How to set up backflush system in 6890A GC?

[Anthony_Ng]

Hello all,

Today I read an application note of Agilent (5989-8664EN, Capillary Flow Technology for GC/MS: a simple Tee configuration......) to set up a backflush system.

Our 6890A has installed split/splitless in front inlet but nothing in back inlet. What hardware should I buy (e.g. another inlet?) for the back inlet position so that I can perform backflush function?

By the way, other than setting up at back inlet, can I use auxillary pressure as backflush pressure source? Can I (and how) to do it myself?

Thanks in advance.

[Schmitty]

We got a new 7890/5975 with the quick column switch option, which I think included the AUX EPC and a bunch of extra hardware. I think the backflush system is very similar to this.

I bet Agilent would provide you with the necesary part numbers (or whole kit necessary) if you ask for the list nicely.

[GCPRO]

You will need a separate pressure source. I prefer using an Aux module because it has three channels (You get two extra channels for use as backups or for more complicated experiments). The other choice is a PCM module. They cost about the same.

You will also need a Capillary Flow Technologies device of some sort. These include QuickSwap (only relevant for an MSD), purged splitter plate, purged Ultimate Union (anything with a purge). Do you have an MSD?

The biggest decision for you is what form of backflush makes the most sense to start with. I find post-column backflush to be most straight forward and easiest to set up because you simply backflush post-run after the last peak of interest elutes. Agilent also has the most documentation on this approach, and there is a backflush wizard screen in ChemStation to help set conditions. To do post-column backflush you'd use Quickswap or a purged splitter device between the end of the column and the detector, with a short FS restrictor going from the device to the detector. Pressure is increased at the device at the end of the run while the split inlet pressure decreased. Flow through the column reverses. Simple. For background principles, read:

“Optimizing capillary column backflush to improve cycle time and reduce column contaminationâ€

[aldehyde]

Hi, backflush owns--I agree. Here is a useful article from Agilent:

http://www.chem.agilent.com/Library/tec ... 5111EN.pdf

The two part numbers you will want are:

6890 Aux EPC module 19231-60610 [You want a brown dot restrictor for your purpose, I think it comes with this but I'd make sure 19231-60610]
Two way splitter with makeup gas G3180B

Here is an installation/operation guide for the two way splitter. Has some useful info: http://www.chem.agilent.com/Library/use ... 045610.pdf

[Anthony_Ng]

Thanks a lot!

We cannot use post-column backflush technique because we are using diffusion pump in MSD. If we use quickswap device, as Agilent said, some oil will "blow" into the MSD.

Therefore, I think connecting a splitting tee device in the mid way is fine. (i.e. changing a 30m column into 2*15m columns, connect the Tee in the mid way and aux EPC to the tee)

The most difficult thing is: how can I know the analytes of interest passed to the 2nd 15m column but high boiling dirty stuff still stay in column 1 so that I can start the blackflush at a correct time?

[GasMan]

There are two ways to set this up.

1. You make a guess with the backflush time, and from the results you either increase or decrease the backflush time until you get the reults that you want. This can be time consuming.

2. You need to replace the second column with a restrictor tube that has the same restriction as the column. Agilent I believe supplies a calculator that enables you to do this. It will also give you the 'holdup' time, which you will need to know. You then inject your sample, and you will see on your detector the times that the compounds elute from the first column. To get the correct backflush time, take the end time of the last peak of interest and subtract the holdup time. This will give you the correct backflush time.

Replace the restrictor with the second column and you are ready to go.

Gasman

[AICMM]

It seems to me that with two 15 meter columns of the same phase with the backflush in between should be pretty straight forward. Run the standard (without the junk), take the time to get off peaks of interest and cut that in half as a starting point. Fine tune from there. Of course it is a bit more problematic since the 2nd 15 meters has significant vacuum influence but that seems like the easy way to start.

Won't hold up for different columns in this configuration but that is not what you have noted here.

Best regards.

[Peter Apps]

Dividing retention by two to get the time when the peaks pass the mid point only works for isothermal analysis. In a programmed run most of the time that a peak spends in the column it is sitting close to the inlet waiting for the temperature to get high enough for it to start moving.

Peter

[GCPRO]

The idea of using a post-column restrictor is OK, but uncoated (even deactivated) fused silica will never be as inert as coated columns. This approach would require a fairly long post-column restrictor.

Using an uncoated but deactivated FS column in front of the analytical column, with purged device in between, would be a better choice because a shorter piece could be used and crud never makes it to the analytical column. Analytes move from the uncoated tubing to the analytical column at something around 140 C below elution temperature from the analytical column and are focused. This allows some slop in backflush timing, triming of pre-column, etc..

For split column approach, void time in the first half (in the described scenario with MS) is approximately 2X that in the 2nd section of column. Elution temperature is dependent on dimensionless ramp rate (C/tM), with faster rates eluting compounds at higher temperatures. At 10 C/tM (total column length), compounds elute at ~ k = 2, for reference. So when solutes are passing the midpoint (at a lower temperature in the program), they might be at k ~ 3 or 4. Then by the time they elute, they are at k = 2. So maybe retention in the second half was only 2 - 4 void volumes (tM the second column section).

If last compound of interst X elutes at 200 C from the total column, for example, then calculate (based on the midpoint pressure at that time and vacuum outlet ) the void time of the 2nd column section at the 200C temp, multiply by 3 or 4 (pick 3 and see what happens, who knows at this point?) and subtract that time that from the RT of compound X. That might be a good starting point for initiation of backflush in the first column section. Fine tune (maybe by void time increments or decrements) based on whether you want to see 100 % of Compound X or 0% of compound X. Of course all this depends on your actual dimensionless ramp rate as well as relative column lengths, etc. But it gives you an idea of one possible approach. I am sure there are others that would work.

[Anthony_Ng]

Thanks for all inputs.

After reading the recent paper: J. Sep. Sci. 2009, 32, 3133-3143, I got a bit confused with the concept of "Dimensionless ramp rate"

Surely it does not refer to the ramp rate (e.g. 50oC/min) in the temperature program setting in Chemstation. Can anyone explain a bit more about this?

By the way, is the "elution temperature" means the temperature at which the compound elute out the analytical column?

[GCPRO]

Elution temperature is the actual oven temperature at the time of compound's elution.

Dimensionless ramp rate is the fundamental parameter that influences elution temperature in a temperature programmed analysis with a specific column. Dimensionless ramp rate is a combination of hold-up time and temperature ramp rate. Dimensionless ramp rate has the units C/tM. To calculate, multiple the current ramp rate (C/min) times the initial void time of the analytical column (tM).

Faster dimensionless ramp rates will cause a given compound to elute at a higher temperature. That is is why you can see some selectivity changes by adjusting oven ramp rate independently from flow (or vice versa); compounds of different polarities have different retention vs. temperature dependencies. The concept of Method Translation was developed to allow predictable scaling of methods. In the MTL approach, oven ramp is scaled proportionally to hold-up time such that dimensionless ramp rate is held constant. In this way, retention times are predictably scaled, individual elution temperatures do not change and relative retention remains constant.