廣瀬 和之

We studied HfAlOx(N)/SiO2/Si films which were fabricated by the layer-by-layer deposition and annealing (LL-D&amp A) method with different annealing conditions. In this time, in-situ annealing was performed at various temperatures in an NH3 ambient. In addition, post-deposition annealing (PDA) was performed for some samples. For each sample, the interfacial lattice strain was evaluated using extremely asymmetric X-ray diffraction and the local dielectric constant near the Al atoms was measured by X-ray photoelectron spectroscopy (XPS). Observation of the strain field was done by measuring the X-ray rocking curve of the Si 113 reflection of the Si (001) substrate under grazing incidence conditions. It was found that in the case of the samples without PDA, for higher in-situ annealing temperatures compressive strain is introduced and the local dielectric constant becomes lower. For the samples with PDA, the differences of the lattice strain and the local dielectric constant are small for different in-situ annealing temperatures. © 2007 IOP Publishing Ltd.

We studied HfAlOx(N)/SiO2/Si films which were fabricated by the layer-by-layer deposition and annealing (LL-D&amp A) method with different annealing conditions. In this time, in-situ annealing was performed at various temperatures in an NH3 ambient. In addition, post-deposition annealing (PDA) was performed for some samples. For each sample, the interfacial lattice strain was evaluated using extremely asymmetric X-ray diffraction and the local dielectric constant near the Al atoms was measured by X-ray photoelectron spectroscopy (XPS). Observation of the strain field was done by measuring the X-ray rocking curve of the Si 113 reflection of the Si (001) substrate under grazing incidence conditions. It was found that in the case of the samples without PDA, for higher in-situ annealing temperatures compressive strain is introduced and the local dielectric constant becomes lower. For the samples with PDA, the differences of the lattice strain and the local dielectric constant are small for different in-situ annealing temperatures. © 2007 IOP Publishing Ltd.

A scan flip-flop (FF) is designed to observe both single event transient (SET) and single event upset (SEU) soft errors in logic VLSI systems. The SET and SEU soft errors mean the upset caused by latching an SET pulse that originates in combinational logic blocks and the upset caused by a direct ion hit to the FF, respectively. An irradiation test method using the scan FF is proposed to obtain SET and SEU soft-error rates at each FF distributed in logic VLSI systems. A test chip is designed using a 0.2-mu m fully-depleted silicon-on-insulator standard cell library. The basic concepts have been validated with Verilog timing simulations. The cell-level implementation costs of the proposed scan FF are estimated to be reasonable.

We propose a new method to estimate the local dielectric constant of an ultrathin gate insulator film formed on Si substrates by using X-ray photoelectron spectroscopy (XPS). First we measure the difference of core-level binding energy shifts for Si Is and Si 2p, ΔE1s - ΔE 2p, for various Si compounds using high-resolution high-energy synchrotron radiation. We find that the ΔE1s - ΔE 2p values are in very good correlation with the dielectric constant values of the Si compounds. Then, using this relation, we deduce the local dielectric constant for ultrathin SiO2 film formed on Si substrates. The results are in good agreement with values predicted by a first-principles calculation. © 2006 IEEE.

A scan flip-flop (FF) is designed to observe both single event transient (SET) and single event upset (SEU) soft errors in logic VLSI systems. The SET and SEU soft errors mean the upset caused by latching an SET pulse that originates in combinational logic blocks and the upset caused by a direct ion hit to the FF, respectively. An irradiation test method using the scan FF is proposed to obtain SET and SEU soft-error rates at each FF distributed in logic VLSI systems. A test chip is designed using a 0.2-mu m fully-depleted silicon-on-insulator standard cell library. The basic concepts have been validated with Verilog timing simulations. The cell-level implementation costs of the proposed scan FF are estimated to be reasonable.

We describe state-of-the-art photoelectron spectroscopy studies of SiO(2)/Si interfaces that play fundamental roles in metal-oxide-semiconductor (MOS) field-effect transistors. We show comprehensive photoelectron spectra Of SiO(2)/Si interfaces, which were taken from SiO(2)/Si samples of extremlye high-quality by the high-resolution photoelectron spectroscopy technique with either synchrotron or laboratory X-ray sources. The spectra discussed here include the Si 2p, the Si 1s, the O 2s, and N 1s core-level spectra, and the valence-band spectra. We perform quantitative analysis using selected values of the photoionization cross-section and the electron escape depth, which is governed by both inelastic scattering and elastic scattering in SiO(2). On the other hand, we analyze peak energies by considering peak energy shifts that are due to several factors. Atomic structures are discussed in terms of intermediate oxidation states at SiO(2)/Si(1 0 0) interfaces and strained Si-O-Si bonds near the interfaces, while electronic structures are discussed in terms of valence-band offset at the interfaces and dielectric constants near the interfaces. Applications of photoelectron spectroscopy study to advanced oxide formation are also shown in terms of depth profiling of oxynitride films and interface structures of low-temperature oxide. (C) 2006 Elsevier Ltd. All rights reserved.

The statistical transient response of floating body SOI and bulk devices is measured under proton and heavy ion irradiation. The influence of the device architecture is analyzed in detail for several generations of technologies, from 0.25 mu m to 70 nm. The effects of the measured transients on SET sensitivity are investigated. The amount of collected charge and the shape of the transient currents are shown to have a significant impact on the temporal width of propagating transients. Finally, based on our measured data, the threshold LET and the critical transient width for unattenuated propagation are calculated for both bulk and floating body SOI as a function of technology scaling. We show that the threshold LETs and the critical transient widths for bulk and floating body SOI devices are similar. Body ties can be used to harden SOI ICs to digital SET. However, the primary advantage of SOI technologies, even with a floating body design, mostly lies in shorter transients, at a given ion LET, for SOI technologies than for bulk technologies.

Current components of heavy-ion-induced transient currents in a 0.2-mu m fully-depleted SOI MOSFET are analyzed in the time domain. The analysis demonstrates that the transient currents have another slow-decay current component that is different from the two conventional current components: a prompt discharge current and a slow-decay parasitic bipolar current. The slow-decay component revealed here is a flow of deposited carriers stored in the body region to maintain quasi-neutrality, and it drastically widens the transient pulse.

The statistical transient response of floating body SOI and bulk devices is measured under proton and heavy ion irradiation. The influence of the device architecture is analyzed in detail for several generations of technologies, from 0.25 mu m to 70 nm. The effects of the measured transients on SET sensitivity are investigated. The amount of collected charge and the shape of the transient currents are shown to have a significant impact on the temporal width of propagating transients. Finally, based on our measured data, the threshold LET and the critical transient width for unattenuated propagation are calculated for both bulk and floating body SOI as a function of technology scaling. We show that the threshold LETs and the critical transient widths for bulk and floating body SOI devices are similar. Body ties can be used to harden SOI ICs to digital SET. However, the primary advantage of SOI technologies, even with a floating body design, mostly lies in shorter transients, at a given ion LET, for SOI technologies than for bulk technologies.

Heavy-ion-induced SET-pulse widths in NOR-logic cells fabricated by a 0.2-mu M FD-SOI technology are directly measured by using an on-chip self-triggering Flip-Flop circuit. The pulse widths are distributed. from 0.3 to 1.0 ns under a constant LET of 40 MeV center dot cm(2)/mg.

Current components of heavy-ion-induced transient currents in a 0.2-mu m fully-depleted SOI MOSFET are analyzed in the time domain. The analysis demonstrates that the transient currents have another slow-decay current component that is different from the two conventional current components: a prompt discharge current and a slow-decay parasitic bipolar current. The slow-decay component revealed here is a flow of deposited carriers stored in the body region to maintain quasi-neutrality, and it drastically widens the transient pulse.

The authors measure the difference of core-level binding energy shifts for Si 1s and Si 2p, Delta E-1s-Delta E-2p, for various Si compounds using high-resolution high-energy synchrotron radiation. They find that the Delta E-1s-Delta E-2p values are in very good correlation with the dielectric constant values of the Si compounds. Using this relation, they deduce the local dielectric constant for each of the Si intermediate oxidation states formed at the SiO2/Si interface. The results are in good agreement with values predicted by a first-principles calculation. (c) 2006 American Institute of Physics.

The authors measure the difference of core-level binding energy shifts for Si 1s and Si 2p, Delta E-1s-Delta E-2p, for various Si compounds using high-resolution high-energy synchrotron radiation. They find that the Delta E-1s-Delta E-2p values are in very good correlation with the dielectric constant values of the Si compounds. Using this relation, they deduce the local dielectric constant for each of the Si intermediate oxidation states formed at the SiO2/Si interface. The results are in good agreement with values predicted by a first-principles calculation. (c) 2006 American Institute of Physics.

MOSデバイスの微細化に伴い、ゲートSiO_2膜の厚さは0.8nmにまで達した。このような極薄膜の領域では、金属/SiO_2界面のバリアハイトは測定されていない。本研究ではXPSを用いてSiO_2膜1.0nm以下のAu/SiO_2界面のバリアハイトを測定する。界面バリアハイトをXPSで測定する方法として、XPSでSi2p束縛エネルギーから求める方法と、カットオフエネルギーから求める方法の二つを適用した結果を報告する。

We measure the relative chemical shift between Si 1s and Si 2p, Delta E-1s - Delta E-2p, for 0.20-1.96nm-thick SiO2 films formed on Si substrates using high-resolution high-energys x-ray radiation. It is found that Delta E-1s - Delta E-2p is independent of SiO2 film thickness for films thicker than 1.0nm. whereas it is smaller for films thinner than 0.5 nm. The result, in conjunction with first principles molecular orbital calculations, indicates that the valence charge of the Si atom is larger for in ultrathin SiO2 films than in the thicker SiO2 films.