Search

Nuclear spin hyperpolarization provides a promising route to overcome the challenges imposed by the limited sensitivity of nuclear magnetic resonance. Here we demonstrate that dissolution of spin-polarized pentacene-doped naphthalene crystals enables transfer of polarization to target molecules via intermolecular cross relaxation at room temperature and moderate magnetic fields (1.45$\,$T). This makes it possible to exploit the high spin polarization of optically polarized crystals while mitigating the challenges of its transfer to external nuclei, particularly of the large distances and prohibitively weak coupling between source and target nuclei across solid-solid or solid-liquid interfaces. With this method, here we inject the highly polarized mixture into a benchtop NMR spectrometer and observe the polarization dynamics for target $^1$H nuclei. Although the spectra are radiation damped due to the high naphthalene magnetization, we describe a procedure to process the data in order to obtain more conventional NMR spectra, and extract the target nuclei polarization. With the entire process occurring on a timescale of one minute, we observe NMR signals enhanced by factors between -200 and -1730 at 1.45$\,$T for a range of small molecules., Comment: 8 pages, 4 figures

Recent experimental results have shown that vibro-fluidized, binary granular materials exhibit Rayleigh-Taylor-like instabilities that manifest themselves in rising plumes, rising bubbles and the sinking and splitting of granular droplets. This work explores the physics behind the splitting of a granular droplet that is composed of smaller and denser particles in a bed of larger and lighter particles. During its sinking motion, a granular droplet undergoes a series of binary splits resembling the fragmentation of a liquid droplet falling in a miscible fluid. However, different physical mechanisms cause a granular droplet to split. By applying particle-image-velocimetry and numerical simulations, we demonstrate that the droplet of high-density particles causes the formation of an immobilized zone underneath the droplet. This zone obstructs the downwards motion of the droplet and causes the droplet to spread and ultimately to split. The resulting fragments sink at inclined trajectories around the immobilized zone until another splitting event is initiated. The occurrence of consecutive splitting events is explained by the re-formation of an immobilized zone underneath the droplet fragments. Our investigations identified three requirements for a granular droplet to split: 1) frictional inter-particle contacts, 2) a higher density of the particles composing the granular droplet compared to the bulk particles and 3) and a minimal granular droplet diameter.

We report a new lift force model for intruders in dense, granular shear flows. Our derivation is based on the thermal buoyancy model of Trujillo & Hermann[L. Trujillo and H. J. Herrmann, Physica A 330, 519 (2003).], but takes into account both granular temperature and pressure differences in the derivation of the net buoyancy force acting on the intruder. In a second step the model is extended to take into account also density differences between the intruder and the bed particles. The model predicts very well the rising and sinking of intruders, the lift force acting on intruders as determined by discrete element model (DEM) simulations and the neutral-buoyancy limit of intruders in shear flows. Phenomenologically, we observe a cooling upon the introduction of an intruder into the system. This cooling effect increases with intruder size and explains the sinking of large intruders. On the other hand, the introduction of small to mid-sized intruders, i.e. up to 4 times the bed particle size, leads to a reduction in the granular pressure compared to the hydrostatic pressure, which in turn causes the rising of small to mid-sized intruders., Comment: 13 pages

MXenes are two-dimensional materials with a rich set of remarkable chemical and electromagnetic properties, the latter including saturable absorption and intense surface plasmon resonances. To fully harness the functionality of MXenes for applications in optics, electronics and sensing, it is important to understand the interaction of light with MXenes on atomic and femtosecond dimensions. Here, we use ultrafast electron diffraction and high-resolution electron microscopy to investigate the laser-induced structural dynamics of Ti3C2Tx nanosheets. We find an exceptionally fast lattice response with an electron-phonon coupling time of 230 femtoseconds. Repetitive femtosecond laser excitation transforms Ti3C2Tx through a structural transition into a metamaterial with deeply sub-wavelength nanoripples that are aligned with the laser polarization. By a further laser illumination, the material is reversibly photo-switchable between a flat and rippled morphology. The resulting nanostructured MXene metamaterial with directional nanoripples is expected to exhibit an anisotropic optical and electronic response as well as an enhanced chemical activity that can be switched on and off by light.

We compute the leading corrections to the differential cross section for top-pair production via gluon fusion due to dimension-six operators at leading order in QCD. The Standard Model fields are assumed to couple only weakly to the hypothetical new sector. A systematic approach then suggests treating single insertions of the operator class containing gluon field strength tensors on the same footing as expli\-citly loop suppressed contributions from four-fermion operators. This is in particular the case for the chromomagnetic operator $Q_{(uG)}$ and the purely bosonic operators $Q_{(G)}$ and $Q_{(\varphi G)}$. All leading order dimension-six contributions are consequently suppressed with a loop factor $1/16\pi^2$.

Germanium telluride (GeTe), a phase-change material, is known to exhibit four different structural phases: three at room temperature (one amorphous and two crystalline, $\alpha$ and $\gamma$) and one at high temperature (crystalline $\beta$). Because transitions between the amorphous and crystalline phases lead to significant changes in material properties (e.g., refractive index and resistivity), GeTe has been investigated as a phase-change material for photonics, thermoelectrics, ferroelectrics, and spintronics. Consequently, the temperature-dependent phase transitions in GeTe have been studied for bulk and thin-film GeTe, both fabricated by sputtering. Colloidal synthesis of nanoparticles offers a more flexible fabrication approach for amorphous and crystalline GeTe. These nanoparticles are known to exhibit size-dependent properties, such as an increased crystallization temperature for the amorphous-to-$\alpha$ transition in sub-10\,nm GeTe particles. The $\alpha$-to-$\beta$ phase transition is also expected to vary with size, but this effect has not yet been investigated for GeTe. Here, we report time-resolved X-ray diffraction of GeTe nanoparticles with different diameters and from different synthetic protocols. We observe a non-volatile amorphous-to-$\alpha$ transition between 210$^{\circ}$C and 240$^{\circ}$C and a volatile $\alpha$-to-$\beta$ transition between 370$^{\circ}$C and 420$^{\circ}$C. The latter transition was reversible and repeatable. While the transition temperatures are shifted relative to the values known for bulk GeTe, the nanoparticle-based samples still exhibit the same structural phases reported for sputtered GeTe. Thus, colloidal GeTe maintains the same general phase behavior as bulk GeTe while allowing for more flexible and accessible fabrication. Therefore, nanoparticle-based GeTe films show great potential for applications, such as in active photonics.

An external load on a particle packing is distributed internally through a heterogeneous network of particle contacts. This contact force distribution determines the stability of the particle packing and the resulting structure. Here, we investigate the homogeneity of the contact force distribution in packings of highly non-convex particles both in two-dimensional (2D) and three-dimensional (3D) packings. A newly developed discrete element method is used to model packings of non-convex particles of varying sphericity. Our results establish that in 3D packings the distribution of the contact forces in the normal direction becomes increasingly heterogeneous with decreasing particle sphericity. However, in 2D packings the contact force distribution is independent of particle sphericity, indicating that results obtained in 2D packings cannot be extrapolated readily to 3D packings. Radial distribution functions (RDFs) show that the crystallinity in 3D packings decreases with decreasing particle sphericity. We link the decreasing homogeneity of the contact force distributions to the decreasing crystallinity of 3D packings. These findings are complementary to the previously observed link between the heterogeneity of the contact force distribution and a decreasing packing crystallinity due to an increasing polydispersity of spherical particles., Comment: 28 pages, 13 figures. This version of the manuscript has been accepted for publication in Physical Review E

Certifying the robustness of neural networks against adversarial attacks is essential to their reliable adoption in safety-critical systems such as autonomous driving and medical diagnosis. Unfortunately, state-of-the-art verifiers either do not scale to bigger networks or are too imprecise to prove robustness, limiting their practical adoption. In this work, we introduce GPUPoly, a scalable verifier that can prove the robustness of significantly larger deep neural networks than previously possible. The key technical insight behind GPUPoly is the design of custom, sound polyhedra algorithms for neural network verification on a GPU. Our algorithms leverage the available GPU parallelism and inherent sparsity of the underlying verification task. GPUPoly scales to large networks: for example, it can prove the robustness of a 1M neuron, 34-layer deep residual network in approximately 34.5 ms. We believe GPUPoly is a promising step towards practical verification of real-world neural networks., Comment: M\"uller and Serre contributed equally to this work

In this paper we propose a novel regularization strategy for the local discontinuous Galerkin method to solve the Hamilton-Jacobi equation in the context of level-set reinitialization. The novel regularization idea works in analogy to shock-capturing schemes for discontinuous Galerkin methods, which are based on finite volume sub-cells. In this spirit, the local discontinuous Galerkin method is combined with an upwind/downwind finite volume sub-cell discretization, which is applied in areas of low regularity. To ensure the applicability on unstructured meshes, the finite volume discretization is based on a least squares approach., Comment: 31 pages, 6 figures

Traditional compilers operate on a single generic intermediate representation (IR). These IRs are usually low-level and close to machine instructions. As a result, optimizations relying on domain-specific information are either not possible or require complex analysis to recover the missing information. In contrast, multi-level rewriting instantiates a hierarchy of dialects (IRs), lowers programs level-by-level, and performs code transformations at the most suitable level. We demonstrate the effectiveness of this approach for the weather and climate domain. In particular, we develop a prototype compiler and design stencil- and GPU-specific dialects based on a set of newly introduced design principles. We find that two domain-specific optimizations (500 lines of code) realized on top of LLVM's extensible MLIR compiler infrastructure suffice to outperform state-of-the-art solutions. In essence, multi-level rewriting promises to herald the age of specialized compilers composed from domain- and target-specific dialects implemented on top of a shared infrastructure., Comment: 12 pages, 16 figures

In this paper we present a technique for constructing robust solvers for stiff algebraic source terms, such as those typically used for modelling relaxation processes in hyperbolic systems of partial differential equations describing two-phase flows, namely models of the Baer-Nunziato family. The method is based on an exponential integrator which employs an approximate linearised source term operator that is constructed in such a way that one can compute solutions to the linearised equations avoiding any delicate matrix inversion operations.