Tuesday, June 19, 2012
Week 4 Day 2
Today I ran two trials in the spectrometer. After looking at the data, I realized I completely botched the calibrations for set 5, but set 6 was fine. This completely threw off the data for set 5, but set 6 was consistent with the data I took on Friday. After looking at the errors and correlations, I have determined that our machine is just bad at measuring very small concentrations. What we can do is make sure our calibration sets are dead on, because that improves error significantly. Also getting the internal standard in the sample right on seems to help. But with the lowest concentration of sodium, I'm still getting a 20% error. The other concentrations are almost all in the single digits for percent error. I'm hoping that when we actually take data, we can run the calibration first, look at the correlation in sodium and decide of its going to be good enough to run with the samples. We'll just have to make enough to run the calibration set twice if need be.