It seems I can’t go a month without seeing at least one article decrying the end of Moore’s Law and another which shows that it’s still on track. Ultimately this dichotomy comes from the fact that we’re on the bleeding edge of material sciences with new research being published often. At the same time however I’m always sceptical of those saying that Moore’s Law is coming to an end as we’ve heard it several times before and, every single time, those limitations have been overcome. Indeed it seems that one technology even I had written off, Extreme Ultraviolet Lithography, may soon be viable.
Our current process for creating computing chips relies on the photolithography process, essentially a light that etches the transistor pattern onto the silicon. In order to create smaller and smaller transistors we’ve had to use increasingly shorter wavelengths of light. Right now we use deep ultraviolet light at the 193nm wavelength which has been sufficient for etching features all the way down to 10nm level. As I wrote last year with current technology this is about the limit as even workarounds like double-patterning only get us so far, due to their expensive nature. EUV on the other hand works with light at 13.5nm, allowing for much finer details to be etched although there’s been some significant drawbacks which have prevented its use in at-scale manufacturing.
For starters producing the required wattage of light at that wavelength is incredibly difficult. The required power to etch features onto silicon with EUV is around 250W, a low power figure to be sure, however due to nearly everything (including air) absorbing EUV the initial power level is far beyond that. Indeed even in the most advanced machines only around 2% of the total power generated actually ends up on the chip. This is what has led ASML to develop the exotic machine you see above in which both the silicon substrate and the EUV light source work in total vacuum. This set up is capable of delivering 200W which is getting really close to the required threshold, but still requires some additional engineering before it can be utilized for manufacturing.
However progress like this significantly changes the view many had on EUV and its potential for extending silicon’s life. Even last year when I was doing my research into it there weren’t many who were confident EUV would be able to deliver, given its limitations. However with ASML projecting that they’ll be able to deliver manufacturing capability in 2018 it’s suddenly looking a lot more feasible. Of course this doesn’t negate the other pressing issues like the interconnect widths bumping up against physical limitations but that’s not a specific problem to EUV.
The race is on to determine what the next generation of computing chips will look like and there are many viable contenders. In all honesty it surprised me to learn that EUV was becoming such a viable candidate as, given its numerous issues, I felt that no one would bother investing in the idea. It seems I was dead wrong as ASML has shown that it’s not only viable but could be used in anger in a very short time. The next few node steps are going to be very interesting as they’ll set the tempo for technological progress for decades to come.
Traditional computing is bound in binary data, the world of zeroes and ones. This constraint was originally born out of a engineering limitation, designed to ensure that these different states could be easily represented by differing voltage levels. This hasn’t proved to be much of a limiting factor in the progress that computing has made however but there are different styles of computing which make use of more than just those zeroes and ones. The most notable one is quantum computing which is able to represent an exponential amount of states depending on the number of qubits (analogous to transistors) that the quantum chip has. Whilst there have been some examples of quantum computers hitting the market, even if their quantum-ness is still in question, they are typically based on exotic materials meaning mass production of them is tricky. This could change with the latest research to come out of the University of New South Wales as they’ve made an incredible breakthrough.
Back in 2012 the team at UNSW demonstrated that they could build a single qubit in silicon. This by itself was an amazing discovery as previously created qubits were usually reliant on materials like niobium cooled to superconducting temperatures to achieve their quantum state. However a single qubit isn’t exactly useful on its own and so the researchers were tasked with getting their qubits talking to each other. This is a lot harder than you’d think as qubits don’t communicate in the same way that regular transistors do and so traditional techniques for connecting things in silicon won’t work. So after 3 years worth of research UNSW’s quantum computing team has finally cracked it and allowed two qubits made in silicon to communicate.
This has allowed them to build a quantum logic gate, the fundamental building block for a larger scale quantum computer. One thing that will be interesting to see is how their system scales out with additional qubits. It’s one thing to get two qubits talking together, indeed there’s been several (non-silicon) examples of that in the past, however as you scale up the number of qubits things start to get a lot more difficult. This is because larger numbers of qubits are more prone to quantum decoherence and typically require additional circuitry to overcome it. Whilst they might be able to mass produce a chip with a large number of qubits it might not be of any use if the qubits can’t stay in coherence.
It will be interesting to see what applications their particular kind of quantum chip will have once they build a larger scale version of it. Currently the commercially available quantum computers from D-Wave are limited to a specific problem space called quantum annealing and, as of yet, have failed to conclusively prove that they’re achieving a quantum speedup. The problem is larger than just D-Wave however as there is still some debate about how we classify quantum speedup and how to properly compare it to more traditional methods. Still this is an issue that UNSW’s potential future chip will have to face should it come to market.
We’re still a long way off from seeing a generalized quantum computer hitting the market any time soon but achievements like those coming out of UNSW are crucial in making them a reality. We have a lot of investment in developing computers on silicon and if those investments can be directly translated to quantum computing then it’s highly likely that we’ll see a lot of success. I’m sure the researchers are going to have several big chip companies knocking down their doors to get a license for this tech as it really does have a lot of potential.
For as long as we’ve been using semiconductors there’s been one material that’s held the crown: silicon. Being one of the most abundant elements on Earth its semiconductor properties made it perfectly suited to mass manufacture and nearly all of the world’s electronics contain a silicon brain within them. Silicon isn’t the only material capable of performing this function, indeed there’s a whole smorgasbord of other semiconductors that are used for specific applications, however the amount of research poured into silicon means few of them are as mature as it is. However with our manufacturing processes shrinking we’re fast approaching the limit of what silicon, in its current form, is capable of and that may pave the way for a new contender for the semiconductor crown.
The road to the current 14nm manufacturing process has been a bumpy one, as the heavily delayed release of Intel’s Broadwell can attest to. Mostly this was due to the low yields that Intel was getting with the process, which is typical for die shrinks, however solving the issue proved to be more difficult than they had originally thought. This is likely due to the challenges Intel faced with making their FinFET technology work at the smaller scale as they had only just introduced it in the previous 22nm generation of CPUs. This process will likely still work down at the 10nm level (as Samsung has just proven today) but beyond that there’s going to need to be a fundamental shift in order for the die shrinks to continue.
For this Intel has alluded to new materials which, keen observers have pointed out, won’t be silicon.
The type of material that’s a likely candidate to replace silicon is something called Indium Gallium Arsenide (InGaAs). They’ve long been used in photodetectors and high frequency applications like microwave and millimeter wave applications. Transistors made from this substrate are called High-Electron Mobility Transistors which, in simpler terms, means that they can be made smaller, switch faster and more packed into a certain size. Whilst the foundries might not yet be able to create these kinds of transistors at scale the fact that they’ve been manufactured at some scale for decades now makes them a viable alternative rather than some of the other, more exotic materials.
There is potential for silicon to hang around for another die shrink or two if Extreme Ultraviolet (EUV) lithography takes off however that method has been plagued with developmental issues for some time now. The change between UV lithography and EUV isn’t a trivial one as EUV can’t be made into a laser and needs mirrors to be directed since most materials will simply absorb the EUV light. Couple that with the rather large difficulty in generating EUV light in the first place (it’s rather inefficient) and it makes looking at new substrates much more appealing. Still if TSMC, Intel or Samsung can figure it out then there’d be a bit more headroom for silicon, although maybe not enough to offset the investment cost.
Whatever direction the semiconductor industry takes one thing is very clear: they all have plans that extend far beyond the current short term to ensure that we can keep up the rapid pace of technological development that we’ve enjoyed for the past half century. I can’t tell you how many times I’ve heard others scream that the next die shrink would be our last, only to see some incredibly innovative solutions to come out soon after. The transition to InGaAs or EUV shows that we’re prepared for at least the next decade and I’m sure before we hit the limit of that tech we’ll be seeing the next novel innovation that will continue to power us forward.
It may come as a surprise to you to find out that Australia is a predominately service base industry. Whilst it’s hard to argue that we’ve enjoyed the benefits of the current mining boom Australia’s GDP is still predominately derived from our service industry, to the tune of 69% (pg. 134). Still the current prosperity and insulation from global economic crises that Australia has received from the growing mining sector won’t last forever and now is the time for us to start looking towards the future so we can ensure future economic prosperity. I strongly believe that we’ve already undertaken the first steps towards achieving this with the implementation of the National Broadband Network.
Australia as it stands today suffers from an incredible amount of skill drain to other countries. Well over half of the Australian residents who leave Australia for over a year or permanently were skilled workers and whilst the trend has gone down in recent times (thanks wholly to Australia’s isolation from the global economic turmoil) that hasn’t stemmed the flow of talent leaving our shores. For the high technology sectors at least there is the potential to recreate the hot bed of innovation that led to the creation of Silicon Valley on the back of the NBN. This would not only stem the brain drain overseas but would produce large and sustainable gains to the Australian economy.
Right now the public view of the NBN varies wildly. Businesses by and large have no idea what benefits it can bring them, public opinion is mixed (although Senator Conroy says differently) and even the federal government seems at a loss to what it could mean for Australia’s future, doling out cash to local governments in the hope they’ll be able to sell it for them. To combat this the government should instead provide incentives and seed capital to high-tech start ups who are looking to leverage Australia’s upcoming ubiquitous high speed Internet infrastructure, in essence building an Australian Silicon Valley.
Doing this requires co-ordination with entrepreneurial communities, venture capitalists and the willing hand of the government. They could easily make investment in these kinds of companies more desirable by extending tax breaks that are currently enjoyed by other asset classes to investment in NBN based high-tech start ups. This would also make Australian based startups incredibly attractive for overseas investors, pumping even more money into the Australian economy. As the sector grows there would also be an increasing amount of ancillary jobs available, ones that accompany any form of corporation.
Australia would then become a very desirable location for both established and aspiring businesses looking to expand into the Asia-Pacific region. It also works in the reverse, giving Asia-Pacific businesses (and nations) a more local launch pad into the western business world. Establishing Australia as a high tech hub between our strong local ties and western allies abroad would provide a massive economic boost to Australia, one to rival that of the current mining boom.
Of course it’s not like this hasn’t been tried before in Australia, indeed many have tried to recreate the success of the valley with little results. Indeed I believe this is due to a lack of co-operation between the key players, namely the government, entrepreneurs and investors. The NBN represents a great opportunity for the government to leverage the industry not only to ensure Australia’s future economic prosperity but also to establish Australia as a leader in technology. I believe that the government should be the ones to take the first steps towards fostering such an environment in Australia as once the industry knows they have the support they’ll be far more willing to invest their time in creating it.
Not leveraging the NBN in such a way would leave the NBN as a simple infrastructure service, woefully underutilised given the capabilities that it could unlock. Make no mistake the NBN puts Australia almost at the top in terms of ubiquitous, high speed Internet access and that makes a lot of services that are currently infeasible to develop attractive targets for investigation. Indeed since the same level of broadband access is almost guaranteed throughout the country it is highly likely that benefits will stretch far past the borders of the CBD, even as far as regional centres.
As someone who’s group up on and made his career in technology it’s my fervent hope that the Australian government recognizes the potential the NBN has and uses that for the betterment of Australia. As a nation we’re well positioned to leverage our investment in infrastructure to provide economic benefits that will far exceed its initial cost. Creating a Silicon Valley of the Asia-Pacific region would elevate Australia’s tech industry to rival those throughout the rest of the world and would have massive benefits far beyond Australia’s borders.