USP <467> on Agilent G1888

[GregK]

Has anyone gotten this to work?

I have to admit, I'm very new to headspace analysis. My boss insisted that we purchase one because we had to have for residual solvents. It has been sitting on the bench for close to a year now. I developed a method for testing residual solvents in another API, using direct injection, but it won't work for the API I'm working with now.

My problem: I don't know what values to set for all the conditions not specifically stated in the USP (loop fill, loop equilibration, etc.). All the references I can find on the 'net are using different conditions than the USP all together. We need to stay within USP acceptable limits to save time by not validating another method.

Any advice, comments? We've got a 3 day weekend coming up, so I've got some time to think about what experiments to run when we get back...

Thanks for any and all help.

-Greg K.

[chromatographer1]

Greg,

Depending upon the port sizes in your sampling valve the loop fill time will depend upon the size of your vial and the amount of pressure with which it is subjected. Normally, a minimum of three volumes of your sample is required to properly flush and fill your loop. One can decide to flush more, but it is an operator determined variable. Don't forget the dead volume ahead of the sampling valve, following the vial, as this volume must also be flushed through the sample loop to vent.

The loop equilibration time allows any pressure in the sampling line to return to atmospheric pressure (no bubbles from the exit line of the sample path).

These are best determined experimentally as no two HS analyzers may be configured exactly the same.

These variable factors are why they are not included in the Chapter's directions.

If you ever decide you could save money by hiring a knowledgeable consultant, I know of one who could give you excellent training and one who has good information for optimizing your analysis.

best wishes,

Rodney George
(ahem.....) consultant

[krickos]

Hi

as a starting point these settings usually works for us, regardless if 7694 or G1888 model for a 0,32mm id, 30m column of USP/EP standard for residual solvents, split ratio 1:5-10.

Lopp fill time: 0,2-0,40 min, below 0,17min time is usulally too short

Loop equilibration (done before injection): In fact I tend dto keep this one really short like 0,05min as it is a risk for smaple intecation with loop walls.

Injection time: Here I always use a bit more then necessary like 0,5-1min. I f you use standard 1ml loop (guess so as it is USP) it is easy to see how many times the loop is flushed based on total flow (split+column) however this depends a bit on your set up, works fine with volatile inlet and EPCs.

Vial pressurization time: well in short do not approve of the general USP/EP settings, feels like a "sample" diluter , might depend on different HS models but start with USP setting.

Agilent and USP/EP settings do not really mix well when it comes to loop/transfer line temperatures (USP/EP settings seems based on PE 40HS models). Especially if using solvents like DMF as sample solvent or testing for high boiling polar stuff. So here I strongly recommend that you set the loop temp 10°C higher than the boiling point of your sample solvent/highest boiling residual solvent (whichever highest), if not you sooner or later will build up residues in the sample loop, causing problems seen often in this forum. Transfer line 10 degrees above loop.


Good luck

[chromatographer1]

krickos made an excellent post since it appears he has the same equipment as GregK.

The concern with loop interactions is valid, especially if the loop is not fused silica coated, a change in hardware that is highly recommended.

The increase of transfer line of 10°C above the loop temperature is also highly recommended.

Loop equilibration being short is fine as long as the fill time is long enough to allow the pressure to come to equilibration, the purpose for the loop equilibration time setting. If the loop fill is long enough, then pressure equilibration is achieved with no loop equilibration time required.

Having the injection time even as long as the run time is OK, because all you will be doing is allowing the loop to be flushed with clean carrier gas until it is time to fill the loop for the next injection.

One point should be made clear to you. Unless you require very few methods for HS analysis, you will be money and time ahead if you design a method that will work for a large number of solvents and for a select number of dissolution solvents. This can be validated in less than a week of overnight runs and a day or two of writing the validation. Then the only additional work needed for a new drug is to show that the drug matrix when added to the dissolution solvent does not affect the linearity and repeatability of the results.

I have done this 'generic method' validation for a very large and reputable pharmaceutical company and it has saved them large amounts of time and money when residual solvent analysis was required for a new drug, using either timed injection or sample loop headspace analyzers.

The USP generic method is not always kind to laboratory personnel who need results quickly.

I have completed a validation report for a new drug residual solvent method in less than a day, once the generic method had been developed and documented.

best wishes,

Rodney George
consultant

[krickos]

I have done this 'generic method' validation for a very large and reputable pharmaceutical company and it has saved them large amounts of time and money when residual solvent analysis was required for a new drug, using either timed injection or sample loop headspace analyzers.

The USP generic method is not always kind to laboratory personnel who need results quickly.

I have completed a validation report for a new drug residual solvent method in less than a day, once the generic method had been developed and documented.

Indeed, a very good point there by Rodney. Seen the same thing as well. Worth the time if you are running residual solvents frequently.

To Clarify a few points. USP 33 (finally!) as of 1st supplement actually allows you to increase, tranferline-syringe/loop temperatures and adjust vial pressurizations if needed. Also note that procedure A-B is more of a limit/identity test while procedure C is refered to for quantitative purposes.

[jane bench]

Hi, I hope somebody can help!
Vial and carrier pressure programing and also vial pressurisation time working residual solvent analysis with GC 6850 and HSS G1888 gives me quite a headacke. Does anyone know some general rule about setting this parameters on because I ussualy have situation that injection is done but aux 4 (vial pressure) is not ready and it takes long about 1 min. Should I set injection time to be longer or something else ?

[krickos]

Hi, I hope somebody can help!
Vial and carrier pressure programing and also vial pressurisation time working residual solvent analysis with GC 6850 and HSS G1888 gives me quite a headacke. Does anyone know some general rule about setting this parameters on because I ussualy have situation that injection is done but aux 4 (vial pressure) is not ready and it takes long about 1 min. Should I set injection time to be longer or something else ?

Hi Jane

Think we need more details around the not ready message. When does it appear in the log ie what is the instrument doing at the moment? What is your setting in kPa or psi?
It can be "normal" to get a breif error/not ready message in conjuction with pressurization of vial as the EPCs notice the start of vial pressurization, however 1 min seems too long. An old rule around my place for Agilent Headspace samplers was to set vial pressure to initial column pressure+100kPa (example 70+100=170kPa).

[GregK]

Thanks everyone. My home computer died, so I was unable to reply over the holiday...

chrom1, I did create a generic method, as you suggested last time I was trying to figure this stuff out. Unfortunately, the generic method does not work with this API. I'd have to change too many variables to get it to work, so I figured I'd just use the USP and skip validation again. (Report is due by the end of the week, I was told last week and then not able to use the GC until this week.)

I'll try krickos's suggestions above regarding fill and equilibration times and try and report back.

Thanks again.

-Greg

[chromatographer1]

Greg,

Your comment is interesting about your API and your Generic method.

It sounds like there was not enough work done on the Generic method, thus my questions:

Why does it not work?

Non consistent recovery?

Dissolution issues?

Can you elaborate?

Rodney George

[GregK]

Well, really there are two problems. First, I need to test this API for chloroform and I could not get a good response for chloroform in my generic method. Second, when this API is dissolved in DMSO (used in my generic), it gives a couple large peaks in the GC run, effectively swamping out the solvent peaks. I should have validated the generic with DMSO and water, and I could maybe have used it for this API... Maybe I still can...

Another thing: the generic is built around direct injection. The boss ordered and fought to get the HSS because "we have to have it for residual solvents." To save face, I'd like to have at least one residual solvent method that actually uses the HSS.

BTW, just figured out that DMF is in the API, so I'll have to use something other than just Class 2 A & B mixes anyway...

[chromatographer1]

When I spoke of the generic method I meant using Headspace, not direct injection. So I guess you misunderstood.

Your DMSO is not pure. You need quality solvents to perform any analysis.

Since you have a non-polar solvent with a high boiling point (DMF) you will need DMAc or some other higher boiling point solvent to dissolve your API like 1-methyl-2-pyrrolidinone.

I hope you get this report done. There are technical consultants (other than myself) who might save your company time and money. Most refuse outside help and would rather waste time and money than have someone from outside assist them.

Good luck, you are about to discover that experience IS valuable and that reinventing the wheel costs a lot more than your boss realizes.

Rodney George

[GregK]

I assumed that you meant headspace last time, but the pressure was on to get a method that worked. We had a copy of the method that the manufacturer used to quantify the residual solvent, so I expanded on it a little and added some other solvents to the standard mix. The DMSO blanks gave clean chromatography, so the solvent was/is pure (enough, anyway). I tried to use it for some of the other APIs, but it gave bad chromatography for this particular one (two others were okay).

I'd love to hire someone in to do this work, but that is beyond my control (and my boss's, for that matter). I do know that experience is valuable, that is why I come to this message board. I've tried giving advice where I can, but I'm new and learning. I really appreciate the help that people give here.

I'm not trying to reinvent the wheel, I'm trying to follow the USP...

All that being said, I probably won't get the report done. My boss will take the heat, and explain that if I wasn't working on three other projects, I could have gotten this one done.

I'm really just trying to understand our HSS better. I don't understand why I can't decent sized peaks. The USP standard preps are giving me tiny tiny peaks for the system suitability injections. And standards that I prepared at the USP's standard concentrations give me little to no detectable peaks.

Should I increase the loop fill time? Last time I was dealing with this, someone suggested making that a small value (~0.1), to capture the first bit of gas that escapes the vial.

Does the loop equilibration time matter? I thought that was the time the gas from the vial sits in the loop, before injection; smaller=better=less time to react?

Longer injection time doesn't matter, as it is just the time the valves stay open and flush the loop into the injector?

[shaun78]

I have always used PE HSS but ...

If you are injecting standards and not getting large enough peaks, then is it possible for you to increase vial pressurization time and potentially the anmount of pressure in the vial?

Additionally, are you sure that the crimper you are using is tightly crimping the vial so that it is able to be pressureized?

I believe one if able to increase method sensitivity by optimizing vial and loop pressures.

[chromatographer1]

Greg,

Many labs have found the USP test difficult to use. Small peaks and peaks that cannot be seen are complaints that many have voiced. That means from a practical point of view that too little of the analytes are getting onto the analytical column.

Let's look at causes:

If your vial temperature is too low or your analyte has too low a vapor pressure then the concentration of the analyte in the headspace will be low.

If your dissolution solvent has too strong an attraction for the solvent of interest then the partition of the analyte will be inadequate to have an appropriate concentration in the vapor phase.

If you pressure the vial too little you may not be able to flush the sample transfer line and sample loop adequately to fill the loop with gas from the headspace.

If you pressure the vial too much you may dilute the concentration of the analyte in the sample as well as inhibiting the partition of the analyte into the headspace phase.

If your headspace vial leaks you won't get consistent results. It isn't as easy as you might think to get a proper seal.

If your loop is too small you will not get enough sample onto the column.

If your split is set too high you will not get enough sample onto the column.

If the amount of sample solution in your headspace vial is inappropriate to the size of the vial the partition of the volatile analytes will be too low.

If you have bare metal in your sample transfer line any amine or amide like DMF will have a difficult time to get onto the column for analysis.

Headspace isn't rocket science but there is science involved and to understand the process you really need the experience of experimentation
to learn how your unit works. Loop injection and timed injection analyzers while similar require difference parameters and conditions to get similar results.

Sadly, the art of headspace analysis is not something that can be learned from a web forum asking questions. It requires time to learn it on your own or it takes money to hire someone to teach you. Bosses and bosses of bosses need to learn that if they really want to avoid wasting time and money.

I have only scratched the SURFACE of issues with headspace analysis of difficult matrices. I had years to study and to develop the art in my former lab as there was no one with experience available. I ended up teaching the vendors how to optimize the art, as even they, at that time, had much to learn. Today they are your best option for help. A lot of vials have been injected in the GC since the early 1990s when I began my study of HS and their knowledge is more complete today than it was then.

I sympathize with you in your situation. And I do hope you can resolve the situation.

Rodney George
consultant

[Peter Apps]

I have used four different equilibrium headspace analysers, based on two different plumbing systems; loop and syringe injection. All of them had design quirks that made their performance much poorer than it could have been, and none of them interfaced optimally with the GC. Given the huge (and growing ?) number of samples that are run with headspace I would have thought that it was worth somebody's while to sit down and design a good one.

On a more concrete note; chloroform gives a weak signal on FID because it has only one C-H bond per molecule.

Peter