GC-FID-MS: what's the catch?

[Rachel-G]

Hi all,

I'm a PhD student in the process of helping set up a new lab and we're debating what GC to purchase. The Agilent rep suggested a combined GC-FID-MS model: a GC-MS 5977B single quad + a 7820 GC with TCD and FID detectors.

Anyone have any experience with that kind of setup? Sounds like the best of both worlds, but I'm worried that there's some kind of catch-- the splitter is robust, reliable, etc? I haven't come across combined systems before-- in the past we used two separate machines, and the other systems in my university are also either MS or other detectors.

Thanks in advance,

Rachel

[salincoln]

Hi Rachel,

There's no real catch that I can see. I'm using a T connector post column that sends ~90% to the MSD and ~10% to the FID/ECD. You have to work out the balance using fused silica capillary tubing (in our case the line to the MS is much smaller ID and shorter than the line to the FID). If Agilent suggested a split FID-MS configuration to you they probably have a scheme for setting it up effectively.

Of course the T connections are subject to leaks, but those are easily detected and fixed.

Apart from intended benefits, it's really useful to have the simultaneous MS and FID traces when troubleshooting chromatography.

One potential, minor catch: There is a slight shift in % of given compounds going in each direction across a temperature ramp (running constant flow), and this varies with compound class. I have not worried about it enough to really figure out, because I generally have standards I can use for quantification, that are subject to the same discrimination.

Good luck,
Sara

[rb6banjo]

If you are buying a new system, I would recommend getting a system with their "dean's switch - heartcutting" option. I don't know anything about your samples or their complexity but if you are analyzing real-world samples, they are generally complex and separations are not easily carried out using any single stationary phase. With my system, I can get a separation (I use a nonpolar phase, ZB-5 from Phenomenex) roughly based on boiling point on the the first column then cut what I want (using the switch) to a more selective phase (Simplicity Wax, Supelco) on the second column. I monitor the separation on the first column with a flame and the second column with the mass spec. You can cut the entire first-column separation to the second column if you want.

Here's an example of what is possible. It's difficult to separate menthol from camphor (sample is an extract of a methanol extract of a petrolatum rub from my medicine cabinet at home, more discussion at: viewtopic.php?f=2&t=30531) and some of the other components in the matrix. The target molecule was menthol so I set up a cut to just transfer the menthol to the second column. They are easily resolved from each other on the second phase.

https://onedrive.live.com/redir?resid=4 ... file%2cpdf

If you're a novice chromatographer, it might take some effort to get it lined out but if you're a fairly well-versed user, it's not too bad to get it going. There haven't been too many times where I've found my target analyte to coelute with something on both phases. You can check the Agilent.com website for other app notes that employ the "Dean's Switch".

[GOM]

Hi Rachel

What analyses do you think that you will be trying to cover?

Regards

Ralph

[chemstation]

Hello,
The Microfluidic Splitter (MFS) is solid and reliable, we have two GCMSs with them, and use them to align detector responses to MS spectra.

The MFS is easier to setup than a T-Connector, also the split ratio is also constant during an ramped oven run.

It appears that the rep is offering a 5 port MFS, there is a caveat that the MS detection limit is degraded when using a MFS, so need to inform and discuss your detection limits if the lab is doing trace analysis.
Otherwise you can connect your primary column directly to the MS, or using
the blanking plugs, to blank of the detectors to force the all the column eluent
into the MS.

kind regards
Alex

[Rachel-G]

Hi all,

First of all thanks for all of your input!

My samples are large collections of complex biological samples, so far my metabolites of interest have mostly been short, long, and medium chain fatty acids, amino acids, sugars, and methane, oxygen, carbon dioxide, and hydrogen gases.

Some of our analysis will be targeted and quantitative (not as hardcore as an analytics lab but certainly relative quantities are important). I'm also interested in trying some untargeted comparisons of the overall metabolomic profile of samples, but that we still haven't tried. Trace analysis is less relevant I think.

I talked to the Agilent rep and actually the current price quote he sent me has two channels in parallel: one with a packed inlet and a molecular sieve column for looking at the fixed gases connected to TCD-FID sequentially, and next to it a split/splitless inlet with a capillary column connected to the MS. I don't see any disadvantage in this setup, it's basically what we had before just combined-- this is relatively standard right?

I asked him about the option of splitting also, they indeed sell the following: https://www.agilent.com/cs/library/broc ... 9667EN.pdf
This is the microfluidics splitter you referred to, Alex?

It sounds great on paper, and if you get the EPC model, you can change columns without venting, which I like. I'm tempted to try to convince my PI to go for the splitter option, it seems like a shame to have the TCD/FID and the MS sitting there next to each other and not be able to utilize them both simultaneously. However, I am a little cautious about the splitter, so if anyone else has worked with this and can give input I'd love to hear.

Re: the 2D heartsplitter option, it sounds great but so far we haven't had resolution problems for any metabolites we tried to detect (that I could see at least), and for the more untargeted analyses I'm not sure what I would do with the mountains of data it would generate. I'll look into it though, thanks.

Thanks again!

Best, Rachel

[GOM]

Hi Rachel



--
OK

You have had several good suggestions.

I originally started out with 17 GCs with a variety of general and selective and specific detectors and several staff.

Before benchtop MS I had to rely on experience and recognising chromatogram patterns and peak shapes.

Over the years with reduction in staff and budgets, I was able to come down to 5 instruments because of benchtop MS and the use of the Gerstel type multiple application autosamplers. These had to cope with a huge range of analyses from oxygen to triglycerides in really wide variety of sample types and analytes and levels.

I guess what I am trying to say is

a) Often it is the sampling and sample prep and sample introduction that is the difficult bit - once you have the sample on the column the separation is the easiest bit







b) I found that the more detectors and injectors that I had on an instrument, the more problems that I was likely to encounter.

c) Without knowing what range of samples that you will want to analyse, if you wish to get as much out of one instrument as possible, I would suggest keeping to just FID and MS, dropping the TCD and investing that money in a good autosampler that can give you a variety of injection techniques from liquid to headspace - e.g. Gerstel MPS autosampler.

d) will you have staff backup and when you have finished your PhD - will others be able to pick up your work easily for their work?

Regards

Ralph

edit - your reply came just as I was posting

I am slightly concerned that you are trying put too many detectors and column types into one instrument and thus making its operation overcomplicated.

[itspip]

Another thing to consider is that the columns are in the same oven. The disadvantage here is that you could only utilize one analysis at a time. You mentioned that this setup is similar to what is currently in your lab just put into one GC. Just be aware that if you're running a sample into the MS, you will have to wait to run the next sample into the FID/TCD.

[chemstation]

Hello,

yes the link was for the MFS I was writing about.

After you had clarified your system setup, it seems that your requirement for gas analyses with the proposed configuration by the Agilent rep is appropriate.

Depending on how much there is in the budget, maybe ask for the costs for a deans switch and a MFS, as they both run of the same EPC module. As you can always keep them in reserve until you need them, and then connect them up if necessary, as the saying goes better to have and not need, than need than not have ( and maybe save the day too! )

kind regards
Alex

[Rachel-G]

Hi all,

I'm indeed concerned that we're trying to combine too much into one system, but the Agilent rep actually seemed very positive about it, which was encouraging (especially because it goes against his natural incentive to tell us to buy two separate systems..).

Assuming we don't buy two separate systems, the other option would be to get the MFS and change the column/inlet when moving between systems. Our lab wouldn't need to change so frequently I think, there are periods when we are analyzing fixed gases and then periods when we look at other metabolites. Is changing the column and inlet easy?

I definitely want an MFS, I just need to convince my supervisor Although actually how likely is it that we'd see compounds in the FID that don't show up in the MS? And quantification on the FID is generally more reliable than on the MS? Those are the benefits I thought of, what else?

Thanks again for all the input,

Rachel

[rb6banjo]

I can tell you that I'm seeing things with the 5977 MS that I could never see with my 5973's. Let's say you see something in the flame that's not in the MS, what are the odds that you'll figure out what it is anyway? I would always get a MS over a flame - especially if identification of unknowns is your game.

The FID is superior to the MS for some selected compounds. Usually small molecules that have mass spectra that look like almost everything else. There was a post on the Restek chromatography blog not long ago that discussed "when not to use a MS detector":

http://blog.restek.com/?p=6662

In my experience, quantitative analysis is no worse with a MS than it is with a flame. Actually, it's generally better because you can sort the data by specific masses which makes your S/N ratio dramatically better. Detection limits can be quite low because you can often times make other "junk" vanish. If you go SIM, you can get detection limits that the FID can only dream about. Lots more flexibility with the MS.

[Rachel-G]

I can tell you that I'm seeing things with the 5977 MS that I could never see with my 5973's. Let's say you see something in the flame that's not in the MS, what are the odds that you'll figure out what it is anyway? I would always get a MS over a flame - especially if identification of unknowns is your game.

The FID is superior to the MS for some selected compounds. Usually small molecules that have mass spectra that look like almost everything else. There was a post on the Restek chromatography blog not long ago that discussed "when not to use a MS detector":

http://blog.restek.com/?p=6662

In my experience, quantitative analysis is no worse with a MS than it is with a flame. Actually, it's generally better because you can sort the data by specific masses which makes your S/N ratio dramatically better. Detection limits can be quite low because you can often times make other "junk" vanish. If you go SIM, you can get detection limits that the FID can only dream about. Lots more flexibility with the MS.

Thanks for the link, great blog! And good point about identifying something I only see in FID. I guess it's not essential to use them simultaneously, though I'll still suggest it. In any case I'll make sure we get a system where a splitter/heart-cutter can be added later.

[rb6banjo]

If you want to add it on, you can do it later. Having the right number of EPC modules is the key. Good luck!!

[James_Ball]

The one thing you have to keep in mind when doing something like this is if you have two columns installed at once, the maximum oven temperature is dictated by column with the lowest maximum temperature. If one can only go to 200C and the other can go to 350C, then the higher temperature column will be limited the the lower temperature so it may not be able to do all it could possibly do.

I have a setup with both injected semivolatile analysis and purge and trap volatiles analysis both on the same instrument using two injection ports and combining into a tee fitting to go to one MS. This way I can do either analysis without switching out columns, but thing that made it possible was finding the new volatiles column that maxed out at 315C so I only lost the last 10C of the usable limit of the semivolatiles column.

[BMU_VMW]

The one thing I miss here in the responses (or I just looked over it) is the choise of the injector. if possible I would go with one SSL and one large volume PTV.