Scalable silicon nanostructuring for thermoelectric applications

E. Koukharenko, S. A. Boden, D. Platzek, D. M. Bagnall, N. M. White

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)

Abstract

The current limitations of commercially available thermoelectric (TE) generators include their incompatibility with human body applications due to the toxicity of commonly used alloys and possible future shortage of raw materials (Bi-Sb-Te and Se). In this respect, exploiting silicon as an environmentally friendly candidate for thermoelectric applications is a promising alternative since it is an abundant, ecofriendly semiconductor for which there already exists an infrastructure for low-cost and high-yield processing. Contrary to the existing approaches, where n/p-legs were either heavily doped to an optimal carrier concentration of 1019 cm-3 or morphologically modified by increasing their roughness, in this work improved thermoelectric performance was achieved in smooth silicon nanostructures with low doping concentration (1.5 × 1015 cm-3). Scalable, highly reproducible e-beam lithographies, which are compatible with nanoimprint and followed by deep reactive-ion etching (DRIE), were employed to produce arrays of regularly spaced nanopillars of 400 nm height with diameters varying from 140 nm to 300 nm. A potential Seebeck microprobe (PSM) was used to measure the Seebeck coefficients of such nanostructures. This resulted in values ranging from -75 μV/K to -120 μV/K for n-type and 100 μV/K to 140 μV/K for p-type, which are significant improvements over previously reported data.

Original languageEnglish
Pages (from-to)2114-2118
Number of pages5
JournalJournal of Electronic Materials
Volume42
Issue number7
DOIs
Publication statusPublished - Jul 2013
Externally publishedYes

Keywords

  • e-beam lithography
  • low doping concentrations
  • nanostructures
  • silicon

Fingerprint Dive into the research topics of 'Scalable silicon nanostructuring for thermoelectric applications'. Together they form a unique fingerprint.

Cite this