Mechanism-data fusion for quantitative prediction of transmitted wavefront error in multi-point bonded optics

A mechanism-data fusion method is reported for the quantitative prediction of transmitted wavefront error (TWE) in planar optical components subjected to multi-point adhesive bonding. In this framework, TWE sources are explicitly decoupled into two competing physical mechanisms: refractive index variations governed by the photoelastic effect and surface deformations driven by the Poisson effect. To address the measurement inaccessibility of the rear surface under loading, a bounding analysis strategy is implemented to establish physically consistent theoretical prediction intervals. Experimental results show that, for the investigated N-BK7 optical component and loading conditions, the TWE contribution associated with Poisson-induced surface deformation is approximately one order of magnitude larger than that associated with the photoelastic effect. This result quantitatively clarifies the relative importance of these competing mechanisms in the assembly-induced wavefront error. The predictive model is validated through interferometric measurements, where measured wavefront RMS values generally fall within the predicted bounds, with a mean relative prediction error of approximately 10%. These results indicate that the proposed framework can quantitatively capture the dominant physical mechanisms governing assembly-induced wavefront error and offer a mechanism-informed basis for the predictive analysis of bonded opto-mechanical assemblies.