Symposium EN03—Thermal Materials, Modeling and Technoeconomic Impacts for Thermal Management and Energy Application

Phonon scattering plays a central role in the quantum theory of phonon linewidth, which in turn governs important properties including infrared spectra, Raman spectra, lattice thermal conductivity, thermal radiative properties, and also significantly affects other important processes such as hot electron relaxation. Since Maradudin and Fein's classic work in 1962, three-phonon scattering had been considered as the dominant intrinsic phonon scattering mechanism and has seen tremendous advances. However, the role of four-phonon scattering had been persistently unclear and so was ignored. The tremendous complexity of the formalism and computational challenges stood in the way, prohibiting the direct and quantitative treatment of four-phonon scattering. In 2016, a rigorous four-phonon scattering formalism was developed, and the prediction was realized using empirical potentials. In 2017, the method was extended using first-principles calculated force constants, and the thermal conductivities of boron arsenides (BAs), Si and diamond were predicted. The predictions for BAs were later confirmed by several independent experiments. Four-phonon scattering has since been investigated in a range of materials and established as an important intrinsic scattering mechanism for thermal transport and radiative properties. After this tutorial, the audience is expected to (1) be familiar with the background of three- and four-phonon scattering and their roles in thermal and radiative properties of materials, (2) understand the various characteristics of four-phonon scattering mechanism in different systems and scenarios, (3) understand the broad impact of four-phonon scattering on thermal transport and radiative properties in various materials, (4) be able to tell in which types of materials and scenarios will four-phonon scattering be critical, (5) be able to use the open-source code Four-Phonon together with Sheng BTE to calculate four-phonon scattering rates (linewidth) and thermal conductivities for materials.

High and low thermal conductivities in polymers are needed for new materials for thermal management. While the thermal conductivity of simple inorganic crystals can be calculated from first principles with good accuracy, a predictive understanding of the relationship between the thermal conductivity and the molecular structure of polymers is not yet established. We recently embarked on an extensive experimental study of the molecular structure/thermal-conductivity relationships within the class of materials known as epoxy thermosets, i.e., mixtures of epoxide monomers and amine hardeners. We measured 32 formulations to-date. For each formulation, we measure the thermal diffusivity using frequency domain probe beam deflection (FD-PBD), thermal effusivity by time-domain thermoreflectance (TDTR), acoustic velocity by picosecond interferometry, density by buoyancy, and for selected compositions, crystalline and liquid-crystalline order by small and wide-angle x-ray diffraction (SAXS and WAXS). We compare the results to predictions based on the model of the minimum thermal conductivity. The thermal conductivity of these 32 epoxy resins spans the range from 0.15 to 1.0 W/m-K. Thermal conductivity is enhanced by extended aromatic structures, high molecular density, and liquid crystalline and crystalline order. We have discovered a striking odd-even effect in the dependence of thermal conductivity on the molecular length of an aliphatic linker segment in epoxide monomers.

Crystalline materials with amorphous-like ultralow thermal conductivity are highly coveted for thermal barrier coating and thermoelectric applications. However, such behavior is rarely observed in high-quality, bulk crystals with simple crystal structures. In this study, we report on the amorphous-like ultralow thermal conductivity of high-purity, bulk, crystalline chalcogenide perovskite BaZrS₃ and its ruddlesden popper phase Ba₃Zr₂S₇. Extensive microstructural characterizations and first-principles calculations are performed to explain these exceptional thermal behaviors. Both BaZrS₃ and Ba₃Zr₂S₇ exhibit very high phonon scattering rates as indicated by some of the highest ever reported elastic modulus/thermal conductivity ratios. The low number density of the acoustic phonon modes, strong gruneisen parameter, and short phonon lifetimes lead to a thermal conductivity of 1.21 ± 0.18 W m^-1 K^-1 in BaZrS_3. In addition, this material also exhibits an amorphous-like thermal conductivity trend which likely stems from the phonon-domain boundary scattering. The highly ordered layered crystal structure of Ba₃Zr₂S₇, on the other hand, leads to a thermal conductivity of 0.46 ± 0.08 W m^-1 K^-1, nearly one third of that of the BaZrS₃. Moreover, the thermal conductivity of Ba₃Zr₂S₇ is in agreement with the diffuson-mediated thermal conductivity limit indicating the dominance of diffuson like vibrational modes. As a result, at 100 K, the thermal conductivity of Ba₃Zr₂S₇ is the lowest reported to date among bulk crystalline materials. Introduction of disorder by ion bombardment gradually reduces the thermal conductivity of BaZrS₃. This is in contrast to Ba₃Zr₂S₇ where no change in thermal conductivity is observed with irradiation doses. Our study provides conclusive evidence that highly ordered nanostructuring can lead to diffuson like lattice vibrations in already low thermal conductivity crystalline materials.

Most thermodynamic modeling of hydrogels focused on predicting their final volumes in equilibrium with water, built on Flory theories for the entropy of mixing, rubber elasticity, and the Donnan equilibrium. This work will focus on water and ions in and outside hydrogels. Deficiencies and confusions in the Donnan equilibrium theory and Flory hydrogel model will be discussed and amended. It turns out that Donnan’s equilibrium criteria established over 100 years ago neglected the coupling between ion concentrations and the osmotic pressure. The revised model will be employed to study thermodynamic properties of water in hydrogels including (1) the latent heat of evaporation, (2) the ability of hydrogels to retain water and to absorb water from the atmosphere, and (3) the use of hydrogels for desalination. Such thermodynamic studies will also be useful for understanding polymer electrolytes and ion transport in membranes in electrochemical systems.

Polymers typically have low thermal conductivities, ~0.2 W/(m K), which limits their application in systems that must manage high heat loads. Dynamic covalent networks are a class of polymer networks studied for their self-healing and stress relaxation properties and they have potential applications as shape memory polymers, flame retardants, solid electrolytes and shockwave dissipation materials. Thermal conductivity of dynamic polymer network is measured by time-domain thermoreflectance (TDTR) and frequency domain probe beam deflection (FD-PBD). At room temperature, thermal conductivity of dynamic polymer network varies from 0.05 W/(m K) to 1.0 W/(m K) by a factor of 20. Dynamic polymer network with crystal structure can have thermal conductivity of 1.0 W/(m K). Thermal conductivity of amorphous ethylene dynamic networks, synthesized from benzene diboronic acid and alkane diols, decreases from 0.19 W/(m K) to 0.095 W/(m K) at room temperature as ethylene linker length increases from 4 to 12. Thermal conductivity of these amorphous ethylene dynamic networks also decreases from 0.29 W/(m K) to 0.095 W/(m K) by a factor of 3 as the temperature increases across glass transition temperature. Minimum thermal conductivity model (MTCM) successfully predicts thermal conductivity of dynamic polymer network with different ethylene linker length and under different temperatures.

Lithium-intercalated layered transition-metal oxides, Li_xTMO_2, brought about a paradigm change in rechargeable batteries in recent decades and show promise for use in memristors, a type of device for future neural computing and on-chip storage. Thermal transport properties, although being a crucial element in limiting the charging/discharging rate, package density, energy efficiency, and safety of batteries as well as the controllability and energy consumption of memristors, are poorly managed or even understood yet. Here, for the first time, we employ quantum calculations including high-order lattice anharmonicity and find that the thermal conductivity κ of Li_xTMO₂ materials is significantly lower than hitherto believed. More specifically, the theoretical upper limit of κ of LiCoO₂ is 6 W/m-K, 2–6 times lower than the prior theoretical predictions. Delithiation further reduces κ by 40–70% for LiCoO₂ and LiNbO₂. Grain boundaries, strain, and porosity are yet additional causes of thermal-conductivity reduction, while Li-ion diffusion and electrical transport are found to have only a minor effect on phonon thermal transport. The results elucidate several long-standing issues regarding the thermal transport in lithium-intercalated materials and provide guidance toward designing high-energy-density batteries and controllable memristors.

Machine learning is trending to be an integral part of thermal science. In this talk, I will introduce our efforts in utilizing machine learning (ML) techniques to predict phonon and thermal properties. I will talk about the use of a new ML method, called transfer learning, to establish accurate models based on limited data for predicting phonon and thermal transport properties, such as phonon frequency gap, heat capacity, speed of sound and lattice thermal conductivity.

As the need for energy-efficient devices has become more pressing, so has interest in finding ways to enhance the thermoelectric performance of materials. While the Seebeck effect has been intensely studied, the Nernst effect has been rarely investigated computationally. Recent results, such as those showing enhanced thermopower in compensated semimetals via the Nernst effect, suggest that it also has potential to provide enhancements in thermoelectric performance in the right circumstances. Additionally, established results such as those for bismuth, which shows a low-temperature peak in the Nernst coefficient, could be better understood and further engineered for potential applications.

However, these interesting and possibly useful behaviors can be difficult to interpret from experimental results alone. Using first-principles solutions of the Boltzmann transport equation as implemented in our new open-source code package, Phoebe [1], we calculate the Seebeck and Nernst coefficients to better understand the physics underlying these effects. Time permitting, we will also discuss the role of phonon drag on these coefficients at low temperatures.
[1] https://mir-group.github.io/phoebe/

The thermal conductivity (k) of semiconducting nanomaterials is determined by the geometry-dependent phonon boundary scattering mean free path (MFP). Although prior work has investigated phonon transport in periodically corrugated rectangular nanowires, the detailed relationship between geometric parameters and the MFP in recently fabricated diameter-modulated cylindrical nanowires is not known. For example, it is not clear whether enhancing the larger cylinder diameter would reduce the MFP due to the phonon backscattering effect, or enhance the MFP due to the larger average path length between phonon collisions in the larger cylinder. In addition to these fundamental questions about the phonon transport physics, understanding the thermal conductivity in modulated diameter nanowires may be important for future applications of modulated silicon nanowires as photodetectors or optomechanical sensors.

Here, we use Monte Carlo phonon ray tracing simulations to comprehensively study the effect of geometric parameters on the boundary scattering MFP in periodic coaxial cylindrical nanowires. We find that for a fixed smaller cylinder diameter, the MFP can be maximized or minimized via geometric control of the pitch, length, and diameter ratios. Our simulations show that the phonon backscattering framework is appropriate for small-pitch nanowires and gives rise to the minimum in the dimensionless MFP, while an enhanced path-length mechanism is appropriate for long-pitch nanowires and leads to the maximum in the dimensionless MFP.

Combining our calculations with analytical phonon dispersion and bulk scattering models, we predict that the thermal conductivity of periodic silicon nanowires with a fixed smaller diameter and cylinder length ratio can be tuned by up to 34% via geometric control of the larger diameter and pitch. A stronger geometric enhancement can be observed in the thermal conductance than in the thermal conductivity. Our results provide insight into the mechanisms of thermal conductivity suppression and enhancement in periodically modulated nanowires, and can be used by experimentalists to explore phonon backscattering and path-length effects in modulated nanowires.

Theoretical modeling of phonon transport process in strongly anharmonic materials at a finite temperature needs to accurately capture the effects of lattice anharmonicity. The anharmonicity of potential energy surface would result in not only strong phonon scatterings but also shifts of phonon frequencies and eigenvectors. In this work, we evaluated the roles of anharmonicity-renormalized phonon eigenvectors in predicting phonon transport properties of anharmonic crystals at high temperatures, using molecular dynamics-based normal mode analysis (NMA) methods in both time domain and frequency domain. Using PbTe as a model of strongly anharmonic crystal, we analyzed the numerical challenges to extract phonon lifetimes using NMA methods when phonon eigenvectors deviate from their harmonic values at high temperatures. To solve these issues, we proposed and verified a better fitting strategy, Sum-up Spectrum Fitting Method (SSFM), than the original frequency-domain NMA method. SSFM is to project the total spectrum energy density data of all phonon modes onto an inaccurate (harmonic or quasi-harmonic) eigenvector base and then manually sum up the peaks that belong to the same phonon mode (at the same frequency). The SSFM relaxes the requirement for accurate temperature-dependent eigenvectors, making it robust for analyzing strongly anharmonic crystals at high temperatures.

Two-dimensional (2D) materials have the potential to be a breakthrough platform for designing next-generation solid-state devices with enhanced performance and novel functionalities. However, the ongoing design concern of heat dissipation is further exacerbated in 2D-based systems due to their weak interlayer van der Waals (vdW) atomic bonding with surrounding three-dimensional (3D), particularly the substrate. Since the primary pathway for heat removal in 2D-devices is suspected to be across the 2D/3D interface and through the 3D substrate, the uniquely poor thermal performance of 2D/3D interfaces is a pressing issue for the development and implementation of 2D-based devices. Therefore, there have been several efforts to experimentally measure 2D/3D TBC as well as to provide explanations of the underlying physical dynamics through theoretical modeling. Simultaneously, materials informatics based on machine learning (ML) algorithms has seen tremendous success in new material discovery and application-specific material selection. However, the existing literature does not hold enough TBC data (both experimental and theoretical) to construct a diverse set of 2D/3D interface pairs which is required to reliably train ML algorithms.
In this work, we use our first-principles-driven TBC model to expand upon available data in the literature by computing the TBC for 156 unique interface pairs, including technologically relevant substrates such as HfO2 and CaF2. Then we employ machine learning to develop a streamlined predictive model that is computationally cheap and demonstrates strong transferability. For single-layer 2D materials our dataset includes dispersions of graphene, silicene, boron nitride, boron arsenide, black phosphorus, blue phosphorus, and 1H-MX₂ transition metal dichalcogenides (with M=Mo, W and X=S, Se, Te). For 3D materials, our dataset includes dispersions of silicon, germanium, diamond, CaF₂, titanium, AlN, GaN, 6H-SiC, Al₂O₃, SiO₂, AlO_x, HfO₂, and PMMA. We then use our 2D/3D thermal boundary conductance (TBC) model to calculate the TBC of each 2D/3D pair from the above lists. We then vary the substrate coupling strength, which is a crucial and highly sample-dependent property, resulting in a total observation pool of 5460 TBC values which are then randomized and split 70/30 for training and testing datasets. We train a range of machine learning models and find that fine tree and wide-neural network (WNN) regression algorithms perform the best out of an array of algorithms (linear, step-wise linear, gaussian process, and support vector) due to their unbiased residuals. We then train fine tree and WNN models on the dependence of TBC on the van der Waals spring coupling constant (K_a) enabling the machine learning algorithm to accurately predict the TBC vs. K_a curve for any 2D/3D interface provided the known list of material descriptors are available and within the extrema of the parameter space.
Our results show that fine tree ensembles and wide neural networks can predict the 2D/3D TBC within +/- 1.41 MW.m^{-2}.K^{-1} and 1.046 MW.m^{-2}.K^{-1}, respectively, with a descriptor set comprised of structural data, select thermal transport properties (e.g., thermal conductivity, Debye temperature, etc.), and the vdW spring coupling constant. The reported R-squared values for either model are 0.98 (fine tree) and 0.999 (WNN). We further show that fine tree ensembles demonstrate good transferability by withholding select materials from the training pool and subsequently testing the trained models on the unseen data. In most cases, the model retains an RMSE < 10 MW.m^{-2}.K^{-1} and R-squared values greater than 0.65. Lastly, we use sensitivity analysis and Pearson correlation coefficients to identify the vdW spring coupling constant, amorphous flag, and resonant frequency gap as the descriptors that are most correlated with the TBC.

Controlling thermal conductivities of materials is important for a wide range of applications, from thermoelectrics for clean energy generation to electronic devices and thermal barrier coatings. The thermal conductivity is commonly estimated using molecular dynamics simulations within the Green-Kubo formulation. This requires a force field that is both 1) an accurate estimate of the interatomic interactions and 2) fast enough to allow simulations with sufficiently large length and time scales. Traditionally, only empirical force fields have fulfilled both of these requirements, which severely limits the applicability of the method.

In this work, we employ the Gaussian Process-based FLARE force field, which automatically learns the interactions of more complex materials than empirical force fields. The resulting model can then be mapped to a low-dimensional, computationally efficient model. Through GPU-acceleration with LAMMPS and the Kokkos library, we achieve excellent performance and obtain well-converged estimates for the thermal conductivity. Furthermore, we investigate state-of-the-art sampling and spectral denoising methods for further acceleration of the simulations.

Understanding thermal transport processes is essential to design materials and devices with desired thermal transport properties for thermal energy conversion and management. Despite the significant progress in thermal transport of simple inorganic crystals, our understanding of thermal transport in complex structures and across interfaces remains limited. In this talk, I will present my research group’s efforts to understand thermal transport in complex structures and across interfaces, using atomic modeling and experimental characterization. I will start with our study on boron suboxides, which have unusually high thermal conductivity despite their complex structures. I will then cover phonon dynamics in 3D and 2D hybrid organic-inorganic perovskites using inelastic x-ray scattering, where we gained useful insights into their ultralow thermal conductivity and remarkably weak anisotropy in 2D hybrid. Last but least, I will share our development of rigorous formalism for anharmonic atomistic Green’s function to capture the contribution of inelastic scattering at the interfaces, direct observation of phonon Anderson localization in aperiodic superlattices, and machine-learning-assisted interface design.

While phase change material (PCM) heatsinks and energy storage modules have been shown to act as highly efficient transient energy storage components in thermal management systems, they have traditionally been limited by the rate at which heat is absorbed into the PCM. Composite thermal energy storage (TES) materials, consisting of highly thermally conductive elements and PCMs can overcome some of the challenges presented by the slow rate of heat absorption by most PCM phases. However, the application of these components is limited by a lack of coherent design guidelines. Here, we present a design framework based around the concept of phase change material (PCM) composites, where the effective thermal properties of a composite are dictated by constitutive relationships relating the composite structure to its transport properties. Under this approximation, relatively simple approximate analytical solutions or numerical solutions reveal the response of PCM composites to different transient heat loads. Furthermore, that response can be optimized, given a particular performance metric of interest (total heat absorbed, heat absorbed per unit mass, etc.).
First, we develop the concept of composite PCM volumes, and demonstrate that the defining criteria of a composite PCM is the critical length scale separating conductive elements. For systems below this length scale, the composite can be defined by effective thermal properties, which dramatically simplifies the description of the system, and computation of transient thermal response. Both numerical and experimental results illustrate that this length scale is on the order of 1 mm for many systems of interest, and thus is easily achievable by a number of manufacturing techniques.
Next, we demonstrate the role of metal volume fraction in governing transient thermal performance by considering heat transfer in both planar and cylindrical systems. Systems which are designed to optimize the total rate of heat transfer into the composite volume (or similarly, to minimize the transient temperature range at an interface) have dramatically higher volume fractions of conductive component than are generally considered. However, if the performance metric is modified to take into account the system volume and mass, the resulting optimal composite structure varies as a function of 1) the thermal boundary condition, 2) the period over which heat is absorbed, 3) material properties, and 4) composite geometry. The outcome of this task is a set of design guidelines for thermal storage composites, valid across a range of geometries, orientations, and material thermophysical properties. Such design guidelines will enable order of magnitude improvements in cooling power densities for TES composites.
Finally, the concept of objective oriented design of PCM composites is presented, and is illustrated with examples based around two design objectives for a finite volume cylindrical PCM composite: 1) Maximizing thermal buffering capacity at a given time, and 2) Maximizing the time the system can achieve a minimum thermal buffering capacity threshold. To this end, we demonstrate that there exist two design regimes where the thermal buffering capacity is either limited by the rate at which the system can absorb thermal energy, or by the total thermal capacitance of the system. We present expressions of the optimal volume fraction for each combination of design objectives, coordinate systems and boundary conditions based on appropriate analytical solutions for the melting problem.

High temperature thermal transport processes are important for a variety of thermal energy conversion and management processes, such as concentrating solar power, thermophotovoltaic, thermoelectric, and thermal storage. While characterization and understanding of high-temperature thermal conductivity of various materials have been studied for several decades, there are still challenges associated with many materials in different forms (bulk, particles, films). In particular, complex high- and medium- entropy ceramic materials have recently emerged as an appealing class of materials due to their various unusual attributes, such as thermal and chemical stability, but their thermal conductivity mechanisms and potential applications as thermal materials are yet to be fully explored. In this work, we study thermal transport properties of both simple and complex (high- and medium- entropy) metal oxides. We use both traditional methods (such as laser flash) and new optical technique to measure thermal conductivity of these oxides in the form of bulk solids, coatings, and particles. The measured thermal conductivity is understood through intrinsic phonon scattering mechanisms from room to high temperature in bulk solids as well as extrinsic effects in coatings and particles.

Many applications such as effective thermal barrier coatings require ceramic materials to have low thermal conductivities at high temperature. In these crystals, phonons are the main heat carriers and thermal conductivity is largely determined by phonon-phonon interactions. We predict this key property of ceramic materials using first-principles calculations, where we consider recent theoretical advances: four-phonon scattering and phonon renormalization. We find these two effects to be significant at high temperatures, a condition where ceramics are expected to work. Our predictions agree well with available experimental data and our own measurements, and can guide thermal barrier coating design for high temperature applications.

A fundamental understanding of thermal boundary conductance is critical for device
cooling, and with ion bombardment as a prominent method of doping of semiconductors, the impact on TBC must be explored. We present a two-part study in which we experimentally measure the TBC across defected / pristine crystalline-crystalline interfaces, and use molecular dynamics to help explain the results.
Experimentally, we measure Gallium Nitride bombarded with varying doses of Helium, Carbon, Nitrogen, and Gallium ions, with an Aluminum capping layer deposited, and we observe a consistent increase in TBC with increasing ion dose. These measurements are performed using Time Domain Thermoreflectance (TDTR), which is a pump-probe technique that measures the change in thermoreflectance on the surface of a sample as a function of pump-probe delay time from picoseconds to nanoseconds. This change in thermoreflectance is related to both the temporal temperature decay from impulse heating driven by the sub-picosecond pump pulse and the frequency-dependent temperature rise induced from the modulated pump pulse train. Thus, TDTR is well suited to measure the thermal conductance across near surface interfaces. Our TDTR apparatus includes an 800nm 80MHz femtosecond pulsed laser, the pump is modulated at 8.4MHz, and all samples included an aluminum transducer deposited on the surface. Computationally, we explore a Silicon / Heavy Silicon system in which Frenkel defects are introduced. We note an increase in TBC when defects are added, and point to an increased level of lattice deformation (in addition to the sparse defects themselves) as leading to an increase in phonon scattering within the material near the interface. This portion of the study uses Non-Equilibrium Molecular Dynamics (NEMD) simulations using Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS). A Stillinger Weber potential is used for the
Silicon - Heavy Silicon system, and the system is initialized with pairs of void / interstitial defects present.

Any power generation through energy harvesting or from renewable sources should be combined with a suitable energy storage system to complement the intermittency of energy sources. Lithium-ion batteries (LIBs) are by far the most popular energy storage option in the global energy storage market. Thermal management is indispensable to control the temperature of LIB within a suitable range for energy-storage applications. For a proper thermal management of LIB, it is important to establish a thermal modeling methodology which is computationally efficient. An LIB cell consists of multiple alternating layers of positive and negative electrode materials, aluminum and copper current collectors, and porous separators impregnated with electrolyte. For the thermal modeling of LIB, we need a systematic way to obtain the thermal properties of the various compartments of the cell components as well as the thermal boundary resistances of the interfaces between the cell components.
In this work, a thermal modeling methodology is presented with a special emphasis to investigate the effect of thermal boundary resistance on the thermal behaviors of an LIB cell. A 63 Ah LIB cell fabricated by LG Chem. is modeled. The density and specific heat capacities are obtained by using the volume-averaged values of the various compartments of the cell components and the effective thermal conductivities are calculated by using the values estimated based on the equivalent networks of parallel and serial thermal resistances of the cell components including the thermal boundary resistances of the interfaces between the cell components. The non-uniform distribution of the heat generation rate in an LIB cell is calculated based on the modeling results of the potential and current density distributions of the battery cell. Thermal modeling of an LIB cell is validated by the comparison between the experimental measurements and modeling results. The modeling methodology presented in this work can be used to estimate accurately the thermal behaviors of various devices used for thermoelectric and thermophotovoltaic energy conversion.

The Raman peak position and linewidth provide insight into phonon anharmonicity and electron-phonon interaction in materials. For monolayer graphene that has emerged as a potential electronic and optoelectronic material, prior first-principles calculations have yielded opposite temperature dependence of the linewidth compared to measurement results. Here, we explicitly consider four-phonon anharmonicity, phonon renormalization, and electron-phonon coupling, all found important to successfully explain both the G peak frequency shift and linewidths in our suspended graphene sample at a wide temperature range. Our calculation shows that the Raman linewidth can be widely tuned by varying carrier doping.

Beta-phase gallium oxide (β-Ga₂O₃), the most thermally stable phase of Ga₂O₃, has stimulated great interest in power electronics due to its ultra-wide bandgap (∼4.9 eV) and high breakdown electric field. The relatively low thermal conductivity of β-Ga₂O₃, however, limits the device performance due to excessive temperature driven by self-heating. Recently, integrating β-Ga₂O₃ thin films on substrates with high thermal conductivities has been proposed to improve heat rejection and device reliability. In this work, we prepare high-quality single-crystal β-Ga₂O₃ thin films by mechanical exfoliation of bulk crystals and study their thermal transport properties. Both the anisotropic thermal conductivity of β-Ga₂O₃ bulk crystals and the thickness-dependent thermal conductivity of β-Ga₂O₃ thin films are measured using the time-domain thermoreflectance technique. The reduction in the thin-film thermal conductivity, compared to the bulk value, can be well explained by the size effect resulting from the enhanced phonon-boundary scattering when the film thickness decreases. This work not only provides fundamental insight into the thermal transport mechanisms for high-quality β-Ga₂O₃ thin films but also facilitates the design and optimization of β-Ga₂O₃-based electronic devices.

From the energy sector to the pharmaceutical industry, there has been a growing need for characterization of thermal properties of materials at the nanoscale. In particular, measuring the specific heat capacity of a material is critical to quantifying heat dissipation in thermal management research as well as optimizing additive manufacturing routines and pharmaceutical development. While traditional methods for determining specific heat capacity, such as differential scanning calorimetry, rely on bulk samples for measurement, it can be both extremely expensive and time-consuming to produce bulk volumes of sample materials.

In this work, we introduce a novel pump-probe thermoreflectance technique for measuring thermal properties of thin films, on the order of nanometers thick. More specifically, we use a boxcar averager to monitor the transient temperature rise of the sample surface as it is periodically heated by the “pump” laser beam. By measuring the laser-induced heating over multiple modulation frequencies we are able to measure not only specific heat capacity, but also thermal conductivity and thermal boundary conductance. We demonstrate the application of the proposed measurement technique by measuring and analyzing standard "film on substrate" samples as well as a collection of low-temperature phase change materials.

Solar energy can alleviate our dependence on traditional energy sources like coal and petroleum. In this regard, the design and performance of solar absorbers are crucial for capturing energy from sunlight. Specifically, for applications relying on solar-thermal energy conversion, it is desirable to construct solar absorbers using scalable techniques that also allow a variation in optical properties. In this study, we demonstrate the ability to tune the spectral absorptance of nickel-infused nanoporous alumina using a scalable and inexpensive fabrication procedure. With simple variations in the geometry of the nanostructures, we enable broadband absorption with a net solar absorptance of 0.96 and thermal emittance of 0.98 and spectrally-selective absorption with a net solar absorptance of 0.83 and thermal emittance of 0.22. The simple manufacturing techniques presented in this study to generate nanoengineered surfaces can lead to further advancements in solar absorbers with well-controlled and application-specific optical properties.

Molten chloride salts are a promising option for heat transfer use in concentrating solar power plants due to their ability to transfer heat efficiently and remain stable at a variety of temperatures. However, these salts are corrosive and therefore break down pipes in the plant. In this study, the aim is to further investigate the solubility of magnesium in MgCl₂-KCl-NaCl salt to reduce the formation of corrosive hydroxide/oxide ions. Redox potential was used to indicate the dissolution of the magnesium in the molten chloride salts at different temperatures. A previous study reported that the addition of the magnesium to the MgCl₂-KCl-NaCl salt resulted in a lower redox potential than that of pure MgCl₂-KCl-NaCl. This study will focus on refining the relationship between the solubility of magnesium and operating temperature. Tests were conducted at 500°C, 600°C, and 650°C. With this data correlating to temperature, concentrating solar power plants can know precisely how much magnesium needs to be added to the molten chloride salt at a certain temperature in order to adequately lower the redox potential and thus result in a less corrosive substance.

Modern electronic and data storage technologies combine shrinking feature sizes with ever-increasing operating speeds, leading to transient nanoscale hotspots that can limit device performance. Quantifying hotspots in operating devices thus requires a non-invasive thermometer with sub-100 nm spatial resolution and sub-100 ns temporal resolution. One promising approach is to optically probe the time-resolved, temperature-dependent luminescence of an individual nanoparticle placed at a specific critical location. Nitrogen vacancy (NV) centers are bright, photostable luminescent defects found in diamond nanoparticles (nanodiamonds, or NDs) that have emerged as exemplar optical sensors with wide-ranging applications, including nanoscale magnetometry, bioimaging, quantum computing, stress mapping, and thermometry. The spin state of NV centers can be coherently manipulated using resonant microwave pulses and read out via the emitted luminescence, an approach known as optically detected magnetic resonance (ODMR). The ODMR signal has well-known temperature dependence, enabling ODMR-based thermometry. However, the temporal resolution of ODMR thermometry is limited to ~10 ms. Furthermore, ODMR measurements require a microwave antenna, an experimental complication that can also interfere with device operation.

To overcome the limitations of ODMR thermometry while retaining the advantages of ND-based sensing, we utilize the temperature-dependent luminescence lifetimes of NV centers in individual NDs for thermometry. NV centers have luminescence lifetimes of ~25 ns, naturally facilitating high temporal resolution measurements. At elevated temperatures, the probability of phonon-assisted non-radiative relaxation increases, thereby decreasing the luminescence lifetime. To demonstrate single-ND luminescence lifetime thermometry, we first developed robust strategies for cross-identifying single NDs via luminescence, atomic force microscope (AFM), and scanning electron microscope (SEM) imaging, since luminescence emission intensity is not a reliable indicator for distinguishing single NDs from aggregates. We then measured the luminescence lifetimes of individual ~70 nm NDs as a function of temperature using an approach called time-correlated single photon counting (TCSPC). TCSPC relies on the repeated detection of single photons over many excitation and decay cycles, allowing us to overcome the challenge of low photon counts associated with small NDs and high temperatures. We assessed the temperature dependence, repeatability, and ND-to-ND uniformity of single-ND luminescence lifetimes. We find that single-ND lifetimes show reliable, sensitive temperature dependence that can be calibrated and employed for thermometry. We also show that AFM-based nanomanipulation can be used to position individual NDs with ~20 nm precision. The combination of high spatial and temporal resolution along with precise positioning control can enable practical applications of single-ND luminescence lifetime thermometry for transient hotspot characterization.

Surface phonon-polaritons (SPhPs), electromagnetic modes excited by optical phonons, are investigated to promote heat spreading in nanostructures.
In the first stage, it will be shown that SPhPs can carry heat more efficiently than phonons in ultra-thin SiN amorphous membranes by measuring thermal conductivity as a function of temperature and thickness by using Time-Domain ThermoReflectance. The significant contribution of SPhPs is proven in the 400K-800K temperature range.
By using the 3w technique, we then show the presence of SPhPs in the 300K-400K temperature range and their quasi-ballistic contribution to thermal conduction over a distance larger than 100 microns to be compared to the few hundred100nms of phonon mean path.
In a second stage, we show that SPhPs also impact heat conduction in silicon devices by measuring the SPhP heat flux in SiO₂/Si/SiO₂ multilayers with a gap. In this latter case, a predominant SPhP contribution is highlighted.

Understanding nanoscale thermal transport is critical for nano-engineered devices such as quantum sensors, thermoelectrics, and nanoelectronics. However, despite overwhelming experimental evidence for non-diffusive heat dissipation from nanoscale heat sources, the underlying mechanisms are still not understood. In this work, we model nanoscale heat flow with both advanced mesoscopic and microscopic approaches to elucidate the fundamental physics of an unexplained and counter-intuitive behavior: decreasing nanoscale heat source spacing actually enhances cross-plane thermal conduction [1].

Using advanced atomic-level simulations, we obtain the microscopic information needed for a fundamental investigation of nonequilibrium phonon scattering amidst interacting heat sources [2]. We observe a reduction in in-plane phonon lifetimes with decreasing nanoheater spacing, as phonons from neighboring heat sources scatter downwards into the substrate. The increased scattering is accompanied by elevated cross-plane thermal conduction – a novel channeling phenomenon that lies beyond many of the traditional Boltzmann transport equation approximations and opens the door to nanoscale engineering of thermally anisotropic materials. We also note the striking similarities between the non-diffusive temperature profiles calculated from the atomistic data and our second, mesoscopic approach.

The kinetic-collective model (KCM) is a hydrodynamic transport model which uses finite element analysis to solve the Guyer-Krumhansl equation over large, device-relevant geometries [3]. While resistive scattering in bulk silicon is traditionally thought to obscure hydrodynamic effects, we observe good agreement between KCM and experimental data for the thermal relaxation of 1 and 2D nanoscale hot spots [4]. In the case of small, effectively isolated heaters, the temperature decay is double-exponential, with distinct time scales for interface resistance and hydrodynamic-like phonon transport. By tuning a nonlocality parameter in KCM to the limiting length scale, we can also predict the increased conduction observed experimentally for closely-packed sources. Moreover, we predict the full thermomechanical physics in ultrafast laser-based experiments, even in complex and general geometries, without the need for geometrical fit parameters, an advance beyond traditional simulations. The combined models explain the counterintuitive experimental observations of enhanced cooling for close-packed heat sources, from atomic to >micron length scales and represent a new fundamental behavior in materials, with far reaching implications for electronics and future quantum devices.
[1] Frazer et al., Phys. Rev. Appl. 11, 024042 (2019)
[2] Honarvar et al., PNAS, accepted (2021)
[3] Beardo et al., Phys. Rev. Appl 11, 034003 (2017)
[4] Beardo, Knobloch et al., ACS Nano 15, 13019 (2021)

Actinide and lanthanide fluorite oxides, UO2, CeO2, ThO2, form an important class of ceramic energy materials with applications ranging from nuclear fuels to solid oxide fuel cells. Of these systems, ThO2 (thoria), has been the least explored and offers unique features including an extremely high melting temperature and a fixed tetravalent oxidation state. Many potential application environments for thoria include high radiation fields or temperatures which either promote or directly generate lattice defects. Such defects drastically influence thermal transport, a controlling safety and performance property in many applications. To date, little experimentally-validated understanding has been generated on the role of defect formation and clustering on thermal transport in thoria due to the lack of high-quality starting material in which these effects may be explored. In this work, high-quality oriented single crystals of thoria are grown using hydrothermal synthesis and then exposed to proton irradiation to generate extrinsic lattice defects (both nanoscale point defects and extended dislocation loops). The thermal conductivity of these damaged crystals is measured direction using a spatial domain thermoreflectance (SDTR) technique and defects are quantified using optical spectroscopy and electron microscopy. Experimental values are compared with the calculation of thermal conductivity from the linearized Boltzmann transport equation for phonons taking into account the relevant scattering mechanisms. This combined experimental and computational approach provides a framework for predicting the thermal performance of this class of materials from first principles in a variety of environmental conditions.

The Seebeck coefficient and electrical conductivity are two central quantities to be optimized simultaneously in designing thermoelectric materials, and they are determined by the dynamics of carrier scattering. We uncover a new regime where the presence of multiple electron bands with different effective masses, crossing near the Fermi level, leads to strongly energy-dependent carrier lifetimes due to intrinsic electron-phonon scattering. In this anomalous regime, electrical conductivity decreases with carrier concentration, Seebeck coefficient reverses sign even at high doping, and power factor exhibits an unusual second peak. We explain the origin and magnitude of this effect using a general simplified model as well as first-principles Boltzmann transport calculations in recently discovered half-Heusler alloys. We identify general design rules for using this paradigm to engineer enhanced performance in thermoelectric materials.

Grain boundaries have a remarkable effect on the thermal and electrical transport properties of polycrystalline materials but are often ignored by prevailing physical theories. Grain boundaries and interfaces can adversely alter the properties of Solar Cells, Batteries and Thermoelectrics. To devise strategies for improving the thermoelectric performance of materials, it is essential to understand the coupled charge and thermal transport mechanisms including an interfacial Seebeck effect. The inhomogeneous nature of materials, such as that caused by grain boundaries, must be taken into account to rethink engineering strategies based on Mathiessen’s rule which interprets scattering homogeneously.
Prevailing models for thermal transport treat interfaces and grain boundaries as structureless even though at the atomic scale they are better described as arrays of linear defects of various types. Allowing for this inherent structure, several fundamental characteristics of heat transport arise, such as diffraction conditions when heat carrying phonons scatter off the periodic, linear defect arrays that should be present in grain boundaries. Furthermore, a dimensionality crossover is observed in diffusive heat transport where phonons with a wavelength longer than the linear defect spacing see the interface simply as a structureless planar defect, and phonons with see the interface as a collection of independently scattering linear defects.

Thermoelectric devices offer the potential to convert wasted heat into electricity, and they can be used for thermal management by pumping heat. The shape of thermoelectric devices, particularly the active semiconducting material units (or thermoelectric legs) within the device, has remained largely unchanged – thus limiting our ability to customize device thermal resistances. The advent of new manufacturing techniques such as additive manufacturing enables customizable device shapes, but there is little knowledge about what shapes are desirable. In this work, we show the effect of different (non-rectangular) thermoelectric leg designs on thermoelectric device performance. Various leg shapes (rectangular prisms, rectangular prisms with interior hollows, trapezoids, hourglass, and Y-shapes) were modeled numerically to determine their thermal and electrical performance under constant temperature and heat flux boundary conditions. Two thermoelectric materials, bismuth telluride and silicon germanium, were modeled to capture low and high temperature applications, respectively. An hourglass-shaped thermoelectric leg subjected to a fixed temperature boundary condition has the best thermal and electrical performance: the hourglass-shaped leg results in more than double the electrical potential and maximum power compared to the conventional rectangular shape. Under a fixed heat flux boundary condition on the hot side, a trapezoid-shaped leg results in almost double the electrical potential and a 50% increase in the power output compared to the conventional leg shape. Varying boundary conditions, which reflect different device operating conditions, result in different performance values for the same leg shapes. These findings underscore the importance of leg geometry on electrical and thermal performance of a thermoelectric leg, as well as the importance of considering the device operating condition when selecting the best leg shape. The results are relevant as new processing techniques emerge and enable topological optimization of thermoelectric devices.

Unique class of 3D framework materials with large polyhedral cages, clathrates, is a promising platform for the development of novel thermoelectric materials due to intrinsically low thermal conductivity. Traditionally the tetrahedral clathrate frameworks are based on tetrels, group 14 elements, Si, Ge, and Sn. In this work we will present recent advances in the design, synthesis, order-disorder structural transitions, and properties of unconventional clathrates with three-dimensional frameworks comprised of oversized metal-pnictide (P, As, and Sb) polyhedral cages that encapsulate guest cations.

Around 60% of primary energy consumption is wasted as heat energy, representing an untapped source of energy with huge potential to reduce our overall consumption.¹ Thermoelectric materials convert heat energy into electricity through the use of temperature gradients. The materials historically used for this purpose include bismuth chalcogenides and lead tellurides. However, issues relating to their natural abundance and toxicity, combined with the increasing demand for environmentally friendly energy sources, is driving research into novel materials with potential thermoelectric properties to replace these.
The effectiveness of a thermoelectric is measured by the dimensionless figure of merit ZT. Materials with a high ZT need to balance a high electrical conductivity and a low thermal conductivity. Oxides generally exhibit properties valuable for thermoelectric applications, including low cost, thermal and chemical stability and environmental benignity. However, their thermoelectric performance has been hindered by their high lattice thermal conductivities. Novel ternary oxides have advanced research in this field as their complex crystal structures possess lower lattice thermal conductivities than traditionally seen in oxides.
In this study, we conduct a detailed investigation into the thermoelectric properties of the ternary wide band semiconductor Sr₂Sb₂O₇, which has previously been synthesised under high temperature conditions and shown to be thermally stable.^2,3 Our study uses the latest approach to separately calculate the different contributions to overall scattering rates, in order to more accurately predict thermal conductivity as compared to the constant relaxation time approximation (CTRA).⁴ We benchmark this approach against the CTRA and show that Sr₂Sb₂O₇ possesses a ZT greater than many existing oxide thermoelectric materials, and has the potential to perform as a high-performance n-type thermoelectric component. We go on to investigate the defect chemistry of this exciting new thermoelectric candidate to further assess its suitability.


¹ M Kanatzidis, Chem. Matter. 2010, 22, 648
² Sato et al., J. Photochem. Photobiol., A 2002, 148, 85
³ Xue et al., J. Phys. Chem. C 2008, 112, 5850
⁴ Ganose et al., Nat. Commun. 2021,12, 2222

Unlike the longitudinal 2-leg thermoelectrics, (p × n)-type transverse thermoelectrics (TTE) function with a single leg hosting orthogonal p- and n- Seebeck behavior. The anisotropic ambipolar thermoelectric behavior is originated from the anisotropic effective masses of electrons and holes, leading to opposite dominant transport types along different crystal axes despite the parallel transport of electrons and holes along both directions. Whereas doped unipolar thermoelectrics have activated conduction to a single band, these (p × n)-type materials can have activated conduction to both bands since both electrons and holes simultaneously contribute to the anisotropic conduction in all directions. As such an important concept here is that of the “partial gap”, i.e. the gap between the Fermi energy and the electron or hole band edges for the n- and p-type partial gaps, respectively. The partial gap is defined as positive for non-degenerate doping, designated n or p, and negative for degenerately doping, designated n⁺ or p⁺. Consequently, we categorize intrinsic semiconductors with the label p × n given the positive electron and hole partial gaps, while the intrinsic semimetal is labelled p⁺ × n⁺ given the heavy doping resulting in two negative partial gaps. There remain two intermediate hybrid cases, namely p⁺ × n and p × n⁺ where one band is degenerately and the other non-degenerately doped. Theoretical analysis and simulations show that among all four types of bulk TTE materials, the semiconducting (p × n)-type TTEs have the largest transverse power factors. By comparison, the metallic transport of both electrons and holes of (p⁺ × n⁺)-type TTE leads to the lowest transverse power factors, though these might be the most promising TTE candidates in extreme low-temperature applications. The dependence of transverse Seebeck coefficient on electron and hole effective mass anisotropy is explored as a function of Fermi energy for different directional effective mass ratios for both electrons and holes. All categories of semiconductor, semimetal, and hybrid (p × n)-type TTE are modeled with band gap E_g = ± 3 k_BT to illustrate peak TTE performance, including Seebeck tensor terms and transverse power factors. We point out unusual scenarios in the hybrid case where, for example, the valence band can be degenerately doped, yet the Seebeck coefficient in one of the transport directions can still be n-type, dominated by activated conduction to the electron band. The introduction of this classification method will greatly help the characterization and performance enhancement of (p × n)-type TTE materials, allowing broader application of this type of emerging TTE material.

Despite the extraordinarily high maximum thermoelectric figure of merit (ZT) ~2.2-2.8 discovered in single-crystal SnSe materials, their performance has been severely debated. Many research groups have observed paradoxically higher thermal conductivity in polycrystalline samples than the reported values for single crystals, leading to much lower ZT. Accordingly, the widely sought goal in the thermoelectric society has been to realizing polycrystalline SnSe samples with comparable to or even higher performance than the single-crystals. In this talk, I will discuss the origin of such higher thermal conductivity and poorer ZT of polycrystalline SnSe samples. Then, I will introduce our facile synthesis process to remedy this problem. Our properly prepared, hole-doped SnSe polycrystalline samples show a record-breaking peak ZT greater than ~3.0 at 783 K and average ZT ~2.0 between 300 and 783 K. Their lattice thermal conductivity is finally ultralow at ~0.07 W m^–1 K^–1 at 783 K, much lower than that of the single crystals.

From the standpoint of pure crystalline materials, phonons have been long considered a solved problem. The standard model is sometimes termed the phonon gas model (PGM), and it provides a useful physical picture for both understanding and predicting behavior of phonons in pure crystalline materials. However, a variety of studies over the last decade have highlighted the fact that when symmetry is broken, for example by compositional or structural defects (e.g., random alloys, amorphous materials and interfaces) the mode shapes can change from what is traditionally envisaged for phonons, namely plane waves. The physics of these alternative types of phonons, how they interact, and how they impact phenomena such as chemical reaction is still being explored, and this talk will review progress made on these questions by the Atomistic Simulation & Energy (ASE) research group at MIT. This talk will also review recent progress on techniques to model the interactions between atoms and phonon transport.

In this talk we will present some of our efforts for characterizing materials transport properties from first-principles simulations. First, we will discuss how first-principles calculations, typically limited to semiclassical models, are unable to capture the electronic transport properties of Bi2Se3, a narrow-gap semiconductor. In fact, we show that transport in this material at low doping concentrations is dominated by Zener tunneling [1], a phenomenon in which carriers tunnel between the valence and the conduction band, rather than just diffusing under the action of the electric field. This transport mechanism can be described using a novel first-principles model based on the Wigner distribution. Surprisingly, we find that Zener tunneling is not limited to low-energy carriers, but occurs also between band subvalleys of energy larger than the band gap.
Next, we will introduce Phoebe [2], a software for computing thermoelectric properties by solving the electron and phonon Boltzmann equation. Phoebe can compute electron-phonon and phonon-phonon scattering properties from first-principles simulations, allowing a fully ab-initio prediction of thermoelectric properties. Additionally, we implemented an efficient mixed MPI and OpenMP parallelization and GPU acceleration, allowing us to take advantage of contemporary computing infrastructure.
[1] A. Cepellotti & B. Kozinsky, Mater. Tod. Phys. 19, 100412 (2021)
[2] https://mir-group.github.io/phoebe/

Topological phase transitions occur when the electronic bands change their topological properties, typically featuring the closing of the bandgap. While the influence of topological phase transitions on electronic and optical properties has been extensively studied, its implication on phononic properties and thermal transport remains unexplored. In this work, we use first-principles simulations to show that certain phonon modes are significantly softened near topological phase transitions, leading to increased phonon-phonon scattering and reduced lattice thermal conductivity. We demonstrate this effect using two model systems: pressure-induced topological phase transition in ZrTe₅ and chemical composition induced topological phase transition in Hg_1-xCd_xTe. We attribute the phonon softening to emergent Kohn anomalies associated with the closing of the bandgap. Our study reveals the strong connection between electronic band structures and lattice instabilities and opens up a potential direction towards controlling heat conduction in solids. This work is based on research supported by DOE under the award number DE-SC0019244 and the UC Santa Barbara NSF Quantum Foundry funded via the Q-AMASE-i program under award DMR-1906325.

Thermal management is essential in many electronic devices, the components of which have been scaled down to the nanometer. When the characteristic size becomes comparable to the mean free path of energy carriers, interactions with boundaries can dominate thermal transport and the quasi-ballistic heat conduction regime arises: Fourier’s diffusive law is no more valid to describe thermal transport and ‘temperature jumps’ appear at contacts. This heat conduction regime takes place for both phonons and electrons, however these two types of energy carriers do not have the same mean free path. In addition, it is known that electron and phonon temperatures can differ under some circumstances, when they are driven out of equilibrium.
In this contribution, we investigate the deviation from the diffusive regime in thin films. To this end, we solve the Boltzmann Transport Equation (BTE) both for phonons and electrons using the Discrete Ordinate Method (DOM) [1-4], in particular in the isotropic-crystal and single relaxation time approximations. A coupling term reminiscent of the two-temperature model is added in the BTE equations [5]. This method accounts for the directional and non-local aspects of the energy flux, which can be tuned by the electron-phonon coupling constant. We analyze the impact of the electric field and the geometry on the flux transport, and compare the results with diffusion-based energy/charge conservation equations.
These results shed light on the relation between local energy non-equilibrium and transport properties.


Références :
[1] A. Majumdar, J. Heat Transfer 115, pp. 7-16 (1993)
[2] S. Volz, D. Lemonnier, J.-B. Saulnier, Microscale Thermophys. Eng. 5 (3), pp. 191-207, (2001)
[3] Thermal transport phenomena beyond the diffusive regime, P.-O. Chapuis et al., Proceedings of MIXDES (2016)
[4] A. Alkurdi et al., Phonon Boltzmann Transport Equation and its EPRT approximation: lessons learned and outlooks, submitted
[5] A. Pattamatta and C.K. Madnia, Journal of Heat Transfer, 131(8), 082401 (2009)

Acknowledgements:
The support of project EFINED (H2020-FETOPEN-1-2016-2017 766853) is acknowledged. We address a special thank to D. Lemmonier (Pprime), S.Merabia (ILM), A. Varpula and M. Prunnila (VTT).

In recent years, there have been numerous efforts to synthesize materials with ultralow thermal conductivities, a crucial parameter in the development of thermoelectric materials, memory devices, and thermally protective coatings. It has been generally believed that amorphous solids possess the lowest thermal conductivity possible. The heat transport mechanisms in ultralow thermal conductivity materials are often described using formalisms originally put forth by Einstein, and later refined by others, which account for some degree of localization of the vibrational modes or strong suppression of vibrational scattering length scales. These concepts, which partially form the basis of analytical minimum thermal conductivity models, are able to successfully predict the thermal conductivity of a wide range of amorphous solids. With advances in nanofabrication, several studies have shown that, by reducing the mean free paths of the vibrational modes, the thermal conductivity of crystalline materials can be significantly lower than this aforementioned minimum limit. In amorphous materials, however, manipulating the atomic structure to reach values below the minimum limit is more complicated, as it is already in its highest disordered state. Thus, a question remains: how can the thermal conductivity of a fully dense amorphous solid be further reduced?

Here, we report on a mechanism to suppress the thermal transport in a representative amorphous chalcogenide system, silicon telluride (SiTe), by nearly an order of magnitude via systematically tailoring the cross-linking network among the atoms. As such, we experimentally demonstrate that in fully dense amorphous SiTe the thermal conductivity can be reduced to as low as 0.10 ± 0.01 W m⁻¹ K⁻¹ for high tellurium content with a density nearly twice that of amorphous silicon. Using ab-initio simulations integrated with lattice dynamics, we attribute the ultralow thermal conductivity of SiTe to the suppressed contribution of extended modes of vibration, namely propagons and diffusons. This leads to a large shift in the mobility edge - a factor of five - towards lower frequency and localization of nearly 42% of the modes. This localization is the result of reductions in coordination number and a transition from over-constrained to under-constrained atomic network.

The frequency domain perfectly matched layer method is used to study phonon scattering and localization in two and three dimensional domains of randomly embedded nanoparticles. We show the FDPML method has the capability of capturing mode-by-mode phonon transport in domains with up to 100 million atoms with the use of parallel computing. From simulations of phonon-nanoparticle scattering across a wide range of volume fractions of nanoparticles, in the Mie and Rayleigh regime, we show that the elastic mean free path scales inversely with the number density of nanoparticles, for volume fractions as high as ~23% before exhibiting dependent scattering effects which causes the mean free path to increase again. We also observe phonon localization by using the Landauer method in 2D and 3D domains of random nanoparticles, at volume fractions up to ~40%. The modal decomposition method is also used to study localized modes in 2D and the results of the two approaches are compared.

Since both sides of a metal-metal interface can support high intensity of electron emission and absorption, thermal transport at metal-metal interface is dominated by electronic carriers. Metal-metal thermal interface conductances (ITC) are at least 1-2 orders of magnitude higher than for interfaces that contain at least one dielectric material, yet the current state of knowledge of thermal transport across metal-metal interfaces remains surprisingly underdeveloped, with just a single simplified model - the electron diffuse mismatch model (eDMM) in the parabolic band approximation - having been employed and compared against experimental data. Here we develop several more advanced models of electronic transmission that can handle diffuse and coherent transport, using realistic band structures derives from first principles calculations. Using a tetrahedron interpolation method, we develop an accurate full-dispersion, multiband implementation of the electronic diffuse mismatch model that is particularly useful for capturing transport in transition metals, intermetallics, and transition metal compounds (XN, XC, XB2, X= transition metals); even in the case of Al-Cu we show that previous apparent experimentaly agreement with the eDMM is less convincing when using realistic band strucure, and that agreement found in previous studies on Ir-Pd, hinges on the inclusion of electron-phonon coupling corrections to the electronic heat capacity. We give eDMM preditions for ITCs of a variety of interfaces types both with and without inclusion of electon-phonon coupling corrections. We also present two coherent electron transmission frameworks: (1) an analytic approach valid for nearly free-electron metals and (2) and NEGF approach appropriate for epitaxial metal-metal interfaces.

Identifying new materials that have a large thermoelectric response would be beneficial for global energy production by converting waste heat into useful electric power. While conventional thermoelectric have regularly faced challenges with regards to vanishing electronic entropy at lower temperatures, topological materials offer an alternative pathway to surpass them through the topological protection of electronic states. Notably, theories have proposed that topological semimetals can large to a large, non-saturating thermopower and quantized thermoelectric Hall conductivity at the quantum limit. We experimentally demonstrate these unique features in topological Weyl semimetal tantalum phosphide (TaP), showcasing an ultrahigh longitudinal thermopower and a giant power factor which are predominantly attributable to the quantized thermoelectric Hall effect. Our findings demonstrate a confluence in the pursuit of effective thermoelectrics and novel topological materials in addition to the realization of the quantum thermoelectric Hall effect towards low-temperature energy harvesting applications.

Han, F., Andrejevic, N., Nguyen, T., Kozii, V., et al. Quantized thermoelectric Hall effect induces giant power factor in a topological semimetal. Nat Commun 11, 6167 (2020).

Hyperbolic materials (HM) have unique properties of opposite signs in the principal components of dielectric permittivity in frequency ranges called Reststrahlen bands. These properties enable a significantly greater number of propagating radiation modes inside the bulk material compared with other materials, called hyperbolic phonon polaritons. Hexagonal boron nitride (hBN) is one of the natural HMs with low loss and supports propagation of hyperbolic phonon polaritons in the Reststrahlen bands. Thermal energy can excite hyperbolic phonon polariton modes in hBN at around room temperature and above. We study contributions of hyperbolic polariton modes to energy transport inside bulk hBN. Using the fluctuation-dissipation theorem, a theoretical model is derived to compute the energy transport by the hyperbolic modes. This model employs the many-body scattering method to account for radiative interactions within the bulk, across an interface, and across a small gap between two surfaces in the near-field radiation regime. The calculation results showed that hyperbolic polaritons in the Reststrahlen bands contribute to a significant amount of energy transport inside the bulk. The model also revealed that the equivalent radiative thermal conductivity by phonon polariton increases with temperature increase and can be comparable to phonon thermal conductivity at elevated temperatures. Experimental measurements conducted over a temperature range between 300 K - 600 K agree with the result of the theoretical model. Additionally, enhancement of near-field radiation due to the bulk phonon polaritons is also illustrated. In additional to hBN, other hyperbolic materials are also being investigated. This work reveals the significance of radiative transfer as a new energy transfer mode inside the hyperbolic materials and across an interface.

Dynamic management of heat loads in building envelopes can utilize polymeric photonic mirrors as dynamic thermal switches. This work investigates both the tunable thermally radiative and thermally conductive properties of photonic mirrors as a function of fabrication parameters and humidity. In particular, a sequential sol-gel process is used to fabricate distributed Bragg reflectors (DBRs) consisting of periodic multilayers of hydrophobic Polymethyl methacrylate (PMMA) and hydrophilic Poly Vinyl Alcohol (PVA). Variable volume fractions of optically transparent titania hydrate (TiO_x) are added to the PVA, forming a hybrid material that increases the refractive index contrast between layers and is moisture sensitive. UV-vis spectroscopy is employed to quantify the dynamic optical responses from polymer swelling obstructing the periodic conditions required to achieve constructive (reflective) interference. Corresponding changes to the films thermal conductivity using a differential 3 Omega technique are also observed beyond film swelling resulting from the degree of moisture uptake that is hypothesized to be effect by the degree of titania hydrate crosslinking. The results of this work provide novel insight into the dynamic nature of polymeric photonic mirrors and suggest applications for these materials as energy saving building technologies.

Solar thermal energy has been one of the major energy sources to realize zero energy buildings for a sustainable future. The high efficiency and low cost make it particularly attractive to reduce the energy consumption for hot water and space heating. On the other hand, the radiative cooling uses the rest of the sky, i.e. the outer space, as the cold source to generate cooling that can reduce the usage of refrigeration and space cooling. Recent progress in photonics and thermal science have further pushed this field to daytime subambient cooling, which created vast opportunities for sustainability and energy. Both technologies showed rapid improvement in the past decade, which become the enablers for the next generation smart net zero energy buildings.
In this talk, I will introduce our recent research progress of switchable building envelope that can change between solar heating and radiative cooling. The dual-mode switchable device can be applied to buildings at various climate zones (spatial variation) under different weather or demand (temporal variation). I will give two examples of the dynamic envelope: (i) The mechanically switchable device can produce ~600 W/m² of solar heating or ~100 W/m² of radiative cooling with less than one minute switching time. This high performance is the outcome of heat transport analysis and thermal interface engineering. we perform building energy simulation for all the climate zones in the U.S. The results show our dual-mode device, if widely deployed, can save 45.8% of the heating, ventilation and air conditioning (HVAC) energy, which is 2.46 times higher than cooling-only and 1.43 times higher than heating-only approaches. (ii) The electrochromic device has continuous tunability and operates without any moving part. Rational photonic modeling and electrochemical cell design were implemented to achieve synergistic multispectral (solar + mid-IR) radiative heat management. At solar heating state, the device exhibits high solar absorptivity and low thermal emissivity; whereas at cooling state, the device shows high solar reflectivity and high thermal emissivity.

Polymeric stabilizers derived from imidazolium-based ionic liquids have been developed to prepare high concentration nanocarbon dispersions in water to yield aqueous dispersions up to 17 percent in MWCNT (and dispersions 1.4 percent in SWCNT and 6.4 percent in graphene). Stable MWCNT dispersions require a stabilizer/MWCNT weight ratio of 0.2 - 0.3 for effective stabilization in water. MWCNT concentration variations demonstrate slow equilibration effects that affect stabilizer demand, which increases with concentration. Highly scalable and anisotropic templated coatings have been prepared by drawdown-bar coating these aqueous dispersions on mulberry paper and fiberglass tissue substrates. Electrical and thermal conductivities for such coatings exhibit in-plane/through-plane anisotropies of up to 200 and 2,000, respectively. Electrical conductivities measured by a four-point method and broadband dielectric spectroscopy yield high but unremarkable in-plane conductivities in the 1 - 10 S/cm range, consistent with stabilizer-resistive connectivity between MWCNT. Ultrahigh thermal diffusivities (1,000 - 6,000 mm²/s) were obtained by a Fourier-transform analysis of Xenon flash data. Heat capacities of coatings measured by modulated differential scanning calorimetry and coating densities measured gravimetrically along with such diffusivities yield thermal conductivities in the range of 0.5 - 3 kW/m/K. These conductivities appear to be the largest observed for MWCNT assemblies produced outside a vacuum chamber and 500-fold higher than available commercially. The stabilizer-induced electrical resistance at MWCNT junctions appears not to impede phonon transport across such junctions significantly. This thermal transport is likely related to stabilizer binding to the MWCNT sp² surfaces by π-orbital overlap. But bulk phonon transport through stabilizer assemblies seems unlikely. Photoacoustic experiments to examine thermal transport by such stabilizers are proposed. However, the de facto highly efficient phonon transport between MWCNT is more likely due to stabilizer rearrangement, promoting physical contact, and near contact at connecting junctions.

Cooling consumes a great portion of energy in both residential and commercial applications. Passive radiative cooling can provide cooling power without any energy consumption by directly emitting heat to deep space. Conventional air conditioners only move the heat from the inside of the space to the outdoors while consuming electricity. Hence, passive radiative cooling holds great promises for addressing the global warming effect. A recent review paper on commercial heat reflective paints shows solar reflectance around 80% to maybe 92%, but none of these paints has achieved full daytime below-ambient cooling. To achieve passive daytime radiative cooling, current state-of-the-art solutions often utilize complicated multilayer structures or a reflective metal layer, limiting their applications in many fields. Attempts have been made to achieve passive daytime radiative cooling with single-layer paints, but they often require a thick coating or show partial daytime cooling. Therefore, it is a pertinent task to develop radiative cooling paints similar to commercial paints, since they provide ease of use, low cost, scalable fabrication, and are able to take advantage of the reliability engineering in the paint industry.

Here we have demonstrated that with the right fillers under certain concentrations and a broad particle size distribution, we can achieve below-ambient cooling in paints under direct sunlight. Monte Carlo simulations show consistent results with experimental characterization, validating the strategies we adopted to achieve high radiative cooling power. One of the designs showed a solar reflectance of 98% among paints and a high sky-window emissivity of 0.95. Outdoor tests exhibited that our paints showed comparable or better cooling performance than the other state-of-the-art approaches while offering unprecedented benefits including the convenient paint form, a fraction of the cost of the other approaches, and compatibility with the commercial paint fabrication process. A system figure of merit is defined in this work to evaluate the cooling power performance of different radiative cooling systems regardless of weather conditions. Additional testings of viscosity, abrasion resistance and outdoor weathering indicate the potential of the paints as a replacement for commercial white paints. Due to the abundance of the filler materials and a similar fabrication cost with commercial paint, our cooling paints are expected to save $0.7 to $1.5 for a moderate 100m2 apartment during a summer sunny day compared to commercial white paints.

Recent advances in integrated circuits require integration of multiple chips in the same package as well as 3D chip modules with multi-layer metallization, copper through silicon vias and more complex heterogeneous integration. Heat dissipation in such devices require careful optimization of bonding interfaces, thermal interface materials and heat sink integration. During operation, small defects or voids can create hot spots and overtime reduce the chip performance and reliability. It is advantageous if the integrity of complex ICs could be characterized during manufacturing before full packaging and powering of the device which require integration on test boards and complex electrical work load. Laser heating and transient thermoreflectance full field imaging is used to characterize heat flow in 3D ICs. While inverse problems in heat diffusion are ill-posed, transient temperature profile provide valuable information about in-plane and out-of-the-plane material and interface properties. Applications to material characterization during manufacturing will be also discussed.

The recent push for the “materials by design” paradigm requires synergistic integration of synthesis, characterization, theory, computation and data-driven approaches. Among these, techniques for efficient measurement of thermal transport are still challenging, especially for diverse materials with a range of roughness, porosity, and anisotropy. Traditional contact methods like steady-state, 3ω, and transient plane source methods suffer from relatively slow throughput, sometimes the requirement of large sample volumes, and the lack of spatially resolvable property mapping capability for parallel measurements. Non-contact pump probe laser methods (e.g., TDTR, FDTR) often need mirror smooth sample surfaces and serial rastering for property mapping which can be inefficient when applied to larger sample areas. Here, we present a new thermal characterization tool which has the potential to significantly improve the measurement throughput based on Structured Illumination with Thermal Imaging (SI-TI). SI gives convenient software control of the shape and frequency of heating light pattern, and thermoreflectance based TI enables efficient large area temperature mapping with good temperature and spatial resolutions. Experimental results will be presented showing how the SI-TI method enables: (1) paralleled measurement of multiple samples in a single field of view without any raster scanning; (2) software control of the dynamically adjustable heating pattern, which can be tuned to optimize the measurement sensitivity in different directions for anisotropic materials; and (3) can tolerate porous, rough, and/or scratched sample surfaces. This work highlights a new avenue in efficient thermal characterization for “materials by design.”

Granular media such as Carbo ceramic particles and silica sands have been used as heat transfer media in heat exchangers for solar-thermal energy storage systems. However, thermal transport processes in the particulate heat exchangers are still not fully understood. In particular, it is still not clear whether and how the thermal conductivity of the particles in the near-wall region could be impacted by the particle -wall interaction. While numerous models exist, there is a lack of suitable experimental technique to probe the thermal transport processes of granular media when they are flowing in a heat exchanger. In this work, we report the instrumentation development of high-temperature thermal conductivity measurement of flowing ceramic particles within a confined channel. We use an existing technique called modulated photothermal radiometry (MPR) and first develop it for high temperature measurements (up to 800 deg. C). MPR is a non-contact technique where the surface of the sample is heated by an intensity-modulated laser and it utilizes the intrinsic thermal emission from the specimens for thermometry, which is favorable for the measurement at high temperature in harsh environment. The high-temperature MPR setup is validated by measuring bulk and thin coating materials with known thermal conductivity. We then develop the technique for measuring both stationary and flowing ceramic particles. The measurement results are compared with modeling based on the discrete element method (DEM) to understand thermal transport in both bulk and near-wall regions of flowing granular media.

As an updated version of the ultrafast pump-probe laser technique, the time-resolved magneto-optical Kerr effect (TR-MOKE) metrology can capture the magnetization dynamics of materials with superb temporal (sub-picosecond) and spatial (diffraction-limited beam spot) resolutions. With optical excitation and detection, TR-MOKE can probe spin precession at high resonance frequencies (up to a few hundreds of GHz), beyond those achievable by conventional Ferromagnetic Resonance (FMR) approaches. Thus, it is a powerful tool to reveal the rich physics of magnetization dynamics of magnetic thin films, in addition to thermal transport properties. In this talk, I will highlight several representative applications of the TR-MOKE metrology for capturing the magnetization dynamics in technologically important spintronic materials. Examples include the studies of temperature-dependent Gilbert damping, the spin-strain coupling, and the interlayer exchange coupling of materials with strong perpendicular magnetic anisotropy (PMA). The structure-property relationships of these emerging materials established by TR-MOKE open up opportunities of tailoring material properties by structural engineering for applications in spintronic and memory devices with low energy consumption, high thermal stability, and fast switching.

Quantitative mapping of temperature fields at the nanoscale is critical in various areas of scientific research and technology, such as nanoelectronics, surface chemistry, plasmonic and quantum systems. Scanning Thermal Microscopy (SThM) techniques, developed in the past, enable temperature mapping with ~10 nm spatial resolution. The main challenge in achieving quantitative thermal imaging with SThM is a lack of knowledge of the tip-sample thermal resistance (R_TS) which is critical for quantifying the sample temperature. A recent advance in SThM enabled the quantification of R_TS in situations where the temperature field is modulated. However, such an approach is not directly applicable to situations where the device operates under conditions where it is not readily possible to modulate the temperature field. Here, we present a novel SThM measurement technique using custom-fabricated scanning thermal probes (STPs) with a sharp tip and an integrated heater/thermometer where the steady-state temperature field of the device can be quantitatively mapped in a single scan with <5 nm spatial resolution and ~50 mK temperature resolution. This is accomplished by introducing a modulated heat input to the STP and measuring the AC and DC response of the probe temperature which allows for simultaneous mapping of tip-sample thermal resistance and sample surface temperature. The advances presented here can greatly facilitate temperature mapping of a variety of microdevices under practical operating conditions.

Molten salts are being used or studied for thermal energy conversion systems, such as concentrating solar power and nuclear power. Thermal conductivity is an important thermophysical property dictating the performance of these systems but its accurate measurement of molten salts has been challenging. The corrosive and conductive nature of these fluids makes it difficult to use traditional contact methods such as the hot wire method. Techniques that require a large volume of the fluids, such as the steady-state method, could also be prone to errors caused by natural convection. Commonly used non-contact optical thermoreflectance techniques are difficult to operate at high temperatures. Here, we report the thermal conductivity measurement of molten salts using a modulated photothermal radiometry (MPR) technique. MPR is a laser-based, non-contact, frequency domain method that has been used to measure thermophysical properties of various materials. We develop the technique for the measurement of molten salts at both stationary and flow states. Stationary molten salts, including nitrate and chloride salts, have been measured using a high-temperature MPR holder. The measurement results are compared with those obtained from a laser flash analyzer (LFA). Our results demonstrate the reliability of the MPR method to measure the thermal conductivity of stationary molten salts. We further extend the technique to perform in-operando measurements on flowing fluids and show that the intrinsic thermal conductivity can be reliably obtained under certain experimental conditions. Finally, the in-operando technique is applied to a laboratory-scale molten salt loop to demonstrate the capability of measuring flowing molten salts in real-time. Our work shows that the MPR technique is a convenient tool to provide reliable molten salt thermal conductivity data and could serve as a diagnostic tool for flowing molten salts.

This work describes a new material deposition process to synthesize graphitic layers using a custom system that employs a concentrated xenon light source (3.5 kW optical) that simulates solar irradiance. A peak heated zone of approx. 3 cm diameter (q_max = 3.2 MW/m²) can heat samples up to 1000°C in less than 1 sec, orders of magnitude faster than electric heaters integrated within conventional CVD systems. We have found clear signs of surface graphitization on organic samples exposed in vacuum. Recent tests on pulsed heating at ~1 Hz indicates further graphitization via a controlled heating/cooling process that avoids pyrolysis. Significant advantages of this unique approach include: (1) ultrafast heating and cooling of substrates, allowing more flexibility to tune physical and electrical properties of the graphitic layer; (2) improved graphitic growth speed because of the ultrafast response to concentrated light heating; (3) ability to build a substantial depth of graphitic layering due to the optical transparency that emerges in the upper layers that transform under pulsed heating; and (4) low cost and easy serviceability of concentrated light as compared to laser-induced graphene (LIG) systems that in theory could produce similar results. Process optimization for this problem will be important, and as such, we apply adaptive Bayesian experimental design as in prior work to maximize a heat spreading metric (a combination of thermal conductivity and layer thickness). Parameters to be optimized include: gas composition and pressure; light pulse intensity, duration, and pulse duty cycle; and substrate type. The output metric, thermal spreading capability, is captured to first order by the product of film thickness and thermal conductivity. Here, we employ an extensive set of capabilities for heat conduction testing and modeling, including a highly accurate, rapid screening approach for heat spreading.

Polymer linked nanoparticle networks are important to many applications, such as advanced composite materials and nano-phononics. Developing networked materials capable of emulating neuronal networks for information processing is critical to the construction of a brain-like cognitive computing system. In this work, we elaborate heat transfer mechanisms between polymer linked gold nanoparticle (AuNP) dimers. We find the effects of different polymer types on heat flux, using 6 different polymers, viz. polyethylene (PE), polyacetylene (PA), poly(p-phenylene) (PPP), polyacene (ACE), polythiophene (PT) and poly(3,4-ethylenedioxythiophene) (PEDOT). We also uncover the effects of AuNP size on heat flux, using three different AuNP sizes: 2, 4 and 8 nm diameter. By steering the AuNPs to different distances, the heat flux is compared between a straight polymer conformation and a loose (or bended) polymer conformation. The effect of heat transfer channels on the total heat flux between AuNP dimers is revealed using 1, 2 and 3 polymer links. The discussion of AuNP facet effect is also included using different linked facets: (111)-(111), (111)-(001) and (001)-(001). Dimers made of different AuNP sizes --- viz. 2 nm AuNP linked 8 nm AuNP --- can have different heat fluxes towards two directions like a thermal diode. Finally, we use four simple networks constructed from 3, 4, 5 and 6 AuNPs, to mimic neuronal networks for information processing. In our models, the information is recorded by the thermal potential (temperature)—that is, the driving heat flux—in analogy to the neuronal networks where the electric potential drives the ion flux. We expect that this work can help improve our understanding of heat transfer mechanisms inside polymer linked nanoparticle networks. Our simulations can also provide guidance for developing new materials and devices of use in brain-like computing.

With progressing climate change and global warming, we face the challenge to reduce our energy consumption and CO₂ emission. In Europe, more than one third of the total energy consumption and CO₂ emission result from the built environment, with energy for heating and cooling of buildings as a major share. A recent study analyzing the energy and CO₂ emission savings potential, has shown that by equipping all buildings in Europe with commercially available energy-efficient glazing the total energy consumption of the building sector can be reduced by over 30% for most countries. Current energy efficient glazing systems are static and therefore best suited for either hot or cold climates. In climates that vary between hot summers and cold winters, the windows have to adapt to the environment for optimized energy efficiency. Thermochromics can be used for adaptive glazing since they can autonomously switch their solar infrared (IR) transmission depending on the glazing temperature. In a recent study we have shown that thermochromic windows are perfect for intermediate climate regions. Using building energy simulations, we demonstrated that energy savings of up to 22% could be achieved with our thermochromic smart window, outperforming commercially available energy-efficient glazing systems.
For a more detailed analysis of the performance of thermochromic smart windows in the Netherlands, we here present a study on the 4 main building types: free-standing buildings, duplex houses, terraced houses and apartments. In our labs we have developed a thermochromic coating with excellent optical properties for application in smart windows. The coating shows high visible transmission (T_vis) of 74% combined with a large solar modulation (ΔT_sol) of 10%, which as a combination is unrivaled by thermochromic VO₂ based coatings reported in literature to date.
For the building energy simulation study we modeled the annual energy consumption for heating, cooling and lighting of the 4 selected building units equipped with our smart window, where our thermochromic coated glass is combined with a commercial low-e coated glass to yield a double glass window, and compared the results to building units with uncoated double glazing.
The free-standing building with the largest floor area and all 4 outside walls in contact with the environment showed the largest benefit of the smart window on annual energy consumption. Here annual energy savings of 21.2% could be reached. With reduced floor area and either three or two outside walls being in contact with the environment, for duplex and terraced houses, respectively, the impact of the smart window on total energy consumption was lower. Here annual energy savings of 16.1% and 15.4% were achieved for the duplex and terraced house, respectively. The apartment was a more special case, since it is a one floor unit with large window surfaces on the two outside walls in contact with the environment. Here the large window surface area in relation to the total floor area leads to a big impact of the thermochromic smart window on annual energy consumption. Here annual energy savings of 29.4% could be reached in comparison to non-coated insulating glass units (IGUs).
Suggesting that natural gas was used for heating and electricity for cooling and lighting, the average cost price for gas and electricity in the Netherlands was used to calculated annual cost savings for each building type. Here annal cost savings of up to 445 € could be calculated. Due to the cheap material and processing costs a return on invest within 7 years will be possible with attributed annual cost savings of 6-7.5 €/m².
Furthermore, we have analyzed the impact on CO₂ emissions the smart window can have in the Netherlands. By extrapolating the data to the complete Dutch built environment, we could show that annual CO₂ emissions for building operations can be reduced by 46%. This means annual CO₂ emission savings of 5.8 Mt for the Dutch built environment.

With most of the primary energy generated in the USA being thermal in nature, developing materials that can reversibly store thermal energy is critical to more efficiently using and managing thermal energy. Metal–organic phase-change materials, which consist of meltable metal coordination complexes with organic ligands, offer an opportunity to expand the library of existing phase-change materials for thermal energy storage and further establish fundamental structure-property relationships to design energy-dense thermal energy storage materials. Herein, we study an isostructural series of divalent metal amide complexes featuring expanded hydrogen bond networks that can undergo tunable, high-enthalpy melting transitions over a wide temperature range. We investigate the influence of coordination bonds, hydrogen bonds, neutral ligands, and outer sphere anions on the phase change thermodynamics, studying how metal dependent hydrogen and coordination bond breaking across the melting transitions alters the melting thermodynamics of metal–organic phase-change materials. Importantly, despite having lower volumetric hydrogen bond densities than the structurally similar metal-salt hydrate systems, the metal–organic phase-change materials investigated here show similar volumetric energy densities to metal-salt hydrates, indicating that metal–organic phase-change materials offer alternative pathways to designing new thermal energy storage materials.

Large components of the primary circuit of nuclear reactors, made with 16MND5 low alloy bainitic steels, operate at 300°C under a pressure of 155 bars, undergoing thermal aging and irradiation, and must retain their mechanical properties throughout the plant life. A thorough understanding of the links between microstructure and mechanical properties is therefore crucial to be able to predict their evolution under operating conditions. Carbides are known to play an important role in explaining the mechanical properties of the steels. [1–3] Their population, size and distribution, evolve mainly during temper heat treatments performed after a step of austenitisation and quenching. However, after quenching, chemical heterogeneities can appear, locally affecting the microstructure, the carbide population, hence the mechanical properties. Therefore, it is necessary to understand the impact of alloying elements on carbide populations to be able to predict the mechanical properties of these heterogeneous areas. Furthermore, understanding the impact of alloying elements on the carbide precipitation sequence can lead to an optimisation of the heat treatments, and the design of steels with enhanced mechanical properties, as well.
Amongst the three main alloying elements (Mn, Ni, Mo) of 16MND5 steels, molybdenum, and manganese to a lesser extent, contribute to define the carbide population. In this context, to isolate the contribution of each alloying element, high-purity model alloys, FeCMo, FeCMn and FeCMoMn were investigated, which compositions were chosen close to the industrial alloy.
Two tempering temperatures were explored, 650 and 700°C, for times of several seconds up to two months. The starting microstructures of the model alloys were selected to be as close as possible to the 16MND5 steel bainite, using continuous cooling transformation diagrams and the associated examinations.
The carbide crystallography and chemistry were determined by TEM and STEM/EDS on carbon extractive replicas, while their size distributions and morphologies were obtained by SEM image analysis. Their volume fraction and lattice volume were measured by ex-situ synchrotron X-ray diffraction.
It resulted in the complete characterization of the precipitation sequences and in the quantification of the impact of Mo and Mn on these precipitation sequences. The use of thermodynamic calculations gave valuable information on the carbides thermodynamical stability.
At last, these precipitation sequences were modelled using the Prisma module from the Thermocalc suite and compared to the experimental results obtained within this study.

[1] Y. R. Im, Y. J. Oh, B. J. Lee, J. H. Hong, and H. C. Lee, Effects of carbide precipitation on the strength and Charpy impact properties of low carbon Mn-Ni-Mo bainitic steels, J. Nucl. Mater., 297 (2001) 138–148, doi: 10.1016/S0022-3115(01)00610-9.
[2] P. Bowen, S. Druce, and J. Knott, Effects of microstructure on cleavage fracture in pressure vessel steel, Acta Metall., 34 (1986) 1121–1131, .
[3] K. J. Irvine and F. B. Pickering, The tempering characteristics of low-carbon low-alloy steels, J. Iron Steel Inst., 194 (1960) 137–153, .

A model is developed to predict how the effective thermal boundary conductance (TBC) of a metal cap/metal contact/dielectric junction varies with the contact thickness. Two-temperature model molecular dynamics (MD) simulations are applied to qualitatively recreate the experimental observation that the junction TBC increases and then saturates as the contact thickness increases. The MD simulations reveal a strong correlation between the junction TBC and the fraction of the electron-phonon coupling that occurs in the contact versus that in the cap. This correlation is then combined with insights gained from an analytical solution to the two-temperature model in the cap and contact to propose a model that predicts how TBC varies with contact thickness. The model, which contains no fitting parameters, is validated against more than one hundred experimental measurements from the literature on a variety of cap/contact/dielectric junctions. By normalizing the TBC and the contact thickness, all the experimental data collapse onto a single curve, with 92% of it lying within 10% of the model. Through physically-motivated approximations, the model reduces to a simple thermal circuit that maintains high predictive ability. Population-weighted phonon density of states extracted from the MD simulations suggest that the TBC contact thickness dependence is strongly influenced by electron-phonon coupling in the contact. The model provides guidance for streamlining the design of thermally-efficient electrical contacts.

Accurate measurements of interfacial thermal conductance (G) of two-dimensional (2D) materials are imperative for effective thermal management of 2D material devices. Despite the importance, there is a large discrepancy in the literature on G of interfaces of 2D materials. Here, we accurately measure G of oxide/MoS₂/oxide and oxide/graphene/oxide interfaces for three oxides (SiO₂, HfO₂, Al₂O₃) by differential time-domain thermoreflectance (TDTR). We find that TDTR measurements of G of these interfaces are 2-to-4 times larger than previously reported G of the similar interfaces measured by Raman thermometry. Using a simple model, we show that the observed discrepancy can be explained by an additional internal thermal resistance between non-equilibrium phonon modes in Raman thermometry measurements. Moreover, the additional internal resistance in Raman thermometry measurements also leads to a stronger temperature dependence of G of 2D materials. Our work provides important guidelines for correct interpretation and analysis of Raman measurements of heat transport across interfaces of 2D materials.

The fabrication of nano-architected materials has improved considerably over the last decade. It is now possible to synthesize series of nanoporous materials with functionalized internal surfaces. However, the impact of humidity on heat transfer at the nanoscale is not fully understood. The structural and vibrational properties of nanoconfined water differ from those of the bulk near the solid/liquid interface [1]. Thus, for materials with nanoscale features, macroscopic models such as Fourier's law or the Effective Media Approach (EMA) can not predict the effective thermal properties of solid/liquid composites [1]. In addition, experimental measurements are difficult due to humidity. During Scanning Thermal Microscopy (STM) experiments, the presence of a water meniscus at the mechanical contact between the tip and the sample creates capillary forces that impact the thermal conductance [2,3].

In this study, we simulate solid/liquid interfaces using molecular dynamics (MD) simulations. The interfacial thermal resistance (Kapitza resistance) is calculated between a film of water and various solids such as crystalline and amorphous silicon or amorphous silica with functionalized surfaces. The impact of the average temperature, the heat flux and the thickness of the water slab on the Kapitza resistance is investigated. Finally, we modified the hydrophobicity level of the silica by grafting hydrophobic molecules on the hydroxylated surface. The results of this study provide interesting insights about the nanoscale heat transfer through solid/liquid interfaces.


[1] Isaiev, Mykola, et al. "Thermal transport enhancement of hybrid nanocomposites; impact of confined water inside nanoporous silicon." Applied Physics Letters 117.3 (2020): 033701.
[2] Assy, Ali, et al. "Analysis of heat transfer in the water meniscus at the tip-sample contact in scanning thermal microscopy." Journal of Physics D: Applied Physics 47.44 (2014): 442001.
[3] Assy, Ali, and Séverine Gomès. "Temperature-dependent capillary forces at nano-contacts for estimating the heat conduction through a water meniscus." Nanotechnology 26.35 (2015): 355401.

Understanding thermal transport across interfaces is crucial for the thermal management of electronic devices. Phonons, the dominant heat carriers across solid-solid interfaces, have been shown to transport elastically in most experiments, except for a few extreme cases where inelastic phonon transport occurs when the two materials of the interface are highly dissimilar. However, there is still limited information on when the inelastic transport will occur.
Here, we measured the interface thermal conductance of atomically sharp Al/Si and Al/GaN interfaces, and found that the thermal conductance linearly increases with temperature at high temperatures, suggesting inelastic phonon transport dominates heat conduction across interfaces, though the two materials have similar Debye temperatures. Moreover, we found that when the interface roughness increased to 5-7 atomic layers, inelastic thermal transport diminishes. Our molecular dynamics calculation indicates that phonon non-equilibrium drives the inelastic phonon transport when the interface is atomically sharp. When the interface becomes rough, phonon non-equilibrium reduces, leading to a saturated thermal conductance at high temperatures. Our results provide new insights on phonon transport across interfaces and open up opportunities for rengineering their thermal conductance.

Thermal interface materials (TIMs), usually composed of polymer matrix filled with thermally conductive particulate fillers, are pivotal for heat transfer between the layers in high-density electronic packaging. Numerical simulation provides a versatile tool for understanding the mechanism of thermal transport and design of TIMs; however, the insufficient efficiency and accuracy in modeling the composites with complex microscopic structures hinders its application in engineering the formula of material. Here we present a strategy of modeling the effective thermal conductivity of polymer-based composites with high efficiency and accuracy, which enables a high-throughput screening of the formula of TIMs. The efficiency of the strategy comes from an implementation of a numerical homogenization algorithm based on fast Fourier transform, in which the unit cell problem is solved by means of a polarization-based iterative scheme. The algorithm shows good precision and requires dramatically reduced computational cost compared with finite element modeling. The accuracy of the strategy originates from an accurate determination of the interfacial thermal resistance (ITR) between different components inside the TIM composites, which is achieved by combining high-throughput calculations, machine learning and experiments. As an example of application of the strategy, the ITRs for the interfaces of filler-silicone and filler-filler in the Al₂O₃/silicone composites are determined as 2.2e-08 m²K/W and 1.0e-07m²K/W, respectively. We further conduct a high-throughput computational screening of the gradation scheme for the silicone based composite TIMs filled with Al₂O₃ and AlN fillers with different size distributions, weights, and surface characteristics, and validate the results by comparing the results with experiments. The proposed strategy could be an effective tool for the smart engineering of the formula of TIMs for electronic packaging, as well as other thermally conductive composite materials.

Polymer-based thermal interface materials play key roles in thermal management applications thanks to their unique combination of properties not available from any other known materials.^1-3 They are lightweight, corrosion-resistant, easy to process, and among others. However, common polymers are thermal insulators with low thermal conductivity on the order of 0.1 Wm^-1 K^-1.^4,5 To achieve high thermal conductivity in polymers, highly thermally conductive fillers such as graphene and nano-diamond powder have been added into polymers.^6-9 Although polymers are blended with highly thermally conductive fillers (e.g. silver nanoparticle ~400 Wm^-1 K^-1) at a high loading volume fraction of 70 vol%, thermal conductivity in polymer-based composites is limited to ~5 Wm^-1 K^-1.¹⁰ This is because low thermal conductivities in common polymer matrices, high thermal interface resistance between filler and polymer, and high thermal interface resistance between filler and filler.^3,11 In this poster, we will present our current research on developing polymer-based thermal interface materials. To increase effective thermal conductivity in polymer-based composites, we strive to increase thermal conductivity in pure polymer matrices (e.g. polyethylene, polyvinyl alcohol, and epoxy) and reduce thermal interface resistances between fillers and polymers by enhancing vibrational couplings. We believe that polymer-based thermal interface materials will play a key role in many existing and unforeseen thermal management applications.

1. Qian, X.; Zhou, J.; Chen, G., Phonon-engineered extreme thermal conductivity materials. Nat. Mater. 2021, 1-15.
2. Feng, C. P.; Chen, L. B.; Tian, G. L.; Wan, S. S.; Bai, L.; Bao, R. Y.; Liu, Z. Y.; Yang, M. B.; Yang, W., Multifunctional Thermal Management Materials with Excellent Heat Dissipation and Generation Capability for Future Electronics. ACS Appl. Mater. Inter. 2019, 11 (20), 18739-18745.
3. Xu, Y.; Wang, X.; Hao, Q., A mini review on thermally conductive polymers and polymer-based composites. Compos. Commun. 2021, 24.
4. Choy, C., Thermal conductivity of polymers. Polymer. 1977, 18 (10), 984-1004.
5. Sperling, L. H., Introduction to physical polymer science. John Wiley & Sons: 2005.
6. Huang, Y.; Ellingford, C.; Bowen, C.; McNally, T.; Wu, D.; Wan, C., Tailoring the electrical and thermal conductivity of multi-component and multi-phase polymer composites. Int. Mater. Rev. 2019, 65 (3), 129-163.
7. Burger, N.; Laachachi, A.; Ferriol, M.; Lutz, M.; Toniazzo, V.; Ruch, D., Review of thermal conductivity in composites: Mechanisms, parameters and theory. Prog. Polym. Sci. 2016, 61, 1-28.
8. Ma, H.; Gao, B.; Wang, M.; Yuan, Z.; Shen, J.; Zhao, J.; Feng, Y., Strategies for enhancing thermal conductivity of polymer-based thermal interface materials: a review. J. Mater. Sci. 2020, 56 (2), 1064-1086.
9. Han, Y.; Shi, X.; Yang, X.; Guo, Y.; Zhang, J.; Kong, J.; Gu, J., Enhanced thermal conductivities of epoxy nanocomposites via incorporating in-situ fabricated hetero-structured SiC-BNNS fillers. Compos. Sci. Technol. 2020, 187.
10. Chen, H.; Ginzburg, V. V.; Yang, J.; Yang, Y.; Liu, W.; Huang, Y.; Du, L.; Chen, B., Thermal conductivity of polymer-based composites: Fundamentals and applications. Prog. Polym. Sci. 2016, 59, 41-85.
11. Ruan, K.; Shi, X.; Guo, Y.; Gu, J., Interfacial thermal resistance in thermally conductive polymer composites: A review. Compos. Commun. 2020, 22.

Identifying the ultimate operating limits of electronic devices requires an understanding of how atomic motions couple to changes in electronic transport -- a coupling that is often mediated by energy dissipation (i.e. Joule heating) and thermal transport. One class of electronics where this coupling has been especially challenging to understand is phase-change Mott-memory. Mott insulators (e.g. VO2) have long been known to undergo an insulator-to-metal transition under an applied voltage. However, several aspects of this electric-field-driven transition has remained unclear, in particular, the role of Joule heating in triggering the transition on fast timescales, and the thermal relaxations on slow timescales. We address these questions by developing a new platform for ultrafast, correlative electrical-thermal-structural characterization of phase-change electronics [1]. Following electrical excitation of VO2, we directly probe the role of Joule heating on short timescales, distinguishing it from direct field-induced effects. We discover a new phase of this material under electrical excitation, which points towards a non-equilibrium modification of the free-energy landscape. On long timescales, we uncover a regime of thermally-bottlenecked structural relaxation, which determines the fastest turn-off speed of the device. Taken together, our approach identifies the ultimate speed and energy-efficiency limits of phase-change electronics.

[1] A. Sood et al., Science (2021) "Universal phase dynamics in VO2 switches revealed by ultrafast operando diffraction", https://arxiv.org/abs/2102.06013

Vanadium dioxide (VO₂) is a multipurpose material with notable applications in fields such as energy storage and energy-efficient smart windows. The versatility of VO₂ is related to its extensive crystallography with many polymorphic forms. VO₂ can undergo a reversible metal–insulator phase transition (MIT) at a critical temperature of 68 °C, which is accompanied by a structure transformation from a low-temperature insulator phase VO₂ (M) to a high-temperature metal phase VO₂ (R). This phase transition results in a distinct change of the infrared transmittance, rendering VO₂ suitable as a pigment for thermochromic glazing.

Various methods for the synthesis of VO₂ have been reported, including chemical vapor deposition, atomic layer deposition, pulsed laser deposition and sputtering. Wet chemical synthesis routes are widely considered as valuable alternatives for vacuum based techniques on account of their low cost, limited environmental impact and reasonably simple experimental setup. Also, the versatility of the method allows the direct synthesis of both ceramic (nano)powders as well as thin (nanostructured) films. However, many earlier wet chemical routes intermediately form V₂O₅ and require an additional, reductive anneal to obtain VO₂. This emphasizes the large challenge in the immediate formation of thermochromic VO₂ from an aqueous solution. Given the complex redox chemistry of vanadium, it is therefore essential to gain fundamental insights in the aqueous vanadium precursor chemistry and in their relation to the vanadium oxide stoichiometry and phase formation, obtained via a solution-based strategy.

Our approach starts with the formulation of stable aqueous precursor solutions containing oxovanadium(IV) complexes. To analyze and understand each step, and the associated chemical transformations, during this synthesis, various complementary spectroscopic techniques (UV-Vis, ⁵¹V-NMR, FTIR) are used to probe and identify each vanadium species during the entire synthesis sequence. The structure and stability of the final complexes that are present in the aqueous solution, and their relation to the synthesis parameters will be highlighted. Subsequently, the formation of vanadium oxide, starting from aqueous vanadium precursor solution, requires a thermally-induced transformation of the precursor gel into the oxide. As an illustration, the thermal decomposition pathway of a citrato/oxalato-oxovanadium(IV) gel, obtained by evaporation of a precursor solution, is analyzed by Simultaneous Differential Calorimetry Thermogravimetry (SDT), High temperature X-ray Diffraction (HT-XRD) and ex-situ analysis of the remaining products. The thermochromic nature of the resulting VO₂ products is demonstrated by monitoring the underlying structural phase transition with Differential Scanning Calorimetry (DSC) and UV-Vis spectroscopy.

The authors thank the European Fund for Regional Development (crossborder collaborative Interreg V program Flanders-the Netherlands, project SUNOVATE) for their financial support.

We have fabricated polyethylene nanofibers with high chain alignment and high degree of crystallinity. The nanofibers show an orthorhombic crystal structure with high thermal conductivity at room temperature. However, rotational disorder occurs at a high temperature close to melting temperature. The nanofibers switch to the hexagonal phase with a high-contrast and abrupt thermal conductivity change. Such a solid-state phase transition makes the polyethylene nanofiber intrinsically a thermal regulator. Measurements show a thermal switching ratio in average ~8x with maximum ~10x, which is the highest among solid-solid and solid-liquid phase transitions of all existing materials. Based on the sharp and high-contrast phase transition, an unusual dual-mode solid-state thermal rectification effect is demonstrated using a heterogeneous “irradiated-pristine” polyethylene nanofiber junction as a nanoscale thermal diode, in which heat flow can be rectified in both directions by changing the working temperature. For the nanofiber samples measured here, we observe a maximum thermal rectification factor as large as ~ 50 %, which only requires a small temperature bias of <10 K. The nanoscale thermal regulators and rectifiers open up new possibilities for developing advanced thermal management, energy conversion and, potentially thermophononic technologies.

Dropwise condensation on thin (< 100 nm) hydrophobic coatings enables exceptional condensation heat transfer. However, hydrophobic coatings can be easily delaminated by the formation of water blisters in wet environments or during steam condensation. To prevent delamination while maintaining nanoscale thickness, designing interfaces with high wet adhesion is a key challenge. In nature, cell membranes face the identical challenge where nanometer-thick lipid bilayers require exceptional adhesion in wet environments to maintain their integrity. Nature achieves high inter-layer adhesion by using a lipid interface having two non-polar surfaces, which does not rely on chemical bonding. Instead, physicochemical resistance is used to prevent biofluids from opening and delaminating the interface. Based on this nature-inspired strategy, we develop fluorine-carbon molecular chains on nanostructured surfaces to build lipid-like interfaces for nanometric fluorinated coatings. Our superhydrophobic materials show unprecedented jumping droplet condensation stability as demonstrated from continual one-year condensation experiments. Our work presents a solution to the century-old problem of hydrophobic and superhydrophobic coating durability through bio-inspired materials design with robustness beyond what is currently available.

The sky is a natural heat sink that has been extensively used for passive radiative cooling of households. Past experimental works have focused on maximizing the radiative cooling power of roof coating in hot daytime using static, cooling-optimized material properties. However, the resultant overcooling in cold night or winter times exacerbates the heating cost, especially in climates where heating dominates energy consumption. In this work, we approach the thermal regulation from an all-season perspective by developing a mechanically flexible coating that adapts its thermal emittance to different ambient temperatures. The fabricated temperature-adaptive radiative coating (TARC) optimally absorbs the solar energy and automatically switches thermal emittance from 0.20 for ambient temperatures lower than 15 °C to 0.90 for temperatures above 30 °C, driven by a photonically amplified metal-insulator transition. The TARC is simulated to outperform all existing roof coatings for energy saving in most climates, especially those with significant seasonal variations, yielding a cut in annual source energy consumption up to 3.65 GJ for a typical single-family home in the U.S.

Conventional cooling systems, that is, air conditioners, should be replaced because they consume a substantial amount of energy and cause environmental pollution. In this context, radiative cooling systems, which perform cooling without consuming any energy or causing environmental pollution, are emerging as an alternative. However, most of the radiative coolers explored thus far include metals, such as silver, that are used as solar reflectors, thereby entailing problems in terms of practicality, mass production, cost, and light pollution. Herein, we propose calcium carbonate (CaCO₃) micro-particle-based radiative cooling, which utilizes the high-energy band gap of CaCO₃ for high-performance radiative cooling. As the cooler has only a single layer of a CaCO₃ composite without any metal reflector, it is mass-producible, cheap, and does not cause light pollution. To demonstrate the cooling performance of CaCO₃, optical properties and temperature changes are measured and compared with those of commercial white paint. As a result, it is demonstrated that the CaCO₃-based radiative cooler has cooling power 93.1 W/m² in calculation and can be cooled 6.52 ^oC and 3.38 ^oC under ambient temperature in daytime and nighttime respectively. Thus, it can perform as radiative cooler in entire day.

Metal additive manufacturing (AM) has revolutionized engineering design and production of components, eliciting a wide design space for geometrical, material, and functional optimization. This wide design space has led to the development of innovative functional devices for a plethora of applications including energy, transportation, chemical and bioengineering. The new processing phenomena and requirement for use of different alloying materials in AM poses unprecedented nanostructuring challenges of AM surfaces along with opportunities for the development of novel generations of surface architectures. Here, through the understanding of the grain formation mechanisms and material compositions resulting from the AM process, we developed an ultrascalable and cost-effective method to rationally generate nanostructures on metallic AM surfaces for enhanced droplet repellency, condensation, evaporation, and boiling heat and mass transfer. By combining a facile microdroplet impact method with high speed optical imaging, we quantify that the AM structures significantly lower droplet adhesion as compared to state-of-the-art nanostructured surface on conventional aluminum alloys. We further explored the nanostructure morphology governing the departure size and frequency of condensed droplets and developed an AM structuring strategy that categorically reduces condensate flooding and enhanced condensation heat transfer coefficient by 6X, as compared to untreated surfaces in high supersaturation environments. This work not only shows the potential of using AM structured surfaces to achieve multi-functional performance enhancement, but it also outlines design guidelines and performance optimization strategies for structuring metallic AM materials.

Heavy metal (HM)/magnet bilayers host many magnetoresistance (MR) and spin caloritronic (SC) effects. Here we show that the spin Peltier effect and electron-phonon scattering result in a much larger measured unidirectional MR of an HM on a magnetic insulator than existing theories that neglect the interplay between MR and SC effects. This spin Peltier MR and the frequency-dependent spin Seebeck effect (SSE) reveal Onsager reciprocity and offer experimental probes of elusive properties of the nonequilibrium magnons driving these effects, including the magnon-phonon thermalization length and SSE coefficient.

Structural symmetries determine the underlying physical properties of materials and thereby govern their potential functionalities. Here, I will discuss the role of twist structural symmetry in determining vibrational and thermal transport properties of one dimensional and bulk materials. The developed twist dynamics dictate symmetry-enforced band degeneracies, non-trivial topologies, phonon interactions, and unique scattering observables. This will be demonstrated by first-principles calculations supported by inelastic neutron scattering measurements.

R.J. and L.L. acknowledge support from the U. S. Department of Energy, Office of Science, Basic Energy Sciences, Materials Sciences and Engineering Division.

Cubic boron arsenide (BAs) is a promising compound semiconductor for thermal management applications due to its high thermal conductivity, exceeding 1000 W m⁻¹ K⁻¹ at room temperature in high-quality samples. However, the as-grown BAs crystals usually exhibit large variations in thermal and electronic transport properties. The origin of these large variations has thus far been inconclusive. Here, we investigate the effects of impurities on the thermal and electrical properties of BAs. Time-of-flight secondary ion mass spectrometry and electron probe microanalysis measurements reveal the presence of several impurities in BAs, including Si, C, O, H, Te, Na, and I. Some of these impurities, especially Si, C, and H, could serve as shallow acceptors, leading to the p-type conducting behavior commonly measured in BAs. The thermal conductivity and hole mobility are reduced more in the samples with higher impurity concentrations due to the enhanced impurity scattering of phonons and holes, respectively. Firstprinciples calculations are used to determine the thermal conductivity reduction induced by different impurities. The calculated results confirm the experimental trends. The substitution of O for As leads to a large bond distortion resulting from the breaking of the T_d symmetry, which yields unusually strong phonon scattering with a correspondingly large reduction in thermal conductivity. Our results offer useful insights into the impurity-sensitive transport properties of BAs.

In the past decade, non-diffusive heat transport by phonons at room temperature at the micro/nanoscale was reported in many single crystal materials. However, the transition to the ballistic transport regime, well studied in heat pulse experiments at the liquid He temperatures, remained elusive at room temperature. Motivated by the recent experimental work at the FERMI free electron laser facility in Trieste, Italy, where the extreme ultraviolet (EUV) laser source was used to produce nanoscale transient thermal gratings (TTGs) with heat transfer distances in the tens on nanometers range, we theoretically study heat transport in silicon in the TTG geometry using the Peierls-Boltzmann phonon transport equation based on first-principles calculations of the phonon dispersion and scattering rates. We find that the transition to the ballistic regime takes place in the sub-100 nm range of TTG periods. We compare the calculations done with the relaxation time approximation and the full scattering matrix and investigate the effect of different initial heat source distributions (e.g. optical phonon excitation). The results will be compared with the experimental EUV TTG data from single crystal silicon membranes.

Boron arsenide (BAs) has ultrahigh thermal conductivity because of its unique phonon-phonon scattering properties [1]. At ambient pressure, “bunching” of acoustic phonon branches in BAs leads to a small phase-space for three-phonon scattering [2]. Density functional theory calculations suggest three-phonon scattering rates increase at high pressures because “bunching” of acoustic phonon branches decrease [2]. To test this hypothesis, we measure the thermal conductivity of BAs vs. pressure using time-domain thermoreflectance (TDTR) with a diamond anvil cell (DAC). We observe a pressure independent thermal conductivity from 0 to 25 GPa. We explain our observations as a result of the pressure dependence of the phonon dispersion relation, three-phonon scattering rates, and mass-disorder scattering.
[1] Feng, T., Lindsay, L., & Ruan, X. Physical Review B, 96(16), 1–6 (2017).
[2] Ravichandran, N. K., & Broido, D. Nature Communications, 10(1), 1–8 (2019).

Cooling with magnetocaloric materials has been pursued by numerous agencies and companies, but many of the challenges remain unsolved. Some of these challenges are excessive pressure drop across the magnetocaloric regenerator, mechanical failure from magnetostructural and thermal cycling, and material degradation due to exposure. Presented here is a continuously processable powder-in-tube (PIT) composite in which the tube is a high thermal conductivity metal, and the core is dense GdF3. Using this PIT composite, the magnetocaloric material is not exposed to the external environment. Additionally, the metal sheathing keeps the magnetocaloric material in slight compression to prevent mechanical fatigue. The arrangement of these PIT wires into an axially aligned array results in a high effective thermal conductivity, low demagnetization factor, and low impedance regenerator, overcoming some of the most difficult challenges in proposed magnetocaloric systems.

Conventional gas compression refrigeration technique needs to replace by alternative cooling technology to reducing the depletion of the ozone layer and carbon production. Nowadays, scientists give more interest on magnetocaloric materials (MCM) for refrigeration to replace the vapor compression technique. MCMs are eco-friendly, more efficient, economical, and easy to use. In general, good magnetocaloric materials should have a low Debye temperature value, phase transition near working temperature, large temperature difference (ΔTad) in the vicinity of phase transition, no thermal or magnetic hysteresis, low specific heat, high thermal conductivity, high electrical resistance, resistant to corrosion, and good mechanical properties. Additionally, eddy current heating depends on the surface morphology that is caused by fast varying magnetic fields. Thus, the morphology of materials play an important role to control the magnetocaloric properties. Gd₂O₃ nanomaterial was reported to exhibit a large magnetocaloric entropy change at ~10K which is suitable to liquifying helium useful for cryogenic space application. Literature shows different values of magnetic entropy change (ΔS_m) with the same Gd₂O₃ nanostructures with different morphology like - 8.7 J/kg.K for nanotubes, - 20 J/kg.K for the rod. Several attempts have been made to study MCE on different morphology but comparative study on different morphology is lacking.
In this present study, we investigate the magnetocaloric properties of Gd₂O₃ nanoparticles of different morphology. Single phase Gd₂O₃ nanoparticles were synthesized via the hydrothermal method and a homogeneous precipitation method followed by heat treatment in air at 1100 ^oC for 12 h. Thus Gd₂O₃ nanoparticles of different shape viz. bulk, rod, spheres, and plate were prepared. The morphology of the samples was analyzed using SEM. The SEM images of synthesized Gd₂O₃ showed 1D nano-rods, 2D nano-plates like architecture prepared via hydrothermal method while hollow nanosphere morphology achieved with a homogeneous precipitation method. The average area of plates are 1mm², the average length of rods are ~500 nm and diameter of hollow micro sphere are ~600 nm. The phase identification and lattice parameter of the powders performed using x-ray diffraction showed presence of highly-crystalline single phase cubic crystal structure with JCPDS No.11-0604. The average crystallite size of sample were lies between 20-30 nm calculated by using Scherrer formula. The lattice parameter of the sample using profile fitting is lies between 10.810 to 10.832 Å.
Magnetic measurements were performed using a SQUID magnetometer. Temperature dependence of zero-field-cooled (ZFC) curve exhibits unchanged phase. The material demonstrate typical paramagnetic behavior in the whole measured temperature 5K to 100K. The paramagnetic susceptibility follows Curie-Weiss law and the fit parameters yield a paramagnetic Curie temperature (Tc) value of ~5K for all samples. The M-H data in the temperature range of 3–30 K is used to calculate the isothermal magnetic entropy change. The M-H data shows the paramagnetic behavior. The maximum value of ΔS_m for 50 kOe field change at 5K is 11.2 Jkg−1 K−1 for the nanoplate sample, 9.4 Jkg−1 K−1 for the nano-rod and 9.2 Jkg−1 K−1 for the nano-sphere, and 10.7 J/kg.K for Bulk. The reason for achieving large ΔSm could be the surface area and surface geometry of the samples. More surface area indicates more surface spins that enhanced magnetization value. These results shows the Gd₂O₃ nano plates are the promising materials for magnetic refrigeration technology compare with oxide nanoparticles at 5K. Our study suggests that the Gd₂O₃ nanostructures of various morphologies could exhibit interesting magnetocaloric properties at low temperatures.

References
1. Phan, M. H.et al, Journal of Magnetism and Magnetic Materials, 308 (2007) 325-340.
2. Hazarika, S.et. al, Applied Surface Science, 491,(2019) 779-783.

Suspended micro-island device has enabled researchers to directly measure thermoelectric properties in nanomaterials spanning from nanowires and nanotubes to 2 dimensional materials. The unique capability of this device of probing thermal conductivity at the level of single nanowire and 2D material has significantly contributed to expanding knowledge in nanoscale thermal transport. While the suspended micro-island device remains a powerful tool to measure thermoelectric properties, we further advanced experimental techniques using suspended micro-island devices in order to increase success rates in sample preparation as well as to increase accuracy in thermal conductivity measurements. We will explain the detailed structure of the newly designed suspended micro-island devices and our experimental configurations. Further, we will describe advantages and versatile usages of our experimental technique in various scenarios.

Time-domain thermoreflectance (TDTR) mapping technique has been applied in many fields for detecting the spatial variation of the thermal properties of materials. Here we report the TDTR mapping work with a micro-scale spatial resolution on single carbon fibers and an oxidized SiC/SiC composite. We determine the transverse thermal conductivity of a single IM7 carbon fiber to be ~1.7 W/m-K and obtain uniform thermal conductivities at different radial positions by mapping a carved step structure on the IM7 carbon fiber. Moreover, we determine both the thermal conductance of the oxide layer and the thermal conductivity of the SiC/SiC composite by mapping at different delay times, and figure out the differences between the fibers and the matrix. The effective thermal conductance of the oxide formed on the fibers (27-40 MW/m²-K) is lower than the oxide on the matrix (40-47 MW/m²-K), while the fibers have lower thermal conductivities (15 W/m-K) than the matrix (50-90 W/m-K). TDTR mapping on these two small-size composite materials allows for obtaining accurate thermal property changes that are related to the microstructure discrepancies of materials. In addition, we enable a faster data acquisition rate up to 33 μm/s by introducing heterodyne detection into the conventional TDTR setup, which allows for a 17 times higher scanning speed than the conventional TDTR mapping.

Thermal energy lies at the heart of any irreversible process that constitutes much of the operating world around us. Despite this, there is seldom a unifying call to fashion better thermal materials and understand their phenomenology directly; it is often treated as a secondary problem. However, as technoeconomic modeling and global impact are of increasing prominence in materials design and applied physics, I argue that thermal properties should be a primary consideration. Here, I present recent research on a variety of topics - design of phase change cooling materials, phonons in hybrid materials, water separations, thermoelectrics, hydrogen storage, personal comfort, and computing - all of which orbit around how thermal energy can be better harnessed, guided, and utilized. Some highlights will be design considerations of flexible heat exchangers, interfacial solar evaporation in brine management, thermal storage in molecular-invaded porous crystals, unusual temperature dependent plasmonic catalysis, phonon transport in soft materials, and thoughts on computing limitations.

ReS₂ is a rising star in the transition metal dichalcogenide family. The uniqueness of ReS₂ lies in its distorted 1T triclinic crystal structure where the additional d valence electrons of Re atoms form zigzag Re chains parallel to the b-axis, drastically reducing its symmetry. As a result, ReS₂ possesses extremely weak interlayer bonding and anisotropic optical, thermal, and electronic properties. Unlike black phosphorus, ReS₂ is very stable in air, making it more suitable for optoelectronic applications. Recently, two stacking orders were identified in multi-layer ReS₂: AA stacking where adjacent layers sit right on top of one another with minimal displacement, and AB stacking where adjacent layers shift by one-unit cell along a-axis. With, Ab-initio calculations, the interlayer distance of stacking AB is determined to be 2.59 Å, smaller than that of stacking AA, 2.71Å, which suggests the stronger interlayer interaction in stacking AB. The carrier dynamics, and both optical and vibrational properties in different stacking orders were found to be drastically distinct. In this study, we prepared ReS₂ flakes via mechanical exfoliation with both AA and AB stacking and measured their thickness dependent cross-plane thermal conductivity. We found that the stacking orders of ReS₂ are very robust. Pure stacking order (either AA or AB) can persist up to the thickness of several microns. For both stacking orders, thermal conductivity increases with sample thickness until around 2µm, indicating an extremely long phonon mean free path along cross plane direction. The thickness dependent thermal conductivity in samples with AA stacking increases much faster than those with AB stacking. Our results suggest that stacking order can be an additional degree of freedom to manipulate thermal transport in 2D materials.

Higher-order force constants of Cu₂O, cuprite, were calculated with a perturbative ab initio effective potential method, and used to predict thermal expansion and other temperature-dependent phonon behaviors. Inelastic neutron scattering was used to measure phonon dispersions in single crystal cuprite at 10, 300, 700, and 900 K. Low- and high-order force constants accounted for the negative thermal expansion of cuprite; however, calculations with high-order force constants were more effective in predicting the observed temperature-dependent phonon behavior. None of these calculations sufficiently accounted for an observed broad, diffuse intensity in the measured phonon spectrum. Classical molecular dynamics with machine learning interatomic potentials models will be compared with perturbative anharmonic calculations. A local dynamics model is proposed to explain this diffuse inelastic scattering, analogous to phase noise in oscillators.

Boltzmann transport equation (BTE) is an ideal tool to describe the multiscale phonon transport phenomena, which are critical to applications like microelectronics cooling. Numerically solving phonon BTE is extremely computationally challenging due to the high dimensionality of such problems, especially when mode-resolved properties are considered. In this work, we demonstrate the use of physics-informed neural networks (PINNs) to efficiently solve phonon BTE for multiscale thermal transport problems with the consideration of phonon dispersion and polarization. In particular, a PINN framework is devised to predict the phonon energy distribution by minimizing the residuals of governing equations and boundary conditions, without the need for any labeled training data. Moreover, geometric parameters, such as the characteristic length scale, are included as a part of the input to PINN for learning BTE solutions in a parametric setting. This enables our method to handle structures over a wide range of length scales after a single training. The effectiveness of the present scheme is demonstrated by solving a number of phonon transport problems in different spatial dimensions (from 1D to 3D). Compared to existing numerical BTE solvers, the proposed method exhibits superiority in efficiency and accuracy, showing great promises for practical applications, such as the thermal design of electronic devices.

Two-dimensional (2D) ZrS₂ monolayer (ML) has emerged as a promising candidate for thermoelectric (TE) device applications due to its high TE figure of merit, which is mainly contributed by its inherently low lattice thermal conductivity. This work investigates the effect of the lattice anharmonicity driven by temperature-dependent phonon dispersions on the thermal transport of ZrS₂ ML. The calculations are based on the self-consistent phonon (SCP) theory to calculate the thermodynamic parameters along with the lattice thermal conductivity. The higher-order (quartic) force constants were extracted by using an efficient compressive sensing lattice dynamics technique, which estimates the necessary data based on the emerging machine learning program as an alternative to computationally expensive density functional theory calculations. Resolve of the degeneracy and hardening of the vibrational frequencies of low-energy optical modes were predicted upon including the quartic anharmonicity. As compared to the conventional Boltzmann transport equation (BTE) approach, the lattice thermal conductivity of the optimized ZrS₂ ML unit cell within the SCP + BTE approach is found to be significantly enhanced (e.g., by 21% at 300 K). This enhancement is due to the relatively lower value of phonon linewidth contributed by the anharmonic frequency renormalization included in the SCP theory. Mainly, the conventional BTE approach neglects the temperature dependence of the phonon frequencies due to the consideration of harmonic lattice dynamics and treats the normal process of three-phonon scattering incorrectly due to the use of quasi-particle lifetimes. These limitations are addressed in this work within the SCP + BTE approach, which signifies the validity and accuracy of this approach.

As a consequence of a lattice mismatch between dissimilar materials, misfit dislocations are unavoidable at the phase interfaces in heterostructures when the epilayer thickness exceeds a critical thickness, regardless of the growth mechanisms and conditions. In this study, we report simulation results of the resonant interaction between long-wavelength transverse acoustic (TA) phonons and misfit dislocations at the PbTe/PbSe(100) interface, using the Concurrent Atomistic-Continuum method. It is found that TA phonons with frequency near a specific value excite the resonance behavior of the misfit dislocations, as identified by large-amplitude, out-of-phase vibration of the atoms on both sides of the slip plane of the misfit dislocations. The vibration amplitude can be up to ten times of the amplitude of the incident phonon wave. The resonant interaction leads to a low energy transmission of TA phonons.

To meet the global threat of climate change and ensure that all Americans benefit from the transition to a clean energy economy, the DOE Solar Energy Technologies Office (SETO) works to accelerate the development and deployment of solar technologies to decarbonize the electricity system by 2035 and decarbonize the energy sector by 2050. A fully decarbonized grid will likely require large amounts of energy storage to provide energy when it is needed and increase the grid’s reliability and resilience. While lithium-ion batteries have enabled rapid deployment of energy storage, most commercial applications have been limited to four hours of storage or less. Longer-duration storage can help alleviate the impact of extended periods of cloudy weather, for example, or even seasonal variations of solar energy production.

Concentrating solar-thermal power (CSP) plants use mirrors to capture the sun’s energy as heat, which can be stored in a thermal energy storage (TES) system and used to produce electricity on demand, even when the sun is not shining. Commercial CSP plants with TES have demonstrated the viability of thermal storage to be responsive to grid needs, particularly for long durations of daily storage, up to 17 hours, which is not currently economically viable with lithium-ion batteries. Thermal systems can decouple the storage capacity component (e.g., molten salt stored in tanks) from the power-generating component (heat engine). This enables system designers to increase marginal storage capacity or duration—without having to build additional generating capacity—by increasing the size of the storage tank. The combination of readily scalable TES and conventional turbine technology can allow CSP to provide reliable and flexible renewable electricity production.

SETO has set a cost target of $0.05 per kilowatt-hour electric (kWh_e) for baseload plant configurations with 12 or more hours of storage, which could enable deployment of 25 to 160 GW of U.S. CSP capacity by 2050. To reach this target, SETO is supporting the development of next-generation CSP plants that are capable of delivering heat to the power cycle at temperatures exceeding 700°C, increasing plant efficiency. Current technologies that use molten salt cannot exceed 565°C owing to thermal decomposition of the salt, which is used as a heat-transfer fluid. If successful, these third-generation (Gen3) CSP designs will enable CSP systems to utilize advanced power cycles based on supercritical carbon dioxide (sCO₂), which are much more efficient than existing steam-based cycles.

Projects in SETO’s Gen3 CSP program have been investigating and designing fully integrated thermal transport systems based on three competing concepts for the heat-transfer media (HTM) organized by the phase of matter for leading HTM candidates—gas, liquid, or solid. In March 2021, SETO announced the selection of Sandia National Laboratories to develop a design based on solid particles as the HTM. Sandia will receive $25 million to construct a 2 megawatt (MW) thermal integrated test facility to validate the proposed system.

Beyond electricity generation, solar-thermal technologies may also be able to help meet the national goal of decarbonizing the entire U.S. energy sector by 2050 by providing industrial process heat. Even with more renewable electricity available, many industrial processes will be difficult to electrify because they require high temperatures or have other challenging process characteristics. SETO aims to make solar industrial process heat cost-competitive with fossil fuels to address high-priority applications, including calcination to produce cement, thermochemical water splitting for producing solar fuels, and ammonia synthesis for producing fertilizer. The solid particle heat-transfer pathway being developed for Gen3 CSP may be particularly relevant for a number of these processes, which have solid-phase components or are mediated by solid-phase catalysts.

We are developing a 100-GWh Crushed-Rock Ultra-large Stored Heat (CRUSH) System that couples to nuclear reactors or very large Concentrated Solar Power (CSP) heat sources. Heat storage enables nuclear or CSP plants to operate at full capacity with dispatchable electricity. The electric storage capabilities of CRUSH are similar to the Tennessee Valley Authority Raccoon Mountain pumped hydro facility that can provide 1600 MW(e) for 22 hours to address hourly to weekly storage. Heat is transferred into and out of storage using nitrate salts (<600°C). The CRUSH system capital-cost goals are $2-4/kWh—an order of magnitude less than current heat storage systems and a factor of 50 less than batteries.
Existing solar power towers use clean nitrate salt for heat transfer and storage. Cold salt is heated by the CSP system and sent to hot nitrate storage tanks. The power block, sized for peak electricity demand, takes hot nitrate salt, produces dispatchable electricity and sends the cold salt to cold-salt storage tanks. Three advanced nuclear reactor concepts use nitrate salt in the intermediate loop and propose the same system design as CSP plants—cold salt heated with a reactor operating at base load, hot salt sent to the hot salt storage tank, hot salt from storage goes to the power block to produce variable electricity and the resultant cold salt is sent to the cold-salt storage tank. Existing CSP systems have up to several gigawatt hours of heat storage and have capital costs of $20-30/kWh of heat stored. About half the capital cost is associated with the tanks and a third or more of the cost is associated with the nitrate salts.
With the CRUSH system, heat storage is in a crushed rock pile (200 m by 200 m with heights of 20 to 40 m) with heat transfer by hot nitrate salt that is sprayed on the crushed rock and drains to the catch pan below. Heat is recovered by spraying cold fluid onto the hot crushed rock and recovery of the hot fluid from the catch pan below. To minimize costs, nitrate salt is used for heat transfer, not heat storage. This process has similarities to heap leaching of copper ores where solutions are sprayed on top of large crushed rock piles, flow through the piles, dissolve the copper and are collected by the drain pans below. The crushed rock pile has a flat top surface with sloping sides of crushed rock. This allows free expansion and contraction of the rock pile with changing temperatures. The building over the crushed rock is similar to an aircraft hangar with a gas-tight insulation on the inside—no rock or liquid nitrate salt touches the building structure. Unlike current nitrate-salt heat storage systems with clean nitrate salt in clean tanks, the CRUSH salt is in direct contact with the crushed rock.
We are examining materials reactions between the salt, the crushed rock, heat exchangers and piping. The thermal transients will generate fines; thus, filters are required to remove larger solid particles in the liquid salt. Rock choices must be chosen to (1) withstand thermal cycling, (2) forces by 20+ meters of crushed rock height, (3) dissolution of selected components that may or may not be limited by solubility limits in nitrate salts, (4) dissolution and precipitation of rock components with temperature and (5) corrosion and fouling of heat exchangers and other components. There has been some experimental work associated with nitrate salt/crushed rock in proposed heat-storage thermocline systems with hot salt on top of cold salt in a tank filled with crushed rock. Only some crushed rocks are chemically compatible with the nitrate salts. We review the existing work, our analysis and priorities going forward. Major work, including poly-temperature flowing salt loops, will be required to qualify rock types for this service.

As a complementary technology to photovoltaics in harvesting and converting solar energy, concentrated solar power (CSP) system offers a unique advantage in that it stores solar energy as heat at low cost, which enables dispatchable solar electricity supply even at night. Same as other heat engine systems, the CSP system follows the Carnot’s theorem: the higher the operation temperature of the system, the higher the conversion efficiency. Thus, a key component of the CSP system, the tubing material, is required to be both mechanically strong at high temperatures >873 K and resistant to the corrosion of both the working fluids inside and the oxidation in air outside at this working temperature. Currently, most of the tubing materials employed in power plants could not operate at temperature higher than 873K due to their mechanical and chemical limits. Even though the nickel or titanium-based alloy could have better performance, they would not be economic solutions to this challenge at a cost ~10x that of stainless steels. In this paper, we present FeMnNiAlCr high entropy alloys (HEAs) to address this challenge. Low-cost Fe and Mn constitute >60 at.% of the HEA composition, much more cost-effective than the nickel-based superalloys, while the tensile strength of the HEAs are comparable to Ni-based Inconel alloys at 973K [1,2]. Compared with 304 stainless steels, its yield strength is also twice as high at 973K. Therefore, our FeMnNiAlCr HEA shows promising mechanical properties to satisfy the mechanical requirements of the high temperature tubing material. The other crucial element of the CSP system is the solar selective absorber. High solar absorbance, low thermal emittance loss and stable optical-to-thermal energy conversion efficiency at 1023K for long term are essential properties of the solar selective absorber utilized in this working environment. One of the widely applied solar selective absorber coatings, Pyromark 2500, has high solar emittance loss (~80%) and would need sophisticated fabrication steps to ensure its adherence to the substrate [3]. Its conversion efficiency also degrades after operation at 1023K for just 300h. The native oxide we grown on two phase HEAs, which mostly comprises Mn₂O₃ on the surface and a dense Al₂O₃ at the oxide/HEA interface, provides a facile synergistic solution to solar absorbers. This native oxide layer has high solar absorbance of 91.4%, low thermal emittance of 27% and an optical-to-thermal energy conversion efficiency of 89.7% at 1023K for 1000x solar concentration ratio. The solar absorption is mainly contributed by manganese oxide on the surface, while the dense protective aluminum oxide at the oxide/HEA interface can be engineered to prevent excessive oxidation. Interestingly, the manganese oxide spontaneously forms surface texture upon oxidation, similar to the surface engineering of photovoltaic cells, which enhances the solar absorption while the dense and robust layer itself protects the alloy from both the corrosion of the working fluids and excessive oxidation of the air. In preliminary high-temperature corrosion studies, two-phase FeMnNiAlCr HEAs have sustained bromide molten salts for 14 days at 1023 K with <2% weight loss. Thus, the potential of synergistically addressing mechanical, optical and chemical challenges makes FeMnNiAlCr HEAs a promising cost-effective candidate for high-temperature tubing materials in future CSP systems.
[1] Baker, Ian, et al. "Recrystallization of a novel two-phase FeNiMnAlCr high entropy alloy." Journal of Alloys and Compounds 656 (2016): 458-464.
[2] Z. Wang, M. Wu, Z. Cai, S. Chen and I. Baker, “The microstructure and mechanical behavior of(Fe36Ni18Mn33Al13)100-xTix high-entropy alloys”, Intermetallics 75, 79-87 (2016). http://dx.doi.org/10.1016/j.intermet.2016.06.001
[3] Ho, Clifford K., et al. "Characterization of Pyromark 2500 paint for high-temperature solar receivers." Journal of Solar Energy Engineering 136.1 (2014).

Selection of an appropriate heat transfer agent (HTA) and thermal energy storage (TES) material is important for minimizing the cost of the solar receiver, thermal storage and heat exchangers, and for achieving high receiver and cycle efficiencies. It has also been observed that the use of the molten salt eutectic composition as an HTA / TES fluid in many different areas is effective in increasing system efficiency. Current molten salt HTAs have high melting points (>200°C) and degrade above 600°C. The main problem with this application is that it can easily freeze in the evening or in the winter months and clog the pipeline to make working conditions difficult. For this reason, there is an urgent need for inexpensive salt melt compositions having a lower melting point and a higher thermal stability temperature. To do this, a new nanostructured eutectic molten salt was designed to achieve lower melting point (>200°C) and higher thermal conductivity. In this study, we prepared phase diagrams of molten salts and anodic alumina oxide (AAO) with different pore size (25 to 400nm). Maximum melting temperature depression was observed as ΔT=~170°C for AAO/KNO₃ composites for 35nm pore size. Furthermore, a drastic increase on thermal conductivity (73%) was recorded for the same composite. It was observed that this increase was not systematic in AAO / KNO₃ composite structures, but it was systematic in AAO / NaNO₃ composite structure. The highest thermal conductivity increase is seen in AAO / KNO₃ composite structure with pore diameter of 35nm.

Ab-initio design or engineering of systems and processes may be considered a holy grail in computational sciences, spanning across the fields of materials science, chemistry, biology, and engineering. The goal here is to integrate multiple simulation modalities, from quantum mechanical calculations to continuum finite element simulations, for a fully ab-initio prediction of device or process operation with limited experimental input. Such an ab-initio predictive workflow, a digital twin, can complement experimental investigation by narrowing the relevant design space.

Specifically, thermal management of electronics is an important aspect in the design of semiconductor devices, where certain applications require device operation in temperature extremes. Chips with electromagnetic detector arrays often have to operate at 77 K and frequently cycle back to room temperature when not in use. One of the design goals here is, thus, to optimise device geometry to minimize thermal stresses upon temperature cycling and improve operational reliability of the resulting integrated detector array.

In this contribution, we introduce the design of Modified Direct Band Interconnect (MoDiBI) detector arrays using artificial-intelligence inspired optimization heuristics integrated with a finite element thermo-mechanical model. While fully ab-initio device engineering is not a reality yet, we use this system as a template to discuss the effect changes in the material properties of device components have on the optimized device geometry. This, in turn, paints a picture on the accuracy needed in the prediction of thermo-mechanical material properties in a fully ab-initio workflow.

Further, we discuss some case studies to evaluate if current state-of-the-art in thermal property prediction from atomistic simulation is sufficient to enable a robust predictive workflow for thermo-mechanical device optimization. Finally, a perspective is provided on recipes to transition toward robust industry-grade predictive workflows for thermal management applications.

Condensation has significant applications in various industries. Commercially condensation can be used for thermal management, power generation, refrigeration, air conditioning, water desalination etc. Condensation can be distinguished in to two distinct modes commonly termed as filmwise condensation (FWC) and dropwise condensation (DWC). Among these DWC is the preferred mode of condensation in heat transfer applications due to the inherently higher heat transfer coefficient (almost one order of magnitude). However, condensation still manifests as FWC on most of the condenser surfaces (copper, aluminium, steel etc.) used in industry. This is because the metal surfaces are hydrophilic, and an additional hydrophobic coating is required to achieve DWC. An ideal hydrophobic coating should have high droplet shedding efficiency, mechanical durability and low thermal resistance. However, it is difficult to attain all these properties in one coating.
In this work, we report DWC with stable heat transfer performance on poly-[1H,1H,2H,2H- perfluorodecyl acrylate] (pPFDA) coating grafted on flat copper and aluminum substrates using initiated chemical vapor deposition (iCVD) technique. iCVD is environmental friendly as it is solvent-free and vapor phase process. The condensation heat transfer performance and durability of this coating is compared with 1H,1H,2H,2H-Perfluorodecyltriethoxysilane (Silane) functionalized copper and aluminium substrates.
The grafted polymer coating using iCVD was obtained in two steps. First, we coated flat copper or aluminum substrate with trichlorovinylsilane (TCVS) in a custom-made chemical vapor deposition chamber at 10 bar. Post TCVS coating, we utilized the iCVD tool (iLab Coating System, GVD Corporation) to deposit an ultra-thin layer of pPFDA (Advancing contact angle (ACA) – 127°±3°, contact angle hysteresis (CAH) – 10.3°±3.4°). The thickness of the coating was found to be ∼40 nm. For silane coating, flat copper and aluminium samples were dipped in a mixture of Silane and hexane for 1 hour followed by baking it on a hot plate (ACA – 123.3°±1°, CAH – 21.3°±1.2°).
The condensation heat transfer performance of both the coatings was evaluated in a harsh condensation environment. The surfaces were mounted in a high saturation pressure chamber (1.42 bar) and exposed to superheated steam (111°C). These are extreme conditions to test the performance of the hydrophobic coatings as compared to a low saturation pressure conditions typical for industrial condensers. The superheated steam was flowing on the test substrates at 3 m/s and 9 m/s during the experiments by exerting a vapor shear on the coating. The heat transfer performance was quantified in terms of the heat transfer coefficient (HTC).
Apart from heat transfer performance, the mechano-chemical robustness of both the coatings was tested for their practical applicability with a protocol consisting of eleven different durability tests. The grafted polymer coating via iCVD outperformed the Silane coating in terms of mechano-chemical durability. Additionally, it survived exposure to superheated steam during condensation experiment and exhibited DWC with a stable HTC during the experiment. This can be attributed to the higher chemical strength of the hydrophobic coating with the substrate deposited via iCVD. Vinyl groups in TCVS are covalently bound to the surface and the reaction between the surface vinyl groups and the vinyl groups in PFDA monomer binds them chemically to the substrate. On the other hand, Silane coating is a monolayer (∼2-3 nm) and hence is susceptible to deterioration in extreme environments. Consequently, the silane coating failed in the condensation chamber, and we observed FWC on the surface.
These findings make iCVD coating an ideal candidate for practical applications.

Energy storage technologies are marked by tradeoffs in energy and power density; a Ragone plot is a standard graphical representation used to investigate the relationship between these two quantities. Absolute upper limits on energy density generally derive from intrinsic thermodynamic limits. In contrast, the power density of a substance represents kinetic limits on heat, mass, or charge transfer which limit the rate of energy storage. Viewed through this lens, the relationship between these two quantities represents how much of the potential or theoretical energy storage density must be sacrificed to achieve a particular rate of energy storage. Ragone relationships offer an approach to critically analyze the dependence of material performance on different design parameters.
Phase change materials (PCMs) are of intense interest due to their ability to passively store thermal energy in order to mitigate a transient temperature rise, or to improve the overall efficiency of a system. While PCMs possess high effective thermal capacitance, due to the latent heat of transformation, many suffer from low intrinsic thermal conductivity. To increase conductivity, PCMs can be combined with materials with higher thermal conductivities (e.g., metals, graphite, etc.) to create composite PCMs. In this case, the maximum energy density of a composite PCM volume is defined primarily by the intrinsic material properties of the PCM phase, the fraction of the composite composed of PCM phase, and the temperature rise. In contrast, the power density of a composite PCM volume is a non-trivial quantity, as it depends strongly on the many factors that affect heat transfer rate, including intrinsic material properties, operating conditions, the geometry of thermal energy storage (TES) modules, the coupling between the composite PCM volume and the design of the PCM composite itself.
In this study we develop and validate a numerical finite difference model to interrogate heat transfer from a heat transfer fluid into a planar or cylindrical volume composed of a composite PCM. From this, we extract energy density and power density Ragone relationships over the period of charging or discharging (5 to 5000 s). Here, we focus on the effects of composite structure, including the relative volume fraction of PCM and conductive components, the role of component orientation and anisotropic transport, and the interaction between different PCMs (paraffins, salt hydrates, low melting point metals) and conductive elements (aluminum, graphite). Systems which are optimized for longer timescales allow the system to approach its maximum energy storage density, as minimal mass and volume is devoted to enhancing heat transfer into the composite PCM bed. As an example, a planar module composed of aluminum and hexadecane (a paraffin wax) optimized to store energy over a period of 500 s can achieve energy densities (180 J/g; 175 J/cm³) that are approximately 70% of the intrinsic limit, corresponding to relatively low power densities (0.36 W/g, 0.35 W/cm³). In contrast, a module of the same geometry that is optimized to store energy over a period of 50 s can achieve energy densities (130 J/g; 155 J/cm³) that are approximately 50% of the intrinsic limit, corresponding to an order of magnitude larger energy storage rates (2.6 W/g, 3.1 W/cm³).[SPJ1] These results reveal the interactions between different thermophysical properties and composite design parameters, and also allow for efficient optimal design of energy storage components to meet application specific requirements.

Materials informatics (MI), developing or studying materials with an aid of informatics or machine learning, has become popular in academia and industry. Use of MI for heat transfer started later than other fields but now there are growing number of reports showing good compatibility. A general methodology is to train a black box model that relates basic descriptors (structure, composition, etc) and figure of merit (target properties) and predict or design a material with largest FOM. At Thermal Energy Engineering Lab at University of Tokyo, together with the collaborators, we have been working on MI for heat transfer for about 6 years, starting from nanostructure design using Bayesian optimization[1, 2]. We have applied the methodology to computationally design and experimentally realize aperiodic superlattice that optimally impedes coherent thermal transport[3] and multilayer metamaterial with ultranarrow-band wavelength-selective thermal radiation[4]. While Bayesian optimization has limitation in the size of search space for global optimization, considering the complex system with many dimensions or parameters needed for further MI, the challenge now is to develop a scheme that can efficiency handle massive material search space. Quantum annealing is a powerful tool since it can quickly solve combinatorial optimization problems of the size that scales with the power of number of qubits, and number of qubits commercially available for use is increasing year by year. However, there are some practical challenges. One is that the quantum annealing can only solve the problem in form of Ising model, and this has been overcome by using factorization machine [5]. Another is that even if optimization is quick and large, it is only useful when the speed and size are well balanced with those of data generation. In that sense, radiative heat transfer problem, such as designing metamaterial for wavelength-selective thermal emitter, is most suited to go massive because calculation of electromagnetic field is light and easy to parallelize. In the talk, we will introduce these progresses in applying annealers to thermal radiation problems and discuss the capability and remaining challenges for further development.

[1] Ju, Shiga, Feng, Hou, Tsuda, Shiomi, Physical Review X, 7, 021024 (2017).
[2] Yamawaki, Ohnishi, Ju, Shiomi, Science Advances, 4, eaar4192 (2018).
[3] Hu, Iwamoto, Feng, Ju, Hu, Ohnishi, Nagai, Hirakawa, Shiomi, Physical Review X, 10, 021050.
[4] Sakurai, Yada, Simomura, Ju, Kashiwagi, Okada, Nagao, Tsuda, Shiomi, ACS Central Science 5, 319-326 (2019).
[5] Kitai, Guo, Ju, Tanaka, Tsuda, Shiomi, Tamura, Physical Review Research, 2, 013319 (2020).

Near-field heat radiative transfer (NFRHT) is based on the classical Fluctuation Dissipation Theorem that requires the heat produced in macroscopic bodies upon absorbing EM radiation creates temperature fluctuations that is reciprocated by temperature fluctuations producing EM radiation. But NFRHT in nanoscale gaps between hot and cold bodies is governed by quantum mechanics (QM) and not classical physics. Indeed, the Planck law precludes temperature fluctuations in nanoscale gaps upon absorbing heat. However, all known NFRHT theories are based on the difference of surface temperatures of nanoscale gaps. In effect, the Planck law requires theories of NFRHT that do not depend on temperature which means all known NFRHT theories are invalid.

One such temperature independent NFRHT theory is simple QED based on the Planck law that conserves heat by creating temperature independent EM standing or travelling waves across the evacuated nanoscale gaps. But the heat requires EM confinement across the gap to a half-wavelength having the dimension d of the gap. In simple QED, the EM confinement is provided by the momenta of the thermal kT energy of atoms participating in the heat flow between hot and cold bodies, the hot and cold momenta directed toward their respective gap surfaces because of the thermal gradient with gap surface temperatures at absolute zero, the gap surface temperatures taken at absolute zero because the Planck law requires Brownian motion to cease everywhere in the nanoscale gap. Under EM confinement by opposing thermal momenta, heat flows if the hot momentum exceeds its cold counterpart

Simple QED highlights the stark difference between QM and classical physics illustrated by the emission of bright white-light from a Joule heated single graphene layer (0.337 nm) over a nanoscale trench reported in the literature. The application of NFRHT to the nanoscale trench is difficult to implement as the Joule heat from the graphene layer into the nanoscale trench is not known. Instead, the Fourier equation was applied to the graphene layer assuming Joule heating produces electron temperatures of 2800 K giving graphene temperatures near 2000 K. But bright-white light may require higher temperatures, say up to 5000 K.

In contrast, simple QED argues the graphene remains at ambient temperature. Instead, 1.8 keV soft X-rays are produced across the graphene thickness that fluoresce down to about 200 eV, the heat of which is either lost to the surroundings or enters the trench, the latter by simple QED creating IR waves standing in the trench, the overtones of the IR producing the observed bright white-light absent high temperatures. Consistent with simple QED, the graphene does not acquire 2800 K electron temperatures evidenced by the observation of white-light at 10 K. Indeed, there is a stark difference between QM and classical physics and should be expected whenever classical physics including NFRHT is used to explain observations at the nanoscale.

In an effort to protect metal substrates from extreme heat, polymer–clay multilayer thin films are studied as expendable thermal barrier coatings. Nanocomposite films with a thickness ranging from 2 to 35 μm were deposited on steel plates and exposed to the flame from a butane torch. The 35 μm coating, composed of 14 deposited bilayers of tris(hydroxymethyl)aminomethane (THAM)-buffered polyethylenimine (PEI) and vermiculite clay (VMT), decreased the maximum temperature observed on the back side of a 0.32 cm thick steel plate by over 100 °C when heated with a butane torch. Upon exposure to high temperature, the polymer and amine salt undergo pyrolysis and intumesce, subsequently forming a char and blowing gas. The char encases the nanoclay platelets, and a ceramic bubble is formed. The macro-scale bubble, in tandem with the nanocomposite coating properties, increases resistance to heat transfer into the underlying metal substrate. This heat shielding behavior occurs through radiative effects and low aggregate through-plane conductivity resulting from multilayer nanodomains and intumesced porosity (i.e., conduction through the gas as the film expands to form a ceramic bubble). These relatively thin and lightweight films could be used to protect important metal parts (in automobiles, aircraft, etc.) from fire-related damage or other types of transient high-temperature situations.

Enhancing evaporation processes via nanoengineered materials design offers unique opportunities for a range of energy applications. In this work, we show our ability to enhance evaporation processes and new device designs with various engineered materials sets. First, we show how we achieve a passive high-temperature solar steam generator for medical sterilization. The high performance leverages optimized ultra-transparent silica aerogel to achieve efficient thermal concentration to locally increase the heat flux and temperature. We show nearly a 2x higher energy efficiency compared to past work, and successful sterilization in field tests in Mumbai, India. Next, we show novel architectures that take advantage of thermal localization and multiple stages for vaporization enthalpy recycling to achieve ultra-high efficiency passive solar desalination. With a ten-stage design with scalable and low-cost materials, a solar-to-vapor conversion efficiency of 385% and total production rate of 5.78 Lm^-2 h^-1 under one-sun illumination was achieved. We also present techno-economic analysis that highlights the balance between water production, reliability, and cost. The works presented here promises new opportunities for the development efficient solar thermal systems in energy conversion, storage, and transport, as well as water applications.

Wide bandgap electronics are currently under development due to their potential to create future power electronics. The growth of materials based on gallium nitride and more recently gallium oxide is expected to help create technological advancements that may yield a range of devices that operate with more efficiency, higher operational frequency, and smaller form factors.
In this talk, we will discuss a range of materials and device architectures that are being developed to enable efficient heat dissipation from both GaN and Ga₂O₃ devices starting at the device level. We will discuss materials integration issues that will enable lower thermal resistance device architectures to assist with thermal management strategies. This will include bonding of heterogenous layers as well as direct growth methods. Finally, we will discuss new aspects of materials processing and properties to create a low CTE mismatch cooled power substrate for packaging power devices. At each step, we will show how considerations for materials development and methods for scalable manufacturing are necessary to help transition these advancements to applications.

Power densification of modern electronics is greatly shaped by emerging thermal management technologies. Heat spreaders represent a thermal routing technique implemented to reduce thermal resistance and limit temperature variations. Here, we demonstrate the monolithic integration of copper (Cu) conformal coatings directly on electronics to serve as heat spreaders and temperature stabilizers. We developed a fabrication recipe to electrically insulate the electronics with Parylene C and coat them with the Cu heat spreaders, demonstrating better proximity to heat generating elements, removal of thermal interface materials, and enhanced cooling performance compared to standard thermal components. Our results show that 8 µm thick Parylene C can handle up to 600 V, and that the Cu coatings provide device-to-ambient specific thermal resistance as low as 2.5 (cm².K)/W in quiescent air and 0.8 (cm².K)/W in quiescent water, with time constants of 110 s in air and 5.61 s in water. Our cooling technology has the potential to enable the continued miniaturization of electronics, potentially extending Moore’s law and greatly reducing the energy consumption used for electronics cooling. Furthermore, by removing the need for large external heat sinks, our approach should enable the realization of very compact and highly power dense co-designed power converters.

One major challenge in printing conductive metal inks onto heat-sensitive plastic substrates is the adverse effects of high temperatures experienced during the sintering process and during operation as a result of joule heating, which leads to substrate damage and current carrying capacity limitations. To combat this issue, incorporating an electrically insulating film interposed between the substrate and the printed feature can aid in spreading heat laterally away from the substrate-feature interface without sacrificing conductive performance or rigidity of the plastic substrate. Boron nitride nanotubes (BNNTs) in particular exhibit relatively high thermal conductivity and impressive thermal stability, and therefore present a promising platform to minimize thermal degradation of the printed feature and the substrate over time.

We report on a highly transparent BNNT heat-spreading interlayer for thermal management of silver components screen-printed onto polyethylene terephthalate (PET) substrates. Printed features of varying widths are investigated to optimize the transparency and conductivity of the resulting design. Upon exposure to increasing photonic energies, the BNNT layer actively dissipates the heat through the tube network, resulting in morphological improvements in both the feature and the substrate and reducing sheet resistance by up to 200%. Thermal imaging demonstrates a reduction in localized areas of excessive heat accumulation, achieving greater uniformity in heat distribution throughout the printed feature upon applied current. In summary, the BNNT thin film provides a simple approach to shield the substrate from deformation and damage at high temperatures while simultaneously improving electrical performance of the printed features and structural integrity of the substrate.

Traditionally it has been assumed that battery thermal management systems should be designed to maintain the battery temperature around room temperature. That is not always true as Lithium-ion battery (LIB) R&D is pivoting towards the development of high energy density and fast charging batteries. As the energy density and charge rate increases, the optimal battery temperature can shift to be higher than room temperature. To improve the temperature uniformity and avoid excessive internal temperature rise, heat transfer inside the battery needs to be enhanced, and reducing the thermal contact resistance between the electrodes and separator can significantly increase the effective thermal conductivity of batteries. In this talk various challenges and latest developments related to heat generation, thermal transport and properties of LIBs will be discussed. The importance of heat of mixing due to ion diffusion during fast charging will be also highlighted. Finally operando measurement of Lithium concentration inhomogeneity and plating under fast charging conditions using thermal wave sensing will be presented.

A novel ceramic-ceramic fiber composite for use at cryogenic temperatures is presented in this research. Thermal properties, mechanical properties, and electrical properties are compared to more standard polymeric-ceramic fiber composites such as G-10 and CTD-101K/S-glass. It is shown that this composite can function in certain areas to reduce thermal impedance, while retaining sufficient mechanical properties and electrical isolation capability. Multiphysics simulations are used to predict improvements to applications such as cryogenic motors and magnets. Processing routes will be presented to incorporate this composite into windings.

When a resistive random access memory (RRAM) memory crossbar array is switched repeatedly, a considerable amount of Joule heat is dissipated in the cells of the array [1,2]. A conductive nanofilament (CF) is heated in one cell by switching the cell frequently on and off and the electrical conductance of the CFs of the neighboring cells is measured. The neighboring cells have been set to a conductive low resistance state (LRS). The CF-to-CF heat transfer comprehends nanoscale heat transport from CF to the mesoscopic electrode and from mesoscopic electrode to another CF. When, under such heating conditions the electric properties of the neighboring cells are tested immediately after the heating of the heated cell has ended, we find that in most cases the neighboring cells are in in high resistance state (HRS) despite having been set to an LRS state prior to the heating. Thus remote heating ruptures the CFs, i.e. erases the neighboring cells. The probing of the neighboring cell is done at voltages much smaller than the minimum set voltage. The key findings are: (i) only for robust CFs (low resistance R_on) formed at high compliance current I_cc>0.2mA we find that the probed cell remains in an LRS state just after the heating of the heated cell. In most cases, we find slightly degraded resistance R_on compared with the original R_on value. (ii) For all probed cells that have been set at I_cc<0.2 mA (high resistance R_on), we find that the probed cell is in HRS immediately after completion of the heating of cell. However, when those cells are probed again, after 2 min or longer, we find the cell does no longer persist in HRS but returns of its own accord back to the original LRS state, with R_on in most cases the same as the original R_on. Thus, during the heat dissipation period, the CF in the probed cell spontaneously recovers and restores itself to the original R_on value to which they have been set prior to the heating. (iii) In a few cases we managed to capture the time evolution of the spontaneous restoration of the CF to LRS demonstrating the transient, non-equilibrium behavior of the heat transfer phenomenon. The heat dissipation time constant for our devices is of the order of 15s and depends strongly on the thermal conductivity of the metal electrodes. This behavior can be explained in terms of 3D resistor network with a unit resistance R_o=1/G_o with G_o=(2e²/h)×t, where t is the electron tunneling probability from Cu atom to Cu atom of the CF and depends on the distance L between the two Cu atoms, thus t~exp(2kL) where the wave vector k depends on the barrier height between the two Cu atoms. The truncated pyramidal resistor 3D network describes accurately the resistance of the CF as a function of the number of surface Cu atoms at the vertex of the pyramid. We demonstrate experimentally the presence of the quantum conductance nature of the filament’s resistance for this kind of Cu nanofilaments. The experiments point to a tight interaction between phonons of the oscillating Cu atoms and the electron tunneling rate between the Cu atoms. Harmonic oscillator model argues that the tunneling probability t is proportional to exp(-K√T) with a constant K>0, implying that t decreases exponentially with increasing temperature and thus explains the transient effect of the heat dissipation. As shown elsewhere the local temperature T in a CF can be as high as 1500K [3]. We also show that the same heat transfer effects are substantially mitigated when the thermal conductivity of the electrodes is substantially increased. With modified electrodes replacing a 150nm thick Cu electrode by a 250nm thick and 50 nm thick Pt electrode by a Pt(50nm)/Cu(200nm) composite electrode the time dissipation constant can be reduced to less than 1 second.

M.Al-Mamun, M.Orlowski, , Electronics & Photonics 4(48), 2593-2600, 2019
M.Al-Mamun, M.Orlowski, Electronics, 9. 127, p.2-11, 2020
[3] M. Al Mamun and M. Orlowski, Scientific Reports, 11:7413, 2021

Machine Learning (ML) and feature engineering have had success in predicting the thermoelectric (TE) performance of inorganic materials. Generally, there are two steps: the first involves featurizing the materials into numerical inputs in machine-readable format. The second step is the use of regression models to predict their TE performance, typically the powerfactor given by S²σ. We have used this methodology to screen for potential undiscovered materials using a known database of compounds; we found that this ML screening process self-discovers Half-Heuslers as good materials with no prior knowledge. In addition, going beyond screening, there are a wide range of unexplored inorganics, which are not in any existing databases. I will describe our use of generative models, similar to generating new images learning from existing ones, that are expected to have high TE performance¹. Finally, I will talk about our efforts at synthesizing some of these new predicted compounds and describe our progress in defining synthesizability of such new materials.
¹https://arxiv.org/abs/2005.07609

Certain metal phosphides have recently been proposed as potential thermoelectric materials [1]. In particular, the simple binary CuP₂ possesses multi-valley degeneracy in the valence band and low-frequency acoustic phonon modes. These features are expected to enhance its electrical properties and lower its thermal conductivity respectively.

As CuP₂ has not been experimentally studied as a thermoelectric material yet, we have synthesized CuP₂ thin films for rapid thermoelectric characterization on specialized chips. The films are grown by reactive sputtering of a Cu target in a PH₃-containing atmosphere under a high phosphorus chemical potential.

Thermoelectric characterization is performed on CuP₂ films with different Cu/P ratios and thermal treatments. The films are deposited on multi-purpose chips for simultaneous characterization of electrical conductivity, Seebeck coefficient, and thermal conductivity as a function of temperature. We generally find high Seebeck coefficients, low thermal conductivities, and p-type conductivity. The electrical conductivity is rather low, owing to a low hole concentration in the films without external doping. We suggest that suitable p-type dopants should be found to increase the hole concentration. If these efforts are successful, CuP₂ would be a highly attractive earth-abundant thermoelectric material.

[1] Pöhls et al., J. Mater. Chem. C 5, 12441 (2017).

In this work, we assess the full range of n-type and p-type dopability achievable in the group IV-VI binary chalcogenides with A1B1 stoichiometry, well-known to be excellent thermoelectric materials. Despite being well-studied, they are more commonly observed as p-type thermoelectric materials while n-type performance is more rarely reported in the literature. Yet, n-type thermoelectric performance is expected to be nearly as good as (SnSe, SnTe) and even better (SnS) than p-type, which has driven recent experimental interest in realizing n-type material. Therefore, we use first-principles methods to assess the limits of dopability in three Sn-based chalcogenides, SnS, SnSe, and SnTe, and establish whether there is room for improvement in reported n-type performance. By analysis of phase stability and defect chemistry, we assess the full range of dopability ranging from maximum p-type to maximum n-type that is possible in these compounds. While SnTe is degenerately p-type in all environments, we find that SnS and SnSe exhibit windows for both p-type and n-type dopability in the proper growth environment. We consider Bi and Br as candidate n-type dopants. This reveals that n-type carrier concentrations of 5(10¹⁷) cm^-3 and 1.3(10¹⁸) cm^-3 for SnS and SnSe respectively are achievable, which are in good agreement with experimentally reported concentrations. The differences in the dopability across all three compounds are attributed to differences in band edge positions and the allowable ranges of chemical potentials.

Thin film evaporation is important in the design of heat pipes, desalination, lubrication, air conditioning, and medical devices. Existing theories suggest that with smaller film thickness, the local evaporation mass flux will increase due in part to an increased interfacial temperature that results from lower conduction resistance in the liquid. However, this mass flux will sharply decrease to zero as the film gets even thinner due to the increased interface forces that inhibit the escape of interfacial vapor molecules.
In this project, we present the first experimental effort to locally measure the evaporation mass flux in a liquid meniscus. The microscale lateral extent of the meniscus and sub-Kelvin liquid-vapor temperature differences (superheats) challenge conventional thermometry. Here, Frequency domain thermoreflectance (FDTR), a non-contact laser-based method with a lateral resolution of micrometers, is used. A non-monotonic experimental signal change is observed as a function of position in the thin film portion of the meniscus, where a high evaporation heat flux is expected. This experimental technique has been hitherto widely employed with micron-scale lateral resolution in thin solid films, but not when evaporation is present. To extract the evaporation heat transfer coefficient in the meniscus region from the obtained signal, we modify the analytical solution to the heat diffusion equation to account for the phase change. This modification provides a new paradigm for investigating heat flux discontinuity in thermal models. Furthermore, to account for the thickness variation in the meniscus we also analyze the data using a finite element simulation performed with ANSYS. The ANSYS simulations are first validated against the modified FDTR equation for uniform liquid thicknesses. The ANSYS model is also able to capture account for the variation in evaporative heat transfer coefficient that occurs within the region interrogated by the 3 ��m laser spot. A machine learning framework that uses neural networks is then used to find the optimal combination of meniscus thickness and evaporative heat transfer coefficient distribution to match the ANSYS model to the experimental data.

Heat transfer across organic-inorganic heterojunctions and solid/liquid interfaces, is crucial for applications such as photothermal cancer therapy and flexible hybrid thermoelectric materials. However, fundamental knowledge of heat transport at atomic scale is still limited, due to challenges to experimentally probe heat transport with an atomic spatial resolution. In this talk, we will discuss our recent results to experimentally measure the heat transfer across the Au/thiol heterojunctions. We integrate two pump-probe techniques, the picosecond transient absorption and time-resolved Raman spectroscopy, to concurrently monitor heating and cooling of gold nanorods and bonds in the conjugated ligands, with a sub-picosecond time resolution and an atomic spatial resolution. We find that bonds in the conjugated ligands are heated almost instantaneously and reach a peak temperature within ~1 ps after the nanorods were heated by the laser pulse. We attribute this fast heating to direct heating of bond by the non-equilibrium electrons in gold nanorods, due to the remote coupling across the Au-thiol heterojunction. Our analysis suggests that the remote coupling could contribute substantially to heat transport across Au-thiol heterojunctions.

Resistive thermometer materials enable transduction of temperature to a measurable electrical signal. Detection of this signal is limited by the material’s temperature coefficient of resistance (TCR), the degree to which its fractional resistance changes in response to a change in temperature. AuGe has been demonstrated to yield a high TCR at low temperatures, which lends itself towards applications in cryogenic thermometry. Here, we investigate the role of annealing temperature and film thickness on the TCR of Au_0.17Ge_0.83, a composition of particular interest. Temperature-dependent electrical resistance measurements spanning 10 K to 300 K provide insight into the performance of the film, demonstrating orders-of-magnitude enhancement in sensitivity over commonly used materials such as pure metallic platinum. Furthermore, we observe that a movement towards thinner films provides additional enhancement in TCR with lower required annealing temperatures.

This work was performed, in part, at the Center for Integrated Nanotechnologies, an Office of Science User Facility operated for the U.S. Department of Energy (DOE) Office of Science. A portion of this work was supported by the Laboratory Directed Research and Development program at Sandia National Laboratories. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

I will summarize several collaborative efforts to develop new non-contact methods for heating, thermometry, and thermal property measurements. The main focus will be on nanoscale techniques using SEM. The incident electron beam can serve as a point heat source, and in certain conditions the local sample temperature can also be deduced from the electrons leaving the sample. I also will briefly introduce an all-optical “structured illumination, thermal imaging” (SI-TI) technique which: can tolerate rough, porous, and/or scratched samples; enables parallel measurement of multiple samples in a single field of view; and uses software-controllable heating patterns which can be tuned to extract directional properties like the thermal conductivity tensor.

The 3 omega (3ω) method is a trusted technique for measuring thermal conductivity – a fundamental material property of critical importance in a broad range of applications. However, traditional 3ω sensor processing requires some form of physical vapor deposition, such as metal evaporation or sputtering. Such 3ω sensor deposition techniques limit the materials and sample sizes applicable to the 3ω method. This work demonstrates an aerosol jet printing method to directly print silver 3ω sensors that yield accurate temperature-dependent measurement up to 300°C on materials with thermal conductivity ranging from 13 to 150 W/m-K. Thermal conductivity measurement with 3ω sensors conventionally sintered at 300°C agrees to independent laser flash measurement within 4% from room temperature to 150°C. Beyond 150°C, printed sensors display thermally activated tunneling effects that introduce significant errors in 3ω measurements. An unconventional rapid high-temperature sintering method is shown to eliminate the tunneling effects, producing sensors that agree within 3% of laser flash measurements from room temperature to 300°C. The rapid sintering profiles also reduced the sensor-substrate thermal boundary resistance of the printed sensors by as much as 88%. The direct printing of 3ω sensors creates opportunities for measurement of thermal transport properties in applications previously inapplicable to the 3ω method while the ultrafast sintering technique to eliminate tunneling effects and significantly reduce the sensor-substrate thermal contact resistance has implications for many types of aerosol jet printed sensors.

Phonon transport properties, such as phonon relaxation-time and free path distributions, play an essential role in the study of nanoscale solid-state heat transfer. Reliable descriptions of these properties are required for modeling heat transport at the kinetic level, which becomes necessary due to the failure of Fourier-based analyses at micron/sub-micron scales. Applications of practical interest include improved heat management in nanoelectronic circuits and devices, microelectromechanical sensors, and nano-structured materials for improved thermoelectric conversion efficiency.

Theoretical approaches such as density functional theory (DFT) have been widely used to calculate phonon transport properties, but have yet to gain widespread acceptance, in part due to their inability to consistently reproduce experimental observations. An alternative class of approaches relies on extracting material constitutive information from thermal spectroscopy experiments. However, the analysis of thermal spectroscopy data remains a challenging task. One such analysis technique seeks to extract the cumulative distribution function of thermal conductivity as a function of the phonon free path, F(Λ), from the experimentally measured thermal responses using the "effective heat conductivity" concept. This approach attempts to match the experimentally measured response to solutions of the heat conduction equation with the thermal conductivity (or thermal diffusivity) treated as an adjustable, "effective", property. Unfortunately, as it was discussed previously (see [Physical Review B 94, 155439 (2016)] and [Physical Review B 97, 195440 (2018)]), among many other approximations, this procedure implicitly assumes that heat transport is Fourier-like, which is only justified under fairly restrictive conditions of late times and large scales that are not always satisfied under experimental conditions.

In this work, we discuss some fundamental issues associated with the analysis of thermal spectroscopy experiments, such as the uniqueness of reconstructed quantities (e.g., F(Λ)), as well as their ability to reproduce the experimental results (e.g., temperature profiles). In particular, we show that approaches that are based on reconstruction of F(Λ) (as a means to defining "effective heat conductivities") fail to guarantee a unique thermal response. This can be understood by noting that it is the relaxation-time and not F(Λ) which appears as a primary material-parameter input to the Boltzmann transport equation (BTE). Specifically, we show that the same F(Λ) can be arrived at from multiple relaxation-time functions, which, in general, will correspond to multiple thermal responses. This observation seriously questions the reliability of approaches which use F(Λ) in the reconstruction. This implies that in the context of thermal transport at the nanoscale within the Boltzmann relaxation-time approximation framework, the "fundamental" property that should be reconstructed from the thermal spectroscopy experiments is the frequency-dependent relaxation-times function (provided group velocities are known), since it explicitly appears in the BTE as the input material property.

In response to the failure of effective heat conductivity-based methods, we developed and extensively validated (on both synthetically generated and experimentally measured data) an optimization-based methodology for directly reconstructing frequency-dependent phonon relaxation-times from thermal spectroscopy data [Physical Review B 94, 155439 (2016); Physical Review B 97, 195440 (2018); Applied Physics Letters 114, 026103 (2019)]. Extensive validation, including global optimization studies, shows that this formulation provides numerically unique solutions and is robust to the noise in the measurement.