USP 467 HS method

[Mike H.]

I having tailing problems when using the USP 467 separation using their parameters.
I'm trying to separate MeOH, IPA,MeCN and Pyridine using DMSO as the solvent. The polar components MeOH and MeCN are tailing in excess of 2.5. I've cut and reset the column (Brand new RTX-624 30m,0.53mm id, 3µm). What else should I try. I've follow the USP totally and can't figure out why my chromatogram is crap and theirs isn't.

[chromatographer1]

Do you have uncoated bare metal in your sample path?, ie: needle, sample loop, transfer line?

Are you comparing their chromatography using water as a diluent and yours using DMSO?

Do you realize that no one is forbidden to validate their own method, and the USP method is meant to settle disputes between parties that disagree in their results each using their own method, not to be the best possible method for all parties at all times?

best wishes,

Rod

[Mike H.]

Do you have uncoated bare metal in your sample path?, ie: needle, sample loop, transfer line? Not 100% sure but we have good instruments Agilent 6890 and a PM service contract with them. If I had to guess I'd say no to this.
Are you comparing their chromatography using water as a diluent and yours using DMSO? Yes. I also tried the prep in DMAC with the same result.

Do you realize that no one is forbidden to validate their own method, and the USP method is meant to settle disputes between parties that disagree in their results each using their own method, not to be the best possible method for all parties at all times? Yes but it was bothering me that I could reproduce their results after checking and double checking everything I could thing of. I'm trying a wax column now. What column would you select if you were trying to resolve these componds? thx
best wishes,

Rod

[chromatographer1]

I do not believe the Agilent HS unit has deactivated components. They did not at one time, but I have not checked lately to know what they provide as basic equipment. I would suggest that if you do not know you do have fused silica components, then you don't.

Again, using a non-polar solvent in a bare metal environment is quite different from using water as a solvent. Do you know the temperatures at which you are trying to perform this analysis should be different depending upon the diluent used. DMAc, DMF, DMSO, and other non-aqueous require higher temperatures. Using the USP parameters designed for use with water diluent will produce different results when attempted using non-aqueous diluents, and so you have discovered. The RTX-624 column should perform the test well. You might use a short wax column as a precolumn.

http://www.restek.com/pdfs/PHFL1018A.pdf shows the USP test using DMSO as a diluent.

The problem lies before the column, not in the column. You need professional help and I hope your company will provide it. Take consolation in that you are not the first to have problems with the USP test procedure for residual solvents. Search the archives here and you will many posts on this subject.

best wishes,

Rod

[krickos]

agree with rod, but to cut to some basics, default settings in USP an EP do not suit well wit agilent HS/GC basiclly as temp settings in valve/needle/transferline in general is set too low. So needle/valve/ transferline should have a temp above bp of DMSO to start with.

EP recognize this issue more clearly than USP currently.

[Mike H.]

Why does the transfer line have to be set above boiling poing of DMSO when I'm not assaying DMSO? (I don't have a good fundamental knowledge of headspace technique and have been searching for a good online resource to learn it.) I was under the impression that the loop temps and the transfer line temp setting are set with the analytes in mind and not the solvent. Is this true? My highest boiling point component is pyridine and it's BP is ~115C. Can someone please verify that the transfer line temp has to be > solvent boiling point? thx

[chromatographer1]

Mike,

If you have bare metal in your lines the problem with DMSO, DMF, DMAc, etc is that a residual amount remains, coating the lines and affecting the separation and retention times of the analytes. Essentially, it behaves as a coated column.

If you have fused silica coated or glass coated lines then depending upon the time between injections and the flow through the lines this problem is avoided.

Usually I have found that temperatures of 110C to 120C are more than adequate to clear the lines of residual heavy carrier solvents without causing deposits of chemical reactions from temperature when using fused silica coated lines.

Your hardware will determine the minimum temperature you require. Of course, higher is always better as long as you do not degrade your analytes, or cause artifacts in your analysis.

I hope this is helpful.

best wishes,

Rod

[Mike H.]

I'm using an Agilent G1888 Network Headspace Sampler connected to an Agilent HP6890 GC. I've never thought to ask what type connections metal or silica lined the headspace has. How do I determine this? The only line I can see is from the transfer line to the GC inlet. It is metal, but I can't tell if it's silica lined or not. How can I verify this?

[chromatographer1]

you have a manual?

Rod

[Mike H.]

Someone has it. I called the vendor, don't know why I didn't think of this first?, and found out it was silica steel so it's coated. I increased loop temp to 130°C and transfer line temp to 140°C. The separation looks good. The retention times are stable, but the only thing that's needs to improve is the precision of the pyridine peak. I'm resolving MTBE,MeOH,ACN,IPA and Pyridine. They elute in this order and my precision (6 injs) for MTBE,MeOH,ACN,IPA are<1% RSD but the Pyridine precision is ~7% RSD. I'm guessing I need to increase the respective temps another 10°C? I'll try this next unless you have another suggestion.

One additional question I have is that is there a maximum rule of thumb for loop temps and transfer line temp with respect to the boiling point of an analyte? I'm not sure about this but am curious to find out if there is a temp ceiling that if we exceed will cause precision problems for the lower boiling point compunds?

[chromatographer1]

I am guessing but are you using a needle through septum at the injector?

When measuring pyr I used a union from the transfer line to the column and got 2-3% RSD.

Pyr is a good test probe and is greatly affected by dirt in the flow path. Rotor composition also effects RSD.

110C was always enough for my inert system, a Varian Genesis (Tekmar 7000)

best wishes

Rod

[krickos]

Hi

As I recall Agilent stopped putting Ni as default material in needle/loop/transferline a few years ago in favour of silcosteel which we always had to "add on" previously.

I have used similar temp settings for pyridine however in DMAC/DMAA instead of DMSO.I Avoid DMSO whenever possible due to the high bp and potential contamination of the Agilent HS.

It is possible that you still suffer from residual DMSO contamination or you need to raise temps further closer to bp of DMSO. You can always try exchanging needle and loop first (can be cleaned) or steam cleaning.
As Rod stated pyridine but usually also butanol/pentanol are good indicators on active sites/contamination/too low temp settings.

Think there is a max setting of about 200°C, have not used that but been around 170 without issues with initial column temp at 35-40°C for typically early eluting residual solvents.

[Mike H.]

I finally jacked up the loop and transfer line temps to 180°C. Precision for pyridine on the next attempt was 3.5%. It took about a day before the blanks were clean enough to run though. I have a couple addition questions to which I couldn't find the answers from google.

#1. I downloaded a copy of Agilent's steam cleaning procedure. It states that the HS transfer line needs to be disconnected before the procedure is followed, but it seems like it was written for use with a MS unit. I just have a regular 6890 agilent GC using a DB-Wax column. Will this column be damaged from the water vapor if I didn't disconnect the HS transfer line?

#2. If the transfer line is disconnected from the GC, will the GC sense a leak condition and fault out? If so how do I avoid this?

#3. Is it ok for the headspace oven temp to exceed one or more of the boiling points of the analytes to be tested? I'm unclear about this. In my assay there are 5 components who's boiling points range from ~80°C-115°C. What oven temp do you guys suggest? What is the rule of thumb?

It's tough being a novice and I appreciate the help. thx

[krickos]

I finally jacked up the loop and transfer line temps to 180°C. Precision for pyridine on the next attempt was 3.5%. It took about a day before the blanks were clean enough to run though. I have a couple addition questions to which I couldn't find the answers from google.

#1. I downloaded a copy of Agilent's steam cleaning procedure. It states that the HS transfer line needs to be disconnected before the procedure is followed, but it seems like it was written for use with a MS unit. I just have a regular 6890 agilent GC using a DB-Wax column. Will this column be damaged from the water vapor if I didn't disconnect the HS transfer line?

#2. If the transfer line is disconnected from the GC, will the GC sense a leak condition and fault out? If so how do I avoid this?

#3. Is it ok for the headspace oven temp to exceed one or more of the boiling points of the analytes to be tested? I'm unclear about this. In my assay there are 5 components who's boiling points range from ~80°C-115°C. What oven temp do you guys suggest? What is the rule of thumb?

It's tough being a novice and I appreciate the help. thx

Hi
Thinks its clear that you had residual DMSO residues in your system based on the number of blank runs, seen that before. To monitor steam clean effectivness I never used a WAX column more sensitive than others(always did dedicate for DMSO/DMAA/DMF use or water), typically used a like a DB-5 or similar that is less sensitive starting at 100°C in column oven. Even a retention gap would prob do.

1. Yes thats for MS detector. When I had a MSD to work with we always had a FID as well as 2nd detector so we just shifted detector to monitor steam clean effectivness.

2. I prefer to monitor the steamclean effectivness so not a problem if you keep it connected to FID.

3. Absolutely!. The point is that the gas sample should ideally NOT interact with anything in the sample path way until it reaches the column. A good rule is to kept needle/valve/loop at +10°C above the highest bp of analytes including the sample diluent (DMF/DMSO/DMAA....). PArticulary for HS like Agilent that has a valve/loop injection system.

Finally, it can be adviseble to consider regular steamcleaning based on your sample throughput as a preventive maintenance schedule.