Chromatography Forum

I think you had success, just because you stop to use mixture of water and DMSO. But you should lose a lot in sensitivity. You'll get 10 times bigger benzene peak in water, than in DMSO. But if it is working in DMSO, who cares: congradulation.

hi,
you just check your gas cylinder pressure for analysis.
it will creat a big problem like that
Hardik

I think, except water and benzene, you have other components. Benzene almost insoluble in water. Other components, if present, affect benzene peak response. Prepare solution of water and benzene only and you should improve reproducibility

there are six components in vial in all。why try with only wate rand benzene，which is not my need。i want

assay the six components at the same time。why the other components affect benzene peak response，based

on science？

these days i found one big mistake i made，the first benzene solution 400ppm much bigger than USP467 class1 standard mixture first dilution of 100 ppm，while i use the same size volumetric flask and same volume DMSO with USP467。therefore my subsequent dilution not consistent with class1 solution。result：peak area much bigger than the benzene in class1 standard。also lead to my mixture benzene high concentration and not homogenous，and surplus benzene easily float on water，aspirated by pipette first，maybe usually lead to first vial benzene peak area bigger than following。

hi,
you just check your gas cylinder pressure for analysis.
it will creat a big problem like that
Hardik

usually before i start a new sequence，i will check the gas cylinder ，make sure enough nitrogen storage。

Headspace method for Benzene determination works in water very well without any special care. I developed and validated few methods by myself. Reproducibility is within 2-3%. Benzene dissolves in water good enough to prepare 2 ppm and, of course, 0.1 ppm solutions. Check Merck index for solubility.
But you have right to believe to some one who sad it does not.
And yes, USP method for Benzene will not work, because of presence of DMSO in the system. It will screw up linearity, recovery and may be reproducibility.

are you sure DMSO is bad guy，bad data maker？

what's ur incubation parameter。now i suspect my parameter 105 degree，45 min，too fierce，which may lead to occationally very bad rsd 10%-15%。

NOW bad rsd still emerge occationally，i hardly gain ideal data to valid this method，especially linar and recovery data。

sometime only benzene express bad rsd。 sometime all the six bad rsd，even water-solve component。 hardly to boost my validation work。

Good technique must be learned.

I have always found that trying to dilute an already dilute solution that easily partitions in headspace is an accident waiting to happen.

It is much easier to make a DMAc solution containing the solvent analytes by weight per 50 microliters at the amount I wish to add to 100mL of water and then adding the 50 microliters to 99.95mL of water after which I vigorously mix the aqueous solution before sampling into a headspace vial.

Making up different DMAc solutions at different concentrations, each having a precise weighed amount of solvent(s) was more accurate than trying to dilute a dilute solution that may not be fully dispersed.

Headspace will expose any mistakes in sample preparation. It is more accurate than any preparation technique. In that aspect it is a blessing.

Good luck in preparation of your samples.

Rod

Good technique must be learned.

I have always found that trying to dilute an already dilute solution that easily partitions in headspace is an accident waiting to happen.

It is much easier to make a DMAc solution containing the solvent analytes by weight per 50 microliters at the amount I wish to add to 100mL of water and then adding the 50 microliters to 99.95mL of water after which I vigorously mix the aqueous solution before sampling into a headspace vial.

Making up different DMAc solutions at different concentrations, each having a precise weighed amount of solvent(s) was more accurate than trying to dilute a dilute solution that may not be fully dispersed.

Headspace will expose any mistakes in sample preparation. It is more accurate than any preparation technique. In that aspect it is a blessing.

Good luck in preparation of your samples.

Rod

thanks for sharing your experience.

but sometimes the concentration i need very low, hard to weigh by balance, just one step gain solution i need,

not avaiable, i have to make a high concentration solution first, then dilute into the concentration i need.

0.1ppm for example. very low concentration solution hard to make. i have to say.


now i suspect the incubation temp of 105 degree very much which is the main factor ,leading to my bad rsd,

random，like a ghost。i don't know when it emerge leading to failure once again。although we are not critical，

just rsd of six vial below 10% is OK。

is most pharmaceutical company in drug residual solvents assay using usp467 method？

what's the difference using 467 method or using method self-developed？

method validation and verification，big difference in amount of work？

I think my statements were possibly misunderstood.

If I need to add 1 microgram (1000ng) of benzene to water, I would make up a solution of benzene in DMAc, 1mg/mL, or 25mg per 25mL. 25 mg is easily weighed out into ~10mL and then an additional amount of DMAc to bring the volume to 25mL.

I could then add 1 microliter of this solution to 100mL of water to achieve a 10ng/mL solution of benzene in water.

or 50 microliters to give a 500ng/mL of benzene in water. How low do you need to go?

Most large companies will validate their own methods, which are not low limit methods but quantitative tests to accurately measure residual solvents.

For example I was able to measure residual solvents in drugs at 1 ppm, plus or minus 0.1 ppm routinely.

Once validated, it would take three vials to test a drug sample and only 45 minutes of lab time to achieve accurate results, not a limit test.

A validation does take time, but once done it does not need to be repeated unless new equipment is purchased. USP 467 gives a test that disputing companies can use to resolve conflicting results.

To design and validate a method may be beyond the capabilities of a small company.

best wishes,

Rod

I think my statements were possibly misunderstood.

If I need to add 1 microgram (1000ng) of benzene to water, I would make up a solution of benzene in DMAc, 1mg/mL, or 25mg per 25mL. 25 mg is easily weighed out into ~10mL and then an additional amount of DMAc to bring the volume to 25mL.

I could then add 1 microliter of this solution to 100mL of water to achieve a 10ng/mL solution of benzene in water.

or 50 microliters to give a 500ng/mL of benzene in water. How low do you need to go?

Most large companies will validate their own methods, which are not low limit methods but quantitative tests to accurately measure residual solvents.

For example I was able to measure residual solvents in drugs at 1 ppm, plus or minus 0.1 ppm routinely.

Once validated, it would take three vials to test a drug sample and only 45 minutes of lab time to achieve accurate results, not a limit test.

A validation does take time, but once done it does not need to be repeated unless new equipment is purchased. USP 467 gives a test that disputing companies can use to resolve conflicting results.

To design and validate a method may be beyond the capabilities of a small company.

best wishes,

Rod

Rod，u are so kind。i still have question，if i prepare solution of very low benzene as u say，

when i transfer into 9 vials parallel，can make sure good rsd？ the solution more steady than

solution i used to prepare，can gain better data？

actually recently i change solution preparation method。first 100mg benzene to 10ml

volumetric flask，DMSO as solvent。then 1ml to 100ml flask containing 9ml DMSO，using

water volume。 then 1ml to 50 ml volumetric flask all water，then 5ml this solution to 100ml

flask，at last gain a solution of 0.1ppm。same method with method of 467 in making class1

standard solution。 maybe more steps，mean more loss of solvents。

now i suspect the incubation temp of 105 degree very much which is the main factor ,leading to my bad rsd, random，like a ghost。i don't know when it emerge leading to failure once again。although we are not critical，just rsd of six vial below 10% is OK。

Rod what's ur point？sometimes six components rsd all bellow 2%。since i change solution preparation method，now my benzene peak area only about 7，equivalent with benzene‘s in class1 standard。 the first vial benzene bigger than latter，not happen again。

Benzene partitions from water so very easily. Leaving a flask open between samplings, leaving a vial uncovered or unsealed after transfer, using a glass pipette to sample instead of a hand pipetter, so many opportunities for a loss of benzene.

If you routinely get a first vial with higher peak area than others it is indicative of loss from the other vials. Septa and handling errors account for this. As long as you get the data you need and your samples pass and you are not making stds at too high a level somehow, then all should be well.

best wishes,

Rod

why the other components affect benzene peak responsebased
on science？

Benzene is almost insoluble in water. If you have other components in the system (like DMSO), Benzene "moves" completely in DMSO. When you try to do linearity for Benzene in Water/DMSO, you see a problem.
why is DMSO a bad guy?

DMSO is a great solvent for GC headspace. May be, to aggressive sometimes. If Benzene has atleast NMT 20 ppm spec, but it is only 2 ppm.
Problem is that Benzene is very well soluble in DMSO. For headspace it is giving you much weaker response. For spec 2 ppm it is usually critical

Are you headspace oven is at 105? If water is in the sysytem it should be maximum 85