High-throughput Parasitic-independent Probe Thermal Resistance Calibration for Robust Thermal Mapping with Scanning Thermal Microscopy

TL;DR

This paper presents a novel probe thermal resistance calibration method for scanning thermal microscopy, enabling high-resolution thermal mapping and accurate thermal conductivity measurement of nanostructured films.

Contribution

It introduces a parasitic-independent calibration technique that improves the accuracy of local thermal resistance and conductivity measurements at nanoscale.

Findings

Achieved sub-100 nm thermal resistance mapping of 15 nm Al film.

Determined Al film thermal conductivity as 45.1 W/(m·K), significantly lower than bulk aluminum.

Results align with existing experimental and theoretical models.

Abstract

Nanostructured materials, critical for thermal management in semiconductor devices, exhibit a strong size dependence in thermal transport. Studying thermal resistance variation across grain boundaries is critical for designing effective thermal interface materials. Frequency-domain Thermoreflectance (FDTR)-based techniques can provide thermal resistance mapping at the micrometer ({\mu}m) scale. Scanning Thermal Microscopy (SThM) enables quantification of local thermal transport with significantly higher spatial resolution (<100 nm). However, challenges in quantifying the raw signal to thermal conductivity and surface sensitivity limit its widespread adoption for understanding nanoscale heat transport and defect-mediated thermal properties in nanostructured films. Here, we introduce a circuit-based probe thermal resistance (R_p) calibration technique independent of parasitic heat pathways, enabling accurate determination of probe heat dissipation and tip temperature rise, thereby allowing extraction of local thermal resistance. SThM achieved sub-100 nm spatial resolution in mapping thermal resistance across a 15 nm-thick Al film deposited via e-beam evaporation on SiO_2 substrate. The thermal resistance maps are converted to thermal conductivity using robust analytical and finite element models that account for tip-sample geometry, lateral heat spreading, and buried interface effects. Gaussian distribution fitting of pixel-level thermal resistance values yields k_Al = 45.1(+4.7/-3.6) W/(m.K) for the ultra-thin Al film (13-15 nm), representing a 5.3-fold reduction from bulk aluminum (237 W/(m.K)). These results agree with published experimental data and theoretical frameworks explaining thickness-dependent heat transport in ultra-thin metallic films.