Scanning Transmission Electron Microscopy Facility
Advantages of the BNL STEM
1. Quantitative imaging- (insert Blocker Fig)
All aspects of the STEM design are optimized for quantitative imaging.
The beam optics are such that detectors collect the same scattering angles even if the specimen is not flat.
Scan sizes (magnification-1) are calibrated to within 0.1%. Detector signals are normalized to give %
scattering independent of beam current (see #3, below). All signals are recorded digitally with low-dose
methodology (see #5, below).
2. No phase grain (insert simulated images)
Normal TEM imaging includes a large component of phase contrast that depends strongly on defocus, even reversing in some cases.
This is useful on specimens with low contrast (such as those embedded in ice), but severely complicates quantitative interpretation.
The STEM large angle detector integrates over a large volume of diffraction pattern (reciprocal) space, averaging out phase
contrast to leave only scattering contrast. This feature is highly valued in materials science since it gives images of atoms
exactly where the corresponding atoms are in the specimen. Furthermore, defocusing the STEM beam gives a blurred image with no
net change in contrast. The small angle STEM detector can show some diffraction effects on crystalline specimens and is not
usually used for mass mapping.
3. All transmitted electrons are detected with quantum-efficient detectors
STEM detectors are europium-doped calcium fluoride scintillators mounted on quartz light pipes coupled to photomultiplier tubes.
More than 99% of electrons transmitted through the specimen strike one of the dark field detectors or the bright field detector.
Summing the detector signals gives a direct measure of beam current (dose) and dividing each signal by the sum gives scattering as
a percentage (the number needed for mass measurement).
4. Cold specimen stage (insert mass loss graph)
In the BNL STEM the specimen stage is cooled by contact with the cold pole pieces and separator block, creating a
very clean environment for the specimen. Cooling the specimen to Ė160C reduces the rate of mass loss per unit dose
by a factor of four. STEM operation at room temperature (especially with biological specimens) causes buildup of
contamination in the scanned area. The same happens in TEM over a much larger area, but in the STEM it is much more
obvious when viewing a lower magnification image covered with contamination squares. Cooling the sample below Ė40C
eliminates such contamination. At Ė160C, repeated scans at the highest magnification show no buildup or etching of
the substrate. Unfortunately the biological specimen does deteriorate and we routinely check dose-response curves for
mass loss and resolution loss to stay within limits.
5. High visibility of heavy atoms & clusters - (insert gold cluster picture)
The STEM large angle dark field detector signal gives the highest signal-to-noise ratio on single heavy atoms and
clusters, as well as the fewest false positives. The carbon substrate is not truly amorphous but contains small clumps of
ordered atoms that can give bright spots similar in size and intensity to atoms. The small angle dark field detector is much
more prone to this and actually provides an easy way to tune up the probe size (resolution) even when there are few of no heavy
atoms present. The STEM display provides side by side viewing of both images. Clusters are much easier to see than atoms,
even in internal cavities of a comples or in the presence of negative stain.
6. Low dose first-scan imaging -(insert AOI picture)
The fact that the beam only irradiates the area scanned makes low-dose methods much simpler than in TEM.
Since the dose is proportional to magnification squared, the idea is to scan a large area at the lowest
magnification where specimen objects or TMV controls can be seen (very low dose). From that base image one
selects areas of interest (AOIís) for focusing (high mag.) next to areas for recording image data by positioning
boxes and clicking with the mouse. The range of DC offset for AOIís at full resolution is +/-8u. The STEM control &
data acquisition scans the AOIís in sequence each time a key is pressed. If an image is worth saving, an additional
keystroke saves it to disk along with a header containing all relevant STEM operating parameters. At any point in the
sequence one can override using manual mode to follow up on interesting features. The computer also keeps track of
the accumulated dose at every 1u square on the grid to avoid returning to an area already imaged.
7. Efficient searching -(insert LoMag screen view, share with 6?)
The BNL STEM has a LoMag mode obtained by switching off the objective lens and focusing the beam on the specimen with
the condenser lens. This permits viewing the entire grid (2mm field) with 10nm resolution. This is recorded at
2048x2048 resolution whenever a new grid is inserted. A variation of the automatic search feature of PCMass traces
all holes in the holey film (usually ~3,000), measures the thickness of each and displays the result with empty,
single thickness and multiple thickness carbon areas color coded. Pointing to an area of that image with the mouse
causes a zoomed view of the selected area to be displayed below the first image and a mouse click causes the specimen
stage to move to that area. This makes selecting promising areas for imaging very rapid. The computer also gives a
warning if one returns to an area previously scanned.
8. TMV internal control in every specimen
Tobacco mosaic virus (TMV) is very stable and has a well-know cylindrical structure with 2.3nm per turn,
18nm diameter, 300nm length and a mass per unit length of 131kDa/nm. It is distinctive in shape so not to
be confused with an unknown specimen and is very pure. It serves as a useful control on multiple fronts.
Its attachment to normal carbon is quite reproducible so any aberrant distribution of TMV is a warning that
the substrate may have unusual properties in that area. Departure from cylindrical symmetry or filling of the
central hole (see radial profile analysis) is usually a sign of salt accumulation (also increasing M/L and
usually increasing SD above the typical 2%). The corners at the ends of the cylinders have a radius of
curvature of ~2nm at low dose, rounding off as dose increases and providing an additional monitor of
accumulated dose. The most important use is checking the M/L calibration of the STEM.
9. Image simulation to compare with theory
We use the Kirkland software suite to simulate any electron optical system. Most often we use atomic coordinates for
known specimens (such as TMV), random coordinates with nearest neighbor constraints for amorphous materials and known
microscope parameters to calculate simulated images for comparison to experiment. This is also useful for simulating
labeling experiments to see if the S/N would be high enough to justify the effort. We can study the effect of different
STEM operating parameters or aberration correction (available on instruments at the NanoCenter). It is informative to
compare STEM and TEM with optimal operating conditions on mathematically identical specimens, either freeze dried,
embedded in ice or in stain.
10. PCMass software is optimized for STEM images
Image data and software are available for download on our FTP site
The PCMass software is optimized for critical viewing of images in comparison to models and for
quantitative mass mapping (see Analyzing STEM Images).
Last Modified: June 12, 2009
Please forward all questions about this site to: