Ascorbic and citric acids analysis with preservative

[randy]

I developed a method for the analysis of ascorbic and citric acids in water. Conditions are 98% phosphate buffer at pH 2.6/2% MeOH at 1mL/min using a Waters XBridge C18 150x4.6mm column. Detection is UV at 210nm.

It worked well except for the fact that ascorbic acid degrades quickly in water at room temperature. I found a suggestion in a Dionex column manual to add chromate to samples which would act as a preservative. I remade the cal standards and added enough potassium chromate to the cal samples to result in a potassium chromate concentration of 10ug/mL. When I ran the cal samples, the ascorbic acid response was a fraction of what it was before. The citric acid response declined about 50%. I then made and ran a fresh cal standard without chromate, and the response was where it should be.

Am I doing something wrong?

[tom jupille]

chromate?? That's a strong oxidizing agent, and should do exactly what you saw: chew up the ascorbic acid. What you need is a good reducing agent (antioxidant). Either that or make up your samples using degassed water and keep them away from air.

[HW Mueller]

There is no problem (in the context of HPLC) with stability of ascorbic in water when the solution is acidic (for instance if you dissolve pure ascorbic acid in pure water).

[randy]

Back to the original problem then:

I made a cal curve for ascorbic acid from 3 to 60 ppm. The curve looked great. Twenty injections and about 4 hours later, the midpoint calibration check results (injected from the same vial) were 88% of target for ascorbic acid. Later cal check results got progressively worse.

I made the standards in HPLC water from the neat chemical. If ascorbic acid should be stable under these conditions, that leads me to think that there is something in the sample degrading the ascorbic acid, but it must be either inorganic or inactive at 210nm because nothing is showing up in my blank sample. Or it could be the vials.

[HW Mueller]

One has to keep things quite clean even at pH below 4. The air oxidation of ascorbic is quantum mechanically forbidden, it goes when catalysts, namely metal ions, are present. Normally the ion concentration and pH of a "neat" solution should be such that oxidations are not detectable in a practical sense. Repeated sampling might have added more ions, might have changed the pH, or might have made it more light succeptible (a highly pure solution of undissociated ascorbic absorbs quite low as you saw, ascorbate absorbs at around 240nm, if I remember, but you may have introduced traces of colored material).

If one has a metal ion problem the addition of some EDTA can eliminate it.

(It is quite interesting to see that an aqu. acidic solution of ascorbic shows no radicals in an ESR apparatus. As soon as the pH goes near 7 the ascobic radicals show up. This also indicates that some metal ions were always present. Also, I was not able to get the O2 concentration low enough via purging with ~99.999 N2 to eliminate radicals at a neutral pH. The ascorbic radical rapidly disproportionates to ascorbic acid and dehydroascorbic acid, reversibly, but the latter is decomposing quite rapidly.)

[rhaefe]

Another helpful piece of information:
Stability of Ascorbic Acid in Solutions Stored in Autosampler Vials
Sam A. Margolis and Edward Park

1 NIST, Analytical Chemistry Division, Gaithersburg, MD 20899-8392

author for correspondence: fax 301-977-0685, e-mail sam.margolis{at}nist.gov

http://www.clinchem.org/cgi/content/full/47/8/1463

[randy]

Thanks for the reference. I managed to find that same one by searching this morning.

I have found that by using polypropylene autosampler vials and a fresh vial for each cal check, I can maintain a fairly stable sample.

[HW Mueller]

A cursory reading of the article prompts me to caution about it. I didn´t see mention of the pH of samples in the vials, also, the oxidation of ascorbic can hardly be due to oxidizing metal ions (amounts!), but is known to be due to oxidation by O2, catalyzed by these ions. (Of course super small trace amounts of ascorbic might be reacted away by such trace ions, but the concentrations as I saw in the article were quite high).

Serum has a pH of around 7.2, still Ascorbic oxidizes only slowly (with a long delay) in such samples, because they are loaded with antioxidants which are stronger than ascorbic. As soon as you modify this situation (as in a workup), without lowering the pH, you can loose the ascorbic quite rapidly. There are other problems with HPLC of ascorbic in blood, for instance the very low concentration. I gave up on doing this after extensive trials, including derivatization, as I didn´t get the accuracy I thought to be necessary.
Later someone asked me to help with HPLC of ascorbic in tomatoes and I couldn´t believe how easily we got it going.

[randy]

Even in a pH 2.6 phosphate buffer, I found the ascorbic acid to be fairly reactive. I had to use a fresh vial for each continuing calibration check every 3 hours, but even when I did this with cal check samples prepared more than 6 hours prior, I still had some failing cal checks.

One thing I did observe was that the degradation was much more pronounced at 14ppm than at 30ppm.

[HW Mueller]

3 h, 6h, some failures? What gives, what are you doing?

[randy]

Groundwater analysis. I had over 100 samples to analyze, and they would run overnight. I would prepare a set of cal check samples, and most of them would analyze within 10 percent of target, but a couple of cal checks near the end of the sequence were outside the 10% window.

[HW Mueller]

Your cal checks are in ground water? Still don´t understand the 3h and 6h.

[randy]

oops, sorry if I was unclear. I've prepared cal checks in HPLC water, DIUF water, and a phosphate buffer @ pH 2.6. For every 10 samples analyzed, a cal check is required. At 9 min per run and with 2 injections per sample, that's a cal check every 180 minutes.

If I prepare the cal check samples all at once, some will be analyzed closer to prep time than others. I can get one, maybe two injection pairs per cal check sample, but definitely not three. So to be on the safe side, I would prepare a separate cal sample for each cal check. For an overnight sequence, I would still have fresh cal check samples that were prepared 9-12 hours before analysis. These samples would not analyze within 10% of target at the 14ppm level, even in the acidic buffer.

[HW Mueller]

Now there is a new question: When you reuse a prepared standard it shows degradation, but not if you let it sit unused (unopened)? You can rule out, then, that something accumulating on the column degrades ascorbic?
Anyway: If you have trouble with standards how can you expect that ground water samples with all kinds of potentially catalyzing/reacting gunk will remain stable??

[randy]

An unused sample will not degrade as quickly, but after about 15 hrs, a new sample was still 87% of target at 14ppm.

To your second question: The short answer is, we couldn't expect stability. Indeed, I prepared several different matrix spikes, and most had recoveries of less than 60% for ascorbic acid. A couple even had 0% recovery! And they were analyzed within a few hours of preparation. So there was a qualification in the report for the ascorbic data.