Session We-A1

a-Si/a-Ge: Alloys and Clathrates I

Chair: Gurinder K. Ahluwalia, College of the North Atlantic

We-A1.1 8:20–8:40

A Method to Evaluate Explosive Crystallization Velocity of Amorphous Silicon Films During Flash Lamp Annealing

Keisuke Ohdaira

Japan Advanced Institute of Science and Technology, 1-1 Asahidai, Nomi, Ishikawa 923-1292, Japan

PRESTO, Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan

Flash lamp annealing (FLA) is a method to anneal samples for a few milliseconds using pulse light emitted from Xe lamps, and has thus been widely used for the activation of ion-implanted dopants and the crystallization of amorphous materials on substrates with low thermal tolerance. In particular, micrometer-order-thick amorphous silicon (a-Si) films can be fully crystallized by a single shot of flash lamp light, which is applicable for the formation of thick polycrystalline Si (poly-Si) for a photovoltaic material. We have so far observed explosive crystallization (EC), high-speed lateral crystallization driven by the release of latent heat, of a-Si films induced by FLA [1]. There can be several EC modes, such as EC governed by liquid-phase epitaxy (LPE) and EC involving solid-phase process, and it is essential to know the velocity of ECs in order to discuss the mechanism of the ECs quantitatively. Unlike the case of laser or other scanning annealing techniques, it is difficult to equip an in-situ observation system for the EC in a FLA system because of large-area irradiation. Instead, we have developed an easy method to evaluate EC velocity by using multi-pulse FLA system, in which one millisecond-order pulse consists of a number of sub-pulses and the emission frequency of the multi-pulses can be changed systematically. The irradiation of multi-pulse leads to the cyclic modulation of the temperature of a Si film, resulting in the formation of macroscopic stripe patterns on poly-Si films formed. We can thus easily estimate the velocity of EC from sub-pulse emission frequency and the width of stripe patterns (a length of EC for one sub-pulse).

We performed the multi-pulse FLA for a-Si films prepared by three deposition methods: sputtering, catalytic chemical vapor deposition (Cat-CVD), and electron-beam (EB) evaporation. All the films are crystallized by FLA, and have macroscopic stripe patterns on their surfaces, indicating that EC occurs during the irradiation of the multi-pulse. The widths of the macroscopic stripe patterns are changed systematically, depending on multi-pulse emission frequency. EC velocities estimated can be divided into two groups: ~4 m/s and ~14 m/s. Cross-sectional transmission electron microscopy (TEM) images of these poly-Si films show that the former EC leaves behind microscopic (~1 μm) periodic microstructures and 10 nm sized fine grains, while the latter EC forms poly-Si films consists of large grains stretched along EC directions. The EC velocity of ~14 m/s corresponds to an LPE speed around the melting point of a-Si. This fact means that the large-sized grains are formed through LPE-based EC. EC with a velocity of ~4 m/s is, on the other hand, involves solid-phase nucleation (SPN), since nucleation rate to form densely packed 10 nm sized fine grains cannot be explained by nucleation from Si melt. The EC velocities obtained by the multi-pulse method are therefore consistent with the phenomena taking place in the Si films. We have also found that EB-evaporated a-Si films or thinner a-Si films tend to be crystallized by LPE-based EC.

[1] K. Ohdaira, et al., J. Appl. Phys. 106, 044907 (2009)

Keywords: flash lamp annealing, crystallization, amorphous silicon, polycrystalline silicon, explosive crystallization

We-A1.2 8:40–9:00

Rectifying and Schottky Characteristics of a-SixGe1–xOy with Metal Contacts

Md Muztoba (1), Donald Butler (2), and Mukti Rana (1)

1. Department of Physics and Engineering and Optical Science Center for Applied Research (OSCAR), Delaware State University, 1200 N DuPont Highway, Dover, DE, USA

2. Department of Electrical Engineering and Nanotechnology Research and Education Center, University of Texas, Arlington, TX, USA

Metal-semiconductor contacts are vital part of semiconductor devices as they can form a Schottky barrier or an Ohmic contact. The nature of the contact plays an important role to determine the electrical and physical characteristics of the device and hence is of paramount importance in device operation. In current work, we report the design, fabrication and current-voltage (I-V) characteristics involving a-SixGe1–x and a-SixGe1–xOy sensing layers. We also report the formation of both Schottky and Ohmic contacts involving a-SixGe1–xOy with Ti, Cr, Al, Au, Ni or Ni0.80Cr0.20 metals. Eight different types of bolometers were fabricated with a-Si0.15Ge0.85, a-Si0.15Ge0.85 (Si n-doped) and a-Si0.15Ge0.85O0.039 sensing layers. Bolometers with a-Si0.15Ge0.85 sensing layer formed Schottky contact with Ti, Ti-Au, Cr, Cr-Au and Al metal contacts while a-Si0.15Ge0.85 (Si n-doped) formed an Ohmic contact with Au. For Si0.15Ge0.85O0.039 sensing layer, both Ni and Ni0.80Cr0.20 contact metals formed Ohmic contact. The I-V characteristics of the microboloemters were analyzed with a thermionic emission model. Linear dependence on the Ge composition was approximated to find the effective Richardson constant. The theory predicts Richardson constants of 112 A/cm2K2 and 50 A/cm2K2 for Si and Ge respectively. Microbolometers with Si0.15Ge0.85 sensing layer formed Schottky contact with barrier heights of 0.7039 V, 0.7376 V, 0.7555 V and 0.7062 V for devices with Ti, Au, Al and Cr contact metals respectively. Microbolometers with Si0.15Ge0.85 (lightly doped n-type Si) formed Ohmic contact with Au while microbolometers with Si0.15Ge0.85O0.039 sensing layers formed Ohmic contacts with Ni and Ni0.80Cr0.20 metals.

Keywords: silicon germanium, silicon germanium oxide, ohmic contact, Schottky contact, Schottky barrier height

We-A1.3 9:00–9:20

Crystallization of Silicon-Germanium Induced by Aluminum-induced Layer Exchange

Masao Isomura (1), Masahiro Yajima (1), and Isao Nakamura (2)

1. Course of Electrical and Electronic System, Graduate School of Engineering, Tokai University, 4-1-1 Kitakaname, Hiratsuka, Kanagawa 259-1292, Japan

2. Tokyo Metropolitan Industrial Technology Research Institute, 1-20-20 Minamikamata, Ota, Tokyo 144-0035, Japan

We have studied crystallization of silicon-germanium (a-SiGe) by aluminum (Al) induced layer exchange (ALILE). The ALILE was conducted with a starting structure of glass/Al/Al-oxide/a-SiGe, and the a-SiGe is crystallized during the layer exchange of SiGe and Al caused by thermal annealing. In the literature [1], the thermal annealing below the eutectic temperature of binary Al-Ge system (420°C) is necessary for the ALILE of SiGe to avoid alloying of Al and Ge. However, the ALILE below the eutectic temperature takes more than 10 days. We examined the ALILE at 450°C slightly higher than the eutectic temperature, and the ALILE time could be significantly shortened to a few hours, although a little alloying effect was observed [2]. In this report, we investigated the dependence of the ALILE on thicknesses of Al and a-SiGe, and successfully controlled the formation of poly-crystalline SiGe on glass substrates.

The thickness of Al layers evaporated on the glass substrates was varied from 2000Å to 3200Å. The a-SiGe was deposited on the oxidized Al surface by radio-frequency (RF) magnetron sputtering, and its thickness was varied from 2000Å to 2800Å. The Al-oxide layers have an important role for the ALILE and formed in the atmosphere at RT for a few days. We observed the Al-oxide layers at the same position after the ALILE. It was confirmed that the SiGe and Al are exchanged through the Al-oxide layers.

When the Al thickness is thin like 2000Å, the ALILE occurs in small areas due to insufficient amount of Al. On the other hand, the ALILE clearly occurs when the Al thickness is sufficiently thick like 2800Å and more. In the case that a-SiGe and Al thickness are both 2800Å, the ALILE occurs in almost all region. After elliminating the surfaced Al, unform poly-crstalline SiGe layers are formed on the substrates.

When the Al layer is thicker than the a-SiGe layer, the ALILE patially occurs and formes island structures in Al matrix due to the segregation of extra Al. The ALILE and Al segregation were confirmed by depth profiles of Si, Ge and Al compositions measured by depth profiles of the Auger electron spectroscopy. The ALILE islands become smaller with thicker Al layers because of more extra Al. After elliminating the surfaced and segregated Al, island structures of poly-crystalline SiGe are formed on the substrates and size of the island can be controlled by the Al thickness.

The results indicate that the configuration of poly-crystalline SiGe can be controlled from uniform films to island structures with various sizes. This technique might be useful to form poly-crystalline SiGe on low cost substrates. Especially, the island poly-crystalline SiGe is expected as seed crystals to grow crystalline films.

[1] M. Gjukic, M. Buschbeck, R. Lechner and M. Stutzman, Appl. Phys. Lett. 85, 2134 (2004)

[2] T. Iwasa, T. Kaneko, I. Nakamura and M. Isomura, Phys. Status Solidi A 207, 617 (2010)

Keywords: Silicon germanium, Aluminum induced layer exchange, Crystallization

We-A1.4 9:20–9:40

Electron-spin Resonance Studies on Na-doped Type II Si Clathrates

Mitsuo Yamaga (1), Masato Aoki (1), Takumi Kishita (1), Shogo Sunaba (1), Fumitaka Ohashi (1), Takayuki Ban (1), Tetsuji Kume (1), Kouhei Goto (2), Genki Shimizu (2), and Shuichi Nonomura (2)

1. Faculty of Engineering, Gifu University, Gifu 501-1193, Japan

2. Graduate School of Engineeering, Gifu University, Gifu 501-1193, Japan

Silicon clathrates are classified into type I and type II with the chemical compositions of Na8Si46 and NaxSi136 (0≤x≤24), respectively. Type I Na8Si46 clathrate shows a metallic property, whereas electric conductivity of type II NaxSi136 clathrates changes from a semiconductive to metallic behavior with an increase of Na concentration of x. In this paper, we report electron-spin resonance (ESR) results observed in type II Si clathrates, giving information on electronic structure of localized or delocalized electrons of Na atoms in type II Si clathrates.

Type II clathrates NaxSi136 (0.9≤x≤11.2) were synthesized by chemical reaction of Si and Na in an Ar atmosphere at 923K for 48 hours and thermal annealing in vacuum at 673K for 3 hours. Concentrations of Na in the clathrates were controlled by further thermal treatment under vacuum with annealing temperature at 673K or 723K and annealing duration varied from 0 to 66 hours. ESR spectra were measured at 77K and in the temperature range from 133K to 323K.

The ESR spectra for NaxSi136 (0.9≤x≤11.2) consist of several lines when measured at 77K. The four hyperfine-structure (hfs) lines with g=2.040 and a width of 0.7 mT are due to nuclear spin (I=3/2) of an isolated Na. The intensities decreased gradually when x increased from 0.9 to 4.6 and were negligibly small above 4.6. Three weak and sharp hfs lines were also observed in the middle of the adjacent four hfs lines. These lines are assigned to an electron trapped at two Na+ ions in the closest proximity. A single broad line with the same g value (g=2.04–2.05) as the four hfs lines and a width of 5 mT was observed. The intensity increased gradually with an increase of x up to 4.6. These results suggest that such broad line may be due to an exchange-coupled pair of Na atoms. In addition, a sharp single line with a width of 0.5 mT was observed for all concentrations of x. The g values are very close to that (g=2.003) of a free electron and were almost independent of temperature. Thus, this single line may be assigned to an electron localized in the form of dangling bonds. The g values of the ESR lines from the isolated Na atom and the exchange-coupled Na pair are obtained to be 2.04 close to each other. The positive (negative) g-shift of the 3s-electron in Na from the g value (g=2.0023) of free electrons is caused by a partial electron transfer through the spin-orbit interaction from (to) Si ligands to (from) a central Na atom in a large cage of Si28. In order to explain the positive g-shift, we calculate the spin-polarized band structure of Na-doped type II Si clathrate with periodic Na configurations and the contribution of the wavefunctions of the valence and conduction bands to those of the Na intermediate band.

Keywords: Si clathrate, semiconductor, Sodium, Electron-spin resonance

We-A1.5 9:40–10:00

Synthesis of Si Clathrate Films via Thermal Decompositions of Zintl Phase NaSi on Si Substrates

Fumitaka Ohashi (1), Masashi Hattori (1), Yoshiki Iwai (1), Takuya Ogura (2), Akihiro Noguchi (1), Tetsuji Kume (2), Takayuki Ban (2), and Shuichi Nonomura (1)

1. Environmental and Renewable Energy Systems Division, Graduate School of Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan

2. Materials Science and Technology Division, Graduate School of Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan

Type I (Si46) and type II (Si136) Si clathrates are materials composed with polyhedral cages, Si20, Si24 and Si28, which share their faces with neighbors. Such cage structures are generally formed by thermal decompositions of Zintl phase materials such as sodium silicide (NaSi). The evaporation of Na from the NaSi gives rise to construction of a cage-like Si framework in which void spaces accommodate residual Na atoms as guests. The Na atoms provide electrons to Si framework due to their ionization, and thus the Na doped Si clathrates typically show metallic characteristics. Among the clathrate materials, the Na doped type II Si clathrate (NaxSi136) has been known as a clathrate of which Na concentrations vary in the range x = 0 to 24 by post-annealing under vacuum. As the Na concentration decreases, the electrical property of NaxSi136 changes from metallic to semiconducting. When the guest atoms are removed completely, Si136 has been suggested to have a direct band gap of about 1.8 eV. The attractive property motivated us to develop the Si clathrates as novel photo absorption materials for solar cells. However, the investigations on the properties have been limited so far because the synthesized samples were in powdery forms. In order to investigate its optical and electronic properties in detail, it is highly required to establish the synthesis techniques of Si clathrate as film forms. In this paper, we attempted to synthesize the film of Si clathrate, by using single-crystalline Si substrates as starting materials.

Si(100) or Si(111) substrates are used as a material source to form Si clathrate film, together with Na. The Si substrate underwent a chemical reaction in Ar atmosphere with Na vapor in a closed vessel heated at 500°C ~ 650°C for 12 ~ 96 hours. These processes resulted in formation of NaSi film on the surface of Si substrate. Further annealing under vacuum at 400°C for 3 hours induced thermal decomposition of NaSi into Si clathrate. X-ray diffractions (XRD), scanning electron microscope (SEM) and Raman spectroscopy were used for their characterizations.

The growth rate of NaSi depended strongly on the crystal plane of the surface of Si substrate; Si(100) substrate more highly reacted with Na than Si(111). Higher reactivity of Si(100) substrates allowed us to synthesize NaSi films at lower annealing temperature. NaSi was grown on the substrate as a film highly orientated, and the direction of the crystal orientation depended on that of the surface of Si substrate. Further annealing under vacuum changed the NaSi films to Si clathrates films. Although NaSi films have been highly orientated, the resultant Si clathrate films were polycrystals with random crystal orientation. SEM images showed formations of Si clathrate film of 3–50 μm in thickness which depended on the annealing temperature and duration on synthesizing NaSi.

Keywords: Si clathrate, film, Zintl phase, NaSi