Adhesion of graded elastic materials: A full self-consistent model and its application

A full self-consistent model (FSCM) of the axisymmetric adhesive contact between a rigid punch with an arbitrary surface shape and a power-law graded elastic half-space is developed.The self-consistent equation between the surface gap and the surface interaction (e.g., the Lennard-Jones force law) involves a nonlinear singular integral, posing a great challenge to numerical calculations.By applying the properties of Gauss's hypergeometric function, the integral singularity is eliminated in the numerical calculation through Riemann-Stieltjes integral.Case studies for power-law punch profiles are performed and the self-consistent equation can be expressed in a dimensionless form with three dimensionless parameters, namely a shape index, a gradient exponent, and a new generalized Tabor number.The FSCM results are obtained by solving the self-consistent equation through the surface central gap control method and Newton-Raphson iterative method.For large generalized Tabor numbers, the forcedisplacement curves are 'S-shaped' and condense to the extended JKR limit in the high-load branch.As the generalized Tabor number decreases, a continuous transition from the extended JKR model to the Bradley model for the adhesion of power-law graded materials is obtained.It is found that the pull-off force of a graded material usually depends on the three dimensionless parameters, but for some cases of the shape index, it is not sensitive to the gradient exponent when the generalized Tabor number is fixed.Asymptotic solutions are derived to predict the unstable jump points, which coincide well with the FSCM predictions.The FSCM is applied to validate the extended Maugis-Dugdale (M-D) model of graded materials and it is found that the accuracy of the original M-D-- model using the maximum strength condition to determine the cohesive stress is limited.By introducing the rigid-limit-consistency condition of the pull-off force to determine the cohesive stress, the M-D-- model is improved and its predictions show good consistency with the FSCM results.