Nanoscale elemental analysis and applications using STM combined with brilliant hard X-rays

Analyses by scanning tunneling microscopy (STM) combined with brilliant X-rays from synchrotron radiation (SR) can provide various possibilities of original and important applications. The STM observation under inner-shell excitation at a specific core-level enables us to analyze the elements or control the local reaction with the high spatial resolution of STM [1]. We have recently demonstrated the elemental analyses with a spatial resolution lower than 2 nm on semiconductor surfaces [2]. The principle of our analyses is not to collect the secondary electrons by STM tip (that may damage the spatial resolution), but to extract the element-specific modulation of the ”tunneling current” succeeding the core-excitation process, which contains truly local information. A key to accomplish successful results is to effectively increase the signal to noise (S/N) ratio. On this purpose, we developed a special SR-STM system. The experimental setup is shown in Fig.1. To surmount a tiny core-excitation efficiency by hard X-rays, we focused two-dimensionally an incident beam having the highest photon density at the SPring-8. Many problems derived from the high brilliance (thermal and electrical noise, damage of STM scanner, instability such as thermal drift, etc.) were solved by the special apparatus and system [1]. Furthermore, we developed a special tip [3] (that can eliminate the noisy electrons coming from a wide area) and signal acquisition system that realizes a high signal to noise ratio to obtain a small modification of the tunneling current originating from the core excitation. After first results on a semiconductor hetero-interface (Si(111)7x7-Ge) [1], second results on the nanoscale elemental analysis were acquired for metal-semiconductor interface (Ge(111)-Cu nanodomains) [2]. For both cases, the spatial resolution of the analysis was estimated to be 1~4 nm, and it is worth noting that the measured domains had a thickness of less than one atomic layer (Fig.2). After progresses of the measurement system and techniques, we succeeded in obtaining a series of successive STM images at an atomically same area without serious drift or sample damages. Accordingly, we could acquire a linear dependence of the element contrast on the incident photon density. The photon density dependence of the elemental contrast will give an important clue to know the origin of the element contrast. Actually, our result on the linear dependence of the element contrast on the photon density suggests that we can deny a possibility of the local potential change derived from the core excitation, because the potential should give an exponential dependence of the contrast on the incident photon density. Also we could recently measured scanning tunneling spectroscopy (STS), which have long been impossible because of instability due to brilliant X-ray irradiation. STS information gives us more direct hint to approach the mechanism of contrast to obtain a higher resolution. It is notable that the image in Fig. 2(c) shows the contrast originating from the chemical difference (that is not based on the surface step height), presenting the structures different from the conventional topographic (Fig. 2(b)) image.