Boffins looking for ways around Moore’s Law have decided that transistors are a bad idea and they need to go back to the vacuum tubes – read valves – of the early computers.
Of course, the vacuum tubes are a little smaller than those used in the codebreaking WW2 computers – those under development at Caltech’s Nanofabrication Group are a million times smaller.
Alan Huang, a former electrical engineer for Bell Laboratories, told the New York Times. “Some of the same algorithms that were developed for the last generation can sometimes be used for the next generation.”
Dr. Axel Scherer, head of the Nanofabrication Group said silicon transistors employed in most computing gadgets only take us so far. Some of the best minds in the world are working on smaller-than-ever transistors.
However, at this size silicon becomes more elastic, and starts to give out light. Silicon transistors also leak electrons, which can be embarrassing in a public place.
Vacuum tubes could prove a better solution at these sizes. The tubes can be made out of a variety of metals, and allow for innovative solutions that could consume less power than silicon chips.
The tech has attracted big investments. Boeing has been putting its money into researching vacuum tube chip research, possibly appearing in the aviation industry before 2020, but it may be a very long time before we see the fruits of Caltech research.
Engineers at the University of California Santa Barbara have shown off a new transistor that switches at 0.1 volts and reduces power dissipation by over 90 percent, compared to the latest silicon transistors (MOSFETs).
Continuing miniaturisation of MOSFETs poses a power dissipation challenge because of their fundamental turn on characteristics.
The engineers at UC Santa Barbara have used the quantum mechanical phenomenon of “band to band” tunneling to make a tunnel field effect transistor.
Professor Kaustav Banerjee at UC Santa Barbara, said: “We restructured the transistor’s source to channel junction to filter out high energy electrons that can diffuse over the source/channel barrier even in the off state, thereby making the off state current negligibly small.”
The engineers said that the global electronics industry loses billions of dollars a year because of the power dissipation effect.
He said: “This transates into lower battery lifetime in personal devices like cell phones and laptops and massive power consumption of servers in large data centres.”
The transistor the engineers have designed used a layered two dimensional material called molybdenum disulphide, with the current carrying channel played over a highly doped germanium as the source electrode.
“We have engineered what is, at present, the thinnest channel subthermionic transistor ever made,” said Banerjee.
Scientists at the National University of Singapore (NUS) said they have developed a hybrid sensor more sensitive than sensors currently available.
The scientists said that the invention will lead to the development of smaller and cheaper censors for electronics, comms, automotive and biotechnology applications.
The hybrid sensor developed by the scientists may meet the needs of high sensitivity to low and high magnetic fields, tunability and small resistance variations due to temperature.
The team claims that the sensor, which is made of graphene and boron nitride is 200 times more sensitive than commercially available sensors today.
Existing sensors are commonly made of silicon and indium antimonide, but these new sensors have an eight fold gain over previously reported laboratory results.
The magnetoresistance sensor industry is expected to be worth nearly $3 billion by the year 2020. One advantage over existing sensors is that there won’t be a need for silicon wafers or temperature corretion circuity, with production cost for graphene based sensors much lower.
The research wing of IBM claimed it has made an engineering breakthrough which could mean carbon nanotubes will displace silicon transistors for future computers.
The scients said they have managed to “shrink” transistor contacts without reducing carbon nanotube performance and that means more powerful microprocessors way beyond existing semiconductors.
IBM said that as devices get smaller, increased electrical resistance has prevented serious performance gains.
But a discovery by the scientists mean carbon nanotube chips will vastly improve the performance of high end computers, increase power and battery life of mobile devices and let cloud datacentres deliver services more economically.
Moore’s Law, said IBM, “is running out of steam”.
IBM thinks that its approach overcomes a major hurdle to the adoption of carbon nanotubes.
Silicon wafers, used by manufacturers to create microprocessors and other semiconductor devices, are facing a slowdown.
That’s emerged as GlobalWafers Co, the sixth biggest wafer supplier in the world, said that the whole semiconductor industry is being affected by a combination of forces including the devaluation of the Chinese yuan and slower general overall growth.
The Taipei Times reported that CEO Doris Hsu predicted “headwinds” for the entire industry because of financial uncertainties.
GlobalWafers supplies silicon wafers to a number of major chip makers including TSMC, UMC, Infineon and Texas Instruments.
She predicted that the slowdown will be short term rather than long term because there’s much less inventory than when the semi industry saw its last crisis, during 2008.
She believes that the automotive industry will generate more semiconductor sales. The firm is set to increase eight inch wafer production by 10 percent or so.
A team of researchers at the Australian National University (ANU) has hit on a way that it could mean ultrathin and ultralight solar cells and LEDs will be available.
The discovery is based on the element phosphorous which the researchers showed could produce single atom layers called phosphorene using simply sticky tape.
Phosphorene acts like a semiconductor much as silicon does and is light too.
Lead scientist Dr Yuerui Lu from ANU said: “It shows very promising light emission properties. It creates possibilities for making lots of interesting devices such as LEDs or solar cells.”
The phosphorene is manufactured by using sticky tape to peel layers of crystals from the black crystalline form of phosphorous, Lu said.
By varying different layers, engineers can control the band gap, so changing the colour of the LED made.
Lu said phospherene at very thin layers is better than silicon – which doesn’t like repeated applications of sticky tape.
Spending on capital investment in the semiconductor industry is set to rise by 2.5 percent this year, well down on the numbers predicted.
Gartner said it had projected a growth of 4.1 percent for the year, but weakness in both the euro and the yen have skewed the picture. Total capital spending will amount to $66.1 billion for 2015.
Bob Johnson, a research vice president at the semiconductor firm, said: “With over half of all equipment being produced by either Japanese or European suppliers, the weakness in their currencies has been the primary factor in our reducing our overall outlook for 2015.”
He said that there will be pause in the equipment market growth next year, because DRAM will go through its normal cyclical downturn.
Foundries will outspend the logic integrated device manufacturers (IDMs) this year, with spending set to increase by 17.2 percent. But saturation of the smartphone market will subdue the necessity to create fresh capacity.
Capital spending in memory will only be 3.2 percent for the year – Gartner had predicted that that it would rise by 10.2 percent. But major manufacturers have revised their spending plans.
Silicon wafer spend will only increase by 0.1 percent during the year – that’s due to a slowdown on fab construction.
Researchers at McGill University and Université de Montréal have come with an alternative to graphene for the next Moore’s Law bending chip material.
Graphene is a one-atom thick material with high electrical conductivity, but as a zero bandgap semiconductor, it acts like a metal. This means that transistors made of the material cannot be easily turned on and off.
However black phosphorus can also be separated into one-atom thick layers known as phosphorene and it does not have the switching problems.
This means that the energy needed to power transistors would be incredibly low.
Thomas Szkopek wrote in Nature that transistors work more efficiently when they are thin, with electrons moving in only two dimensions. Nothing gets thinner than a single layer of atoms.
Of course black phosphorous is not ideal. It is damaged by light, which spells trouble for the creation of single layer transistors.
But the experiments at the highest-powered magnet laboratory in the world, the National High Magnetic Field Laboratory in Florida, the research team discovered that even in thicker sheets of black phosphorous, the electrons still move only in two dimensions.
This is surprising as the electrons are able to be pulled into a sheet of charge which is two-dimensional, even though they occupy a volume that is several atomic layers in thickness. Finding that the 2D electronic structure of black phosphorous was not dependent on the thickness means that it could be manufactured on a large scale.
Of course all that is a few years off, while graphene is closer to production.
Scientists at Stanford School of Engineering believe they have devised a method to speed up data transport in semiconductors by as much as 30 percent in the future.
While a typical semiconductor has millions of transistors connected by tiny copper wires, it’s the sheath that has given the Stanford scientists inspiration.
They said that tantalum nitride has been used to sheath the copper wires within chips.
But experiments show that if graphene is used as a sheath, electrons can fly through the copper wires faster.
The sheath over the tiny copper wires prevents them from interacting with the silicon and also conducts electricity.
But the Stanford team shows that a graphene layer would be eight times thinner than one made from tantalum nitride. But using graphene as a layer can also act as an auxiliary conductor of electrons as well as isolating the copper from the silicon.
But while the method holds promise, there are still hurdles to adopting graphene to this purpose. That would include methods to grow graphene directly onto the wires during mass production.
A group of researchers has started to implement security in silicon that can help thwart nosey parkers or criminals from understanding what data is in your cloud.
The MIT researchers said two years ago they proposed a method for preventing outsiders by checking the way computers access memory banks.
The researchers said that they’ve already tested their methods on reconfigurable semiconductors and are moving into manufacturing these devices.
The chip improves security by checking that when data is fetched from a memory address, it will query other address too.
Although this puts stress on a system because extra data is involved, the MIT team said they store the memory addresses in a tree-like data structure, with every address randomly assigned to a path through the tree.
The chip they’ve designed avoids a performance overhead by having an additional memory circuit, with storage slots mapped onto the nodes in any path through the tree.
It discards all redundant or decoy data.
The circuits the MIT scientists have designed can be easily added to existing semiconductor designs and switched off or on as needed. So software engineers may activate it only when it’s needed, while other applications could use it all the time.