C18 column problems due to ammonium bicarbonate buffer?

[DM909]

Hi all

I am using the following mobile phase:

Mobile Phase A: 10 mM aqueous ammonium bicarbonate containing 0.2% triethylamine, pH 7.5 (adjusted with formic acid)
Mobile Phase B: Acetonitrile
Gradient: 90% A (0 min) - 90% A (6 min) - 70% A (21 min) - 70% A (50 min) - 90% A (55 min) - 90% A (65 min)

We are using Waters Xterra C18 and Xbridge C18 columns (4.6mm, 150mm). After less than a 100 runs we start experiencing peak splitting or peak-shoulders on the high RT end of the peak and increase in column pressure, rendering the column useless. We've gone through 3 columns in so many months.

Anyone have any insight? Is there something wrong with this buffer?

[Uwe Neue]

My first suggestion is to measure the pH again after a few hours. It may have crept into the alkaline range, and you are not really using a buffer. I had believed myself that it is possible to make a stable carbonate buffer at pH 7, but it did not work (how do you degas, for example).

The other thing that could happen is the adsorption of contaminants in your sample matrix.

[JGK]

I have worked on methods using a filtered sodium bicarbonate solution where, on preparation, a specific pH (8.2 ± 0.1) was a requirement . However, this solution was not usable for more than 1 Day as the pH had "drifted" outside the valid range.

Not sure, but similar could be happening with your Ammonium Bicarbonate solution.

[DM909]

Uwe: We degas the aqueous buffer and also use an in-line degasser on our HPLC.
Yes, it is true that we do get pH drift but I would think that this would only effect RT, if anything, and RTs are relatively stable.
What I am concerned about is the short lifetime of our columns. We've busted 3 columns in three months and each one shows the same characteristics - peak shouldering, followed by pressure increase. In one case, we've had to replace a column after only one week of use. Since then we've switched from Xterra to Xbridge and the situation is still problematic. I can only conclude that it is the buffer or maybe incompatibility between buffer and analyte.

[danko]

.....maybe incompatibility between buffer and analyte.

That’s a step in the right direction. Backpressure increase indicates – clearly - that something is sticking on the column inlet frit or the stationary phase. So does peak deformation.
I don’t know what your analyte is, neither what the rest of the sample consists of, but similar problems are not at all unusual in protein separation context – especially when the mobile phase pH and the pI of the protein are too close.
If I were you I’d perform a simple solubility investigation (visually for a start) and see whether solubility is the problem.
The simplest procedure is to mix mobile phase and sample solution in a test tube and observe the mixture for a while. If the problem is solubility, you’ll see some precipitation going on – or maybe some cloudiness.
Ones you’ve arrived to a conclusion, the fix is quite easy/obvious.

Best Regards

[Uwe Neue]

I agree with Danko. Peak shouldering followed by pressure increase indicates contamination of the column with something from the sample matrix. The fact that the buffer is not stable may be irrelevant.

[DM909]

Thanks guys. We will look into this in detail. Of course, part of our analytical method development is the stability of the sample and its solubility in the diluent and mobile phase. We're working with an pharmaceutically active ingredient which is not proteinatious. It's an organic molecule of FW~500. I wasn't sure what the peak shoulders implied but if you think that they could be due to analyte sticking to column, then the problem seems evident. I'll see if washing the columns might restore them too.

[mohan_2008]

DM909:

As ealier Danko stated:

Column backpressure increase suggests a possible sample matrix problem, sticking to the column head or inlet frit.

You have a pretty aggressive combination of the mobile phase with 0.2% TEA that can take your pH to ~14.0. A considerable amount of formic acid will be used to bring it down to pH 7.5. This seems to make your mobile phase pretty viscous - owing to higher back pressures.

Please use a calibrated pH meter to ensure a right pH range of operation, otherwise which can dissolve the column stationary phase.

Waters X-Bridge or X-Terra are not promising when it comes to peak shape (polymeric) and less silica based. Try Agilents Extend C-18 instead.
Peak tailing, shouldering is more on these Water's columns than original silica based (such as Extend C-18) - specially at neutral pH conditions where silanol effects dominate.

The other more devastating problem is that, your mobile phase is generating voids at the column head - which might result in split peaks.
If the peak splitting is aggravated after multiple injections, then this definitely is the problem.

[Uwe Neue]

I disagree with Mohan. XBridge and XTerra are not "polymeric" packings, they are hybrid packings with a very low silanol activity. XTerra gvies good peak shapes for most basic analytes, much better than classical silica-based packings. XBridge has no disadvantage compared to silica-based packings with respect to plate count and column efficiency. In the alkaline pH range, XBridge is the most stable packing compared to silica-based or hybrid packings or similar things.
I have the data to back up every one of my statements.

[mohan_2008]

I am sorry Uwe.

You are right. They are hybrid packings. Xbridge also should give good pH stability at the higher end. I did run some assays on Xbridge earlier, at neutral pH around 7.5.

The peak shape was a broad split peak (unacceptable) and I had to switch the column with an Extend which worked great under the same hplc conditions.

At neutral pH, peak shape is a primary concern due to enhanced silanol activity. Even though, Xbridge appears to be rugged towards high pH use - I don't think it can outbeat any silica based column regarding efficiency and peak shape.

[Uwe Neue]

You don't get split peaks from silanols, but you do get split peaks from using an incorrect injection solvent, or lack of a buffer or a few other things.
So if you have had strange problems with the XBridge column, I don't see that being a problem of the column. By the nature in which the packing is made, and by the surface coverage, it has a better peak shape at pH 7 then classical silica-based packings. pH 7 is our standard column testing procedure... Don't forget, I have the data...

[danko]

Hi Mohan,
If everything else was equal (i.e. equipment, eluents, method parameter, etc.) and the only difference was switching from the XBridge to Zorbax, the explanation could be the difference of column inlet fitting format. The capillary goes deeper in Waters columns compared to Zorbax. So if the capillary didn’t fit completely in the XBridge, there could’ve been a void causing peak deformation and even peak splitting.
Otherwise I wouldn’t expect this kind of symptoms regardless of silanol activity, degree of derivatization etc.

Best Regards