Precision Oxide Layer Control Could Extend Semiconductor Road Map

From oemagazine January 2005
02 January 2005
By Charles Whipple

The blue plume inside the University of Tokyo crystal growth chamber is caused by the pulsed laser deposition process.

Two or three years hence, semiconductors will face a major problem. In the quest for ever-smaller elements insulated by ever-thinner layers, silicon will reach its limit, and without new technology or materials, quantum leaps in semiconductor performance will be impossible.

The possibility of overcoming this roadblock lies in complementing silicon and silicon oxides with other materials. Scientists currently concentrate on oxides as prime candidates because of their properties, which range from complete insulation to superconducting. A team of researchers from the University of Tokyo (Tokyo, Japan) and Cornell University (Ithaca, NY) recently showed that the electronic properties of strontium titanate (SrTiO3) can be controlled to nanometer precision, an announcement that caused considerable interest within the information industry.

Harold Hwang, University of Tokyo spokesman for the research team, says, "Oxygen vacancies are particularly important in SrTiO3 thin films because they tend to retain high carrier mobilities, even at high carrier densities." Hwang and his colleagues successfully fabricated SrTiO3 superlattice films with supernanometer-precise oxygen doping profiles, and demonstrated absolute detection sensitivities of one to four oxygen vacancies. "With our findings," Hwang says, "we open the way to microscopic study of individual vacancies and their clustering. And, with our methodology, this can be done with oxides and with other crystalline materials as well."

Jochen Mannhart of the Centre for Electronic Correlations and Magnetism, Institute of Physics, University of Augsburg (Augsburg, Denmark) says, "Heretofore, doping usually meant replacing one cation with a different cation, that is, an impurity. But the ability to dope films without impurities is an enticing advantage. It makes one wonder if the technique will be applicable to other ionic materials."

Darrell Schlom of the Department of Materials Science and Engineering at Penn State University (University Park, PA) adds, "In the relatively few years since Hwang entered the field of oxide thin films, he has stepped to the top and has done and continues to do fantastic work."

Hwang and Akira Ohtomo of Tohoku University's Institute for Materials Research (Sendai, Japan) recently delved further into growth mode control of free carrier density in SrTiO3­δ films. "We found the two dominant factors determining growth mode in these films are the kinetics of surface crystallization and oxidation," Hwang says. "By matching these factors, we can grow persistent two-dimensional layers, and by tuning the factors, we can freeze oxygen vacancies in thin films. That means we can achieve controlled, systematic doping across a metal-insulator transmission. In fact, we grew metallic films that exhibited Hall mobilities a high as 25,000 cm2/Vs."

According to Hwang, precision control of SrTiO3­δ free carrier density can change these thin films from dielectric insulator, to doped semiconductor, to metal, to superconductor within the first 0.03% of oxygen vacancies.

Hwang's team mapped growth mode transitions from 3-D island growth to 2-D layer-by-layer growth by varying the temperature (Tg) and the oxygen partial pressure (PO2) during growth. "We found that optimal conditions for extended layer-by-layer growth occur when the time constant for cystallization (τcyrst) of the adatom species matches the surface oxidation time constant (τox). This also marks a key threshold for the electronic properties of the films, because when τcyrst < τox, oxygen vacancies are frozen in the growing film, resulting in free electron carriers."

The team grew homoepitaxial SrTiO3 films 1000 Å thick by pulsed laser deposition (PLD) on atomically flat, titanate-terminated (001) SrTiO3 substrates. The repetition rate and fluence of the laser were held constant while Tg and PO2 were varied. They found that optimal layer-by-layer growth was realized on the verge of 3-D growth when τcyrstox. Also, free carriers suddenly appeared as τcyrst < τox, and the resulting carrier density reflected the competing rates for crystallization and oxidation, which dominated the electrical properties of the films. "In all, our test results indicated a nontrivial modification of the lattice arising from the non-equilibrium kinetics of PLD growth," Hwang says.

Recent News