Fig. 2: Definition of a feasible parameter space through automated pipetting robot and support-vector machine (SVM) model. | Nature Communications

Fig. 2: Definition of a feasible parameter space through automated pipetting robot and support-vector machine (SVM) model.

From: Predictive design of stretchable electrodes with strain-insensitive performance via robotics- and machine learning-integrated workflow

Fig. 2: Definition of a feasible parameter space through automated pipetting robot and support-vector machine (SVM) model.

a Generation of G1/G2 stretchable nanocomposites via sequential deformations. SEM images illustrate the evolution from a filtered nanocomposite with planar surface to various microtextured nanocomposites: a G1–1D stretchable nanocomposite with parallel wrinkles, a G1–2D stretchable nanocomposite with isotropic crumples, a G2–2D1D stretchable nanocomposite with large parallel wrinkles and small isotropic crumples, and a G2–2D2D stretchable nanocomposite with large curvy wrinkles and small isotropic crumples. b Influence of fabrication parameters on the strain responses of G1/G2 stretchable nanocomposites. Left: Resistance–elongation curves of various G2–2D1D stretchable nanocomposites with different MXene/SWNT/AuNP/PVA ratios. All G2–2D1D stretchable nanocomposites were at the same thickness of 1200 nm and the same applied pre-strain of 100%. Right: Resistance–elongation curves of stretchable nanocomposites with different deformation sequences. All stretchable nanocomposites were at the same thickness of 1200 nm and the same MXene/SWNT/AuNP/PVA ratio of 45/45/8/2. The applied pre-strain of the G2–2D1D and G2–2D2D stretchable nanocomposites was the same at 100%. c Demonstration of an automated pipetting robot (i.e., OT-2 robot) capable of preparing a library of 286 aqueous mixtures with varying MXene/SWNT/AuNP/PVA ratios. Followed by vacuum filtration, 286 filtered nanocomposites were obtained. Thicknesses of filtered nanocomposites were controlled to be 800 nm. Created in BioRender. Shrestha, S. (2025) https://BioRender.com/g8co23b. d Left: Electrical conductance values of 286 filtered nanocomposites with varying MXene/SWNT/AuNP/PVA ratios. Right: A 3D heatmap representing the model-predicted \({S}_{{filtered}}\) values.

Back to article page