better interdiffusion barrier characteristics than those of the well-established TiW. However, for step coverage over high aspect ratio contacts of via holes, conformal depositions such as from CVD are needed. The ULSI fabrication needs further dictate deposition temperatures below about 700 °C. Sherman presented encouraging results for low-temperature, PECVDdeposited TiN using organometallic precursors. However, the suitability of their process for diffusion barriers and conformal deposition are yet to be demonstrated. An alternative TiN film deposition technique--electron cyclotron resonance (ECR) plasma CVD-was demonstrated by T. Akahori and A. Tanihara of Sumitomo. Using TiCl and N2 as the constituent gases, they were able to obtain Cl-free TiN films with low resistivity (40 μ-cm), interdiffusion stability for up to 650 °C, 30-minute anneals (for an Al/TiN/Si contact system), and excellent conformality. According to the paper from K. Mori et al. of the Mitsubishi LSI Laboratory, thermal LPCVD of TiN gives excellent contact plug filling capability. Using the same system, they also deposited TiSi, ohmic contacts prior to the TiN plug deposition. A low-temperature tungsten CVD process was described by H. Goto et al. from the Hitachi Central Research Laboratories. Using difluoro-silane as a reducing gas additive to WF, and surface-reaction-limited deposition at 270 to 395 °C, they were able to obtain conformal W films with low resistivity (10.4 μ-cm) and minimal encroachment on Si. The principal new dielectric materials covered in this symposium were ferroelectric thin films. The develop ment of ferroelectric thin films currently proceeds along two fronts: (1) the application of the high dielectric constant in conventional DRAM memory as it evolves to gigabit levels and (2) the ferroelectric hysteresis that offers the exciting possibility of nonvolatile memories, with attendant radiation resistance, high density, and low voltage operation. N. Abt et al. from the National Semiconductor Corporation reviewed the current status and electrical requirements for gigabit level memories. They pointed out that the ferroelectric process is being developed as an addendum to the standard CMOS, so that it could prove versatile enough to be tagged on to different technologies. Sol-gel or sputter-deposited lead zirconate titanate (PZT) is the material mostly used in current ferroelectric memories, and some of the problem areas involve patterning, grain size control, and longterm ferroelectric endurance. Formation of PZT films by a novel MOCVD process was described by a group from Mitsubishi LSI Laboratory, and formation by a single-target sputtering process was described by researchers from Hitachi Central Research Laboratories. The latter found that precise compositional control was possible by varying the RF power (which controls the Pb content of the film), and they obtained well-crystallized films with a dielectric constant of 1180 after a 590 °C anneal. Another Mitsubishi group reported on a sol-gel synthesis of lead-lanthanum zirconate-titanate (PLZT) films with an order of magnitude reduction in leakage current relative to PZT. They found the equivalent SiO, film thickness to reduce with the SiO, film thickness to reduce with the PLZT thickness and had a value of 0.67 nm for a PLZT thickness of 100 nm. In a paper by H. Shinriki et al. from the Hitachi Central Research Laboratories, extremely thin films of CVD Ta,O, with a two-step annealing process with UV-O, and dry O, were reported to have adequate properties as the storage capacitor dielectric in 1.5 V operated 64-Mbit DRAMs. They attributed the improved properties to the control of oxygen vacancies. A novel Si-based alternative to Ta,O, and the ferroelectrics was proposed by a group from the Matsushita Central Research Laboratories. Based on the low leakage current and large dielectric endurance of SiO, and high dielectric constant of titanium oxide, they strived to optimize the dielectric by forming a silicon titanium oxide. The emergence of ferroelectric thin films is quite an interesting development and brings possible integration of nonvolatile memory, a feature unavailable in the days of the nonvolatile ferromagnetic core memory. ADVANCED OPTOELECTRONIC DEVICES This symposium was highlighted by an SSDM plenary lecture by T. Ikegami of NTT Optoelectronics Laboratories and a series of invited talks. Dr. Ikegami traced the evolution of fiber optic device performance over the past decade and a half and focused on the laser diode operating wavelength vis-a-vis the optical fiber requirements and the vast reduction in the laser threshold current density. Distributed feedback laser diodes have been commercially used at 1.6 Gb/s rates, and multiple quantum well structures have given extremely narrow linewidths. In the area of detectors, the "Separated Absorption and Multiplication" (SAM) concept with several bandgap-engineered structures has given rise to an excess noise factor as low as 3 and a high gain bandwidth (BW) factor of 90 GHz. Monolithic integration of detector and front-end amplifier is also being widely investigated. The recently developed Er-doped fiber optic amplifiers (EDFA) have attracted a great deal of attention due to their high current gain at a wavelength of 1.5 μm. This has also resulted in the rush for developing new pumping light sources (laser diodes). Optoelectronic integrated circuits (OEIC) based on InP and planar light wave circuits with patterned SiO, waveguides on Si wafers are also being pursued vigorously. The developments in the field of optoelectronics are ushered in with both the advanced device concepts originating from quantum-size effects and the outstanding developments in material growth. In the first talk of this symposium, M.E. Prise of AT&T Bell Laboratories described their new developments in free space interchip optical interconnects between integrated circuits. He also appraised the audience on the use of microlasers and GaAs self-electrooptic device (SEED) based modulators in the optical interconnect technology. T. Kobayashi and B.Y. Lee from Osaka University made a critical assessment of electro-optic (E-O) and optooptic (O-O) devices and concluded that in order to fully realize their speed potential in the picosecond to subpicosecond range, it is essential to shorten any electrical interconnects. Thus, they favor integration of the total system including the optoelectronic devices. A review of the physics and technology of quantum dots and wires was presented by K.J. Vahala et al. from the California Institute of Technology. Apart from enhancing the performance of existing devices (e.g., quantum dot lasers), entirely new concepts for process architectures and synthetic dopants could emerge from the physics at zero and one dimension. They also outlined some of the key material fabrication technologies that help realize quantum wires (selective epitaxy) and dots (single-crystal GaAs cluster in the 40to 100-Å size range). Y. Yamamoto of T.L. Paoli and R.L. Thornton of the Xerox Palo Alto Research Center presented a talk on the integration of optical and electronic devices by impurityinduced layer disordering (IILD). IILD occurs because the interface between different III-V alloys is unstable against solid state diffusion of the group III elements in the presence of a high impurity concentration. Thus, under appropriate conditions, it is possible to intermix (without melting) layers whose resultant composition will be an average of those of the initial layers. With lateral patterning on a wafer, lateral bandgap engineering is possible. Si IILD has thus far been applied to lowthreshold diode lasers and lateral heterojunction bipolar transistors with exceptionally high gain. The authors believe that the applications will multiply as the IILD technology is refined further. The characteristics of visible semiconductor lasers based on wavelength shortening in AlGaInP by forming multiple quantum well (MQW) active layers were reviewed by Y. Mori et al. of the Matsushita Semiconductor Research Center. They demonstrated room temperature continuous-wave (CW) lasing at 643.5 nm (bright red) and suggested that orange and yellow CW operation may be on the horizon. Integrated optoelectronic devices NTT Basic Research Laboratories and into ordered films on substrates with- Besides the above principal sympo- (4) Focused ion beam (FIB) etching for micromachining and FIBinduced lateral solid phase epitaxy for eventual three-dimensional integration. (5) Hot ion implantation (up to 500 °C) into CVD a-Si to increase conductivity. (6) Cryogenic BF2+ ion implantation to suppress defect annealing during implantation and thereby alter fluorine trapping by defects. The emerging areas of Si nanostructures and heterojunction bipolar transistors (HBT) were addressed in a session on advanced devices. An evening "rump" session also dealt with the latter topic. There were also a few sessions devoted to advances in III-V electron and optical devices, silicon processing, and reliability. S. Ashok is presently a Professor of Engineering Science at the Pennsylvania State University. He received his B.E. degree from P.S.G. College of Technology (University of Madras), Coimbatore, India, an M.Tech. degree from the Indian Institute of Technology, Kanpur, and a Ph.D. degree from Rensselaer Polytechnic Institute, Troy, NY, all in electrical engineering. His research interests are in semiconductor surface modification, ion-assisted processing, Schottky barriers, interface phenomena, radiation effects, thin films, and photovoltaics. He has authored or coauthored over 100 publications in these areas and has held visiting positions at Uppsala University, Sweden, Indian Institute of Science, Bangalore, University of Erlangen-Nürnberg, Germany, and Technical University, Aachen, Germany. He is the principal organizer of a forthcoming Materials Research Society symposium on "Defect Engineering in Semiconductor Growth, Processing, and Device Technology" to be held in San Francisco in April 1992. ASPECTS OF SOLID STATE/ SEMICONDUCTOR PHYSICS RESEARCH IN CHINA This article describes three laboratories in China that are making substantial INTRODUCTION As an academic writing on aspects of solid state/semiconductor physics in China, I feel it is appropriate to open with a few pertinent background remarks. American universities and research laboratories have witnessed the precursor of an important physics establishment in mainland China through the influx of very well trained, diligent, and bright Chinese graduate students, postdoctoral students, and visiting scientists. Starting around 1981, through the efforts of Prof. T.D. Lee, CUSPEA (China-U.S. Physics Examination and Application program) students appeared at our universities, revitalizing the graduate student bodies in physics departments across the nation. The ensuing years brought us nonCUSPEA students whose excellence was also impressive and refreshing. Furthermore, many of us have come to have a keen appreciation of the exceptional talent and depth of knowledge of visiting Chinese scholars at all levels. Behind the strong, beneficial impact that this has had on the American-indeed the world--physics community lies the substantial intellectual power by Norman J. Horing of the Chinese physics establishment. While experimental physics in China may suffer from a lack of expensive, modern laboratory equipment, the modern laboratory equipment, the strength of the theoretical base is evident in the fine work of the many Chinese students and researchers in China, in America, and throughout the world. Moreover, their dedication is epitomized by the generation of senior physicists who struggled to maintain and improve their skills in the face of the privations of the cultural revolution. The rigors of those times gone by have given way to rigorous classroom training of a new generation of scientists who are fully competent to lead China into the new technological age. As a poor nation with a vast population to support, China has wisely invested its resources in cultivating its ability to "live by its brains." Thus, there stands a fine Chinese physics establishment which, in time, can help to develop Chinese technology toward the point of world competitiveness. This article is concerned with three laboratories in China that are making substantial progress in engaging important current research in various aspects of condensed matter physics, showing real strength both theoretically and experimentally. NATIONAL LABORATORY FOR SUPERLATTICES AND MICROSTRUCTURES (NLSM), INSTITUTE OF SEMICONDUCTORS, ACADEMIA SINICA, BEIJING Research at this well-equipped national research laboratory is focused on semiconductor superlattices and lowdimensional microstructures, encompassing materials, physics, and device standpoints. State-of-the-art fabrication technologies are employed, including molecular beam epitaxy (MBE), metalorganic chemical vapor deposition (MOCVD), etc. Launched in 1986 and re-established in 1989, this laboratory has achieved significant progress in many areas, including theoretical and optical spectroscopy studies of electronic structure, lattice dynamics, optical transitions and energy relaxation of photo-excited electrons in superlattices and quantum wells, quantum transport and other electronic response properties of low-dimensional electronic systems, growth technologies and |