Crosstalk‐Free, Stretching‐Insensitive Sensor Based on Arch‐Bridge Architecture for Tactile Mapping with Parallel Addressing Strategy toward Million‐Scale‐Pixels Processing

Abstract In the field of biomimetic electronics, flexible sensors with both high resolution and large size are attracting a lot of attention. However, attempts to increase the number of sensor pixels have been thwarted by the need for complex inner circuits and the resulting interferences with the output. Technological challenges, such as real‐time spatiotemporal mapping and long‐time reliability, must be resolved for large‐scale sensor matrices. This paper reports a simple and robust sensor with an arch‐bridge architecture (ABA) to address these challenges. The device, which consists of an anti‐icing all‐transparent material system, is fabricated by immobilizing ABA ionic arrays on predefined grooves on the substrate. It systematically integrates ABA structure‐designing, resistance‐position‐sensing, and parallel‐addressing logic, allowing for an improvement of three orders of magnitude in the scanning speed (million‐scale pixels) without logical “diagnose confusion.” In addition, it can withstand 100 000 stretching cycles without functional failure. It is also resistant to interferences from stretching. humidity, wet surfaces, and power lines. The proposed strategy is envisaged to serve as a general solution for high‐density, large‐area tactile sensors in various applications.


Supplementary Text
The anti-icing performance for transparent double-phase hydrogels: By using oil-water mixture solvent instead of a water phase, the anti-acing materials also demonstrates exceptional temperature tolerance down to −30 ℃ and allows operation in a wide temperature range and under various circumstances. Fig. S1 shows the cold-resistance performance relative to the gradient volume concentration of glycerin to water from 0%-50%. The ionic materials remains transparent and mechanically soft at a temperature of −40 ℃ at a volume ratio of 40%-50% (the performance in other temperature see Fig. S1a,b,c), and its electronic characteristics are shown in Fig. S1d by a volume ratio of 40%. By further increasing the ratio of the glycerin phase up to more than 40%, the toughness of the APM hydrogel can be abruptly impaired, possibly because of the high ratio of glycerin mixed in water, which hinders the cross bonding of acrylamide. By considering both the anti-freezing property and mechanical performance of the conducting layer, the ratio of glycerinwater is optimized by 40% for the sensor fabrication.
A 24-channel matrix is fabricated and preserved in a commercial deep-frozen refrigerator at -40 ℃ for 24 hours. And then the matrix is taken out for observation by eyes and characterization by LCR meters.As Figure S7 shows, under -40℃ , the matrix remains flexible and the ionic bars are still conducitive. Once taken out from deep fozen surroundings, there is a thin layer of ice quickly condensed on its surface (air humidity in Shenzhen at that time is about 95%). It is found that after frozen, the resistance of the ionic bars decreased slightly and gradully recovered to original value after it is warmed up. However, the resistance along the bars are still homogenous and the sensing properties of the device remains the same.
The outcome of toughness interface bonding test is displayed in Figure S3. The test is carried out as 180° peeling-force applied by a universal testing machine. Its result clearly shows that compared with the former reported approaches the bonding strength of glycerin-existed interface between APH and PDMS is largely improved by 10 times around. The peeling test is performed using samples with 100 mm in length, 18 mm in width and 2 mm in thickness. The interfacial toughness is calculated by dividing the steady-state peeing force with the sample width.

FEM Simulation of ABA deformation under finger-mimic press:
To elucidate the underlying mechanism between contact resistance and pressure visually, we further establish a 3D sensor model with material systems in ABAQUS, varying the height of H1 at 200 and 500 μm. The models of sensor's structure are established in a subarea of 1/4 for a minimum computation. Due to the hyperelasticity property in the materials, our sensor can be modeled as an integrated platform with Arruda-Boyce hyperelastic characteristics. Here, for a small deformation relatively, the Arruda-Boyce model is used without a correction term to define its distortion boundary. The model equation in ABAQUS can be expressed as: where U represents the energy potential, C 1 = 1/2, C 2 = 1/20, C 3 = 11/1050, C 4 = 19/7000, and C 5 = 519/673750, I 1 refers to Partial strain, and J e1 is the volume ratio. μ and λ can be calculated by measured stress-strain data and stand for the Arruda-Boyce parameter, represents a series of principal stretches. D is a constant related to the bulk modulus of elasticity in which D=0 means an incompressible property.
By the Curve Fitting routine in ABAQUS, the uniaxial test data as mean to fit it into the Arruda-Boyce model is displayed in Supplementary Fig. S5. Meanwhile the output of fitted parameters is listed in Supplementary table S1, where μ 0 is represented as the consistent shear modulus of elasticity.

Figure S1
Three different density on ABA matrix with APH hydrogels dying in red. a) 25dpi; b) 12.5dpi; c) 6.25dpi.

Movie S2
A ABA touch pad is served as human-computer interfaces on playing computer piano games.

Movie S3
A ABA touch pad is served as human-robot arm interfaces to deliver different controlling commands.

Movie S4
The sectional view of FEM simulation to elucidate the inner deformation of sensors with the 200 μm height of H1.

Movie S5
The top-view of FEM simulation to elucidate the inner deformation of sensors with the 200 μm height of H1.

Movie S6
The sectional view of FEM simulation to elucidate the inner deformation of sensors with the 500 μm height of H1.

Movie S7
The top-view of FEM simulation to elucidate the inner deformation of sensors with the 500 μm height of H1.