午夜福利1000集合

A PURE silicon laser has been demonstrated for the first time, offering the prospect of 鈥減hotonic鈥� silicon chips that communicate using light, and operate faster than electrical chips. No previous laser has inexpensively produced the elusive frequency band generated by the silicon laser, and the technology could soon be used in defence and communications.

Researchers have long tried to get silicon to emit light, and have had some success by doping the material with other light-emitting elements. But until now they have not been able to produce laser light, in which all the photons are in phase, or coherent. The feat was finally achieved by 鈥減umping鈥� a sliver of silicon with another laser.

Standard semiconductor lasers and LEDs emit light when the voltage across them causes electrons to jump between 鈥渉oles鈥� in the semiconductor鈥檚 crystal lattice. In doing so, an electron has to jump the material鈥檚 inherent energy level 鈥渂andgap鈥�, releasing energy as it does so. Gallium arsenide and related compounds are the most popular materials for this task in LEDs, as their bandgap ensures the energy is released as light. But silicon鈥檚 bandgap allows electrons to settle at a halfway-house energy level, with the result that silicon tends to dissipate recombination energy as heat instead of light.

Researchers have already made silicon emit some light by doping it with elements such as erbium (New Scientist, 16 November 2002, p 15), or etching microstructures in the crystal to encourage electrons to jump the full bandgap. However, they have not been able to get silicon to produce strong coherent laser light, which involves exciting a large number of atoms in the crystal to high energy levels so that they emit light when they return to lower energy levels.

Now Bahram Jalali of the University of California at Los Angeles has found a way to excite the silicon atoms using what is know as the Raman effect. When light passes through a transparent material, atoms absorb and re-emit the photons, usually at exactly the same energy (and therefore wavelength) as the ones they absorb. However, some photons donate energy to atoms in the crystal lattice, and this energy loss means the re-emitted photons are 鈥淩aman-shifted鈥� to a longer wavelength.

The Raman effect is weak for low-intensity light, but a strong light source like a laser can pour so many photons into the material that the Raman excitation pushes many atoms to a high-energy state. In the latest edition of Optics Express (vol 12, p 5269), Jalali reports that he has done exactly that to stimulate coherent silicon laser light. He fired pulses of near-infrared light with a wavelength of 1540 nanometres from a cheap solid-state erbium laser into a narrow strip of silicon 2 centimetres long. The silicon duly emitted Raman-shifted laser light in the mid infrared at 1675 nanometres 鈥� a wavelength that is hard to generate without cryogenically cooled lasers.

By pumping the silicon with different frequencies, Jalali could force the laser light further into the mid infrared, up to around 5000 nanometres. Inexpensive lasers in this band could be used for applications that have been impractical until now, Jalali says.

One application the Pentagon is considering is affordable anti-missile systems for civil aircraft. The silicon laser could blind the guidance sensors on heat-seeking ground-to-air missiles, which home in on hot jet exhaust by looking for bright spots in the 3000 to 5000-nanometre range. Modulating the laser beam in certain ways 鈥� the details are classified 鈥� can fool a missile鈥檚 guidance system into thinking the light source is moving, drawing the missile away from the plane.