About

Hailiang Wang is a Professor of Chemistry and the Director of Graduate Studies at Yale University, a member of the Yale faculty since 2014. His research interests encompass Materials Chemistry, Inorganic Chemistry, and Physical Chemistry, with a focus on developing systems for the precise interconversion of chemical and electrical or solar energy. His work aims to address global challenges related to clean energy, carbon emission mitigation, and environmental protection by developing catalysts and catalytic processes for energy and environmental applications. His research group investigates heterogeneous molecular catalysts for electrochemical conversion of small molecules such as CO2, nitrate, and volatile organic compounds, as well as photochemistry for solar energy utilization, nanoscale interactions for cooperative electrocatalysis, interface chemistry in high-energy batteries, and applications of electrocatalysis and photocatalysis in air and water treatment.

Research topics

• Chemistry
• Organic chemistry
• Nanotechnology
• Physical chemistry
• Composite material
• Combinatorial chemistry
• Materials science
• Inorganic chemistry
• Chemical engineering
• Metallurgy

Selected publications

• Robust Perovskite Solar Cells for Extreme and Emerging Applications

  Advanced Functional Materials · 2025-11-20 · 4 citations

  article1st authorCorresponding

  Abstract Perovskite solar cells (PSCs), with their lightweight nature, ultrahigh power conversion efficiency, and tunable optoelectronic properties, offer unprecedented opportunities beyond the scope of traditional photovoltaics. These advantages have accelerated their exploration in emerging and extreme‐use scenarios, including space‐based systems, indoor light harvesting, concentrated photovoltaics, and flexible or wearable electronics. However, current studies on PSCs deployment in such specialized environments remain fragmented, and a critical and in‐depth understanding of their resilience under coupled external stressors is still lacking. This review pioneers a hierarchical dissection spanning application‐specific demands, device design principles, and perovskite material fundamentals, enabling the systematic identification of dominant degradation pathways across diverse operational contexts. It is critically assessed, for the first time, how extreme conditions (such as high/low temperatures, ionizing radiation, variable illumination, and mechanical deformation in flexible systems) impact photovoltaic performance and the underlying mechanisms. The discussion further unravels key challenges in the field, including long‐term material stability, interface failure, and mechanical fatigue, and explores future directions such as advanced degradation studies, machine learning assisted material design, multifunctional interface engineering, and planar‐to‐fiber architectures evolution. This review aims to bridge fundamental understanding with application‐specific needs, guiding the development of robust PSCs tailored for next‐generation, mission‐critical applications.

  PublisherDOI

• The importance of CO supersaturation and surface area—not geometry—for tandem single-catalyst CO₂ reduction to CH₃OH

  ChemRxiv · 2025-10-17

  preprint

  Effective electrochemical CO2 reduction to liquid fuels requires that the local catalytic environment facilitates the desired reactivity, yet a microscopic understanding of this environment is difficult to achieve from experiment alone. In this work, a 3D reaction-diffusion model was developed to explore the effects of electrode surface area and local geometry on the performance of a heterogeneous catalyst that performs a two-step CO2 reduction cascade reaction to CO and then methanol under aqueous conditions. Kinetic parameters for the model were directly motivated by experimental results that used carbon electrodes functionalized with cobalt phthalocyanine (CoPc) catalysts. 3D architectures composed of arrays of square pillars with varying dimensions and either smooth or periodically modulated surfaces were tested, revealing the extent to which geometry modulates the performance of the cascade reactions. Surprisingly, the model illustrates that electrochemically-active surface area, regardless of electrode geometry, is the primary factor dictating the overall cascade reaction yield. Moreover, the results suggest that supersaturation of CO, with concentrations up to ten-fold higher than the equilibrium solubility limit, is critical for more efficient conversion to methanol. For any given geometry, the spatially averaged ratio of [CO] to [CO₂] is dictated by the electrochemically active surface area and determines the yield of methanol. For a fixed surface area, geometries that spatially confine the electrolyte yield moderate local [CO] to [CO₂] ratios within small volumes. In contrast, less confining geometries result in a broader distribution of local ratios spread over larger volumes with both configurations yielding the same spatially averaged [CO] to [CO₂] ratio. These insights provide valuable design principles—highlighting the critical importance of surface area and CO supersaturation—for engineering advanced electrode architectures that leverage intermediate trapping and CO supersaturation to enhance overall performance in tandem CO₂ reduction systems.

  PublisherDOI

• Bond performance between hooked-end steel fibers and hydrophobic fibre reinforced concrete under freeze-thaw cycling

  Construction and Building Materials · 2025-09-08 · 4 citations

  article
  PublisherDOI

• Fault Diagnosis of HCSY-MG Microsource Power Abnormal Fluctuation Based on Multi-Feature Fusion CNN

  2025-03-28

  articleSenior author

  In a Half-bridge Converter Series Y-connection Microgrid (HCSY-MG), each bridge arm is connected with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$N$</tex> Generation modules (GM) in series. The three bridge arms are connected and combined in a star configuration to form a three phase system. To accurately diagnose the system imbalance problem caused by each micro-source fault in HCSY-MG, this paper adopts a Fast Fourier Transform (FFT). In this method, the in circuit signals such as voltage and current in HCSY-MG are taken as input signals for FFT and input to a one-dimensional CNN, and also as input signals for Wavelet Time Frequency Representation (WTFG) and input to a two-dimensional CNN. Finally, the obtained voltage and current signals are enhanced by the multi-head attention mechanism (MHAM). The network can learn information in different feature spaces in parallel, thus improving the accuracy and comprehensiveness of fault diagnosis.

  PublisherDOI

• Vacuum thermal activation-driven surface morphological evolution and hydrogen sorption kinetics of amorphous TiZrV non-evaporable getter films

  Applied Surface Science · 2025-11-07

  article
  PublisherDOI

• Construction of a novel prediction model based on albumin‑hemoglobin score and serum microRNA-497‑5p for prognosis of patients with stage II‑III colorectal cancer

  Oncology Letters · 2025-12-17

  articleOpen access

  Stage II-III colorectal cancer (CRC) represents the initial phase of tumor invasion and lymph node metastasis. Surgery is the preferred treatment option for patients at this stage, however, due to the complex progression mechanisms of the disease, incomplete tumor resection, postoperative cancer cell metastasis, and the development of drug resistance to chemotherapy often lead to poor prognosis. The present study aimed to investigate the status of poor postoperative prognosis in patients with stage II-III CRC following radical surgery, analyze the impact of preoperative albumin-hemoglobin (ALB-Hb) score and serum microRNA-497-5p (miR-497-5p) levels on adverse outcomes and develop a predictive model for poor prognosis. Baseline data of 154 patients with stage II-III CRC treated at Tianjin Haihe Hospital (Tianjin, China) or Guangzhou First People's Hospital (Guangzhou, China) between December 2021 and December 2024 were retrospectively selected. Patients were stratified into the poor and good prognosis groups based on clinical outcomes. Univariate analysis was performed on baseline characteristics and laboratory parameters. Variables demonstrating notable differences were subsequently evaluated for multicollinearity. Factors without collinearity were incorporated into a Cox proportional hazards regression model to identify determinants of postoperative prognosis in patients treated with radical resection. These predictors were then used to construct a prognostic nomogram, with model accuracy verified through calibration curves. Among the 154 patients with stage II-III CRC, 63 cases (40.91%) had poor prognosis and 91 cases (59.09%) had good prognosis. Univariate and collinearity analyses revealed marked differences in preoperative levels of miR-497-5p, Kallikrein-related peptidase 5 (KLK5), angiopoietin-2 (Ang-2) and ALB-Hb scores, with no collinearity observed (P<0.05; variation inflation factor ≤10; tolerance ≥0.1). Cox proportional hazards regression model demonstrated that all these indicators were independent factors influencing poor prognosis following radical surgery in patients with stage II-III CRC (all P<0.05). Based on these findings, a nomogram was constructed and calibration curves closely approximated the ideal curve. Preoperative ALB, Hb, miR-497-5p, KLK5, Ang-2 levels and ALB-Hb scores were notable prognostic factors for patients with stage II-III CRC following radical resection, demonstrating high predictive value for poor postoperative outcomes. The present study provided clinically relevant indicators to screen high-risk patients with potential poor prognosis following radical surgery for stages II-III CRC.

  PublisherOA PDFDOI

• The Reliability Fallacy: How Label Ambiguity Undermines AI Hate Speech Detection

  Preprints.org · 2025-11-25 · 1 citations

  preprintOpen access1st authorCorresponding

  Automated content moderation is a critical AI security task. However, models often fail in the nuanced, subjective task of distinguishing “hate” from “offensive” speech. The influential ‘HateXplain‘ benchmark attributed this poor performance to a lack of model explainability, proposing rationale-based training as a solution. In this paper, we challenge this premise. We hypothesize that the models’ unreliability stems from a more fundamental, unaddressed security flaw: a crisis of data integrity caused by high label ambiguity. The original dataset relies on a “majority vote” to assign groundtruth labels, which masks significant annotator disagreement and introduces noise. To test our hypothesis, we isolate this variable. We partition the ‘HateX- plain‘ dataset into two cohorts: (1) a “noisy” Majority-Label set (using standard 2-1 majority votes) and (2) a “clean” Pure-Label set (using only 3-0 unanimous-consensus votes). We then rigorously benchmark five models (Logistic Regression, Random Forest, LightGBM, GRU, and AL- BERT) on both datasets. Our results are conclusive. All models trained on the “Pure-Label” data achieved statistically significant and substantially higher performance. The ALBERT model’s weighted F1-score, for instance, rose from 0.7447 on the “noisy” data to 0.8126 on the “clean” data. This demonstrates that label ambiguity is a more dominant performance bottleneck than the architectural factors previously considered. We conclude that for building secure and reliable AI safety systems, addressing foundational data integrity and label consensus is a more critical challenge than model-level explainability.

  PublisherOA PDFDOI

• Output Power Allocation Control Strategy Considering Strong Nonlinearity of HCSY-MG Grid-Connected System

  IEEE Journal of Emerging and Selected Topics in Power Electronics · 2025-08-04

  article

  The half-bridge converter series Y-connection microgrid (HCSY-MG), as a novel type of series-connected microgrid, has received limited attention regarding output power allocation and control. Moreover, existing power allocation and control strategies for microgrids are not directly applicable to HCSY-MG grid-connected system. To achieve the output power distribution and control of the HCSY-MG grid-connected system across various generation modules (GMs), while considering the system’s strong nonlinearity constraints, a control strategy for output power distribution based on the Lagrange-proximal policy optimization (Lagrange-PPO) algorithm was proposed. First, the output power mathematical model of the HCSY-MG grid-connected system was established under the carrier disposition sinusoidal pulse width modulation (CD-SPWM) method. Based on the output power range of each GM and microsource output power, the corresponding relationship between each GM and the carrier layers was derived. Then, considering the strong nonlinearity in the output power constraints of each GM under the CD-SPWM strategy, the output power control problem is reformulated as a constrained Markov decision process (MDP) to reduce algorithmic complexity. To address this constrained MDP, a Lagrange-proximal policy optimization algorithm is proposed. Finally, the feasibility and effectiveness of the proposed strategy, along with its superiority over conventional methods, are validated through both simulations and experimental comparisons.

  PublisherDOI

• Roof damage detection and evaluation using aerial image based on improved DeepLabv3+

  Nondestructive Testing And Evaluation · 2025-08-24

  articleSenior author
  PublisherDOI

• Two-dimensional materials based two-transistor-two-resistor synaptic kernel for efficient neuromorphic computing

  Nature Communications · 2025-05-09 · 26 citations

  articleOpen access

  Neuromorphic computing based on two-dimensional materials represents a promising hardware approach for data-intensive applications. Central to this new paradigm are memristive devices, which serve as the essential components in synaptic kernels. However, large-scale implementation of synaptic matrix using two-dimensional materials is hindered by challenges related to random component variation and array-level integration. Here, we develop a 16 × 16 computing kernel based on two-transistor-two-resistor unit with three-dimensional heterogeneous integration compatibility to boost energy efficiency and computing performance. We demonstrate the 4-bit weight characteristics of artificial synapses with low stochasticity. The synaptic array demonstration validates the practicality of utilizing emerging two-dimensional materials for monolithic three-dimensional heterogeneous integration. Additionally, we introduce the Gaussian noise quantization weight-training scheme alongside the ConvMixer convolution architecture to achieve image dataset identification with high accuracy. Our findings indicate that the synaptic kernel can significantly improve detection accuracy and inference performance on the CIFAR-10 dataset. He et al. report a two-transistor-two-resistor core unit based synaptic kernel with monolithic 3D heterogeneous integration compatibility, showing compelling performance, yield (91.2%), and component uniformity. Synaptic kernel supporting differential weight achieves beyond 85% CIFAR-10 recognition.

  PublisherOA PDFDOI

Recent grants

• Developing Sulfur Cathode Materials for Electrochemical Energy Storage

  NSF · $522k · 2019–2022

• Interparticle Metal-Metal Interactions in Electrocatalytic Carbon Dioxide Reduction Reactions

  NSF · $445k · 2020–2023

• CAREER: Heterogeneous Molecular Catalysts for Electrochemical CO2 Reduction

  NSF · $700k · 2017–2023

• NSF-BSF: Deciphering Molecule-Carbon Nanotube Interactions for Environmental Remediation Reactions

  NSF · $547k · 2021–2025

• CAS: Nitrogen-Coupled Carbon Dioxide Conversion to Methylamine: Molecular Level Understanding and Tailoring of the Electrocatalysis

  NSF · $547k · 2022–2025

Frequent coauthors

• Hongjie Dai

  Southwest University

  62 shared

• Yongye Liang

  Southern University of Science and Technology

  45 shared

• Wen Liu

  Collaborative Innovation Center of Chemical Science and Engineering Tianjin

  44 shared

• Yueshen Wu

  ShanghaiTech University

  43 shared

• Yiren Zhong

  Beijing Normal University

  43 shared

• Zishan Wu

  China Agricultural University

  33 shared

• Conor L. Rooney

  Yale University

  29 shared

• Gábor A. Somorjai

  University of California, Berkeley

  27 shared

Education

• Postdoctoral Fellow, Chemistry

  University of California, Berkeley

  2014

• PhD, Chemistry

  Stanford University

  2012

• BS, Chemistry

  Peking University

  2007

Awards & honors

• Young Investigator Award, Division of Inorganic Chemistry, A…
• IUPAC Prize for Young Chemists (2013)
• Web of Science Highly Cited Researcher (Chemistry) (2016-pre…
• Emerging Investigator, J. Mater. Chem. A, Royal Society of C…
• NSF CAREER Award (2017)