Micro-spring force sensors using conductive photosensitive resin fabricated via two-photon polymerization

Abstract

The rapid miniaturization of electronic devices has fueled unprecedented demand for flexible, high-performance sensors across fields ranging from medical devices to robotics. Despite advances in fabrication techniques, the development of micro- and nano-scale flexible force sensors with superior sensitivity, stability, and biocompatibility remains a formidable challenge. In this study, we developed a novel conductive photosensitive resin specifically designed for two-photon polymerization, systematically optimized its printing parameters, and improved its structural design, thereby enabling the fabrication of high-precision micro-spring force sensors (MSFS). The proposed photosensitive resin, doped with MXene nanomaterials, combines exceptional mechanical strength and conductivity, overcoming limitations of traditional materials. Using a support vector machine model in machine learning techniques, we optimized the polymerizability of the resin under varied laser parameters, achieving a predictive accuracy of 92.66%. This model significantly reduced trial-and-error in the TPP process, accelerating the discovery of ideal fabrication conditions. Finite element analysis was employed to design and simulate the performance of the MSFS, guiding structural optimization to achieve high sensitivity and mechanical stability. The fabricated MSFS demonstrated outstanding electromechanical performance, with a sensitivity coefficient of 5.65 and a fabrication accuracy within ±50 nm, setting a new standard for MSFS precision. This work not only pushes the boundaries of sensor miniaturization but also introduces a scalable, efficient pathway for the rapid design and fabrication of high-performance flexible sensors. The development of flexible, high-performance microscale force sensors remains a critical challenge for next-generation biomedical and wearable electronics. Here, we report a novel micro-spring force sensor fabricated via two-photon polymerization using a custom-designed conductive photosensitive resin doped with MXene nanosheets. The resin formulation was optimized to balance mechanical strength and electrical conductivity while ensuring high-resolution printability. To accelerate parameter optimization, a support vector machine model was trained to predict polymerization outcomes based on laser conditions and material compositions, achieving a prediction accuracy of 92.66%. Finite element analysis guided the design of the MSFS structure, enabling tunable electromechanical performance. The fabricated MSFS exhibited excellent sensitivity high fabrication precision and long-term stability. These results demonstrate the potential of integrating machine learning, functional nanomaterials, and TPP microfabrication to enable scalable, high-precision production of intelligent microsensors for biomedical and soft robotic applications.

The development of flexible, high-performance microscale force sensors remains a critical challenge for next-generation biomedical and wearable electronics. Here, we report a novel micro-spring force sensor fabricated via two-photon polymerization using a custom-designed conductive photosensitive resin doped with MXene nanosheets. The resin formulation was optimized to balance mechanical strength and electrical conductivity while ensuring high-resolution printability. To accelerate parameter optimization, a support vector machine model was trained to predict polymerization outcomes based on laser conditions and material compositions, achieving a prediction accuracy of 92.66%. Finite element analysis guided the design of the MSFS structure, enabling tunable electromechanical performance. The fabricated MSFS exhibited excellent sensitivity high fabrication precision and long-term stability. These results demonstrate the potential of integrating machine learning, functional nanomaterials, and TPP microfabrication to enable scalable, high-precision production of intelligent microsensors for biomedical and soft robotic applications.

Fig. 1. Overview of dispersibility and printability of photosensitive resins doped with different conductive nanomaterials.

The status of photosensitive resin doped with AgNPs (a), CNT(b), and MXene (c), respectively. The printing process of photosensitive resin sample mixed with CNT (d), and MXene (e), respectively. Scale bar: 100 µm

Fig. 2. SVM-based prediction workflow and model performance evaluation for MSFS cured.

a Representative optical images of MSFS structures under different curing outcomes: uncured, cured, and damaged. b Workflow of the SVM-based prediction process for printability evaluation. c Confusion matrices of SVM classifiers using different kernel functions (Linear, Rbf, Poly, Sigmoid), showing prediction performance on test datasets. Scale bar: 100 µm

Fig. 3. Mechanical and electrical characterization of MSFS.

a–c Mechanical properties at different PEGDMA mass fractions: (a) stress–strain curves; (b) Young’s modulus; (c) maximum stress. d–f Mechanical and structural responses at different MXene mass fractions: (d) stress–strain curves; (e) Young’s modulus. f Schematic of a monolayer MXene sheet. g Geometry of the nanocomposite unit cell with randomly doped MXene; (h) Geometry of the nanocomposite unit cell considering the tunneling effect. i Schematic of effective electrical conductivity calculation in the nanocomposite. j Curve of effective conductivity versus height variation percentage of the unit cell. k Electrical conductivity at different MXene mass fractions. l Schematic diagram of MSFS geometry and structural parameters

Fig. 4. Structural parameter variations and their influence on the electromechanical response of MSFS.

a–c Geometry of the MSFS with wire diameter $r$ = 0.01 mm, 0.015 mm, and 0.02 mm (a), as well as their force-displacement curve (b) and force-resistance curve (c), where (c) includes both linear plot and logarithmic subplot. d–f Geometry of the MSFS with outside of diameter $R$ = 0.08 mm, 0.016 mm, and 0.24 mm (d), as well as their force-displacement curve (e) and force–resistance curve (f), where (f) includes both linear plot and logarithmic subplot. g–i Geometry of the MSFS with axial pitch $L$ = 0.06 mm, 0.09 mm, and 0.23 mm (g), as well as their force–displacement curve (h) and force-resistance curve (i), where (i) includes both linear plot and logarithmic subplot

Fig. 5. Electromechanical performance and structural validation of the MSFS.

a–l Progressive deformation of the MSFS under applied force: (a, c, e, g, i, k) simulation results and (b, d, f, h, j, l) corresponding optical microscopy images captured at applied forces of 0, 50, 100, 150, 200, 250, and 300 $\mathrm{\mu }\mathrm{N}$, respectively. m Simulation and experimental displacement-resistance curves. n Force-resistance response curves of MSFSs with and without MXene, highlighting the critical role of MXene in conductivity enhancement. o Force-resistance response of the MSFS at 1, 15, and 30 days. p Relative resistance changes versus strain curve. q 3D surface morphology map of the MSFS. Scale bar: 100 µm

Fig. 6. Mesh convergence analysis of the MSFS model.

Top displacement of the MNSF structure as a function of the number of mesh elements