Vacuum ultraviolet field emission lamp utilizing KMgF₃ thin film phosphor

Due to the wide range of existing and potential applications such as surface treatment, optical cleaning of semiconductor substrates, and sterilization considerable attention has been focused on the development of UV lamps, not only in conventional physics but also in industry,^1,2 and there has been continuous research on lamps operating in the deep and vacuum ultraviolet (VUV) regions. However, in VUV region, there are only a few light sources, such as low-pressure mercury lamps, excimer lamps, and deuterium lamps.^2,4 These light sources have disadvantages in terms of lifetime and stability because of utilizing a gas-state phosphor. In addition, mercury and excimer lamps use toxic gases that could pollute the environment and harm people.

To overcome these drawbacks, a solid-state phosphor is required. The solid-state phosphor has advantages, such as reducing the device size and design flexibility. These merits enable new industrial applications, which may be used as substitutes for conventional light sources.

Oxide-based II–VI and nitride-based III–V compound semiconductors, which can be used for solid-state phosphors, have been well studied. However, even when using AlN and BN, which have been extensively studied as UV phosphors, the operating wavelength has not been extended to the VUV region.^5–8 Oxide-based compounds, such as BeO and MgZnO, have wide band gaps, and several UV emitting devise have been reported by applying MgZnO.^9–13 However, BeO is difficult to handle because of its toxicity, and MgZnO is difficult to grow because of different structures of ZnO and MgO. MgO has been reported to demonstrate fluorescence in the VUV region; however, its luminosity was extremely low in our experiment.¹⁴ 

Fluoride is a promising VUV phosphor candidate. Fluoride composite materials have been extensively studied as laser materials, scintillation materials, and optical materials because of its extremely wide band gap.^15–17 Until recently, single crystals of MgF₂, CaF₂, and LiCaAlF₆ have been used as windows, lenses, and prisms in the VUV region.^3,18,19 Scintillation and laser oscillation based on Ce³⁺, Eu²⁺, and Nd³⁺ ion-doped fluorides have been reported.^20–24 Nd³⁺ ion-doped LuF₃, LaF₃, and LiLuF₄ have been reported as VUV phosphors.^25,26 In addition, KF binary compounds, such as KMgF₃ and KCaF₃, have been studied as scintillators for VUV fluorescence.²⁷ 

We have applied KMgF₃ as a VUV solid-state phosphor and have grown a KMgF₃ thin film for device fabrication. KMgF₃ has a fluorescence range of 140–200 nm and is a phosphor with high conversion efficiency in the VUV region. KMgF₃ has a cubic structure and does not require rare-earth ion doping; therefore, it is relatively easy to grow.²⁸ 

For the growth of the fluoride thin film, supply of fluorine by the atmosphere control is difficult. The harmful F₂ gas could pollute the human body and corrode a vacuum chamber. CF₄ gas is also difficult to handle and may cause the problem of contamination by carbon. In this experiment, we grew the thin film in vacuum atmosphere by applying pulsed laser deposition (PLD). This preparation method has a prominent advantage that there are few composition gaps between a target and a thin film. Thus, we can obtain the fluoride thin film without inflow of toxic gasses.^29,30 PLD also has other features such as vaporize the target with high melting points, not be polluted by a melting pot and pulsingly controllable process.³¹ Therefore, we can apply PLD to KMgF₃ which has high melting point of 1070 °C.

The substrate was a (001)-oriented MgF₂ crystal mounted on a rotating holder, and the target was fused KMgF₃ (KF:MgF₂ = 1:1). It was melted under Ar:CF₄ (95:5) atmosphere at 1220 °C. After melting, it took 1 h to cool to 900 °C and an additional 48 h to cool to room temperature. The thin film deposition was accomplished by irradiating the target with a focused beam from the third harmonic (355 nm) of a Nd:YAG laser operating at a repetition rate of 10 Hz. The laser spot size focused on the KMgF₃ target was 1 mm in diameter with a laser fluence of 4.2 J/cm². The deposition was performed under a base pressure of 2 × 10⁻⁴ Pa for 4 h. MgF₂ was selected as a substrate because of its high transmittance in the VUV region. The substrate was placed parallel to the target surface at a distance of 50 mm. The temperature of the substrate was controlled at 400 °C.

The surface morphology and a cross-sectional view of the thin film were observed using the scanning electron microscope (SEM), as shown in Fig. 1. The surface of the film contained an irregular array of droplets (Fig. 1(a)), and the particles were 10–500 nm in diameter. Figure 1(b) shows that the thickness of the thin film was approximately 300 nm.

The cathodoluminescence (CL) spectrum was obtained using a spectrometer attached to the SEM. Figure 2(a) shows the CL spectra of the KMgF₃ thin film. The spectra were observed from approximately 140 to 220 nm. VUV fluorescence (crossluminescence) was observed with two peaks at approximately 155 nm and 180 nm. These peaks are attributed to transitions from the valence anion band to the outermost core cation band and correspond well with emission peaks previously reported from a single KMgF₃ crystal.^22,32 Therefore, we considered that the KMgF₃ thin film was successfully grown by PLD with few chemical composition differences.

Figure 2(b) shows the dependence of emission intensity on acceleration voltage for the KMgF₃ thin film and single crystal. Intensity was estimated from the dominant emission peak at 180 nm. The acceleration voltage was changed from 1 to 10 kV while the current value was approximately 6 × 10⁻⁶ A/m². The open and closed squares in Fig. 2(b) indicate the emission intensity at 180 nm of the single crystal and the thin film, respectively. Below an acceleration voltage of 5 kV, the intensity of the dominant peak of the thin film increased with higher acceleration voltage. The performance of the thin film was comparable with that of the single crystal under low acceleration voltage even though its morphological properties were inferior, as we have described above. Emission from the thin film was saturated at acceleration voltage above 6 kV. This was because of increased electronic penetration beyond the film thickness. These results indicate that we obtained a thin film of sufficient quality in comparison to the single crystal. It should be noted that it may be possible to increase the emission intensity by increasing the film thickness. In addition, even when we increased acceleration voltage up to 20 kV and repeatedly perform such experiment, the same results were obtained without any degradation. Therefore, this thin film has enough strength as a phosphor of an electron-pumping device.

The crystallographic orientation of the thin film was also investigated by X-ray diffraction. The diffraction pattern showed four peaks as shown in Fig. 3. Three peaks at 30°, 39°, and 45° correspond to the (110), (111), and (200) planes of the KMgF₃ thin film, respectively. The full widths at half maximum of every peak are approximately 0.2°. The peaks appeared at 61° originated from the MgF₂ substrate.

We fabricated a VUV field emission lamp consisting of a KMgF₃ thin film as the phosphor and carbon nanofibres (CNFs) as the field electron emitter. Unlike a powder phosphor, a binder is unnecessary when using a thin film as a phosphor because of the film's strong adhesive properties. In addition, excellent heat conduction is expected. By employing a field electron emitter as a cold cathode, a heat-free lamp with low power consumption should be obtained.

Figure 4 shows a schematic diagram of the lamp. The lamp was constructed by stacking CNFs, two mesh electrodes, two spacers, and the KMgF₃ thin film on an aluminium sheet. The CNFs were grown on a glassy carbon substrate by bombardment with Ar⁺ ions at room temperature.^33–37 The surface of carbon plate was polished and irradiated with ion beam at an incidence angle of 45° from the normal to the surface for 45 min. The diameter and energy of the ion beam were 6 cm and 1 keV, respectively. Briefly, sputter-ejected carbon atoms from the carbon surface were redeposited onto the sidewall of the ioninduced conical surface protrusions. Next, the redeposited carbon atoms diffused toward the cone tip to form the CNFs without any catalyst even at room temperature. The system was evacuated by rotary pump and turbo molecular pump. The basal and working pressures were 1.5 × 10⁻⁴ Pa and 5.0 × 10⁻² Pa, respectively. The density, length, and diameter of the CNFs were approximately 5 × 10⁸ cm⁻², 0.3–2 μm, and 20 nm, respectively. The current density at a field emission threshold of 3.8 V μm⁻¹ was 5 mA m⁻². The electrodes were constructed of copper, with a mesh width of 0.1 mm. The spacers were constructed of Teflon, and their thickness for extraction and acceleration was 300 μm and 1 mm, respectively. By applying extraction voltage, electrons were released from the tip of CNFs. The electrons, which excited the KMgF₃ thin film, were accelerated by applying acceleration voltage. VUV fluorescence from the thin film was emitted through the MgF₂ substrate. The working pressure in the lamp was approximately 9 × 10⁻⁵ Pa.

Figure 5 shows the output spectra of the lamp. The emission peak corresponds approximately to the CL spectrum of the thin film described above. In the spectra, the peak intensity at 180 nm was approximately equal to that at 155 nm because of calibration of the spectrometer. The calibration of the spectrometer in the VUV region was difficult because we could not obtain a standard sample. This is the shortest operation wavelength for a lamp employing a solid-state phosphor.

The output power of the lamp was investigated using a photomultiplier tube (Hamamatus R976), as is shown in Fig. 6. As the acceleration voltage increased, the output power increased nonlinearly. This dependence is attributed to an increase in the interaction volume of electrons in the KMgF₃ thin film. The output power can be more effectively enhanced by applying higher acceleration voltage. The output of the plane emission was obtained at a range of 64 mm². Current was estimated at approximately 0.32 μA. It should be noted that the VUV power reached to 2 μW at an extraction voltage of 800 V and acceleration voltage of 1800 V. The output power is closely related to the design of the lamp and structural properties of the CNFs. For example, the emission area was not effectively used because of the non-uniform length of the CNFs; thus, higher output power should be obtained by further improvement of the CNFs. Additionally, there is the optimal value of mesh width of the electrodes according to density of the CNFs.

In summary, we have successfully demonstrated a VUV emitting lamp by employing a KMgF₃ thin film as a phosphor and CNFs as a field electron emitter. The thin film was prepared by PLD, and its fluorescence performance was nearly equal to that of a single crystal. A CNF emitter can be grown at room temperature and on flexible materials. A VUV output power of 2 μW was obtained at the extraction voltage of 800 V and the acceleration voltage of 1800 V. Furthermore, it was observed that the emission intensity effectively increased with acceleration voltage. Unlike deuterium lamps, a lamp constructed with a CNF field electron emitter does not need to block thermal radiation. In addition, a solid-state phosphor brings many advantages, such as stability, safety, and longevity compared with gas-state phosphor. In study of field emission display, large-sized screen is well-established technology. By applying this technology, therefore, our lamp is also expected to obtain the larger fluorescence area. Such VUV light is a powerful tool for sensing, material processing, and decomposition of chemical materials.

This research was partially supported by a Grant-in-Aid for Scientific Research C (40370126), from the Japan Society for the Promotion of Science (JSPS), Nippon Sheet Glass Foundation, and Nanotechnology Platform Program (Molecule and Material Synthesis) of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. The authors would like to thank Professor Y. Ichikawa of the Nagoya Institute of Technology (NIT) for invaluable discussions. The authors are grateful to Dr. S. Nakao of the Institute for Molecular Science (IMS) for important technical support in this research. The authors would like to thank Enago (http://www.enago.jp) for the English language review.