Preliminary Remarks and Historical Overview

Similar to electron photoemission into vacuum, which is universally observed at the surfaces of condensed phases, the phenomenon of internal photoemission (IPE) represents an equally common property of solid/solid or solid/liquid interfaces. It consists of the optically assisted transition of a charge carrier, an electron or a hole, from one phase into another across the interface energy barrier. This chapter reviews observations of IPE phenomena at interfaces of different types, including interfaces between metals and semiconductors with wide bandgap insulators, metal-semiconductor contacts, semiconductor heterojunctions, interfaces of molecular and organic materials, and solid-electrolyte interfaces. Specific features of the IPE reproducibly observed in these systems are compared and the underlying physical mechanisms are discussed in order to be developed into analytical models in subsequent chapters.

Keywords

Photoemission; internal photoemission; photoinjection; interface barrier; optical excitation; charge injection

1.1 General Concept of IPE

In the most simple terms, internal photoemission (IPE) can be defined as a process of optically induced transition of a mobile charge carrier, electron or hole, from one solid (the emitter) into another condensed phase (the collector) across the interface between these phases. Except for the fact that electrons do not leave the condensed phase, IPE is quite similar to the classic photoemission of electrons from a solid into vacuum (external photoemission) because the optical excitation of a carrier in the emitter and its transport to the emitting surface (or interface) are basically the same. This similarity allows the use of the multi-step model developed for external photoemission as the departure point in describing IPE, as will be done in Chapter 2.

Differences between external and IPE processes are predominantly related to the different nature of the potential barriers at the surface and at the interface of a solid, respectively, which require modifications of several descriptions. First, there are differences in carrier transport associated with the different nature of wave function in collectors and different barrier properties. Second, the photon energy h? required for IPE transition may be significantly (sometimes by one order of magnitude) lower than for photoemission into vacuum, as illustrated in Fig. 1.1. This figure shows schematically the transitions corresponding to photoemission of electrons from a metal (Au) into vacuum, a wide bandgap insulator (SiO2), and a semiconductor (Si) in panels (a-c), respectively. The energy onsets of electron emission correspond to the experimentally determined photoemission threshold (work function) of the metal Fvac (Rhoderick, 1978), and the barrier heights (Deal et al., 1966) and FSi (Tung, 2001). Finally, thanks to the presence of occupied electron states in the collector material, IPE of holes becomes possible, which has no analogue in vacuum photoemission. The corresponding electron transitions are schematically shown in Fig. 1.1(d) using the barrier parameters pertinent to the case of a PtSi/p-Si Schottky diode (Mercer and Helms, 1989).

Fig. 1.1 Schematic of optically excited transitions corresponding to photoemission of electrons from the states near the Fermi level of a metal (EF) into vacuum (a), insulator (b) and semiconductor (c). The shown threshold energies of transitions correspond to experimentally determined values for the surface of Au (the energy level of an electron resting in vacuum is indicated as EVACUUM), and Au/SiO2 and Au/n-type Si interfaces. The energies EC and EV correspond to the edges of the conduction and the valence bands, respectively. Zero on the energy scale is placed at the Fermi level of the metal. (d) Photoemission of holes from the states near the Fermi level of PtSi into the valence band of p-type silicon.

Despite the aforementioned close similarity between IPE and external photoemission, the general understanding of the IPE process and, related to it, the development of IPE-based spectroscopic methods came almost half a century after the classic photoemission picture had been established. The most significant difficulty in IPE is the need for sufficient understanding of the spectrum of electron states inside a solid to clarify the origin of the energy barriers at interfaces. These barriers are generically related to the occurrence of forbidden energy gaps (bandgaps) in a solid. Therefore, transport of charge carriers across the interface could only be adequately addressed when the quantum theory of solids had been sufficiently developed. In fact, the concept of IPE was first introduced by Mott and Gurney to illustrate formation of conduction bands in rock salt crystals by comparing energy thresholds of electron photoemission from metallic sodium into the salt and into vacuum (Mott and Gurney, 1946) (cf. Fig. 1.1). Since then, thanks to its extremely rapid development (for an overview of early work, see, e.g., Mead, 1966 and Williams, 1970), IPE spectroscopy has emerged as the most physically sound and reliable tool for characterizing energy barriers between condensed phases and for determining the transport properties of excited charge carriers in the near-interface region. The "older sister" of IPE, external photoemission, gave numerous hints to the development of modern physics ranging from the quantum theory of light to the band theory of electronic states in condensed phases. In its turn, IPE deals with intricate electron transfer interactions at interfaces of solids, which in many cases still cannot be adequately described even at the present level of quantum theory because the atomic structure of interfaces is not known precisely. Thus, by using this kind of spectroscopy one often addresses fundamentally novel elements in condensed matter physics.

1.2 IPE and Materials Analysis Issues

In addition to fundamental physics, great impetus to development of IPE spectroscopy came from the practical application of solid-state electronics, primarily semiconductor-based heterostructures. Electron transport through and near semiconductor interfaces plays a crucial role in the operation of most solid-state electronic and optoelectronic devices. Essential features of this transport are determined by the density, relative energy and quantum-mechanical coupling between electron states in the contacting materials, which ultimately determine the rate of electron transition(s), i.e., the electric current. Therefore, to understand the details of electron transport phenomena in device-relevant heterostructures, the energy spectrum of electron states at the interface requires quantitative characterization so one can control technologically the electronic properties of the interfaces. Furthermore, knowledge of interface barrier properties is often needed for numerical simulations of electronic transport, which become increasingly important during device design.

The results of studies carried out over past 50 years strongly indicate that the spectrum of electron states at an interface cannot be immediately derived from the known bulk band structure of two contacting solids. Moreover, in many cases the properties of solid materials in the vicinity of their interfaces appear to be very different from the corresponding bulk parameters. These differences indicate the significance of interface chemistry and bonding configurations on composition and structure of the near-interfacial layers of a solid (for a review, see, e.g., Mönch, 2004). With the continuing trend to reduce the size and dimensionality of functional elements in solid-state electronic devices, incorporation of new, often surface-stabilized materials in the device design, as well as the extension of the solid-state electronics to new areas of functionality, the need to understand interface properties of solid materials and related nanostructures is acute as never before.

This need, in turn, raises question about reliable sources of information concerning electron states at interfaces of solids. More specifically, physical methods capable of probing the interface-relevant portion of electron-state energy distribution appear to have the focus of attention. As the picture of the observed physical process or phenomenon must be unambiguous and transparent to enable straightforward and reliable interpretation of the results, the experimental characterization of electron states at the interfaces must go far beyond the conventional electrical characterization of the interface commonly applied in the semiconductor industry. This brings up the issue of designing experimental physical methods suitable for detecting and characterizing the interface-specific portion of electron-state density.

When developing a characterization technique of this type, one might follow two different paths to isolate interface-related contributions to the electron density of states (DOS). As the partial DOS is proportional to the number of atoms encountered in a particular bonding configuration, the bulk component(s) of DOS will be dominant (at least in the energy range outside the fundamental bandgap) unless the analysis is confined to a narrow near-interface layer of a solid. To enhance the sensitivity of the technique to electron states at the interface, the studied volume of the sample can be limited to its very surface layer by using surface-sensitive measurements. The...