Poor Recovery for residual solvents by GC-FID-HS

[vdnsrinivas]

Hello,
I am running a GC-FID-HS method for residual solvents@ Ethanol,Acetone, IPA, DCM. Working std 1ml in 20ml HS Vial. The areas were arround 400 (EtoH), 200 (Acetone and IPA) and 12 (DCM). In the spiked sample the recoveries for acetone IPA and DCM were arround 70-80%. I compared the volume to be tested 2ml and 5ml in HS vial the areas are good in working std but recoveries were still like 70-80%. Please give me some ideas.

Thanks
Srinivas

[chromatographer1]

You are using large samples. Large samples decrease the amount of solvent that partitions into the vapor phase of the vial (the headspace).

You should be doing std addition for accurate results, keeping the total amount of liquid volume consistent in all vials. External std methods are often done, but it is not a good practice unless it is unavoidable.

Let's assume you have 10mg samples of a 'drug'. Prepare 3 of them. 10 mg = 10,000 micrograms

Let's assume you are looking at 1% of each residual solvent or 100 micrograms: 100mcg/10,000 mcg = 1%

Prepare std addition solution. I prefer DMAc as the carrier solvent. 10 g/L or 10 mcg/mcL of each solvent
is a possible working concentration (adjust as needed)

Add 0.5mL of solvent of choice for 'drug' to each vial.

Add 20mcL of DMAc to Vial 1. Cap vial.

Add 10mcL of DMAc and 10mcL of std addition solution to Vial 2. Cap vial.

Add 20 mcL of std addition solution to Vial 3. Cap vial.

Run each preparation through HS analyzer.

Your three samples should give you linear results for each residual solvent. The amount in the 'drug' is the negative value for the x-axis intercept (a negative of a negative becomes a positive number.)

If the linearity of the results are not linear, then

your HS method is inadequate
or
your vials are leaking
or
your samples are not homogeneous in residual solvent content.

Doing external stds for a single sample, you will never know if the last point is true or not.
By doing Std Addition you prove your analysis is linear over the range in which you have prepared your samples.

Homogeneity and linearity are important factors that Regulators LOVE to see demonstrated in your analysis.

You will under less than often situations have ext stds that are consistent in recovery to actual samples. It is much more common to find that recoveries are within 50-90% range. So either your answer is inaccurate or you have to demonstrate the recovery is consistent over the range of measurement you seek to perform your HS analysis. This is WAY TOO MUCH WORK AND TIME TO CONSUME.

It is much quicker and cheaper to do std addition, running three samples and be done, than a sample and an external std. Plus two more ext standards to demonstrate linearity and recovery when the numbers don't make sense.

But, hey, what do I know?

best wishes,

Rod

[chromatographer1]

I should add that sometimes recoveries are MORE than 100%.

I have had situations were recoveries were as much as 120%.

This is why the large pharmaceutical company I worked for took my advice and always prepared std addition HS methods.

The FDA was ALWAYS happy with our (MY) work. No complaints. Everything was clearly demonstrated with good Scientific methodology. RSD values were around 2%. Linearity was 0.999+ for 18 common solvents. (from 1 ppm to 1% residual solvent)

See my paper in Analytical Chemistry June 1997 to see how linear the results can be using the technique I described in my previous post.

best wishes,

Rodney George

[vdnsrinivas]

Thanks for your constructive information. I will let you know if their are any quires.

I should add that sometimes recoveries are MORE than 100%.

I have had situations were recoveries were as much as 120%.

This is why the large pharmaceutical company I worked for took my advice and always prepared std addition HS methods.

The FDA was ALWAYS happy with our (MY) work. No complaints. Everything was clearly demonstrated with good Scientific methodology. RSD values were around 2%. Linearity was 0.999+ for 18 common solvents. (from 1 ppm to 1% residual solvent)

See my paper in Analytical Chemistry June 1997 to see how linear the results can be using the technique I described in my previous post.

best wishes,

Rodney George

[vdnsrinivas]

Dear Rodney George,

I am running the residual working standard and drug substance spiked with residual working standard with matrix agent @ Nacl (1.5gm==because I ran the samples with 0.5gm Nacl and 1 gm Nacl but recoveries for 1.5gm Nacl were good) but the recoveries for methanol and DCM were good but for Acetone and IPA there is only 80% recovery. Can you give me some suggestions and concepts regarding matrix agents.

Thanks,
Srinivas

[chromatographer1]

I assume you are calculating recovery against a single external std.

In my paper I calculated recovery of a sample containing 50 ng of solvent from a regression line (considered to be 100% recovery) formed from a series of samples containing solvents from 0.5 ng to 1000 ng per sample. My recoveries for 17 solvents were 95 to 99 %, with pyridine at 89%. Even so the regression line linearity (R squared) for the solvents were all > 0.999 except for pyridine which was 0.995.

Recoveries from matrices are not always 100%. Trying to hit the ideal recovery from a std matrix that differs from the dissolved drug solution is a difficult solution to achieve, especially for multiple solvents.

This is why I again repeat over and over that it is a better practice to NOT use external stds or internal stds when doing HS analysis. Use a std addition method instead.

If you prove you have linear recoveries (linear does not mean 100%) from your drug matrix by doing std addition samples, then you have demonstrated your determination is accurate over the range of addition. Use the X axis intercept to determine the actual content of the solvent in your drug. Three sample vials are needed for an analysis. Additional samples can be run at the same time without the std additions as long as the drug weights and dissolution solvent volumes are the same.

You may NEVER find a perfect dissolution solvent and salt mix that will give you the same recovery as from a solution of the drug for multiple solvents. The fact that you found TWO solvents that gave you 100% recovery (not 70% and not 110%) is remarkable in itself.

I never used salt enhanced techniques as I never needed them. I found that reducing the sample size and the volume of drug solution improves recoveries and speeds the HS equilibration. Sensitivity is improved by reducing the vial size. The vials that I used in my paper were 3.35mL in internal volume, not 12 mL and not 20 mL. And I was able to see 1 ppm of solvent for a 1 mg sample of drug only using 25 microliters of solution. I NEVER used more than 250 microliters of solution in my work. NEVER, and I was able to see much less than 1 ppm with 1-3% RSDs for over twenty common solvents.

Bigger vials and bigger samples do NOT mean lower limits of detection or improved reproducibility.

Other salts commonly used to saturate the drug solution are K or Na carbonates or sulfates.

I prefer to avoid ANY salts as they might contribute absorbed solvents to your analysis that may interfere with your analysis or make it inaccurate. If you require them as a buffer that is a different situation.

Good luck with your work. My work was extended to validate HS analysis for a large number of different matrices: narcotics, peptides, salts, waste water, and biologicals.

I hope your work goes well.

best wishes,

Rod

[kalidassa]

Dear Mr. Srinivas can yoyu tell me the diluent what you are using and about your sample matrix( It is water soluble), Because diluent playes main role in Recovery. We can change the polarity nature of the diluent by just adding water 5 %, 10% for eg DMF: Water (95:5,90:10). You try you may get good recovery

[vdnsrinivas]

Dear Rod,
Here are the specifications of the residual solvents in the XX Drug substance,
Methanol (NMT 500 ppm)
Acetone (NMT 100 ppm)
Isopropanol (NMT 100 ppm)
Methylene Chloride (NMT 50 ppm)

Even though you mentioned the procedure in the previous post, Can you please once again give me a detailed step by step procedure @ preparation of samples and calculations that would be greatful.

Thanks,
Srinivas

[chromatographer1]

I assume your drug is soluble @ 5mg per 100µL of water.

Prepare std addition solutions in DMAc containing

1. 12.5µg/mL, 25µg/ml, 25µg/mL and 50µg/mL (DCM, IPA, DMK, and MeOH)

2. 25µg/mL, 50µg/mL, 50µg/mL, and 100µg/mL (DCM, IPA, DMK, and MeOH)

3. 50µg/mL, 100µg/mL, 100µg/mL and 200µg/mL (DCM, IPA, DMK, and MeOH)



10µL of 12.5µg/mL = 125 ng = 25 ppm of 5 mg

10µL of 25.0µg/mL = 250 ng = 50 ppm of 5 mg

10µL of 50.0µg/mL = 500 ng = 100 ppm of 5 mg

10µL of 100µg/mL = 1000 ng = 200 ppm of 5 mg

10µL of 200µg/mL = 2000 ng = 400 ppm of 5 mg



Weight exactly 5mg of drug into 3 vials.

Add 100µL of water and 10µL of one of each of the std addition solutions to the three vials.

Do this quickly and do not allow the vials to remain uncovered for any length of time.


It is not critical that you measure exactly the prescribed weights of each solvent into your std addition solution but that you exactly record the weights that were added.

For example: If you prepared a 12µg/mL and a 26µg/mL and a 52µg/mL solutions for DCM that is fine.
Just record that you added 24 ppm and 52 ppm and 104 ppm of solvent to the vials.

Plot the ppm added against the peak area of the solvent. Calculate the equation of the regression line and determine the x-axis intercept. The negative value of it will be the amount of solvent in the drug sample.

If you do not have the skills or understand the math I can send you a spreadsheet that will do the calculations for you.


This procedure is more accurate and reproducible than the ability of the chemist to prepare and perform the analysis. This I and several chemists learned when they were being certified to perform this procedure.

When I was sloppy I got R squared values of 0.998 and when I was careful I got values of 0.99995 or better.

You may wish to run samples without the drug being present until you get the linear and reproducible results that show you can do the procedure properly.

Best wishes,

Rod

[vdnsrinivas]

Thanks very much Rod for your detailed explaination, I will try and let you know the output. I would be even happy if you send me d excel calculation to my email:: vdnsrinivas@gmail.com


Thanks,
Srinivas

[chromatographer1]

The Excel spreadsheet file was sent to you.

Happy Headspace analysis !

best wishes,

Rod

[chromatographer1]

Reviewing my posts I may not have been clear about the additional samples run at the same time as the std addition samples.

The dissolution solvent for the additional samples must be the SAME as for the std addition samples.

So, as in the example I provided, the drug samples must also contain 100µL of solvent AND 10µL of DMAc containing 0 ppm of the residual solvents that were added to the other samples. The weight of the drug must also be the same (5mg).

The pure dissolution solvent then consists of 110µL of the 10:1 co-dissolution solvents, water and DMAc.

I hope I caused no confusion.

It is also possible that one of the std additions can be 0 ppm. Thus only 2 std addition solutions are prepared and a zero ppm addition (pure DMAc) is one of the three points, but I would prefer that three std addition solutions are prepared. This makes the procedure more rugged and more likely to prove errors clearly if any are made (non-homogeneity of the samples or errors in the preparation of std addition solutions, which of course, never happen in the real world, right?

best wishes,

Rod

[chromatographer1]

Srinivas

I received no word from you whether or not you received the spreadsheets I sent you.

I hope you received them and were able to use them.

best wishes,

Rod

[vdnsrinivas]

Dear Rod,

Sorry for the delay in replying for your email regarding Excel. It helped me a lot in understanding the calculations.

Thanks,
Srinivas

[chromatographer1]

Srinivas

You are welcome.

Feel free to share your experience using the std addition procedure after first using it without the drug being present.

Did your solution preparation and crimping demonstrate accuracy or show handling errors which can be less than obvious without a good diagnostic test?

The std addition method will also demonstrate inadequate chromatography hardware or inappropriate chromatography conditions.

That is a great advantage because unless the equipment and the operators are performing as they should your data results will be poor. And when your data results are good (good linearity and low RDS values) then you have high confidence that your end solvent results are good as well. And your reviewers will have high confidence in your work, and everyone will be happy.

Getting the accurate answer is why you are doing the work to begin with, right?

Good luck and please let us know how things work out for you.

best wishes,

Rod