Unusual peak shape problem.

[Jam]

Dear All,

I have a peak that is very fronting looking like a triangle, the symmetry factor criteria is set 0.6-1.4, we normally just pass the lower criteria. I am sorry, but I don't know how to attach a picture of it.
The molecule is dorzolamide hydrcohloride diluted in mobile phase A.

The UPLC method is as follows:
Flow rate: 0.5 mL/min
Wavelength: 254nm
Column: C18 (2.1*100 1.8µm) also tested another C18, slight different chemistry (2.1*100 1.6µm)
Column temperature: 35°C
Injector Temperature: 20°C
Run time: 15min
Injection volume: 1µL, so 0.4mg/mL standard solution injected into column gives 0.04µg.
Active preheater: enabled
Gradient:
0 > 2.7min 100% A
2.7min > 6.7min 100% A to 50% A
6.7min > 12.7min 50% A to 100% A
12.7min > 15min 100% A
Mobile phase A: acetonitrile : buffer solution (8:92, v/v). Buffer is 10mM phosphate buffer (pH not adjusted).
Mobile phase B: acetonitrile

Injecting even smaller volume 0.5µL didn't help at all. I know that basic analytes might have severe tailing problem sometimes if the pH of mobile phase is not correct, but very fronting peaks are unfamiliar to me. Do you have any advice?

Best regards,
JAM

[Rndirk]

What has changed between the time that it did work (under the same conditions (?)), and now?

It's explained in the sticky how you can include a chromatogram. It would be nice to see the situation how it is now and how it was before.

Fronting is a rare problem. If it's not due to overloading or inadequate conditioning, the column might be damaged or just end of lifetime.

[tkubowicz]

Hello

I'd run this method with mobile phase buffer pH adjusted to about 3-3.5
Also I think that 1ul of 0.4mg/ml std gives 0.4ug on column.

Regards

Tomasz Kubowicz

[Jam]

Dear Tomasz,

you're correct, there was a typo, ot should be 0.4microg on column.
Unfortunately the method was developed to current pH maybe to analyze all degradation peaks also and the method is validated, the change of pH would trigger extra-validation again that I would like to avoid.

[tkubowicz]

Hello

Your std is prepared in ACN (according to information you gave) so try to dilute stock std not with ACN but with water:ACN 80:20%
If your std is preapred in solvent stronger than mobile phase you can have problems with peak shape.

Regards

Tomasz Kubowicz

[Jam]

Standard is prepared in mobile phase A.
Mobile phase A: acetonitrile : buffer solution (8:92, v/v). Buffer is 10mM phosphate buffer (pH not adjusted).

[tkubowicz]

Hello

Just paste chromatogram so we can see peak.
What is the pH of your 10mM buffer?

Regards

Tomasz Kubowicz

[Hollow]

posting a chromatogramm would indeed be very helpful.

Beside this, may you provide more information regarding
retention time
Peak height
pH of buffer (or detailed preparation)
symmetries of the other peaks? Are they affected too?

even if one would expect tailing peaks, re-check all connections and diameter of tubings including and upwards injector. Maybe theres is a void somewhere or a 'big-bore' tubing installed.

[Jam]

I hope you can see the pictures. So the main peak looks triangle, but degradation peaks have nice gaussian shape. The main peak height is 0.34AU.

[HPLC chemist]

I have only encountered 'fronting' a few times. The solution is experimentation. In short the API (main peak) is not 'happy' with the conditions of mobile phase B. Try as Mobile Phase B;

1. 50% ACN. Is the peak sharper?

2. Try 25% IPA. Is the peak sharper?

3. Try 50% IPA. You know the question. You may have to use a weaker solution than 50% ACN or even go 'isocratic' and not gradient.

4. Optimize the pH.

[tkubowicz]

Hello

As I mentioned before I'd check pH and optimized it. Also I think that you could check higher column temperature (60 deg)

Regards

Tomasz Kubowicz

[Hollow]

as your compound elutes within the isocratic step with 100%A, tuning on eluent B seems useless to me

looks like overload too me and/or is really a pH issue (they may also come together)
So please specify the buffer preparation and measure pH (before addition of ACN).

Also would be nice to see the UV spectra of the main peak (and of the impurites).
Maybe the wavelenght was selected for the impurites and not for the main compound, which in fact is overloaded on column.

So next step I would do:
- measure the pH of buffer -> if it's close to or above the pka of compound, adjust to about 2 units below pka (just to see if it get's better. If you will change it definitely and revalidate is another question...)

- look at the UV spectra of the peaks -> compare with the actual detection wavelenght

- do a series of about 7-8x 1:5 dilution (can be semi-quant: to 1 ml of sample solution add 4 ml of mob phase A, mix; from this take again 1 ml +4 ml etc).
-> if it gets better, then overload of column or buffer capacity/missmatch is evident
->-> maybe two runs are then needed: one in high concentration for the impurites (with no attention to the main peak), and a low conc one for the main compound.

[Hollow]

after having a look at some wise books in my bookshelf, a pH issue seems most probable.
mass overload would be triangle with 'steep front-straight tail'

if your mobile phase pH is too close to the pka of your compound, secondary equilibria effects may/will occur, often leading to fronting peaks.

from web, dorzolamide has one pka of 6.4, which is very close to unadjusted equimolar H2PO4/HPO4 buffer with a pH of 6.8.

So some molecules are in the ionized form and are moving faster through the columns than the rest, leading to the fronting seen.

therefore trials with other mobile phase pHs are most promising in resolving the root cause. desirable would be a pH below 4.4.
But between 3.5-6, phosphate has quite a low buffering capacity. So maybe try a phosphate buffer pH at 3-3.5 (just like Tomasz suggested in first reply)
Or just try the 'generic' water with 0.1% formic acid or acetic acid.
Just as a trial, not as final method change (yet).

References
Uwe Neue, HPLC Columns, 1997, Wiley-VCH, page362-363
Snyder, Kirkland, Glajch, Practical HPLC Method Development, 2nd, 1997, Wiley-VCH, page 296-297

[Jam]

Dear Hollow,

Thank you for your multiple replies. I also thought first that the pH might be the case, but the buffer pH (measured) is 4.8, so the difference between pH of buffer and API pKa is over 1.5. And the amount of acetonitrile in Mobile phase A is quite small, so I assume it doesn't change the pH of mobile phase A much.

Best regards,
JAM

[Hollow]

you're welcome. it's also nice to learn from other's problems

ok, pH seems reasonable.

but anyway, I would give those suggestions (dilutions and run at pH 3-3.5) a try.
You know, the thing between theory and praxis...