SPME sensitivity

[k.schmidt]

hello,

I am investigating volatile organic compounds produced by cells in vitro with the use of SPME-GC-MS. I sample cell culture media (with cells while they are cultivated or cell-free culture media collected after cell cultivation).

And I cannot detect any VOCs from my samples...

Everything leads to a low sensitivity of my instrument. For example, I was following methodology in one of the publications. I used the same fiber, matrix, extraction and desorption conditions, culture conditions and GC column. The difference would lie in the cells I am using. However, my cells (A549s) were used in some of the other studies were VOCs were extracted by SPME. These people could detect the VOCs. I do not think it matters (especially that I tried different cell lines too).

Unfortunately, most of the papers I can refer to, do not give their LODs (lower limits of detection) for the VOCs detected in cell culture media. They just say which VOCs they identified. I used some of these compounds as standards. The paper that showed LODs gave these values as low as 0.004 µM (ppt range). For example LOD for acetophenone was 0.033 µM. I could detect it at the concentration of about 5 µM. So it is a huge difference in sensitivity. This team used different fiber (2 cm DVB/CAR/PDMS) and column but the same cells and I also tried to use their methodology with 1cm DVB/CAR/PDMS. I am aware that here 2 cm fiber may have an impact on sensitivity but what about the first publication when I used the same fiber???

The last thought I have and going to check it soon is my GC inlet liner. I'm using an SPME inlet liner 0.75 mm I.D. I found an information that sometimes liners with larger inner diameter like 2 mm are recommended for liquid SPME applications due to the fact that water may be trapped between the SPME fiber rod and the protective rod. It would result in a lost of the VOCs. This could be the case as I usually observe a drop of water at the tip of my guiding rod after withdrawal of the fiber from the vial with a sample. Has anyone experienced this problem?

There surely are aspects I did not consider. I would be very grateful for any comments/thoughts on this topic.

[Don_Hilton]

How is the sensitivity of the instrument you are working on? Are you using the same type of instrument and detector as in the publications? What sensitivity test have you used to demonstrate the GC system is working at the expected sensitivity? The tune report will show the mass spectrometer is able to obtain a signal for compounds once they are in the ion source - but you need to be sure the compunds can make it from the inlet, through the column, and into the right region in the ion source.

[rb6banjo]

I'm with Don. You need to verify that you can easily detect something as a model system (1-2 ppb toluene in water, something like that) so that you know for sure your chromatographic system is up-to-snuff. That should be fairly easy (but low) to get. By my calculations that's about 1/92 = 0.011 µmole/L.

If you can't see toluene in water at this concentration, figure out why.

[k.schmidt]

Sorry for the late reply.

In a meantime, it appeared that I was working on the old GC column... I was assured it was a new one so I excluded an impact of GC column on the sensitivity. I am going to change it soon. If it won't improve the sensitivity I will take SPME off line and check sensitivity of my instrument by running a sensitivity standard. The only universal sensitivity standard for EI-MS I found is OFN. What do you think about it?

I am using the same type (but not brand) of instrument as in one of the publications. The second research team used MS-tof (not quadrupole).

[Don_Hilton]

There should be a sensitivity specification for your instrument. And, unless you have some other benchmark to go by, your best bet is to compare your instrument to the manufacturer's specification. The standard may be OFN or HCB. Some older instruments used methyl sterate and probably other things as well. (Methyl sterate has cary over in the inlet - just in case you have one of those older instruments.)

Note the column and instrument condtions used for checking sensitivity. You need the peak to be the specified peak witdth. If you have a wider peak, it will be shorter in height and the signal to noise will be too low - just because the area is spread out along the baseline.

[k.schmidt]

Hi
As a result of changing the column, the sensitivity was improved by at least an order of magnitude. However, I am still having problems to detect the VOCs from my matrix.
The instrument have some 'hardware diagnostics warning' (seaprate topic viewtopic.php?f=2&t=23479).

So I want to buy OFN to run a performance check. But here I am not sure about my reference. I got manufacturer's performance specifications for my instrument which is:

Performance Specifications*
Mode Test Standard S/N†
EI Scan 1 pg Octafluoronaphthalene 20:1
EI SIM 50 fg Octafluoronaphthalene 10:1

*Note: These specifications apply with the CP-1177 Injector in splitless mode and the 1079 PTV Injector
in cold-on-column mode.
†Based on RMS noise.

I have the 1079 Injector but I do not know anything about on-column mode. What I found is that 'In the on-column modes the column is sealed to the glass insert (0.18 - 0.53 mm columns). This ensures that there is maximum transfer of sample to the column'. What does 'sealed' mean?

Can I just inject an OFN sample into the cold injector and then ramp the temperature of the injector?

Don, you also mention to note the column used. Manufacturer does not give this information.

Thank you very much for any comments.

Kamila

[Peter Apps]

Ideally you need a Uniliner from Restek - it has a tapered lower end that seals to the polyimide coating on the column and gives (very nearly) 100 % transfer of sample to column.

However, for troubleshooting you can do an ordinary splitless injection - from the figures that you quote your sensitivity is orders of magnitude too poor so the little bit of test substance that you lose when you open the split will not be an issue.

Which prompts a question - if you have a leak, or are using split when you should be using splitless during the SPME desorption you will get exactly the loss of sensitivity that you describe. Have you leak checked your inlet ?

Peter

[Don_Hilton]

On the column used for checking instrument performance - peak shape is important. A wide, flat peak will have poorer signal to noise (peak hight relative to noise amplitude) than a narrow, sharp peak. Alos a peak acquired on a relatively fast temperature program will tend to be sharper than a peak obtained under isothermal conditions. For OFN, my guess would be a DB-5 type column, narrow enough to give good pressure drop across the column (.25 or .18 id to maintain the inlet above atmospheric pressure with optimal linear velocity - an issue only with mass spectrometers as detectors) thin film - and not too long. Short column, narrow diameter, and thin film work in the instrument's favor. A look in the papers with the instrument (if they have not been lost or discarded) should have a copy of the chromatogram that was used to demonstrate the instrument met specifications at installation. If not - a call to the vendor for performance test conditions may help.

[k.schmidt]

The figures I quoted were obtained with the use of SIM mode (I didn't realize before...). And with the new GC column I obtained the same or even better detection limits in SIM mode (for example I got 0.013 µM in comparison to 0.033 µM for acetophenone from the paper). But I still cannot see any VOCs in full scan mode from the headspace of my samples where other researchers could. (Although I do not know their detection limit values, they simply didn't give them). But what is more with a new column I get lots of contaminants (probably siloxanes) in chromatograms (when I sample cell media or standards in water). Blank SPME injections are ok though.

I am using splitless injection mode with SPME. But Peter you are right. I performed auto tune and the ratio of m/z 28 : 18 is much too high (percentages of 100% for 28 and 6.1 % for 18, which makes the ratio of 16.4 : 1, and I found that the limit is 1.2 : 1). Therefore, there is a need to check the instrument for possible leaks. I do not know too much about it and how significant the leak might be?I just wonder why the SIM mode results with the new column are so good if there is a leak...

Anyway. We're changing the mode from SPME to normal GC injections. First, I need to find the leak and then will do OFN performance check. Uniliner from Restek is not an option (lack of funds...). I'll go for a splitless injection.

The papers were lost. I found somewhere else the same performance specifications for my instrument with a note that they were performed on 5% phenyl ms column (30 m, 0.25 mm id, 0,25 film). I have Rxi-5ms (Restek) with the same dimensions. I hope I can compare the two.

Kamila

[rb6banjo]

You might have a leak but don't discount the fact that your helium might be contaminated with nitrogen. I've had it happen a couple of times in the last six months.

When in scanning mode, what mass range are you covering?

[MSCHemist]

I perform SPME on biological materials (flavors) all the time. I use the 2cm carb/dvb/pdms and a restek 20973 liner with a 2:1 split. though you can do splitless as well.

I would think a good wax like Inno or Stabilwax would be better for your app than a db-5. Carboxylic acids especially give very poor peak shape on a db-5.

Also for cell culture media look at methyl or ethyl chloroformate derivitization they are very big in metabolomics right now. They allow you to see many biological materials that are not normally volatile such as amino acids, Kreb's cycle acids. DB-5 is actually very good for this app.