Potential Sources of Error in Lactate Absorption Measurements

[nobahar]

Hello!
I've got a question that asks what is the most likely reason that my experimental absortion plot for lactate differs from the theoretical. I figured it's simply incorrect volumes transferred via the pippette. Could it be the other chemicals in the solution? Although they're 'not supposed to' absorb at that wavelength, is it still a possibility?
Any ideas welcome. Thanks in advance.
Nobahar.

 

[chemisttree]

What absorbance values did you record? Remember that Beer's Law is only linear for dilute solutions. More concentrated solutions are non-linear and will give you somewhat lower values than you might expect using molar extinction coefficients determined for dilute solutions.

 

[nobahar]

I didn't know that. In fact, that hasn't been mentioned wherever I've turned, thankyou.
My results ended up above the theorectical curve (which is supposed to be the maximum!); but aren't too bad'. For example, at 1.48 micromoles/ml, it should be... 9.2056?! (6.22 x conc.).
If this is correct, I'm going to be p****d off! I swear I tried this and it didn't come out this high, this is approximately what I got from the results...

 

[chemisttree]

1.48 micromoles/mL is 1.48 mmol solution. That's fairly dilute and you shouldn't see nonlinearity.

Were you measuring in serum? If so, hemolysis is a known interference which will give you spuriously high values.

 

[epenguin]

As you do not state just what you were measuring and how (with preferably some way of reader knowing background theory or method e.g. quotation from your manual or something) nor even what the nature of the discrepancy is, you are rather lucky to have found a mentor who intuited some of this.

 

[nobahar]

Point taken, my apologies; that's also testament to the brilliance of chemisttree! :) Thanks chemisttree! You've responded to many of my posts, hopefully the majority of the others were perspicuous. Unlike this one!
Okay, one mistake was that hydrazine and NAD⁺ react, but, according to the handout, this is accounted for by subtracting absorbance values at each reading of the absorbance (which was done) from a 'blank' sample (without lactate). I wasn't aware of the dilution playing a role in the linearity of the graph; but the 1.48 concentration was also the highest concentration of all the samples, and as chemisttree has pointed out, this should be fine.
If I can recall correctly, there was perchloric acid (concentration of 6M and 1000 microlitres is used in each sample), distilled water (500, 375, 250, 125, 0 microlitres for samples 1, 2, 3, 4, 5 respectively) and lactate (0, 125, 250, 375, 500 microlitres in samples 1, 2, 3, 4, 5 respectively). Therefore there are five samples, one without lactate (the 'blank'), of 1.5ml.
Then 200 microlitres of each of these mixtures is added, separately, to 2500 microlitres of glycine-hydrazine buffer pH 9.0, then 200 microlitres of NAD⁺ is added, then 20 microlitres of lactate dehydrogenase (LDH) - to all samples. The absorbance is read prior to the addition of LDH (referred to as 0 mins), then 30 minutes later. Subtracting the absorbances of the samples from the one without lactate, the blank (0 mins sample from 0 mins blank, and 30 mins sample from 30 mins blank).
I hope this is clear enough. I think that's everything!
P.S. I identified the nature of my discrepancy, my absorbances were above the theoretical curve. All my results for the samples were above.

 

[nobahar]

Additionally, hemolysis is an interesting point, it was subsequently used to test the lactate concentration of blood samples, although it was deproteinised (N.B. I'm jumping to the conclusion here that this would mean hemolysis won't influence the absorbances, is this the case?). I wasn't aware of that though, thankyou!

 

[chemisttree]

What were the absorbance values (raw data) that you measured?

 

[nobahar]

Sorry, I thought some general possibilities would be thrown at me. Thanks for taking the time to help with this. The practical is handed in but I'm interested to know for my own benefit, also because it's interesting!

μmol Lactate...at 0 min...at 30 min...Change in absorbance (30 min - 0 min)
S1...0.074...-0.127.....0.241....0.368
S2...0.148...-0.009.....0.567....0.576
S3...0.222...-0.01...0.71.....0.72
S4...0.296...-0.007.....0.91...0.917

The first absorbances come out minus since the 'blank' sample at 0 min is taken off, the 30 min absorbances is also 'corrected' by taking off the 30 min 'blank' sample. Absorbances for the blanks were 0.245 at 0 min and 0.290 at 30 min.
The final column is the 'final' absorbance reading if you will, what went on the graph. The micromoles of lactate is what is present in each sample, since I used 4.44 micromole/ml concentration at the outset. Using sample 4 as an example, I added 500 microlitres, that's 2.22 micromoles in the sample, now with a concentration of 1.48 micromoles/ml (as perchloric acid, 1000 micromoles, is also present). Then 0.2 microlitres is extracted, and so there is 0.296 micromoles. However, I was wrong in saying that there was 1.48 micromoles/ml concentration, as the 0.2microlitres of the sample is added to the glycine-hydrazine buffer, NAD+, LDH, etc. But this would simply dilute the concentration further. This was taken into account for the theoretical curve; and so the graphs should match.

 

[chemisttree]

Absorbances at or above 0.6 are in the nonlinear range regardless of the concentration. Subtracting two samples in the nonlinear range can give you weird results.

Think of it this way...

If you are counting trees in a slice of forest, can you accurately count them if there are many trees? Don't some of the trees in the foreground block you from seeing the trees in the background?

Imagine the same forest with far fewer trees. It is now easy to accurately count all the trees since the trees in the foreground don't block the ones in the background.

Now imagine you are counting only the Aspen trees in a forest of Pinion Pines. If there are a lot of Pines in front of the Aspens, you can't accurately count the Aspen, even though the concentration of Aspen is fairly low (in the 'linear' range). What is important is the overall concentration of trees (medium + analyte) rather than the overall concentration of the Aspen alone (analyte).

Getting back to chemistry...
You provided some reduced (not raw) data that showed that all of the 30 minute samples in S2, S3 and S4 were in the nonlinear range. The raw data would have been something like
S2₃₀ = 0.857
S3₃₀ = 1.00
S4₃₀ = 1.20

before you subtracted the blank. They are all in the nonlinear range.

 

[nobahar]

Thanks Chemisttree. I like the anology. I think I understand the idea that as the concentration of the analyte (thanks for the new vocab!) increases then this will affect the accuracy of absorbances reading, but I thought that the other 'types of tree' wouldn't be important, in a sensethey're invisible...