DMSO

[mojo]

We are developing a method for a very poorly soluble compound. A method exists for a 200mg dosage, but it involves use of 2 litres of acetonitrile to test 1 dosage unit. We do not want to use 10 litres of ACN to test each dosage unit as we progress to a 1000 mg dosage. We have looked at DMSO as an alternative solvent (dissolve in DMSO then dilute in ACN), and inital results looked promising. However, when we tried to determine accuracy, the recoveries were 96%, 97% and 92% at 70, 100 and 130% of the nominal concentration. I know DMSO can be problematic, but is it likely to be responsible for poor recoveries? We are looking at increasnig the DMSO volume to aid dissolution of the API.

[JGK]

Are your STDs prepared in ACN only?

If they are, try preparing STDs with the same percentage v/v DMSO present in the diluted samples and testing recovery.

It may be that the DMSO content in the sample is "depressing" analyte peak response.

[mojo]

Cheers. We are using the same DMSO v/v in std. Currently reviewing extraction technique, looks like it may be the culprit!

[adam]

I've always wondered how DMSO would work as a diluent for reversed phase HPLC. It's generally a more powerful solvent than methanol or acetonitrile. But I wonder if there would be problems with the injection volume. In other words: would you be able to inject the same volumes as you would with ACN or MeOH before you see peak distortion?...would you have to inject less?.....could you inject more?

I've never gotten around to trying it - has anyone else got experience with this?

[Uwe Neue]

Due to the very pronounced viscosity maximum of DMSO-water mixtures, DMSO is only good for low-volume injections. Dilutions with acetonitrile have been used to alleviate the problem without causing solubility issues.

In the worst cases that I have seen, a large DMSO injection carried a significant amount of analytes with it, eluted as unretained peaks, while the remainder of the (gradient) chromatogram showed little peak distortion. Such a situation can throw you off completely, unless you use MS detection.

[mojo]

The original method we used for this compound used acetonitrile as the solvnt only. When we had dmso present at about 1%v/v, the peak shape was poor, so we halved injection volume and it was fine.

[Uwe Neue]

Going back to your original question: I can't see why DMSO at such a low concentration would cause the accuracy problems described above.

[totosr]

@mojo

When possible try to change the pH of the sample acidify, for ex.
Many compounds are poorly solube for ex. in water but in protonated form much more easily.
Have tried to dissolve your compound in THF? Normally THF don't makes distortions similar to DMSO.

[JGK]

Just a thought: You're not temperature controlling the sample compartment are you?

If you are DMSO may give you problems with its high mp