Sample Requirements
Like all experiments in the TEM, data quality is only as good as the sample put in the TEM. In general, samples should be thinned to electron transparency using a method that does not introduce artifacts. The sample needs to be electrically conductive and rigidly supported by the sample holder. Surface damage layers should be minimized. For STEM experiments in particular, any contamination of the sample, its holder or the TEM vacuum by hydrocarbons, silicone oils or similar chemicals will lead to rapid sample contamination under the beam and must be avoided.
Select appropriate sample thickness
A successful EELS experiment begins with a clean, electron transparent sample. Depending on the data required, the sample needs to be between 0 – 6 times the local mean free path (mfp). For 200 kV electron, the mfp of most materials is on the order of 80 – 120 nm; however it has a strong variation with Z, density and the TEM accelerating voltage.
The sample geometry can also affect the optimal sample thickness simply due to sampling considerations. Consider the case of precipitates in a wedge-shaped sample. For thick regions, the precipitates will be abundant, but they represent a small fraction of the volume the beam interacts with. In addition, the EELS data may have high background due to multiple scattering in thick regions. In very thin regions, there may be no precipitates to measure. Grain boundaries and interfaces pose similar constraints.
Below are sample thickness guidelines that correlate to the type of EELS analysis you may want to perform.
Application |
Sample thickness (mfp) |
---|---|
Qualitative mapping (edges <800 eV) |
0.1 – 1.2 |
Qualitative mapping (edges >800 eV) |
0.3 – 2.5 |
Quantification |
0.1 – 1.0 |
Quantification (edges >1.5 keV) |
<1.5 – 2.0 |
Energy loss near edge structure (ELNES) mapping |
~0.0 – 1.5 |
Thickness maps |
<6.0 |
Plasmon energy analysis |
0.1 – 6.0 |
Low-loss EELS analysis |
0.1 – 3.0 |
Low-loss EELS analysis with deconvolution |
0.5 – 6.0 |
Sample's contribution to image drift
Common environmental factors, including vibration, thermal variations, and stray electromagnetic fields, can cause your sample to drift during an experiment. This drift may cause image blurring that impacts the resolution and quality of your data. Below are frequently encountered sources of drift related to the sample itself:
-
Charging
-
Symptoms – Sample drift or jumping will increase as more beam is applied or when the objective aperture is inserted or removed from the system
-
Solution – Coat your sample with carbon to make the sample conductive
-
-
Improper sample mounting
-
Symptoms – Sample vibrates or appears loose on the grid or in the holder
-
Solution – Verify how securely fastened the sample is to the grid or holder
-
-
Beam sensitivity
-
Symptom – Sample appears to be shrinking
-
Solutions
-
Coat your sample with carbon to make the sample conductive
-
Condition (controllably stabilize) the sample with a large area, low-intensity beam prior to starting your experiment
-
Reduce the electron dose
-
-
References
Malis, T.; Cheng, S. C.; Egerton, R. F. “EELS log ratio technique for specimen-thickness measurement in the TEM” J. Electron Microscope Technique 8 (1988) 193