Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

distribution (dark line) does not describe the distribution of off-time spectralshifts.*Moreover, this correlation differs from blinking caused by spectralshifting observed in single molecules such as pentacene in a p-terphenylmatrix [23,24]. In single-molecule experiments, the chromophore is resonantlyexcited into a single absorbing state, and a spectral shift of the absorbing stateresults in a dark period because the excitation is no longer in resonance. In ourexperiments, we excite nonresonantly into a large density of states above theband edge [25].The initial model for CdSe QD fluorescence intermittency [12,26]adapted a theoretical model for photodarkening observed in CdSe-QD-dopedglasses [27] with the blinking phenomenon under the high excitation intensityused for single-QD spectroscopy. In the photodarkening experiments, Chepicet al. [28] described a QD with a single delocalized charge carrier (hole orelectron) as a dark QD. When a charged QD absorbs a photon and creates anexciton, it becomes a quasi-three-particle system. The energy transfer fromthe exciton to the lone charge carrier and nonradiative relaxation of thecharge carrier (f100 ps) [29] is predicted to be faster than the radiative recombinationrate of the exciton (100 ns to 1 As). Therefore, within this model,a charged QD is a dark QD. The transition from a bright to a dark QD occursthrough the trapping of an electron or hole leaving a single delocalized hole orelectron in the QD core. The switch from a dark to a bright QD then occursthrough recapture of the initially localized electron (hole) back into the QDcore or through capture of another electron (hole) from nearby traps. Whenthe electron–hole pair recombines, the QD core is no longer a site for excitonelectron(exciton–hole) energy transfer. Concomitantly, Empedocles andBawendi [13] showed evidence that spectral diffusion shifts were caused bya changing local electric field around the QD, where the magnitude of thischanging local electric field was consistent with a single electron and holetrapped near the surface of the QD.We can now combine both models to explain the correlation shown inFig. 4. Using the assumption that a charged QD is a dark QD [30], there arefour possible mechanisms, illustrated in Figs. 6a–6d, for the transition back toa bright QD. Electrostatic force microscopy (EFM) studies on single CdSeQDs recently showed positive charges present on some of the QDs, y even after* The on-time histogram also has weak tails on top of the Gaussia distribution because ofapparent on-time spectral shifts that may have occurred during an off time faster than the timeresolution (100 ms) allowed by our present setup.y Note we make a distinction between a charged QD where the charge is delocalized in the coreof the QD (a dark QD) and a charged QD where the charge refers to trapped charges localizedon the surface or in the organic shell surrounding the QD (not necessarily a dark QD).Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.