Effect of Mechanical Deformation on Thermoelectric Properties of p-Type ( Bi 0 . 225 Sb 0 . 775 ) 2 Te 3 Alloys

The effect of mechanical deformation and annealing on thermoelectric properties of p-type (Bi 0.225 Sb 0.775 )Te 3 was performed. The ingots were prepared by melting, followed by quenching method using source materials with compositions of (Bi 0.225 Sb 0.775 ) 2 Te 3 . Rectangular shaped specimens (5×5×12mm) were cut from ingots and then cold-pressed at 700MPa for 2 to 20 times by changing the press direction perpendicular to previous one.The cold-pressed samples have been annealed in a quartz ampoule at 573K.The grain size of the samples was controlled by the number of cold-pressing process and annealing time. Fine grain structure with a grain size of not more than 10μm is obtained in highly deformed samples. The Seebeck coefficient of the deformed samples were gradually increased with annealing and converged to the similar value of about 225 μV/K after 30 hrs.The small grain size in highly deformed sample enables a rapid increase of Seebeck coefficient with annealing time (∼2 hrs.), indicating that the thermal energy needed to recrystallize in highly deformed specimens is lower than that in low deformed specimens. Z values are rapidly increased with annealing time especially in highly deformed alloys, and converge to about 3.0 × 10/K at room temperature. A higher thermoelectric performance could be expected by the optimization of composition and microstructural adjustment. The present study experimentally demonstrates a simple and cost-effectivemethod for fabricating Bi-Te-based alloys with higher thermoelectric performance.


Introduction
Bi 2 Te 3 -related compounds such as solid solutions of Bi 2 Te 3 and Sb 2 Te 3 have been widely studied over the past decades due to their excellent properties for use in thermoelectric cooling and power generation near the room temperature [1][2][3].Currently, one stream of thermoelectric researches follows the improving thermoelectric dimensionless Figure-of-Merit ( =  2 /, where , , , and  are the Seebeck coefficient, electrical conductivity, thermal conductivity, and absolute temperature) by fabricating nanostructures with existing thermoelectric (TE) materials [4][5][6].Thermoelectric nanostructures can be produced by many techniques, such as hydrothermal methods [7], wet chemical reactions [8], and ball milling [2,4].Conventional ball milling has been employed to produce large quantities of fine particles with a size of one to several microns and this process can be readily scaled up for commercial use at reasonable cost.In order to obtain TE materials, the TE particles should be consolidated into a dense solid using various methods, such as spark plasma sintering [5,9], hot-pressing [2,4,7,10], and extrusion methods [11,12].
Generally cold-pressing only mechanically compacts the particles and thus the density of cold-pressed composites tend to be low, resulting in a material with poor mechanical and thermoelectric properties.To improve the mechanical and electrical properties of the composites, the cold-pressing should be followed by sintering at appropriated temperatures.Thus, the techniques used to produce TE nanocomposites can be any combination of a powder preparation method and a composited consolidation methods.As a powder preparation method, high energy ball milling is generally used with source materials of crystalline Bi-Sb-Te alloy ingots which are prepared by zone melting and Bridgman methods [5,9,11,13].
In this paper, we introduce a simple fabrication method of randomly oriented polycrystalline Bi-Sb-Te based alloys which is not employing conventional powdering process.(Bi 0.225 Sb 0.775 ) 2 Te 3 compound was prepared by melting the source elements followed by serial processing with coldpressing process and recovery annealing.The effects of degree of deformation and annealing time on the thermoelectric properties were investigated.Thermoelectric properties are discussed on the point of view of microstructural evolution of cold-pressed sample with annealing.

Experimental Procedures
Elemental bismuth (Bi), antimony (Sb), and tellurium (Te), with purity of 99.99% were used as starting materials for synthesis of (Bi 0.225 Sb 0.775 ) 2 Te 3 compounds.Each of three materials was weighed according to the chemical formula of (Bi 0.225 Sb 0.775 ) 2 Te 3 , loaded into quartz ampoules, and sealed under a vacuum of 10 −6 Torr.Before loading the mixture, the quartz ampoule, whose inner diameter was 11 mm, was chemically etched with 10% HF + 90% HNO 3 and then washed with deionized water to remove any surface impurities.Bi, Sb, and Te were melted and rocked at 1023 K for 4 hrs and then rapidly cooled to room temperature.The cylindrical ingots were then cut into bar-shaped specimen (5 × 5 × 12 mm 3 ) for the measurement of thermoelectric properties and cold-pressing.Cold-pressing was performed by applying the mechanical pressure of 700 MPa using toolsteel die.The press direction was perpendicular to the 5 mm × 5 mm plane of the specimen.After the first pressing step, the specimen was removed from the press die and pressed again with the press direction perpendicular to newly formed 5 mm × 5 mm face of deformed specimen.These cold-press steps were carried out for 2 to 20 times to change the amount of mechanical deformation and then annealed at 573 K up to 30 hrs.
X-ray diffraction patterns of quenched ingot, compacted pallets after pressing and annealing process were obtained using a BRUKER D8 diffractometer.Specimen density was calculated from precise measurements of the specimen's dimension and mass.The electrical resistivity () was measured by the conventional AC four-probe method.The Seebeck coefficient was measured by a method based on the slope of a voltage versus temperature-difference curve.The thermoelectric properties of the same samples were also measured by using commercial equipment (ZEM-2M, Ulvac Inc.) and two sets of measurements are within ±1% of each other.Microstructural analysis was carried out using a scanning electron microscope (Hitachi S-4300) equipped with an Electron Back-Scatter Diffraction (EBSD, Bruker e-Flash) system.The Figure-of-Merit () was determined by the Harman method in vacuum chamber (10 −5 Torr).Four Pt-wires with 40 m in diameter were bonded to both ends of rectangular-shaped sample, two for current flow and the other two for voltage measurement.

Results and Discussion
Figure 1 shows the X-ray diffraction patterns of the asquenched (Bi 0.225 Sb 0.775 ) 2 Te 3 ingot and cold-pressed pallets on the surface perpendicular to the press direction.The peak intensities were normalized by (015) diffraction peaks of the samples.All diffraction peaks were assigned to a rhombohedral Bi 2 Te 3 -Sb 2 Te 3 structure.The presence of strong diffraction peaks in the cold-pressed samples such as (006), (009), (0015), and (0018) indicates the alignment of grains along the [00ℓ] direction, that is, reorientation of the ab planes of the grains took place during the repressing process even at room temperature.Inset of Figure 1 is magnified (015) diffraction peak.The more re-pressing process, the more broadened the peaks are in the patterns suggest that fine grain-sized particles are formed and/or complicated natures of the residual stress and strain remain in pallets.
Figure 2 shows the X-ray diffraction patterns of coldpressed samples taken perpendicular to the direction of press with different annealing times.Solid dots in the figure are the positions of peak heights with standard intensities of powder patterns for rhombohedral (Bi 0.2 Sb 0.8 ) 2 Te 3 phases (JCPDS number 72-1836).In 10 times cold-pressed and 30 hrs annealed sample, the peaks position and intensities exhibited almost the same as that of standard powder patterns, which indicates that the ab oriented grains which are aligned perpendicular to press direction in cold-pressed specimen were randomly rearranged by recrystallization during the annealing process.It was also noted that the diffraction peaks of the annealed samples were narrowed with the increase of annealing time (shown in the inset of Figures 2(a   of cold-pressing and annealing time was carried out.Figure 3 illustrates an EBSD image of the samples which were coldpressed 2 and 10 times followed by annealing for 2 and 30 hrs, respectively.It is clear that the more repetition of pressing process of the sample, the smaller the grain can be found.Dark areas in Figures 3(a  increase of grain size for the 10 times cold-pressed samples.The change of the microstructure during annealing may cause changes in thermoelectric properties such as resistivity and Seebeck coefficient.The relative density of the samples with the number of cold-pressing is shown in Figure 4, and the electrical resistivity with different annealing time is shown in Figure 5. Initially, a pronounced electrical increase is observed in the highly deformed samples.Heat treatment process produces a strong decrease of electrical resistivity especially for highly deformed sample and the values were saturated after 6 hrs of heat treatment.The relative density values were decreased with an increase of the number of cold-pressing process.The annealing of the specimen does not lead to densification in the whole observed region.
In general, when a metal is deformed below its recrystallization temperature, it is said to be "cold-worked." All the properties of a metal that are dependent on the lattice structure are affected by plastic deformation or cold-working [14,15].Cold-pressing process in this work produces very high deformation enough to fracture the ingot into fine particles.This is related to relatively low ductility and mechanical strength of the Bi-Te-based alloy compared to a metal and the presence of cleavage planes.The decrease of relative density in Figure 4 may be attributed to the increase of fine particles with cold-working since large grains have broken into smaller pieces and as they are smaller their porosity constantly increases.
The electrical resistivity in polycrystalline structures can depend on a grain size, porosity, carrier density, and its mobility [16,17].It is well known that the electrical resistivity is determined competitively by carrier concentration and mobility as described by the relationship:  = 1/, where , , and  are the carrier concentration, carrier mobility, and electron charge, respectively.The decrease of electrical resistivity with annealing time, especially in highly deformed samples (10 and 20 times cold-pressed samples in Figure 5, can be explained by an increased electrical mobility because of the grain growth during the annealing process.This would be associated with a decrease in total scattering of charge carriers and annihilation of deformed stress in the sample, leading to enhanced conductivity.
Figure 6 shows the Seebeck coefficient of cold-pressed samples as a function of annealing time.The positive values of Seebeck coefficient indicate that all the samples are p-type materials.All the as-deformed specimens showed the same Seebeck coefficient of about 140 V/K.However, these values were increased with increase of annealing time.Interestingly, the Seebeck coefficient of the samples cold-pressed 10 and 20 times repetition were sharply increased in relatively short annealing periods of about 2 hrs, and saturated to the value of about 220 V/K.The Seebeck coefficient of the samples coldpressed 2 or 5 times repetition increased slowly compared to that of the samples cold-pressed 10 and 20 times.
It is well known that Seebeck coefficient is mostly governed by effective mass and carrier concentrations, and independent of the grain size.In highly deformed samples, it should be pointed out that large amount of dislocations, point defects, and antistructure defects may increase the electrical carriers with very low mobility, and these are responsible for high electrical resistivity.More detailed analysis on exact structure of generated defects according to cold-pressing and how these defects affect thermoelectric properties is beyond the scope of this paper.Accordingly, it is thought that the time needed for recrystallization of crushed particles in the coldpressed Bi-Sb-Te compounds depend on the number of coldpressing process.The enhancement of Seebeck coefficient with annealing can be considered by two aspects: the reduction of carrier concentration with annihilation of residual defects and recrystallization.
Room temperature Figure-of-Merit () of the coldpressed alloys as a function of annealing time are plotted in Figure 7, respectively. values are rapidly increased with annealing time especially in highly deformed alloy.The highest Figure-of-Merit of the sample reaches about 3.0 × 10 −3 /K at 300 K, which is lower than the zone-melted one and that of state-of-art Bi 2 Te 3 -Sb 2 Te 3 -based materials [2,18].The power factors of as-quenched ingot and 10 times deformed sample are plotted as the function of the annealing time in Figure 8.As shown in the figure, the power factors of annealed ingot exhibited similar values to those in deformed samples after annealing.Even though the low Seebeck values are in as-quenched ingot, the low resistivity of the ingot resulted in similar values for power factors with highly deformed samples after annealing.Thus, the improvement of  value in a highly deformed sample is mainly attributed to reduced thermal conductivity.During the annealing process, new grain structure with a submicron size is generated by the recrystallization and grain coarsening.Although the  value is still low, the fabrication method of bulk Bi 2 Te 3based compounds introduced here is more cost effective.It is also much easier compared to other powder-base fabrication techniques for bulk thermoelectric materials and it can be easily scaled up for commercial applications.

Conclusions
The effects of the repetition of cold-pressings followed by recovery annealing on the thermoelectric properties were investigated for (Bi 0.225 Sb 0.775 ) 2 Te 3 alloys.Generally, the plastic deformation increases dislocation density and changes grain size distribution in single and polycrystalline materials.In this experiment, the mechanical pressure of 700 MPa is high enough to fracture the as-quenched ingot into fine particles.Thus, high density of dislocation as well as the reduction of grain size may be expected in our cold-pressed samples.The grain size of cold-pressed samples was controlled by the number of cold-press processing and annealing times.The high electrical resistivity in highly deformed samples is mainly attributed to the increase of grain boundary scattering due to the reduction of grain size and high porosity.The enhancement of the Seebeck coefficient with annealing is mainly related to the recovery and recrystallization of the deformed samples.The recrystallization process can be obtained by diffusion process or annihilation of generated defect.The time needed for recrystallization with annealing depend on the grain size and porosity; the smaller the grain size, the shorter the diffusion distances.The fast saturation of Seebeck coefficient in highly deformed samples with annealing indicates that the Seebeck coefficient is independent of the grain size.However, further decrease in electrical resistivity

Figure 1 :
Figure 1: X-ray diffraction patterns of as-quenched ingot and (Bi 0.225 Sb 0.775 ) 2 Te 3 alloys with different number of cold pressing.

Figure 2 :
Figure1shows the X-ray diffraction patterns of the asquenched (Bi 0.225 Sb 0.775 ) 2 Te 3 ingot and cold-pressed pallets on the surface perpendicular to the press direction.The peak intensities were normalized by (015) diffraction peaks of the samples.All diffraction peaks were assigned to a rhombohedral Bi 2 Te 3 -Sb 2 Te 3 structure.The presence of strong diffraction peaks in the cold-pressed samples such as (006), (009), (0015), and (0018) indicates the alignment of grains along the [00ℓ] direction, that is, reorientation of the ab planes of the grains took place during the repressing process even at room temperature.Inset of Figure1is magnified (015) diffraction peak.The more re-pressing process, the more broadened the peaks are in the patterns suggest that fine grain-sized particles are formed and/or complicated natures of the residual stress and strain remain in pallets.Figure2shows the X-ray diffraction patterns of coldpressed samples taken perpendicular to the direction of press with different annealing times.Solid dots in the figure are the positions of peak heights with standard intensities of powder patterns for rhombohedral (Bi 0.2 Sb 0.8 ) 2 Te 3 phases (JCPDS number 72-1836).In 10 times cold-pressed and 30 hrs annealed sample, the peaks position and intensities exhibited almost the same as that of standard powder patterns, which indicates that the ab oriented grains which are aligned perpendicular to press direction in cold-pressed specimen were randomly rearranged by recrystallization during the annealing process.It was also noted that the diffraction peaks of the annealed samples were narrowed with the increase of annealing time (shown in the inset of Figures2(a) and 2(b)), supporting that the grain growth took placed during the heat treatment.To obtain detailed information on the microstructures, especially the grain-boundary characteristics, Electron Back-Scatter Diffraction (EBSD) analysis according to the number
) and 3(b) are due to the lattice distortions, such as dislocations and vacancies.The image in Figure3(a) exhibits unclear grain structure and further annealing leads to grain structure with the size of a few tens of micrometers (Figure3(b)).The decrease of the dark areas especially inside the 2 times cold-pressed grains means the recovery happened during the annealing process.However, it is easily recognized from Figures3(c) and 3(d) that the annealing led to dynamic recrystallization significant 4

Figure 4 :Figure 5 :
Figure 4: Change of relative density for (Bi 0.225 Sb 0.775 ) 2 Te 3 alloys with the number of cold-pressing.

Figure 6 :
Figure 6: Change of the Seebeck coefficient of (Bi 0.225 Sb 0.775 ) 2 Te 3 alloys with the number of cold-pressing and annealing time at 573 K.

Figure 8 :
Figure 8: Change of the power factor of (Bi 0.225 Sb 0.775 ) 2 Te 3 alloys of as-quenched ingot and the 10 times cold-pressing with annealing time at 573 K.