« iepriekšējāTurpināt »
better interdiffusion barrier characteris- nonvolatile memories, with attendant Research Laboratories. Based on the tics than those of the well-established radiation resistance, high density, and low leakage current and large dielectric TiW. However, for step coverage over low voltage operation. N. Abt et al. endurance of Sio, and high dielectric high aspect ratio contacts of via holes from the National Semiconductor Cor- constant of titanium oxide, they strived conformal depositions such as from poration reviewed the current status to optimize the dielectric by forming a CVD are needed. The ULSI fabrica- and electrical requirements for gigabit silicon titanium oxide. tion needs further dictate deposition level memories. They pointed out that The emergence of ferroelectric thin temperatures below about 700 °C. the ferroelectric process is being films is quite an interesting developSherman presented encouraging results developed as an addendum to the ment and brings possible integration for low-temperature, PECVD- standard CMOS, so that it could prove of nonvolatile memory, a feature unavaildeposited TiN using organometallic versatile enough to be tagged on to able in the days of the nonvolatile ferprecursors. However, the suitability of different technologies. Sol-gel or romagnetic core memory. their process for diffusion barriers and sputter-deposited lead zirconate titanate conformal deposition are yet to be (PZT) is the material mostly used in ADVANCED demonstrated. An alternative Tin film current ferroelectric memories, and OPTOELECTRONIC DEVICES deposition technique--electron cyclo- some of the problem areas involve tron resonance (ECR) plasma CVD-- patterning, grain size control, and long- This symposium was highlighted by was demonstrated by T. Akahori and term ferroelectric endurance.
an SSDM plenary lecture by T. Ikegami A. Tanihara of Sumitomo. Using Tici, Formation of PZT films by a novel of NTT Optoelectronics Laboratories and N, as the constituent gases, they MOCVD process was described by a and a series of invited talks. Dr. Ikegami were able to obtain Cl-free Tin films group from Mitsubishi LSI Labora- traced the evolution of fiber optic device with low resistivity (40 u 12-cm), inter- tory, and formation by a single-target performance over the past decade and diffusion stability for up to 650 °C, sputtering process was described by a half and focused on the laser diode 30-minute anneals (for an Al/TiN/Si researchers from Hitachi Central operating wavelength vis-a-vis the contact system), and excellent confor- Research Laboratories. The latter found optical fiber requirements and the vast mality. According to the paper from that precise compositional control was reduction in the laser threshold curK. Mori et al. of the Mitsubishi LSI possible by varying the RF power (which rent density. Distributed feedback laser Laboratory, thermal LPCVD of TiN controls the Pb content of the film), diodes have been commercially used at gives excellent contact plug filling and they obtained well-crystallized films 1.6 Gb/s rates, and multiple quantum capability. Using the same system, they with a dielectric constant of 1180 after well structures have given extremely also deposited TiSi_ohmic contacts prior a 590 °C anneal. Another Mitsubishi narrow linewidths. In the area of detecto the TiN plug deposition.
group reported on a sol-gel synthesis tors, the “Separated Absorption and A low-temperature tungsten CVD of lead-lanthanum zirconate-titanate Multiplication” (SAM) concept with process was described by H. Goto et al. (PLZT) films with an order of magni- several bandgap-engineered structures from the Hitachi Central Research tude reduction in leakage current rela- has given rise to an excess noise factor Laboratories. Using difluoro-silane as tive to PZT. They found the equivalent as low as 3 and a high gain bandwidth a reducing gas additive to WF, and Sio, film thickness to reduce with the (BW) factor of 90 GHz. Monolithic surface-reaction-limited deposition at PLZT thickness and had a value of integration of detector and front-end 270 to 395 °C, they were able to obtain 0.67 nm for a PLZT thickness of 100 nm. amplifier is also being widely conformal W films with low resistivity In a paper by H. Shinriki et al. from investigated. (10.4 u82-cm) and minimal encroach- the Hitachi Central Research Labora- The recently developed Er-doped ment on Si.
tories, extremely thin films of CVD fiber optic amplifiers (EDFA) have The principal new dielectric mate- Tazo, with a two-step annealing pro- attracted a great deal of attention due rials covered in this symposium were cess with UV-0, and dry O2 were to their high current gain at a waveferroelectric thin films. The develop- reported to have adequate properties length of 1.5 um. This has also resulted ment of ferroelectric thin films cur- as the storage capacitor dielectric in in the rush for developing new pumprently proceeds along two fronts: (1) the 1.5 V operated 64-Mbit DRAMs. They ing light sources (laser diodes). Optoapplication of the high dielectric attributed the improved properties to electronic integrated circuits (OEIC) constant in conventional DRAM the control of oxygen vacancies. A based on InP and planar light wave memory as it evolves to gigabit levels novel Si-based alternative to Tao, and circuits with patterned Sio waveguides and (2) the ferroelectric hysteresis the ferroelectrics was proposed by a on Si wafers are also being pursued that offers the exciting possibility of group from the Matsushita Central vigorously. The developments in the
field of optoelectronics are ushered in T.L. Paoli and R.L. Thornton of the The following outlines a few of the with both the advanced device con- Xerox Palo Alto Research Center pre- most striking results presented in these cepts originating from quantum-size sented a talk on the integration of optical effects and the outstanding develop- and electronic devices by impurity- Hot carrier reliability of metal oxidements in material growth.
induced layer disordering (IILD). IILD semiconductor field-effect transistors In the first talk of this symposium, occurs because the interface between (MOSFETs) under scale-down was M.E. Prise of AT&T Bell Laboratories different III-V alloys is unstable against discussed in a number of papers. Some described their new developments in solid state diffusion of the group III of the interesting results are as follows: free space interchip optical intercon- elements in the presence of a high nects between integrated circuits. He impurity concentration. Thus, under (1) 77 K operation to realize higher also appraised the audience on the use appropriate conditions, it is possible to speed and reliability. of microlasers and GaAs self-electro- intermix (without melting) layers whose optic device (SEED) based modulators resultant composition will be an averin the optical interconnect technology. age of those of the initial layers. With (2) Exponential distribution of trapped
T. Kobayashi and B.Y. Lee from lateral patterning on a wafer, lateral holes away from the Si/Sio, interOsaka University made a critical assess- bandgap engineering is possible. Si IILD face in thin oxide p-MOSFÉTS. ment of electro-optic (E-O) and opto- has thus far been applied to lowoptic (0-0) devices and concluded that threshold diode lasers and lateral hetero- (3) No X-ray irradiation effect on in order to fully realize their speed junction bipolar transistors with excep- interface-trap generation during hot potential in the picosecond to subpico- tionally high gain. The authors believe carrier injection; of interest with second range, it is essential to shorten that the applications will multiply as the use of x-ray lithography. any electrical interconnects. Thus, they the IILD technology is refined further. favor integration of the total system The characteristics of visible semi- In the area of thin gate dielectrics, including the optoelectronic devices. conductor lasers based on wavelength reoxidation of the oxynitride was shown
A review of the physics and technol- shortening in AlGaInP by forming to be an effective way of reducing leakogy of quantum dots and wires was multiple quantum well (MQW) active age. The practical tunneling limit on presented by K.J. Vahala et al. from the layers were reviewed by Y. Mori et al. the gate dielectric thickness is thought California Institute of Technology. of the Matsushita Semiconductor to be around 2.5 nm. Apart from enhancing the performance Research Center. They demonstrated Advanced silicon processing generof existing devices (e.g., quantum dot room temperature continuous-wave ally focuses on low-temperature epilasers), entirely new concepts for pro- (CW) lasing at 643.5 nm (bright red) taxy, low-energy ion implantation for cess architectures and synthetic dopants and suggested that orange and yellow shallow junctions, and an overall reduced could emerge from the physics at zero CW operation may be on the horizon. thermal budget. Some of the most interand one dimension. They also outlined Integrated optoelectronic devices esting findings are as follows: some of the key material fabrication based on organic/inorganic heterojunctechnologies that help realize quan- tions were discussed by S.R. Forrest
tions were discussed by S.R. Forrest (1) Low-temperature silicon epitaxy tum wires (selective epitaxy) and dots and F.F. So of the University of Southern without substrate heating, using (single-crystal GaAs cluster in the 40- California. Organic films, due to their an ECR PECVD; selective etching to 100-Å size range). Y. Yamamoto of weak Van der Waals bond, can be grown deposition achieved by controlling NTT Basic Research Laboratories and into ordered films on substrates with- the H, gas addition. G. Björkofthe Royal Institute of Tech- out meeting lattice match requirements. nology, Stockholm, presented their Organic films can also be active semi- (2) Profound influence of low-energy) theoretical calculations on microcavity conductors, thus enabling junction ion flux on the conductivity and lasers to demonstrate that lasing could formation and eventually formation of mobility of low-T grown Si epi films. occur without inversion. This fascinat- devices based on organic MQWs. ing result is attributed to a photon
(3) A rapid vapor phase direct doping recycling mechanism without energy OTHER TOPICS
technique for ultra-shallow juncloss. The authors believe that this scheme
tions (<50 nm), with surface concould open up new possibilities of Besides the above principal sympo- centration controlled by time, coherent lightwave generation in sia, certain sessions were devoted to temperature, and dopant gas flow. wavelength regions where population topics of high technological interest. inversion is hard to achieve.
(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-Velectron 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 Engincering in Scmiconductor Growth, Processing, and Device Technology" to be held in San Francisco in April 1992
ASPECTS OF SOLID STATE/ SEMICONDUCTOR PHYSICS RESEARCH
This article describes three laboratories in China that are making substantial
at the Modern Physics Institute, Fudan University; and the Ion Beam
by Norman J. Horing
of the Chinese physics establishment. real strength both theoretically and
While experimental physics in China experimentally. As an academic writing on aspects may suffer from a lack of expensive, of solid state/semiconductor physics in modern laboratory equipment, the NATIONAL LABORATORY FOR China, I feel it is appropriate to open strength of the theoretical base is evi- SUPERLATTICES AND with a few pertinent background dent in the fine work of the many Chinese MICROSTRUCTURES (NLSM), remarks. American universities and students and researchers in China, in INSTITUTE OF research laboratories have witnessed America, and throughout the world. SEMICONDUCTORS, the precursor of an important physics Moreover, their dedication is epito- ACADEMIA SINICA, BEIJING establishment in mainland China mized by the generation of senior physthrough the influx of very well trained, icists who struggled to maintain and Research at this well-equipped diligent, and bright Chinese graduate improve their skills in the face of the national research laboratory is focused students, postdoctoral students, and privations of the cultural revolution. on semiconductor superlattices and lowvisiting scientists. Starting around 1981, The rigors of those times gone by have dimensional microstructures, encomthrough the efforts of Prof. T.D. Lee, given way to rigorous classroom train- passing materials, physics, and device CUSPEA (China-U.S. Physics Exami- ing of a new generation of scientists standpoints. State-of-the-art fabricanation and Application program) stu- who are fully competent to lead China tion technologies are employed, includdents appeared at our universities, into the new technological age. As a ing molecular beam epitaxy (MBE), revitalizing the graduate student bodies poor nation with a vast population to metalorganic chemical vapor deposiin physics departments across the nation support, China has wisely invested its tion (MOCVD), etc. Launched in 1986 The ensuing years brought us non- resources in cultivating its ability to and re-established in 1989, this laboraCUSPEA students whose excellence "live by its brains." Thus, there stands tory has achieved significant progress was also impressive and refreshing. a fine Chinese physics establishment in many areas, including theoretical Furthermore, many of us have come to which, in time, can help to develop and optical spectroscopy studies of have a keen appreciation of the excep- Chinese technology toward the point electronic structure, lattice dynamics, tional talent and depth of knowledge of of world competitiveness.
optical transitions and energy relaxavisiting Chinese scholars at all levels. This article is concerned with three tion of photo-excited electrons in superBehind the strong, beneficial impact laboratories in China that are making lattices and quantum wells, quantum that this has had on the American- substantial progress in engaging impor- transport and other electronic response indeed the world--physics community tantcurrent research in various aspects properties of low-dimensional electronic lies the substantial intellectual power of condensed matter physics, showing systems, growth technologies and
material characterizations of MBE- 6. Physicalstudies of processes under- scientists from all over the world are grown superlattices and multilayered lying superlattice/quantum well welcome to work in the laboratory. heterostructures, and new electronic based devices.
Limited research grants are available and optoelectronic devices based on
(only to cover the costs of expendable superlattices and quantum wells. A tangible representation of research supplies and usage of expensive labo
Having twin objectives in the devel- at NLSM is provided in the list of ratory facilities involved in research at opment and refinement of MBE and recent research paper titles given in NLSM). Application forms may be MOCVD technologies and in the basic Appendix A.
obtained from: physical research on new features of The principal facilities available at low-dimensional superlattices and NLSM include:
Chief Secretary microstructures and their potential
Dr. Jian Liu device applications, this laboratory's (a) VAX 3500 computer system.
The National Laboratory for principal scientific programs are:
Superlattices and (b) Time-resolved optical spectroscopy
structures, elementary excitations, nously pumped mode-locked dye Academia Sinica
Tel: 288131-287 (domestic) structures.
Fax: 86-01-2562389 (c) Raman spectroscopic system and 2. Spectroscopic investigations, includ- photoluminescence spectrometer SURFACE PHYSICS ing conventional/time-resolved
LABORATORY (SPL), photo-absorption and photo
MODERN PHYSICS reflection spectroscopies, Raman (d) Modulation/infrared spectrometer INSTITUTE, FUDAN spectroscopy, magneto-optic spec- composed of monochromator, 10-K UNIVERSITY, SHANGHAI troscopies, etc., used in combination cryogenic refrigerator, and detector/ with temperature, pressure, and computer data acquisition system. The Surface Physics Laboratory at electric/magnetic field modulation
Fudan University in Shanghai is one of techniques. Research topics consist (e) Far infrared cyclotron resonance/
(e) Far infrared cyclotron resonance/ China's prestigious State Key Laboraof energy band structures, intrinsic
magneto-opticalspectroscopysys- tories. It is dedicated to the study of and extrinsic recombinations, dynam- tem constructed by far infrared response and interaction properties of ics of photo-excited electrons, and molecular laser and 8-T supercon- surfaces, interfaces, superlattices, and nonlinear optical properties of ducting magnet system.
heterostructures. It is equipped with semiconductor superlattices and
one MBE machine, with another to be quantum wells.
(1) Different computerized deep level added soon. The major research pro
transient spectroscopy (DLTS) grams of this laboratory include: 3. Quantum transport (both parallel junction spectrometers. and perpendicular to interfaces) and
1. Theoretical studies of electronic its relevance to dimensional size and
structures and vibrational properunderlying physical processes. (g) 8-T/12-T superconducting magnet ties of semiconductor surfaces, inter
system, including 0.3-K He low faces, and superlattices. 4. Investigation of electronic proper- temperature insert, 4.2- to 300-K
ties and behavior of impurities, variable temperature insert, and 2. Theoretical studies on the electronic defects, and deep-level centers computer data acquisition systems. properties of metal-semiconductor related to MBE- and MOCVD
interfaces. grown materials.
(h) Molecular beam epitaxy growth
3. Electron spectroscopy studies of the 5. Material growth and technologies
polar surfaces of III-V compound of various low-dimensional semi- NLSM has about 20 senior semiconductors. conductor structures with artifi- researchers and administrators and a cially tailored band structures. total staff of about 40 people. Visiting