Feed aggregator

Magneto-ferroelectric effects in van der Waals EuO/graphene heterostructures emerge via a topotactic method, inducing high compressive strain. This strain stabilizes a ferroelectric state up to room temperature, coexisting with the magnetic proximity effect in graphene. These intertwined magneto-electric effects enable control of magnetization and polarization, offering the potential for next-gen memory and neuromorphic devices.

Abstract

2D van der Waals materials and their heterostructures are a fantastic playground to explore emergent phenomena arising from electronic quantum hybridization effects. In the last decade, the spin-dependant hybridization effect pushed this frontier further introducing the magnetic proximity effect as a promising tool for spintronic applications. Here the uncharted proximity-controlled magnetoelectric effect in EuO/graphene heterostructure is unveiled. This is obtained while creating a new multiferroic hybrid heterostructure with multifunctional properties. Using a topotactic method magnetic insulating EuO thin films on graphene is grown under high compressive strain, which induces the appearance of an additional ferroelectric order, with an electric polarization that reaches up to 18 µC cm−2 at room temperature. This observation therefore quantitatively confirms the theoretical predictions made 15 years ago of a strain-induced ferroelectric state in EuO. Moreover, the EuO induces a magnetic proximity state into the graphene layer by interfacial hybridization. This new ferroelectric state in the EuO/graphene heterostructure is stable up to room temperature where it coexists with the EuO/graphene magnetic state. Furthermore, intertwined magneto-electric effects are shown in these strained heterostructures which can facilitate the manipulation of magnetization and electric polarization in future memory and neuromorphic devices.

Energy Environ. Sci., 2025, Accepted Manuscript
DOI: 10.1039/D4EE04861J, PaperXintong Li, guanzhen chen, Yan Liu, Ruihu Lu, Chao Ma, Ziyun Wang, Yunhu Han, Dingsheng Wang
Precisely regulating the electron transfer capacity and Ru-O covalency of RuO2-based catalysts is crucial and challenging for resolving the inadequate performance of RuO2-based acidic oxygen evolution reaction (OER) catalysts in...
The content of this RSS Feed (c) The Royal Society of Chemistry

Nature Materials, Published online: 13 March 2025; doi:10.1038/s41563-025-02182-1

An inhalable nanoplatform responds to inflamed lung tissues by self-assembling into catalytically active fibrillar structures that locally decrease reactive oxygen species, relieve inflammation and alleviate viral pneumonia symptoms.

Nature Energy, Published online: 13 March 2025; doi:10.1038/s41560-025-01742-8

The performance of inorganic, wide-bandgap perovskite solar cells is hindered by unsuitable electron transport layers. Han et al. design an acidic magnesium-doped tin oxide quantum dot layer, improving efficiency and stability in single-junction and tandem cells.

Nature Nanotechnology, Published online: 13 March 2025; doi:10.1038/s41565-025-01874-8

Networking remote superconducting quantum computers requires low-noise microwave-to-optical photon conversion. A transducer based on an integrated silicon electro-optomechanical resonator now achieves below one photon of added noise referred to the transducer input while operating continuously under laser drive.

From two to three dimensions: Impact of stacking order on electronic properties of graphene and graphite

Despite many decades of intense studies, graphite still surprises us by the richness of its electronic and optoelectronic properties. This is largely due to the van der Waals nature of its interlayer bonding coexisting with a substantial hybridization between the electronic states in the consecutive layers. In my talk, I will introduce some of the structures utilizing graphene, a single honeycomb layer of carbon, as a basic building block: from two-dimensional multilayer and twistronic graphene to three-dimensional Bernal and rhombohedral graphite. I will focus on how electronic coupling between the layers allows for engineering of the low-energy electronic properties of the resulting stacks, including formation of topological surface states in rhombohedral graphite. Motivated by this and the experimental breakthroughs in the fabrication of graphene stacks, I will then explore electronic properties of interfaces between semi-infinite crystals: systems with one foot in both the flatland of graphene and the bulk world of graphite. In particular, I will discuss how interfaces between rhombohedral graphite crystals might allow tracking of the transition between topologically non-trivial and trivial electronic phases.

• Speaker: Marcin Mucha-Kruczynski (Bath)
• Friday 14 March 2025, 14:00-15:30
• Venue: TCM Seminar Room.
• Series: Theory of Condensed Matter; organiser: Bo Peng.

Add to your calendar or Include in your list

Multimodal AI in Spatial Biology

The increasing availability and resolution of spatially resolved sequencing on human tissue samples, such as Spatial Transcriptomics (ST), provides rich and spatially resolved molecular information to diagnose and analyse tumours beyond the morphological information routinely available to pathologists through Whole Slide Images. Complex morphological and molecular spatial information becoming available at scale requires building robust multimodal AI architectures that take advantage of such high-dimensional information.

This talk will cover recent advancements from our group in building such models, looking at the intersection of robust multimodal learning, learning from hierarchical structures, and representation learning for spatially resolved transcriptomics.

You can also join us on Zoom

Note: unusual date compared to regular AI Seminars

• Speaker: Konstantin Hemker (University of Cambridge)
• Thursday 13 March 2025, 13:00-14:00
• Venue: Lecture Theatre 2, Computer Laboratory, William Gates Building.
• Series: Artificial Intelligence Research Group Talks (Computer Laboratory); organiser: Mateja Jamnik.

Add to your calendar or Include in your list

From two to three dimensions: Impact of stacking order on electronic properties of graphene and graphite

Despite many decades of intense studies, graphite still surprises us by the richness of its electronic and optoelectronic properties. This is largely due to the van der Waals nature of its interlayer bonding coexisting with a substantial hybridization between the electronic states in the consecutive layers. In my talk, I will introduce some of the structures utilizing graphene, a single honeycomb layer of carbon, as a basic building block: from two-dimensional multilayer and twistronic graphene to three-dimensional Bernal and rhombohedral graphite. I will focus on how electronic coupling between the layers allows for engineering of the low-energy electronic properties of the resulting stacks, including formation of topological surface states in rhombohedral graphite. Motivated by this and the experimental breakthroughs in the fabrication of graphene stacks, I will then explore electronic properties of interfaces between semi-infinite crystals: systems with one foot in both the flatland of graphene and the bulk world of graphite. In particular, I will discuss how interfaces between rhombohedral graphite crystals might allow tracking of the transition between topologically non-trivial and trivial electronic phases.

• Speaker: Marcin Mucha-Kruczynski (Bath)
• Friday 14 March 2025, 14:00-15:30
• Venue: TCM Seminar Room.
• Series: Theory of Condensed Matter; organiser: Bo Peng.

Add to your calendar or Include in your list

Asymmetry in Supposedly Equivalent Facts: Pre-training Bias in Large Language Models

Understanding and mitigating hallucinations in Large Language Models (LLMs) is crucial for ensuring reliable content generation. While previous research has primarily focused on “when” LLMs hallucinate, our work explains “why” and directly links model behaviour to the pre-training data that forms their prior knowledge. Specifically, we demonstrate that an asymmetry exists in the recognition of logically equivalent facts, which can be attributed to frequency discrepancies of entities appearing as subjects versus objects. Given that most pre-training datasets are inaccessible, we leverage the fully open-source OLMo series by indexing its Dolma dataset to estimate entity frequencies. Using relational facts (represented as triples) from Wikidata5M, we construct probing datasets to isolate this effect. Our experiments reveal that facts with a high-frequency subject and a low-frequency object are better recognised than their inverse, despite their logical equivalence. The pattern reverses in low-to-high frequency settings, and no statistically significant asymmetry emerges when both entities are high-frequency. These findings underscore the influential role of pre-training data in shaping model predictions and provide insights for inferring the characteristics of pre-training data in closed or partially closed LLMs.

• Speaker: Zifeng Ding (University of Cambridge)
• Friday 21 March 2025, 12:00-13:00
• Venue: Room SS03 with Hybrid Format. Here is the Zoom link for those that wish to join online: https://cam-ac-uk.zoom.us/j/4751389294?pwd=Z2ZOSDk0eG1wZldVWG1GVVhrTzFIZz09.
• Series: NLIP Seminar Series; organiser: Suchir Salhan.

Add to your calendar or Include in your list

Sex differences in the placenta and autism

Autism and related neurodevelopmental conditions have high heritability and are often attributed to Genetics. Yet males are more likely to be diagnosed, even when the general spectrum of traits or alternative presentations are considered in females. Sex differences in the prenatal environment may then be interacting with genetic variance to ultimately affect neurodevelopment. Recent findings indicate that the placenta may be the key mediator of this interaction and an understudied source of neurodiversity in humans. This is supported by the following lines of evidence. First, steroid hormones such as estradiol are elevated in the fetal and maternal circulation of autistic males and correlate with the development of autistic traits. Second, subtle sex differences in placental function (e.g. in the levels of the placental growth factor) mediate sex differences in the future autistic traits of the offspring. Third, sex differences in placental gene expression are enriched for genes implicated in autism. Fourth, recent assessments of large population registries, such as the MBR in Sweden, show that males are more likely to have placental complications and complicated labour, compared to females, who, in turn, are more able to adjust their growth patterns prenatally. Finally, evolutionary adaptations in the primate lineage may show ‘changes of degree’ in humans, such as increased steroidogenesis, which may be linked to cortical expansion and for understanding the association between the placenta and conditions such as autism.

• Speaker: Dr Alex Tsompanidis, University of Cambridge
• Wednesday 19 March 2025, 11:30-12:30
• Venue: https://us02web.zoom.us/j/87076030035?pwd=XUpJuh8jiR0mae1AhkV79qbg8MtlSM.1.
• Series: ARClub Talks; organiser: Simon Braschi.

Add to your calendar or Include in your list

User-extensible and Productive Programming of Specialized Hardware

As single-core performance has reached its limit, exploiting the peak performance of heterogeneous accelerators and specialized instructions has become crucial in many applications. Compilers struggle to keep pace with the diverse and rapidly evolving hardware targets, and automatic optimization often fails to guarantee state-of-the-art performance. Consequently, high-performance libraries are still commonly coded and optimized by hand, at great expense, in low-level C and assembly. User-schedulable languages (USLs) have been proposed to address this challenge by decoupling algorithms and scheduling. I will share our work on Exo, a USL based on the principle of exocompilation, which externalizes hardware-specific code generation and scheduling library implementation in the user code, decoupled from the compiler. Additionally, I will discuss other projects that borrow ideas from USLs and the lessons we have learned from the industry adoption of Exo.

• Speaker: Yuka Ikarashi (MIT CSAIL)
• Wednesday 09 April 2025, 11:00-12:00
• Venue: Computer Laboratory, William Gates Building, FW11.
• Series: compiler socials; organiser: Emma Urquhart.

Add to your calendar or Include in your list

Title to be confirmed

Abstract not available

• Speaker: Bishakh Gayen, University of Melbourne
• Friday 04 April 2025, 13:00-14:00
• Venue: BAS Seminar Room 1.
• Series: British Antarctic Survey - Polar Oceans seminar series; organiser: Dr Birgit Rogalla.

Add to your calendar or Include in your list

Anisotropic conducting polymer films can exhibit optical hyperbolic properties which here are found connected to large birefringence and linear dichroism. Off-axis twist-stacking of only two such subwavelength films can therefore enable large and broadband circular dichroism (CD), with CD bandwidth closely related to the hyperbolic bandwidth. The study demonstrates the general concept using twist-stacked stretched PEDOT films.

Abstract

Chiral-specific interaction of light with organic materials is important but typically arises from circular polarization-dependent absorption of specific optical transitions, resulting in narrow effective wavelength ranges. This study presents a scalable and universal concept for broadband circular dichroism (CD) enabled by strained conducting polymer thin films that possess in-plane hyperbolic optical behavior (i.e., optically metallic and dielectric properties along orthogonal directions). It is shown that off-axis stacking of two or more such thin films provides broadband CD that varies with the hyperbolic bandwidth and stacking geometry. By contrast to traditional chiroptical materials, the CD can also be modulated by redox-tuning of the hyperbolic polymer properties, opening for broadband dynamic chiroptical components.

A high interfacial compatibility of ionic and electronic conductors is achieved via molecular chain interpenetration. With the benefit of the interface, the fabricated reversible adhesive electrode exhibits a high shear strength of 317 kPa at a low voltage of 4 V, and the ionic diode and transistor maintain stable semiconductor characteristics under arbitrary deformation.

Abstract

Ionic devices find applications such as flexible electronics and biomedicines and function by exploiting hybrid circuits of mobile ions and electrons. However, the poor interfacial compatibility of hard electronic conductors with soft ionic conductors in ionic devices leads to low deformability, sensitivity, electromechanical responses, and stability. Herein, an interpenetrating interface between silicone-modified polyurethane/carbon nanotube electronic conductors and ionoelastomers in an ionic device using in situ polymerization is fabricated. A robust interpenetrating electronic/ionic conductor interface is realized through molecular chain entanglement and molecular forces (such as ion-dipole interactions and H-bonds), effectively enhancing the bonding strength and contact area between the components and resulting in an excellent flexibility, stability, and device performance. The electroadhesive prepared based on this strategy exhibits a superrobust shear strength of 317 kPa under a reduced voltage input of −4 V, and the diode and the transistor can undergo arbitrary deformation while maintaining the semiconductor device characteristics, including rectification and switching. In addition, electromechanical transducers exhibit sensitive electrical responses to various deformation signals. This solution to the interfacial compatibility problems of electronic and ionic conductors holds promise for the development of multifunctional ionic devices.

A compressible, stretchable conductor combining liquid (GalnSn) and solid (BilnSn) metals is designed. The solid layer resists compression, while the liquid layer self-heals cracks under strain. Embedded micropillars enhance durability, enabling stable conductivity under extreme pressure (38.2 MPa) and cyclic loads. This biphasic design allows high-resolution manufacturing of robust antennas and circuits for soft electronics.

Abstract

Compression strongly degrades the electrical conductivity of the liquid-metal-based circuits because the liquid state is prone to be squashed. Here, a new compressible and stretchable biphasic liquid-solid self-healing circuit is proposed by filling GalnSn-BilnSn biphasic metal into micropillar-embedded channels. The underlying BilnSn solid alloy layer serves as a compression resistance layer, while the upper GalnSn liquid metal layer enables the real-time filling of the cracks in the solid layer under large deformations, resulting in autonomous self-healing and maintenance of conductivity under both stretching and compression. The embedded micropillars further improved the compression durability by providing additional mechanical support. The synergistic effect between the biphasic materials and embedded micropillar enables the designed stretchable conductor to show stable performance (R/R0<10) under pressure of 38.2 MPa (≈389.5 Kgf cm−2) and cyclic pressure of 15.8 MPa over 7000 cycles (R/R0<0.48%) without compromising the stretchability, whereas the liquid metal-based conductor can only endure pressure up to 2.5 MPa (25.49 Kgf cm−2). The stretchable antenna and hybrid circuits fabricated using the designed biphasic metal conductor showed enhanced compression durability. The structure-confined filling strategy enabled high-resolution and scalable manufacturing. Overall, robust stability under compression significantly expands the range of possible applications of liquid-metal-based conductors in soft electronics.

An ion-confined transport strategy is developed to construct a bio-inspired retina that processes information through synaptic mechanisms, enabling precise modulation of optical signals. By integrating both inhibitory and excitatory artificial synapses, the system effectively processes various vision tasks, including image recognition, dynamic motion analysis, and path planning for robot vehicles.

Abstract

The effective and precise processing of visual information by the human eye primarily relies on the diverse contrasting functions achieved through synaptic regulation of ion transport in the retina. Developing a bio-inspired retina that uses ions as information carriers can more accurately replicate retina's natural signal processing capabilities, enabling high-performance machine vision. Herein, an ion-confined transport strategy is proposed to construct a bio-inspired retina by developing artificial synapses with inhibitory and excitatory contrasting functions. By fine-tuning the ionic hydrogel structures to control ion transport across the heterogeneous interfaces, inhibitory and excitatory synapses are realized to negatively or positively modulate the optical signal. The integration of these synapses facilitates advanced tasks such as image recognition and motion analysis. Moreover, as a proof of concept, guiding robot vehicles to perform path planning is demonstrated. This work offers a new idea for constructing the bio-inspired retina by precisely regulating ion transport, allowing it to reach a level closer to the biological retina.

A shape change-free and lithium-free anode composed of highly ordered hollow ZnO matrix with surface-protective LiPON layer is developed, which achieves a highly reversible Li plating/stripping inside the matrix cavity for long life-span anode-free lithium metal batteries with high energy density.

Abstract

Anode-free lithium metal batteries are promising toward high-energy-density power sources with low-cost, but their practical applications are challenged by poor cycling stability and low rate capability. Herein, a shape change-free and lithium-free anode that well controls the reversible Li plating-stripping is reported, which is composed of a highly-ordered hollow ZnO matrix with a surface-coated lithium-phosphorus-oxynitride (LiPON) layer. The ZnO matrix supplies sufficient cavities and lithiophilic sites to facilitate uniform Li plating/stripping within the hollow cavity, while the LiPON layer maintains stable solid-electrolyte interphase from mechanical and electrochemical damage. As a result, lithium is constrained within the cavity and the overall anode shape is effectively controlled during long-term and high-rate cycling. The assembled half-cell stably works at 1.2 mA cm−2 for 335 cycles with Coulombic efficiency of 98.8%. Without Li pre-deposition, full-cells using modified-LiFePO4 and LiNi0.5Co0.2Mn0.3O2 cathodes demonstrate 150-cycles lifespan and high energy density of 617 Wh kg−1 at 2C rate.

This work demonstrates an on-chip non-volatile reconfigurable THz topological silicon chip that integrates a waveguide-cavity coupled system with phase-change material, Ge2Sb2Te5, enabling persistent and efficient functionality without constant power input. This advancement significantly paves the way for integrating phase change materials into silicon topological chips for programmable photonic devices, including interconnects, modulators, phase shifters, and logic circuits.

Abstract

Programmable on-chip terahertz (THz) topological photonic devices are poised to address the rising need for high-capacity data systems, offering broad bandwidth, minimal loss, and reconfigurability. However, current THz topological chips rely on volatile tuning mechanisms that require continuous power to function. Here, a nonvolatile, programmable THz topological silicon chip is demonstrated that integrates a waveguide-cavity coupled system with phase-change material, Ge2Sb2Te5 (GST), enabling persistent and efficient functionality without constant power input. Through precise tuning of the intermediate phase states of GST between amorphous and crystalline forms, a stable, non-volatile reconfiguration of the topological cavity is achieved, enabling transitions across over-coupling, critical coupling, and under-coupling states. Multi-level modulation of resonance transmission with a modulation depth of 70 dB is demonstrated, enabling precise control over the onset and disappearance of resonance modes and dynamic tuning of critical coupling states. The THz topological chip facilitates phototunable, volatile modulation across nonvolatile configurations, allowing controlled resetting of the coupling states of the cavity. Here, the first nonvolatile, programmable terahertz topological integrated chip is demonstrated, offering flexible control over resonance modes. This advancement significantly paves the way for integrating phase change materials into silicon topological chips for programmable photonic devices, including interconnects, modulators, and logic circuits.

The actuation of stomata—referring to their opening and closing—is crucial for photosynthesis because it regulates gas exchange and water balance. Here, it is investigated if this life-like property could enhance the photocatalytic properties of soft materials. A hybrid supramolecular and covalent polymer hydrogel is developed to integrate photocatalytic chromophores and stimuli-responsive properties for controlling or improving light-driven superoxide or hydrogen peroxide generation.

Abstract

Mechanical expansion and contraction of pores within photosynthetic organisms regulate a series of processes that are necessary to manage light absorption, control gas exchange, and regulate water loss. These pores, known as stoma, allow the plant to maximize photosynthetic output depending on environmental conditions such as light intensity, humidity, and temperature by actively changing the size of the stomal opening. Despite advances in artificial photosynthetic systems, little is known about the effect of such mechanical actuation in synthetic materials where chemical reactions occur. It is reported here on a hybrid hydrogel that combines light-activated supramolecular polymers for superoxide production with thermal mechanical actuation of a covalent polymer. Superoxide production is important in organic synthesis and environmental remediation, and is a potential precursor to hydrogen peroxide liquid fuel. It is shown that the closing of pores in the hybrid hydrogel results in a substantial decrease in photocatalysis, but cycles of swollen and contracted states enhance photocatalysis. The observations motivate the development of biomimetic photosynthetic materials that integrate large scale motion and chemical reactions.

A novel multifunctional fiber energy storage device consisting of LMO-LTP-AC is developed by the coating-extrusion method. Due to the continuous preparation process, devices exhibited an ultra-high production rate of 6000 km year−1. This innovative approach significantly enhanced the device performance, achieving a specific capacity of 89.4 mAh g−1 for fiber batteries and rate performance of 20 C for fiber hybrid supercapacitors.

Abstract

The rise of wearable electronics demands flexible energy storage solutions like flexible fiber energy storage devices (FESDs), known for their flexibility and portability. However, it remains difficult for existing fabrication methods (typically, finite-coating, thermal-drawing, and solution-extrusion) to simultaneously achieve desirable electrochemical performances and fast production of FESDs. Here, a new scalable coating-extrusion method is developed, utilizing a novel extruded spinneret with tapered apertures to create dual pressure zones. These attributes reduced porosity, enhanced electrode materials loading, and stabilized the interface between the fiber electrode and gel electrolyte of FESDs, enabling the integration of three functional electrodes for the fabrication of both fiber LMO-LTP batteries and fiber LMO/LTP-AC hybrid supercapacitor within a single energy storage device. The resultant multifunctional device achieved a high specific capacity of 89.4 mAh g−1 in battery mode and demonstrated excellent rate performance of 20 C with nearly 50% capacity retention in supercapacitor mode, with a production rate of 6000 km year−1.