PE HS40 Method for USP Residual Solvents

[wired1000]

Hello All,

I have been struggling for a couple months, on and off, to get our new GC up and working to analyze for USP <467>.

The column, temp program, and gas pressures are as specified in the USP monograph. Actually I'm doing everything they say, with the exception of Headspace operating parameters. Instead of 30sec pressurization time, I'm using 2 minutes because that was recommended by the HS40 user manual.

The problem is, for the Class 1 mix, using Procedure A (haven't even gotten to "B" yet...) I can't see any kind of peaks eluting. I've tried turning the detector attenuation to -5, but that only makes it a really amplified, noisy baseline. For Class 2A, the peaks come out nicely.

Does anyone have some suggestions for fine-tuning? Anybody working with PE equipment (I've got a Clarus 500 and a HS40 headspace unit. Both are new so unlikely to be malfunctioning)

I'd appreciate any and all help or ideas!

Thanks!

[wired1000]

Other deviations from USP:

My headspace parameters are as follows:

Oven: 90
Needle: 100
Transfer Line 110
Carrier Pressure 15psi
Equilibration (Thermostatting time): 20min (This is shorter because of all the trial and error... is 45-60min really necessary?)
Injection: 1mL (this takes approx 1 min)


The USP lists 3 options for Headspace settings. It seems the only one that makes sense is "B", because wouldn't there be condensation issues with water at 80 degrees? My boss won't let me set the oven to 105 b/c he's afraid the water will boil and explode them... I assume they won't...


Any hints?

[Peter Apps]

This sounds like exactly the problem that Terry posted very recently - try searching the forum to save us going over the same ground again.

Peter

[Peter Apps]

This sounds like exactly the problem that Terry posted very recently - try searching the forum to save us going over the same ground again.

Peter

[krickos]

Hi

Here is the link to Terrys issues:
viewtopic.php?t=9845

In short he had problems doing about the same analysis for Benzene with the Ph Eur method (more or less identical). I think it was concluded in Terry case that the DMF caused the low sensitivity so he had to adjust sample/standard preparations. It is stated in Ph Eur that the proposed sample/std preparation might have be adjusted.

So have a look at that thread first.

Secondly a recommendation: As discussed in the other thread, prolonged pressurization may cause leaks and "dilution" of sample, I think you should reduce the pressurisation time to a minimum.

As for the equilibration time: Well I would not change too much when trouble shooting but understand that you do. Agree that normally a equilibration time of 20 min at 80°'C normally would be enough but recall that this test of class 1 solvents has a caracter of trace analysis.


From Terrys thread:

Elevated HS oven temperature (105°C) did not make a difference to the peak response.

I guess it is the composition of the diluent, 5 ml of DMF plus 1 ml of water, to blame. I have tried 5 ml of water and 1 ml of DMF, with the same amount of benzene in the vial, and the peak is way much higher. The composition is pretty much like the USP <467> method for water-insoluble substance.

[wired1000]

Thanks for the help. I've just skimmed through Terry's thread and AH I don't want to write/validate an in-house method! I'll try some things today...

Thanks again

[mbicking]

I just helped a pharmaceutical company complete a validation for three solvents.

Painful doesn't begin to describe it.

If you follow the method verbatim, you will probably not succeed. There are numerous technique issues, but also some procedural things that need to be changed.

We used the Agilent 7694 and 6890, so I can't help with specifics of your instrument, but we can talk about the general settings. I agree with the previous comments also.

[krickos]

Hi

As a more general comment. The general methods in Ph Eur/USP might be OK for screening, but as they are GENERAL and is made to cover a lot of different solvents they are absurdly time consuming if you doing it as a routine analysis.

I would say that your always better off to validate a method based on either of the two columns used in Ph Eur/USP and optimise it for your solvents of intrest. The time spent on validation is quite quickly saved up by the considerble faster analysis. This approach is quite common regardless if you work in "big pharma" or ín a small generic API firm.

[Jana P]

I am not sure of this helps but when setting up the USP method for our laboratory i could not get the class 1 peaks to elute well under the same headspace parameters as for the class 2 solutions. Thus we run two headspace samples when analysing, 1 for class 1 and 1 for class 2 (A and B)

[Vince T]

Sorry to bump an old thread, but I'm running into the same issues.

Standard solutions were prep'd per the USP requirements. I'm using a Clarus 500 GC w/ FID coupled with a Turbomatrix 40 Headspace Sampler (I coupled the Headspace to the capillary injector port).

Column: 0.32mm x 30m x 1.8 micron film of USP Phase G43
Carrier Gas (through Clarus): 35 cm/s He (as per USP)
Split ratio: 5:1
Temp Program: per USP

I originally used Headspace Operating Parameter Set 3 which called for 80 degree oven temp, 90 degree needle temp, and 105 degree transfer temp, 45 minute thermostat time, and 1 mL injection volume. I lowered the USP recommendation of 60 minute pressurization time to 1 minute.

Using this set of headspace parameters, I don't see any true peak resolution in both the class 1 and class 2 mix b standards, but I do get most of the expected peaks from class 2 mix a (the first few peaks are somewhat distorted and wide, though).

Today, I switched to parameters similar to set 2 for the headspace settings: 105 degree oven temp, 115 degree needle temp, 120 degree transfer temp, 45 minute thermostat time, 1 minute pressurize time, 1 mL injection.

Using this new set, I am starting to see very small peaks in class 1 and class 2 mix b standards, but somewhat broad and ugly. Class 2 mix a, I get even more peaks than I did with the original headspace settings.

It looks like I'm on the right track using higher temperatures on the headspace, but are there any modifications that anyone else has used that can help me improve the chromatography? I have a set of pamphlets my supervisor gave me showing the expected chromatography and so far, my chromatograms don't look anything like those using the same column and parameters.

[mbicking]

Exactly how is the headspace connected? Does the transfer line go in through the septum, or does all the carrier go through the HS unit before the carrier? If you are going through the septum, what is the flow from the HS? I used a 6890, and this meant that the extra flow went out the split vent, and the effective split ratio was actually much larger than the setting.

I used an HS oven temp of 90 C for 30 minutes, and a 140 C transfer line temp. After that I made some adjustments for individual equipment variations (the USP method is really just a guideline, and as long as you set and validate the final conditions, you should be OK with compliance).

The Class 1 solvents are a problem, and I could not see CCl4 at all. The final dilution is also not stable for more than a few hours, so if you are using old standards (more than a day), you may not see much.

Try some of these things if they seem appropriate and let us know how it goes.

[Vince T]

I'll try and describe it, but I'm still relatively new to the GC, so bear with me.

The headspace transfer line is screwed into the injector port. The transfer line itself pokes through a septum and into the injector liner. I dialed in a 15 psi carrier pressure of Helium on the headspace end, and have the Clarus GC set to 35 cm/s. All gases on the Clarus are controlled by the PPC, so the split is also automated (though I will play around with the split to see how it affects the chromatography).

So far, it seems that class 2A peaks always show up. I definately get more peaks to resolve with the overall higher headspace temperature settings vs. the lower ones. The chromatography isn't as pretty as the examples I've seen, though. Also, even with the split at 5:1, some of the peaks actually max out the detector, so I don't think the split ratio being too high is the problem...if anything, it could probably be raised.

My standards are at least a day old, so I'll try reprepping. You say only the final dilution is unstable? I hope so, cus those standards are expensive, if even the stocks are that unstable. In any event, I don't seem to get any chromatography out of the class 2B standard either. It looks pretty similar to the class 1 standard. I get a decent baseline, with a few lumps here and there where peaks might be. From all the examples I've seen, class 1 and 2B peaks should all resolve within the first 10 minutes, while 2A should have plenty of peaks through 35 minutes.

I'll play around with the headspace parameters some more. Maybe my settings are still somewhat low and I have some condensation issues? I'll let you know how it goes.

Also, I've been reading that a 1mm id split injection liner can also help with peak resolution vs. the normal 4mm id injection liner. Do you use it, or you think it's perfectly fine w/o?

[krickos]

Hi

Some general thoughts:

Liner: YES you should definitly use the HS liner (200µL?), will improve your peaks. Do not use the standard 990µl one.

Sample preparation per USP/EP: Both are more or less identical however EP has been gracious enough to add that the default dilution may not be good and you may have to change it.

So first user the right liner. Then increase sample concentration if needed. If you use only DMF in sample/std vial you "must" increase sample/std concentration because the deafult one is not enough (or poor) to spot benzene (others?) in USP default settings.
DMF generally retains solvents more than water (see comments/link to Terrys issues).

[Vince T]

Ok, I ordered the liner on Friday so we'll see how that goes.

I've looked at the old thread and I guess I'll cross that bridge when I get to it, but right now I'm still just using the water-soluble part of the chapter. So even with the standards in water, I don't get much in the way of peaks in class 1 and 2b standards and the 2a standards give somewhat wide, oddly-shaped peaks. I can probably attribute the peak shapes to the type of liner, so I guess we'll see about that.

Here are links to what my chromatograms are looking like just for your information.

Class 1 STD
Class 2A STD
Class 2B STD

[Vince T]

I'm not sure if this is what could be causing my issues, but I might have identified another problem.

As far as I can see, the 624 series columns adhere to the USP Phase G43 requirement, but it looks like most validations I'm seeing in my research are using 1301. Do you think that would make an impact on the chromatography I'm seeing?