Fig.4. Voltage tunneling spectra on a 20nm Ti02 film on titanium (a) and on a nitrogen-implanted region of the same sample. Initial conditions were -3 V (-1 nA setpoint), scan rate 3 V/s; W tip, samples were dried with nitrogen.

This is an additional evidence for the above mentioned partial penetration of the tip into the surface. Figure 5(b) shows distance tunneling spectra on a 5 nm oxide-covered titanium. The spectrum was started from the normal STM conditions (-1 V bias, -1 nA setpoint), and the tip was then retracted from the surface. According to the Gamov formula exponential behavior is expected. With a barrier of some electronvolts the tunnel current should decrease within 0.1-0.2 nm to 1/10 of the initial value. On these surfaces however, the tunnel current decreases to 1/10 of the initial (-InA) value at more than 2 nm, validating the above theory of partial penetration. If the apparent barrier height is calculated according to the Gamov formula, a value of only some millielectronvolts is obtained. It is interesting to compare the spectra under dry nitrogen and wet air (I and H in Fig. 5(b)). The spectra were recorded directly after each other with identical parameters, except that between the recording of spectra the surface/tip system was flushed with a stream of dried nitrogen for 10 min. The wet air leaves a water film on the surface and the tip will be surrounded by this water film as well water due to capillary forces. Even while the tip is retracted for the distance spectrum, this water film will allow tunnel current to flow over a larger distance than without this film, resulting in a smaller dependency on the distance, as seen in Fig. 5(b).

Bias /V

Fig. 5. A) Voltage tunneling spectra on 5/15/40 rnn oxide covered Sputter-titanium recorded with Au-tip, under dry nitrogen atmosphere. B) Distance tunneling spectrum on an 5 nm oxide covered titanium, under dry nitrogen (I) and wet air (II), scan rate 3 nm/s, bias -IV.

Rel. distance / am

Fig. 5. A) Voltage tunneling spectra on 5/15/40 rnn oxide covered Sputter-titanium recorded with Au-tip, under dry nitrogen atmosphere. B) Distance tunneling spectrum on an 5 nm oxide covered titanium, under dry nitrogen (I) and wet air (II), scan rate 3 nm/s, bias -IV.

4 Discussion

The titanium surfaces were investigated to gain insight into mechanical and electronic properties. With AFM and related methods it was possible to identify grains (and their orientation in correlation to optical data) and grain boundaries. The surface could be prepared in such a way that the single grains had a flat surface with substructures of only some nanometers. The height differences between the grains were located at the grain boundaries with dz/dx gradients of only 5-10%, if the surface was carefully electropolished. However, if the titanium was strongly electropolished the grain boundaries showed height differences of several hundreds of nanometers and gradients of more than 50%. Thin anodically prepared TiC>2 films have a typical roughness (in the z-direction) of 1-3 nm and substructure in the xy-dimension of 20-50 nm. Other investigations in the literature have found the same substructures [3, 8, 10] on similar surfaces. Only Rohrer et al. [18] have found atomic resolution on a reduced single crystal of Ti02, with special preparation. Since most surfaces were prepared in a different way, it is difficult to compare these results. The chemical and physical properties of modified Ti02 surface films (either by reducing in hydrogen or implating nitrogen) are not well defined. It is necessary to apply all the possible analytical tools in order to obtain a full picture of these surfaces. That includes all SPM methods, but also XPS, Auger, classical electrochemistry, photoelectrochemistry, or ellipsometry. In this paper we have demonstrated the SPM part of these investigations. The grains exhibited different mechanical properties (as seen in the dF/ds and LFM pictures). Comparisons with experiments in the light microscope, microelectrochemistry, and microellipsometry [12, 13, 19] have shown that grains with different orientation (either on a single crystal with defined orientation or on fine-grain titanium with approximate orientation of the individual grains) have different mechanical, electronic, and electrochemical properties. The SPM data add a new dimension to these experiments due to the increased resolution and mechanical and electronic data of individual grains. It is impossible to determine surface properties in the nanometer range (xy- and z-dimension) with any of the other methods mentioned. The tunnel spectra have shown that the tip is in contact with the surface, resulting in a more or less symmetrical spectrum in the anodic as well as in cathodic range. However, it is also possible to record spectra with n-type semiconducting behavior (as would be expected for anodically prepared Ti02 films), if the bias is changed to higher values [17]. For these experiments the initial tip-sample distance was increased, resulting in less or no penetration of the tip into the oxide. All tunnel spectra were recorded at one position, without movement of the tip in the xy-directions during any of the measurements. However, STM experiments (as seen in Fig. 2) require this movement, resulting in bad or unstable conditions. If the tip is not moved, then the feedback system has time to drive the tip gently into the surface to get a tunnel current. But while it is moving over the surface this is not possible and the tip crashes (damaging the tip and the surface, resulting in irreproducible conditions). We have shown here that with all SXM techniques it is possible to obtain data even on these surfaces. In addition to information on the topography, we achieved a correlation between grain orientation, mechanical, electronic and electrochemical properties.

Acknowledgments. We thank the Fonds des Verbandes der Chemischen Industrie for the grant for C. Kobusch and the Ministerium für Wissenschaft und Forschung (NRW) for the financial support of this work.

0 0

Post a comment