18 March 2024 in Nature Nanotechnology
One possible solution against the accumulation of petrochemical plastics in natural environments is to develop biodegradable plastic substitutes using natural components. However, discovering all-natural alternatives that meet specific properties, such as optical transparency, fire retardancy and mechanical resilience, which have made petrochemical plastics successful, remains challenging. Current approaches still rely on iterative optimization experiments. Here we show an integrated workflow that combines robotics and machine learning to accelerate the discovery of all-natural plastic substitutes with programmable optical, thermal and mechanical properties. First, an automated pipetting robot is commanded to prepare 286 nanocomposite films with various properties to train a support-vector machine classifier. Next, through 14 active learning loops with data augmentation, 135 all-natural nanocomposites are fabricated stagewise, establishing an artificial neural network prediction model. We demonstrate that the prediction model can conduct a two-way design task: (1) predicting the physicochemical properties of an all-natural nanocomposite from its composition and (2) automating the inverse design of biodegradable plastic substitutes that fulfils various user-specific requirements. By harnessing the model’s prediction capabilities, we prepare several all-natural substitutes, that could replace non-biodegradable counterparts as exhibiting analogous properties. Our methodology integrates robot-assisted experiments, machine intelligence and simulation tools to accelerate the discovery and design of eco-friendly plastic substitutes starting from building blocks taken from the generally-recognized-as-safe database.
1 June 2024 in Nature Communications
Designing ultralight conductive aerogels with tailored electrical and mechanical properties is critical for various applications. Conventional approaches rely on iterative, time-consuming experiments across a vast parameter space. Herein, an integrated workflow is developed to combine collaborative robotics with machine learning to accelerate the design of conductive aerogels with programmable properties. An automated pipetting robot is operated to prepare 264 mixtures of Ti3C2Tx MXene, cellulose, gelatin, and glutaraldehyde at different ratios/loadings. After freeze-drying, the aerogels’ structural integrity is evaluated to train a support vector machine classifier. Through 8 active learning cycles with data augmentation, 162 unique conductive aerogels are fabricated/characterized via robotics-automated platforms, enabling the construction of an artificial neural network prediction model. The prediction model conducts two-way design tasks: (1) predicting the aerogels’ physicochemical properties from fabrication parameters and (2) automating the inverse design of aerogels for specific property requirements. The combined use of model interpretation and finite element simulations validates a pronounced correlation between aerogel density and compressive strength. The model-suggested aerogels with high conductivity, customized strength, and pressure insensitivity allow for compression-stable Joule heating for wearable thermal management.
19 January 2026 in Nature Communications
Innovations in wearable electronics and soft robotics hinge significantly on the development of stretchable electrodes. However, a persistent challenge lies in balancing high stretchability, functional performance, and strain insensitivity. Conventional approaches, such as design of experiments and trial-and-error methods, often rely on time-consuming and labor-intensive experiments to navigate a vast and complex parameter space. To overcome this, we establish an integrated workflow merging robot-automated experimentation, machine learning predictions, and finite element simulations to enable the predictive design of stretchable electrodes with strain-insensitive performance. Initially, we construct an ensemble of artificial neural networks through a two-stage workflow, including feasible parameter space definition and active learning loops. Leveraging the prediction model and two-scale simulations, a microtextured stretchable nanocomposite is discovered as a strain-stable platform. Conformal deposition of a thin gold layer showcases metal-like conductivity, high resistance-insensitive stretchability, and robust durability. Furthermore, electrodeposition of Zn and MnO2 on gold conductors enables a stretchable Zn||MnO2 battery, exhibiting large elongation and strain-insensitive electrochemical performance. This machine intelligence-driven approach expedites the multi-parameter optimization of stretchable electrodes, achieving strain-invariant functionalities.
3 October 2024 in Research Square
Innovations in wearable electronics and soft robotics hinge significantly on the development of stretchable electrodes. However, a persistent challenge lies in balancing high stretchability, functional performance, and strain insensitivity. Traditional methods rely on time-consuming and labor-intensive iterative experiments to navigate a vast parameter space. To overcome this, we establish an integrated workflow merging collaborative robotics, machine learning, and finite element simulations to enable the predictive design of ultrastretchable electrodes with strain-insensitive performance. Initially, an automated pipetting robot generates 286 nanocomposites, and their electrical conductance is assessed to train a support-vector machine regressor. Through 7 active learning loops, we fabricate 146 conductive interlayers to construct an ensemble model of artificial neural networks. Leveraging the prediction model and two-scale simulations, we discover a microtextured conductive interlayer as a strain-stable platform. By evaporating a thin gold layer, we develop an ultrastretchable gold conductor with metal-like conductivity, surpassing 1,000% resistance-insensitive stretchability, and robust durability. Furthermore, electrodepositing Zn and MnO2 on gold conductors enables fabrication of a Zn//MnO2 battery showcasing >300% stretchability and strain-insensitive performance. This machine intelligence-driven approach expedites the multi-parameter optimization of stretchable electrodes, achieving strain-invariant functionalities.
30 October 2026 in Science Advances
Reliance on petroleum-derived plastic packaging exacerbates environmental burdens and lacks intrinsic antimicrobial protection against food spoilage. While biobased materials offer a sustainable alternative, optimizing their formulations is hindered by vast component libraries, inefficient trial-and-error experimentation, and complex multiproperty requirements. Herein, we develop an integrated workflow combining robotic automation, machine learning predictions, density functional theory, and life cycle assessment to accelerate biobased packaging discovery. Leveraging data augmentation and ensemble modeling, we use a hierarchical neural network to predict ∼1 billion candidate formulations, balancing mechanical properties, optical transparency, and environmental footprint. The resulting biobased packaging demonstrates enhanced hydrostability, effective antimicrobial activity, and high barrier performance, significantly extending the shelf life of cucumbers and avocados. Last, we release a data-sharing platform featuring forward prediction, inverse design, and environmental impact functions to support community use. This integrative framework provides sustainable, antimicrobial packaging materials alternative to conventional plastic films.