Downward peaks

[rima patel]

can someone tell me when does the sample show downward peaks and what does it signify, in an HPLC analysis.

[tom jupille]

Depends on the detector and the separation conditions. We would need more details regarding your system.

[rima patel]

the column i am using is c-18 and the detector is UV

[DR]

go on....

(wavelength, mobile phase composition, sample diluent composition, inj. vol., flow rate etc.)

[Srinivas SK]

Hello Rima

If by "downward peak" you mean a negative peak, then it means that your peak of interest has a UV absorbance that's lower than the mobile phase it is in.

One can check this up by changing the detection wavelength to a higher wavelength.

If you still have negative peaks, try changing the mobile phase, or if you're running a buffer, try changing the pH.

Sometimes if you're running a gradient that's too steep, you will get weird peaks, owing to refractive index effects in the UV flow cell. Slow down the gradient if so and try.

If you still have several negative peaks and if this occurs with other samples as well, one needs to check the connection from the detector output to the data system. Some detectors use analog outputs and if so, the output leads may have been reversed. Sounds silly, but it does happen.

Hope this helps.

Best rgds,

S.K. Srinivas, MPharm
Bangalore

[rima patel]

the colum is C-18, the detector is UV , the wavelength is 254nm, the flow rate is 1.00ml/minute, the injection volume is 10ul, and the mobile phase is water:acetonitrilr:methanol.
the drug is dissolved in methanol as mentioned in a concentration of 2mg/ml.
the peaks are showing a negative response.

i ran another drug sample under these same conditions they gave positive responses.

[JA]

I wonder if the component of interest has a chromaphore.

Srinivas: Can you elaborate on the statement "One can check this up by changing the detection wavelength to a higher wavelength." with respect to how this helps in troubleshooting.

My understanding is that the deflections often seen early in a run, and commonly referred to as the solvent front, will disappear at high wavelength since typical solvents and additives are unlikely to absorb up there. While this tidies up the front of your chromatogram, you run the risk of moving too far away from the absorption maxima of your sample leaving you with a nice flat baseline.

If the deflections are down to some sort of refractive index phenomenon, are they also likely to diminish at high λ?

Open question: Is it actually possible for a chromaphoric analyte to elute with a negative UV peak under typical conditions? Naively, I had just assumed anything eluting (which absorbs say, somewhere in 210-400 nm) would give positive peak due to additive nature of sample + eluent absorbance vs. eluent alone. Any examples?

[tom jupille]

As far as I know, there are four things that can cause negative peaks with UV detection (most of them have already been suggested, but I'm trying to be comprehensive here )

1. Polarity of the output has been accidentally switched. Rima, if all of you peaks are now negative, that is the first thing I would check.

2. You have a transparent analyte (at a particular wavelength) which displaces a UV-absorbing mobile phase component during the separation. This is sometimes done on purpose when things like Cl- or SO4= are separated using an aromatic buffer. The technique is often referred to as "indirect UV detection". Given that Rima is using only water/MeOH/ACN and detecting at 254 nm, this seems like an unlikely possibility. Just to be sure, though: Rima, you have run UV spectra of these analytes and know they absorb at 254 nm, right?

3. You have a transparent analyte which changes the refractive index of the mobile phase enough to generate a negative deflection. I know, Rima is using UV, but UV detectors do respond to RI changes. As with the previous paragraph, I'm assuming that Rima's analytes do absorb at 254 nm.

4. You have a fluorescent analyte, with an emission wavelength at which your photodetector is more efficient that it is at the excitation wavelength. This is rare.

[amaryl]

[quote="Srinivas SK"]Hello Rima

If by "downward peak" you mean a negative peak, then it means that your peak of interest has a UV absorbance that's lower than the mobile phase it is in.

One can check this up by changing the detection wavelength to a higher wavelength.

Could you please elaborate on this sir.

Regards,

Amaryl.

[Srinivas SK]

Hello All:

Thought I'd clarify on the "shifting to a higher wavelength" issue I brought up, since more than one person has asked about it.

Rima is using a detection wavelength of 254 nm, at which wavelength she's getting negative peak(s). If this is happening only with the peak of interest, it usually means the UV absorbance of her analyte is lower than the mobile phase, at that particular wavelength. So, when this sample band passes thro' the detector flow cell, you may get a negative signal.

One can confirm this by changing the detection wavelength to a higher wavelength, say, 270 nm, for instance. If the compound has a chromophore that absorbs at this wavelength or thereabouts, we'll get a normal peak, confirming our theory.

If one has a dual wavelength detector, like the old Gilson 119, then it's easier to do this. Or if one has a time-programmable ( or tunable) detector, then one can change the wavelength during the run.

If one is lucky to have a PDA, then of course there's no need to re-inject the sample, since the PDA will yield absorbance data at all wavelengths.

If all else fails, and negative peaks still appear, then it's time to look at the detector output leads and see if they are connected correctly.

If the leads are OK and we still have -ve peaks, then one can use the software to cheat a little and change the polarity of the signal during the run, at the right time. If I'm not mistaken, I think one can do this on both Waters and Shimadzu. Not sure though.

If one still gets negative peaks, no matter what, well then, I suppose we'll just have to grin and bear it ... 'tis the season to be jolly after all!!

Hope this clarifies the issue.

Warm rgds and Seasons Greetings to All.

S.K. Srinivas

[JA]

Srinivas, thanks for your repost but as they say in America "I'm not buying it"

The way it was presented, I read your suggestion to move to higher λ as a troubleshooting solution to the negative peak. I think if an operator did this and it worked, they got lucky. Perhaps they were at the wrong wavelength to begin with, or are already pushing the long wavelength side of the analytes' absorption spectrum (assuming, as I previously wrote, that it absorbs in UV/vis). Without knowing what this spectrum looks like I might equally suggest someone move to shorter wavelength to look for an advantage over the eluent.

The remaining points in your post from Dec 16th I read as you having direct experience of circumventing this type of issue. Fortunately, I am yet to see a negative UV peak from a chromaphoric compound. Out of further interest do you recall any examples?

Thanks

[Srinivas SK]

Hello JA:

I do have some amount of experience in LC problems - about 20 years of it. And still learning.

Actually, I did forget to mention that one can also shift to a lower wavelength, as well. Problem is, as one shifts to low UV, baseline noise tends to increase. Which is why I prefer to move upwards, from 254 nm.

Of course, the logical thing to do would be to take a UV spectrum in the first place, but if one has a PDA, this can be done during the LC run itself.

I've tackled negative peaks before - with amino acids in serum, which I found was due to the gradient we used. Same thing happened when I was developing methods for multivitamin analysis. Baseline went down, down, down. Same reason - overambitious gradients.

We'd get neg peaks with catecholamines in serum, using electrochemical detection. I spent a whole week trying everything in the book and finally had the sense to look at the rear panel - and found the detector output leads were reversed.

In UV detection, neg peaks turn up with steep gradients, abrupt changes in flow rate and dramatic changes in mp composition, because of refractive index effects. UV detectors are sensitive to RI changes, no matter how clever the flow-cell design is.

And they happen a lot with RI detectors of course.

Tom Jupille is right when he points that this may happen when the analyte is a UV quencher. Simple way to check this out would be to run a TLC of the sample on a GF 254 tlc plate.

Rima, I'm sure you've got much more than you bargained for !!

Hope all this is useful.

Merry Xmas all

[amaryl]

If the leads are OK and we still have -ve peaks, then one can use the software to cheat a little and change the polarity of the signal during the run, at the right time. If I'm not mistaken, I think one can do this on both Waters and Shimadzu.


Please elaborate this point Sir.


Merry Christmas to all

Regards

Amaryl

[jitender]

@ Rima- Not relevent with this discussion though...........What you are doing with the concentration of 2mg/ml.

I was of opinion we can integrate the negative peaks in chromatographic software.

Negative Peaks remined me of "VACANCY CHROMATOGRAPHY".

[amaryl]

@ Rima- Not relevent with this discussion though...........What you are doing with the concentration of 2mg/ml.

I was of opinion we can integrate the negative peaks in chromatographic software.

Negative Peaks remined me of "VACANCY CHROMATOGRAPHY".

Good observation Jiten. Only Rima can highlight what concentration she is injecting ...quite possible thats her stock solution which is diluted further down before injection.

What's this Vaccany chromatography? some good link to this?

Regards,

Amaryl.