Various works on silicon (Si)-based luminescent materials utilizing the quantum-confinement effect, such as porous Si [1] [2] and Si nanocrystal (Si-nc) [3] -[5] , have been reported. Ultraviolet (UV) light emission can be obtained from such materials fabricated using various methods such as laser ablation [6] and sputtering [7] [8] . UV photoluminescence (PL) spectra having a peak wavelength around 370 nm were obtained from sputtered SiO₂ films including Si-ncs after annealing [7] . The UV light emission might originate from emission centers of interface layers (SiO_x) between Si-ncs and SiO₂ matrices [7] .

Furthermore, Si/SiO₂ multilayers produced by alternately stacking several-nanometer-thick Si and SiO₂ layers using the sputtering method are very important structures from which we can controllably and effectively produce such Si-ncs and SiO_x [8] . In these multilayered structures, Si-ncs seem to be produced from nanometer- order-thick Si layers, and SiO_x can clearly be formed after annealing (see Figure 1) [8] . In our previous study, we fabricated Si/SiO₂ multilayered films having nanometer-order-thick Si and SiO₂ layers using radio-frequency (RF) magnetron sputtering and observed UV PL having a sharp peak at a wavelength of 370 nm after annealing at 1150˚C to 1250˚C [8] . Such UV-light-emitting materials may be used in future higher-density optical disk systems.

However, the UV peak intensities from the Si/SiO₂ multilayered films should be stronger than those in our previous study in order to be used as light-emitting materials. We considered that the Si-layer-thickness ratio of the film was a very important parameter for enhancing the peak intensity because the ratio was concerned with the size of Si-nc, which might affect the intensity of the UV light emission [7] . Therefore, in this study, we investigated effects of the Si-layer-thickness ratios on the UV peak intensity of the Si/SiO₂ sputtered and multilayered films.

A RF magnetron sputtering apparatus (ULVAC, SH-350-SE) was used in our fabrication processes. Si and SiO₂ plates having diameters of 100 mm were used as sputtering targets, and Si/SiO₂ multilayered films were deposited by alternately supplying RF power to the Si and SiO₂ targets. The flow rate of Ar gas introduced into the chamber was 10 sccm, and the RF powers supplied to the Si and SiO₂ targets were 50 and 100 W, respectively. Commercial fused-silica substrates (ATOCK Inc., 11.5 mm × 24.5 mm × 1 mm^t) were used and not heated during deposition. One hundred pairs of Si and SiO₂ layers were alternately stacked on the substrates. The deposition rates of Si and SiO₂ thick films were 0.05 and 0.12 nm/s, respectively. Switching of sputtering targets and opening or closing of two shutters positioned above the targets were automatically controlled by digital timers. The total thicknesses of multilayered films were measured using a surface profiler (ULVAC, Dektak 3ST). We can estimate the Si and SiO₂ layer thicknesses (d_Si and, see Figure 1) from the measured total thicknesses of the films and the ratio of their deposition rates (0.5:1.2). In this study, we prepared seven samples with estimated Si-layer-thickness ratios d_Si/(d_Si +) of 0.15, 0.23, 0.24, 0.26, 0.27, 0.29, or 0.30 (Table 1). Four specimens were prepared from one as-deposited sample by cutting it using a diamond-wire saw. We subsequently annealed the specimens in ambient air at 1150˚C, 1200˚C, or 1250˚C for 25 min using a Siliconit electric furnace after sputter deposition.

PL spectra from the annealed samples were measured at room temperature with excitation using a He-Cd laser

(Kimmon, IK3251R-F, wavelength l = 325 nm). A monochromator (Nikon, P250), a photomultiplier (Hamamatsu, R2658), and a lock-in amplifier (NF, LI-572B) were used in our measurements. The intensities of PL spectra were calibrated by using a white-light spectrum measured with an optical spectrum analyzer (Anritsu, MS9701C + MS9030A) at wavelengths from 350 to 1750 nm because sensitivities of the monochromator and the photomultiplier depend on input light wavelengths.

Figure 2 presents typical PL spectra of specimens prepared from the sample with a Si-layer-thickness ratio of 0.29 (sample 6) after annealing at 1150˚C, 1200˚C, and 1250˚C. The total thickness of the multilayered film was 463 nm, and the thickness of the Si and SiO₂ layer can be estimated to be 0.9 and 2.2 nm, respectively. UV PL peaks around a wavelength of 370 nm were observed from all the specimens regardless of annealing temperature. The maximum arbitrary intensity of the UV PL peak was 8.45.

Figure 3 plots the maximum intensities of the UV PL peaks versus the Si-layer-thickness ratios d_Si/(d_Si +) of the samples. The data in our previous work (plotted as a rectangle in Figure 3) [8] agreed well with the present graph. From the result, we can estimate that the proper Si-layer-thickness ratio to obtain the strongest UV peak from our Si/SiO₂ multilayered films is around 0.29.

We investigated the effects of the Si-layer-thickness ratios on the UV peak intensities of Si/SiO₂ multilayered films produced by alternately stacking several-nanometer-thick Si and SiO₂ layers using RF sputtering for the first time. The proper Si-layer-thickness ratio to obtain the strongest UV peak from our Si/SiO₂ multilayered films was estimated to be around 0.29. The sizes of Si-ncs seem to be controllable by selecting both the amounts

of Si and SiO₂ deposition and the annealing conditions [8] . We will try to optimize the conditions (annealing temperature and time) in order to further increase the UV light-emission intensity of the Si/SiO₂ multilayered films.

Such sputtered films can be used as constituent materials of autocloned photonic crystals that can be applied to novel light-emitting devices [9] -[11] . UV-light-emitting materials and devices are expected to be used in future higher-density optical-disk systems.

This work was supported by JSPS KAKENHI Grant Number 26390073. Part of this work was conducted at the Human Resources Cultivation Center (HRCC), Gunma University, Japan.