HS-GC RSD problem with residual solvents

[chromatographer1]

How proportional were your areas? Was 3000ng sample ten times as large as the 300ng sample?

You might compare samples heated for 60 minutes with those heated 10 minutes:

area versus area : RSD vs RSD.

Let us know the results please. Your customer might find it interesting as well.

FYI

My paper on HS solvents spiked IPA at 25 ng per vial with a 2.5% RSD. I did not use MEK but I used MIBK. 25 ng with 3.3% RSD.

I used 25 µL of sample solution. I heated and equilbrated the vials for 7 minutes.

But that was on a Varian GC that was over 5 years old in 1995. The Tekmar 7000 was two years old and had the NEW deactivated transfer lines, fused silica lined SS tubing and sample loop. My RSD values for pyridine and esters would have been slightly better had I used the new PE HS40 at 25 ng samples.

I could have used 1mL sample solutions. My RSD values would have been at least 6% if not 10% or more.

good luck in your research,

Rod

[Peter Apps]

Hi Tinsang

You must decrease the volume of headspace taken up and injected by the syringe to 0.1 ml from 1 ml - otherwise you will get a wide, flat-topped peak with the 3 ml/min injection speed.

You also need to look at your units - you talk about ppm, which is presumably a concentration, but then give it as equivalent to a mass in nanograms. If you want to make these comparable you have to have the concentration expressed as a mass per volume or a mass per mass.

Peter

[tinsang]

Hello Rod and Peter,
I’m back to report you the result of my work:
Standard solution (6 Inj.): 100 ppm ≡ 10000 ng of 1- propanol and 35 ppm ≡ 3500 ng of MEK
Limit of detection solution (6 Inj.): 0,5 ppm ≡ 50 ng of 1-propanol and 0,5 ppm ≡ 50 ng of MEK.

1-propanol (100 ppm):
Std. 1: 50 mg 1-propanol/50 ml H2O
Std. 2: 50 mg MEK/50 ml H2O
Solution of standard: 10 ml of Std. 1 and 3,5 ml of Std. 2/ 100 ml = 100 ppm of 1-propanol and 35 ppm of MEK
Solution of LOD: 1 ml of Std. 1 and 1 ml of Std. 2 / 2000 ml = 0, 5 ppm of 1-propanol and 0,5 ppm of MEK
The HS-Parameter as you have suggested:
Incubation time: 12 min.
Inj.speed: 3 ml/min.
Sample draw: 0,1 ml
Sample volume in vial: 100 µl
Sample size: 10 ml
1-Propanol (100 ppm)
Areas: 4077530/3631392/4004675/3932052/4754663/3965183
Mean: 41109456
RSD %: 7,786

MEK (35 ppm):
Areas: 1252910/1277745/1277128/1256814/1201962/1265504
Mean: 1255344
RSD %: 2,2

LOD of 1 –propanol and MEK: It’s not able to see! I assume that the volume to inject is too low (0,1 ml instead 1 ml)
And now I have decided to use another parameter as below:
Incubation time: 15 min.
Inj.speed: 30 ml/min. Because the first time I had very good results when I used this speed
Sample draw: 1 ml
Sample volume in vial: 100 µl
Sample size: 10 ml
1-Propanol (100 ppm)
Areas: 12716042/14008992/13457191/13882085/13129651/13667652
Mean: 12716042
RSD %: 3,61

MEK (35 ppm):
Areas: 8961327/9315132/8975384/9119346/8027827/8894020
Mean: 444681
RSD %: 5,00

LOD of 1 –propanol ( 0,5ppm): This time I’m able to see the peaks but the RSDs are too high.
Areas: 48962/52363/54794/49095/52059/65933
Mean: 53368
RSD %: 11,87
LOD of MEK ( 0,5ppm):
Areas: 80580/95340/103690/117563/100274/128617
Mean: 104344
RSD %: 16,19.
Do you have any idea how I can improve?
Regards.

[chromatographer1]

When you say sample size, I assume you mean vial volume is 10mL. Correct me if I am wrong.

You are injecting at 19mL/min carrier flow (316.7µL/sec)

and are injecting 1mL at 30mL/min or 500µL/sec, ( initial sample plug is 2 seconds wide) then

your sample injection is faster than your carrier flow. This can cause a pressure - flow disturbance.
This could be a cause for your poor RSD values at low concentrations. It is borderline.

Try injecting at 250µL/sec (15 mL/min) and have the peaks 4 seconds wide initially. This keeps the sample injection flow LOWER than the carrier flow (250 µL/sec versus 316.7 µL/sec)

I would consider increasing your solution volume in the vial to 200µL and heating time to 15 minutes.

This may improve your numbers. You are only injecting 100µL on column. 50ng is a reasonable LOD with that size split and vial volume. Hopefully the 200µL won't be out of linearity when the amount of solvent increases to 10µg and 3.5µg.

I noticed your IPA values INCREASE with time. This suggests that the sample plug may be backflashing in the injector (too much sample injected too quickly for the volume of the injection liner and carrier flow rate.

This allows the solvent to cling to active sites that may cause inconsistency.

Slow the injection speed and decrease the injection volume if the peaks are too wide.

best wishes,

Rod

[tinsang]

Hello Rod,
I’ve repeated my analysis according to your advice (sample volume : 200 µl and inj. speed:15 ml/min) with 2 different vials because I wasn’t familiar with this vial size (10 ml).
Solution of standard: 100 ppm of 1-propanol & 35 ppm of MEK:
Vials with short neck: 5 Inj.
Areas of 1 propanol: 14521940/14802619/14714464/14343803/14339351; Mean: 14544435; RSD%: 1, 45
Areas of MEK: 9552507/9332997/9623039/9582515/8617240; Mean: 9341660; RSD%: 4, 49
Vials with long neck:5 Inj.
Areas of 1 propanol: 14871818/14659797/14531129/14753137/16008906/; Mean: 14964957; RSD%: 3, 98
Areas of MEK: 10418253/9989669/10047582/10172710/10135031/; Mean: 10152649; RSD%: 1, 6
I can’t say which vial is better because 1 propanol with short neck is better than with long neck, as well as MEK with long neck is better than with short neck. Even though both analytes are in the same solution.
Could you explain to me this phenomenon?
Do you have some experience with HS-vials? Which vials (name and material number) do you prefer?
Solution of LOD: I have only used vials with short neck:
RSD% of 1-propanol (0, 5 ppm): 5, 9
RSD% of MEK (0, 5 ppm): 5, 5
RSD% of 1-propanol (1 ppm): 2, 5
RSD% of MEK (1 ppm): 1, 5
As you see, your suggested parameter works very well. Thank you very much for your advice
P.s: You asked me about sample sizes: yes, that means vial volume.
Can you explain to me how you calculated the peak wide initially?
Regards,

[chromatographer1]

I have only used the Perkin Elmer HS-40 and the Tekmar 7000 (with the fused silica flow path deactivation) which used 6mL 12mL or 20mL vials. These do not use syringe but fixed loop sampling.

I demonstrated with both instruments that if your sample is limited in size then smaller vials give lower LOD values, bigger peaks. See the June 1997 Analytical Chemistry journal article where I compared 3.35mL vials. PE went ahead and still offers 2mL vials following my recommendation.

Ketones are generally more inert to active sites than are alcohols. I have no other explanation for the peaks you describe. Remember your metal needle is an active site.

The 'width' of the plug is the linear width of the sample which depends upon its volume and the flow rate of the carrier gas (assuming no mixing and perfect injection). Convert your carrier flow to µL per sec.

If your injection volume were 600 µL and your carrier flow was 0.150 mL (150µL) per sec then your plug width would be 4 seconds wide: 600/150 = 4 sec.

(0.600cc and 0.150cc/sec, or 0.6mL and 0.9mL/min)

This assumption ignores the 'focusing' of the sample plug on the phase of the column.

best wishes,

retired old Rod

[chromatographer1]

tinsang

Did you follow my explanation?

I hope your HS issues are resolved now and your boss is 'off your back'.

Happy headspace !

Rod

[tinsang]

Hello Rod,
Thank you for your explanation. I think that I'm starting about HS to understand it step by step.
Last week I've sent my validation plan to the customer and I'll see whether they agree with it or not.
There's a possibility that I may consult you again.
Happy headspace to you, too, and have a good time