Spectroscopy operation

In spectroscopy, an image of the projector lens crossover is magnified and projected onto the camera. When properly adjusted, the observed CCD image consists of sharp vertical lines that represent electrons of particular energy. Displacement of lines in the horizontal (or dispersive) direction indicates a difference in energy between electrons.

The projector crossover is present under typical transmission electron microscope (TEM) or scanning (S)TEM operating conditions (in some dedicated STEMs a virtual cross over is formed). Movement of the crossover up and down in the column slightly depends on the exact operating mode, but you can easily compensate for this movement by making small adjustments of the pre‐prism quadrupole Focus X. Thus you can tune the spectrometer for any operating condition of the microscope.

The spectra you acquire will correspond to the feature in the center of the microscope viewing screen; this is the area selected by the spectrometer entrance aperture when you lift the screen up. This area can be a precipitate image in TEM (e.g., a Bragg beam from a diffraction pattern) depending on the operational mode. Three basic operation modes are possible: TEM Imaging, TEM Diffraction, and STEM that will be discussed below.

Note: Due to mechanical variability of the TEM, viewing screen and spectrometer mount, the center of the TEM screen (or viewing camera) is not at the exact position of the spectrometer entrance aperture. This is normal and will not affect the operation of the system. For STEM electron energy loss spectroscopy (EELS) acquisition, the alignment of the ADF STEM detector and the entrance aperture is important. Detectors from Gatan should be concentric with the entrance aperture and are adjusted at installation. For detector from 3rd party or TEM suppliers, please contact the appropriate organization for assistance.

TEM imaging mode

With the TEM in the Imaging mode, an image is formed on the viewing screen, while the projector crossover contains a small diffraction pattern. Because the spectrometer projects an energy dispersed diffraction pattern onto the detector, this mode is also known as diffraction‐coupled. The camera length \(L\) of the pattern in the projector crossover is given by

\(L = \frac{h}{M}\)

and the diameter of the projector crossover, \(d_{p}\), is

\(d_{p} = \frac{2 ß h}{M}\)

where

  • \(h\) = distance from the projector crossover to the viewing screen
  • \(M\) = image magnification at the viewing screen
  • \(ß\) = half acceptance angle as defined by the objective aperture

In practical terms, with \(h\) = 50 cm and at \(M\) = 10,000x, \(L\) = 50 μm. A diffraction pattern that encompasses angles to 50 mrad is therefore only 5 μm in diameter at the projector back focal plane and it becomes even smaller at larger microscope magnifications. This means that you can attain good energy resolution in the TEM imaging mode while it accepts practically all the scattered electrons from a particular specimen area. If you require a smaller range of scattering angles, you can select this when you use the objective aperture.

Operating in the TEM imaging mode is convenient because you can easily modify the spectrum intensity via the illumination setting and/or size of the condenser aperture. The selected specimen area is directly visible on the viewing screen and you change it by translating the specimen, shifting the image electronically, changing the microscope magnification and/or the spectrometer entrance aperture. The collection efficiency can be very high in this mode. These characteristics make the TEM imaging mode ideal for the initial examination of any specimen.

Note: Only operate in TEM imaging mode for spectroscopy during the initial examination, as it may give misleading analysis results.

The TEM imaging mode at first sight appears to allow the selection of a small specimen area for microanalysis by the spectrometer while a larger area of the sample is illuminated. This might seem a useful way to get a spectrum from a small particle or interface when it is impossible to get a small enough probe. A user with a LaB6 instrument may believe they can do analysis similar to an owner of a FEG instrument. However, they would be incorrect. The image seen on the TEM screen is from the region of the spectrum that has the most electrons, on a thin sample, the zero-loss. Unfortunately electrons that have lost energy are focused by the TEM objective at a different location due to its chromatic aberration (\(C_{c} \)). This means the image at the observed energy loss in the spectrum will come from a different area of the specimen. Thus this method will always give a false result, as shown below. Electrons that have been scattered by large angles and also suffer an energy loss will be displaced laterally in the image by a distance \(d\) given by

\(d = q\cdot \Delta f\)

where

  • \(q\) = scattering angle
  • \(\Delta f\) = defocus error of the objective lens of the microscope

The defocus error depends on the energy difference between electrons for which the objective lens was focused, and the energy loss electrons of interest. It is equal to

\(\Delta f = C_{c} \cdot \frac{\Delta E}{E_{o}}\)

where

  • \(C_{c}\) = chromatic aberration coefficient of the objective lens
  • \(E_{o}\) = primary energy

Focus is normally done while you look at the image formed by the zero-loss electrons. For a 100 kV primary energy electron which lost 1000 eV while being scattered over 5 mrad, the displacement (referred back to the sample) amounts to 100 nm (\(C_{c}\) = 2 mm). Thus, to obtain information from areas smaller than 100 nm, you should use a small probe, as in the STEM mode.

In addition, quantitative analysis is made nearly impossible in the TEM imaging mode for the same reasons. Electrons of different energies are spread over areas of different size in the image. If the illumination area is too small, or the specimen is not homogeneous, the intensity ratio between low and high energy edges can change dramatically. It is therefore almost always better to collect spectra for quantitative analysis in the diffraction mode unless the specimen is homogeneous over tens of µm and the illumination area size on the viewing screen is much larger than the spectrometer entrance aperture. In this case, the electrons that miss the spectrometer entrance aperture due to chromatic aberration will be replaced by a similar number of electrons that enter the aperture in error also due to chromatic aberration. Operation under these conditions is sometimes useful on contamination‐prone specimens, or when one needs to spread the illumination over a large area to minimize radiation damage.

TEM diffraction mode

With the TEM in Diffraction mode, a diffraction pattern is formed on the viewing screen, while the projector crossover contains a small image of the illuminated specimen area. Because the spectrometer projects this energy dispersed image onto the detector, this mode is also known as image‐coupled. If a small area of the sample is illuminate, the size at the projector crossover, \(d_{p}\), is given by

The magnification, \(M\), of the image at the projector crossover is given by

\(d_{p} = Md_{s}= \frac{h}{L}d_{s}\)

where

  • \(d_{s}\) = diameter of the illuminated specimen area
  • \(M\)= distance from the projector crossover to the spectrometer entrance aperture
  • \(h\) = camera length at the spectrometer entrance aperture
  • \(L\) = spectrometer energy dispersion and magnification factor

and you therefore determine the diameter\(D_{p}\)in eV at the beam trap of the projector crossover by

\(D_{p} = d_{p}K_{s} = \frac{h}{L}d_{s}K\)

where

  • \(K\) = spectrometer energy dispersion and magnification factor (~0.3 eV/μm for the GIF Quantum® system at 200 kV)

For example, a 3 μm diameter illuminated specimen area will give rise to a 3 μm projector crossover size for \(L\) = 70 cm and \(h\) = 70 cm. This permits better than 1 eV energy resolution at 200 kV. The TEM Diffraction mode is therefore useful to acquire spectra from large areas, although you must carefully monitor the size of the illuminated area when you use a small camera length, if the energy resolution is not to worsen considerably.

In the TEM diffraction mode, the spectrometer angular acceptance range is limited by the entrance aperture and can be varied when you change the camera length or select a different aperture size. The diffraction pattern visible on the viewing screen makes it possible to monitor the specimen phase, thickness and diffraction condition. When you do not use a selected area diffraction aperture, the spectrum is collected from the whole illuminated specimen area, and the spatial resolution is therefore determined purely by the probe‐forming performance of the microscope. In this case, unlike in the imaging mode, chromatic effects do not appreciably change the collection efficiency of the spectrometer in the Diffraction mode. This makes quantitative chemical analysis of small sample areas much more reliable.

Diffraction mode is ideal to acquire the high quality spectra necessary to look for low concentration elements, or when you want to perform extended fine structure (EXELFS) analysis. It is also very useful to closely monitor the diffraction condition and/or collection angles, as for instance in channeling experiments.

STEM mode

In a correctly set up STEM mode, a diffraction pattern is displayed on the viewing screen and the projector crossover contains an image of the probe. Electron‐optically, this mode is therefore similar to the TEM Diffraction mode.

In the STEM mode, it is easy to image a large area of the sample even though the illumination is focused into a small probe (via collection of a STEM image), and to position the probe accurately on the feature of interest (when you use the scan coils). This makes the STEM mode ideal to acquire EELS data from precise areas of the sample (e.g., precipitates, interfaces, cell membranes).

When the STEM probe scans over large areas of the specimen, the probe motion will transfer to the spectrum, thus giving a motion of the energy loss peaks. This effect is not apparent at high magnifications but becomes increasingly important as the field of view increases. The use of descan coils is recommended to remove the motion of the probe in the energy loss spectrum.