July 31, 2015

Tutorial: Mass spectrometry in plant science, part 4 – measuring ethylene gas by GC-MS

In this post I would like to discuss a specific application of mass spectrometry to plant science, that of measuring the concentration of ethylene gas in head space samples. Ethylene gas is a plant hormone with numerous roles in plant development and defense. Climacteric fruits produce ethylene when they enter the ripening phase, and so for agronomists it can be very important to measure ethylene in order to monitor the progression of fruit development. Its chemical structure can be represented as CH2=CH2 and is the simplest alkene in nature. Because it is so volatile and present at such low concentrations, it is difficult to measure. A number of dedicated infrared and laser detectors can be found on the market. They may or may not offer the kind of stability and reproducibility that is required for precision analysis. For labs that do not have the funds to buy a dedicated, high end detector,  the default analytical technique to measure ethylene is a standard GC with a packed column connected to a FID (flame ionization detector). This technique has the virtue of being highly reproducible and sensitive enough for most applications. It takes advantage of the fact that ethylene produces an electrical signal when it burns at the detector, unlike the oxygen and nitrogen in the air which it is dissolved in. This is a tremendous advantage over mass spectrometry in this case because due to an unfortunate cosmic coincidence, both ethylene and nitrogen gas have a molecular mass of 28, meaning they are indistinguishable at unit resolution. This can make it nearly impossible to use mass spectrometry to pick out low levels of ethylene in a head space sample where the nitrogen is a million times as concentrated. And in practice it isn't so easy to separate the two on normal stationary column phases. An aluminum oxide column is probably the best phase currently available. However, nowadays, a GC system hooked up to a mass spectrometer is probably far more common that one connected to a FID, so many labs that might like to analyze ethylene production from plants have an analytical tool that is perhaps too sophisticated to get the job done, at least without a few tricks.  Here I offer a modest technical innovation that makes GC-MS nonetheless a viable technique that is comparable to a GC-FID for this task and represents another way that mass spectrometry can enrich our investigations in the plant sciences.

 First of all, here is the spectrum of nitrogen gas:

Nothing too interesting to see here. The base peak comes at m/z 28 and the only real fragmentation that can occur is the loss of one of the N atoms, leaving 14. A small isotope peak at m/z 29 is visible due to 15N whose abundance is only 0.37% of that of 14N.

Now let's look at the spectrum of ethylene produced by electron impact.

It should be immediately obvious why GC-FID has traditionally been the preferred method for analyzing ethylene in head space samples. All the nitrogen is invisible to a FID, whereas it totally saturates the signal of ethylene at m/z 28, and using SIM won't help us here for the same reason.

One solution in this case is to monitor ions only at m/z 26 and 27, which are still very abundant in the ethylene spectrum and totally absent from the nitrogen spectrum. When we look at a compound as concentrated at nitrogen in air, it will still produce a small signal when monitoring 26 and 27, probably because its concentration is so high that a tiny portion of its pool at the extremes of its energy distribution overlaps with those energy windows. But the situation is much improved by eliminating m/z 28. Consider the following two total ion chromatograms.

Here we are monitoring m/z 26-28 of an ethylene standard (20 ppm) in nitrogen. The nitrogen signal at m/z 28 totally overwhelms the signal and the ethylene cannot be detected.

Here however, we can clearly see a minor peak for ethylene when we limit the SIM to m/z 26 and 27. Another improvement implemented included here is to run the purge flow to the split vent from the moment of sample injection rather than waiting 20-60 s as is normally done. This favors the entrance of sample gas into the column. This has nothing to do with mass spectrometer in this case but can be helpful for the analysis of samples that are already in the gas phase, versus a more typical scenario where a liquid sample is vaporized in the injection port liner. In the latter case, we wouldn't want the purge flow to run to the split vent immediately. Furthermore, because all that nitrogen gas in a 1 mL headspace sample is hard on the filament, it is best to use a solvent delay to keep the filament turned off until the nitrogen has passed through the detector, in this case about 2 min. One additional change included below is the use of pulsed splitless injection.

This injection technique requires a slightly more elaborate piece of hardware such as a multi-mode injection module but can be very helpful for gas injection on a GC system. It effectively provides a short pulse of additional carrier gas pressure to the sample during injection which prevents sample loss out the split vent. For comparison, this analysis was done by scanning m/z 26-28 in SIM mode, and we see the background is high even after the nitrogen gas has already passed. That's because there is always some nitrogen in the system, and this produces an unacceptable level of background.

When we now limit monitoring to the m/z 26 & 27 non-base peak ions, we can appreciate the combined effect of injection conditions that favor gas samples and selective detection of ions to favor our signal to noise ration for ethylene. This analysis achieves a good separation of ethylene from nitrogen by running isocratically for 1.5 min at 30 C before ramping to 100 C on this Al2O3 capillary column. The choice of SIM ions minimizes nitrogen background interference. The end result:

The ethylene peak eluting at 2.3 minutes has a great signal to noise ratio, the entire analysis only takes 5 minutes, and while this is a high purity standard, it is at a physiologically relevant concentration typical of climacteric fruits such as melon and tomato. Subsequent air sampling of ripening melon fruit showed that it gives an identical result. This method takes advantage of the specificity and selectivity of a mass spectrometer. At the same time, we can easily quantify the ethylene produced by plant tissue under these conditions, since the secondary mass peaks at m/z 26 and 27 still provide plenty of sensitivity.

This entry intends to provide another example of how mass spectrometry can be applied to the study of plant biological processes using commonly available laboratory instruments, in this case a capillary gas chromatograph and quadrupole mass spectrometer. Even a highly volatile, low abundance gas hormone like ethylene can be reliably and rapidly quantified with GC-MS by combining the right conditions at the injection port and in the analyzer.

Instrumental conditions:
Injection port T: 100 C
Liner: Single taper empty
Injection vol: 250-1000 uL
Standard: 20 ppm ethylene in N2
Oven program: 30 C for 1.5 min, then 20 C/min to 100
Column: 0.25 mm x 30 m Al2O3/KCl
Detector: Agilent MSD quadrupole operating in positive mode, SIM m/z 26 and 27 100 msec dwell time

No comments:

Post a Comment