,
H. C.
Wikle
 III
, and
D.-J.
Kim
,
Sens. Actuators, A
154
,
103
(
2009
).
10.
S.
Kim
,
S.
Priya
, and
I.
Kanno
,
MRS Bull.
37
,
1039
(
2012
).
11.
R.
Elfrink
,
T. M.
Kamel
,
M.
Goedbloed
,
S.
Matova
,
D.
Hohlfeld
,
Y.
van Andel
, and
R.
van Schaijk
,
J. Micromech. Microeng.
19
,
094005
(
2009
).
12.
T. T.
Yen
,
T.
Hirasawa
,
P. K.
Wright
,
A. P.
Pisano
, and
L.
Lin
,
J. Micromech. Microeng.
21
,
085037
(
2011
).
13.
E.
Meng
,
P.-Y.
Li
, and
Y.-C.
Tai
,
Sens. Actuators, A
144
,
18
(
2008
).
14.
Y.
Saito
,
H.
Takao
,
T.
Tani
,
T.
Nonoyama
,
K.
Takatori
,
T.
Homma
,
T.
Nagaya
, and
M.
Nakamura
,
Nature
432
,
84
(
2004
).
15.
K.
Nilsson
,
J.
Lidman
,
K.
Ljungstrom
, and
C.
Kjellman
, U.S. patent 6526984 B1 (4 March
2003
).
16.
L.
Wang
,
W.
Ren
,
W.
Ma
,
M.
Liu
,
P.
Shi
, and
X.
Wu
,
AIP Adv.
5
,
097120
(
2015
).
17.
S. Y.
Lee
,
C. W.
Ahn
,
J. S.
Kim
,
A.
Ullah
,
H. J.
Lee
,
H.-I.
Hwang
,
J. S.
Choi
,
B. H.
Park
, and
I. W.
Kim
,
J. Alloys Compd.
509
,
L194
(
2011
).
18.
Y.
Kizaki
,
Y.
Noguchi
, and
M.
Miyayama
,
Appl. Phys. Lett.
89
,
142910
(
2006
).
19.
B.-Y.
Kim
,
I.-T.
Seo
,
Y.-S.
Lee
,
J.-S.
Kim
,
S.
Nahm
,
C.-Y.
Kang
,
S.-J.
Yoon
,
J.-H.
Park
, and
Y.-H.
Jung
,
J. Am. Ceram. Soc.
98
,
119
(
2015
).
20.
J. F.
Scott
,
J. Phys.: Condens. Matter
20
,
021001
(
2008
).
21.
A. L.
Kholkin
,
C.
Wutchrich
,
D. V.
Taylor
, and
N.
Setter
,
Rev. Sci. Instrum.
67
,
1935
(
1996
).
22.
K.
Prume
,
P.
Muralt
,
F.
Calame
,
T.
Schmitz-Kempen
, and
S.
Tiedke
,
J. Electroceram.
19
,
407
(
2007
).
23.
D. Y.
Wang
,
N. Y.
Chan
,
S.
Li
,
S. H.
Choy
,
H. Y.
Tian
, and
H. L. W.
Chan
,
Appl. Phys. Lett.
97
,
212901
(
2010
).
24.
Y.
Nakashima
,
W.
Sakamoto
,
H.
Maiwa
,
T.
Shimura
, and
T.
Yogo
,
Jpn. J. Appl. Phys., Part 2
46
,
L311
(
2007
).
25.
A.
Kupec
,
B.
Malic
,
J.
Tellier
,
E.
Tchernychova
,
S.
Glinsek
, and
M.
Kosec
,
J. Am. Ceram. Soc.
95
,
515
(
2012
).
26.
M.
Abazari
,
E. K.
Akdoğan
, and
A.
Safari
,
Appl. Phys. Lett.
92
,
212903
(
2008
).
27.
W.
Hofman
,
S.
Hoffmann
, and
R.
Waser
,
Thin Solid Films
305
,
66
(
1997
).