column died: why?

[ece]

Questions about GC:

Our last column, about $500, was a metal column that we killed in a couple of weeks. We didn’t even get to use it. I’m trying to backtrack the many steps that we did wrong to determine what was likely the root cause of stationary phase breakdown. At the time we didn’t have a leak detector and we didn’t have a replacement ferrule for the one that was cracked, but we installed without leak checking it, b/c we didn’t know. But we did a lot of the following things wrong:

1) If you use a cracked ferrule on the column and run it at high temperatures, can you get air into it and kill the stationary phase?

2) If you have a helium leak somewhere can you get air into the column and kill the stationary phase when you run it at high temperatures?

3) If you have a bent column (like with the metal) that doesn’t go in straight at the inlet, can that affect column performance? If so, how?

4) If you didn’t get a clean cut when you score and break away the column, can that lead to helium leak?

5) If you didn’t get a clean cut when your score and break away the column, and you accidentally get some metal pieces in the column, can that contaminate it?

6) If you install the column backwards, with the guard at the detector end and the activated portion at the inlet side, how does that affect column performance?

7) We were initially told to condition the column at 20 deg. below the maximum temperature the column could handle or 20 deg. below the method’s maximum temperature, whichever was lower, for a period of 20 minutes. This seemed too short to me, until later I found out it is supposed to be 2 hours. I did condition it for 20 minutes the first time. The second time I conditioned it for 2 hrs but it turned out that the column’s stationary phase was pretty much already spent at that point. I cleaned the detector parts in between conditioning runs, but I am concerned that improper conditioning was part of its demise. What is the proper conditioning procedure for a new, 15 m column anyway?

If you handle the column and parts with nitrile gloves on, does it contaminate it? How does that affect column performance? Is it that easily contaminated? We didn’t have tweezers when we started handling, but should you pretty much always use tweezers to handle all the parts when they are handled and cleaned? How easily contaminated are all the parts?

Root cause I would say is #1, not leak checking the system and having a cracked ferrule. What do you think? What are the consequences of all the other ones?

[Consumer Products Guy]

What type of phase on that column? What temperature did you take it to? Do you have an oxygen trap on your carrier gas?

OK, we don't use tweezers. And we rarely wear gloves. A non-perfect break will likely not affect the ferrule sealing surface, I'd rule that out. Also, rule out contamination from nitrile gloves killing the column. Heating with some air in the column will kill a PEG phase, not significantly affect some other types. Why do you think that the column is dead?

[ece]

I should clarify that we had an MXT-TG from Restek with an Integral connection b/w the guard column and the activated column portion, and the guard was made of the same material as the column, so the inlet was labeled with a tag that was misplaced when it was installed (in other words, we didn't know where the inlet was and where the outlet was). Another clarification is the ferrule that was cracked was the detector ferrule to the 1/4" nut.

The carrier gas does have oxygen/moisture traps in parallel and they are both fully functional. The highest temperature we take it to specified by the method is 380 deg. C for about 10 minutes. The highest temperature we were told the column could handle is 420 deg. C. So it was conditioned at 400 deg. C. I don't know what the stationary phase is, some kind of polysiloxane I imagine but haven't officially confirmed it.

[ece]

Why do you think that the column is dead?

incredibly high baseline that won't go down and noise when i did a heptane blank run. baseline continues to rise as the temperature rises and never stabilizes. clean the detector/inlet really well, replaced the jet, still get same results.

[Bruce Hamilton]

1. If you keep the carrier gas at a pressure/flow greater than the leak when the column is at any temperature above ambient, and outlet flows are reasonably consistent during the run, you are unlikely to get oxygen in. However, it's like playing russian roulette with all chambers filled - you're now relying on your own bad aim...

2. Same answer as 1.....

3. Discrimination in split injection, but you should be using On-column. Possibility of activity from exposed internal surfaces. Not good, but the revolver chamber has two rounds..

4. Very unlikely, would have to affect the ferrule, and lead to a leak. Revolver chamber has one round.

5. Yes, metal surfaces will be quite active for some lipid derivatives, but mainly would mainfest as variable retention ( backpressure) and poor peak shapes. Revolver chamber has two rounds.

6. Very significant. The precolumn is designed to cope with the underivatised gunk from your samples, and is 0.53 mm ID, instead of 0.32 so instead of retaining gunk, it smears it over the main column - reducing resolution and increasing baseline. the increased internal volume at the back would cause the peaks to be much broader, and separtion would be reduced. Revolver chamber has 5 rounds.

7. Follow the manfacturer's guidelines, but it's usually until the baseline in Ok, which for 0.1 um would be at least a couple of hours.

However, the most critical requirement is that the column is heated slowly and the carrier is free of oxygen, and that the samples are as clean as possible. Under no circumstances should a column be heated when there is no carrier flowing out the end.

I've handled columns with hands, issue is that outer polyamide coating becomes brittle after tiem at high temperature, and columns break. My oils don't seem to have much effect on life, but I always try to handle the cage, rather than the column. Handling with your paws doesn't affect the separation.

Because of the instability of lipid derivatives, it's a good idea to avoid contamination as much as possible, by solvent cleaning components.

I'd also suggest that your problem may still have some sample preparation issues, as oxygen is likely to kill the column stone dead in hours, so you may have had poor aim, but there's more revolvers on the table to test...

Please keep having fun,

Bruce Hamilton

[jf]

thanks for the advice bruce!

if any of you guys can comment on our chemstation thread, thats hit a dead end, that would be great. I feel that we were very close to resolving an answer

http://www.sepsci.com/chromforum/viewtopic.php?t=7628

thanks!

[ece]

we didn't actually run any samples, just heptane blanks. we did the butane test using a cigarette lighter but at 50 deg. C only.

but we did run the helium at really low flow rates initially b/c we thought it was supposed to flow at 3 mL/min, not really considering at the time that it was a larger diameter column throughout and thus required a higher flow rate. we did a few blank runs at that and maybe that caused oxygen to flow through the cracked ferrule. then we realized we should go by linear velocity and not necessarily by flow rate and had to adjust it for the new column dimensions. i think by then it was too late, the column was already dead.

i'm not sure if we scored the the column cleanly or not, but i'm a lot more comfortable with the idea of using a tubing cutter than a scoring wafer as it is really difficult to get a clean break on metal tubing, and there's a higher possibility metal shavings can get in and contaminate it.

[Peter Apps]

If this is the same column that had a very low and fluctuating inlet pressure in another thread, with issues about how to measure flow rates etc than it could have been killed by having air diffusing into the detector end while there was no flow through the column, or into the carrier gas when the inlet pressure was very low.

Another possibility is that the very low inlet pressure allowed backflash of solvent vapours into places that they shouldn't go, leading to contamination of the plumbing upstream of the inlet and a constant stream of heptane into the carrier gas. Can you replace the column with a piece of empty tube - even metal will do. If the detector signal stays high under these circumstances then the column is not the (or not the only) cause of the problem.

Another vague possibility - if you have a metal column couldn't it be shorting the flame tip to ground ? Try moving the column end of the detector down by 5 mmm or so.

Peter

[ece]

If this is the same column that had a very low and fluctuating inlet pressure in another thread, with issues about how to measure flow rates etc than it could have been killed by having air diffusing into the detector end while there was no flow through the column, or into the carrier gas when the inlet pressure was very low.

Bingo, Pete. That's pretty much what I think happened, too. We operated at too low of a pressure and air got into the detector side, while we were running at high temperatures, with a cracked ferrule, and no leak detector to leak check the system (which we have now)

I'm not an expert chromatographer, but to me that seems to be the most obvious answer. Anyone else think so?

[Bruce Hamilton]

You will find out soon enough, but I'm dubious.

This post is not to dispute your conclusions, but because if back diffusion isn't your problem, you could kill another column quickly...

The TAG columns are reasonably robust, and I've never managed to kill one by back diffusion. It's possible, but even the smallest flow of a carrier through the column should exclude oxygen and protect the column.

I assume you have duplicated the original performance test, and it's showing the peaks are tailing badly, and may have reduced resolution. If not, you probably should, just to confirm the column is dead.

I would be concerned about the front end of the GC, ensuring that your carrier gas line are always filled with carrier, and that the column and injector have carrier flowing through them before any heating occurs.

That especially includes at start up and cool down ( before shutting the carrier off ) - if you are regularly closing down the instrument.

I tend to leave GCs at equilibrium with a reduced carrier gas flow when not used. During a sequence, I also watch carrier pressures and flows regularly for the possibility of leaks at the inlet side, especially during injection, as the instrument is going through major thermal swings for each analysis.

A leak detector is great for setting up systems. Recognising that instrument manufacturers prohibit using soap, my perception is that any leak test is better than no test. If you have to use soap, always ensure the system is at relatively high pressure, and rinse the soap off with distilled water, and thoroughly dry before allowing the gas pressure to drop.

Please keep having fun,

Bruce Hamilton

[ece]

Hi Bruce,

I think that's a good policy. I wouldn't want to rule anything out, as if that isn't the problem then we may have another $500 column we won't be able to use if we kill it.

Be that as it may, we were using a cracked ferrule for awhile b/c at the time we didn't have a replacement, nor did we have a leak detector, and so we just went with it assuming it would probably work. Although we had a new ferrule installed a couple of weeks before we did actually get a leak detector (electronic, not soap!), we had still been operating at really low pressures (2 psi) which I found out later was too low for that column's .53 mm diameter. When we got the new leak detector I saw a marginally small leak at the detector end when it was hot, even with the new ferrule in there. Once I had corrected it, the column's performance was still poor as far as bleed and baseline noise, so I concluded that was what had done it in.

Although I did not duplicate the column's original performance test exactly, I did use similar chemicals that we had on hand (heptane instead of hexanediol) combined with an internal standard test. I got a peak for the heptane and very low peaks for the internal standards b/c they were intercepted with the baseline noise from bleed. The point was to check for resolution only, and how much resolution was achieved, and at what retention time. It didn't seem to matter what chemicals were used as we don't use the chemicals indicated in the performance test for our method anyway.

I generally always make sure there is helium flowing through the column before I start up every day and that the flow rate is correct. Also, before I conditioned it, I had the end detached and submerged it in a little methanol/acetone and confirmed that there was flow through the end of the column (bubbles).

I have personally leak checked the system with the dead column by uninstalling/reinstalling, heating and cooling, for practice when I install the new one. When I do see a leak, it's usually the detector side when it's really hot. I've rarely seen the inlet leak. I've also personally leak checked the helium lines many times to see if there were any leaks, in and around the fittings and valves, and have not seen a leak yet.

I NEVER turn the instrument unit off when the oven and unit are hot, I cool all the parts down to 25 deg. C before maintenance. The only time where something like that has happened was when the power went out and even then the oven wasn't very hot (50-80 deg. C). I believe at that point too the column may have already died.

I see some regular swings on the pressure control but they are usually very minimal (.5-1 psi). That I believe has more to do with the way that pressure control self adjusts to get within a defined range of pressure than it does with a leak. But I suppose I can see how that would introduce air into the column, although I fail to see how that could be easily controlled. No matter how sharp the needle is or how well the septa seals, once you penetrate it air can get in there. I guess I'll just make sure the sample preparation involves blanketing the headspace with helium as best I can to drive out the oxygen and use a sharp needle.

Thanks for the support!