Sunday, July 26, 2026

Team develops battery component designed to make lithium metal batteries safer and more powerful



Novel organic-inorganic composite electrolyte




Tsinghua University Press

Asymmetric quasi-solid-state composite electrolyte 

image: 

An asymmetric quasi-solid-state composite electrolyte was constructed by evaporation-induced self-assembly of 1D molecular brushes and 0D hairy ceramic nanoparticles, followed by in situ polymerization. The resulting ultrathin quasi-solid-state composite electrolyte delivers a high ionic conductivity and a lithium-ion transference number of 0.67, and enables lithium iron phosphate cells to operate for 200 cycles with 94% capacity retention at 0.5 C.

view more 

Credit: Energy Materials and Devices, Tsinghua University Press





A research team has developed a new type of battery component that results in safer and higher-capacity batteries. It can improve lithium metal batteries’ safety, reliability, and electrochemical performance. Their work provides a feasible molecular engineering strategy for developing safe, high-performance lithium metal batteries.

 

This research was published in the journal Energy Materials and Devices on June 24, 2026.

 

“Simply put, we aimed to make a safer, thinner, and faster-charging lithium metal battery free of internal short-circuit risks,” said Yin Cui, Sun Yat-sen University. The development of electric vehicles and smart grids has led to a growing demand for batteries that are safer and have greater energy density. Traditional lithium-ion batteries cannot meet these needs. Lithium metal batteries can store far more energy than conventional lithium-ion batteries, but they are dangerous because lithium metal tends to form dendrites that can pierce the separator and trigger internal short circuits.

 

In recent years, scientists have explored solid-state composite electrolytes. These hold promise as electrolyte systems for solid-state lithium-metal batteries. However, their application in lithium-metal batteries has been restricted because of challenges, such as poor interfacial compatibility between different functional components, low ionic conductivity at room temperature, and difficulty in completely inhibiting the growth of lithium dendrites.

 

Quasi-solid-state composite electrolytes have distinctive multilayer architectures that allow function customization based on the cathode and anode requirements. Yet the current quasi-solid-state composite electrolytes lack the qualities needed to attract and guide the lithium ions. These electrolytes also work poorly in colder temperatures. There are still significant challenges in the current development of asymmetric quasi-solid-state composite electrolytes.

 

The research team set out to design a novel organic-inorganic composite electrolyte for longer-lasting, safer batteries. Their core goal was to develop an asymmetric quasi-solid-state composite electrolyte through evaporation-induced self-assembly of molecular brushes and hairy nanoparticles, followed by in situ cationic ring-opening polymerization. “This asymmetric electrolyte can not only achieve intimate interfacial contact with the cathode, but also effectively suppress lithium dendrites growth on the anode side, significantly improving the overall safety, reliability, and electrochemical performance of lithium metal batteries,” said Cui.

 

The team notes that constructing an asymmetric double-sided electrolyte represents an extremely practical route to developing safe and high-performance lithium metal batteries. Their ultrathin electrolyte (19 μm) has two different functional sides: a rigid ceramic-rich side blocking lithium dendrites physically and a flexible polymer composite side ensuring tight contact with the cathode. These polymer chains on the surface of nanomaterials can evenly distribute lithium ions and simultaneously speed up lithium-ion transfer.

 

“This design delivers three standout real-world advantages: great ionic conductivity, a high lithium-ion transference number, and stable long-cycle battery life,” said Cui.  By integrating 1D molecular brushes and 0D hairy ceramic nanoparticles, the team created a continuous 3D ion transport network. The new material they created helps the charged particles move much more efficiently. This molecular synergy is the core innovation behind their excellent electrochemical data.

 

The team’s study showed that designing unique asymmetric quasi-solid-state composite electrolytes via molecular engineering is a promising research direction to fundamentally promote lithium-ion conduction and stabilize the lithium anode. This method provides a viable strategy for the practical application of high-performance solid-state lithium-metal batteries.

 

Looking ahead, the team’s next step is to scale up this fabrication process and test the electrolyte in larger‑format batteries, not just coin cells. They also aim to further optimize the composition to boost ionic conductivity and verify long‑term stability under practical operating conditions, especially during fast charging. “Our ultimate goal is to enable practical, safe lithium metal batteries that outperform today’s lithium‑ion systems in electric vehicles, drones, and portable electronics, especially in cold climates where current batteries struggle. In the longer term, we hope this molecular self‑assembly approach can be adapted to other solid‑state battery chemistries, such as sodium or potassium, broadening the impact beyond lithium,” said Cui.

 

The research team includes Shenghao Lin, Yin Cui, Guofang Yu, Dongtian Miao, and Dingcai Wu from the School of Chemistry, Sun Yat-sen University, Guangzhou, China, and Ruliang Liu from the School of Chemistry and Materials Science, Guangdong University of Education, Guangzhou, China.

 

This research is funded by the National Key Research and Development Program of China; the National Natural Science Foundation of China; the Guangdong Major Project of Basic and Applied Basic Research; the Natural Science Foundation of Guangdong; the Fundamental Research Funds for the Central Universities, Sun Yat-sen University; the Science and Technology Program of Guangzhou; and the Guangdong Basic Research Center of Excellence for Functional Molecular Engineering. 

 

DOI Link:

https://doi.org/10.26599/EMD.2026.9370095

 

Can large language models capture human risk preferences?



Tsinghua University Press
Distributions of estimated risk attitudes across real and LLM-simulated data 

image: 

Distributions of estimated risk attitudes across real and LLM-simulated data

view more 

Credit: Communications in Transportation Research





Large language models (LLMs) are increasingly used as agents to simulate human behavior, yet their fidelity in complex decision-making under uncertainty remains insufficiently understood. To address this gap, we develop a comparative framework that benchmarks LLM-simulated risk preferences against empirical human behavior. Using demographic profiles from surveys conducted in Sydney, Hong Kong, and Nanjing, we construct role-playing prompts and evaluate three LLMs on abstract lottery-choice tasks. We adopt the classical Constant Relative Risk Aversion (CRRA) framework as a domain-neutral “standard ruler” to compare risk attitudes. The analysis yields three main findings. First, off-the-shelf LLMs do not exhibit a universal risk profile: the two GPT models are more risk-averse than human benchmarks, whereas Gemini is more risk-seeking. Second, prompt language systematically affects simulated risk attitudes, with English-to-Chinese switching inducing a more conservative shift in most cases. Third, LLMs do not reliably reproduce the empirical heterogeneity of human risk preferences, tending either to generate overly concentrated distributions or unrealistically large dispersion. Taken together, these findings show that off-the-shelf LLMs remain vulnerable to model-family-specific miscalibration, language-sensitive distortions, and failures in distributional fidelity. Rigorous empirical calibration is therefore necessary before off-the-shelf LLMs can be reliably deployed in computational social science and choice modeling.

The team published their study in Communications in Transportation Research (https://doi.org/10.26599/COMMTR.2026.9640025).

Our findings highlight an important limitation of using off-the-shelf LLMs as tools for behavioral prediction. Since different model families exhibit different baseline calibration biases, the choice of model can materially affect the inferred pattern of public risk preferences. In practice, one model family may overstate conservatism, whereas another may overstate willingness to accept risk. Without empirical calibration, such biases can distort inference and lead researchers to draw policy conclusions that do not accurately reflect observed human behavior.

This limitation is particularly consequential in transportation research, where risk perception is central to decision-making under uncertainty. Travel behavior frequently involves probabilistic trade-offs, including route choice under unreliable travel times, mode switching during service disruptions, and the adoption of emerging mobility technologies under safety and performance uncertainty. If the underlying risk preference parameters are systematically miscalibrated, demand forecasts, welfare evaluation, and policy design may all be biased. For example, using uncalibrated LLM-generated data to infer willingness to adopt safety-critical systems such as autonomous vehicles or low-altitude mobility services could yield either overly conservative or overly optimistic projections, depending on the model family used.

The multilingual results add a further layer of caution. In linguistically diverse settings, prompt language is not a neutral implementation choice: it can systematically perturb the behavioral calibration of the model. This is especially relevant for transportation systems serving multilingual populations, where researchers may be tempted to use native-language prompting as a straightforward way to improve realism. Our results suggest that such an assumption is unwarranted unless the model has first been validated against human benchmarks in the relevant linguistic context.

Overall, the implications of this study are methodological as much as substantive. Off-the-shelf LLMs should not be treated as direct substitutes for human respondents in risk-sensitive behavioral applications. Instead, they should be regarded as tools whose outputs require domain-specific and language-sensitive calibration. Rigorous validation against human ground truth remains a necessary prerequisite for deploying these models in transportation and other cross-cultural social science settings.

 

DOI Link:

https://doi.org/10.26599/COMMTR.2026.9640025

About Communications in Transportation Research

Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University, and Professor Xiaopeng (Shaw) Li from University of Wisconsin–Madison. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.

It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan Phase Ⅱ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in "TRANSPORTATION" category. In 2026, its 2025 IF was announced as 12.7, maintaining the Top1 position (1/66, Q1) in the same category.

From Volume 6 (2026), Communications in Transportation Research will be published by Tsinghua University Press on the SciOpen platform with the official journal website at https://www.sciopen.com/journal/2097-5023. We kindly request that all new manuscript submissions be made through the journal’s submission system at https://mc03.manuscriptcentral.com/commtr. For any submission-related inquiries, please contact the Editorial Office at commtr_e@mail.tsinghua.edu.cn.

 

Exploring the role of knowledge graphs and large language models in autonomous vehicle safety




Tsinghua University Press
The framework of the review 

image: 

 The framework and structure of this review

view more 

Credit: Communications in Transportation Research





In a new comprehensive review, researchers from Tsinghua University, Xiaomi EV, Beijing Institute of Technology, and University of Electronic Science and Technology of China examine how these two AI paradigms are being applied across the development, validation, and operation of autonomous driving systems, discuss their strengths and limitations, and further review their emerging synergy for more robust, interpretable, and trustworthy autonomous driving systems. 

The team published their study in Communications in Transportation Research (https://doi.org/10.26599/COMMTR.2026.9640023).

Two complementary AI paradigms for autonomous driving safety

Autonomous vehicles must operate in highly dynamic and uncertain traffic environments. Although recent advances in sensing and learning have improved perception and control, current systems still face major challenges in scene understanding, risk reasoning, rare scenario handling, and explainability.

In the study, the research team examined two important AI approaches with complementary strengths. Knowledge graphs represent structured knowledge such as traffic rules, object relationships, causal chains, and expert driving experience, enabling explicit reasoning and traceable decision support. Large language models, in contrast, are more flexible in semantic understanding, contextual generalization, and reasoning over open-ended situations.

Comparison and synergy of these two AI paradigms for autonomous driving safety

The review shows that knowledge graphs and large language models offer different advantages for autonomous driving safety. Knowledge graphs are well suited for organizing structured knowledge and supporting explicit, interpretable reasoning, especially in tasks involving traffic rules, causal relations, and expert knowledge. Large language models, in contrast, are more flexible in semantic understanding, contextual reasoning, and handling open-ended situations, which gives them strong potential in complex and uncertain traffic environments.

Rather than treating them as competing approaches, the study further highlights their growing synergy. Researchers are increasingly exploring how structured knowledge can improve the transparency and reliability of large language models, while the flexible reasoning ability of large language models can expand how knowledge graphs are queried, understood, and applied. This emerging combination points to a promising direction for building more robust, interpretable, and trustworthy autonomous driving systems.

DOI Link:

https://doi.org/10.26599/COMMTR.2026.9640023

About Communications in Transportation Research

Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University, and Professor Xiaopeng (Shaw) Li from University of Wisconsin–Madison. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.

It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan Phase Ⅱ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in "TRANSPORTATION" category. In 2026, its 2025 IF was announced as 12.7, maintaining the Top1 position (1/66, Q1) in the same category.

From Volume 6 (2026), Communications in Transportation Research will be published by Tsinghua University Press on the SciOpen platform with the official journal website at https://www.sciopen.com/journal/2097-5023. We kindly request that all new manuscript submissions be made through the journal’s submission system at https://mc03.manuscriptcentral.com/commtr. For any submission-related inquiries, please contact the Editorial Office at commtr_e@mail.tsinghua.edu.cn.