Carnegie Mellon Logo

Zoey Su — 2011-12 Fellow

Zoey Su photo

Since over 20% of the electricity in US is consumed by lighting. Herein I propose to study heat generation and dissipation in nitride semiconductors for solid-state lighting and power electronics. Heat generation and removal in LEDs has come to the forefront as high operating temperatures degrade efficiency, shift the emission spectrum, and reduce the lifetime of GaN and InGaN LEDs. For every 10 degrees increase in operating temperature, the lifetime is approximately halved. While package-level thermal management strategies have been called upon to mitigate these concerns, substantial thermal resistance exists within the device itself due to imperfections inherent to state-of-the-art growth techniques (e.g., dislocations, vacancy clusters, and grain boundaries). Thermal transport is further complicated by the thin film device structure, which contains nucleation layers, buffer layers, and superlattices, each having interface spacing comparable to the mean free path of energy carriers (phonons and electrons).

I will study the nature of (i) thermal transport and (ii) heat generation in blue and green LEDs using the Frequency Domain Thermoreflectance (FDTR) technique that I have developed with my advisor, J. Malen, during my first and second years at CMU. Our collaborator, Robert Davis (MSE), as well as our industrial collaborators, Kyma Technologies and Cree, will supply LED samples.

(i) Thermal Transport. I will expand my study of thermal transport in GaN and InGaN LEDs through measurement of their thermal conductivity (k). A frequency domain thermoreflectance technique (FDTR) will be used to measure thermal conductivity with high accuracy for a range of samples. FDTR uses two lasers with different wavelengths to test the sample’s thermal properties. A 488nm pump laser is modulated with an electro-optic modulator (EOM) and focused onto the sample. A collinear 532nm probe laser is used to sense the temperature change on the sample surface through the temperature dependent optical properties. This signal is compared with an analytical model to determine the unknown thermal properties of the sample. LEDs at various points in the growth process will be measured to isolate the thermal conductivity of the various layers.

(ii) Heat Generation. The heat generation mechanism in GaN and InGaN LEDs still remains unclear. Researchers usually measure the light output to study efficiency, while my plan is to measure the heat output. Images taken with an IR camera indicate that the surface temperature of LED is inversely proportional to efficiency. Inefficiencies that generate heat including Shockley-Read-Hall and Auger non-radiative recombination, as well as joule heating, have different dependency on the drive current in an operating LED (e.g., Joule heating ~I2). We will exploit this fact, paired with surface temperature measurements using thermoreflectance, to determine the relative contributions of each factor.

Solid-state light emitting diodes (SSL-LEDs) hold considerable potential to be a source of artificial light that is more efficient and effective than existing technologies. At present, red LEDs surpass current lighting techniques in overall efficiency. The need for artificial white light requires that LEDs generate light at shorter wavelength. Blue and green LEDs, typically made from Gallium Nitride (GaN) and Indium Gallium Nitride (InGaN) superlattices, are being developed to meet these requirements. Alloys and multi-layer structures of nitrides lead to high electronic mobility, a continuum of accessible band gaps, and strain-sensitive band structures, all of which have enabled these technologies. While nitride-based devices have the potential to revolutionize lighting and to outperform silicon-based electronics, their thermal properties play an important role in improving the efficiency and lifetime. My research will help people understand and better design LED structures.