Thesis advisor: Zhifeng Ren / Thermoelectric (TE) technology is an environment-friendly one due to reduction of carbon emission, which can be widely used either for power generation or for refrigeration. Basically applications of TEs are based on TE effects, which involve the transition between heat and electricity. Despite the superior advantages of being solid state and providing a clean form of energy, TE technology so far only finds its niche area of application due to the relatively less efficiency compared to traditional methods. The efficiency of a thermoelectric device is solely determined by the dimensionless figure-of-merit (ZT) of thermoelectric materials. According to the definition, ZT is equal to square of Seebeck coefficient times electrical conductivity times absolute temperature divided by thermal conductivity. Therefore, a good thermoelectric material should possess high Seebeck coefficient and electrical conductivity while low thermal conductivity, so called phonon glass electron crystal (PGEC). In bulk materials, it is challenging to further improve ZT or independently vary individual parameters without affecting others, mainly due to the interrelated relationships among these three parameters. Fortunately, nano approach gives us some independent control in parameters adjustment. One important aspect of nano idea lies in the fact that enhanced boundary scattering due to the increased intensities of interfaces arising from nano-sized grains could reduce the thermal conductivity more than the electrical conductivity, which is practically realized in our material system. Since the introduction of nano idea, large ZT as high as above two has been achieved in the superlattice system. Due to the high fabrication cost of superlattices, they are not scalable for mass production. Theoretical calculations indicate that thermal boundary resistance is the main mechanism for the low thermal conductivity in superlattices, rather than the periodicity. Basically, we hope to achieve the supplattice-like ZT in the less costly bulk nanograined materials, based on the idea that reduction of thermal conductivity which is responsible for ZT enhancement in superlattices can be realized in bulk materials with embedded nanostructures as well. Inspired by the nanocomposite idea, in my thesis work I applied the technique of ball milling and then hot press to various thermoelectric materials, from low temperature to high temperature, demonstrating the feasibility of the approach. By ball milling alloyed ingot into nanopowders and DC hot pressing them, we have achieved a 62-89% ZT improvement for p-type half-Heusler samples, mainly due to the significantly enhanced Seebeck coefficient and partially due to the moderately reduced thermal conductivity. Microstructure studies indicated that increased boundaries due to smaller nano-sized grains is the cause for change of parameters. For our ball milled samples, the trend of decreasing thermal conductivities with increasing ball milling time is observed, further substantiating our nano-approach idea because longer ball milling time gives rise to smaller grain sizes and thus stronger boundary scattering. By applying the same technique to n-type half-Heuslers, we also successfully obtained pronounced enhancement in ZT especially at medium and low temperature ranges, which might be useful in medium temperature power generation. By ball milling a mixture of individual constituent elements into alloyed nanopowders and then DC hot pressing them, we did not gain improvement in ZT initially for n-type BiTeSe system mainly due to the simultaneous reduced power factor with the thermal conductivity. Considering anisotropic properties of the n-type BiTeSe single crystal and randomization effect of ball milling process, we repressed the as-pressed bulk samples in a bigger diameter die, during which lateral flow took place, resulting in preferred grain orientation. As a result, a 22% improvement in the peak ZT from 0.85 to 1.04 at 125 oC in n-type Bi<sub>2</sub>Te<sub>2.7</sub>Se<sub>0.3</sub> has been successfully achieved, arising from the more enhanced power factor than the thermal conductivity. Compared with single crystal, we benefit from the small nano-sized grains in bulk materials. Taking into account the in-plane power factor of single crystal, we still have much room for further ZT improvement if more ab orientation is promoted into the disk plane and/or the crystal plate size and thickness are reduced. By applying our technique of ball milling and then hot press to p-type skutterudites system, we have achieved a peak ZT of 0.95 at 450 <super>o</super>C in NdFe<sub>3.5</sub>Co<sub>0.5</sub>Sb<sub>12</sub>, which is comparable to that of the state-of-the-art ingot. Our approach has the advantage of being less costly and more time-efficient compared to traditional fabrication methods. Besides, even lower thermal conductivity and hence higher ZT can be expected, provided that the nanosize of the precursor powder is preserved during hot press. The nanocomposite idea has been substantiated and the feasibility and generality of our ball milling and then hot press approach has been demonstrated, based on the thermoelectric properties data we obtained and the microstructure studies we carried out from various thermoelectric material systems, from low temperature to high temperature. We believe that continued effort in the area of thermoelectrics by our approach should be paid with superlattice-like ZT if ingenious methods are devised to control the grain growth during consolidation. / Thesis (PhD) — Boston College, 2010. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Physics.
| Identifer | oai:union.ndltd.org:BOSTON/oai:dlib.bc.edu:bc-ir_101357 |
| Date | January 2010 |
| Creators | Yan, Xiao |
| Publisher | Boston College |
| Source Sets | Boston College |
| Language | English |
| Detected Language | English |
| Type | Text, thesis |
| Format | electronic, application/pdf |
| Rights | Copyright is held by the author, with all rights reserved, unless otherwise noted. |
Page generated in 0.0025 seconds