Abstract

The effect of the spreading resistance on the bileg thermoelectric generator (TEG) performance was experimentally evaluated. In planar bileg-TEGs, the width ratio of the p- and n-type legs should be carefully selected to compensate for the impedance mismatch between them and to maximize thermoelectric power generated from a unit area. In the bileg-TEG at the μm-scale, the electrical resistance becomes larger than a simple estimate using lumped parameter circuit model, which is caused by the spreading resistance; when a current flows from a narrower leg to a wider leg. A distance of greater than about 10 μm is required to distribute the electric current over the entire region of the wider leg. At shorter leg lengths, it is better to align the widths of p- and n-legs to maximize the areal power density of TEG. Decreasing the electrical resistance of the wiring between the p- and n-legs is also effective in enhancing the performance of the miniaturized TEG. The width of the p- and n-type legs in the bileg-TEG at the μm-scale should be carefully selected.

Abstract

The performance of a thermoelectric (TE) generator consisting of GeSn wire is experimentally found to be higher than that of a TE generator fabricated by Si wire. The TE generators are developed in a cavity-free architecture, where the wires are directly placed on the substrate without forming a cavity space underneath. In the cavity-free structure, the heat current flows perpendicularly to the substrate and the TE generator is driven by a steep temperature gradient established around the heater inlet. With an identical patterning design, the TE performance of both generators is characterized by varying lengths. The maximum Seebeck coefficient of the generator consisting of GeSn is −277 μV K⁻¹ and that for the Si is −97 μV K⁻¹. The GeSn-TE generator achieves a higher power factor of 31 μW· K⁻²· cm⁻¹ than that of the Si-TE generator of 12 μW· K⁻²· cm⁻¹. The maximum areal power density of the GeSn-TE generator is intrinsically higher than that of the Si-TE generator by approximately 2.5 to 6 times considering the wire thickness difference. The obtained results support the superiority of the GeSn-TE generator over the Si-TE generator.

We report on the behavior of an acoustic phonon spectral linewidth of bulk single-crystalline Si_{1− x}Ge _x alloy with the x of 0.16, 0.32, and 0.45 in the phonon dispersion relation along the Γ–X ([00 q]) direction. Broadening of both transverse acoustic (TA) and longitudinal acoustic (LA) modes of the bulk Si_{1− x}Ge _x alloy was directly observed using inelastic x-ray scattering (IXS) with increasing momentum (from Γ to X points in the Brillouin zone), which cannot be observed in pure Si or pure Ge. The IXS spectral linewidth of the TA mode indicated Ge dependence, which suggests the overlapping of a low-energy local vibration mode (LVM) caused by Ge clusters surrounded by Si atoms around the X point. Although the behavior of the IXS spectral linewidth of the LA mode showed almost no dependence on Ge fraction, the IXS spectra of the LA mode indicated broadening after crossing with a low-energy LVM with increasing momentum. The results obtained by molecular dynamics showed almost the same behavior of the acoustic phonon spectral linewidth. These results suggest that a change in the acoustic phonon spectral linewidth between the Γ and X points indicates a reduction in the acoustic phonon lifetime caused by the appearance of a localized mode originated from a random atom position in the alloy structure, leading to suppression of the thermal transport in the SiGe alloy.

Abstract

We investigate Sn incorporation effects on the thermoelectrical characteristics of n-type Ge-rich Ge_1−x−ySi_xSn_y layers (x ≈ 0.05−0.1, y ≈ 0.03) pseudomorphically grown on semi-insulating GaAs(001) substrates by molecular beam epitaxy. Despite the low Sn content of 3%, the Sn atoms play a role in suppressing the thermal conductivity from 13.5 to 9.0 Wm⁻¹ K⁻¹ without degradation of the electrical conductivity and the Seebeck coefficient. Furthermore, a relatively high power factor (maximum: 14 μW cm⁻¹ K⁻² at room temperature) was also achieved for the Ge_1−x−ySi_xSn_y layers, almost the same as the Si_1−xGe_x ones (maximum: 12 μW cm⁻¹ K⁻² at room temperature) grown with the same conditions. This result opens up the possibility of developing Sn-incorporated group-IV thermoelectric devices.

Ruthenium may replace copper interconnects in next-generation very-large-scale integration (VLSI) circuits. However, interfacial bonding between Ru interconnect wires and surrounding dielectrics must be optimized to reduce thermal boundary resistance (TBR) for thermal management. In this study, various adhesion layers are employed to modify bonding at the Ru/SiO2 interface. The TBRs of film stacks are measured using the frequency-domain thermoreflectance technique. TiN and TaN with high nitrogen contents significantly reduce the TBR of the Ru/SiO2 interface compared to common Ti and Ta adhesion layers. The adhesion layer thickness, on the other hand, has only minor effect on TBR when the thickness is within 2-10 nm. Hard X-ray photoelectron spectroscopy of deeply buried layers and interfaces quantitatively reveals that the decrease in TBR is attributed to the enhanced bonding of interfaces adjacent to the TaN adhesion layer, probably due to the electron transfer between the atoms at two sides of the interface. Simulations by a three-dimensional electrothermal finite element method demonstrate that decreasing the TBR leads to a significantly smaller temperature increase in the Ru interconnects. Our findings highlight the importance of TBR in the thermal management of VLSI circuits and pave the way for Ru interconnects to replace the current Cu-based ones.

<p>Thermal deposition of n-dopant onto SWCNT sheet (p-type) using patterned mask can fabricate p–n patterns with high special resolution. Thermoelectric generator using patterned SWCNT sheets exhibited power density of 60 nW cm⁻² at Δ<italic>T</italic> = 25 °C.</p>

The large anisotropic thermal conduction of a carbon nanotube (CNT) sheet that originates from the in-plane orientation of one-dimensional CNTs is disadvantageous for thermoelectric conversion using the Seebeck effect since the temperature gradient is difficult to maintain in the current flow direction. To control the orientation of the CNTs, polymer particles are introduced as orientation aligners upon sheet formation by vacuum filtration. The thermal conductivities in the in-plane direction decrease as the number of polymer particles in the sheet increases, while that in the through-plane direction increases. Consequently, a greater temperature gradient is observed for the anisotropy-controlled CNT sheet as compared to that detected for the CNT sheet without anisotropy control when a part of the sheet is heated, which results in a higher power density for the planar-type thermoelectric device. These findings are quite useful for the development of flexible and wearable thermoelectric batteries using CNT sheets.

Direct bonding of GaAs and diamond was successfully achieved by surface activated bonding (SAB) method at room temperature. The structures of GaAs/diamond interface before and after annealing at 400 ℃ were investigated by transmission electron microscope (TEM). A 3-nm-thick crystal defect layer was formed at the bonding interface, the change in the crystal defect layer thickness was not observed after annealing. There were no nanovoids and micro-cracks observed at the interface with annealing at temperature 400 ℃. These results indicated that the GaAs/diamond interface has high thermal stability and can withstand the temperature rise of power devices during operating.

GaAs-based power devices have excellent electron transport properties and make it suitable for high frequency operation at high frequency. The output power and the lifetime of GaAs devices are largely degraded by the temperature rise of the active region during operating. The thermal conductivity of GaAs is very small, so that the generated heat by self-heating cannot be sufficiently dissipated through the substrate. Diamond has the highest thermal conductivity among materials and is an ideal material to suppress the temperature rise of the devices. The integration of the GaAs-based devices and diamond will be a more promising approach for improving the heat dissipation ability of the devices. However, since there is a large mismatch between the thermal expansion coefficients and lattice constants of GaAs and diamond, the direct growth of GaAs on diamond is quite difficult and vice versa. We have achieved the direct bonding of diamond and Si at room temperature using surface activated bonding (SAB) method and obtained the excellent thermal stability bonding interface.^1-3 In this study, we examine the structures of the diamond/GaAs bonding interface and effects of thermal annealing on the interfacial structure of the interface by transmission electron microscopy (TEM).

GaAs epitaxial layer grown on GaAs substrate was bonded to diamond by SAB method at room temperature. GaAs epitaxial substrate was composed of a 200 nm thick GaAs and a 100 nm thick InGaP layers grown on GaAs. After bonding, the GaAs substrate and InGaP layer were removed by mechanical polishing and selective wet etching to obtain 200 nm thick GaAs layer bonded to diamond substrates. The structures of the GaAs/diamond bonding interface were investigated by TEM observation. The TEM samples were fabricated by using a focused ion beam (FIB) technique.

The cross-sectional TEM images of the GaAs/diamond bonding interface without and with annealing at 400°C for 5 min are shown in Fig 1(a) and 1(b), respectively. A crystal defect layer with a thickness of about 3 nm was observed in the as-bonded interface. The thickness of the crystal defect layer did not change, but no voids or cracks were observed at the bonding interface after annealing. These results indicate that the bonding interface of diamond and GaAs has an excellent thermal stability, which is extremely qualified for the heat dissipation of the devices.

Acknowledgements This work was partly supported by Hirose International Scholarship Foundation. The fabrication of the TEM samples and part of the TEM observations were respectively performed at The Oarai Center and at the Laboratory of Alpha-Ray Emitters in IMR under the Inter-University Cooperative Research in IMR of Tohoku University (NO. 18M0045 and 19M0037).

References Liang, S. Masuya, M. Kasu, N. Shigekawa, Appl. Phys. Lett.2017, 110, No.111603.

Liang, S. Masuya, S. Kim, T. Oishi, M. Kasu, N. Shigekawa, Appl. Phys. Express2019, 12, No. 016501.

Liang, Y. Zhou, S. Masuya, F. Gucmann, M. Singh, J. Pomeroy, S. Kim, M. Kuball, M. Kasu, N. Shigekawa, Diam. Relat. Mater.2019, 93, 187 – 192.



Figure 1

<p></p>

A phonon transport in Si wire structures were simulated based on a Monte Carlo method to clarify the influence of the wire geometry and the surface roughness on thermal conductivity and the phonon-drag component of Seebeck coefficient. The mean free path (MFP) spectrum was estimated by tracing the simulated phonons. The MFPs of 1 THz phonons which mainly contribute to Seebeck coefficient become shorter with a decrease of the wire width for rough surfaces. This agrees with experimental observation of Seebeck coefficient. The MFPs of 3 THz phonons which mainly contribute to thermal conductivity were influenced even by small-roughness surfaces.

Copyright © 2020 American Chemical Society. In microthermoelectric generators (μTEGs), parasitic thermal resistance must be suppressed to increase the temperature difference across thermocouples for optimum power generation. A thermally conductive (TC) layer is typically used in μTEGs to guide the heat flow from the heat source to the hot junction of each thermocouple. In this study, we investigate the effect of the thermal boundary resistance (TBR) in metal/dielectric TC layers on the power generation of silicon nanowire (SiNW) μTEGs. We prepared various metal/adhesion/dielectric TC layers using different metal, adhesion, and dielectric layers and measured the thermal resistance using the frequency-domain thermoreflectance method. We found that the thermal resistance was significantly different, mainly due to the TBR of the metal/dielectric interfaces. Interface characterization highlights the significant role of the interfacial bonding strength and interdiffusion in TBR. We fabricated a prototype SiNW-μTEG with different TC layers for testing, finding that the power generation increased significantly when the thermal resistance of the TC layer was lowered. This study helps to understand the underlying physics of thermal transport at interfaces and provides a guideline for the design and fabrication of μTEGs to enhance power generation for effective energy harvesting.

Copyright © 2020 American Chemical Society. Temperature increase in the continuously narrowing interconnects accelerates the performance and reliability degradation of very large scale integration (VLSI). Thermal boundary resistance (TBR) between an interconnect metal and dielectric interlayer has been neglected or treated approximately in conventional thermal analyses, resulting in significant uncertainties in performance and reliability. In this study, we investigated the effects of TBR between an interconnect metal and dielectric interlayer on temperature increase of Cu, Co, and Ru interconnects in deeply scaled VLSI. Results indicate that the measured TBR is significantly higher than the values predicted by the diffuse mismatch model and varies widely from 1 × 10-8 to 1 × 10-7 m2 K W-1 depending on the liner/barrier layer used. Finite element method simulations show that such a high TBR can cause a temperature increase of hundreds of degrees in the future VLSI interconnect. Characterization of interface properties shows the significant importance of interdiffusion and adhesion in TBR. For future advanced interconnects, Ru is better than Co for heat dissipation in terms of TBR. This study provides a guideline for the thermal management in deeply scaled VLSI.

We measured the influence of the germanium (Ge) nearest neighbor atom on the lattice vibration of silicon germanium (SiGe) thin films on silicon substrate, which is expected as a material for next-generation electronic and thermoelectric devices, by x-ray absorption fine structure measurement. The amount of changes in the Debye-Waller factor (Δσ2) of each sample were estimated from the obtained extended x-ray absorption fine structure spectra, and it was experimentally clarified that the lattice vibration of the SiGe films were different from that of Ge. On the other hand, it is considered that the lattice vibration of the SiGe films were suppressed by the compressive strain, since Δσ2 have hardly changed with the Ge concentration. The Einstein temperature estimated from Δσ2 decreased with increasing Ge concentration, suggesting that the thermal conductivity of SiGe film can be controlled by Ge concentration.

Thermal conductivity characteristics of Si nanowires (SiNWs) treated with thermal oxidation before and after a subsequent Ar+ ion irradiation process were evaluated by UV Raman spectroscopy, in order to investigate the impact of interfacial oxide-induced lattice disorder. Laser-powerd-ependent Raman spectroscopy showed that the rise in temperature caused by laser heating of SiNWs is suppressed by the Ar+ ion irradiation process. It is considered that this suppression of an increase in temperature is caused by the Ar+ ion irradiation breaking bonds at the SiO2/SiNW interface. These results indicate that not only roughness and defects but also bonding characteristics at SiO2/SiNW interfaces should be carefully considered to achieve a low value of thermal conductivity for next-generation SiNW thermoelectric devices. To realize phonon scattering in SiNWs efficiently, optimization of thermal oxidation is necessary. (c) 2019 The Japan Society of Applied Physics

© 2019 Electrochemical Society Inc.. All rights reserved. We present a new design of silicon-based micro-thermoelectric generator, which utilizes silicon nanowires as the thermoelectric leg. It is driven by a steep temperature gradient exuding around a heat flow perpendicular to the substrate, and the silicon nanowires are not suspended on a cavity etched on the substrate. The power density is scalable by shortening the silicon nanowire to sub-μm length, which was experimentally demonstrated and tens of µW/cm2-class power generation was achieved at an externally applied temperature difference of only 5 K. A numerical discussion shows that the thermoelectric power can be drastically enhanced by suppressing the thermal resistance at the entire substrate. Thus, there is a plenty of room at the micro- or submicrometric scales for realizing thermal energy harvesting devices with high power densities.

© 2018 IEEE. We propose a planar device architecture compatible with the CMOS process technology as the optimal current benchmark of a Si-nanowire (NW) thermoelectric (TE) power generator. The proposed device is driven by a temperature gradient that is formed in the proximity of a perpendicular heat flow to the substrate. Therefore, unlike the conventional TE generators, the planar short Si-NWs need not be suspended on a cavity structure. Under an externally applied temperature difference of 5 K, the recorded TE power density is observed to be 12 μW/cm2 by shortening the Si-NWs length and suppressing the parasitic thermal resistance of the Si substrate. The demonstration paves a pathway to develop cost-effective autonomous internet-of-things applications that utilize the environmental and body heats.

© Published under licence by IOP Publishing Ltd. We report the enhancement of thermoelectric power of Si nanowire micro thermoelectric generator (SiNW-μTEG) by improving the thermal conductivity of aluminum nitride (AlN) thermally conductive film. This work builds on the previous results that have been reported at SSDM 2017. [1] This work shows the measured thermal conductivities of the two different AlN films to understand the underlying mechanisms of the enhancement of the thermoelectric power.

A simple structure of planar silicon nanowire (SiNW)-based thermoelectric (TE) generators (TEGs) is presented in this paper, which has the ability to sustain temperature difference along SiNW under ultrashort channel length for achieving mW/cm2-class power output from environmental heat energy. The TE performance of the proposed SiNW-based TEGs was evaluated by finite-element simulation and analytic modeling. The channel length, pad length, and the thickness of the SiO2 layer were varied in the model while keeping a series of constant proportions toward finding the optimal SiNW-based TEG structure for high power generation. Based on the analysis, we demonstrated that decreasing the dimension of SiNW-based TEG is beneficial to improving the total TE power density. A very high power density of 4.2 mW/cm2 is possible to be achieved at SiNW length of μm and pad length of μm under a temperature difference of 5 K across the hot and cold regions. The miniaturized SiNW-based TEG has a great potential to obtain the output power density which is 100-1000 times higher than conventional planar TEGs with air cavity but has a simple structure and can easily be fabricated.

For harvesting energy from waste heat, the power generation densities and fabrication costs of thermoelectric generators (TEGs) are considered more important than their conversion efficiency because waste heat energy is essentially obtained free of charge. In this study, we propose a miniaturized planar Si-nanowire micro-thermoelectric generator (SiNW-TEG) architecture, which could be simply fabricated using the complementary metal-oxide-semiconductor-compatible process. Compared with the conventional nanowire TEGs, this SiNW-TEG features the use of an exuded thermal field for power generation. Thus, there is no need to etch away the substrate to form suspended SiNWs, which leads to a low fabrication cost and well-protected SiNWs. We experimentally demonstrate that the power generation density of the SiNW-TEGs was enhanced by four orders of magnitude when the SiNWs were shortened from 280 to 8m. Furthermore, we reduced the parasitic thermal resistance, which becomes significant in the shortened SiNW-TEGs, by optimizing the fabrication process of AlN films as a thermally conductive layer. As a result, the power generation density of the SiNW-TEGs was enhanced by an order of magnitude for reactive sputtering as compared to non-reactive sputtering process. A power density of 27.9 nW/cm(2) has been achieved. By measuring the thermal conductivities of the two AlN films, we found that the reduction in the parasitic thermal resistance was caused by an increase in the thermal conductivity of the AlN film and a decrease in the thermal boundary resistance.

We provide the parameters of Stillinger-Weber potentials for GeSiSn ternary mixed systems. These parameters can be used in molecular dynamics (MD) simulations to reproduce phonon properties and thermal conductivities. The phonon dispersion relation is derived from the dynamical structure factor, which is calculated by the space-time Fourier transform of atomic trajectories in an MD simulation. The phonon properties and thermal conductivities of GeSiSn ternary crystals calculated using these parameters mostly reproduced both the findings of previous experiments and earlier calculations made using MD simulations. The atomic composition dependence of these properties in GeSiSn ternary crystals obtained by previous studies (both experimental and theoretical) and the calculated data were almost exactly reproduced by our proposed parameters. Moreover, the results of the MD simulation agree with the previous calculations made using a time-independent phonon Boltzmann transport equation with complicated scattering mechanisms. These scattering mechanisms are very important in complicated nanostructures, as they allow the heat-transfer properties to be more accurately calculated by MD simulations. This work enables us to predict the phonon- and heat-related properties of bulk group IV alloys, especially ternary alloys.

We demonstrate that the nickelidation (nickel silicidation) reaction rate of silicon nanowires (SiNWs) surrounded by a thermally grown silicon dioxide (SiO2) film is enhanced by post-oxidation annealing (POA). The SiNWs are fabricated by electron beam lithography, and some of the SiNWs are subjected to the POA process. The nickelidation reaction rate of the SiNWs is enhanced in the samples subjected to the POA treatment. Ultraviolet Raman spectroscopy measurements reveal that POA enhances compressive strain and lattice disorder in the SiNWs. By considering these experimental results in conjunction with our molecular dynamics simulation analysis, we conclude that the oxide-induced lattice disorder is the dominant origin of the increase in the nickelidation rate in smaller width SiNWs. This study sheds light on the pivotal role of lattice disorders in controlling metallic contact formation in SiNW devices. Published by AIP Publishing.

We have found experimentally an anomalous thermoelectric characteristic of an n-type Si nanowire micro thermoelectric generator (mu TEG). The mu TEG is fabricated on a silicon-on-insulator wafer by electron beam lithography and dry etching, and its surface is covered with a thermally grown silicon dioxide film. The observed thermoelectric current is opposite to what is expected from the Seebeck coefficient of n-type Si. The result is understandable by considering a potential barrier in the nanowire. Upon the application of the temperature gradient across the nanowire, the potential barrier impedes the diffusion of thermally activated majority carriers into the nanowire, and it rather stimulates the injection of thermally generated minority carriers. The most plausible origin of the potential barrier is negative charges trapped at the interface between the Si nanowire and the oxide film. We practically confirmed that the normal Seebeck coefficient of the n-type Si nanowire is recovered after the hydrogen forming gas annealing. This implies that the interface traps are diminished by the hydrogen termination of bonding defects. The present results show the importance of the surface inactivation treatment of mu TEGs to suppress the potential barrier and unfavorable contribution of minority carriers. Published by AIP Publishing.

The evaluation of strain states in silicon nanowires (Si NWs) is important not only for the surrounding gate field-effect transistors but also for the thermoelectric Si NW devices to optimize their electric and thermoelectric performance characteristics. The strain states in Si NWs formed by different oxidation processes were evaluated by UV Raman spectroscopy. We confirmed that a higher tensile strain was induced by the partial presence of a tetraethyl orthosilicate (TEOS) SiO2 layer prior to the thermal oxidation. Furthermore, in order to measure biaxial stress states in Si NWs accurately, we performed water-immersion Raman spectroscopy. It was confirmed that the anisotropic biaxial stresses in the Si NWs along the length and width directions were compressive and tensile states, respectively. The Si NW with a TEOS SiO2 layer on top had a larger strain than the Si NW surrounded only by thermal SiO2. (C) 2017 The Japan Society of Applied Physics

We experimentally investigated the dipole layer formation at Al2O3/AlFxOy (x:y = 1:1 and 1:2.5) interfaces, which would be explicable by considering the anion density difference as the key parameter to determine the dipole direction at the dielectric interface with different anions. Molecular dynamics (MD) simulation of Al2O3/AlF3 demonstrates a preferential migration of O from Al2O3 to AlF3 compared with F to the opposite direction which suggests that anion migration due to the density difference could determine the direction of the dipole layer formed at this interface. In addition, charge separation due to the difference in the anion valences could have certain effect simultaneously. Published by AIP Publishing.

We clarified the mechanism of oxygen (O-)-ion migration at a high-k/SiO2 interface, which is a possible origin of the flat-band voltage shift in metal/high-k gate stacks. The oxygen density difference accommodation model was reproduced by a molecular dynamics simulation of an Al2O3/SiO2 structure, in which O- ions migrate from the higher oxygen density side to the lower one. We determined that the driving force of the O--ion migration is the short-range repulsion between ionic cores. The repulsive force is greater in materials with a higher oxygen density, pushing O- ions to the lower oxygen density side. (C) 2017 The Japan Society of Applied Physics

Nano composite (NC) materials have a great potential to improve characteristics of electrical rotating machinery insulation systems. In order to understand the contribution of the NC materials on the insulation characteristics, a fundamental study which used semi-conductor field technologies is shown in this paper. The thin NC film is formed on a silicon substrate by a spin coating method. The silica particle dispersion in epoxy is analyzed by SEM. Furthermore, the insulation breakdown is measured and the nanoscale breakdown spots are investigated by STM. The results show that a NC which has a potential of actual application is developed and that the nanoscale method seems to be effective for evaluating NC materials.

A new device architecture of micro thermoelectric generator (mu-TEG) is proposed. The mu-TEG utilizes silicon nanowires as the thermoelectric (TE) material, and it can be fabricated by the CMOS-compatible process. It is driven by an "evanescent thermal field" exuding around a heat flow perpendicular to the substrate. We demonstrate experimentally that the TE power increases in the shorter TE leg lengths. The results show that the TE power density is scalable by miniaturizing and integrating the proposed structure.

We show the electric dipole layer formed at a high-k/SiO2 interface can be explained by the imbalance between the migration of oxygen ions and metal cations across the high-k/SiO2 interface. Classical molecular dynamics (MD) simulations are performed for Al2O3/SiO2, MgO/SiO2, and SrO/SiO2 interfaces. The simulations qualitatively reproduce the experimentally observed flatband voltage (V-FB) shifts of these systems. In the case of the Al2O3/SiO2 interface, a dipole layer is formed by the migration of oxygen ions from the Al2O3 side to the SiO2 side. By way of contrast, opposite dipole moments appear at the MgO/SiO2 and SrO/SiO2 interfaces, because of a preferential migration of metal cations from the high-k oxide toward the SiO2 layer in the course of the formation of a stable silicate phase. These results indicate that the migrations of both oxygen ions and metal cations are responsible for the formation of the dipole layer in high-k/SiO2 interfaces. (C) 2016 The Japan Society of Applied Physics

Silicon nanowire Schottky barrier tunnel field effect transistors (NW-SBTFETs) are promising structures for high performance devices. In this study, we fabricated NW-SBTFETs to investigate the effect of nanowire structure on the device characteristics. The NW-SBTFETs were operated with a backgate bias, and the experimental results demonstrate that the ON current density is enhanced by narrowing the width of the nanowire. We confirmed using the Fowler-Nordheim plot that the drain current in the ON state mainly comprises the quantum tunneling component through the Schottky barrier. Comparison with a technology computer aided design (TCAD) simulation revealed that the enhancement is attributed to the electric field concentration at the corners of cross-section of the NW. The study findings suggest an effective approach to securing the ON current by Schottky barrier width modulation. (C) 2016 The Japan Society of Applied Physics

We have newly developed the interatomic potential of Si, Ge or Ge, Sn mixed systems to reproduce the lattice constant, phonon frequency, and phonon dispersion relations in the bulk pure group IV crystal and group IV alloys by molecular dynamics (MD) simulation. The phonon dispersion relation is derived from the dynamical structure factor which is calculated by the space-time Fourier transform of atomic trajectories in MD simulation. The newly designed potential parameter set reproduces the experimental data of lattice constant and phonon frequency in Si, Ge, Sn, and SiGe. Furthermore, the Sn concentration dependence of the phonon frequency, which are not yet clarified, is calculated with three type assumptions of lattice constant in GeSn alloy. This work enables us to predict the elastic and phonon related properties of bulk group IV alloys.

The presence of a SiO2 layer on Si nanowires (SiNWs) has been found through molecular dynamics simulation to reduce their thermal conductivity (kappa), with kappa approaching the amorphous limit of Si as the oxide layer thickness is increased. Through analysis of the phonon energy dispersion and vibrational density of states (VDOS) spectrum, this decrease in kappa was attributed to dispersionless vibrational states that appear in the low energy range below 4 THz as a result of the lattice vibration of Si atoms near the SiO2/Si interface. The SiO2 layer also induced a low-frequency tail in the VDOS spectrum, the length of which was more closely correlated to the reduction in kappa than the frequency-integrated value of the VDOS spectrum. These findings provide a more refined explanation for the decrease in kappa than has been previously observed, and contribute to providing a greater understanding of the anomalistic vibration near the interface that is critical to determining the heat conductivity in nanoscale materials.

We demonstrate that the parallel computing with graphic processing unit (GPU) effectively accelerates a particle-based carrier transport simulation called EMC/MD method. The simulation speed is increased by parallelizing the point-to-point Coulomb's force calculation, which is sufficient to accomplish a device characteristic simulation of nanostructured metal-oxide-semiconductor field effect transistor (MOSFET) including source and drain diffusion regions. The EMC/MD simulation powered by GPU computing is a useful tool to investigate the statistical variability analysis of nano-scale transistors.

We show that electric dipole layer at Al2O3/SiO2 interface is reproduced by classical molecular dynamics simulation with a simple two-body rigid ion model. The dipole layer was spontaneously formed by the migration of oxygen ions from Al2O3 side to SiO2 side. Built-in potential at the Al2O3/SiO2 is estimated to about 0.35 V, which roughly compares with the experimental value of the flat band voltage shift. Contrary, no significant dipole layer appeared at Y2O3/SiO2 interface. The simulation results are explained in terms of the difference in the magnitude of multipole moments around cations of these oxides. (C) 2014 the Japan Society of Applied Physics

We have found that the thermal history of the fabrication process of Si nanowires (NWs) has a strong impact on the Ni silicidation rate. We compared the Ni silicidatioh rates in Si NWs fabricated by two different types of processes: the "Doping First" process, in which dopant activation annealing is completed before the lithography of NW structures, and the "Patterning First" process, in which NWs are firstly fabricated and then subjected to heat treatment entailing thermal oxidation and dopant activation. The Ni silicidation rate was appreciably higher in the Doping First process than in the Patterning First process. The difference is attributed to the residual stress rather than to the dopant concentration in Si-NWs. To control the silicidation rate in NWs, particular attention to the thermal history is necessary. (C) 2014 The Japan Society of Applied Physics

A new type of silicon-based Tunneling FET (TFET) using semiconducting silicide Mg2Si/Si hetero-junction as source-channel structure is proposed and the device simulation has been presented. With narrow bandgap of suicide and the conduction and valence band discontinuous at the hetero-junction, larger drain current and smaller subthreshold swing than those of Si homo-junction TFET can be obtained. Structural optimization study reveals that low Si channel impurity concentration and the alignment of the gate electrode edge to the hetero-junction lead to better performance of the TFET. Scaling of the gate length increases the off-state leakage current, however, the drain voltage (V-d) reduction in accordance with the gate scaling suppresses the phenomenon, keeping its high drivability. (C) 2014 Published by Elsevier Ltd.

The phonon dispersion relation in (100) Si nanowire ( SiNW) is calculated by employing a realistic atomistic model surrounded by thin SiO2 layer. We performed molecular dynamics simulations to calculate the dynamical structure factor by the space-time Fourier transform of atomic trajectories, and extracted the phonon dispersion relations. In the SiNWs, low energy phonon branches spread into broad spectra due to the presence of the SiO2 film, which is considered as the origin of the thermal conductivity degradation. A softening of the transverse optical mode also appears due to the lattice strain induced by the outer oxide film. This work suggests that the presence of amorphous oxide layer is crucial factor to characterize phonon vibration properties in practical SiNWs. (C) 2014 The Electrochemical Society.

Electric dipole layer formation at high-k/SiO2 interface is reproduced by classical molecular dynamics simulation based on a simple two-body rigid ion model (1). The dipole layer was spontaneously formed by the migration of oxygen ions across the high-k/SiO2 interface. In the case of Al2O3/SiO2, a part of oxygen ions of Al2O3 penetrated into the SiO2 side, resulting in the formation of a built-in potential of about 0.5 V. The opposite migration of oxygen ions, from SiO2 side to high-k oxide side, is also reproduced by using different potential parameters of ionic radius and effective charge. The simulation result suggests that the dipole is not merely formed by the oxygen density difference. Rather, oxygen ions are driven by some interatomic forces at the interface. We discuss the origin of the driving force of the oxygen migration in terms of the multipole moments around cations in the oxides.

We have realized the full-scale whole device EMC/MD simulation including source and drain regions by utilizing graphic processing unit. The transfer characteristic of a gate-all-around nanowire Si MOSFET is simulated by reproducing the field effect of the surrounding gate electrode with spreading charged particles on the gate insulator layer. We have found an appreciable impact of the random dopant distribution (RDF) in source and drain regions on the drain current variability. Furthermore, the dynamic fluctuation of the drain current is found to be increase as the channel length decreases. The EMC/MD simulation powered by GPU is a useful method to investigate the dynamic fluctuation as well as the statistical device-to-device variability of nano-scale FETs.

We demonstrate that the image force effects in low-dimensional Si are highly controllable to achieve the best possible performance of the gate-all-around (GAA) Schottky barrier tunneling FET (SB-TFET). Our finite element electrostatic calculation shows that the image potential lowers near the metal source/drain, whereas it rises in the proximity of the gate insulator. Moreover, the drain induced barrier lowering (DIBL) of GAA-SB-TFET is suppressed by the image forces in a thin Si nanowire of about 4.0nm diameter.

We numerically demonstrate that a random dopant fluctuation (RDF) in a source region causes a noticeable variability in the on-current of Si nanowire (NW) transistors, and its effect is much larger than that of a random telegraph noise (RTN). This work assesses the static and dynamic variability of NW device characteristics using the ensemble Monte Carlo/molecular dynamics (EMC/MD) simulation, which employs parallel computing technique using a graphic processing unit (GPU). The current flow in a one-dimensional NW device is determined by the number of dopants at the source edge, indicating the importance of forming an abrupt source-channel boundary to suppress the variability.

We demonstrate that the image force effects in low-dimensional Si are highly controllable to achieve the best possible performance of the gate-all-around (GAA) Schottky barrier tunneling FET (SB-TFET). Our finite element electrostatic calculation shows that the image potential lowers near the metal source/drain, whereas it rises in the proximity of the gate insulator. Moreover, the drain induced barrier lowering (DIBL) of GAA-SB-TFET is suppressed by the image forces in a thin Si nanowire of about 4.0nm diameter.

The influence of structural parameters, including the Schottky barrier height for electron (phi(Bn)) and channel doping (N-a), on the electrical characteristics of a scaled Schottky barrier tunneling FET (SBTFET) have been clarified by numerical device simulation. The thermionic emission current (I-TH) as well as the tunneling current (I-TN) have been considered as the main electron injections at the source edge. Simulation results have revealed that the main conduction is I-TN in the region near and above the threshold voltage (V-th). As tunneling probability is determined by phi(Bn) and the width of the triangular potential barrier at the source edge, a lower phi(Bn) with higher N-a results in a better subthreshold swing (SS) with high on-state drive current (I-ON) at a gate length (L-g) of 50 nm. With L-g scaling down to 10 nm, however, a lower phi(Bn) has shown an increased off-state leakage current (I-OFF) due to the short-channel effect (SCE), while a larger phi(Bn) can suppress the I-OFF at the cost of I-ON. Therefore, considering SS with I-ON and I-OFF ratio, it can be concluded that an optimum phi(Bn) exists for short-channel devices. The SBTFET showed good subthreshold performance and higher I-ON/I-OFF than the conventional silicon-on-insulator (SOI) MOSFET in 10 nm region with the Schottky barrier height optimization. (C) 2013 The Japan Society of Applied Physics

Impacts of atomic disorder on avalanche multiplication in a one-dimensional silicon nanodot (SiND) array have been theoretically studied. The disorder lifts the degeneracy of the energy levels and reduces the impact-ionization threshold. This leads to a larger carrier multiplication factor in the disordered SiND array compared to an ideal SiND array without disorder or strain. (C) 2013 The Japan Society of Applied Physics

We theoretically investigate effects of atomic disorder existing near the Si/SiO2 interfaces on the impact ionization rate of a Si nanodot (SiND). We find that the impact ionization rate of a disordered SiND becomes higher near the threshold energy and approaches that of an ideal SiND for higher energy region. © 2013 AIP Publishing LLC.

Fluctuations of device characteristics due to random discrete dopant (RDD) distribution are numerically investigated in ultra-small Si nanowire transistors. Kinetic Monte Carlo process simulation is performed to obtain realistic RDD distributions, whose effects on the transport characteristics are then analyzed by using a non-equilibrium Green's function (NEGF) method. Fluctuations due to atomic disorder near the Si/SiO2 interface are also investigated by performing molecular dynamics oxidation simulation for realistic atomic structure models and NEGF device simulation for transport characteristics.

The phonon dispersion relation in &lt; 100 &gt; Si nanowire (SiNW) is calculated by employing a realistic atomistic model surrounded by thin SiO2 layers. We performed molecular dynamics simulation to calculate the dynamical structure factor by the space-time Fourier transform of atomic trajectories, and extracted the phonon dispersion relations. Although the bulk dispersion relations are maintained in the SiNWs on the whole, acoustic phonon branches are diffused beyond recognition, which is considered as the origin of the thermal conductivity degradation in SiNWs. A red shift of the transverse optical mode also appears probably due to the lattice strain induced by the outer oxide film. These results provide a foothold to estimate the electron-phonon scattering rates and the heat transport processes in realistic SiNWs.

We propose a new strategy for in-situ observation of dopant activation process using scanning tunnelling microscopy (STM) technology. Two factor techniques are established within our original STM system combined with liquid-metal-ion-source ion-gun (LMIS-IG/STM): (1) in-situ real-time observation of Si(001)-2x1 surface modified with dopant ions, and (2) in-situ hydrogen passivation of Si(001)-2x1 surface. Sequential STM images of the surface morphological changes are successfully obtained with keeping the original observation area. The hydrogenated Si(001)-2x1 surface is stable enough even under the degraded vacuum conditions which are unavoidable in ion irradiation processes.

This article reports the results of real-time scanning tunneling microscopy (STM) observation of Au⁺ ion irradiation effects on high-temperature Si surface, which was achieved by our original ion gun and STM combined system. Sequential STM images of a Si(111)-7×7 surface kept at 500^oC were obtained before, during, and after Au⁺ ion irradiation with 3 keV. Vacancy islands, which are two-dimensional clusters of surface vacancies, and 5×2-Au structures were formed on the surface and their size were changed during the subsequent thermal treatment. This method enables us to count exact numbers of vacancies and Au atoms on the surface by measuring the sizes of vacancy islands and 5×2-Au reconstructions. The timescale of the growth of the 5×2-Au domain suggests that the implanted Au atoms diffuse to the surface almost without interacting with point defects induced by the ion irradiation.

The impact of current fluctuation due to discreteness in carrier numbers on high-frequency noise amplitudes is numerically investigated, focusing on the comparison to the impact of a single trapped charge in the oxide layer for gate-all-around nanowire structures. The variation in the amount of the charge transporting through the channel within a single clock cycle is estimated. The transported charge variation due to the current fluctuation clearly shows the universality with respect to the total amount of the transported charge. It concludes that the current fluctuation becomes a dominant noise source over 100 GHz range.

A framework of a new reactive molecular force field is proposed. It is designed within the framework of an extended classical mechanical system that describes not only the motion of atomic nuclei but also the motion of additional degrees of freedom, which determine bond orders among atoms. The bond order determination is clearly distinguished in the potential energy formulation, and the parametrization in the new force field can be performed in the same way as that in non-reactive force fields. The new reactive force field is highly transferable to various multicomponent materials. In this article, two specific applications are described: (1) modeling a SiO2/Si interface and (2) the molecular dynamics simulation of the proton transfer reaction in water.

Effects of oxidation-process-induced atomic disorder on extended electronic states in the channel region of narrow Si nanowire (NW) field-effect- transistors (FETs) are theoretically investigated by using the molecular dynamics, empirical tight-binding, and non-equilibrium Green's function methods. Simulation results show that the injection velocity in n-type Si NW FETs is less affected by the disorder compared to p-type devices, which can be attributed to differences in the in-plane carrier profile. © 2011 IEEE.

Effect of the channel shape on the nano-scale carrier transport is studied by using the ensemble Monte-Carlo - molecular dynamics method (EMC/MD). Carrier transport in hone-shaped asymmetric channels which widen from source to drain sides is simulated by comparing that in the conventional straight channels. The obtained conductance of the horn-shaped channels is larger than that of the straight channel, as a result of the enhancement of the carrier mobility in the hone-shaped channel. This can be attributed to two reasons: the collimation effect of the asymmetric channel peculiar in the quasi-ballistic carrier transport regime, and the suppression of carrier-carrier interaction due to widening of the channel. © 2011 IEEE.

Hot phonon generation and its impact on the current conduction in a nanoscale Si-device are investigated using a Monte Carlo simulation technique. In the quasi-ballistic transport regime, electrons injected from the source lose their energies mainly by emitting optical phonons in the drain. Due to the slow group velocity of the optical phonons, the efficiency of the heat dissipation is so poor that a region with a nonequilibrium phonon distribution, i.e., a hot spot, is created. In this study, we have implemented the hot phonon effect in an ensemble Monte Carlo simulator for the electron transport, and carried out the steady state simulations. Although it is confirmed that the optical phonon temperature in the hot spot is larger than that of acoustic phonons by &gt; 100 K, the electron current density is not significantly affected. The local heating would degrade the hot electron cooling efficiency and the parasitic resistance in the drain, but they have a minor impact on the quasi-ballistic electron transport from the source to the drain.

A series of molecular dynamics (MD) simulations have been conducted to investigate the heat transport in terms of the phonon dynamics in nanoscale silicon (Si). This work is motivated by a concern over the stagnation of heat at the drain region of nanoscopic transistors, owing to this, a large amount of optical phonons with a low group velocity are emitted from hot electrons, which are ballistically transferred through channel region. The point of this work is the explicit inclusion of the SiO2 film in the MD simulation of the Si lattice. The calculation results show that longitudinal optical (LO) phonons decay faster as Si lattice thickness decreases and turn into acoustic phonons. In contrast, thermal diffusion rate decreases with Si lattice thickness. Both the decay rate of LO phonons and thermal diffusion rate are not governed by oxide thickness. These results imply that the phonon scattering at the SiO2/Si interface is enhanced by thinning the Si layer. In nanoscopic devices, a thin Si layer is effective in diminishing the optical phonons with a low group velocity, but it hinders the subsequent heat transport. (c) 2011 The Japan Society of Applied Physics

A large-scale molecular dynamics simulation was carried out in order to investigate the adsorption mechanism of ribosomal protein L2 (RPL2) onto a silica surface at various pH values, RPL2 is a constituent protein of the 50S large ribosomal subunit, and a recent experimental report showed that it adsorbs strongly to silica surfaces and that it can be used to immobilize proteins on silica surfaces. The simulation results show that RPL2, especially domains I (residues 1-60) and 3 (residues 203-273), adsorbed more tightly to the silica surface above 7. We found that a major driving force for the adsorption of RPL2 onto the silica surface is the electrostatic interaction and that the structural flexibility of domains 1 and 3 may further contribute to the high affinity.

Two numerical simulation techniques are presented to investigate the heating issues in nanoscale Si devices. The first one is the Monte Carlo simulation for both electron and phonon transport, and the transient electrothermal analysis is carrier out in n+-n-n+ device with the n-layer length of 10 nm. The second is the molecular dynamics approach for simulating the atomic thermal vibration in the nanoscale Si/SiO2 systems. It is shown that the lattice temperature at the drain edge is raised by the hot electron injection from the source after turning on the device, and the impact of this phenomenon becomes more significant in the smaller devices due to the worse heat conductivity. ©2010 IEEE.

A series of molecular dynamics (MD) simulations is conducted to investigate the dynamics of longitudinal optical (LO) phonon in Si nano-structure confined with oxide films. This work is motivated by heat issues in nanoscopic devices; it is considered that the LO phonons with low group velocity are accumulated in the nanoscopic device and the electric property deteriorates. We estimate the relaxation time of the LO phonon and investigate its dependency on the oxide thickness. The calculation results show that the LO phonon decays faster as the oxide thickness increases and turns into acoustic phonon. The result indicates that the presence of SiO2 films promotes the scattering of the phonon and this is effective to diminish the optical phonon.

The stress relaxation mechanism at the GeO2/Ge interface is studied by means of classical molecular simulation employing empirical force-fields. In general, the chemistry in GeO2 is characterized by weaker bonds and softer bond angles than that in SiO2, which has been considered to lead to the relaxation of the GeO2 film on Ge substrate. However, Ge-O-Ge angle is stiffer than Si-O-Si angle, and has a narrower equilibrium angle of 133 degrees than that of Si-O-Si of 144 degrees. The present simulation results show that the narrow Ge-O-Ge bond contribute the reduction of the compressive stress in the GeO2 films. If the Ge-O-Ge bond angle had the same equilibrium angle with Si-O-Si angle, a higher residual stress would remain in the GeO2 films.

Using a coupled Monte Carlo method for solving both electron and phonon Boltzmann transport equations, the transient electrothermal behaviors of nanoscale Si n-i-n device are simulated. The nonequilibrium optical phonon distribution is characterized by a temperature different from that of the acoustic phonons, and these two temperatures show different characteristics not only in the steady state, but also in transient conditions. It has been also suggested that the simulated transient response of the phonon temperatures can be practically described by the equivalent thermal circuit model, which is useful for, e. g., projecting the NBTI lifetime during the realistic circuit operations.

The real-time scanning tunneling microscopy (STM) observation of Au+ ion irradiation effects on a high-temperature Si surface has been performed using our original ion gun and STM combined system. Sequential STM images of a Si(111)-7 x 7 surface kept at 500 degrees C have been obtained before, during, and after Au+ ion irradiation with 3 keV. Vacancy islands, which are two-dimensional clusters of surface vacancies, and 5 x 2-Au structures were formed on the sample surface, and their size were changed during the heat treatment after the ion irradiation. This method enables us to count exact numbers of vacancies and Au atoms on the surface by measuring the sizes of vacancy islands and 5 x 2-Au reconstructions. The timescale of the growth of the 5 x 2-Au domain suggests that the implanted Au atoms diffuse to the surface almost without interacting with point defects induced by the ion irradiation. (C) 2010 The Japan Society of Applied Physics

We perform a series of molecular dynamics (MD) simulations to investigate the heat transport in Si nano-structures, while explicitly including oxide cover layers in the simulation system for the first time. The dependences of thermal diffusion velocity on the thicknesses of the SiO2 film and Si lattice are investigated. The results show that thermal diffusion velocity decreases with Si lattice thickness and does not depend on SiO2 film thickness. (C) 2010 The Japan Society of Applied Physics

Enhancement of transconductance for strained nanowire transistors (s-nwFETs) on (001) and (110) planes are demonstrated by evaluating I-ds-V-bg curves of the devices. Normalized transconductance, g(m)*, for &lt; 100 &gt; direction s-nwFETs is enhanced by a factor of 2.16 for n-type on (001) plane and 1.83 for p-type on (110) plane. This is due to the lighter effective mass of electrons/holes along the selected channel direction.

Strained Si nanowire FETs of nanowire width (W) of W=155nm and W=5000nm, are evaluated by I-d-V-bg characteristics at various temperatures. Transconductance (g(m)) and subthreshold slope are obtained from the I-d-V-bg characteristics. The normalized g(m) (g(m)*) increases by a factor of 1.38 for W=155nm and 3.13 for W=5000nm. Subthreshold slope decreases 22% for W=5000nm and 42% for W=155nm. This improvement is due to suppression of electron-phonon scattering at low temperature. This also indicates that the influence of electron-phonon interaction on g(m) enhancement is different compared to that in bulk Si.

Potential energy distribution of interstitial O2 molecule in the vicinity of SiO2/Si(001) interface is investigated by means of classical molecular simulation. A 4-nm-thick SiO2 film model is built by oxidizing a Si(001) substrate, and the potential energy of an O2 molecule is calculated at Cartesian grid points with an interval of 0.05 nm in the SiO2 film region. The result shows that the potential energy of the interstitial site gradually rises with approaching the interface. The potential gradient is localized in the region within about 1 nm from the interface, which coincides with the experimental thickness of the interfacial strained layer. The potential energy is increased by about 0.62 eV at the SiO2/Si interface. The result agrees with a recently proposed kinetic model for dry oxidation of silicon [Phys. Rev. Lett. 96, 196102 (2006)], which argues that the oxidation rate is fully limited by the oxidant diffusion. © 2008 The American Physical Society.

The effect of siloxane bonding topology on in-plane X-ray diffraction (XRD) profiles of octadecylsilane self-assembled monolayers (SAMs) on SiO2 was investigated by large-scale atomistic simulation. Equilibrium structures of the octadecylsilane SAM/SiO2 systems were obtained by Metropolis Monte Carlo simulations, sampling SAM/SiO2 structures with different interfacial bonding topologies. Lateral ordering of the octadecylsilane molecules was evaluated by calculating in-plane XRD profiles from the structural models. Analyses of the diffraction profiles revealed that molecules retain a hexagonal order at thermal equilibrium and that the structural order decreases as the number of siloxane bonds increases at the SAM/SiO2 interface. (c) 2008 The Japan Society of Applied Physics

Potential energy distribution of interstitial O-2 molecule in the vicinity of SiO2/Si(001) interface is investigated by means of classical molecular simulation. A 4-nm-thick SiO2 film model is built by oxidizing a Si(001) substrate, and the potential energy of an O-2 molecule is calculated at Cartesian grid points with an interval of 0.05 nm in the SiO2 film region. The result shows that the potential energy of the interstitial site gradually rises with approaching the interface. The potential gradient is localized in the region within about 1 nm from the interface, which coincides with the experimental thickness of the interfacial strained layer. The potential energy is increased by about 0.62 eV at the SiO2/Si interface. The result agrees with a recently proposed kinetic model for dry oxidation of silicon [Phys. Rev. Lett. 96, 196102 (2006)], which argues that the oxidation rate is fully limited by the oxidant diffusion.

An ion beam alignment system has been developed in order to realize real-time scanning tunneling microscope (STM) observation of "dopant-ion" irradiation that has been difficult due to the low emission intensity of the liquid-metal-ion-source (LMIS) containing dopant atoms. The alignment system is installed in our original ion gun and STM combined system (IG/STM) which is used for in situ STM observation during ion irradiation. By using an absorbed electron image unit and a dummy sample, ion beam alignment operation is drastically simplified and accurized. We demonstrate that sequential STM images during phosphorus-ion irradiation are successfully obtained for sample surfaces of Si(111)-7 x 7 at room temperature and a high temperature of 500 degrees C. The LMIS-IG/STM equipped with the developed ion beam alignment system would be a powerful tool for microscopic investigation of the dynamic processes of ion irradiation. (C) 2008 American Institute of Physics.

The bonding network of an alkylsilane self-assembled monolayer (SAM)/SiO(2) substrate interface is investigated by means of canonical Monte Carlo (MC) simulations. SAM/SiO(2) systems with different interfacial bonding topologies are sampled by the Metropolis MC method, and the AMBER potential with a newly developed organosilicon parameters are used to obtain an optimized structure with a given bonding topology. The underlying substrates are modeled as hydroxy-terminated (100) or (111) cristobalites. The SAM/SiO(2) interface is characterized by a polysiloxane bonding network which comprises anchoring bonds and cross-linking bonds, namely, molecule-substrate and molecule-molecule Si-O-Si bonds, respectively. We show that at thermal equilibrium, the ratio of the number of anchoring bonds to cross-linking bonds decreases as a total Si-O-Si bond density increases, and that nevertheless, number of anchoring bonds always dominate over that of cross-linking bonds. Moreover we show that the total Si-O-Si bond density strongly affects the lateral ordering of the alkylsilane molecules, and that increase in the Si-O-Si bond density disorders the molecular packing. Our results imply that a lab-to-lab variation in the experimentally prepared SAMs can be attributed to different Si-O-Si bond densities at the SAM/SiO(2) interface. (C) 2008 American Institute of Physics.

The interactions between the luciferase surface area around the active site and the Si surface were estimated by molecular dynamics simulations. The results show that the luciferase surface area around the active site is more unstable than the luciferase surface area with relatively high hydrophobicity when luciferase adsorbs on the Si surface. The main factor of the decrease in activity of luciferase on the Si surface would be adsorption-induced structural changes in active site. Si surface should be designed focusing on structural changes in active site. (c) 2008 Elsevier B. V. All rights reserved.

Transconductance (g(m)) enhancement in n-type and p-type nanowire field-effect-transistors (nwFETs) is demonstrated by introducing controlled tensile strain into channel regions by pattern dependant oxidation (PADOX). Values of g(m) are enhanced relative to control devices by a factor of 1.5 in p-nwFETs and 3.0 in n-nwFETs. Strain distributions calculated by a three-dimensional molecular dynamics simulation reveal predominantly horizontal tensile stress in the nwFET channels. The Raman spectra features in the strain controlled devices display an increase in the full width half maximum, and a shift to lower wavenumber confirming that g. enhancement is due to tensile stress introduced by the PADOX approach.

Transconductance (gm) enhancement in n-type and p-type nanowire field-effect-transistors (nwFETs) is demonstrated by introducing controlled tensile strain into channel regions by pattern dependant oxidation (PADOX). Values of gm are enhanced relative to control devices by a factor of 1.5 in p-nwFETs and 3.0 in n-nwFETs. Strain distributions calculated by a three-dimensional molecular dynamics simulation reveal predominantly horizontal tensile stress in the nwFET channels. The Raman lines in the strain controlled devices display an increase in the full width half maximum, and a shift to lower wavenumber confirming that gm enhancement is due to tensile stress introduced by the PADOX approach © The Electrochemical Society.

Electron transport in bulk silicon with ordered dopant arrays is studied using an ensemble Monte Carlo (EMC) technique coupled with molecular dynamics (MD) method. This work is motivated by a recently developed single-ion implantation (SII) technique, which enables us to fabricate a semiconductor device with an ordered dopant array. We numerically estimate the carrier mobility in silicon with such an ordered dopant array comparing to that with conventional random dopant distribution. The calculation results show that electron mobility can be enhanced in the ordered dopant array if the fluctuation of dopant position is less than 5 nm.

Transconductance (g(m)) enhancement in n-type and p-type nanowire field-effect-transistors (nwFETs) is demonstrated by introducing controlled tensile strain into channel regions by pattern dependent oxidation (PADOX). Values of g(m) are enhanced relative to control devices by a factor of 1.5 in p-nwFETs and 3.0 in n-nwFETs. Strain distributions calculated by a three-dimensional molecular dynamics simulation reveal predominantly horizontal tensile stress in the nwFET channels. The Raman lines in the strain controlled devices display an increase in the full width at half maximum and a shift to lower wavenumber, confirming that g(m) enhancement is due to tensile stress introduced by the PADOX approach. (c) 2007 American Institute of Physics.

We have performed three-dimensional molecular dynamics simulations to investigate strain and stress distributions in silicon nanostructures covered with thermal oxide films, by using our original molecular force field for Si, O mixed systems. We have modeled a wire-shaped nanostructure by carving a Si(001) substrate, and then an oxide film with a uniform thickness was formed by inserting oxygen atom into Si-Si bonds from the surface. The simulation results show that a compressive stress is concentrated on the oxide region in the vicinity of the side SiO2/Si interface of the nanowire. At the top interface, there is also a compressive stress in the [110] direction, whereas the [001] component of the normal stress tensor is almost relaxed. These results suggest that the oxidation is strongly suppressed at the side faces of the silicon nanowire.

The binding energies between luciferin and luciferase adsorbed on a Si surface were estimated by docking simulations. These binding energies were more stable on a hydrophilic Si surface than on a hydrophobic Si surface, but their difference was small. Luciferase adsorbed more strongly on the hydrophobic Si surface than on the hydrophilic Si surface. The luciferase active site was not covered by the Si surface in this adsorption state. The hydrophobic surface would be suitable for the immobilization of luciferase both from the viewpoints of the binding of luciferin to luciferase and the adsorption of luciferase to the Si surface. (C) 2007 Elsevier B.V. All rights reserved.

A newly formulated kinetic theory for thermal oxidation of silicon is reviewed. The new theory does not involve the rate-limiting step of the interfacial oxidation reaction, instead it is supposed that the diffusivity is suppressed in a strained oxide region near the SiO2/Si interface. The expression of the parabolic constant is the same as that of the Deal-Grove model, while the linear constant makes a clear distinction with the model. The estimated thickness using the new expression is close to 1 nm, which compares well with the thickness of the structural transition layer. The origin of the deviation from the linear-parabolic relationship observed at initial oxidation stages can be explained by the enhanced diffusion hypothesis, which is opposite conclusion to the Deal-Grove theory. © The Electrochemical Society.

A formulated kinetic theory for thermal oxidation of silicon is presented in detail. The theory does not involve the rate-limiting step of the interfacial oxidation reaction, instead it is supposed that the diffusivity is suppressed in a strained oxide region near the SiO2/Si interface. The expression of the parabolic constant is the same as that of the Deal - Grove model, while the linear constant makes a clear distinction with the model. The estimated thickness using the expression is close to 1 nm, which compares well with the thickness of the structural transition layer. The origin of the deviation from the linear-parabolic relationship observed at initial oxidation stages can be explained by the enhanced diffusion hypothesis, which is the opposite conclusion to the Deal - Grove theory. (c) 2007 The Electrochemical Society.

We propose a new oxidation rate equation for silicon supposing only a diffusion of oxidizing species but not including any rate-limiting step by interfacial reaction. It is supposed that diffusivity is suppressed in a strained oxide region near the SiO2/Si interface. The expression of a parabolic constant in the new equation is the same as that of the Deal-Grove model, while a linear constant makes a clear distinction with that of the model. The estimated thickness using the new expression is close to 1 nm, which compares well with the thickness of the structural transition layers.

Molecular mechanics (MM) and molecular dynamics (MID) simulations have been performed to investigate the two-dimensional structure of organosilane self-assembled monolavers (SAMs). Unlike alkanethiol SAMs, the arrangement of molecules in organosilane SAMs is not crystalline, and their molecular structure yet remains undetermined. AMBER 8 is employed with our newly developed Si parameters for the MM/MD simulations. Simulations performed for structures with different bonding networks in the polysiloxane layer shows that the ratio of hydrogen bonds has a profound effect on conformations and strain energies of optimized structures. Our results suggest that alkylsilane SAMs formed on substrates are not perfectly uniform but may have some defects.

A series of molecular dynamics (MD) simulations have been performed to investigate the interactions between luciferase and Si substrates. The results show that luciferase adsorbs directly on the hydrophobic Si substrate, and via water molecules on the hydrophilic one. The adsorption-induced changes in conformation of luciferase are smaller on the hydrophilic Si substrate than oil the hydrophobic one. The dynamic atom motions in luciferase are larger on the hydrophilic Si substrate than on the hydrophobic one. Inside the active site, the adsorption-induced changes in distances between the atoms forming hydrogen bonds to substrate luciferin are smaller on the hydrophilic Si Substrate than the hydrophobic one. In order to prevent the denaturation of luciferase caused by immobilization, the solid surface should be hydrophilic. For higher thermostability, after immobilization, however, a hydrophobic surface is preferable since the dynamic atom motions in luciferase are smaller on a hydrophobic surface. The solid surface should be prepared delicately both from the viewpoint of preventing the denaturation caused by immobilization and improving the thermostability.

We propose a new oxidation rate equation for silicon supposing only a diffusion of oxidizing species but not including any rate-limiting step by interfacial reaction. It is supposed that diffusivity is suppressed in a strained oxide region near the SiO2 Si interface. The expression of a parabolic constant in the new equation is the same as that of the Deal-Grove model, while a linear constant makes a clear distinction with that of the model. The estimated thickness using the new expression is close to 1 nm, which compares well with the thickness of the structural transition layers. © 2006 The American Physical Society.

We have performed a series of molecular dynamics (MD) simulations on interactions between green fluorescent protein (GFP) and Si substrates. The results show that GFP adsorbs directly on the hydrophobic substrate, and via water molecules on the hydrophilic substrate. The adsorption-induced changes in the conformation of GFP are smaller on the hydrophilic substrate than on the hydrophobic substrate. On the other hand, the dynamic atom motions in GFP are larger on the hydrophobic substrate than on the hydrophilic substrate. In order to prevent the denaturation of proteins caused by immobilization on a substrate, the Si surface should be prepared from the viewpoints of both conformation and dynamic atom motions.

To address the reactions and diffusion of atomic and molecular oxygen in SiO2, the modification of the SiO2 network on exposure to an atomic or molecular oxygen atmosphere is investigated by measuring the x-ray-diffraction profile of the residual order peak emanating from the oxide. Analyses of the peak intensity and its fringe pattern provide experimental evidence for the recent theoretical predictions, indicating that atomic oxygen is incorporated into the SiO2 network near the surface and diffuses toward the interface along with modifying it even at a low temperature of 400 degrees C, whereas molecular oxygen diffuses without reacting with the bulk SiO2 even at a temperature of 850 degrees C that is sufficiently high for oxidation reaction at the interface.

Using a low-energy ion gun/high-temperature scanning tunneling microscope combined system (lG/STM), we observed the atomic scale behavior of a Si surface kept at 500 degrees C before, during and after Ar ion irradiation. We found that Si islands grew up within ion irradiation selectively along the boundaries of domains of 7 x 7 dimer-adatom-stacking (DAS) faults. The Si islands were 1 double atomic layer high and had the 7 x 7 DAS reconstruction. The area of the islands increased linearly with ion doses up to 2.5 x 10(14) cm(-2).

We investigated the atomic structure of the SiO(2)/Si interface and the initial oxidation process of Si surfaces using our developed large-scale atomistic simulation technique for Si, O mixed systems. We constructed large-scale SiO(2)/Si(0 0 1) interface models (now up to 12,536 atoms in size) by inserting O atoms into Si-Si bonds in crystalline Si substrates from the surface of the models. The resulting SiO(2)/Si Models exhibited a compressively strained oxide region near the interface, and reproduced X-ray diffraction peaks compatible with experimental results. Using the large-scale modeling technique, we simulated an atomistic oxidation process where the O atoms were introduced into the Si substrate in one by one so as to minimize the strain energy caused by the insertion of the O atoms. A mostly abrupt change in the composition at the SiO(2)/Si interface was reproduced in this energetic scheme, though the oxidation did not proceed layer by layer as previously reported by many other reports. We found out that the layer-by-layer oxidation phenomenon can be explained by the kinetics of oxidants arriving at the interface through the oxide film. (C) 2004 Elsevier B.V. All rights reserved.

We propose an improved formula of a previous interatomic potential for Si, O mixed systems. The new potential is designed so as to more accurately reproduce the structural property of compressively strained SiO2 structures, by reducing unnatural steric hindrance caused by a long-range part of a three-body term. As the results of the improvement, (1) compressive stress in SiO2 film, which was highly overestimated to be 13 GPa by the earlier potential, is reduced to 2.7 GPa, and (2) a spurious peak in Si-O pair correlation function of SiO2 film disappeared. A limitation of the conventional interatomic potentials and its solution are also discussed. (C) 2004 Elsevier B.V. All rights reserved.

Large-scale SiO(2)/Si(111) models were constructed by introducing oxygen atoms in c-Si models in an atom-by-atom manner. Molecular dynamics calculation at a constant temperature was repeatedly carried out for the growing oxide model. By comparing the oxidation simulation of Si(111) substrate with that of Si(001) substrate performed previously, the influence of substrate orientation on the oxide structure was discussed. Owing to the significant feature of bonding arrangement within a Si bilayer in the Si(111) substrate, the inherent stress induced at the SiO(2)/Si interface by oxygen insertions is originally higher for the Si(111) oxidation than for the Si(001) oxidation, resulting in frequent changes in the bonding network. The resulting structure of bulk SiO(2) on Si(111) has less strain and a lower density than that on Si(001), but involves a larger number of dangling bonds. The X-ray diffraction pattern calculated for the SiO(2)/Si(111) model exhibits a diffraction peak with a Laue-function-like profile on each of the crystal-truncation-rods from the 111 and 11 (1) over bar points, agreeing well with experimental results. These diffraction peaks indicate that the thermally grown SiO(2) retains the residual order emanating from the {111} atomic planes in the original c-Si. Because of differences in the angles between the surface and the {111} atomic planes, the residual order within the SiO(2) differs depending on the substrate orientation.

The origin of x-ray diffraction peaks observed on the crystal truncation rods (CTR's) in reciprocal space for thermally grown SiO2 films has been investigated by large-scale atomistic simulation of silicon oxidation. Three models of SiO2 on Si(001), Si(111), and Si(113) were formed by introducing oxygen atoms in crystalline Si from the surfaces in an atom-by-atom manner. The SiO2 structures are classified as being amorphous in conventional characterizations, but retain the residual order originating from the {111} atomic planes in their parent crystals. The calculated diffraction patterns exhibit intensity peaks with Laue-function-like fringe profiles along the CTR's, at positions depending on the substrate orientations, agreeing quite well with experimental results.

By investigating the interactions between Green Fluorescent Protein (GFP) and Si substrates using molecular dynamics simulation, we got the following results. GFP adsorbs on the hydrophobic substrate directly, and on the hydrophilic substrate via water molecules. GFP-hydrophobic substrate attractions are stronger than GFP-hydrophilic substrate attractions. Flexibility of GFP is reduced when GFP adsorbs on the Si substrate.

By simulation of silicon oxidation and measurement of X-ray crystal-truncation-rod (CTR) scattering, the structures of silicon dioxide films grown at different temperatures and the structural changes due to thermal annealing have been investigated. Large-scale SiO2/Si(001) models were formed by introducing oxygen atoms, atom-by-atom, in crystalline Si from the surfaces. Molecular dynamics (MD) calculation at a constant temperature was repeatedly carried out for the growing oxide model. The intensity and position of the extra diffraction peak observed for the oxide, correlating with the residual order emanating from the parent Si crystal, depend on the growth temperature and change after thermal annealing. The peak intensity becomes smaller with increasing growth temperature. Thermal annealing monotonically decreases the peak intensity and shifts the position along the CTR,. toward the lower angle side. There is a good agreement between the results of simulation and experiment. It is shown that (1) the oxide grown At a higher temperature has a lower degree of residual order, (2) thermal annealing decreases the residual order, ultimately leads to complete amorphization and never restores the ordering, and (3) the peak shift along the CTR corresponds to the volumetric expansion of the SiO2 in the surface-normal direction.

Quantum chemical calculations were performed to investigate the probabilities of the existence of molecular and atomic oxygen inside Si and SiO2. Optimized configurations were obtained for both Si and SiO2 clusters including an O-2 molecule or an O atom in the spin singlet or triplet state. The spin triplet O-2 molecule is the most favorable in SiO2, while the spin singlet 0 atom is dominant in Si. The diffusion of an O-2 molecule from the outside to the inside Of SiO2 Was computed with its potential energy change. The dissociation reaction of an O-2 molecule at the SiO2/Si interface and the transfer reaction of an O atom across the SiO2/Si interface were also examined by estimating for their energy barriers.

Density functional calculations using model clusters were performed to clarify the atomic-scale diffusion mechanism of an O-2 molecule or an O atom in SiO2. The activation energy required for the atomic O diffusion was estimated to be 1.3 eV, whereas that for the molecular O-2 diffusion was revealed to be fairly low, 0.3 eV. This strongly suggests that the diffusing oxygen in SiO2 is primarily in a molecular form. The computational results were confirmed to be consistent between two SiO2 configurations of the cristobalite and quartz structures. Diffusion pathways and the related activation energies are shown to be well compatible with many recent works. [DOI: 10.1143/JJAP.42.3560].

We propose an estimate to quantitatively evaluate the Hausdorff dimension of a self-similar set based on a system of weak contractions each of whose contraction coefficient is not a constant but a function of a parameter. Using the estimate, we investigate the topological structures specific to this weak self-similar set. (C) 2001 Elsevier Science Ltd. All rights reserved.

Large-scale modeling of ultra-thin SiO₂ films on Si(001) surfaces has been performed by means of molecular dynamics utilizing our original inter-atomic potential energy function for Si, O mixed systems. The SiO₂ film is formed by layer-by-layer insertion of O atoms into Si-Si bonds in a Si wafer from the surface. The obtained models reproduce quantitatively the structural transition layers near the interface. Through a modeling of vicinal SiO₂/Si(001) model including atomic steps, it has been found that oxide film near the step-edge is preferentially amorphized. For a more advanced modeling method, we propose a new simulation procedure where O atoms are introduced into the substrate in one-by-one manner. In the calculation, the oxidation is started from the surface and abrupt change in composition at the SiO₂/Si interface is reproduced. Thus, the classical molecular dynamics is a powerful method together with a simplified inter-atomic potential function applicable to mixed systems.

First-principles quantum chemical calculations have been performed to clarify the nucleation site of Cu on the H-terminated Si(111) surface. The adhesion energies of a Cu atom on various sites have been obtained accurately. We have examined the following surface species as candidates for the Cu nucleation site: Si dihydride and monohydride species, fluoride, chloride, and hydroxide species, and locally oxidized sites. The basis set for Cu, which is one of the transition metals, has been chosen suitably. Results of our calculation indicate that (1) a Cu atom migrates freely on the H-terminated Si(111) surface, (2) it adheres selectively on the OH-terminated site, which is considered to exist mainly at a kink site and at the intersection of step edges, and (3) the Cu atom adhering on the OH-terminated site easily grows to a cluster.

Large-scale modeling of a SiO2/Si(001) structure including a couple of monoatomic steps has been performed by using a novel inter-atomic potential energy function for Si, O mixed systems. The SiO2 film was formed by layer-by-layer insertion of O atoms into Si-Si bonds of Si(001) 2 x 1 reconstructed surface misoriented from surface normal by 3.2 degrees. Just after the backbond oxidation of the dimer structures, a clear difference in the structural order of oxide films appeared between two types of terraces on which the dimer rows run perpendicular and parallel to the step lines. The difference is explained by the difference in stretching directions of Si-Si intervals into which O atoms are inserted. It was also found that the oxide structure near the step preferentially became amorphous after further oxidation. These results suggest a possibility that the structure of thin oxide film can be controlled by the initial Si surface preparation. (C) 2000 Elsevier Science B.V. All rights reserved.

Nucleation and growth kinetics of dimer-adatom-stacking-fault (DAS) structures on laser-quenched Si(1 1 1) surfaces have been investigated by measuring the time evolution of DAS domain size distribution at 320-440 degrees C with a scanning tunneling microscope (STM). The time evolution of the distribution suggests two different kinetics of DAS reconstruction At a very early stage (immediately after laser irradiation), the surface contains many voids and formation of DAS structures is quite rapid. In the next stage (after the disappearance of the voids), the DAS domains develop at constant nucleation and growth rates in a measurable time scale. The temperature dependence of nucleation and growth rates is discussed based on a typical theory of the two-dimensional structural phase transition.

A structural transition region near the SiO2/Si interface has been considered to play an important rule with respect to gale oxide reliability. We clarify the effects of the structural transition region on the time-dependent dielectric breakdown (TDDB) characteristics, particularly the activation energy of the oxide breakdown for ultrathin gate oxides formed by different oxidation processes, i.e., pyrogenic oxidation, rapid thermal O-2 oxidation and N2O oxynitridation. Furthermore, we investigate the properties of the structural transition region, such as the density of SiO2 as measured by the grazing incidence X-ray-scattering reflectivity (GIXR) method, the Si-O-Si bond angle by Fourier-transform infrared attenuated total reflection (FTIR-ATR), the etching rate by chemical etching and X-ray photoelectron spectroscopy (XPS). Through these investigations, it is clarified that the oxide breakdown tends to occur at the Si-O-Si network with a lower bond angle (&lt;115 degrees) and that the strain in the structural transition region reduces the barrier to the oxide breakdown. A 1-nm-thick strained layer is round to have a strong effect on the oxide reliability and to limit oxide scaling in future ultra-large-scale integrated circuits (ULSIs).

A novel interatomic potential energy function is proposed for condensed systems composed of silicon and oxygen atoms, from SiO2 to Si crystal. The potential function is an extension of the Stillinger-Weber potential, which was originally designed for pure Si systems. All parameters in the potential function were determined based on ab initio molecular orbital calculations of small clusters. Without any adjustment to empirical data, the order of stability of five silica polymorphs is correctly reproduced. This potential realizes a large-scale modeling of SiO2/Si interface structures on average workstation computers.

Large scale modeling of ultrathin SiO2 films on Si(100) surfaces has been performed using our original potential, which was developed to simulate both Si and SiO2 crystal systems. A SiO2 film was formed by layer-by-layer insertion of oxygen atoms into Si-Si bonds in a Si wafer from one of the surfaces. The thickness of the obtained SiO2 layer was about 17.2 Angstrom, and it showed the presence of the structural transition layer; the average Si-O-Si bond angle becomes smaller in the region closer to the SiO2/Si interface. The peak of Si-O-Si bond angle distribution is shifted toward a narrower angle from the equilibrium angle of 144 degrees, in agreement with experimental results reported so far. (C) 1999 Published by Elsevier Science Ltd. All rights reserved.

Coverage of the 7 x 7-reconstructed region on quenched Si(111) surfaces has been compared for two types of Si wafers with different oxygen concentration, Czochralski and modified float zone (m-FZ) wafers. The m-FZ wafer clearly showed the lower coverage, which suggested that oxygen had some influence on the formation of a 7 x 7 structure. The activation energy of formation of the 7 x 7 structure has been estimated to be 2.4 eV. [S0163-1829(98)01939-0].

Time-independent and -dependent Monte Carlo simulations of a two-dimensional lattice gas were performed to investigate the effect of fixed particles on periodic adatom arrangements on Si(111) 1 X 1 surfaces. In both the simulations introducing fixed particles, domains of the periodic structures were divided into small fragments around them. This can be explained by the fixed particles reducing the degree of freedom of the adatom configurations so the number of ways of reaching the uniform configuration is highly restricted. Based on the simulations, we predict that a large 1 X 1 region without any obstacle impurities is desirable to obtain a large periodic structure of Si adatoms. (C) 1998 Elsevier Science B.V. All rights reserved.

The formation process of stacking-fault (SF) triangles during the Si(111) 1 x 1--&gt;7 x 7 reconstruction has been studied using the quantum chemical theoretical calculations. On the basis of the propagation and subsequent merging mechanism of the SF areas proposed by one of authors (I.O.), several reaction paths to complete a single SF triangle have been examined. The most probable formation process of SF triangles has been determined from the their lowest energy reaction path. A comparison of total energy changes along the SF triangle formation both with and without oxygen atoms indicated a preference of oxygen incorporation in the formation of 7 x 7 dimer-stacking-fault (DS) structure, which was compatible with the recent experimental results suggesting the important role of oxygen in the 7 x 7 reconstruction. The effect of the step edge has also been discussed. (C) 1997 Elsevier Science B.V.

Monte Carlo simulations of a two-dimensional lattice gas in the grand canonical ensemble are performed to analyze the behavior of Si adatoms on Si(111)-1 x 1 surfaces. 2 x 2, c2 x 8 and c2 x 4 structures appear at 300 K under the condition that only Coulomb repulsion is introduced as a pair interaction between adatoms. The adatom density increases with the temperature, and root 3 x root 3 structures appear at 550 degrees C. These results agree with our recent STM observations. (C) 1997 Elsevier Science B.V.

Semiempirical molecular orbital calculations were performed to investigate the mechanism of H-2 desorption from dihydride species on Si(001) surfaces. The lowest energy pathways were calculated with respect to three different mechanisms which have been proposed previously. We performed additional calculations under the different H coverage conditions to examine the dependence of activation energy on the varieties of surrounding hydride species. The new transition state structure was obtained by the calculation of the recombinative desorption of two H atoms from adjacent Si dihydrides. We have found that the activation barrier of the recombinative desorption mechanism was the lowest of all and it was hardly influenced no matter what the surrounding hydride species is.

Si(111) surface reconstructions are classified into two families, the 2 x 2 family (2 x 2, c2 x 4, c2 x 8 and root 3 x root 3) and n x n DAS family. By in-situ atomic scale observation of Si(111) surface reconstruction and by a statistical argument on the nucleation of a daughter phase in Si(lll) matrix, we have found that the 2 x 2 family is a result of random motion of adatoms on a Si(lll)-1 x 1 substrate, while the n x n DAS family can never be formed only by the movement of adatoms but some cooperative movement of substrate Si atoms is necessary.