Micropyramid Array Bimodal Electronic Skin for Intelligent Material and Surface Shape Perception Based on Capacitive Sensing

Abstract Developing electronic skins (e‐skins) that are comparable to or even beyond human tactile perception holds significant importance in advancing the process of intellectualization. In this context, a machine‐learning‐motivated micropyramid array bimodal (MAB) e‐skin based on capacitive sensing is reported, which enables spatial mapping applications based on bimodal sensing (proximity and pressure) implemented via fringing and iontronic effects, such as contactless measurement of 3D objects and contact recognition of Braille letters. Benefiting from the iontronic effect and single‐micropyramid structure, the MAB e‐skin in pressure mode yields impressive features: a maximum sensitivity of 655.3 kPa−1 (below 0.5 kPa), a linear sensitivity of 327.9 kPa−1 (0.5–15 kPa), and an ultralow limit of detection of 0.2 Pa. With the assistance of multilayer perceptron and convolutional neural network, the MAB e‐skin can accurately perceive 6 materials and 10 surface shapes based on the training and learning using the collected datasets from proximity and pressure modes, thus allowing it to achieve the precise perception of different objects within one proximity‐pressure cycle. The development of this MAB e‐skin opens a new avenue for robotic skin and the expansion of advanced applications.


Note S2. Construction of MLP neural network model for intelligent material perception.
MLP is constructed by the Keras open-source artificial neural network library, which has five layers, including one input layer, three hidden layers, and one output layer.The 46 neurons in the input layer represent the 46 data points collected by objects of different materials in the process of approaching the MAB e-skin, and the 6 neurons in the output layer represent the output recognition probability of the 6 materials.Regarding the hidden layers, the number of neurons in each layer is set to 64, 128, and 64, respectively (Figure 3d).The input and hidden

Conventional parallel-plate capacitive sensing:
For the parallel-plate capacitive e-skin, its capacitance (C) can be expressed as where ε represents the effective dielectric constant; A is the effective area of parallel plates; k stands for electrostatic constant; d means the separation distance between parallel plates.According to the above formula, ε, A, and d can all cause changes in capacitance.However, as the e-skin is not stretchable (A remains constant), the capacitance will be primarily determined

Figure S1 .
Figure S1.Chemical structures of the two components used in the ionic gel: PVDF-HFP

Figure S2 .
Figure S2.a) Top-view and b) cross-sectional SEM images of the ionic gel with micropyramid

Figure S3 .
Figure S3.Enlarged image of the anti-micropyramid array template.

Figure S4 .
Figure S4.a,b) Schematic and c) photograph of the PI/Cu/Au electrode array.

Figure S5 .
Figure S5.COMSOL simulation for the potential distribution of an object near the single-

Figure S6 .
Figure S6.Relative capacitance variations in response to an approaching palm at different a)

Figure S7 .
Figure S7.Response/recovery time curve of proximity sensing in the high-frequency

Figure S8 .
Figure S8.Sensing principles of the conventional parallel-plate capacitive e-skin.
by the ε and d, as illustrated in Figure S6.Further, following the general Lichterecker mixing rule,[1,2] the expression for ε between parallel plates is ε α = Vair air  + VP P  where α is the parameter determining the type of mixing rule; εair and εP are the dielectric 6 constant of the air and PVDF-HFP, respectively; Vair and VP are the volume fractions (Vair + VP = 1) of air and PVDF-HFP, respectively.As the dielectric between parallel plates consists of air with low εair (= 1) and PVDF-HFP with high εP (~ 9), a large amount of air is squeezed out upon applied pressure, thus leading to an enhanced ε α .

Figure S9 .
Figure S9.a) Stress distributions of COMSOL FEA simulation results for the single-

Figure S10 .
Figure S10.Equivalent circuit of the MAB e-skin in pressure mode.

Figure S12 .
Figure S12.Pressure-dependent relative capacitance changes at different a) temperatures and

Figure S13 .
Figure S13.Photographs of different positions approached by the fingertips (one (i) and two

Figure S14 .
Figure S14.3D mappings of capacitive responses at two different heights between the object

Figure S15 .
Figure S15.Photograph of 3D measurement and metallic objects such as sphere (i), ring (ii),

Figure S16 .
Figure S16.a) Photograph of the anti-micropyramid array template.b) SEM images (top-view

Figure S17 .
Figure S17.Capacitance response of 6 materials at various distances from the MAB e-skin.

Figure S18 .Figure S19 .
Figure S18.Sensitivities of 6 materials with the MAB e-skin in the distance range of 0.5-2.5

Figure S25 .
Figure S25.Distribution diagram of 10 sets of capacitance response-pressure mapping data