Let’s get physical-organic. A big topic of research in recent years has been the properties of liquids and solids under boundary conditions. By that sweeping statement, I mean questions such as “When does a small cluster of metal atoms start to act like a small piece of bulk metal? Why is there a transition, and what happens below it?” Or “What’s different about atoms/molecules in a crystal when they’re at a corner and not surrounded by so many of their kind? Are they the first ones to dissolve or react?” And another: “How do water atoms act differently when they’re in the first layer of a surface – against air, against the wall of a container, against an immiscible liquid, or in the first layer of a hydration shell? Do things dissolved in such zones change reactivities?”
There are plenty of such topics, and plenty of reasons, as you do the thought experiments, to expect interesting and unusual behavior. But for many years it was next to impossible to get experimental data on such things. Pioneers like Langmuir and Blodgett (among others) opened up surface chemistry as a field, but it’s fair to say that for a long time after them it remained pretty empirical. Good examples are industrial catalyst design and pharmaceutical formulations. We know for certain that particle sizes and compositions make huge differences in catalyst turnover numbers, fouling, and selectivity, as well as in drug dissolution rates and bioavailabilities. But a lot of what was discovered over the years was done by sheer experimentation while hunting for trends and empirical rules. Science was waiting on the tools to study physical behavior and structure at these difficult scales.
Here’s a current example from a group at Manchester (a university with a storied history in structural and nano-chemistry). It uses MCM-41, an industrial catalyst and catalyst support developed at Mobil, which is a bit like a zeolite (although those have aluminum in their structure and thus contain acidic centers, while MCM-41 is pretty much all silicate). It’s full of long empty pores, several nanometers across, a stack-of-tubes arrangement that you also see in some metal-organic frameworks and mesoporous materials in general. These hollow tubes fill with solvent, naturally, but on that scale, the solvent is not exactly just plain bulk liquid. It doesn’t have the room to be. So what is it?
People have studied this sort of thing with simulations, but when you take X-ray diffraction structures of such things (as with MOFs), you usually just get disordered solvent that can’t be refined (and whose noisy data indeed sometimes have to be tossed out, computationally, to get the rest of the structure to solve well). When the solvent molecules do refine well, they tend to be rather closely associated with the surface of relatively narrow pores and cavities, a bit more like the structure of a freestanding metal complex. But MCM-41 is too large for that and too small to be “normal” solvent. Simulations have suggested various poses for solvents like benzene – sometimes they’re flat against the pore walls in the first layer, and sometimes they come out tilted, and there are various proposals for how they stack in the layers after that.
This paper take benzene-soaked MCM-41 and gets neutron scattering data, which as I understand it is not a lot of fun (and can be obtained at relatively few locations around the world as well). The model that fits the experimental data best has benzene molecules arranged in the pore as concentric cylindrical shells, which needless to say is not what you would expect to see if you could peer inside a bottle of bulk benzene. In an 18A pore, there’s room for four of these layers. They’re definitely affected by contact with the pore wall, but in a complicated way. The closest benzenes to the wall tend to be canted at about a 40-degree angle relative to the silicate surface, but the ones a bit further out, while also definitely more constricted than a bulk phase, tend to be more perpendicular to the wall (plus or minus some wiggle room). And the ones beyond that are flat parallel to the (now distant) wall, and by this point you’re talking about benzene molecules that are mainly reacting to the presence of other structurally ordered (or semi-ordered) benzene molecules. In fact, that last interaction sounds like the “edge-to-face” orientation often seen with the aromatic rings of drug structures in protein binding pockets.
Shown is a simulation fitted to the experimental data, to give you an idea of the scale we’re talking about. The red stuff is the silicate of MCM41, and the grey-and-white are the benzenes. We’re looking down a pore, with a slice taken out of the structure. If you squint, you can actually see some of the layering – it’s messy, but it’s a far less random mess than bulk solvent. Diffusion of solute molecules is going to be a very different thing in such an environment, and reactivities can change because the energetic background of solvation has changed. Moving solvent molecules around is a big influence (both enthalpically and entropically) on the thermodynamics of a reaction, and this sort of confinement is a direct way to affect that. If we can get a better handle on it, we have possibilities to do types of chemistry we might never be able to access otherwise.