Hello All, I am working on a method for doing caffeine and metabolites by HPLC. Have tried several methods and haveing good separation and peak symatry with the method described below:

Varian 9010 tert pump with diode array.
Mobile phase: A= H2O c 0.25mL TFA /L
B= MeOH C= ACN
Column Supelco C18 Discovery, 25 cm X 4.6mm.
Program:
0.75 mL / min
Start Phase 90% A : 8% B : 2% C
Equilibrate 15 min
Inject
Hold 3 min
Ramp to 62% A : 30% B : 8% C by 20 min.
Hold 1 min.
Ramp back to 90:8:2 and Hold 3 minutes
Column Heated at 28 C.

Caffeine is the last compound to elute. A run this afternoon, 8 sequential injections gave the following times:
15.95; 16.44; 16.25; 15.55; 16.13; 16.00; 15.83; 15.85

Things I have done: replaced all filters, resivoir and mixer. Replaced the check valve.
Measured flows at each pump at 100% and at different mixes. Measured at 1mL / min. Flows were 1.01mL/min +- .01 mL in all cases.

Run at 1.0 mL / min.
Run at temps of 30, 33 and 35.
Run with TFA in all solvents at 0.25 mL / L.
Have tried other solvents (acetic acid and binary with H2O (acetic acid and TFA)and MeOH. Bad variability and not as good separation and peak symetry. Have tried isocratic with H2O-TFA, MeOH and ACN. Have done this with pump making mix and also by manually mixing.

Everything working well except variability of retention times.

Any ideas??? TIA
Rick Wood
thewoods@netpci.com

The first thing that occurs is that your re-equilibration time is too short. Silica based reverse-phase columns need 10-15 column volumes to equilibrate. You have 18 min at initial conditions at 0.75 ml/min. A 25x4.6 column has a void volume of approx 2.5 ml. Therefore, you're re-equilibrating with 5.4 column volumes. Run the flow up as much as reasonable pressures allow as soon as you've stopped acquiring data and the system is back to initial conditions. You might also have to increase the time.
Secondly, you may try putting the same concentration of TFA in each of your mobile phase components. It may be the TFA that is causing the problem.

measure the total volume pumped with each run. Does this correlate with retention times?
Cheers
Bryan Wallwork

What is the gradient delay volume of your system?

I am not familiar with the Varian 9010 pump, but I have seen similar symptoms with HPLC systems that mix mobile phases on the low pressure side of the pump using a gradient proportioning valve.

What can happen is either a buildup or particulate can prevent one of the channels of the gradient valve from fully seating. Thus during the portion of the pump stroke when the pump is attempting to draw from channel A it is actually drawing from channel A plus some from the leaking channel.

We have 20 low pressure mixing systems (HP 1100s and Waters Alliance) and I have seen this maybe 4 times in the last three years. I am not sure what causes the problem but it occurs in the channel that is being used for acetonitrile and can be cleared by pumping hot water through the channel or replacing the multi channel gradient valve.

To test for this problem or to retest following correction you can perform a gradient proportioning test with 0.5% Acetone as a tracer and monitor at 265nm. Good luck.

Tom,

Thanks, you just solved a problem that I have been working on for three days - with Agilent!

I have just taken over operational management of a pharmaceutical QC lab with 12 HP1100s any advice or warnings regarding the 1100s would be greatly appreciated.

Specifically, what kind of maintenance contract do you suggest. Do you recommend Agilent's calibration service. Our operating/calibration/maintenance SOPs are in a sad state, written for different equipment I would appreciate any advice any of you have.

Just a Status Report. Thanks for all of the comments, has really been helpful. I increased the equilibration time and saw some improvement. I then added a 50:50 MeOH:ACN step at the end of the analytic run for a few min then went back to starting gradient all at increased flow and equilibrate for 25 min at analytic flow before injection.

Much improved retention times. Have only run a few so far. Acceptable now, if it keeps up. Are there any secretes to reducing equilibration time? Is my step at 50:50 MeOH:ACN really helping? seems to be.

I am very plesently amazed at how helpful this board is - more helpful than my calls to Varian.

Regards,
Rick

Try adding 3% n-propanol to each of your mobile phase components. This should cut your re-equilibration time by 1/3 or more. Sometimes the addition changes selectivity slightly. I've only heard it referred to as "enhancing" selectivity. Also increasing the flow during re-equilibration enables you to pump more column volumes in a shorter amount of time, thus decreasing re-equilibration time.

TFA possesses some ion-pair activity and needs a well equilibrated system. With a gradient elution, it is difficult to coppy the same equilibria from run to run. With A and B and C mixing, you will need a mixing system with a supper proportionong valve (call me if you find one). Sample solvente different from the eluente and a high injection volume may also "help" disturbing the equilibrium.
Try mixing A and B only, where B is a premix (8:2)of your B and C. Add an equal concentration of TFA (may be less than 0.25 ml/L) in both A and B. Addition of 3% n-PrOH is a good idea as well. If posible inject NMT 50µl of sample dissolved in the initial mobile phase. Increasing the column temperature reduces equilibrium time.
Is your step at 50:50 MeOH:ACN really helping? I think no.