An investigation of heat transfer between a microcantilever and a substrate for improved thermal topography imaging

This paper reports the numerical and experimental investigation of heat transfer from a heated microcantilever to a substrate and uses the resulting insights to improve thermal topography imaging. The cantilever sensitivity, defined as change in thermal signal due to changes in the topography height, is relatively constant for feature heights in the range 100-350 nm. Since the cantilever-substrate heat transfer is governed by thermal conduction through the air, the cantilever sensitivity is nearly constant across substrates of varying thermal conductivity. Surface features with lateral size larger than 2.5 μm can induce artifacts in the cantilever signal resulting in measurement errors as large as 28%. These artifacts arise from thermal conduction from the cantilever in the lateral direction, parallel to the surface. We show how these artifacts can be removed by accounting for this lateral conduction and removing it from the thermal signal. This technique reduces the measurement error by as much as 26%, can be applied to arbitrary substrate topographies, and can be scaled to arrays of heated cantilevers. These results could lead to improvements in nanometer-scale thermal measurements including scanning thermal microscopy and tip-based nanofabrication.