1. Introduction

The high-resolution optical microscope, which enables us to observe specimens with subwavelength resolution, has attracted great interests in many fields in science and technology, and it is surely indispensable for research and development in biomedicine, nanotechnology and materials science. Multiphoton optical microscopes and near-field optical microscopes are important examples [1–5]. The spatial resolutions of these optical microscopes are still inferior to those of conventional electron microscopes. However, specimens can be observed in an air atmosphere with these advanced microscopes. Taking advantage of their properties, we can perform nondestructive in-vivo imaging of biological materials.

We, recently, proposed the Electron-beam-eXcitation-assisted (EXA) optical microscope as a new high-resolution optical microscope [6]. Schematics of the EXA optical microscope are shown in Figs. 1(a) and 1(b). The ultrasmall light source used in the EXA microscopes was cathode luminescence (CL) excited by high energy focused electron beams in phosphor thin films of nanometric thicknesses. The specimens were illuminated by CL light with nanometric spot sizes on the film surfaces. The EXA microscope probed the optical properties of the surface of the specimens in contact with the phosphor thin films.

 

Fig. 1 (a) Schematics of EXA optical microscope. (b) Enlarged schematic view of specimen irradiated with nanometric cathode luminescence light from Y₂O₃:Eu³⁺ phosphor films. (c) Top and side views of Y₂O₃:Eu³⁺ phosphor films on Si₃N₄ membrane. PMT: photomultiplier tube.

Download Full Size | PDF

The CL spot sizes are determined by the trajectories of inelastically scattered electrons and the subsequent diffusions of excited electron-hole pairs in the phosphor films. According to Monte Carlo simulations in a previous study, the spatial broadening of the focused electron beams with keV-order kinetic energies was narrower than 100 nm when they were transmitted through films less than 100 nm thick [7,8]. Hence, in principle, the EXA microscope is able to construct the images of the specimens with sub-wavelength spatial resolutions, supposed that the specimens are thinner than the wavelength of the illumination lights.

In our previous study, 50-nm-thick silicon nitride (Si₃N₄) membranes were used as the phosphor films for the EXA microscope. The mechanical strength of the Si₃N₄ membrane was high enough to sustain an air atmosphere and vacuum. Furthermore, the membranes were transparent to the high-energy electron beams. Taking advantage of these unique properties, Jonge et al. proposed atmospheric pressure scanning transmission electron microscopes, and they succeeded in recording gold nanoparticles 1.0 nm-diameter in an air atmosphere [9]. In our studies, the Si₃N₄ membranes were used as the emitter for the ultrasmall illumination light sources of the EXA microscope. We recorded poly styrene latex with average diameters of 50 nm with sufficiently high spatial resolutions.

The EXA microscope resembles to the CL microscope, where the microscope images are constructed by mapping the CL lights emitted from the specimens. The CL microscope is useful only for observing specimens tolerant of exposure to high-energy electron beams. Hence, the objects to be measured with CL microscopes were virtually limited to inorganic materials, such as semiconductors, ceramics or petroleum [10–12]. In contrast, the specimens are not directly irradiated by electron beams in the EXA microscope. We can observe organic materials with low tolerance to high-energy electron beams without damaging them by using the EXA microscope.

Our previous EXA microscope took advantage of the ultraviolet (UV) CL emissions from the Si₃N₄ membrane. It is important to tune the wavelengths of the nanometric CL lights so as to broaden the applicability of the EXA microscope in science and technology. In particular, it is crucial to prepare ultrasmall visible light emitters for biomedical applications because UV lights are not friendly to biological materials, including cells and living tissue.

In this study, Y₂O₃:Eu³⁺ phosphor thin films were produced to prepare ultra-small visible light emitters for the EXA microscope. Y₂O₃:Eu³⁺ phosphor materials have attracted great interests as red luminescence materials with high quantum yields. Their application to flat panel displays has been studied extensively [13–15]. The film preparation conditions for high brightness with a nanometric spot size are examined in terms of the post-annealing temperatures and film thickness. The performances of the EXA microscope containing the phosphor thin film are also reported.

2. Experimental

The Y₂O₃: Eu³⁺ phosphor thin films were deposited on the Si₃N₄ membrane substrates (Silicon nitride membrane, Silson Ltd. 25 × ) by an electron beam evaporation method [16]. The schematic structure of the phosphor films on the Si₃N₄ membrane substrate is shown in Figs. 1(b) and 1(c). First, Y₂O₃ and Eu₂O₃ powders (Kojundo Chemical Laboratory Co.) were mixed and pressed into pellet. The pellets were sintered at 1000°C. The concentration of Eu³⁺ ions in the mixture was 2 mol%. The materials were evaporated in a vacuum chamber at the pressure of ~10⁻⁶ Torr. Electron beams with 50-mA currents were accelerated at a voltage of 4 kV. The film thickness deposited on the Si₃N₄ membrane was monitored with a quartz crystal microbalance method.

The crystallinity of the as-deposited thin films was low, and the quantum efficiency was much smaller than that of films with higher crystalllinity. A post-annealing procedure was performed at temperatures between 500 and 1000°C under an oxygen gas flow for 3 h to improve the crystallinity. The crystallinity of the thin films was characterized by their powder X-ray diffraction (XRD) patterns (Rint Ultima-III, Rigaku Co.). The XRD patterns of the as-prepared and annealed samples were recorded using the CuKα radiation (λ = 1.54056 Å) operating at 40 kV and 150 mA at a scanning rate of 0.05° per step in the 2θ range of 10° < 2 θ <80°. The topography of the film surfaces was examined using an atomic force microscope (AFM) (SPA-400, SEIKO Instruments Inc.).

The EXA microscope was built by modifying a thermal-type scanning electron microscope (SEM) (JSM-6390, JOEL Ltd.) The electron beams were applied from the Si₃N₄ membrane side to the Y₂O₃:Eu³⁺ phosphor film side for CL excitations. The CL emitted from the phosphor films was detected by a photomultiplier tube (R7400-U20, Hamamatsu Photonics K. K.) The images were reconstructed from the signals detected during a raster scan of the electron beams using a computer. The spatial resolutions of the present EXA microscope were characterized using gold nanoparticles with diameters of 100 ± 8 nm (EMGC100, Funakoshi Co. Ltd.) dispersed on the Y₂O₃:Eu³⁺ films as specimens.

3. Results and discussion

Figure 2 shows the CL spectrum of the Y₂O₃:Eu³⁺ phosphor films prepared by electron beam evaporation. Their thickness was 200 nm, and they were annealed at 1000°C. The films were excited by electron beams accelerated at voltages of 10 kV. The CL spectrum has a sharp peak at 611 nm. Its shape is in good agreement with that of the CL spectrum due to the ⁵D₀–⁷F₁ transition in previous studies [17, 18].

 

Fig. 2 Cathode luminescence spectrum of 200-nm-thick Y₂O₃: Eu³⁺ phosphor film on Si₃N₄ membrane.

Download Full Size | PDF

The luminescence quantum efficiency of the as-prepared films is fairly low because of the poor crystallinity due to the disorder of the Y₂O₃ crystal lattices. A post-annealing procedure was done to improve the crystallinity and increase the luminescence quantum efficiency. Figure 3(a) shows the CL intensities of 200-nm-thick phosphor films that were annealed at different temperatures. Here, the measurements were performed under exposures to the electron beams with 10-kV-kinetic energies. We cannot see any clear improvements in the CL quantum efficiency of the materials annealed at 500°C. The materials annealed at higher temperatures exhibited clear improvement, and brighter CL emissions. The post-annealing procedure at higher temperatures is more effective for improving the crystallinity of the Y₂O₃:Eu³⁺ phosphor films.

 

Fig. 3 (a) Cathode luminescence intensities from Y₂O₃:Eu³⁺ phosphor films on Si₃N₄ membrane annealed at different temperatures. (b) X-ray diffraction patterns of as-deposited and post-annealed Y₂O₃:Eu³⁺ phosphor films on Si₃N₄ membranes.

Download Full Size | PDF

Figure 3(b) shows the XRD patterns of the Y₂O₃:Eu³⁺ phosphor films annealed at 500 and 1000°C. The XRD patterns of the as-deposited films are also shown as a reference. The diffraction patterns are in a good agreement with the body centered cubic structure of Y2O3:Eu³⁺ reported previously [19,20]. Diffraction peaks appear at 2$\theta$ = 29.2, 33.8, and 48.5°, corresponding to the (222), (400), and (440) planes, respectively. All the peak intensities were higher at elevated annealing temperatures. The results also indicate that the post-annealing procedure at higher temperatures is more useful for improving the crystallinity of the phosphor thin films.

Figure 4(a) shows the CL intensities of Y₂O₃:Eu³⁺ phosphor films of different thicknesses. All of the films were annealed at 1000°C. The measurements were made with electron beams accelerated at a voltage of 10 kV. The 80-nm-thick films emitted almost no CL lights. The films with thickness greater than 80 nm emitted a significant amount of CL lights, and the brightness increased with increasing film thickness.

 

Fig. 4 (a) Cathode luminescence intensity from post-annealed Y₂O₃:Eu³⁺ phosphor films of different film thickness on Si₃N₄ membrane. (b)–(e) AFM images of as-deposited and 1000°C -annealed Y₂O₃:Eu³⁺ phosphor films on Si₃N₄ before and after post-annealing procedure. Images (b) and (c) correspond to the data for the as-deposited and annealed films 200 nm in thickness, respectively. The images (d) and (e) correspond to the as-deposited and annealed films 85 nm in thickness, respectively.

Download Full Size | PDF

Figures 4(b)–4(e) shows AFM images of the Y₂O₃:Eu³⁺ phosphor films with thicknesses of 80 and 200 nm before and after the post-annealing procedures. In the as–deposited condition, the 200-nm-thick film consisted of aggregates of nanoparticles with average diameter of 20 ± 5 nm (Fig. 4(b)). The particles grew in size during the annealing procedure and the average particle diameter became 70 ± 5 nm after annealing (Fig. 4(c)). The neighboring particles interdiffused during the annealing procedure, causing particle growths. The disorders in the crystalline lattice probably relaxed during particle growths, resulting in the improved crystallinity.

The 80-nm-thick films consisted of nanoparticles with average diameter of 30 ± 5 nm in the as-prepared conditions (Fig. 4(d)). The morphology of the particles changed to the smooth grain structures during the post-annealing procedures (Fig. 4(e)). Although the particles interdiffused during annealing, the CL quantum efficiency was not improved significantly. The unexpectedly small CL emissions may be related to interdiffusions between the Y₂O₃:Eu³⁺ and the Si₃N₄ membranes; Si atoms or N atoms were partly substituted in the Y₂O₃ lattice or stepped into the interstices of the Y₂O₃ lattices. The substituted or inserted atoms may prevent improvement of the Y₂O₃ lattice.

The CL intensities of the 1000°C-annealed, 200-nm-thick films were measured by irradiation with electron beams accelerated with different voltages (Fig. 5). The electron beams were applied from the Si₃N₄ membrane side to the Y₂O₃:Eu³⁺ film side. The data for the as-deposited Y₂O₃:Eu³⁺ film and pristine Si₃N₄ membrane are also shown as references. The peak intensity of the annealed films was 12 times higher than that of the as-deposited one and 900 times higher than that of the pristine Si₃N₄ films.

 

Fig. 5 Cathode luminescence intensity of 200 nm-thick 1000°C-annealed Y₂O₃:Eu³⁺ phosphor films on Si₃N₄ membrane with as a function of electron beam accelerating voltage (open circles). Data for as-deposited Y₂O₃:Eu³⁺ phosphor films (open triangles) and pristine Si₃N₄ membrane films (filled circles) are also shown as references.

Download Full Size | PDF

The phosphor films emitted almost no CL lights under exposure to electron beams accelerated by voltages smaller than 2 kV. The kinetic energies of the electron beams were too low for transmission through the Si₃N₄ membrane. All the kinetic energy was consumed by inelastic scattering inside the Si₃N₄ membrane, and the electrons could not reach the Y₂O₃:Eu³⁺ film layers. On the other hand, the CL intensities dropped off at kinetic energies higher than those at the 6-kV accelerating voltage. In the higher kinetic energy regimes, inelastic scattering occurred less frequently, and most of the electrons passed through the phosphor films without colliding with the crystalline lattices and creating electron–hole pairs.

Finally, the performance of the EXA microscope with the present Y₂O₃:Eu³⁺ phosphor film was examined. Figure 6(a) shows the EXA microscope image of gold nanoparticles with average diameter of 100 ± 8 nm dispersed on the Y₂O₃:Eu³⁺ phosphor film. Electron beams with a 6-pA current were accelerated using 20-kV-voltages and irradiated from the Si₃N₄ membrane side. Figure 6(b) shows a SEM image that was obtained simultaneously with the EXA image in Fig. 6(a). Here, the SEM image was constructed by detecting the electrons from the phosphor films or the gold nanoparticles reflected through the Si₃N₄ membrane. The acquisition time of EXA image was 16 min at 1024 × 1024 pixels.

 

Fig. 6 (a) EXA microscope images of gold nanoparticles scattered on Y₂O₃:Eu³⁺ phosphor films. (b) SEM images of area in (a). (c) Intensity distribution on solid line in (a).

Download Full Size | PDF

The features of the gold nanoparticles appeared at the same positions in the EXA and SEM images. The intensity distribution on the solid line in Fig. 6(a) is shown in Fig. 6(c). The gold nanoparticles were clearly spatially resolved. The profiles were reproduced with a Gaussian- function with a least-square method. The FWHM of the dip structure was 94 ± 14 nm, while the amplitude error was 15%. Hence, we conclude that the present EXA microscope can observe specimens with the spatial resolution better than 100 nm. The contrast of the image was calculated with Eq. (1).

4. Conclusions

This paper presented a EXA optical microscope containing Y₂O₃:Eu³⁺ phosphor thin films. The microscope’s nanometric illumination light source was the red CL light spots emitted by exposure to high-energy focused electron beams. The CL quantum efficiencies were examined in terms of the post-annealing temperature and film thickness. Gold nanoparticles with average diameters of 100 nm were spatially resolved by the EXA microscope with the phosphor films. Rare-earth-doped Y₂O₃ phosphor materials is known to emit luminescence of different wavelengths depending on the type of rare-earth dopant; for example, Tm- and Tb-doped Y₂O₃ materials emit blue (~450 nm) and green (~520 nm) luminescence lights, respectively [21]. We are preparing Y₂O₃ phosphor films doped with different rare-earth elements. The performance of the modified EXA microscope will be reported elsewhere in the future.