Computational Study of Solvent Incorporation into a Porphyrin Monolayer

Abstract

The density functional theory (DFT) is used to investigate the conversion from a solvent incorporated pseudopolymorph into a single-component monolayer. Calculations of thermodynamic properties for the surfaces in contact both with the gas phase and with the solvent are reported. In the case of wetted surfaces, a simple bond-additivity model, first proposed by Campbell and modified here, is used to augment the DFT calculations. The model predicts a dramatic reduction in desorption energies in the solvent as compared to the gas phase. Eyring's reaction rate theory is used to predict limiting desorption rates for guest (solvent) molecules from the pockets in the pseudopolymorph and for cobalt octaethylporphyrin (COEP) molecules in all structures. The pseudopolymorph studied here is a nearly rectangular lattice (REC) composed of two CoOEP and two molecules of either 1,2,4-trichlorobenzene (TCB) or toluene (TOL) supported on 63 atoms of Au(111). At sufficiently high initial concentrations of CoOEP, only a hexagonal unit cell (HEX) with two molecules of CoOEP, supported on 50 atoms of gold, is observed. Experimentally, the TCB-REC structure is more stable than the TOL-REC structure existing in the solution at initial mM concentrations of CoOEP in TCB as opposed to the initial μM concentration of CoOEP in toluene. Calculations here show that the HEX structure is the thermodynamically stable structure at all practical concentrations of CoOEP. Once the REC structure forms kinetically at low concentrations because of the vast excess of solvent on the surface, it is difficult to convert it to the more stable HEX structure. The difference in stability is primarily due to the difference in electronic adsorption energy of the solvents (TOL or TCB) and to the very low desorption rate of CoOEP. The adsorption energy of TCB has two important contributors: the adsorption energy onto Au alone and the intermolecular interactions between TCB and the CoOEP host lattice. Neither factor can be neglected. We also find that the planar adsorption of both TOL and TCB on Au(111) is the energetically preferred orientation when space is available on the surface. Rates of desorption are very sensitive to the solvent free activation energy and to the thermodynamic parameters required to convert the solvent free activation energy to one for the solvated surface. Small changes in the computed energy (of the order of 5%) can lead to a 1 order of magnitude change in rates. Further, the solvation model used does not provide the barrier to adsorption in the solution needed to determine values for the desorption activation energy in solution. Thus, the rates computed here for desorption into the solvent are limiting values.