1322 wileyonlinelibrary.com applications in fi elds of healthcare monitoring, human-computer interaction, and electronic skin. [ 12 ] The relative resistance Δ R normalized by the initial resistance R 0 depends on Poisson's ratio ( ν ) and resistivity variation (Δ ρ ) normalized by its initial resistivity ρ 0 through the expression ΔR / R 0 = (1 + 2ν) ε + Δ ρ / ρ 0.[ 13 ] The sensitivity revealed by gauge factor (GF, defi ned as ( ΔR / R 0 )/ ε ) depends on both intrinsic property and structural feature. According to this formula, graphene-based strain sensors have shown low sensitivities due to the rigid and stable structure of intrinsic graphene. [ 14 ] With hardly opened band gap, the GF of a suspended graphene is only about 1.9 under moderate uniaxial strains. [ 15 ] Therefore, structural engineering of graphene is needed to boost the sensitivity of graphene-based strain sensors.Adjustment of the connection channels in graphene is an effective way to alter its resistivity for improved sensitivity in strain sensors. Two common methods for the structural construction of graphene include high temperature processing based chemical vapor deposition (CVD) and solution processing based sheets/fl akes assembly. As for CVD, the resistivity of graphene would be affected by its grain boundary, grain size, and the defect density. [16][17][18] Continuous graphene fi lms grown by CVD could sustain 1% strain with a GF of only 6.1, [ 19 ] and the GF increases to 151 for a 5% strain due to the morphological Large-Area Ultrathin Graphene Films by Single-Step Marangoni Self-Assembly for Highly Sensitive Strain Sensing ApplicationXinming Li , Tingting Yang , Yao Yang , Jia Zhu , Li Li , Fakhr E. Alam , Xiao Li , Kunlin Wang , Huanyu Cheng , Cheng-Te Lin , * Ying Fang , * and Hongwei Zhu * Promoted by the demand for wearable devices, graphene has been proved to be a promising material for potential applications in fl exible and highly sensitive strain sensors. However, low sensitivity and complex processing of graphene retard the development toward the practical applications. Here, an environment-friendly and cost-effective method to fabricate large-area ultrathin graphene fi lms is proposed for highly sensitive fl exible strain sensor. The assembled graphene fi lms are derived rapidly at the liquid/air interface by Marangoni effect and the area can be scaled up. These graphene-based strain sensors exhibit extremely high sensitivity with gauge factor of 1037 at 2% strain, which represents the highest value for graphene platelets at this small deformation so far. This simple fabrication for strain sensors with highly sensitive performance of strain sensor makes it a novel approach to applications in electronic skin, wearable sensors, and health monitoring platforms.

Nucleation plays a critical role in many physical and biological phenomena ranging from crystallization, melting and evaporation to the formation of clouds and the initiation of neurodegenerative diseases 1-3. However, nucleation is a challenging process to study in experiments especially in the early stage when several atoms/molecules start to form a new phase from its parent phase. Over the years, a number of experimental and computational methods have been used to investigate nucleation processes 4-17 , but it remains unachievable to experimentally determine the 3D atomic structure and dynamics of early stage nuclei. Here, we develop 4D atomic electron tomography (AET) to study early stage nucleation at atomic resolution. Using FePt nanoparticles as a model system, we reveal that early stage nuclei are irregularly shaped, each has a core of one to a few atoms 2 with the maximum order parameter, and the order parameter gradient points from the core to the boundary of the nucleus. We capture the structure and dynamics of the same nuclei undergoing growth, fluctuation, dissolution, merging and/or division, which are regulated by the order parameter distribution and its gradient. These experimental observations are corroborated by molecular dynamics simulations of heterogeneous and homogeneous nucleation in liquid-solid phase transitions of Pt. Our experimental and molecular dynamics results differ from classical nucleation theory (CNT) 1,2,18 , indicating a theory beyond CNT is needed to describe early stage nucleation at the atomic scale. Looking forward, we anticipate that 4D AET opens the door to study many fundamental problems in materials science, nanoscience, condensed matter physics and chemistry such as phase transition, atomic diffusion, grain boundary dynamics, interface motion, defect dynamics and surface reconstruction with 4D atomic resolution. AET is a powerful method to determine the 3D atomic structure of materials without the assumption of crystallinity 19 and has been applied to study dislocations, stacking faults, grain boundaries, atomic displacement, strain tensor, chemical order/disorder and point defects with unprecedented 3D detail 20-26. But all of these studies were of static structures. To probe the 4D atomic structure of early stage nucleation, we have tracked the same nuclei at different times and applied AET to determine their 3D atomic coordinates and species at each time (Methods). We used FePt nanoparticles as a model system because binary alloys have been widely used to study phase transitions 2 and FePt is a very promising material for next generation magnetic recording media 25,27. As-synthesized FePt nanoparticles form a chemically disordered face-centred cubic (fcc) structure (A1 phase) 27. With annealing, the A1 phase Author contributions J.M. conceived and directed the project; F.S. and H.Z. prepared the samples; J.Z.,

A graphene/n-type silicon (n-Si) heterojunction has been demonstrated to exhibit strong rectifying behavior and high photoresponsivity, which can be utilized for the development of high-performance photodetectors. However, graphene/n-Si heterojunction photodetectors reported previously suffer from relatively low specific detectivity due to large dark current. Here, by introducing a thin interfacial oxide layer, the dark current of graphene/n-Si heterojunction has been reduced by two orders of magnitude at zero bias. At room temperature, the graphene/n-Si photodetector with interfacial oxide exhibits a specific detectivity up to 5.77 × 10(13) cm Hz(1/2) W(-1) at the peak wavelength of 890 nm in vacuum, which is highest reported detectivity at room temperature for planar graphene/Si heterojunction photodetectors. In addition, the improved graphene/n-Si heterojunction photodetectors possess high responsivity of 0.73 A W(-1) and high photo-to-dark current ratio of ≈10(7) . The current noise spectral density of the graphene/n-Si photodetector has been characterized under ambient and vacuum conditions, which shows that the dark current can be further suppressed in vacuum. These results demonstrate that graphene/Si heterojunction with interfacial oxide is promising for the development of high detectivity photodetectors.

CO 2 electroreduction facilitates the sustainable synthesis of fuels and chemicals 1 . Although Cu enables CO 2 -to-multicarbon product (C 2+ ) conversion, the nature of the active sites under operating conditions remains elusive 2 . Importantly, identifying active sites of high-performance Cu nanocatalysts necessitates nanoscale, time-resolved operando techniques [3][4][5] . Here, we present a comprehensive investigation of the structural dynamics during the life cycle of Cu nanocatalysts. A 7 nm Cu nanoparticle ensemble evolves into metallic Cu nanograins during electrolysis, before completely oxidizing to single-crystal Cu 2 O nanocubes upon post-electrolysis air exposure. Operando analytical and four-dimensional (4D) electrochemical liquid-cell scanning transmission electron microscopy (EC-STEM) reveals the presence of metallic Cu nanograins under CO 2 reduction conditions. Correlated high-energy-resolution time-resolved Xray spectroscopy suggests that metallic Cu, rich in nanograin boundaries, supports undercoordinated active sites for C-C coupling. The quantitative structure-activity correlation shows a higher fraction of metallic Cu nanograins leads to higher C 2+ selectivity. A 7 nm Cu nanoparticle ensemble, with a unity fraction of active Cu nanograins, exhibits 6 times higher C 2+ selectivity than the 18 nm counterpart with one-third of active Cu nanograins. The correlation of multi-modal operando techniques serves as a powerful platform to advance our fundamental understanding of the complex structural evolution of nanocatalysts under electrochemical conditions.Editor's one-sentence summary: By investigation of structural dynamics during the lifecycle of Cu nanocatalysts, correlation of multimodal operando techniques was found to serve as a powerful platform to advance understanding of their complex structural evolution. Main text:1 Copper remains the only heterogeneous electrocatalyst to selectively catalyze CO 2 reduction reaction (CO 2 RR) to multicarbon (C 2+ ) products, including ethylene, ethanol, and propanol at appreciable rates 1,2 . Recent developments in operando/in situ methods, including advanced electron microscopy and synchrotron-based X-ray methods, provide powerful non-destructive tools to probe active sites and structural changes of electrocatalysts under reaction conditions 3-5 . However, there remains a lingering debate over the active state of Cu catalysts, regarding valence state or coordination environments under CO 2 RR. For instance, some reports have proposed Cu + species and subsurface oxide as possible active sites of oxide-derived Cu electrocatalysts [6][7][8][9] , while others suggested the active state of bulk Cu catalysts is metallic [10][11][12] as subsurface oxides are not stable under negative potentials [13][14][15] . Another possible structural descriptor of locally enhanced CO 2 RR activity has been reported to be micrometer-sized grain boundaries (GBs) on bulk metal electrodes [13][14][15][16][17][18] . Those studies probed the local activity at GBs with a µm-leve...