pi-conjugated compounds have been intensively studied for their use in plastic electronics', an alternative to inorganic semiconductor devices. In particular, research in the use of pi-conjugated compounds as the active layer in electroluminescent devices has advanced rapidly: potential applications include e.g. flat, flexible display panels. In organic electroluminescent devices the annihilation of positive and negative charge carriers to the ground state proceeds via two fundamental steps: (i) formation of a neutral exciton, either spin singlet or spin triplet, and (ii) the radiative or non-radiative decay of the exciton. In most pi-conjugated materials only the spin singlet exciton is emissive, and this is why the spin statistics of step (i) determines the maximum achievable electroluminescent efficiency. A number of recent studies indicate that exciton formation is spin dependent and that more singlets form in organic electroluminescent devices than what is expected from spin degeneracy. We review recent experimental and (to a lesser degree) theoretical results. In particular we focus on our own recent work, where we have measured the ratio, r = sigma(S)/sigma(T), of the spin-dependent formation cross section, sigma, of singlet (S) and triplet (T) excitons in pi-conjugated oligomer and polymer films, using a spectroscopic/magnetic resonance technique. The experimental results reveal a distinct difference between exciton formation in molecular devices compared to polymer devices.
