Study on 3D characterized profile and point accuracies of ground micro-pyramid-structured Si surface

It is very difficult to evaluate 3D profile accuracy of micro-machined surface due to the 3D characterization of its measured point cloud. Hence, three ideal point cloud models, which are grid point model, slash profile model and horizontal profile model, were constructed to register 3D measured point cloud of micro-machined surface and evaluate the accuracies of characterized profile and points. First, #400 and #600 diamond wheel V-tips were employed to fabricate non-integrated and integrated micro-pyramid-structured surfaces with 173 A mu m in depth and 0.865 in aspect ratio in Computer Numerical Control (CNC) grinding system, respectively; then a white light interferometer was utilized to measure the micro-ground surfaces; finally, registration accuracy and registration efficiency were investigated with regard to characterized profile and point accuracies. The results show that registration accuracy increases and registration efficiency decreases with increasing model point number, but they have little relation with the posture of measured point cloud. Registration error may converge to micro-ground form error when model point number is larger than measured point number. Moreover, the slash profile and horizontal profile models may register characterized profile and point more precisely than grid point model. The slash profile model can produce the best registration accuracy and efficiency. It is confirmed that the micro-ground form errors are 23.8 A mu m and 7.9 A mu m, the characterized profile errors are 37.2 A mu m and 19.0 A mu m and the characterized peak errors are 51.2 A mu m and 34.1 A mu m for non-integrated surface and integrated surface, respectively. As a result, a #600 diamond grinding wheel may be used to perform a precision micro-grinding in CNC grinding system.