Laser-Fabricated Ordered Tri-Level Hemispherical Microstructures for Flexible Pressure Sensors with Ultrahigh Sensitivity over a Wide Linear Range

The rapid advancement of embodied intelligence and soft robotics demands flexible pressure sensors that simultaneously achieve high sensitivity, a broad linear response range, and excellent long-term stability. However, the inherent compression hardening effect of soft materials imposes a fundamental trade-off between sensitivity and linear range, severely limiting their practical applicability in complex dynamic environments. Inspired by the hierarchical tactile perception mechanism of human skin and guided by Archard's contact theory, we propose a piezoresistive sensor incorporating an ordered tri-level hemispherical microstructure array. The sensing layer, composed of a graphite powder/multiwalled carbon nanotubes/polydimethylsiloxane (GP/MWCNT/PDMS) conductive composite, is patterned via a rapid, scalable, and mask-free three-dimensional (3D) dynamic focusing laser processing technique. Finite element analysis (FEA) guides the optimization of geometric parameters, confirming that the tri-level configuration exhibits the highest compressibility and the greatest rate of contact area change among single-level, dual-level, and tri-level designs. The sensor operates through a graded deformation mechanism: under low pressure, the smallest hemispheres deform first to rapidly increase contact area, under medium pressure, the intermediate hemispheres engage to sustain sensitivity; under high pressure, the largest hemispheres provide structural support, ensuring continuous contact area expansion over a broad pressure range. Consequently, the sensor achieves an ultrahigh sensitivity of 12.8 kPa-1 over a wide linear range of 0-369 kPa (R2 = 0.998), with fast response/recovery times (20/21 ms) and excellent cycling stability (15,000 cycles). Beyond single-point sensing, we further construct a 4 × 4 sensor array that successfully discriminates spatial pressure distributions, as demonstrated by recognition of "M"-shaped and "U"-shaped loading modules. Practical applications include precise monitoring of finger joint bending angles and real-time plantar pressure distribution during walking and running, and shape-resolved tactile imaging, highlighting its great promise for applications in embodied intelligence, wearable health monitoring, human-machine interaction, and related fields.