Light scattering is the preferred technique (ELSD or CAD) to use and when combined with a good gradient method, will successfully resolve them apart with little effort.

I have lots of experience analyzing phospholipids by HPLC using many types of detection systems (e.g. UV/VIS, RID, ELSD, CAD, MS). Your approach of using an RID and/or PDA detection system with underivatized phospholipids is not likely to be successful. Most phospholipids will not be detected at all, even at high concentrations at low UV (i.e. 205nm) wavelengths. Response by UV will also be dramatically different for each one (from high to none at all).  Some of the preferred mobile phases needed to dissolve these compounds have a high UV cut off limiting detection.  An RID can detect many of them, but their concentrations will need to be very high. However, RID can only be used with isocratic methods and you really need a proper gradient method to resolve different phospholipids apart. Without knowing exactly which phospholipids you are trying to resolve apart, it will be difficult to suggest how to proceed with the tools you have available to you.

• I have had some success using a plain silica column and a simple isocratic mobile phase of ACN/MeOH/Formic Acid (90/9/1) on a silica column (UV 205nm). This may be your only initial option to try. A gradient will work much better, but if you run isocratically, then you can use the UV/VIS detector inline with your RID.
• Derivatization may also be an option for you. This would allow you to see them using your photodiode array detector. 

*One tip I will share for use with these types of methods is to always wash down any silica column (cyano too, but a cyanopropyl column is a poor choice for this application. Try a plain silica or amino instead) after each run with a methanol or IPA wash mobile phase. The lipids quickly clog up the columns and this alcohol wash helps reduce the fouling over time. This step can be incorporated into the end of the gradient, but I prefer to implement it as a separate wash step, after the analysis method is complete.

Also of concern is this statement; "The problem is I always get a very high peak around half way of the run. It almost reached maximum absorbance even when I injected solvent only, without any analyte." - This implies that you have a fundamental error in your technique. No peaks should be seen when injecting the mobile phase into the column (except a pressure pulse or RID blip) unless the column still has retained sample or is fouled. Try and also use a stronger solvent as a wash solution after running each method to make sure you are eluting off all compounds (as noted above). A well developed HPLC method must elute ALL peaks off the column.

Article A Rapid Method For Phospholipid Class Separation by HPLC usi...

Light scattering is the preferred technique (ELSD or CAD) to use and when combined with a good gradient method, will successfully resolve them apart with little effort.

I have lots of experience analyzing phospholipids by HPLC using many types of detection systems (e.g. UV/VIS, RID, ELSD, CAD, MS). Your approach of using an RID and/or PDA detection system with underivatized phospholipids is not likely to be successful. Most phospholipids will not be detected at all, even at high concentrations at low UV (i.e. 205nm) wavelengths. Response by UV will also be dramatically different for each one (from high to none at all).  Some of the preferred mobile phases needed to dissolve these compounds have a high UV cut off limiting detection.  An RID can detect many of them, but their concentrations will need to be very high. However, RID can only be used with isocratic methods and you really need a proper gradient method to resolve different phospholipids apart. Without knowing exactly which phospholipids you are trying to resolve apart, it will be difficult to suggest how to proceed with the tools you have available to you.

• I have had some success using a plain silica column and a simple isocratic mobile phase of ACN/MeOH/Formic Acid (90/9/1) on a silica column (UV 205nm). This may be your only initial option to try. A gradient will work much better, but if you run isocratically, then you can use the UV/VIS detector inline with your RID.
• Derivatization may also be an option for you. This would allow you to see them using your photodiode array detector. 

*One tip I will share for use with these types of methods is to always wash down any silica column (cyano too, but a cyanopropyl column is a poor choice for this application. Try a plain silica or amino instead) after each run with a methanol or IPA wash mobile phase. The lipids quickly clog up the columns and this alcohol wash helps reduce the fouling over time. This step can be incorporated into the end of the gradient, but I prefer to implement it as a separate wash step, after the analysis method is complete.

Also of concern is this statement; "The problem is I always get a very high peak around half way of the run. It almost reached maximum absorbance even when I injected solvent only, without any analyte." - This implies that you have a fundamental error in your technique. No peaks should be seen when injecting the mobile phase into the column (except a pressure pulse or RID blip) unless the column still has retained sample or is fouled. Try and also use a stronger solvent as a wash solution after running each method to make sure you are eluting off all compounds (as noted above). A well developed HPLC method must elute ALL peaks off the column.

Article A Rapid Method For Phospholipid Class Separation by HPLC usi...

Hi Bill,

Thank you very much for your clear answer. I actually looking on all phospholipid components in my sample. I want to know the composition of each phospholipid class especially Phosphatidyl inositol, Phosphatidyl ethanolamine, Phosphatidyl chloline, Phosphatidyl serin, Phosphatidyl glycrol. I will try the mobile phase that you suggested. Using the mobile phase you suggested, what gradient can I use? can I increase the methanol in the mobile phase to sepate the phospolipid?

Could you tell me the best derivatization method so that I can get the signal for those compounds? I read in some paper that the derivating agents are different for each phospolipid class.

Hi! I haven't used cyanoproyl columns yet but I adopted and adpated a DIOL column to work under HILIC conditions for the separation of lipid as classes.

You can have a look at my pubblication "Anesi A. and Guella G. 2015. A fast LC-MS methodology for membrane lipid profiling through HILIC chromatography. J. Chrom. A. 44-52. DOI: 10.1016/j.chroma.2015.01.035". You might try to use your CN is HILIC mode under gradient elution and adapt the method just to phospholipids!

Best regards

Andrea

Hi,

I have a lot of experience in phospholipid analysis. In my profile you can find a paper of a method published in Journal of Chromatography A that I developed using a Chromolith column, but it works perfectly also with a silica column of 250 mm. My advise is to forget the amino column because it needs to be regenerated every few injections.

Let me know if you try it!! Have a good day.

Regards,

Monica 

I'd argue that you need to test different solvent systems. You say that in the first system the material wasn't bound to the column and then changed systems to something else entirely.  Why did you do that?  Not being bound to the column would be a factor of the material and not the solvent system.  These are the things I'd recommend thinking about when you consider how to change it.  What I'd recommend is a slower solvent system than the first one you tried and a one with better separation than the one you tried later on.

Furthermore, I am not sure hexane is a good basis really.  It is the classic apolar solvent and phospholipids have an odd relationship with it.  Can your system cope with chlorinated solvents?  If you don't have something that can give you structural information on your compounds, even if you do resolve them, I think youa re going to have a hard time convincing a reviewer that you have separated compounds well, however good the LC trace looks.  Sorry, but I have seen it too many times, checked it and found it wasn't as rosy as I thought :(