Enhancing Thin-Film Wafer Inspection With A Multi-Sensor Array And Robot Constraint Maintenance

TL;DR

An autonomous inspection system using a robotic manipulator with a novel learned constraint manifold to control a wafer to its calibration point, and a novel multi-sensor array with high potential for scalability into large substrate areas is presented.

Abstract

Thin-film inspection on large-area substrates in coating manufacture remains a critical parameter to ensure product quality; however, extending the inspection process precisely over a large area presents major challenges, due to the limitations of the available inspection equipment. An additional manipulation problem arises when automating the inspection process, as the silicon wafer requires movement constraints to ensure accurate measurements and to prevent damage. Furthermore, there are other increasingly important large-area industrial applications, such as Roll-to-Roll (R2R) manufacturing where coating thickness inspection introduces additional challenges. This paper presents an autonomous inspection system using a robotic manipulator with a novel learned constraint manifold to control a wafer to its calibration point, and a novel multi-sensor array with high potential for scalability into large substrate areas. We demonstrate that the manipulator can perform required motions whilst adhering to movement constraints. We further demonstrate that the sensor array can perform thickness measurements statically with an error of $<2\%$ compared to a commercial reflectometer, and through the use of a manipulator can dynamically detect angle variations $>0.5^\circ$ from the calibration point whilst monitoring the RMSE and $R^2$ over 1406 data points. These features are potentially useful for detecting displacement variations in R2R manufacturing processes.

Figures (10)

• Figure 1: Schematic of the sensor array with seven sensors: (a) Sensor sub-assembly cross section showing the STM Nucleo L432KC board, the DC converter, the pin socket, the sensor C12666MA and an LED; (b) 3D models of the array sensor holder and the lateral and front locks of the sensor devices. The top picture shows sensor zones A1-A2, B and C, and sensor positions 1-7; (c) Front and back isometric views of the array full assembly; (d) Front view of the multi-sensor array in static measurement configuration.
• Figure 2: Reflectance curve formed by the model formula in equation \ref{['eq:target']} compared against the measured reflectance from equation \ref{['eq:measured-reflec']}.
• Figure 3: Model of the VAE system for generating a latent space Riemannian manifold. The latent space $q_\zeta(z | \theta)$ is used to generate the Riemannian metric $\mathbf{M}$, which is used to determine whether the manipulator is experiencing joint drift. The estimate of the constraint function $\hat{\mathbf{f}}(\theta)$ is computed from decoding the latent space and computing the manifold constraint function
• Figure 4: Array height and angle experiments. (a) Sensor array calibration point at 2mm above wafer surface (red dotted line) and height variations from the calibration point -1mm/+2mm (green arrows). Thorlabs base PY005/M, reproduced with permission. The numbers are the sensor numbers i.e. SENSOR1 = 1. (b) Angle variations from 0° to 0.498° (rounded to 0.5°) in steps of 0.166°. (c) The hardware setup.
• Figure 5: Sensor array inspection box