2.4.1 Technology Background

In the early 1980's the STM was developed at IBM Zürich as an instrument to image surfaces at the atomic level [13]. A schematic of the basic operations of a STM is shown in Figure 2.14. After the initial patent, the scientists involved won a Nobel Prize in Physics for their design in 1986 [15]. Countless publications followed, as the STM was perfected as a tool to utilize tunneling current between a conductive metal tip and a sample surface in order to detect depressions and protuberances on a nanometer sized section of the test surface. The test surfaces initially considered include CaIrSn$_4$(110) [16], Au(110) [16], Si(111) [17], and GaAs(111) [14]. The STM allows for a 0.1nm lateral resolution and a 0.01nm depth resolution [7]. Shortly after, the discoverer of the STM, Binnig was involved in the discovery of the AFM [12], which has a much higher resolution and is today one of the foremost tools for imaging surfaces at the atomistic level [74], [114], [116], [184], some even suggesting that different bond orders of individual carbon-carbon bonds can be visualized using an AFM [67], [173], [223]. The AFM monitors the surface by sensing the van der Waals force between the tip and the surface, in order to achieve a much finer resolution than the STM. A basic schematic of AFM operation is shown in Figure 2.15, where a cantilever is brought close to a test surface, while the movements of the cantilever needle are recorded by the photodiode. The AFM has been used extensively, not only in the semiconductor industry, but also in physics, chemistry, biology, biochemistry, and other disciplines where the chemical or physical properties of a surface are required [136], [207]. An additional advantage of AFM over STM is that AFM offers the advantage of realizing local oxidation and reading the topography of the generated pattern [197].

Figure 2.14: Typical STM schematic for surface imaging.

Figure 2.15: Typical AFM schematic for surface imaging.

LON of semiconductor surfaces was first suggested by Dagata et al. at the National Institute of Standards and Technology, where a STM was used in order to chemically modify a hydrogen-passivated Si(111) surface in an air ambient with 100nm precision [38]. By 1993, Day and Alle used an AFM in order to generate sub-100nm SiO$_2$ lines on a silicon wafer, thereby introducing the AFM as a lithographic tool and solving the issues that were persistent with STM lithography [40]. The main disadvantage of STM is that the tip sample spacing cannot be chosen independently from the tunneling current and tip voltage. Therefore, the tip can often crash into the insulating structure, because it is frequently not possible to keep the tip above the highest insulating feature while simultaneously having the optimum bias voltage and tunneling current for ultrathin resolution [40]. Both methods worked in such a way that a potential difference was applied between the tip (STM or AFM) and the sample substrate. A positive bias voltage is applied to the STM tip, while a negative was applied to the AFM tip, with respect to the silicon substrate. The negative voltage results in the ability to produce thicker insulating oxides [4], [40], [73], [191], [231].