different retention times on repeat runs of the same sample

[officed144]

I am an newbe at GC so please bear with me.
Here at work we have an older GC with a packed column and a FID detector that I am trying to use to quantify some additives and contaminates in our mill rolling oil. The peaks I am having trouble with are from an additive - lauric acid. I see the peak fine, but the retention time isn't always the same. For the sample I was running today, the peaks were sometimes 10 sec apart which is playing havoc with my peak analysis efforts.

I inject the sample manually, but I know I have less than 1 second variation in my injection technique - certainly not 10 seconds.

My heating profile lets it set for 5 minutes then it ramps up evenly to max temp over the next 25 minutes. Is this something I should change? Or is there somewhere else I should be looking to solve this problem?

Could some of you more experienced folks give me some ideas on how to reduce the variation in retention time?

Thanks,
Ron

[Don_Hilton]

If your oven temperatures are consistant, flow becomes the next guess.

Check for gas leaks. Use a leak sniffer (if you are using helium or hydrogen as a carrier. They do not work well with nitrogen.) If your GC uses needle valves, check to be sure that you have adequate pressure to the back of the GC (at least 20 lb above the maximum pressure you need at the head of the column). A leaking septum can give variation - but 10 minutes is a bit much for a leaking septurm. (Check it any how.)

Make sure that your GC run is long enough that the compund comes out in one run. Otherwise, it will stay on the column and appear in the next run. (I've done this. My first sample had no peak and my blank at the end of the queue had a peak... And because the time between manual runs varries - so does the retention time of the carried over peak.)

[Peter Apps]

Gas flow or temperature are the obvious suspects, and on an old (how old, which model ?) GC a 10 s variation in retention is not too bad. What is the expected retention - several tens of minutes ? On a packed column I would expect the peaks to be several tens of seconds wide, so a 10s shift should not move a peak out of its expected retention window.

What is the maximum temperature ?, and how long do you allow for the oven to stabilise after it has cooled down ?

Peter

[Ron]

Is there any pattern to the variation in retention times? Is the first run retention time consisent, and following runs have a slightly earlier retention time? If so then the equilibration time between runs is probably too short. Packed columns have larger thermal mass than capillary columns, and it taked longer for the column to heat and cool.

If the method is used to analyze only one compound, I would suggest an isothermal method for the analysis if you can get the proper resolution running isothermal. The most stable retention times are achieved with isothermal methods, assuming a leak-free system and good flow control.

[officed144]

I want to say thanks to all you who have replied thus far. I am still learning and your comments are all helpful.

For the person who asked, this is a Shimadzu 14A GC. Yes, it is old and pretty simplistic, but as far as I can tell, everything on it works as intended. It has a mass flow setup for the He carrier that I don't know how to check. It works, but I don't know how to check the actual flow to see if it is high, low or consistent. The only problem is that the Class VP software version we have only works on a Windows 95 setup. The PC is old and I am unsure as to how to replace it as new software is pretty pricy. Still thinking on this.

Right after I posted yesterday, I remembered that I hadn't changed the septum recently and pulled it out. There were some serious chunks of septum rubber inside the glass liner. Enough that I think they could have been hindering the vaporized sample from getting into the column consistently. I cleaned out the liner and replaced the septum and the peak retention time seemed to be more consistent. I will just have to try it for a while and see.

There were so many tips that it will take me a while to digest all of them and try different things. Thanks

[Ron]

There should be an ugrade price for the VP software that is significantly less than the cost of a new license. You might want to contact your local Shimadzu rep to see how much an upgrade would be.

[Peter Apps]

So this was a leaky septum - bits fo septum in the inlet would gibe broad, tailing peaks rather than a retention shift. Rather than spend money on updated software to run an old but good instrument, invest in a leak seeker, then you can check for leaks as part of routine trouble shooting and maintenance. To measure the flow through the column, turn off the detector gasses, connect a flow meter to the outlet of the detector and measure.

Peter