Abstract
The method and conditions of Ni plating were optimized to maximize the output of a betavoltaic battery using radioactive 63Ni. The difference of the short circuit currents between the pre- and postdeposition of 63Ni on the PN junction was 90 nA at the characteristics. It is suspected that the beta rays emitted from 63Ni did not deeply penetrate into the PN junction due to a Ni seed layer with a thickness of 500 Å. To increase the penetration of the beta rays, electroless Ni plating was carried out on the PN junction without a seed layer. To establish the electroless coating conditions for 63Ni, nonradioactive Ni was deposited onto a Si wafer without flaws on the surface. This process can be applied for electroless Ni plating on a PN junction semiconductor using radioactive 63Ni in further studies.
1. Introduction
A betavoltaic battery that converts the decay energy of beta- (β-) emitting radioisotopes into electricity is characterized by specialized properties such as a long service life and being free of maintenance [1]. The origin of nuclear batteries using beta-particle was the pacemaker related to pacing industry in 1973. These devices offered young patients the possibility of having a single pacemaker implant that could last their whole life. The β-cell for pacemaker had promethium (147Pm) sandwiched between semiconductor wafers. The impact of the β-particles on PN junction causes a forward bias in the semiconductor similar to the behavior of photovoltaic cell. The β-cell had an open circuit voltage of 4.7 V and a short circuit current of 115 μA. However, it has a half-life of 2.6 years, at most. So, nuclear pacemakers were displaced by devices powered by lithium cells in mid-1980, due to longevity of lithium-powered batteries with approximately 10 years. The important factor affecting the performance of a betavoltaic battery is a radioisotope used as a power source. 63Ni is pure beta-emitter with a low energy spectrum of keV and significantly long half-life of 100.1 years and thus is widely used as the power source of betavoltaic batteries. Tritium (3H) has shorter half-life (12.3 years) than those of 63Ni. However, this isotope has enough longevity, the same as lithium-powered batteries. The beta spectrum of 63Ni is below the radiation damage threshold (approximately 200 keV for Si) of semiconductors such as Si and SiC. Also, 63Ni is easier to handle than other beta-particles such as 3H, 90Sr, and 147Pm because of its low energy spectrum and solid-metal form. For this reason, it is suitable for the power source of a betavoltaic battery to be within the nano- to microwatt range [2–4]. There are several methods for the formation of a Ni deposit onto a substrate such as electroplating, electroless plating [5, 6], and chemical vapor deposition (CVD). Among them, the electroplating process is most commonly used for Ni deposition when using 63Ni as a power source for a battery. The Ni seed layer on a semiconductor is necessary for the deposition of Ni during electroplating, but it can cause a decrease of electric efficiency of the battery because it absorbs beta rays emitted from 63Ni and causes a self-shielding effect of the beta rays. In order to prevent the self-shielding and improve the electrical efficiency of a betavoltaic battery, electroless Ni plating that can be carried out without a seed layer was performed in this study.
Electroless plating is a process for the deposition of metals, such as nickel, onto a substrate via a catalyzed chemical reduction of metal ions at the surface of the substrate without the use of an electric current. Unlike electroplating, an electroless plating process has some advantages such as a thickness uniformity and independence of the electrical properties of the substrate [7, 8]. In addition, it is possible to form a Ni deposit directly on a substrate without a Ni seed layer when using electroless plating, which can solve the problem in which beta rays emitted from radioactive 63Ni are absorbed on the seed layer.
The aim of this work is to optimize the method and conditions of Ni plating to maximize the electrical efficiency of a betavoltaic battery using radioactive 63Ni. In this study, a betavoltaic battery using 63Ni coated on a PN junction semiconductor was fabricated and the characteristics were measured using Probe Station. In addition, an electroless Ni plating process using nonradioactive Ni was established to minimize the decrease in electric efficiency. This process can be applied for electroless plating on a PN junction semiconductor using radioactive 63Ni.
2. Experimental Technique
2.1. 63Ni Electroplating and Measurement of Characteristics
The process of electroplating using radioactive 63Ni in a hot cell (Bank-2, HANARO research reactor in KAERI) was carried out using a two-step process: the preparation of an ionic solution including 63Ni and electroplating on the PN junction semiconductor as a substrate. The spacing of the PN junction semiconductor was 5 μm, and the specific radioactivity of electroplated 63Ni was estimated to be about 2.5 mCi. To perform electroplating 63Ni on the PN junction, the minimum thickness of the seed layer should be 500 Å, for a semiconductor fully covered with metal. The characteristics of a 63Ni-coated semiconductor were investigated using Probe Station of the Precision Source/Measure Unit, B2911A.
2.2. Electroless Ni Plating on Si Wafer
Table 1 shows the composition and operating conditions of an electroless Ni plating bath. The solution for electroless Ni plating was made up of 0.1 M nickel sulfate, 0.23 M sodium hypophosphite, 0.07 M sodium citrate, 0.06 M sodium acetate, and 4 nM lead acetate used as stabilizer. Nickel sulfate is the main salt, sodium hypophosphite is a reducing agent, and sodium citrate and sodium acetate are used as a complexing agent. The pH value of the plating solution was adjusted to 10 with the addition of NaOH, and the bath is heated to a temperature of 75°C. Si wafers with dimensions of 5 × 5 mm3 were used as the substrates. Prior to electroless Ni plating, the wafers were dipped in an HCl solution (10%) for 40 s and washed with distilled water. The cleaned Si wafers were put into a plating beaker for 20 min without any pretreatment, activation, or sensitization. The Ni-coated Si wafers obtained by the electroless plating were cleaned with distilled water and dried for 1 h at 90°C in a vacuum.
The surface morphology of the Ni deposits was observed using field emission scanning electron microscopy (FE-SEM) of JSM-7100F, and the composition and content of deposited Ni was analyzed by X-ray diffraction (XRD). XRD investigations were carried out using a Philips X’Pert-Pro instrument operated at 40 kV and 30 mA with CuKα radiation ( Å) at a scanning rate of 2°/min.
3. Results and Discussion
To evaluate the PN junction prepared by the Electronic Telecommunications Research Institute (ETRI), the electron beam induced current technique has been employed to experimentally simulate the beta emission of 63Ni and to estimate the total device current [9]. The open circuit voltage was found to be 0.29 V. The short circuit currents were 3.3 A. The power output was found to be 66.5 W/cm2. From the e-beam illumination test, we confirmed the good operation of the PN absorber. Figure 1 shows the performance characteristics of the 63Ni-coated PN junction semiconductor with a Ni seed layer with a thickness of 500 Å. A prototype of the electroplating 63Ni was carried out in a glove box in a hot cell (Bank-2, HANARO Reactor in KAERI). The specific radioactivity of the electroplated Ni including 63Ni was estimated to be about 2.5 mCi. The nominal specific activity was measured by a portable activity meter (PAM 1704). The curves of both dark and deposited 63Ni show almost the same values. The difference between the predeposition (dark) and deposited 63Ni can be obtained through a magnification of the curve [4]. The difference of the short circuit current () between the predeposition and postdeposition of 63Ni with a thickness of about 3 μm was found to be 90 nA. The output power was found to be 8.0 pW (picowatt) at an open circuit voltage () of 0.08 mV. The performance of betavoltaic battery prepared by Ulmen et al. was achieved as 2.5 pW of maximum power output at a voltage of 0.4 mV using 4 mCi of 63Ni. [4]. Recently, the maximum performance was achieved at 5~500 nW of stack structure (package) prepared by Widetronix Co. [10]. We obtained a relatively enhanced power output at a single cell, though the specific radioactivity was a lower value than those of other groups. From the characteristics analysis, we confirmed that there are no large differences in the power output between pre- and postdeposition, though the power output is increased as compared with other research groups. This is thought to be due to beta rays emitted from 63Ni not penetrating the Ni seed layer because of the self-shielding of the beta rays.
The penetration depth of the particles in the silicon device was reported in a Katz-Penfold range equation [11]. This equation considers only the density of the materials and energy of the particles. This effectively determines the depth of the depletion region required. We modeled the energy deposition as a function of the depth in the silicon with/without a seed layer using the Monte Carlo code. The self-shielding effect at the seed layer of the radioisotope layer was also investigated. Figure 2 shows the results of the relative energy deposition as a function of thickness for the deposit layer of 63Ni on a Si wafer (a) with a seed layer and (b) without a seed layer. Energy absorption in silicon with a seed layer as a function of 63Ni thickness was decreased compared with those without a seed layer.
(a)
(b)
To prevent the self-shielding, which causes a decrease in the output of the betavoltaic battery, electroless Ni plating that can be carried out without the Ni seed layer was performed. The reaction temperature was changed from 50 to 80°C. The reaction was inactive at and below 70°C. The coating layer was delaminated at and above a reaction temperature of 80°C because of a tremendous massive reduction. A large amount of colloids and precipitation was observed at a high reaction temperature of above 80°C. The coating layer was formed at a pH between 9 and 11. The pH and reaction temperature of the solution were determined at pH 10 and 75°C, respectively. Lead acetate was added to the solution as a stabilizer. The adhesion strength between substrate and coating layer was enhanced, as the stabilizer prevents overabundance of reduction on the substrate. The reaction was inactive, though the stabilizer was added above 0.004 mM in the solution. Figures 3(a) and 3(b) present SEM images of Ni deposits after 20 min of electroless plating. As shown in Figure 3, well-distributed spherical particles are spread on the Si wafer under a condition of 75°C at pH 10. The Ni deposits were uniformly and continuously formed, and there were no flaws on the surface. If the temperature of the plating solution is too high, many more hydrogen bubbles were generated, and thus the coating layer was peeled from the Si wafer because of an unstable surface formation. Figure 3(b) shows the SEM images for thickness of the Ni-coated Si wafers. The thickness of the Ni deposit is about 1.5 μm after 20 min of electroless plating. The agglomeration of Ni particles is observed on the surface layer.
(a)
(b)
Figure 4 shows the results of XRD for the deposited Ni on a Si wafer under the plating condition of 75°C at pH 10. XRD patterns were observed showing that the crystal structure of the Ni deposits is pure FCC (face-centered cubic) Ni. Five characteristic peaks at around 2θ are 44.5°, 51.8°, 76.4°, 93°, and 98.5°, which is consistent with the (111), (200), (220), (311), and (222) lattice plane diffractions of Ni crystals, respectively, which are definitely observed for the Ni particles. Crystal orientations of the films were estimated by the high degree (200) in the XRD patterns, as shown in Figure 4. The main peak of the bulk Ni is presented generally at (111). However, the plane orientation was formed, because of the thin thickness of the coating layer [12].
This process can be applied for electroless plating on a PN junction semiconductor using radioactive 63Ni. The solution for electroless plating using 63Ni is made up of radioactive Ni sulfate dissolved in a sulfuric acid solution. The whole process will be performed in a hot cell in Bank-2, the HANARO reactor in KAERI. An investigation of the surface morphology, component of the deposits, and characteristic of the PN junction semiconductor coated with radioactive 63Ni will be carried out in further studies. However, we could prepare a beta source of 63Ni on substrate with the nominal specific radioactivity on the surface of 0.45 mCi. Figure 5 includes a photograph of the prepared beta source. We attached prepared beta source on the PN junction without seed layer using vacuum. The difference of the short circuit current () between the predeposition and postdeposition of 63Ni was found to be 6.6 nA. The output power was found to be 0.29 nW at an open circuit voltage of 0.044 V. The power output of a single cell was significantly enhanced compared with previously prepared single cell with seed layer.
4. Conclusion
Radioactive 63Ni was coated onto a PN junction semiconductor with a Ni seed layer of 500 Å and the performance characteristics of this semiconductor were measured. There are no significant differences of the short circuit current () between pre- and postdeposition. This is suspected to be due to the self-shielding of beta rays on a Ni seed layer. Thus, electroless Ni plating that can be carried out without the seed layer was performed using nonradioactive Ni. We confirmed that the Ni deposits were well coated onto a Si wafer using SEM and XRD studies. The coating condition was determined by the optimum temperature, pH, and amount of stabilization. In a further study, the process for electroless Ni plating on a PN junction semiconductor using radioactive 63Ni will be optimized. Also, the power output was significantly increased, as the seed layer on the PN junction was removed.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgment
This work is supported by the National Strategic R&D Program for Industrial Technology (10043868), funded by the Ministry of Trade, Industry and Energy (MOTIE).