Tailing Alcohols - Headspace Analysis

[Schmitty]

Hi everyone,

I am having an issue with light alcohols (MeOH, EtOH, IPA, and to a lesser extent Acetone), where I have poor sensitivity and poor peak shape. The detector is MS, but I think the detection type is irrelevant in this instance and would rather have more input from those with more knowledge about GC.

The instruments are: a Tekmar HT3 (in dynamic mode), 6890 GC, 5973 MS. The phase in trap is Supelco "K", and desorbed at 220°C. GC Column is a 60x0.25, 1.4µ DB-624. Temp prog is 40°C for 1 minute, then 1°C/min to 60°C, 10° to 120, 15° to 240, hold at 240 for 5. Typical volatile program.
The peak shape for all components is great except for those listed above. The liner was new when I istalled the HT3 and new column. I have tried changing the liner without success (it really has no purpose because of the HT3 anyway).

I also reduced the desorbing time on the HT3 from 1 minute to 0.2 minutes without any improvement.

The samples are water, and the desired concentrations of each of my standards is from 0.05 to 50 ppm) Essentially I have everything working except the methanol, ethanol and isopropanol. They seem to be "sticking" somewhere in the system. Let me know if you need more information about the system to make a diagnosis, however I do not want to give away too much.

Thanks!

Example:


edit: added more details

[JI2002]

For K trap, I would use parameters recommended by Supelco:

Preheat: 245C
Desorb temp: 250C
Bake temp: 260C

Another thing to consider is split ratio which affects carrier gas flow rate during desorb. If you have a split ratio 50:1 and column flow rate of 1 ml/min, that gives you a total desorb flow of about 53 ml/min. If the flow is less than 10 ml/min, you will have peak shape issue.

other than that, check the heated zones to make sure they are at their set points and no cold spot. Also, are you using dry purge?

[chromatographer1]

It looks like you may have a cold spot in your transfer line into your GC injector port. Do you have a piece of cold metal in the sample flow path?

While increasing the temperature from 220 to 250°C might help, I suspect that may not.

best wishes,

Rod

[Schmitty]

For K trap, I would use parameters recommended by Supelco:

Preheat: 245C
Desorb temp: 250C
Bake temp: 260C

Another thing to consider is split ratio which affects carrier gas flow rate during desorb. If you have a split ratio 50:1 and column flow rate of 1 ml/min, that gives you a total desorb flow of about 53 ml/min. If the flow is less than 10 ml/min, you will have peak shape issue.

other than that, check the heated zones to make sure they are at their set points and no cold spot. Also, are you using dry purge?

My split is 60:1. All semi-heated zones (from end of heating tape to injection port) are wrapped in Al foil, but I will double check and double-insulate. I am dry purging at 50mL/min at 25°C for 2 minutes. I will try the higher temps.

Thanks

[Peter Apps]

Hi Schmitty

Apart from the dramatic tailing your peaks are a good size, and that is with a 60:1 split. Given that you have sensitivity to spare, why not try simple equilibrium mheadspace, at least it will tell you whether the tailing is coming from the trp or from the rest of the sytem.

Peter

[Schmitty]

I tried a static injection for you, Peter. However my higher boilers (DMF, DMSO) are not present (as expected). Here is an example of the progress I've made. The things that I've changed are:
-Decreasing the desorb time from 1 min to 0.2 min.
-Increasing desorb preheat to 240°C.
-Increasing desorb temp to 245°C.
-Doubled the insulation at the end of the transfer line.


I think the results are much improved, and are probably "good enough" for now.

[Peter Apps]

Hi Schmitty

That's what I call moving the goal posts ! from light alcohols to DMF and DMSO How did the alcohols look with static HS ?

I am not familiar with the HT3, but the stepped peak shapes suggest that there is more going on than simple adsorption. My guess would be uneven heating of the trap or its adjacent plumbing, so that the alcohols are coming off in two bands. Does the trap load and desorb with the gas flowing in the same direction (bad) or is the gas flow through the trap reversed for desorption ?. Is the plumbing, including the transfer line, deactivated (Silcosteel or similar). How is the packing held in the trap, glass wool ? and if so is it deactivated.

Reducing the column start temperature might help (but with a cost in cycle time), as might holding the column for longer at the initial temperature.

Peter

[Schmitty]

Hi Peter,

The HT3 loads the purged samples in one direction, then reverses flow for the desorbsion. The peak shape for the static HS was similar to the above dynamic TIC (stepped).

I tried bumping the transfer line temp from 105 up to 150, without any change. I would think that the HT3 would be constructed from the materials that would be best suited for the job at hand.

The Supelco "K" trap should be made to Supelco's specifications. I could not find the construction in the '05-'06 catalog, however.

Now I am trying to ramp the sample temperature through successive runs, in order to increase the relative sensitivity of the DMF and DMSO to the other compounds.

[Peter Apps]

Hi Schmitty

If static headspace gave the same stepped peaks the problem is clearly not with the trap.

The ether peak is also stepped, which points away from silanol active sites as the (sole) cause of the problem.

There might be a problem with gas flows - at a 60:1 split you are trying to push at least 60 ml/min of gas down the transfer line. Depending on how you have the pneumatics plumbed in, and what GC you have, there might be some pressure or flow disruptions as the multiport valve switches. Does changing the split ratio make any difference ? - you will proably have to change it quite dramatically, say reduce to 10:1.

For reasons of cost and profit margins instrument manufacturers do not always put top end technology into their hardware. Often things like deactivated transfer lines and internal plumbing come as an option (with a higher cost), so you might not have a deactivated flow path. When doing aqueoues ethanol and minor voaltiles in beverages on an old Varian Genesys, which was a Tekmar clone, I threaded a length of deactivated fused silica through the transfer line and attached it direct to the 6-port valve to get an inert flow path.

Peter

[Schmitty]

Regarding the split of 60:1: The 6890 does not seem to like it if I drop the flow too low. It will error and stop my sequence. I can try lower, possibly 20:1, and see what happens.

Thanks Peter

[JI2002]

Make a direct injetion into the GC injector port and see how the peaks look, by doing that, you can isolate the problem. If you get same peak shapes, GC is the root cause, otherwise look into HT3.

[Peter Apps]

Hi Schmitty

Strange, we have a 6890 that complains if we set split flows too high, seems that they have minds of their own.

Two other things occurred to me - if you have the end of the transfer line threaded through the GC inlet septum, where is the tip of the line in relation to the tip of the column ?. And what is the inlet temperature ?

Peter

[JI2002]

I'm not sure what the column flow rate is, but if it's about 1 ml/min, and GC stops at split ratio of 10:1, it clearly indicates a leak or clog in the carrier gas flow path.(from GC to HT3, lines and valves inside HT3 and from HT3 back to GC injector port)

[Schmitty]

A direct injection of 0.2 µL of my mixed standard in dimethylacetamide produces excellent peaks.
The injection port temperature is 250°C.
Column flow is 1mL/min.
The transfer line is not threaded into anything. The weldment going to the inlet was cut, and taken off to go to the HT3. The return line to the inlet goes back into the inlet through the previously cut line. This is where we have the bare tubing, that is now wrapped in 2 layers of Al foil. There is about 4-5" total, with an inch or so under the tower, with no insulation. I think it is all 1/16" SS.

I will talk with one of our "GC guys" about testing for possible leaks in the transfer line. I would not suspect anything, though, because the MS tune shows very little air. The HT3 has a self test for leaks, that passes without issue.

[chromatographer1]

From what I understand from your description you could very well be in need of insulation around your lines. Foil is good but you need better.

Try and find some thermal insulation and wrap your lines or 'cover' them thoroughly and see if things don't improve for you.

best wishes,

Rod