Patent Application
20140027632
This invention relates to detection of cathodoluminescence in a charged particle microscope.
This invention addresses the measurement of the angular distribution of emitted luminescence produced by a sample in an electron microscope in one image over a large solid angle. Such measurements were performed in various ways in the prior art. For example, in one approach, the 2D detector is orthogonal and external to the electron microscope. In this approach, extraction of the light requires the use of a 90 degree off-axis parabolic mirror. Coenen et al., Angle-resolved Cathodoluminescence Spectroscopy, Applied Physics Letters 99, 143101 (2011). All references cited herein are incorporated by reference.
Yamamoto and co-workers developed a technique using the same geometry as Coenen. In Yamamoto, however the angular distribution of CL was measured by translating a pinhole in front of a CCD camera, thus collecting point-by-point angular data. This approach was to screen out rays which were not in focus at the detector, and investigate the polarization of the luminescence. Yamamoto et al., “Gao and Mie Plasmons in Individual Silver Nanospheres near a Silver Surface” 11 Nano Letters 91-95 (2011).
This will include plain language versions of the claims once they are settled.
An embodiment of the invention is described as follows with respect to
The mirror is preferably a rotationally symmetric parabolic mirror and will ideally be mounted on the SEM chamber port wall 10. This facilitates actuated insertion and retraction with fine scale control. In an alternative embodiment the mirror is mounted to the pole piece of the SEM.
It is critical that the mirror has a well-defined optical focus and that the electron beam irradiated part of the specimen is at this precise location. Variations in either the surface roughness or curvature of the mirror from an ideal parabola by even a fraction of a wavelength of light will cause a phase related interference pattern on the detector image. This will impair the ability to deconvolve the desired angular information.
The spatial resolution of the device is related to focused electron beam spot size, the electromagnetic modes supported by the specimen and the quality of the focusing in the experiment.
Bandpass filtering 50 can be incorporated into the parallel light path between the mirror and the detector. Polarization filtering 60 can also be inserted instead or as well as bandpass filtering into the same light path.
The novel geometry of the collection optics and detector position provides access to a significantly larger solid angle than prior art devices which use an open off-axis parabola. Close to 2 pi steradians solid angle is exposed in the present invention as opposed to 1.2 to 1.4 pi steradians in prior art. Other mirror geometries are possible as long as they are rotationally symmetric and the specimen resides at a focal point. For example if a hemispherical mirror is employed, then a different image will be recorded and a different algorithm will be employed to deduct the angular luminescence.
The symmetry of the apparatus precludes the need to rotate and reposition the specimen with respect to the measurement system. However rotation of the stage 41 may still be desirable for some embodiments.
The algorithm to convert the recorded image on the detector into an angular emission profile is simpler than that required where there is no 360 rotation symmetry and the emitted light is parallel to the surface normal of the specimen.
Non-transparent specimens cast a shadow over the detector, so utilization of larger specimens will require scaling of the distances, mirror and detector hardware. This is possible without losing resolution or sensitivity. This is within the limits of the Z space available on the stage.
The electron beam focus can be kept stationary with different parts of a specimen irradiated by translating the specimen on a stage independent of the mirror optics. For a flat specimen, this can be achieved without altering the focus of the device.
The mirror has precision adjustment in the (X,Y) plane orthogonal to the axis of the column and zoom (working distance) axis of the SEM. This is because the mirror's rotationally symmetric axis and electron beam aperture should be concentric with the zoom axis of the SEM. A test specimen with known uniformity of angular emission can be used to align the system.
The 2D detector is mounted on the SEM stage. In an embodiment, this is a modular design which is installed when the chamber is vented or, in a further embodiment, attached to the stage using an airlock. The signal from the detector, either processed or unprocessed depending on the design is be fed outside the SEM chamber through an hermetic-feed through. In case of an airlock loading solution, then the electrical connection is by a socket.
If the parabolic mirror occupies excessive solid angle such that normal SEM imaging techniques are overly impacted for the purposes of the experiment, then a cut away in certain locations may be employed.
The field of view in an SEM can be large because there are few restrictions to the beam path and no optical foci conditions to be met. For this technique the field of view and beam mapping magnification is limited by these conditions. For this reason, in an embodiment, the specimen is translated on the SEM stage without altering the Z height in order to study specimens over a wider area. The electron beam focus can be kept stationary with different parts of a specimen irradiated by translating the specimen on a stage independent of the mirror optics. For a flat specimen, this can be achieved without altering the focus of the device.
Using the described apparatus, experiments can be performed in conjunction with bandpass filtering and polarization filtering. For example the same region of interest can be probed on a specimen and alternative filters are employed to record the image.
The algorithm which turns the 2D image into an angular representation (Theta/Phi) is a known mathematical co-ordinate transformation such as is detailed in Ju Ren and Jianlin Zhao “Measurement of a bidirectional reflectance distribution and system achievement based on a hemi-parabolic mirror” Optics Letters 35, 1458 (2010).
In an embodiment where rotation of the specimen without impacting other variables is required, then a mechanism for eucentric rotation is used, wherein there is no movement of the region of interest in the X/Y plane. This can be achieved on an X/Y stage using automated corrective translation. If the specimen or stage is tilted with regard to the rotation axis, then correction of this tilt will be required. As the focus must remain constant, then eucentric tilt of the region of interest will be required.
To tilt the specimen, in a further embodiment, eucentric tilt about one or two axis may be provided. These eucentric tilt and rotation movements will provide a more flexible system. For example, in case the specimen emits luminescence of a certain polarization, then a second image of defined specimen rotation should provide confirmation of the result. With all other variables constant, the difference between two images recorded with the specimen rotated by a set degree will prove the lack of rotational symmetry in the emission of the specimen.
While the invention has been described in detail and with reference to specific examples thereof, it will be apparent to one skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope thereof.
