Combipal reproducibility

[Charles A. Burger]

I'm running an EP 2.4.24 residual solvents system. I'm trying to quantitate and to do so one must run 3 sets of test solution and spiked test solution in order to meet a NMT 15% RSD.

My solvents of interest; Ethanol, Acetone, Chloroform and Toluene. The %RSD's for the first sample set of 3 are 28%, 8%, 10% and 13% respectively. To be honest with you, my biggest worry was the Toluene which makes the Ethanol RSD that much more alarming. Aside from the failing 28% Ethanol, are these typical RSD's or can/should they be better?

Headspace conditions
vial oven = 80°C
Syringe = 85°C
Injection size = 1000µL
2 fill strokes
Injection speed = 290µl / min

GC conditions
I'm using a Restek Rxi-624sil MS 30m x 0.32mm 1.8µm film
Linear Velocity = 35cm/s
split ratio = 5:1
Carrier gas = Helium

This is my first attempt with %RSD on this headspace unit. I'm much more familiar with the fixed loop and pressure balanced systems. Any tips are welcome.

Thanks

[chromatographer1]

If I ever got more than 5% RSD it was because the solvents were difficult, not these easy ones like EtOH, CHCl3, Tol, and DMK,which I usually got around 2%.

Can you describe how you are sampling from the vials in more detail?

Why have two fill strokes? Once you take out 1mL of headspace you disrupt the equilibria of the system. Do you pump the sample back into the vial, or do you expel it into atmosphere?

I would expect RSD values of less than 5%.

Rod

[Peter Apps]

What GC are you using, what inlet, what inlet liner and what inlet temp ?

If your injection speed really is 290 ul/min you are taking over 3 min to inject the sample. I suspect that you mean 290 ul/s and depending how the inlet is set up that might be a problem.

What is the sample equilibrium time ?

The sample teperature control on the combipal and its clones is not the best, but your RSDs are about an order of magnitude bigger than they could be if everything is running right.

Peter

[chromatographer1]

I assumed that 290µL/min was a typo. Who would inject at 5µL / sec ?

right, Charles?

Rod

[Peter Apps]

Hi Rod

I am almost sure that you are right about the typo, but only almost. Too fast and injection speed can upset the flow control on the brand of GC that sales statistics says Charles is likely to be using, so I wanted to be sure sure.

Petr

[Charles A. Burger]

You are both correct on the typo it is sec not min.

I long for the days of a fixed loop or pressure balanced system. Those are the systems I'm most familiar with.

My system is a Shimadzu 2014, the carrier flows are controlled electroniclly.
I'm using a straight insert with a 1mm ID. The injection speed is a carry over from when I was running a 0.53mm ID column with a higher total flow. My goal was to be at or just under the total flow in µL/sec. Now having realized that fact I'm actually faster than my total flow which happens to be 11.9ml/min or 198µl/s

There is no venting of the first stroke, it goes back into the vial. I can't remember my initial thought on why I chose 2 strokes. I'd be happy to use just one if that was the only reason for my issue.

Draw speed was set at default which I do not know. After looking around for information, I now have it at 50µL/s

The Pull up Delay was also default, which as a guess was 1 sec. Now it's 5 s.

Pre injection delay - 200ms
post injection delay - 500ms
Incubation time - 60min ( could this be 30 for the solvents at hand?)
agitation speed 250 rpm
on - 2s
off - 5s

And a little extra info, the total volume of liquid, in this case water, required is 6 ml. Presently, I am using 10 ml headspace vials so I can have better sensitivity for my Chloroform. I have a suspicion something may be said about that because I myself am wondering if I have to much static pressure built up in the vial.

Let me know if I've forgotten everything else.

Thanks guys

[chromatographer1]

Charles,

Since you are using a split injection, you should not inject faster than your total flow. Perhaps you should use a smaller syringe volume, like 1/2 or 1/4 mL.

Now, consider: 60 min to equilibrate a vial of 6 mL ?

You should reach equilibration if a volume of 0.5mL is used in about 10 min. You should reach a reproducible situation (one in which can be validated in performance) in 5 min.

Question for thought:

What is the pressure inside the vial when you sample it?

What is the pressure inside the syringe after you fill it?

What happens to the gas in the syringe when you pull it out of the vial?

Does the pressure remain the same pressure inside the vial? Or does the gas 'leak' out of the syringe until ambient pressure is reached?

If the syringe is injected before the syringe contents reach ambient pressure and you inject the syringe contents into the injector, what kind of reproducibility should you expect?

Next question for thought:

Vial sealing and permeability:

Do you assume your septa never leak slowly over time, that a perfect seal is ALWAYS attained, every time.

I DON'T.

Do you assume your septa never absorbs the volatile solvents and that a secondary and changing equilibrium is NEVER attained for EACH of the solvents, with a different partition value for EACH of the solvents ?

Note: If you are not using septa with teflon lined silicone rubber base you are doing something wrong.

I DON'T.

Recommendations:

Use less than 0.5 mL total liquid in a 10mL vial. (20mg sample)
Equilibrate at 80°C for no more than 10 min.

Simple experiment: Inject 4 samples after 5, 10, 15, and 20 min in the equilibration oven.

I found RSDs for toluene at those conditons would vary 2%, 2%, 3%, 6%, and that is having 'perfect' septa and sealing. With common septa and sealing, the values were 3% 4% 8% and 12%. When heated for 60 min my RSDs were closer to 15-20%.

If you have more than 1000µg (5%) of each solvent of each solvent then you have too much sample to partition the solvents. REDUCE SAMPLE SIZE. The greatest cause for poor reproducibility in headspace is too large or too small a sample. Too small is easy to see (where is that peak?) Too big is not. (forest for the trees)

As long as you are trying to measure solvent amounts that the unit is capable of handling you should be ok. But if you are trying to reach ppm levels, I would use SPME headspace and accept the 5-10% RSDs. Of course you should be able to measure down in the ppb range with SPME headspace, but still with less than 15% RSDs.

Good luck, Chuck.

Rod

[Charles A. Burger]

Where to start? Here we go.

Changing syringes is not a short term option because I need to buy the whole assembly which includes the heating block to fit the syringe.

I agree that 60min is too long, but I'm stuck with that parameter also. The EP has 3 choices of temp setings for the system, unfortunately they are really not designed for the heated syringe platform. I could reduce my heating time to 45 min, but the EP then states that the "transfer line is to be at 105°C" which I would take as the syringe, which is the transfer device, would need to be at that temp. Now we have water vapor at a higher temp, more chance of a drastic pressure equilization.

I don't have a way to measure the pressure inside the vial as of yet. My guess it may be around 6 psi, again, big guess at the moment. I lost that data to my last job when I didn't take it with me.

I'm sure the pressure inside the syringe less than but moderately close to that of the vial.

The Vial is going to equalize to atmosphere.

I wouldn't hold out a lot of hope of great reproducibility.

Vial setting an permeability:

What is your prefered vial of use if it's 10 or 20 ml. I used crimp cap vials in the past on the other 2 types of systems and for most solvents attained 1-2% RSD's. Reading some literature on the Combipal injector, Screw cap vials are the only recommended vial. The assumption is that they seal better then the crimp cap? I can't answer that at this time. I can say that I am using teflon lined slicone rubber base septum on screw cap vials. The septa are thinner then any of the crimp caps I've used. About half as thick if I was to estimate.

I have 250µg of Ethanol and Acetone each, 3µg of Chloroform and 44.5µg of Toluene in each headspace vial.

Overall, I would love to try all of the suggestion you've made, unfortunately, at the clients wishes I'm stuck with polishing a turd. I may get it to shine, but it doesn't change what it is.

[chromatographer1]

I prefer a smaller vial to a larger, but I also prefer a larger ratio of headspace volume to liquid volume so equilibria can be reached more quickly and more reproducibly. I would certainly try the 20mL vials to see if you are saturating the headspace with solvent vapor. I would also suggest you try lowering the temperature of the vial oven to see if you can equilbrate more reproducibly.

Your guess of 6psi is pretty close to the truth. Hopefully, the needle never gets blocked and you have enough time to allow the syringe volume to reach ambient pressure before you inject. The pressure in the syringe should be the same as the vial allowing for the small drop due to loss of vapor during withdrawal.

Allowing the syringe to lose vapor solvent sample with the metal syringe 'cooler' than 80°C gives opportunity for variability of measurement.

All you can do is to show your bosses that the equipment you have may not be the best for the purpose at hand. Vary the variables and if nothing works, then that is what you must report. This is a routine analysis but sometimes with slightly different hardware the routine becomes impossible. Having extensive experience with both Tekmar/Varian/Agilent fixed loop and PE timed injection hardware, I don't like the syringe headspace equipment.

I hope someone in the Forum can be of more assistance. You have plenty of solvent to measure. Doing it reproducibly is the catch, and your polishing may not save the day, but may only make it smell all the more.

I do wish you success.

But it is good not to be you..... right now, anyway.

Good luck,

Rod

[Peter Apps]

Hi Charles

To save posting suggestions that you cannot implement because you are locked into a method that was developed on a completely different system, can you post a list of all the parameters that you ARE able to change, then we'll work on them.

Thanks

Peter