The interface between silicon and a high-k oxide

The ability of the semiconductor industry to continue scaling microelectronic devices to ever smaller dimensions (a trend known as Moore's Law) is limited by quantum mechanical effects: as the thickness of conventional silicon dioxide (SiO(2)) gate insulators is reduced to just a few atomic layers, electrons can tunnel directly through the films. Continued device scaling will therefore probably require the replacement of the insulator with high-dielectric-constant (high-k) oxides, to increase its thickness, thus preventing tunnelling currents while retaining the electronic properties of an ultrathin SiO(2) film. Ultimately, such insulators will require an atomically defined interface with silicon without an interfacial SiO(2) layer for optimal performance. Following the first reports of epitaxial growth of AO and ABO(3) compounds on silicon, the formation of an atomically abrupt crystalline interface between strontium titanate and silicon was demonstrated. However, the atomic structure proposed for this interface is questionable because it requires silicon atoms that have coordinations rarely found elsewhere in nature. Here we describe first-principles calculations of the formation of the interface between silicon and strontium titanate and its atomic structure. Our study shows that atomic control of the interfacial structure by altering the chemical environment can dramatically improve the electronic properties of the interface to meet technological requirements. The interface structure and its chemistry may provide guidance for the selection process of other high-k gate oxides and for controlling their growth.     

Revised rates for the stellar triple-alpha process from measurement of 12C nuclear resonances

In the centres of stars where the temperature is high enough, three alpha-particles (helium nuclei) are able to combine to form 12C because of a resonant reaction leading to a nuclear excited state. (Stars with masses greater than approximately 0.5 times that of the Sun will at some point in their lives have a central temperature high enough for this reaction to proceed.) Although the reaction rate is of critical significance for determining elemental abundances in the Universe, and for determining the size of the iron core of a star just before it goes supernova, it has hitherto been insufficiently determined. Here we report a measurement of the inverse process, where a 12C nucleus decays to three alpha-particles. We find a dominant resonance at an energy of approximately 11 MeV, but do not confirm the presence of a resonance at 9.1 MeV (ref. 3). We show that interference between two resonances has important effects on our measured spectrum. Using these data, we calculate the triple-alpha rate for temperatures from 10(7) K to 10(10) K and find significant deviations from the standard rates. Our rate below approximately 5 x 10(7) K is higher than the previous standard, implying that the critical amounts of carbon that catalysed hydrogen burning in the first stars are produced twice as fast as previously believed. At temperatures above 10(9) K, our rate is much less, which modifies predicted nucleosynthesis in supernovae.