NEXT month Japan will launch a major programme to become a world leader in the science of the split second. Because the government believes that the ability to control events at the femtosecond scale will have far-reaching implications for the communications and computer industries, it is prepared to see its Ministry of International Trade and Industry (MITI) spend more than 拢120 million (20 billion yen) on the project. And in the usual Japanese style, the project is a long-distance effort lasting some ten years and probably involving more than a dozen companies, including such major players as Fujitsu, Hitachi and NEC.
So what exactly is the excitement about? The answer lies in the femtosecond itself. It is vanishingly small, one millionth of a billionth of a second (10鈭�15 seconds). It is the time it takes an atomic bond to break during a reaction and the period of vibration of a single atom in a lattice. If time were slowed down so that such events lasted a second, the world record for the hundred metres sprint 鈥� a touch under 10 seconds 鈥� would be equivalent to 320 million years. Until recently, it was impossible even to time such events. Now scientists say they can not only measure these events, but control them with femtosecond pulses of laser light.
Most of the work on femtosecond phenomena has been academic research carried out at European and American universities. Now Japan wants to join in.
Advertisement
In February, the Agency of Industrial Science and Technology (AIST), the branch of MITI which is overseeing the project, outlined its research programme at an international workshop on ultrafast science and technology in Tsukuba Science City, some 60 kilometres north of Tokyo. The programme will focus on four areas. One group will work on ultrafast lasers, another team will build electronics devices that can switch on and off at femtosecond speeds. A third will develop materials that can convert rapid pulses of light into rapidly oscillating electric fields and vice versa. These optoelectronic materials could be used to measure short pulses of light or the tiny electric fields in microchips as they operate. And a fourth team will apply these ideas to medical imaging techniques such as scanning systems that can measure the electromagnetic signals associated with brain function.
But Japan has some catching up to do. Femtosecond technology was born in 1982, when a group led by Charles Shank at AT&T Bell Labs in New Jersey demonstrated a simple way of producing pulses lasting 100 femtoseconds from a dye laser. In effect, Shank鈥檚 group had created an ultrafast stroboscope that could be used to freeze the motion of femtosecond phenomena, such as the reactions between individual atoms and molecules, in the same way that a camera flash can freeze the motion of a speeding bullet.
Ahmed Zewail, a chemist at the California Institute of Technology in Pasadena, pioneered the use of femtosecond lasers to look at the formation and breaking of chemical bonds. His method is to fire two pulses of light at an isolated molecule in quick succession. The first pulse splits the molecule into its constituents. A few femtoseconds later, the second pulse is absorbed by these parts and then re-emitted as a characteristic spectrum of light that can be analysed to work out which by-products are present.
Encouraged by their ability to watch molecules at work, chemists now think that they may also be able to control the way molecules react with each other. Many chemical reactions involve a large number of intermediate steps, each of which produces a different by-product in different quantities. Most of the time chemists are only interested in the by-product from one of these steps, which may be produced in only small quantities. According to Philippe Fauchet, an electrical engineer who leads a group working on ultrafast phenomena at Rochester University in New York State, it may be possible to use a laser to control the appropriate step in the reaction so as to force it to produce more of a particular by-product. He says that the method could even be used to create entirely new compounds and these ideas have generated keen interest in femtochemistry over the past five years.
This interest has grown with the invention of cheaper, smaller and easier to run lasers that can produce femtosecond pulses. In 1991, groups at the Massachusetts Institute of Technology and the University of St Andrews showed that solid-state lasers made from titanium-doped sapphire could be used to produce femtosecond pulses instead of the complicated dye lasers that require highly trained staff to operate them. According to Fauchet, the result is that scientists no longer need to be laser experts to carry out femtosecond research.
Several firms now sell reliable solid-state femtosecond lasers that cost $100 000. But even this is too much for many laboratories. Roy Taylor, who heads the Femtosecond Optics Group at Imperial College, London, is developing lasers that will cost only $10 000 and fit on a small trolley rather than the large table-top that current models require (see Technology, 15 January 1994). 鈥淲e really believe that the widespread availability of low-cost lasers will revolutionise femtosecond research by giving more scientists the chance to enter the field,鈥� Taylor told scientists at the workshop.
In Japan, femtotechnology is already being used to test semiconductor devices. The technique is called electro-optic sampling and relies on a transparent electro-optic crystal that changes the polarisation of light passing through it according to the voltage applied to it. The crystal does not even need to be in contact with the device because it is sensitive enough to detect the tiny fields generated close to the surface of semiconductors. By measuring the polarisation of femtosecond pulses before and after they pass through the crystal, researchers can assess the currents in the semiconductors over very short periods of time. 鈥淵ou can reconstruct in real time where these currents go, how a device switches, and how quickly it switches,鈥� explains Fauchet. He points out that this information is impossible to obtain from conventional methods of testing electronic devices, which measure the signals with tiny electrodes that must be in contact with the circuits.
The technique should save time and money. Chips are manufactured in batches on circular wafers of silicon or gallium arsenide. At the moment, high frequency chips can only be tested individually after the wafer has been cut up and the chips mounted in packages 鈥� an expensive and time-consuming process. Electro-optical sampling, on the other hand, could be applied to the wafer itself, allowing dud chips to be identified and discarded at an earlier stage. Hamamatsu Photonics, a Japanese firm based in Hamamatsu, which specialises in making measuring equipment, has already developed an electro-optical sampler that operates with femtosecond pulses and is capable of testing integrated circuits.
According to Yutaka Tsuchiya, the firm鈥檚 deputy director of research, Hamamatsu is ready to sell the device to anyone prepared to pay the price tag of several millions of dollars.
Many of the other technologies will only produce commercial benefits in the distant future. This is why Western researchers find it hard to get funding for femtosecond technology. 鈥淎lthough it has great potential, no companies are willing to invest in this sort of research in the US,鈥� laments Chi Hsiang Lee, an electrical engineer at the University of Maryland in College Park, who has received most of his funding for femtosecond research from the US government.
In Britain, the government鈥檚 requirement that research be 鈥渨ealth creating鈥� makes it difficult for Roy Taylor鈥檚 group at Imperial College to get public support. And, as for the private sector, 鈥測ou can鈥檛 get anybody in the UK interested鈥�, he says. Taylor has had to look abroad for funding.
This indifference leads the Japanese to believe they can play an important role in the femtosecond field. At the moment, however, even Japan鈥檚 best labs lag way behind their Western counterparts. 鈥淣obody in this country can compete,鈥� says Takashi Yagi, who runs the laser laboratory at the AIST鈥檚 Institute of Research and Innovation in Tsukuba. The shortest pulses that Yagi has been able to generate are 20 femtoseconds, a figure achieved in the US four years ago. The current record, held by Bell Laboratories, is 6 femtoseconds.
Japan has some frontrunners on the equipment side, however, most notably Hamamatsu. For several years, the company has been lobbying the Japanese government to put serious money into ultrafast phenomena research. 鈥淲e need big breakthroughs,鈥� says Tsuchiya, in a wide range of basic fields such as optics, electrical engineering and materials science. He points out that being able to generate 6 femtosecond pulses is all very well, but such pulses are of no practical use because other parts of the system, such as the specialised measuring equipment, still lag behind.
The kind of studies Tsuchiya has in mind are a far cry from the work MITI enthuses about. To convince Japan鈥檚 finance ministry that the project was worth funding, MITI officials are reported to have argued that a femtosecond switch would be necessary to deliver high-definition video-on-demand to Japanese homes. 鈥淚t鈥檚 easier to get money if you tell them that there are wide applications in information processing,鈥� confesses Yagi 鈥淚t鈥檚 sad, but that鈥檚 how it is.鈥�