Analysis Conditions

[sadsal123]

Can anyone help please: I am running a temperature program to seperate 22 solvents. These are the conditions:

Instrument: Clarus 500 Perkin Elmer AutoSystem GC with built in Autosampler

Injection Method: Auto Injection.
Syringe Capacity: 0.5ul
Injection Volume: 0.3ul
Injection Speed: FAST

Sampling Rate: 6.25pts/s
Inlet A: CAP
Detector: FID

Pre-injection Solvent Washes= 0
Pre-injection Sample Washes = 3
Post-injection Solvent Washes = 0

Sample Pumps: 10
Viscosity Delay: 0

Injection Temp: 200oC
Carrier Gas: Nitrogen at 0.5ml/min
Split Ration: 250:1
Column = Elite 624 (6% Cyanopropylphenyl -94% Dimethyl Polysiloxane), 30m, diameter 250um

Detector Temp: 300oC
Attenuation: -2

Oven Program:
Initial Temp: 30o for 17mins
Ramp 1: 8oC/min to 140oC and hold for 4.25mins
Ramp 2: 25oC/min to 180oC and hold for 1.4mins

After some attempts, I found that o- & p- Xylenes cannot be separated but this is not such a big issue for us as this is not absolutely necessary to achieve. However, I have THF and MEK eluting very close together and I can only achieve decent separation if I set the initial temp to 30oC and hold for 17mins.

However, I dont have anything in place to cool down the GC to 30oC so on really hot days, I cant get it to cool down to that temp. I tried to start at 40 and hold for longer but I still cannot achieve a decent separation. I have also tried to reduce the ramp rate to from 8 to 5 but thats not worked either. Everything comes off before the second ramp. Thats just in place to make sure that everything has eluted.

Has anyone got any other suggestions? Pete? By the way, my 21 components are all solvents because we work with waste solvents.

The obvious ones are to change column type and install a cooler but I dont want to do either if I can help it as I want to avoid costs. Can anyone suggest anything with the method?

Thank you all in advance

[Peter Apps]

About the only thing that you can do (except get an air conditioner ) is to use a column with a thicker film. You do not give the film thickness but I would guess that is is 0.25 micron or less from the elution times.

Peter

[sadsal123]

the film thickness is 1.4um. So nothing I can do with the method then?

[chromatographer1]

Care must be taken concerning film thickness. You may find the selectivity may change with a thicker film, although 1.4 micron on a 0.25mm ID tube is pretty thick to begin with.

p- and o- isomers of xylene usually separate quite well on almost any phase. Perhaps you meant m- instead of o- ? Do you have your peaks labeled correctly?

Also you may lose some separation using a thicker film and then other solvent separations may decrease as well.

A beaker of ice in the oven can work wonders in getting the oven temperature down below 35°C. Remember things go better with Coca-cola. (Coca-cola is optional in the oven)

AC in the lab is also a good idea. In the summer I have problems getting my ovens down to or below 30°C but I can reach 27°C consistently with the ice in the oven trick.

best wishes,

ROd

[Peter Apps]

How close is "very close", and how wide are the peaks themselves, or in other words what is the resolution that you get at the moment ? Could you post a chromatogram ?

Ice in the oven will do the cooling, but is difficult to automate . I've never even seen a Clarus GC, but its worth checking that the hot air coming out of the oven when it cools does not go back into the cool air inlet. I have had this problem with a GC standing a bit close to the wall, and with GCs back to back on a bench. With one of them a baffle to redirect the hot air gave me 10 degrees of extra usable cooling.

Dropping the detector temperature to 250 will also help. You could even decrease the final programme temperature and decrease detector temperature even further.

Good luck Peter

[MikeD]

I don't have a solution with your present setup - if you continue to use a non-selective detector. The point is that with 22+ components the probability of no co-eleutions on a single column is low. The probability can be calculated given the column resolution and N peaks although I can't now find the reference.

We used to have a similar problem to you, but with 60+ solvents. We used a split to two columns in parallel and dual FIDs. One column was like your 624 phase and the other was methylsilicone. There were no cases where an identification could not be made on one or other phase. The job was impossible on a single phase with FID, but we don't use dual column setups anymore. It's all MS detection.

[JI2002]

How about change carrier gas to helium or hydrogen?

[chromatographer1]

Peter has good suggestions worth trying.

I have never seen co-elutions of p- and o- xylene except on certain packed carbon support columns. Even short MS and Cwax capillaries will separate these two analytes. (search a host of web sites that offer chromatograms)

And to let you know, the small amount of cooling from the ice will bring the oven temperature down and the controlling thermostat will add enough heat to maintain the temperature long enough to hold a lower isothermal temperature ONCE. As the temperature ramp proceeds the ice will all melt and require replacement for the next injection.

Not a great solution but it has worked for me in the past. (not a bucket of ice, but a small beaker)

Sometimes necessity is the mother of invention.

Please check your peak identifications and let us know how it works out.

best wishes,

Rod

[Bruce Hamilton]

Lots of good suggestions already. The Ice / coke cola trick really works, as does squirting liquid CO2 into th ovrn for a minute. Don't use a fire extinguisher - use a small noozle that sprqays the liquid.

Sticking some alumina wool under the injector and detector also helps - just secure with some thick metal foil. Such insulation can also help instruments where the injector base and fittings are directly in the oven. I've found it gives more consistent injections when running at low oven temperatures ( below 25C ), but YMMV.

Provided you have the sensistivity, and acknowledging that you already have a relatively high split ratio, I'd look to reduce the amount you are injecting to improve resolution of critical peak pairs, perhaps even trying diluting the samples with a solvent you're not interested in.

You are unlikely to fully resolve the meta-para xylene pair, unless there are very small amounts injected. Not all 624 columns are equal, I've seen major drifts of aromatic RRT between brands, but not improved resolution of m-p xylene.

You seem to have a good handle on your options, so I'd keep playing tunes using all the good advice you've got.

Please keep having fun,

Bruce Hamilton

[sadsal123]

Thank you everyone, some of these suggestions are great and I will definitely try them, and yes Peter, the back of my GC is very close to the wall so I will try shifting it a little and will also try to post on a chromatogram.

Till then, thank you all once again.....

Salma