-
- Posts: 32
- Joined: Fri Aug 05, 2011 11:50 am
Hello, I would like to describe a rather difficult to solve UV background problem that I could use some advise on. First I will describe the system we are using, then the problem we are having, and the attempts I have made to troubleshoot the problem.
System- Thermo Accela autosampler, 600 pump, and PDA detector. Column is C18 material and the mobile phase consists of methanol / water with 0.1% formic acid. We use gradient elution. The samples we are analyzing with this instrument are mostly plant extracts with relatively non-polar phenolic type compounds. We use a number of UV wavelengths depending on the compound but mostly 228 nm and 270 nm. The methods and samples have been used successfully in the past on other HPLC systems (Agilent for example). I am a bit new to Thermo equipment. The PDA detector was used but the rest of the equipment was new.
Problem- During gradient elution as the concentration of methanol (0.1% formic acid) increases the baseline goes up as normal but when one examines the UV spectra (200-400 nm) there is a peak with a maximum at ~236-237 nm and what looks like noise at lower wavelengths at beginning of run which eventually smooths out as another maximum at ~217 nm towards end of run. The absorbance of this peak goes up and up as the methanol (0.1% formic acid) increases. The problem with this background peak is it that it is close to 228 nm which I prefer to use for quantification of most of the compounds I am dealing with. It also makes it impossible to get a clean UV spectrum for peak purity analysis / identification because the background peak is large and gets in the way of seeing normal UV spectrum of compounds. This problem occurs in every sample I inject on this system whether its a blank, standard, or sample.
Troubleshooting- My first inclination was that this problem was an issue with contamination. So I set about a series of troubleshooting procedures to see if this was the issue and if I could solve it.
1- New mobile phases. No difference. I used different stocks of methanol and water which didn't make a difference. The formic acid is LC/MS grade and has been stored for about a year at 4 C. I have not yet tried a new batch of formic acid.
2- Flow individual solvents through PDA detector. Pure H20 = no background peak. Pure MeOH = no background peak. Pure H20 + 0.1% formic acid = no background peak. 50% MeOH with H20 = no background peak. MeOH + 0.1% formic acid = background peak. If I zero detector the peak will go away as expected but it always comes back if you start a gradient after zeroing.
3- Test each solvent channel in case there is contamination in line normally used for MeOH (0.1% formic acid). No difference the peak always appears with MeOH (0.1% formic acid).
4- Remove / replace column. New column or no column made no difference the peak is still there with MeOH (0.1% formic acid).
5- Replace UV and visible lamps. Lamps passed all usual calibration tests. No difference peak still there.
6- Replace solvent inlet filters. No difference peak still there.
7- Flush system with isopropyl alcohol to remove less polar impurities. No difference peak still there.
8- Flush system with 20% nitric acid in H2O. No difference peak still there.
After all this troubleshooting I am rather stumped. There are a few more possibilities left to try:
1- Get new formic acid. I find it hard to believe that LC/MS grade formic acid stored properly would be causing this problem. Has anyone else experienced problems like this from formic acid?
2- Replace PDA detector flow cell. This is an expensive part to replace and if it doesn't solve the problem it would be a total waste of money so I am reluctant to try this unless I am sure its a flow cell problem. Again the PDA detector was used and the problem has been with the system from day 1 it seems so that could be the problem.
3- Use different type of UV lamp. We using a CTS-A123 TSP UV 6000 lamp assembly. For some strange reason these lamp assemblies don't have the proper wire connections to connect directly to Thermoe Accela PDA detector and had to be rewired to connect. The reason I bring this up is I wonder if this wiring is part of the problem? I can't see how and I would expect if wiring was a problem the lamp wouldn't even turn on. I just found it weird that a lamp is sold as being a lamp for Thermo equipment yet doesn't directly connect to the detector without some rewiring.
4- Software problem? I again find this hard to believe but perhaps there is some setting that is not correct? Thermo software does have some detector settings I am a bit unfamiliar with such as rise time but I don't see how this could be an issue. But at this point I consider anything.
5- Pump contamination problem? I would have imagined nitric acid cleaning would have removed any organic impurities. I have been reluctant to take apart the pump and look for signs of contamination or breakage because everything else about this pump is pretty normal and I don't want to risk damaging the pump. Even if pump was problem I am unsure which parts would require replacement first. Although we do get occasional solenoid errors but I don't see how this could be connected?
Any other ideas? This is by far the most difficult and strangest problem I've ever dealt with on an HPLC. Any advise would be greatly appreciated.
System- Thermo Accela autosampler, 600 pump, and PDA detector. Column is C18 material and the mobile phase consists of methanol / water with 0.1% formic acid. We use gradient elution. The samples we are analyzing with this instrument are mostly plant extracts with relatively non-polar phenolic type compounds. We use a number of UV wavelengths depending on the compound but mostly 228 nm and 270 nm. The methods and samples have been used successfully in the past on other HPLC systems (Agilent for example). I am a bit new to Thermo equipment. The PDA detector was used but the rest of the equipment was new.
Problem- During gradient elution as the concentration of methanol (0.1% formic acid) increases the baseline goes up as normal but when one examines the UV spectra (200-400 nm) there is a peak with a maximum at ~236-237 nm and what looks like noise at lower wavelengths at beginning of run which eventually smooths out as another maximum at ~217 nm towards end of run. The absorbance of this peak goes up and up as the methanol (0.1% formic acid) increases. The problem with this background peak is it that it is close to 228 nm which I prefer to use for quantification of most of the compounds I am dealing with. It also makes it impossible to get a clean UV spectrum for peak purity analysis / identification because the background peak is large and gets in the way of seeing normal UV spectrum of compounds. This problem occurs in every sample I inject on this system whether its a blank, standard, or sample.
Troubleshooting- My first inclination was that this problem was an issue with contamination. So I set about a series of troubleshooting procedures to see if this was the issue and if I could solve it.
1- New mobile phases. No difference. I used different stocks of methanol and water which didn't make a difference. The formic acid is LC/MS grade and has been stored for about a year at 4 C. I have not yet tried a new batch of formic acid.
2- Flow individual solvents through PDA detector. Pure H20 = no background peak. Pure MeOH = no background peak. Pure H20 + 0.1% formic acid = no background peak. 50% MeOH with H20 = no background peak. MeOH + 0.1% formic acid = background peak. If I zero detector the peak will go away as expected but it always comes back if you start a gradient after zeroing.
3- Test each solvent channel in case there is contamination in line normally used for MeOH (0.1% formic acid). No difference the peak always appears with MeOH (0.1% formic acid).
4- Remove / replace column. New column or no column made no difference the peak is still there with MeOH (0.1% formic acid).
5- Replace UV and visible lamps. Lamps passed all usual calibration tests. No difference peak still there.
6- Replace solvent inlet filters. No difference peak still there.
7- Flush system with isopropyl alcohol to remove less polar impurities. No difference peak still there.
8- Flush system with 20% nitric acid in H2O. No difference peak still there.
After all this troubleshooting I am rather stumped. There are a few more possibilities left to try:
1- Get new formic acid. I find it hard to believe that LC/MS grade formic acid stored properly would be causing this problem. Has anyone else experienced problems like this from formic acid?
2- Replace PDA detector flow cell. This is an expensive part to replace and if it doesn't solve the problem it would be a total waste of money so I am reluctant to try this unless I am sure its a flow cell problem. Again the PDA detector was used and the problem has been with the system from day 1 it seems so that could be the problem.
3- Use different type of UV lamp. We using a CTS-A123 TSP UV 6000 lamp assembly. For some strange reason these lamp assemblies don't have the proper wire connections to connect directly to Thermoe Accela PDA detector and had to be rewired to connect. The reason I bring this up is I wonder if this wiring is part of the problem? I can't see how and I would expect if wiring was a problem the lamp wouldn't even turn on. I just found it weird that a lamp is sold as being a lamp for Thermo equipment yet doesn't directly connect to the detector without some rewiring.
4- Software problem? I again find this hard to believe but perhaps there is some setting that is not correct? Thermo software does have some detector settings I am a bit unfamiliar with such as rise time but I don't see how this could be an issue. But at this point I consider anything.
5- Pump contamination problem? I would have imagined nitric acid cleaning would have removed any organic impurities. I have been reluctant to take apart the pump and look for signs of contamination or breakage because everything else about this pump is pretty normal and I don't want to risk damaging the pump. Even if pump was problem I am unsure which parts would require replacement first. Although we do get occasional solenoid errors but I don't see how this could be connected?
Any other ideas? This is by far the most difficult and strangest problem I've ever dealt with on an HPLC. Any advise would be greatly appreciated.
.
