Chromatography Forum

Hello everybody

I'm doing an experiment on a GC-FID Varian 3400 GC. Condition are 150°C injector temperature, 1ml/min carrier gas, 1:50 ml split, column program starting at 50°C and going till 250°C and the detector at 270°C

We are trying to complete a peak table of hexane and octane because those two components are the reference peaks. We mixed the concentration of those two components and try to find everytime the right identification. When this works we can start measuring real samples.

But here something doesn't work. Hexane is always detected as a regular peak but octane sometimes isn't. We found out it is because of the way the octane peak is looking. It has a sloping front face and a sharp back face (I would like to post a image about it but i don't know how) which is called a leading peak.

We checked the manuel but all the solutions they suggest aren't working. It's a too high volume but we tried first 5ul, than 1ul and even 0.5ul.
Another thing is the column temperature. We start at 50°C and go to 250°C so I guess the temperature of the column is not too low. I tried and changed the method to start at 100°C but than the peaks are stil the same (now the hexane peak shows a little bit of the same but starting with a sharp front and a sloping back).
They also state that the carrier velocity could be too low. Like I said it's 1ml/min for the moment and most of the literature states 1ml/min. We tried to increase it but it was actually worse.

Because the solutions of the manual doesn't give any real solution we are desperate to find a real solution because this way it isn't possible to measure anything on a decent way.

Anyone an idea?

Thanks in advance.

Your column phase thickness is too thin. You will have to GREATLY reduce your sample size if you continue to use the column you presently have in your GC. What is your split ratio? and what are the actual column specifications of the column you are using.

Decrease your beta and you will improve the peak shape. The octane plug acts like a phase on itself which is why the peak capacity is increased on the back part of the plug versus the front part of the sample plug. Thus you are seeing a 'bearding' peak, the opposite of a tailing peak.

Changing the flow won't help much, as you have found out. An increased temperature may give the appearance of an improved peak shape but you may lose all separation

Use a 1µm or thicker film and you should see an improved symmetry, especially if you greatly reduce your sample size by increasing your split ratio.

best wishes,

Rod

If you are injecting a mixture of hexane and octane in roughly equal proportions then you have classic concentration overloading - there is just too much analyte for the column to handle. Increasing the loadability of the column (Rod's suggestion) will help, as well as increasing the film thickness you need to go for an increased diameter. The higher the split ratio the less the peaks will overload, but you will need ratios of hundreds to one, which uses a lot of carrier gas and can cause problems with retention time and split ratio stability. Injecting smaller volumes will also help, you will probably need to go to a plunger-in-needle syringe. The repeatability of injection volume will not be as good as with a conventional plunger in barrel syringe.

What you are able to do will depend to an extent on what your real samples are - having hexane and otane as reference peaks suggests that you are analysing for very volatile analytes. Can you tell us what the target analytes and the matrix are ?

Peter

Thanks for the replies. To start I'll answer your questions

Your column phase thickness is too thin. You will have to GREATLY reduce your sample size if you continue to use the column you presently have in your GC. What is your split ratio? and what are the actual column specifications of the column you are using.

Column specifications:
Film thickness: 0,25
Length: 25 m
ID: 0.32 mm
OD: 0.45 mm
Non-polar

Split-ratio: 1/50

Because of your suggestion about an overload, we tried to dilute the n-octane by mixing it with a solvent. We made a micture of 1ml solvent with 10ul octane and injected 1ul of this mixture. So the concentration of octane is very low, and stil the "bearding" peak is present. So i think the problem of overload should be pretty much be gone.

After this some other guys suggested that the liner was dirty so we placed a new one, but still the same problem.

We are going to measure Volatile organics and how they photocatalytic degrade.

But nothing of the things people suggested worked untill now. So do you really think that changing the column will change the problem?

I made a link to one of our spectra. This is 1ul injection of the mixture 1ml ether - 10ul octane. The right peak is the problem peak.
http://www.ua.ac.be/download.aspx?c=bir ... 5&e=155705

Thanks for your advice and help.

You have shown a classic example of column overload in your chromatogram.

Notice your peak at 1.6 minutes is very symmetrical? there is less of it on the column and the temperature used to elute it at the same temperature and retention time as the octane would be very much less than for this injection.

You can likewise improve the octane peak shape by increasing the temperature of the oven.

OR

You can improve the peak shape by increasing the film thickness and/or increasing the the oven temperature.

Of course if you do not increase the oven temperature after increasing the film thickness you will increase your retention time of octane a great deal.

Of course you can dilute your sample 10 or 30 times and try it again.

The peak shape is rather nice. You can also improve your symmetry by using a slightly more polar phase like a 5% Phenyl 95% Methyl silicone or a OV-1301 type phase if you chose to use the same temperature and film thickness.

best wishes,

Rod

With that thin film column I think you are still injecting 10 x too much octane. It's best for trace analysis at nanogram level. Change to a 1.0 um film thickness if possible.

If not then be advised that your retention times will shift significantly when you put more than 30 nanograms on that column. By significant I mean more than +2 RI units on the Kovat system relative to say 1 nanogram. It sounds like you are depending on retention time for identification and your scheme may be compromised with the loads you are putting on.

30 ng at 50:1 split equates to about 0.2 uL of your current 1% v/v mix, however the fronting is pretty bad at 1 uL (160 ng) which makes me think that the split is not really as much as 50:1. When the solvent flash evaporates in the injector strange things can happen to the split. Try 100:1 and 0.2 uL.