One other troubleshooting you might try is varying the injection volume. If the contamination is in the water you should see the peaks change size in proportion to your injection size. If the peaks are from your system the size should remain the same.  You could also find another source of water to inject (no change in mobile phase water).  Unless that water had the exact same contaminants, it should change what you see.  If there is no change, your problem likely lies with your system.

Hi Juliana,

I guess you have looked already at some earlier discussions on researchgate about negative peaks on HPLC, and also did a search for troubleshooting on internet. Still not found an answer? The only thing I can imagine for the moment is the use of water: are mobile phase and your blank both degassed? But I'm not sure whether this could cause your problem. Another solution - if possible - might be to change your settings to shift your sugar peaks to greater RT's.

Hi Marcel,

Yes, I checked all the discussions, but my case is really particular because I have water over water. So, all the explanations they mentioned in those discussion do not apply. 

Yes, they are both degassed and my equipment also has an online degasser, so that should not be a problem. 

I supposed I can make that change, but I would like to know the reason, because I have used this setting before with no problems, so I think something has changed in my equipment that is making this to happen and I want to fix it. 

There might be a possible contamination in the water you are using. I faced similar kind of problem once, and changing the eluent and the water thoroughly to HPLC grade water (that comes in bottles), was the solution.

As Dr. Mukherjee has suggested  it could be a contamination problem both in sample and mobile phase. But to my mind it could be due to other reasons also. I need some more information for this.  Are you getting the negative peak for the sample sample at the same position in different runs? Or is it at randompositons?

Hi Juliana,

Are you connected to MS? Sugars usually are detected in neg mode in MS.

If its HPLC question(when not connected to MS): Do you see the drop in absorbance at elution time of sugars. Do you use DAD or RI detector? (Most probably RI).

RI will measure the refractive index of anything in the cell and it will bend the light if the refractive index is changed (more in one way and less in another way). That is why is possible to see a negative peak. You assume that the water in the blank is the same as the water in the mobile phase. When I have an issue in my experiment, I would not assume anything until I prove it. For what I read, you use only water in the mobile phase and only water in the blank. I don't know how you make the "blank". To troubleshoot, I would take the water from the mobile phase and inject to the system as blank. I should see no peak at all. Then inject "your blank". If you see negative peak in your blank, you have to ask yourself or explain to me more of how you make the blank, so we can go to the next step. 

Hi Juliana. Could you let me know what is your detection wave length?

Wael

The detector is RI not the UV, so no specific wavelength.

Oh. Thank you Narong. Sorry I did not notice that!

I didn't read it explicitly stated above, so I'll ask if the water being injected is from the same source as the mobile phase? What happens if you inject your mobile phase?

If injecting mobile phase water causes the same negative peak, there is something in either your pump lines, pump, or someplace prior to your injector. The clean water is displacing the contaminant and causing the drop in baseline causing a negative peak. If something was bleeding off the column, or anyplace after the injector, I would not expect a change since the injected water would dissolve whatever is bleeding off after the injector too.

If injecting some of the mobile phase does _not_ cause the negative peak, it suggests that the mobile phase is contaminated somehow and should be replaced.

Have you considered the formation of Cl(-) adducts -if you employ HPLC/MS-?.

Thank you all for helping. The negative peaks (two of them) are at the same position each time I inject. The "blank' was a samples of water that I took of the same bottle of the mobile phase, filter through a 0.2 um filter and inject it, so it is exactly the same thing. Water used is HPLC grade. As many of you notice, it is RI detector

Ok, you use mobile phase as the blank but you filter it. try not to filter, put the solvent from the bottle into the autosampler vial and inject. You may extract somehting from the filter. Leave the filter out of the equation. Also inject air to see if you see anything, because it may come from the injector vial. Let me know how it goes.

So, I think the contamination before the injector makes a lot of sense as Jack mentioned. I am cleaning/replacing tubing and filters, and purging autosampler to see if I can solve it. I will let you all know if it works. Thank you and if you have any other suggestion, it would be my pleasure to explore it. 

Injecting air will rule out that anything comes from your water or filter. Again do both, it is a quick check. Isn't it fun to do method troubleshooting?  Keep your neuron firing up.

please degas your injected blank and samples . it could be dissolved gases .

if this will not work use an other HPLC method

Hey, Juliana, I think the negative peak might come from water. Sometimes it happens. If these peaks did not disturb other peaks, just leave them there. It's okay. 

If your problem persists after cleaning try reversing your column or use different column.

Prepare samples in mobile phase and then please check

One other troubleshooting you might try is varying the injection volume. If the contamination is in the water you should see the peaks change size in proportion to your injection size. If the peaks are from your system the size should remain the same.  You could also find another source of water to inject (no change in mobile phase water).  Unless that water had the exact same contaminants, it should change what you see.  If there is no change, your problem likely lies with your system.

As mentioned by several persons, dissolved air can cause a negative peak with an RID.  However, if it is interfering with your analysis then you do not have sufficient retention of your analyte.  Air should elute at the dead volume.

I like Jeffrey idea.

The travelling solute occupy a space on column, and displace the solvent; so you have two things reach the detector; [1] the vacant spot at which displacement happen, and you got negative peak relative to the start autozero point; [2] peak due to your compound.

(1) I would talk to the people who normally service your HPLC instrument

(2) Changing the column might eliminate the overlaps that bother you

There may be various reasons for negative peaks in HPLC diagrams, like impurities or air bubbles in the eluent.

In case of RI detection, negative peaks are quite common as a difference in reading (eg conductivity or Refractive Index) is measured. The main reason for the occurrence of the negative peak is that the refractive index of the solute is less than that of the mobile phase. To avoid negative peaks, you could reverse the recorder leads.

Using a refractive index detector, an eluent with as low a refractive index as possible is used, so that the majority of peaks come out as positive peaks.

Hey guys,

I want to tell you want I did. I did inject air as Dr. Chamkasem said and I got the negative peaks too. I injected different sources of water and I got the same peaks. So, I cleaned all the equipment upstream from the injector and I almost got rid of the peaks. They are still there, but they are almost imperceptible now and they are not altering my measurements anymore (they got reduced by almost 50X). So, I think it really was a contamination at some point upstream of my injector.  

I did not vary the injection volume because I have already started cleaning when I say the answer of Dr. Rohrer. But I will take that suggestion into account next time. 

Thank you all for the help. 

Using an RI detector is very sensitive to refractive index changes in anything.  Jeffery's ideas are good and would add the fact that if peak height does not change with injection volume then possibly a seal in the valve could be going, or if this is a manual injector try turning the valve very fast and slow to eliminate a valve problem, see how the detector responds to the pressure change during injection.

So, looks like the issue is coming from anything before the column. You can do is method of elimination. I don't think it is from the pump because it occurs "only" when you do the injection. The pump is pumping all the time. I would assume you run isocratic (RI will change response if you do gradient because the refrative index of the mobile phase will change over the gradient run). You can switch the injector or use manul injector (Rheodyne) that you know it is clean.  Once you swap the part, you will know if that part is a problem. When you injected air and see the peak, it may beccaused by the movement of the actuator that may dislodge something from the injector part or contaminat that remain in the injector.  It is a process of elimination. Sound like a fun project that all the experts all over the world trying to help you. What's an honor.

Hi!! I am having exactly the same issue. What do you mean exactly by cleaning before the injector. I am thincking about purging the autosampler line, but what else? You mean replacing tubing?

Hi Anna Burniol-Figols . If I remember correctly, I ran a solvent (methanol or isopropanol maybe?) through the equipment without connecting it to the column. I think I might have purge the autosampler with it too. Then, I changed back to my mobile phase (water), and purge some more, to get rid of the solvent. I think I run the solvent at a high flow rate, to make sure it cleaned very well.

Sorry if I don't remember the details very well, but you can try that. Good luck!

Ok! I understand... It is been a while.... I will try! Thanks!!

Anna

Let me know how it goes Anna Burniol-Figols . If needed, I look for my old lab notebooks to figure out exactly what I did.

I actually discovered a leak in the autosampler line. We had a service from the company changing the rotor of the low pressure valve, and he accidentally forgot one of the connections loose.... I am still running samples and making sure that it is solved, but by now it seems so. Thank you so much for your input though! These forums always have way more answers that the HPLC providers itself!!

I agree with @ Jeffrey S Rohrer