Improving the Positional Accuracy of the Goniometer on the Philips CM200 TEM

We have developed a system¹ for automatically acquiring large numbers of high quality transmission electron micrographs under low dose conditions. This system has been implemented on a Philips CM200 Transmission Electron Microscope equipped with a Gatan MSC CCD camera. Implementing the automated data acquisition system requires that target locations be identified in a low magnification image and then accurately located at the center of the viewing area. The magnification is subsequently increased, the image is focused on an area adjacent to the target area and the final image is acquired. Centering an identified target location for subsequent high magnification imaging typically requires moving the specimen by many thousands of nm and accurately locating the target to within a few hundred nm. This movement is too large to be achieved using the image shift coils, which would be very accurate, and instead must be achieved using the goniometer.

We have measured the accuracy of the goniometer on the Philips CM200 and the results are shown in fig 1. Data were obtained by selecting a target area from a low magnification image [660x], moving to this target and then measuring the accuracy of the requested movement by cross correlation. The results show that the targeting error increases as the distance to the target increases. For example the error can exceed 1500 nm at a target distance from the origin of ~10,000 nm. Errors of this magnitude mean that the targeting area may be missed completely when the final image is acquired at high magnification. The targeting accuracy may be improved by a series of refinement steps in which the target location in the original image is compared against the currently located center. However these steps are time consuming and also require that the magnification be switched between the low and high settings for every target to be imaged and this can contribute to hysteresis instabilities in the beam.

Figure 1. The dashed curve illustrates the accuracy of the goniometer under the usual assumptions of linear behavior. The solid line illustrates the improvement in accuracy gained by modeling the non-linear behavior of the goniometer.

To address these problems we have modeled the behavior of the goniometer and used this model to improve the positioning accuracy. The behavior of the goniometer was characterized by measuring the image displacement resulting from a requested goniometer movement. The results for the y axis of the goniometer are shown in fig. 2 and indicate that the behavior of the goniometer is non-linear. The data show a distinct periodic function, which can be modeled with a Fourier series. The x and y axes exhibit similar behavior although the shape and periodicity of the response functions are unique. This was found to be true for a second instrument also (the CM200 TEM located at The Research Institute at Scripps Clinic). This non-linear behavior of the goniometer can be modeled and used in calculating the required goniometer motion, resulting in a substantial improvement in the targeting as illustrated in fig. 1. For example the targeting error now remains on the order of 200 nm even at a distance of 10,000 nm from the center. We expect to improve this accuracy still further by refining the model of the goniometer behavior.

Figure 2. The behavior of the goniometer was characterized by measuring the image displacement resulting from a requested goniometer movement. The results shown are for a portion of the y axis.

Implementing this method for improving the accuracy of the goniometer is relatively straightforward. Acquiring the data necessary for modeling the takes a few hours and should not need to be repeated unless the goniometer undergoes major repairs. Calculation times necessary to implement the model are trivial compared to the time taken to physically move the goniometer.

References

1. Potter et al., Submitted to this conference.
2. Support is provided by the NSF (9730056) and the IBM Shared University Research program.

Page 2 of 2 -