SPME Fundamentals: Are longer SPME extraction times really worth it?

Many laboratories approach Headspace (HS) – Solid-Phase Microextraction (SPME) as follows:  During SPME workflows, the extraction step is carried out until an equilibrium is reached between extracted analytes and phase coating. In this way, an attempt is made to extract the maximum amount of analyte and achieve the maximum sensitivity. In many publications, extraction times of 30 min or more are used!

But is the juice worth the squeeze!?

I am currently completing my fall internship at Restek and the project that I am working on deals with the development of a HS-SPME method for the analysis of nitrosamines using a SPME Arrow. In order to determine the optimal conditions for sample preparation, I have conducted a series of experiments for the analysis of the following nitrosamines:

NDBA	N-Nitrosodibutylamine	NMEA	N-Nitrosomethylethylamine
NDEA	N-Nitrosodiethylamine	NPIP	N-Nitrosopiperidine
NDMA	N-Nitrosodimethylamine	NPYR	N-Nitrosopyrrolidine
NDPA	N-Nitrosodipropylamine

To determine the best possible extraction conditions, I varied the extraction time (60, 120, 240, 480, 960, and 1920 s) and plotted it against the compound response. That’s when I noticed an interesting trend, which can also be seen in one of Jason’s blog posts here.

From the graphs, you can see that as we doubled the extraction time our compound responses increase until they hit an equilibrium point, as expected. What is interesting, is that the greatest increase in compound response happens between the shortest extraction times. NDBA, for example, increases from 2.29 million counts at an extraction time of 60 s to 7.73 million counts at 120 s. This corresponds to a response improvement of 238%! However, if you increase the extraction time from 960 s to 1920 s, the response increases from 44 million to 48 million counts. This corresponds to only a 9% improvement.

So, is the small gain in response needed at such long extraction times? We say that you simply don't get the return on investment by extracting longer beyond a certain point, but that it can even be counterproductive if you overdo it because longer extraction times (i.e., 1920 s) lead to decreased responses for NDPA, NMEA, NPYR, NDEA and NPIP when compared to 960 s. You can see what appears to be the exact same phenomenon for traditional SPME and SPME Arrow in Jason’s blog post published five years ago. Our hypothesis for the decrease in response over long extraction times is that volatiles are escaping the HS vial due to a leak. Where is this leak coming from? Our best theory is that as the analytes are being extracted by the SPME coating, the SPME fiber/Arrow is stretching the vial’s septum (thereby creating leaks) due to all the extended fiber/Arrow motion while in the Heatex/Agitator.

If you then also look at the time needed for sample preparation, it becomes clear very quickly that longer extraction times are not always the best idea for increased sensitivity. In fact, we see unjustifiably long extraction times far too often, even though numerous papers show that short extraction times work perfectly well. ^{[1] [2]}

Next, I assumed that at least the reproducibility would increase with higher extraction times. However, as you can see from the distribution of the %RSDs in the figure below, this is not the case either.

I achieved the best average reproducibility after 960 s (16 min), and not after 1920 s (32 min) as I would have expected. However, since the %RSD is almost always in the single digits even at 480 s (8 min), an economically working laboratory should ask itself whether this small improvement in reproducibility, which can be seen in the table below, is worth extracting every single sample for 16 min instead of 8 min. Especially considering that other officially used methods, such as EPA method 521 for the determination of nitrosamines in drinking water, view %RSDs up to 20% as sufficiently good.

Compound	%RSD
	480 s	960 s
NDMA	7.8 %	2.4 %
NMEA	5.2 %	6.7 %
NDEA	9.3 %	2.1 %
NPYR	11.4 %	8.8 %
NDPA	5.2 %	0.6 %
NPIP	5.5 %	3.6 %
NDBA	1.3 %	0.1 %

Of course, this does not apply if the chromatography following the extraction takes at least 16 min anyway. But then again, do you think it is beneficial to the SPME fiber/Arrow lifetime to have the device unnecessarily shaking around and bending in the Agitator or longer, just to squeeze out an insignificant amount of sensitivity? Ultimately, only you can determine the best cost-to-benefit analysis for your laboratory.

Conclusion:

When building a SPME method, extraction times not only need to be evaluated by overall analyte response, but also reproducibility. In the majority of cases, our experience indicates that SPME extraction times should be shortened, which goes against much of the published literature. Time is money and as a German, I have a strong appreciation for efficiency, so why waste more time if your detection limits can be met with shorter, reproducible extractions!?

References

1. C. Myers, J. S. Herrington, P. Hamrah and K. Anderson, "Accelerated Solvent Extraction of Terpenes in Cannabis Coupled With Various Injection Techniques for GC-MS Analysis," Frontiers in Chemistry, vol. 9, pp. 2296-2646, 2021.
2. D. Patel, T. Roychowdhury, D. Shah, C. Jacobsen, J. S. Herrington, J. Hoisington, C. Myers, B. G. Salazar, A. V. Walker, D. S. Bell and M. R. Linford, "6-Phenylhexyl silane derivatized, sputtered silicon solid phase microextraction fiber for the Parts-Per-Trillion detection of polyaromatic hydrocarbons in water and baby formula," Journal of Separation Science, vol. 44, p. 2824 – 2836, 2021.