MOS technology. Opening statement

276

World Abstracts on Microelectronics and Reliability

technologies to a number of consumer areas. It discusses the relevant advantages of MOS and Bipolar technology; a wide range of actual devices are included and brief details of the operating performance.

Linear and interlace circuits. A. J. PAYNE. Microelectron. & Reliab. 13, 373 (1974). The scope of linear and interface circuits covers such a vast field that this paper will be limited to brief descriptions of some recent IT1~ Semiconductors developments in the field of custom design and the 7100 series of standard devices. These devices encompass TTL, MOS and Bipolar technology and therefore provide a broad technology view across a limited number of products. Briefly, the circuits are divided into three groups:

(a) Input interface circuits. (b) Output interface circuits. (c) Miscellaneous peripheral circuits. C-MOS--a status report. A. BISHOP and P. JONES. Microelectron. & Reliab. 13, 363 (1974). Today's electronic design engineer has before him a bewildering choice of digital "building blocks". The range of speed, complexity, power dissipation, etc. available is indeed very wide. The vast majority of these "building blocks" in the past have been based on bipolar technology, but another option is increasingly making its presence felt in the digital component field. This option is the one presented by Complementary Metal-Oxide-Silicon components. C-MOS has the high packing density of MOS combined with logic speeds approaching that of TTL. The general characteristics of C-MOS have been fully described elsewhere ([1, 2] and Appendix) and as a contrast this paper examines two specific aspects of C-MOS. Firstly on the processing side a comparison is made of today's approaches. Secondly a comparison between C-MOS and TTL is illustrated by a typical equipment design.

Mixed-discipline monolithic integrated circuits--bipolar, MOS and COS/MOS technologies and future trends. L. R. AVERy. Microelectron. & Reliab. 13, 349 (1974). This paper describes the basic bipolar integrated-circuit process, its advantages and its limitations. The benefits of close matching and temperature tracking are discussed. The current mirror concept is developed and its various forms and its value in circuit design are discussed, culminating in the operational transconductance amplifier (OTA). The paper then deals with the shortcomings of p - n - p lateral transistors and their replacement by p-MOS devices to achieve wideband performance in operational amplifiers. MOS-input operational amplifiers, together with techniques for reducing offset voltage and offsetvoltage drift, are mentioned. The special requirements of COS/MOS processing for linear-mode devices are examined, with special reference to the CA3600E. The possibility of circuits with zero popcorn noise and ultra high input impedance, and the combination of linear and digital functions on a single chip, are also discussed. COS/MOS-bipolar interface circuits are analysed, followed by future trends in design, including dielectric, air isolation and beam-lead technology.

Bipolar memories. P. ALFKE. Microelectron. & Reliab. 13, 339 (1974). Bipolar Read-Write Memories (RAMs) are very complex and sophisticated devices using the latest advances in semiconductor technology to achieve high speed and high packing density. They are more expensive and consume more power than MOS circuits of comparable complexity, but they are significantly faster and are much easier to understand and to use.

Bipolar RAMs use the same supply voltages as the

associated logic (+5 V for "ITL, -5.2 V for ECL) and their inputs and outputs are level compatible with the logic, no special interface elements are required. LIC t~dmolo~. N. DOYLE. Microelectron. & Reliab. 13, 315 (1974). The first monolithic integrated circuit, constructed in 1961, was an RTL flip--flop. Three years later the first linear IC followed. Like the flip--flop, this was also a basic building-block---an operational amplifier. Since that time the growth of digital ICs into MSI and LSI systems has been paralleled by the evolution of complex linear sub-systems. The typical linear IC of the 1970s is no longer a simple analog gain block but rather a complex device performing a number of precise functions. This seminar will review the present status of linear IC technology and discuss future trends. IC voltage regulators. N. DOYLE. Microelectron. & Reliab. 13, 325 (1974). The power supply is the one design area common to all systems and, at the same time, that portion of the system most likely to be neglected in design effort or foreshortened on design time due to continual changes in the system power requirements. The purpose of this paper is to familiarize the designer with the Fairchild range of monolithic voltage regulators and to show some of the more popular applications of these devices. Monolithic voltage regulators have made giant strides since their introduction in 1967, the most significant step probably being the advent of the three-terminal fully protected regulator in 1969.

MOS Wdmolol~--MOS a ma|or taetor. S. FORTE. Microelectronics 6, 14 (1974). MOS has long been transformed from a laboratory curiosity to a major factor in the semiconductor industry. We have moved a long way since the time when MOS products seemed an attractive proposition from the point of view of higher density and relative process simplicity but suffered what appeared to be an inherent lack of stability and were thus relegated to the laboratory. The advent of the electronic calculator and the interreactive impact of MOS on calculators and of calculators on MOS cannot be overstated. Since the calculator aspect has been fully covered by Dr. Carlson other aspects of MOS which have helped to put it in its present position will be examined.

Bi-polar teelmolo~. Integrated injection Iogk---present and tuture. N. C. DE TROYE. Microelectronics 6, 9 (1974). During the 1972 International Solid State Circuit Conference two papers were presented which discussed a new form of bi-polar logic circuit using vertical npn transistors with multiple collectors as inverters and lateral pnp transistor current sources as loads. This novel logic has been named Integrated Logic (I2L) or Merged Transistor Logic (MTL). High packing densities and lowspeed-power products for low and medium currents have been obtained. Further developments have taken place in the direction of I2L gates with low speed-power products and chips containing other types of logic and analogue circuits in addition to I2L. This paper will discuss specifically the state of the art of VL gates. After a short description of the basic VL structure, different possible ways of producing VL gates with low speed-power products will be shown. A layout comparison with MOS logic will be presented and finally a possible way of producing faster circuits will be proposed.

MOS teelanolog,v. Opening ~ m e n t .

R. S. CARLSON. Microelectronics 6, 5 (1974). When discussing the relative merits of two technologies it is necessary to stress at the outset that a technology has value only if it can be used to

World Abstracts on Mieroelectronics and Reliability produce useful economic products. The issue at stake here is to attempt to assess the business potentials of metal oxide silicon (MOS) and bi-polar transistor technologies. It must be emphasized that complex equations must be solved in order to evaluate the particular business potential of a new technology. Thus is it only possible at the best, in comparing the merits of two rival technologies, to establish a qualitative model by which judgments can be made. Such a result should not be regarded as one of minor significance since it is upon this evidence that the development and use of the chosen technology has to be made.

adequate description of a single all-embracing fabrication method. The purpose of this paper is not to question the success to date of MOS. It is to question the continued dominance of MOS into the future, and to suggest that it will become necessary for the potential user/designer to assess more carefully the relative trade-off between MOS and Bi-polar technology before making his decision.

Bi-polar technology. A re-evaluation. D. H. ROBERTS. Microelectronics 6, 18 (1974). In the mid-1960's P-channel metal gate MOS technology burst on the scene offering overnight the facility to achieve certain digital functions at a fraction of the cost of that required for bi-polar implementation of the same function. Now, a decade later, one can see that the early promise of MOS has been achieved in offering low cost/high complexity slow speed digital functions that have resulted in the widespread use of MOS in such applications as hand-held calculators, shift registers for serial memories, read only memories and random access memories. Indeed, the situation has been reached where approximately 10,000 active elements per silicon chip have become the accepted state of the art. At the same time the initial simple approach to MOS technology has been eroded by the gradual penetration of: (i) N-channel processes (if) complementary channel processes (iii) ion implantation (iv) alternative gate conductors such as silicon (v) alternative gate dielectrics, for example silicon nitride (vi) sapphire substrates The successful application of one or other of this. vast range of MOS processes to the economic achievement of complex digital functions has led to the situation where LSI and MOS are often considered to be synonymous to the exclusion of hi-polar, and also where MOS is treated as an 7. S E M I C O N D U C T O R

INTEGRATED

Effect of disorder on direct and indirect band gaps of semiconductor alloys. M. ALTARELLI. Solid State Commun. Vol. 15. p. 1607. A simple analysis of the effect of disorder on different band edges of ternary III-V alloys is performed in terms of the symmetry properties of Bloeh functions. The effect of disorder on the X~ conduction band minimum is shown to be strongly dependent on the valence of the alloyed element, in agreement with available experimentaF data. The implications for indirect gaps and for the calculations of crossover compositions are briefly discussed.

The gate-controlled diode Someasurement and steady-state lateral current flow in deeply depleted MOS structures. R F. PIERRET. Solid State Electron. 1974. Vol. 17. p. 1257. A detailed analysis of steady state lateral current flow in thL deeply depleted gate-controlled diode structure is performed primarily to ascertain the minority carrier quasiFermi level positioning and lateral variation near the oxide-semiconductor interface. From a consideration of limiting case solutions it is quantitatively established that, under typical conditions, the semiconductor surface will be weakly inverted over all but a small fraction of the surface channel for gate biases far below the voltage required to strongly invert the surface. It follows from this weak inversion positioning of the minority carder quasi-Fermi level that the So-value deduced from the previously published interpretation of gate-controlled diode data will be less, typically much smaller, than the true depleted surface recombination velocity. A method for ascertaining the

277

The impact of technological change in the semiconductor industry. D. J. ELLIOT. Microelectronics 6, 29 (1974). On June 30, 1948, one of the most important innovations of this century was announced. Scientists of Bell Laboratories revealed the invention of an electronic device to replace the vacuum tube--the transistor. Since that point in time, the electronics industry has grown from fortieth to fourth in annual dollar volume, with sales increasing from 1.7 billion dollars to nearly 40 billion today. The invention of the transistor marked the beginning of a revolution in electronics. The transistor was the springboard for numerous technologies that have had tremendous impact on our lives. Almost every phase of modern life has been, or is being, improved by the transistor. In order to appreciate the significance and impact of the transistor and the technological change it produced, we must first examine the causes of this revolution. Secondly, we can examine the nature of growth in the semiconductor industry, tracing the chronology of events and technology acceleration that led to the current state of the art. We will also examine technology change in the semiconductor industry as a function of device complexity, market expansion and dollar volume growth. Finally, we will look at the implications of semiconductor industry growth for consumer products, computer technology and for future semiconductor technology. The overall impact of the transistor reaches into social, political, economic and technological spheres. An attempt will be made to identify both the current and future impact of semiconductor technology.

CIRCUITS,

DEVICES

AND

MATERIALS

minimum error in the previously published so determination procedure is presented and illustrated. Ohmic hole mobility in cubic semiconductors. M. COSTATO, G. GAOLIANI, C. JACOBONI and L. REGGIANL J. Phys. Chem. Solids. 1974 35, p. 1605. Lattice ohmic mobility of holes in cubic semiconductors has been theoretically evaluated accounting for most fundamental aspects of the physical problem. Solutions sufficiently simple for the interpretation of experimental data have been obtained. (1) The deformation potential approach has been used in the description of the hole-acoustic and hole-optical phonon scattering mechanism. (2) The overlap corrections introduced in the scattering rates by the symmetry of the valence band wave functions, which are predominantly p-like, have been taken into account. (3) Intra-band and inter-band transitions within and between the heavy and light mass bands have been considered. (4) Nonparabolicity and warping of equienergetic surfaces has been omitted in favour of spherical parabolic bands. Results show that (2) doubles the mobility with respect to the case in which overlap is omitted, and (3) halves the mobility when the two valence bands have equal effective masses with respect to the case in which inter-band scattering is omitted. Comparison with experimental mobility is good for Ge and fair for Si due to the strong non-parabolicity of its valence bands.

Method of calculation of the impurity obtained by a multiple diffusion process. W. MALY. Electron Tech. (Warsaw) 7 1/2 p. 19. The purpose of this work is to present a method of calculation of the impurity distribution obtained in the