Vitamin B12 analysis

10 posts / 0 new
Last post
Lysander
Lysander's picture
Vitamin B12 analysis

Good morning guys
Hope everyone is doing well

Had a question
How would you go about running an assay on Vitamin B12 with UV-Vis?
I put this topic under Chromatography, because there is a long method of running water soluble vitamins with HPLC- I was wondering if there is a short method of running an assay on B12-

Thanks in advance
Hope to get some responds

Regards
-Lysander

Dr. Analytical
Dr. Analytical's picture
I was able to find an

I was able to find an absorbance spectrum online at:

http://omlc.ogi.edu/spectra/PhotochemCAD/html/vitaminB12.html

There is a maximum at 361 nm, but also another band near 550 nm.  (Note: pH 10). 

The longer wavelength would offer good selectivity, depending on the matrix.  Depending on your sample and instrument, you should be able to adjust the pH and read immediately.   Of course, you would have to validate to ensure stability, specificity, accuracy, etc.

So, in theory, this should be possible for a simple matrix.  If you have other interferences, then you could separate the mixture or use multi-component analysis software.

Might be worth trying.  Write back and tell us how it goes.

Lysander
Lysander's picture
Dr. A

Dr. A
Thank you so much for replying sir
I have to conduct this test next week- So please stand by and I will get back to you-
Thanks you sir

Lysander
Lysander's picture
Dr. A

Dr. A
I ran this analysis by making standards and running them at 361 nm
These are the following Standards
 
Std A  7.67 mg/L
Std B  15.34 mg/L
Std C  30.68 mg/L
Std D  61.35 mg/L
Std E  122.7 mg/L
 
The sample was diluted 1/5 and ran on the UV and gave a result within the mean of the  
5 calibration point linear curve
The equation reads   Y= 0.009X + 0.0015
 
Can this be implemented as running an Assay on Vitamin B12 ?
 
Thanks Dr.A

Dr. Analytical
Dr. Analytical's picture
Can you send the absorbance

Can you send the absorbance data at each level?  That would help with the evaluation.  Also, you may have just simplified the equation for your post, but make sure you have more significant figures for the slope and intercept.

Lysander
Lysander's picture
Dr. A

Dr. A
Here are the responses of absorbance for each and every concentration

Std A  7.67 mg/L    Concentration 11.596 ppm        Absorbance 0.106
Std B 15.34  mg/L                               19.429 ppm                             0.176 
Std C 30.68  mg/L                               42.959 ppm                            0.387
Std D 61.35  mg/L                               93.209 ppm                            0.838
Std E  122.7 mg/L                               182.44 ppm                            1.639
Sample gave me                                   72.799                                   0.655                      
 Dr.A
With regards to the equation - I got what the instrument gave me-I'll make sure I'll get
more significant figures-

Dr. Analytical
Dr. Analytical's picture
I used your concentration

I used your concentration values and absorbance to look at the results.

Initial comments:
I am a little worried about the highest standard with an absorbance of 1.6.  Although some people claim they can work in this range, it is always a little risky because not much light is getting through the solution.

Regression analysis (Excel) gave me the follwoing results:

 
Coefficients
Lower 95%
Upper 95%

Intercept
0.00168
0.00123
0.00213

Conc.
0.00897
0.00897
0.00898

The Lower/Upper columns reflect the intercept and slope test limits.  The slope is different from zero (which is expected), but so is the intercept, because the range does not include zero (fails the test).  It is a positive intercept, so the curve is raised at the lower end of the scale.

r-square is very good.

The residuals data and plot are below:

Observation
Residuals
Standard Residuals

1
0.000260
1.4168

2
-0.000033
-0.1817

3
-0.000192
-1.0464

4
-0.000137
-0.7450

5
0.000102
0.5563

You can see that there is a curving trend in the residuals.  This particular trend indicates that a linear fit is not appropriate, and a second order polynomial would be better.  Also, unlike more common situations where the curve flattens out at the top, your curve is actually curving up at the bottom (the lower two points are higher than they should be, maybe due to scattering or other background correction problems).

At this point I would normally stop and say that there is a problem with the fit type.  However, when you recalculate the standards as samples, you get:

Conc.
Abs.
Calc. Conc.
% Error

11.596
0.106
11.62
-0.249%

19.429
0.176
19.43
0.019%

42.959
0.387
42.94
0.050%

93.209
0.838
93.19
0.016%

182.44
1.639
182.45
-0.006%

With the exception of the lowest standard, the relative errors are all quite small. 

So, although there are some issues with the fit, I think this curve would probably provide very good data.  A few validation experiments would provide the necessary data.

Either way, this gives you some idea of how to critically work through a data set.

More questions?

Lysander
Lysander's picture
Dear Dr. A

Dear Dr. A
Thank you so much for putting the time to calculate all the data-
I appreciate it very much- I really do

I will run some more tests and follow your suggestions-
1st- I will take out the highest standard-
2nd- I will run this test couple of more times to see if the data is reproducible

and hopefully this will resolve the problem

Thank you Dr. A
and have a wonderful holidays sir
Sincerely
-Lysander

nelunga_sonali
nelunga_sonali's picture
Hi,

Hi,

Were you trying to analyze vitamin B12 quantitatively? Was it successfull? Because i need to analyze a vitamin Syrup fpr B12. Will you be able to help me.

Waiting for your reply.
Rgds,
Sonali

Lysander
Lysander's picture
Yes-

Yes-
I'm sorry getting into this so late-

The answer is no- We can do it with LC- but not with UV Vis-