Mannerisms

Sitting on a panel in front of 350 technologists at the MEDEA+ Forum in Monte Carlo, I was suddenly asked a question of which I had less than five minutes notice.

The question was: should pan-European R&D programmes, like MEDEA+, invite participants from other regions like China, the USA and Korea.

Thinking fast I stumbled out with the view that it might be useful to invite the USA but, seeing that neither China nor Korea had a university rated among the top 200 universities in the world, their participation was probably not going to add a lot.

The truth is that Europe and the USA are level-pegging in microelectronics R&D these days with the rest of the world lagging.

IMEC and the Albany cluster are neck-and-neck on EUV (with Japan a year behind, and they are much-of-a-muchness on 45nm.

Some would say IMEC has the advantage in terms of having a more catholic input of view, while Albany is dominated by IBM, but Albany can probably out-gun IMEC in the scale of its finance.

The opening afternoon of the 2006 MEDEA+ Forum, the best attended forum ever, took place on a warm, sunny, shirt-sleevey sort of day, and ended with a slap-up meal at Monte Carlo’s poshest hotel, the Hotel de Paris.

A word of warning. If you’re ever having rack of lamb at the Hotel de Paris, don’t bother to ask for mint sauce. They haven’t got any.

MEDEA+, the European microelectronics R&D initiative, has a tricky dilemma. As an organization which works on R&D which is four to five years away from the market, it has to make a judgment about when nanotechnology will become commercially feasible in relation to electronics.

In the case of nanotechnology, i.e. manipulating material at the atomic or molecular level, MEDEA, has to try and do better than the Americans who appear to have jumped in too early.

Biomimetics, it transpires at a pre-Christmas party, is the next big thing. Nature, the great exemplar of design efficiency, has evolved design efficiencies which we must follow to produce high performing products.

NXP, formerly Philips Semiconductors, is said to be considering leaving the Crolles 2 Alliance for IC R&D, and IBM or TSMC are being considered as possible replacements.

Samsung, has produced an LCD which can display different images
on the front and the back which is only 1mm thicker than a conventional display.

At Las Vegas airport I thought I had seen the future. ‘Send a digital postcard’ said the sign on top of a bunch of these things, with one in the middle showing a changing display of scenes of the city.

In the world's laboratories, which are increasingly tightly budgeted and directed, are there still places for the solitary geniuses who have historically made great contributions to technology progress?

It is a major blow for European high-tech as Crolles2 unravels with first NXP, and now Freescale, pulling out, leaving STMicroelectronics on its own with TSMC as a kind of junior partner.

With ST pulling out of Crolles2 for basic semiconductor process technology development, advanced CMOS process technology development in Europe will stop at the end of this year. Does this matter?

With NXP and STMicro pulling out of basic chip process development at Crolles, and with Texas Instruments deciding to abandon basic process development, the semiconductor industry is taking a dangerous route.

Sometimes, though not often, one feels that top execs earn their large modern rewards packages.

When did the accountants take over the semiconductor business? When we look back we’ll probably say Q1 2007, when NXP, Freescale, STMicroelectronics and Texas Instruments all decided to give up doing fundamental CMOS process R&D.

Most of semiconductor process development work is done by the production equipment companies nowadays, according to Wim Roelandts, CEO of Xilinx and a director of the world's largest equipment company Applied Materials, so it doesn't really matter giving up basic process development.

NXP has now said that all its production will be sourced from TSMC for chips with more advanced processes than 90nm.

Will PETeC, the proposed Plastic Electronics Technology Centre, get the creative juices flowing or will it turn out to be a bureaucratic condom?

The worst two whingers I've had this year have both been American, and have both been whingeing about the same subject, the pull-out from the basic semiconductor process R&D project at the European R&D centre at Crolles2

Before 45nm generation wafers are hot from the furnaces, the consortia for the 32nm generation are gearing up for commercial introduction around the end of the decade.

If ever there was a case for not giving up pursuing basic process R&D, as NXP, Freescale and STMicroelectronics have said they will do, it is the recent announcement of Fujitsu’s success in reducing leakage current at 45nm.

If anything goes to show the importance of pedigree and geography when it comes to chip technology, it is the link from Inmos, the 1970s start-up, to the fact that Bristol has more chip designers than anywhere else in Europe.

The scattering of the Crolles2 partners is now complete, with STMicroelectronics saying it will join up with IBM for basic semiconductor process R&D, following the example of its former Crolles compatriot Freescale which announced it was joining the IBM group some months ago. The third member of the Crolles2 triumvirate, NXP, departed Crolles to join TSMC earlier in the year.

It was tough to leave out some great institutions, like the University of Toronto, UCLA , and the Korea Advanced Institute of Science and Technology, but here are the top ten microelectronics R&D bodies.

Trawling around the antiseptic halls and corridors of the Novotel Congress Hotel in Budapest during the MEDEA+ conference, I learn that Europe gave up its microelectronics independence for a lousy $150m.

Why is pharmaceutical R&D so much slower to produce results than semiconductor R&D? In a recent speech to the Society for Neurosciences, Andy Grove, Chairman Emeritus of Intel, argued that the pharmaceutical companies have the wrong attitude.

God, how the Continentals love complexity. A new European R&D project, called ENIAC is about to be approved by the EC authorities.

Over the chip industry's history there have been an increasing number of R&D collaborations as the cost of developing the technology has become increasingly expensive. The ten most effective examples of these have been:

Europe’s share of worldwide chip markets has fallen below the level which, in 1983, triggered the thinking behind the Megaproject which, followed by JESSI and MEDEA, allowed Europe to catch up with the US and Japan in chip technology.

In 1981, when Siemens Semiconductor brought out its 64K DRAM, it was four years behind the Americans - then the best makers of memory in the world.

If the rumours are true that Crolles is to be revived as a great microelectronics R&D centre, then the decision by STMicroelectronics to invest heavily in the site, if confirmed will be magnificent news for European microelectronics capability.

Water remains the best liquid for immersion lithography is the conclusion of the recent SPIE Advanced Lithography conference in San Jose.

What comes after silicon and who will make it? Almost certainly the Americans will make it as US companies like IBM and Nantero of Boston appear to be leading the way in getting tangible results from research into graphene and carbon nanotubes as alternatives to silicon.

Why have we been led constantly to expect that consumer fuel cells are an imminent proposition?

If only they could speed up the development of all this carbon nanotube, molecular switching, graphene transistor technology.

“The industry needs a new transistor, a new device”, reckons Dr Dwight Decker, Chairman of the Global Semiconductor Alliance and Chairman of Conexant, “we’ve been lucky to put it off for 20 years. We need something to reduce the cost of manufacturing dramatically, maybe a completely new material.”

April was an excellent month for new memory technologies. First we get IBM saying its Racetrack technology can increase memory density by ten times; then we get Glasgow University saying they can make a molecular switch which could implement a Petabyte memory on a square inch substrate; then Manchester University says it’s made a transistor one atom thick and ten atoms wide out of graphene; Duke University, using a masking technique, makes the highest density cluster of carbon nanotubes ever achieved - ten nanotubes a few atoms thick, and then H-P comes up with its ‘memristor’, a resistor with storage capabilities which could be used in FPGAs and for ultra-low power memory.

The reason why Samsung, Intel and TSMC announced last week that they were combining to push 450mm wafer manufacturing technology is because no one else wants it, according to Dr Mark Pinto, executive vice president and CTO of Applied Materials, the world’s largest manufacturer of semiconductor production equipment.

There seems to be a schism opening up between the semiconductor industry and the semiconductor production equipment industry.

Carbon nanotubes will be used commercially in the form of a fabric, called buckypaper, within a year, according to researchers at Florida State University quoted in the San Jose Mercury News.

Interesting at Electronica to hear the semiconductor companies all banging on about applications. Not so many years back they were all banging on about breakthroughs. The next generation of chip, a new process technology or a new material..

Instead of narrowing the market applicability of its products by focussing on specific customer needs, the programmable logic industry would do better to address its fundamental problems : FPGAs use too much power and are too expensive.

One of the pet hates of GPs is patients who come in and say: "I read  that bit called 'A Dr writes . . .' in the Daily Mail/Sun/Mirror/Express and  it says I've got a terminal case of bubonic plague."

Last week in Euroland, it was interesting to see that, although the big companies seem to be calling the shots on Euro-R&D, the kind of R&D the electronics industry will increasingly be doing is well suited to small companies - i.e. software.

What's another simile for sluggishness apart from the movement of glaciers and the mating of tortoises? How about the gestation of technology? Unity Semiconductor has been seven years in the development of its new non-volatile memory technology which it calls CMOx.

Intel has unveiled some futuristic product concepts at the Computer History Museum in Mountain View, California