The Chelate Effect

Teaching can be a disappointing enterprise. The other TA's and I worked fairly hard on updating a lab procedure to eliminate a lot of the systematic error that was persistent in the lab. I taught this same course last year in this same semester, so I was pretty familiar with what went wrong. The procedure is pretty simple: you make up some electrochemical copper cells and then measure the voltages at various temperatures. Plot the data and you can get delH, del S, Keq, del G and all this fantastic information. Although that last comment carries a whiff of sarcasm, this lab is pretty instructional if you don't screw it up. In spite of our best efforts to idiot-proof the lab, a substantial amount of idiotic data remains. As you can see from the plot below, the AB and BC data points are quite similar to what a grouping might look like for an epileptic firing a shotgun. (What we're looking for is more like the grouping of an epileptic firing an Uzi, preferably in one big spasm.) The R^2 value is supposed to indicate roughly the goodness of fit for the data. An R^2 of 1 is the gold standard, but 0.90 or up is good for this lab. As you can see, we have missed by an order of magnitude. Idiots.