GC Method for determination of MeOH, isopropanol and DMF

[chromatographer1]

For some matrices it can 'leave', but it is not guaranteed. Solvents can be trapped in the crystal structure or only partially released.

Sloppy science is not good science.

I would perform other non-headspace residual solvent tests to verify the total release of the solvent before expressing any confidence in my numbers.

I would not make any claims unless I had proof. Of course, I never did enjoy a stay in jail, or paying fines, or losing my job. Just my personal preference.

best wishes,

Rod

[alemaggot]

My product can be dissolved only with NaOH solution. That it's not good for GC. But what you say Rod it's correct, I know. I'll try one direct injection.

Then today I've found in my lab an old chaney adapter. I've learn how to use it. I've clean it and I've try 5 injection of standard. My RSD is about 1 %. It's good

As mentioned above it's the first time that I use this equipment. I've got a question. What is best choice between a slow syringe piston release and fast syringe piston release, when I transfer the solution in the vials?
Some one of you used a Shandron repro-jector? It's the type of my chaney adapter.

[Peter Apps]

My product can be dissolved only with NaOH solution. That it's not good for GC. But what you say Rod it's correct, I know. I'll try one direct injection. Do not do a direct injection of sodium hydroxide solution. You said that you were doing headspace analysis - dissolve the product in sodium hydroxide solution and run a headspace on that. I anticipate problems with chemical reactions but it is worth a try. One question - if the product only dissolves in sodium hydroxide solution how does it get taken up by the body of the person that takes the drug ?

Then today I've found in my lab an old chaney adapter. I've learn how to use it. I've clean it and I've try 5 injection of standard. My RSD is about 1 %. It's good

As mentioned above it's the first time that I use this equipment. I've got a question. What is best choice between a slow syringe piston release and fast syringe piston release, when I transfer the solution in the vials? Try both and see what works the best for you
Some one of you used a Shandron repro-jector? It's the type of my chaney adapter.

Peter

[alemaggot]

I don't know what is the final pharmaceutical forms that our client product. I only know that it is not soluble in water and is pratically insoluble in other solvent, such as DMSO.
If I inject a NaOH solution in headspace method, how can I ensure the total DMF extraction, if I can't overcome 100°C of sample solution?

[chromatographer1]

I have used NaOH (aq) solutions in headspace. They don't pose a problem per se.

I would use a minimum volume of solvent and of drug weight to dissolve the drug and completely evaporate your water at your final temperature for equilibration for headspace.

Water expands about 1500-2000 times of its liquid volume when evaporated. So 25 microliters of water makes about 50mL of steam when evaporated. This would increase your vial pressure perhaps to an adequate level when heated to 190C.

I would check to see how much drug would dissolve in NaOH at 90C and use that amount for your testing. Then use FET as before and see if your DMF values change.

You can decrease/increase your drug amount and aqueous NaOH solvent volume with trials to achieve your best results.

Good luck,

Rod

[Peter Apps]

I don't know what is the final pharmaceutical forms that our client product. I only know that it is not soluble in water and is pratically insoluble in other solvent, such as DMSO.
If I inject a NaOH solution in headspace method, how can I ensure the total DMF extraction, if I can't overcome 100°C of sample solution?

You do not need total DMF extraction, you need a repeatable equilibrium with detectable quantities of DMF in the headspace.

Peter

[chromatographer1]

There lies the rub, a repeatable equilibrium from solids is unprovable and is based on assumptions which may not hold true all the time. That is why to have a valid method for residual solvent (as is required in pharmaceuticals) the sample must be in a liquid state for Henry's Law to apply. While you CAN SOMETIMES get repeatable results from extracting solvents from a solid it is impossible to prove, anymore that sight unseen you can demonstrate two samples of solid are in exactly the same condition.

melt or dissolve is the only solution (pardon the pun)

Rod

[Peter Apps]

Hi Rod

If I understand correctly the problem with headspacing solids is that there could be a fraction of the analyte content that is not exchangeable between the two phases - locked up in crystal structures or sealed into encapsulated powders for e.g., so a headspace method gives a measurement only of the exchangeable fraction.

The function of melting the solid, or dissolving it in something is to make the whole of the analyte content exchangeable - but is it ever really possible (without doing an independent, non-headspace analysis) to demonstrate that there is no "hidden" analyte - a portion of it that for some reason does not exchange ? The only way that I can think of is complete evaporation.

The issue of exchangeable and "fixed" volatiles can be quite important in semiochemistry - many of the volatile chemical signallers are tightly bound to specific proteins which modulate their release, so headspace methods underestimate total content and extraction methods overestimate accessible signal.

Back to alemaggot's desire for "total DMF extraction" I was just trying to reassure him that he does not have to go to very high temperatures because there is no need for "total extraction" (as long as you have exchangeable analytes).

Peter

[chromatographer1]

Peter wrote:

"Hi Rod"

Hi Peter,

"If I understand correctly the problem with headspacing solids is that there could be a fraction of the analyte content that is not exchangeable between the two phases - locked up in crystal structures or sealed into encapsulated powders for e.g., so a headspace method gives a measurement only of the exchangeable fraction."

Absolutely correct. And the problem of reproducibility is getting two samples of the same 'powder' which have identical characteristics. One can come close, and come often, but it is the 'oh, my, it is different this time' problem which gets the chemist into trouble with the government for those working in regulated industries.

"The function of melting the solid, or dissolving it in something is to make the whole of the analyte content exchangeable - but is it ever really possible (without doing an independent, non-headspace analysis) to demonstrate that there is no "hidden" analyte - a portion of it that for some reason does not exchange ? The only way that I can think of is complete evaporation."

Henry's Law is defined as an equilibrium of vapor in a gas phase and in a LIQUID phase. There has to be the possibility of movement of an analyte THROUGH a liquid phase and contact a physical presence of a gas phase. The equilibrium may be one sided for those analytes with a low vapor pressure, and the equilibrium may take a long time with larger samples, but in theory according to Henry's Law, it can be achieved. I have never read, never seen, or even heard of any violation of this theory with any liquid. Please give me an example of a non-exchangeable analyte if you know of one, I am all ears.

"The issue of exchangeable and "fixed" volatiles can be quite important in semiochemistry - many of the volatile chemical signallers are tightly bound to specific proteins which modulate their release, so headspace methods underestimate total content and extraction methods overestimate accessible signal."

A chemical that is connected by a chemical bond to another analyte, or is hindered by some physical or chemical means, "modulated", is outside the definition of a headspace experiment. A solid with an entrapped liquid is one example of this.

"Back to alemaggot's desire for "total DMF extraction" I was just trying to reassure him that he does not have to go to very high temperatures because there is no need for "total extraction" (as long as you have exchangeable analytes)."

Concerning the goal of 'total extraction' you are correct, there is a moving target of a reproducible equilibrium when extracting liquids from a solid matrix, which if done VERY carefully, with a reproducible sample of solid material (good luck with that), it is possible to estimate a solvent content with a solid. But another problem is how does one determine absolutely HOW MUCH of the entrapped liquid is released at any point in time based on a weight percent of the solid?

Smaller solid particles will release more liquid per weight of sample per time, will they not? as the shorter path for escape with more particles per weight should give a faster rate of escape from the solid.

I have history with solid samples that have been heated to release trapped liquids. Sometimes the matrix is destroyed by oxidation or self-reactivity decomposition before the liquid entrapped is mostly released. Now I am not saying that with research and testing it can never be stated that 'X' material can be tested for liquid content by headspace analysis accurately, I am only saying that there is a whole heck of a lot of work involved before making that claim.

And many of the folks who ask questions here tend to not have the time to perform a rigorous study of their matrix. I would always wish to inform the 'unaware / less experienced' chemist of problems with any analysis that could get them into trouble with their bosses or the government.

Solid reliable analytical chemistry that would stand up in court is my soapbox, and I will continue to stand on it. Getting a 'measurement' and making a reasonably correct and precise determination are two very different situations in the pharmaceutical industry. Theoretical approximations are great when one's career and reputation are not on the line.

But let's remember, when working with solids to perform residual solvent determinations, we are dealing with approximations which only have the dignity of accuracy after a lot of research showing that our assumptions are valid.

best wishes,

Rod

[Peter Apps]

Peter wrote:

"Hi Rod"

Hi Peter,

"If I understand correctly the problem with headspacing solids is that there could be a fraction of the analyte content that is not exchangeable between the two phases - locked up in crystal structures or sealed into encapsulated powders for e.g., so a headspace method gives a measurement only of the exchangeable fraction."

Absolutely correct. And the problem of reproducibility is getting two samples of the same 'powder' which have identical characteristics. One can come close, and come often, but it is the 'oh, my, it is different this time' problem which gets the chemist into trouble with the government for those working in regulated industries.

"The function of melting the solid, or dissolving it in something is to make the whole of the analyte content exchangeable - but is it ever really possible (without doing an independent, non-headspace analysis) to demonstrate that there is no "hidden" analyte - a portion of it that for some reason does not exchange ? The only way that I can think of is complete evaporation."

Henry's Law is defined as an equilibrium of vapor in a gas phase and in a LIQUID phase. There has to be the possibility of movement of an analyte THROUGH a liquid phase and contact a physical presence of a gas phase. The equilibrium may be one sided for those analytes with a low vapor pressure, and the equilibrium may take a long time with larger samples, but in theory according to Henry's Law, it can be achieved. I have never read, never seen, or even heard of any violation of this theory with any liquid. Please give me an example of a non-exchangeable analyte if you know of one, I am all ears. I suppose the simplest example of what I had in mid would be a weak organic acid in an aqueous solution at very high pH - the acid is there in the liquid phase, it is even in equilibrium with its dissolved unionized form, which in turn is in equilibrium with the headspace, so Henry's law is not violated, but you could not do a headspace analysis and the only way to demonstrate that there was the acid in the liquid phase would be to change the pH. And I'll agree with you if you say the ions and the unionized molecules are chemically different so the example is not valid

"The issue of exchangeable and "fixed" volatiles can be quite important in semiochemistry - many of the volatile chemical signallers are tightly bound to specific proteins which modulate their release, so headspace methods underestimate total content and extraction methods overestimate accessible signal."

A chemical that is connected by a chemical bond to another analyte, or is hindered by some physical or chemical means, "modulated", is outside the definition of a headspace experiment. A solid with an entrapped liquid is one example of this.

"Back to alemaggot's desire for "total DMF extraction" I was just trying to reassure him that he does not have to go to very high temperatures because there is no need for "total extraction" (as long as you have exchangeable analytes)."

Concerning the goal of 'total extraction' you are correct, there is a moving target of a reproducible equilibrium when extracting liquids from a solid matrix, which if done VERY carefully, with a reproducible sample of solid material (good luck with that), it is possible to estimate a solvent content with a solid. But another problem is how does one determine absolutely HOW MUCH of the entrapped liquid is released at any point in time based on a weight percent of the solid?

Smaller solid particles will release more liquid per weight of sample per time, will they not? as the shorter path for escape with more particles per weight should give a faster rate of escape from the solid.

I have history with solid samples that have been heated to release trapped liquids. Sometimes the matrix is destroyed by oxidation or self-reactivity decomposition before the liquid entrapped is mostly released. Now I am not saying that with research and testing it can never be stated that 'X' material can be tested for liquid content by headspace analysis accurately, I am only saying that there is a whole heck of a lot of work involved before making that claim. I was certainaly not advocating that alemaggot use a solid sample - he was talking about having his sample dissolved in sodium hydroxide but seemd to think that he had to heat it to very high temperatures just in order to get the DMFinto the headpsace

And many of the folks who ask questions here tend to not have the time to perform a rigorous study of their matrix. I would always wish to inform the 'unaware / less experienced' chemist of problems with any analysis that could get them into trouble with their bosses or the government.Robust and reliable should indeed be close to the top of the list for any analysis that has to stand regulatory scrutiny

Solid reliable analytical chemistry that would stand up in court is my soapbox, and I will continue to stand on it. Getting a 'measurement' and making a reasonably correct and precise determination are two very different situations in the pharmaceutical industry. Theoretical approximations are great when one's career and reputation are not on the line.Metrologists would argue vehemently that any measurement must have an uncertainty to go with it, and to determine the uncertainty you have to fully validate the method, and run replicates - in other words do the kind of careful chemistry that you advocate

But let's remember, when working with solids to perform residual solvent determinations, we are dealing with approximations which only have the dignity of accuracy after a lot of research showing that our assumptions are valid.So we agree, if you can't melt it or dissolve it don't headspace it

best wishes,

Rod

Peter

[chromatographer1]

It appears that I misunderstood you. I supposed you were recommending that a good DMF measurement from the solid was a possibility. I missed the dissolve in NaOH part. Sorry about that.

Yes. We agree.

For what it is worth: I once developed a validated method to measure a volatile nitrogen base, present as a salt, in a solid matrix, a peptide in fact, using a total evaporative technique. It wasn't perfect, but was accurate enough to determine a minimum of 50 ppm reliably and accurately. This was adequate for regulatory needs at the time. And at $100 a milligram, using less sample was better. I was hoping for my usual 1 ppm, but sometimes the chemistry is what it is.

I also have done residual solvent analysis on nearly insoluble zinc salts by heating them to their melting temperature.

But I have never gotten reproducible results trying to extract solvents from solids, although how heartily I have tried. (Did get close once, but no cigar.)

Using NaOH (aq), temperature performs wonders, and sometimes a lot less than expected will get the job done.

DMF still is a trouble maker, as it loves to stick wherever it wants which is mostly everywhere. Temperature is a great help here too.

best wishes,

Rod

[alemaggot]

Yesterday evening I was read once again the B. Kolb books "Static head space analysis. Theory and practice".

If you want, you can work with solid sample. But, before to start, you must verify if you work in "Partition system" or in "Adsborment system". That is you must verify if your extraction coefficent (Q) is indipendent from the analyte concentration. It must be function of it.
If you work in partiton system you're sure that this thing is present. In adsorbmente system no.

Another important thing is the particle size of your matrix. If you work with powder with particle size minor than 300um (my case) your extraction more probaly is very well.

For test the system you must apply a MHE analysis. You must extract for n time from the same vial and verify your Q costant.
Here I've got a problem. Because I can't develop my results. I don't understand how I calculate if I work in partition system or not.

Can you help me?