HPLC CALIBRATION

[NC - CALIBR8R]

We are in the process of updating our lengthy HPLC Calibration procedures and are wondering if any of you have recommendations as it pertains to streamlining the testing from start to finish. It was also asked:

-does isporpanol have to used to clean the clean during the PM,

-can pressure, or weight, or volume at different flow rates be employed for Gradient Testing

As far as I can tell, these are the only tests that really need to be performed:

Auto-Sampler
-Flow Rate Test {1.0 ml/min @ a known wavelengths. Other flow rates can also be checked if need be: 1.5 ml/min, etc.) Note: All solvent lines need to be checked.

-Linearity

-Precision or Reproducibility

-Carryover (is this required or necessary?)

Detector: Wavelength Accuracy

Column Oven: Temperature Accuracy

Please let me know....

[NC - CALIBR8R]

Also, has anyone ever used volume delivered over time for gradient accuracy verification? The expected total volume can be calculated from the flow rate(s) as determined by the gradient profile.

[DR]

IPA is less toxic, more amenable to being sewered in small quantities than the alternatives are.

Gradient testing is best done with water/0.3% Acetone so that the entire gradient (steps or linear) can be evaluated over time. Any other approach will likely leave you staring at unoccupied train tracks wishing you knew wich way the train went.

Flow rate is best checked with a volumetric flask or graduated cylinder and a stop watch (while a column is in place).

With the column oven - some experimentation may be required with respect to probe placement.

[Uwe Neue]

The gradient test should include the measurement of the gradient delay volume for each system. Then you know what adjsutements you may need to make going from system to system. If you set up the test properly, you get this info automatically.

[ksharp]

As far as I can tell, these are the only tests that really need to be performed:

Auto-Sampler
-Flow Rate Test {1.0 ml/min @ a known wavelengths. Other flow rates can also be checked if need be: 1.5 ml/min, etc.) Note: All solvent lines need to be checked.

I don't think you need a specified/known wavelength to measure flow rate, you don't even need your detector on (please correct me if I am wrong). Test flow rate with water as mobile phase; collect mobile phase in a weighed volumetric flask for exactly ten minutes, as it leaves the system (after the column) and determine the difference.

-Precision or Reproducibility

Injector precision is verified by %RSD over six injections.

-Carryover (is this required or necessary?)

Yes. Inject a sample with high concentration of analyte, then perform injection containing zero analyte but in the same solvent. Area/amount in the second sample relative to the first is the carryover.

Detector: Wavelength Accuracy

Agilent sells certified caffeine reference standards, local minima and maxima are measured on your instrument and compared with known values (244 nm min., 272 nm max.)

Column Oven: Temperature Accuracy

We measure with a external thermometer (NIST traceable), set instrument to a known temperature and measure it, compare that with the instrument setpoint.

[unmgvar]

NC - CALIBR8R,

you must understand that you do not calibrate the system.
what you are doing is a verification that the system is within a set of specifications that for your work environment satisfies it's use.
in case that a result falls out of the specs then you have to recalibrate it in order to meet your requirements
most important in the eyes of the auditor you are responsible.

most HPLC vendors provide a ready build set of tests for that, but it is your job to see that they meet your needs.
for example most fow test are done at 1.0 ml/min, what if you do mostly 0.3-0.5 ml/min, you need to adress the matter and check for it.
most autosampler tests of RSD are for 100 ul loops, if you have another one then no vendor will test it.

from my experience the best is this:
for the pump: flow precision where depends of what you do, we do 2 flows for some systems (in one channel is enough), pressure/leak test, gradient accuracy and reproducability for all 4 channels
for the autosampler- reproducability, linearity. if you have a pulled loop system, you must use calibrated standards to do it, weighing the vials is useless.
for the column oven- do a temp, again, depending on what you do one temp point is not necesseraly sufficient, we do 3 points, and see reproducability for each one.
detector-Drift, noise, linearity are a must, depending on the type of detector check lamp energies as well.

preferably do not use a column, you will need a restriction capillary or some sort of pressure inducer for the pump to perform correctly