Tsinghua–Berkeley Shenzhen Institute

Abstract. Accurate assessment of anthropogenic carbon dioxide (CO2) emissions andtheir redistribution among the atmosphere, ocean, and terrestrial biospherein a changing climate is critical to better understand the global carboncycle, support the development of climate policies, and project futureclimate change. Here we describe and synthesize data sets and methodologies toquantify the five major components of the global carbon budget and theiruncertainties. Fossil CO2 emissions (EFOS) are based on energystatistics and cement production data, while emissions from land-use change(ELUC), mainly deforestation, are based on land use and land-use changedata and bookkeeping models. Atmospheric CO2 concentration is measureddirectly, and its growth rate (GATM) is computed from the annualchanges in concentration. The ocean CO2 sink (SOCEAN) is estimatedwith global ocean biogeochemistry models and observation-baseddata products. The terrestrial CO2 sink (SLAND) is estimated withdynamic global vegetation models. The resulting carbon budget imbalance(BIM), the difference between the estimated total emissions and theestimated changes in the atmosphere, ocean, and terrestrial biosphere, is ameasure of imperfect data and understanding of the contemporary carboncycle. All uncertainties are reported as ±1σ. For the year 2021, EFOS increased by 5.1 % relative to 2020, withfossil emissions at 10.1 ± 0.5 GtC yr−1 (9.9 ± 0.5 GtC yr−1 when the cement carbonation sink is included), and ELUC was 1.1 ± 0.7 GtC yr−1, for a total anthropogenic CO2 emission(including the cement carbonation sink) of 10.9 ± 0.8 GtC yr−1(40.0 ± 2.9 GtCO2). Also, for 2021, GATM was 5.2 ± 0.2 GtC yr−1 (2.5 ± 0.1 ppm yr−1), SOCEAN was 2.9 ± 0.4 GtC yr−1, and SLAND was 3.5 ± 0.9 GtC yr−1, with aBIM of −0.6 GtC yr−1 (i.e. the total estimated sources were too low orsinks were too high). The global atmospheric CO2 concentration averaged over2021 reached 414.71 ± 0.1 ppm. Preliminary data for 2022 suggest anincrease in EFOS relative to 2021 of +1.0 % (0.1 % to 1.9 %)globally and atmospheric CO2 concentration reaching 417.2 ppm, morethan 50 % above pre-industrial levels (around 278 ppm). Overall, the meanand trend in the components of the global carbon budget are consistentlyestimated over the period 1959–2021, but discrepancies of up to 1 GtC yr−1 persist for the representation of annual to semi-decadalvariability in CO2 fluxes. Comparison of estimates from multipleapproaches and observations shows (1) a persistent large uncertainty in theestimate of land-use change emissions, (2) a low agreement between thedifferent methods on the magnitude of the land CO2 flux in the northernextratropics, and (3) a discrepancy between the different methods on thestrength of the ocean sink over the last decade. This living data updatedocuments changes in the methods and data sets used in this new globalcarbon budget and the progress in understanding of the global carbon cyclecompared with previous publications of this data set. The data presented inthis work are available at https://doi.org/10.18160/GCP-2022 (Friedlingstein et al., 2022b).

Lithium-sulfur (Li-S) batteries have attracted tremendous interest because of their high theoretical energy density and cost effectiveness. The target of Li-S battery research is to produce batteries with a high useful energy density that at least outperforms state-of-the-art lithium-ion batteries. However, due to an intrinsic gap between fundamental research and practical applications, the outstanding electrochemical results obtained in most Li-S battery studies indeed correspond to low useful energy densities and are not really suitable for practical requirements. The Li-S battery is a complex device and its useful energy density is determined by a number of design parameters, most of which are often ignored, leading to the failure to meet commercial requirements. The purpose of this review is to discuss how to pave the way for reliable Li-S batteries. First, the current research status of Li-S batteries is briefly reviewed based on statistical information obtained from literature. This includes an analysis of how the various parameters influence the useful energy density and a summary of existing problems in the current Li-S battery research. Possible solutions and some concerns regarding the construction of reliable Li-S batteries are comprehensively discussed. Finally, insights are offered on the future directions and prospects in Li-S battery field.

Two-dimensional (2D) materials have attracted increasing research interest because of the abundant choice of materials with diverse and tunable electronic, optical, and chemical properties. Moreover, 2D material based heterostructures combining several individual 2D materials provide unique platforms to create an almost infinite number of materials and show exotic physical phenomena as well as new properties and applications. To achieve these high expectations, methods for the scalable preparation of 2D materials and 2D heterostructures of high quality and low cost must be developed. Chemical vapor deposition (CVD) is a powerful method which may meet the above requirements, and has been extensively used to grow 2D materials and their heterostructures in recent years, despite several challenges remaining. In this review of the challenges in the CVD growth of 2D materials, we highlight recent advances in the controlled growth of single crystal 2D materials, with an emphasis on semiconducting transition metal dichalcogenides. We provide insight into the growth mechanisms of single crystal 2D domains and the key technologies used to realize wafer-scale growth of continuous and homogeneous 2D films which are important for practical applications. Meanwhile, strategies to design and grow various kinds of 2D material based heterostructures are thoroughly discussed. The applications of CVD-grown 2D materials and their heterostructures in electronics, optoelectronics, sensors, flexible devices, and electrocatalysis are also discussed. Finally, we suggest solutions to these challenges and ideas concerning future developments in this emerging field.

Abstract Although the rechargeable lithium–sulfur battery is an advanced energy storage system, its practical implementation has been impeded by many issues, in particular the shuttle effect causing rapid capacity fade and low Coulombic efficiency. Herein, we report a conductive porous vanadium nitride nanoribbon/graphene composite accommodating the catholyte as the cathode of a lithium–sulfur battery. The vanadium nitride/graphene composite provides strong anchoring for polysulfides and fast polysulfide conversion. The anchoring effect of vanadium nitride is confirmed by experimental and theoretical results. Owing to the high conductivity of vanadium nitride, the composite cathode exhibits lower polarization and faster redox reaction kinetics than a reduced graphene oxide cathode, showing good rate and cycling performances. The initial capacity reaches 1,471 mAh g −1 and the capacity after 100 cycles is 1,252 mAh g −1 at 0.2 C, a loss of only 15%, offering a potential for use in high energy lithium–sulfur batteries.

Stabilizing monolayer nitrides with silicon Transition metal carbides and nitrides are nonlayered materials that in monolayer form have potentially useful electronic and chemical properties. These monolayers are usually made by chemical etching that produces flakes with surface defects that have poor air and water stability. Hong et al. report that introducing silicon during chemical vapor deposition growth of molybdenum nitride passivates the surface and prevents island formation. Centimeter-scale monolayer films of the semiconductor MoSi 2 N 4 form as a MoN 2 layer sandwiched by two Si-N bilayers. These layers possess high mechanical strength and ambient stability. Science , this issue p. 670

Twinborn TiO₂–TiN heterostructures enable smooth trapping–diffusion–conversion of polysulfides and produce ultralong life lithium–sulfur batteries.

Abstract With the prevalence of pre-trained language models (PLMs) and the pre-training–fine-tuning paradigm, it has been continuously shown that larger models tend to yield better performance. However, as PLMs scale up, fine-tuning and storing all the parameters is prohibitively costly and eventually becomes practically infeasible. This necessitates a new branch of research focusing on the parameter-efficient adaptation of PLMs, which optimizes a small portion of the model parameters while keeping the rest fixed, drastically cutting down computation and storage costs. In general, it demonstrates that large-scale models could be effectively stimulated by the optimization of a few parameters. Despite the various designs, here we discuss and analyse the approaches under a more consistent and accessible term ‘delta-tuning’, where ‘delta’ a mathematical notation often used to denote changes, is borrowed to refer to the portion of parameters that are ‘changed’ during training. We formally describe the problem and propose a unified categorization criterion for existing delta-tuning methods to explore their correlations and differences. We also discuss the theoretical principles underlying the effectiveness of delta-tuning and interpret them from the perspectives of optimization and optimal control. Furthermore, we provide a holistic empirical study on over 100 natural language processing tasks and investigate various aspects of delta-tuning. With comprehensive study and analysis, our research demonstrates the theoretical and practical properties of delta-tuning in the adaptation of PLMs.

Abstract Solid composite electrolytes (SCEs) that combine the advantages of solid polymer electrolytes (SPEs) and inorganic ceramic electrolytes (ICEs) present acceptable ionic conductivity, high mechanical strength, and favorable interfacial contact with electrodes, which greatly improve the electrochemical performance of all‐solid‐state batteries compared to single SPEs and ICEs. However, there are many challenges to overcome before the practical application of SCEs, including the low ionic conductivity less than 10 −3 S cm −1 at ambient temperature, poor interfacial stability, and high interfacial resistance, which greatly restrict the room temperature performance. Herein, the advances of SCEs applied in all‐solid‐state lithium batteries are presented, including the Li ion migration mechanism of SCEs, the strategies to enhance the ionic conductivity of SCEs by various morphologies of ICEs, and construction methods of the low resistance and stable interfaces of SCEs with both cathode and anode. Finally, some typical applications of SCEs in lithium batteries are summarized and future development directions are prospected. This work presents how it is quite significant to further enhance the ionic conductivity of SCEs by developing the novel SPEs with the special morphology of ICEs for advanced all‐solid‐state lithium batteries.

Abstract Electrochemical water splitting technology for producing “green hydrogen” is important for the global mission of carbon neutrality. Electrocatalysts with decent performance at high current densities play a central role in the industrial implementation of this technology. This field has advanced immensely in recent years, as witnessed by many types of catalysts designed and synthesized toward industriallyrelevant current densities (>200 mA cm –2 ). By discussing recent advances in this field, several key aspects are summarized that affect the catalytic performance for high‐current‐density electrocatalysis, including dimensionality of catalysts, surface chemistry, electron transport path, morphology, and catalyst‐electrolyte interplay. The multiscale design strategy that considers these aspects comprehensively for developing high‐current‐density electrocatalysts are highlighted. The perspectives on the future directions in this emerging field are also put forward.

A 3D network gel polymer electrolyte (3D-GPE) is designed for lithium metal batteries and prepared by an initiator-free one-pot ring-opening polymerization technique. This 3D-GPE exhibits an unprecedented combination of mechanical strength, ionic conductivity, and more importantly, effective suppression of Li dendrite growth. The produced lithium-based battery presents long life, high rate, and excellent safety. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Abstract Large-scale implementation of electrochemical hydrogen production requires several fundamental issues to be solved, including understanding the mechanism and developing inexpensive electrocatalysts that work well at high current densities. Here we address these challenges by exploring the roles of morphology and surface chemistry, and develop inexpensive and efficient electrocatalysts for hydrogen evolution. Three model electrocatalysts are flat platinum foil, molybdenum disulfide microspheres, and molybdenum disulfide microspheres modified by molybdenum carbide nanoparticles. The last catalyst is highly active for hydrogen evolution independent of pH, with low overpotentials of 227 mV in acidic medium and 220 mV in alkaline medium at a high current density of 1000 mA cm −2 , because of enhanced transfer of mass (reactants and hydrogen bubbles) and fast reaction kinetics due to surface oxygen groups formed on molybdenum carbide during hydrogen evolution. Our work may guide rational design of electrocatalysts that work well at high current densities.

Graphene oxide is highly desired for printing electronics, catalysis, energy storage, separation membranes, biomedicine, and composites. However, the present synthesis methods depend on the reactions of graphite with mixed strong oxidants, which suffer from explosion risk, serious environmental pollution, and long-reaction time up to hundreds of hours. Here, we report a scalable, safe and green method to synthesize graphene oxide with a high yield based on water electrolytic oxidation of graphite. The graphite lattice is fully oxidized within a few seconds in our electrochemical oxidation reaction, and the graphene oxide obtained is similar to those achieved by the present methods. We also discuss the synthesis mechanism and demonstrate continuous and controlled synthesis of graphene oxide and its use for transparent conductive films, strong papers, and ultra-light elastic aerogels.

The increasing demands of energy storage require the significant improvement of current Li-ion battery electrode materials and the development of advanced electrode materials. Thus, it is necessary to gain an in-depth understanding of the reaction processes, degradation mechanism, and thermal decomposition mechanisms under realistic operation conditions. This understanding can be obtained by in situ/operando characterization techniques, which provide information on the structure evolution, redox mechanism, solid-electrolyte interphase (SEI) formation, side reactions, and Li-ion transport properties under operating conditions. Here, the recent developments in the in situ/operando techniques employed for the investigation of the structural stability, dynamic properties, chemical environment changes, and morphological evolution are described and summarized. The experimental approaches reviewed here include X-ray, electron, neutron, optical, and scanning probes. The experimental methods and operating principles, especially the in situ cell designs, are described in detail. Representative studies of the in situ/operando techniques are summarized, and finally the major current challenges and future opportunities are discussed. Several important battery challenges are likely to benefit from these in situ/operando techniques, including the inhomogeneous reactions of high-energy-density cathodes, the development of safe and reversible Li metal plating, and the development of stable SEI.

The inkjet technique has the capability of generating droplets in the picoliter volume range, firing thousands of times in a few seconds and printing in the noncontact manner. Since its emergence, inkjet technology has been widely utilized in the publishing industry for printing of text and pictures. As the technology developed, its applications have been expanded from two-dimensional (2D) to three-dimensional (3D) and even used to fabricate components of electronic devices. At the end of the twentieth century, researchers were aware of the potential value of this technology in life sciences and tissue engineering because its picoliter-level printing unit is suitable for depositing biological components. Currently inkjet technology has been becoming a practical tool in modern medicine serving for drug development, scaffold building, and cell depositing. In this article, we first review the history, principles and different methods of developing this technology. Next, we focus on the recent achievements of inkjet printing in the biological field. Inkjet bioprinting of generic biomaterials, biomacromolecules, DNAs, and cells and their major applications are introduced in order of increasing complexity. The current limitations/challenges and corresponding solutions of this technology are also discussed. A new concept, biopixels, is put forward with a combination of the key characteristics of inkjet printing and basic biological units to bring a comprehensive view on inkjet-based bioprinting. Finally, a roadmap of the entire 3D bioprinting is depicted at the end of this review article, clearly demonstrating the past, present, and future of 3D bioprinting and our current progress in this field.

Extensive research on two-dimensional (2D) materials has triggered the renaissance of an old topic, that is, the intercalation and exfoliation of layer materials. Such top-down exfoliation produced 2D materials and their dispersions have several advantages including low cost, scalable production capability, solution processability, and versatile functionalities stemming from the large number of species of layer materials, and show promising potential in many applications. In recent years, many new methods have been developed for exfoliating layer materials to 2D materials for different application purposes. In this review the different exfoliation approaches are first systematically analyzed from the viewpoint of methodology, and the advantages and disadvantages of each method are compared. Second, the assembly of exfoliated 2D materials into macrostructures by solution-based techniques is summarized. Third, the state-of-the-art applications of 2D material dispersions and their assemblies in electronics and optoelectronics, electrocatalysis, energy storage, etc., are discussed. Finally, insights and perspectives on current research challenges and future opportunities regarding the exfoliation and applications of 2D materials in dispersions are considered.

Abstract Single‐atom catalysts (SACs) are the smallest entities for catalytic reactions with projected high atomic efficiency, superior activity, and selectivity; however, practical applications of SACs suffer from a very low metal loading of 1–2 wt%. Here, a class of SACs based on atomically dispersed transition metals on nitrogen‐doped carbon nanotubes (MSA‐N‐CNTs, where M = Ni, Co, NiCo, CoFe, and NiPt) is synthesized with an extraordinarily high metal loading, e.g., 20 wt% in the case of NiSA‐N‐CNTs, using a new multistep pyrolysis process. Among these materials, NiSA‐N‐CNTs show an excellent selectivity and activity for the electrochemical reduction of CO 2 to CO, achieving a turnover frequency (TOF) of 11.7 s −1 at −0.55 V (vs reversible hydrogen electrode (RHE)), two orders of magnitude higher than Ni nanoparticles supported on CNTs.

The ever-increasing demands for batteries with high energy densities to power the portable electronics with increased power consumption and to advance vehicle electrification and grid energy storage have propelled lithium battery technology to a position of tremendous importance. Carbon nanotubes (CNTs) and graphene, known with many appealing properties, are investigated intensely for improving the performance of lithium-ion (Li-ion) and lithium-sulfur (Li-S) batteries. However, a general and objective understanding of their actual role in Li-ion and Li-S batteries is lacking. It is recognized that CNTs and graphene are not appropriate active lithium storage materials, but are more like a regulator: they do not electrochemically react with lithium ions and electrons, but serve to regulate the lithium storage behavior of a specific electroactive material and increase the range of applications of a lithium battery. First, metrics for the evaluation of lithium batteries are discussed, based on which the regulating role of CNTs and graphene in Li-ion and Li-S batteries is comprehensively considered from fundamental electrochemical reactions to electrode structure and integral cell design. Finally, perspectives on how CNTs and graphene can further contribute to the development of lithium batteries are presented.

Abstract. Global anthropogenic emission inventories remain vital for understanding the sources of atmospheric pollution and the associated impacts on the environment, human health, and society. Rapid changes in today's society require that these inventories provide contemporary estimates of multiple atmospheric pollutants with both source sector and fuel type information to understand and effectively mitigate future impacts. To fill this need, we have updated the open-source Community Emissions Data System (CEDS) (Hoesly et al., 2019) to develop a new global emission inventory, CEDSGBD-MAPS. This inventory includes emissions of seven key atmospheric pollutants (NOx; CO; SO2; NH3; non-methane volatile organic compounds, NMVOCs; black carbon, BC; organic carbon, OC) over the time period from 1970–2017 and reports annual country-total emissions as a function of 11 anthropogenic sectors (agriculture; energy generation; industrial processes; on-road and non-road transportation; separate residential, commercial, and other sectors (RCO); waste; solvent use; and international shipping) and four fuel categories (total coal, solid biofuel, the sum of liquid-fuel and natural-gas combustion, and remaining process-level emissions). The CEDSGBD-MAPS inventory additionally includes monthly global gridded (0.5∘ × 0.5∘) emission fluxes for each compound, sector, and fuel type to facilitate their use in earth system models. CEDSGBD-MAPS utilizes updated activity data, updates to the core CEDS default scaling procedure, and modifications to the final procedures for emissions gridding and aggregation. Relative to the previous CEDS inventory (Hoesly et al., 2018), these updates extend the emission estimates from 2014 to 2017 and improve the overall agreement between CEDS and two widely used global bottom-up emission inventories. The CEDSGBD-MAPS inventory provides the most contemporary global emission estimates to date for these key atmospheric pollutants and is the first to provide global estimates for these species as a function of multiple fuel types and source sectors. Dominant sources of global NOx and SO2 emissions in 2017 include the combustion of oil, gas, and coal in the energy and industry sectors as well as on-road transportation and international shipping for NOx. Dominant sources of global CO emissions in 2017 include on-road transportation and residential biofuel combustion. Dominant global sources of carbonaceous aerosol in 2017 include residential biofuel combustion, on-road transportation (BC only), and emissions from the waste sector. Global emissions of NOx, SO2, CO, BC, and OC all peak in 2012 or earlier, with more recent emission reductions driven by large changes in emissions from China, North America, and Europe. In contrast, global emissions of NH3 and NMVOCs continuously increase between 1970 and 2017, with agriculture as a major source of global NH3 emissions and solvent use, energy, residential, and the on-road transport sectors as major sources of global NMVOCs. Due to similar development methods and underlying datasets, the CEDSGBD-MAPS emissions are expected to have consistent sources of uncertainty as other bottom-up inventories. The CEDSGBD-MAPS source code is publicly available online through GitHub: https://github.com/emcduffie/CEDS/tree/CEDS_GBD-MAPS (last access: 1 December 2020). The CEDSGBD-MAPS emission inventory dataset (both annual country-total and monthly global gridded files) is publicly available under https://doi.org/10.5281/zenodo.3754964 (McDuffie et al., 2020c).

Lithium–sulfur (Li–S) batteries are promising next-generation energy storage technologies due to their high theoretical energy density, environmental friendliness, and low cost. However, low conductivity of sulfur species, dissolution of polysulfides, poor conversion from sulfur reduction, and lithium sulfide (Li2S) oxidation reactions during discharge–charge processes hinder their practical applications. Herein, under the guidance of density functional theory calculations, we have successfully synthesized large-scale single atom vanadium catalysts seeded on graphene to achieve high sulfur content (80 wt % sulfur), fast kinetic (a capacity of 645 mAh g–1 at 3 C rate), and long-life Li–S batteries. Both forward (sulfur reduction) and reverse reactions (Li2S oxidation) are significantly improved by the single atom catalysts. This finding is confirmed by experimental results and consistent with theoretical calculations. The ability of single metal atoms to effectively trap the dissolved lithium polysulfides (LiPSs) and catalytically convert the LiPSs/Li2S during cycling significantly improved sulfur utilization, rate capability, and cycling life. Our work demonstrates an efficient design pathway for single atom catalysts and provides solutions for the development of high energy/power density Li–S batteries.

Backdoor attack intends to embed hidden backdoors into deep neural networks (DNNs), so that the attacked models perform well on benign samples, whereas their predictions will be maliciously changed if the hidden backdoor is activated by attacker-specified triggers. This threat could happen when the training process is not fully controlled, such as training on third-party datasets or adopting third-party models, which poses a new and realistic threat. Although backdoor learning is an emerging and rapidly growing research area, there is still no comprehensive and timely review of it. In this article, we present the first comprehensive survey of this realm. We summarize and categorize existing backdoor attacks and defenses based on their characteristics, and provide a unified framework for analyzing poisoning-based backdoor attacks. Besides, we also analyze the relation between backdoor attacks and relevant fields (i.e., adversarial attacks and data poisoning), and summarize widely adopted benchmark datasets. Finally, we briefly outline certain future research directions relying upon reviewed works. A curated list of backdoor-related resources is also available at https://github.com/THUYimingLi/backdoor-learning-resources.