Standard Preparation

[claudette]

Hello All,
First I would like to thank everyone for the information I've already gathered searching these forums! I've just graduated from "Technician" to "Analyst," and am quite new to the method development side of HPLC.

Presently, the method I am working on has a calibration curve made from 2 component stock standards diluted into 5 working standards. However, I was recently told that volumetric dilutions are 'frowned upon.' This is quite contrary to my experience, and was quite disconcerning. I searched around here, and googled a bit, but I can't find any information to that effect. Does any one have comments/resources? They would quite appreciated.

Also, is there a target range for response factors (ie: slope of the calibration curve)? I'm currently around 5.1e6 for component #1 and 1.3e7 for component #2. These are repeatable (I've already done linearity validation for these curves) but are they in an 'acceptable' range?

Thanks,
Claudette

[Matt Savage]

I'll answer this as someone not involved in a FDA-type regulated industry.

However, I was recently told that volumetric dilutions are 'frowned upon.' This is quite contrary to my experience, and was quite disconcerning.

We measure everything out using pipettes and volumetric flasks but... all recorded measurements are done on a 4dp analytical balance for the simple reason that I pretty sure that the balance is much more accurate than my eye!

Also, is there a target range for response factors (ie: slope of the calibration curve)? I'm currently around 5.1e6 for component #1 and 1.3e7 for component #2.

I may be corrected for this but there's nothing particlularly wrong with those figures but it does suggest that perhaps some of the scales you are using on your measurements could be more appropriate i.e. use ppm instead of ppb

Matt

[DR]

Not having had the benefit of reading your SOPs dealing with standard curves & external standard calculation preferences where you work, I'll hazard a guess...

The basic idea that is "frowned upon" is having the low end of your standard curve based on umpteen serial dilutions. Each time you do a serial dilution, you introduce error. These errors can be cuumulative in nature, hence the frowning.

If you want to use a single stock to generate a standard curve, make sure that
-you can weigh the standard accurately enough (ideally, 5 sig figs going directly into your highest standard which is also your stock for other dilutions)
-you don't have to perform more than 2 (or maybe 3) serial dilutions to get to each step in the curve (if you can't do these because of a poor balance and/or several orders of magnitude curve range, see what you can do by preparing 2 stocks - one for the high end of the curve, one for the low end).

[Bruce Hamilton]

I'd suggest that if your calibration curves are linear, there's not much wrong with the dilutions!.

However, I'd agree that multiple dilutions are likely to attract enhanced scrutiny by auditors, whether by mass or volumtric. It also depends on your solvent, if it's aqueous and not viscous, volumetric pipettes and air displacement pipettors will be fine. If there is volatile or viscous solvent present, life gets very messy, and I prefer to use syringes or positive displacement pipettors for volumetric work.

Whilst balances are possibly more accurate for many solvents, the convenience of making up a 100mL standard stock that's also your highest std, then diluting 5 mL into 10, 25, 50, 100, 250 mL volumetric flasks etc. has always been easier for me, especially when dealing with predominantly aqueous solutions. Phaermacopeia still use volumetric dilutions as well.

Everything you do has to be defended. Provided you show that the methods work as intended, and have documentory evidence, auditors should accept the procedure.

There's far too much rumour and personal preference surrounding FDA auditors, if the science and data are good, they should be happy, and concentrate on more important things.

With regard to your other question, each detector has a linear range and, providing you have stayed within it you should be fine. For example, with a UV detector you may use 0 - 200 mAU, or 0 - 1200 mAU - some people prefer to stay in the lower region, others the wider, depends on detector and personal preferences. There are threads here debating the issue.

It's working within the qualified range ( hopefully linear - but doesn't have to be ) of the detector that's important, not the actual integrator counts.

Please keep having fun,

Bruce Hamilton

[claudette]

Thank you for your replies. The standard curves are linear and the dilutions are made by adding varying amounts of the stocks to each working standard (ie: not serial dilutions). The weights are precise enough, 4-5 places, depending on amount. I'm going to stick with the current mode of preparation, as nothing in the APC characterization points to any problems.

Thanks again,
Claudette