Search

The fabrication of three-dimensional silicon nitride (Si3N4) fiber-reinforced silicon nitride matrix (3D Si3N4f/BN/Si3N4) composites with a boron nitride (BN) interphase through precursor infiltration and pyrolysis (PIP) process was reported. Heat treatment at 1000–1200 °C was used to analyze the thermal stability of the Si3N4f/BN/Si3N4 composites. It was found after heat treatment the flexural strength and fracture toughness change with a pattern that decrease first and then increase, which are 191 ± 13 MPa and 5.8 ± 0.5 MPa·m1/2 respectively for as-fabricated composites, and reach the minimum values of 138 ± 6 MPa and 3.9 ± 0.4 MPa·m1/2 respectively for composites annealed at 1100 °C. The influence mechanisms of the heat treatment on the Si3N4f/BN/Si3N4 composites include: (Ⅰ) matrix shrinkage by further ceramization that causes defects such as pores and cracks in composites, and (Ⅱ) prestress relaxation, thermal residual stress (TRS) redistribution and a better wetting at the fiber/matrix (F/M) surface that increase the interfacial bonding strength (IBS). Thus, heat treatment affects the mechanical properties of composites by changing the properties of the matrix and IBS, where the load transfer efficiency onto the fibers is fluctuating by the microstructural evolution of matrix and gradually increasing IBS.

Silicon nitride fibers-reinforced silicon nitride matrix (Si3N4f/BN/Si3N4) composites with a BN interphase were fabricated by chemical vapor infiltration (CVI) and precursor infiltration pyrolysis (PIP). It is shown that PIP Si3N4f/BN/Si3N4 composites possessed better mechanical properties and achieved effectively interface debonding and fiber pull-out, whose flexural strength and fracture toughness can achieve 189 ± 17 MPa and 5.6 ± 0.6 MPa·m1/2, about twice and 1.5 times as much as those of CVI Si3N4f/BN/Si3N4 composites, respectively. Combining the micro-nano mechanical and macro-mechanical properties investigations, it is revealed that weakening the interfacial bond and achieving proper modulus matching relationship between fibers and matrix by PIP process could remarkably increase the mechanical properties of the Si3N4f/BN/Si3N4 composites, compared with CVI process.

In this study, three-dimensional silicon nitride fiber-reinforced silicon nitride matrix (3D Si3N4f/BN/Si3N4) composites with a boron nitride (BN) interphase were fabricated through chemical vapor infiltration. Through comparing the changes of microstructure, thermal residual stress, interface bonding state, and interface microstructure evolution of composites before and after heat treatment, the evolution of mechanical and dielectric properties of Si3N4f/BN/Si3N4 composites was analyzed. Flexural strength and fracture toughness of composites acquired the maximum values of 96 ± 5 MPa and 3.8 ± 0.1 MPa·m1/2, respectively, after heat treatment at 800 °C; however, these values were maintained at 83 ± 6 MPa and 3.1 ± 0.2 MPa·m1/2 after heat treatment at 1200 °C, respectively. The relatively low mechanical properties are mainly attributed to the strong interface bonding caused by interfacial diffusion of oxygen and subsequent interfacial reaction and generation of turbostratic BN interphase with relatively high fracture energy. Moreover, the Si3N4f/BN/Si3N4 composites also displayed moderate dielectric constant and dielectric loss fluctuating irregularly around 5.0 and 0.04 before and after heat treatment, respectively. They were mainly determined based on the intrinsic properties of materials system and complex microstructure of composites.

Cryo-electron tomography (cryo-ET) with subtomogram averaging (STA) is indispensable when studying macromolecule structures and functions in their native environments. Due to the low signal-to-noise ratio, the missing wedge artifacts in tomographic reconstructions, and multiple macromolecules of varied shapes and sizes, macromolecule localization and classification remain challenging. To tackle this bottleneck problem for structural determination by STA, we design an accurate macromolecule localization and classification method named voxelwise particle detector (VP-Detector).VP-Detector is a two-stage particle detection method based on a 3D multiscale dense convolutional neural network (3D MSDNet). The proposed network uses 3D hybrid dilated convolution (3D HDC) to avoid the resolution loss caused by scaling operations. Meanwhile, it uses 3D dense connectivity to encourage the reuse of feature maps to reduce trainable parameters. In addition, the weighted focal loss is proposed to focus more attention on difficult samples and rare classes, which relieves the class imbalance caused by multiple particles of various sizes. The performance of VP-Detector is evaluated on both simulated and real-world tomograms, and it shows that VP-Detector outperforms state-of-the-art methods.The experiments show that VP-Detector outperforms the state-of-the-art methods on particle localization with an F1-score of 0.951 and a precision of 0.978. In addition, VP-Detector can replace manual particle picking in experiment on the real-world tomograms. Furthermore, it performs well in classifying large-, medium-, and small-weight proteins with accuracies of 1, 0.95, and 0.82, respectively. Finally, ablation studies demonstrate the effectiveness of 3D HDC, 3D dense connectivity, weighted focal loss, and training on small training sets.VP-Detector can achieve high accuracy in particle detection with few trainable parameters and support training on small datasets. It can also relieve the class imbalance caused by multiple particles with various shapes and sizes.

Analysis and prediction of drug-target interactions (DTIs) play an important role in understanding drug mechanisms, as well as drug repositioning and design. Machine learning (ML)-based methods for DTIs prediction can mitigate the shortcomings of time-consuming and labor-intensive experimental approaches, while providing new ideas and insights for drug design. We propose a novel pipeline for predicting drug-target interactions, called DNN-DTIs. First, the target information is characterized by a number of features, namely, pseudo-amino acid composition, pseudo position-specific scoring matrix, conjoint triad composition, transition and distribution, Moreau-Broto autocorrelation, and structural features. The drug compounds are subsequently encoded using substructure fingerprints. Next, eXtreme gradient boosting (XGBoost) is used to determine the subset of non-redundant features of importance. The optimal balanced set of sample vectors is obtained by applying the synthetic minority oversampling technique (SMOTE). Finally, a DTIs predictor, DNN-DTIs, is developed based on a deep neural network (DNN) via a layer-by-layer learning scheme. Experimental results indicate that DNN-DTIs achieves better performance than other state-of-the-art predictors with ACC values of 98.78%, 98.60%, 97.98%, 98.24% and 98.00% on Enzyme, Ion Channels (IC), GPCR, Nuclear Receptors (NR) and Kuang's datasets. Therefore, the accurate prediction performance of DNN-DTIs makes it a favored choice for contributing to the study of DTIs, especially drug repositioning.

We report our investigation on the electronic properties of Cr-N codoped rutile TiO2(110) single crystal thin films, homoepitaxially grown by pulsed-laser-deposition method, and characterized using scanning tunneling microscopy and spectroscopy (STM/STS), X-ray/ultraviolet photoemission spectroscopy (XPS/UPS), in combination with first-principles calculations. Our results show that the bandgap reduction of the TiO2(110) surface is mainly contributed by the delocalized states whose position is at 2.0 eV below the Fermi level, introduced by the substitutional codoped Cr-2N pair, which is evidenced by the accordance of the results between the STS spectra and the calculated DOS. The codoped Cr-N pair contributes the gap state at about 0.8 eV below the Fermi level, in consistent with the theoretical calculations. While, the monodoped Cr contributes the states either close to the valence band maximum or the conduction band minimum, which should not contribute to the bandgap reduction too much. Our experimental results joint with theoretical calculations provide an atomic view of the bandgap reduction of the rutile TiO2(110) surface, which indicates that the excess substitutional N atoms should be important to efficiently narrow the bandgap by introducing the Cr-2N pairs.

At the core of RNA interference, the Argonaute proteins (Ago) load and utilize small guide nucleic acids to silence mRNAs or cleave foreign nucleic acids in a sequence specific manner. In recent years, based on extensive structural studies of Ago and its interaction with the nucleic acids, considerable progress has been made to reveal the dynamic aspects of various Ago-mediated processes. Here we review these novel insights into the guide-strand loading, duplex unwinding, and effects of seed mismatch, with a focus on two representative Agos, the human Ago 2 (hAgo2) and the bacterial Thermus thermophilus Ago (TtAgo). In particular, comprehensive molecular simulation studies revealed that although sharing similar overall structures, the two Agos have vastly different conformational landscapes and guide-strand loading mechanisms because of the distinct rigidity of their L1-PAZ hinge. Given the central role of the PAZ motions in regulating the exposure of the nucleic acid binding channel, these findings exemplify the importance of protein motions in distinguishing the overlapping, yet distinct, mechanisms of Ago-mediated processes in different organisms.

The multi-instance dictionary plays a critical role in multi-instance data representation. Meanwhile, different multi-instance learning applications are evaluated by specific multivariate performance measures. For example, multi-instance ranking reports the precision and recall. It is not difficult to see that to obtain different optimal performance measures, different dictionaries are needed. This observation motives us to learn performance-optimal dictionaries for this problem. In this paper, we propose a novel joint framework for learning the multi-instance dictionary and the classifier to optimize a given multivariate performance measure, such as the F1 score and precision at rank k. We propose to represent the bags as bag-level features via the bag-instance similarity, and learn a classifier in the bag-level feature space to optimize the given performance measure. We propose to minimize the upper bound of a multivariate loss corresponding to the performance measure, the complexity of the classifier, and the complexity of the dictionary, simultaneously, with regard to both the dictionary and the classifier parameters. In this way, the dictionary learning is regularized by the performance optimization, and a performance-optimal dictionary is obtained. We develop an iterative algorithm to solve this minimization problem efficiently using a cutting-plane algorithm and a coordinate descent method. Experiments on multi-instance benchmark data sets show its advantage over both traditional multi-instance learning and performance optimization methods. Different multivariate performance require different optimal multi-instance dictionaries.We propose the problem of learning performance-optimal multi-instance dictionary.We learn dictionary and classifier jointly to optimize a target performance measure.It outperforms both independent dictionary learning and performance optimization methods.

Understanding the response of vegetation spring phenology to climatic factors in arid regions is crucial for projecting the land-climate interactions of arid ecosystems under climate change. The Xinjiang region, an important part of northern China with complex climatic features, was selected as a case study. We investigated the impact of temperature, precipitation, and solar radiation on the start of the vegetation growing season (SOS) using a long-term series of the normalized difference vegetation index, climate data from 1982 to 2014, and the partial least squares regression (PLS) method. The primary findings were following: (1) the regional scaling of the mean SOS ranged from 95 to 113 day of year (DOY) and showed distinct spatial heterogeneity for different vegetation types. The relative percentage of the area that exhibited advanced SOS (65.80%) was larger than that of the area that exhibited delayed SOS (34.20%), with a regional advancing tendency of 0.19 days per year during the study period. Different vegetation types showed an advancing SOS, which shifted to delaying after 2005. (2) Temperature significantly affected the SOS for various vegetation types. Generally, the warming spring temperatures led to advanced SOS, whereas reduced warming in spring after 2005 reversed the SOS trends. The higher solar radiation occurring from late winter to spring also contributed to the advance in the SOS. Conversely, the increasing precipitation during this period caused a delay in the SOS. These results can provide a useful reference for improving the vegetation phenology models of arid regions.

The protein structure classification problem, which is to assign a protein structure to a cluster of similar proteins, is one of the most fundamental problems in the construction and application of the protein structure space. Early manually curated protein structure classifications (e.g., SCOP and CATH) are very successful, but recently suffer the slow updating problem because of the increased throughput of newly solved protein structures. Thus, fully automatic methods to cluster proteins in the protein structure space have been designed and developed. In this study, we observed that the SCOP superfamilies are highly consistent with clustering trees representing hierarchical clustering procedures, but the tree cutting is very challenging and becomes the bottleneck of clustering accuracy. To overcome this challenge, we proposed a novel density-based K-nearest uphill clustering method that effectively eliminates noisy pairwise protein structure similarities and identifies density peaks as cluster centers. Specifically, the density peaks are identified based on K-nearest uphills (i.e., proteins with higher densities) and K-nearest neighbors. To our knowledge, this is the first attempt to apply and develop density-based clustering methods in the protein structure space. Our results show that our density-based clustering method outperforms the state-of-the-art clustering methods previously applied to the problem. Moreover, we observed that computational methods and human experts could produce highly similar clusters at high precision values, while computational methods also suggest to split some large superfamilies into smaller clusters.

The ‘4 per mille’ initiative, launched in 2015, highlighted the crucial role of soil organic carbon (SOC) sequestration in food security and climate change adaptation and mitigation. The objectives of this study were to investigate the changes in the SOC stock (SOCs) with depth and time and explore the SOC accumulation process, in a millennium chronosequence derived from marine sediments. Chronofunctions were formulated based on SOCs at different depths to evaluate the effects of sampling depth on chronofunctions and the rates of SOC sequestration. The depth distribution of SOCs showed that SOC accumulation occurred mainly in the top 40 cm, with 69% of the SOC being distributed in the upper 0–40 cm layers. The chronofunctions yielding the best fits changed from linear to power and then to logarithmic, suggesting that sampling depth affected the chronofunction type. The SOCs for the entire profile (0–100 cm) could be estimated from that in the 0–20 cm or 0–50 cm, which can potentially aid the estimation of legacy SOC data. The SOC sequestration rates were found to be high at the beginning of development, gradually slowing down with time. The soil in this area can sequestrate SOC for more than 2000 years, at a rate greater than 0.4%. Spectroscopy and digital soil mapping methods can provide accurate and acceptable SOC data pertaining to agriculture and climate change and could be a potential technology for global SOC monitoring.

Accompanying the boom in the global economy, CO2 emissions have soared over the past several decades, with the developing world exhibiting higher emission growth rates than the developed world. Emissions transfers between regions, which represent a significant fraction of total emissions, are assumed to be a primary factor contributing to this difference. It is important to understand these transfer figures and the resulting consumption-based emissions in order to evaluate the emissions drivers and establish climate policies. Existing studies, however, have merely estimated figures over a 20 years span (post-1990) using a traditional input–output analysis (IOA) framework. To broaden the data coverage (to pre-1990) of these transfer figures and to further analyze their impacts on total emissions in the long term, a new model called the Long-term Consumption-based Accounting model (LCBA), which is directly based on statistics, is developed to span the period from 1948 to 2011. The results are consistent with the magnitudes and trends of existing studies over the validation (post-1990) period. We use Monte Carlo methods to calculate upper and lower bounds on the LCBA for each country and year, and find that 3 existing time series are almost fully included within these boundaries from 1990. Furthermore, the LCBA model is succinct enough to be easily expanded for future GHG estimations or to analyze other ecological footprints related to “the flow of materials”. It can be assumed that the soaring emissions transfers will seriously jeopardize the current climate policies such as Kyoto Protocol. The Durban Platform for Enhanced Action (ADP) under which all parties are legally bound will require a consumption-based accounting method together with the territorial one in order to achieve an equitable agreement. However, more researches are still needed to facilitate the use of these figures to better support decision making.

We evaluate the effectiveness and the regional inequalities of solar radiation management (SRM) in compensating for simultaneous changes in temperature and precipitation caused by increased greenhouse gas concentrations. We analyze the results from Earth System Models under four Geoengineering Model Intercomparison Project (GeoMIP) experiments with a modified form of the Residual Climate Response approach. Each experiment produces 50 model yrs of simulations: 13 models completed experiment G1 (offsetting 4 × CO 2 via solar reduction); 12 models completed experiment G2 (offsetting CO 2 that increased by 1% per year); 3 models completed experiment G3 (offsetting increasing radiative forcing under RCP4.5 with increasing stratospheric aerosol); and 7 models completed experiment G4 (injection of 5 Tg SO 2 a − 1 into the stratosphere). The regional inequalities in temperature and precipitation compensation for experiments G1 , G3 and G4 are significantly different from their corresponding noise backgrounds for most models, but for G2 they are not significantly different from noise. Differences in the regional inequalities and the actual effectiveness among the four SRM scenarios are not significant for many models. However, in more than half of the models, the effectiveness for temperature in the solar dimming geoengineering scenarios ( G1 and G2 ) is significantly higher than that in the SO 2 geoengineering scenarios ( G3 and G4 ). The effectiveness of the four SRM experiments in compensating for temperature change is considerably higher than for precipitation. The methodology used highlights that a large across-model variation in the treatment of key geoengineering processes (such as stratospheric aerosols) and the quantification of damage caused by climate change creates significant uncertainties in any strategies to achieve optimal compensation effectiveness across different regions.

The Mongolian Plateau (MP) steppe is one of the largest steppe environments in the world. To monitor the terrestrial vegetation dynamics on the MP and to ascertain what the driving forces, this study examined the vegetation dynamics in Republic of Mongolia (M) and the Inner Mongolia Autonomous Region (IM) of China from the period 1982 to 2011, based on the satellite-derived GIMMS NDVI3g (Normalized Difference Vegetation Index) data across three biomes (desert, grassland and forest). The results are as followed: (1) Vegetation coverage in IM was generally greater than that in M. Before 2002, time series of NDVI over the MP increased at an average rate of 0.05% yr −1 . Additionally, after 2002, the NDVI increased at a rate of 0.21% yr −1 . From 1982 to 2011, the area of IM and M with positive anomalies in the NDVI increased at a separate rate of 1.82% yr −1 and 1.76% yr −1 , respectively. (2) At the biome scale, the inter-annual forest NDVI variation in IM and desert NDVI for the entire MP had a significant increasing trend (0.06% yr −1 and 0.04% yr −1 , respectively). (3) Climate forcing was a dominant controlling factor affecting the vegetation, and the anthropogenic behavior exhibited no significant value in the whole region. However, overgrazing was the most important reason for the regional degradation, particularly in IM. (4) In the future, the forest biome will go to recovery, whereas both the grassland and desert biomes are predicted to degrade continuously.

Drought is projected to become more prevalent in the future due to climate change, and its impact on the fate of terrestrial ecosystems has aroused great concern in the scientific community over the past decade. Mounting evidence suggests that drought may be the most important physical stress of terrestrial ecosystems: drought limits vegetation growth, increases wildfires, and induces tree mortality, among other impacts. Drought not only weakens the carbon sink function of terrestrial ecosystems but also may interfere directly or indirectly with biosphere–atmosphere interactions, further exacerbating climate change. This paper reviews the current evidence of the impacts of drought on terrestrial ecosystems, with particular emphasis on the ways in which drought alters the biological, biogeophysical and biogeochemical processes underlying the interaction between the biosphere and the atmosphere.

Based on statistical physics and graph theory, the research paradigm of a complex network, which has sprung up in the last decade, provides us with new global perspective to discuss the topic of international trade. In this paper, we engage in the issue of countries' roles and positions in international trade using the latest complex network theories. On a mid-level structure, countries are classified into three communities that reflect the structure of the “core/periphery” using the weighted extremal optimisation algorithm and the coarse graining process. On a micro-level, countries' rankings are provided with the aid of network's node centralities, which presents world trade as a closed, imbalanced, diversified and multi-polar development. Further, we firstly introduce the improved bootstrap percolation to simulate cascading influences following the breaking down of bilateral trade relations. We find that the breakdown of EU's export relations can more easily form a cascading reaction, which would result in a global collapse of world trade. All the results highlight the important positions of the EU, USA and Japan in the international trade system, which plays a positive role in promoting the world economy.

China's land use has undergone many changes over the past 300 years due to the significant transformations caused by natural and human factors and their impact on regional climate and the environment. This comprehensive review of recent state-of-the-art studies of China's land-use changes during that period concentrates on cropland, forest, grassland and urban areas. While most small-scale studies have reconstructed information from historical archive data and focused on a specific time period, large-scale studies have tended to rely on inverse modeling techniques to interpret land-use change dynamics based on remote-sensing data for example, the global land-use products of the History Database of the Global Environment (HYDE) and Center for Sustainability and the Global Environment (SAGE) datasets. All studies have shown that the cropland areas in China increased between 1700 and 1950, although they indicate different magnitudes and rates. A decrease in forest coverage was also reported in all studies. Little information was available on urban and grassland areas over the same period. Rapid urbanization in China has been particularly evident in the past 50 years. Meanwhile, spatially explicit reconstructions of historical land-use change in China since 1700 remain highly uncertain due to the lack of reliable data. Extensive work on primary data collection is required, including land-use records and drivers for future change.

We present the preparation of C54 TiSi 2 nanoislands on Si (1 1 1) with a method of the pulsed laser deposition of titanium oxide thin films. The TiO 2 thin films with nominal thicknesses of 1 nm on Si (1 1 1) were annealed at 850 °C for about 4 h in situ. The X-ray diffraction patterns and the X-ray photoelectron spectra indicate that the nanoislands are in C54 TiSi 2 phase. The characterization using a scanning tunneling microscope shows that the nanoislands with triangular, polygonal and rod-like shapes on Si (1 1 1) exhibit the Volmer–Weber growth mode. The sizes of the polygonal islands distribute in two separated ranges. For the small islands, they have a narrow lateral size distribution centered at 4 nm and a height range in 0.6–3.6 nm, while for the large islands, their lateral sizes are in the range of 12–40 nm and the heights in the range of 4–9 nm. The sizes of the well-shaped triangular islands are intermediate with the lateral sizes in range of 5–20 nm and the heights of 2–3.5 nm. The rod-like islands are about 50–200 nm in length, 5 nm in height and about 15–20 nm in width. The origination of the various shapes of the nanoislands is attributed to the symmetry of Si (1 1 1) substrate and the lattice mismatch between the C54 TiSi 2 and the Si (1 1 1) surface.

A general atmospheric circulation model (ECHAM5) has been applied to investigate the impact of land use changes on the Tibetan Plateau (TP) on local and global climate. The “control” simulation with current land cover reasonably represents the large-scale circulation and state of the atmosphere over the Tibetan Plateau and the surrounding region. Modifying the land cover to a hypothetical non-anthropogenically-influenced vegetation cover shows significant modifications to the local and remote climate. Compared to this scenario, the TP is warmer and drier under present conditions. The Indian summer monsoon is intensified and the East China summer monsoon is weakened due to human-induced land cover change on the TP. The mean global temperature has almost no variation, whereas precipitation slightly increases. Our study indicates that human-induced land use changes on the Tibetan Plateau have had a significant impact on local to regional, and to a lesser extent global, climate.

We present the growth of textured TiO2 thin films on muscovite mica using pulsed laser deposition. Atomic force microscopy, scanning electron microscopy, X-ray photoemission spectroscopy, and transmission electron microscopy were used to characterize the TiO2 films. Quasi-periodic wavy and comb-like buckles were observed. Below a critical thickness of about 25 nm, TiO2 films were relatively smooth, and buckles began to form when the nominal thickness of TiO2 films was larger than 25 nm. Co-existence of wavy and comb-like quasi-periodic buckles was observed when the nominal thickness of TiO2 films was larger than 100 nm. The film stress is compressive due to the competition of the tensile stress from the lattice mismatch and the growth stress. The buckle delamination occurred when the compressive stress overcome the adhesion strength of the TiO2 films on mica. A value of adhesion strength around 0.9 MPa for TiO2 on mica is obtained. The symmetric domains of buckles are assigned to the anisotropic lattice mismatch for TiO2 on mica.

Tin oxide (SnO 2 ) nanowires have been synthesized in bulk quantities at lower temperature (680 °C) by thermal evaporation of SnO powder. These as-synthesized SnO 2 nanowires are single crystals with [3 0 1] growth axis, with diameters ranging from 10 to 190 nm and lengths extending to tens of micrometers. The growth of SnO 2 nanowires synthesized by lower-temperature evaporation is dominated by a self-catalytic vapor–liquid–solid (VLS) mechanism. The self-catalytic effect in VLS process has been described.

This paper is to highlight the threatening of climate change on the water shortage problem in North China, a semi-arid and arid area. Observations show that the runoff of the major rivers is decreasing dramatically especially in the lower reaches where water is highly demanded. Most noticeable are the dry-outs of the Yellow River, the so-called ‘mother river of China’. Water scarcity would be the most front challenge for the society and economy in China, especially the urban area with extremely high population density. The government already paid attention to the deepening of the underground water level at Beijing. Sustainable water resource managementshouldthetoppriorityforthelocalcommunitytoadapttoclimatechange.Studiesshowthathumanactivitiesheavilyimpact on the local climate and environment in China and a dramatics amount of clean water is wasted or polluted.