Solv. Calibration Curve Problem for Unknowns: A-E

[kynephrus]

I don't know if this is the right place to post such a question, but I saw a couple of other similar questions, so I figured I would give it a shot.


My problem has to do with a calibration curve. My lab partner and I just did an analysis using UV-Vis on caffeine. We constructued a calibration curve using linear regression, the equation of the line being:

y = 0.0002x + 1.5033

We are trying to extrapolate the concentration of an unknown using this equation and the absorbencies obtained for the unknown. I know theoretically I should just plug the absorbancy into y and solve for x, but the absorbance is so small, whenever I try I get a really negative number, in the 6000. Am I doing something wrong? Is there a way to use this number even though it is negative?

The absorbancy of the various unknowns are:
a=0.118988
b=0.225891
c=0.305755
d=0.234054
e=0.226928

if someone can help me with this is would be greatly appreciated. I've pretty much run out of sources and ideas.

 

[Gokul43201]

What are all the units (y and x)? What were the extremum points of the calibration data? How good was the linear fit (do you have a value for RMS error or chi-square)?

Right off the bat I notice that the slope of your regression line is good to only 1 sig fig. That doesn't look encouraging. Consider fixing that.

 

[chemisttree]

It is worse than that, I'm afraid. The slope of this line is close to zero and the offset is 1.5 absorbance units! 1.5 absorbance units is waaay into the nonlinear range. You will never be able to do anything with this calibration line. Your calibration should be forced to go through zero since Beer's law states that absorbance = (path length)X(concentration)X(absorptivity). the path length is constant as is the molar absorptivity so combining them gives you an equation something like absorption = (constant)X(concentration) + 0