6564
J. Phys. Chem. C 2008, 112, 6564-6570
Defect-Dominated Charge Transport in Si-Supported CdSe Nanoparticle Films Shaibal K. Sarkar,† Gary Hodes,† Leeor Kronik,† and Hagai Cohen*,‡ Department of Materials and Interfaces and Chemical Research Support, Weizmann Institute of Science, RehoVoth 76100, Israel ReceiVed: December 11, 2007; In Final Form: February 10, 2008
Investigations of charge transport mechanisms in thin CdSe nanoparticles films deposited on silicon, using surface photovoltage and chemically resolved electrical measurements, reveal a strongly nonlinear optical response and a negative differential resistance. Here, both phenomena are rationalized within a phenomenological model consisting of two spatially separated types of traps, one related to the CdSe nanoparticles and the other to the nanoparticle/substrate interface. The model successfully explains a broad range of both old and new experimental observations and is used as a numerical framework for showing how the interplay between hole and electron processes dominates the photoelectrical properties of nanoparticle films.
I. Introduction Studies of charge transport mechanisms through aggregated nanocrystalline layers are of interest from both a basic and an applied science point of view.1-3 From the applied science perspective, nanocrystalline films are important for novel solar cells.4,5 From the basic science perspective, the huge surface to volume ratio means that the standard drift-diffusion transport typical for crystalline semiconductors is not directly applicable to nanocrystalline semiconductors. The latter are often depleted, and charge transport is then usually dominated by charge hopping and trapping.1,6-12 As a consequence of the greatly increased role of surface trapping, manipulation of surface states (e.g., by mild etching and/or ambient modification) has been shown to control major qualitative features of the currentvoltage behavior of nanocrystalline films. Notable examples are a reversal of the direction of the current13,14 and the appearance of negative differential resistance (NDR).15 Understanding carrier transport across an interface between a nanocrystalline film and one or more crystalline layers, as well as the system response to optical excitation, is of generic interest for optoelectronic devices. Generally, one expects interface properties to affect transport, possibly significantly. However, elucidation of interface contributions to carrier and photocarrier transport is greatly complicated by the presence of competing mechanisms at different spatial locations and by the lack of experimental tools able to selectively probe the different contributions. A technique proposed recently, known as chemically resolved electrical measurements (CREM),16-18 is well-suited for overcoming these difficulties. CREM is based on the detection of voltage-induced shifts of core-level features observed in X-ray photoelectron spectroscopy (XPS). Therefore, examining different chemical elements allows for probing the potential at different inner regions within the sample. CREM has already been shown to be useful in probing the electrical potential at inner regions within heterostructures.6,17 Combined with optical excitation, CREM studies have been demonstrated to be a * Corresponding author. E-mail:
[email protected]. † Department of Materials and Interfaces. ‡ Chemical Research Support.
powerful tool for broad in situ characterization of photoactive multicomponent systems.16,17 In this article, we explore CdSe nanocrystalline films deposited on Si wafers as a prototypical two-component photoactive system. This system has already been shown to exhibit unusual electrical and photovoltaic characteristics, including both negative differential resistance (NDR) and photovoltage reversal.6 Here, we present CREM and conventional (Kelvin probe) surface photovoltage measurements of this system in light of a simple phenomenological model that rationalizes all experimental observations, the new as well as earlier ones, by means of competition between two spatially distinct trap populations that dominate transport: one at nanocrystalline surfaces and the other at the film/substrate interface. II. Experimental Section 1. Sample Preparation. CdSe films were deposited by chemical solution deposition from an aqueous bath containing CdSO4 (80 mM), potassium nitrilotriacetate (120 mM), and Na2SeSO3 (80 mM Se; stock solution made by dissolving 0.2 M Se in 0.4 M Na2SO3 at ca. 60 °C for 1 h). Films with an interface layer of Cd(OH)2 were deposited by adding Na2SeSO3 to the deposition solution after the substrate was already immersed in the solution. For films without this interface layer, the substrate was immersed typically 7-10 min after selenosulphate addition, when a light yellow-orange coloration had already appeared in the solution. The pH was adjusted to ca. 10.5. The substrates used were Si (arsenic-doped n-type, resistivity in the range 0.001-0.005 Ω‚cm). Formation of a thin (1-2 nm) silica layer was detected below the CdSe layers (observable for less than 15 nm thickness of the latter). Deposition was carried out in a thermostated bath at 10 °C in the dark. The crystal size of the CdSe films was ca. 4 nm under these conditions, with a moderately narrow size distribution (standard deviation ca. 0.8 nm).19 This corresponds to a bandgap of ca. 2.2 ( 0.2 nm. The film thickness (deposition time) was varied depending on the specific experiment. Approximate values ranged from
