While the efficiency of polymer solar cells has increased into the 9-10% range, it is still clear that there is much room for improvement in efficiency, lifetime, and cost-effective synthetic and fabrication approaches. A major obstacle to higher efficiencies is the development of donor-acceptor pairs that are optimal for light harvesting, charge generation and collection. Several avenues toward developing such donor-acceptor pairs will be discussed. A new family of semi-random hexyl-thiophene based donor-acceptor copolymers was synthesized where the restricted, yet randomized linkage pattern of monomers retains a high degree of structural order in the polymers preserving attractive properties of rr-P3HT while also generating broadband absorption. In many cases, efficiencies exceeding that for P3HT are observed in solar cells. We are also investigating ternary blend solar cells based on two donor components and one acceptor component, which have been recognized as a potential route to increase the absorption breadth of a solar cell and consequently the short-circuit current density. Recently, we demonstrated for the first time that the open-circuit voltage of ternary blend solar cells is composition dependent and can be tuned across the full range defined by the corresponding limiting binary blends without negatively impacting the fill factor. As a result, with judicious choice of components, the attainable product of short circuit current and open circuit voltage (and by extension the efficiency) in a single-layer ternary blend solar cell can be higher than is achievable with a standard binary blend solar cell. We have successfully demonstrated higher efficiencies in ternary blends based on two donor polymers and one fullerene acceptor than could be achieved in either of the limiting binary blends. Efforts toward developing a deeper understanding of the mechanism of operation in these ternary blends will also be discussed.