Symposium NM08—Nanoscale Heat Transport—Fundamentals

As the dominant thermoelectric material around the room temperature, Bi₂Te₃-based alloys have been intensively studied for its applications in power generation and refrigeration. Nanograined bulk Bi₂Te₃ can be synthesized by grinding down bulk Bi₂Te₃ or mixed element ingot into nanopowder and then hot pressing the nanopowder into a bulk disc. The obtained nanostructured bulk material has even better thermoelectric performance due to the largely suppressed lattice thermal conductivity by high-density grain boundaries (GBs) inside the material. In kinetic theory, lattice thermal conductivity reduction is attributed to the effective phonon mean free path (MFP) reduced from the bulk value due to the restriction of grain sizes. In this aspect, atomistic simulations suggest bulk phonon MFPs in Bi₂Te₃ are below 100 nm along major axis directions. Therefore, a reduced effective phonon MFP cannot explain the lattice thermal conductivity reduction observed in micro-grained Bi₂Te₃. To better understand thermal transportation in crystalline Bi₂Te_3, high-quality single crystals for accurate thermal measurements and modeling is needed.
In this work, high-quality Bi₂Te₃ nanoflakes are synthesized using chemical vapor deposition on a SiO₂/Si substrate. Thickness-dependent thermal conductivity has been measured along the cross-plane direction using the time-domain thermoreflectance (TDTR) method at room temperature. The film thickness ranges from 20 nm to 300 nm, as confirmed using the atomic force microscopy.
To explain the experimental data, first-principles calculations are carried out to extract both in-plane and cross-plane phonon MFP distributions for bulk Bi₂Te₃. Calculations based on the computed cross-plane phonon MFPs and measurement data agree well with each other.
These validated bulk phonon MFPs are then used to explain the reported lattice thermal conductivities of various polycrystalline Bi₂Te₃ and its alloys. It is found that the observed lattice thermal conductivity reduction is mainly attributed to the GB thermal resistance that results from the GB reflection and transmission of phonons. Particularly, optical phonons are largely filtered out by GBs to explain the data for micro-grained samples. This result could provide important guidance for thermal studies of general polycrystalline materials.

At the nanoscale, enhancing the propagation distance of heat carriers in solids attracts intensive attention for improving the thermal performance in micro- and nanoelectronics. However, the major heat carriers, phonons, propagate ballistically over the distances shorter than one micrometer above room temperature. To compensate this limited propagation range, we propose a novel heat carrier, surface phonon polariton (SPhP), which is generated from the hybridization of the optical phonon and the photon and propagate in the range of a few hundred micrometers. In this study, we conduct a 3ω measurement of SiN suspended submicron films to investigate the thermal transport by SPhP in hundreds of micrometers range.

As a result, we report the thermal conductivity (TC) measurements in silicon nitride (SiN) films with different thicknesses (h = 30, 50, 100, and 200 nm) measured in the 300 - 400 K temperature range. We detect the higher TC with the sensor further from the heater (d = 200 μm) in thin films (h = 30, 50 nm). This phenomenon suggests that the heat propagation due to SPhPs remains quasi-ballistic in the hundreds of micrometers range, which is consistent with our previous study [1]. When the 30-nm-thick film heats up to 400 K, we detect approximately 13% enhancement of the TC compared to 300 K. Meanwhile, for 100- and 200-nm-thick membranes, the TC is nearly independent of both temperature and the distance. The results presented here have potential applications in the fields of heat transfer, thermal management, near-field radiation, and polaritonics.

[1] Y. Wu et al. “Enhanced thermal conduction by surface phonon-polaritons”, Science Advances 6, eabb4461 (2020)

Surface phonon-polaritons (SPhPs) are surface waves generated by the coupling of electromagnetic waves and optical phonons at the interface of polar materials. In suspended nanofilms, SPhPs propagate over typical distances of a few hundreds of micrometers along their interfaces [1], which positions them as powerful energy carriers to transport energy along the plane of nanomaterials. However, such nanometric films are mechanically unstable and hence their practical application is limited. In this work, we study the SPhP propagation along a three-layer structure made up of a Si film, acting as a mechanical support, sandwiched by two nanofilms of SiO₂, which is the SPhP active material. This is done by determining the SPhP propagation length as a function of the spectral frequency and thicknesses of the Si and SiO₂ films. Longer propagation lengths are obtained for thinner SiO₂ films, due to the reduction of energy absorption [2]. By contrast, as the Si film thickness reduces, the propagation length increases (decreases) for thicknesses smaller (greater) than the SPhP wavelength. This behavior is generated by the tradeoff between the energy absorption and coupling of SPhPs propagating along the top and bottom SiO₂ layers. It is shown that the resulting SPhP contribution to the in-plane thermal conductivity of the considered structure exhibits a minimum value for a given Si film thickness. The obtained results thus shed light on the physical understanding of the propagation of SPhPs and the modulation of their thermal transport inside a layered and mechanically stable structure.
[1] Chen D.Z.A., et al., Phys. Rev. B 72, 155435 (2005)
[2] Tachikawa S., et al., Nanomaterials 10, 1383 (2020)

Radiative heat transfer is said to be in the near-field regime when the separation distance between the heat exchanging media is smaller than or comparable to the dominant thermal wavelength (λ_max ≈ 10 mm at room temperature). Near-field radiative heat transfer can be quasi-monochromatic and exceed the blackbody far-field limit by several orders of magnitude. The enhancement beyond the blackbody limit is due to the extraneous contribution to radiative heat transfer by evanescent waves that are confined to a distance of about the thermal wavelength from the surface of the thermal emitter. Near-field radiative heat transfer can be quasi-monochromatic when the heat exchanging media support surface phonon and plasmon polaritons. These unique properties of near-field regime of radiative heat transfer are highly promising for many advanced and state-of-the-art technologies such as waste heat recovery, nanoscale-resolution imaging, nanomanufacturing, thermal management of electronic devices, and thermal rectification through a vacuum gap. The quasi-monochromatic behavior of near-field radiative heat transfer plays a central role in these technologies. Development of the near-field applications requires materials with plasmonic or phononic resonances in the mid-infrared where these modes can be thermally excited.
The number of materials with plasmonic/phononic resonances in the mid-infrared is very limited, and the spectral location (wavelength) of these resonances does not often match the ones required by practical applications. Graphene supports surface plasmons in the mid-infrared and terahertz ranges. It is thermally very stable, and due to the two-dimensional nature of graphene, free electrons or holes can effectively be induced into this material using chemical doping or electrical gating rendering dramatically tunable optical properties. Due to these advantages, graphene has recently gained significant attention for increasing or modulating thermal radiation in the near-field regime. In this presentation, the possibility of tuning the spectrum of near-field radiative heat transfer using graphene-based materials is discussed. Radiative heat transfer between graphene sheets of various chemical potentials, temperatures, and separation distances is analyzed. The dispersion relation of graphene plasmons is derived, and the conditions under which graphene plasmons can result in quasi-monochromatic near-field conductance is discussed.

Phonon hydrodynamics is the framework to understand thermal transport in materials where momentum conservation is important. This is the case of 2D materials like graphene or transition metal dichalchogenides but also 3D materials like graphite and diamond. In these materials, momentum conservation is granted by the dominance of the normal phonon-phonon collision.
In the last decades, phonon hydrodynamics has been shown to offer also an explanation for the anomalous thermal behaviour of semiconductors like silicon or germanium, where normal scattering is not dominant. Despite also based in the momentum conservation, the appearance of hydrodynamics in materials like silicon is related to surface and interface physics. In this case, a boundary layer is created that is the responsible to transfer momentum between the surfaces and the inner regions of the materials. In that layer, conservation rules are also crucial and modify the heat equations. The final result is that a hydrodynamic equation like Guyer and Krumhansl is also valid to describe these materials.
We present a multiscale analysis going from the microscopic fundamentals for the appearance of a hydrodynamic behaviour in silicon to the experimental evidences in favour of using the Guyer and Krumhansl equation to describe this kind of materials. We will also demonstrate that in the samples where this happens, traditional approaches based on an effective Fourier equation are completelly ruled out.
Moreover, the proposed framework allows to understand why using a Fourier approach, a size dependent thermal boundary resistance at nanoscale contacts between metals and substrates is required. It will be shown that this size dependence understood through the hydrodynamic behaviour of the substrate near the metal line.

Recent experiments revealed non-Fourier thermal conduction in general semiconductors like Silicon [1-3]. The use of a hydrodynamic heat transport equation has shown to capture thermal transport beyond Fourier’s law through new phenomena like phonon vorticity and viscosity [4-6].

We focus on the process of energy release from a nanoscale heat source towards a Silicon substrate. By monitoring the heater thermal relaxation through EUV-scatterometry techniques [2], we demonstrate that the heater temperature follows a double exponential decay [7]. This system response cannot be explained if using Fourier’s law with effective parameters to model the substrate conduction, which provides a single exponential thermal decay followed by a power law. In contrast, the hydrodynamic model with geometry independent parameters naturally predicts the two exponential decay and provides analytic expressions for each characteristic time scale. The shorter time scale (~100ps) is dominated by the interfacial resistance and the second one (~1ns) is dominated by hydrodynamic effects in the substrate. The geometry dependencies of the second characteristic time accurately predict the observed slower relaxation of the heater by reducing its size. Furthermore, the model provides a simple analogy based in an electrical circuit to illustrate the heater thermal evolution.

References:
[1] A. Ziabari et.al., Nat. Comm. 9, 255 (2018).
[2] K. M. Hoogeboom-Pot et. al., PNAS 112 16 4851 (2015)
[3] R.B. Wilson, D. Cahill, Nat Comm. 5, 5075 (2014)
[4] S. Alajlouni, A. Beardo et al. Nano Research 14 945–952 (2020)
[5] P. Torres et al. Phys Rev. Materials 2, 076001 (2018)
[6] A. Beardo et al. Phys Rev. B 101, 075303 (2020)
[7] A. Beardo, J. Knobloch, et al. submitted

We present a new formalism that solves the Boltzmann Transport Equation (BTE) and derives a hydrodynamic heat transport equation (Guyer-Krumhansl equation) [1] similar to Navier-Stokes, which generalizes the Fourier’s law [2]. This generalization applies from systems dominated by momentum-preserving collisions to kinetic materials dominated by resistive collisions. Therefore, it can be applied to general semiconductors like Silicon, in contrast to the common belief [1]; and arbitrary complex geometries by using Finite elements methods at a low computational cost.
This contrasts with other approaches, where efforts have been made to use Fourier’s law with effective coefficients depending on the physical situation [3,4] or through alternative solutions of the phonon BTE [5,6], whose complexity only allows to face simple geometries.
Our model allows the understanding of the experimentally observed Fourier’s law breakdown at small length and time scales by using geometry independent parameters calculated from ab initio, resulting in a predictive model [7-9].
Furthermore, it supplies a connection between microscopic variables (phonon properties) and mesoscopic ones (temperature, heat flux, …), providing an explicit solution for the nonequilibrium phonon distribution function. Furthermore, this approach allows to derive consistent boundary conditions for arbitrary complex surfaces and interfaces between different materials.

References:
[1] R. A. Guyer and J. A. Krumhansl, Physical Review 148, 766 (1966).
[2] L. Sendra et al., submitted to PRB.
[3] K. M. Hoogeboom-Pot et al., Proceedings of the National Academy of Sciences 112, 4846 (2015).
[4] K. T. Regner et al., Nat. Comm. 4, 1640 (2013).
[5] A. Cepellotti and N. Marzari, Phys. Rev. X 6, 041013 (2016).
[6] C. Hua et al., Phys. Rev. B 100, 085203 (2019).
[7] A. Beardo et al., Phys. Rev. Applied 11, 034003 (2019).
[8] S. Alajlouni et al., Nano Res. (2020), 10.1007/s12274-020- 3129-6.
[9] A. Beardo et al., Phys. Rev. B 101, 075303 (2020).

Understanding the thermal conductance in molecular junctions offers important insights in designing future thermoelectric devices. One of the most interesting features in molecular junctions is the possibility to exert external control through mechanical forces. We use non-equilibrium molecular dynamics simulations to study the steady state thermal conductance of different molecular junctions consisting of finite macromolecules at various end-to-end distances/electrode displacements. We find that the nature of the thermal conductance along such polymer chains strongly depends on mechanical tuning, leading to significantly different heat conductions and temperature profiles along the chain in the compressed-chain and stretched-chain limits. This transition between modes of behaviors appears to be a threshold phenomenon: at relatively small end-to-end distances/electrode displacements, the thermal conductance remains almost constant as one stretches the polymer chain. At given critical end-to-end distances/electrode displacements, thermal conductances start to increase, reaching the fully extended chain values. In addition, we find that the thermal conductances of molecular junctions sandwiched between metal leads significantly depend on the nature of metal leads and are much smaller than the corresponding intrinsic thermal conductances, i.e. in the absence of metal leads. We then address how the thermal conductance is correlated with the vibrational density of states and study the metal-molecule coupling and anharmonic effects upon stretching. We also conduct similar studies for disordered polymer chains and bundles of several polymer chains, which show remnants of the same behaviors of the aforementioned systems.

Geometry-modulated nanostructures consist a class of metamaterials promising for efficient thermoelectric applications and nanoscale heat management. In these metamaterials, it is possible to achieve enhanced thermoelectric effects and decreased thermal conductivity by designing the geometry-modulation in the nanometric scale. We indicated this possibility in 2010 when we proposed diameter-modulated nanowires for enhanced thermoelectric efficiency [1]. We showed that coupling between propagating waves and resonances of the modulation units modifies the carriers transmission coefficients and results in enhanced thermoelectric power factor and decreased electron and phonon thermal conductions [2,3,4]. Electron and phonon scattering can be controlled by designing the shape of the geometry-modulation [5,6]. Ballistic or diffusive transport effects may dominate depending on characteristic dimensions of the geometry modulation [7,8]. We will discuss physics aspects of electron and phonon transport in geometry-modulated metamaterials based on our work and the evolution of the research in this field that is currently gaining increased interest for advanced applications.

[1] X.Zianni, Applied Physics Letters 97, 233106 (2010)
[2] X.Zianni, Nanoscale Research Letters 6, 286 (2011)
[3] X.Zianni, Journal of Solid State Chemistry 193 53 (2012)
[4] X.Zianni, Microelectronic Engineering 112, 235 (2013)
[5] X.Zianni and P.Chantrenne, Journal of Electronic Materials, DOI: 10.1007/s11664-012-2304-2 (2012)
[6] X.Zianni, V. Jean, K. Termentzidis and D. Lacroix, Nanotechnology 25, 465402 (2014)
[7] X.Zianni, K.Termentzidis and D.Lacroix, Journal of Physics: Conf. Series 785, 012011 (2017)
[8] X.Zianni, J. Phys. D: Appl. Phys. 51 114003 (2018)

Layered cobalt oxide such as Na_0.75CoO₂ and Ca₃Co₄O₉ are known as good candidates for oxide thermoelectric material. Although layered cobalt oxides have several advantages such as environmental compatibility and thermal-chemical stabilities compared to heavy metal-based chalcogenides like Bi₂Te₃ and PbTe, their ZT is low as compared to that of heavy metal-based chalcogenides, mainly due to rather high thermal conductivity. In order to overcome the difficulty, we hypothesized that the thermal conductivity of layered cobalt oxide can be reduced by heavier ion substitution^[1]. Firstly, we fabricated Na_0.75CoO₂ epitaxial films on sapphire or YSZ substrate by the R-SPE method^[2]. Then, the Na ions were exchanged with Ba ions by the ion exchange treatment. The resultant film was Ba_1/3CoO₂, which was clarified from the XRD patterns and HAADF-STEM images. Although Na_0.75CoO₂ is unstable at high temperature in air, the resultant Ba_1/3CoO₂ epitaxial films show excellent thermal stability at high temperature (700 °C) in air. The room temperature thermal conductivity of Ba_1/3CoO₂ epitaxial films along the layered structure was ~3 W m^-1 K^-1, which is approximately half of that of Na_0.75CoO₂ (~5.5 W m^-1 K^-1), indicating that our hypothesis is true. Since the power factor was almost unchanged, the ZT was greatly improved (ZT = 0.11^[3] at room temperature and 0.2 at 200 °C). These ZT values are the highest among oxides ever reported as a “reliable value”. The present result clearly indicates that Ba_1/3CoO₂ shows rather large ZT at high temperatures in air, exerting a potential high temperature application prospect.

References
[1] H.J. Cho*, Adv. Mater. Interfaces 7, 1901816 (2019).
[2] H. Ohta et al., Cryst. Growth Des. 5, 25 (2005).
[3] Y. Takashima, Y. Zhang* et al., J. Mater. Chem. A 9, 274 (2020).

Thermoelectric materials enable direct conversion of waste heat into electrical energy. The conversion efficiency is inversely proportional to the thermal conductivity, which is generally dominated by phonons in semiconductors. Zintl compounds AMg₂X₂ constitute a class of new thermoelectric compounds with excellent thermoelectric performance in n-type Mg₃(Sb,Bi)₂ alloys, with zT values up to 1.6 reported so far. Mg₃Sb₂ exhibits very low lattice thermal conductivity (~1-1.5 W/m/K at 300K), comparable with PbTe and Bi₂Te₃, despite a much lighter average ionic mass. We report on neutron scattering and first-principles studies of the lattice dynamics of AMg₂X₂. Inelastic neutron scattering measurements provided the temperature dependence of the phonon density of states (DOS). Extra peaks and overall softer phonons were found at low frequency in Mg₃Sb₂ and Mg₃Bi₂ compared to CaMg₂X₂ or YbMg₂X₂. Combined with simulations, we highlight the importance of a specific soft Mg-X chemical bond that suppresses phonon group velocities and drastically enlarges the scattering phase-space, enabling the threefold suppression in thermal conductivity.

High-resolution thermometry is essential in calorimetry and bolometry with broad applications in studies of dissipation in electronics and quantum systems [1, 2], metabolism in biological systems [3], and millimeter-wave sensing for astronomy [4]. Achieving high-resolution thermometry combined with small device size, at room temperature, is highly desirable. In this presentation, we will describe a novel thermometer capable of <100 nK resolution in a 0.1 Hz measurement bandwidth. Our thermometer is composed of a microfabricated suspended GaAs device with lateral sensing dimensions of <200 µm and employs an optical readout scheme. The temperature sensing mechanism is based on the Fabry-Pérot resonance at the Urbach edge of direct band-gap semiconductors, where the optical properties of the medium, namely the absorption coefficient and refractive index, change abruptly by variation of incident photon energy relative to the band gap [5]. We show that the temperature dependence of band gap provides a highly temperature-sensitive reflection or transmission signal at the Fabry-Pérot resonance peaks. The device we describe here represents a first demonstration of a micro-thermometer for all-optical, nano-Kelvin thermometry at and above room temperatures.

[1] K. Kim, W. Jeong, W. Lee, and P. Reddy, Acs Nano 6, 4248 (2012).
[2] D. Halbertal et al., Nature 539, 407 (2016).
[3] S. Hur, R. Mittapally, S. Yadlapalli, P. Reddy, and E. Meyhofer, Nature Communications 11, 1 (2020).
[4] R. Kokkoniemi et al., Communications Physics 2, 1 (2019).
[5] S. Johnson and T. Tiedje, Journal of Applied Physics 78, 5609 (1995)

As structures in nanodevices approach the size of the phonon mean-free-path of materials, studying thermal transport properties becomes increasingly difficult due to the need for nanoscale thermal probes. Many different techniques exist to study these nanoscale effects, however, measurement speed and/or the diffraction limit of light is often a limiting factor. Here, we present new methods for measuring temperature and thermal conductivity of semiconductor nanostructures rapidly with nanometer resolution using cathodoluminescence spectroscopy in a scanning electron microscope (SEM).

First, we demonstrate the first measurements, to our knowledge, of cathodoluminescence (CL) nanothermometry in semiconductors. In semiconductor CL, a 5-30 keV electron beam from a scanning electron microscope excites electrical carriers, that recombine and emit radiation at energies near their bandgap, which is then collected and analyzed [1]. We perform CL imaging of 7-micron-long, 200-nm-diameter GaN nanowires, made by a top-down approach based on sublimation under vacuum, which are suspended across a 6.5 micron hole in a TEM membrane. The wires are thermally clamped on both sides using focused ion beam deposition of Pt and we record the bandgap emission near 3.4 eV at every position in the SEM image. By precisely measuring shifts in the bandgap, we map the temperature as the electron beam is scanned over the wire by using the Varshni relation fit to calibration data. We find a 6 K temperature resolution with our samples, limited primarily by small doping variations in the wires which can be corrected for by mapping the doping profile with a low-current electron beam with nm-resolution[2]. Using low-current electron beams which deposit less than 1 µW of power into the wire, we can measure temperature in a non-perturbative way. Using large currents of 70 nA which can deposit 100’s of μW of power, we can heat the wires by over 400 K.

Next, we will discuss three techniques for measuring thermal conductivity using this novel method of CL nanothermometry. The first two use a continuous electron beam current scanned over the wires with the beam acting as both a delta-function-like heat source and local temperature probe with resolution of the electron cascade. We fit the temperature rise as a function of position, correcting for doping variations in the wire. The final method uses an electrostatic beam blanker[3], with which we measure the frequency dependence of the thermal response of the GaN nanowires to electon beam heating in the range of 100 Hz to 5 MHz in a matter of minutes. We develop a theoretical model to fit the data and find a cut-off frequency of ~500 kHz from which we derive a thermal conductivity of 54 ± 13.2 W/m*K, which agrees within error with the other techniques we discuss. We will discuss the advantages and limitations of each technique and future directions. With minor modifications, the underlying principles of methods like time-domain thermo-reflectance are applicable to CL nanothermometry experiments, especially with the rise of electron microscopy with fs-photoexcited field-emission guns[4]. Because of the nanoscale nature of these measurements, CL nanothermometry methods could prove useful for probing thermal transport and phonon properties at small dimensions, including ballistic thermal transport, near-field heat transfer, or in studies of thermal contact resistance.

[1] Edwards P. R. and Martin R. W., Semicond. Sci. Technol. 26 064005 (2011)
[2] Chen, H.-L. et al. Nano Lett. 17, 6667–6675 (2017)
[3] S. Meuret, et al., Ultramicroscopy 197, 28 (2019)
[4] M. Solà Garcia, et al., ACS Photon. 7, 232 (2019)

Electrocaloric (EC) effect refers to the isothermal entropy or adiabatic temperature changes of a dielectric material induced by an external electric field. This phenomenon has been largely ignored for application because only small EC effects (2.6K) have been detected in bulk materials such as Pb_0.99Nb_0.02(Zr_0.75Sn_0.20Ti_0.05)_0.98O₃. The discovery of a giant electrocaloric effect by Mischenko et al. in thin films (up to 12°C with E=480kV.cm^-1 ) renewed interest in electrocaloric materials and related cooling technologies. It is indeed the development of thin structures that has made it possible to obtain high dielectric rigidity and significant electrocaloric effects.

Several methods have been developed to characterize EC effect. “Direct methods” measure the EC temperature change or released heat in the material on application or withdrawal of the electric field. ‘‘Indirect method’’ employs Maxwell equations to calculate the EC effect from the change of polarization with temperature.

In fact, ferroelectric thin films exhibit significant dielectric losses (tanδ>0.1). Furthermore, EC thin layer being deposited on a solid substrate, the thermal capacity of the assembly is very high compared to that of the thin layer alone. Thus, even if the change in specific entropy following the application of an electric field is very large, the change in total entropy is almost undetectable because of the very low volume of the thin films. The temperature variation will also be hardly detectable because of the rapid dissipation of the heat in the substrate that is behaving as a heat sink. Due to these limitations, till now it has been difficult to present experimental values of electrocaloric properties in thin films from direct measurements, only estimates by indirect characterization are accessible. It is therefore imperative to study these materials using instruments offering high resolution.

To overcome these difficulties, we propose a novel characterization method based on thermoreflectance for measuring directly the electrocaloric effect. This method allows the study of systems such as thin films, multilayers, and nanoparticle. It is a non-contact and non-destructive measurement method with a very large temporal dynamic, which allows the study of extremely rapid phenomena. It exploits the temperature-dependent reflectivity of a surface which is probed using the intensity variation of a reflected laser beam. We report laser-based direct measurements of ECE in various solid ferroelectric materials in the form of multilayered capacitor and thin films as a function of the applied electric field. Using appropriate model, the signal can be used to retrieve Electrocaloric properties of embedded or stacked electrocaloric layers.

Acknowledgment: We are grateful to Université du Littoral Côte d’Opale and the National Council for Scientific Research in Lebanon for the support of this work

The chiral anomaly in Weyl semimetals has a thermal analog [1]. In electrical measurements, an electric field applied parallel to a magnetic field and parallel to the direction of separation of the Weyl points produces a charge transfer between the Weyl points that results in an anomalous charge current and a negative magnetoresistance; this is the chiral anomaly. In thermal measurements, there is no charge transfer between the Weyl points, but an energy transfer, which produces an anomalous heat current and an increase in thermal conductivity in magnetic field. We unambiguously demonstrate this experimentally in the topological insulator bismuth-antimony alloys (Bi₈₉Sb₁₁ and Bi₈₅Sb₁₅) driven into an ideal Weyl semimetal state by a Zeeman field, with the chemical potential pinned at the Weyl points, and in which the Fermi surface has no trivial pockets. The experimental signature is a 300% enhancement of the electronic thermal conductivity in a 9T applied magnetic field parallel to the thermal gradient that follows the Wiedemann-Franz law above 60 K. The same material shows an equally strong decrease in thermal conductivity in a magnetic field applied perpendicularly to the direction of Weyl point separation. The talk will emphasize how the material is potentially useful in heat switches in applications, such as adiabatic demagnetization refrigeration, where an external magnetic field is already being used to induce ordering in a paramagnetic salt.

[1] Vu et al., arXiv:1906.02248 (2019).

Funding: Center for Emergent Materials, NSF-DMR-2011876.

Understanding the thermal generation of spin currents in magnetic materials is an important goal for the field of spin caloritronics. Among magnetic materials, rare earth iron garnets (REIG) display intriguing magnetic transport properties as result of strong antiferromagnetic exchange interactions and low magnetic damping. We report on ultrafast longitudinal spin Seebeck effect (LSSE) experiments on thin film REIG / heavy metal (HM) heterostructures (RE: Y, Tm, Eu, Tb; HM: Au). We use time-resolved magneto optic Kerr effect measurements to directly observe the transfer of magnetization from the REIG into the HM on femto-second timescales, allowing us to selectively probe the interfacial SSE. We observe a factor of 4 difference in the magnitude of the LSSE among the different REIG samples. Our results provide insight regarding the different contributions to the spin current from the different REIGs and the relevance of the interface between the REIG and the HM layers.
This work was supported by the U.S. Army Research Laboratory and the U.S. Army Research Office under contract/grant number W911NF-18-1-0364; and SHINES, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences under Award SC0012670.


References
[1] J. Kimling, G. M. Choi, J. T. Brangham, T. Matalla-Wagner, T. Huebner, T. Kuschel, F. Y. Yang and D. G. Cahill, "Picosecond Spin Seebeck Effect," Phys. Rev. Lett., vol. 118, no. 5, 2017.

We present new data on the polarization-dependent thermal conductivity of ferroelectric materials, and a new hypothesis on its possible physical origin. Hopkins and colleagues [1] report that piezoelectric materials have a thermal conductivity that is a function of the external electric field. Further, they assert that the decrease as the field is applied is due to an increase in ferroelectric domain wall density, which increases the phonon scattering. While their experiment is conducted on a thin film, we investigate the field and temperature dependency of thermal conductivity on a PZT stack with interwoven electrodes. We present data that are consistent with [1], but with a sensitivity to the applied field that is an order of magnitude larger. The thermal conductivity changes are hysteretic and follow the polarization of the material, but the signal is too large to be explained by volume or sound velocity changes. Our data hint at an alternate hypothesis, in which local thermal fluctuations of the polarization in the material may carry heat, by analogy to magnons carrying heat in ferromagnets [2].
[1] Nano Lett. 15, 1791-1795 (2015); [2] Phys. Rev. B 91 226401 (2015)
Funding: NSF DMR 2011876

The traditional normal mode analysis, which employs complex normal mode coordinates, cannot distinguish the normal (N) phonon interactions from the Umklapp (U) interactions. In our recent work [Sci. Rep. 9, 7982 (2019)], we have demonstrated that both N and U processes can be resolved from atomistic simulations by adopting a set of real asymmetric normal mode amplitudes. In the current work, we have developed an object-oriented code – Phononica, which when dynamically paired with the popular molecular dynamics (MD) simulator – LAMMPS, transforms the atomistic trajectories, on the fly, to real normal mode amplitudes that can be used for analyzing phonon interactions, and phonon modes in non-equilibrium thermal transport. Additionally, Phononica can read-in the dynamical trajectories from ab initio MD codes such as VASP or Quantum Expresso for phonon analysis; using the recently developed ZTR-2 method [Comp. Phys. Comm. 238, 124 (2019)], both crystalline and disordered materials can be analyzed even when the mean displacements are not defined or known. Phononica is particularly suited for (i) computing phonon dispersion curves, (ii) analyzing elementary phonon-phonon interactions, (iii) computing phonon modal conductivities from equilibrium and non-equilibrium simulations using our approach outlined in J. Chem. Phys., 151, 104110 (2019), (iv) analyzing defect-phonon interactions, and (v) modeling phonon transport across low dimensional molecular junctions and interfaces. In this talk, we summarize the capability of Phononica by elucidating the phonon and heat current modes in representative bulk and low-dimensional materials in both equilibrium and non-equilibrium conditions.

Cellulose is an inexhaustible naturally occurring plant-based polymer. Because of its sustainability and renewable properties, it has drawn significant research interest recently including simulation-based studies of thermal [1] and mechanical properties [2]. However, the phonon behavior in Crystalline Cellulose Iβ, especially the anharmonicity, is less studied. In this work, we investigate the intrinsic anharmonicity of Cellulose Iβ by calculating both harmonic and anharmonic phonon behaviors using molecular dynamics simulation using a reactive molecular forcefield (ReaxFF). To improve the performance of geometry optimization and get better energy conservation, a modified version of ReaxFF [3] is used in our study. The contributions of different phonon modes towards thermal conductivity and the mediating effects of anharmonicity are shown using both diffusive and ballistic models. Comparisons of the results using different models reveal the unconventional way heat is transported by phonons in Cellulose Iβ and help to understand nanoscale thermal transport mechanisms in this highly complex crystal. Moreover, similar to other organic molecular crystals, such as α-RDX [4], we show the unusual role of certain phonon-phonon scattering processes. Unlike ordered crystal, our study will demonstrate the diffusive behaviors of phonon in complex organic crystals. These results indicate complex materials like Cellulose Iβ can diverge strongly from accepted multiscale and multiphysical notions of phonon contributions to the macroscale thermal transport behavior.

References
[1] R.-Y. Dong, Y. Dong, Q. Li and C. Wan, Ballistic-diffusive phonon transport in cellulose nanocrystals by ReaxFF molecular dynamics simulations, International Journal of Heat and Mass Transfer 148, 119155 (2020).
[2]X. Wu, R. J. Moon and A. Martini, Tensile strength of Iβ crystalline cellulose predicted by molecular dynamics simulation, Cellulose 21, 2233-2245 (2014).
[3] D. Furman and D. J. Wales, Transforming the Accuracy and Numerical Stability of ReaxFF Reactive Force Fields, The Journal of Physical Chemistry Letters 10(22), 7215-7223 (2019).
[4] G. Kumar, F. G. VanGessel and P. W. Chung, Bond Strain and Rotation Behaviors of Anharmonic Thermal Carriers in α-RDX, Propellants, Explosives, Pyrotechnics 44, 1-9 (2019).

The development of nanomaterials has revolutionized the materials industry. In order to exploit the full industrial potential of these materials, it is necessary to have a reliable mathematical model to simulate the physical processes that are relevant for specific applications. Such a model is needed, for example, for their characterization and testing and for virtual experimentation, useful for designing new devices.
Thermal transport in nanomaterials is a subject of strong topical interest because of its wide-ranging practical applications. Examples are thermoelectric energy conversion as well as heat management systems. Almost all solid-state electronic devices are sensitive to heat and require efficient heat dissipation mechanisms. The conventional modeling techniques based upon the continuum approximation of solids are not applicable to nanomaterials. This is because phonons play a major role in the thermal transport and their mean free path is comparable to the interatomic spacing and the device dimensions.
Another modeling challenge is that the thermal transport characteristics of solids, specifically their thermal conductivities, are sensitive to the low frequency (GHz - MHz) phonons. These quasiparticles require modeling of a many particle system over extended times. The iterative numerical techniques, based upon the use of conventional molecular dynamics (MD), suffer from poor temporal convergence. They are, therefore, limited to small times and are more suitable for simulating high frequency phonons. For example, the basic MD formulation is limited to a few picoseconds, whereas simulation of GHz -MHz phonons need the simulations to extend to nano-micro seconds.
At NIST, we have developed a causal phonon Green’s function (CGF) method, which is very efficient for modeling the phononic response of nanomaterials at the atomistic length scales and large temporal scales. In the conventional MD, a typical function Z(r,t) in the Hamiltonian is expanded to first order in the space variable ‘r’, as well as the time variable ‘t’. The differentials are approximated by finite differences. In our CGF method, we first make the problem mathematically more complicated by expanding the Z(r,t) to quadratic powers in ‘r’ but not in t. Then we exploit the fact that the CGF for Hamiltonian with quadratic dependence on r can be calculated analytically. This enables us to construct the CGF to obtain an exact solution in ‘t’. It breaks the limitation of the small time steps in temporal integration.
We still need to iterate because of the need to include the anharmonic (cubic and higher) terms in Z(r,t). This introduces some constraints on the iteration steps but they are much less severe than those in the conventional MD. For some idealized cases, we had shown that our CGF technique can model the temporal processes over an unprecedented wide range from femto to microseconds.
One problem was that the calculation of CGF needs diagonalization of an N x N matrix, where N is the number of atoms in the model. This can be done for symmetric lattices but becomes computationally expensive for lattices with defects and disorders. We have now developed an ICGF (Iterative CGF) scheme for calculating the CGF for realistic nanomaterials without invoking any symmetry. We have tested the validity of the ICGF by applying it to a model problem, for which exact results are available for comparison.
The CGF is useful for a variety of modeling problems in nanomaterials. For example, it gives pair correlation functions and is useful for solving problems on scattering of quasi particles. The poles of the CGF give phonon frequencies and the imaginary part of its diagonal elements give the frequency spectrum. In this talk I will briefly describe our CGF and ICGF techniques and discuss their efficacy and validity by giving a numerical example.

Interfaces impede phonon transmission and cause heat accumulation. This heat accumulation degrades the performance of nanoscale transistors and microprocessor chips used in our everyday life. Hence effective thermal management strategies are desirable to control the heat carriers both at an atomic scale and at the device level. Several phenomenological elastic phonon transmission models, such as the acoustic mismatch model, diffuse mismatch model, atomistic Green's function, are proposed to spectrally resolve the phonon conductance across the interface¹. More recently, the conductance is shown to have a significant (~50%) contribution from inelastic phonon transmission assisted by interfacial modes². Moreover, within both elastic and inelastic phonon transmission, phonons can retain or lose their coherency³. Simulation of these elastic and inelastic models at the device level is necessary to design effective thermal management strategies.
To this end, we have developed a variance-reduced Monte Carlo solver for linearized Peierls Boltzmann transport equation in three-dimensions⁴. Using phonon dispersions and lifetime from inelastic scattering experiments, ab-initio simulations, or empirical models, we simulate domains ranging from a few nanometers to hundreds of microns. Simulations are performed on both periodic and non-periodic geometries. Since phonons scatter differently from physical domain boundaries, interface, impurity, and other phonons, we simulate boundary scattering, three-phonon processes, two-phonon processes, and impurity scattering independently. A frequency-dependent transmission probability and interface roughness (specularity) capture the size-effects and device-level conditions. Grain-boundary scattering models and interface transmission models are incorporated separately. Our interface modeling strategy is not limited to elastic energy transport and allows us to include inelastic transport across the interface by energy exchange between high- and low-frequency modes. We compare our interface conductance results in quasi-ballistic and diffusive regimes measured using near-IR pump EUV/X-ray probe experiments^5,6,7. The spectrally-resolved contribution to temperature and flux at the interface provides critical insights into the heat accumulation. We are presently simulating interfacial conductance in layered materials to compare with available near-IR pump and X-ray probe experiments⁸. We anticipate our code will further find applications in evaluating the nano-structuring and nano-patterning efficiency to thermal conduction properties of mesoscale devices.

References:
1. Monachon, C., Weber, L., & Dames, C. (2016), Annual Review of Materials Research, 46, 433–463.
2. Feng, T., Zhong, Y., Shi, J., & Ruan, X. (2019). Physical Review B, 99(4), 045301.
3. Ravichandran, N. K., & Minnich, A. J. (2014). Physical Review B, 89(20), 205432.
4. MCBTE-v0.1. Nov. 2020. https://github.com/abhipath90/MCBTE
5. Siemens, M. E., Li, Q., Yang, R., Nelson, K. A., Anderson, E. H., Murnane, M. M., & Kapteyn, H. C. (2010). Nature Materials, 9(1), 26–30.
6. Hoogeboom-Pot, K. M., Hernandez-Charpak, J. N., Gu, X., Frazer, T. D., Anderson, E. H., Chao, W., Falcone, R. W., Yang, R., Murnane, M. M., Kapteyn, H. C., & Nardi, D. (2015). Proceedings of the National Academy of Sciences of the United States of America, 112(16), 4846–4851.
7. Frazer, T. D., D., Knobloch, J. L., Hoogeboom-Pot, K. M., Nardi, D., Chao, W., Falcone, R. W., Murnane, M. M., Kapteyn, H. C., & Hernandez-Charpak, J. N. (2019). Physical Review Applied, 11(2), 1.
8. Nyby, C., Sood, A., Zalden, P., Gabourie, A.J., Muscher, P., Rhodes, D., Mannebach, E., Corbett, J., Mehta, A., Pop, E. and Heinz, T.F., (2020). Advanced Functional Materials, 30(34), p.2002282.

Michele Diego, Romain Rouxel, Fabio Medeghini, Aurélien Crut, Francesco Rossella, Paolo Maioli, Fabrice Vallée, Francesco Banfi and Natalia Del Fatti

Single-particle optical spectroscopy methods have enabled quantitative investigations of the optical [1], electronic [2], and vibrational [3,4] responses of nano-objects in the recent years. In this work, we exploit single-particle pump-probe optical spectroscopy to investigate the cooling dynamics of individual gold nanodisks supported on a sapphire substrate. Although this topic has already been theoretically investigated in the past [5,6], quantitative measurements have so far proved very challenging. We here show time-resolved signals encompassing the temporal evolution of the single nanodisk temperature following its impulsive excitation [7]. By comparing the experimental signals (measured on individual nanodisks with different sizes) to the results of analytical [4] and finite-element simulations, we are able to retrieve the thermal boundary resistance at the single gold nanodisk-sapphire interface, which limits the nanodisk cooling rate. The present findings pave the way for quantitatively characterizing the temperature-dependent optical response of a single nano-object by investigating the probe-wavelength dependence of time-resolved signals, and for correlating the nano-object thermal properties to its specific morphology and composition. Moreover, it bears great relevance for post-processing inspection of nanopatterned metal nano-objects [8].


[1] A. Crut, P. Maioli, N. Del Fatti, and F. Vallee; Chem. Soc. Rev. 43, 3921 (2014)
[2] H. Baida, D. Mongin, D. Christofilos, G. Bachelier, A. Crut, P. Maioli, N. Del Fatti, and F. Vallee; Phys. Rev. Lett. 107, 057402 (2011)
[3] P. V. Ruijgrok, P. Zijlstra, A. L. Tchebotareva, and M. Orrit; Nano Lett.. 12, 1063 (2012)
[4] A. Crut, P. Maioli, N. Del Fatti, and F. Vallee; Phys. Rep. 549, 1 (2015)
[5] F. Banfi, F. Pressacco, B. Revaz, C. Giannetti, D. Nardi, G. Ferrini, and F. Parmigiani; Phys. Rev. B 81, 155426 (2010)
[6] M. Gandolfi, A. Crut, F. Medeghini, T. Stoll, P. Maioli, F. Vallee, F. Banfi, and N. Del Fatti; J. Phys. Chem. C 122, 8655 (2018)
[7] R. Rouxel, M. Diego, F. Medeghini, P. Maioli, F. Rossella, F. Vallee, F. Banfi, A. Crut, and N. Del Fatti; J. Phys. Chem. C 124, 15625 (2020)
[8] S. Peli, A. Ronchi, G. Bianchetti, F. Rossella, C. Giannetti, M. Chiari, P. Pingue, F. Banfi, and G. Ferrini; Scientific Reports 10, 16230 (2020)

Materials near structural phase transitions usually have large phonon anharmonicity and low lattice thermal conductivity. The proximity to structural phase transitions in IV-VI materials is one of the main reasons for their low thermal conductivity and high thermoelectric figure of merit [1,2]. However, experiments indicate that the lattice thermal conductivity of germanium telluride (GeTe) actually increases near the ferroelectric phase transition around 600-700 K. In this work, we elucidate this unexpected effect using first principles calculations [3].

We first compute the lattice thermal conductivity of GeTe for a range of temperatures using the temperature dependent effective potential method [4] and the standard Boltzmann transport equation. We find that, although anharmonicity increases near the phase transition in the low symmetry phase (below the Curie temperature), the phonon group velocities increase as well, leading to a rise in the lattice thermal conductivity. Furthermore, anharmonicity decreases in the high symmetry cubic phase, further enhancing the phonon lifetimes and thermal conductivity.

Strong anharmonicity near the phase transition induces non-Lorentzian shapes of the phonon power spectra, which exhibit large softening of the peaks for soft phonon modes near the phase transition. To account for these effects, we implement a novel method of calculating lattice thermal conductivity based on the Green-Kubo approach [3]. Using this approach, we find that the Boltzmann transport equation underestimates the lattice thermal conductivity of GeTe near the phase transition. Our findings elucidate the influence of structural phase transitions on lattice thermal conductivity and provide guidance for design of better thermoelectric materials.

This work is done in collaboration with Djordje Dangic, Olle Helman and Stephen Fahy. The work is supported by Science Foundation Ireland PI Award 15/IA/3160.

[1] T. Shiga et al., Phys. Rev. B 85, 155203 (2011).

[2] R. M. Murphy, E. D. Murray, S. Fahy, and I. Savic, Phys. Rev. B 93, 104304 (2016).

[3] D. Dangic, O. Hellman, S. Fahy, and I. Savic, in preparation.

[4] O. Hellman and I. A. Abrikosov, Phys. Rev. B 88, 144301 (2013).

In this work, we determine how a multilayer structure made of different types of nanoscale phase change materials results in a highly non-linear and asymmetric heat flow depending on thermal bias directionality. This will set the basis of solid-state thermal diodes that can be scaled up for integration in energy management, conversion or storage applications. Thermal diodes (TD) or rectifiers are capable to manage heat in a similar manner as how electronic diodes control electricity, propagating heat preferably in one direction.^[1] The rectification ratio is defined as, RR = (Q_fwd-Q_rev)/Q_rev, where Q_fwd and Q_rev correspond to the heat flux in the forward and reverse direction, respectively. Unfortunately, current solid-state thermal diodes typically present moderate RRs, complex fabrication designs, and/or lack of operating temperature tuneability.^[1]

Phase change materials (PCM) have become popular for the development of TDs due to their thermal conductivity (k) change during the phase transition at a critical temperature (T_crit), e.g. VO₂ (k_low = 1.5 W/(m×K) to k_high = 3.5 W/(m×K) at T_crit~340 K).^[1,2] These materials are often combined with phase invariant materials (PIM) to develop TDs.^[1] As an example, in a PIM/PCM structure based on VO₂, when the heat source is at temperatures T> T_crit and close to the PCM side, both PIM and PCM conduct the heat well (Q_fwd). The situation is reversed when the heat source is applied to the PIM (Q_rev). This typically leads to RR larger than those found for individual PCMs.^[1]

Here, we used finite element modeling (COMSOL®) to develop a versatile TD based on a novel multilayer PCM/PIM structure. Our design includes an alternating combination of two PCM layers with two PIM layers. The total length of the multilayer structure was set to 1 μm, while the thickness of the individual layers was varied at the nanoscale. We selected carefully different types of PCM and PIM materials, whose experimental thermal properties were extracted from literature, to find the optimum PCM/PIM configuration that led to the highest RR. Then, we applied a temperature difference across the two ends of this structure. RRs larger than 100% were observed when we used Si and SiO₂ as PIMs and Ag₂Te and Ag₂S_0.6Se_0.4 as PCMs with transition temperatures at 420 K and 360 K, respectively.^[3] Compared to the state of the art^[1] of PCM/PIM diodes, this configuration represents one of the highest RRs at room temperature as well as offers new possibilities for advanced thermal control. Additionally, from the point of view of applicability, this thermal diode offers numerous advantages over previously reported TDs, including simple design, scalability and operating temperature tunability.

To determine the potential of this thermal device for solid state refrigeration, we analyzed the effect of integrating this TD into a magnetocaloric (MC) device. For that purpose, we used a one-dimensional MC device consisting of gadolinium as MC material with two TDs at their ends, sandwiched between the heat sink and the heat source. The presence of TDs avoids the heat to flow back to the heat source improving the efficiency of the MC device.^[4] We selected a PIM/PCM structure that worked optimally for the MC system operating temperature and we considered quasi-steady-state operation in an alternating magnetic field with 1 T change. We observed that the integration of this TD enables higher operating frequencies compared to the conventional active magnetic regeneration process, increasing the cooling power density. Beyond MC refrigeration, this versatile and unique TD can be used in other solid state refrigeration, heat pump and energy harvesting technologies to improve their performances or efficiencies.^[1,4]

References
[1] Swoboda et al., Adv. Electron. Mater., 2020 (Accepted Review) DOI: 10.1002/aelm.202000625
[2] Oh et al., Appl. Phys. Lett., 96, 2010
[3] Hirata et al., J. Electron. Mater., 49, 2020
[4] Kitanovski, Adv. Energy Mater., 10, 2020

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.

Colloidal nanocrystal superlattices that are made of uniform-sized quantum dots (QD) connected by organic ligands have shown promising applications for photon detectors, solar cells, and thermoelectrics, among many others. Due to the secondary periodicity, the nanocrystal superlattice is also an ideal platform for phonon engineering, and the existence of coherent superlattice phonon mode is confirmed through the inelastic neutron scattering measurement recently. [1] However, it is also found that such coherent phonons are sensitive to the rotational motions of the building block [2], [3], and the surface morphology. [4] In this work, based on molecular dynamics simulation techniques, we explored the preconditions that are required to sustain the coherent phonons in the lead sulfide nanocrystal superlattice and the possible margins to engineering the thermal conductivity by manipulating such coherent phonons. We found that the mean free paths of the low-frequency coherent modes are tens of nanometers. However, those coherent modes, of which the frequency is less than 0.2 THz, carrier negligible amount of heat due to their small weights in the overall vibrational spectrum. We have shown that phonon in the frequency range of 0–5 THz dominates the thermal transport. In this frequency range, the interatomic vibrations of QDs have the maximum overlapping with the translation and bending modes in the ligands. We also identified the ligand coverage ratio as the key parameter that will affect the morphology of the nanocrystal superlattice and thus dictate the thermal transport.

[1] N. Yazdani et al., “Nanocrystal superlattices as phonon-engineered solids and acoustic metamaterials,” Nat Commun, vol. 10, no. 1, pp. 1–6, Sep. 2019, doi: 10.1038/s41467-019-12305-3.
[2] W.-L. Ong et al., “Orientational order controls crystalline and amorphous thermal transport in superatomic crystals,” Nature Materials, vol. 16, no. 1, p. 83, Jan. 2017, doi: 10.1038/nmat4739.
[3] S. Kumar, C. Shao, S. Lu, and A. J. H. McGaughey, “Contributions of different degrees of freedom to thermal transport in the C60 molecular crystal,” Phys. Rev. B, vol. 97, no. 10, p. 104303, Mar. 2018, doi: 10.1103/PhysRevB.97.104303.
[4] K. J. Schnitzenbaumer and G. Dukovic, “Comparison of Phonon Damping Behavior in Quantum Dots Capped with Organic and Inorganic Ligands,” Nano Lett., vol. 18, no. 6, pp. 3667–3674, Jun. 2018, doi: 10.1021/acs.nanolett.8b00800.

Heat conduction due to phonons is usually dominated by diffuse scattering processes from crystal defects and anharmonicity. As such, doping and grain boundary engineering have been the primary approach for controlling the heat conduction in solids. While the contributions from coherently scattered low frequency phonons are not significant, they are challenging to control since they are not sensitive to crystal defects and therefore, remain as a major challenge for mastering the conduction of heat in solid. As such, understanding the propagation of coherent phonons is of fundamental interest in thermal management technologies. Superlattices in this regard provide an excellent platform since its heat conduction is attributed to coherent phonons. Experimental studies are typically performed on artificial superlattices, but they exhibit dislocations at interfaces, which increases anharmonicity and affect coherent scattering processes. For this reason, there has been some discrepancies between experimental and theoretical research efforts. In this study, we investigated the heat conduction in defect-free InGaO₃(ZnO)_m superlattice films to experimentally demonstrate several important concepts in phonon engineering, including the discrepancy between coherence length and mean free path, scattering mechanism changes according to the coherence length, specular phonon propagation along the in-plane direction, and the suppression of Umklapp processes.

Environmental scanning electron microscopy (ESEM) is an important micro-to-nanoscale inspection tool to image diverse wet or insulating samples, such as liquid, polymer, and biomaterials in a gaseous environment. However, the presence of gas induces intensive electron-gas collisions, which leads to significant degradation of imaging quality. For this reason, the operating pressure of ESEM is typically below 1000 Pa. In this work, we enhanced the performance of ESEM by referring to the phase and scattering force of electron. Much higher contrast of ESEM images was obtained from the phase and scattering force reconstruction without any hardware changes. The limit of operating pressure was also extended up to 2500 Pa. More importantly, a close connection between the sample morphology and scattering force distribution was observed, which promises to advance the quantitative analysis using ESEM. To demonstrate the effectiveness of this enhanced ESEM technique, we studied the nucleation site distribution during droplet condensation, which cannot be precisely probed in previous studies due to the micro/nanoscopic nature of nucleation. We show that the population of nucleation sites is governed by the Poisson distribution while the nearest neighbor distance follows the Rayleigh distribution, instead of the commonly used Poisson distribution. Our work provides a simple approach to enhance the performance of ESEM and can be widely used to study a variety of material dynamic processes in gaseous environments.

Fast evaporation of liquids is a process of great importance to many industrial applications such as evaporative cooling of gas turbines, efficient combustion in engines, and superfine inkjet printing. Accurate prediction of the maximum evaporation flux from a condensed phase is one of the basic problems in the study of fast evaporation processes. The maximum evaporation flux occurs during the evaporation into a vacuum. In spite of the extensive work on the process of evaporation into a vacuum in the past decades, most of modeling studies only focused on the evaporation of monatomic substances. Whether these modeling results on monatomic fluids can be applied to predict the maximum evaporation flux of more complex fluids such as water and polymers is questionable. To address this question, one must have a fundamental and quantitative understanding of how the internal motions including rotations and vibrations of polyatomic fluid molecules affect the evaporation process and the maximum evaporation flux.

In this work, we use the kinetic theory of gases (KTG) to develop a theoretical model to understand the role of internal motions of molecules on the maximum evaporation flux from a planar liquid surface. The kinetic theory is applied to study the evaporation of molecular fluids into a vacuum and predict the dimensionless maximum evaporation flux (J_R,max, i.e. the ratio of the maximum evaporation flux to the molar flux emitted from a liquid surface). The key assumptions regarding the velocity distribution function (VDF) of polyatomic molecules in the highly non-equilibrium vapor near the evaporating surface are validated by the VDF obtained directly from molecular dynamics (MD) simulations. Our KTG-based analysis show that J_R,max is affected by the specific heat (c_V,int) associated with internal degrees of freedom of fluid molecules. When the maximum evaporation flux is reached, the isotropic evaporating vapor far from the liquid surface moves at its speed of sound regardless of whether it is a monatomic vapor or polyatomic vapor. To fundamentally understand the evaporation of a molecular fluid into a vacuum, we solve the Boltzmann transport equation (BTE) to obtain the temperature, density, and flow speed distributions in the highly non-equilibrium evaporating vapor flow. Our BTE solutions indicate that there are several universal features of the evaporating vapor when the maximum evaporation flux occurs. In particular, we find that the evaporating vapor flow speed reaches the maximum value of sqrt(1.5) times the most probable thermal speed in the vapor flow direction at the vacuum boundary, and this maximum value is independent of fluid properties. All theoretical predictions in this work are verified by the MD simulation results of the evaporation of the model liquid Ar and the model liquid n-dodecane into a vacuum, and existing experimental data.

Interfaces often govern the thermal performance of nanoscale devices and nanostructured materials. As a result, accurate knowledge of thermal interface conductance is necessary to model the temperature response of nanoscale devices or nanostructured materials to external heating. We report the thermal boundary conductance between metals and insulators that are commonly used in spin-caloritronic experiments. We use time-domain thermoreflectance to measure the interface conductance between metals such as Au, Pt, Ta, Cu, and Al with garnet and oxide substrates, e.g. NiO, YIG, EuIG, Cr₂O₃, and sapphire. We find that, at room temperature, the interface conductance in these types of material systems range from 50 and 300 MW m^-2 K^-1. We also measure the interface conductance between Pt/YIG at temperatures between 80 and 350 K.

Correct estimation of the thermal resistance of the PCM memory structure is crucial due to the optimization of its parameters. The thermal resistance of PCM memory consists of thermal boundary interface resistances and GST bulk resistance. After entering the reset state, the crystallization of bulk GST begins and as a result a composite of GST crystallites distributed in an amorphous GST matrix is formed and thermal conductivity of such composite depends on the stage of the crystallization. Influence of the time-spatial distribution of the function defining the volume of the transformed fraction of the GST on the thermal resistance of the GST-based PCM structure is presented.

Control of heat conduction through the manipulation of phonons as coherent waves have been attracted great interest for the advanced thermal management. Although smooth interfaces were reported to be obtained in artificial superlattices, preparation of coherent interfaces for terahertz phonons having nanoscale periodicity with atomic-scale perfection are still challenging since MOCVD and MBE were nonequilibrium growth process. Therefore, we focus on natural superlattice titanium oxides, in which periodic structures are obtained as thermodynamically stable phases. A homologous series of the titanium oxides is so-called crystallographic shear (CS) structure, in which planar faults are periodically introduced in the mother rutile structure with their spacing depending on the oxygen. Recently we have revealed that the titanium oxide with CS structure possesses the periodicity and the interface with high perfection and the coherent interface for almost all phonons are formed. In the present study, we demonstrated the control of the period of the CS structure in titanium oxides by the addition of chromium oxides to form oxygen deficiency.
The polycrystalline crystals of the titanium oxides with the addition was fabricated by the following process: firstly mixture of titanium and chromium oxides were sintered and obtained ceramics were melted by arc melting in an argon atmosphere and subsequently annealed in air for homogenization and oxidization of the crystals. The structure of the crystals were investigated by x-ray diffraction (XRD) as well as scanning transmission electron microscopy (STEM) with atomic-scale observation.
Judging from the XRD measurement, it seemed that the obtained crystals were of single phases and the structure was systematically changing according to the chromium content. Atomic-scale STEM observation revealed that the interval of the interfaces in the crystals were decreasing with increasing chromium content and we obtained the titanium oxides with the different interval of the interfaces ranging from 1.2 to 2.0 nm. From the STEM observation, not only the interval but also the atomic structure of the interfaces were changing depending on the chromium content. We have successfully control the periodic structure in a homologous series of titanium oxide bulk crystals. The current results indicates that the tuning of phonon band gap formation at higher frequency is possible in titanium oxide natural superlattices.

While coherent effects in solid heat conduction have recently been exemplified in superlattices and disordered structures [1-3], those evidences have remained essentially qualitative because the understanding of the intrinsic coherence of thermal phonons remains unclear. We propose a formalism supported by theoretical arguments and direct atomic simulations, which takes into account both the conventional phonon gas model and the internal wave nature of thermal phonons. By naturally introducing wavepackets in the heat flux from fundamental concepts, we derive an original thermal conductivity formula including coherence times and lifetimes. We then estimate the deviation between the predictions of the conventional model and ours by using atomic simulations to highlight significant effects in graphene. This contribution establishes a new frame for understanding and quantifying the intrinsic coherence of phononic heat carriers, which has a general impact on the estimation of the thermal properties in nanostructures as well as in bulk materials.
References
[1] Luckyanova, M. N., et al., Science (2012) 338 (6109), 936
[2] Isaeva, L., et al., Nat. Commun. (2019) 10 (1), 3853
[3] Simoncelli, M., et al., Nat. Phys. (2019) 15 (8), 809

Surfactants are amphiphilic molecules that can be used to control the evaporation rate on a liquid-vapor interface. Previous experimental work has shown that molecular monolayers can reduce the evaporation rate on a flat liquid-vapor interface, while surfactant-modified sessile droplets show an increased evaporation rate due to changes in their wetting properties. A surfactant-induced increase in evaporation rate on a flat interface has yet to be demonstrated.

We applied molecular dynamics (MD) simulations to study the evaporation of a Lennard-Jones liquid with a variety of surfactants. A non-equilibrium MD (NEMD) simulation with a temperature difference was designed to directly extract the evaporation mass flux. Surfactants are added to the liquid-vapor interface. We found that an anchor-shaped surfactant with a flat hydrophobic top and hydrophilic chains enhances the evaporation rate by a factor of around 1.4 compared to the pure liquid.

To determine the origin of the evaporation rate enhancement, additional equilibrium MD simulations of the pure liquid and liquid-surfactant systems were performed to calculate thermal conductivities, diffusion coefficients, latent heats, vapor pressures, and mass accommodation coefficients. The results cumulatively point towards the evaporation enhancement being a result of liquid structuring around the surfactant, which lowers the thermal resistance to evaporation. Different surfactant sizes and structures are currently being investigated to further elucidate the origin and limits of the enhanced evaporation rate.

Understanding the heat transport in intercalated layered materials has attracted significant interest in Li-ion battery (LIB)-based electrochemical energy storage devices because they act as electrode materials. Heat dissipation affects a battery lifetime. Moreover, thermal runaway is one of the main faults of modern LIBs [1]. Therefore, tuning thermal conductivity in these devices becomes a key challenge. Chemical intercalation has proved to be an effective approach to control thermal transport in this application [2,3].
In our recent work, we have demonstrated thermal conductivity variation across amorphous Si thin films during Li intercalation [4]. In the present work, we further extend this research on titanium oxide and lithium titanate electrode materials, which were proved to be good alternatives for high power applications due to high cycling stability and long lifetime [5]. To study the effects of Li intercalation on their subsurface thermal conductivity, a picosecond time-domain thermoreflectance (TDTR) [4,6] was employed. Corresponding thin films were deposited by RF magnetron sputtering system and galvanostatic cycling was used between 0.5 and 3 V for Li insertion. Raman microscopy measurements were exploited to study chemical evolution during Li insertion into lithium titanate samples.
Thermal conductivity measurements revealed an average 50% increase in as-deposited TiO₂ thin films upon lithiation. Raman spectra indicated the formation of Li_xTiO₂ phase formation. Also, samples with different annealing temperatures have been prepared to study the effect of crystallinity on thermal conduction during Li intercalation.

Reference

[1] Q. Wang, P. Ping, X. Zhao, G. Chu, J. Sun and C. Chen, Journal of Power Sources 208 (2012): 210-224.
[2] J. Cho, M. D. Losego, H. G. Zhang, H. Kim, J. Zuo, I. Petrov, D. Cahill and P. V. Braun, Nature Communications 5 (2014): 1–6.
[3] Y. Xiong, N. C. Lai, Y. C. Lu and D. Xu, International Journal of Heat and Mass Transfer 159 (2020): 120077.
[4] A. Abdullaev, A. Mukanova, T. Yakupov, A. Mentbayeva, Z. Bakenov and Z. N. Utegulov, Materials Today: Proceedings 25 (2020): 88-92.
[5] M. V. Reddy, G. V. Subba Rao and B. V. R Chowdari, Chemical reviews, 113(7) (2013): 5364-5457.
[6] D.G. Cahill, Review of Scientific Instruments 75 (2004): 5119–5122.

Displaying unique thermal, mechanical, electrical and magnetic properties [1], nitride ceramics have been widely used in nuclear engineering [2] and microelectronics applications [3]. In particular, the effect of swift heavy ion (SHI) irradiation on thermal transport of radiolysis-resistant nitride ceramics is a matter of interest [4]. However, there has been no prior study of the impact of SHI irradiation on the thermal conductivity of these materials on the subsurface scale.

AlN and Si₃N₄ polycrystalline ceramics were irradiated at the U-400 FLNR JINR cyclotron facility in Dubna, Russia by 167 MeV Xe and 700 MeV Bi ions ions at fluences between 1×10¹¹ and 1×10¹⁴ cm^-2. The projected ion ranges were 13.6 (Si₃N₄) and 13.2 (AlN) µm as calculated using SRIM 2013 in full cascade mode. Structural analysis with corresponding molecular dynamics (MD) simulations revealed formation of ion induced tracks in Si₃N₄, while no tracks were observed in AlN.
We report the results of cross-plane heat transport measurements of SHI irradiated aluminum nitride and silicon nitride ceramics as a function of ion dose by femtosecond laser pulse-based time-domain thermo-reflectance (TDTR) and continuous laser wave-based modulated thermoreflectance (MTR) techniques on nano- and micro- meter subsurface scales, respectively. TDTR technique revealed thermally probed region extending down to hundreds of nanometers, where the electronic stopping / ionization processes dominate over nuclear stopping, while MTR approach has revealed thermal wave penetration depths down to microns. Thermal transport in these sub-surface domains irradiated by Xe and Bi ions for the given fluences has degraded by the factor of ~ 3 compared to non-irradiated as-received sample for both nitrides.
Using theoretical modeling based on Debye approximation and Klemens analytical thermal transport model [5], we analyzed defects concentrations extracted from experimentally measured thermal property data fitted to thermal diffusion model. In addition MD simulations were employed to model phonon scattering at ion tracks in Si₃N₄ ceramics. Given TDTR probed sub-surface region overlapped with ionization depth determined by SRIM calculation, whereas MTR probed region spanned nuclear displacive domain.


References

[1] G. A. Slack, R. A. Tanzilli, R. O. Pohl, and J. W. Vandersande, “The intrinsic thermal conductivity of AIN,” J. Phys. Chem. Solids, vol. 48, no. 7, pp. 641–647, 1987, doi: 10.1016/0022-3697(87)90153-3.
[2] A. Janse Van Vuuren, V. Skuratov, A. Ibrayeva, and M. Zdorovets, “Microstructural Effects of Al Doping on Si₃N₄ Irradiated with Swift Heavy Ions,” vol. 136, 2019, doi: 10.12693/APhysPolA.136.241.
[3] A. J. Griffin, F. R. Brotzen, and P. J. Loos, “The effective transverse thermal conductivity of amorphous Si 3N4 thin films,” J. Appl. Phys., vol. 76, no. 7, pp. 4007–4011, Oct. 1994, doi: 10.1063/1.357347.
[4] S. J. Zinkle, V. A. Skuratov, and D. T. Hoelzer, “On the conflicting roles of ionizing radiation in ceramics,” Nucl. Instruments Methods Phys. Res. Sect. B Beam Interact. with Mater. Atoms, vol. 191, no. 1–4, pp. 758–766, May 2002, doi: 10.1016/S0168-583X(02)00648-1.
[5] A. Koshkinbayeva, A. Abdullaev, Z. Nurekeyev, V. A. Skuratov, Y. Wang, M. Khafizov, Z. N. Utegulov “Thermal transport and optical spectroscopy of swift heavy ion irradiated LiF Crystals”, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 475, 14-19 (2020).

We report on the heat conduction properties of smooth layers prepared from dried colloidal pure crystalline silicon (Si) nanoparticle (NP), which were probed by weak continuum wave (cw) laser light-induced heating with simultaneous Raman scattering detection under atmospheric conditions. The same approach of heating measurement was applied for the bulk silicon at high laser powers [1, 2]. Low thermally conductive Si NP films are envisioned as heat insulators for thermal management of optoelectronic devices. Laser heating causes softening of the first-order Raman Si-Si transverse optical phonon mode and a decrease in the corresponding phonon lifetimes [3]. The temperatures of laser-heated Si NP films were assessed from measured ratio of Stokes-to-Anti-Stokes integrated peak intensities of the Raman probed phonon mode and were compared with the external heater-heated temperature and temperature evaluation for cw-laser heated bulk crystalline Si [4].
It was revealed that the evaporation rate of the colloidal Si NP droplets and heating of deposited thin layers of variable thickness were heavily influenced by the thermal conductivities of the substrates, on which the films were deposited. It is suggested that the heat conduction of nanostructured layers is governed by the ballistic transfer of phonons across the boundaries of nanoparticles and varied with an increase in the layer thickness [5]. Our experimental results are in good agreement with theoretical molecular dynamics calculations [6] and Raman measurements in meso-porous nanocrystalline Si layer [7]. Micro-Raman-based thermal conduction measurements in our Si NP films are discussed in the light of variation of deposited colloidal NP film thickness, NP size, heating laser beam diameter and the substrate type.





References:
1. M. Balkanski, R.F. Wallis, and E. Haro, “Anharmonic effects in light scattering due to optical phonons in silicon”, Phys. Rev. B, 28 (4), 1928-1934 (1983).
2. M. Yamada, K. Nambu, Y. Itoh and K. Yamamoto, “Raman microprobe study on temperature distribution during CW laser heating of silicon on sapphire”, J. Appl. Phys., 59(4), 1350-1354(1986).
3. P. Mishra and K. P. Jain, “Temperature-dependent Raman scattering studies in nanocrystalline silicon and finite-size effects”, Phys. Rev. B, 62 (22), 790-795 (2000)
4. J. Ready “Effects of High-Power Laser Radiation”, Academic Press (London), 1971.
5. Z. Wang, J. E. Alaniz, W. Jang, J. E. Garay and C. Dames. “Thermal Conductivity of Nanocrystalline Silicon: Importance of Grain Size and Frequency-Dependent Mean Free Paths”, Nano Lett., 11, 2206–2213 (2000).
6. S. Ju and X. Lianga, “Thermal conductivity of nanocrystalline silicon by direct molecular dynamics simulation”, J. Appl. Phys., 112, 064305 (2012).
7. V. Lysenko, S. Perichon, B. Remaki, D. Barbier, B. Champagnon, “Thermal conductivity of thick meso-porous silicon layers by micro-Raman scattering”, J. Appl. Phys, 86 (12), 6841-6846 (1999).

Beyond the Fourier diffusion theory on heat conduction, the classical size effects---the Casimir regime---caused by phonon boundary scattering is well known and extensively studied. However, over the last three decades, new regimes beyond the Fourier and the Casimir pictures of heat conduction have been demonstrated. In this talk, I will discuss different phonon heat conduction regimes, including the Knudsen regime, the hydrodynamic regime, the quantization regime, the coherence and localization regimes, and the divergence regime. The Knudsen regime expands Casimir’s picture to many other quasi-ballistic transport geometries, and is being exploited to develop phonon mean free path spectroscopy techniques. Phonon hydrodynamic transport happens when the normal scattering dominates over the resistive scattering, which is a condition difficult to satisfy and only observed at a narrow temperature range less than 20K. However, recent experiments have observed second sound---a consequence of phonon hydrodynamic transport---at much higher temperatures, while simulations point to possibility of observing hydrodynamic heat conduction even at room temperature. Quantized phonon transport was observed at very low temperatures. Signatures of coherent heat conduction, including localization, will be discussed, together with experimental evidences. Divergent thermal conductivity, implying thermal superconductors, is predicted to be possible in low-dimensional materials, although no experiments have provided conclusive evidence. These different phonon heat conduction regimes will be summarized in a regime map, demonstrating the rich phonon transport physics rivaling that of electrons.

Manipulation of thermal fluxes is essential in modern science and technology. Modern thermal phononics tries to achieve coherent control of heat conduction based on the wave interference of phonons that occurs in periodic nanostructures, called phononic crystals. However, the past decade of the research demonstrated that such a coherent control of heat is truly efficient either at sub-kelvin temperatures or in nanostructures with features at atomic scale.
In this work, we propose an alternative concept of heat manipulation based on the particle picture of phonons and their ballistic transport. We call this concept ray phononics as a thermal analogy of ray optics. Our simulations demonstrate the formation of directional thermal fluxes in phononic crystals due to the stochastic phonon motion in phononic crystals. We illustrate possible application of such directional fluxes for emitting directional heat rays, filtering the phonon spectrum, and even protecting a specific region from a thermal gradient.
Finally, we show that the proposed concept is not bound to only low temperatures, and ray-phononic nanostructures can control heat fluxes even at room temperature using modern materials like boron arsenide. Moreover, ray phononics is free from limitations of the traditional thermal phononics and enables creating and guiding thermal fluxes in realistic nanostructures regardless of their surface roughness. This makes us believe that advances in material synthesis may soon enable creation of devices capable of thermal guiding, emitting, filtering, and shielding using the proposed ray-phononics principles at room temperature.

Reference:
Anufriev and Nomura, "Ray phononics: thermal guides, emitters, filters, and shields powered by ballistic phonon transport", Materials Today Physics 15, 100272 (2020)

Applying a first-principles theory that includes both the lowest-order three-phonon and the higher-order four-phonon processes to zinc blende semiconductors, we find that the lowest-order theory of phonon-phonon interactions drastically overestimates the measured thermal conductivities, k, for many of these materials, while inclusion of four-phonon scattering gives significantly improved agreement with measurements [1]. We identify new selection rules on the phase space of three-phonon processes that help explain many of these failures in terms of anomalously weak anharmonic phonon decay rates predicted by the lowest-order theory competing with four-phonon processes. We also show that zinc blende compounds containing boron (B), carbon (C), or nitrogen (N) atoms have exceptionally weak four-phonon scattering, much weaker than in compounds that do not contain B, C, or N atoms. This new understanding helps explain the ultrahigh k in several technologically important materials like cubic boron arsenide, boron phosphide, and silicon carbide. At the same time, it not only makes the possibility of achieving high k in materials without B, C, or N atoms unlikely, but it also suggests that it may be necessary to include four-phonon processes in many future studies. Finally, we show that the hidden influence of selection rules on three-phonon scattering can be exposed through anomalous signatures in the pressure (P) and temperature (T) dependence of the k of certain compounds [2]. In particular, boron phosphide reveals such underlying behavior through an exceptionally sharp initial rise in k with increasing P, which may be the steepest of any material, and also a peak and decrease in k at high P. These features are in stark contrast to the measured behavior for many solids, and occur at experimentally accessible conditions.

This work was supported by the Office of Naval Research under MURI Grant No. N00014-16-1-2436. N.K.R. acknowledges a start-up fund from the Indian Institute of Science, Bangalore, that supported a part of this work.

[1] Navaneetha K. Ravichandran and David Broido, Phys. Rev. X 10, 021063 (2020).
[2] Navaneetha K. Ravichandran and David Broido, https://arxiv.org/abs/2009.01464

For decades, the three-phonon scattering process had been considered to govern thermal transport in solids, while the role of higher-order four-phonon scattering had been persistently unclear and so ignored. However, recent quantitative calculations of three-phonon scattering had often overestimated thermal conductivity as compared to experimental data. We developed a rigorous perturbation method to predict four-phonon scattering rates and investigated its significant impact on thermal conductivity and infrared thermal radiative properties. We show that four-phonon scattering is generally important in solids. For silicon and diamond, the predicted thermal conductivity is reduced by 30% at 1000 K after including four-phonon scattering, bringing predictions in excellent agreement with measurements in the entire temperature range. For the projected ultrahigh-thermal conductivity material, BAs, a competitor of diamond as a heat sink material, four-phonon scattering is found to be strikingly strong as three-phonon processes have an extremely limited phase space for scattering. The four-phonon scattering reduces the predicted thermal conductivity from 2200 to 1400 W/m K at room temperature. The reduction at 1000 K is 60%. Our prediction on BAs has been confirmed by three independent experimental breakthroughs. Four-phonon scattering also significantly reduces the thermal conductivity of ZA phonons in graphene due to the fact the reflection symmetry does not restrict four-phonon scattering. Moreover, we find that four-phonon scattering universally plays a significant or even dominant role on the zone-center optical phonon linewidth over three phonon scattering, for a wide range of materials including diamond, Si, Ge, GaAs, boron arsenide (BAs), cubic silicon carbide (3C−SiC), and α-quartz. This is enabled by the large four-phonon scattering phase space of zone-center optical phonons. Raman measurements on BAs were conducted, and their linewidth verifies our predictions. The predicted infrared optical properties through the Lorentz oscillator model, after including four-phonon scattering, show much better agreement with experimental measurements than those three-phonon-based predictions.

How and how much we can control transport of thermal phonons by nanostructures utilizing their wave nature is a central issue in science and engineering of nanoscale heat conduction. Phonon-particle scattering can impede heat conduction but, for that, large disorder needs to be introduced, which makes it inherently difficult to balance the impact with those on other quasi particles like electron and magnons. Furthermore, in the best scenario, the wave interference should be superior to particle scattering in impeding phonon transport as the former allows localization. The general difficulty in realizing interference by artificial nanostructure particularly for phonon transport is that the coherent lengths of phonons are typically quite short at room temperature, but thanks to the efforts of the researchers in the field, the control of phonon coherent heat conduction can be realized in heterostructures with nanometers lengths and clear interfaces like the semiconductor superlattices. Over the last few years we Thermal Energy Engineering Lab (TEEL) at the University of Tokyo, together with the collaborators, have been pushing the limit of controllability of coherent phonon heat conduction by exploring the effect of “non-random aperiodicity” in the superlattice. Since the design of aperiodic superlattice involves vast degrees of freedom, one of the key tools has been the machine learning, where we are now capable of efficiently handing extremely large search space using various machine learning schemes [1, 2] and hardware such as quantum annealer [3]. We performed a thorough demonstration from designing the structure by coupling coherent phonon transport calculation and machine learning to experimental fabrication and characterization for the GaAl/GaAs superlattice [4]. We further found a large controllability by the non-random aperiodicity in graphene/MoS₂ heterostructure [5]. An interesting finding here is that non-random aperiodic structure performs better than random aperiodic structure. Each part of the aperiodic structure selectively imposes interference of specific phonon band, allowing the overall structure to act on the entire thermal phonon spectrum. Here, competition with Anderson localization, which works best with full randomness, becomes important and we find that non-random structure is superior when the sample length/thickness is at nanoscale.

[1] S. Ju, etal, Physical Review X, 7, 021024 (2017)
[2] T. Dieb, etal, Science and Technology of Advanced Materials, 18, 498 (2017).
[3] K. Kitai, etal, Physical Review Research, 2, 013319 (2020).
[4] R. Hu, etal, Physical Review X, 10 (2), 021050 (2020).
[5] S. Hu, etal, (submitted).

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.

Grain boundaries (GBs) often control a variety of properties of polycrystalline materials, which are widely used for practical applications. Atomistic simulations are helpful for understanding the relationship between GBs and materials properties. However such a simulation is largely limited by the lack of accurate and computationally efficient interatomic potentials. Machine learning has recently demonstrated its ability to develop interatomic potentials at first-principles level accuracy while costs a few orders of magnitude cheaper computational expense. However, for disordered structures like GBs, developing a high-fidelity machine learning interatomic potential (MLIP) usually requires a large training dataset to cover most of atomistic configurational space. Building the training dataset typically from first-principles calculations is very time-consuming. Here, we propose an efficient scheme for developing MLIPs for GBs from small dataset, while covering the entire atomistic configurational space. Using graphene GBs as an example, we show that the trained MLIP can achieve first-principles accuracy in the prediction of a broad range of structures and properties, including grain boundary energies, elastic constants, phonon spectra, etc. Our MLIP shows the almost same level of accuracy even for GBs not included in the training dataset, demonstrating an excellent transferability. We expect the proposed scheme will facilitate the efficient and rationale development of MLIPs of general GBs, enabling large-scale simulations of GB systems with high accuracy and low computational cost.

A century of experiment and theory have produced a thorough understanding of heat conduction by phonons in simple inorganic crystals. By contrast, basic understanding of heat conduction by molecular vibrations in soft materials (amorphous and crystalline polymers, small molecule solids, and biological materials) is much less mature. Complex, non-periodic structures spanning multiple length scales are difficult to characterize and model. Low thermal conductivity, fiber morphologies, poor control of defects, and anisotropy created by molecular order create daunting challenges for experiment. I will discuss our past work on the thermal conductivity and elastic constants of a wide variety of polymeric materials in the form of thin films and fibers that span a factor of 300 range in thermal conductivity, 0.06 to 20 W/m-K. Time-domain thermoreflectance (TDTR) provides a common experimental platform for these studies; varying the thickness and modulation frequency changes the relative sensitivities of the TDTR measurement to thermal conductivity and heat capacity. Our recent work has employed light-activated changes in the morphology azo-polymers to switch the thermal conductivity by a factor of 3 between a low conductivity amorphous state and higher thermal conductivity crystalline state. We are developing frequency-domain probe beam deflection and optical-fiber based TDTR measurements to provide new capabilities for measurements of the thermal effusivity and thermal diffusivity of small volumes of soft materials.

Amorphous chalcogenide alloys are key materials for optical/electronic data storage and thermoelectric applications due to their threshold switching and low lattice thermal conductivity. Here, we report on a mechanism to suppress the thermal transport in a representative system, silicon telluride (SiTe), nearly by an order of magnitude via systematically engineering the cross-linking network between the atoms. As such, we experimentally demonstrate that in fully dense amorphous silicon telluride the thermal conductivity can be reduced to as low as 0.1 ± 0.01 W/m/K for high tellurium content with a density nearly twice that of amorphous silicon. With no fitting parameter, our estimation for thermal conductivity using ab-initio simulations integrated with lattice dynamics shows excellent agreement with the empirical measurements. We attribute the ultralow thermal conductivity of SiTe to suppressed contribution of propagons and a large shift in the mobility edge -a factor of five- towards lower frequency modes, leading to strong localization of extended modes of vibration, namely diffusons. This localization is the result of large atomic mismatch and a transition from over-constrained amorphous network to under-constrained atomic structure.

Several recent experimental and computational studies show that the thermal conductivity of amorphous silicon varies with sample size.^1–4 This suggests that phonon-like propagating vibrational modes (propagons) carry a significant amount of heat in amorphous silicon. The previous computational studies used the continuous random network (CRN) atomistic structure that represents an ideal random structure without medium- and long-range orders. However, recent spectroscopy results reveal that the actual atomistic structure of amorphous silicon contains medium-range order (MRO) that the CRN structure does not.⁵ In this work, we show the dependence of the propagon thermal conductivity on the MRO which has not been discussed in the previous studies. We compare the extent of MRO and the propagon thermal conductivity in several model amorphous silicon structures using simulated fluctuation electron microscopy, dihedral angle distribution, and molecular dynamics simulation. The results indicate that the structures with MRO show significantly larger propagon thermal conductivity (over 50 %) than the structures without MRO. As the extent of MRO depends on the material preparation methods, our study suggests that the thermal conductivity of amorphous Si also should depend on the material preparation method.⁶


References
¹ H.-S. Yang, D.G. Cahill, X. Liu, J.L. Feldman, R.S. Crandall, B.A. Sperling, and J.R. Abelson, Phys. Rev. B 81, 104203 (2010).
² S. Kwon, J. Zheng, M.C. Wingert, S. Cui, and R. Chen, ACS Nano 11, 2470 (2017).
³ J.L. Braun, C.H. Baker, A. Giri, M. Elahi, K. Artyushkova, T.E. Beechem, P.M. Norris, Z.C. Leseman, J.T. Gaskins, and P.E. Hopkins, Phys. Rev. B 93, 1 (2016).
⁴ K.T. Regner, D.P. Sellan, Z. Su, C.H. Amon, A.J.H. McGaughey, and J.A. Malen, Nat. Commun. 4, 1640 (2013).
⁵ M.M.J. Treacy and K.B. Borisenko, Science 335, 950 (2012).
⁶ A. Hashemi, H. Babaei, and S. Lee, J. Appl. Phys. 127, 45109 (2020).

A continuous increase in power density and miniaturizing electronics makes heat dissipation one of the most important and complex technological challenges [1]. A major issue is that vast amounts of waste heat generated during device operation leads to overheating problems [2, 3]. A need for new materials to manage overheating arises. Polymers have been used for electronics packaging, thanks to their unparalleled properties: light weight, electrical insulation, chemical resistance and easy processability [4]. However, common polymers are generally regarded as thermal insulators with low thermal conductivity and are undesirable for efficient heat dissipation [5]. Understanding thermal transport in polymers and turning polymer insulators into heat conductors are needed for better thermal management [5, 6]. In this poster, we will present the recent work on improving thermal conductivity of polymers and understanding the relationships between thermal conductivity and polymer structures [5, 6]. We have developed polyethylene films with a high thermal conductivity of ~62 W m^-1 K^-1, which is two orders of magnitude greater than that of traditional polymers (~0.1 W m^-1 K^-1) and is higher than that of many metals (e.g. 304-stainless steel ~15 W m^-1 K^-1) and ceramics (e.g. aluminum oxide ~30 W m^-1 K^-1). Unlike conventional approaches focusing on crystalline phase in polymers that can only marginally increase the thermal conductivity, we have engineered the non-crystalline chains through disentanglement and alignment and achieved a remarkably high thermal conductivity of ~16 W m^-1 K^-1 [5]. We will also discuss our current work on investigating the relationships between thermal conductivity and polymer structures such as chain morphology and radius of gyration [6].

Reference:
1. Goyal, V. and A.A. Balandin, Thermal properties of the hybrid graphene-metal nano-micro-composites: Applications in thermal interface materials. Applied Physics Letters, 2012. 100(7): p. 073113.
2. Ball, P., Computer engineering: Feeling the heat. Nature, 2012. 492(7428): p. 174-6.
3. Waldrop, M.M., The chips are down for Moore's law. Nature, 2016. 530(7589): p. 144-7.
4. Peplow, M., The plastics revolution: how chemists are pushing polymers to new limits. Nature, 2016. 536(7616): p. 266-8.
5. Y. Xu, D. Kraemer, B. Song, Z. Jiang, J. Zhou, J. Loomis, J. Wang, M. Li, H. Ghasemi, X. Huang, X. Li & G. Chen, Nanostructured polymer films with metal-like thermal conductivity. Nature Communications, 2019. 10(1): p. 1771.
6. Y. Xu, X. Wang, and Q. Hao, A mini review on thermally conductive polymers and polymer-based composites. Composites Communications, 2021. 24: p. 100617.

Two-dimensional materials offer a wide range of options to study thermal transport in the nanoscale. The current understanding of thermal transport by phonons will be reviewed considering a few exemplary cases. We will discuss the case of free-standing polycrystalline membranes of MoS2 where the grain size, their orientation and its distribution all play important roles to control the thermal conductivity, for which we have carried out experiments and simulations. We explore the case of single crystal membranes containing few monolayers in which the sample preparation is crucial to eliminate effects of imperfections and contamination focusing on the thickness dependence of the thermal conductance. The role of the surface and interfaces will be discussed using the example of MoS2/hBN heterostructures. Recent laser Raman thermometry data obtained from MoS2 crystalline membranes supported on patterned substrates, with patterns covering the crossover range of phonon confinement to ballistic transport and from supported single-layer membranes of PtSe will be reported. In the former the focus is on the impact of phonon hybridization between the membrane and the patterned substrate, which has been observed and discusses in free-standing Si membranes on a phononic crystal, on the thermal transport. In the latter material, new challenges emerge and these will be discussed. On the methodological part, the sources of errors in the determination of thermal properties and outstanding questions concerning phonon transport in TMDCs will be discussed with view to find reliable figures of merits for a future benchmarking.
Work in collaboration with M. Sledzinska, E. Chavez Angel, A. El-Sachat, X. Peng, M-J Esplandiu, R. Quey, B. Mortazavi, B. Graczykowski, M. Placidi, D. Saleta Reig, F. Alzina, L. Colombo, S. Roche, F. Bonell and M. Jamet.

Molybdenum disulfide (MoS₂) is a promising material for optoelectronics that belongs to the Transition Metal DiChalchogenide (TMDC) family and that can be grown in monoloyers. Studies of such materials have led to a broad range of values for thermal conductivity and thermal contact conductance, both on simulation [1] and experimental [2] sides, probably due to the variety of configurations considered. Scanning Thermal Microscopy (SThM) is an Atomic Force Microscopy-based technique dedicated to the thermal investigation of micro and nanometric structures such as membranes, grains, films, etc, especially useful for cross-plane transport studies [2]. In this work, we apply the technique to characterize MoS₂ monolayers grown by Chemical Vapor Deposition (CVD) on Silicon Dioxide SiO2_300nm/Si substrates. Our technique employs a self-heated resistive probe that is brought into contact with the surface of the sample to be characterized. Scanning the sample surface provides simultaneously the topography and thermal contrast images of sample. Using either the Wollaston microprobe or the Palladium nanoprobe in ambient environment, we find that SThM images show a clear thermal contrast between the substrate and MoS2. Based on a calibration of the SThM probe using reference bulk materials and an original set of {oxide films on silicon substrate} samples of varying thickness, we are able to trace back to the effect of the MoS2 layers. We find that the boundary dominates heat conduction on supported samples, and provide estimates of associated thermal contact resistance. Simulations reproducing the full {SThM probe + MoS2 sample} experiment allows to clarify heat conduction in such systems.

[1] D.D. Fan et al., Applied Physics Letters, 103(13), 133113 (2014); A.N. Gandi et al., EuroPhysics Letters, 113(3), 36002 (2016); W. Li et al., Applied Physics Letters, 103(25), 253103 (2013)
[2] R. Yan et al., ACS nano 8.1 (2014): 986-993 (2104); X. Zhang et al., ACS applied materials & interfaces, 7(46), 25923-25929 (2015); A. Taube et al., ACS applied materials & interfaces, 7(9), 5061-5065; S. Vaziri et al., Sci Adv. 5(8): eaax1325 (2019)
[3] S. Gomès, A. Assy, P.O. Chapuis, Physica status solidi (a), 212(3), 477-494.(2015).

Understanding and controlling interfacial heat transport is crucial for several applications in nanoelectronics, photonics, and energy conversion. Over the past few years, ultrafast optical techniques have enabled high-accuracy measurements of the thermal boundary conductance. Typically, these measurements are performed on systems that are thick compared to atomic lengthscales (at least several nm), and lack the ability to distinguish materials on both sides of the interface. The emergence of two-dimensional (2D) materials and heterostructures has revealed the need for techniques that can provide angstrom-scale, material-selective characterization of heat flow on picosecond timescales. Here, using mega-electron-volt ultrafast electron diffraction (MeV-UED), we present time-resolved measurements of thermal transport across type-II heterojunctions of monolayer 2D materials. Such heterostructures are of significant interest due to ultrafast (<100 fs) charge separation that is known to occur following photoexcitation. However, it has remained unclear how these electronic processes dissipate energy to the lattice. In this work, we prepare electron-transparent samples of 2D heterostructures, and selectively excite layer(s) within the stack using a tunable wavelength source. Layer-specific heating is subsequently monitored through the time-dependent dimming of Bragg peak intensities, thereby enabling picosecond tracking of energy flow across angstrom-scale junctions. Using ab initio calculations, we uncover the fundamental coupling between the electronic and lattice degrees of freedom in van der Waals 2D heterostructures.

Two-dimensional (2D) materials and their heterogeneous integration promise to advance the applications in electronics and photonics. However, the high thermal resistance at the interfaces bonded by weak van der Waals (vdW) interaction largely limits the energy dissipation, which may induce significant temperature rise and large thermal mismatch, and therefore degrades the device performance. Considering the practical tunability of vdW binding for 2D heterostructures reported by experimental studies, we investigated the thermal transport through an MoS₂-amorphous silica (a-SiO₂) heterostructure using molecular dynamics (MD) simulation, where the effects of vdW binding on both the in-plane and the cross-plane heat transfer are considered. We model the vdW interaction between MoS₂ and a-SiO₂ using the Lennard-Jones (LJ) potentials and scale the binding energy of the LJ force field. The in-plane thermal conductivity of the MoS₂ monolayer was calculated using the Green-Kubo formalism from the equilibrium MD simulation. The cross-plane thermal conductance between the MoS₂ monolayer and the a-SiO₂ substrate was calculated by the thermal circuit model using the approach-to-equilibrium MD. We show that the increase of vdW binding energy leads to a decrease of in-plane thermal conductivity and increased cross-plane thermal conductance. In particular, we observed saturation of the cross-plane thermal conductance at high vdW binding energy, which is attributed to the substrate-induced localized phonon modes in the MoS₂ with high phonon relaxation times and low participation ratios. The insights gained at the atomistic level lead to a device level optimization strategy of heat dissipation in the solid-state 2D nanodevice thermal management. Through the finite element analysis based on MD-calculated thermal properties, we show that three length scales, i.e., the hotspot size, the size of the 2D film, and characteristic length of the lateral heat spread, determine the device level thermal transport and the optimal vdW binding energy that maximizes heat dissipation. We identified two distinct heat transfer regimes that are characterized by the coupling of in-plane and cross-plane heat transfer and the cross-plane heat transfer, because of the interplay among the three length scales. At each regime, the optimal heat transfer for device operation was analyzed. Our study elucidates how the vdW binding affects the thermal transport through the 2D heterostructure from the atomistic level to the device level, which provides a general guideline to optimize the heat dissipation of high-performance 2D nanoscale devices.

Understanding nanoscale thermal transport processes is essential to design nano-materials and devices with desired thermal transport properties for thermal energy conversion and management. Despite the significant progress in thermal transport of inorganic crystals, thermal transport in more complicated materials and across material interfaces remains largely unexplored. In this talk, I will present my research group’s efforts to understand nanoscale thermal transport in hybrid organic-inorganic materials and across interfaces, using atomic modeling and experimental characterization. I will start with our study on phonon dynamics in hybrid perovskites using inelastic x-ray scattering. It provides useful insights into their ultralow thermal conductivity. I will then share our recent development on anharmonic atomistic Green’s function, which allows us to capture the importance of inelastic scattering at the interfaces. Lastly, I will share our latest work on phonon Anderson localization and quantum thermal rectification.

Top-side device cooling methods, e.g., diamond heat sinks on top of the metal contacts, facilitate heat removal from hot spots near the gate, directly lowering local and global temperature rises. Because the top-side surface of modern devices is often complex and rarely flat, conformal nanocrystalline diamond (NCD) growth techniques are critical to maximize the contact area between the device and the heat sink. Weak adhesion at this interface will cripple thermal transport, acting as a large thermal resistance. As a result, thin adhesive metal layers are often deposited between the device and the heat sink. The effectiveness of heat sinks is directly dependent on the thermal boundary conductance of the metal/NCD interface. In this study, we seek to optimize thermal transport across Au/Cr/Pd/NCD stacks. Cr was chosen for its thermal merits. Excellent adhesive properties and high maximum phonon frequencies has enabled Cr to produce some of the largest thermal boundary conductance interfaces for materials like Diamond, Al2O3, and AlN. Despite this, Cr is predicted to react with NCD to form enthalpically favorable carbides, which will lower the thermal boundary conductance. We hypothesize that a thin non-reactive Pd layer can eliminate these reactions and preserve the metal/NCD thermal boundary conductance in heated conditions. Pd and then Cr layers are deposited as perpendicular wedges with thickness of 0 to 20 nm on a single NCD sample, allowing us to test all thickness combinations of Pd and Cr. The thermal boundary conductance is measured using a non-destructive pump-probe optical technique, frequency domain thermoreflectance (FDTR). By scanning the FDTR across the sample surface, a 2D map of thermal boundary conductance is created for all Pd/Cr thickness combinations. A local maximum in thermal boundary conductance is observed for a specific Cr/Pd thickness combination, providing evidence that our hypothesis is correct. Using Tunneling Electron Microscopy (TEM), we inspect a lift-out sample taken from our original NCD results, at the location of the TBC local maximum. Using image filtering, we identify the structure of this metal interface and determine the thickness of each metallic and reacted-carbide layers. The sample was then annealed and remeasured multiple times. Changes in the 2D thermal boundary conductance map following each heating cycle is used to quantify the ability for thin Pd layers to improve the thermal stability of metal contacts like Cr. Since large temperature rises will alloy and chemically react metal films and interfaces, it is crucial that we evaluate the thermal performance of metal stacks throughout the entire device’s lifetime, not just when it’s first made.

Interfacial resistance between solids has drawn great interest due to its importance for various applications such as thermal management in electronic devices and energy conversion. The simple Landauer formula, which had been widely used, fails in describing interfacial resistance because it neglects internal phonon scattering. Past studies have shown that the modified Landauer formula proposed by S. Simons could give a much smaller interfacial resistance by considering internal phonon scattering [S. Simons, Journal of Physics C: Solid State Physics, 1974]. However, it assumes that phonons near the interface follow the distribution of bulk phase (i.e., distribution in an infinitely large sample with homogeneous temperature gradient), neglecting the non-equilibrium phonon distribution due to the interface and the complex interplay between the interface scattering and internal phonon scattering. We study the phonon transport across an interface between silicon and germanium by solving the Peierls-Boltzmann transport equation with inputs from first-principles. The phonon distribution is solved in both real and reciprocal spaces, and the local effects can be well captured. The results show that both the Landauer formula and its modified version fail for interfacial phonon transport due to their assumptions neglecting highly non-equilibrium phonon distribution near an interface. The highly non-equilibrium phonon distribution decays with a distance from the interface and finally recovers to the phonon distribution of bulk phase at the distance which we define as an effective interfacial region. We find that the internal phonon scattering within the interfacial region provides an important contribution to overall interfacial resistance. Our study can help to understand the large discrepancy between experimentally measured interfacial resistance and that from the Landauer formula that is often found in literature, providing a useful way to interpret experimental data.

Effective heat dissipation from a two-dimensional semiconducting transition metal dichalcogenide (TMD), such as MoS₂ or WS₂, to its substrate is constrained by the thermal boundary conductance (TBC) of their planar interface which can exhibit significant variability with: (1) the charge carrier density, (2) surface roughness, (3) substrate elastic property, and (4) TMD thickness. Hence, insights into the factors affecting the TBC variability can lead to superior thermally aware design and more effective thermal management of TMD-based nanoelectronic devices through the optimal use of materials. We probe the physical origins of the TBC variability by modeling the underlying electronic and phononic processes in the TBC.
We quantify the temperature-dependent electronic TBC G_el between a single-layer TMD (MoS₂ or WS₂) and different dielectric insulating substrates (SiO₂, HfO₂ or Al₂O₃) and find that it varies significantly with the dielectric material and electron density, indicating that G_el is strongly tunable with a metal gate especially with Al₂O₃ [1]. To account for the effect of surface roughness on the phononic TBC G_ph, we introduce a phenomenological theory for diffuse phonon scattering between the TMD and its substrate [2]. The G_ph from diffuse scattering is compared to and found to be substantially higher than the G_ph from coherent scattering as well as comparable to experimentally measured TBC values for the MoS₂-SiO₂ interface, suggesting that the TMD-SiO₂ interface is highly disordered. To understand the effects of anisotropic elasticity in the substrate, we formulate an elasticity theory-based model of stacked thin plates to model the phononic TBC between a TMD and h-BN, a layered substrate [3]. Good agreement between the model and experimental TBC data is obtained for the MoS₂-BN interface. Our results reveal that h-BN is substantially more effective for heat dissipation than SiO₂, especially when the number of layers in the MoS₂ is large, because of the effective coupling of flexural modes across the MoS₂-BN interface.
References
1. Z.-Y. Ong, G. Zhang, Y.-W. Zhang, and L.Y. Cao, “Gate-tunable cross-plane heat dissipation in single-layer transition metal dichalcogenides,” Phys. Rev. Research 2(3), 033470 (2020).
2. Z.-Y. Ong, Y. Cai, G. Zhang and Y.-W. Zhang, “Theoretical analysis of thermal boundary conductance of MoS₂-SiO₂ and WS₂-SiO₂ interface,” Submitted.
3. Z.-Y. Ong, G. Zhang and Y.-W. Zhang, “Theory of phononic heat dissipation to a hexagonal boron nitride substrate,” Submitted.

Nanostructured thermoelectric materials are generally highly polycrystalline, meaning that defects such as interfaces are prevalent. Advancing beyond the current state-of-the-art nanostructured thermoelectric materials requires a detailed understanding of the impact of interfaces on their thermal properties. In addition to the importance of nanoscale engineering, it is highly desirable to achieve high thermoelectric performance at room temperature for Internet of Things, energy harvesting, and active cooling applications. Bismuth telluride (Bi₂Te₃) is the best performing thermoelectric material at room temperature. One route to achieving highly efficient thermoelectrics at room temperature is to understand the impact of interfaces in materials such as Bi₂Te₃ on thermal transport at an atomistic level.

In this work, we utilise reverse non-equilibrium molecular dynamics (rNEMD) simulations, with a recently developed empirical interatomic potential [1], to elucidate the impact of specific twin boundary structures on thermal transport in Bi₂Te₃. Three basal plane twin boundaries terminated at different layers in Bi₂Te₃ and the 60° twin boundary [2] are investigated. The interfacial thermal resistance (Kapitza resistance) is calculated, along with an analysis of the structural changes observed at interfaces. We find that interfacial thermal resistance increases with decreasing stability of the interface, suggesting that the most effective interfaces may require more effort to fabricate. Despite large structural differences between the basal twin boundaries and the 60° twin boundary, we do not see significant variance in interfacial thermal resistance when comparing these interfaces. A comparison of the properties of these four twin boundaries is carried out, enabling us to identify those that may be the most impactful in terms of suppressing lattice thermal conductivity at room temperature.

[1] B. Huang et al., J. Phys. D Appl. Phys. 52, 425303 (2019)
[2] K. Kim et al., Nat. Commun. 7, 1-6 (2016)

This work is supported by a research grant from Science Foundation Ireland (SFI) and the Department for the Economy Northern Ireland under the SFI-DfE Investigators Programme Partnership, Grant No. 15/IA/3160. This work is also supported by SFI Grant No. 13/RC/2077.

The ultrahigh lattice thermal conductivities reported near the room temperature for boron arsenide (BAs) and carbon nanotubes (CNTs) make them useful for both thermal management applications and fundamental studies of phonon transport. The peak thermal conductivity of a single crystal usually appears at a low temperature and provides insight into the competition between extrinsic phonon-defect scattering with intrinsic phonon-phonon processes. However, existing measurements have yet to obtain the peak thermal conductivity of BAs and reported a much higher peak temperature for CNTs than for high-quality graphite. Here we report highly sensitive electro-thermal bridge measurements of the peak thermal conductivities of both bulk BAs crystals and individual CNTs. Theoretical calculations are used to correlate the thermal transport measurement results with the structure and chemical composition characterization data.

Multinary chalcogenides are diverse systems, which can be derived from simpler binary compositions by cation substitution and preserving the octet rule. One can achieve different classes of materials with more complicated compositions, which share some common features and yet have distinct properties. By using first principles simulations we study the electronic and vibrational properties of quaternary chalcogenides from the I₂-II-IV-VI₄, I-II₂-III-VI₄, and I-IV-V-VI₃ families. These systems have inherently low thermal conductivity, which is directly related to the lattice structures and underlying anharmonic phonon scattering mechanisms. Our first principles simulations help us understand from a microscopic perspective the phonon conduction processes in these classes of materials. We also find that there are various possibilities for different atomic arrangements which can lead to distinct structural phases. Also, various types of doping are possible, which can further be used for electronic property optimization. The inherently low thermal conductivity and the available earth-abundant chemical compositions make these materials attractive for applications in thermoelectricity and solar cells among others.
Support from the US National Science Foundation under Grant No. DMR-1748188 is acknowledged.

Clathrate compounds are promising candidates for thermoelectric materials in terms of the “electron-crystal phonon-glass” concept. Clathrate compounds are composed of guest atoms encapsulated in cagelike structures. It is expected that electrons can smoothly transport in the frame work while phonons are scattered by guest atoms. Particularly, in type-I clathrate compounds, guest atoms vibrate in a strong anharmonic potential because they are composed of a Weaire-Phelan structure, which divides a space with the maximum volume with the same cross-sectional area. While off-center clathrates such as type-I Ba₈Ga₁₆Sn₃₀, in which guest atoms vibrate in a double-well anharmonic potential, have low thermal conductivity, silicon-based clathrates are preferable for actual use in terms of the material cost.
In this study, we show that aluminum substitutions can enhance thermoelectric performance of silicon-based clathrates, type-I Ba₈Ga₁₆Si₃₀. Comparing calculated thermoelectric figure of merits (ZT) with those obtained in experiment, we concluded that (a) aluminum additive enhances ZT with lowering doping level and (b) further suppression of doping level can enhance ZT significantly. To investigate thermoelectric properties, we obtained possible structures at finite temperature with using a cluster expansion and Monte Carlo method [M. Ångqvist, P. Erhart et al., Chem. Mater. 29, 7554 (2017) and 28, 6877 (2016)] and thermoelectric properties were averaged for different possible structures. Phonon properties of the clathrates were calculated with using a self-consistent phonon (SCP) theory based on first-principles calculations [Tadano et al. Phys. Rev. Lett 120, 105901 (2018)]. Our state-of-the-art analysis method shows a good agreement with experimental data.

Understanding the ultrafast energy dynamics in a metal after laser heating is critical for many engineering applications, including nanoscale thermal engineering, photocatalysis, and spintronics. Despite the topic’s importance, and decades of study, fundamental questions about the topic remain unanswered. In this talk, I discuss my group’s experimental work to answer two such questions. (i) Do photoexcited electrons transport energy ballistically, super-diffusively, or diffusively before thermalizing with phonons? (ii) Can we control the energy dynamics in metal alloys by using the effect of alloy composition on electronic band structure to control the strength of quasi-particle interactions? To answer the first question, we combine wavelength-dependent time-domain thermoreflectance experiments and time-resolved magneto-optic Kerr effect experiments to study nanoscale heat-transfer in Au/Co multilayers. We perform measurements on a series of samples with varied Au thickness. Our experiments show that, while the energy dynamics in Au after photoexcitation are nonthermal, heat is still transported diffusively. To answer the second question, we perform pump/probe experiments on Au_1-xAl_x and Co_1-xFe_x alloys. We find the ultrafast energy dynamics in these alloys depends strongly on composition. In Au_1-xAl_x alloys, we credit the observed relationship between alloy composition and electron-phonon energy relaxation time to changes in the strength of electron-phonon interactions. In Co_1-xFe_x alloys, we credit the observed composition dependence of the electron-magnon energy relaxation time to changes in the strength of electron-magnon interactions. I conclude my talk by discussing the implications of our findings to the plasmonics and spintronics communities.

Light–matter interactions that induce charge and energy transfer across interfaces form the foundation for photocatalysis, energy harvesting and photodetection, among other technologies. One of the most common mechanisms associated with these processes relies on carrier injection. However, the exact role of the energy transport associated with this hot-electron injection remains unclear. Plasmon-assisted photocatalytic efficiencies can improve when intermediate insulation layers are used to inhibit the charge transfer or when off-resonance excitation is used, which suggests that additional energy transport and thermal effects could play an explicit role even if the charge transfer is inhibited. This provides an additional interfacial mechanism for the catalytic and plasmonic enhancement at interfaces that moves beyond the traditionally assumed physical charge injection. In this work, we report on a series of ultrafast plasmonic measurements that provide a direct measure of electronic distributions, both spatially and temporally, after the optical excitation of a metal/semiconductor heterostructure. We explicitly demonstrate that in cases of strong non-equilibrium, a novel energy transduction mechanism arises at the metal/semiconductor interface. We found that hot electrons in the metal contact transfer their energy to pre-existing free electrons in the semiconductor, without an equivalent spatiotemporal transfer of charge. Further, we demonstrate that this ballistic thermal injection mechanism can be utilized as a unique means to modulate plasmonic interactions. These experimental results are well-supported by both rigorous multilayer optical modelling and first-principle ab initio calculations.

Photoexcitation of a metal results in hot electrons that cool on femto to picosecond time-scales because of electron-phonon interactions. The energy relaxation times of hot electrons can vary a lot depending on the metal. For example, the electron-phonon energy relaxation time in Pt is less than 200 fs ^[1]. Alternatively, in Au experiments suggest the energy relaxation is between 1 and 3 ps ^[2-5]. In this work, we experimentally study how energy relaxation times in metals are determined by the interplay of electron-electron (e-e) interaction and electron-phonon (e-p) interaction strengths. We report on pump/probe measurements of picosecond acoustic signals in Au, Al, and their alloys. We perform these measurements on wedge samples with varied thicknesses. These metals have a wide range of e-e and e-p interaction strengths, and therefore provide a rigorous test on existing theories for hot electron dynamics. The picosecond acoustic signals we measure are caused by the rapid change in lattice temperature as a function of depth that occurs following laser excitation. Therefore, our experimental signals are a sensitive probe of electron-phonon energy relaxation times. Our results provide insight into how quasi-particle interactions govern nanoscale heat transfer in metal systems, which is important to many disciplines, including nanophotonics, plasmonics, and ultrafast magnetism.

Reference
1. Hyejin Jang, Johannes Kimling, and David G. Cahill. Nonequilibrium heat transport in Pt and Ru probed by an ultrathin Co thermometer. Phys. Rev. B 101, 064304
2. Mueller, B. & Rethfeld, B. Relaxation dynamics in laser-excited metals under nonequilibrium conditions. Physical Review B 87, 035139 (2013).
3. Groeneveld, R. H. M., Sprik, R. & Lagendijk, A. Femtosecond spectroscopy of electron-electron and electron-phonon energy relaxation in Ag and Au. PRB 51, 11433-11445, doi:10.1103/PhysRevB.51.11433 (1995).
4. Del Fatti, N., Voisin, C., Achermann, M., Tzortzakis, S., Christofilos, D. & Vallée, F. Nonequilibrium electron dynamics in noble metals. PRB 61, 16956-16966, doi:10.1103/PhysRevB.61.16956 (2000).
5. Gusev, V. E. & Wright, O. B. Ultrafast nonequilibrium dynamics of electrons in metals. PRB 57, 2878-2888, doi:10.1103/PhysRevB.57.2878 (1998).

We introduce thermally insulating corrugated spacers that establish micron-scale electrode gaps without electrical shorting in thermionic energy converters, allowing temperature differences of hundreds of kelvins to be achieved with only a few watts of conductive heat transfer per square centimeter. These films’ microscale corrugation serves both structural and thermodynamic functions, namely, to increase their robustness^1,2 and to increase their overall thermal resistivity to limit heat transfer between the electrodes^3,4. Our thermomechanical testing reveals that these films have thermal resistance values high enough to reduce parasitic conduction to levels below that of thermionic heat transfer even at moderate (~1300 K) emitter temperatures, resulting in great enhancements in output current density and thermionic conversion efficiency while minimizing heat loss. These results represent a significant step toward high power density, mass producible, commercially viable direct heat-to-electricity energy conversion.
Thermionic energy converters are solid-state heat engines in which electrons are produced by a hot cathode, cross a vacuum gap, and impinge on a cold anode, delivering electric current. Maximizing efficiency and power density in these devices requires minimizing their vacuum gap thicknesses to roughly 2 μm. Establishing gaps of this scale by intricate external positioning mechanisms is impractical for mass-produced devices, thereby necessitating intra-gap spacers to internally maintain spacings. Previous methods of establishing these gaps have had limited success for reasons related to their high conductive losses or other manufacturing difficulties^5,6.
Therefore, we have developed spacers made through alumina atomic layer deposition with micron-scale thicknesses and square centimeter-scale areas that can be mass-produced using standard microfabrication techniques and subsequently transferred to thermionic energy converter electrodes. These films include hexagonal apertures to allow high electron currents (several A/cm²), can tolerate compressive pressures of several atmospheres to ensure electrode parallelism, and have high thermal resistances (hundreds of cm²K/mW) to minimize conductive losses. Our films’ low conductance values stem from the high contact resistance they exhibit where they touch the electrodes³ and their carefully architected nanoscale corrugation, which reduces the area through which conduction can occur⁴.
We fabricate our spacer films by (1) constructing a silicon mold via sequential cycles of photolithography and deep reactive ion etching, (2) conformally coating the mold with insulating materials such as alumina using atomic layer deposition, (3) cutting hexagonal apertures in the insulating layer using inductively coupled plasma etching, and (4) releasing the film using XeF₂ etching. In addition, we have developed an ultra-high vacuum testing apparatus to probe the interactions between the mechanical and thermal properties of our films and in addition are conducting extensive electrical testing of these spacers in thermionic devices.

1) Davami Nat. Commun. 6(2015)10019
2) Jiao Extreme Mech. Lett. 34(2020)100599
3) Nicaise Microsyst. Nanoeng. 5(2019)31
4) Campbell J. Microelectromech. Syst. 29(2020)637
5) Littau Phys. Chem. Chem. Phys. 15(2013)14442
6) Belbachir J. Micromech. Microeng. 24(2014)085009

Thermionic energy conversion is the direct conversion of heat into electricity by the spontaneous emission of high-energy electrons from a surface. Although the physical mechanism of thermionic emission has been suggested to be used for energy conversion applications for over century now, thermionic energy converters (TEC) have not been widely implemented due to low conversion efficiency even at high temperatures. Recently, it has been suggested that by reducing the inter-electrode gap distance in a thermionic converter, electron emission can increase remarkably mainly due to the mitigation of excessive potential barriers (i.e., negative space charge). However, by shrinking the inter-electrode dimensions, other physical phenomena now need to be considered including near-field thermal radiation and image-charge effects. Therefore, significant knowledge gaps remain in the comprehensive modeling and analysis of thermionically induced electron and photon emission, particularly across nanometer distances. In this study, we present a comprehensive model describing the performance of a tungsten-based TEC by developing an energy balance with various heat flux inputs to the emitter and considering the effects of negative space charge, image charge, and near-field thermal radiation between the two electrodes. The steady-state temperatures of the emitter and collector are found to satisfy the energy balance of the TEC and the effects of inter-electrode gap distance, operating voltage, and work functions of the electrodes on the operating performance are further investigated for different heat input scenarios. The dependence of work output and efficiency on gap distance is highlighted to show the maximum efficiency and maximum power points, which are helpful in selecting gap distance for optimum performance. By clear illustration of the trade-off between the aforementioned physical phenomena at the nanoscale regime using this comprehensive model, our results prove that the power output and the conversion efficiency of a nanoscale TEC system can surpass those of conventional TECs and other competitive solid-state converters.

With miniaturizing electronic devices, thermal management has attracted great interest for overcoming various heat issues in nanoscale world: dissipating heat, energy harvesting from the wasted heat, and so on. Therefore, much effort has been made to reveal the phonon transport physics in nanoscale leading to tailoring heat conduction. Nowadays, some studies experimentally demonstrated the interesting coherence property of phonon using superlattice structures [1, 2]. On the other hand, in the thermoelectric research field, phonon transport in nanostructured materials has also drawn much attention because of the thermal conductivity reduction. However, in general, the materials including nanostructures have unintentional defective structures: uncontrolled crystal orientation, composition deviation, strains and crystal defects. It needs the atomically-controlled nanostructured materials to elucidate the phonon transport physics in the material system containing nanostructures.
We have developed various kinds of unique well-controlled nanostructures: IV group element-based nanodot structures, etc [3-6]. This talk will discuss the phonon transport physics in the unique nanomaterial films containing atomically-controlled nanocrystals where the thermal conductivity is drastically reduced. One example of our nanostructures is Si-based films including ultrasmall Ge nanodots with sharp interfaces and without strain unlike the conventional Stranski Krastanov nanodots. When Ge nanodots get smaller than 10 nm, the thermal resistance of Ge nanodot layers in Si films decreases, while that of Ge nanodot layers in SiGe films does not depend on Ge nanodot size. The tendency can be explained by phonon scattering theory. In this talk, we will introduce phonon transport in some kinds of well-controlled nanostructured materials. These findings will give a new insight for thermal management.
This work was supported by Grant-in-Aid for Scientific Research A (19H00853), Exploratory Research (19K22110). This work was also supported by JST CREST.

References
[1] M. N. Luckyanova, et al., Science 338, 936 (2012).
[2] J. Ravichandran, et al., Nat. Mater. 13, 168 (2014).
[3] S. Yamasaka, et al., Sci. Rep. 5, 14490 (2015).
[4] Y. Nakamura, et al., Nano Energy 12, 845 (2015).
[5] Y. Nakamura, et al., Sci. Technol. Adv. Mater. 18, 31 (2018).
[6] T. Taniguchi, et al., ACS Appl. Mater. Interfaces 12, 25428 (2020).

Distributions of the phonon mean free path (MFP) in materials and nanostructures are important for understanding and engineering their thermal properties. In the past years, the phonon MFP in bulk semiconductors has been extensively studied both theoretically and experimentally. However, modern MEMS devices often rely on thin membranes rather than bulk material, yet the MFP spectra in thin membranes have not been measured experimentally.
In this work, we demonstrate an experimental method to probe the phonon MFP spectra in suspended semiconductor membranes. The method consists in measuring the thermal conductivity of membranes with arrays of slits and extracting the accumulated thermal conductivity as a function of the MFP through a fully analytical approach based on the semi-analytical procedure proposed by Hao et al. [1]. In silicon, we find that the phonon MFP in 145-nm-thick membranes is much shorter than that in bulk due to the scattering at the membrane boundaries [2]. At room temperature, the phonon MFP spans from 10 nm up to only 400 nm. Yet, at 4 K, the MFP becomes longer, and most phonons can travel ballistically for more than 100 nm. These results thus shed light on the long-lasting question of the range of ballistic and coherent phonon transport at different temperatures in nanostructures based on semiconductor membranes.

References:
[1] Hao et al, Materials Today Physics 10, 10012 (2019)
[2] Anufriev et al, Physical Review B, 101, 115301 (2020)

Scanning Thermal Microscopy (SThM) allows the thermal characterization of materials with a sub-micrometric spatial resolution [1]. While obtaining qualitative trends, e.g. by means of images, is straightforward, determining reliable quantitative thermal data from the experiments is challenging because the probe-sample heat exchange depends strongly on parameters such as the size, geometry and surface states of probe and sample. In this work, we study first the impact of sample roughness on the thermal conductance at the probe/sample contact and then the impact of thin fims of oxide on the surface.
Rough samples consisting of several sets of silicon surfaces with out-of-plane Sq roughness parameters of ~0, 0.5, 4, 7 and 12 nm were prepared by anodic oxidation [2] and measured with three types of probe (Wollaston, palladium and doped Si probes) in vacuum conditions. Sample roughness is clearly shown to reduce the thermal conductance very significantly compared to a flat Si sample, up to 40-95% depending on the probe. SThM sensitivity is low for highly conductive materials, so erroneous thermal conductivity can be deduced from a calibration curve performed for flat samples [3].
In a second step, we analyze the impact of thin oxide films on the determined thermal conductivity. It was shown that heat conduction in the cross-plane direction is strongly affected by native oxide in nanometre-thin suspended silicon membranes [4]. Here we extend the study to supported films of thickness varying between few nanometres and one micrometre. We find a logarithmic dependence on thickness for the apparent thermal conductivity and show that arbitrarily-small thicknesses may be detected. Nanoprobes are found to be less sensitive to the oxide thickness, as they probe only at the very surface [5]. We discuss the impact of thermal boundary resistances and ballistic dissipation in the samples.
All these results allow to envisage precise calibration of nanometre-scale thermal studies.

[1] S. Gomes, A. Assy, P.O. Chapuis, Physical status solidi (a) 3, 447 52015)
[2] J. Zemek, K. Olejnik, P. Klapetek, Surface Science, 602, 7, 1440-1446 (2008)
[3] E. Guen, P.O. Chapuis, N.J. Kaur, P. Klapetek, S. Gomes, submitted (2020)
[4] A.M. Massoud, J.-M. Bluet, V. Lacatena, M. Haras, J.-F. Robillard, P.-O. Chapuis, Applied Physics Letters 111, 063106 (2017)
[5] E. Guen, P.O. Chapuis, R. Rajkumar, P. S. Dobson, G. Mills, J. M. R. Weaver, S. Gomes, submitted (2020)

We acknowledge support from EU project QuantiHeat.

Nanostructuring has been successfully used to improve thermal transport performance across a wide range of materials, whether a high or low thermal conductivity is desired. This includes applications ranging from latent heat storage to thermoelectrics and phononic crystals. Nanoscale porosity has been shown to substantially decrease the thermal conductivity of semiconducting materials — thereby improving their thermoelectric efficiency — and, in some cases, even enhance transport. In this work [1], we identify an unusual anticorrelated specular phonon scattering effect which yields an additional ~ 80% decrease (at maximum efficiency) in the thermal conductivity of specific nanoporous geometries and which lends itself to a high degree of control by manipulating pore size and distribution.

Large-scale equilibrium molecular dynamics (MD) simulations show the presence of negatively correlated (or anticorrelated) heat flux for specific nanoporous geometries. Wave-packet MD simulations show how the heat becomes trapped between large pores with narrow necks due to specular scattering. We further reproduce the behavior observed in the MD simulations and demonstrate how specular scattering can give rise to negatively correlated heat flux using a heuristic ray-tracing analytical model. While specular scattering does not necessarily indicate the presence of coherence effects it is a necessary requirement for coherent interference, and our simulations support the notion that coherence effects could in principle be present at room temperature. We show with MD how the pore size/neck ratio and pore periodicity can be manipulated to control the strength and duration of the correlated reflections and thus the reduction in thermal conductivity. In addition to thermoelectrics, we believe this work could have a broader application in phononics.

[1] de Sousa Oliveira, L., Hosseini, S. A., Greaney, A., & Neophytou, N. Heat current anticorrelation effects leading to thermal conductivity reduction in nanoporous Si. Phys. Rev. B. Accepted October 2020.

Thermal rectification is one of the next steps for a better thermal management. In this study, this effect has been assessed, in an original nanostructuration, thanks to non equilibrium molecular dynamics and wave packets propagation simulations. The direction dependent thermal conductivity results from the use of asymmetric crystalline-core/amorphous-conical-shell nanowires that induces variable axial and radial phonon propagation/confinement. This effect is linked to the combination of a crystalline/amorphous interfaces parallel to the heat flux, and with the variable amount of amorphous coating due to the conical shell of the nanowire. Rectification is analyzed terms of mean free path of phonons, axial and radial energy transfer, energy diffusivity, and vibrational density of states restricted to different constitutive elements of the nanowire.

When the characteristic size becomes comparable to the mean free path of phonons, the quasi-ballistic heat conduction regime arises while interactions with boundaries become significant. Both thermal boundary resistance and ballistic transport flatten the temperature profile and induce temperature jumps in comparison to macroscopic heat conduction: it is therefore difficult to disentangle the two effects by simply observing a temperature profile. We solve the phonon Boltzmann Transport Equation using the Discrete Ordinate Method, which accounts for the directional and non-local aspects of the energy flux, in order to analyze independently the effects and observe their interplay in a system of two contacted thin films. In particular, we find that the effective thermal conductivity for in-plane heat conduction is a function on the nondimensional figure Kn_T/tau, where Kn_T is the mean Knudsen number (ratio of average mean free path to sum of the thicknesses of the two films) and tau the mean transmissivity at their interface. The impact of the specularity of the film boundaries and interface is also evidenced, especially on the angular distribution of phonons.
These results shed light on the relation between local energy non-equilibrium and transport properties.

The support of project EU project EFINED is acknowledged. We thank E. Nefzaoui (ESIEE), D. Lemmonier (Pprime), and C. Abs Da Cruz.

An emerging class of methylammonium lead iodide (MAPbI₃) based Ruddlesden-Popper (RP) phase perovskites, BA₂MA_n-1Pb_nI_3n+1 (n = 1 – 7), exhibit enhanced stability to environmental conditions relative to MAPbI₃, yet still degrade at elevated temperatures. We use Frequency-Domain Thermoreflectance (FDTR) to experimentally determine the thermal conductivities of these layered RP phases for n = 1 – 4 thin films and n = 3 – 6 single crystals. We find that the initially decreasing trend of thermal conductivity in similarly oriented perovskites with increasing n results from the transport properties of coherent phonons, emergent from the superstructure, that do not scatter at the interfaces of organic butylammonium (BA) chains and perovskite octahedra. With the exception of n = 1, the perovskites’ thermal conductivities are lower than the range of 0.34 – 0.50 W/m/K reported for single crystal MAPbI₃. Reduced-order lattice dynamics modeling suggests that reduced group velocity of coherent phonons in n = 3 – 6, a consequence of band flattening in the phonon dispersion, is primarily responsible for their ultralow thermal conductivities. Similar effects on thermal conductivity have been experimentally demonstrated in deposited superlattices, but never in naturally defined materials like RP phases. GIWAXS measurements reveal that higher n RP phase thin films are less orientationally controlled, and therefore possess apparently elevated thermal conductivities relative to single crystals of the same n.

Electron beams with sub-10 nm diameters can be used as nanoscale heaters or thermometers, enabling new techniques to probe heat transfer in nanostructures. However, since the electron beam diameter is often smaller than the energy carrier mean free path, using Fourier’s law will underpredict the (quasi-)temperature rise due to electron-beam induced heating. Here, we solve the Boltzmann transport equation (BTE) under the relaxation time approximation to quantify electron-beam induced heating in thin dielectric samples. In both thin and thick samples, the full BTE solution is in agreement with a simple “resistors in series model” that sums the ballistic resistance and the typical Fourier conduction resistance (considering the thin-film classical size effect). For nanoscale thermometry applications, the temperature rise under typical imaging conditions is predicted to remain negligibly small (<1 K), even when considering the dominant ballistic thermal resistance. Therefore, this BTE modeling indicates that nanoscale thermometry methods are not expected to display notable electron beam heating artifacts, even though the ballistic resistance is the dominant contibutor to the total thermal resistance in experiments using nanoscale heat sources.

Ref: Wehmeyer, J. of Appl. Phys. 126, 124306 (2019).
*Support from the Welch Foundation (Grant No. C-2023).

Superatoms are atomically-precise and chemically-stable clusters that can emulate atoms, but at the nanometer scale. Ordered arrangements of superatoms are called superatomic crystals (SACs), which have potential use as electronic, thermoelectric, and phononic materials. In this talk, I will describe experiments and molecular dynamics (MD) simulations performed on SACs built from C₆₀ fullerene, Co₆Se₈(PEt₃)₆, and Co₆Te₈(PEt₃)₆ superatoms to characterize their thermal transport properties.

Thermal conductivity measurements on the [C₆₀] and [Co₆Se₈(PEt₃)₆][C₆₀]₂ SACs show a decreasing thermal conductivity with increasing temperature, as it typical for a crystal. The thermal conductivity of the [Co₆Te₈(PEt₃)₆][C₆₀]₂ SAC, on the other hand, is lower and temperature-independent, as is typical for an amorphous solid. X-ray diffraction measurements suggest that the temperature dependencies result from the C₆₀s being either locked-in-place (crystalline behavior) or freely-rotating (amorphous behavior).

The MD-predicted thermal conductivities agree with the experimental trends and offer additional insight into the underlying mechanisms. The C₆₀ rotational dynamics and orientation are quantified using the root mean square displacements of the carbon atoms and the relative orientations of the C₆₀s. At low temperatures, the C₆₀s exhibit small rotations around equilibrium positions (i.e., librations). When the librating C60s are orientationally-ordered, as in the [C₆₀] and [Co₆Se₈(PEt₃)₆][C₆₀]2 SACs, thermal conductivity decreases with increasing temperature. When the librating C₆₀s are orientationally-disordered, however, as in the [Co₆Te₈(PEt₃)₆][C₆₀]2 SAC, thermal conductivity is temperature independent. At higher temperatures, where the C₆₀s in all three SACs freely-rotate and are thus dynamically disordered, thermal conductivity is temperature independent. The abrupt changes driven by the C₆₀ dynamics suggest that fullerene-based SACs can be designed to be thermal conductivity switches based on a variety of external stimuli.

Low-dimensional materials are the focus of intense research due to their unusual and tailorable properties, including their thermal transport characteristics. Examples of distinctive features include flexural acoustic phonons with quadratic dispersions and, in the case of graphene, selection rules that prohibit certain 3-phonon scattering processes. Often the bulk counterparts to these 2D materials are layered van der Waals materials, with the only difference between bulk and monolayer originating from weak inter-layer coupling. Thus, 2D materials can be expected to retain much of the bulk character, and vice versa. In this talk, I will present our work investigating how phonon transport evolves in van der Waals materials when going from 3D to 2D, using graphite/graphene as a case study. Focus is placed on identifying the features impacting thermal conductivity that may be shared with other layered materials. Our calculations are carried out using density functional theory in combination with a self-consistent solution of the phonon Boltzmann transport equation.

Unlike previous studies that compare the thermal properties of mono- and few-layer to bulk, in this work the transition from 3D to 2D is investigated by gradually increasing the inter-layer distance in graphite to form graphene. This approach provides a different perspective on how the phonon and scattering characteristics continuously evolve from bulk to monolayer. We find that changes in the phonon dispersion play a significant role in the lower thermal conductivity of graphite, versus graphene. Specifically, as the graphene sheets are brought closer together, the acoustic flexural phonons form low-energy optical phonons with lower velocities, density-of-states, and phonon occupation, which reduce thermal conductivity. Interestingly, our findings on the 3-phonon scattering properties indicate that the selection rules are mostly upheld as the atomic layers are brought together to form graphite. While the selection rules only strictly apply to graphene, in practice similar scattering behavior is displayed due to the weak inter-layer coupling. Lastly, we attempt to identify key features in the 3D-to-2D transition, that impact thermal conductivity, which may be shared with a broader class of van der Waals materials.

Environmental barrier coating (EBC) systems are applied to the surface of silicon-based composites exposed to high temperature combustion gas flow paths in gas turbine engines. They reduce the rate of composite oxidation, its volatilization by reactions with water vapor, and the temperature of the composite as their thermal conductivity decreases. Current EBC systems consist of a silicon bond coat covered by an ytter- bium disilicate (Yb₂O₃●2SiO₂: YbDS) permeation resistant and low silica activity barrier. When applied by atmospheric-plasma spray deposition, this outer layer contains 10–15% of a secondary ytterbium monosilicate (Yb₂O₃●SiO₂: YbMS) phase. YbDS has a coefficient of thermal expansion (CTE) of 4–5×10−6 °C⁻¹, similar to those of the silicon and the silicon-based composite, but a relatively high thermal conductivity of 5–7 W m⁻¹ K⁻¹. YbMS has a higher steam volatility resistance than YbDS, and it has a much lower thermal con- ductivity (∼2–2.5 W m⁻¹ K⁻¹) at ambient temperature compared to YbDS, but its higher, highly anisotropic CTE (3–11×10−6 °C⁻¹) results in channel cracking which reduces environmental protection. Here, we use a combination of scanning electron beam and laser-based thermoreflectance methods to spatially map the distribution of the silicate phases and thermal conductivity at ambient temperature in a Si-ytterbium disilicate EBC system exposed to thermal cycling in water vapor. We show that during thermal cycling, diffusion of silica from the thermally grown oxide on the Si bond coat surface to nearby YbMS regions transforms this phase to YbDS, thereby reducing the risk of thermomechanical coating failure but decreasing its effective thermal resistance. We also show that as silica is volatilized at the water vapor–ytterbium silicate interface, YbDS is transformed to YbMS, restoring some of the thermal protection of the coating system lost by its reduction in thickness and the YbMS to YbDS transformation near the bond coat.

Cellulose and cellulose nanomaterials (CNMs) can be used to produce light-weight foams and aerogels with heat transport properties that can be phonon-engineered for thermal management and thermal insulation applications^[1].
Cellulose possesses an intrinsic anisotropy of the thermal conductivity, which has been utilized to produce foams made with aligned CNM particles that exhibit highly anisotropic thermal conductivities. Anisotropic foams based on cellulose nanofibrils (CNF) made by ice-templating exhibit very low radial – perpendicular to the fibrils – thermal conductivity^[2,3], which suggests that CNM-based foams could reduce both the carbon footprint and the energy consumption in buildings.
Cellulose is a hygroscopic material, and it is thus essential to understand how the heat transfer mechanisms are influenced by water uptake and the relative humidity in air. We have investigated how the thermal conductivity of anisotropic CNM-based foams depends on moisture uptake and temperature, together with the diameter and degree of alignment of the fibrils, and the density and nanoporosity of the foams. Combining molecular simulations, thermal conductivity measurements, and moisture uptake experiments we have identified the importance of swelling in reducing the radial thermal conductivity. Tuning the surface properties of cellulose nanomaterials or increasing their degree of crystallinity and temperature has a significant effect on the moisture uptake which plays important role in governing the heat transfer within CNM foams at humid conditions. Furthermore, here we show potential routes to achieve very low thermal conductivity by tuning the gaseous and solid conduction contributions. Apart from the degree of the alignment and the sizes of pores, the CNM entanglement and the plethora of solid/solid and solid/air interfaces are crucial parameters to consider. To this end, enhancing the diffuse phonon scattering by tuning the length and number of interfaces seems to be the most efficient way to reduce the thermal conductivity below the superinsulating level in highly porous (>95%) CNM foams.

References
[1] V. Apostolopoulou-Kalkavoura, P. Munier, L. Bergström, Adv. Mater. 2020, 2001839, 1.
[2] B. Wicklein, A. Kocjan, G. Salazar-Alvarez, F. Carosio, G. Camino, M. Antonietti, L. Bergström, Nat. Nanotechnol. 2015, 10, 277.
[3] P. Munier, V. Apostolopoulou-Kalkavoura, M. Persson, L. Bergström, Cellulose 2019, 1.

Resonance hybridization (RH) can be described as the coupling between wavenumber-independent vibrons in a substructure and wavenumber-dependent modes (phonons) of the host medium. This mechanism showed remarkable consequences for thermoelectric applications such as the ~40-fold reduction of thermal conductivity (TC) of silicon (Si) membranes due to resonances induced by native oxide at membrane surfaces [1]. In particular, nanopillar resonators on silicon thin-films have been illustrated to exhibit unique subwavelength phonon properties at the nanoscale, resulting in strong TC reduction. Our recent study reported the first remarkable observation that the RH mechanism can enforce directional mode propagation, and therefore anisotropic TC at room temperature, in otherwise isotropic single-crystal thin films of materials with isotropic bond symmetries (e.g. cubic or hexagonal. In Si membranes with surface nanofins, TC is ~100% greater in the direction parallel to the fins than in the perpendicular direction [2]. Anisotropic in-plane TC has been observed or predicted in 2D and 3D materials due to direction-dependent bonding environments and anisotropic pore spacing, and in two to four monolayer thick silicene due to surface reconstruction, the effect decreasing with thickness. However, anisotropic in-plane TC in the Si membranes beneath nanofins is remarkable as it is the first time such an effect has been observed in isotropic semiconductors at room temperature. This method of inducing anisotropy is particularly valuable, since modern fabrication techniques allow for more precise control over the addition of surface resonators than embedding Bragg scattering mechanisms within the material. This study further suggested that hybridized modes within the base material of such structures, deemed nanophononic metamaterials (NPMs), may not retain their full crystalline character. In fact, RH potentially induces ``glass-like" phonon transport, rendering the host membrane or film a disordered "phonon-glass" for low frequency modes. Heat conduction mechanisms in crystals and glasses are fundamentally different. The structural periodicity of crystals allows propagating phonons (propagons) to become excited and carry heat. In glasses, a lack of periodicity breaks the quasi-particle nature of heat carriers so that heat is transported mainly by the coupling of vibrational modes (diffusons) as described by the Allen and Feldman (AF) model. In NPMs, some hybridized modes retain a propagating character while others become localized with ~0 group velocity and display glassy heat transport behavior. The presence of both types of modes within `simple' crystalline materials has not been adequately probed, and therefore further research is essential to describe the phenomenon. In this work, we use a combination of classical molecular dynamics, harmonic and anharmonic lattice dyanmics and the newly proposed lattice dynamics based quasi-harmonic Green-Kubo approach to provide a complete microscopic description of the transport processes of the hybridized modes. We aim to establish the extent to which established phonon physics models remain valid for these systems.

1. S. Neogi, et al., ACS nano, vol. 9,no. 4, pp. 3820–3828, (2015).
2. S. Neogi, S. and D.Donadio, Physical Review Applied, 14(2), p.024004 (2020).

Fluorine-rich polymers incorporated with aluminum particles are of interest to the energetics community due to their enhanced flame propagation characteristics. Recent investigations have shown that PVDF-Al demonstrates better ignition and sustained combustion when compared to similar fluoropolymer-metal composites such as Viton-Al and THV-Al [1]. Enhanced thermal conduction in PVDF-Al nanocomposites due to the Al nanoparticles is speculated to contribute to their superior energetic properties. My talk will focus on nanoscale thermal transport in PVDF-Al films and their influence on flame propagation. Using time-domain thermo-reflectance (TDTR), I present high resolution 2-dimensional thermal conductivity maps of thin PVDF-Al films. The PVDF-Al composite samples have varying volume fractions of Al. The TDTR maps provide spatially correlated thermal conductivity and heat capacity measurements across the polymer, metal particles, and heterogenous metal-polymer interfaces. I also report the results of TDTR measurements as a function of heating frequency. Heating frequency dependent measurements allow us to study how homogenous the transport properties of the composites are as a function of thermal penetration depth. My investigation provides important insights into the nanoscale thermal transport in Al/PVDF nanocomposite systems, which could be used to predict their flame propagation and combustion characteristics.
Ref:
[1] Wang, Haiyang, et al. "Comparison study of the ignition and combustion characteristics of directly-written Al/PVDF, Al/Viton and Al/THV composites." Combustion and Flame 201 (2019): 181-186.

In this talk we report on recent progress in modeling phonon transport in bulk and across interfaces at high temperatures beyond the harmonic approximation. In bulk crystalline materials the self-consistent phonon theory has been extended within a variational formalism to include thermal strain and phase transitions. The anharmonic hamiltonian parameters of the high-symmetry phase are first extracted from a set of total-energy and force calculations (from first-principles DFT calculations) in a supercell. Then a variational free energy is constructed which when minimized, produces the equilibrium structure and its lattice dynamical properties at any temperature, beyond the quasi harmonic approximation.
As for the interfacial thermal conductance, within the Landauer formalism, an expression is derived for the non-equilibrium thermal current of the anharmonic system based on the Green’s function formalism. This non-perturbative expression is made to satisfy the detailed balance and current conservation constraints. Examples illustrating both methodologies will be shown.

Recently, it has been shown computationally and experimentally that the inclusion of higher-order scattering among four phonons is crucial to accurately represent the phonon thermal conductivities of several strongly-bonded materials like boron arsenide and indium phosphide [1-3], as well as weakly-bonded solids like sodium chloride (NaCl) [4]. Here we show, from rigorous first-principles calculations, that four phonon scattering also critically affects the phonon lineshapes of NaCl and other weakly-bonded solids typically observed in inelastic neutron/x-ray scattering and Raman scattering experiments. These calculations are performed by building on the unified first-principles framework [4], which already includes the effects of the lowest-order three-phonon and higher-order four-phonon scattering processes on the thermal conductivity of materials, a many-body self-consistent anharmonic phonon renormalization step to address the possible ill-defined nature of certain phonon quasiparticles in weakly-bonded solids, and the determination of the stable lattice constants at finite temperatures, by explicitly evaluating and minimizing the fourth-order accurate free energy of the crystal. By performing these calculations over a broad range of temperatures, we show that the four-phonon scattering processes significantly broaden the phonon lineshapes compared to their three-phonon counterparts - particularly at high temperatures, and their inclusion into the calculations are pivotal to explain the experimental data.

This work has been supported by a start-up fund from the Indian Institute of Science, Bangalore, India.

[1] Navaneetha K. Ravichandran and D. A. Broido, Phys. Rev. X 10, 021063 (2020).
[2] Navaneetha K. Ravichandran and D. A. Broido, Nature Comm. 10, 827 (2019).
[3] Fei Tian, Bai Song, Xi Chen, Navaneetha K. Ravichandran et al., Science 361, 6402 (2018).
[4] Navaneetha K. Ravichandran and D. A. Broido, Phys. Rev. B 98 (8), 085205 (2018).

Density functional theory (DFT) is an ab-initio method that has been successfully used to predict the phonon dispersion and thermal conductivity of a variety of insulating and semiconducting materials. DFT-based calculations are attractive because, unlike calculations based on empirical potentials, they are free from fitting parameters. A DFT calculation is made computationally tractable by describing the complicated exchange-correlation (XC) electron interactions with an effective potential, called an XC functional. There is no a priori method of choosing an XC functional for a given material, so that choosing any single XC functional introduces uncertainty in any quantity derived from that calculation.

We present a framework for quantifying the uncertainty that results from the choice of XC functional in predictions of phonon properties and thermal conductivity that use DFT to calculate the atomic force constants. The framework uses the energy ensemble capabilities of the BEEF-vdW XC functional to predict ensembles of interatomic force constants, which are then used as inputs to lattice dynamics calculations and a solution of the Boltzmann transport equation. We apply the framework to isotopically-pure silicon and find that the ensembles bound property predictions (e.g., phonon dispersion, specific heat, thermal conductivity) from other XC functionals and experiments. We distinguish between properties in silicon that are correlated with the predicted thermal conductivity [e.g., the transverse acoustic branch sound speed (Pearson correlation coefficient, R2, of 0.89) and average Gruneisen parameter (R2 = 0.85)] and those that are not [e.g., longitudinal acoustic branch soundspeed (R2 = 0.23) and specific heat (R2 = 0.00)]. We find that differences in ensemble predictions of thermal conductivity are correlated with the behavior of transverse acoustic phonons with mean free paths between 100 and 300 nm. This framework systematically accounts for XC uncertainty in phonon calculations and should be used whenever it is suspected that the choice of XC functional is influencing physical interpretations.

Models and tools devoted to the theoretical modeling of heat transport in new materials and nanostructures are of very different natures. Strong interest for DFT based methods has raised to compute thermal properties of bulk material from scratch, just knowing the atom organization in a supercell. In this field, tools such as those provided by the Phonopy [1] environment developed by A. Togo have paved the way to the discovery of new compounds with optimal thermal properties. The main strength of such an approach is its intrinsic reliability that usually matches very well with the available experimental data [2]. Yet, such an approach also has drawbacks, the most prominent is related to computing resources needed to carry out calculations. In addition, this approach basically addresses bulk materials. Dealing directly with nanostructures is currently impossible with such methods, as they remain too large for the actual computing facilities and probably even for the next generation ones. Nevertheless, we can circumvent this limitation by the coupling of ab-initio outputs with mesoscale modelling tools such as Monte Carlo solution of the Boltzmann transport equation (BTE) for phonons. For the latter technique, MC simulation of BTE is often limited by the knowledge of phonon dispersion properties and lifetimes of any kind of material.
However, what is a weakness for one method is a force for the other. Thus, coupling both techniques make sense and should allow to overcome intrinsic limitations of DFT and MC approaches. Within the group, but also in collaboration with the “Centre de Nanosciences et Nanotechnologies” (Université Paris-Saclay), we work on this coupling. As a first challenge, we investigate how to combine full band modelling approach in K-space, given by DFT calculations. Furthermore, for MC model the scattering term of the BTE was treated in the frame of the relaxation time approximation, considering the Matthiessen rule and statistics on phonon populations to define relevant lifetimes. Considering DFT coupling, a direct access to the mode lifetime with respect to the K-grid discretization is now possible.
On this basis, we have modified our existing MC tool for phonon transport in order to take into account full band phonon properties and lifetimes given by the DFT calculations. Among the main improvements of this new ab-inito/MC model, there is the resolution of BTE along characteristic lines. With this method there is a rigorous balance between energy and momentum as we follow each phonon bundle sampled on the K-grid. This new technique is much more rigorous as this is no longer necessary to compute scattering probability of phonon to restore equilibrium. Results concerning those calculations have been recently published in Applied Physics Letters [3], demonstrating the ability of the method to handle phonon lifetimes extracted from DFT calculations.
In addition, for the first time, we extend the methodology to the coupling of phonon-phonon and electron-phonon interactions in doped Si, Ge and C compounds using DFT simulations that provide the relevant lifetimes. Application to such coupling are numerous, especially for materials where electron transport can hindered/increase heat transfer.

1 - https://atztogo.github.io/phonopy/

2 - A. Seko et al., Prediction of Low-Thermal-Conductivity Compounds with First-Principles Anharmonic Lattice-Dynamics Calculations and Bayesian Optimization, Phys. Rev. Lett. 115, 205901 (2015)

3 - L. Chaput et al., Ab initio based calculation of the thermal conductivity at the micron scale, Applied Physics Letters, vol. 112, (3), 033104 (2018)

Using detailed measurements of inelastic scattering of neutrons and x-rays, complemented with first-principles simulations of atomic dynamics, we investigate strongly anharmonic effects on phonons near or across phase transitions, and their impact on thermal transport. This presentation will focus on transitions ranging from soft-mode condensation to sublattice melting in the case of superionic compounds. In particular, we will highlight systems undergoing ferroelectric transitions or related Jahn-Teller distortions (e.g. SnS and SnSe across their high-temperature structural transition [1]), which strongly renormalize group velocities and scattering phase-space, as well as “phonon-liquid electron-crystal” systems, in which selective phonon modes collapse to relaxational dynamics, as in Cu or Ag-based superionics [2,3].

[1] T. Lanigan-Atkins*, S. Yang*, J. L. Niedziela, D. Bansal, A. F. May, A. A. Puretzky, J.Y.Y. Lin, D. Pajerowski, T. Hong, S. Chi, G. Ehlers, and O.Delaire, Nature Communications 11 (1), 1-9 (2020).
[2] J. Ding, J. L. Niedziela, D. Bansal, J. Wang, X. He, A. F. May, G. Ehlers, D. L. Abernathy, A. Said, A. Alatas, Y. Ren, G. Arya, and O. Delaire, PNAS 117 (8) 3930-3937 (2020).
[3] J. L. Niedziela, D. Bansal, A. F. May, J. Ding, T. Lanigan-Atkins, G. Ehlers, D. L. Abernathy, A. Said, and O. Delaire, Nature Physics, 15, 73–78 (2019).

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 μm 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.

A thermophotovoltaic (TPV) device converts infrared radiation from hot sources into electrical power via a specifically designed photovoltaic cell. In order to increase the generated electrical power, the cell can be coupled with an infrared emitter brought in its near field (< 4 µm at 700 K), where the emitter-cell radiative heat exchange can be enhanced by several orders of magnitude. This enhancement is well described theoretically but near-field TPV devices have been demonstrated experimentally only recently [1-3], unfortunately with very low electrical output power densities (<10 W.m^-2) and conversion efficiencies (<1 %).
In this work, we first characterize experimentally fully-functional InSb TPV cells [4] that operate below room temperature (77 K). Indium antimonide (InSb) is an attractive material for TPV cells since its very low energy band gap (~5 µm at 77 K) matches well the blackbody spectrum of infrared emitters. We then demonstrate strong electrical power photogeneration by placing a graphite spherical emitter in the near field of the InSb TPV cells. In order to maximize the electrical power, large temperature differences are investigated: the emitter can be heated up to temperatures higher than 1000 °C. Impacts of emitter material and cell design on the near-field radiative heat transfer and near-field TPV conversion are investigated as a function of emitter-cell distance down to a few tens of nanometers. We measure a 7-fold enhancement of the generated electrical power when bringing the emitter in the near field, leading to an electrical power as high as 7500 W.m^-2, three orders of magnitude higher than previous works [1-3], and near-field conversion efficiency larger than 14 % [5].

[1] A. Fiorino et al., Nature Nanotechnology (2018). [2] T. Inoue et al., Nano Letters (2019). G. Bhatt et al., Nature Com. (2020). [4] D. Cakiroglu et al. Solar Energy Materials and Solar Cells (2019). [5] C. Lucchesi, et al., arXiv:1912.09394 (2019)
Financial support by the French National Research Agency (ANR) under grant No. ANR-16-CE05-0013 and partial funding by the French "Investment for the Future" program (EquipEx EXTRA ANR-11-EQPX-0016 and IDEXLYON ANR-16-IDEX-0005) and by the Occitanie region are acknowledged.

Heat transfer in the extreme near field has shown recently an increased interest driven by the development of scanning probe techniques allowing to measure thermal transport across a nanoscale gap and across single molecule junctions. On the one hand, recent experimental studies reported giant heat flux transfer between gold surfaces separated by nanometer distances [1]. These results showed discrepancies with the conventional theory which are attributed to phonon tunneling neglected.in the standard picture. On the other hand, the thermal conductance of single molecule junctions has been recently determined experimentally [2,3] opening the door to confirm the high thermoelectric efficiency displayed by these junctions. All these experiments call for models capable of describing accurately phonon tunneling and phonon thermal transport across molecules.

In this contribution, we model heat transfer across nanometer gaps and across single-molecule junctions using a combination of molecular dynamics and ab-initio simulations. First; we demonstrate, using a combination of ab-initio lattice dynamics [4] and molecular dynamics simulations, that phonons dominate heat transport at nanometer distances, even in the presence of molecules in the gap [5]. This conclusion is shown to hold for nanogaps separating different types of materials. The computed thermal conductance turn out to be several orders of magnitude higher than the predictions of acoustic continuous models. We show that these discrepancies originate from the contribution of intermediate phonon frequencies which are not accurately described by acoustic models. Moreover, we highlight the leading role played by phonon scattering as compared with electron tunneling and electron/phonon processes at interfaces [6].
The second issue to be be discussed concerns thermal transfer across molecular junctions [7]. We consider junctions made of OPE3 derivatives, as these molecules are commonly synthesized by experimentalists and already extensively studied for electronic transport.
The phonon thermal conductance of the corresponding gold-molecule-gold junctions are computed here using a combination of molecular dynamics and ab-initio calculations. In particular, we parametrize the interaction between the gold surface and the OPE3 derivative using an ab-initio adiabatic connection scheme [8]. For Au-OPE3-Au junctions, the computed thermal conductance obtained is in good agreement with the experimental value [3]. Next, we characterize the phonon thermal conductance of cross-linked molecules junctions, as a function of their length and the nature of the cross-linker. In particular, we show that constrained cross-linked molecules display relatively low thermal conductance, which makes them promising for thermoelectric applications. These low conductance levels are interpreted based on phonon transmission calculations.

References:
[1] K. Kloppstech et al., Nat. Comm. 8 (2017) 144505
[2] Cui et al., Nature 572 (2019) 628
[3] Nico Mosso et al., Nano Letters, 19 (2019) 7614–7622.
[4] A. Alkurdi, S. Pailhès and S. Merabia, Appl. Phys. Lett. 111 (2017) 093101
[5] A. Alkurdi, C. Adessi, F. Tabatabaei, S. Li, K. Termentzidis and S. Merabia, Int. Jour. Heat Mass Transf. 158 (2020) 119963
[6] J. Lombard, F. Detcheverry and S. Merabia, J. Phys. Cond. Mat. 27, 015007 (2015)
[7] F. Tabatabaei, T. Niehaus and S. Merabia, in preparation
[8] Zou, W., Kalescky, R., Kraka, E., & Cremer, D. The Journal of chemical physics, 137 (2012), 084114.

Acknowledgment: This work has been financially supported by the iMUST Labex project ATTSSEM and by the EU project EFINED (Project No. 76685).

Understanding radiative heat transfer in nanoscale gaps and devices is of considerable interest for creating novel energy conversion devices. In this talk, I will first describe ongoing efforts in our group to experimentally elucidate nanoscale radiative heat transfer. Specifically, I will present our recent experimental work where we have addressed the following questions: Can existing theories accurately describe radiative heat transfer in single nanometer sized gaps? Can radiative thermal conductances that are orders of magnitude larger than those between blackbodies be achieved? In order to address these questions we have developed a variety of instrumentation including novel nanopositioning platforms and microdevices, which will also be described. Further, I will discuss possible applications of near-field thermal radiation for energy conversion and photonic cooling. Finally, I will briefly outline how these technical advances can be leveraged for future investigations of nanoscale heat transport and near-field thermophotovoltaic energy conversion.

Thermophotovoltaic (TPV) systems are based on solid-state energy conversion technique and involve a hot emitter to radiate photons that excite electron hole pairs in a photovoltaic (PV) cell. Recent TPV studies in the far-field demonstrated impressively high efficiencies, but at low power densities. Several theoretical proposals have shown that dramatic enhancement in power density can be achieved by reducing the gap to tens of nanometers. However, such demonstrations have been elusive due to the challenge of maintaining such nano-separations between a hot emitter (~1000 K) and a cold cell (~300 K). Here, we will describe how high resolution calorimetry and a custom-built nano-positioning platform have made possible the first demonstration of ~40-fold increase in the power output, compared to far field, when the gap between the PV cell and thermal emitter at 650 K is reduced from ~10 µm to 60 nm. Further, we will discuss our recent work aimed at achieving larger power densities and efficiencies by developing thermal emitters that can operate at high temperatures (~1300 K) that may lead to power densities exceeding 1 W/cm².

We investigate the near field radiative heat transfer between two slabs made of high temperature superconductors, namely, YBCO. We find a three orders of magnitude decrease in the total heat flux when the temperature of at least one of the slabs is below the critical temperature. We interpret this suppression as a consequence of the detailed balance of the charge carrier density: the increased density of the condensed electron pairs is accompanied by the simultaneous decrease of the density of the unpaired electrons. This effectively increases the conductivity of the superconductor by suddenly suppressing the electron damping mechanisms. Consequently, we observe a reduction of the contribution of the p-polarized waves compared to the s-polarized ones to the spectral heat flux.
The difference between using free-standing films and bulk meda is also discussed. Thin-films allow for more modes that modify the radiative heat transfer making it larger than the case of bulk-materials.

Two-phase cooling is one of the most powerful approaches for removal of large amounts of thermal energy, which is required in a number of applications ranging from high power electronics to nuclear reactor safety. The analysis of the two-phase cooling processes is typically based on continuum modeling of heat and mass transport equations. While the constitutive laws for these transport processes in the bulk are well established, the quantitative, physics-based description of the transport processes at the liquid-vapor interface is still a challenge. A typical assumption here is that the interface is in a local thermal equilibrium. While such assumption might be appropriate for modeling systems with large feature sizes, is might lead to inaccurate predictions for micro/nanoscale systems, particularly for highly driven systems.

In our presentation I will describe an approach that enables high fidelity, predictive modeling of the two-phase flow systems based a continuum modeling framework informed via physics-based and molecular dynamics simulations verified “interfacial constitutive laws”. These interfacial laws provide quantitatively the rate of phase change in terms of local thermodynamic variables, such as temperature and pressure, and allow for non-equilibrium vapor pressure and temperature discontinuity at the interface. We demonstrated the fidelity of this approach in the case of planar heat pipe geometry, even in the limit of nanoscale systems and highly transient heating processes. We will discuss avenues and challenges for the development of this approach for more complex systems and processes.

Most previous studies of interfacial heat transport focus on transmission of vibration modes across the interfaces. It is generally believed that heat is predominantly carried by vibration modes, even for metal/dielectric interfaces in which electrons in metals could be coupled remotely to the polar vibration modes in dielectrics due to the long-range nature of Coulomb force. In this talk, we will discuss our recent attempts to experimentally measure the interfacial heat transfer due to this remote coupling. In the first work, we measured the heat transport across van-der-Waals graphene interfaces, due to remote coupling of induced charge couriers in graphene and polar phonons in oxide substrate, by a newly developed technique called voltage-modulated thermoreflectance (VMTR). We find that the thermal conductance of SiO₂/graphene/SiO₂ interfaces increases by up to ≈1.5 MW m^-2 K^-1 under electrostatic fields of <0.2 V nm^-1, and hence establish an upper limit for direct heat transfer from graphene to SiO₂ substrates via remote interfacial phonon (RIP) scattering. In the second work, 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. Our work could provide a new pathway to enhance and control heat transport across interfaces and heterojunctions, through remote coupling of charge carriers and polar vibration modes.

The interface conductance between two strongly bonded materials is determined by the vibrational stiffness of the constituent materials. In theory, because diamond is very vibrationally stiff, interfaces between diamond and other materials should have a high interface conductance. In practice however, the interface conductance values observed in diamond systems are small. Observed conductance values in diamond systems range from 30-60 MW/m²-K for Pb and Au on diamond to ≈ 150 MW/m²-K for Al or Pt on diamond [1]. Low thermal boundary conductance in diamond systems is an obstacle to efforts to improve thermal performance in nitride electronic devices by using diamond substrates. In this work, we report time-domain thermoreflectance (TDTR) measurements of the interface conductance between nitride metals and diamond. Nitride metals are amongst the most acoustically stiff metals and are therefore good candidates for investigating the upper limits to the conductance of metal/diamond interfaces. We find TiN/diamond has a large interface conductance of ≈ 400 MW/m²-K, much larger than typical in diamond systems.

[1] Hohensee, G., Wilson, R. & Cahill, D. Thermal conductance of metal–diamond interfaces at high pressure. Nat Communications 6, 6578 (2015).

Phonons play important roles in quantum information science (QIS) devices through heat dissipation and transduction of signals. To better engineer such devices, it is critical to understand the phonon transport mechanism, especially how they interact with the interfaces. While defects, including vacancies, impurities, oxidations etc., naturally appear in all materials and their interfaces, recent developments in techniques such as molecular beam epitaxy also make possible the formation of atomically perfect interfaces between metals and semiconductors. Therefore, considering interfaces from both perspectives is necessary towards a deeper understanding of phonon transport in realistic devices. However, so far the detailed phonon scattering mechanism at such interfaces is not well understood.

In this study, we applied two complementary classical molecular dynamics (MD) techniques to elucidate an understanding of interfacial phonon transport mechanisms: 1) non-equilibrium MD to identify the overall transport mechanism of the entire phonon spectrum, and 2) wave packet formalism to reveal the individual scattering behavior of a single phonon at an interface. With NEMD we identified the effects of atomic vacancies at a Cu(100)/Si(100) interface, finding that interfacial phonon conductance can be increased by up to 76% with just a 3% concentration of vacancies. Further analysis based on phonon normal modes and density of states showed that the increase is primarily due to high-frequency phonons through enhanced inelastic scattering. On the other end, we revealed how individual phonons interact with atomically perfect Si(111)/Al(111) interfaces using wave packet simulations, considering 1×1, sqrt(3)×sqrt(3) and 7×7 Si surface reconstructions. It was found that interfacial scatterings differ depending on the interfacial bonding, phonon mode and frequency. Our study provides insights on manipulating interfacial phonon transport through defect and atomically precise engineering, and interpreting measurements involving phonon signals in QIS.

Colloidal, two-dimensional nanocrystals (NCs) known as nanoplatelets (NPLs) exhibit many advantages over more traditional spherical NCs while maintaining the benefits of colloidal processability and bandgap tunability. Namely, the lack of size dispersion in the single quantum-confined dimension bestows NPLs with exceptionally narrow emission linewidths that bolster their use in optoelectronic applications such as light-emitting diodes and lasers. However, these applications operate under high-fluence conditions that raise a fundamental concern regarding NC stability and the effect of such excitation conditions on desired properties such as optical bandgap. NPLs in particular may exhibit reduced thermodynamic stability even beyond that of other NCs, owing to the presence of high surface energy corners and edges. While some movement toward understanding impacts of heat on electronic properties of NCs is underway, the influence of NC morphology, and especially dimensionality, to the material stability remain largely unexplored.

To this end, here we employ time-resolved x-ray diffraction at the Advanced Photon Source to study the effect of heating on anisotropic NCs. NPLs are excited using a picosecond pulsed laser and then probed using synchrotron x-rays, which reveals both thermal expansion and transient disordering following photoexcitation. Photoexcitation induces greater out-of-plane than in-plane disordering, with thicker NPLs producing more anisotropic responses. Recovery dynamics suggest that out-of-plane cooling slightly outpaces in-plane cooling, with recrystallization occurring on indistinguishable time scales. This response to high-fluence excitation differs significantly from the response of spherical NCs, which both respond and recover isotropically to photoexcitation and recover on faster timescales. These findings offer significant implications for optoelectronic devices, where the ability to operate at high excitation densities without loss of efficiency is advantageous. Our results suggest that both the composition and structure of the NC must be taken into account when considering crystal lattice stability and optical performance, and that NPLs are likely to exhibit superior performance over traditional, zero-dimensional nanocrystals.