Improving sensitivity on untargeted SPME analysis (dairy)

[nmccready]

I'm running an untargeted analysis (m/z 29-450) on a single quad MSD looking to capture volatile compounds with aromatic properties.

I analyze all sorts of matrices looking at food ingredients, formulations, products.

I'm not seeing many compounds in some products, especially dairy. I suspect there are extraction problems due to matrix effects and fats trapping molecules from entering headspace.

I was hoping to improve sensitivity by switching to a splitless injection (typically have run split 5:1 or higher in the past for this analysis. However, switching to spitless seems to reduce peak area and deteriorate peak shape for some compounds.

Sample prep: ~1g of sample, 5mL saturated salt water in 20mL headspace vial

Method details:
DVB/CAR/PDMS SPME Arrow fiber 50C extraction at 40C, 40min extraction, 1min desorption.
1mL/min He flow, 250C inlet temp
liner: Agilent 5190-2293 Inlet liner, Ultra Inert, splitless, single taper, glass wool (78.5 MM ID 4MM 900 µL)
purge to split: 50mL/min @1min. Gas saver 20ml/min @2min
column: HP5ms 30m x 250um x 0.25um
oven temp ramp ~: 40C -> 250C over 1hr

[MichaelVW]

I'm running an untargeted analysis (m/z 29-450) on a single quad MSD looking to capture volatile compounds with aromatic properties.

What kind of detection limit are you hoping for? You're talking about light, volatile aromatics?

[nmccready]

I'm running an untargeted analysis (m/z 29-450) on a single quad MSD looking to capture volatile compounds with aromatic properties.

What kind of detection limit are you hoping for? You're talking about light, volatile aromatics?

I'm looking at unknowns and hoping to capture as wide a range of volatiles as possible, but yes mostly light aromatics as I imagine I won't get many semi-volatiles using SPME... It's more qualitative than quantitative analysis so I don't have specific LoD requirements (semi-quant using a single ISTD) and not running any calibration curves - just hoping to obtain relative concentrations. My hope would be to detect as many compounds at around their OAV as possible (but I know that can vary substantially from compound to compound and will be limited by the single quad capabilities).

[MichaelVW]

Another thing to try is adjusting your ramp rate. An hour long run seems like a lot. Maybe you need such a slow ramp because stuff goes through your column too quickly? I think a HP-5 column is better suited for semis. You might want to consider a column designed for volatiles.

I don't think you need to scan up to 450 m/z. Nothing heavier than ~200 AMU is going to make it to the GC.

I think it'd be worth it to drop fifty bucks on an analytical standard so you can use different conditions and compare the results for a variety of compounds. For the heck of it, I'd try splitting at 10, 50, and 100, and dial in to find the sweet spot. It's normal for VOC methods to have a high split. Less split means more sample hits the column, but also more water vapor. So it's quite possible that you will get better results with a higher split rather than lower.

[nmccready]

Another thing to try is adjusting your ramp rate. An hour long run seems like a lot. Maybe you need such a slow ramp because stuff goes through your column too quickly? I think a HP-5 column is better suited for semis. You might want to consider a column designed for volatiles.

The longer run time has been helpful for analysis of flavor extracts and some spices which have a lot of compounds, to reduce the number of mixture peaks. Is there any advantage to increasing ramp (aside from time saved?)

Do you have a column recommendation for more volatile compounds? I thought the HP/DB-5 columns were the most commonly used for this application.

I don't think you need to scan up to 450 m/z. Nothing heavier than ~200 AMU is going to make it to the GC.

Is there an advantage to scanning a narrower range of masses, such as increased resolution?

I think it'd be worth it to drop fifty bucks on an analytical standard so you can use different conditions and compare the results for a variety of compounds. For the heck of it, I'd try splitting at 10, 50, and 100, and dial in to find the sweet spot. It's normal for VOC methods to have a high split. Less split means more sample hits the column, but also more water vapor. So it's quite possible that you will get better results with a higher split rather than lower.

I might try that, thank you. This is what I was wondering about... I have typically used higher split to dilute for samples with compounds which oversaturated the column. The water vapor/increased matrix load could be the key. Do you think there could be any advantage to trying different types of liners as well?

[MichaelVW]

I suggest you take a look at Restek's chromatogram modeler. (Under "resources" on their website). It can give you a rough idea of what a chromatogram would look like given a set of GC conditions and the column type. You give it what compounds you want and it gives you suggested columns/conditions, so you'll have to enter a few compounds you expect to see to give it something to work with. You can adjust column length, film thickness, etc. Or just call one of the column vendors - they like selling columns so they'll be happy to suggest one. I think chromatography-specific companies tend to have better customer service than the big ones that do everything.

What I run is EPA method 8260, so you could look that up to see how similar or not it is to what you need. For that I use RTX-VMS or another company's equivalent (they're all pretty much the same as far as I can tell).

I don't think the liner is going to make a big difference. The analytes should already be in the gas phase, right? So the liner's not doing much except getting them the last few inches to the column. I just use a straight liner, no wool, 1mm or 2mm ID.

Scanning a tighter M/Z range won't make a huge difference - it might not even be noticeable - but I like the idea of my detector spending most of its time looking for masses I care about. A smaller mass range = more scans per peak. That can give you a slightly better peak shape. Each scan gives an x/y coordinate on the chromatogram, and the software connects the dots. More dots, better shape.

[James_Ball]

The long run is not bad if you do not need to run very many samples per day.

For volatiles and if you are looking at possibly have some high boiling components the Rtx-624SilMS column from Restek is good as it can go up to over 320C and still give good resolution on the light weight components. Most good volatiles columns will have a thicker film to help with the light weight resolutions.

You probably should do an initial hold for 2 minutes if you are desorbing for 1 minute, just to let most of the analytes form up on the head of the column and let the water begin to pass through. Only things like Methanol and permanent gasses will elute before water and the ones near the water peak can have distorted peak shapes but the rest should be ok. The reason you may see smaller responses with splitless is because there is less carrier flow around the fiber while it is desorbing which would require a longer desorb time, but that can also cause broader peaks for early eluting compounds. If the first compound of interest is Benzene, then you should be able to hold the oven temp low and desorb longer, the cold oven will allow the analytes to focus on the head of the column and the thicker film will act as another adsorbent capturing them before ramping the oven.

[Peter Apps]

SPME will never give really low LODs because there is so little ab/adsorbent on the fibre.

You might get better peak shapes with spiltess transfer if you use a liner with a taper at the bottom that seals to the polyimide on the column - Restek makes them I believe.

Peter

[James_Ball]

SPME will never give really low LODs because there is so little ab/adsorbent on the fibre.

You might get better peak shapes with spiltess transfer if you use a liner with a taper at the bottom that seals to the polyimide on the column - Restek makes them I believe.

Peter

Restek calls them Uniliners, they fit to the column just like a press tight connector.
https://www.restek.com/en/products/acce ... tr=Agilent

[Peter Apps]

Those are the ones I was thinking of.