How to Choose a WDS

There are several types of Wavelength Dispersive Spectrometers (WDS) for electron beam micro-analysis and this discussion is the opinion of the inventor and manufacturer of the Parallel Beam Spectrometer so it is necessarily biased.

Why do you need WDS

In the past, WDS had several reasons to exist, for awhile it was the only available x-ray spectrometer for micro-analysis but in the late 1900s Energy Dispersive Spectrometers (EDS) became available and represent nearly the entire market for x-ray spectrometers on electron microscopes. However, there remains a need for WDS due to its strengths relative to EDS.

Originally, EDS had poor performance for light elements because ultra-thin windows were not available so WDS had to be used for light element analysis. The excellent energy resolution of WDS relative to EDS was also a reason for it to exist. Today, with ultra-thin windows, EDS does a reasonable job of detecting light elements down to Be so this reason has mostly gone away for conventional Rowland Circle WDS but a modern Parallel Beam Spectrometer WDS has considerably higher count rates/nano-amp for light elements than an EDS so light element performance remains one of the reasons for using a parallel beam WDS.

EDS has always been count rate limited with high count rates causing high “dead time” so EDS could not be used with ultra-high beam currents whereas the proportional counters of WDS were essentially unlimited in count rates. Today, Silicon Drift Detector EDS also has high count rate ability so this reason for WDS is also mostly gone. High count rate ability remains a reason for WDS for use on micro-probes where beam currents may be very high.

Energy resolution has always been the REAL reason for using WDS and EDS cannot hope to match the spectral resolution of WDS. An EDS might have a resolution of 126 eV while a WDS has a resolution of better than 20 eV and often below 10 eV. With such high resolution, it possible to do spectroscopy on closely spaced spectral lines that would not be resolved with EDS. This is particularly true for the very low energy spectral region where K, L, and M lines are all very close together so that EDS becomes useless. If you really want to do analysis using low electron beam voltage, EDS can be nearly useless because the high energy K and L lines are not available and the low energy lines are too closely spaced to resolve. If you do much work with electron beam voltages of 5 KV or less, you really need WDS.

Rowland Circle or Parallel Beam WDS

Rowland Circle WDS is the original type that uses curved diffractors. Parallel Beam WDS only became available in 1996 when practical x-ray collimating optics became available. The Rowland Circle instruments use curved diffractors that move along a circular path and a detector that moves to intercept diffracted x-rays. Parallel Beam WDS uses an x-ray optic that collects a diverging beam of x-rays and re-directs them into a parallel beam and then onto flat diffractors and then into a detectors that moves.

Rowland Circle

Advantages
Very good performance for higher energy
Relatively insensitive to sample height (not true for vertical types)

Disadvantages
Very poor performance for lower energy
Horizontal inclination takes up a lot of room around the SEM
Requires a WDS port
Often massive enough to require customizing the SEM supports

Considering that nearly all of the higher energy problem areas such as closely spaced transition element lines can be resolved by using lower energy L lines, the problems of these older instruments with performance at lower energy where you really want WDS and the requirements for a special port and the size, this author cannot understand why these instruments still exist.

Parallel Beam WDS

Examples include the Thermo MaXRay, the Parallax Research LEXS, HEXS and the EDAX LambdaSpec and EDAX TEXS.

Parallel Beam WDS using only grazing incidence collimators
This includes the original MaXRay, the LambdaSpec, LEXS, and HEXS.

MaXRay
Although this author invented the MaXRay, he is not familiar with how Thermo implemented it.

Advantages
Very good performance for lower energy (below 2 KeV)
Vertical inclination takes up little SEM room
Does not require a WDS port

Disadvantages
It seems fairly large for what it does
Probably limited to less than 2 KeV with the original grazing incidence optics
Medium sensitivity to sample height.

LEXS
Advantages
Very good performance below 2.2 KeV
Can be used for energies between 2200 and 3300 eV.
Vertical inclination takes up little room
Small package
Does not require WDS port
Auto-gate valve and optic positioning

Disadvantages
Medium sensitivity to sample height
Cannot be used above 3300 eV

LambdaSpec
Advantages
Very good performance below 2.2 KeV
Vertical inclination takes up little room
Small package
Does not require WDS port

Disadvantages
Medium sensitivity to sample height
Cannot be used above 2200 eV

HEXS
Advantages
Very good performance below 2500 eV
Can be used up to 4500 eV
Vertical inclination takes up little room
Small package (not as small as LEXS/LambdaSpec but smaller than MaXRay)
Does not require WDS port
Auto-gate valve and optic positioning

Disadvantages
Medium sensitivity to sample height
Cannot be used above 4500 eV

Parallel Beam WDS using Polycapillary Optics

TEXS
Advantages
Medium performance below 1000 eV
Good performance above 1000 eV
Vertical inclination takes up little room
Small package
Does not require WDS port

Disadvantages
Extreme sensitivity to sample height.
Only moderate performance below 500 eV and poor performance for Be
Use of a diffractor for higher energy requires even further performance degradation below 1000 eV, usually for the Be and B region.

MaXRay ER
The author is not familiar enough to assess this spectrometer.