GC Newbie - please be gentle!

[HPLC_Help]

We had an incident where our methanol peak (the diluent for our samples) moved about 6 minutes out from its normal RT. The Varian support person (we have a Varian GC) suggested we cut the column, and said we may have to do so regularly because of our dirty samples.

Does this make sense? I'm an HPLC guy, so I'm learning on the fly.

Thanks!

HH

[Don_Hilton]

I'll have to answer like my lawyer: "It depends."

Are we talking earlier or later in the run? What GC conditions are you using and have you made any changes (like head pressure or column temperature)?

Cutting a meter or two off of a capillary column makes sense - if your column is degrading. Generally you have other things going on like bad peak tailing. I would want to be sure that there is nothing else going on like a lower (or higher) carrier gas flow rate, for instance.

[Peter Apps]

Anything that moved a methanol solvent peak by 6 min would be very likely to have affected everything else on the chromatogram as well. I doubt that deterioration of the first part of the column is the (only) cause. As Don says; give us all the details.

Peter

[HPLC_Help]

Thanks, Don and Peter. Here is what I know.

We have fermentation samples that we're analyzing for ethanol content. The samples come in quite dirty, with all the media and other fermentation by-products (sugars and organic acids) present.

The sample prep is 1 part sample to 9 parts methanol. Centrifuge and filter.

The column is a Restek Rt-Q-Bond, 30 meters x 0.53 mm. It's direct-inject, 1 uL injection. I think they sometimes split the injection.

I THINK the flow is 10 ml/min and the temp is 250, but I have to confirm that. (I don't run the instrument, obviously).

The methanol peak has always come off at about 2.5 min (very broad, obviously). The ethanol peak came at about 4 to 5 min (I'm writing this from memory). Suddenly, all peaks went away. We examined the column, and the first three inches were blackened. After cutting the column and re-attaching, the methanol moved to about 8 min. The analyst re-inserted the column and the peak was back to normal.

SO MY ORIGINAL POST HAD AN ERROR! I think the shifted RT was due to the column not being attached correctly, not because the column was bad. We cut the column because all the peaks DISAPPEARED.

However, the "lost" peaks ... is this due to direct-inject of dirty samples? Should we come in every Monday and snip off 5 inches, as was suggested to us by the GC vendor?

Thank you!

H_H

PS: I'm looking for a good day-long GC course in the Boston area for myself and some analysts. Obviously, we're beginners!

[Consumer Products Guy]

Check your carrier flow by disconnecting the column at the detector end. If below what you expect, check the inlet for leaks.

[Peter Apps]

The complete disappearance of all the peaks could have been due to the column being blocked with crud. The late elution after the column trimming was probably due to a leak at the inlet end. With samples this dirty you need to find some way to protect the column from the involatile muck. Rather than a direct injection try an ordinary split-splitless injection, or even split if your ethanol concentration is high enough to give you an acceptable peak size. I would also put in a retention gap - 1/2 m of deactivated uncoated fused silica connected to the column with a pressfit connector. You then replace the retention gap rather than trimming the column.

How often are you changing inlet liners and septa ?

Peter

[HPLC_Help]

We're changing septa and liners at the recommended intervals, but I think we may have to do it more frequently.

Would the retention gap you described come in addition to a liner, or instead of a liner?

I think we may also have an unresolved injector variability issue.

Thanks, everyone!

H_H

[Peter Apps]

You always have to have a liner in the inlet - a retention gap connects to the inlet in the same way that the column would, and is then connected to the column.

When you say that you are doing "direct injection" does this mean that you have the column sealed into a tapered section at the bottom of the inlet liner, or that you are actually injecting sample into the column itself rather than into an inlet liner, or something else ?

These samples are crying out for equilibrium headspace analysis, any chance of that ?

Peter

[HPLC_Help]

I would love to do headspace, but we don't have the appropriate setup. We need a heated sample compartment for it, right? To vaporize the sample? I found a paper online that describes exactly this technique for these types of samples.

"direct inject" in our case does indeed mean injecting straight onto the column with no liner.

H_H

[Ron]

There are some things here that don't add up. You said the injections are done directly into the column, which implies you are using an on-column injector. In the first post sometimes using split injections was mentioned, which is not consistant with an on-column injector.

If you are really are doing on-column injection for this type of sample the column will have to be trimmed regularly, it will become contaminated quickly.

[AICMM]

HPLC_help,

If you really are doing on column, and it is really directly on the Q-plot, your needle is probably scraping the column packing off the walls. All the more reason to use a retention gap like Peter says because it will survive the needle insertion better than the Plot probably would.

Best regards.

[GregK]

HPLC_Help-

You've come to the right place for getting answers. The people here have helped me a lot in the past, and I love coming here to read up on problems I might have in the future.

My 2 cents: We run a pseudo-headspace analysis here on methylene chloride for one product. We put the sample in a headspace vial (large vial with septum top), add a salt and some water, place it in a hot water bath to volatilize the sample, then pull 1mL of the headspace out and do a manual injection. It requires a lot of "babysitting", but works. If you don't have too many samples, or don't require extensive system suitability, or have a short run-time (lots of ifs), maybe give that a try.

Otherwise, yeah, what everyone else has said. Get a retention gap (guard column) to save the column.

Hope that helps.

-Greg

[Peter Apps]

Hi H_H

What model of Varian GC do you have, and what model of inlet ? On-column injections require a special inlet, direct injections require a specific liner. What temperature is the inlet set to ?.

With the dirty samples that you have, and a PLOT column, on-column injections are the worst possible option.

What concentrations of ethanol are in your samples ?

Peter

[Don_Hilton]

You say that you are changing the liner at the reccomended interval. Two questions: 1) how often is that reccomended interfal in number of injections and/or time. If time how many injections do you make in that period of time? 2) what does the liner look like when you replace it? is it discolored?

I am curious as to how long this method has been running in your lab. Is this a method that your laboratory has been trying to develop or is it one that was developed and worked well until reciently. My thinking in this is that if this method was developed some time back and ran well, but the experienced operator moved on, the problem may be a change in some step in the method - and resotration to original procedures and supplies could resolve the problem. If not, I would suggest that you give us as much detail as you can including the type of GC, inlet, type of liner, conditions, and all that. The more details you post, the more likely it will be that suggestions posted here will be on target.

[HPLC_Help]

Hi, everyone -

Thanks for the info. I'll try and clear things up a bit. Again, I'm new to GC and I'm transmitting info from other people to you, so sometimes bits get lost in the translation!

Ron: indeed, the split injections were being done in the past, but have not been tried in a while. We ordered a mega-bore column which (I've been told) can't be used with split/splitless. SO - we had to switch to the on-column injector.

AICMM: are retention gaps available from column vendors?

GregK: I don't think we'll have the time to do the pseudo headspace regularly, but it might be worthwhile to try as part of a plan to justify getting a headspace injector.

Peter Apps: We have a Varian 450-GC with a Varian CP-8400 autosampler. Ethanol concentrations are quite high - anywhere up to 30 mg/mL. We're getting strong signal. Inlet temp is 250.

Don_Hilton: we are indeed developing this method. The methods for our application that I've seen online have been headspace. After talking to some people, I'm not sure that we're even using the best column we could be.

I think my plan will be: 1) confirm I'm using the best column for my application; 2) get the correct column for a split/splitless so that we aren't injecting as much gunk onto the column; 3) price out headspace samplers.

I thank you for your help. Please keep the suggestions coming, and let me know what you think of my plan. My alternative plan is to tear out all my hair.

H_H

PS: Oops, too late for the last option. I guess I'm stuck.