An innovative optical approach to measuring thermal conductivity may eventually assist in the development of next-generation, high-temperature turbine blades based on niobium alloys, which have been proposed as a replacement for currently used nickel-based materials.
David Cahill and colleagues at the University of Illinois (Urbana-Champaign, IL) as well as at GE Global Research (Schenectady, NY) are using time-domain thermoreflectance (TDTR) to produce time-resolved imagery of thermal transport in a cross section of a material sample containing niobium, silicon, and titanium. The procedure involves taking individual measurements at thousands of locations across a sample and mapping the data distribution with micron-scale spatial resolution.1
"Viewing the image, we were surprised to find significant variations in the thermal conductivity of individual crystal grains," Cahill said (see figure). "The grains were randomly oriented in the sample, and how effectively each grain conducted heat depended on its orientation." Traditional approaches to measuring thermal transport have depended upon contact methods based on atomic-force microscopy (scanning thermal microscopy) and noncontact methods based on temperature-induced changes in optical reflectivity (modulated thermoreflectance microscopy).
The need for calibration of cantilever response along with sensitivity to sample morphology and heat flow between cantilever and sample inhibit the acquisition of quantitative data using the former method. And prior implementations of thermoreflectance microscopy have lacked the desired spatial resolution and speed because the probe beam must be scanned laterally relative to the focused laser that heats the sample.
The TDTR approach developed at the University of Illinois relies on a single measurement of thermoreflectance at each spatial location with a fixed delay time, on the order of 100 ps. Lateral heat flow is limited by modulating the pump beam at about 10 MHz, and artifacts created by scattering of the pump beam due to surface roughness are eliminated by modulating the probe beam at audio frequencies. They have achieved a lateral spatial resolution on the order of 3.4 µm and measurement of 100 × 100-pixel images for time scales on the order of an hour.
Next, the researchers plan to image the cross section of an actual turbine blade. "The image should reveal how the thermal properties of the metal, bond coat, and thermal-barrier coating vary by position as we go through the structure," Cahill said. "This information may help improve the thermal performance and operating efficiency of future turbine blades."
REFERENCE
- S. Huxtable et al., Nature Materials, in press.