Solubility on Google Books and the ONS challenge

Andrew Lang has brought something very interesting to my attention: there are some solubility measurements on Google's Library project, providing access to the full text of books that are old enough to be part of the public domain. This means that certain measurements can be referenced directly and openly, without excluding people (and machines) who don't have subscriptions.

For example, Atherton Seidell compiled extensive collections of solubility information in 1907 and 1919. Specific pages can be referenced directly, which makes it easy to track upon inclusion into our Open Notebook Science solubility challenge.

Consider the solubility of benzoic acid in methanol at room temperature. Either using Rajarshi's drop-down menus or Andy's direct URL method, we find that there are currently 5 measurements, 4 from the ONS project and one from the Seidell compilation.


Clicking through the last link takes you directly to the table in the book where the information was obtained.
It is comforting to see that all the values are consistent. But to view all the measurements expressed as molar concentrations required a bit of fiddling. The Seidell book reports the solubility of benzoic acid in methanol at 23C as 71.5g/100g solvent. In order to convert to molar we have to make two assumptions.

First we assume that the volume of the solute and solvent are additive. This is clearly not the case but until someone proposes a better model that's what we will use. Second we must estimate the density of the solute. For a liquid that is not a problem but experimental values of the density of solid organic materials is scarce. Luckily, ChemSpider provides such estimates - and for benzoic acid that turns out to be 1.197 g/ml.

We're performing all the calculations directly in the SolubilitySum spreadsheet (columns N-W) and making these assumptions explicit. If anyone has a better model let me know.

Optimally, we would like to provide links to the public domain as much as possible when comparing our results with the literature. But that probably won't be the case. For example, many of the values for 4-nitrobenzaldehyde are from toll-access sources and the best we can do freely is provide a text citation: Maccarone, E; Perrini, G. Gazzetta Chimica Italiana vol 112 p 447 (1982).

By the way this particular search nicely showcases Andy's graphical representation of the temperature dependence of solubility in various solvents. Just scroll to the bottom.