UCLA

Organic solar cells have gained much attention as an inexpensive alternative

to traditional inorganic cells. While experimental efforts have steadily improved

the efficiency of organic devices, a large portion of the improvements have been

the result of trial-and-error. Therefore, it would be ideal to be able to use theory to predict which types of materials would produce the most efficient devices.

This dissertation presents a series of theoretical studies designed to improve understanding of what makes certain solar cells perform well and to serve as a predictive

tool to screen potential new materials.

First, a study of electron transfer in pentacene dimers is presented. The study

compares several methods for calculating the electronic transfer integral, including

time-dependent density function theory, a time dependent semi-empirical method,

and several static calculations. The results demonstrate that at large separations,

static calculations can underestimate the strength of coupling.

Next, electronic coupling in fullerenes is calculated. In this section, a method

for mimicking bulk chemical environments in film through the use of solvation and

application of electric fields is developed. The method is a applied to a number

of fullerenes used in organic solar cells, and compared with experimental data on

local and bulk electron mobilities. Comparing the theory and experiment allows

one to identify beneficial self-assembly behavior in the fullerenes studied.

This method is then extended to a calculation method we have termed direct

delocalization. In this method, a field is applied directly to the Fock matrix in

order to delocalize frontier orbitals across a dimer. Once this is accomplished,

electronic transfer time is calculated within the standard Marcus theory framework. The results are compared to the more thorough methods described above,

and found to be in agreement.

Next, the formulation of a stochastic approach to the GW approximation is

presented. In this section, a method for calculating the polarization self-energy

with stochastic orbitals is introduced. The method is highly efficient, achieving

near linear scaling with respect to system size, compared with the theoretical

fourth order scaling. The method is applied to large silicon clusters and several

fullerenes to accurately calculate quasiparticle energies.

Finally, in the last two chapters, several methods for studying plasmons are

presented. The first presents a method for studying the interaction between

molecules and plasmonic materials. The method interfaces a semiempirical quantum mechanical calculation (to study the molecules) with a finite- ifference time-

domain (FDTD) calculation (to study the plasmonic material). The study shows

that plasmon propagation can be heavily influenced by the presence of a molecule.

In the last section, an alternative FDTD method is presented. This method, labeled near-field, is a time-dependent version of the quasistatic frequency-dependent

Poisson algorithm. This approach is advantageous in that it allows for much large

time steps in the propagation, greatly expediting the calculation.