Cover Story: Work in progress?

A quote attributed to Gordon Moore, claims that if the car industry had matched the advances of the semiconductor industry, a Rolls Royce would do 1million mpg, travel at thousands of miles per hour and would be cheaper to throw away than park. While no one seriously expects other industries to mimic semiconductors, there is no question that some technologies have simply not fulfilled their promise – or at least not yet. At the same time, there are many examples of electronic products and systems that simply failed, either technologically, or in terms of market success, and sometimes both. Together, these two aspects show that advances in technology are far from inevitable. Take the first aspect – technologies that have not yet achieved what many – perhaps naively – had hoped for from them. To cite just three: solar power; batteries; and wireless power distribution. Clearly, it would be wrong to describe solar power as a failed technology. But, equally, there would seem to be huge potential for improvement, if only because of the typical efficiency of solar cells in converting solar energy to electricity. One estimate claims that converting just 0.00016% of the energy coming from the sun into electricity would provide mankind's total energy requirements. So if we could get solar cells' efficiency up to, say, 90%, we could solve the world's energy problems. Instead, the most efficient cells typically achieve a conversion rate of around 22%, according to Dan Davies, director of technology for Solarcentury, a leading solar power supplier. Many conventional solar cells do not even achieve this, measuring around 16 to 17%. If the semiconductor industry had faced such a limitation, it would by now have just about managed to build the first simple calculator. But there is hope. Just a few weeks ago, US based Spire announced a record solar cell efficiency of 42.3%. This is clearly an important advance, but are we likely to see dramatic improvements? Davies thinks it unlikely because the efficiency limitations stem from fundamental physical features. "The key issue is the spectral response of the cell. Sunlight contains a wide range of frequencies and the band gap in a photovoltaic cell can only efficiently use part of that spectrum." At the low end, the energy of the photons is not enough to create a conducting pair (hole and electron); at the higher end, it is in effect too much and energy is lost as heat. Multijunction cells try to overcome this, but standard PV cells are single junction, so therefore limited. But no one knows that a radical breakthrough is impossible – not long ago the idea of high temperature superconductivity would have been laughed at, but it happened. A similar point can be made about cost. While it is true that costs have decreased over the last decade, encouraging increasing numbers of people to install solar panels on their homes, the payback time for the investment is still many years. "The desire for a radical breakthrough is quite common, with people looking for a step change in solar technology, but I don't see the success of solar depending on that," Davies says. Today, the market is dominated by conventional crystalline silicon photovoltaic technology, which accounts for more than 90% of systems in use. The efficiency of this technology is increasing, but slowly – by a few tenths of a per cent per annum, Davies says, and costs are coming down too, thanks to increases in volumes. The first systems Davies designed in 1999 cost around £8000/kW. Today, that has fallen to as low as £3000/kW. The layman's view of batteries is similar. Yes, they have improved significantly. Many of today's consumer items would be unfeasibly large and heavy without lithium batteries, for example. A couple of decades ago, mobile phones were like bricks, principally because of the battery they needed, and even then talk times were nothing like those of today. But the feeling remains: couldn't we have achieved more? It is now 210 years since Alessandro Volta invented the first modern electric battery. Today's electronic and IT industries have been going effectively for, say, 60. So batteries have had 150years lead! If batteries had advanced like the semiconductor industry, we could be launching space vehicles with them or, at the very least, riding around in high performance electric cars. Who knows, we may eventually do that. But it may be that battery technology as a whole will remain constrained. As with solar power, it looks as if the capacity of batteries is limited by fundamental constraints imposed by the chemical and electrical properties of atoms. Other question marks surround the battery's future. Many depend on relatively rare elements, such as lithium itself, and there are serious doubts that enough is available to support hundreds of millions of electric vehicles. Also, batteries in future may face more competition from devices like fuel cells, for example, and even wireless power. Wireless power is a technology dream that has so far failed to materialise. That dream goes back to the 19th Century, when Nikola Tesla, inventor of the alternating current, envisioned huge power stations giving the Earth's crust such an electric charge that power could be obtained by plugging into the ground. Tesla spent millions on this vision, building an electrical tower so powerful that the fields around it glowed when it was turned on, but it never progressed and wires have dominated ever since. Yet wireless power may be about to arrive, at least for low power levels. Work at MIT, Intel and elsewhere has demonstrated its feasibility. Intel has exhibited a wirelessly powered iPod speaker, which was attached to a copper coil and powered by magnetic fields from a second coil nearly a metre away. And the Wireless Power Consortium has established an interoperability standard called Qi. Higher power levels are also being achieved. US based LaserMotive specialises in laser power beaming; the wireless transfer of energy over distances using laser light. It has already won a $900k NASA sponsored competition, transmitting more than 1kW over a distance of several hundred metres, and recently demonstrated a laser powered remote controlled mini helicopter. It is now planning to develop a full scale laser power system for unmanned aerial vehicles. Compared with semiconductors, the technologies discussed could be said to have failed to an extent – or have been limited – but they have an excellent excuse: it has proved difficult to make dramatic advances for fundamental physical reasons. However, the electronics and computing industries have plenty of examples of failure for other reasons. One recurring theme is that technological failure is not necessarily absolute: often, a brilliant idea was simply before its time and the technology of the period wasn't up to what a visionary inventor wanted to achieve. Perhaps the best single example is the Father of Computing, Charles Babbage. Equipped with electronics – even valves, let alone transistors – he would probably have built his extraordinary analytical engine. Limited to purely mechanical workings, it was impossible. There are some current examples of the same thing: the Tablet PC was launched with huge fanfare by Microsoft in 2001, but it has hardly become a device used by millions. Then Apple released the iPad earlier in 2010 and sold 3m in 80 days. Similarly, predictions that we will all be reading books electronically go back decades, but it is only now that the visual quality of the devices on offer is making those forecasts look plausible. Another device that arguably tried to do more than was technically feasible at the time was the Apple Newton, one of the first PDAs, introduced in 1993. It featured ambitious functions like handwriting recognition, which was too difficult to achieve, at least to the point of being genuinely useful. It was also very expensive. Even so, it paved the way for PDAs – the Palm Pilot came only a couple of years later and was a major success. Consumer electronics has seen its fair share of technological failures, for various reasons – a bad idea, competing incompatible standards, even just sheer irrelevance. Take post cd audio standards like Super Audio CD and DVD-Audio, said to be significantly superior to the cd. While hifi fanatics may enthuse, most listeners would struggle to tell the difference, and probably not care much even if they could. Today's music revolution is not about sound quality, but convenience and storage capacity, thanks to MP3 players. The audio world has seen lots of other failures: quadraphonic sound never really caught on and was ousted by surround sound standards for cinema in the home systems; Sony's MiniDisc was introduced in 1992 to try to capture the portable music market, but people stuck to portable cd players and cassettes, until MP3 players arrived; digital audio tape (DAT) players failed partly because of the recording industry itself, which imposed Draconian copy protection requirements to stop DAT players becoming the perfect tool for music piracy; and, going back even further, the eight track audio cartridge. Admittedly, a lot of these were sold but it gave way ultimately to the superior cassette. Video recording contains perhaps the most frequently cited technology failure. Sony's Betamax video recorders lost out to JVC's VHS system, even though Betamax was acknowledged to be superior. A virtual replay has just been staged, with Sony's Blu-ray HD disk system defeating Toshiba's HD-DVD. Another spectacular technological failure was in the satellite communications business – Motorola's Iridium project. It aimed to build and launch 66 satellites to provide a worldwide wireless phone service. But the service was hugely costly to implement and expensive to use. Meanwhile, the cellular phone business had started to take hold. The result: Iridium had to default on $1.5billion and the system that cost Motorola more than $5bn to build ultimately sold for $25m. But Iridium has now turned itself around. Other failures are rather more generic – the paperless office, for example. Email has supposedly caused a 40% increase in paper use and, even though sales of plain white paper are levelling off, it's a pretty safe bet that offices without paper are not going to happen. And what of today's potential technologies that won't make it? Fully immersive virtual reality is still little more than a dream, will it stay that way? Will web tv become the norm? And will we finally start to talk to our computers? Speech recognition is now far better than it used to be but there are few signs that it is taking the world by storm. Maybe we prefer our machines to stay quiet.