Gary Snyder Research
My research group is interested in making molecules that violate classical structure theory and exist with "broken" π-bonds. A molecule with a broken bond must have two unpaired electrons, and is thus called a biradical. Most biradicals are very high-energy compounds that have only a fleeting existence as intermediates in thermal or photochemical reactions. But some have been studied by making them photochemically at extremely low temperatures (≤ 77K), imprisoned in a cage of frozen solvent molecules. In addition to being intrinsically interesting species, biradicals are great tests of the theoretical methods that chemists use to describe processes that make or break bonds. Chemists have recently become interested in extending what they've learned about biradicals to designing and building polyradicals with many unpaired electrons that communicate in predictable ways. This could ultimately lead to an organic magnet or other material with unique combinations of magnetic and optical properties that could be useful in organic electronic devices. It may also be possible to make a compound that switches from magnetic to non-magnetic when irradiated with different colors of light or in response to an electrical voltage change.
Virtually all of the biradicals that have been studied and proposed as building blocks for magnetic materials are biradicals because they have no choice — they are compounds designed so that they cannot pair electrons to make a bond. After all, if a molecule can pair two electrons to make a bond it will always do that — that's the classical principle of maximum bonding.
But is it possible to design a compound that can easily make a bond but for some reason chooses not to? The compound below, for example, can be represented by a structure (black) that has all of its π-electrons paired into bonds; but an alternative biradical structure (red) is possible. Compared to the fully covalent structure, the biradical form is destabilized because it has one fewer π-bond, but is stabilized by the presence of several aromatic rings. So does the biradical form end up being more or less stable than the covalent form? High-level electronic structure calculations predict that this compound is actually happier — by several kcal/mol — with two electrons unpaired than with them all paired into π-bonds!
Stated more precisely, what we're comparing is two electronic states. One is the triplet, which has three magnetic sublevels (one of these, e.g., has both spins parallel, α α) and is described pretty well by structures like the red one above. The other is the singlet, in which those two electrons have opposite spin (α β) and for which the covalent and biradical forms shown are resonance structures. Thus the singlet state should be thought of as being somewhere between the two forms shown — it is a covalent compound with biradical character, called a biradicaloid. So to use the proper terminology, this molecule is predicted to be a ground state triplet biradical with a covalent singlet biradicaloid state 4 kcal/mol higher in energy.
But that's just theory. Could such a wild claim really be true? A slightly more elaborate relative of this compound is, in fact, the only experimentally known Kekulé biradical — a compound that has been shown conclusively to leave two π-electrons unpaired in its most stable (triplet) state.1
But there are even simpler compounds that might also do this. The first compound below is called isobenzofulvene. Here, the biradical form has additional resonance stabilization involving the extra cross-conjugated π-bond, and this structural feature is known to favor the triplet over the singlet in simple biradicals. The compound with one aromatic ring is an ordinary, boring, singlet polyene, and the one with two aromatic rings is predicted to be so as well. But with just one more aromatic ring (3 in all) calculations find that the triplet biradical is the ground state.2
Over the past couple of years, my undergraduate research students at Amherst College made good progress on the synthesis of compounds that we think will form the first two members of this series in fluid solution, and the synthetic route they used can easily be extended to make a precursor of the third compound. By studying the reactions of these compounds in solution we should be able to tell whether they are, in fact, very reactive singlet biradicaloids or very reactive triplet Kekulé biradicals. A student interested in organic synthesis and reaction mechanisms would be a good fit for this project.
Perhaps even more bizarre, some fairly low-level calculations done by an undergraduate collaborator at UMass have found that attaching the proper substituents to the first member if the isobenzofulvene series also results in a triplet ground state! (Note that no sane organic chemist would ever think of drawing the red structure instead of the black one for this molecule.) We now know how to apply much higher level calculations to study how substituents effect the energies of the electronic states of these molecules — that project is eagerly awaiting an interested, ambitious, theoretically inclined student. And possibly even more exciting... there is an easy way to make such a compound photochemically, under cryogenic, matrix isolation conditions — if it is, in fact, a ground state triplet, electron paramagnetic resonance (EPR) spectroscopy would be able to verify that.
Several interesting biradicaloid π-systems (acenes, bis-phenalenyls, indacenes, zethrenes, etc) have been made and studied over the past decade, and in some cases the energy gap between the singlet ground state and the higher energy triplet biradical has been measured experimentally. It's clear that the most popular computational method, density-functional theory, is incapable of providing reliable singlet-triplet gaps for these compounds; however, our recent results demonstrate that the new RASSCF-RASPT2 method probably can.2 More importantly, this method can provide an understanding of why the singlet is the ground state in some cases and the triplet is the ground state in others.2 Several research projects are possible that extend this new computational methodology to these biradicaloids.
Students interested in doing research should e-mail email@example.com and tell me (1) what chemistry courses you have completed, how you did, and what topics you found most interesting and enjoyable, (2) whether you've had previous research experience, and if so, when, with whom, and on what topic, (3) what you're thinking about doing after graduation, (4) what project you're interested in and why. (5) In addition, please tell me how many independent study credits you are interested in, and what days/times during the week you would be available to do research.
- D.R. McMasters, J. Wirz, G.J. Snyder J.Am.Chem.Soc., 1997, 119, 8568-9.
- G.J. Snyder J.Phys.Chem.A, 2012, 116, 5272-91.