Recently, I published a paper in the Journal of Physical Chemistry, A with lead author Kyle Grice at DePaul University. He’s an inorganic chemist studying catalytic transformations using transition-metal complexes . One active area in catalysis is the development of systems that are photoactive. Using light to activate a chemical reaction (think photosynthesis) is interesting because the process is considered environmentally friendly. There are other research areas that seek to develop and better understand photochemically active systems, such as organic light-emitting diodes and solar cells. Yes, you read that correctly, better blinky-lights through chemistry.

Traditional metals used in photochemically active systems include ruthenium and iridium, both of which are reasonably expensive. Aluminum can also be used to create photoactive compounds; however, the current technology used for mining and processing the metal is energy intensive and not all that environmentally friendly. Zinc, when coordinated to the ligand 8-hydroxyquinoline, exhibits photochemical properties and may be a promising alternative to the more expensive metals that also has a smaller environmental impact.

8-hydroxyquinoline and the complexes it can make with zinc and aluminum.

The current problem is that the structure of bis(8-hydroxyquinolinato)zinc is a bit ambiguous. There have been a few studies performed to determine its structure and the results suggest it may be octahedral (with solvent ligands binding either in cis or trans configurations) and it may form small tetrameric clusters. My role in the project was to assist in determining the solution structure of bis(8-hydroxyquinolinato)zinc.

The structure of 8-hydroxyquinolinato complexes of zinc is influenced by solvent and other environmental factors.

The technique I used is called Diffusion Ordered SpectroscopY (DOSY). It is an NMR technique that measures the rate of diffusion of molecules in solution. Because larger molecules will have a harder time moving through a sea of solvent molecules, they exhibit lower rates of diffusion. In some cases, it is even possible to predict this rate using the Stokes-Einstein equation:

$$D=\frac{k_B T}{6\pi\eta r}$$

In the Stokes-Einstein equation $k_B$ is Boltzmann’s constant, T is the temperature, $\eta$ is viscosity and $r$ is the radius of the molecule. Together, these terms can be used to predict $D$, the diffusion coefficient. Because both radius and viscosity are in the denominator, increasing these values results in a decrease in the diffusion coefficient. This model works best on molecules that are (or are nearly) spherical, which is the case for the zinc compound we are exploring.

We performed these experiments in dimethyl sulfoxide, DMSO, (actually, deuterated DMSO) because the zinc complex dissolves readily into this solvent. DOSY doesn’t create pretty pictures (it can, and we have some in the supplemental information, but it isn’t terribly exciting) so the results are a bit anticlimactic. We obtained a diffusion coefficient for bis(8-hydroxyquinolinato)zinc relative to ferrocene of 0.5. We use ferrocene as an internal standard to minimize the introduction of uncertainty due to systematic errors in the experimental procedure. Note that temperature also affects the diffusion coefficient, and small temperature fluctuations can give rise to unexpected results in the DOSY measurement. Since ferrocene is included in the sample along with the zinc complex, both compounds will be influenced (more-or-less) equally, and taking the ratio of the two values will remove these effects from the final result.

Additionally, we performed the same experiment on a similar compound, tris(8-hydroxyquinolinato)aluminum, which has been well studied and has a known chemical structure. It also has a diffusion coefficient in DMSO of 0.5 relative to ferrocene. We draw two conclusions from the similar (relative) diffusion coefficients. First, the results support the idea that the zinc complex is monomeric in DMSO; otherwise the relative diffusion coefficient would be much smaller. Second, the similarity in relative diffusion coefficients suggests that the sizes of bis(8-hydroxyquinolinato)zinc and tris(8-hydroxyquinolinato)aluminum are similar. Since the zinc complex only has two ligands – whereas the aluminum complex has three – the zinc complex is likely also bound to two molecules of solvent.

Both of the claims made above are supported by another experiment described in the paper (called EXAFS) and theoretical calculations made on proposed chemical structures. We believe this evidence is sufficiently strong to support the conclusion that bis(8-hydroxyquinolinato)zinc exists as a monomeric octahedral complex in solution, with the two solvent molecules likely occupying a trans configuration. What didn’t make it in to the paper is that we also tried to perform the same experiments in a different solvent, dichloromethane, in which the zinc complex is proposed to exist in an oligomeric form. We ran in to problems with solubility (and possibly sample decomposition) which made DOSY measurements very difficult. If I continue on this project, it would be to identify ways to make diffusion measurements in dichloromethane or another solvent that isn’t as likely to coordinate to the metal and determine if the polymeric species are present.