Extreme ultraviolet (EUV) tends to be associated with lithography for semiconductor chips. But photonics in the extreme ultraviolet has a much more diverse and interesting range of applications than chip lithography. This is, perhaps, just as well because no microchip has yet gone into production using this technology.
Scientists are already looking to extend the wavelength range down through what is called the ‘water window’ – wavelengths between 2nm and 4nm – towards the soft x-ray region. At these sorts of wavelengths, researchers are finding new applications for EUV, also known as XUV, from cutting operations, to spectroscopy, imaging of biological samples, to data storage.
In developing novel materials for use in ultra-high speed data storage, scientists need to be able to analyse what is going on inside the materials. The ultra high-speed data storage would use the laser to switch electrons’ state between what would represent a one and what would represent a zero. Extreme UV is of interest as a research tool, because it can examine events occurring at very short timescales. Professor John Collier, director of the central laser facility at UK’s Science and Technology Facilities Council (STFC), points out that the beams are ‘getting down into a few nanometres in wavelength space and it is also extremely short, just a few tens of femtoseconds long. That then goes into an experimental station and we can use that pulse of light as a way of diagnosing what is going on inside materials.’
‘EUV opens the way for new devices and technologies such as ultra high-speed data storage. Imagine you could find a way of switching magnetic systems on a femtosecond timescale. That would have a very big impact on data storage,’ Collier explaines.
Collier’s facility uses high harmonic generation to generate very high frequency (and hence very short wavelength) short-pulse radiation by multiplying up the frequency of an initial, longer wavelength laser. The process starts with a conventional titanium sapphire laser that produces a pulse that is 25 to 30 femtoseconds long and runs at 1kHz with 10mJ pulses. The laser is directed into a noble gas where it kicks the gas’ electrons, out of their nuclear orbits into the vacuum. But half an optical cycle after this ionisation, the electron will reverse direction as the electric field changes, and will accelerate back towards the parent nucleus. Upon returning to the atom, it will emit bremsstrahlung (braking radiation) as this atom returns to its ground state. The result is very short wavelength radiation that the scientists can direct at the samples to be analysed.
For high-resolution imaging of biological samples using EUV sources, a laser-plasma is used in the 2.3nm to 4.4nm range. This range largely covers the ‘water window’, where the radiation passes through water, but is strongly absorbed in carbon-containing materials such as protein. Thus the wavelength range is ideal for making high-contrast images of cells and other biological samples. This region is where the ultraviolet spectrum at its extreme end meets with what are referred to as soft x-rays.
‘Some interesting biological imaging has been done with water-window x-rays produced by a laser plasma. It’s very early days though,’ says professor Greg Tallents. Tallents works in the physics department of the University of York and studies EUV and x-ray laser development.
Tallents explains that free-electron lasers are another source of EUV, and this time it is not for imaging, but for analysing materials. The Deutsches Elektronen-Synchrotron (DESY) near Hamburg, Germany, has a free-electron laser called Free-Electron Laser in Hamburg, or FLASH. It delivers intense ultra-short femtosecond coherent radiation in the wavelength range between 44nm and 4.1nm without recourse to high harmonic generation.
Tallents expects a growing number of applications for EUV: ‘There is going to be an opening up of applications in the EUV spectral region. For example, I think EUV will be useful for cutting solid materials. You don’t heat the material too much and there is high penetration depth into the target, so it is useful if you want to produce deep features.’
Tallents points out that a longer-wavelength laser interacts only with a very thin surface layer of the target material, generating a plasma of the ablated material, with which the laser then interacts. This cloud of material that is evaporating from the lased surface hinders the laser’s ability to get through to the target’s surface. EUV light, however, will go through the plasma cloud and continue altering the surface.
XUV Lasers, a spin-out company from Colorado State Univeristy, is developing applications of these lasers to ablate materials for mass spectroscopy. The two-year old firm is developing compact extreme ultraviolet lasers and optical systems. The product is a desktop-size laser that produces a highly coherent light source at 46.9nm. Its short wavelength is combined with a high energy per pulse, 10µJ, and short pulse duration, 1.5ns. This beam will ablate the surface of a sample and then the plasma will be ionised, stripped of its electrons, for analysis in the mass spectrometer.
In the mid-1990s, professor Jorge Rocca at Colorado State University demonstrated the feasibility of obtaining laser amplification of soft x-ray wavelengths by fast capillary discharges in plasma columns. Subsequent work developed capillary discharge lasers into the tabletop coherent soft x-ray sources with the highest average power presently available. Capillary discharge soft x-ray lasers developed in Colorado have been installed and are used in laboratories in the United States and in Europe.
The president of XUV Lasers is Carmen Menoni, a professor of electrical and computer engineering at Colorado State University’s college of engineering. Menoni says: ‘Our laser is the smallest at this wavelength. Applications include microscopic imaging.’ The technology’s properties have made it possible to carry out tabletop experiments in nanoscale imaging, nanopatterning, single photon ionisation mass spectroscopy, interferometry, dense plasma diagnostics, and nanoscale ablation of materials.
Recently a tiny silver funnel has also been demonstrated to produce EUV. The funnel, only a few micrometres long, is made out of silver and is filled with xenon gas. Like the high harmonic generation used with noble gases in professor Collier’s facility, the funnel’s xenon emits radiation across a spectral range that includes EUV.
Created by scientists from South Korea’s Advanced Institute of Science and Technology, Germany’s Max Planck Institute of Quantum Optics, and the Georgia State University in the USA, the funnel turned incident infrared pulses into EUV pulses that last for a few femtoseconds. The EUV pulses in this case are repeated 75 million times per second. The funnel also acts as a filter. Its 100nm aperture stops 800nm IR light but allows EUV pulses with wavelengths down to 20nm to go through. The scientists expect the new technology will help them measure the movement of electrons with a higher resolution, so they can better understand materials.
High harmonic generation through to silver funnels are all heady stuff. The technology currently being investigated for lithography, while still extreme UV, is at slightly longer wavelengths of up to 30nm. Lithography depends on either laser-produced plasmas or discharge-produced plasma, where the plasma of choice is tin. In a laser-produced lithography process, a laser is fired into droplets of tin to create a tin-plasma.
Ultimately, for lithography, professor Tallents from the Univeristy of York expects systems with wavelengths of less than 13.5nm to be developed. He says: ‘They are talking about half the wavelength of 13.5nm as the next step, after the 13.5nm lithography, which of course isn’t in production yet, but 5.5 to 6.5nm would extend the technique to enable finer features to be produced in lithography.’
Instead of tin or any of the other elements or compounds mentioned, terbium and gadolinium have been studied, because their electron structure is similar to that of tin. Work with gadolinium produced emissions in the 5.5nm to 10nm range.
As for microchip production at the 5.5nm wavelength, Trumpf’s head of laser technology Dr Peter Leibinger is sanguine. Speaking to Electro Optics at the company’s headquarters, he said: ‘There could be [5.5nm], in ten years maybe. Let’s first get 13.5 working, and then we can talk about the next step’. Getting 13.5nm to work is part of Trumpf’s focus right now, as the start of chip production using this wavelength begins next year. Leibinger adds: ‘The first machines are now installed. [Lithography technology manufacturer] ASML has publicised that. They have shipped, I think, six pilot systems to chip manufacturers and we now think that in 2012 the first production systems will go in.’
EUV development has taken longer than expected. Companies such as Jenoptik got out of EUV altogether, selling their stakes in the technology. Leibinger does not understate the scale of the task. He says: ‘It’s an extremely tough project. This is a very early-stage technology [and] we’ve been involved in this development project for eight years now.’
Sources of EUV are expanding. For non lithography applications, alternatives to tin plasmas include methanol and liquid nitrogen. Here a pulsed infrared laser is focused into a jet of methanol or liquid nitrogen to create EUV sources at 3.37nm with methanol and 2.48nm with liquid nitrogen. Lasers employing high harmonic generation, free electron lasers, tin-plasma, methanol and liquid nitrogen plasmas, and now silver funnels of xenon gas, water windows, producing wavelengths between 5nm and 13.5nm – the toolkit is impressive. Researchers ranging from biologists to material scientists are queuing up to take advantage of the possibilities that this developing technology offers.