PGPhilippe H. Geubelle

Philippe H. Geubelle

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Philippe H. Geubelle has a total of 197 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Reduced-Order Model for Frontal Polymerization with Application to Growth Printing

    by Minjiang Zhu, Adham Abdelmaguid, Yun Seong Kim, Rohan Shah, Derrick M. Sanders, Yigit Hergul, Varda F. Hagh, Sameh Tawfick, Philippe H. Geubelle

    Accurate 3D simulation of frontal polymerization (FP) remains computationally prohibitive due to the sharpness of the propagating reactive front, which imposes severe spatial and temporal resolution requirements in the numerical implementations of the reaction–diffusion formulation. This limitation restricts predictive modeling to simplified 1D or 2D/axisymmetric settings and hinders the analysis of FP processes involving more complex 3D geometries.In this work, we introduce a modification of the reaction–diffusion model of FP that enables the thickening of the propagating polymerization front while maintaining key propagation characteristics such as the front speed and the dependence of the front speed on the front curvature. The approach artificially thickens the polymerization front by a factor of up to four, allowing the use of coarser space and time discretizations. We verify the reduced-order model by comparing the forward and inverse solutions of axisymmetric growth printing problems. Finally, a 3D inverse strategy for growth printing is developed, implemented, and validated against experiments. The proposed method provides a computationally tractable approach for simulating thin reaction fronts in 3D.

  • Open Access

    Addressing the Seam: Autonomous Control of Multi-Front Frontal Polymerization

    by Grayson Schaer, Qibang Liu, Ivan Wu, Angelina Yan, Jeffery W. Baur, Philippe H. Geubelle, Timothy Bretl

    Frontal polymerization (FP) is a self-sustaining process that converts resin into polymer through a propagating reaction front, thereby providing a rapid and energy-efficient alternative to conventional manufacturing approaches for thermoset polymers and composites based on the bulk polymerization of the resin. The initiation of multiple fronts can further reduce the cure time at little additional energy cost, but also results in a thermal spike leading to the creation of a residual vertical deformation or `seam' of the resin where the fronts merge. Prior work suggests that the magnitude of this vertical deformation is a predictor of reduced mechanical performance and lifespan, attributed to the aggregation of heat and compressive strain during the merging of the fronts. In our work, we show for the first time that the vertical deformation of the seam can be reduced through the application of model-based closed-loop control. First, we model the vertical deformation associated with the propagating polymerization front by combining the reaction-diffusion model of FP with lubrication theory to capture the evolution of the free surface of the resin under the combined effects of Marangoni flow, buoyancy, and cure shrinkage. We then demonstrate that, by shaping fronts with closed-loop control to enhance energy dissipation at the front merging location, we can reduce the average vertical deformation of the seam by more than 50\%. These results suggest that our framework provides a way to enhance the mechanical performance and lifespan of components manufactured by multi-front FP.

  • Open Access

    Embedded Direct Ink Writing of Thermoset and Elastomeric Polymers via Frontal Polymerization

    by Mohammad Tanver Hossain, Yun Seong Kim, Pallab Layak, Pranav Krishnan, Young Bum Lee, Minjiang Zhu, Shubh Singh, Philippe H. Geubelle, Paul V. Braun, Jeffery W. Baur, et al.

    Direct ink writing (DIW) using frontal ring‐opening metathesis polymerization (FROMP) enables rapid, energy‐efficient fabrication of thermoset and elastomeric architectures through a self‐propagating exothermic curing reaction. However, FP‐DIW is constrained by stringent rheological requirements, feature‐size limits from front quenching, and the need for the reaction front to track the nozzle. Here, we overcome these constraints by leveraging embedded 3D printing to implement FP‐DIW with delayed solidification, thereby decoupling shape retention and solidification from ink chemistry and rheology. A yield‐stress support medium stabilizes low‐viscosity inks, suppresses gravitational and capillary instabilities, mitigates quenching at small diameters, and enables complex overlapping and mechanically interlinked features to fuse after deposition. Two thermal initiation strategies are introduced: volumetric dielectric heating via microwaves and surface heating at the support‐bath boundary. Dicyclopentadiene (DCPD), cyclooctadiene (COD), and mixed formulations provide tunable mechanical properties with glass transition temperatures from –50°C to 160°C. The approach enables lattices, springs, mechanically interlocked structures, and multimaterial architectures, substantially expanding the accessible material, feature‐size, and geometric design space compared with FP‐DIW in air.


  • Open Access

    Probabilistic delamination estimation in composite laminates from thermography using a deep-learning surrogate

    by Mumen Hamadneh, Ivan C.L. Wu, Jeffery W. Baur, Philippe H. Geubelle

  • Open Access

    Multi-Physics Modeling of Frontal Polymerization with Thermally Gated Catalyst Release

    by Fabio Sozio, Kazi Zahir Uddin, Po-Wen Wang, Sameh Tawfick, Nancy R. Sottos, Philippe H. Geubelle

    Embedding the catalyst into polymer microparticles to control its release within the surrounding monomer is a successful strategy to prolong the shelf life of resins undergoing ring-opening metathesis polymerization. In this work, we develop a thermal activation model for catalyst release in heterogeneous systems and integrate it into a reaction-diffusion framework for frontal polymerization that accounts for both composition heterogeneity and catalyst transport. From a chemical kinetics point of view, we adopt a hybrid approach blending mechanistic and phenomenological elements. Numerical predictions of front speed and temperature profile are validated against experiments conducted on dicyclopentadiene - ethylidene norbornene mixtures in which a second-generation Grubbs catalyst is embedded into polysulfone microparticles. The proposed model successfully captures the essential features of front propagation in resins with catalyst-filled microparticles.

  • Open Access

    Machine-learning surrogates for inverse design of frontal-polymerization-induced patterns in cyclooctadiene

    by Sagar Vyas, Qibang Liu, Philippe H. Geubelle

    Thermo-chemical instabilities in the frontal polymerization (FP) of cyclooctadiene are studied numerically using a one-dimensional (1D) reaction-diffusion model. The cure kinetics parameters, initial conditions, and boundary conditions are systematically perturbed to determine their effect on the magnitude and spacing of unstable FP-induced temperature spikes that define the generated pattern. A reduced-order representation is implemented to record the numerical results, which are arranged in a dataset together with their corresponding cure kinetics, initial conditions, and boundary conditions. This dataset is then used to train artificial neural network surrogates that bypass costly numerical simulations. The trained surrogates are embedded in a traditional gradient-based optimization loop to solve the inverse problem, i.e., given a target pattern, how to predict the cure kinetics, initial conditions, and boundary conditions that result in said pattern. A key advantage of such an approach is the ability to selectively constrain some of the design variables, i.e., cure kinetics, initial conditions, or boundary conditions. In the test cases considered here, the surrogates predict peak temperature amplitudes and wavelengths within about 10% of the full numerical simulations for almost all test points, and the inverse design procedure accurately reproduces a dense-region target pattern while highlighting the lower accuracy of the surrogate in sparsely sampled regions of the pattern space.

  • Open Access

    Trade-off between curing energy and rate in incremental composite manufacturing by frontal polymerization

    by Sanghun Shin, Ivan Wu, Tyler C. Price, Nawaf Rasheed, Michael Zakoworotny, Jillian J. Kouzmanoff, Jeffery W. Baur, Nancy R. Sottos, Philippe H. Geubelle, Sameh H. Tawfick

  • Additive Manufacturing of Continuous Carbon Fiber Thermoset Composites Based on Frontal Polymerization

    by Michael Zakoworotny, Nadim S. Hmeidat, Gavin DeBrun, Jeffery W. Baur, Sameh H. Tawfick, Nancy R. Sottos, Philippe H. Geubelle

  • Open Access

    Geometry-informed neural operator transformer for partial differential equations on arbitrary geometries

    by Qibang Liu, Weiheng Zhong, Hadi Meidani, Diab Abueidda, Seid Koric, Philippe Geubelle