Tailing Alcohols - Headspace Analysis

[Schmitty]

Thanks all for the replies,

AICMM,
Yes, I was having good (better) luck with this until I went to 80°C. The compounds I was working on were sort of a pre-updated <467> test for some of our clients. Since we now have the new <467> to work from, I am modifying our method to analyze for the Class 1, 2A and 2B categories. I am not concerned with the delay of the <467> update, since we can offer the analysis to those clients that want it, instead of the normal OVI analysis. So, the 80°C came from the updated <467> and I think a deviation back to 60 would probably put any results into question, because of insufficient heating.

For our clients that want to continue with our current residual solvent profile, I want to improve the peak shapes and reproducibility I am seeing, while moving towards the method modifications necessary to meet the new <467>.

The current sweep time is 2 minutes. The longer the sweep, the further the molecule will travel and the more water will accumulate. Reducing this is next on my list of things to try.

The last chromatogram on the previous page is from a trap analysis. I only mention the headspace analysis in passing. I will again try static sampling with the analytes in our current profile (since the DMF and DMSO have been excluded).

Rod,
That is another good idea with an additional portion of a -wax column. However I would be concerned about the possibility of air interfering with my MS. Is there a good solution to glass press-fit connections? I would assume that I should start with ~1m and increase if necessary. Or should I start with 5m right away?

I also think I forgot to mention that the above chromats (one marginal, one poor) are from the same sequence, just separated by 10 samples.

Thanks,
Schmitty

[chromatographer1]

Besides the press fit unions there are unions made of fused silica coated metal (by a competitor company who will remain nameless ) which I have used with graphite ferrules quite successfully in this application.

One can also use polyimide resin to seal the fused silica connections as one would do with a Supercritical fluid capillary column attached to a restrictor frit. Supelco sells the resin.

I would use a 3 meter piece of Supelcowax 1µm film 0.53mm ID as this will move 1,2-DCE away from benzene and put ethanol behind ether, both of which are valuable modifications of the selectivity of the G43 phase. Pyridine is also well separated from toluene with this combination.

best wishes,

Rod

[Schmitty]

My initial attempt for analysis without water was very positive. The peak shape was much improved and symmetrical, however the alcohols were a bit wide. This is most likely attributed to the 2 minute sweep time.

I have an automated method modification run going now that will ramp up the sweep time. We'll see how that looks tomorrow.

Rod,
I found something in the Restek catalog for a coated metal union. I don't even see glass cap unions in the Supelco catalog (I have some on hand anyway). Product 23817 is what you suggest if I use the press-fit?

[chromatographer1]

Schmitty

That is correct, 23817, is 5g of polyimide sealing resin, page 290 of the 2007-8 Supelco Catalog.

On the same page are 23627 (5ea) or 23628 (25ea) of the glass tapered fit butt connectors for FSOT columns.

If weight (stress) on your columns is a critical issue then by all means go with the glass tapered fit. I used to prepare my own SFC columns. It took a little practise but I learned

'Y' connectors are also available on the same page if needed.

best wishes,

Rod

[Schmitty]

Well, the multiple methods with varying sweep time did not produce any change in the peak shape or width. The only change was in response for each of the analytes, i.e. less sweep=smaller peaks. I don't recall if I mentioned this before, but the response for each of the analytes is not usually consistent between multiple injections of the same level standard. I would suspect the conditioning of the trap for this problem.

I did try a loop method (1mL loop) and achieved acceptable results for the majority of the analytes. The peak shape is not as poor as the trap method, but not as good as a liquid injection. I increased the transfer line temperatures, again with no change. Also, the split was reduced from 60:1 to 20:1 and further to 10:1, without much change in peak shape, and little increase in signal. The reproducibility of multiple injections is excellent for this loop method.

As before I think I will jinx myself and say that this is good enough for now (again).

[chromatographer1]

It appears that performing a dynamic trapping of alcohols with a large amount of water does not allow a good tight focusing of the alcohols in this trap.

As trapping without water showed improved peak shape I would expect this carbon trap would be more effective for non-polar analytes and less effective for polar analytes. Unless you refocus the purge plug by some manner you will have broad peaks for the alcohols as they won't be trapped as a tight plug on the trap but as a broad plug, and that is how they will be transferred to the column.

Perhaps static headspace will be a better solution for you. You should be able to see less than 1ppm of methanol in water by static headspace and orders of magnitude lower if you use SPME headspace.

I published a static headspace chromatogram showing 0.5ng of methanol in 25 mg of water in 1997 (20 ppb). You should be able to get the same from 250ng of sample in most headspace analyzers today.

Good luck, Schmitty.

Rod

[Schmitty]

Yeah, I haven't found the time to add that short wax column as a guard in front of the -624. That might be more ideal, but it is probably still pretty rough on the detector to have so much water vapor traveling through it.

I think the issue I would run into doing the small sample volume would be in the preparation of standards and the various matrices we would need to cover. We are trying to follow the USP as much as is reasonable, and having that water/DMSO matrix has significant importance to their method.

Do you recall your injection volume for your methanol in water presentation? Also the vial size (2/6/10/20mL)?

This is a screen method, both for what we are trying to provide to our clients, and what the USP would like to implement (next year). So SPME should nicely cover the qualitative aspects of the requirement. But you have to be concerned with how many steps away from the method you choose to take. In my case, I am using an MS, and mixing the compounds.

Also regarding the SPME. We have several of the CTC combipals with the SPME attachment, however I have not been very pleased with the operation of the instrument. Bending needles and breaking fibers is not the way to my heart

[AICMM]

Rod,

If I understand Schmitty's dilemna though, SPME is not a 467 technique? Your point about re-focus is well taken though and the suggestion of a polar "guard" column still seems a potential solution to re-focus the polars.

Was this tried Schmitty?

Best regards.

[Schmitty]

Yeah, I haven't found the time to add that short wax column as a guard in front of the -624. That might be more ideal, but it is probably still pretty rough on the detector to have so much water vapor traveling through it.

Not yet. I don't like venting the instuments so very often.

[chromatographer1]

Schmitty,

My injection loop was 1mL The vial was a prototype 3.35mL, now reduced by PE to 2mL I believe. But even a 6mL might be possible.

You will need to heat the vial for 8-10 minutes for maximum recovery of alcohols at 85°C.

The flow rate was 20cc/min ( a problem using MS).

I used a 3 meter G16 column followed by a 30m G43 column followed by a 5 meter 0.25mm ID piece of FSOT.

Concerning SPME, you might wish to ask Technical Service or R+D Scientist Robert Shirey at Supelco to assist you with the SPME application. He has already consulted with large pharmaceutical companies concerning your application. With the new flexible metal SPME fibers and Merlin Microseal inlets, breakage and bent needles should be history.

Tech Service 800-359-3041

Rod

[chromatographer1]

I wrote:

I published a static headspace chromatogram showing 0.5ng of methanol in 25 mg of water in 1997 (20 ppb). You should be able to get the same from 250ng of sample in most headspace analyzers today.

That should be 250 mg (250µL) of sample in most headspace analyzers today.


SORRY.

Rod

[AICMM]

Scmitty,

In my opinion, you don't have to vent to add a pre-column. Your column acts like a big restrictor on air flow into the MS, especially a 0.25 id. Environmetal analysts do front end maintenance on the column and inlet all the time with- out venting. Pull the column out of the injector and put a septum on the end. Then install the pre-column in the inlet along with the union and then re-connect the column to the union. The time the MS is open should be minimal and as long as the filament is off the recovery time should be brief.

Best regards.

[chromatographer1]

Schmitty,

If the trap is giving you a broad plug for the alcohols then even putting a precolumn of Cwax won't be of much help, the only help will be to trap for a shorter amount of time which will minimize the plug spreading and will sharpen your peaks.

If this affects your limit of detection, well, that will have to be determined.

best wishes,

Rod

[Schmitty]

AICMM,

True enough and good points. I have done septum replacement and liner replacement without venting.

Rod,

I did try altering the trap times. I did 0.25 minute increments from 0.25 up to 2 minutes. All were very similar in peak shape, but the higher boilers were not nearly sensitive enough at the shorter times, if I recall correctly. Perhaps the "K" trap is not best for this mix of analytes.

Thanks,
Schmitty