When Gary Patton pulls a thin, round wafer-like sheet from his coat pocket, he makes no attempts to conceal his excitement. “It’s cutting-edge innovation!” says the man who heads IBM’s research and development operations worldwide, as he holds out what seems like a thin, metallic sheet. Then, with childlike excitement, he folds the circular sheet into a neat triangle, then unfolds it and rolls it into a thin cigarette.
The metallic sheet that Dr. Patton is gung-ho about represents a major breakthrough in material sciences, one that will drive high-performance computing to a new level, ushering in a new paradigm in silicon and electronic circuits. A closer look at the round patch reveals that beneath the shiny surface is an intricate network of circuits that looks much like your regular printed circuit board.
Made of silicon, this wafer, he says, holds the future. Also called flexible circuits, these, for now, are a huge step for portable computing, but the possibilities are endless.
They can power your mobile devices, you can have embedded electronics in your body, say to monitor health parameters, or even actively make biomedical interventions, Dr. Patton told The Hindu on the sidelines of the Indian Electronics and Semiconductor Associations (IESA) summit here last month. This, he says, will allow for products that will usher in the era of ‘anywhere computing’, the stuff that we’ve only read abut in futuristic literature or sci-fi.
What’s it all about?
So what’s this flexible wafer all about? Silicon wafers are typically hard and thick structures, processed in large fabs to make integrated circuits and other micro-devices. In IBM’s research centre, Dr. Patton explains, scientists have selectively cleaved the silicon substrate into a “very, very thin layer, thinner than paper”. “When we did this, we found that the characteristics of the material did not change significantly. It’s really a major breakthrough; it’s the stuff of sci-fi,” he gushes. The research, he said, suggests that such wafers can be made with conventional processes at room temperature, and the process is entirely scalable. But how long will it be before something like this takes the form of a product? “Not too long.”
This was among the key recent breakthroughs demonstrated by IBM at the Common Platform Technology Forum earlier this year. Other exciting systems on insulator (SOI) research showcased included breakthroughs in the field of carbon nanotubes and silicon nanophotonics.
Dr. Patton said that the semiconductor industry is at “a significant transition point”, where process engineering has reached the limit of being able to engineer material by changing fundamental properties of silicon by introducing strain or using High K metal gates to change the properties of gate oxides; all technical processes that were used up until now to fuel growth in computing power. Currently, Dr. Patton said, the semiconductor industry is in what he calls the 3D era, where devices known as Fin-Fets are being integrated in 3D.
“This is going to result in a real explosion in the next decade in the mobile interconnect of devices. That will take us into 2020, but at some point 3D will run its course and we’ll hit the atomic dimension limit. By then, we’ll be going to new types of materials, and this is where we’ll come into the nanotechnology era… things like carbon nanotubes, silicon nanowires or even the use of photonics where light will replace conventional electrical signals. These will have tremendous benefits in improving the speed of the circuit.”
But what was driving this explosion in demand for computing power, asked Dr. Patton, speaking at the IESA summit. He said it was “about simple economics, at the end of the day.” He explained: “You can make smaller features, you get better performance, you get better cost per function and you can do more applications. What you can do on the phone has grown dramatically over the last few years and, guess what, we have an explosion in the market.
“Consider the difference in transistor price. The relative price of a transistor has gone down by five orders of magnitude over the last 30 years, while the number of transistors consumed per year has gone up by six orders of magnitude over the same period. At the end of the day, it’s all about economics.”
For instance, he compares IBM’s Systems 360, its mainframe computer system announced in the late 60s to its Watson, its high-performance artificial intelligence computer system of Jeopardy fame. At the heart of the IBM 360, a major breakthrough back then, was a module that had six transistors and four resistors, just 10 devices, said Dr. Patton. Today, Watson is powered by 360 power chips, and each one of them has 1.2 billion transistors, so there is a tremendous amount of integration.
But this advancement, he pointed out, is not just restricted to high-end computing. It has resulted in a huge explosion in consumer devices. From smartphones and games to digital TV, and, of course, behind all of it, the network. “Today, there’s literally a trillion devices that are connected in some way to the Internet, and in the last three years, the amount of data has doubled.”
Dr. Patton says that collaboration is key. Take, for instance, Samsung’s line of phones and tablets that compete with iPhones and iPads. These products use a low-power version of IBM’s 32 nm technology, used in high-end servers and networks. “This optimisation was done by working very closely with Samsung’s product team and ARM to figure out how to optimise the full stack to deliver the best for the mobile space. And that’s paid off, as you can see.
“At the Common Platform Technology forum (a collaborative platform of IBM, Samsung Electronics and Global Foundries) over 75 per cent of products with High K metal gate, are our products. But we didn’t develop all these by ourselves. We partner with product companies and co-develop the technology to deliver the best product. That’s a key theme: cooperation between product design teams and technology teams to deliver the best solution.”
The entire industry is seeing large collaborations today. This wasn’t so earlier, but the regime has now changed, Dr. Patton explains. And given the material challenges ahead, companies have no option but to continue to innovate, but these breakthroughs, he emphasised, are not going to happen in silos.