ACN/Acidic Buffer baseline "junk" problem

[randy]

Hello,

I'm having a bit of a problem with a separation I've developed. I have developed a gradient, using the scouting method from Tom Jupille, to separate the following compounds:

• maleic acid
• phthalic acid
• o-toluic acid
• benzoic acid
• citraconic anhydride
• phthalide

I was able to develop a good baseline separation for the compounds, but I've also developed a method with a very poor baseline. Here is the gradient:

A = 15mM KH2PO4, adj to pH 2.45 w/ phosphoric acid
B = ACN
flow rate = 2.0 mL/min

Time %A %B
0 100 0
1 100 0
9 65 35
12 65 35

Post time: 5 min

Column: Waters XBridge 4x150mm C18 3.5u @ 30°C
Detection: UV @ 210nm

and here is a 25uL injection of water on an Agilent 1100:



I replaced the eluents, containers and inlet frits for both eluents. Since I was using a previously used guard column, I removed it to see if that was the problem. Same results.

I've read here and there in this forum about problems with acidic phosphate buffers and acetonitrile. Should I try a gradient with methanol instead?

PS - Pardon my poor gradient table formatting. I'm not sure how to make a nice orderly table.

[Mark Tracy]

We do gradients like this also for organic acids. You probably need to condition the column with 60:40 buffer:MeCN until the baseline stabilizes. Then be very suspicious of your water quality. Have you tried an alternate source of water?

[randy]

I haven't tried an alternate source for water, but when I obtained a full-range UV spectrum for the prepared buffer, the spectrum looked clean.

[Mark Tracy]

The trouble is that junk in the aqueous mobile phase will accumulate on the column and elute later in a more concentrated lump. Whenever you work at 210 nm, the requirements for clean solvents and reagents become more rigorous. Even a dirty pH electrode can sabotage you.

[tom jupille]

randy, I'd also say that it looks like garbage building up on the column. The standard diagnostic is to run a set of three "dummy" (no injection) gradients with different pre-gradient equilibration times:
1. any equilibration time (this is just to clean junk off the column)
2. your regular equilibration time
3. longer than your regular (2x or 3x)

Compare the traces from runs 2 and 3. If the problem is coming from contamination in the A reservoir, the long-equilibration run will have significantly larger peaks, as in this example:


John Dolan wrote this up a few years ago in one of his LC Troubleshooting columns: LC-GC 16(11) 992 (1998)

[randy]

OK, I performed the suggested experiment, and it appears that the contamination is coming from the A (buffer) reservoir. So in order to determine if the contamination was indeed coming from the buffer, I decided to perform the same experiment with the same gradient program using just water from a separate reservoir. I got similar results!

The contamination peaks don't match up completely, but a majority of the stuff is in the water also. It looks as though I need to try a different source of water. We don't have an in-house filtration system, so we have to order our water. Up until now, the water has been satisfactory for our IC and LC work.

I am also going to obtain a 190-400nm UV spectrum for each compound to determine if I can increase the detector wavelength and maybe minimize the problem.

Comments?

[Richard987x]

The quality of water is the largest source of analytical error while your buffer introduces additional variability- My suggestion for you is:

1. Buy commercially available water, B& J-
2. Use highest HPLC grade ACN from Fishers
3. If (1) and (2) are not sufficient, you can filter your mobile phase A offline using 3 M Empore carbon Disks and use the filtered mobile phase.

[DR]

Minor change to ^ - Use Empore extraction disks to filter your water, then prepare your A-phase. Having anything else in the water seems to help get the stuff you don't want to pass through the Extraction disk, which defeats the purpose of using it - it is most effective if used on pure water.

Also - I'd try the Empore disks first, especially if you have more time than $ (B&J, Optima et al brands of water get expensive and frequently don't solve the problem at <215nm).

LC grade Phosphates are usually pretty clean, but it wouldn't hurt anything to filter your A phase through a nylon filter after dissolving the PO4.

Also - isn't a pH <3 a bit low for PO4 to have any buffering capacity?

[randy]

Not according to this http://www.mac-mod.com/tr/pbprep-tr.html

[randy]

Also, I tried the gradient at 240nm and am still getting garbage out around 7-10 min.

Is there a way to start a gradient on an Agilent 1100 controlled by Chemstation without making an injection? I haven't yet found a way. I had to make injections when performing the suggested experiment above, so i guess I can't eliminate the injector as a contamination source.

[danko]

Just inject 0 μL

Best Regards

[randy]

I found another way as well. In the Sample Info option in the Run Control menu, leave the Sample Location blank and then click Run Method.

[randy]

Well, I've been busy in the lab but finally found the solution to the problem a few weeks ago. Turns out the problem was in the water. Once I used HPLC-grade water the junk peaks were drastically reduced and in most cases eliminated completely. Luckily the few remaining peaks did not interfere with my target analytes.

[tom jupille]

Glad you got the problem solved, and thanks for letting everyone know the outcome.