A Feasibility Study of Fabrication of Piezoelectric Energy Harvesters on Commercially Available Aluminum Foil

Abstract

In this paper, we present zinc oxide (ZnO)-based flexible harvesting devices employing commercially available, cost-effective thin aluminum (Al) foils as substrates and conductive bottom electrodes. From the device fabrication point of view, Al-foils have a relatively high melting point, allowing for device processing and annealing treatments at elevated temperatures, which flexible plastic substrate materials cannot sustain because of their relatively low melting temperatures. Moreover, Al-foil is a highly cost-effective, commercially available material. In this work, we fabricated and characterized various kinds of multilayered thin-film energy harvesting devices, employing Al-foils in order to verify their device performance. The fabricated devices exhibited peak-to-peak output voltages ranging from 0.025 V to 0.140 V. These results suggest that it is feasible to employ Al-foils to fabricate energy-efficient energy harvesting devices at relatively high temperatures. It is anticipated that with further process optimization and device integration, device performance can be further improved.

1. Introduction

With much progress in a variety of small-scale sensors, various energy harvesting technologies have also been developed. Zinc oxide (ZnO), in particular, having a large piezoelectric constant has drawn much attention, since it can be deposited in the form of high-quality thin films [1]. Further, lead-free ZnO is a nontoxic, environmentally friendly, and biocompatible material [2], thus, it can be used to enable piezoelectric energy harvesters that are implanted into the body or in contact with the body skin. Unlike most ferroelectric materials, it has the advantage that ZnO does not need any poling process after depositing the ZnO thin film [3,4,5,6]. However, lead-free ZnO-based piezoelectric energy harvesters produce lower output voltages than lead zirconate titanate (PZT)-based harvesters, containing toxic piezoelectric materials [7]. Many research results have reported on various piezoelectric energy harvesters using polyethylene naphthalate (PEN), polyethylene terephthalate (PET), various metal wires, various flexible substrates, etc., in order to enhance the output voltages [8,9,10,11,12,13,14,15,16]. Also, various piezoelectric energy harvesters have been fabricated which exploit insulators to inhibit the so-called screening effect of free carriers, compensating for the positively polarized parts of ZnO, thus reducing the piezoelectric effect [7,8,11,13,14]. To suppress the screening effect, wide-band-gap or p-type semiconductor materials have been combined with ZnO [7,8,11,12,15,16,17]. Moreover, mathematical analysis and simulation studies have been theoretically performed to evaluate and interpret the piezoelectric characteristics or phenomenon of piezoelectric materials and devices [18,19].

Some researchers have developed piezoelectric energy harvesters using polyimide (PI) flexible substrates or stainless steel (STS) substrates in an effort to ensure that the substrate will endure any heat processing treatments implemented at temperatures above 400 °C, because plastic substrates become unstable during the annealing process, even below this temperature [20,21]. We have also tried to find a way of annealing devices at relatively high temperatures with the hope of enhancing device performance. Considering candidate substrates, we have chosen to use an aluminum (Al) foil which is highly cost-effective, commercially available, and easy to use. There has been an attempt to make a ZnO/Al stacked structure by depositing a ZnO film on an Al-foil by the sol-gel method, focusing on the photocatalytic effect [22] instead of observing its piezoelectric effect. In this work, we fabricate piezoelectric energy harvesters by depositing ZnO films on Al-foils in order to observe their piezoelectric energy harvesting performance. The Al-foil was cut into a certain size, on which various thin films were deposited, which enabled post-annealing up to ~500 °C (a temperature which PEN substrates cannot withstand), leading to device healing [23,24]. This device annealing treatment may lead to an output voltage improvement and a film durability enhancement, simultaneously.

In this work, to further improve Al-foil-based energy harvesters, we deposited chromium (Cr) as an interlayer to enhance the adhesion between the two adjoined film layers in the device. Other researchers have used chromium (Cr) as a metal electrode [10,25] as well as a thin film layer [26,27] to improve the adhesion quality. Considering these advantages, we deposited a Cr interlayer below the zinc oxide (ZnO) thin film and above the indium tin oxide (ITO) bottom electrode.

We also fabricated various energy harvesters using an aluminum nitride (AlN), n-type a-Si, and SiO₂ layer beneath or on the Cr layer which were used in our earlier studies to form a potential energy barrier. Finally, to simplify the device structure and processing steps, and lower the cost, we fabricated and characterized the energy harvesters (stacked with Ag/ZnO/Cr/Al layers) without the ITO bottom electrode by depositing only the Cr and ZnO films on the Al-foil. These simplified devices have been demonstrated to function well through performance tests, indicating that the newly adopted Al-foils in this work could act as a substrate and bottom electrode, potentially replacing conventional ITO/PEN substrates. Figure 1a shows the device structure of the five-layered films (Al, ITO, Cr, ZnO, Ag) stacked on top of each other and the corresponding energy band diagram (Figure 1b). The work functions (qΦ) of Al, Cr, Ag are, respectively, 4.08 eV, 4.5 eV, 4.26 eV [12,28,29]. And ITO has a bandgap (Eg) of 4 eV and an electron affinity (qΧ) of 4.73 eV [30,31]. On the other hand, ZnO has a bandgap of 3.4 eV and an electron affinity of 4.35 eV [12].

2. Materials and Methods

In this work, 18 μm-thick aluminum foil (Samjin foil, Daegu, Korea) was used as a substrate material to fabricate various energy harvesters. The Al-foil sheet is low-cost, commercially available, and frequently used for cooking, wrapping, crafting, and packaging purposes. A single sheet of Al-foil was cut into pieces each 2.5 × 2 cm² in size. The Al-foils substrates were dipped in hydrofluoric (HF) acid (H₂O:HF = 51:49) (J.T.Baker^TM, Center Valley, PA, USA) for 10 min to eliminate native oxide [32], followed by rinsing three times with deionized (DI) water, drying with a nitrogen blow gun and a halogen lamp. All thin films were deposited by a radio frequency (RF) magnetron sputtering system (HVS-340, Hanvac, Daejeon, Korea). The process details of each film deposition are summarized in

Table 1. Target materials and radio frequency (RF) sputtering conditions used to deposit thin films.

Material	RF Power (W)	Film Thickness (nm)	Deposition Time (min)	Gas/Flow Rate (sccm)
ITO	200	446	30	Ar/20
		1390	90
AlN	80	33	40
n-type a-Si		57
SiO2		52
Cr	160	192	30
ZnO	200	485	90	Ar/18 + N2O/12

. Prior to each film deposition, all targets were presputtered for 20 min under the same conditions (RF power of 150 W, working pressure of 1 × 10⁻² Torr, Ar gas flow rate of 20 sccm), to remove undesirable residues and contaminants. An ITO (4N, iTASCO) target was sputtered with an RF power of 200 W for 30 min to form a Schottky barrier bottom electrode [16]. A ZnO (5N, iTASCO) target was sputtered with an RF power of 200 W for 90 min. A Cr (4N, iTASCO) target was sputtered with an RF power of 160 W for 30 min to improve the adhesion between both ZnO and ITO films.

All the fabricated samples were loaded into an electric furnace (Daesung Electric, Pocheon, Korea) and annealed for a fixed length of time in an Ar atmosphere. The formation of the top electrode using Ag paste (CANS, ELCOAT P-100, Jinchemical, Busan, Korea) completed the device fabrication. The measurement of device output voltages was conducted using a bending machine (Labtech Science, Daejeon, Korea), to which a sheet of resilient PET was attached as a bending platform for the device to be measured. All output voltages were measured with a digital multimeter (34401A, Keysight Technologies, Santa Rosa, CA, USA) while bending with 0.2519% strain, at a 0.17%s⁻¹ strain rate, as shown in Figure 2. Many experiments were carried out to attach the Al-foil-based devices to the PET bending platform efficiently by using adhesives, cyanoacrylate, electric iron, and double-sided tape.

The deposition conditions of all the materials used in this work are described in detail, as summarized in Table 1. It is noted that all the films were deposited under the working pressure of 1 × 10⁻² Torr, while the target-to-substrate distance was 12 cm, and the substrate rotation speed was 8 rpm.

For the ZnO deposition in particular, a mixture of 18 sccm Ar and 12 sccm N₂O gas was flowing during the sputtering process to reduce the free-carrier concentration in the ZnO film, leading directly to the improvement of device performance [16].

Many previously reported devices fabricated on flexible plastic substrates [10,12,13,14,16,17] have an intrinsic limitation to increasing process temperatures. Thus, in this work, instead of flexible plastic substrates, we have attempted to employ the Al-foil as a substrate to further increase the process temperatures, which helpfully allows for annealing heat treatments at higher temperatures.

As illustrated in Figure 3, to understand the effect of annealing on device performance, some devices were annealed once before ZnO deposition (one-step annealing) or twice immediately before and after ZnO deposition (two-step annealing), while others were not annealed at all (nonannealing or as-deposited). The formation of the top electrode using Ag paste completed the device fabrication.

We have confirmed the feasibility of using Al-foil sheet as a substrate on which we fabricated several Ag/ZnO/ITO/Al-foil structured devices by depositing both ZnO and ITO films under various process conditions. However, the output voltages of these devices were not stably generated or as high as expected, possibly because of wrinkling or cracking of the stacked films. To resolve this issue, a thin layer of Cr was deposited to enhance the adhesion between the ZnO and ITO films [26,27], and Ag/ZnO/Cr/ITO/Al-foil structured devices were also fabricated. The surface morphologies of the two different structures, those with and without the Cr layer, after a bending test were investigated using a microscope (S39A, Microscopes Inc., Saint Louis, MO, USA). The devices including the Cr layer showed better output voltages (Figure 4), apparently due to a significant reduction in the wrinkling or cracking phenomenon of the stacked films, as shown in the micrographs of Figure 5.

The output voltages for all devices were measured and compared, as shown in Figure 6, Figure 7, Figure 8 and Figure 9. Interestingly, in this work, it was found that the one-step annealing resulted in better performance than the two-step annealing, however, the reason remains unclear and requires a further study.

To further verify the feasibility of using the Al-foil as a substrate for the fabrication of energy harvesters, we fabricated modified devices by sputtering AlN (5N, iTASCO), n-type a-Si (5N, iTASCO), or SiO₂ (4N, iTASCO) targets and depositing interlayers between the ZnO and Cr films, or Cr and ITO films, to form a total of six different device structures. These devices were measured and compared, as shown in Figure 10. The device with the stacked layers of Ag/ZnO/AlN/Cr/ITO/Al-foil showed higher output voltages than the other devices.

In this work, to fabricate the devices in a more cost-effective way, we replaced the ITO bottom electrode with Al-foil because the Al-foil is a conductive material, acting as a substrate as well as an electrode. From this standpoint, we have fabricated four different device structures on Al-foil by substituting the conventional ITO electrode with the Al-foil electrode. Measurements of their respective output voltages (as mean peak-to-peak output voltage histograms) were conducted for comparison. The fabricated devices and their measured output voltages are shown in Figure 11.

3. Results and Discussion

In this work, the piezoelectric devices fabricated on Al-foil substrates generate substantial output voltages. As shown in Figure 4, we fabricated and measured two sample devices, with and without a Cr layer. The output voltages of (a) Ag/ZnO/ITO/Al-foil and (b) Ag/ZnO/Cr/ITO/Al-foil structured devices are compared. The addition of a Cr layer resulted in a more stable output waveform and larger peak-to-peak voltage, increased from approximately 0.035 V to 0.070 V, while the device without a Cr layer was found to have a less stable waveform with a smaller mean peak-to-peak voltage of 0.035 V. We have investigated and compared the microscopic surface morphologies of both (a) ZnO/ITO/Al-foil and (b) ZnO/Cr/ITO/Al-foil by optical microscopy, after completion of 20 bending cycles (bending was performed prior to the formation of the Ag top electrode), as shown in Figure 5. Each sample shows significantly different morphologies in the magnified micrographs (×400). More crack propagation seems to occur in the stacked film structure without a Cr layer. By depositing a thin Cr interlayer, the adhesion between the ITO and ZnO layers is apparently improved, leading to an enhancement of the device output voltage.

As shown in Figure 6, for further understanding of annealing effects, the six fabricated devices were annealed using either the one-step or two-step annealing processes described in Figure 3. The detailed annealing conditions and sample names are summarized in

Table 2. Sample names and annealing conditions of the eight devices included in Figure 6 and Figure 7.

Sample Name	SA-1	SA-2	SA-3	SA-4	SA-5	SB-1	SB-2	SB-3
Annealing process	as-deposited	one-step	two-step	one-step	two-step	as-deposited	one-step	two-step
Temperature		300 °C	300 °C	400 °C	400 °C		300 °C	400 °C
Duration		1 h	1 h	1 h	1 h		1 h	0.5 h

. The samples which were annealed at 300 °C for 1 h in the “one-step annealing” showed significantly better output voltages than the other devices. The output voltages of the eight samples included in Figure 6 are compared, as shown in Figure 7.

To analyze the crystallinity and composition of the deposited material, the ZnO/Cr/ITO/Al-foil samples were prepared by focused ion beam (FIB). As shown in Figure 8, a low-resolution transmission electron microscope (LRTEM) cross-sectional image of the sample was measured, and the selected area electron diffraction (SAED) patterns of each layer were also acquired by 200 KeV LRTEM. The SAED patterns of the four layers are labeled, respectively, from (a) to (d). All the layer materials appear to exist in polycrystalline forms, where the Al-foil has a face-centered cubic structure [33].

Energy-dispersive X-ray spectroscopy (EDS) depth profiling was performed, as shown in Figure 9. Compositional analysis confirmed the constituent materials of each layer in the ZnO/Cr/ITO/Al-foil structure.

Similarly to other reports [12,13,14], we fabricated additional devices including each material (AlN, SiO₂, or a-Si) used in our earlier researches for a further study, as shown in Figure 10. Here, each target of AlN (5N, iTASCO), n-type a-Si (5N, iTASCO), or SiO₂ (4N, iTASCO) were used to fabricate the devices shown in Figure 10. The deposition conditions along with film thicknesses are described in Table 1. The output voltages were extracted from the fabricated devices, and are compared in Figure 10. In particular, the sample B of the Ag/ZnO/AlN/Cr/ITO/Al-foil structure showed the highest output voltage.

To design the energy harvesting devices in a more cost-effective way, we have fabricated various kinds of similarly formed devices with and without an ITO layer on the Al-foil substrate.

As shown in Figure 11, the devices without an ITO layer have shown more than a two-fold increase in output voltage over those with an ITO layer. It is speculated that with the elimination of the ITO layer, the energy barrier at the Cr/ITO interface may not exist, thus making it easier for the electrons to flow toward the top electrode, as shown in Figure 1b.

Overall, the Al-foil substrate-based devices have shown relatively low output voltages when measured after 180 bending cycles. The exact mechanism of device operation has not been clearly explained yet, however, we attribute the relatively low performance to the bending-cycle-induced strain, causing an interface adhesion deterioration, finally leading to possible wrinkling, delamination, or even cracking of the stacked film layers.

4. Conclusions

We have fabricated and characterized various piezoelectric energy harvesting devices employing a commercially available Al-foil. The adoption of a thin Cr film as an interlayer may enhance the adhesion between two layers (ZnO and ITO), reducing the formation and/or propagation of cracks in devices during bending cycle tests, and thus improving the output voltage stability. Moreover, the annealing effects on the device output voltage performance were investigated under various annealing conditions. In addition to the annealing conditions, the existence of the ITO bottom electrode layer was found to significantly affect the device performance. Overall, the fabricated devices have shown peak-to-peak output voltages ranging from 0.025 V to 0.140 V. The devices without the ITO electrode layer showed two-fold higher output voltages than those with the ITO electrode layer. These findings suggest that it may be possible to employ various kinds of commercially available, cost-effective Al-foils to fabricate more energy-efficient energy harvesting devices, requiring further integration and optimization of device fabrication processes. This work could contribute to developing more advanced energy harvesting devices adopting Al-foil substrates.