Analysis problem with UV

[mijnadarian]

Hello guys
I have a problem with my analysis- Thought maybe someone has a solution

I am trying to get a concentration of the Vitamin B12 with UV - Vis spectrophotometer in a finished product. So, I ran the UV spectra of the B12 and came up with 361 and 550 nm wavelengths.
Here is the problem

The finished product showed 4 times higher B12 content than it should.
I have a perfect linear curve based on 6 point standard calibration.
I made my standards in deionized water.

Then I thought-
I'll get the placebo and run that without any B12.
Here was the mystery problem.
The Placebo had about same concentration B12!! . It was about 4 times high as well. ( which it shouldn't be there)
I ran the placebo in 3ml/10 ml water dilution.
( I wish I could post the spectra for you- but I couldn't)
The spectra ran from 190 nm to 700 nm and it took a big curve right about 380 nm to all the way to the left of the spectrum.
The placebo has some Riboflavin in it which looks orange- The final product is maroon red

here is what I did to test myself

The UV that I use is Shimadzu 1700 series-
I ran the UV on two different modes
Mode 1 is Spectrum
I ran the analysis on this mode- Got the calibration points and Absorbances and the curve looked perfect- Got the Absorbance of the sample and (A-Intercept/Slope) gave me the same answer as what I did on Mode 3 which is Quantitation
Same thing again
I constructed a calibration curve and ran the sample and the placebo
and both gave me four times higher than expected result.
I'm thinking somwhere I have interferring wavelengths-
Anyway- I'm left with reading the manual to see whether I can do something else..

[Don_Hilton]

And perhaps the answer is in the forum you posted in - chromatography.



It would appear that there is some other ingreedient (or ingredients) with an absorbance at the wavelengths you are using. In analytical chemistry, one must always separate the analyte from interferences before attempting to quantify the analyte. Find out what other components are in the mixture and you may have a cadidate for chromatographic separation followed by UV detection.

[mijnadarian]

Thank you for the feedback Don

I'm checking the pH of my sample ( 3.85) and pH of my B12 standard in deionized water ( 5.75) and obviously there is a difference-
Could this be a problem?

[Uwe Neue]

Well - what is the spectrum of riboflavin, and how much will it interfere with B12 at the wavelength that you are using to quantitate the B12. You may need chromatography after all, or algebra...

[Don_Hilton]

If your molecule has a chromophore with a pKA or pKB in the range of the differences, it will indeed make a difference. Not only the Vitamin B12, but anything else in the mixture.

Is the final product buffered? Do the ions in the buffer adsorb in the UV?

The question that I would look at first is: What other compunds are in the mixture that would interfere with the measurement of the analyte of interest? What is the buffer? preservatives? other components?

Uwe suggets algebra as a solution - which could be a great way around the need to chemically separate species. But you must have good knowledge of what kinds of things are present in the mixture because if the math does not account for all of the components that contribute to the spectrum, you will get some bad answers down the road. I've watched chemometics do some really nice things and I've watched it used without sufficient care - resulting in methods that would not remain calibrated.