coeluting or comigrating impurities

[juddc]

I agree heartily with those suggesting alternate methods of analysis for the compound in question, however I have one small suggestion that while not perfect, might give an indication as to the nature of the problem more rapidly than developing a second separation.

If you can, take a compound - any compound - whose purity is both well known (and high) and separate it under the conditions you're using, then process as you normally would and see if you come up with a similar error. Anything that you would consider using as an internal standard might be a good candidate.

If you have a compound that is known to be 99.99% pure, for instance, you have good signal to noise, you're not saturating your detector, your blank and gradient aren't causing any spurious peaks in the neighborhood of the peak of interest, and you're still seeing peak purity issues, you might want to investigate your data processing settings. I don't know Chemstation (I'm a Waters guy), but if there is a place in the processing method in which you have to specify an area of "quiet" baseline to subtract (as you do for Millennium / Empower: called the noise interval time) and you have that set in a place where there are peaks or dips, then you'll drive yourself insane looking for pure peaks and you'll never find any.

Much luck!

[ntruong]

Hi All,
I really need your expertise to help me solve this problem because I'm back to square one as I mentioned on my original post. The UV profile of the weak organic base compound (des-methyl) I have been working on is maxed at 247nm (secondary) and 306nm (primary) during method development. I chose 306nm for better sensitivity. When I made a test injection at the concentration of 1mg/ml, the peak profile was the same (247 and 306nm). However, when I set up sequence to run overnight (all gradient throughout) for the whole range 0.01mg/ml-1.0mg/mL, the peak profile changed (primary 230nm, secondary 306nm). I repeated the run with ACN blank inserted in between for couple of injection to check for carry-over and found none. I then switched the colum, prepared new mobile phase and repeated test. Test injection showed 247nm and 306nm profile. Sequence run showed peak profile maxed at 230nm, not 306nm observed from concentrated to diluted samples. It happened three days in a row. Any idea?
Mobile phase: initial at 65% 10mM AMA, pH 3.00 and 35% ACN
after 20 minutes: 60% buffer, 40% ACN
after 23 minutes: 20% buffer and 80% ACN, holds for 12 minutes to flush out junk.
after 40 minutes: back to initial condition
Column: Zorbax Eclipse XDB C8
Could the peak profile change due to gradient effect or some
impurity peak get squeezed underneath the peak of interest?
Thanks in advance,[/code]

[HW Mueller]

Could be that you are accumulating some gunk on the column that elutes (almost?) continually with a max at ~230.

[JM]

hi ntruong,
To me it seems stability problem of your sample. have you checked UV spectra of freshly prepared sample vis-a-vis old sample solution at same concentration ?? I would suggest to carry-out this experiment saparatly on UV-spectrophotometer as they are more accurate than DAD. may be some sort of isomaric conversion. The issue of co-eluting peak seems to be remote.

The UV Max can look different at different test concentration due to UV absorbance of mobile phase ( even more in gradients). Are you worried about the peak purity or the UV spectra of peaks in DAD??
As discussed at length in this forum the peak purity can not be confirmed by any one technique but it just give you some sort of assurance at best . I have seen even USP/BP standards fail in purity factors as it depends on how you have set the values of other factors like noise, peak thresh hold etc.

Hope this will help.

JM

[ntruong]

Hi All,
HW Mueller:

I washed the column real good with 100% water for 50 minutes, 100% methanol for another 50 minutes, and 50:50 for another 50 minutes after each sequence. I injected sample couple times, same profile (247 and 306). sequence, new profile (230 and 306). repeated the above steps, same thing.
JM:

I ran both fresh and aged sample today using UV/VIS and profile looks the same. I don't see anything unusual. I'm pulling my hair right now

[ntruong]

I guessed I am not familiar with the "quote" thing yet!

[JM]

OK that is what i expected, this indicates you dont have any stability issue . The difference in UV spectra is just cos of mobile phase and sample concentration change. Dont wory about it too much , just set the peak purity parameter correctly like noise , peak threshold and select the wevelengths range above the absorbing range of mobile phase etc.

Thats it.

JM

[ntruong]

Thanks to you all,
I guessed I shouldn't pull my hair for a little thing like this. I will move on with method validation.
Respectfully,

[HW Mueller]

If one suspects dirt coming off a column (more or less continuously) one doesn´t just "clean" it and hope that the dirt went away, but since this is a science one checks the effluent: Compare the UV spec of mobile phase (pumped through the detector without the column) to that of water and then to that of mobile phase passing the column.
(In other words, this forum would still be much more interesting if the heads were used first before asking questions, rather than the other way around).

JM, what UV spectrometer do you have? I have not yet seen one that isn´t about a 100 times less sensitive than a HPLC detector.

[ntruong]

Hi HW Mueller,
I couldn't agree more. You brought up a very good point. I cannot sweep dirt and hide it under the carpet. I will continue to look into it since it will come back and haunt me later. Why not do it right the first time? Even though my timeline for this project is due at the end of this month.
Respectfully,

[JM]

HW,
i was not talking about sensitivites but accuracy with respect to the interference from mobile phase components and lower conc of analyte during HPLC run in DAD ( No surprise why they are more sensitive?) but a stand alone UV spectrophotometer can give you " Accurate UV spectra" on account of high conc , single suitable solvent etc.

JM

[HW Mueller]

No problem if you can get your "dirt" concentrated ~ 100 times.