Chemical Forums
Chemistry Forums for Students => Analytical Chemistry Forum => Topic started by: austanius on December 13, 2016, 02:51:30 PM
-
I use a chemilulecence analyzer that measures nitorgen but it has no units/not calibrated. So I use standards of known concentrations before and after to create a linear line to plug my measurements in. Problem is the instrument gradually changes over time(let's assume linearly) so my standards have different measurements than the beginning. I want to create one standard curve that I could use that would adjust to the instrument change over time. Let's say the fist set of standards has a slope of y=2x and at the end it's y=3x and it took 2 hours. I have a time stamp of when each sample went in as well. What would be the value for a measurement of let's say 30, 45 minutes into my run? I can almost visualize how it should work but am not quite sure how to do this.
-
For an analytical application, I wouldn't try to do it that way. First of all, we should try to understand the drift: random, or rising or falling? All three have different explanations possible. What happens if you leave the instrument powered for a long time? Drift may stop if you do. Perhaps the instrument specifies a warmup time, 15 minutes or an hour. And perhaps you discover that it needs more -- that happens.
If the drift is linear, then you can develop a curve for standards that flank your samples. If your drift is random, then the average of your standards may be the best answer you can get.
But a lot of this is outside of how most people like to report their analytical results. You should give us some information on your application.