Rutin&Diosmin HPLC UV reproducibil.probl. [Solved: Gradient]

[dap]

Hello to all! Sorry for my bad English

There is a Diosmin impurities&assay HPLC method in the European Pharmacopoeia. It is isocratic of 11% Acetic acid, Methanol and Acetonitrile. Diosmin is dissolved in DMSO. Nucleosil C18 column is used.
I used a consistent Nucleodur column and found the method rather fine. However any compound have only a relative time and it fluctuates a lot imho. Diosmin is eluted from 15 to 19 minute, chromatography must continue 6 more times because of Diosmin impurity F; and it is not so significant for incoming control but I have to develop an assay method for Rutin&Diosmin capsules.

I switched to H3PO4 solution with pH=3 (pH from 2 to 4 seems to change nothing, Et3N and KH2PO4 solutions too) and no Methanol and got almost the same chromatogram, of course I decreased organic content to about 15%. I added a gradient to 75% Acetonitrile after Diosmin eluted to flush all "long-playing" impurities and got rapid 20 minutes method I got some problems with separating of Rutin and fast Diosmin impurities but it isn't a message.

The real problem is Rutin&Diosmin elution times non-reproducibility: my method has the same bug as EP one has. And now it is really important... I failed to try Zorbax Rq, CN and NH2 columns and I failed to switch to Methanol. Zorbax C18 provides the same result and also the bug.

There are ACQ Method, Test and Reference solutions prints:
https://dl.dropboxusercontent.com/u/162 ... m/meth.pdf
https://dl.dropboxusercontent.com/u/162 ... m/test.pdf
https://dl.dropboxusercontent.com/u/162 ... sm/ref.pdf

I ask for any help, any information, from personal experience or literature or any other source, any assumptions and ideas on the matter, any advice or suggestion on what to do or what to look at.

[Gerhard Kratz]

There are 2 major things to say:
A) RP is never the same
B) EP, USP methods are always challenging.

To A) A standard Nucleosil C18 column is a complete different thing then a Nucleodur C18 material.
The manufacturer can explain it in more detail.

To B) Be happy that in the EP method is mentioned a column material name, usually it is only mentioned L1 column is used.

Try to dissolve your sample in your mobile Phase, not only in DMSO.
Good luck

[dap]

Gerhard Kratz, great thanks for a response!

A) Amide phase is highly recommended by lots of analytical articles writers for quercetin glycosides so I tried it. I was very happy to get an almost ideal peak (asymmetry ~ 1,1) but it was two-in-one Rutin&Diosmin peak.

B) Sad but true.

To A) Macherey-Nagel's authorized reseller answered that Nucleosil was taken out of production and that Nucleodur was next generation consistent replacement. Anyway I got the same issue with Zorbax C18 and I can't just order a new column.

To B) http://www.edqm.eu/en/Knowledge-Database-707.html usually provides such an information, see https://extranet.edqm.eu/4DLink1/4DCGI/ ... /mono/1611 (diosmin) f.e.

[dap]

Sorry, and PS.
Diosmin (0,25 g) can be dissolved only in DMFA (not very quickly), DMSO (very fine) and mixtures of DMSO and Methanol or Acetonitrile from about 33% of DMSO content to 100 (50-100 ml). Diosmin hates water at all, it can be moistened only with alkaline solutions at very hight pH value. I use 1:1 DMSO/Methanol mixture.

[James_Ball]

I don't know about the Diosmin, but we analyze for Quercetin using a Kinetics C18 column with a mobile phase of 0.2% Phosphoric Acid in Water and Acetonitrile and gradient elution beginning with a 90:10 then ramping to 70:30. We adjusted the beginning concentrations and gradient until we achieved good separation of the Quercetin peal. This column does give very sharp peaks and reproducible retention times for Quercetin(variations less than .1 minutes) over a nights run. It may work for the Diosmin too, or at least allow you to separate the Quercetin from the Diosmin.

[dap]

Thank you, James_Ball! This idea is very interesting. Long fluent gradient can make retention times consistent and also might decrease asymmetry.

[Gerhard Kratz]

Which country are you located? M&N website has still Nucleosil C18 to offer. This reseller does not do a good job.

[dap]

Belarus. Seems to be so.
Gradient looks fine, plates number increased a lot, asymmetry is fine.

[dap]

I switched to Poroshell 120 EC-C18 4,6x50 column and to gradient from 95% acidified water to 60. There are new chromatogramms of Rutin, Diosmin and new ACQ Method prints.
https://dl.dropboxusercontent.com/u/162 ... sm/rut.pdf
https://dl.dropboxusercontent.com/u/162 ... sm/dio.pdf
https://dl.dropboxusercontent.com/u/162 ... sm/asq.pdf
The method is better, time will show if it is more reproducible.

[dap]

Gradient really rocks. Thanks! Let it be solved.

[James_Ball]

Glad it worked!

Most times you will find that a gradient will produce sharper peaks than isocratic. Isocratic though does have the advantage of not needing equilibration time between runs. So the trade-off is run time versus peak shapes normally.

[lmh]

for what it's worth, I almost always use gradient separation on C18 phases for flavonoids. I've done rutin in various contexts, but never diosmin, so I can't say how they'd separate. Currently I'm using Kinetex a lot, but I don't think it's uniquely magic: any decent C18 should do.

On occasions where I get inconvenient coelution, typically I change to a phenyl phase because although it will still give coelution, different things will coelute!

In biological samples of any realistic complexity, I am always dubious about expecting perfect separation on any single HPLC phase; phenolics are just such a complex group. But often a handful of major phenolics account for most of what's present, and these can be separated, so if you're not bothered about the small ones, hplc might be enough. LC-MS is much safer, ideally with combined UV and mass spec detection.