Abstract:

Objective To further understand the surface adhesion mechanism between cotton fibers in needle-punched nonwoven fabrics， this study aims to establish a reliable finite element model capable of quantifying interfacial adhesion forces at microscale， and to reveal how surface roughness modulates adhesion behavior through a "shielding effect" that cannot be captured by classical contact mechanics theories alone.

Method A 2-D axisymmetric finite element model with a "cellulose sphere-cellulose film" contact configuration was built in Abaqus to simulate the pull-off process. A bilinear cohesive traction-separation relation was used to captured interfacial damage evolution. The sphere was modeled as a Mooney-Rivlin hyperelastic solid （E = 12 MPa， ν = 0.3）， and film surface roughness （RMS = 0.15 μm） was introduced based on AFM measurements. Cohesive parameters were calibrated against colloidal-probe AFM data. Six sphere radius （R=3.30-16.54 μm） were tested， each repeated five times under randomized roughness.

Results The simulation results showed strong agreement with experimental data across all tested radius， with a maximum relative deviation of 14.88%. The adhesion force exhibited a clear non-monotonic dependence on sphere radius. In the range of 3.3 μm-13.18 μm， adhesion increased monotonically from 0.157 15 μN to 0.659 39 μN， consistent with the JKR prediction that adhesion force scales with contact area. The sphere at R=13.18 μm produced the highest adhesion force of 0.659 39 μN （simulated） versus 0.576 61 μN （experimental）. However， when the radius exceeded 14.35 μm， adhesion dropped markedly to 0.416 83 μN （R = 14.35μm） and 0.505 82 μN （R = 16.54 μm）. This phenomenon can be attributed to the presence of microscale rough structures on the membrane surface. Stress field visualization revealed that large spheres made contact predominantly with surface asperities rather than penetrating into surface valleys， substantially reducing the effective contact area. This mechanism-termed the "shielding effect" of surface roughness-explains the observed adhesion reduction that the micro-scale topography of the rough cellulose film prevents large spheres from achieving conformal contact， thereby weakening van der Waals interactions. The full loading-unloading-pull-off sequence was successfully reproduced by the cohesive model， capturing interface damage initiation and progressive softening up to complete separation.

Conclusion This study demonstrates that a bilinear cohesive zone finite element model， combined with explicit surface roughness representation， can accurately reproduce the adhesion behavior between cellulose fiber surfaces with deviations below 15%. The results confirm that adhesion between cellulose microspheres and films follows JKR scaling at moderate radius but is governed by roughness-induced contact shielding at larger radius. The identified "shielding effect" provides a quantitative micro-mechanical explanation for the radius-dependent non-monotonic adhesion behavior observed experimentally. The proposed modeling framework offers a practical tool for quantitative analysis and optimization of fiber-fiber bonding in needle-punched cotton nonwovens， with potential applicability to broader cellulose-based fibrous systems.

Key words: needle-punched nonwoven fabric, cotton fiber, surface adhesion, cohesive zone model, finite element simulation