1. Introduction
Thermoelectric power generation is attractive from an environment-friendly aspect. Nowadays the solar battery and the wind electric power generation are well accepted and spread out as a clean generation method. For the practical applications of thermoelectric power generation, it is necessary to develop new materials to pursue better effectiveness. At the same time, thermoelectric materials can be applied to electronic cooling and heating devices on the other hand.
Among various materials, Bi2Te3 has been best known and studied as well as a Bi-(Te,Se) alloy. Because of a large thermoelectric figure of merit around 3 × 10–3 K–1 at room temperature [1], the Bi2Te3 system has been deeply investigated so far. On the basis of its crystal structure, the system shows a large asymmetry in resistivity, Hall coefficient, magnetic resistance and thermal conductivity. Researchers reported the existence of the second conduction band [2,3], the asymmetry of relaxation time of carriers [4], and the determination of the scattering coefficient of carriers [5]. In addition, anisotropic galvanomagnetic and thermoelectric properties have been studied [6]. In recent years several group reported that a superlattice structure enhances thermoelectric figure of merit [7,8]. This is actually an excellent idea though forming superlattice requires high-cost manufacturing methods such as a molecular beam epitaxy.
The Bi-Sb alloy is another good candidate for a thermoelectric material in a room temperature range. Further, the Bi-Sb alloy conveniently has the maximum figure of merit at room temperature. V. D. Das et al. reported [9] the dependence of resistance on temperature, in a Bi-Sb alloy film deposited by a vacuum deposition method. R. Tolutis et al. [10] observed negative magneto-resistance in a Bi-Sb alloy film. In addition, superlattice nanowire arrays of Bi and Sb was recently demonstrated by means of a pulsed electrodeposition method [11]. While the Bi-Sb system has been widely studied, a Bi-Sb alloy film was deposited by an electrodeposition method [12]. This method only needs a container of electrolyte, where soaked a substrate and a counter electrode connected with an electric power supply. It is why this method actually has an advantage of low cost for the deposition. An electrode made of Bi-Sb alloy has been usually used for the deposition. We tried to use a Cu electrode for the deposition for the purpose of better availability of the electrode and analyzed the obtained Bi-Sb alloy films.
2. Experimental
Electrolyte consisted of the mixture of BiCl3 (12 g) and SbCl3 (1.5 g) dissolved in aqueous HCl (50 ml) diluted with water (100 ml). The amount of SbCl3 was quite small because of its hard solubility. A substrate electrode and a counter electrode were situated in the electrolyte at a distance of 1.5 cm to each other. Both electrodes were made of Cu plates masked with polymer films where the square (1 × 1 cm) area was exposed to the electrolyte. A direct current-constant power supply (40 mA) was connected to the substrate as a cathode and the counter electrode as an anode. The deposition was performed continuously for 20 hours. The voltage appeared across the electrodes was 0.11 V at the beginning of the deposition.
The amount of the Bi-Sb alloy deposited on the substrate was about 1.8 g.
The obtained samples were investigated with an x-ray diffraction (XRD) apparatus (RAD-IIA by Rigaku), a scanning electron microscopy (SEM: S-570 by Hitachi) equipped with an electron probe micro analysis (EPMA: EMAX-5770 by Horiba). The main issue of the investigations is elemental constitution of the obtained material.
The surfaces of grown Bi-Sb alloy samples were too rough to measure x-ray diffraction pattern. The grown Bi-Sb alloy samples were cleaved off from the Cu substrate and then the interface to the substrate was analyzed. The grounded powder of Bi-Sb alloy in the surface region is also measured. As for an EPMA measurement, the cross sections of the Bi-Sb alloy samples were studied for a depth profile.
3. Results and Discussion
The SEM photos of the cleaved interface and the surface of a BiSb sample are shown in Figures 1(a) and (b), respectively. The obtained BiSb evidently consisted of polycrystals. The grain size at the interface was under 10 μm but that at the surface was several ten μm. The fact argues that the crystal grows gradually as the film was deposited thicker.
Figure 2(a) shows an XRD pattern obtained from the cleaved interface of Bi-Sb. Eleven sharp peaks marked with open circles in the figure were observed, corresponding to Bi-Sb were observed, with a small peak situated at 43.5 degree in 2θ assigned to (111) reflection by Cu. A relatively strong peak at 67.7 degree was not well assigned yet. Figure 2(b) shows a powder XRD pattern from the powder obtained by grinding a part of a surface region only. In this pattern, the same 11 peaks to Bi-Sb were observed with a little shift (from 0.2 to 0.4 degree) in the lower side. It can not be determined that the shift derived from the compositional change or the change in stress caused by the grind to powder. It should be mentioned that the structure at 43.5 degree (Cu) was not seen in the Figure 2(b), which argues the reflection by Cu (111) in Figure 2(a) derived from a Cu trace that trans-