Patterned fabrication depends on selective deposition that can be best achieved with atomic layer deposition (ALD). For the growth of TiO2 by ALD using TiCl4 and H2O, X-ray photoelectron spectroscopy reveals a marked difference in growth on oxidized and hydrogen-terminated silicon surfaces, characterized by typical and predictable deposition rates observed on SiO2 surfaces that can be 185 times greater than the deposition rates on hydrogen-terminated Si(100) and Si(111) surfaces. Large-scale patterning is demonstrated using wet chemistry, and nanometer-scale patterned TiO2 growth is achieved through scanning tunneling microscopy (STM) tip-based lithography and ALD. The initial adsorption mechanisms of TiCl4 on clean, hydrogen-terminated, and OH-terminated Si(100)-(2 × 1) surfaces are investigated in detail through density functional theory calculations. Varying the reactive groups on the substrate is found to strongly affect the probability of precursor nucleation on the surface during the ALD process. Theoretical studies provide quantitative understanding of the experimental differences obtained for the SiO2, hydrogen-terminated, and clean Si(100) and Si(111) surfaces.

In this work, the authors used density-functional theory methods and x-ray photoelectron spectroscopy to study the chemical composition and growth rate of HfO 2 , Al 2 O 3 , and TiO 2 thin films grown by in-situ atomic layer deposition on both oxidized and hydrogen-terminated Si(001) surfaces. The growth rate of all films is found to be lower on hydrogen-terminated Si with respect to the oxidized Si surface. However, the degree of selectivity is found to be dependent of the deposition material. TiO 2 is found to be highly selective with depositions on the hydrogen terminated silicon having growth rates up to 180 times lower than those on oxidized Si, while similar depositions of HfO 2 and Al 2 O 3 resulted in growth rates more than half that on oxidized silicon. By means of density-functional theory methods, the authors elucidate the origin of the different growth rates obtained for the three different precursors, from both energetic and kinetic points of view. V

Lithographic precision is as or more important than resolution. For decades, the semiconductor industry has been able to work with Ϯ5% precision. However, for other applications such as micronanoelectromechanical systems, optical elements, and biointerface applications, higher precision is desirable. Lyding et al. ͓Appl. Phys. Lett. 64, 11 ͑1999͔͒ have demonstrated that a scanning tunneling microscope can be used to remove hydrogen ͑H͒ atoms from a silicon ͑100͒ 2 ϫ 1 H-passivated surface through an electron stimulated desorption process. This can be considered e-beam lithography with a thin, self-developing resist. Patterned hydrogen layers do not make a robust etch mask, but the depassivated areas are highly reactive since they are unsatisfied covalent bonds and have been used for selective deposition of metals, oxides, semiconductors, and dopants. The depassivation lithography has shown the ability to remove single H atoms, suggesting the possibility of precise atomic patterning. This patterning process is being developed as part of a project to develop atomically precise patterned atomic layer epitaxy of silicon. However, significant challenges in sample preparation, tip technology, subnanometer pattern placement, and patterning throughput must be overcome before an automated atomic precision lithographic technology evolves.

No abstract