Several Key Issues to Understand in SEM-EDS Analysis

If you need to analyze the types and contents of elemental components in a material's micro-region, there are often multiple methods available. EDS analysis is the most commonly used approach.

EDS is simple to operate, provides fast analysis, and delivers intuitive results. Most importantly, compared to high-end electron microscopes, it is relatively inexpensive, making it a standard feature in modern electron microscopes.

This article compiles various common questions about EDS (Energy Dispersive Spectroscopy) in the hope of providing useful insights.

Q1: Is the abbreviation for Energy Dispersive Spectroscopy EDS or EDX?

Initially, there were multiple abbreviations for energy dispersive spectroscopy, including EDS, EDX, and EDAX. People generally understood that "ED" stood for Energy Dispersive, but differences in the use of "X-ray Analysis" and "Spectrum" led to variations in the abbreviations. Similarly, different Chinese translations emerged, such as energy-dispersive spectroscopy and energy-scattering spectroscopy.

However, around 2004, relevant associations established that EDS refers to the spectrometer or the spectroscopy technique, while EDX refers to the study of spectroscopy. The term "Dispersive" was left untranslated.

Thus, EDS is the standard usage in academic papers. However, many articles still use alternative terms due to established conventions, and as long as people understand the meaning, it's generally acceptable.

Q2: Does TEM EDS have less error than SEM EDS?

Many people assume that since TEM (Transmission Electron Microscopy) has higher resolution, the resolution of its associated EDS is also higher than SEM (Scanning Electron Microscopy). However, this is not necessarily true.

For the same manufacturer and product generation, the resolution of an EDS system used with TEM is typically a few eV lower than that of SEM. While TEM can observe finer details, this pertains to spatial resolution rather than the resolution of EDS itself.

SEM samples are easier to prepare, and the electron beam typically penetrates a few microns into the sample. During quantitative analysis, standard reference materials (e.g., pure Si for silicon analysis, MgO for magnesium oxide analysis) can be used for calibration. Heavier elements, such as metals and rare earth elements, can often be quantified with reasonable accuracy.

Professor Xiangting Li from the Shanghai Institute of Ceramics has systematically compared EDS analysis results from SEM and electron probe microanalysis (EPMA). Below are some key points:

The detection limit for EDS analysis is approximately 0.x%, where "x" varies depending on the element.
The "National Standard for Quantitative Analysis of EDS in Electron Probe and Scanning Electron Microscopy" specifies that, for flat, dry, dense, stable, and well-conductive samples, the total quantitative error should be less than ±3% (excluding samples containing ultra-light elements).

Additionally, relative error allowances for EDS analysis are as follows:

(a) For major elements (>20 wt%), the relative error should be ≤ ±5%.

(b) For elements with 3–20 wt%, the relative error should be ≤ ±10%.

(d) For elements with 0.5–1 wt%, the relative error should be ≤ ±50%.

For non-flat samples, results can be averaged over three analysis points. If the total quantitative error is ≤ ±5%, normalization can be applied, assuming no undetected elements. If the error exceeds ±5%, the results should only be considered semi-quantitative.

In many cases, standard reference materials are unavailable, so database values are used for quantification. This method, called standardless quantification, is widely applied in routine analysis but does not yet have an official national standard. However, efforts are underway to establish one.

For TEM samples, which are mostly thin films, reduced interference is an advantage, but accurate thickness measurement is required for quantification. Since determining micro-region thickness is challenging, quantification in TEM becomes problematic. Even with standard reference materials, it is difficult to prepare TEM samples of equivalent thickness for comparison. To date, there are no officially recognized national TEM reference standards.

Therefore, for many samples, TEM-EDS analysis is semi-quantitative, and for light elements, it may only provide qualitative results. The best practice is to choose an appropriate analysis tool, minimize interference, and average multiple data points (preferably over 20) to reduce errors.

Q3: If an element in an EDS spectrum has multiple peaks, does it mean its content is high?

Understanding EDS principles clarifies this question.

EDS works by bombarding a sample with electrons, which eject inner-shell electrons, leaving vacancies. When outer electrons transition to fill these vacancies, they emit characteristic X-rays. The energy differences between different electron shells create distinct spectral peaks.

Thus, the number of peaks correlates with the number of electron shells involved, not the element's abundance. Quantitative analysis relies on the intensity of specific spectral lines and corresponding response factors for each element, not simply the number of peaks.

Q4: If peaks for light elements are absent in an EDS spectrum, does it mean they are not present in the sample?

Not necessarily. Several factors could obscure the detection of light elements:

Presence of large particles or thick materials nearby can absorb the already weak X-rays from light elements, leading to interference.
In such cases, compare different sample areas, check the K and L line series of transition elements, or tilt the sample to assess changes in the spectrum. This helps evaluate the reliability of the results.

Q5: Why do EDS spectra sometimes show elements that shouldn't be present in the sample?

Several factors could cause this:

(a) Carbon (C) and Oxygen (O):

Organic contaminants (e.g., airborne oils) often adhere to sample surfaces.
Both TEM and SEM may show C and O peaks due to contamination.
In TEM, carbon film is commonly used as a support, making C peaks inevitable.

(b) Aluminum (Al) and Silicon (Si):

SEM samples are often mounted on aluminum stubs or glass substrates, leading to background signals.

These are specific to TEM.
Cu originates from TEM grids.
Cr often comes from the sample holder or chamber materials.

(d) Boron (B):

If B suddenly appears with an unusually strong peak, check for beam drift or heating effects. Large sample movements during scanning can also introduce false B signals.

(e) Rare earth elements or actinides (La, Ac series):

These may result from noise artifacts in the instrument's detection system.
Software post-processing can usually remove such spurious peaks.

By understanding these principles and limitations, users can better interpret SEM-EDS and TEM-EDS results and minimize errors in elemental analysis.