phosphate solubility contributing to my noisy blank?

[mikear]

Hello,
I've been troubleshooting a rather noisy blank for almost two weeks now, basically large peaks ranging in the 10K area count. This problem is apparent regardless of system used. I'm not allowed to change the organic modifier. I think I have the problem identified, but would like to have some input if what I am seeing sounds correct. This is a HPLC method being run on a agilent 1100 (binary pump),
mobile phase A is 5mM (total) phosphate buffer, pH7 (prepared by combining 2mM monobasic and 3mM dibasic potassium salts)
mobile phase B is 100% ACN.
Sample diluent is 80% ACN/20%water (v/v).
Injection vol is 5ul
The gradient is linear from 5%B to 75%B in 13 minutes.
Peak of interest elutes at ~12 minutes (~70% organic in the gradient).

When I replace the mobile phase A with 100% water, the baseline looks fine, so I thought we had a contaminate in one of the salts or in the water. We've selectively changed one variable at time ,either repreparing the salt with a new bottle or using bottled water. The baseline always reverts to a similar level of noise when the buffer is used as Mobile phase A.

Then, I read an article from Schellinger and Carr (LCGC Northamerica, vol22, number 6, June 2004) regarding the solubility of various phosphates in 3 different organic modifiers. One conclusion reached in the paper was that phosphate buffer (at pH7) has 5mM solubility at 80% ACN. Hmmm...my diluent does not have any buffer in it, BUT it does have 80% ACN. I tried to add (dropwise) some of the buffer in a small beaker of the sample solvent and I do NOT see any precipitation...BUT I am wondering that the combination of the high organic in the diluent PLUS the organic being contributed in the gradient could be perturbing the salt enough to make the baseline noisy.

I then prepared two additional diluents: 50:50 ACN:water and 30:70 ACN: water, keeping the 5mM PO4 as mobile phase A. The purpose of these was to in a stepwise fashion, increase the water content in the diluent (blank) to test my hypothesis that the phosphate needs more water. There is a reduction in noise as I increase the water content in the diluent, but it is not completely gone.

Hypothesis: There is no contaminate causing the noise, but there is an apparent threshold I am approaching that even at 5mM PO4, which is causing the salt to "crash out" and perturb the baseline with a combination of 80% organic (or as low as 30% organic) in the diluent PLUS the organic being contributed in the gradient (~70%) . As I type this, perhaps if I am correct in this hypothesis, I could try to premix the mobile phase B to 80:20 ACN:water to additionally increase the water content at the point of mixing in the pump.

Question(s): What are the limits in terms of organic content people typically use with potassium phosphate salts? My next step was to try to use ammonium phosphate which according to the article, has greater solubility. Would the use of ammonium phosphate adequately buffer ~pH7. Would there be any significant change in chemistry of the separation going from a PO4 to a NH4 salt?

Thanks in advance for your assistance. Comments to my thoughts would be greatly appreciated.

[Gerhard Kratz]

With Phosphate buffer you need always >50% water in your mobile phase. Otherwise precipitation of the phosphate will lead to higher backpressure, noisy baseline and sometimes the column is destroyed. Please try to use an organic buffer like MOPSO to avoid that.

[carls]

It unlikely you'll see any diffrence in chromatographic behavior when switching from K+ to NH4+ salt of phosphate.

Do you see noise throughout the chromatogram, i.e. even at low %ACN? Remember, at 70% ACN you only have 1.5mM phosphate.

At pH 7 phosphate potassium phosphate has very low solubility above ~75% ACN . Can you stop your gradient at 75%? Using ammonium phosphate might solve the problem you're seeing.

Also, did you immerse the pH electrode directly into the buffer solution during preparation? If so, you contaminated your buffer with salts, etc from the electrode. Instead of contaminating the entire solution, remove an aliquot and check the pH of the aliquot (dispose of aliquot).

Try making the buffer by titrating 5mM H3PO4 with NH4OH and see if the problem goes away. You can also try filtering your buffer to see if there is any insoluable matter in your salt. I usually establish a new buffer recipe by first titrating it with the electrode in the buffer. Then dispose of this solution and remake it following the recipe. Remove an aliquot to verify the pH.

For best perfromance premix your mobile phase, i.e. A 5/95 ACN/buffer, B-= 80/20 ACN/buffer and run your gradient from 100% A to 100% B.

[tom jupille]

With Phosphate buffer you need always >50% water in your mobile phase. Otherwise precipitation of the phosphate will lead to higher backpressure, noisy baseline and sometimes the column is destroyed. Please try to use an organic buffer like MOPSO to avoid that.

Actually, that runs counter to both my experience and to the solubility data in the paper cited by the OP. That said, it is simple enough (and a good idea!) to check by mixing the proposes buffer and ACN in test tubes and then holding them up to the light to look for "sparklies".

[cody84]

We've had plenty of problems with USP method for CHG. Mostly the analysts would prime the ACN directly after the buffer and it would crash out. 2 one week old Waters systems and one older waters system all were out of business for a while because of this. Lesson, ALWAYS prime water between each solvent.

[Mattias]

Slightly off-topic: It is quite strange that phosphoric acid is soluble in 100% ACN. I guess that the H2PO4 ion has a much better solubility in ACN compared to the HPO4 or PO4 ion?

[austen]

Yes, I guess the same Mattias