Surface states, surface potentials, and segregation at surfaces of tin-doped ${\mathrm{In}}_2{\mathrm{O}}_3$

Surfaces of ${\mathrm{In}}_2{\mathrm{O}}_3$ and tin-doped ${\mathrm{In}}_2{\mathrm{O}}_3$ (ITO) were investigated using photoelectron spectroscopy. Parts of the measurements were carried out directly after thin film preparation by magnetron sputtering without breaking vacuum. In addition samples were measured during exposure to oxidizing and reducing gases at pressures of up to $100\mathrm{Pa}$ using synchrotron radiation from the BESSY II storage ring. Reproducible changes of binding energies with temperature and atmosphere are observed, which are attributed to changes of the surface Fermi level position. We present evidence that the Fermi edge emission observed at ITO surfaces is due to metallic surface states rather than to filled conduction band states. The observed variation of the Fermi level position at the ITO surface with experimental conditions is accompanied by a large apparent variation of the core level to valence band maximum binding energy difference as a result of core-hole screening by the free carriers in the surface states. In addition segregation of Sn to the surface is driven by the surface potential gradient. At elevated temperatures the surface Sn concentration reproducibly changes with exposure to different environments and shows a correlation with the Fermi level position.

Figure 1

Surface potentials of an $n$ type semiconductor in flat band condition (a). The work function $\varphi$ can change either by modification of the surface dipole preserving flat bands but modifying the electron affinity $\chi$ and ionization potential $I_P=\chi +E_g$ (b). The work function might also change by bending of the bands at the surface $q{V}_b$ (c). Surface dipole and band bending could also change simultaneously.

Figure 2

Carbon $1s$ core level spectra taken during heating of an ITO thin film sample stored in air (a). The first spectrum recorded at room temperature is shown by circles. Small arrows indicate the time where oxygen has been introduced into the chamber. The corresponding pressures are given in mbar. The intensity of the C $1s$ line in dependence on time is shown in (b) together with the temperature of the substrate.

Figure 3

X-ray photoelectron core level and valence band spectra (top) and UPS spectra (bottom) of ITO (a,b) and ${\mathrm{In}}_2{\mathrm{O}}_3$ (c–e) thin films prepared by radio frequency magnetron. Substrate temperatures were $400°\mathrm{C}$ for samples (a–c) and $200°\mathrm{C}$ for samples (d) and (e), respectively. The films were deposited with pure Ar (a,c,d) and with a $90∕10$ $\mathrm{Ar}∕{\mathrm{O}}_2$ gas mixture (b,e). A magnified view of the UPS spectra close to the Fermi energy (zero binding energy) is also shown ($E_F$ UPS). In this part of the figure the intensity at binding energies $\gtrsim 1\mathrm{eV}$ are mostly due to excitation of valence band states by a satellite of the He discharge lamp.

Figure 4

Line fits of the $\mathrm{In}3d_{5∕2}$ and $\mathrm{O}1s$ lines recorded with monochromatic $\mathrm{Al}\mathrm{K}\alpha$ radiation from an ITO thin film deposited at $400°\mathrm{C}$ substrate temperature in pure argon [spectra (a) in Fig. 3]. The energy difference between the three components is identical. Intensity ratios are $1:0.55:0.14$ for $\mathrm{In}3d$ and $1:0.23:0.05$ for $\mathrm{O}1s$, respectively.

Figure 5

(Color online) Binding energy with respect to the valence band maximum ${\mathrm{BE}}_{\mathrm{VB}}\left(\mathrm{In}3d\right)$ of ${\mathrm{In}}_2{\mathrm{O}}_3$ and ITO thin films in dependence on valence band maximum energy $\mathrm{BE}\left(\mathrm{VB}\right)$ (a). The variation is induced by different deposition parameters. Low values of $\mathrm{BE}\left(\mathrm{VB}\right)$ correspond to more oxidizing conditions. In (b) the same plot is shown for an ITO sample measured in situ during exposure to different atmospheres at different temperatures (indicated in $°\mathrm{C}$). All binding energies are determined from the maximum of the emission line. The lines in (a) are drawn to show the anticipated dependence. Lines in (b) are reproduced from (a).

Figure 6

Concentration of Sn determined from XP spectra of sputter deposited ITO films versus valence band maximum binding energy $\mathrm{BE}\left(\mathrm{VB}\right)$. Low values of $\mathrm{BE}\left(\mathrm{VB}\right)$ correspond to more oxidizing conditions. Values in (a) are absolute concentrations in cation % Sn $\left[I\right(\mathrm{Sn})∕(I\left(\mathrm{Sn}\right)+I\left(\mathrm{In}\right)]$ determined in situ from samples after deposition of films. The films were deposited at $400°\mathrm{C}$ substrate temperature with varying oxygen content in the sputter gas. Values in (b) are relative concentrations determined from integrated intensities of core levels recorded with $h\nu =600\mathrm{eV}$ synchrotron radiation of an ITO sample exposed to different environments at different substrate temperatures (indicated in $°\mathrm{C}$). The lines are drawn to guide the eye.

Figure 7

Core level and valence band photoelectron spectra of an ITO sample hold at different temperatures and atmospheres. The spectra were recorded with $h\nu =600\mathrm{eV}$ synchrotron radiation during exposure to $5\times {10}^{-4}$ mbar oxygen (a,b) and $5\times {10}^{-4}$ mbar hydrogen (c,d). Sample temperatures were $100°\mathrm{C}$ for (a) and (c) and $500°\mathrm{C}$ for (b) and (d), respectively. The spectra are normalized to the electron current in the storage ring and referenced to the Fermi edge emission of the sample.

Figure 8

Possible surface potential distributions leading to the observation of photoelectron emission at the Fermi level from a degenerately $n$-doped semiconductor. (a) Flat band situation: the Fermi level at the surface is inside the conduction band as in the bulk of the material; (b) Surface potential drop: the emission can only occur if a high density of surface states is present at the Fermi energy compensating for the space charge.