Welcome to the New Restek.com! Take a tour of the new site and feel free to send us your feedback.

Your web browser will no longer be supported by Restek.com as of 30 June 2021.
To avoid any interruption in access or functionality, install a current-generation web browser now. Learn more.

SPME Fundamentals: Don’t forget the salt for HS VOCs!

  • Jason S. Herrington
  • #VOC
  • #Method Optimization
  • #SPME
  • #Blogs
  • #GC
  • #Injection Techniques
  • #Solid-Phase Microextraction (SPME)
  • #Sample Preparation Products
  • #Sample Preparation & Air Sampling
  • Share:

Long story short: We were comparing head space (HS)-SPME data with some colleagues, when they asked us “how are we using a 4 minute extraction time on our brewed coffee, when they need 10 minutes to achieve comparable results?” I told you coffee was on the horizon in my last blog. After comparing the 12 or so HS-SPME extraction and desorption parameters from our method and our colleagues' method, we could not find anything very divergent. Of course, you know where this story goes based on the title of the blog… Our colleagues were not adding salt to their HS samples. In fact, they were surprised to hear that we were doing this.

Why would we add salt to our HS-SPME samples? It does not matter if we are using HS-Syringe or HS-SPME; the addition of salt to the sample matrix (coffee in this example) will often lower the partitioning coefficient (K) for some target analytes, in particular polars. So, for all of the HS-SPME samples we run, we add sodium chloride. How much salt? Several articles have indicated ~20 – 30% wt/wt salt is optimum [1], but of course you should determine the optimum amount for your particular application.

In an attempt to show our colleagues the power of salt, we analyzed the headspace of brewed coffee samples with and without NaCl, all other variables being equal. We added 30% wt/wt NaCl to 10 mL to achieve saturation, which would help ensure consistency across samples. In addition, all samples were incubated (2 min) and extracted (various times) with a shaker speed of 250 rpm (with and without salt); but we threw in a wild card of 1000 rpm. The results of all this may be found in the following figure:

As you saw in the previous teaser, we observed numerous VOCs in the HS of brewed coffee; however, more on this in a future blog. For today, we are looking 2-furanmethanol, which has been shown to be an excellent differentiator between coffee bean roasts [2]. Here are the take-away messages:

  1. As you may see by the red trace, using a shaker speed of 250 rpm on samples with no salt correlates to 2-Furanmethanol (and other VOCs not shown) not reaching equilibrium until perhaps 960 seconds. I say perhaps, because it is hard to say when equilibrium was reached, as we did not extract longer.
  2. You will also see by the green trace, using a shaker speed of 1000 rpm on samples with no salt indicates equilibrium was reached at 480 seconds.
  3. Finally, the purple trace, using a shaker speed of 250 rpm on samples with salt clearly shows equilibrium was achieved at 240 seconds and the overall response was higher than the other two scenarios

Hopefully it is clear why our colleagues were extracting for 10 minutes and still not able to achieve what we observed in 4 min extractions (it was clear to them). It was all about the salt! Now, you could say something like “weighing out salt is time consuming, messy, and the juice is not worth the squeeze.”  To which I would say “stay tuned for our Life Hack on Weighing Out Salt for HS Samples in an up-coming blog.”  Till next time…


  1. S. W. Myung, H. K. Min, S. Kim, M. Kim, J. B. Cho and T. J. Kim, "Determination of amphetamine, methamphetamine and dimethamphetamine in hman urine by solid-phase microextraction (SPME)-gas chromatography/mass spectrometry," J Chromatogr B Biomed Sci Appl, vol. 716, no. 1-2, pp. 359-65, 1998.
  2. C. I. I. Rodrigues, C. M. Hanson and J. M. F. Nogueira, "Coffees and Industrial Blends Aroma Profile Discrimination According to the Chromatic Value," Coffee Science, pp. 167-176, 2012.


Tue, Dec 03, 2019

Hi Jason, For the green line in the graph, is it a general observation that a high speed shake with no salt(e.g. the green line here) results in progressively decreasing sensitivity after reaching the max(and never reaching a plateau equilibrium)??