2018 Volume 59 Issue 7 Pages 1041-1045
Mg2Si has attracted interest as a potential thermoelectric material that can convert waste heat into electricity. To improve thermoelectric performance of Mg2Si, Al/Mg2Si composite materials with nominal composition of xAl/Mg2Si (x = 0.25, 0.5, 0.75, 1.0, or 1.5) were fabricated, i.e., an Al metal phase was introduced to an Mg2Si matrix. Reflecting an increase in the electrical conductivity and a decrease in the Seebeck coefficient with increasing x, the power factor was successfully enhanced by the incorporation of the Al phase. The increase in electrical conductivity was discussed in terms of electron carrier density and carrier mobility.
Recently, thermoelectric materials have attracted much attention because of their ability to convert heat to electricity.1–3) The thermoelectric conversion efficiency of a thermoelectric material is evaluated in terms of the dimensionless figure of merit, zT = σS2T/κ, where σ, S, T, and κ are the electrical conductivity, Seebeck coefficient, absolute temperature, and thermal conductivity, respectively. Additionally, the power factor (PF (= σS2)) is used as a measure of the electric power obtained from the thermoelectric material. Especially, Mg2Si is a promising thermoelectric material because of the natural abundance of Mg and Si as well as its low toxicity, low weight, and high thermoelectric performance in the middle temperature range from 600 K to 900 K.
Numerous studies have been conducted to improve the thermoelectric performance of Mg2Si by means of partial substitution.4–17) In contrast, we focus on another method for the improvement; fabrication of composite materials which has proven to be an effective method for increasing the electrical conductivity and/or decreasing the thermal conductivity.18–22) Quite recently, our group prepared Mg-excess Mg2Si samples with nominal composition Mg2+xSi by means of spark plasma sintering at 1123 K for 10 min under an applied pressure of 30 MPa, and found that the samples were Mg/Mg2Si composite materials where the Mg phase was dispersed in the Mg2Si matrix.23) Because the dispersed Mg is a metal, the existence of the Mg phase led to an increase in the electrical conductivity of Mg2Si, as depicted in Fig. 1. However, we could not further increase the amount of Mg phase because of the high vapor pressure of Mg, i.e., the increase in electrical conductivity was limited for the Mg/Mg2Si composite material. In this study, we attempt to disperse the Al metal in the Mg2Si matrix. Because Al metal has a higher electrical conductivity and lower vapor pressure than Mg metal,24) we expect that the amount of the Al phase in the Mg2Si matrix will become higher than that of the Mg phase, leading to a higher electrical conductivity.
Relation between the electrical conductivity, σ, of Mg-excess Mg2Si measured at 300 K and the area ratio of Mg in scanning electron microscope (SEM) images.23) The bold line is only guide to the eyes.
The powders of Mg (Kojundo Chemical Lab., purity 2Nup, particle size 180 µm), Si (Kojundo Chemical Lab., purity 3N, particle size 5 µm), and Al (Kojundo Chemical Lab., purity 3N, particle size 30 µm) were weighed with an Al molar ratio to the Mg2Si matrix equal to x:1 (x = 0, 0.25, 0.5, 0.75, 1.0, or 1.5), i.e., Al:Mg:Si = x:2:1, and were mixed for 10 min in an Ar atmosphere. Hereafter, we use the notation xAl/Mg2Si to denote the nominal composition of the Al/Mg2Si composites. The mixed powders were synthesized in vacuum by spark plasma sintering (Fuji Electric Industrial, SPS-520S) at 873 K for 10 min under an applied pressure of 30 MPa. To characterize the synthesized samples, powder X-ray diffraction (XRD) measurements using a CuKα radiation source (BRUKER, D8 ADVANCE) and Raman spectroscopy using a laser source at the wavelength of 532 nm (JASCO, NRS5100) were performed. The morphology and composition of cut surfaces of the samples were analyzed using a scanning electron microscope equipped with an energy-dispersive X-ray spectrometer (SEM-EDX; JEOL, JSM-IT100LA). The cut surfaces were ion-etched (Gatan, PIPS Model691) or polished. The electrical conductivity and Seebeck coefficient were simultaneously measured at temperatures ranging from 300 K to 900 K in vacuum using an automated Seebeck tester (Ozawa Science, RZ2001i). Additionally, the carrier density and carrier mobility were measured at room temperature under a magnetic field ranging from −5.0 kOe to 5.0 kOe using a Hall measurement system (Quantum Design, Physical Property Measurement System).
Figure 2(a) shows the XRD patterns of the xAl/Mg2Si samples. Almost all peaks could be assigned to the Al and Mg2Si phases. The XRD peaks at 38° and 44° corresponded to the Al phase, while the indexed peaks could be attributed to the Mg2Si phase. Minor amounts of Mg and SiO2 phases were found for the Mg2Si and 0.25Al/Mg2Si samples, respectively. Additionally, the XRD peaks of the Al phase were detected when x ≥ 0.5. Using the Le Bail analysis of the XRD patterns, the lattice constant of the Mg2Si phase was evaluated, as shown in Fig. 2(b). The lattice constant decreased linearly as x approached 0.75 and subsequently remained almost constant when x was greater than 0.75. Such decrease in the lattice constant has been reported theoretically for an Al-doped Mg2Si system.25) Thus, we can assume that some of the Al in the xAl/Mg2Si samples was used for the partial substitution in the Mg2Si matrix, and that the substitution amount increased up to x = 0.75.
(a) Powder X-ray diffraction patterns of the xAl/Mg2Si samples. (b) Lattice constant, a, of Mg2Si in the xAl/Mg2Si samples as a function of x. The bold and dashed lines are guide to the eyes.
To confirm the Al-substitution in the Mg2Si matrix, Raman spectra were measured as shown in Fig. 3(a). A sharp Raman peak was observed at approximately 260 cm−1 for the Mg2Si sample, corresponding to the optical phonon mode (F2g) of Mg2Si.26,27) There were also sharp peaks in the spectra of the Mg2Si matrix in the 0.75Al/Mg2Si and 1.5Al/Mg2Si samples. It was found that the peaks for the 0.75Al/Mg2Si and 1.5Al/Mg2Si samples were located at a slightly lower energy of approximately 259 cm−1 compared with Mg2Si (see the inset of Fig. 3(a)). This shift can be explained by the partial substitution of Al in the Mg2Si matrix. Since an Al atom is heavier than an Mg atom, the energy of the optical phonon mode of Mg2Si was lowered by the Al partial substitution. Peak positions of the optical phonon mode of the xAl/Mg2Si samples are shown in Fig. 3(b). Due to the dispersion of the Al phase in the Mg2Si matrix, it was difficult to set a laser spot just at the Mg2Si matrix for the Raman spectroscopy. For this reason, the peak positions varied with the location of the laser spot; however, one can find that they tended to decrease, and kept almost constant above x = 0.5, which was a similar trend of the lattice constant shown in Fig. 2(b). Thus, we conclude that the increase in the amount of the Al-substitution occurred from x = 0 to x = 0.5–0.75.
(a) Raman spectra of the Mg2Si, 0.75Al/Mg2Si, and 1.5Al/Mg2Si samples. The inset shows a part of these spectra. (b) Raman shift of the xAl/Mg2Si samples as a function of x. The bold kinked line is guide to the eyes.
Figure 4(a) shows a typical SEM image of the 0.25Al/Mg2Si sample. Bright areas corresponded to the Al phase, while remaining areas corresponded to the Mg2Si phase. A small amount of the Al phase in the 0.25Al/Mg2Si sample, which could not be identified using the XRD measurement because of its low detection limit, was found in the SEM image. We can observe from Fig. 4(b) that the amount of the Al phase in the 1.5Al/Mg2Si sample was larger than that in the 0.25Al/Mg2Si sample. The Al-substitution in Mg2Si in the Mg2Si matrix was confirmed in a SEM image around the Al phase in the 1.5Al/Mg2Si sample (Figs. 4(c), 4(d), 4(e), and 4(f)). In the Mg2Si matrix, Al was homogeneously dispersed, an indication of the uniform Al-substitution. From the results of XRD measurements, Raman spectroscopy, and SEM observations, it was concluded that not only the Al-substitution in Mg2Si but also the incorporation of the Al phase in the Mg2Si matrix occurred in the xAl/Mg2Si samples.
Scanning electron microscope (SEM) images of (a) 0.25Al/Mg2Si and (b) 1.5Al/Mg2Si samples. (c) A SEM image around the Al phase in the 1.5Al/Mg2Si sample with line scan analyses of Mg, Si, and Al, and element mapping of (d) Mg, (e) Si, and (f) Al in the image.
Area ratio of the Al phase in SEM images was evaluated and plotted in Fig. 5. It was found that the area ratio of the Al phase increased monotonically as a function of x, indicating the successful fabrication of Al/Mg2Si composite materials. Although the rate of increase of the area ratio decreased with increasing x, a larger amount of the dispersed phase was observed to remain in the Al/Mg2Si composite material as compared to that in the Mg/Mg2Si case shown in Fig. 1.
Area ratio of Al in SEM images as a function of x. The grey curve is only guide to the eyes.
It has been predicted that the doped Al substitutes for the Mg site, leading to electron carrier doping.7,25) Additionally, Al metal possesses a high electron carrier density of 1.89 × 1023 cm−3.24) Therefore, it is expected that both the Al-substitution in Mg2Si and incorporation of the Al phase in the Mg2Si matrix increase the electrical conductivity of the xAl/Mg2Si samples. Figure 6(a) shows the electrical conductivity, σ, of the xAl/Mg2Si samples. As expected, the electrical conductivity increased with increasing x. The highest electrical conductivity was obtained for the 1.5Al/Mg2Si sample over the whole temperature range. The electrical conductivity of the 1.5Al/Mg2Si sample ranged from 6.77 × 102 S·cm−1 to 1.84 × 103 S·cm−1, which was higher than that of Al-doped Mg2Si reported in previous studies6,7,10,11,16) as well as that of the Mg/Mg2Si composite materials depicted in Fig. 1. Generally, an increase in electrical conductivity results in a decrease in Seebeck coefficient. Such tendency was apparent in Fig. 6(b); the Seebeck coefficient, S, of the xAl/Mg2Si samples decreased with increasing x. The 1.5Al/Mg2Si sample exhibited the lowest Seebeck coefficient throughout the entire temperature range of the measurement. Consequently, the highest power factor, PF, of 2.0 × 10−3 W·m−1·K−2 was obtained for the 0.75Al/Mg2Si sample at 800 K (Fig. 6(c)), which was as high as that of the Al-doped Mg2Si reported in the previous studies.6,7,10,11,16)
Temperature dependence of (a) electrical conductivity, σ, (b) Seebeck coefficient, S, and (c) power factor, PF, of the xAl/Mg2Si samples. The grey curves are guide to the eyes.
Figure 7(a) shows the electrical conductivity of the xAl/Mg2Si samples measured at 300 K. The rate of the increase in electrical conductivity decreased with increasing x. To determine the effects of Al-substitution in Mg2Si and incorporation of the Al phase in the Mg2Si matrix on the electrical conductivity, we performed Hall measurements. The sign of the Hall coefficient was negative, indicating that the majority carriers of the xAl/Mg2Si samples were electrons. Figures 7(b) shows electron carrier density, n300K, of the xAl/Mg2Si samples. The electron carrier density increased with increasing x. The saturation tendency of the electron carrier density well explained the x dependence of electrical conductivity shown in Fig. 7(a). It is noted that the electron carrier density of the 0.75Al/Mg2Si sample (1.4 × 1020 cm−3), where the Al-substitution and the incorporation of Al phase both occurred, was an order of magnitude higher than the literature values of Al-doped Mg2Si.6,7,10,16) Considering the high electron carrier density of Al metal,24) the incorporation of the Al phase is the predominant factor influencing the increase in electron carrier density of the xAl/Mg2Si samples.
(a) Electrical conductivity, σ300K, (b) electron carrier density, n300K, and (c) carrier mobility, μ300K, of the xAl/Mg2Si samples measured at 300 K as a function of x. The grey curves are only guide to the eyes.
The carrier mobility, μ300K, of the xAl/Mg2Si samples is shown in Fig. 7(c). The carrier mobility of the xAl/Mg2Si samples gradually decreased with increasing x. It has been reported that Al-doped Mg2Si exhibits lower carrier mobility than non-doped Mg2Si.6,7,10,16) Additionally, a decreasing trend for carrier mobility is generally caused by the introduction of a secondary phase. From the above considerations, it was concluded that the decrease in carrier mobility was caused by the incorporation of the Al phase as well as the Al-substitution. Unfortunately, we could not quantitatively discuss the effects of the Al phase and the Al-substitution on the thermoelectric properties, electron carrier density, and carrier mobility. To discuss the effect of the Al phase separately, Al-substituted Mg2Si should be first synthesized and then sintered with Al, i.e., the preparation of Al/(Mg, Al)2Si composite materials is required. Our priority is to control the microstructure of the Al phase in the Mg2Si matrix without decreasing the carrier mobility. Therefore, we will optimize the preparation conditions for the Al/Mg2Si composite material to further improve its thermoelectric performance.
We successfully fabricated Al/Mg2Si composite materials with nominal composition xAl/Mg2Si, which was confirmed by phase determination methods using X-ray diffraction measurements and Raman spectroscopy, and by a microstructure observation using a scanning electron microscope. The electrical conductivity and Seebeck coefficient of the xAl/Mg2Si samples increased and decreased, respectively, with increasing x. The changes in electrical conductivity and Seebeck coefficient could be attributed to an increase in the electron carrier density caused by the incorporation of the Al phase in the Mg2Si matrix. However, the carrier mobility of xAl/Mg2Si samples decreased that can be related to the incorporation of the Al phase as well as the Al-substitution in Mg2Si. Thus, the 0.75Al/Mg2Si sample exhibited the highest power factor, PF, of 2.0 × 10−3 W·m−1·K−2 at 800 K. This high power factor was a result of the increased area ratio of the Al phase to the Mg2Si matrix as compared to that of the Mg phase. To enhance thermoelectric performance of the Al/Mg2Si composite material, the microstructure of the Al phase in the Mg2Si matrix will be optimized.
This study was partly supported by a JSPS Grant-in-Aid for Scientific Research (B) (No. 17H03398).
