Offset expected for GC-MS with split detection to FID?

[IdiotFool]

Greetings, all.

I've got a Shimadzu QP2010-S GC-MS with FID. Recently, I've acquired a splitter so that samples can be analyzed through both detectors. Outside of software issues I'm having (cannot set up a batch to automatically quantitate through the FID - have to manually open the file and quantitate) I'm seeing a retention time offset of 6-7 seconds at the beginning of a run and 3 seconds towards the end.

If anybody else has a split detection system: how do these offsets compare to yours? Is this normal, poor, middling? Do you think there'd be any issues with proceeding with calibration or should I make adjustments?

Thanks!

[Peter Apps]

Does the discrepancy in retention times make it difficult to relate peaks on the FID chromatogram with peaks on the MS chromatogram ?, or are you worried that the FID-MS split ratio is changing during the run ? (which it looks as if it is).

I am intrigued as to why you would want to split between an FID and an MS, and what kind of splitter you have.

Peter

[IdiotFool]

Does the discrepancy in retention times make it difficult to relate peaks on the FID chromatogram with peaks on the MS chromatogram ?, or are you worried that the FID-MS split ratio is changing during the run ? (which it looks as if it is).

I am intrigued as to why you would want to split between an FID and an MS, and what kind of splitter you have.

Peter

Peaks are relatable since the order of retention does not seem to change and I wouldn't expect it to. I don't see anything indicative of gas leaks, so I'm unsure why the retention time differences change over time. I think it may have something to do with the fact that a good 15 cm of MS restrictor column are outside of the oven and are with the MS chamber, which should be about 250 °C, constant. As the temperature of the oven gets more comparable to the MS chamber, the retention times get more in line.

I want to split to the FID and MS so that I can do qualitative analysis via FID and quantitative via MS. I need to analyze volatile content of materials for VOC testing (I work at a paint manufacturing facility) and the results need to have higher levels of accuracy than what I've been able to obtain via MS (+/- 0.5% results). I didn't want to have to manually move the column from the MS to the FID and I wanted a permanent setup.

Shimadzu sold me what is basically a glass-lined T with restrictor columns of different bore and inert film thicknesses. Theoretically, these should have been developed so that retention times would be equal between the two, due to the split ratio, but there are obviously other factors at work here that are causing shifts.

[Consumer Products Guy]

MS detector operates at vacuum.

[AICMM]

Idiotfool,

I'll try this again....

To begin with, you did not say if your MS is early or late....

I would start off with the difference in viscosity of the carrier gas as the oven warms up. This increase will affect performance and this might account for the difference. If you are constant flow I would not expect to see as much difference as if you are constant pressure.

Your comment about unheated transfer line is also astute. My try some heating tape to see if this makes a difference.

Finally, you might consider un-coated transfer lines if you are truly using coated lines.

Best regards,

AICMM

[Peter Apps]

I'll support what AICMM said - try uncoated transfer lines. Any stationary phase in the transfer lines can only be a source of problems. You might still see some small effects of the impact of changing temperature on gas viscosity but with no stationary phase the peaks should be going through the lines in about 1 s, and differences and shifts in retention should them be less than 1 s.

I doubt that with your set-up the FID will give any more repeatability than the MS in terms of peak area per unit mass of analyte. The FID's linear range is wider, which might help, depending on where the problem lies. 0.5% repeatability rsds (which is what I interpret you to have with the MS) is not too bad, and probably mostly due to injector and inlet imprecision.

Peter

[IdiotFool]

The transfer lines are as follows:
FID Restrictor column is 150 / 363 and 1.0 M
MS Restrictor is 110 / 320 x 1 M

These should be free of stationary phase coatings, and just fused silica columns. Splitter set came from SGE: http://www.sge.com/products/gc--lc-supp ... splitters3

Sorry, yes, the MS spectra is coming out faster. Yes, the run is set to constant flow. I know that MS is done at vacuum, but the interface and ion source are at 260 °C and I would assume that the area of column at this section is significantly warmer than the column in the oven, which is at 60 ° with the six second offset and about 180 °C at the three second offset. If it's not the restrictor columns that are causing the issues, then I don't know what else it could be.

I was already upgrading the auto-sampler to a rail system with Headspace capabilities and I thought I'd throw the splitter in to see what sorts of improvements I could garner from using the FID. If it's not any better then it's not a huge issue as I really didn't spend much to get this capability.

Peter: do you have a split system and are you saying that retention times are within 1 second between detectors?

To anyone: I'd like to have the spectra superimposable but sometimes I go OCD on pointless things. In your opinion: as long as peak order remains constant and retention times are at least somewhat similar (within 10 s), would you find this acceptable or would you try to get the spectra aligned? Why?

Thanks, again.

[Peter Apps]

The dimensions (?) that you give for the transfer lines are meaningless because you do not say what is length and what is diameter, and there are no units - and what is the mysterious 1M ?

I agree that the transfer lines are causing the difference in retentions, but the scale of the differences, and the amount by which they change suggests either that the gas flows through the lines are very slow or that there is stationary phase in the transfer lines.

Are you by any chance using a narrow bore (say 0.1 mm diameter) column ?.

I have a post-column split between an FID and a PFPD, retention times are identical but both arms are at the same temperatures all the time since they are both in the GC oven and the two detectors are set to the same temperature. Quite soon I need to set up a PFPD - MS post-column split - and I need superimposable chromatograms so retention discrepancies such as you report are a concern to me.

Peter

[GasMan]

If you are seeing a difference in retention times with and without the splitter, this is to be expected. See the topic below

viewtopic.php?f=2&t=16269&start=0

Gasman

[chemstation]

Hi,
For that type of splitter, not only do you get different RT, you also get different split ratios in a GC running a oven gradient programme. (less split into FID at higher temperatures)
as the ~.12 cm of .1mm column into the 280°C MS transfer line, is not long enough to behave as a constant flow restrictor at typical flow rates, and the measured pressure (even in constant flow mode) at the split is different at low to high oven temperature programmes,
I can comment that Agilent sells their MicroFludic Splitter (MFS) system, that needs the EPC module so as to supply a source of Helium at constant pressure,
http://www.chem.agilent.com/Library/Sup ... 0E6FBE.pdf
Also, the EPC ensures there is enough Helium for the varying volumes of makeup Helium depending on the oven temperature. eg if memory serves me correctly, @ 40°C the MFS makeup was ~2.8 mL/Min, @ 280°C it was ~0.8 mL/Min (the makeup dilution adversely effected the MDL on both detectors, so I canned the MFS, but the split ratio was constant).

You could also try a twin hole ferrule for the injection port, and run dual columns, the RTs won't be the same, but as you will be using the FID for quantitation, you will only be reporting on peak areas for a particular RT.

Have you trying running your MS in SIM mode, as this may improve your precision (%RSD) as it creates more points under a peak, which is what I assume you are calling accuracy when stating (+/- 0.5% results).

Alex

[IdiotFool]

Hi,
For that type of splitter, not only do you get different RT, you also get different split ratios in a GC running a oven gradient programme. (less split into FID at higher temperatures)
as the ~.12 cm of .1mm column into the 280°C MS transfer line, is not long enough to behave as a constant flow restrictor at typical flow rates, and the measured pressure (even in constant flow mode) at the split is different at low to high oven temperature programmes,
I can comment that Agilent sells their MicroFludic Splitter (MFS) system, that needs the EPC module so as to supply a source of Helium at constant pressure,
http://www.chem.agilent.com/Library/Sup ... 0E6FBE.pdf
Also, the EPC ensures there is enough Helium for the varying volumes of makeup Helium depending on the oven temperature. eg if memory serves me correctly, @ 40°C the MFS makeup was ~2.8 mL/Min, @ 280°C it was ~0.8 mL/Min (the makeup dilution adversely effected the MDL on both detectors, so I canned the MFS, but the split ratio was constant).

You could also try a twin hole ferrule for the injection port, and run dual columns, the RTs won't be the same, but as you will be using the FID for quantitation, you will only be reporting on peak areas for a particular RT.

Have you trying running your MS in SIM mode, as this may improve your precision (%RSD) as it creates more points under a peak, which is what I assume you are calling accuracy when stating (+/- 0.5% results).

Alex

Thanks for the response, Alex.

I saw, after my purchase, that Shimadzu also has a pneumatically controlled system. I'm not sure why I was sold this less expensive T-splitter that was made by a different company when the pneumatic system is guaranteed to be a larger margin and dollar profit for them. Oh well, done is done. I'll check to see the repeatability with the system when I get some support to help me calibrate the instrument via FID. If it sucks, step one will be to berate the salesperson for selling me an inferior system.

I set the MS up for FASST (scan and SIM, sequentially) monitoring, which brought me to the accuracy I'm at, now. Before, I'd say there was +/- 2% repeatability in running a sample (injecting even the same sample vial for two consecutive runs gave 2% repetability). I probably could've played with the scanning rates more to improve accuracy but, since we have an FID on the instrument, I wanted to use it.

I'd had experience with an agilent dual-detection system in the past and it seemed the best option to improve results further, but I didn't delve into it and was unaware on how it was set up. Unfortunately, Shimadzu does not seem particularly strong in this area and both software and hardware limitations have lead to frustration. Software issues are that I cannot do a snapshot analysis on the FID, I can't set a batch to automatically quantitate FID results and I'll never be able to print results onto one report.

[chemstation]

Hi,
If the splitter the salesman sold you does what you asked/hoped for, then you should be happy. A MFS will affect your detection limits, I don’t know if Shimadzu has the option for dual injection ports, so one injection port can go to the FID and the other to the MS ($$$ though), assuming the two hole ferrules splitting doesn’t work out. Any thing you add to a system introduces complications that can be difficult to trouble shoot, particularly for air leaks into the MS and active sites at connections.

As you work for a private company, chances are you are responsible for the output of the GCMS, If I could make a suggestion, it would be to keep things simple, and thus robust, as you still need to be able to go on holidays to relax, and not have your name mentioned due to some other operators’ lack of ability, and overheard by management.

Your aim, IMHO, should be able to write an SOP, that allows other co-workers to use a robust system, without complication, whilst you are holidaying and they should be able to analyse the results.

As for your software issues, sorry can not help with that, if possible, see if other users are running different software that can acquire &/or data process Shimadzu Instrumentation ($$$ though) or Google for Shimadzu CDS. You also cannot do a snapshot analysis on the FID using MSD chemstation either

Alex

[IdiotFool]

Hi,
If the splitter the salesman sold you does what you asked/hoped for, then you should be happy. A MFS will affect your detection limits, I don’t know if Shimadzu has the option for dual injection ports, so one injection port can go to the FID and the other to the MS ($$$ though), assuming the two hole ferrules splitting doesn’t work out. Any thing you add to a system introduces complications that can be difficult to trouble shoot, particularly for air leaks into the MS and active sites at connections.

As you work for a private company, chances are you are responsible for the output of the GCMS, If I could make a suggestion, it would be to keep things simple, and thus robust, as you still need to be able to go on holidays to relax, and not have your name mentioned due to some other operators’ lack of ability, and overheard by management.

Your aim, IMHO, should be able to write an SOP, that allows other co-workers to use a robust system, without complication, whilst you are holidaying and they should be able to analyse the results.

As for your software issues, sorry can not help with that, if possible, see if other users are running different software that can acquire &/or data process Shimadzu Instrumentation ($$$ though) or Google for Shimadzu CDS. You also cannot do a snapshot analysis on the FID using MSD chemstation either

Alex

The system is guaranteed to be complicated since our other injection port is for pyrolysis (which was also never utilized and needs to be set up). The nice thing about being the only Analytical Chemist here is that the high-end testing will just wait for me to get back or priority samples would be sent out to an analytical lab for testing; nobody else will ever use anything in my lab except for the oven and microscopes.

I'm told Shimadzu is working on updating their software, so I'm hopeful that some of the limitations can be overcome. Interesting to know that MSD chemstation also does not allow a snapshot analysis.

Thanks, again.

[mkigls]

Tuning a MS/FID split is tricky - while in theory you can calculate it we've never been able to, and always had to rely on manually adjusting the length of the restrictors. It is a much greater split than you would assume - instead of 50:50 (by length) it is closer to 90:10 due to the vacuum in the MS.

I've never gotten it to completely correlate without using mathematical offsets - you will see some changes during programmed runs, similar to what you are seeing, even with EPC. Your 3-7 second offsets aren't bad - I'd be tempted to call it good and calibrate. Besides, you're really only using the MS data for qualitative purposes, so as long as you have the correlation you're golden.

As far as using SIM for quantification, be aware that you're still in the 5-10% precision range, even with SIM. If you need 0.5% (I assume absolute, since your syringe is no better than 1% relative) then you'll probably need to stick with the FID for quantification - that is precisely why we set up our system like you have.