Phys Techno1 Vol 16 1985 Printed in Northern Ireland

HISTORY OF RESEARCH LABORATORIES: 3

GEC HIRST RESEARCH CENTRE

Nearly 70 years ago a small laboratory was planned for research into incandescent lamps. Since then, during peace and war, it has grown to become one of the premier research centres in the UK, with R and D activities across a wide range of new and not-so-new technologies

R Clayton and J Algar

76

The idea for the research laboratory which became the GEC Hirst Research Centre had originated by 1916 when the management of the Osram Lamp Works in the UK asked Clifford Paterson, then a principal assistant at the National Physical Labora- tory, to recommend someone to organise a research department for them. The Osram Lamp Works had been founded jointly by DGA of Berlin and GEC of England, but the first world war made the GEC management recognise the dangers of necessary scientific resources being outside this country and they decided that the company must no longer be in this position. Implementation of the idea had to await the end of the war. Then, when Paterson asked the Osram management if they wished to reconsider the project, he was told that, far from there being any weakening of intention, a bigger scheme was envisaged. The German holding in Osram in the UK had been purchased and the company was to be amalgamated into GEC. It was now proposed that the new research laboratory should serve the whole of GEC, and Paterson was asked to become its first Director.

The Laboratories were to be located on a site within 20 miles of Hammersmith, and with a good train service to London. Luton and Wembley were considered but Paterson objected to Luton on the grounds of the distance which staff would have to travel to attend scientific meetings, and Wembley

0 The General Electric Company Ltd Middlesex 1985

1 Figure 1 Section through the Laboratories at the time of their opening

was chosen. Paterson took a considerable hand himself in the design of the building, stipulating that there must be facilities for piping and wiring supplies to individual laboratories, offices for thinking and writing, and even space for junk (see figure 1). A nucleus of staff started work at Hammersmith and moved to East Lane, Wembley in 1922. The Laboratories were formally opened on 27 February 1923.

The 1920s The first work of the Laboratones was connected with filament lamps, but even before the new building had been designed, extensions were proposed for work on batteries, transmitting valves and high voltages. Approaches had also been made by the electromedical, mechanical engineering and instrument units in GEC. Work on illumination and the application of lamps started at Wembley towards the end of 1922. Although there was deliberately little formal organisation. much of the work was in teams concentrating on lamps, valves and fundamental vacuum physics; the valve team subsequently divided into a further group on communications and broadcasting.

GEC and the Marconi Company set up a valve manufacturing company at Hammersmith, based on GEC expertise, and the Wembley laboratories were responsible for research and some develop- ment on its behalf. Work on vacuum physics, metallurgy and cathode emission led to a series of successful glass envelope valves for communica- tions and broadcasting. A major breakthrough was the cooled anode valve depending on glass-metal seals. By 1930 world leadership had been achieved and the CAT12 transmitting valve was operating the transatlantic telephone link. The even larger CAT14 valves, each weighing about 130 kg. were first made in the Laboratories. They made their debut in the BBC Droitwich long-wave transmitter

in 1934 and for at least ten years were the most powerful valves in the world.

Work on heating and domestic appliances, involving a mixture of physics and engineering design, started in the mid 1920s. At about the same time, demands for cable loading coils and telephone transmission equipment led to the establishment of the telecommunications research group. Early in the history of the Laboratories support groups were set up for metallurgy, glass, engineering design and workshops, materials characterisation and patents. Apart from these the organisation was kept as flexible as possible, groups corresponding to the needs of particular G E C units being brought together or reoriented from time to time. Close liaison between a research group and its opposite numbers in the unit was considered to be of paramount importance, helping the works to produce new and improved products and processes and to reduce costs, while in turn providing a stimulus for the research group.

The 1930s The beginning of the second decade coincided with the industrial depression. While the Laboratories were called upon to exercise all possible econo- mies, their work was not allowed to be affected, and was actually expanded. This decade produced advances in discharge lamps, their application to street lighting, in fluorescent tube light sources, and in colour photometry. Meanwhile, research was expanding on line communications. radio com- munications, broadcasting and television. Work in the metallurgy group led to advances in industrial furnace techniques, and another group applied aerodynamic studies to turbine problems.

Research which was to prove of value in the future was J W Ryde’s work on translucent media, as was B P Dudding’s and W J Jennett’s applica- tion of statistical methods to industrial products

77

and operations. The work of the valve groups was extended to cathode-ray tubes, higher fre- quencies and new types of valve. The microwave frequencies work of E C S Megaw and his colleagues, including early magnetrons operating down to wavelengths of 3 c m , was also to be particularly important. Another feature dating from the earliest days which was to be of unexpected significance was the concept of laboratory-factories or try-out units at Wembley. In 1938 the Admiralty asked Paterson if the Laboratories would take part in the development of some special valves for secret applications. This proved to be one of the first contacts with what is now the Components, Valves and Devices (CVD) organisation of the Ministry of Defence.

The 1939-45 war O n the outbreak of war the G E C management put the Laboratories at the disposal of the government to carry out work for the services. All the groups engaged on valves and cathode-ray tubes now turned to devices for radar and communications. Among wartime developments were vacuum and gas-filled radar modulator valves, rugged valves for use in proximity fuses in shells, and disc-seal triodes and associated circuits for both cw and pulse operation, some of which were capable of working up to 3000MHz, while pairs of others produced peak pulse powers of 200 kW at 600 MHz. It soon became clear that the needs of the services were so urgent that, for many programmes, there was no time to transfer designs to a factory. at least until early demands had been met. Preproduction units, based on the Laboratories’ earlier experience of try-out units, were set up; a thousand newcomers joined the staff and they produced over 300000 valves of 45 new types.

Wembley staff first heard of the work of Randall and Boot on magnetrons at Birmingham in April 1940 and realised that it was a major advance which, when married with existing techniques, could lead to still greater progress. E C S Megaw suggested ways in which the design could be improved and simplified and the magnet weight greatly reduced. A sealed-off design was prepared which incorporated, among other modifications, gold seals in place of soldered joints for the copper end plates of the envelope. Its performance proved to be similar to that of the Birmingham cw demountable valve.

Urgent need for a lOcm pulse transmitter for airborne radar led to consideration of whether a design based on the Birmingham valve could provide a more powerful and lighter source than that which was already available. A design was worked out using a block with cross-sectional

78

dimensions nearly the same as those of the Birmingham valve but with the ratio of anode length to diameter chosen to give a compromise between power output and magnet weight. The relative dimensions of this valve set a standard which was not widely departed from throughout the war. By June 1940 an output of about 1 kW peak at 9.8 cm wavelength was achieved at Wembley, shortly followed by 5 kW, and by the end of 1940 pulse outputs up to 100 kW were being obtained. The Laboratories continued to play a significant part in the national effort on magnetron develop- ment throughout the war, producing valves with pulse powers of 2 M W peak at lOcm, 200kW at 3 cm and even a few kilowatts at 8 mm wavelength.

Radar In November 1939 Sir Henry Tizard visited the Laboratories and stressed the vital importance of ‘radio location’ (radar) as a war weapon. A meeting with R A Watson-Watt provided information and terms of reference for work on a pulsed radar for airborne interception (AI) at wavelengths not greater than 30cm with a range of 5-10 miles (8-16 km). By February 1940 complete schematic and circuit arrangements had been drawn up for a scanning aerial system, and construction followed immediately. By March 1940 a 25 cm system was demonstrated to the Air Ministry.

By that time work on mixers using silicon crystals had started, and in early April 1940 the first comparisons were made between crystal mixers and concentric diode mixers in a pulse system receiving reflected echoes. Significantly better performance was observed with the crystal mixer and from then on it steadily improved its lead. In June 1940 sleeve couplings and contactless switches were invented and patented by D C Espley. These were to be of considerable importance in rotating centiinetric scanning aerials and in other microwave systems both during and after the war. In mid-1940 the efforts of the Wembley team were combined with those of a team at TRE Swanage (now RSRE Malvern) in an endeavour to get results as quickly as possible, and by February 1941 Research Laboratory staff had participated in the first flight trials of centimetric AI from Christchurch. The aerial scanners ultimately used in AI were based on a design by A L Hodgkin of TRE (now Sir Alan Hodgkin FRS) but a G E C scanner was installed in an aircraft and was the first to be used for the development of the bombing navigational system H2S. Airborne radar work continued with the development of units for a second engineered system (AI Mk VIII) which were delivered to Christchurch in November 1941. Members of the Wembley radar team later contributed to a

centimetric version of the blind bombing system, Oboe, and to the microwave aerial design of a number of other bombing and navigational aids.

Torpedo guidance In 1942 some RAF officers from Technical Training Command described to Lord Hirst , the Managing Director of G E C , their ideas for combating the U-boat menace. H e put them in touch with Paterson who arranged discussions with Wembley staff. It was found that most of the ideas were not valid, but there was one possibly feasible idea which used spaced ultrasonic transducers. transmit- ting and receiving short pulses, and measuring the phase difference between the echoes received from the target to determine its direction. An audio- frequency model was built for demonstration in air and shown to senior R A F officers and to the Admiralty. It was agreed that G E C should work on the feasibility of phase-controlled homing, either passive (i.e. using submarine-generated noise) or active, in an underwater environment. The R A F officers were seconded to work with the G E C team. Investigation proved that the scheme was feasible and, before the end of the war. models of a battery-driven passive homing torpedo had been completed and had homed successfully on an artificial target. G E C subsequently manufactured several hundred sets of electronic equipment for this weapon which was the first homing torpedo carried by the RAF.

Post-war changes In 1948 the then Ministry of Supply suggested to G E C that there should be a substantial increase in defence research and development at Wembley. The G E C management rejected this proposal but after considerable discussion said that it would be prepared to set up and manage a laboratory on behalf of the government. This offer was accepted and a large part of the team which had worked on centimetric radar and, since the war, on broad- band microwave communication became the nucleus of a new laboratory at Stanmore. This remained part of the Research Laboratories for a number of years. It then became one of GEC’s Applied Electronics Laboratories before also becoming the headquarters of Marconi Space and Defence Systems, and now of the Marconi Company. A further evolution came from the recognition in the second half of the 1950s of the importance of digital computers. G E C made an agreement with ICT (now ICL) for joint development of the ICT 1301. This took substantial effort in the Research Laboratories and. in addition, a number of staff transferred to a joint GEC-ICT system development company.

Although in the immediate post-war period the Laboratories had re-established close links with G E C units, it was recognised in the late 1950s that the size of the Laboratories and diversity of G E C interests called for a change in organisation. Sir Olliver Humphreys, who had succeeded Sir Clifford Paterson on the latter’s death in 1948, introduced the concept of a ‘research centre’ with laboratories for valves, telecommunications, semi- conductors, and lamps and lighting, each having a close relationship with the corresponding G E C operating unit. A Central Research Laboratory was to undertake most of the substantial amount of research still paid for by G E C centrally, and provide scientific and technical services for the whole Centre.

About ten years later GEC acquired A E I in 1967 and merged with English Electric in 1968. Inevitably there were overlaps of programmes and facilities, and changes in company interests and organisation as a result of these mergers. Eventually, under Sir Eric Eastwood FRS as G E C Director of Research, the Hirst Research Centre and the Marconi Research Laboratories became the GEC’s two major research centres, with, in addirion, Mechanical Engineering Laboratories at Whetstone and Stafford, an Electrical Engineering Laboratory in Stafford, and a number of individual research laboratories in operating units.

Contemporary research Ending research programmes in an establishment is said to be one of the most difficult tasks for research management, even when related research programmes are to be pursued elsewhere. Follow- ing the merger with English Electric, GEC had two valve companies (EEV and MOV). E E V had its own strong research and development teams; MOV relied on a large research laboratory at Wembley, although most of this was not on the main site. It was decided that the time had come for valve research as well as development to be concentrated in the two operating companies; many of the research staff moved to MOV at Hammersmith and valve research at Wembley substantially came to an end in 1971. A similar decision on lamp research was made in 1980 and work on this subject (for which the Laboratories had originally been conceived) came to an end after sixty years, programmes being taken up by the Osram Development Laboratories, although the Research Laboratories continue to provide scientific support and services for both valves and lamps. The following sections describe some research program- mes undertaken at Wembley in various fields, and conclude with notes on present day activities.

Work on x-ray crystallography started at Wembley

79

in the 1920s under D r F S Goucher, who had worked at Manchester with W L Bragg, and centred on the properties of tungsten, particularly changes which took place as crystals in wire deformed under load. Goucher left to join Bell Telephone Laboratories leaving H P Rooksby to continue the work. In the 1920s the work was done on x-ray diffraction equipment built in the Laboratories since no suitable equipment was commercially available. There was then no ASTM file and the Laboratories assembled their own library of standards which was still consulted on occasion in the 1970s.

In the early 1930s studies were made in attempts to elucidate the structure of glasses; on character- isation of ruby selenium glass (leading to explanation of its variability, depth of colour and the necessary constituents); and on barium and strontium oxide cathodes which showed the composition and structure for best emissive performance. The latter work was continued during and after the 1939-45 war. Also in the early 1930s, x-ray diffraction, spectroscopy and chemical analy- sis were first used in the integrated approach to materials characterisation which has been one of Wembley’s strengths. X-ray diffraction was an important tool in the Laboratories’ work on phosphors leading to the key G E C patent on halophosphate phosphors in 1942. In this field the Laboratories started using diffraction spacing data to identify the subject of a patent specification-a novel idea which was accepted by the Patent Office. A noteworthy observation was the identification of what proved to be the first halophosphate phosphor, a crucial point which might otherwise have been missed.

During the war, studies started on nickel-iron powder and orientation of quartz oscillator crystals, and later on barium titanate and ferroelectrics. Observation of the abnormality of structure of barium titanate led to a letter to Nature in 1945 on structure distortions in members of the perovskite family. The investigations on nickel-iron powders led to observation of an anomaly in the structure of nickel oxide which was later recognised to be an antiferromagnetic substance. In the 1950s, work started on carbon and graphite, including lattice thermal expansion measurements from - 190 “C to 2500 “C. Investigations on structure changes in a number of oxides at high and low temperatures were also made. In the 1960s it became necessary to examine the perfection of many types of crystals, particularly semiconductors, and work on diffrac- tion topography was started, including the develop- ment of a special goniometer for topography.

Although the x-ray crystallography group has continued to progress in a number of areas, the

80

emphasis has been on development of instrumenta- tion and processes for characterisation of single crystals in bulk and thin film forms. X-ray topography is still being developed to keep pace with requirements to handle larger crystals and faster throughput of samples, requirements which are met by the Hirst Topographical Camera now sold commercially (figure 2).

The increasing perfection of crystals now necessitates the use of double-crystal techniques which permit the study of crystals with strain gradient sensitivity of 1 part in lo7 as compared with 1 part in lo4 achievable with conventional single-crystal instruments. This improved sensitivity can only be achieved and exploited with micro- processor-controlled precision goniometers and novel x-ray optics. In recent years thin films- both single and polycrystalline - have become more widely used, and a range of specialised x-ray instruments and techniques has been evolved for characterising them. These include the G E C X-ray Texture Camera (primarily for the assessment of preferred orientation in polycrystalline films and of hetero-epitaxial relationships in single-crystal layers). X-ray multiple diffraction procedures have been developed for detailed symmetry and lattice dimensional studies on hetero-epitaxial systems; and double-crystal-based techniques are being developed to study structural modifications caused by ion implantation in semiconductor crystals and hetero-epitaxial systems of thickness 1-10 nm.

The function of the x-ray crystallography group continues to be to respond to problems in G E C product units and research centres. At the present time there are short- and long-term studies on a wide range of materials from amorphous metals and glasses to large crystals. Looking to the future, plans are advanced for exploring ‘critical angle x-ray diffraction’ for characterisation of surfaces and very thin films on the atomic scale.

Atmospheric scattering of radio waves As early as 1933 J W Ryde carried out theoretical studies on the diffusion of light waves by opal glasses, i.e. glasses containing particles which are themselves transparent but are of different refractive index from the host medium. In one of his papers he commented ‘In the same way, cloud or thick mist may strongly diffuse the light falling on it, although the individual water droplets are clear’. When centimetric radar was being de- veloped it became important to be able to predict attenuation and backscattering in the atmosphere for different weather conditions. Ryde started a theoretical investigation in 1940 and by the summer of 1941 had provided detailed estimates of the magnitudes of the effects in the centimetre and

Figure 2 Hirst x-ray diffraction topography system

decimetre wavebands for the cases of fog, cloud, rain, hail, and desert sand and dust storms.

The work was later continued in collaboration with other establishments which undertook work to determine the complex dielectric constants for water and ice and to provide meteorological data. When the first theoretical work was undertaken there was no apparatus readily available for an experimental approach, nor time to await the arrival of the necessary weather conditions for each experiment. When, however, subsequent observa- tional determinations were made, the results tended to confirm the early estimates, at least as regards their orders of magnitude. Although the work was subsequently published, the early reports of which J W Ryde and Mrs Ryde were authors remained of considerable interest. Copies have been requested from all over the world, inquiries being received at Wembley as recently as April 1984. Ryde was elected a Fellow of the Royal Society in 1948.

Interest in radio wave propagation in the atmosphere continued after the war. In particular, R G Medhurst extended Ryde's work during the 1960s and commented on reasons for some discrepancies between theoretical and observed results on attenuation of centimetre waves due to rain, while E M Hickin undertook both calculation and experiments on atmospheric attenuation of radio signals at frequencies above 10 GHz.

Phosphors Phosphor research began in 1934, interest having been created by the introduction of cold-cathode neon sign tubes coated with sulphide phosphors which extended the range of colours available for display purposes. Such phosphors were available commercially largely as a result of German work, and the first GEC efforts were directed to preparing materials equivalent to those from Germany. It took some years to equal their light

output but, in the meantime, effort had been extended to the non-sulphide (oxygen-dominated) phosphors which were potentially of greater importance because they responded to the short-ultraviolet emission from low-pressure mer- cury discharges whereas the sulphide phosphors were excited by long-ultraviolet radiation.

Attempts to synthesise non-sulphide phosphors were originally restricted to the known fluorescent minerals such as willemite and scheelite. In attempts to synthesise willemite the Wembley laboratories produced a very pure form of silica which was also very reactive chemically, and this in turn gave rise to a highly efficient zinc silicate phosphor. Lamps coated with this phosphor raised the efficacy from about 20lmW-' with sulphide phosphors to about 50 lm W-l, and established that fluorescent lamps given the right phosphors would be capable of a high light output.

Research was then concentrated on this type of material with the object of extending the colour range. Tungstate phosphors based on scheelite such as calcium tungstate (deep blue) and later magnesium tungstate (pale blue) were prepared. This was especially important as this colour combined with that of pink-fluorescent zinc beryllium silicate gave the effect of white light in fluorescent lamps. Within about three years the Laboratories had produced the main phosphor ingredients for the fluorescent lamp (zinc beryllium silicate and magnesium tungstate). The American General Electric company with its much greater resources overcame the practical difficulties of producing fluorescent lamps commercially and announced them in about 1938, the first GEC production being in 1940.

Research into other oxygen-related phosphors led to the discovery at Wembley of the first halophosphate. By chance a new phosphate phosphor, cadmium chlorophosphate, was pro- duced which although commercially not important was the forerunner of the important main halophosphate phosphors. The structure of the new cadmium chlorophosphate was elucidated by using x-ray analysis. In 1942 further work on halophos- phates led to the discovery of the new and very important calcium halophosphate phosphors, acti- vated with antimony and manganese. These were capable of giving a range of white colours of considerably higher efficacy than current lamps. The maintenance of light output during life compared very favourably with the then current magnesium tungstate-zinc beryllium silicate mix- tures. Moreover, they were relatively easy to manufacture from cheap starting materials, their disadvantage being the relative lack of red emission in their spectrum. GEC was the first company to

81

Figure 3 Repeater station for the reversible London- Birmingham television link

introduce the halophosphate phosphors in commer- cial fluorescent lamps in 1946.

Broadband microwave communications The group which had worked in television broadcasting before the war had been the nucleus of the team on airborne centimetric radar. After the war it returned to television broadcasting, the majority of the team embarking on research in which their wartime techniques could be applied. When television broadcasting was restarted it was still only available in the service area of the London transmitter. Before it could be received in Birmingham, for example, the television signal had to be carried there by cable or radio transmission. The research programme concentrated on radio transmission at carrier frequencies of the order of 1 GHz, using disc-seal triodes, and coaxial circuits, feeders and aerials. This research continued for some two years until in 1947 the Research Laboratories bid for and won the Post Office contract for the first wideband repeatered radio link in this country. For the next two years the team, built up in the Laboratories, undertook research and development, supervised manufac- ture and was responsible for installation and testing. The target date of December 1949 was

82

achieved with a one-way switchable, reversible link and the BBC was able to open the Birmingham service on schedule (figure 3).

This work was important not only for the success of the project itself but because it was the beginning of a Wembley programme on wideband microwave radio-relay systems for television, telephony and data, which has continued until today, and which has been the basis of a successful GEC business, both at home and overseas. In the 1950s interest was particularly keen in the 2 GHz band. Propagation trials were carried out and R G Medhurst published theoretical papers on perform- ance and echo distortion in FM systems which led to the award of the IEE Heaviside Premium. In the 1960s the research programme was extended to the 6GHz band, and propagation tests were carried out at 19 GHz in the belief that this band would be important in the future.

In the early 1970s there were studies of the application of new microwave devices. Theoretical studies were carried out on the performance of microwave radio systems carrying digital signals and of possible interference between such systems on adjacent transmission frequencies. In 1975 a 19 GHz system carrying a 140 Mbit/s digit stream was demonstrated. The work on microwave radio-relay systems is an excellent example of the benefits of a multidisciplinary research centre in a company such as GEC. There were in the Laboratories, in addition to the telecommunica- tions groups, teams working on materials and on active and passive components for microwave applications; their programmes benefited greatly from the frequent interaction between them.

Work started on ferrites in the 1950s and a range of low-loss ferrites with properties optimised for use at different frequencies was produced. Ferrite

Figure 4 A down converter from RF to 140 MHz intermediate frequency for an 11 GHz radio repeater

isolators were developed for use in 2 GHz systems and then waveguide isolators and circulators for 6GHz systems. A world lead was established in the early 1970s with hybrid microwave integrated circuits incorporating non-reciprocal circulators based on ferrite pucks in alumina substrates. All the radio-frequency stages in microwave systems at 6, 11 and 19GHz have since been converted to microstrip on substrates incorporating embedded ferrite circulators and isolators (figure 4).

Other teams at Wembley were engaged on work on microwave semiconductor devices. Work on semiconductor diodes, for example, started as early as the beginning of 1940 (when the team on aerials for airborne radar sought a mixer, other than the thermionic diodes which were in short supply, for their polar diagram equipment) and has continued ever since. Although much of the impetus for the work on microwave semiconductors has come from defence requirements, devices have been de- veloped for application in microwave communica- tion systems, culminating in hybrid microwave integrated circuits for use at frequencies from 2-19GHz, and more recently in monolithic microwave integrated circuits.

Optical communications Research on optical fibres started at the end of 1970. The programme has produced successful results on materials, fibres, cables, jointing and necessary measurements. A 'pull' of silica to fibre was achieved in 1973. In'the next three years steady improvement was achieved on both fibre and cables, fibre attenuation being progressively re- duced to 10 dB km-'. In November 1975 the first European 2 dB km-' high-strength fibre had been produced and the first optical fibre cables were made. Preform production and fibre-pulling equip- ment was then installed in a GEC factory (figure 5). In the same year an asynchronous Cambridge ring system using GEC's optical fibres was installed and a record-breaking demonstration was also given of 140Mbitls over 10km. The development of low-loss fibres, both multimode and monomode, and manufacturing techniques for them has been based on GEC's own research, the only coopera- tion having been with British Telecom.

The work on fibre and cables has been paralleled by research and development on equipment and systems in the telecommunications groups. In 1976 a 2Mbit/s system was supplied to British Rail for evaluation, and in 1977 a 2 Mbit/s field trial system was installed for British Telecom. By 1981 the Laboratories were working with other GEC units on installation of a 140Mbit/s system; at the same time work was started on monomode systems. The results of this work have shown that separations of

Figure 5 Pulling optical fibre from a preform

30 km between repeaters are possible, eliminat- ing the need for underground repeaters in this country. The research and development on optical fibres, optical fibre cables, and optical cable communication systems, again demonstrates the value of the existence at Wembley of groups in a variety of fields which interrelate on interdisciplin- ary programmes. The work was, in its early stages, supported by GEC central research funds and illustrates the value of such funds being available to the Research Centre.

The present The Hirst Research Centre is one of three centres which now constitute the GEC Research Labor- atories, the others being the Marconi Research Centre at Chelmsford and the Engineering Research Centre with laboratories at Whetstone and Stafford. The present programmes fall under the three headings: applied research, exploratory research and enabling technologies.

Applied research is carried out mostly in. laboratories with close links with GEC operating units. Examples of current programmes are millimetre-wave subsystems, nuclear magnetic resonance applied to medical diagnosis, continuing work on optical fibre based systems, and digital microwave transmission. Exploratory research is

83

Figure 6 One of two installations for molecular beam epitaxial deposition

undertaken to establish feasibility and relevance which would justify embarking on applied research and then development. It also forms bridges with the academic and industrial research communities throughout the world. The present programme includes work on submicrometre device physics, molecular electronics, laser photochemistry, sen- sors, systems architecture, intelligent knowledge- based systems, optical signal processing and perception. The third‘category of research is on ‘enabling technologies’, which are essential for, and common to, a wide range of new products and production processes. The technologies of informa- tion, manufacturing, control and materials are of particular interest to GEC.

The GEC Research Laboratories are heavily committed to participation in both the UK Alvey and the EEC ESPRIT programmes on information technology, which also have the theme of pre-competitive collaborative research with both universities and other companies. The GEC work includes microelectronics, software engineering, mass storage, man-machine interfaces, and exten- sion of the frequency band by use of microwaves, infrared and optoelectronics. Over f20m has been invested at Wembley in the last three years on research and facilities for the computer-aided design of silicon integrated circuits (SICS), electron beam mask-making, clean rooms and the special- ised equipment for making SICS - together with research on device physics, laser processing and novel computer systems architecture. Similar investment is now being made on compound semiconductors for optoelectronic and microwave applications (figure 6).

84

The programme on manufacturing technology includes work on computer-aided design and manufacture, flexible manufacturing systems, robo- tics, jointing and assembly, and inspection. The control technology programme covers work on control theory, sensors and actuators; the materials technology programme includes materials science, biotechnology, nondestructive testing, tribology, composite materials and surface chemistry. An important role for the Laboratories is to offer an attractive route for recruitment of first-class scientists and engineers, many of whom will not remain in research for their whole careers but will move on to GEC operating units.

Conclusion The Laboratories took little part in work on silicon integrated circuits in the 1960s because of a commercial decision that GEC would have a joint company with Philips who undertook the necessary research. The arrangement came to an end after GEC merged with the English Electric Company and acquired Marconi Elliott Microelectronics. By then there was substantial overcapacity in the world integrated circuit industry and the desired tech- nology was not that in which Marconi Elliott Microelectronics had been strong. Although research activity in SICS became the responsibility of Wembley, its outlet was only a small specialist company. Nevertheless, there were some worth- while research contributions, notably in charge- coupled device imagers, MNOS memory devices and CMOS on sapphire technology.

The arrival in 1979 of D H Roberts FRS FEng as Director of Research (and subsequently Technical Director of GEC), and the establishment at Lincoln of Marconi Electronic Devices Ltd as a production and commercial outlet, provided the impetus and opportunity for major expansion of work at Wembley on solid state research and development, and particularly on silicon integrated circuits which were of increasing importance for GEC‘s future. Since that time there have been major increases in staff and facilities at Wembley. A number of high-grade well known scientists have joined the staff from government service, the universities and other companies, from overseas as well as from this country. The Laboratories are now stronger than they have been for many years and should make a major contribution to GEC’s success in the next decade.

Acknowledgment We wish to acknowledge the help of many past and present colleagues who have drawn upon records and personal experiences.