廣瀬 和之

The energy distributions of the occupied surface states of GaAs(001) surfaces with several kinds of surface superstructure are measured using photoemission yield spectroscopy. The surfaces are prepared by molecular beam epitaxy. The surface states are found to change in both distribution and density depending on the surface superstructure and the surface preparation conditions. It is concluded that not all observed surface states originate from ordered surface atomic structures.

Surface superstructures (reconstructed structures) have been observed by many techniques. However, it has not been easy to confirm that a superstructure does exist at an interface between two solid layers. The article reports superstructures at semiconductor interfaces studied by a grazing incidence X-ray diffraction technique with use of synchrotron radiation. At metal-semiconductor interfaces, different superstructures give different Schottky-barrier-height values. Superstructures are also found to exist at insulator-semiconductor interfaces fabricated by the metalorgamc chemical vapor deposition (MOCVD) method. In addition, a boron-induced superstructure is found to exist even at the interface between an epitaxial Si layer and a Si (111) substrate.

The energy distributions of the occupied surface states of GaAs (001) surfaces are measured using photoemission yield spectroscopy. The surfaces are prepared by different kinds of techniques, including molecular beam epitaxy, As decapping, and chemical etching. The surface states are found to change in both distribution and density depending on the surface preparation techniques. The origins of the surface states are discussed in terms of surface atomic structures for the atomically clean surfaces. A great reduction by about one order of magnitude in the density of the surface states is revealed for the surface covered with native oxide compared with the other atomically clean surfaces: the surface state electron density is estimated to be about 3 x 10(13) cm-2 for the latter, and approximately 10(12) cm-2 for the former.

The energy distributions of the occupied surface states of GaAs (001) surfaces are measured using photoemission yield spectroscopy. The surfaces are prepared by different kinds of techniques, including molecular beam epitaxy, As decapping, and chemical etching. The surface states are found to change in both distribution and density depending on the surface preparation techniques. The origins of the surface states are discussed in terms of surface atomic structures for the atomically clean surfaces. A great reduction by about one order of magnitude in the density of the surface states is revealed for the surface covered with native oxide compared with the other atomically clean surfaces: the surface state electron density is estimated to be about 3 x 10(13) cm-2 for the latter, and approximately 10(12) cm-2 for the former.

Al/GaAs Schottky barriers are fabricated with 2.5-20-angstrom thick doping layers of Ce of concentrations 10(20) cm-3 and 10(21) cm-3 situated below the GaAs surface. Schottky barrier heights (SBHs) are determined from current- and capacitance-voltage measurements. n-type SBHs decrease with increasing Ce doping layer thickness, while p-type SBHs increase, but to a lesser degree. A cross-sectional image taken by high-resolution transmission electron microscopy shows that Ce is located in the substitutional sites of the doping layers. The changes in the SBHs are attributed to strain induced by Ce atoms in the substitutional sites. The difference in the magnitudes of change observed for n- and p-type SBHs is discussed in relation to the inhomogeneity observed in the density of Ce atoms at the interface regions.

Al/GaAs Schottky barriers are fabricated with 2.5-20-angstrom thick doping layers of Ce of concentrations 10(20) cm-3 and 10(21) cm-3 situated below the GaAs surface. Schottky barrier heights (SBHs) are determined from current- and capacitance-voltage measurements. n-type SBHs decrease with increasing Ce doping layer thickness, while p-type SBHs increase, but to a lesser degree. A cross-sectional image taken by high-resolution transmission electron microscopy shows that Ce is located in the substitutional sites of the doping layers. The changes in the SBHs are attributed to strain induced by Ce atoms in the substitutional sites. The difference in the magnitudes of change observed for n- and p-type SBHs is discussed in relation to the inhomogeneity observed in the density of Ce atoms at the interface regions.

The interfacial superstructures of Sb/GaAs(001) contacts are observed by use of the grazing-incidence x-ray-diffraction method. The Schottky-barrier heights for these contacts are also determined by using the current-voltage or the capacitance-voltage methods. The Schottky-barrier height of the contact with the (1x6) interfacial superstructure is found to be larger than that of the contact with the (1x4) interfacial superstructure by as much as 0.1 eV. This result indicates the importance of the local electronic structure for Schottky-barrier formation at the metal-GaAs interfaces. The present results are discussed in connection with the current Schottky-barrier models.

The interfacial superstructures of Sb/GaAs(001) contacts are observed by use of the grazing-incidence x-ray-diffraction method. The Schottky-barrier heights for these contacts are also determined by using the current-voltage or the capacitance-voltage methods. The Schottky-barrier height of the contact with the (1x6) interfacial superstructure is found to be larger than that of the contact with the (1x4) interfacial superstructure by as much as 0.1 eV. This result indicates the importance of the local electronic structure for Schottky-barrier formation at the metal-GaAs interfaces. The present results are discussed in connection with the current Schottky-barrier models.

Different 4×1 superstructures are found for Yb/GaAs(001) contacts with different Yb thicknesses. To clarify the role of Yb atoms in the 4×1 structure, an x-ray anomalous-dispersion effect has been used in the grazing-incidence x-ray-diffraction geometry. Diffraction intensities measured in the x-ray energy range near the Yb LIII absorption edge show a characteristic variation, which reflects the energy dependence of the real part of the anomalous scattering factor (f) for Yb atoms. As a result, for all the samples, Yb atoms are found to order in 4×1 structures. The arrangements of the Yb atoms in the 4×1 unit cell are derived from Buergers decomposition method of the Patterson function. By comparison of the calculated diffraction intensities with observed intensities, two different (4×1) structural models for the thin and thick Yb-layer samples are found. Different Schottky-barrier-height values are also found for these samples. © 1991 The American Physical Society.

Different 4×1 superstructures are found for Yb/GaAs(001) contacts with different Yb thicknesses. To clarify the role of Yb atoms in the 4×1 structure, an x-ray anomalous-dispersion effect has been used in the grazing-incidence x-ray-diffraction geometry. Diffraction intensities measured in the x-ray energy range near the Yb LIII absorption edge show a characteristic variation, which reflects the energy dependence of the real part of the anomalous scattering factor (f) for Yb atoms. As a result, for all the samples, Yb atoms are found to order in 4×1 structures. The arrangements of the Yb atoms in the 4×1 unit cell are derived from Buergers decomposition method of the Patterson function. By comparison of the calculated diffraction intensities with observed intensities, two different (4×1) structural models for the thin and thick Yb-layer samples are found. Different Schottky-barrier-height values are also found for these samples. © 1991 The American Physical Society.

The ionization energies for GaAs(001) are determined by photoemission-yield spectroscopy as a function of surface superstructure. The ionization energy changes by as much as 0.5 eV in accordance with the surface superstructure. © 1990 The American Physical Society.

Epitaxial (111)A1/(111)GaAs Schottky contacts are formed using molecular beam epitaxy. The epitaxial relationship is determined by transmission electron microscopy. The interface is found to be abrupt and of an atomic order. Schottky barrier heights are measured by current-voltage and capacitance-voltage methods. The Schottky barrier height for a (111) surface is found to be stable under 450 °C annealing in a N2 atmosphere.

The ionization energies for GaAs(001) are determined by photoemission-yield spectroscopy as a function of surface superstructure. The ionization energy changes by as much as 0.5 eV in accordance with the surface superstructure. © 1990 The American Physical Society.

金属/半導体界面の典型的な電気特性である, ショットキー障壁に関する研究の現状を, SiとGaAsについて簡単に紹介する。理想的なシリサイド/Siショットキー障壁においては, ショットキー障壁高さ (SBH) がシリサイドバルクの性質によって変化しており, 弱いピニング現象しか起きていない。エピタキシャルNiSi₂/Siショットキー障壁に対する理解は, 大型計算機物理の進歩と共に進んでいる。一方, 金属/GaAsショットキー障壁に対しては, 金属誘起準位と界面欠陥準位の両界面準位がショットキー障壁決定によるというモデルが支持されている。しかし, 元来のショットキー理論が成立するという結果も報告されており, その機構は未だに解明されていない。GaAsの上に多結晶金属膜が形成され, 化学反応や相互拡散が起こっている界面に, 界面超構造の存在と共に界面超構造に依存したSBHの違いが見いだされる。さらに, 分子線エピタキシャル成長法により, Al/GaAsショットキー界面に, 希土類金属を含む中間層あるいはドーピング層を挿入することによって, SBHの制御が可能であることが見いだされる。超LSIのコンタクト特性の制御においては, 障壁形成前の半導体表面の清浄化も重要な課題である。従来清浄表面とされてきたSi (111) -7×7超構造は, 吸着原子のバックボンドに酸素原子を含む構造である疑いが生じている。