Poor reproducibility of ethanol by GC-HS

[GT]

Dear all,
I'm trying to analize ethanol in Active Pharmaceutical Ingredient by GC-HS method, according to European Pharmacopoeia conditions.
GC-HS conditions are:
GC6890 + HS G1888 Agilent
Equi. temperature of vial 80 °C
Equi. time 20 min.
Pressurizzation time 30 sec.
With these parameters I obtain a poor RSD% of ethanol area both for standard solutions (about 0.2 mg/ml) and for samples, with value around 5 to 10 %
For others solvents like acetone I obtained good RSD%, around 1%.
How can improve reproducibility for ethanol?
Many thanks in advance.

[krickos]

Hi GT

Could you please post the full HS parmameters and column and splitt flow?

The Sample valve vent time should normally be around 0,15-0,4min, proberly not the issue as acetone works fine.

How are the temp settings in Valve and Transfer line?

One issue with Agilents valve/loop solution is that is it sensitive for contamination especially for polar solvents that can cause poor reproducibility and loss of sensitivity.
Valve/loop temp should be equal to or exceed the highest bp of the solvents including sample sample solvent (Water, DMF DMSO etc)

[GT]

Hi krickos,
here the others parameters:
split ratio 5:1
loop temp 85 °C
transfer line 90 °C
inj time 0.20 min
loop equil time 0.05 min
loop fill time 0.20 min.
carrier N2

thanks

[Peter Apps]

You do not tell us what column you are using, but guessing that it is not a megabore, and seeing that you are using N2 as carrier you have a total flow of about 6 ml/min (5 down the split and 1 down the column). The default loop on the G1888 is 1 ml, which takes a minimum of 10s to flush assuming no mixing or diffusion. You inject time is 12s, so you need to increase the split ratio, or the inject time, use a wider bore column or switch to Helium carrier (higher vol flow rate).

Peter

[krickos]

Agreeing with Peter here.

Also would like to add that when using chemstore on the 7694 headsampler there is a minimum flow required ~ 10ml in total (used helium) when HS is attached to a split/splitless injector or volatile inlet.

Have not used the new Agilent HS so much yet but in that sence they seem pretty equal.

[GT]

Dear All,
thanks a lot for your advices, and sorry again for incomplete information about method.
I'm using a 30 m, 0.53 column, DBwax, 1 um film, split flow 24 ml/min, total flow 31 ml/min.
Why with acetone and methanol I obtain a good RSD% and not with ethanol?

[Peter Apps]

OK, so it is not flow related, or at least not directly.

Try increasing the loop and transfer line temperatures by 20 C, with the loop at only 5C hotter than the vial there is a chance that there is a cool spot somewhere that is cuaing condensation.

Peter

[krickos]

Hi

Given the information I am putting my money on contamination of the valve/loop

One issue with Agilents valve/loop solution is that is it sensitive for contamination especially for polar solvents that can cause poor reproducibility and loss of sensitivity.
Valve/loop temp should be equal to or exceed the highest bp of the solvents including sample sample solvent (Water, DMF DMSO etc)

If you have been runiing samples on the instrument dissolved in DMAA, DMF, DMSO, or DMI with this method or previous other method with these valve/loop temperature setting, you WILL build up contamination in the valve/loop over time.
I have seen this so many times, had the valve/loop temp been like 50-55°C you would also seen it on methanol but now the valve temp exceeds the bp of methanol with a margin. I have seen the exact problem with ethanol on similar parameters on several instruments on our sites. And it almost always boils down to a contamination issue.


The down side with temperature setting in Ph Eur and USP is that they are most likely derived from methods with a heaspace using flow injection rather than a loop thus less sensative for contamination.
If running the Ph Eur or for that matter USP general methods on an Agilent sytem with loop/valve please follow my advice about temperature in the quote, and you will see less contamination issues and need less frequent steam cleaning or manual cleaning of loops.

[GT]

Hi again,
here below are reported the peak area of 6 injections (from 6 different vials) of same water standard solution of acetone and ethanol with new conditions Loop T 105 °C and T. line 110°C:
acetone
797, 797, 804, 787, 807, 826 RSD% = 1.7
ethanol
593, 596, 627, 597, 631, 681 RSD% = 5.4
What do you think?
Thanks

[krickos]

Hi

Intresting, and WATER, that was new information or did I miss ?

Anyways that makes it a bit more intresting, perhaps not contamination at all, I am not at the office currently so I can not check up a few things but maybe you should consider try a lower headspace oven temperature like 60°C?

As a rule especially when using water you try to have the oven temperature below the bp of the residual solvents, practically this is not always possible, but in water the polar solvents respons increase exponentially with the temperatuer in the oven. If you push the temperature too high strange things (instability) may happen to the gas/liquid equilibrum. Now you have a mixture of solvents but I recall some incidents with ethanol behaving strange.

Apart from that I am out of ideas until I get back to office.

Good luck!

EDIT: Stilll out of office, but what I am trying to say is that at 80°C you also get alot of water vapour in the gas phase that may cause you problems for certain solvents. I usually stay away from hight temperatures in water due to this. We have done some testing on various temperatures on solvents in water and generally for polar solvents we obvious get a higher sensitivity with higer temperatures but at 80°C we generally tend to get poorer RSD that at 60-65°C.
For common halocarbons like chloroform etcetera in water this is a well know fact, running at 95°C instead of 75°C signifcantly also reduces the sensitivity for the halocarbons(varies from compound to compound).

[Terry]

GT,

I got a similar problem a couple of months ago. See the link below:

http://www.sepsci.com/chromforum/viewto ... highlight=

The conlusion was that sample size and incubation teperature had a significant impact on RSD of MeOH and IPA in water.

[AICMM]

GT,

Are those the actual area counts or are they in thousands? Oxygenates respond poorly especially out of water but they are very low. What does your signal to noise look like?

I would suggest decreasing your split ratio slightly. Right now, if I read correctly, you are about 4:1 so I would try 3:1. Again, as pointed out earlier, the game is the balance between enough sweep flow and enough on column. Alternatively, if your separation is good, increase your column flow a bit for the same total flow....

Best regards, keep us posted...

[Peter Apps]

At 80C you have plenty of pressure in the vial, without having to add any more during the pressurization step. Depending on what pressure you are setting it is even possible that the vapour pressure of the water is higher than the pressurization gas and that you are getting backflow into the plumbing, with all sorts of "interesting" results. Try setting the pressurization time to zero (you do not actually get no pressurization on the G1888 becuase the solenoid valve does cycle for some reason), and make sure that the gas pressure is above the vial pressure.

Having siad this, it does not fully explain why you get good results for acetone and poor for ethanol, you still need to be looking for adsorption somewhere in the system, especially if the peaks are as small as the areas suggest.

Peter

[GT]

Dear All,
thanks a lot for all your advices.
To AICMM
All area are expressed "as is", not in thousand, with unit pA*s. LOQ method based on S/N ratio around 10 with EP calculation is about 0.8 mg/l ethanol. the actul concentration for reproducibility test is around 100 mg/ml.
To Krickos
I've tried with lower T oven (60 °C, same 20 min.) with following results:
Acetone (6 inj) RSD 1.7%
Ethanol (6 inj) RSD 4.7% (again!)

Next step Terry's way, try to decrease ethanol amount.

Thanks