Question 1

Which spectrometer entrance aperture should I use?

Accepted Answer

In the case of compositional analysis for elemental distribution or quantification, the 5 mm aperture is better suited because it allows a bigger collection angle and more signal into the spectrometer. For chemical analysis, the 2.5 mm is better suited since it delivers a slightly higher energy resolution.

Question 2

Which dispersion should I use?

Accepted Answer

The dispersion you use depends on the energy resolution you need. Like imaging, the higher the resolution the higher the dispersion (magnification). However, as the dispersion increases the energy range decreases.

For compositional analysis, the lowest dispersions are necessary to achieve the largest energy range, which will maximize the number of edges that can be acquired at one time. A larger dispersion would be give better energy resolution for ELNES analysis. A chart of dispersions, and their applications, for the GIF Continuum is displayed below.

Camera	Dispersion (eV/ch)	Energy Range	Chemical Mapping	ELNES Analysis	Low-Loss
GIF Continuum	1.5	3,000	Preferred	Not Suitable	Not Suitable
K3	0.9	3,000	Preferred	Not Suitable	Not Suitable
Stela	3	3,000	Preferred	Not Suitable	Not Suitable
GIF Continuum	0.3	614	May Work / Context Dependent	Preferred	Not Suitable
K3	0.18	622	May Work / Context Dependent	Preferred	Not Suitable
Stela	0.6	617	May Work / Context Dependent	Preferred	Not Suitable
GIF Continuum	0.05	102	Not Suitable	May Work / Context Dependent	Preferred
K3	0.09	311	Not Suitable	May Work / Context Dependent	Preferred
Stela	0.1	103	Not Suitable	May Work / Context Dependent	Preferred

Question 3

Which camera length should I use?

Accepted Answer

In general, the camera length that you use should give you a convergence to collection angle of 1:2 or 1:3. On most microscopes, a camera length between 1.5 – 4 cm will give you the correct ratio, when using the largest entrance aperture (5 mm). As the convergence angle increases, the camera length will need to decrease to maintain this ratio. For more information please read the section on collection angle considerations on EELS.info.

Learn how to measure your convergence to collection angle ratio by taking our free self-paced course EELS-STEM Alignment Tool.

Question 4

Why is the spectral acquisition rate slow?

Accepted Answer

Use higher vertical binning in the camera setup. Ideally you want to run most of the experiments with the camera set with the highest vertical binning.

Question 5

Why is the low energy portion of my electron energy loss spectrum flat?

Accepted Answer

The electron energy loss spectrum is probably saturated, since the low energy portion shows the highest number of counts. Reduce the exposure time or the binning.

Question 6

Why is my spectrum noisy?

Accepted Answer

More signal is needed. Use a larger spectrometer aperture, increase vertical binning, have a longer exposure or higher beam current (provided the sample can stand the electron beam).

Question 7

In my STEM EELS acquisition, the interfaces that were straight in the survey image are tilted or wavy.

Accepted Answer

The waviness in the image is likely caused by sample drift or charging of the sample during acquisition. To mitigate artifact caused by sample drift you should enable drift correction. Also try to increase the acquisition speed by reducing the pixel dwell time in the SI, and use multiple passes to acquire the data rather than one, long single pass. If the drift rate is too fast for drift correction to keep up with, then it is recommend letting the sample sit in the microscope to allow for sample rod and goniometer to mechanically settle.

For sample charging, you can coat the TEM sample with a very thin later of amorphous carbon using a standard vacuum carbon coater. A customer scan pattern such as spiral scanning can also be used to mitigate charging on the sample.

Question 8

How do I improve my energy resolution?

Accepted Answer

The measure of energy resolution is the full width half maximum (FWHM) of the zero-loss peak (ZLP). This is ultimately limited by the type of electron source that you have on your microscope, and whether the microscope is equipped with a monochromator.

To improve the energy resolution start by using the AutoFocus routine in the tuning pallet. You can also manually adjust first and second order focus (FX, FY, SX, and SY) while viewing the 2D spectrum view to try and make the ZLP as sharp as possible. You should only need to adjust first order focus, FX and FY. For more information watch the free Aligning the Spectrometer self-paced course.

The dispersion you use will also affect the energy resolution as every detector has a point-spread function; how much the signal from a point source is spread throughout neighboring pixels. To resolve a spectral feature, the sampling interval should be 3x the feature width; to resolve a feature in the ELNES that is 0.3 eV wide you should use a dispersion of 0.1 eV or higher.

Question 9

How often should I correct for the spatial drift during a STEM SI acquisition?

Accepted Answer

This depends on the drift rate. The faster the rate the more frequently you need to correct for drift. It should be noted that the more frequent drift needs to be measured and corrected for, the longer the acquisition will take.

Question 10

Why is the map that I acquired fuzzy or have poor resolution?

Accepted Answer

A common cause is excessive chromatic aberration in the image. Follow the below steps to solve this issue:

Use a smaller objective aperture (typical aperture sizes are 8 – 24 mrad at 200 kV)
Limit chromatic aberration by using a smaller energy selecting slit

Note: Both of these methods will reduce your image intensity. Increase the beam current or exposure time to compensate.

Poor focus and image stigmation are common contributors. Follow the below steps to solve this issue:

Stigmate the image in normal phase contrast mode using the Diffractogram method if possible. Be sure to have the same objective aperture inserted while stigmating as it will be used for the EFTEM data acquisition.
Acquire a View mode camera image at an energy offset of 400 eV (the inelastic is more sensitive than the elastic image to poor alignment)
- Slit width should be about 50 eV
- The energy and slit width can be configured via the Filter State control of the GIF and EELS Control UI (aka AutoFilter)
Focus the image at this energy loss. Look to optimize the sharpness of the features. You can condense the beam smaller than the field of view to get more current if needed. If you do, be sure to optimize the beam to just larger to image field of view before data acquisition.

Question 11

Why do I see bright and/or dark bands at interfaces?

Accepted Answer

When you see bright and/or dark bands at the interfaces of computed maps, this means your sample may be drifting:

Decrease exposure time (should not exceed 30 s for each frame)
If more than one frame is chosen select align images to remove the spatial drift
Ensure that spatial drift has been removed by aligning the images correctly before extracting elemental maps

Another cause is the application of excessive smoothing of the power law background exponent. This option is set in the EFTEM Mapping Preferences dialog. Consider reducing the smoothing half width or changing to a straight pixel-by-pixel power law model. The maps can be recalculated using the menu item EFTEM / Compute elemental maps without reacquiring the data.

Question 12

What size objective aperture should I use?

Accepted Answer

The main function of the objective lens aperture is to limit chromatic blurring, so your elemental maps are not out of focus. It also helps increase the jump-ratio (signal-to-background) of the energy loss edge. Generally, an aperture on the order of 8 – 16 mrads at 200 kV is adequate. For more information about collection angles see the EFTEM-Imaging Mode section in the Collection Angle Considerations page on EELS.info.

Question 13

My map and/or images are noisy. What can I do?

Accepted Answer

There are a few variables which could cause your maps/images to be noisy.

Sample thickness
- Acquire a thickness map to determine the sample thickness. If the sample is thick (e.g., relative thickness >1), try moving to another sample area. Increase the number of collected electrons by one (or all) of the following:
Low-Beam Intensity
- The number of electrons decreases with increasing energy loss. The strength of the condenser lens may need to be adjusted with the brightness knob.
Short exposure time
- Since there are fewer electrons at higher energy losses you can increase the exposure/live time of the camera. You can also implement hardware or software binning on the camera. However, this will degrade the resolution of the image.
- With the K3 or K2 camera you can increase the number of frames counted.
Narrow slit width
- You can increase the number of electrons in the image by widening the slit width. However, you may notice some image blurring at higher magnifications.

Question 14

I see very sharp, periodic contrast that is not from the sample in bright areas, why?

Accepted Answer

If you see very sharp contrast that is not from the sample in bright areas, you may have saturated the detector in ones of the images. Decrease the detector exposure time and/or the sample illumination, then repeat.

Question 15

Why is the intensity of the EFTEM SI negative at high energy losses?

Accepted Answer

This is typically caused by afterglow of the phosphor. To minimize this effect:

Acquire the data in reverse direction using the “Acquire high to low” option in the spectrum imaging software
Avoid exposing the camera to the unfiltered or elastic (ZLP filtered) image prior to acquiring a core-loss EFTEM SI
Do not use the option “Align ZLP before acquire” option in the spectrum imaging software

Question 16

Why does a strange shadow exist in the images and maps?

Accepted Answer

This is caused by tuning of the Filter or an incorrectly centered zero-loss peak (ZLP). It is good practice to tune the filter (using the Tune for spectrum focus option in AutoFilter) at the magnification you will be using to acquire your maps. Do this at the early stage of the setup to avoid scintillator afterglow.

Question 17

Why does a semicircular shadow appear in a corner of my EFTEM SI (the shadow may shrink into a circle or drift across the image field of view)?

Accepted Answer

This is typically caused by a poorly calibrated Condenser Adjustment. As the HT is changed to acquire the EFTEM SI, an automated adjustment must be made to the condenser focus and deflectors. To run this calibration, make sure the system is in Power User mode and run the menu item “EFTEM | Calibrate | Calibrate Condenser Adjustment …”. This is best done with a sample in the TEM and moderately thick area in the field of view (1 – 2 mean free paths). Simply follow the online prompts.

Question 18

Why does a strange shadow exist in the images and maps?

Accepted Answer

This is caused by tuning of the Filter or an incorrectly centered zero-loss peak (ZLP). It is good practice to tune the filter (using the Tune for spectrum focus option in AutoFilter) at the magnification you will be using to acquire your maps. Do this at the early stage of the setup to avoid scintillator afterglow.

Electron energy loss spectroscopy (EELS)

Which spectrometer entrance aperture should I use?

Which dispersion should I use?

Which camera length should I use?

Why is the spectral acquisition rate slow?

Why is the low energy portion of my electron energy loss spectrum flat?

Why is my spectrum noisy?

In my STEM EELS acquisition, the interfaces that were straight in the survey image are tilted or wavy.

How do I improve my energy resolution?

How often should I correct for the spatial drift during a STEM SI acquisition?

Why is the map that I acquired fuzzy or have poor resolution?

Why do I see bright and/or dark bands at interfaces?

What size objective aperture should I use?

My map and/or images are noisy. What can I do?

I see very sharp, periodic contrast that is not from the sample in bright areas, why?

Why is the intensity of the EFTEM SI negative at high energy losses?

Why does a strange shadow exist in the images and maps?

Why does a semicircular shadow appear in a corner of my EFTEM SI (the shadow may shrink into a circle or drift across the image field of view)?

Why does a strange shadow exist in the images and maps?

Electron energy loss spectroscopy (EELS) 