Engineering Clock Transitions in Molecular Lanthanide Complexes.

Molecular lanthanide (Ln) complexes are promising candidates for the development of next-generation quantum technologies. High-symmetry structures incorporating integer spin Ln ions can give rise to well-isolated crystal field quasi-doublet ground states, i.e., quantum two-level systems that may serve as the basis for magnetic qubits. Recent work has shown that symmetry lowering of the coordination environment around the Ln ion can produce an avoided crossing or clock transition within the ground doublet, leading to significantly enhanced coherence. Here, we employ single-crystal high-frequency electron paramagnetic resonance spectroscopy and high-level ab initio calculations to carry out a detailed investigation of the nine-coordinate complexes, [HoIIIL1L2], where L1 = 1,4,7,10-tetrakis(2-pyridylmethyl)-1,4,7,10-tetraaza-cyclododecane and L2 = F- (1) or [MeCN]0 (2). The pseudo-4-fold symmetry imposed by the neutral organic ligand scaffold (L1) and the apical anionic fluoride ion generates a strong axial anisotropy with an mJ = ±8 ground-state quasi-doublet in 1, where mJ denotes the projection of the J = 8 spin-orbital moment onto the ∼C4 axis. Meanwhile, off-diagonal crystal field interactions give rise to a giant 116.4 ± 1.0 GHz clock transition within this doublet. We then demonstrate targeted crystal field engineering of the clock transition by replacing F- with neutral MeCN (2), resulting in an increase in the clock transition frequency by a factor of 2.2. The experimental results are in broad agreement with quantum chemical calculations. This tunability is highly desirable because decoherence caused by second-order sensitivity to magnetic noise scales inversely with the clock transition frequency.

Search for Toroidal Ground State and Magnetoelectric Effects in Molecular Spin Triangles with Antiferromagnetic Exchange.

Using first-principles methods and spin models, we investigate the magnetic properties of transition-metal trimers Cr3 and Cu3. We calculate exchange coupling constants and zero-field splitting parameters using density functional theory and, with these parameters, determine the ground spin state as well as thermodynamic properties via spin models. Results for Cr3 indicate uniaxial magnetic anisotropy with a magnetic easy axis aligned along the 3-fold rotational symmetry axis and a mostly isotropic exchange interaction. The Cu3 molecule lacks rotational symmetry and our results show strong antisymmetric interactions for three distinct exchange couplings within the molecule. We are able to reproduce experimental findings on magnetic susceptibility and magnetization of Cr3 with the first-principles spin-Hamiltonian parameters. Our results show no presence of a toroidal ordering of spins for Cr3 and a finite toroidal moment for Cu3 in the ground state. We apply an external electric field up to 0.08 V/Å to each system to reveal the field dependence of exchange coupling as magnetoelectric effects. Finally, we scan the parameter space of a spin Hamiltonian to gain insights into which parameters would lead to a sizable toroidal moment in such systems.

High-Performance Acceleration of 2-D and 3-D CNNs on FPGAs Using Static Block Floating Point.

Over the past few years, 2-D convolutional neural networks (CNNs) have demonstrated their great success in a wide range of 2-D computer vision applications, such as image classification and object detection. At the same time, 3-D CNNs, as a variant of 2-D CNNs, have shown their excellent ability to analyze 3-D data, such as video and geometric data. However, the heavy algorithmic complexity of 2-D and 3-D CNNs imposes a substantial overhead over the speed of these networks, which limits their deployment in real-life applications. Although various domain-specific accelerators have been proposed to address this challenge, most of them only focus on accelerating 2-D CNNs, without considering their computational efficiency on 3-D CNNs. In this article, we propose a unified hardware architecture to accelerate both 2-D and 3-D CNNs with high hardware efficiency. Our experiments demonstrate that the proposed accelerator can achieve up to 92.4% and 85.2% multiply-accumulate efficiency on 2-D and 3-D CNNs, respectively. To improve the hardware performance, we propose a hardware-friendly quantization approach called static block floating point (BFP), which eliminates the frequent representation conversions required in traditional dynamic BFP arithmetic. Comparing with the integer linear quantization using zero-point, the static BFP quantization can decrease the logic resource consumption of the convolutional kernel design by nearly 50% on a field-programmable gate array (FPGA). Without time-consuming retraining, the proposed static BFP quantization is able to quantize the precision to 8-bit mantissa with negligible accuracy loss. As different CNNs on our reconfigurable system require different hardware and software parameters to achieve optimal hardware performance and accuracy, we also propose an automatic tool for parameter optimization. Based on our hardware design and optimization, we demonstrate that the proposed accelerator can achieve 3.8-5.6 times higher energy efficiency than graphics processing unit (GPU) implementation. Comparing with the state-of-the-art FPGA-based accelerators, our design achieves higher generality and up to 1.4-2.2 times higher resource efficiency on both 2-D and 3-D CNNs.

Design of Fully Spectral CNNs for Efficient FPGA-Based Acceleration.

Computing convolutional layers in the frequency domain using fast Fourier transformation (FFT) has been demonstrated to be effective in reducing the computational complexity of convolutional neural networks (CNNs). Nevertheless, the main challenge of this approach lies in the frequent and repeated transformations between the spatial and frequency domains due to the absence of nonlinear functions in the spectral domain, as such it makes the benefit less attractive for low-latency inference, especially on embedded platforms. To overcome the drawbacks in the existing FFT-based convolution, we propose a fully spectral CNN using a novel spectral-domain adaptive rectified linear unit (ReLU) layer, which completely removes the compute-intensive transformations between the spatial and frequency domains within the network. The proposed fully spectral CNNs maintain the nonlinearity of the spatial CNNs while taking into account the hardware efficiency. We then propose a deeply customized and compute-efficient hardware architecture to accelerate the fully spectral CNN inference on field programmable gate array (FPGA). Different hardware optimizations, such as spectral-domain intralayer and interlayer pipeline techniques, are introduced to further improve the performance of throughput. To achieve a load-balanced pipeline, a design space exploration (DSE) framework is proposed to optimize the resource allocation between hardware modules according to the resource constraints. On an Intel's Arria 10 SX160 FPGA, our optimized accelerator achieves a throughput of 204 Gop/s with 80% of compute efficiency. Compared with the state-of-the-art spatial and FFT-based implementations on the same device, our accelerator is 4 ×  âˆ¼  6.6 × and 3.0TEXPRESERVE3  âˆ¼  4.4 × faster while maintaining a similar level of accuracy across different benchmark datasets.

Near-Room-Temperature Magnetoelectric Coupling via Spin Crossover in an Iron(II) Complex.

Magnetoelectric coupling is achieved near room temperature in a spin crossover FeII molecule-based compound, [Fe(1bpp)2 ](BF4 )2 . Large atomic displacements resulting from Jahn-Teller distortions induce a change in the molecule dipole moment when switching between high-spin and low-spin states leading to a step-wise change in the electric polarization and dielectric constant. For temperatures in the region of bistability, the changes in magnetic and electrical properties are induced with a remarkably low magnetic field of 3 T. This result represents a successful expansion of magnetoelectric spin crossovers towards ambient conditions. Moreover, the observed 0.3-0.4 mC m-2 changes in the H-induced electric polarization suggest that the high strength of the coupling obtained via this route is accessible not just at cryogenic temperatures but also near room temperature, a feature that is especially appealing in the light of practical applications.

Dipole Switching by Intramolecular Electron Transfer in Single-Molecule Magnetic Complex [Mn12O12(O2CR)16(H2O)4].

We study intramolecular electron transfer in the single-molecule magnetic complex [Mn12O12(O2CR)16 (H2O)4] for R = -H, -CH3, -CHCl2, -C6H5, and -C6H4F ligands as a mechanism for switching of the molecular dipole moment. Energetics is obtained using the density functional theory (DFT) with onsite Coulomb energy correction (DFT + U). Lattice distortions are found to be critical for localizing an extra electron on one of the easy sites on the outer ring in which localized states can be stabilized. We find that the lowest-energy path for charge transfer is for the electron to go through the center via superexchange-mediated tunneling. The energy barrier for such a path ranges from 0.4 to 54 meV depending on the ligands and the isomeric form of the complex. The electric field strength needed to move the charge from one end to the other, thus reversing the dipole moment, is 0.01-0.04 V/Å.

First-principles calculation of gate-tunable ferromagnetism in magic-angle twisted bilayer graphene under pressure.

Magic-angle twisted bilayer graphene (MATBG) is notable as a highly tunable platform for investigating strongly correlated phenomena such as unconventional superconductivity and quantum spin liquids, due to easy control of doping level through gating and sensitive dependence of the magic angle on hydrostatic pressure. Experimental observations of correlated insulating states, unconventional superconductivity and ferromagnetism in MATBG indicate that this system exhibits rich exotic phases. In this work, using density functional theory calculations in conjunction with the effective screening medium method, we find the MATBG under pressure at a twisting angle of 2.88∘and simulate how its electronic states evolve when doping level and electric field perpendicular to plane are tuned by gating. Our calculations show that, at doping levels between two electrons and four holes per moiré unit cell, a ferromagnetic (FM) solution with spin density localized at AA stacking sites is lower in energy than the nonmagnetic solution. The magnetic moment of this FM state decreases with both electron and hole doping and vanishes at four electrons/holes doped per moiré unit cell. Hybridization between the flat bands at the Fermi level and the surrounding dispersive bands can take place at finite doping. On increasing the out-of-plane electric field at zero doping, a transition from the FM state to the nonmagnetic one is seen. An investigation of impurity effects shows that both absorption ofO2molecules and occurrence of Stone-Wales impurities suppress the FM state, and the mechanisms are understood from our calculations. We also analyze the interlayer bonding character due to flat bands via Wannier functions. Finally, we report trivial band topology of the flat bands in the FM state at a certain doping level.

Toward Full-Stack Acceleration of Deep Convolutional Neural Networks on FPGAs.

Due to the huge success and rapid development of convolutional neural networks (CNNs), there is a growing demand for hardware accelerators that accommodate a variety of CNNs to improve their inference latency and energy efficiency, in order to enable their deployment in real-time applications. Among popular platforms, field-programmable gate arrays (FPGAs) have been widely adopted for CNN acceleration because of their capability to provide superior energy efficiency and low-latency processing, while supporting high reconfigurability, making them favorable for accelerating rapidly evolving CNN algorithms. This article introduces a highly customized streaming hardware architecture that focuses on improving the compute efficiency for streaming applications by providing full-stack acceleration of CNNs on FPGAs. The proposed accelerator maps most computational functions, that is, convolutional and deconvolutional layers into a singular unified module, and implements the residual and concatenative connections between the functions with high efficiency, to support the inference of mainstream CNNs with different topologies. This architecture is further optimized through exploiting different levels of parallelism, layer fusion, and fully leveraging digital signal processing blocks (DSPs). The proposed accelerator has been implemented on Intel's Arria 10 GX1150 hardware and evaluated with a wide range of benchmark models. The results demonstrate a high performance of over 1.3 TOP/s of throughput, up to 97% of compute [multiply-accumulate (MAC)] efficiency, which outperforms the state-of-the-art FPGA accelerators.

Tailoring electrocatalytic activity of in situ crafted perovskite oxide nanocrystals via size and dopant control.

Perovskite oxides (ABO3) have been widely recognized as a class of promising noble-metal-free electrocatalysts due to their unique compositional flexibility and structural stability. Surprisingly, investigation into their size-dependent electrocatalytic properties, in particular barium titanate (BaTiO3), has been comparatively few and limited in scope. Herein, we report the scrutiny of size- and dopant-dependent oxygen reduction reaction (ORR) activities of an array of judiciously designed pristine BaTiO3 and doped BaTiO3 (i.e., La- and Co-doped) nanoparticles (NPs). Specifically, a robust nanoreactor strategy, based on amphiphilic star-like diblock copolymers, is employed to synthesize a set of hydrophobic polymer-ligated uniform BaTiO3 NPs of different sizes (≤20 nm) and controlled compositions. Quite intriguingly, the ORR activities are found to progressively decrease with the increasing size of BaTiO3 NPs. Notably, La- and Co-doped BaTiO3 NPs display markedly improved ORR performance over the pristine counterpart. This can be attributed to the reduced limiting barrier imposed by the formation of -OOH species during ORR due to enhanced adsorption energy of intermediates and the possibly increased conductivity as a result of change in the electronic states as revealed by our density functional theory-based first-principles calculations. Going beyond BaTiO3 NPs, a variety of other ABO3 NPs with tunable sizes and compositions may be readily accessible by exploiting our amphiphilic star-like diblock copolymer nanoreactor strategy. They could in turn provide a unique platform for both fundamental and practical studies on a suite of physical properties (dielectric, piezoelectric, electrostrictive, catalytic, etc.) contingent upon their dimensions and compositions.

Molecular junction by tunneling in 1D and quasi-1D systems.

We have investigated electron tunneling through two one-dimensional (1D) molecular junctions based on first-principles simulations using the density functional theory combined with the non-equilibrium Green's functions methodology. The first junction, composed of left and right carbyne wire electrodes with a sodium atom in between, is atomically thin. The second one is quasi-one-dimensional (quasi-1D) and consists of two single-wall carbon nanotube electrodes, closed on the tips and again a sodium atom in the scattering region. Although the bridging atom bonds weakly to the electrodes in both systems, it strongly affects the electronic transport properties, such as electron transmission, current-voltage relation, differential conductance, density of states and eigenchannels. This is demonstrated by comparing with the results obtained from the corresponding systems for both the 1D and the quasi-1D junctions in the absence of the central sodium atom. The revealed transport properties are sensitive to the molecular geometry. This helps future molecular electronic device design.

Orientation and Electronic Structures of Multilayered Graphene Nanoribbons Produced by Two-Zone Chemical Vapor Deposition.

The orientation and electronic structure of multilayered graphene nanoribbons with an armchair-edge (AGNRs) were determined by low-temperature scanning tunneling microscopy in this study. The orientation of AGNRs was found to be an edge-on structure when positioned as a top layer, while previous reports showed a face-on structure for monolayered AGNRs on Au(111). According to density functional theory calculations, AGNRs in a top layer preferentially form as edge-on structures rather than face-on structures due to the balance of CH-π and π-π interactions between AGNRs. Scanning tunneling spectroscopy and density functional theory calculations revealed that the electronic structures of multilayered AGNRs are similar to those in a gas-phase due to the lack of interaction between AGNRs and the Au(111) substrate. The observation of AGNRs in mutilayers might suggest the conformation-assisted mechanism of dehydrogenation when there is no contact with the Au(111) substrate.

Bifunctional Elastin-like Polypeptide Nanoparticles Bind Rapamycin and Integrins and Suppress Tumor Growth in Vivo.

Recombinant protein-polymer scaffolds such as elastin-like polypeptides (ELPs) offer drug-delivery opportunities including biocompatibility, monodispersity, and multifunctionality. We recently reported that the fusion of FK-506 binding protein 12 (FKBP) to an ELP nanoparticle (FSI) increases rapamycin (Rapa) solubility, suppresses tumor growth in breast cancer xenografts, and reduces side effects observed with free-drug controls. This new report significantly advances this carrier strategy by demonstrating the coassembly of two different ELP diblock copolymers containing drug-loading and tumor-targeting domains. A new ELP nanoparticle (ISR) was synthesized that includes the canonical integrin-targeting ligand (Arg-Gly-Asp, RGD). FSI and ISR mixed in a 1:1 molar ratio coassemble into bifunctional nanoparticles containing both the FKBP domain for Rapa loading and the RGD ligand for integrin binding. Coassembled nanoparticles were evaluated for bifunctionality by performing in vitro cell-binding and drug-retention assays and in vivo MDA-MB-468 breast tumor regression and tumor-accumulation studies. The bifunctional nanoparticle demonstrated superior cell target binding and similar drug retention to FSI; however, it enhanced the formulation potency, such that tumor growth was suppressed at a 3-fold lower dose compared to an untargeted FSI-Rapa control. This data suggests that ELP-mediated scaffolds are useful tools for generating multifunctional nanomedicines with potential activity in cancer.

Breaking the diffraction barrier using coherent anti-Stokes Raman scattering difference microscopy.

We propose a method to improve the resolution of coherent anti-Stokes Raman scattering microscopy (CARS), and present a theoretical model. The proposed method, coherent anti-Stokes Raman scattering difference microscopy (CARS-D), is based on the intensity difference between two differently acquired images. One being the conventional CARS image, and the other obtained when the sample is illuminated by a doughnut shaped spot. The final super-resolution CARS-D image is constructed by intensity subtraction of these two images. However, there is a subtractive factor between them, and the theoretical model sets this factor to obtain the best imaging effect.

Oscillating edge states in one-dimensional MoS2 nanowires.

Reducing the dimensionality of transition metal dichalcogenides to one dimension opens it to structural and electronic modulation related to charge density wave and quantum correlation effects arising from edge states. The greater flexibility of a molecular scale nanowire allows a strain-imposing substrate to exert structural and electronic modulation on it, leading to an interplay between the curvature-induced influences and intrinsic ground-state topology. Herein, the templated growth of MoS2 nanowire arrays consisting of the smallest stoichiometric MoS2 building blocks is investigated using scanning tunnelling microscopy and non-contact atomic force microscopy. Our results show that lattice strain imposed on a nanowire causes the energy of the edge states to oscillate periodically along its length in phase with the period of the substrate topographical modulation. This periodic oscillation vanishes when individual MoS2 nanowires join to form a wider nanoribbon, revealing that the strain-induced modulation depends on in-plane rigidity, which increases with system size.

A potent immunotoxin targeting fibroblast activation protein for treatment of breast cancer in mice.

Fibroblast activation protein (FAP) is highly expressed in the tumor-associated fibroblasts (TAFs) of most human epithelial cancers. FAP plays a critical role in tumorigenesis and cancer progression, which makes it a promising target for novel anticancer therapy. However, mere abrogation of FAP enzymatic activity by small molecules is not very effective in inhibiting tumor growth. In this study, we have evaluated a novel immune-based approach to specifically deplete FAP-expressing TAFs in a mouse 4T1 metastatic breast cancer model. Depletion of FAP-positive stromal cells by FAP-targeting immunotoxin αFAP-PE38 altered levels of various growth factors, cytokines, chemokines and matrix metalloproteinases, decreased the recruitment of tumor-infiltrating immune cells in the tumor microenvironment and suppressed tumor growth. In addition, combined treatment with αFAP-PE38 and paclitaxel potently inhibited tumor growth in vivo. Our findings highlight the potential use of immunotoxin αFAP-PE38 to deplete FAP-expressing TAFs and thus provide a rationale for the use of this immunotoxin in cancer therapy.

Density Functional Theory for Steady-State Nonequilibrium Molecular Junctions.

We present a density functional theory (DFT) for steady-state nonequilibrium quantum systems such as molecular junctions under a finite bias. Based on the steady-state nonequilibrium statistics that maps nonequilibrium to an effective equilibrium, we show that ground-state DFT (GS-DFT) is not applicable in this case and two densities, the total electron density and the density of current-carrying electrons, are needed to uniquely determine the properties of the corresponding nonequilibrium system. A self-consistent mean-field approach based on two densities is then derived. The theory is implemented into SIESTA computational package and applied to study nonequilibrium electronic/transport properties of a realistic carbon-nanotube (CNT)/Benzene junction. Results obtained from our steady-state DFT (SS-DFT) are compared with those of conventional GS-DFT based transport calculations. We show that SS-DFT yields energetically more stable nonequilibrium steady state, predicts significantly lower electric current, and is able to produce correct electronic structures in local equilibrium under a limiting case.

Greatly Enhancing Catalytic Activity of Graphene by Doping the Underlying Metal Substrate.

Graphene-based solid-state catalysis represents a new direction in applications of graphene and has attracted a lot of interests recently. However, the difficulty in fine control and large-scale production of previously proposed graphene catalysts greatly limits their industrial applications. Here we present a novel way to enhance the catalytic activity of graphene, which is highly efficient yet easy to fabricate and control. By first-principles calculations, we show that when the underlying metal substrate is doped with impurities, the catalytic activity of the supported graphene can be drastically enhanced. Graphene supported on a Fe/Ni(111) surface is chosen as a model catalyst, and the chemical reaction of CO oxidation is used to probe the catalytic activity of graphene. When the underlying Fe/Ni(111) substrate is impurity free, the graphene is catalytically inactive. When a Zn atom is doped into the substrate, the catalytic activity of the supported graphene is greatly enhanced, and the reaction barrier of the catalyzed CO oxidation is reduced to less than 0.5 eV. Intriguing reaction mechanism of catalyzed CO oxidation is revealed. These studies suggest a new class of graphene-based catalysts and pave the way for future applications of graphene in solid-state catalysis.

Modeling optical properties of silicon clusters by first principles: From a few atoms to large nanocrystals.

Time dependent density functional tight binding (TDDFTB) method is implemented with sparse matrix techniques and improved parallelization algorithms. The method is employed to calculate the optical properties of various Si nanocrystals (NCs). The calculated light absorption spectra of small Si NCs from TDDFTB were found to be comparable with many body perturbation methods utilizing planewave basis sets. For large Si NCs (more than a thousand atoms) that are beyond the reach of conventional approaches, the TDDFTB method is able to produce reasonable results that are consistent with prior experiments. We also employed the method to study the effects of surface chemistry on the optical properties of large Si NCs. We learned that the optical properties of Si NCs can be manipulated with small molecule passivations such as methyl, hydroxyl, amino, and fluorine. In general, the shifts and profiles in the absorption spectra can be tuned with suitably chosen passivants.

Small-Animal PET Imaging of Pancreatic Cancer Xenografts Using a 64Cu-Labeled Monoclonal Antibody, MAb159.

UNLABELLED: Overexpression of the GRP78 receptor on cell surfaces has been linked with tumor growth, metastasis, and resistance to therapy. We developed a (64)Cu-labeled probe for PET imaging of tumor GRP78 expression based on a novel anti-GRP78 monoclonal antibody, MAb159. METHODS: MAb159 was conjugated with the (64)Cu-chelator DOTA through lysines on the antibody. DOTA-human IgG was also prepared as a control that did not bind to GRP78. The resulting PET probes were evaluated in BXPC3 pancreatic cancer xenografts in athymic nude mice. RESULTS: The radiotracer was synthesized with a specific activity of 0.8 MBq/µg of antibody. In BXPC3 xenografts, (64)Cu-DOTA-MAb159 demonstrated prominent tumor accumulation (4.3 ± 1.2, 15.4 ± 2.6, and 18.3 ± 1.0 percentage injected dose per gram at 1, 17, and 48 after injection, respectively). In contrast, (64)Cu-DOTA-human IgG had low BXPC3 tumor accumulation (4.8 ± 0.5, 7.5 ± 0.7, and 4.6 ± 0.8 percentage injected dose per gram at 1, 17, and 48 h after injection, respectively). CONCLUSION: We demonstrated that GRP78 can serve as a valid target for pancreatic cancer imaging. The success of this approach will be valuable for evaluating disease course and therapeutic efficacy at the earliest stages of anti-GRP78 treatment. Moreover, these newly developed probes may have important applications in other types of cancer overexpressing GRP78.

Improved metabolic stability for 18F PET probes rapidly constructed via tetrazine trans-cyclooctene ligation.

The fast kinetics and bioorthogonal nature of the tetrazine trans-cyclooctene (TCO) ligation makes it a unique tool for PET probe construction. In this study, we report the development of an (18)F-labeling system based on a CF3-substituted diphenyl-s-tetrazine derivative with the aim of maintaining high reactivity while increasing in vivo stability. c(RGDyK) was tagged by a CF3-substituted diphenyl-s-tetrazine derivative via EDC-mediated coupling. The resulting tetrazine-RGD conjugate was combined with a (19)F-labeled TCO derivative to give HPLC standards. The analogous (18)F-labeled TCO derivative was combined with the diphenyl-s-tetrazine-RGD at µM concentration. The resulting tracer was subjected to in vivo metabolic stability assessment, and microPET studies in murine U87MG xenograft models. The diphenyl-s-tetrazine-RGD combines with an (18)F-labeled TCO in high yields (>97% decay-corrected on the basis of TCO) using only 4 equiv of tetrazine-RGD relative to the (18)F-labeled TCO (concentration calculated based on product's specific activity). The radiochemical purity of the (18)F-RGD peptides was >95% and the specific activity was 111 GBq/µmol. Noninvasive microPET experiments demonstrated that (18)F-RGD had integrin-specific tumor uptake in subcutaneous U87MG glioma. In vivo metabolic stability of (18)F-RGD in blood, urine, and major organs showed two major peaks: one corresponded to the Diels-Alder conjugate and the other was identified as the aromatized analog. A CF3-substituted diphenyl-s-tetrazine displays excellent speed and efficiency in (18)F-PET probe construction, providing nearly quantitative (18)F labeling within minutes at low micromolar concentrations. The resulting conjugates display improved in vivo metabolic stability relative to our previously described system.