Excitons are elementary excitations consisting of bound electron-hole pairs. The creation and recombination of excitons are the principal mechanisms by which light interacts with semiconductors. In the low density limit, excitons may be treated as weakly interacting bosons, the result of which is that Bose-Einstein condensation (BEC) of excitons is predicted for some regimes. Finite exciton lifetimes have provided an experimental challenge; it has proven difficult to cool excitons sufficiently to observe BEC since the timescale for recombination is much shorter than the time it takes to reach thermal equilibrium.

The most promising experimental setup for solving this problem and realising BEC of excitons is the coupled-quantum-well (CQW, sometimes called bilayer) system. This consists of thin alternating layers of different semiconductors (such as AlGaAs and GaAs) with an applied electric field in the growth direction (perpendicular to the planes). The excitons which form in the CQW system are then composed of electrons and holes which are confined to different layers, hindering recombination and extending exciton lifetimes long enough for BEC to be observed.

Results obtained from CQW experiments remain somewhat ambiguous; there exist quantities which may not be measured directly, but which are needed from theory/modelling in order to say something meaningful about the properties of the system (eg, whether BEC occurred or not).

Tan et al. used QMC to investigate the binding energies of biexcitons in bilayers. More recently, Schindler and Zimmermann, and others have examined the exciton-exciton interaction and the region of biexciton stability. Current work using CASINO aims to more accurately calculate both of these quantities. The bilayer system containing two particles in each layer is an ideal problem for DMC, since the wavefunction is positive everywhere and there is no fixed-node approximation, making our calculations in principle exact.

The region of biexction stability as calculated by 3 different groups; the present DMC data (red circles) agree well with Meyertholen and Fogler's (blue squares) variationally exact results. Both lie some distance from the results of Schindler and Zimmermann (black triangles), who used a model exciton-exciton potential to investigate bound states.

The tail of the biexciton binding energy, where there exists a critical point at which the layer separation becomes large enough to preclude the formation of biexcitons. The deviation of our data from the old DMC results of Tan et al. is due to a systematic error in the old data which was removed for the present work.