Skip to Content

How a Compound Dissolves – One Water Molecule At a Time

Here’s another one of those nanoscale articles that gives me a bit of a shiver, because it shows pictures of something that I had assumed was beyond our ability to see. The authors, from the Ruhr University in Bochum, are looking at a simple organic molecule (an azobenzene, shown at right), adsorbed onto a solid gold surface. Scanning tunneling microscopy reveals (as you’d guess) the two aryl rings to have rather different character – the salicylic acid one seems extended, since the pi-electrons are delocalized out into the carboxylic acid group, while the nitrophenyl actually seems to have a depression at its end, as the nitro group interacts with the adsorption substrate. So far, so good.

They then adsorb a small amount of water (actually deuterated water, to get better contrast) and look at where the individual water molecules end up. And they can see them associating with the test molecule in very defined ways – when there’s just one water, it’s always next to either the carboxyl group or the nitro group (never interacting first with just the phenol). That actually agrees with density function theory (DFT calculations), which confirms one’s chemical intuition that the hydroxy is already tied up in an intramolecule hydrogen bond with the carbonyl of the acid.

Courtesy Wiley/Angewandte Chemie

Shown at right are some examples of this. It’s remarkable that even at an adsorption temperature below 30K the individual water molecules are mobile enough to pick up their preferred spots. Pauling was right; the hydrogen bond does reign supreme. The next step was to increase the amount of water in the experiment, and that showed that (no matter what) the aryl rings, azo group, and salicylic OH just don’t attract waters on their own. Instead, further water molecules join the one next to the nitro group and make a hydrogen-bonded network around it.

Courtesy Wiley/Angewandte Chemie

Down at the salicylic end, things are messier. Once there’s a saturating number of waters around the carbonyl, the hydroxy can get involved, but there are a number of different configurations seen. As the solvation shell grows around the molecule, the final water distributions are a bit different in each case, and the waters around the outside increasingly resemble the bulk state. I have to say, I never expected to be able to see this happening directly, water by water. It’s true that this is a two-dimensional case, with that gold surface down there, but the test molecule is pretty two-dimensional, too, and you’d have to think that this is a good representation of solution reality. You have to wonder if this technique can also show waters interacting with two or more molecules of test compound that are already associated on the surface, to get some idea of how a solid sample is pulled apart by dissolution.

So that’s how it it happens. Individual features on a molecule pick up waters first, in a particular order, and picking up different quantities of water in the process. Then other water molecules join in around them, gradually building a solvation shell around the parts that don’t have any affinity for waters themselves. I look forward to seeing more studies in systems like this – having a mental picture of how this takes place is one thing, but real pictures are something else again!

9 comments on “How a Compound Dissolves – One Water Molecule At a Time”

  1. Ursa Major says:

    This is fascinating, and amazing that it can be seen. As you say, seeing the water structure in 3D or the real-time dissolution of a bulk solid would also be really interesting.

    “Individual features on a molecule pick up waters first, in a particular order, and picking up different quantities of water in the process.” – It’s not something I’d thought about properly, but this makes a lot of sense. As we know, all chemical reactions have a preferred site determined by the interacting orbitals so dissolution shouldn’t be any different.

  2. Peter Kenny says:

    This looks like a very interesting article. I would expect a water molecule to function as a hydrogen bond (HB) acceptor when interacting with the carboxyl group and as an HB donor when interacting with the nitro group. The aza nitrogens will be ‘hidden’ from the water and would be expected to be relatively weak acceptors anyway. The intramolecular HB is likely to make the carboxyl oxygens weaker HB acceptors while making strengthening the carboxyl HB donor. HB basicity and HB acidity can be quantified by measuring association constants for 1:1 complexes in non-polar solvents and molecular electrostatic potential can be usefully predictive. I have linked an article with both measured HB basicity and acidity as the URL for this comment although there is a lot more literature out there (especially for HB basicity).

    1. Wavefunction says:

      Pete, the specific nature of interactions you are talking about (water acting as a HBA to a carboxyl for instance) is an average picture: would that also be seen in case of individual water molecules?

      1. Peter Kenny says:

        Hi Ash, I would expect to see individual water molecules behaving in that way and the HB basicity and HB acidity measurements are for 1:1 complexes. Water (both the molecule and the solvent) tends to be thought of as a better HB donor than acceptor although it not so easy to establish this definitively. Generally measurements of HB basicity and acidity are done using model compounds with either a single HB donor or HB acceptor. However, in some cases it is possible to deconvolute contributions of none-equivalent HB acceptors/donors. Analysis of alkane/water partition coefficients (see ‘polarity of hydrogen bond donors’ section in article linked as the URL for this comment) suggests that the amide HB donor of an amide is significantly less polar than its HB acceptor.

  3. tlp says:

    How would anyone even come up with this idea? I mean it’s so amazingly simple and (I assume) technically difficult! It’s like that peeling-graphite-with-scotch-tape-and-looking-what’s-gonna-happen experiment. No surprise that no grants are mentioned in the paper.

  4. Scott says:

    I’m curious as to how someone thought to do this experiment. I mean, in terms of confirming that Density Function Theory seems to be correct (after all, theories are only ever disproven), this is pretty useful as an experiment.

    Taking pictures of individual atoms/molecules is technically doable, has been so for a while (didn’t IBM write their logo in individual atoms/molecules back in the 1990s to show it was possible?). But dang. I cannot imagine how to set up a system to do this.

    Beers are on me, research team!

  5. Matt says:

    This is fascinating, and is a step further down the road of understanding how things really behave at the molecular level. I do wonder, however, about the differences in behavior between water molecules at very low pressure conditions and 30 K adsorbing onto an Au/NPAS surface like this vs. say, at the surface of a crystal of NPAS placed in water at ambient temp. I also think this is a good opportunity to test MD simulations to see if they are accurately predicting the locations of the water. A lot of that depends on the specific charge model, I am sure. Anyway, good stuff, Derek!

  6. tim says:

    No wonder snow flakes are so unique.

  7. Arnob Endry says:

    Generally measurements of HB basicity and acidity are done using model compounds with either a single HB donor or HB acceptor. However, in some cases it is possible to deconvolute contributions of none-equivalent HB acceptors/donors.

Leave a Reply to tlp Cancel reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.