Practical short-wavelength lasers in the vacuum-UV (VUV) spectral region are in high demand in fields such as photolithography and photochemical science because of their fine spatial resolution and ease in triggering reactions. In addition, because VUV emissions do not exist in the atmosphere, they may lead to interesting new phenomena.1,2 However, few laser media can access this spectral region. Indeed, gas lasers still dominate, especially in the deep VUV (<160nm), where they produce emissions directly. Solid-state laser media, on the other hand, must employ nonlinear wavelength conversion, with a concomitant lowering of efficiency and output. Development of gaseous lasers is further encouraged by the transparent optics at wavelengths down to 105nm. However, the availability of such lasers is limited by our lack of knowledge about their operation. Among existing gaseous lasers, discharge-pumped argon fluoride (ArF) (193nm) and fluorine (F2) (157nm) enjoy the most widespread use.
Figure 1. Laser intensity dependence of the harmonic (integer frequency multiple) intensity (I) and ion signal. Xe: Xenon.
We have been investigating efficient emission sources using rare-gas excimer molecules, such as Ar2* at 126nm and Kr2* (krypton) at 147nm. These deep-VUV lasers are assumed to have unique characteristics in terms of laser-material interactions. The emission wavelength is so short that the penetration depth into materials is also short, usually in the nanometer range. For most materials, the VUV emission energy should thus be absorbed in a very shallow nanorange-thin layer. We have realized nanorange surface alteration of transparent materials such as silicon dioxide (SiO2) using Ar2* excimer lamp emission at 126nm.1 High-quality silicon nitride thin films were produced by irradiating Ar2* excimer emission in silane and ammonia.2
Use of coherent VUV lasers could add to these kinds of nanoscale surface interactions with the submicron spatial resolution attributed to the emission wavelengths. Moreover, because the pulse width of the laser emission is shorter than the specific energy-diffusion time, subpicosecond pulses could be used to regulate the energy diffusion on the material surfaces, achieving very fine processing resolutions of order 100nm.
Accordingly, we are currently constructing a subpicosecond VUV laser system for advanced nanoscale-materials processing. Figure 1 shows the laser-intensity dependence of the seventh-harmonic (i.e., 1/7th of the fundamental laser wavelength) emission intensity, as well as the intensity of the ion signal. The seventh harmonic was generated from a subpicosecond titanium:sapphire (Ti:S) laser at a wavelength of 882nm, resulting in subpicosecond VUV pulses at 126nm. The emission was optimized in low-pressure xenon as a nonlinear medium. The dependence of the seventh-harmonic intensity and other characteristics obeyed classical nonlinear optics.3 The output was limited by ionization, which decreased the number density of the nonlinear medium and deteriorated the phase matching. We treated the short-pulse VUV emission as a seed pulse for the optical-field-induced ionization (OFI) VUV Ar2* (argon dimer) amplifier.
Figure 2. Subpicosecond vacuum UV (VUV) laser system. Ti:S: Titanium:sapphire. CPA: Chirped-pulse amplification. T: Transmission. R: Reflectivity. OFI: Optical-field-induced ionization. Ar2*: Argon dimer. MgF2: Magnesium fluoride. f: Frequency. τ: Pulse length. E: Maximum output energy. λ: Laser wavelength.
The subpicosecond VUV laser system (see Figure 2) uses one high-intensity Ti:S laser to drive both the OFI Ar2* excimer VUV amplifier4 and the subpicosecond VUV seed-pulse generator. The system's maximum output energy was 1mJ at 882nm with a pulse width of 170fs at a repetition rate of 1kHz. The output from the driving laser was split into two beams. One beam was focused inside a VUV amplifier filled with high-pressure argon, producing the OFI Ar2* excimers. The other beam was focused collinearly from the opposite side into the seed-pulse generator to generate a subpicosecond seed pulse at 126nm. This pulse was injected into the gain region of the OFI Ar2* VUV excimer amplifier. A magnesium fluoride lens was manipulated inside high-pressure argon to control the spatial overlap between the seed pulse and the gain region. The optical path difference of the two pump beams provided a temporal delay. The amplified subpicosecond VUV pulse was separated from the driving laser pulse by a dichroic mirror. As a preliminary result, we observed a 40-fold increase of the seed-beam intensity. Simulations predict that the output energy of the VUV seed pulse will be amplified on the order of 1μJ, leading to an average VUV power of 1mW at a repetition rate of 1kHz. Our next steps will focus on understanding the temporal and spatial characteristics of the amplified beam.
In summary, we have developed a subpicosecond VUV laser system in which a single high-intensity laser produces the optimized subpicosecond harmonic seed pulse and the OFI plasma for the Ar2* amplifier at 126 nm. We are applying this unique short-wavelength short-pulse emission to studies of 3D nanoscale surface processing and time-resolved photochemical reactions.
The author acknowledges many collaborators, in particular M. Kaku and M. Katto. Part of this work was supported by the Grant-in-Aid for Scientific Research (B) Program and Advanced Research Driving Program of MEXT (Ministry of Education, Culture, Sports, Science, and Technology), Japan, and by Hamamatsu Photonics K. K., Japan.
Department of Electrical and Electronic Engineering
University of Miyazaki
Shoichi Kubodera is a professor. He received a BS from Keio University, Japan, and his PhD from Rice University, TX. He spent several years at the RIKEN Institute, Japan, as a postdoctoral fellow. His research activities have focused on short-wavelength lasers, laser-plasma interactions, and high-field laser physics.