Mauro Gemmi
1
, Stavros Nicolopoulos
2
1.
Center for Nanotechnology Innovation@NEST, Istituto Italiano di Tecnologia, Pisa Italy
2.
NanoMEGAS SPRL Blvd Edmond Machtens 79 , B-1080 Brussels Belgium
The study of beam sensitive materials (zeolite, hybrid mesoporous crystals, organic crystals,
pharmaceutical and proteins) with electron diffraction (ED) is a challenging task as very often those
materials are difficult to be crystallized in crystals big enough/or in enough quantity to be studied with
X-Ray diffraction. Therefore, if a low dose single crystal ED technique would be available, it should be
a breakthrough in crystal structure analysis. For TEM based crystal structure determination, very
promising results have been obtained with the precession diffraction [1] assisted ED tomography
technique [2] in which a series of patterns are collected every 1° while the sample is tilted around the
goniometer axis. A drawback of the technique is that after each tilt step the crystal moves, therefore a
method of automation for tracking the crystal position must be developed. There is (limited to some
TEM platforms) an automated version of diffraction tomography, where crystal tracking is performed in
STEM mode and the diffraction is collected in nanobeam mode [3]. In this way electron radiation dose
is minimized with respect to a crystal tracking based on conventional TEM imaging. One critical aspect
of the latter procedure is however long data collection times, between 1 and 2 hours, due to dead time
while changing between diffraction / STEM image mode and due to the CCD slow speed data collection
procedure.
To speed up diffraction pattern data collection we tested at first a faster detector. As a proof of concept
we collected an entire ED tomography experiment on a mayenite crystal Ca
12
Al
21
O
33
by recording (tilt
range +/- 60°) a total of 121 patterns on a Zeiss Libra 120 with ultra-fast Stingray CCD camera using the
same experimental configuration for TEM orientation imaging technique (ASTAR) described in [4]. The
Stingray CCD has 256 grey levels dynamic range and a limited number of pixels (configuration used
512x512) but it can record up to 100 frames/sec. For the ASTAR technique set up we used a
commercially available system[5]. Since the Stingray CCD is external to the column and its optical axis
is not normal to the fluorescent screen, it introduces distortions in the collected diffraction patterns.
After correcting for such distortions, the patterns can be processed with the ADT3D software [3] to
obtain unit cell parameters and diffraction intensities. In our experiment, the mayenite unit cell
parameters have been obtained with an error of 2% and the collected ED intensities allowed the
complete structure to be solved with direct methods.
As a second step we tried to automate data collection. We observed that while tilting the TEM holder,
the sample always moved along a definite direction for a specific (quite large) tilt range. However the
crystal movement could be tracked by shifting the beam along the same direction and the diffraction
patterns could be simultaneously recorded with the Stingray CCD while working at high speed (> 40
frames/s). We were able to do an entire diffraction tomography data collection during the time needed
by goniometer to automatically cover a large tilt range. For that we used the ASTAR control unit for
TEM where it is possible to scan a rectangular area with a certain scan step (from 1 to several 100 nm)
and to collect an ED pattern at each point with the Stingray camera. For our purpose we defined the
rectangular area elongated along the direction of crystal movement during the goniometer tilt and we
have been able to track relative crystal movement for a tilt of 40°. A total of 362 diffraction patterns for
40° of tilt have been recorded with an angular resolution of nearly 0.1° where the entire data collection
700
doi:10.1017/S1431927613005497
Microsc. Microanal. 19 (Suppl 2), 2013
© Microscopy Society of America 2013
https://doi.org/10.1017/S1431927613005497
Downloaded from https://www.cambridge.org/core. IP address: 54.163.42.124, on 28 May 2020 at 21:51:23, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.