I have a problem with an HPLC ammonium phosphate/methanol gradient method, wavelentgth of 260 nm. Two unknown peaks are eluting at approximately the same retention time as 2 known peaks. An investigation has been on-going for 3 months to determine the source of the unknown peaks. The peaks are very random and have eluted even while injecting water only. Several sources have been ruled out as the cause of these two peaks: (1) Glassware contamination has been ruled out as the cause because the unknown peaks were observed in injections where water vials were filled directly from the water source (milli-Q) where the water only touched the vial and then capped. (2) Water source has been ruled out as the cause because the peaks are not consistently observed in all water blanks, even though the vials were all filled at the same time, from the same water source, filled directly into the vials, injected on the same run. (3) Other sources of contamination have been ruled out because the peaks are not reproducible in size. The size of the peaks varies, even from multiple injections from the same vial. Also, the peaks may only be present in 1 or 2 injections out of multiple injections from the same HPLC vial. (4) Individual HPLC column has been ruled out as the source. Over 5 different columns (same type and manufacturer) have been used, all showing the peaks. (5) Individual HPLC system has been ruled out. Six different Agilent systems have been used and one Shimadzu, all showing the peaks, randomly. (6) Carryover has been ruled out. The peaks are observed in water injections where the prior blank is clean. Additionally, there is no consistency for which injections would be causing carryover. For example, an entire run of water blanks showed the peaks in only the 35th vial. All others were clean. Also, the spectra showed the peaks not to be related to the known peaks, so the peaks are not carryover. (7) HPLC vials/caps have been ruled out as the source of the peaks. Several different vials and caps were evaluated. All showed the peaks, randomly. (8) Ghost, artifact, or dissolved air peaks: The peaks are not disolved air, as they would not absorb at 260 nm. Also, no correlation was observed with oxygenated air and increased peak size.
Additional information: (1) The two peaks do not always show up together. Sometimes only one or the other are present. (2) A total of over 450 injections of only water were injected with the peaks not showing up at all, over the course of several different runs, different batches of mobile phase, ect.

Strange peaks in gradients may be caused by material carried onto the column with the mobile phase used at the start of the gradient. Common sources are the water, the lines connecting the solvent bottles to the HPLC.
In order to determine that this is the cause, equilibrate your columns with the gradient starting composition for different amounts of time. If the peaks increase with time, this is the source of the problem, and you can troubleshoot it further from there.

Microbiological growth in the In-Line degasser can also be a source of contaminants Uwe Neue mentioned.

Good luck!

Change the purge frit and online filters. Sometimes peaks can leak from filters.

Have you ran any gradients without injecting anything? At least in chemstation you can do it by leaving 'location' empty in a run table. If you still see the peaks they are probably from the mobile phases. Try other methanol and phosphate supplier and other water source.

Change one thing at the time and always go back if you don't find anything.

If you see the peaks try to reproduce them by few injections. Then change the conditions (e.g. mobile phase strenght or column temperature). Check if the ghost peaks move in the same manner as your analytes would. This showes you if they are 'real peaks' from RP chromathography. Some compounds can retain to impurities in column packing material and will not move around on RP basis. Change the column supplier (packed using Fe and Al free silica).

It is almost always the water, not the water you inject, but the water in the mobile phase (as Uwe quite correctly notes). I have also seen material liberated from different batches of membrane filters used to filter solvent. Contaminated methanol and ammonium phosphate are also not unheard of.

I'm very interested in the solution:
did you detect the source of your problem ?

Hello- hoping someone can explain the difference between a gas chromatograph and a liquid? Also if you can give me a briefing on the whole process would be great- I’m a Trainer for a Customs Brokerage and this is being imported alot lately, trying to get info on this subject is difficult. Thanks

Dear Anonymous,

I would first suspect your reagent quality, phosphates buffer reagents are notoriously impure.
Try filtering your aqueous phase through a DVB filter for amazing results - see LCGC, North America, 21 (2) 168 - 178. Also you are providing a superb environment for bacterial growth. Replace your inlet filters, glassware, and column, clean your machine with MeOH, dichloromethane, MeOH. Drop your buffer conc. or try it with no buffer, but your instrument and column are probably already infected. Your water supply could also be infected, you need superb quality water and this can be random. Also if you have a pump seal wash this can also be infected and cause these problems.