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

atomistic approach based on pseudopotentials. Both have been successfullyapplied to various nanocrystal systems [50,51,77–79]. To model the PLE data,one has to calculate the oscillator strength of possible transitions and takeinto account the electron–hole Coulomb interaction which modifies the observed(excitonic) bandgap. In the tunneling case, as discussed earlier, thedevice geometry should be carefully modeled and, in addition, the effect ofcharging on the level structure needs to be considered. The charging may affectthe intrinsic level structure and also determines the single-electron additionenergy.Zunger and co-workers treated the effects of electron charging for a QDembedded in a homogeneous dielectric medium characterized by e out . Theaddition energies and quasi-particle gap were calculated as a function of e out[80,81]. Although this isotropic model does not represent the experimentalgeometry of the tunneling measurements, the authors were able to find goodagreement between the energetic positions of the peaks for several QD sizesusing one value of e out . These authors also noted that the charging energycontribution associated with the bandgap transition may be different fromthat within the charging multiplets in the excited states. This difference is,however, on the order of the peak width in our spectra.In another approach, Niquet et al. modeled the junction parameters C iand R i and used a tight-binding model for the level structure [82]. Thetunneling spectra were calculated using a rate equation method, extendedover the more simplistic approach represented in Section II.C, by allowing forsimultaneous tunneling of electrons and holes. The authors were able toreproduce the experimental tunneling spectra, attributing part of the tunnelingpeaks at negative bias to tunneling through the CB.E. Detecting Surface StatesThe surface plays an important role in determining physical and chemicalproperties of nanocrystals. In particular, the PL is extremely sensitive to thesurface passivation and special care is required to remove potential trap sitesand to achieve high fluorescence quantum yields. Detailed investigation ofsurface states is needed for the understanding of such defects and opticallydetected magnetic resonance has been extensively applied to address theseissues [83]. STM can also be used to probe surface states as demonstrated inFig. 12. Here, the InAs nanocrystals were treated with pyridine, which partlyremoves the capping TOP ligands. The dI/dV curves measured on two suchdots show peaks in the subgap region close to the CB edge. These peaks,absent in ligand-passivated QDs, are tentatively assigned to surface states[84]. Subgap peaks were also observed on unpassivated electrodeposited CdSeCopyright 2004 by Marcel Dekker, Inc. All Rights Reserved.