Electrical thermography via centimetre-scale fiber-based distributed temperature sensing

TL;DR

The work addresses the limitation of infrared thermography for subsurface and cryogenic electronics by introducing a Raman-based distributed temperature sensor (RDTS) that achieves centimetre-scale spatial resolution along a fiber laid on a PCB. Using optical time-domain reflectometry of Raman signals detected by superconducting nanowire single-photon detectors, the approach delivers 3 cm spatial resolution and around 2 °C temperature accuracy with a 5-minute integration, and operates down to 77 K. The method is validated through room-temperature calibration, spatial mapping of multiple heater elements, and cryogenic demonstrations showing substantial reductions in thermal resistance due to enhanced convective cooling in liquid nitrogen, enabling real-time thermography in regimes where infrared imaging fails. The technique offers a non-invasive, high-sensitivity diagnostic tool for identifying hotspots in densely packed or volumetric electronic systems and could extend to cryogenic electronics, detectors, and superconducting devices. Shorter optical pulses could further improve spatial resolution below 1 cm, leveraging the low timing jitter of SNSPDs.

Abstract

We present a Raman-based Distributed Temperature Sensor (RDTS) with centimetre-scale resolution for thermographic analysis of electronic circuits. Temperature is measured along a single-mode fiber routed across a custom printed circuit board (PCB) with 1 cm$^2$ heating elements, using optical time-domain reflectometry of Raman signals detected by superconducting nanowire single-photon detectors (SNSPDs). This approach enables two-dimensional thermal mapping of the PCB under heating configurations with multiple hotspots. A spatial resolution of 3 cm and a temperature accuracy of 2 °C are achieved with an integration time of 5 minutes. Thermography can be performed down to 77 K, revealing that the PCB thermal resistance decreases by nearly an order of magnitude compared to room temperature, due to enhanced convective cooling in liquid nitrogen. These results establish centimetre-scale RDTS as a robust technique for real-time, spatially resolved thermography of electronic circuits, particularly in regimes where infrared imaging is ineffective, such as at low temperatures or within volumetric electronic architectures.

Figures (6)

• Figure 1: Optical setup and printed circuit board (PCB) developed for fiber-based thermography measurements. a) Schematic of the RDTS optical setup. A time-to-digital controller (TDC) triggers a laser emitting pulses at 1548 nm. The pulses pass through a band-pass filter (BPF) and are injected into the fiber under test, which is routed on a custom-designed PCB. A circulator directs the backscattered Raman signal to a wavelength-division multiplexer (WDM) module that separates the AS and S components. These signals are detected by two superconducting nanowire single-photon detectors (SNSPDs), which generate stop signals for the TDC. b) Electrical schematic of the PCB heating elements. c) Footprint of a single heating element on the PCB. d) Top view of the PCB showing the layout of the 15 heating elements. e) Photograph of the experimental setup with the fiber positioned on the PCB.
• Figure 2: Heat dissipation measurements on the PCB. (a) Schematic representation of the PCB layout. (b) Temperature variation on the R9 heating element in ambient air as a function of the electrical current. (c) Temperature variation on the R9 heating element with the PCB immersed in liquid nitrogen (LN$_2$). Quadratic fitting lines are represented in red.
• Figure 3: Temperature calibration of the RDTS in ambient air. (a) AS signal as a function of position $x$ along the FUT, measured for different heating temperatures of the R9 element on the PCB. The grey-shaded area indicates the region used for calibration. (b) Ratio of the AS signal (after subtraction of its value at 23) to the Stokes signal (after subtraction of the dark count noise) within the calibration region, plotted as a function of $\exp(\hbar \Omega / (k_B T_{\mathrm{cal}}))$. The fitting red line is used to extract the calibration constants. (c) DTS-estimated temperature $T_{\mathrm{DTS}}$ as a function of the calibrated thermocouple temperature $T_{\mathrm{cal}}$. The green line represents the ideal one-to-one correspondence. (d) Temperature profiles estimated by the DTS along the FUT for different temperatures in the calibration region.
• Figure 4: Spatial calibration for PCB fiber-based thermography. Spatial coordinates of 1 cm distant points selected along the FUT are first identified (top). Each point is then mapped to a specific 1 cm bin in the temperature profile. A color code is used to distinguish the points (middle-top). A Gaussian hot spot with 1 cm FWHM is generated at each selected position (middle-bottom). Gaussian filtering is then applied to smooth the resulting temperature map (bottom).
• Figure 5: PCB thermography under different spatial and temperature conditions. Temperature maps for different heating temperatures on the R9 element (a-d), on R9 - R3 (e), on R6 - R12 (f), on R4-R11-R14 (g), and on R6-R9-R12 elements (h). The dashed squares indicate the positions of the respective heating elements.