calibration for dual column analyses

[Lila]

Hi there,

I have a question regarding the appropriate calibration technique for a P&T-GC set up. The GC is configured with a single injection port and split between two identical columns types to two different detectors. Carrier flow into the inlet port is from the concentrator only. (OI inlet port) at about 4.5 ml/min. In theory there should be a constant 50/50 split between the two columns for each analytical run, but this is not always the case. The calibration range is 2-50 ppb.

My question is, would you double working concentrations for the calibration curve (5-100 ppb) assuming that half the amount goes to each column or would you use the same calibration range 2-50 ppb regardless of column configuration? Is there a difference between each of these techniques? (In sample analyses using the first method you would have a multiplication factor of 2 for each sample & we account for this in reporting). I have staff members who believe their way is the correct way, so I am hoping that you can help resolve this conflict!

Thank you for your time.

[chromatographer1]

I don't know if there is a right way with this situation.

I do know what I would do if I wanted to explain this clearly to an auditor.

Document your samples as they actually are, whatever ppb you make.

Then state that you are splitting the sample from the P+T between two analytical columns which may vary from 50/50 ideal split.

I would demonstrate reproducibility in my standards ( I would run a minimum of 6, before and after my samples).

Whatever sample you put on the P+T should be described exactly.

Whatever you get from the concentrator and how you split or divide the output you should describe carefully and clearly.

best wishes,

Rod

[JI2002]

As long as you run samples exactly the same way as you run the calibration, in theory, you can do it either way and the results will be identical. But it's easier to use same concentration in your calibration table as in the standards you make(if the concentration of the standard is 2 ppb, then use 2 ppb in the calibration table), then you don't need to multiply 2 to the sample results in term of reporting.

Here is a question: What detectors are you using for the two columns?

[Lila]

the detectors are an FID detector and an XSD detector, manufactured by OI and it is supposed to be comparable to an ECD for sensitivity to halogenated compounds. The XSD detector has been time consuming and required more maintenence than the FID.

In preparing a 2ppb standard, under the first scenario, 1ppb would split on to each column and in the cal file should you label it as 1ppb, because the mass on column is half as much, but in the second situation, the prepared 2ppb standard is called 2ppb in the cal file even though only half of the standard makes it to the detector. I am not sure the best way to approach this, especially from a QA/QC perspective.

Thanks for your help.

[AICMM]

Lila,

If you load a 2 ppb in the concentrator, then it is 2 ppb in the calibration table. No one cares that you are only putting 1/2 the mass on one column and 1/2 the mass on the other. No math needed in the calculations. What it does do is double your detection limit since you need twice as many ppb at the concentrator to see the same mass on each column. One way around this, if you can, is to purge 10 mL of 1 ppb instead of 5 mL of 2 ppb.

Couple other comments. Are you using a micro-trap since you are only desorbing with 4.5 mL/min (seems kind of low?) What are the ids of the two columns? Do you have a way to vent the water before it hits the XSD and finally, what has been your experience with the linearity of the XSD.

Best regards.

[Lila]

The XSD was only calibrated across a small range starting at 4ppb-16 ppb(I didn't set up the criteria) and linearity across that range for the halogenated VOC's was good, consistent, and easy to achieve. The same column was used for both the FID and XSD, a DB-624, 30m, 0.53 id, 3um film. Drawbacks were maintenence related and it could very well be because of the Concentrator/inlet and dual column configuration. There was no venting mechanism prior to the detector other than the water management system on the concentrator.

[Peter Apps]

Hi Lila

I am not clear whether you have a split between the two columns that is nominally 1:1 but that varies from run to run, or whether it is not exactly 1:1 but is constant from run to run. Not having a reproducible split betwen the two columns must surely compromise the reproducibility of the analysis because the peak area from each detector for a given injected mass of analyte will vary with the column:column split. You cannot fix this problem by any calibration scheme. If you have a constant split then calibrate peak area from each detector against concentration of standard without any doubling or halving.

If you have two identical columns, why not split at the end of a single column ?

Peter

[Lila]

Peter,

The ratio to each column is not exactly 1:1, and it is fairly constant and have been able to generate reproducible data.

Your idea of splitting at the end of the single column sounds like a great idea, especially since the columns are identical. Thank you.