A. Heinrich

Thermoelectric properties of β-FeSi2 single crystals and polycrystalline β-FeSi2+x thin films

The resistivity and thermoelectric power of β-FeSi2 single crystals and polycrystalline β-FeSi2+x thin films doped with 3d transition metals, have been investigated for dependence on the deviation from strict stoichiometry and on the purity of the source material. It has been found that in single crystals, the thermoelectric power is characterized by a large phonon drag effect which is more pronounced in samples grown with the source material of higher purity, and by large absolute values of >500 μV/K in a broad temperature range. The values of the resistivity prepared at the Si-rich and Fe-rich phase boundaries do not show any correlation with the deviation from strict stoichiometry, but its temperature dependence ρ(T) is controlled by native defects. Polycrystalline β-FeSi2+x thin films doped with Co and Cr, and prepared by electron beam evaporation and magnetron sputtering, were found to have maximum power factors outside the homogeneity region at approximately x=0.15±0.05. Their thermoelectric power remains below that of single crystals.

Nanodisperse CrSi(O, N) thin films—conductivity, thermopower and applications

Abstract The electrical conductivity σ(T) and thermopower S(T) of CrSiO and CrSiN thin films were investigated as a function of their microstructure for different ratios [Si]/[Cr] > 1 and concentrations co and cN up to 50 at.%. The films were prepared by reactive d.c. sputtering of CrSi targets and non-reactive r.f.-diode sputtering of CrSi/SiO2 targets respectively. The as-deposited amorphous structure a-CrSiOx(SiNy) is changed by stepwise annealing into a nanodisperse structure nc-CrnSimSiOx(SiNy). The temperature dependence σ(T), investigated between 2K and 400 K. Shows a metal-insulator transition with increasing co(cN). In the amorphous state σ(T) is described by weak and strong localization, analogous to doped amorphous semiconductors. In the nanodisperse state a temperature coefficient of resistivity near zero was obtained within the temperature range −50–120°C up to 5 × 104microOhmcm and in a smaller range around 27°C up to 2 × 105microOhmcm. These properties are used in high ohmic and precision thin film resistors. Results are also given for nc-CrnSimSiAlOx thin films. The thermopower of the nanodisperse films is found to be essentially determined by the metallic components. With increasing co (cN) these components are CrSi2, Cr5Si3 and Cr3Si with S(300 K) of 80 microV K−1, 5 microV K−1 and 17 microV K−1, respectively. As a consequence, in the film values between 90 microV K−1 and 10 microV K−1 were obtained. The temperature dependence S(T) above 300 K is reversible at least up to Tann. These results also make the nanodisperse films very suitable for application in thermoelectric sensors.

Electrical transport properties of high resistance CrSiO thin films

Abstract The d.c. conductivity and thermoelectric power of CrSiO thin films were investigated in the temperature region 4–300 K. The samples have a constant silicon-to-chromium ratio of 74 26 but various oxygen contents co. A metal-semiconductor transition which depends on co is observed. As-deposited film samples annealed at high temperatures are also considered. The results are discussed with respect to the structure of the films and are compared with results for the binary system Si1−xCrx.

Single crystal growth of non-stoichiometric β-FeSi2 by chemical transport reaction

Abstract High-purity single crystals are necessary to investigate the intrinsic properties of undoped β-FeSi2 which is dependent on the composition within the homogeneity range of the phase. Therefore, iron was used as the initial material for growing β-FeSi2 single crystals with a high-purity with respect to metallic as well as non-metallic impurities. Chemical vapor transport in a closed system was used for single crystal growth. By optimization of the whole preparation process a final purity of approximately 99.996% by weight could be achieved. The content of the main electrically active elements was lower than 20 ppm. By optimizing the transport process untwinned single crystals with flat surfaces could be obtained. To fix the composition of the crystals at the lower and upper phase boundary FeSi/FeSi2 and FeSi2/Si sources were used only and the crystals were heat-equilibrated at 700, 825, 925 and 975°C for different times. Only n-type single crystals were obtained even in both equilibria, with FeSi and with Si, respectively. Therefore, the p-type conductivity of undoped single crystals reported in the literature result from non-intentional doping by the impurity level of the used source material. The single crystals grown at both phase boundaries are expected to differ in the concentration of native defects and with that in the electrical properties. Four point measurements have shown a clear relation between the temperature dependence of the resistivity and the annealing temperature. However, different values of the resistivity at room temperature of crystals at the lower and upper phase boundary were only found in 975°C annealed crystals. In no case were low ohmic single crystals obtained.

Phase formation process of IrxSi1−x thin films Structure and electrical properties

Experimental data on the phase formation process of amorphous IrxSi1−x thin films with 0.30 ≤ x ≤ 0.41 are presented and discussed in relation to electric transport properties. The structure formation process at temperatures from 300 K up to 1223 K was investigated by means of X-ray diffraction. Distinct phases were observed in the final stage in dependence on the initial composition: Ir3Si4, Ir3Si5, and IrSi3. An unknown metastable phase was found in films with a silicon concentration of 61 at.% to 64 at.% after annealing above the crystallization temperature T = 970 K. The crystal structure of this phase was determined by the combined use of X-ray diffraction and electron diffraction. It was found to be monoclinic, basic-face centred with lattice constants a = 1.027 nm, b = 0.796 nm, c = 0.609 nm, and γ = 113.7°. Additionally, microstructure and morphology of the films were investigated by transmission electron microscopy (TEM). The annealing process was studied by means of mechanical stress investigations as well as by electrical resistivity and thermopower measurements. Correlations between the structure, the phase formation and the electrical transport behaviour are discussed on the basis of conduction mechanism.

Thermopower and electrical resistivity of undoped and co-doped /spl beta/-FeSi 2 single crystals and /spl beta/-FeSi 2+x thin films

Thermopower and electrical resistivity have been investigated of /spl beta/-FeSi/sub 2/ single crystals and /spl beta/-FeSi/sub 2+x/ thin films in the temperature range from 40 K to 1200 K. The single crystals were grown by chemical transport reaction with source material of different purity and with a Si/Fe ratio between 1.5 and 2.5 to obtain crystals from the Fe-rich and Si-rich boundaries of the homogeneity range, respectively. The undoped and co-doped thin films have been prepared in the range -0.06<x<0.25 by coevaporation. It is shown that the conductivity type of the single crystals changes from p-type to n-type with increasing purity of the starting material. Both thermopower and resistivity depend on the deviation from strict stoichiometry. The low temperature thermopower is mainly determined by phonon drag effect. The undoped films are of p-type. It is shown that the films are especially sensitive to doping at x/spl ap/0.13.

Effect of doping on the thermoelectric properties of iridium silicide thin films

Starting from the phase formation process analysed at the binary Ir-Si system the structure formation process and the thermoelectric properties of Fe doped Ir/sub 3/Si/sub 5/ thin films have been investigated. The films were prepared in the vicinity of the stoichiometric chemical composition, Ir/sub 3/Si/sub 5/, by means of different physical vapour deposition techniques particularly magnetron co-sputtering and electron beam co-evaporation. The amount of Fe dopant was varied between 0 and 2.5 at.%. For analysis of doping level and impurity concentration the chemical composition of the as-deposited films was investigated by means of Rutherford backscattering spectroscopy (RBS), energy dispersive X-ray analysis (EDX), and spark source mass spectrometry (SS-MS). The annealing process was studied in-situ by means of high temperature X-ray diffraction (HT-XRD) as well as by measurements of the electrical resistivity /spl rho/ and the thermopower S. The phase formation process depends very sensitively on the volume fractions of the major components Ir and Si, whereas the small concentrations of dopant did not change the sequence of formed crystalline phases. On the other hand, the thermoelectric transport properties correlate strongly with both, the structure formation process and the chemical composition of the films.

High temperature thermoelectric properties of doped iridium silicide thin films

Iridium silicide thin films were prepared as undoped, Fe-doped and Ni-doped material by means of magnetron sputtering and electron beam evaporation. In the as-deposited state the structure of the films was amorphous. Subsequent annealing of binary and of iron-doped films results in crystallization and phase formation passing a metastable state. By contrast the crystallization process of Ni-containing films was characterized by the formation of residual nickel disilicide (/spl delta/-Ni/sub 2/Si) and nickel trisilicide (/spl beta//sub 1/-Ni/sub 3/Si) The electrical resistivity and the thermopower of the films were measured simultaneously during first annealing process at temperatures from 300 K up to 1200 K showing sensitive dependence on the chemical composition of the films and correlations with the phase formation process. In the final state the iron doped films show rather large thermoelectric power factors with maximum values at temperatures >1200 K as a result of the large gap of Ir/sub 3/Si/sub 5/ (1.56 eV). Estimation of thermoelectric efficiency using thermal conductivity of Ir/sub 3/Si/sub 5/ single crystals results in high values of figure of merit at temperatures >1000 K comparable with the best one observed for /spl beta/-FeSi/sub 2/ and MnSi/sub 1.75/. Ni-doped films with a small concentration of nickel silicides showed n-type conduction.

Thermoelectric properties of Re-Si nanocrystalline composites

In this report the results on the phase formation and transport properties of amorphous and nanocrystalline Re-Si thin film composites are presented. We use the composites as a model material to investigate the effect of a reduced grain size and inter-grain interfaces on the thermoelectric properties. We found that, both, the phase composition and the crystalline grain size appear to be important parameters for thermoelectric material optimisation. The results suggest that the thermoelectric parameters of a nanocrystalline composite can be superior to those of the conventional poly- or single-crystalline material of the same composition.

Experimental set-up for thermopower and resistivity measurements at 100-1300 K

In this paper we describe an experimental set-up for simultaneous measurements of thermopower and electrical resistivity at temperature from 100 K to 1300 K. Optimal configuration of electrodes and mechanical contacts of the thermocouples, current leads and potential probes with sample make it possible to measure a large variety of materials and result in a flexibility with respect to the sample form and dimensions. Both, bulk and thin film samples can be investigated. The measurements can be done in vacuum or in an inert gas atmosphere. Samples with resistance varying from 0 /spl Omega/ up to 200 k/spl Omega/ (1 G/spl Omega/ in case of the resistivity measurement) can be measured. High precision and high reliability of the system have been proved during more than two years use. The resistivity and thermopower of pure Ni, CeB/sub 6/ heavy fermion compound and a Cr-Si thin film composite are presented as a test material to demonstrate the possibilities and accuracy of this experimental set-up.