April 22, 2011
As part of my university work, I’ve recently been screening a large number of compounds as part of a relatively simple biochemical assay. The assay takes place in a buffered pH = 7.5 solution and about 30 of the compounds I screened are relatively insoluble in water. So I ran the assay with these compounds after dissolving as much of them as I could in water to make a saturated solution. In order to quantify my results, I need to have a good estimate of these compounds’ solubility in water.
The normal method to determine a compound’s solubility is simple. A saturated solution is prepared, the insoluble remains are filtered off, the solvent is evaporated, and the resulting solute is weighed. The problem with this method for my purpose is that since I need to determine the solubility at pH = 7.5, I need to have large amounts of salts in solution that make up the buffer. Upon evaporation of the water, I would end up with my compound of interest along with a bunch of salts, and since these salts might comprise the majority of the mass of my solids, my error would increase significantly. The second main problem with this gravimetric determination of solubility is that you need large amounts to lessen physical errors associated with recovering all of the solvent after filtering a solution. Since, in some cases, I only had a few milligrams of these compounds, I had to utilize a different method.
I decided to take advantage of the intergrations of 1H NMR spectra in order to determine the solubility of my compounds. The protocol I developed for doing so turned out to work amazingly well- there was little error amongst trials and the procedure was fast and simple. I’m sure there are papers written about this and perhaps it’s even in books, but I haven’t come across the concept too frequently either because it’s too simple or too boring to measure solubilities these days.
Now this method has a few requirements and limitations:
1. You must know the identity of your compound a priori. This is because you need to know how many hydrogen atoms your compound has and have a general idea of where they will appear in your NMR spectrum.
2. The compound must have hydrogen atoms and at least one of these must be detectable by NMR in your solvent system. This requirement is obvious enough for 1H NMR.
3. You must assume your compound has the same solubility in a deuterated solvent system as a solvent system with normal isotopic composition. This may not be as good of an approximation as I had originally thought. A quick search has taught me that ionic salts typically are less soluble in heavy water than in normal water. Good old sodium chloride shows a 6% difference. Relatively insoluble ionic compounds have even greater differences. For example, lead chloride is 34% less soluble in heavy water!
4. I used this method to measure compounds with solubilities between 10-200 mM. I didn’t really push the limits of NMR detection nor did I get close to saturating the signal so the acceptable concentration range is probably large than this.
I measured my solubilities by 1H NMR in a 100 mM pD = 7.5 K2DPO4/KD2PO4 buffer in D2O. The D3PO4 needed to prepare the buffer was made by carefully reacting P4O10 with D2O.
P4O10 + 6 D2O --> 4 D3PO4
I then added my deutuerated phosphoric acid to a solution of K3PO4. The K3PO4 was thoroughly dried in an oven before use to eliminate H2O. While monitoring the solution with a regular pH meter, I then adjusted the apparent pH of the solution to 7.9 since to first approximation, the apparent pH of a heavy water solution subtracted by 0.4 is the pD. All of this is fundamental chemistry, but it's to have it put to good use.
A saturated solution of each compound was prepared by sonicating excess of the solute in buffer (700 uL) for 1 hour in a centrifuge tube. The solution was then centrifuged, and the supernatant (500 uL) was transferred to a NMR tube via syringe. Solids must be removed from the NMR sample in order to get good shimming. Acetone (20 uL) was then added as an internal standard and shook the NMR tube well to mix everything. Then I took the NMR spectrum and integrated the acetone peak at ~2.2 ppm and compared it to some aromatic peaks or other peaks which I could easily identify.
Note that you could use almost anything as an internal standard. I chose acetone because it only has one peak and didn’t interfere with most of the peaks for my compounds. If you use acetone to clean you NMR tubes, then I shouldn’t have to remind you to bake your tubes in the oven for quite some time.
Using this simple method, I was able to determine the solubilities of 20 compounds in about 4 hours. This is sure much faster than any manual gravimetric technique. With variable temperature NMR, this technique could even be adopted to measure solubilities at different temperatures.