Carryover Issue... help!

[HippyLabRat]

Disclaimer: By and large my expertise is in LC, so my GC skills definitely could use some help.

I am performing analyses on residual acrylate monomer in a polyacrylate material.
I'm trying to find an alternative to using Chloroform to dissolve out the monomer. Dichloromethane seems to work really well, but it also seems to leave more noticeable carryover versus the chloroform. (I didn't develop this method initially, I only have to fix it.) Also, the chloroform tends to beat the hell out of the columns, and their use is severely limited once we've injected samples into it. If there's a more resilient column to chlorinated solvents, I'm all ears.
The DCM offers significantly better resolution and there's less disparity between the size of the DCM peak and the analyte peak. Since this is a trace analysis, I think that's probably a good thing, since there has been an issue with the analyte and solvent's resolution.

Column:
Agilent HP5 30m x 0.320mm, 1.0 micron

Conditions:
FID Temp: 220C
Inlet Temp: 210C
Initial Oven temp: 80C
Final Oven Temp: 180C
Ramp: 5C/min

N2 Carrier Flow: 45mL/min
FID Auxiliary (compressed air): 450mL/min
FID Fuel (H2): 40mL/min
Run time: 20 min
Injection volume: 1uL
Split Ratio: 1:1

Thanks for your help. Tried to make sure you all had as much information as possible.
1) What's the best course of action to eliminate carryover?
2) Is there a better column for my analysis that can handle chlorinated solvents better than the one we're currently using?

[LabProARW]

I always thought that a split ratio of 1:1 was not a good idea. The flow is minimal and GC inlets may not be reproducible at that split. I was told 10:1 is a good minimum ratio and that 10-ml/min is a good minimum for split flow. Not exceeding the liner volume is important but if you are using a 78.5 mm x 4 mm straight through liner 1-uL of methylene chloride would not back flash.

[Peter Apps]

A 1:1 split gives you the worst of all possible worlds. The total flow is too low for the EPC to accurately control, and the inlet flushes very slowly, giving you a large tail on the solvent peak that is probably the source of your solvent-analyte resolution problems.

As LabProARW says, go for a 10:1 split if your analyte peak is big enough. If detectability is an issue do a proper splitless injection, with a splitless time of 30s and a purge flow of 20 - 30 ml/min.

Carryover can be all kinds of things, and the cure depends on the cause. It would help to see a chromatogram, instructions how to post are in a sticky on the LC page.

Peter

[HippyLabRat]

Solvent Peak Area Count (CHCl3): 1.638x10^6
CHCl3 RT = 2.22 min
Analyte Area Count: 268 (Yes, that's two hundred)
Analyte RT = 3.28

Solvent Peak Area Count (DCM): 2.707x10^5
DCM RT = 2.150 min
Analyte Area Count = 338
Analyte RT = 3.23 min

Like I said, I didn't write the method... I just have to fix it and validate.
We don't have any of the instruments networked, so getting a shot of actual chromatography might not be feasible, but I'll see what I can do.

[rb6banjo]

I agree with the others. For a liquid injection, you certainly want to have a much higher split - especially if you're using a narrow-bore column like a 0.320 mm.

Is the "polyacrylate" a solid, an aqueous suspension, etc. In what form is the sample? If you have the ability to "fix" the method, you might want to shift gears completely. Solvent-extraction methods are generally not very tidy. It honestly doesn't get much easier than headspace analysis. I use solid-phase microextraction pretty much exclusively in my work to collect the volatile materials in the headspace above my samples.

Polybutylacrylate for instance has a very low glass transition temperature (-55 °C, pp 124 of Kolb and Ettre, Static Headspace Gas Chromatography - Theory and Practice) which indicates that it will behave as a partition system at RT and higher. No solvents involved. I have had good success determining things like methyl acrylate, butyl acrylate, and methyl tiglate in polyethylene/methyl acrylate polymer and also vinyl acetate in polyethylene/vinylacetate copolymer using static headspace with SPME sampling.

[MSCHemist]

http://www.ncbi.nlm.nih.gov/pubmed/22500443

Looks like headspace would be a better choice.

[HippyLabRat]

Its polymethylmethacrylate dissolved in either Dichloromethane or Chloroform. I thought that was more apparent from the Chromatography data. I'm trying to find methyl methacrylate residual in the polymer matrix. The DCM dissolves the polymer the best of all the solvents I tried, has a much better shaped peak compared to the CHCl3 peak, which lends to much better resolution with the MMA.

[chemstation]

Hello,
IMHO
Q1) What's the best course of action to eliminate carryover?
A) At the end of the run, whilst keeping the original parameters, add a rapid temperature ramp to a higher oven temperature.
eg 30°C/min until 250°C, and hold for 2minutes @250; this will add ~5minutes to your~28 minute run.
alternatively you could setup a back flush method and save a couple of minutes.

Thus you haven't changed the initial parameters, and can still do a direct comparison, and 250°C (or lower) is still well within column specifications.

kind regards
Alex

[Peter Apps]

Hello,
IMHO
Q1) What's the best course of action to eliminate carryover?
A) At the end of the run, whilst keeping the original parameters, add a rapid temperature ramp to a higher oven temperature.
eg 30°C/min until 250°C, and hold for 2minutes @250; this will add ~5minutes to your~28 minute run.
alternatively you could setup a back flush method and save a couple of minutes.

Thus you haven't changed the initial parameters, and can still do a direct comparison, and 250°C (or lower) is still well within column specifications.

kind regards
Alex

Hi Alex

This is true if the carryover is heavy muck that moves slowly down the column, but given the conditions of the method I suspect that Hippylabrat's problem is crud in the inlet, in which case heating the column is not going to help. He is dissolving a polymer in a solvent (not actually doing an extraction) so every injection deposits polymethacrylate in the inlet. He has a very low purge rate, so the inlet just stays dirty from run to run. The reason that carryover is worse with DCM than with chloroform is that DCM is a better solvent - so puts more crud in the inlet.

Headspace will solve the carryover as well as the resolution problem.

Peter

[GOM]

Another approach that I have used is to dissolve the sample in the minimum amount of DCM then add methanol to precipitate out the polymer, leaving the monomer in solution.

This reduces the crud.

A splitless injection would be recommended

Ralph

[HippyLabRat]

So what I've gathered is... I'm doing it all wrong. I'm not terribly surprised. I'm not sure that basically redeveloping the entire method is within the scope of my job here, but thank you a ton for the help.

I adjusted the conditions of the method to try to slow down the elution of the methyl methacrylate, and I got some interesting stuff showing up. I slowed the thermal gradient down for the first half of the analysis, and sped it up for the last half. I got mystery peaks! Makes me question the validity of the original method in the first place. I think I'm going to try to take a look at the headspace method, and recommend that to the lab manager.

One way or another, I got what I needed.

[MSCHemist]

The advantage of headspace is that it is very very clean so you will have very long column lifetimes and liner lifetimes so you will save a lot of money there. The disadvantage is matrix effects so you often need to calibrate in a blank matrix or do standard addition. Also some types of headspacers have carryover issues there with polar compounds sticking to tubing and syringes to the point where it is necessary to run blanks between samples.

[tlahren]

I agree with most of the above posts. I believe that your column is going bad due to injecting polymer into it (DCM and Chloroform have never damaged a column of mine). Peak tailing/carry-over is most likely due to residual polymer in your inlet causing unwanted interactions with your analyte.

Headspace, headspace SPME or the precipitation methods all sound like way better ideas.

Coming from a Contract Analysis Lab and going to Research has taught me that many labs that run "Standard Methods" don't care to question the validity of their method so long as what every governing body approves their SOP and they can pass performance checks. You are right to question the method if it is not working consistently.

I would say that headspace is a good place to start as it will be a bit more affordable than SPME setup. The article that MSChemist linked to should be a really good starting point for a good method for you however it looks like it's in Chinese language and you may have trouble accessing it.

This article may be easier reading and they seem to have a similar setup that you do. Their analytical goal is a bit different but the method should work for you. You just need a Headspace Autosampler or dedicated headspace injection setup to do it: http://www.sciencedirect.com/science/ar ... 7312004761

If you have trouble accessing it feel free to email me at guitfiddle07 at gmail.com