Retention Times Shifting

[cawarhur]

I'll preface with the setup of the GC:

Agilent 6890
Rtx-5 105m x .25mm x .25um
Analytes: Hydrocarbons from C5-C14(ish)

I've noticed that the major peaks in my chromatograms are shifting up to as much as 0.2 minutes (on a 105m column with about a 1.5hr run time). This, of course, is a problem, but I can't figure out what is causing the shifting. I've considered all the basic causes and ruled them out: tThe column is new (roughly 2 months old), carrier solvent is and has always been DCM, flow rate and oven temp are constant between runs.

I considered that the sample may be too concentrated--each sample is made by adding 5 drops sample, then filling to 1.5mL with DCM...so with each drop being roughly 25mg and the major component at 90% or above, that puts the concentration of that particular component at 100mg/mL (roughly) which is way too concentrated, I know.

My issue with this is that I did 3 runs each using 1 drop, 2 drops, 3 drops, 4 drops, and 5 drops of sample to get a handle on the way the concentration is affecting the shifting of retention times--there was no significant difference between the different samples. I ran these samples on a shorter column though to save time, so perhaps it wasn't an accurate barometer for my problem. But to give you an idea of the repeatability I'm looking for, between each of those runs, the most any peak shifted was 0.004 minutes.

I'm running out of ideas...help!!

[Peter Apps]

Drops

Do yourself a favour and make up some properly measured dilutions, calculate the concentration that you need to put about 100 ng of the component that makes the biggest peak onto the column, taking into account the injection volume and split ratio.

All all the peaks equally affected in a given run - in other words does the spacing between peaks stay constant while their retentions change by the same amount ?

How do you know that temperatures and flow rates are constant between runs ?

Are you doing manual or automatic injections ?

What is your temperature programme ?

Are you running the carrier gas at constant pressure or constant flow, and if constant flow is it constant linear velocity or constant volume flow ?

Peter

[cawarhur]

I'm injecting 1uL with a split ratio of 20:1 of something that is 100mg/mL with the major component at 90%, doesn't that put me below your prescribed 100ng at ~5ng for the major analyte?

1) The spacing between peaks seems to stay constant.

2) The only way I "know" that the temp and flow rates stay constant is because the method isn't changing. I imagine some people run into situations where retention times shift because they or someone else changed oven programs or flows depending on their needs and never reverted back to the original method. It is something I could look into further if you think the GC itself may be defective.

3) Automatic Injections

4) Oven Program:
C/min Next C Hold Min
Init 36 1.00
R1 8.00 80 0.00
R2 11.00 270 8.00

Injector port @ 160C
Split Ratio: 18:1
Injection Size: 1uL

5) Constant pressure (44.14psi) which puts flow at 2mL/min and avg. velocity at 25.

[Don_Hilton]

What is the condition of the septum on the GC? Are peak intensities changing as the retention times change - if you do multiple injections from the very same GC vial? (on the GC with the problem)

[Peter Apps]

I'm injecting 1uL with a split ratio of 20:1 of something that is 100mg/mL with the major component at 90%, doesn't that put me below your prescribed 100ng at ~5ng for the major analyte? I calculate it at about 5 micrograms i.e. 5000 ng, which agrees with your assessment that it "way too concentrated"

1) The spacing between peaks seems to stay constant. Is it constant or not ? if it is, the problem has to be with the synchronisation of the injection with the programme start, or something wrong right at the beginning of the run. Anything happening later in the run will affect only the peaks eluting after that time, and you will see changes in the relative retention times.

2) The only way I "know" that the temp and flow rates stay constant is because the method isn't changing. I imagine some people run into situations where retention times shift because they or someone else changed oven programs or flows depending on their needs and never reverted back to the original method. It is something I could look into further if you think the GC itself may be defective. Plainly something is defective, and it could well be either the flow control or the temperature control. Check the retention times of methane (or butane from a lighter at a pinch) at various points through the temperature programme (with no sample injected) to check that the pressure does stay consant from run to run - Agilent has a flow calculator that might get you from holdup time to pressure. If you can find a fast reading thermomoter check the real against the actual column temperature as it programmes

3) Automatic Injections

4) Oven Program:
C/min Next C Hold Min
Init 36 1.00
R1 8.00 80 0.00
R2 11.00 270 8.00

Injector port @ 160C maybe a bit low - you might be getting slow and inconsistent transfer from inlet to column
Split Ratio: 18:1
Injection Size: 1uL

5) Constant pressure (44.14psi) which puts flow at 2mL/min and avg. velocity at 25.

Your retention times are shifting by 0.2% of the run time - maybe this is the best that you can expect ?

Peter

[JI2002]

I'll preface with the setup of the GC:

Agilent 6890
Rtx-5 105m x .25mm x .25um
Analytes: Hydrocarbons from C5-C14(ish)

I've noticed that the major peaks in my chromatograms are shifting up to as much as 0.2 minutes (on a 105m column with about a 1.5hr run time). This, of course, is a problem, but I can't figure out what is causing the shifting. I've considered all the basic causes and ruled them out: tThe column is new (roughly 2 months old), carrier solvent is and has always been DCM, flow rate and oven temp are constant between runs.

I considered that the sample may be too concentrated--each sample is made by adding 5 drops sample, then filling to 1.5mL with DCM...so with each drop being roughly 25mg and the major component at 90% or above, that puts the concentration of that particular component at 100mg/mL (roughly) which is way too concentrated, I know.

My issue with this is that I did 3 runs each using 1 drop, 2 drops, 3 drops, 4 drops, and 5 drops of sample to get a handle on the way the concentration is affecting the shifting of retention times--there was no significant difference between the different samples. I ran these samples on a shorter column though to save time, so perhaps it wasn't an accurate barometer for my problem. But to give you an idea of the repeatability I'm looking for, between each of those runs, the most any peak shifted was 0.004 minutes.

I'm running out of ideas...help!!

Are the retention times getting longer or shorter? Did anyone cut the column by any chance? Did you condition the column when installed 2 months ago? What's the time frame for the 0.2 min shift?

[cawarhur]

@Don Hilton: It's typically one injection per vial and the septum is changed weekly (maximum number of injections in a week is around 75 and it rarely reaches that).

@Peter: Arithmetic is my archnemesis, you are correct in your calculation (someone missed a factor of 1000...). The reason our injector port temp is low is because above 160C, we'll crack our major analyte. Is there a best choice for inlet liner for using a low temp? As far as spacing is concerned...what do you consider a significant difference in separation? Are you asking if all retention times shift equally?

@JI2002: Retention times fluctuate between being shorter and being longer. Nobody maintains the columns but me. That being said, the column was not cut. The column was brand new, conditioned, and well-kept through maintenance. Also, I'm not sure what you mean regarding the 'time frame'?

I called Agilent yesterday and on their recommendation, I cut the column, changed the ferrule, liner, o-ring, gold seal, and septa (injection port only where applicable). They seemed to think it was a leak. If not that, they said it could be a cracked column. We'll see where my replumbing and replacing of parts takes me.

EDIT: Additional question: How does concentration affect an FID's life/integrity? If samples are extremely concentrated, how will it affect an FID? I'm wondering if our FID's here need some TLC now...

[aldehyde]

Run a leak test: either install a column nut with a no hole ferrell or trim the column while it is installed and plug it with a septa. Plug the inlet's septum purge vent with an 1/8th inch swagelok plug. Set the inlet to 25 psi and allow it to equilibrate for a minute or so. Turn off the pressure and watch it for a few minutes, it should stay steady.

With the amount of sample you injected you may have contaminated the inlet. Turn off the temperature and wait till it cools, remove the liner, column, and gold seal. Wet a q-tip with the same solvent you dissolve the sample in and wipe the inside of the inlet. If it comes away discolored you need to clean it up. In case of bad contamination you can buy a 38 caliber gun brush and use it to scrub the walls. Use q tips to clean any residue after ward. This isn't a great thing to do often but is helpful in case of severe contamination. This guide details the procedure: http://www.chem.agilent.com/Library/Sup ... a16022.pdf

You may need to clean or replace the split vent line and replace the split vent trap. To do this you have to remove the cover surrounding the inlets using a T-20 screw driver. You can remove the line and rinse it out with solvent or replace it. I'd use a q-tip to clean the port where it connects to the inlet.

Your FID should be fine, anything that makes it through the column is going to keep on trucking and get destroyed in the flame. You may want to replace the jet if you haven't already. This guide shows you how: http://www.chem.agilent.com/en-US/Suppo ... 05099.aspx make sure to select the right jet--if you use the wrong one you might not see any peaks anymore. Look inside the oven at the bottom of the FID to figure out what kind of jet you need (or look at what is in there when you pull it out): http://www.chem.agilent.com/en-US/Suppo ... IDjet.aspx

GC problems usually occur due to the inlet, FID is really rugged. What is your baseline signal?

It is a good idea to disassemble the FID top assembly and wipe the ceramic insulators with a kim wipe (dont use any liquids on them). You can sonicate the collector if it is dirty.

[Peter Apps]

Yes, I am asking if all the retention times shift equally.

While inlet maintenance cannot do any harm (as long as nothing gets broken in the process) I doubt very much that inlet contamination could shift retention times without also degrading peak shapes quite badly, and a contamination problem will not vary from run to run.

If you have a temperature sensitive analyte that stops you increasing the inlet temperature above 160 C you should consider cool on-column injection. As a quick try out step, Restek do a special tapered liner that takes a 530 um retention gap and guides a standard syringe needle into the top of the retention gap.

Intersting that Agilent thought that it was a leak - back pressure regulated inlets are supposed to be able to maintain inlet pressure and column flow even when there is a leak.

Peter

[Bigbear]

Would you consider going to constant flow? A 105M column has significant resistance to flow. If you are using He as a carrier @ 270 there will be a considerable slowing of flow due to the viscosity of He.
For grins try a few injections changing to constant flow only.

[aldehyde]

Eh I mainly recommended the inlet maintenance as precaution since he was injecting such highly concentrated samples. The inlet *should* maintain pressure, but you'll probably still see weird effects on chromatography with a leak.

[Don_Hilton]

A badly cored septum can give enough of a leak that the inlet does not quite hold pressure, I know that I've seen this have an effect on the chromatogram on a 5890 - in the days before the EPC. Even though the EPC will shut down if it can not maintain pressure, there can be some deviation without it shutting down. I don't know that I've run across a 6890 with a cored septum (just a matter of operators, not instrument model).

In either case, the leak should be sufficient to detect wiht a leak sniffer if helium or hydrogen is the carrier gas.

On doing other maintanance, a suggestion: first fix what's broken. It leaves less to chase when figureing out what is not quite right after doing those other tasks.

[cawarhur]

I'm going to perform some of the suggested maintenance this afternoon, including cleaning the inlet and split vent line/trap.

My guess is that the inlet is contaminated due running the inlet at such a low temp and that the split vent line is a big plugged because because I started at this company, it had never been maintained at all (the instrument is several years old now).

I'll post after I get some results.

[cawarhur]

After cleaning the inlet and split vent line, I ran a standard overnight. Peaks differed between runs by about 0.005 minutes and were offset from pre-maintenance runs by roughly 0.05 minutes.

I'm not sure which was the problem, the inlet or the split vent line or a combination of both, but it seems to be running pretty consistently now.

Thanks for the help!

[WillyWonka]

Hello ,

I was just coming in the middle of the conversation. Your problem could have been at the septum. Yeah, small leak is something that the system could not detect and thus not shutting down the system. I personnally recommend changing the septum every analysis. You elimination cross contamination between analysis and small leaks.....yeah i observed it many times.

Everyone knows that 6890 from agilent using a EPC valve does not manage the flow at the end of the column nor the split vent and neither the septum purge but it is mesured before entering the system. So depending on how the split vent purge is clogged you will see difference in retention times because it will keep the same total flow. I assume if the split vent purge was partially clog the retention time should be shorter than usual. The readings on the system for the flows differs from the real flow it should be.

Willy the GC