Microelectronics

microelectronics is a subsection of electro-technology and/or. electronics, which is concerned with the miniaturization of electronic circuits.

Microelectronics has two outstanding features:

  • Integration (in the detail: see 'integrated circuit '): Electronic circuits become on a common substrate in a common manufacturing stepproduced. The components of the circuits consist mostly of transistors, in addition, resistances, condensers and other semiconductor components. The common substrate is a semiconductor material, usually a so-called wafer from monocrystalline silicon. The elements become in and on the base material throughchemical and mechanical processes of transformation of the semiconductor technology produces.
  • Miniaturization: The components of the circuit (and thus the circuit as a whole) are continuously made smaller. The dimensions of a transistor are in the year 2004 clearly under a micrometer. Thus integrated circuits with several millions leave themselves transistors opena piece of silicon with an edge length of few (typically < 10) Millimeters realize.

To table of contents

applications

Elements of microelectronics were originally developed for the requirements of space travel after small and light construction units. They are to be found today in a multiplicity of technical devices and mechanisms:

Here only an exemplary selection can be called - it gives both to thatareas mentioned a multiplicity of further applications and a row here areas of application mentioned, like medical technology, building engineering and much more besides.

elements

the microelectronic elements can be divided first in two large groups: Standard modules and user-specific ones:

  • Standard ICs leave themselves in a multiplicityfrom applications, in large number is produced and is partly also actually by standardisation consortia in their development fixed (demarcation in practice in a diffuse way) microprocessor
    • (MPU - Micro Processor unit) begins - of circuits, which do not have a clearly defined entrance output relation, but a succession of instructions for operation - inProgram - micro CONTROLLERs
    • ( MCU - Micro CONTROLLER unit) process - in chip computer system, which contains a microprocessor, different interfaces (Peripherals) and a memory ( RAM and ROM).
    • Digital signal processors (DSP - digital signal Processor) - similarly a micro CONTROLLER, but specializes in signal processing (digital representing more similarlyData, e.g. Music)
    • dynamic memory modules - inexpensive memory form, those predominantly as main memory use finds
    • bottle EPROMs - non volatile memory, which can be written only block-by-block, frequently in the form of memory cards for the storage of digital music, etc. begun
    • other memory modules: SRAM, ROM, PROM, EPROM
    • Programmable logic components (PLD - programmable logic DEVICE) - integrated circuits, which after completion still for arbitrary specific logic functions can be configured. The FPGAs (field programmable gate array) represents a special case, whose functionality can practically arbitrarily often be changed. But become upthis chips uses non volatile memory or external configuration memory.
  • User-specific ICs: Circuits, those for a certain application (z. B. Motor control in the car) sketched and built are and of its function range no other application permit.
    • Application-specific standard products (ASSP - application specific standard product) - a logic circuit, itsFunction is firmly defined, however by several customers to be used can (z. B. Application processor for mobile phones)
    • customized integrated circuit (ASIC - application specific integrated circuit) - a logic circuit, whose function is likewise firmly defined, however particularly for a customer and/or. an application sketchedand is built (in practice with this term frequently also all fixed logic components are associated - ASIC and ASSP)

apart from this systematics a group by elements is usually specially specified:

  • Similar ICs: integrated circuits, the similar signals, like tensions or rivers, directly - D.h. without transformation into digital signals - process.

finishing technique

one differentiates between the solid technology (also: monolithic manufacturing) of the film technologies.

  • With the solid technology the elements are applied on a semiconductor substrate, usually single-crystal silicon - wafer so mentioned. The structures of the circuitsthereby by photolithography are defined and structured with physical and chemical processes either or submitted of a local change of the material properties. From bringing in foreign atoms (doping) into the silicon for example leading and non conductive areas result, put one over it in suitable form oneConductive strip, those by a very thin insulating layer (z. B. an oxide) is separate from these areas, develops a field-effect transistor. On a wafer several integrated circuits (English usually become. IC or chip) at the same time manufactured. The chips are separated (in the form as thosedesignated) and usually into a housing (English:Package) inserted. The connections of the housing are connected by thin gold or aluminum wires (bonding wire) with the Pads in such a way specified of the integrated circuit (Bonding). The housing protects the chips, exhausts the warmth andalso further ICs can be used, on a printed circuit board as well as other construction units. In the detail: see to semiconductor technology
  • thin-film technology and thick film technique, with which construction units are applied or embedded and connected on a film, having only for special's applications (high-frequency engineering) meaning. Hybrid technology one callsthe combination of the film technology and solid technology. In solid technology manufactured semiconductor chips on a substrate are applied and then in film technology on it z. B. Feeder lines and passive elements realizes.

development of microelectronic elements

with the chip draft goes it around it, the basic elements of the microelectronic circuits - Transistors, passive construction units to locate, to link conductive strips - logically to the desired function geometrically on the silicon surface and to model physically their behavior. The specific characteristics of microelectronics led to the fact that a special draft process was formed.

  • Go to the production of a chipvery high a mark costs (NRE - non recurring engineering costs) ahead (e.g. To mask costs, see photolithography). Also a repair of an integrated circuit is only very reduced possible and productively not practicable. Therefore it is of great importance that the draft only with few revisions (so-called. Redesigns)to the desired product leads. That has the consequence that a substantial portion in the year 2004 it makes simulation - and verification steps the trend determine - them about half of the developping costs for the circuit out with rising tendency.
  • The reduction and rising integration leadto an immense number of realizable functional elements (= usually transistors; several millions in the year 2004) in an integrated circuit. To this extent the transistors no more cannot be converted at justifiable expenditure “by hand” into a circuit. Therefore draft automation continues to gain ever significance.In many cases the chip developer describes the desired circuit only in a 'high-level language '(comparably the higher programming language in computer science, usual developments: VHDL, Verilog), the computer calculates by it the combinatorial circuits (so-called. Synthesis) and platziert the transistors (under human co-operationand control).
  • The progressive miniaturization drives both the structuring processes and the realized functional modules, like transistors and conductive strips, to its physical borders. In order to meet first one, in the draft process to rising extent software is used, those the physical border effects, like e.g. the optical diffraction with that Photolithography simulated and the circuit draft modifies in such a way that these effects become balanced (so-called. OPC, Optical Proximity Correction). In order to work against miniaturization effects with the elements, sequentially new simulation and modelling procedures for the chip draft process are added: for example Simulationen of the voltage drop in long supply networks (IR drop), simulation of the parasitic kapazitativen coupling of neighbouring conductive strips, tools for the static analysis of time conditions in a digital circuit (so-called. STA, statics Timing analysis) and so on.

reasons and consequences of the miniaturization

since Gordon of moorlands 1965 after it 'law 'designated formulated thatthe number of transistors on a chip every twelve months (later eighteen months) doubles itself, microelectronics actually both in view to device complexity and with the reduction of the structures continuous progress made.

The driver for the reduction of the structures is the lowering thatProduction costs. The manufacturing of microchip takes place on disks (wafer) of constant size (there is actually an evolution of the pulley size over the time, but those is sufficiently slow for this view). Therefore the production costs can be described first as sum of the process costs per wafer (alsoa certain abstraction and under neglect of the costs of tests and housings of the ICs). To that extent there are two levers on the production costs per chip:

  • Number of chips per wafer
  • sum of the costs of the structuring and characteristic-changing processes per wafer

first takes by the minimizationthe structure width superproportionally too (surface reduction = Längenmassreduktion^2 + improves utilization of edge - nonlinear effects).

Secondly (the process costs of the structuring processes and the number of necessary process steps) increase however with rising miniaturization usually likewise. Until today (2004) the semiconductor industry thereby has an average production costs reduction of 30% perYear reaches. A possible scenario can be that the increase of the process costs can be compensated during approximation to physical borders no longer sufficiently by the saving of the Chipfläche and the progress in the miniaturization slows down thereby and/or. ends. In addition some special circuit technologies (z leave themselves.B. do not down-scale such to the achievement of higher tensions than the supply voltage (EN) of the chip) today (2006) already no more.

The high constant production costs reduction in microelectronics was a substantial innovation engine of the last thirty years in a multiplicity of industries - not alone in electronics and computer engineering(see applications).

Apart from the reduction of costs gives it in addition, other effects to the miniaturization:

  • Smaller logic elements, which are operated with reduced tension, to have also a substantially reduced energy dissipation (the energy dissipation per surface rises however -> worse heat dissipation). Without reduction and integration would not be battery-operated, mobile electronicsconceivablly, how it is pervasive today:Mobile telephone, Notebook, PDA etc.
  • Smaller transistors have improved switching times, shorter conductive strips shortened signal running times: by the reduction ever faster and thus usually also more efficient circuits become possible. However there are effects also moving in opposite directions: The shortened distances betweenthe conductive strips lead to higher ouple capacities, which work as run time brake.
  • Higher integration (more functions on a chip) means fewer elements on a printed circuit board and thus increased reliability by fewer solder joints.

economics

the branch of industry, that itself with the production of microelectronicConstruction units busily - the semiconductor industry - shows two characteristics, which it from others differentiates.

  • Large scale effects: Semiconductor factories for the mass production of components of the smallest possible in each case structure sizes are only profitable starting from a certain size. Further these factories are more expensive around orders of magnitude than comparable production plants of other industries:today (2004) the costs of building and equipment of a high volume factory are on the state of the art with for instance US$ 2 billion. Both together leads to the pig cycle in such a way specified: There is only a comparatively small number of current semiconductor factories world-wide. If it the industrywell goes (D. h. usually, if the offer at semiconductor components is smaller than the demand), she develops her manufacturing production lines, because most enterprises can apply only then the sums for the development. Each new factory, which goes into production, increases the world market volumeavailable components equal around per cent points, since they must be very large for profitability reasons. The sudden rise of available volume leads to an accordingly strong price purge of the elements, which in-swings again, as soon as the demand caught up the offer again. Are by the price purge many enterprisesnot to develop a time long in the layer their manufacturing production lines - it approaches on the next shortage of the offer. Then the cycle repeats itself.
  • Writing up of strategic meaning: Many states attribute strategic meaning to the semiconductor industry. Usually justifies itself in the “germ cell effect” for othersHigh technologies: In the surrounding field of semiconductor industries highly-qualified suppliers from chemistry and the equipment construction, but also from the customer industries of the elements do not only develop, for example the computer and electronic industry. In some cases the strategic meaning is also militarily justified: Thus the USA estimate thoseMeaning of microelectronics for arms programs so importantly that both devices for the production of current ICs and the circuit drafts and even the circuit development software article of their export control lists are. This highly estimated meaning has the consequence that a multiplicity of States of the settlement of the semiconductor industry in variousWay promote: research nationally promoted by Anschubfinanzierungen, special tax organizations, national credit warranties up to universitären on and industriellen research centers, etc. These promotions are also occasionally subject to economic arguments between states and enterprises - in such a way happen last in the year 2003: after the DRAM - manufacturer Hynix accusedover the masses by the South Korean state in its financial crisis to have been supported, obtained the USA, the European union and last Japan penalty duties on the products of this enterprise, against what South Korea protested.
  • Business models: As in many other industries there is also the full manufacturer - IntegratedDEVICE Manufacturer (IDM) mentioned. A IDM provides the product Design, develops the production technology, manufactures the construction unit and sells it. Besides there is however also still the “Fabless Design Houses” and” Foundries “. Fabless Design Houses provide according to product Design the defaults of the Foundry,it later manufacture and will sell the finished product. The Foundry develops the production technology, makes to its customers available technology-specific aids to the chip draft (EDA) and manufactures the ICs. Combinations of these business models and niche models are to be in practice also found.

History

  • 1904 the first electron tube, a thermionic diode is developed by John Ambrose Fleming.
  • 1906 of Lee de Forest are modified Flemings thermionic diode. It supplemented these around a third electrode and created thereby the triode tube, the forerunner of the transistor.
  • 1946 in the USAthe large computer ENIAC is taken in enterprise. He works with 14468 electron tubes.
  • 1947 John Bardeen, walter Brattain and William Shockley discover the transistor effect and present
  • to 1948 the first transistor from germanium .
  • 1954 the first transistor from silicon one introduces. William Shockley opensLaboratory near the Stanford university in Palo Alto, which is considered as germ cell of silicone Valley. In this year also the contest begins raw material around the better< semiconductor> b -: Germanium or silicon.
  • 1956 William Shockley receives a Nobelpreis for his work. Shockley regards germanium as thatbetter raw material, its pupil silicon.
  • 1957 Shockley' s pupil separate from their teacher. They are called “eight traitors”: Eugene Klinger, Jay load, Victor Ginrich, Jean Hoerni, Sheldon Robert, Julius blank, Gordon E. Moorlands, Robert N. Noyce.The “eight traitors” form the Fairchild Semiconductor corporation with a risk capital.
  • The first transistor based from the raw material silicon makes 1958 Fairchild available Semiconductor in series numbers of items.
  • Consequence pulls 1959 Jack Kilby, employee with Texas Instruments from the fact that transistors such as resistancesfrom only one material to be manufactured. It accommodates for the first time a circuit from different components on only one substrate. Thus the integrated circuit (IC) is invented. In September it demonstrates then the chip, on basis of a germanium substrate. From this work the famous developsKilby patent 320,275. About this patent then approximately ten years before court one argues, because Robert N. Noyce (joint founder of Fairchild Semiconductor) devised a very similar step, this to the patent announces however later. Robert N. Noyce leaves Fairchild, over together with Gordon of moorlands to base the company Intel in Santa Clara.
  • 1960 Jean Hoerni invent the planar process and revolutionize thereby the semiconductor manufacture.
  • 1961 first integrated circuits with few bipolar transistors and resistances are realized.
  • starting from 1970 higher transistor densities with some thousand elements are realized. This becomes as large-scale integration(LSI Large Scale integration) designates. The bipolar transistors are usually replaced by field-effect transistors FETs, in the form by easily producible MOSFETS (Metal of oxides Semiconductor Field Effect transistor).
  • 1979 it begin the grösstintegration (VLSI Very Large Scale integration) with today several millionsTransistor functions and in former times unattainable clock frequencies of several gigahertz. The size of the independent device is thereby a far under square micrometer. Increasingly also whole systems (combination of several building groups, like processors, interface circuits and memory) on an individual chip are realized (SoC system on chip).
Work on []

Future developments

  • nano-electronics: If the structure sizes fall below the 100 nanometer - border (2004 is them already with 90 Nm), then one formally already speaks of nano-technology (definition of US government). Strictly speaking it is however rather meant that special material properties are used, which arise only ifitself the design geometries in the proximity that molecule - and/or. the atomic size move. Among such structures rank for example line courses or transistors from carbon nano-tubes or Isolationen from Self Assembling Monolayern.
  • New elements are developed with resonance tunnel diodes.
  • Integrated optoelectronics: In view of increasing signal running times in particular inlong feeder lines (global Interconnects) large systems on chip one thinks to replace these conductions by optical connections.

literature

  • Simon M. Sze: Physics OF Semiconductor DEVICE 2. Edition. John Wiley and Sons (AS) 1981; ISBN: 0471056618
  • Ulrich Hilleringmann, silicon semiconductor technology, Teubner 2004, ISBN 3519301490
  • Ulrich Tietze, Christoph give, Eberhard Gamm, monolith technology, Springer 2002, 12. Edition, ISBN 3540428496
  • Michael Reisch, semiconductor components, Springer 2004, ISBN 3540213848

see also

 

  > German to English > de.wikipedia.org (Machine translated into English)