pKa of Bromcresol Green lab – feedback

I made two marks on each lab. Report indicates whether the written document was prepared properly. In-lab/Models indicates whether you followed directions reasonably well during the two lab periods.

A number of “unsatisfactory” Report grades were issued. Most of these were related to the graphs. Perhaps a graph was missing. Or perhaps the wrong data was plotted on one of the axes. Or sometimes the data table was missing. (There were many other smaller omissions, like missing lab partner name or missing/substandard reference for published value of pKa, but I didn’t mark off for these.) If your Report grade was “unsatisfactory”, submit a new and improved report to me by Friday, Oct 16, and I will reconsider your grade.
Some other widespread errors that merit comment:

Provide meaningful labels. A huge number of table columns and chart axes were labeled with various sorts of gibberish, sometimes broken into weird combinations, e.g.,


This doesn’t cut it. A report is a serious document and labels, because they communicate information about what is contained in a table or figure, need to be given the respect they deserve. To say, “well, Excel doesn’t let me create labels with chemical formulas” is not really an excuse. No one said that Excel’s labels were required. *Good* labels are required. If Excel can’t provide these, find another way.

Relative %error estimates on pKa aren’t valid. Many of you tried to estimate the error in your pKa values by reporting relative errors. These are not as useful as you might think. For example, if I make an error of 0.5 on a pKa of 10, the relative error is only 5%, but isn’t the error really larger than that? To see this, convert the pKa values to Ka values. The small gap between pKa of 10.5 (my value) and 10 (actual value) becomes the much larger gap of 3.2 x 10-11 (my value) versus 10-10 (actual value). The error is 300%, not 5%.

Another way to think about this: if I had made the same 0.5 error on a pKa of 2, I might have called this an error of 25%, but when I convert my pKa and the actual pKa to the corresponding Ka values, I find an error of 300% again.

Resonance structures, oy and joy. Many good resonance structures were drawn, but many opportunities for drawing good resonance structures were overlooked.

First, a drawing with dashed bonded and partial charges is not a resonance structure. It might be used to show a resonance hybrid, but it can never never be a resonance structure. Draw resonance structures according to the same rules we use for Lewis structures.

Second, the decision to view a molecule as a resonance hybrid is rarely iron-clad. Some judgment is called for. For example, I would say that the last three RNH2 molecules in the table (R = HC=O, NC, NH2-C=NH2+) are all resonance hybrids because I see 1) shorter than normal CN bond distances, 2) larger than normal HNH bond angles, 3) a loss of electron density around N in the potential map, and 4) N that is planar (in two cases) or nearly so (the third case).

Third, don’t forget that the point of drawing resonance structures is to arrive at a more realistic picture of a molecule. Many of you noted that the various benzyl cations delocalized positive charge away from the benzylic carbon, and contained shortened C(benzyl)-C(ring) bonds, but you didn’t draw any resonance structures that delocalized charged or increased the CC bond order.

We will be working on resonance structures throughout the year, but you might want to review your modeling worksheet and see if you need to revise any of your ideas about this strange and wacky topic.

This entry was posted in Labs. Bookmark the permalink.