letters to nature 29. Crone, B et al. Large-scale complementary integrated circuits based on organic Sature loops introduce a local strain field, which modifies the band structure and provides spatial confinement of the charge carriers. 31. verthamer, N. R, Helfand, E. Hohenberg. P. C. Temperature and purity dependence of the It is this spatial confinement which allows room-temperature superconducting field, Hez, Ill. Eledron spin and spin-orbit effects. Phys Rev. 147, 295-303 electroluminescence at the band-edge. This device strategy is highly compatible with ULSI technology, as boron ion implanta tion is already used as a standard method for the fabrication of silicon devices We thank E A Chandross, B Crone, H E. Katz, H Y Hwang, A J Lovinger and T Siegrist There have been great efforts over the past decade to obtain for discussions, and E Bucher for the use of equipment. useful, that is, technologically viable and efficient, light emission Correspondence and requests for materials should be addressed to L.H.S from silicon both in the visible and infrared regions of the spectrum. (e-mail:hendrik@lucent.com In the visible regions, porous silicon and other quantized systems, such as silicon/silicon dioxide superlattices and silicon nanopreci- pitates in silicon dioxide, have been the main emphasis. In the infrared region, systems such as erbium in silicon,, silicon/ and, more recently, iron disilicide offer potential routes. No approach has so far been applied commercially. The An efficient room-temperature reasons for this are a combination of the lack of genuine or silicon-based light-emitting diode in the case of infrared emitters, high thermal quenching giving very Wai Lek Ng, M. A Lourenco, R. M. Gwilliam*, S. Ledaint, G Shot Here we use a new approach-dislocation engineering, using K P. Homewood* conventional ULSI technology-that gives efficient light emission in silicon at room temperature. School of Electronic Engineering, Information Technology e- mathemati Because of its indirect bandgap, silicon is fundamentally a poor t School of Mechanical and Materials Engineering, University of Surrey, Guildford, Surrey, GU2 7XH, UK emitter of light. The main reason for this is that fast non-radiative recombination routes dominate the slower radiative route in this material. Indeed, in bulk silicon, at room temperature, radiative There is an urgent requirement for an optical emitter that is emission is normally entirely absent. However, if recombination ompatible with standard, silicon-based ultra-large-scale integra- through the non-radiative routes can be prevented, the radiative tion(ULSD) technology. Bulk silicon has an indirect energy emission could in principle be enhanced. Non-radiative recombi- bandgap and is therefore highly inefficient as a light source, nation is the result of diffusion of carriers to point defects in the necessitating the use of other materials for the optical emitters. silicon where efficient non-radiative recombination then occurs However, the introduction of these materials is usually incompat- Despite the low defect concentrations in good quality silicon this ible with the strict processing requirements of existing ULSI non-radiative route is always completely dominant. A way of echnologies. Moreover, as the length scale of the devices enhancing the radiative efficiency would be to prevent the carrier decreases, electrons will spend increasingly more of their time diffusion. If silicon can be formed as clusters then strong band-edge in the connections between components; this interconnectivity photoluminescence is possible,, but thus far this has only been problem could restrict further increases in computer chip proces- achieved by incorporating these clusters inside large bandgap sing power and speed in as little as five years. Many efforts have insulating oxides. However, the insulating matrix prevents efficient therefore been directed, with varying degrees of success, carrier injection, making devices difficult to produce. Similar but to engineering silicon-based materials that are efficient light much weaker(with efficiencies up to about 8 x 10) band-edge emitters-. Here, we describe the fabrication, using standard emission has also been observed in laser-recrystallized silicon, but silicon processing techniques, of a silicon light-emitting diode no explanation of its origin was presented LED) that operates efficiently at room temperature. Boron is We have made use of the controlled introduction of dislocation mplanted into silicon both as a dopant to form a p-n junction, as loops using conventional ion implantation and thermal processing well as a means of introducing dislocation loops. The dislocation The dislocation loop array, if appropriately produced, introduces a strain field in three dimensions that modifies the bandgap of the OOOOOOO AuSb Voltage () 8a5 Figure 1 The t for the device measured at room tempe schematic of the light-emitting diode(ED)device. The top and bottom ohmic contacts are Figure 2 Plots of the integrated electroluminescence intensity as a functi formed by Al and AuSb respectively. The infrared light is emitted through the window left forward voltage at various temperatures: 80 K (diamonds), 180 K (trial (circles) 192 A@2001 Macmillan Magazines Ltd NaturEvOl410)8March2001www.nature.comletters to nature 192 NATURE | VOL 410 | 8 MARCH 2001 | www.nature.com 29. Crone, B. et al. Large-scale complementary integrated circuits based on organic transistors. Nature 403, 521–523 (2000). 30. Mooij, J. E. et al. Josephson persistent-current qubit. Science 285, 1036–1039 (1999). 31. Verthamer, N. R., Helfand, E. & Hohenberg, P. C. Temperature and purity dependence of the superconducting field, Hc2, III. Electron spin and spin-orbit effects. Phys. Rev. 147, 295–303 (1966). Acknowledgements We thank E. A. Chandross, B. Crone, H. E. Katz, H. Y. Hwang, A. J. Lovinger and T. Siegrist for discussions, and E. Bucher for the use of equipment. Correspondence and requests for materials should be addressed to J.H.S. (e-mail: hendrik@lucent.com). ................................................................. An efficient room-temperature silicon-based light-emitting diode Wai Lek Ng*, M. A. Lourenc¸o*, R. M. Gwilliam*, S. Ledain†, G. Shao† & K. P. Homewood* * School of Electronic Engineering, Information Technology & Mathematics; † School of Mechanical and Materials Engineering, University of Surrey, Guildford, Surrey, GU2 7XH, UK .................................. ......................... ......................... ......................... ......................... ........ There is an urgent requirement for an optical emitter that is compatible with standard, silicon-based ultra-large-scale integra￾tion (ULSI) technology1 . Bulk silicon has an indirect energy bandgap and is therefore highly inefficient as a light source, necessitating the use of other materials for the optical emitters. However, the introduction of these materials is usually incompat￾ible with the strict processing requirements of existing ULSI technologies. Moreover, as the length scale of the devices decreases, electrons will spend increasingly more of their time in the connections between components; this interconnectivity problem could restrict further increases in computer chip proces￾sing power and speed in as little as five years. Many efforts have therefore been directed, with varying degrees of success, to engineering silicon-based materials that are efficient light emitters2–7. Here, we describe the fabrication, using standard silicon processing techniques, of a silicon light-emitting diode (LED) that operates efficiently at room temperature. Boron is implanted into silicon both as a dopant to form a p–n junction, as well as a means of introducing dislocation loops. The dislocation loops introduce a local strain field, which modifies the band structure and provides spatial confinement of the charge carriers. It is this spatial confinement which allows room-temperature electroluminescence at the band-edge. This device strategy is highly compatible with ULSI technology, as boron ion implanta￾tion is already used as a standard method for the fabrication of silicon devices. There have been great efforts over the past decade to obtain useful, that is, technologically viable and efficient, light emission from silicon both in the visible and infrared regions of the spectrum. In the visible regions, porous silicon2 and other quantized systems, such as silicon/silicon dioxide superlattices3 and silicon nanopreci￾pitates in silicon dioxide4 , have been the main emphasis. In the infrared region, systems such as erbium in silicon5 , silicon/ germanium6 and, more recently, iron disilicide7 offer potential routes. No approach has so far been applied commercially. The reasons for this are a combination of the lack of genuine or perceived compatibility with conventional ULSI technology and, in the case of infrared emitters, high thermal quenching giving very poor room-temperature efficiencies. Here we use a new approach—dislocation engineering, using conventional ULSI technology—that gives efficient light emission in silicon at room temperature. Because of its indirect bandgap, silicon is fundamentally a poor emitter of light. The main reason for this is that fast non-radiative recombination routes dominate the slower radiative route in this material. Indeed, in bulk silicon, at room temperature, radiative emission is normally entirely absent. However, if recombination through the non-radiative routes can be prevented, the radiative emission could in principle be enhanced. Non-radiative recombi￾nation is the result of diffusion of carriers to point defects in the silicon where efficient non-radiative recombination then occurs. Despite the low defect concentrations in good quality silicon this non-radiative route is always completely dominant. A way of enhancing the radiative efficiency would be to prevent the carrier diffusion. If silicon can be formed as clusters then strong band-edge photoluminescence is possible8 , but thus far this has only been achieved by incorporating these clusters inside large bandgap insulating oxides. However, the insulating matrix prevents efficient carrier injection, making devices difficult to produce. Similar but much weaker (with efficiencies up to about 8 3 10 2 6 ) band-edge emission has also been observed in laser-recrystallized silicon9 , but no explanation of its origin was presented. We have made use of the controlled introduction of dislocation loops using conventional ion implantation and thermal processing. The dislocation loop array, if appropriately produced, introduces a strain field in three dimensions that modifies the bandgap of the –1.5 –1.0 –0.5 Voltage (V) 0.0 0.5 1.0 0 2 4 6 8 10 AuSb Al AuSb B (p+) c-Si (n) Current (mA) Figure 1 The current–voltage plot for the device measured at room temperature. Inset, a schematic of the light-emitting diode (LED) device. The top and bottom ohmic contacts are formed by Al and AuSb respectively. The infrared light is emitted through the window left in the bottom contact. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 4 8 12 16 Forward voltage (V) Integrated electroluminescence (a.u.) Figure 2 Plots of the integrated electroluminescence intensity as a function of applied forward voltage at various temperatures: 80 K (diamonds), 180 K (triangles) and 300 K (circles). © 2001 Macmillan Magazines Ltd