Agilent 1100 trouble - fronting / front shoulder on peak

[jmac]

Hello all-

Although I've used this site as a reference many times, this is my first post. Hopefully, an expert/veteran has some advice for me.

So, I've been battling this issue for about 3 weeks now and cannot figure it out. We measure a drug in 80/20 MethOH/H2O. After letting the our Agilent 1100 sit for about a month, I am getting a front/shoulder on the peak. I cannot figure out what the problem is. I have rebuilt pumps (seals and new inlet and outlet valves), new frit/cap, new column, new lines. I have tried measuring other compounds with other mobile phase and fronting occurs with different samples and different mobile phases. I've created new methods and it persists. I've bypassed degasser and damper/pressure sensor and it persists. I've rebuilt rheodyne valve also. I've run the mobile phase on individual channels and premixed on each pump and it persists. I have replaced injector with a spare and it persists. I've replaced inlet filters also.

Any advice?

I'm about to reinstall Chemstation because Agilent tech told me that sometimes methods will continiously get "corrupt" and induce micro pressure flunctuations that don't show up on pressure readout.

System details: Solvent => vacuum degasser (G1322A)=>binary pump (G1312A)=> 3rd party injector/oven (8125 rheodyne valve)=> Agilent C8 guard column=> Agilent C18 column=> UV/Vis DAD (G1314A)=> 100psi backpressure regulator

[LCbob]

Hi

Have you tried a different guard column and main column?

Regs

[mattmullaney]

Hello, given all of the attention to the HPLC, perhaps the problem is not there...

What is the drug dissolved in? If it's neat MeOH or ACN , perhaps it is simply solvent mis-matching.

What is the retention time of the drug, column dimensions and flow? Just to see what k' is for the peak, if it elutes too quickly (say, k' < 2) this could be a source of trouble.

LCBob has a good point, too...if the column is older, maybe it's a little "goofed up" in terms of a void or dirty frits...that can also affect peak shape.

Hope this helps, and good luck!

[jmac]

mattmullaney and LCBob - I've replaced column and guard column problem presists. I dissolve drug in ethanol. Yes I know I should use mobile phase but it has worked in the past and therefore, should work now. I don't want to change my methodology.

Retention time - 7 minutes
Column - 250x3mm 5um Zorbax SB-C18

I'll restate, every disspossable part on the machine has been replaced. Also, it passes leak and pressure tests.

Any more ideas guys? I appreciate your help and suggestions.

[uzman]

Did you change tubings ? Especially the tubing between injector and column ?
Since you are using manual injector , another problematic part is the loop.

What kind of tubing are you using , PEEK , SS , Titanium ?

[mattmullaney]

Wow...that was weird!

Okay, trying again as something happened that caused my post to disappear!!

The k' is roughly five for the drug peak, eluent is 80:20 MeOH/water, unbuffered, drug is dissolved in EtOH.

Uzman has a good point to check the loop, particularly if you do partial-loop injections, as well as the tubing leading from the injector to the column (you may have already swapped this capillary out, though).

After re-reading...if these other methods you tried were different in that the injected compounds were in solvents of weaker elution strength than the eluent composition used to elute them, perhaps my guess of "too-strong of a sample solvent" is not good. Easy way to check would be to either halve your injection volume, and see what happens to the peak shape without otherwise changing your method. Or, if it's possible, dilute the sample solution 1:1 with water and inject twice as much as you normally do...either way, you'll have a better idea if it's the ethanol that is the trouble.

Was also thinking...if you run unbuffered and your drug is polar...you may want to consider doing some buffering. That can help peak shapes, too.

I'm hoping, so that you will have no changes...that the loop or capillary tubing is/are junked up...my suggestions sort of call for changing your method a bit.

Good Luck Once Again!

[jmac]

Lines between purge valve, injector, column, and detector have all been changed and are all PEEK. The loop is SS and has not been changed, but no problems in the past with it.

As for the mobile phase, I don't want to alter to the mobile phase the fix the problem. I have used this method before and it worked perfect for our needs, now it has the shoulder/fronting thing going on. Changing mobile phase is not a fix, there is some type of hardware issue that needs to be addressed.

I'm about to do a "blank" inject and a mobile phase injection and compare. This should give us some info on the injector.

I'll post back in a bit.

[tom jupille]

Yes I know I should use mobile phase but it has worked in the past and therefore, should work now. I don't want to change my methodology.

The "strong diluent effect" (diluent stronger than the initial mobile phase) is sneaky. I've seen it show up on a method that had been working after the instrument was serviced and some of the tubing replaced. The new tubing was shorter and straighter than the old, with the result that there was less pre-column mixing between sample and mobile phase.

[Camisotro]

I don't want to alter to the mobile phase the fix the problem. I have used this method before and it worked perfect for our needs

While I appreciate that you are very attached to following the original method, what you're trying to do here is troubleshoot. Consider that you have been willing to take apart and reassemble or replace every piece of hardware, and have even considered reinstalling software as a long shot. Is it rational to be unwilling to try varying your method as part of the troubleshooting process? If you can find a variation of your original method that resolves the fronting, it could lead you to understand the source of the problem.

Do you experience fronting with all analytes on this system, or only with this one compound?

When you replaced the column and guard - is this with new parts, or other spares you have around in the lab?

If your column and guard are designed to allow back flushing - this can often force the removal of any crud that has collected at the head which does influence peak shape. Flush each one in reverse separately with, say, 10 column volumes of 100% methanol. Then forward flush with a few more column volumes and then re-equilibrate to your original mobile phase.

[JMB]

PHYSICALLY

It does sound like a frit/void problem; if you do the back-flush (v. good recommendation by Camisotro) strongly suggest that you disconnect from HPLC so that any crud does not have the chance to deposit elsewhere, e.g tubing or flow cell.

Is this "ambient", or Temp. controlled ??

CHEMICALLY

Does your API have a structure that may allow isomerization, i.e. Keto=enol tautomerization

If so, then lowering/raising temp. may lock into a single structure.