“Studies such as this serve to highlight the challenges of performing low-concentration technology-to-technology comparisons”


Barry Prince Syft Technologies Director - Global Sales

Dr. Barry Prince

BSc(Hons), PhD


Principal Scientist / Consultant

 

One of the biggest challenges we’ve faced at Syft over the years is benchmarking SIFT-MS accuracy for trace analysis applications.   The challenge with this is well described by the work of Bill Horowitz who showed how interlaboratory comparison data deviates increasing at lower concentrations – with RSDs exceeding 30% below 10 ppb!  If such big deviations are observed with interlaboratory studies, using the same instrument techniques, how much correlation can we reasonably expect from comparisons of trace-level samples using different technologies?

A real-world example of this was recently published as the output of a global research collaboration showing the measurement of the pollutant nitrous acid (or HONO) in Beijing, using a variety of measurement techniques.  HONO is an important compound for air quality researchers due to its role in the formation of radicals that go on to produce other dangerous pollutants.  However, HONO is a challenging species to measure in field conditions due to the reactivity of the compound and the potential for measurement interference from other species.  For example, the researchers report in this work that a PTR-TOF-MS was unable to measure HONO despite using the same measurement principle (H3O+ reagent) as SIFT-MS (presumably due to the high ionization energy employed in the PTR-MS drift tube, compared to the SIFT-MS flow tube).

Due to the challenges of measuring HONO, the researchers employed four instruments they refer to as “established” techniques, utilizing wet chemistry, spectroscopic and offline analysis.  The results presented show significant deviations in reported concentrations – up to two times – between these techniques.  In fact for half the campaign, none of the instruments agreed within their stated measurement errors!

In addition to these so-called established methods, the researchers also carried out measurements using a Syft Voice200ultra.  In the context of the four benchmark measurements, the Voice200ultra performed admirably, with results falling comfortably within the window of those measurements.

But of course, there can be only one true answer to the question of what is the concentration of HONO? To answer that question, we must ask which one of these analytical techniques (if any) is reporting the correct concentration?  That is not an easy question to answer, and indeed the authors are unable to definitively account for the discrepancy.  Nevertheless, it stands to reason that a technique that employs less steps between sample capture and data output should be less prone to measurement bias, since there will be less opportunity for bias to be introduced.

By that guide, it is hard to imagine a better choice than SIFT-MS since the analyte ‘chain of custody’ is extremely short:  Sample is delivered continuously through a heated tube directly to the reaction chamber, and the analysis is governed by the fundamental ion-molecule reaction rates, which are written in stone – independent of sample concentration or matrix effects.

No doubt, benchmarking SIFT-MS for trace-analysis applications will continue to be a challenge for us for sometime to come, but I am encouraged by the appearance of such comprehensive studies as that described by these researchers.  Studies such as this serve to highlight the challenges of performing low-concentration technology-to-technology comparisons, and the magnitude of resources needed to achieve a meaningful result from such an undertaking.

The research paper discussed can be found at: Atmos. Meas. Tech., 12, 6449–6463, 2019