Visible Photoluminescence in Carbon-Implanted Thermal SiO2 Films
533
Fig. 2. PL spectra of argon-implanted
SiO2 films at room temperature
carbon implanted thermal SiO2
films is due to these defects or
not. To address this case, 150 keV
argon ion implantation into 400 nm
thermal SiO2 films on Si at a dose
of 1.4 Â 1016 cm±±2 was performed.
The energy and dose were care-
fully chosen using the TRIM, such
that the projected range, the strag-
gle, and the peak concentration of
atomic displacement profile is cre-
ated by the carbon ion implanta-
tion. Fig. 2 shows the room-tem-
perature PL spectra of thermal
SiO2 films implanted with argon
ions and annealed at the same conditions as shown in Fig. 1. As shown in Fig. 2, the as-
implanted one showed a broad band with a peak at 650 nm. The origin of this band
was related to the damage and defects produced by ion implantation (Non-Bridging
ꢀ
Oxygen Hole center, NBOH) [9]. This band disappeared after annealing over 800 C as
a result of repairing the irradiation-induced damage and defects.
ꢀ
Meanwhile, after annealing over 800 C, a weak and clearly defined band occurred at
the same wavelength as in thermal SiO2 films. The intensity of the band at 460 nm kept
unchanged when the annealing temperature varies from 800 to 1000ꢀC. It seems as if
the irradiation-induced damage passive the pre-exꢀisting defects in the thermal SiO2
more pre-existing defects exist even after the 1000 C treatment. Based upon the data
shown in Figs. 1 and 2, it was found that the structures formed after carbon implanta-
tion in thermal SiO2 and after annealing, which present visible photoluminescence, are
not due to the microstructure change of thermal SiO2 caused by high temperature an-
nealing and defects produced by ion implantation. As the blue PL emission from the
SiC nanocrystals of the porous materials was reported [10, 11], it is also important for
us to discuss whether SiC clusters exist in carbon implanted SiO2 or not. No discernible
traces of Si±C bonds could be detected by infrared (FTIR) measurements. This result
suggests that under these experimental conditions, SiC clusters are not formed or at
least, the amount of the clusters is so small to be undetectable. Furthermore, if we
consider the energy of Si±Si, Si±C and C±C bonds (2.36, 3.21 and 3.70 eV, respec-
tively) and maximum atomic concentrations of the implanted ions in the projected
range, phase separation in SiC and carbon aggregates is thermodynamically favorable
(i.e. Si±C bonds are stronger than Si±Si ones). However, we have to consider the slow
kinetics for SiC formation at 1000 ꢀC from the possible reactions between implanted
ions and the matrix [12]. Then, it was found that the annealing temperature should be
as high as 1300 ꢀC so that the reaction could be kinetically possible. HRTEM was
carried out on the samples similar to those in Fig. 1 in order to find out microstructures