Method and structure for detection of electromechanical problems using variance statistics in an E-beam lithography device

An apparatus and method for detection of electromechanical and mechanical errors in an electron beam device is provided. First the existing subfield is divided into a gridlike structure where each grid can be considered a target. Then a plurality of target points are provided on each grid for measuring combined directional variances. The separated horizontal and vertical variances is also measured for each of the target points. This leads to the performance of a significance tests, based on the F statistic which we refer to as FHV, for horizontal and vertical values of each target points during which FSTITCH values are also obtained. The FSTITCH values are then compared for horizontal and vertical values and an error alert provided when there is a sufficiently large disparity between the separated FSTITCH values. In an alternate embodiment of the present invention, a three dimensional grid is also provided to be used in a similar manner. The severity of the error can also be determined based on the disparity of the values. Lastly we refer to a simple F statistic for testing for flatness of the entire field, based on row and field information and refer to it as FROWS.