LIBS Info: Understanding: What does LIBS (instrumentation) need to do better, if LIBS is going to actually become the “next big thing”?

What does LIBS (instrumentation) need to do better, if LIBS is going to actually become the “next big thing”?



Long Term Reproducibility:
My experience has been that with most LIBS instrumentation, getting consistent results across the an afternoon is a common occurrence. However, take those same samples back a week later and the results are considerably less reliable. This comes down to both engineering and methodology.
In “competing” technologies (eg. ICP-OES, XRF, Arc-Spark OES) there is an expectation that the instrument will hold calibration for months or years. And there are established methodologies for ensuring that:-
  • in ICP-OES, samples of known concentration are interspersed amongst the unknowns and the instrument software corrects the calibrations as the analysis progresses.
  • in XRF or Arc-Spark, while a full calibration may take 20/30/50 standards and happen every couple of months/years, there are a couple of “check standards” which are run daily. The instrumentation software then adjusts the main calibration for the instrument drift based on measurement of these standards.

So, we need to engineer instrumentation to be reproducible:-
  • Construction/spectrometers/lasers/optics that hold their specification over time and are immune to/compensated for changes of temperature and humidity
  • As our Arc-Spark OES colleagues have done over many years, we need to engineer the sample space so that the LIBS spark is consistent
  • We need to develop a “standard methodology” for running samples and checking the calibration that works for LIBS and incorporate that it the instrument software so that it is automated and easy to do.

Instrumentation that facilitates measurements that only LIBS can do
The majority of turnkey LIBS instrumentation available to date has been “Analytical Lab” machines - aimed at replacing an XRF/Arc-Spark OES/ICP in a commercial lab. While there have been some successes in this space, it has been hard as they face entrenched competitors – established techniques backed by large companies. And, once the sample is in the lab, many of the advantages of LIBS have been lost.
Instead we should be developing instrumentation that enables people to solve problems only LIBS can do – out in the field where you can’t take your expensive/heavy ICP or Mass Spec, online where you can’t do XRF, at range or in places that you can only get light in and out of. The handheld LIBS units and conveyor belt analysers are a great start, but no one has made a rig that enables researchers/engineers to try it on their problems.
However, this immediately leads to the next problem.

Solving the Remote Calibration Problem
Getting spectra from a difficult location is only half the problem. We need to be able to turn that into concentrations (with a reasonable estimate of their uncertainty).
Whether this is using the Calibration Free models, or a method that reduces the matrix effects or enables a local measurement to calibrate a remote one. Isn’t particularly important. It just needs to work and work reliably and without much user interpretation.

Got some thoughts - send us a comment