Nenad Stojanovic

Thin-Film Thermal Conductivity Measurement Using Microelectrothermal Test Structures and Finite-Element-Model-Based Data Analysis

We present a new method for measuring thermal conductivities of films with nanoscale thickness. The method combines a micro electrothermal test structure with a finite-element- based data analysis procedure. The test device consists of two serpentine nickel structures, which serve as resistive heaters and resistance temperature detectors, on top of the sample. The sample is supported by a silicon nitride membrane. Analytical solution of the heat flow is infeasible, making interpretation of the data difficult. To address this, we use a finite-element model of the test structure and apply nonlinear least-squares estimation to extract the desired material parameter values. The approach permits simultaneous extraction of multiple parameters. We demonstrate our technique by simultaneously obtaining the thermal conductivity of a 280 mum x 80 mum x 140 nm thick aluminum sample and the 360 mum x 160 mum x 180 nm thick silicon nitride support membrane. The thermal conductivity measured for the silicon nitride thin film is 2.1 W/mK, which is in agreement with reported values for films of this thickness. The thermal conductivity of the Al thin film is found to be 94 W/mK, which is significantly lower than reported bulk values and consistent both with reported trends for thin metallic films and with values that were obtained using electrical resistivity measurements and the Wiedemann-Franz law.

Direct measurement of thermal conductivity of aluminum nanowires

A nanofabricated electrothermal test structure is reported for directly measuring the thermal conductivity of aluminum nanowires near room temperature. Interdigitated nanowires perturb an otherwise symmetric heater-sensor structure analogous to an electrical bridge circuit. Nanowires studied are 100 nm thick with 75, 100, and 150 nm widths. Finite element simulation accounts for complex device geometry. Thermal conductivity and electrical resistivity vary significantly with nanowire dimensions. Electron transport equation models which adequately describe the resistivity data consistently underestimate the thermal conductivity. Incorporating a phonon contribution of ∼21 W/m K to the total thermal conductivity is found to accurately describe the measured values.

Effect of impurity incorporation on emission wavelength in cathodoluminescence spectrum image study of GaN pyramids grown by selective area epitaxy

Metal-organic chemical vapor deposition has been used for selective epitaxy of GaN pyramids ranging in size from over 1μm to 500 nm in length at the base. Pyramids are terminated by {11¯01} crystal facets. The optical properties of the pyramids are investigated by cathodoluminescence (CL) in a scanning electron microscope. CL spectrum imaging reveals the pyramid apices to emit light at ∼363.4 nm corresponding to the emission wavelength of relaxed GaN. As the CL excitation is moved away from the apex a systematic redshift is observed. The redshift is ∼4 nm for pyramids with 3 μm base dimension and ∼2 nm for the 500 nm pyramids. The shift is attributed to diffusion of silicon and oxygen into the GaN pyramids due to SiO2 mask decomposition with negligible contribution from stress. The observations are backed by finite element simulations of diffusion and stress.

Diverse facets of InGaN quantum well microrings grown by selective area epitaxy

InGaN quantum well (QW) microrings were grown using selective area epitaxy on patterned (0001) AlN/sapphire. The well defined shapes are comprised of {112¯2} and {213¯3} facets on inner sidewalls, and {11¯01} facets on outer sidewalls, as well as (0001) top surfaces. The sidewall facets exhibit distinct emission spectra when investigated using cathodoluminescence. Differences in emission spectra are attributed to variations in growth rate and indium incorporation on the facets, with peak emission wavelength as long as 500 nm. The observed weak blueshift with increasing cathodoluminescence excitation current verifies that the internal piezoelectric fields of the semipolar sidewall facets are lower than reference (0001) InGaN QWs.

A Microelectrothermal Bridge Circuit With Complementary Parameter Estimation Algorithm for Direct Measurement of Thermal Conductivity

We describe a novel microelectrothermal test structure and a complementary data analysis algorithm for direct measurement of the thermal conductivity of metallic thin films and nanowires. The device is a thermal analog of an electrical bridge circuit, such as the Wheatstone bridge, as is commonly used to measure electrical impedance. The microelectrothermal bridge circuit addresses the problem of parasitic heat loss to supporting structures-a major obstacle to direct measurement of thermal properties. A nonlinear least-squares parameter extraction method developed specifically for the thermal bridge circuit further reduces the influence of unmodeled heat paths and accounts for complex sample geometry. Thus, accurate measurements may be made when fabrication of suspended structures is difficult or undesirable. The technique is demonstrated on arrays of aluminum nanowires on glass substrates, with excellent results. The Al nanowire experiments are also used to show the improved performance of the parameter extraction algorithm over a standard approach.

Least-squares parameter estimation algorithm for a microelectrothermal bridge circuit

A least-squares parameter estimation algorithm is designed to extract the thermal conductivity of metallic nanowires using a nanofabricated microelectrothermal test structure. The device is a thermal analog of a bridge circuit, such as is commonly used to measure electrical impedance. A resistive heater is positioned symmetrically between two temperature sensors. A nanowire array extends from the heater to one sensor, unbalancing the bridge temperatures. The least-squares cost function exploits the underlying symmetry of the thermal bridge circuit, reducing sensitivity to parasitic thermal conductances and other uncertain parameters. The algorithm is demonstrated on a series of aluminum nanowire arrays with varying nanowire widths. The results are shown to be consistent with predictions made using the electron thermal transport analogy.

Microelectrothermal Bridge Circuits for Thermal Conductivity Measurement of Metallic and Semiconducting Nanowires

We develop a method for direct measurement of thermal conductivity of nanowires, consisting of a microelectrothermal test device and a complementary parameter estimation algorithm. Simulations of a simplified version of the problem show how differential measurements can address the problem of parasitic heat loss, and examine several different parameter estimation schemes. As reported elsewhere, measurements have been performed on aluminum nanowire arrays, with excellent results. Several design modifications are required to accommodate semiconducting samples. A device design for silicon nanowire arrays is presented. A simulation study suggests that these devices will also perform extremely well.Copyright

Direct Study of the Thermal Conductivity in Aluminum Nanowires

Thermal conductivity and electrical resistivity of 1 μm long aluminum nanowires, 75, 100, and 150nm in width and 100nm thick, were measured at room temperature. The method consists of microfabricated electrothermal test devices and a model-based data processing approach using finite-element analysis (FEA). The electrical and thermal properties of the nanowires differ significantly from bulk values while electrical resistivity agrees well with theoretical prediction. Electron transport equation models, which adequately describe the resistivity data, consistently underestimate the thermal conductivity. Incorporating a phonon contribution of ˜ 21 W/m·K to the total thermal conductivity is found to accurately describe the measured values.

Model-Based Data Analysis for Thin-Film Thermal Conductivity Measurement Using Microelectrothermal Test Structures

We present a new method for measuring thermal conductivities of films with nanoscale thickness. The method combines a microelectrothermal test structure with a finite-element-based data analysis procedure. The test device consists of two serpentine nickel structures, which serve as resistive heaters and resistance temperature detectors, on top of the sample. The sample is supported by a silicon nitride membrane. Analytical solution of the heat flow is infeasible, making interpretation of the data difficult. To address this, we use a finite-element model of the test structure, and apply nonlinear least-squares estimation to extract the desired material parameter values. The approach permits simultaneous extraction of multiple parameters. We demonstrate our technique by simultaneously obtaining the thermal conductivity of a 280 μm by 80 μm by 140 nm thick aluminum sample and the 360 μm by 160 μm by 180 nm thick silicon nitride support membrane. The thermal conductivity measured for the silicon nitride thin film is 2.1 W/mK, in agreement with reported values for films of this thickness. The thermal conductivity of the Al thin film is found to be 94 W/mK—significantly lower than reported bulk values, and consistent both with reported trends for thin metallic films and with values obtained using electrical resistivity measurements and the Wiedemann-Franz law.Copyright