The challenge the researchers set out to solve is a familiar one in the field of thermoplasmonics. Gold nanostructures absorb light most efficiently at their localized surface plasmon resonance, the collective oscillation of conduction electrons that turns a gold nanoparticle into a tiny optical antenna. Most temperature-measurement techniques exploit that same resonance, relying on the strong absorption or fluorescence shifts that occur when the laser matches the plasmon peak. But many practical applications, including the self-healing polymer coatings that motivated this study, require heating at wavelengths where the gold is a comparatively weak absorber. At those off-resonance wavelengths, conventional absorption-based thermometry loses sensitivity, and engineers are left guessing at the thermal landscape inside their devices.

The Ningxia team, led by Yanru Xu and Yupeng Qi of the School of Mechanical Engineering, together with Yanqing Wang of the School of Materials and New Energy, turned to a different kind of probe altogether. Rather than measuring absorption, they measured the Raman spectrum of silicon nanoparticles scattered across the surface of their gold nanoarrays. Raman scattering is famously sensitive to temperature: as a crystal warms, its optical phonon peak shifts and broadens in a predictable way. The silicon nanoparticles used here displayed a remarkably stable temperature coefficient of minus 0.023 plus or minus 0.002 wavenumbers per kelvin, and crucially, that coefficient held constant across particle sizes ranging from 20 to 150 nanometers. That size independence means the probes can be deployed without worrying that variations in particle dimensions will corrupt the calibration.

There is a clever twist in how the technique achieves its sensitivity. Although the 785-nanometer excitation laser used in the experiments sits away from the gold nanoarrays’ plasmon resonance, the plasmon still does useful work. The near-field enhancement generated by the localized surface plasmon resonance amplifies the Raman signal from any silicon nanoparticle sitting close to the gold surface, making even subtle temperature-induced spectral shifts detectable. In effect, the gold nanoarray serves double duty: it is both the heat source being studied and the signal amplifier that makes the measurement possible. This plasmon-enhanced Raman thermometry therefore works precisely in the regime where absorption-based methods falter, giving researchers a non-contact, label-free alternative to fluorescence thermometry that requires no dye molecules or genetic tags.

The headline result is a quantitative map of photothermal heating under non-resonant conditions. When the team illuminated their silicon-on-gold-on-silica samples, they measured a temperature rise coefficient of 0.505 plus or minus 0.038 kelvin per milliwatt of laser power, a fourfold enhancement over control samples lacking the gold nanoarrays. At a laser intensity of 5.1 times ten to the seventh watts per square meter, the probes registered a temperature rise of 202 plus or minus 15 kelvin. Those are substantial numbers, hot enough to soften and flow thermoplastic polymers, and they demonstrate that meaningful photothermal conversion does not require driving the plasmon at its resonance peak.

Perhaps the most scientifically interesting part of the study is the mechanistic accounting of where the heat actually comes from. By comparing samples with and without the gold nanoarrays, the researchers decomposed the total heating into two contributions. Roughly 72 percent of the measured temperature rise originates from the non-radiative decay of localized surface plasmons in the gold itself, the process by which the collective electron oscillation relaxes into lattice vibrations and dumps its energy as heat. The remaining 28 percent comes from the plasmon-enhanced intrinsic absorption of the silicon nanoparticles, whose own absorption is boosted by the intensified near field around the gold structures. This kind of quantitative partitioning is rare, and it gives designers a clear rule of thumb for how much heating they can attribute to the metal versus the surrounding materials.

To prove the technique is more than a laboratory curiosity, the team validated it in a practical setting: the photothermal self-healing of polyurethane. Thermoplastic coatings can repair scratches and microcracks if they can be heated locally above their glass transition or flow temperature, allowing the polymer chains to migrate and re-fuse the damaged region. Gold nanoarrays embedded in such coatings act as remotely triggered heaters, and the new Raman thermometry provides the thermal benchmarks needed to design those triggers with precision. Knowing that a given laser intensity produces a given local temperature rise means coating engineers can specify light doses that heal damage without thermally degrading the surrounding material.

The implications extend well beyond self-healing coatings. Plasmonic heating is the engine behind photothermal catalysis, solar-driven water distillation, optical trapping, nanoscale actuation, and biomedical ablation, and in nearly all of these fields the local temperature is the quantity that actually controls performance, yet it is notoriously hard to measure. Techniques such as anti-Stokes luminescence thermometry, photothermal microscopy, and interferometric scattering each have their strengths, but many require resonant excitation, fluorescent labels, or single-particle isolation. The silicon nanoparticle approach is comparatively simple: it uses chemically robust, optically stable probes whose Raman signature is unaffected by photobleaching, and it can be implemented on a standard confocal Raman microscope. Because the probes are non-invasive and label-free, they can be scattered onto virtually any plasmonic substrate without altering its optical properties.

The work also speaks to a broader debate in the plasmonics community about how heat and hot electrons partition after plasmon excitation. Recent studies have examined whether interband or intraband excitation pathways dominate the steady-state population of hot electrons in gold, and how quickly the absorbed energy thermalizes into the lattice. By working at an off-resonance wavelength and quantifying the non-radiative decay contribution directly, the Ningxia measurements add a useful data point: even far from resonance, the plasmon-mediated channel remains the dominant heat source, accounting for nearly three quarters of the observed warming in these nanoarrays. That finding should inform models of photothermal conversion in ordered nanoparticle arrays, where interparticle coupling and substrate effects both shape the absorption profile.

For now, the technique establishes silicon-nanoparticle Raman thermometry as a versatile platform for spatially resolved thermal profiling of plasmonic systems, and the authors suggest it could become a standard benchmarking tool for anyone designing precision photothermal triggers. As smart coatings, thermoplasmonic actuators, and light-driven chemical reactors move from the laboratory toward commercial deployment, the ability to measure, rather than merely estimate, the temperature at the nanoscale may prove to be one of the field’s most valuable enabling technologies. What began as a question about how hot a gold nanoarray gets under the wrong color of laser light has ended with a general-purpose thermometer built from one of the most abundant materials on Earth.

Subject of Research: In situ Raman thermometry using silicon nanoparticle probes to quantify non-resonant photothermal conversion in gold nanoarrays

Article Title: Quantifying non-resonant photothermal conversion in gold nanoarrays via in situ Raman thermometry with silicon nanoparticle probes

Article References: Quantifying non-resonant photothermal conversion in gold nanoarrays via in situ Raman thermometry with silicon nanoparticle probes. (n.d.). https://doi.org/10.1007/s11051-026-06751-6

Image Credits: AI Generated

DOI: 10.1007/s11051-026-06751-6

Keywords: Raman thermometry, silicon nanoparticles, gold nanoarrays, plasmonic heating, photothermal conversion, localized surface plasmon resonance, non-resonant excitation, self-healing coatings, polyurethane, nanoscale thermometry, thermoplasmonics, Journal of Nanoparticle Research