arXiv subjects

Giuseppe Romano

Publications and source records attributed to Giuseppe Romano.

At least 19 recordsLinked to original sources

arXiv2026

When heat goes astray -- non-local heating in a semiconductor

Heating of semiconductor devices limits their performance and lifetime, which must be addressed by thermal management starting at the heat source. It is a common assumption that the heat source and the resulting heat spot locally coincide, if their size exceeds the mean free paths of the main heat carriers, the phonons. We show that this paradigm of heat locality breaks down on length scales spanning several micrometers. As a consequence, non-local heating occurs in contradiction to Fourier's law. Therefore, we heat laterally structured semiconductor membranes that feature a rising number of interfaces with a well-focussed laser and map-out lattice temperatures by Raman thermometry. Remarkably, the non-local heating can exceed the laser-induced local heating, which we attribute to ballistic phonon transport far above cryogenic temperatures.

cond-mat.mes-hall↗

arXiv2026

Differentiating through binarized topology changes: Second-order subpixel-smoothed projection

A key challenge in topology optimization (TopOpt) is that manufacturable structures, being inherently binary, are non-differentiable, creating a fundamental tension with gradient-based optimization. The subpixel-smoothed projection (SSP) method addresses this issue by smoothing sharp interfaces at the subpixel level through a first-order expansion of the filtered field. However, SSP does not guarantee differentiability under topology changes, such as the merging of two interfaces, and therefore violates the convergence guarantees of many popular gradient-based optimization algorithms. We overcome this limitation by regularizing SSP with the Hessian of the filtered field, resulting in a twice-differentiable projected density during such transitions, while still guaranteeing an almost-everywhere binary structure. We demonstrate the effectiveness of our second-order SSP (SSP2) methodology on both thermal and photonic problems, showing that SSP2 has faster convergence than SSP for connectivity-dominant cases -- where frequent topology changes occur -- while exhibiting comparable performance otherwise. Beyond improving convergence guarantees for CCSA optimizers, SSP2 enables the use of a broader class of optimization algorithms with stronger theoretical guarantees, such as interior-point methods. Since SSP2 adds minimal complexity relative to SSP or traditional projection schemes, it can be used as a drop-in replacement in existing TopOpt codes.

eess.SP↗

arXiv2025

Physics Enhanced Deep Surrogates for the Phonon Boltzmann Transport Equation

Designing materials with controlled heat flow at the nano-scale is central to advances in microelectronics, thermoelectrics, and energy-conversion technologies. At these scales, phonon transport follows the Boltzmann Transport Equation (BTE), which captures non-diffusive (ballistic) effects but is too costly to solve repeatedly in inverse-design loops. Existing surrogate approaches trade speed for accuracy: fast macroscopic solvers can overestimate conductivities by hundreds of percent, while recent data-driven operator learners often require thousands of high-fidelity simulations. This creates a need for a fast, data-efficient surrogate that remains reliable across ballistic and diffusive regimes. We introduce a Physics-Enhanced Deep Surrogate (PEDS) that combines a differentiable Fourier solver with a neural generator and couples it with uncertainty-driven active learning. The Fourier solver acts as a physical inductive bias, while the network learns geometry-dependent corrections and a mixing coefficient that interpolates between macroscopic and nano-scale behavior. PEDS reduces training-data requirements by up to 70% compared with purely data-driven baselines, achieves roughly 5% fractional error with only 300 high-fidelity BTE simulations, and enables efficient design of porous geometries spanning 12-85 W m$^{-1}$ K$^{-1}$ with average design errors of 4%. The learned mixing parameter recovers the ballistic-diffusive transition and improves out of distribution robustness. These results show that embedding simple, differentiable low-fidelity physics can dramatically increase surrogate data-efficiency and interpretability, making repeated PDE-constrained optimization practical for nano-scale thermal-materials design.

physics.comp-ph↗

arXiv2025

Hyperparameter-free minimum-lengthscale constraints for topology optimization

The geometric constraints of Zhou et al. (2015) are a widely used technique in topology/freeform optimization to impose minimum lengthscales for manufacturability. However, its efficacy degrades as design binarization is increased, and it requires heuristic tuning of multiple hyperparameters. In this work, we derive analytical hyperparameters from first principles, depending only on the target lengthscale. We present results for both conic and PDE-based filtering schemes, showing that the latter is less robust due to the singularity of its underlying Green's function. To address this, we also introduce a double-filtering approach to obtain a well-behaved PDE-based filter. Combined with our derived hyperparameters, we obtain a straightforward strategy for enforcing lengthscales using geometric constraints, with minimal hyperparameter tuning. A key enabling factor is the recent subpixel-smooth projection (SSP) method (Hammond et al. 2025), which facilitates the rapidly-converging optimization of almost-everywhere binary designs. The effectiveness of our method is demonstrated for several photonics and heat-transfer inverse-design problems.

physics.optics↗

arXiv2025

Phonon Mean Free Path Spectroscopy By Raman Thermometry

In this work, we exemplify on a bulk silicon sample that Raman thermometry is capable of phonon mean free path (PMFP) spectroscopy. Our experimental approach is similar to the variation of different characteristic length scales $l_{c}$ during thermal reflectance measurements in the time or frequency domain and transient thermal grating spectroscopy. In place of $l_{c}$, we vary the laser focus spot size ($w_{e}$) and the light penetration depth ($h_{\alpha}$) during one-laser Raman thermometry (1LRT) measurements. For our largest $w_{e}$ values, the derived effective thermal conductivities $\kappa_{eff}$ converge towards the bulk thermal conductivity $\kappa_{bulk}$ for silicon. However, towards smaller $w_{e}$ values, we observe a pronounced increase for the $\kappa_{eff}$ values, which amounts up to a factor of 5.3 at 293K and even 8.3 at 200K. We mainly assign this phenomenon to quasi-ballistic phonon transport. As a result, we can compare our measured $\kappa_{eff}(w_{e})$ trends with the thermal accumulation function $\kappa_{cum}$ and its dependence on the phonon mean free path $l_{ph}$, which we derive from ab initio solutions of the linearized phonon Boltzmann transport equation (BTE). Since the variation of $w_{e}$ can be experimentally cumbersome, we also suggest varying $h_{\alpha}(\lambda)$ via the applied Raman laser wavelength $\lambda$ during 1LRT. In this regard, we present proof-of-principle 1LRT measurements, yielding a step-like $\kappa_{eff}(\lambda)$ trend for four different $\lambda$ values, which we also interpret in terms of quasi-ballistic phonon transport. Our results shall seed future PMFP spectroscopy based on 1LRT, which can directly be benchmarked against state-of-art theory by comparison of $\kappa_{\text{cum}}$ trends and not only $\kappa$ values, aiming to test our understanding of the intricate phonon transport physics.

physics.optics↗

arXiv2025

Thermal Analog Computing: Application to Matrix-vector Multiplication with Inverse-designed Metastructures

The rising computational demand of modern workloads has renewed interest in energy-efficient paradigms such as neuromorphic and analog computing. A fundamental operation in these systems is matrix-vector multiplication (MVM), ubiquitous in signal processing and machine learning. Here, we demonstrate MVM using inverse-designed metastructures that exploit heat conduction as the signal carrier. The proposed approach is based on a generalization of effective thermal conductivity to systems with multiple input and output ports: The input signal is encoded as a set of applied temperatures, while the output is represented by the power collected at designated terminals. The metastructures are obtained via density-based topology optimization, enabled by a differentiable thermal transport solver and automatic differentiation, achieving an accuracy $>99\%$ in most cases across a pool of matrices with dimensions $2\times2$ and $3\times3$. We apply this methodology -- termed thermal analog computing -- to realize matrices relevant to practical tasks, including the discrete Fourier transform and convolutional filters. These findings open new avenues for analog information processing in thermally active environments, including temperature-gradient sensing in microelectronics and thermal control systems.

cond-mat.mes-hall↗

arXiv2025

DiffChip: Thermally Aware Chip Placement with Automatic Differentiation

Chiplets are modular integrated circuits that can be combined to form a larger system, offering flexibility and performance enhancements. However, their dense packing often leads to significant thermal management challenges, requiring careful floorplanning to ensure efficient heat distribution. To address thermal considerations, layout optimization algorithms concurrently minimize the total wirelength and the maximum temperature. However, these efforts employ gradient-free approaches, such as simulated annealing, which suffer from poor scaling and slow convergence. In this paper, we propose DiffChip, a chiplet placement algorithm based on automatic differentiation (AD). The proposed framework relies on a differentiable thermal solver that computes the sensitivity of the temperature map with respect to the positions of the chiplets. Regularization strategies for peak temperature, heat sources, and material properties enable end-to-end differentiability, allowing for gradient-based optimization. We apply DiffChip to optimize a layout where the total wirelength is minimized while keeping the maximum temperature below a desired threshold. By leveraging AD and physics-aware optimization, our approach accelerates the design process of microelectronic systems, exceeding traditional trial-and-error and gradient-free methods.

physics.comp-ph↗

arXiv2024

Thermal analysis of GaN-based photonic membranes for optoelectronics

Semiconductor membranes find their widespread use in various research fields targeting medical, biological, environmental, and optical applications. Often such membranes derive their functionality from an inherent nanopatterning, which renders the determination of their, e.g., optical, electronic, mechanical, and thermal properties a challenging task. In this work we demonstrate the non-invasive, all-optical thermal characterization of around 800-nm-thick and 150-$\mu$m-wide membranes that consist of wurtzite GaN and a stack of In$_{0.15}$Ga$_{0.85}$N quantum wells as a built-in light source. Due to their application in photonics such membranes are bright light emitters, which challenges their non-invasive thermal characterization by only optical means. As a solution, we combine two-laser Raman thermometry with (time-resolved) photoluminescence measurements to extract the in-plane (i.e., $c$-plane) thermal conductivity $\kappa_{\text{in-plane}}$ of our membranes. Based on this approach, we can disentangle the entire laser-induced power balance during our thermal analysis, meaning that all fractions of reflected, scattered, transmitted, and reemitted light are considered. As a result of our thermal imaging via Raman spectroscopy, we obtain $\kappa_{\text{in-plane}}\,=\,165^{+16}_{-14}\,$Wm$^{-1}$K$^{-1}$ for our best membrane, which compares well to our simulations yielding $\kappa_{\text{in-plane}}\,=\,177\,$Wm$^{-1}$K$^{-1}$ based on an ab initio solution of the linearized phonon Boltzmann transport equation. Our work presents a promising pathway towards thermal imaging at cryogenic temperatures, e.g., when aiming to elucidate experimentally different phonon transport regimes via the recording of non-Fourier temperature distributions.

physics.optics↗

arXiv2024

Optical and thermal characterization of a group-III nitride semiconductor membrane by microphotoluminescence spectroscopy and Raman thermometry

We present the simultaneous optical and thermal analysis of a freestanding photonic semiconductor membrane made from wurtzite III-nitride material. By linking micro-photoluminescence ($μ$PL) spectroscopy with Raman thermometry, we demonstrate how a robust value for the thermal conductivity $κ$ can be obtained using only optical, non-invasive means. For this, we consider the balance of different contributions to thermal transport given by, e.g., excitons, charge carriers, and heat carrying phonons. Further complication is given by the fact that this membrane is made from direct bandgap semiconductors, designed to emit light based on an In$_{x}$Ga$_{1-x}$N ($x=0.15$) quantum well embedded in GaN. To meet these challenges, we designed a novel experimental setup that enables the necessary optical and thermal characterizations in parallel. We perform micro-Raman thermometry, either based on a heating laser that acts as a probe laser (1-laser Raman thermometry), or based on two lasers, providing the heating and the temperature probe separately (2-laser Raman thermometry). For the latter technique, we obtain temperature maps over tens of micrometers with a spatial resolution less than $1\,μ\text{m}$, yielding $κ\,=\,95^{+11}_{-7}\,\frac{\text{W}}{\text{m}\cdot \text{K}}$ for the $\textit{c}$-plane of our $\approx\,250\text{-nm}$-thick membrane at around room temperature, which compares well to our $\textit{ab initio}$ calculations applied to a simplified structure. Based on these calculations, we explain the particular relevance of the temperature probe volume, as quasi-ballistic transport of heat-carrying phonons occurs on length scales beyond the penetration depths of the heating laser and even its focus spot radius. The present work represents a significant step towards non-invasive, highly spatially resolved, and still quantitative thermometry performed on a photonic membrane.

physics.optics↗

arXiv2023

Reduced-Order Model to Predict Thermal Conductivity of Dimensionally-Confined Materials

Predicting nanoscale thermal transport in dielectrics requires models, such as the Boltzmann transport equation (BTE), that account for phonon boundary scattering in structures with complex geometries. Although the BTE has been validated against several key experiments, its computational expense limits its applicability. Here, we demonstrate the use of an analytic reduced-order model for predicting the thermal conductivity in dimensionally confined materials, i.e., monolithic and porous thin films, and rectangular and cylindrical nanowires. The approach uses the recently developed "Ballistic Correction Model" (BCM) which accounts for materials' full distribution of phonon mean-free-paths. The model is validated against BTE simulations for a selection of base materials, obtaining excellent agreement. By furnishing a precise yet easy-to-use prediction of thermal transport in nanostructures, our work strives to accelerate the identification of materials for energy-conversion and thermal-management applications.

cond-mat.mtrl-sci↗

arXiv2022

Inverse Design in Nanoscale Heat Transport via Interpolating Interfacial Phonon Transmission

We introduce a methodology for density-based topology optimization of non-Fourier thermal transport in nanostructures, based upon adjoint-based sensitivity analysis of the phonon Boltzmann transport equation (BTE) and a novel material interpolation technique, the "transmission interpolation model" (TIM). The key challenge in BTE optimization is handling the interplay between real- and momentum-resolved material properties. By parameterizing the material density with an interfacial transmission coefficient, TIM is able to recover the hard-wall and no-interface limits, while guaranteeing a smooth transition between void and solid regions. We first use our approach to tailor the effective thermal-conductivity tensor of a periodic nanomaterial; then, we maximize classical phonon size effects under constrained diffusive transport, identifying a promising new thermoelectric material design. Our method enables the systematic optimization of materials for heat management and conversion and, more broadly, the design of devices where diffusive transport is not valid.

cond-mat.mes-hall↗

arXiv2022

Multifidelity deep neural operators for efficient learning of partial differential equations with application to fast inverse design of nanoscale heat transport

Deep neural operators can learn operators mapping between infinite-dimensional function spaces via deep neural networks and have become an emerging paradigm of scientific machine learning. However, training neural operators usually requires a large amount of high-fidelity data, which is often difficult to obtain in real engineering problems. Here, we address this challenge by using multifidelity learning, i.e., learning from multifidelity datasets. We develop a multifidelity neural operator based on a deep operator network (DeepONet). A multifidelity DeepONet includes two standard DeepONets coupled by residual learning and input augmentation. Multifidelity DeepONet significantly reduces the required amount of high-fidelity data and achieves one order of magnitude smaller error when using the same amount of high-fidelity data. We apply a multifidelity DeepONet to learn the phonon Boltzmann transport equation (BTE), a framework to compute nanoscale heat transport. By combining a trained multifidelity DeepONet with genetic algorithm or topology optimization, we demonstrate a fast solver for the inverse design of BTE problems.

physics.comp-ph↗

arXiv2022

Mode- and Space- Resolved Thermal Transport of Alloy Nanostructures

Nanostructured semiconducting alloys obtain ultra-low thermal conductivity as a result of the scattering of phonons with a wide range of mean-free-paths (MFPs). In these materials, long-MFP phonons are scattered at the nanoscale boundaries whereas short-MFP high-frequency phonons are impeded by disordered point defects introduced by alloying. While this trend has been validated by simplified analytical and numerical methods, an ab-initio space-resolved approach remains elusive. To fill this gap, we calculate the thermal conductivity reduction in porous alloys by solving the mode-resolved Boltzmann transport equation for phonons using the finite-volume approach. We analyze different alloys, length-scales, concentrations, and temperatures, obtaining a very large reduction in the thermal conductivity over the entire configuration space. For example, a ~97% reduction is found for Al$_{0.8}$In$_{0.2}$As with 25% porosity. Furthermore, we employ these simulations to validate our recently introduced "Ballistic Correction Model" (BCM), an approach that estimates the effective thermal conductivity using the characteristic MFP of the bulk alloy and the length-scale of the material. The BCM is then used to provide guiding principles in designing alloy-based nanostructures. Notably, it elucidates how porous alloys such as Si$_{x}$Ge$_{1-x}$ obtain larger thermal conductivity reduction compared to porous Si or Ge, while also explaining why we should not expect similar behavior in alloys such as Al$_{x}$In$_{1-x}$As. By taking into account the synergy from scattering at different scales, we provide a route for the design of materials with ultra-low thermal conductivity.

cond-mat.mtrl-sci↗

arXiv2021

dPV: An End-to-End Differentiable Solar-Cell Simulator

We introduce dPV, an end-to-end differentiable photovoltaic (PV) cell simulator based on the drift-diffusion model and Beer-Lambert law for optical absorption. dPV is programmed in Python using JAX, an automatic differentiation (AD) library for scientific computing. Using AD coupled with the implicit function theorem, dPV computes the power conversion efficiency (PCE) of an input PV design as well as the derivative of the PCE with respect to any input parameters, all within comparable time of solving the forward problem. We show an example of perovskite solar-cell optimization and multi-parameter discovery, and compare results with random search and finite differences. The simulator can be integrated with optimization algorithms and neural networks, opening up possibilities for data-efficient optimization and parameter discovery.

physics.comp-ph↗

arXiv2021

Universal Effective Medium Theory to Predict the Thermal Conductivity in Nanostructured Materials

Nanostructured materials enable high thermal transport tunability, holding promises for thermal management and heat harvesting applications. Predicting the effect that nanostructuring has on thermal conductivity requires models, such as the Boltzmann transport equation (BTE), that capture the non-diffusive transport of phonons. Although the BTE has been well validated against several key experiments, notably those on nanoporous materials, its applicability is computationally expensive. Several effective model theories have been put forward to estimate the effective thermal conductivity; however, most of them are either based on simple geometries, e.g., thin films, or simplified material descriptions such as the gray-model approximation. To fill this gap, we propose a model that takes into account the whole mean-free-path (MFP) distribution as well as the complexity of the material's boundaries in infinitely thick films with extruded porosity using uniparameter logistic regression. We validate our approach, which is called the "Ballistic Correction Model" (BCM), against full BTE simulations of a selection of three base materials (GaAs, InAs, and Si) with nanoscale porosity, obtaining excellent agreement. While the key parameters of our method, associated with the geometry of the bulk material, are obtained from the BTE, they can be decoupled and used in arbitrary combinations and scales. We tabulated these parameters for a few cases, enabling the exploration of systems that are beyond those considered in this work. Providing a simple yet accurate estimation of thermal transport in nanostructures, our work sets out to accelerate the discovery of materials for thermal-related applications.

cond-mat.mtrl-sci↗

arXiv2021

Efficient calculations of the Mode-Resolved ab-initio thermal Conductivity in nanostructures

First-principles calculations of thermal transport in homogeneous materials have reached remarkable predicting power. Modeling deterministically phonon transport in nanostructures, however, poses novel challenges; notably, it entails solving as many algebraic equations as the number of combinations of wave vectors in the discretized Brillouen and polarizations. We show that, within the relaxation time approximation of the Boltzmann transport equation (BTE), this issue is resolved by interpolating the phonon distributions in the vectorial phonon mean free paths (MFP) space. The coupling between structure and mode-resolved heat transport is investigated in terms of angular-resolved bulk thermal conductivity and phonon suppression function, the latter being associated primarily to the material's geometry. Our method, termed the anisotropic MFP-BTE (aMFP-BTE), allows for fast and accurate thermal conductivity calculations in nanomaterials regardless of the number of phonon branches and wave vectors. Furthermore, it naturally blends with first-principles thermal transport calculations, therefore allowing for multiscale, parameter-free simulations. We apply the aMFP-BTE to compute the mode-resolved effective thermal conductivity of porous Si membranes, achieving up to 50x speed with respect to the case with no interpolation. The proposed approach unlocks the engineering of novel nanostructures, with applications to thermoelectrics and heat management.

cond-mat.mtrl-sci↗

arXiv2021

OpenBTE: a Solver for ab-initio Phonon Transport in Multidimensional Structures

Controlling heat flow at the nanoscales is pivotal to several applications, including thermal energy harvesting and heat management. However, engineering nanostructures is challenging because phonon-boundary interaction, not contemplated by Fourier's law, must be taken into account. Nondiffusive models, such as the Boltzmann transport equation (BTE), have been successfully employed to capture size effects in complex structures; however, their widespread has been hindered by the limited offer of open-source solvers in this space. We fill this void by introducing OpenBTE, an efficient solver for the steady-state phonon BTE in multidimensional structures. This tool is interfaced to first-principles calculations, thus it unlocks the calculations of thermal-related properties with no fitting-parameters. As an example, we employ OpenBTE to compute the temperature and flux maps, as well as the mode-resolved, effective thermal conductivity of Si membranes with infinite and finite thickness. By unlocking fast nanoscale heat transport simulations, OpenBTE may help accelerate the design of nanomaterials for thermal energy applications.

cond-mat.mtrl-sci↗

arXiv2021

Mitigating the Effect of Nanoscale Porosity on Thermoelectric Power Factor of Si

The addition of porosity to thermoelectric materials can significantly increase the figure of merit, ZT, by reducing the thermal conductivity. Unfortunately, porosity is also detrimental to the thermoelectric power factor in the numerator of the figure of merit ZT. In this manuscript we derive strategies to recoup electrical performance in nanoporous Si by fine tuning the carrier concentration and through judicious design of the pore size and shape so as to provide energy selective electron filtering. In this study, we considered phosphorus doped silicon containing discrete pores that are either spheres, cylinders, cubes, or triangular prisms. The effects from these pores are compared with those from extended pores with circular, square and triangular cross sectional shape, and infinite length perpendicular to the electrical current. A semiclassical Boltzmann transport equation is used to model Si thermoelectric power factor. This model reveals three key results: The largest enhancement in Seebeck coefficient occurs with cubic pores. The fractional improvement is about 15% at low carrier concentration ($< 10^{20}\ \mathrm{1/cm^3}$) up to 60% at high carrier population with characteristic length around $\sim 1\ \mathrm{nm}$. To obtain the best energy filtering effect at room temperature, nanoporous Si needs to be doped to higher carrier concentration than is optimal for bulk Si. Finally, in $n$-type Si thermoelectrics the electron filtering effect that can be generated with nanoscale porosity is significantly lower than the ideal filtering effect; nevertheless, the enhancement in the Seebeck coefficient that can be obtained is large enough to offset the reduction in electrical conductivity caused by porosity.

cond-mat.mes-hall↗