Publications

2026

1. CVPR-W

   CADLoop: An Equivariant-Aware Skill-Grounded Loop for CAD Data Curation

   Haozhe Zhang, Miaomiao Chen, and Kaichen Liu

   In CVPR 2026 Workshop on Exploring the Next Generation of Data (NeXD) 2026

   Abs URL

   High-quality programmatic CAD datasets are critical for training multimodal models capable of generating executable geometry. However, automatically generated CAD programs frequently contain small geometric inconsistencies that invalidate otherwise correct structures, making large-scale dataset construction difficult. We observe that a large fraction of failures fall into a near-miss regime (IoU approx 0.8), where the generated topology is correct but continuous parameters are misaligned. These cases contain valuable supervision signals but are typically discarded due to strict geometric verification. We introduce CADLoop, a neuro-symbolic data curation pipeline that automatically verifies and repairs such near-miss programs. CADLoop decouples structural reasoning from deterministic geometric execution: a Vision-Language Model (VLM) infers the high-level design intent, while a symbolic geometric module analytically resolves continuous constraints and directly modifies the CAD program’s abstract syntax tree (AST). Evaluated on 1,619 CAD models from the Fusion360 Gallery, CADLoop identifies 226 near-miss cases and successfully repairs all of them, improving the pass rate from approximately 91 percent with Re-Act loops to 100 percent while substantially reducing geometric error. The resulting dataset contains 1,646 verified program-geometry pairs with IoU >= 0.99. Our results demonstrate that deterministic geometric verification can transform imperfect generations into high-quality training assets, enabling scalable construction of datasets for future multimodal CAD generation models.

2. arXiv

   BenchCAD: A Comprehensive, Industry-Standard Benchmark for Programmatic CAD

   Haozhe Zhang, Kaichen Liu, Miaomiao Chen, Lei Li, Shaojie Yang, Cheng Peng, and Hanjie Chen

   arXiv preprint arXiv:2605.10865 2026

   Abs URL Project

   We present BenchCAD, a comprehensive industry-standard benchmark for evaluating programmatic CAD code generation. The benchmark includes over 17,000 execution-verified CadQuery programs spanning 106 industrial part families (e.g., gears, springs, drills) and assesses models across visual question answering, code analysis, image-to-code conversion, and code editing. Our results show that current models often recover coarse outer geometry but fail to produce faithful parametric CAD programs, with common failure modes including missing structural details and oversimplified operations; generalization to unseen part families remains limited even after fine-tuning.

3. IJHMT

   Enhancing device-to-substrate thermal transport by limiting the flexural rigidity of heat-spreading interlayers

   Xiaotong Yu, Kai Chen, Weiyan Chen, Ziqiao Chen, Haozhe Zhang, Baoxing Xu, Rong Chen, and Yuan Gao

   International Journal of Heat and Mass Transfer 2026

   Abs URL

   Heat dissipation across the device-to-substrate interface is a critical bottleneck for high-power-density electronics. This work shows that limiting the flexural rigidity of heat-spreading interlayers enhances conformal contact and significantly improves device-to-substrate thermal transport, providing a design rule for thermal interface architectures in advanced electronic packaging.

2025

1. Nano Lett.

   Degradation of Hydrogen Bonds Enormously Enhances Convective Heat Transfer in Nanofluidics

   Ziqiao Chen, Rui He, Xiaotong Yu, Haozhe Zhang, Rong Chen, Baoxing Xu, and Yuan Gao

   Nano Letters 2025

   Abs URL

   Convective heat transfer in nanochannels can be enhanced by approximately 50% due to the deterioration of hydrogen bonds under internal stress. Degraded hydrogen bonds weaken the intrinsic constraints from local confinement, significantly promoting the mobility of confined liquid molecules and facilitating phonon transmission for rapid heat transfer. A thermal–mechanics scaling law correlates internal stress with convective heat transfer through the Nusselt number and the liquid–solid interaction energy, providing theoretical insights for designing nanofluidics-based cooling strategies for high-power-density electronics.

2. Mech. Mater.

   A buckling mechanics model for pattern transformation of lattice superstructures assembled by soft-hard materials integrated units

   Haozhe Zhang, Yuan Gao, and Baoxing Xu

   Mechanics of Materials 2025

   Abs URL Project

   We develop a buckling mechanics model for the pattern transformation of lattice superstructures assembled by soft-hard material integrated units. The model captures how the soft-hard contrast governs the critical buckling load and post-buckling reorganization of unit rotations, predicting the emergence of distinct ordered patterns. Theory and finite-element analysis agree on how unit geometry and material contrast steer pattern selection, offering a route to program reconfigurable lattice metamaterials.

2024

1. Sci. Adv.

   Extremely durable electrical impedance tomography–based soft and ultrathin wearable e-skin for three-dimensional tactile interfaces

   Kyungbin Kim, Jung-Hoon Hong, Kyungtaek Bae, Kunhee Lee, Dong Jun Lee, Junseong Park, Haozhe Zhang, Mingyu Sang, Ji-Eun Ju, and  others

   Science Advances 2024

   Abs URL Project

   We introduce a thin and soft electrical impedance tomography (EIT) based e-skin for tactile interfaces. The conductive and stretchable sensing domain uses a micropatterned multi-walled carbon nanotube/elastomer composite. The EIT-based tactile interface discerns external pressures to a sub-millimeter degree and vertical deformations of a few hundred micrometers, sustaining stable functions under damage or environmental changes. We demonstrate real-time human–machine interface applications including handwriting recognition and drone control.

2. ACS AMI

   Spongy Ag Foam for Soft and Stretchable Strain Gauges

   Seungse Hong, Haozhe Zhang, Jiwoo Lee, Taemin Yu, Sungjun Cho, Taeyeong Park, Jacob Walsh, Yuhang Ji, Jin Jea Kim, Hyowon Lee, and  others

   ACS Applied Materials & Interfaces 2024

   Abs URL

   We present strain gauges built from spongy Ag foam via a steam-etching process, a simpler and more eco-friendly alternative to conventional fabrication. The gauges achieve a gauge factor exceeding 8 x 10^3 at strains over 100% while preserving stretchability and durability. Systematic studies reveal the structure-property-performance relationships, and we demonstrate the gauges in human-motion monitoring and human-robot interaction.

2023

1. Adv. Mater.

   A Nanoconfined Water–Ion Coordination Network for Flexible Energy-Dissipation Devices

   Yuan Gao, Mingzhe Li, Chuanyong Zhan, Haozhe Zhang, Mengtian Yin, Weiyi Lu, and Baoxing Xu

   Advanced Materials 2023

   Abs URL

   We report a water-ion coordination network of water-water, water-ion, and ion-ion interactions in LiCl solution confined in a hydrophobic nanopore. Confinement enhances the interaction intensity nearly fourfold relative to bulk, strongly couples with the interaction correlation time, and drives formation of local ion clusters and bound water. Unlike localized effects in bulk liquids, the nanoconfined network globally suppresses rotational and translational dynamics. Simulations and experiments demonstrate that this can be leveraged in a reusable flexible liquid-nanopore energy-dissipation system that attenuates pressure during impact unloading.

2. Adv. Mater.

   Mechanical Janus Structures by Soft‐Hard Material Integration

   Haozhe Zhang, Weizhu Yang, Qingchang Liu, Yuan Gao, Zhufeng Yue, and Baoxing Xu

   Advanced Materials 2023

   Abs URL Project

   Engineering Janus structures that possess anisotropic features in functions have attracted growing attention for a wide range of applications in sensors, catalysis, and biomedicine, and are yet usually designed at the nanoscale with distinct physical or chemical functionalities in their opposite sides. Inspired by the seamless integration of soft and hard materials in biological structures, here we present a mechanical Janus structure composed of soft and hard materials with a dramatic difference of mechanical properties at additively manufacturable macroscale. We establish, in the combination of extensive experimental, theoretical, and computational studies, the design principle of soft-hard materials integrated mechanical Janus structures and address their unique rotation mechanism. We further show the systematic studies of assembling the Janus structure units into superstructures with well-ordered organizations by programing their local rotations, providing a direct route of designing superstructures by leveraging mechanical Janus structures with unique soft-hard material integration. Applications are conducted to demonstrate the features and functionalities of assembled superstructures with local ordered organizations in regulating and filtering acoustic wave propagations, thereby providing exemplification applications of mechanical Janus design in functional structures and devices.

2022

1. ACS Nano

   Electrically Suppressed Outflow of Confined Liquid in Hydrophobic Nanopores

   Yuan Gao, Mengtian Yin, Haozhe Zhang, and Baoxing Xu

   ACS nano 2022

   Abs URL Project

   Confining liquid in a hydrophobic nanoenvironment has enabled a broad spectrum of applications in biomedical sensors, mechanical actuators, and energy storage and converters, where the outflow of confined liquid is spontaneous and fast due to the intrinsic hydrophobic nature of nanopores with extremely low interfacial friction, challenging design capacity and control tolerance of structures and devices. Here, we present a facile approach of suppressing the outflow of water confined in hydrophobic nanopores with an electric field. Extensive molecular dynamics simulations show that the presence of an electric field could significantly strengthen hydrogen bonds and retard degradations of the associated networks during the outflow. The outflow deformation and strength are extracted to quantitatively characterize the electrical suppression to outflow and agree well with simulations. This study proposes a practical means of impeding the fast liquid outflow in hydrophobic nanopores, potentially useful for devising nanofluidics-based functional structures and devices with controllable performance.

2. Matter

   Anomalous solid-like necking of confined water outflow in hydrophobic nanopores

   Yuan Gao, Mingzhe Li, Haozhe Zhang, Yue Zhang, Weiyi Lu, and Baoxing Xu

   Matter 2022

   Abs URL Project

   Outflow of water confined in hydrophobic nanopores exhibits an unexpected solid-like necking behavior reminiscent of metals under uniaxial tension. Molecular dynamics simulations and theoretical analysis reveal that hydrogen-bond networks and confinement-induced stress give rise to a tensile-like deformation response during outflow, providing fundamental insights for nanofluidic devices exploiting confined liquid mechanics.

3. APL

   Soft-hard material integration enabled programmable robotic locomotion

   Haozhe Zhang, and Baoxing Xu

   Applied Physics Letters 2022

   Abs URL Project

   Inspired by the opening and closing mechanism of clams, herein, we present a two-dimensional robot by integrating soft-hard materials with structural design and demonstrate a variety of in-plane locomotion modes from a straight line to a circle. A theoretical model is developed to quantitatively predict the locomotion direction and step distance, and the results show remarkable agreement with finite element analysis. Systematic applications are further conducted to demonstrate programmable locomotion behavior by regulating the mechanical stimuli and arranging the order of soft-hard material integrated units. This work provides a rational route for designing diverse strategies of robotic locomotion with soft-hard material integration.

4. Adv. Mater.

   Ultrahigh Sensitive Au‐Doped Silicon Nanomembrane Based Wearable Sensor Arrays for Continuous Skin Temperature Monitoring with High Precision

   Mingyu Sang, Kyowon Kang, Yue Zhang, Haozhe Zhang, Kiho Kim, Myeongki Cho, Jongwoon Shin, Jung‐Hoon Hong, Taemin Kim, Shin Kyu Lee, Woon‐Hong Yeo, Jung Woo Lee, Taeyoon Lee, Baoxing Xu, and Ki Jun Yu

   Advanced Materials 2022

   Abs URL Project

   Monitoring the body temperature with high accuracy provides a fast, facile, yet powerful route about the human body in a wide range of health information standards. Here, the first ever ultrasensitive and stretchable gold-doped silicon nanomembrane (Au-doped SiNM) epidermal temperature sensor array is introduced. The ultrasensitivity is achieved by shifting freeze-out region to intrinsic region in carrier density and modulation of fermi energy level of p-type SiNM through the development of a novel gold-doping strategy. The Au-doped SiNM is readily transferred onto an ultrathin polymer layer with a well-designed serpentine mesh structure, capable of being utilized as an epidermal temperature sensor array. Measurements in vivo and in vitro show temperature coefficient of resistance as high as −37270.72 ppm °C−1, 22 times higher than existing metal-based temperature sensors with similar structures, and one of the highest thermal sensitivity among the inorganic material based temperature sensors. Applications in the continuous monitoring of body temperature and respiration rate during exercising are demonstrated with a successful capture of information. This work lays a foundation for monitoring body temperature, potentially useful for precision diagnosis (e.g., continuous monitoring body temperature in coronavirus disease 2019 cases) and management of disease relevance to body temperature in healthcare.

2021

1. Chem. Mater.

   Three-Dimensional Printable, Extremely Soft, Stretchable, and Reversible Elastomers from Molecular Architecture-Directed Assembly

   Shifeng Nian, Jinchang Zhu, Haozhe Zhang, Zihao Gong, Guillaume Freychet, Mikhail Zhernenkov, Baoxing Xu, and Li-Heng Cai

   Chemistry of Materials 2021

   Abs URL Project

   3D printing elastomers enables the fabrication of many technologically important structures and devices such as tissue scaffolds, sensors, actuators, and soft robots. However, conventional 3D printable elastomers are intrinsically stiff; moreover, the process of printing often requires external mechanical support and/or post-treatment. Here, we exploit the self-assembly of a responsive linear-bottlebrush-linear triblock copolymer to create stimuli-reversible, extremely soft, and stretchable elastomers and demonstrate their applicability as inks for in situ direct-write printing 3D structures without the aid of external mechanical support or post-treatment. By developing a procedure for controlled synthesis of such architecturally designed block copolymers, we create elastomers with extensibility up to 600% and Young’s moduli down to ∼102 Pa, 106 times softer than plastics and more than 102 times softer than all existing 3D printable elastomers. Moreover, the elastomers are thermostable and remain to be solid up to 180 °C, yet they are 100% solvent-reprocessable. Their extreme softness, stretchability, thermostability, and solvent-reprocessability bode well for future applications.

2. Nat. Commun.

   All-printed stretchable corneal sensor on soft contact lenses for noninvasive and painless ocular electrodiagnosis

   Kyunghun Kim*, Ho-Joong Kim*, Haozhe Zhang*, Woohyun Park, Dawn Meyer, Min Ku Kim, Bongjoong Kim, Heun Park, Baoxing Xu, Pete Kollbaum, Bryan W Boudouris, and Chi Hwan Lee

   Nature Communications 2021

   Abs URL Project

   Electroretinogram examinations serve as routine clinical procedures in ophthalmology for the diagnosis and management of many ocular diseases. However, the rigid form factor of current corneal sensors produces a mismatch with the soft, curvilinear, and exceptionally sensitive human cornea, which typically requires the use of topical anesthesia and a speculum for pain management and safety. Here we report a design of an all-printed stretchable corneal sensor built on commercially-available disposable soft contact lenses that can intimately and non-invasively interface with the corneal surface of human eyes. The corneal sensor is integrated with soft contact lenses via an electrochemical anchoring mechanism in a seamless manner that ensures its mechanical and chemical reliability. Thus, the resulting device enables the high-fidelity recording of full-field electroretinogram signals in human eyes without the need of topical anesthesia or a speculum. The device, superior to clinical standards in terms of signal quality and comfortability, is expected to address unmet clinical needs in the field of ocular electrodiagnosis.

2020

1. IJSS

   Rotation mechanics of optical scatters in stretchable metasurfaces

   Haozhe Zhang, Weizhu Yang, and Baoxing Xu

   International Journal of Solids and Structures 2020

   Abs URL Project

   Stretchable metasurface comprised of an array of optical hard-material scatters on the surface of a soft substrate is emerging as an artificially designed structure to manipulate optical wave propagation through a mechanical means. Upon stretching the soft substrate, along with the increase of the spacing distance between hard scatters, the orientations of the hard scatters will vary due to their strong mechanical mismatch with the soft substrate, but the underlying mechanics mechanism is lacking. In the present study, we establish a theoretical mechanics framework to address the rotation of hard scatters bonded on the surface of a soft substrate subjected to a mechanical stretching load. In this model, we introduce a pseudo-interface between the hard scatters and the soft substrate and decouple the effect of in-plane stretching deformation and out-of-plane deformation in the thickness direction of the soft substrate on the basis of moment equilibrium. Besides, the shear stress field along the thickness direction of three-dimensional soft substrate is formulated from that of a two-dimensional plane model by introducing a thickness factor. The effect of spacing distance, geometric shape, and initial orientation of hard scatters and thickness of the soft substrate on the scatter rotation is elucidated. Extensive finite element analyses (FEA) are performed and the results show good agreement with the theoretical predictions in both scatter rotation and stress distribution. We further construct a stretchable metasurface and demonstrate that the refraction of incident lights can be tuned by controlling the rotation of scatters. The mechanics theory established here provides a foundation to design mechanically controllable metasurfaces by leveraging the rotation of hard-material scatters.