HPLC - help, my mobile phase is yellow/green!

[swheelz]

I am trying to analyse human plasma for fat-soluble vitamins and carotenoids using dual channel HPLC as per Thurnham et al, 1988 (modified). I am using 3 mobile phase solvents: methanol, acetonitrile and chloroform (all HPLC grade and recently opened).

However, when I purge the machine as part of the start-up checks, the purged solvent mix comes out a yellowy/green colour. All the solvents are colourless when they enter the mixing chamber.

I initially had problems with the flow rate (set at 1.5ml/min but much less was coming out). However, after much purging this has resolved and the flow rate is now OK. The discolouration issue remains though and one of my baselines (when column is connected and purge valve closed of course) is very erratic.

Attempted solutions thus far have included purging at high flow rates (up to 5ml/min) with all possible ratios of mobile phase solvents, injecting 100ul of chloroform-enriched mobile phase in through the autosampler, vortex-mixing aliquots of the solvents together to check for colourimetric reactions. The problem occurs with all of the solvents, at 100% 0% 0% or at 34:33:33 or any other ratio. We were purging all afternoon with no improvement. Very boring and very frustrating.

I have a feeling that the discolouration and the erratic baseline are due to the same problem. Is is contamination inside the machine? The thing is there isn't really much between the mixing chamber and the purge valve so at which point can the MP turn yellow?

If anyone has any suggestions as to (a) what is causing the discolouration (b) what is causing the erratic baseline and (c) how to solve these problems, your advice would be much appreciated.

I'm a nutrition PhD student, not an analytical chemist (the in-house analytical chemist is equally stumped by the way), so please keep the jargon at novice level if possible.

Thanks very much,

Simon

London, UK

[Bruce Hamilton]

First let me congratulate you for trying a systematic approach, you have apparently identified that it's not your solvents, and presumably not the degasser unit, assuming you are testing the actual lines into the mixer, and that the purge is after the mixer.

You haven't said what brand and model HPLC system you have, and that could be helpful. If there are filters in the solvent inlets and mixing valve or pump outlets, remove and replace or ultrasonicate in warm 60:40 Chloroform : methanol, repeat using warm ethanol as final rinse.

I suspect your solvents are dissolving either rubbish in, or components of, the pump or mixing valve. The chloroform may be swelling polymers ( tubing, seals, etc ) that then leach into the mobile phase with use.

My suggestions are:-
1. Consider the past usage, if similar solvent systems haven't been used recently, then thoroughly flush the whole system with a warm alcohol solvent ( on all solvent lines ). Check if the solvent is coloured.
Try to identify the spectrum of the green material to see if it's a residue from earlier work.
2. Thoroughly flush each solvent line with it's intended solvent, pass through mixer, is colour still present after mixer, but not before?. If it's present before consider the membrane degasser, if your system has one, and you've cleaned the filters as above.
3. Bypass the mixer with a piece of tubing or short column, does the colour go away?. If so, the mixer is your problem.
4. If the colour is present, trace back by bypassing the injector, and if that doesn't work, remove any column and/or column heater
5. If the above fail, then it's your solvents or their containers/filters. Try another brand.

Please keep having fun,

Bruce Hamilton

[Schmitty]

I haven't used one in a while, but I think I recall Perkin-Elmer pumps from the 250/400 series having blueish fluid that would leak out (into the solvent being pumped) when the damper was shot.

[swheelz]

I'm afraid I don't know the name of the machine or the column, which I know sounds a bit dim. The central unit which takes in the solvent is made by a manufacturer beginnning with 'K'. The machine is a hotchpotch of lots of different components (e.g. column heater, cooler, 'control unit', autosampler, external column, processing unit - unlike the Agilent GC analyser I'm used to which is just an all-in-one). The degasser unit doesn't work (leak problem) so instead we are sparging directly in the solvent bottles with lines carrying nitrogen gas that bubbles away at a low rate.

Bruce - thanks for your very useful suggestions - I will certainly try those out next week When you say 'analyse the spectrum' do you mean use a spectrophotometer or something more complicated like GC-MS? Surely it can't be the former as this only measures concentrations of known solutes?

Best Wishes,

Simon

[Bruce Hamilton]

I'm afraid I don't know the name of the machine or the column, which I know sounds a bit dim. The central unit which takes in the solvent is made by a manufacturer beginnning with 'K'.

When you say 'analyse the spectrum' do you mean use a spectrophotometer or something more complicated like GC-MS?

OK, possibly a Knaur, but all I can say is that I have never knowingly worked with an HPLC whose components brand started with K....

If your HPLC detector is a UV diode array type, you should be able to get a copy of the UV-Vis spectrum. The spectrum ( which will extend into the visible region ) may help to eliminate some of the possible sources of yellow green colour, such as oil from unclean tubing, hydaulic fluids from pumps, residual sample rubbish, dissolved polymer etc.

More information on the colour could be obtained by evaporating down some of the solvent ( one coloured, one blank ) and looking at MS, Nmr, IR etc, but that's probably wasted resources, better to try and remove the colour. However, if it's quick and easy, and you suspect that some of your components could be leaking oil or fluid, then evaporating down 50 ml samples after each module ( solvent lines into mixer, after pump, after pusle damper ( if present ), after injector, after column, after detector ) should show where the colour first appears.

I wouldn't botter sparging with nitrogen, althought it may slightly reduce oxidation of your samples. Use helium, it will prevent other problems.
Are you certain that your sparging system is clean, and that you are using suitable grade of gas and clean sinters?.

Use the WWW sites of suplliers to ensure than all of the components, tubing, and fittlings are compatible with your solvents, and that one of the items isn't resistant to your halogenated solvent mix. You could also check to see if they list possible faults that include coloured leaks.

If it is a leak of system fluids ( pump oil, damper fluid ) then the alcohol wash will not be great help, you will have to fix the items and then flush with a suitable solvent. I'm still hoping it's residual rubbish from earlier samples, but .....

Please keep having fun,

Bruce Hamilton

[tom jupille]

The degasser unit doesn't work (leak problem) so instead we are sparging directly in the solvent bottles with lines carrying nitrogen gas that bubbles away at a low rate.

???!!! Air is 70% nitrogen. By sparging with nitrogen you are "re-gassing" instead of degassing. Helium (which is almost insoluble in typical LC solvents) is the more usual sparging choice.

When you say 'analyse the spectrum' do you mean use a spectrophotometer or something more complicated like GC-MS? Surely it can't be the former as this only measures concentrations of known solutes?

I'm pretty sure he meant with a spectrophotometer. Your solvents are colored; maybe the color comes from something in a previous sample. In that case the absorbance spectra should be similar.

[swheelz]

Yes - many apologies. I meant helium for the sparging.

[swheelz]

It it definitely not a Knaur, but thanks for trying!

I think the sparging apparatus is OK as the solvents are clear going into the mixing chamber. The problem is between there and the purge valve.

I left some beakers with various amounts of 'green' in the fume cupboard over the weekend, so they should have concentrated/evaporated by tomorrow.

Bruce - thanks for all your suggestions. Tomorrow I will (a) look at the tubing and mixing chamber in the HPLC machine tomorrow and check for gunk (b) check for oil leaks, (c) compare UV absorbance of mystery substance with solvent mix using the spectrophotometer at carotenoid-specific wavelengths, (d) check compatibility issues, especially re chloroform.

Many thanks and best Wishes,

Simon

[dcree]

If you are still having trouble, I would suggest you contact Dr. Stuart Jones at Laserchrom in Kent. He is an HPLC expert and provides problem-solving services as well as products:

http://www.laserchrom.co.uk/

[DR]

Kratos?

[Mark Tracy]

Kontron?

[swheelz]

Yes, it is a Kontron.

Getting the engineer out..........

[Mark Tracy]

I once worked on a Kontron that had been shipped with acidic citrate/chloride buffers in it. It bled colorful corrosion products. I hope that is not your situation.

[Tomasz]

According to HPLC pump guide from Agilent Technologies chloroform should not be used as a mobile phase constituent. It can form radicals and hydrochloric acid which are reactive to stainless steel. You should consider change in your mobile phase.

Best Wishes,
Tomasz