Application of electroplated Ni–Fe thick film to magnetostrictive material in vibration power generator

Magnetostrictive vibration power generation has attracted much attention as one of energy harvesting technologies. Cantilever is a key component which influences the output power and it is generally prepared by attaching a magnetostrictive plate with thickness in a range of about 25–500 μm on a thicker base plate, where neutral plane exists in the base plate and either tensile or compressive stress is occurred in the magnetostrictive plate when the cantilever is bent by vibration. Recently, magnetic materials with low magnetic anisotropy constants (K₁) and moderately-large saturation magnetostrictive coefficients (λ_s) such as Fe–B^1–3 and Fe–Si^4–6 alloys have been used as magnetostrictive materials as well as those with moderately-low K₁ and large λ_s such as Fe–Ga^7–9 and Fe–Co^10–12 alloys.

Ni–Fe alloy is a typical soft magnetic material and the bulk material shows low K₁ value of −2.0 × 10³ to +1.2 × 10³ J/m³ and moderately-large λ_s value of −43 × 10⁻⁶ to +32 × 10⁻⁶ depending on the composition.^13,14 Therefore, it is one of the strong candidates for magnetostrictive materials. Ni–Fe alloy films of 10–36 μm thicknesses have been prepared by electroplating^15–17 and they seem possible to be applied to magnetostrictive materials. In the present study, cantilevers are prepared by forming Ni–Fe alloy films with different compositions on Cu base plates by electroplating. The influences of magnetic and magnetostrictive properties of electroplated film on the vibration power generation property are systematically investigated.

Ni_100−xFe_x (at. %) alloy films were prepared on Cu plates by using an electroplating system with stirring and heating functions. Two Ni plates (purity: 99.98%) of 140 mm length, 40 mm width, and 2 mm thickness and a Cu plate (purity: 99.9%) of 50 mm length, 10 mm width, and 300 μm thickness were respectively used as anodes and a cathode. One side of the Cu plate was covered with polyimide tape so that only the other side was plated. 1 dm³ of commercial pure water, which were prepared by distillation, hollow fiber membrane filtration, and ion exchange, was used as a solvent. NiSO₄·6H₂O of 0.2(1 − y) mol and FeSO₄·7H₂O of 0.2y mol were employed as reagents supplying Ni and Fe ions, respectively. The mass ratio was varied from y = 0 to 1 to control the film composition. NaCl of 0.85 mol, C₇H₄NNaO₃S·2H₂O of 0.02 mol, and C₆H₈O₇·H₂O of 0.05 mol were used as reagents to promote the anode dissolution, to reduce the internal stress in film, and to control the pH to about 2.2–2.4, respectively. The solution temperature was kept constant at 50 °C. The current density on cathode surface was fixed at 3.2 mA/mm². The plating time was set at 1200 s.

The film composition was estimated by energy dispersive x-ray spectroscopy (EDS). The crystal structure was investigated by x-ray diffraction (XRD) with Cu-Kα radiation (wave length: 0.154 18 nm). The effective film thickness was calculated from the film mass measured by using an electronic balance and the density estimated from EDS and XRD data. Figure 1(a) shows the compositional dependence of film thickness. The film thickness is almost constant at 37 μm despite the composition.

The in-plane magnetization curves were measured by vibrating sample magnetometry (VSM). The samples including Cu base plates were cut into small pieces with the dimension of 10 × 10 mm² for VSM measurements after characterizations of magnetostriction and vibration power generation properties. The magnetostriction was investigated by using a cantilever method¹⁸ under in-plane static magnetic fields up to ±1 kOe (= ±79.6 kA/m). The bending at 35 mm length from the fixed end, which was defined in the present study as magnetostrictive bending (ΔS) value, was measured by using an optical leverage method. The λ_s values were not calculated, since the plated films were not thin enough compared to the base plates, which made it difficult to use an approximate formula for λ_s calculation.

Figure 1(b) shows the measurement system of vibration power generation property. The beam is fixed at 10 mm length from one end and vibrated at the respective resonance frequency in a range of 105–115 Hz with the acceleration of 1.5 G. The bias magnetic field (H_bias) is applied along the length direction by using a Helmholtz coil and the strength is varied in a range from 0 to 300 Oe (= 23.9 kA/m). The output voltage is picked up by using a detection coil whose number is 8000 turns, resistance is 1223 Ω, inductance is 476 mH, and inner width and height are respectively 16 and 12 mm.

Figure 2(a) shows the XRD patterns measured for Ni_100−xFe_x films with different compositions. 111, 200, 220, 311, and 222 reflections from fcc phase are observed for the films with x = 0–57, where fcc single phase is formed. When the Fe composition increases up to x = 61, weak reflections from bcc phase such as 110, 211, and 220 start to be overlapped with those from fcc phase. The film with x = 61 involves a small volume of bcc phase. As the Fe composition increases from x = 61 to 80, the reflection intensity from fcc phase decreases, whereas that from bcc phase increases. The result indicates that the volume ratio of bcc to fcc phase increases. When the Fe composition further increases up to x = 89, reflections from only bcc phase are recognized, similar to the case of nearly pure Fe film with x = 99. The Fe-rich films with x = 89–99 consist of bcc single phase. Figure 2(b) summarizes the phase diagram of plated film, which is compared with those of sputtered film¹⁹ and bulk.²⁰ The compositional regions of fcc and bcc single phases for plated film are respectively narrower and wider than those for sputtered film and bulk.

Figure 3(a) shows the examples of in-plane magnetization curves. Figure 3(b) summarizes the saturation magnetization (M_s) values, where the dotted and dashed lines respectively correspond to those of sputtered films²¹ and bulks.¹³ The M_s value increases from 0.44 to 1.32 kemu/cm³ with increasing the Fe composition from x = 0 to 61. However, when the Fe composition increases up to x = 65, the M_s value slightly decreases to 1.17 kemu/cm³. The result seems to be related to decrease in Curie temperature, similar to the case of bulk.¹³ As the Fe composition further increases from x = 77 to 99, the M_s value increases from 1.43 to 1.62 kemu/cm³. Figure 3(c) summarizes the coercivities (H_c). The compositional dependence of H_c value is similar to those of K₁ values for sputtered film²² and bulk,¹⁴ as shown by the dotted and dashed lines, since the K₁ value is related to H_c value.²³ The Ni (x = 0) and the Fe (x = 99) films respectively show high H_c values of 94 and 40 Oe, indicating that the magnetic anisotropy constants of these films are not so small. On the other hand, low H_c values of around 4 Oe are obtained for the films with x = 41–65.

Figure 4(a) shows the magnetic field dependences of ΔS values of Ni_100−xFe_x films with different compositions. Here, the positive and negative signs correspond to bendings to Cu base plate and Ni_100−xFe_x film sides, respectively. Figure 4(b) summarizes the saturation magnetostrictive bending (ΔS_sat) values. The compositional dependence of ΔS_sat value of plated film is similar to those of λ_s values of sputtered film²⁴ and bulk,¹³ as shown by the dotted and dashed lines. The Ni-rich films with x = 0–14 show large negative magnetostriction, whereas large positive magnetostriction is recognized for the films with x = 57–80.

Figure 5(a) shows the H_bias dependences of peak voltages (V_0-p) obtained by using Ni_100−xFe_x films with different compositions, which are estimated from the output waveforms as shown for example in the inset. As the H_bias value increases, the V_0-p value rapidly increases and then reach a maximum value for the films with x = 10–89. With further increasing the H_bias value, the V_0-p value gradually decreases. On the contrary, a maximum V_0-p value of 0.09 V is observed at the H_bias value of 0 Oe for the Ni film (x = 0) and the V_0-p value slightly decreases with increasing the H_bias value, possibly due to that the film has a high magnetic anisotropy constant as indicated by VSM measurement. Furthermore, the V_0-p value is nearly zero for the Fe film (x = 99), whose magnetic anisotropy constant is not small and magnetostriction is very small. Figure 5(b) summarizes the bias magnetic fields where maximum V_0-p values are observed (H_Vmax). Low H_Vmax values of about 2 Oe are recognized for the films with x = 41–65, which is consistent with the low H_c compositional region as shown in Fig. 5(c). The result suggests that H_Vmax value can be reduced by using a material with low magnetic anisotropy. Figure 5(d) shows the compositional dependence of maximum peak voltage (V_max). V_max values higher than 1 V are obtained for the films of x = 61–65 whose H_c and ΔS_sat are respectively low and high as shown in Fig. 5(c). Therefore in order to obtain a higher output voltage, it is important to employ an Ni−Fe film material not only with large magnetostriction but also with low magnetic anisotropy.

Cantilevers are prepared by electroplating Ni_100−xFe_x films with different compositions on Cu base plates. The influences of magnetic and magnetostrictive properties on the vibration power generation property are investigated. The films with x = 41–65 show low H_c values of about 4 Oe, whereas large positive and negative magnetostrictions are respectively observed at x = 0–14 and 57–80. V_max values higher than 1 V are obtained for the films with x = 61–65. The present study has shown that employment of a magnetic film not only with large magnetostriction and but also with low magnetic anisotropy is effective in obtaining a higher output voltage.

This paper is based on results obtained from a project, 20002152-0, subsidized by the New Energy and Industrial Technology Development Organization. Authors thank Mr. Naoki Yoshihara and Mr. Kazuto Okayasu of Instrumental Analysis Center at Yokohama National University for their technical supports for EDS and XRD measurements, respectively.

The authors have no conflicts to disclose.

Shunsuke Aketa: Conceptualization (supporting); Data curation (lead); Formal analysis (lead); Writing – original draft (equal). Mitsuru Ohtake: Conceptualization (lead); Data curation (supporting); Formal analysis (supporting); Supervision (lead); Writing – original draft (equal). Eishi Ishikawa: Data curation (supporting). Yuta Nakamura: Data curation (supporting); Formal analysis (supporting). Tetsuroh Kawai: Validation (supporting); Writing – review & editing (supporting). Masaaki Futamoto: Validation (supporting); Writing – review & editing (supporting).

The data that support the findings of this study are available from the corresponding author upon reasonable request.