Galactic chemical evolution models require stellar nucleosynthesis yields as input data. Stellar evolution models are used to calculate such yields but do not take into account the fact that many stars are in binaries. The computing time required to explore the binary star parameter space is usually considered to be prohibitively large. Therefore binaries, except for type Ia supernovae and novae which are included in an ad hoc way, are ignored in most Galactic chemical evolution models. In this dissertation synthetic nucleosynthesis models are developed which approximate full stellar evolution models. Cunning methods are employed to model shell burning in low- and intermediate-mass stars while high-mass stars have their surface abundances fitted to their mass. Explosive yields are fitted to published results. The synthetic nucleosynthesis model, with the addition of algorithms to deal with mass transfer in binaries, is coupled to a rapid binary star evolution code. The use of a synthetic model speeds up the calculation of stellar yields by a factor of about 107 and extends the analysis to binary stars.
Single- and binary-star yields are calculated for a range of initial mass and separation distributions. A change in the primary or single-star mass distribution is most significant. Changing the secondary mass or separation distribution has a smaller effect. Consideration is then given to variation of the input physics to determine which free parameters are important for the calculation of yields from single and binary stars. It is found that certain parameters are important for some isotopes. Future prospects are then briefly discussed.