Analysis of white‐luminescent mesoporous carbonized silica under electron irradiation by TEM‐CL system

Abstract

Mesoporous carbonized silica (MPCS) exhibits continuous white photoluminescence (PL) like the solar emission spectrum [Fig.1 (a)], which attracts attention as a new luminescent material, consisting of ubiquitous elements alone. The PL intensity of MPCS, however, is still insufficient for practical use and it is necessary to clarify the origin of the white luminescence to improve its emission intensity. Since MPCS has a nanoscale honeycomb structure and the emission intensity and its color rendering depend on the pore size, it was thought that the nanostructure is responsible for the light emission. In the present study we conducted TEM‐cathodoluminescence (CL) experiments on MPCS in order to investigate relationship between luminescence properties and nanoscale honeycomb structure. An MPCS sample was ground and dispersed in ethanol, a drop of which was dripped on a microgrid. TEM‐CL experiments were carried out by a Jeol JEM2100 STEM equipped with a Gatan Vulcan TEM‐CL system, in which high efficiency light collecting mirrors are implemented, covering both upper and lower sides of the sample over the solid angle of as large as 7.3 steradian. Using this system, it was found that the CL spectral profile was different from that of PL, consisting of several primary peaks around 450 nm, 620 nm and 650 nm [Fig.1 (b)]. Several samples with different synthesis process and carbon content were examined at electron accelerating voltages of 200 kV and 100 kV to investigate the change in the CL spectrum by electron irradiation. It was found that the CL spectral profile changed with the structural change by electron irradiation at an accelerating voltage of 200 kV, as shown in Fig. 2. At the initial stage of irradiation the nano‐structure of MPCS hardly changed, though the intensity of the 620‐650 nm peak was quickly reduced. By prolonged irradiation, on the other hand, the spectral intensity of the intermediate wavelength region (500‐600 nm) between the two primary peaks was increased, followed by the increased integrated intensity of overall spectrum as the honeycomb structure collapsed. It is thus concluded that the origin of the luminescence was attributable to generation and annihilation of point defects rather than nanoscale honeycomb structure itself. Electron irradiation is supposed to break chemical bonds by electronic excitations and/or displacement of host atoms through inelastic and elastic electron scattering, respectively. It is generally known that the inelastic scattering probability is increased and the knock‐on damage probability is decreased with decreasing the accelerating voltage, where a threshold accelerating voltage exists for the displacement of an atom by elastic scattering (knock‐on), depending on the atomic number of the displaced atom. At the accelerating voltages of 200kV and 100 kV, well above and below the threshold for Si and O (~150 kV) in SiO 2 respectively, the intensity of 620‐650 nm peak was decreased both in case of 100 kV 200 kV, though its rate was larger at 100 kV than that at 200 kV, as shown in Fig.3. This suggests that the defect specie giving rise to the 620‐650 nm peak should collapse by an electronic excitation effect rather than knock‐on damage. Besides, the increase in the intensity the intermediate wavelength region observed in Fig. 2 was not observed at 100 kV. It is hence presumed that the origin of the emission in the intermediate wavelength region should be the generation of a defect related to oxygen deficiency by knock‐on damage. Moreover, in order to confirm how carbon incorporated affects the emission spectrum of MPCS, CL spectrum of a non‐carbonized mesoporous silica (MPS), having the same honeycomb structure as MPCS was acquired, as shown in Fig.4. The CL spectral profile of MPS hardly differed from that of MPCS. It has not become clear how the carbon in MPCS affects the light emission. The generation and annihilation of point defects in MPCS by electron irradiation should be modeled by a set of reaction rate equations on the basis of the present observations, which is reported in further detail.