O2 loss in PLOT Columns, Permanent Gases

[nerd75]

Please help me...
I am running parallel PLOT columns (5A MolSiv + PoraBondQ) Helium carrier, and PDHID detector for the analysis of permanent gases within process sample gases. For calibration, I have a 10ppm H2, O2, N2, CO, CH4, CO2 standard injected via a 0.5ml sample loop. Inexplicably, I will see O2 COMPLETELY disappear over time. I can bring it back by injecting 100% O2 (sometimes heating to ~200C also), presumably passivating the system/SS transfer lines, but it is only temporary... O2 ALWAYS gradually and completely disappears. The other gases are not affected.
Strangely, it seems that the O2 loss occurs simultaneously in both columns, suggesting an issue upstream of the columns.
Where is the O2 going?
Can anyone offer an explanation, or a permanent solution? Has anyone else had a similar problem?

[chromatographer1]

O2 loss is not a surprise with a molesieve 5A column. With the Porabond Q however, I have not heard of that before, although that column is made by a different company than the one for whom I work.

The obvious answer is there is something consuming the oxygen. It may be coming from the carrier gas or bleeding down from the pneumatics upstream.

I would simply condition both columns, first at low temperatures, then raising the temperature in steps. I would test the columns after each step and determine if the oxygen loss had disappeared to determine if additional conditioning was necessary.

If simple conditioning doesn't fix the problem ('burning off' whatever is consuming the oxygen) then I would use air for carrier for a short time and see if that fixes the problem. Later you may need to put a guard column ahead of your present columns and see if that traps the problem, I would suggest a thick film Wax column.

While the oxygen loss is not large, the fact that you are working at such low levels makes this loss an issue.

If the reducing sites that consume the oxygen are involved in the chemistry of the materials from which the columns are manufactured then you will have no recourse but to oxidize (condition) the columns routinely.

I suspect the problem lies elsewhere. I assume you are confident that it is not a detector issue by swapping out a new column and finding that the O2 loss went away. Good luck in finding the cause and a permanent solution.

best wishes,

Rod

[AICMM]

Nerd75,

One question, what is the balance gas (if I may ask) and does it pass through the detector?

Best regards.

[nerd75]

The PDHID uses its own supply of discharge gas, in this case gettered He.

[chromatographer1]

Nerd75

I think AICMM meant the balance of the gas in the sample, and "are you venting the sample balance to vent through multidimensional chromatography".

I think he is implying there might be an issue with the HID with certain gases that might desensitize the detector.

best wishes,

Rod

[AICMM]

Nerd75,

You are measuring low ppm levels of fixed gases blended into a balance of some other bulk gas. What is the bulk gas? Could this be creating active sites that are then consuming your low levels of O2? Also, do you vent off (backflush, vent valve, multicolumn...) the balance gas before it reaches the detector?

Best regards.

[nerd75]

i see your point, however the 10ppm standard is prepared in ULSI grade helium (same as my carrier). i would not think this would actively reduce O2 in my system.
The system will ultimately be used for analysis of permanent gases within corrosive matrices (semiconductor gas application), utilizing a silicaplot precolumn, and a backflush to keep these corrosive gases from reaching the analytical columns or detector. But I have not gotten that far yet. I am alrady seeing the O2 disappearance, as described, the only gas my system has ever seen is helium or the occasioanl slug of air.
i am using varian plot columns, do you think it possible that another manufacturer may behave differently?
I do very much appreciate the input, keep the ideas coming!

[chromatographer1]

In what instrument are you developing this method? What kind of gas regulation are you using?

This line of questioning is to determine the possible hardware sources of contamination of the columns. EPC units and septa and seals can offgas contamination that may be the cause of this O2 loss.

best wishes,

Rod

[nerd75]

A good thought as well.
I am using no septa, all lines are SS 1/16" tubing, sample is introduced via closed 0.5ml sample loop (also SS). Using Valco 6 port valves. My regulators have metal seals, no polymers anywhere.
O2 disappreaance also occurs when my baseline is solid, and no other peaks are affected, so i do not think to be a leak. cannot really think of anything to offgas or adsorb O2.

[chromatographer1]

Other than your lines were not clean (and it happens to the best of us) I would simply suspect that this requires oxidation of whatever may be in the system. Flush your lines, columns, connections etc with clean dry air and see if that eliminates the problem.

Sometimes it is not necessary to find a cause for a problem.

It is only necessary to GET RID OF IT.



best wishes,

Rod

[JI2002]

I had a similar problem a while ago. I changed out the packed column used for permanent gas analyses and put in a different packed column for another analyses. When I changed back to my original column for the permanent gas, the oxygen peak disappeared in the first injection of the standard (a mixture of N2, O2, CO, H2, CH4 and CO2 @2000ppm), as I made more injections, O2 peak increased. I ran a bunch of air samples for methane(O2 was not a target compound), by the end of the day, O2 response went back to what it was.

My thought was that O2 was consumed by some active sites, as I ran air samples, O2 in the air saturated the active sites. Because the sample matrix I run is air, not Helium in your case, there are no active sites in the system. I would never have this problem if I hadn't changed out the column. I don't know how active sites were introduced by changing the column. But I think it has something to do with metal parts in the system, such as copper ferrules.

[chromatographer1]

Nerd75

Have you had any success?

Can you share your solution?

best wishes,

Rod