Number 257, February 6, 1996 by Phillip F. Schewe and Ben Stein SELF-ASSEMBLED QUANTUM DOTS . Reducing the dimensionality of semiconductor structures not only saves space, but also brings about beneficial quantum effects. Quantizing allowable electron energies, for example, can make an electronic device more efficient and allow it to operate at lower voltages and higher speeds. Quantum wells, semiconductor sandwiches in which electrons are confined to a plane, are already at work in some devices. Quantum wires (electrons restricted to one dimension) and quantum dots (electrons restricted to a point volume) might prove even more useful. Dotlike structures can be made lithographically (with electron beams and etching) but scientists have tried to achieve greater control and economy by "growing" dots. A group of scientists at the University of Wisconsin (Max Lagally, 608-263-2078), AT&T Bell Labs, and IBM have deposited layers of SiGe on a Si substrate. The strain which arises from the atomic mismatch between the two materials is accommodated by the development of SiGe islands. After several cycles of SiGe and Si layers, the strain built into the material results in the nucleation of a regular array of 100-nm- wide, 3-10-nm-tall SiGe pyramids. A simple theoretical explanation of the phenomenon has now been achieved (J. Tersoff et al., Physical Review Letters, 4 March 1996). These self- organized nanostructures are not exactly point-like (they contain about 10**16 atoms each), but the researchers hope to make them smaller and more isolated by controlling the deposition conditions during growth. Meanwhile, a group at the University of Southern California has made InAs islands on top of GaAs (Qianghua Xie et al., Phys. Rev. Lett., 25 Sept. 1995). What can one do with an array of nanodots? Possibly read and write data into them, probably not through wires, but possibly with tiny bursts of laser light. OPTICAL BULLET HOLES , non-light-absorbing circles on an otherwise absorbing material, have been shown theoretically to be stable and to exist on microscopic scales, offering the possibility for a new type of optical memory scheme. Based on a model developed by researchers at the University of Strathclyde in Scotland (contact Willie Firth, willie@phys.strath.ac.uk), the scheme would involve using a saturable light-absorbing material inside a cavity. Turning on a uniform, constant background light field and then shining very short pulses of light on the material can create two-dimensional bullet holes; essentially, what is happening is that the absorber is being excited at specific locations into a high-energy state, preventing it from taking in any further light. "Unlike real bullet holes, optical bullet holes can be moved around," say the researchers. This is done by locally changing the phase of the background light field, creating a matrix of attracting spots that can pull the bullet holes into new locations. "Of course a message or picture could be written by a volley of bullets, just as in cowboy cartoons," says Firth. But these optical bullet holes, which can be made as tiny as 10 optical wavelengths according to the model and possibly smaller, can potentially be used to store and manipulate computer bits. One advantage of this scheme is that the regularly spaced attracting spots could control the position of the bullet holes, serving as a kind of error correction mechanism. Collaborating teams across Europe are planning to implement these ideas and demonstrate them experimentally in such materials as organic solids and semiconductor microstructures. (W.J. Firth and A.J. Scroggie, upcoming article in Phys. Rev. Letters).