University of Limerick
Browse

On the development of a contactless thermal characterisation technique for micro-scale thermoelectric modules (μTEMs)

Download (7.64 MB)
thesis
posted on 2022-08-25, 08:28 authored by Séamus Hickey
Contemporary internet consumer usage, in the form of social media and wide scale video streaming, has induced an exponential rise in the demand for high speed data. Optical communications infrastructure has had to evolve at a rapid pace to meet the appetite for data. Photonic Integrated Circuits (PICs) are critical components of optical communications equipment that transmit and receive coded light signals of specified wavelengths to transfer high volumes of data over optical fibres. Wavelength is sensitive to thermal fluctuations, however, with variations of as low as ± 0.1°C shifting the wavelength of the encoded signal outside design specifications. Conventional macro-scale thermoelectric modules (TEMs) are currently employed to maintain tight thermal control of PICs, but shrinking device footprints and the resultant higher heat fluxes are driving the need for smaller, micro-scale TEMs. Determining the thermal characteristics (temperature difference across the TEM, ΔT, and heat pumped by the TEM, Qc, for a range of electrical currents through the TEM, I) of these micro TEMs (μTEMs) has emerged as a challenge, however, primarily due to their poor compressive strength (~200 MPa). Conventional characterisation techniques typically use a heat stack configuration, which involves compressing the TEM between a controlled heat source and sink combination to measure ΔT and Qc with minimum losses at the thermal interfaces. The objective of this thesis is to design, commission and demonstrate a novel contactless apparatus to thermally characterise a μTEM (ΔT ~ 20 K, Qc ~ 0.3 W, I ~ 1 A) in a compressionfree fashion. Compressive forces on the upper surface of the thermoelectric devices were obviated by using an infra-red (IR) source to apply a heat load to the upper surface of the TEMs, and an IR sensor was used to measure the upper surface temperature. A calorimeter was used to control the temperature of the lower surface of the TEMs as a constant reference and to determine Qc. Measures were implemented to minimise errors due to reflected radiation within the setup, and an extensive calibration was undertaken on all measurements to minimise uncertainty. The contactless apparatus was benchmarked against a high precision conventional compression apparatus using a macro scale TEM (15 mm x 15 mm x 3 mm) in order to validate the methodology. Then, an array of twelve μTEMs on an aluminium nitride substrate (15 mm x 15 mm x 0.6 mm) was manufactured (4 x 3) in order to produce sufficient heat flow for accurate measurement. The array was characterised in both apparatuses, allowing the thermal characteristics of a single μTEM (0.83 mm x 2.14 mm 1.05 mm) to be isolated from the performance data for the array. The contactless characterisation technique produced values for Qc within 15 – 357 mW (1 – 25.5%) and values for ΔT within 0.4 – 6.2 K (0.5 – 7.6%) of the conventional characterisation apparatus for the macro-scale TEM. The characteristics extracted for a single μTEM measured within 15 – 100 mW (2.5 – 15%) of the conventional characterisation apparatus for Qc, while the values for ΔT were within 0.5 – 1.9 K (1.5 – 6.5%). The thermoelectric figure of merit ZT (0.292), calculated from the Seebeck coefficient (0.0147 V/K), conductance (0.07782 W/K) and resistance (2.83 W) of the μTEM, was within ± 5.2% of the conventional compression method (0.308). It was concluded that the novel contactless characterisation method developed in this thesis could be used to accurately characterise the thermal performance of micro-scale thermoelectric devices in a compression-free manner.

History

Degree

  • Doctoral

First supervisor

Punch, Jeff

Note

peer-reviewed

Language

English

Usage metrics

    University of Limerick Theses

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC