Failing RSD with water stds

[Sandro]

Hello Everyone
The lab I am at is having problems running water calibration stds on a Agilent 7890 with autosampler. We are using a Double Taper Liner and are injecting only 0.2uL. Here is our method setup as of right now:

Oven temp: 45C
Inlet Temp: 210C
Inlet pressure: 2.5psi
Detector (FID): 250C
Liner: Double Taper
Column: 15m x 0.53 Wax
Injection amount: 0.2uL
Split ratio: 1:1
Autosampler injection rate: 30uL/min
Viscosity delay: 2 seconds

We are analyzing Ethylene Oxide, Ethylene Chlorohydrin, and Ethlyene Glycol in water. It has to be analyzed in water, we are not able to change that. Our RSD has been running between 7-15%, but needs to be at 5%. If anyone has any ideas on how to help this, please let me know.

Thank you
Sandro

[Peter Apps]

As the water hits the hot inlet it flash vaporizes and you get a pressure pulse that shoots vapour in all directions. With the short wide bore column you are using the problem is even worse. Can you try a narrower bore column ? Alternatively, Restek have just come out with an on-column liner for Agilent injectors that lets you inject on-column with an ordinary autosampler - definitely worth a try for this problem.

Peter

[krickos]

Hi

Unsure but perhaps this could help.

If your GC has EPCs (Eletronic Pressure Controls) you can try to switch from "constant pressure" to "constant flow". By doing so the EPCs will "disregard" from the pressure/vapour pulse created on injection and continue to feed carrier gas at the set flow.
At "constant pressure" the EPCs will shut off the carrier gas feed breifly until the pressure sets at 2,5psi at injection.

Unsure on the wording but hope it made some sence.

[gcguy]

I agree with Peter, you are way over the vapour volume limit for this liner. You could try another liner with a bigger volume, this would help.
In addition, I am not sure about the 7890, but the 6890 is not very happy with a split of less that 10:1.
The syringe size you are using is also important, if you are not using an internal standard try using a smaller syringe if you wish to inject 0.2ul. There can be a 10-15% variation between injections in some cases and using as much of the syringe volume as you can will help.

GCguy

[krickos]

you are way over the vapour volume limit for this liner. You could try another liner with a bigger volume, this would help.
GCguy

Hmm I get a vapour volume of 375ul = 47% of a normal double tapered Agilent liner (800ul) at the above conditions.

[Peter Apps]

I don't think that it is vapour volume per se that is the problem - rather it is the sudden rise in pressure as the sample vaporizes. This momentarily increases flow into the column and out through the septum purge, while the back-pressure regulator makes a belated attempt to compensate by letting more gas out of the inlet through the split line. Because both the pressure sensor and the pressure regulator are on the ends of long pieces of narrow tubing they cannot keep up with the rapid changes in the inlet itself. Goodbye any hope of a repeatable split ratio.

Oh for the good old days of upstream pressure regulators and splits controlled by needle valves

Peter

[gcguy]

To err is human. My apologies about the vapour volume.



I still stand by my other comments though.

GCguy

[AICMM]

Does the 7890 offer pulsed pressure injection the way the old 5890's did? If so, this might help push the sample out of the port quicker. I agree that the Restek "packed" liner may also be a good alternative.

Best regards.

[Peter Apps]

The Restek liner I was thinking of does not have a packing - a 530 um column is a press fit into a tapered section that guides the needle into the top of the column. The catalogue number is 24976. I have not tried one - I do not have Agilent GCs.

Peter

[krickos]

The Restek liner I was thinking of does not have a packing - a 530 um column is a press fit into a tapered section that guides the needle into the top of the column. The catalogue number is 24976. I have not tried one - I do not have Agilent GCs.

Peter

Perhaps a side note but intresting, seems like Restek "nicked" an old Varian concept, have used a very similar liner on Varians old SPI injector, worked great.

[hotliner61]

Expansion volume and sudden uncompensated pressure burst is clearly the problem
Try use pulsed split by setting for example an initial pulse pressure at 10 psi for 0.1 minute and see if you get any reproducibility.

R.