Scientists devise breakthrough technique for mapping temperature in tiny devices

Overheating is a major problem for the microprocessors that run our smartphones and computers. But a team of UCLA and USC scientists have made a breakthrough that should enable engineers to design microprocessors that minimize that problem: They have developed a thermal imaging technique that can "see" how the temperature changes from point to point inside the smallest electronic circuits. The technique, called plasmon energy expansion thermometry, or PEET, allows temperatures to be mapped in units as small as a nanometer, a unit of measure equal to one-billionth of a meter. This shatters the previous record for thermal imaging resolution, and it could eventually lead to faster and more capable electronics. Reported in the February 6 issue of ("Nanoscale temperature mapping in operating microelectronic devices"), the study reveals -- at the atomic level -- how heat moves along a tiny aluminum wire that is warmed at one end. plasmon energy expansion thermometry Artist’s conception of plasmon energy expansion thermometry showing a focused electron beam penetrating a 100 nanometer wide aluminum wire atop a thin glass window. The wire’s temperature is mapped by scanning the electron beam. Modern microelectronic circuits contain billions of nanometer-scale transistors. Although each transistor generates only a tiny bit of heat as it operates, with that many transistors operating at once, computer chips get very hot, which is why cellphones get warm and computers need fans to run properly. To better understand precisely where the heat is being generated, engineers want to be able to map temperature in tiny electronic circuits. Currently, they use one of two thermal imaging techniques: capturing the infrared radiation the device emits or dragging a tiny thermometer back and forth across the device's surface. But both standard techniques have fundamental limitations. Radiation-based thermometers struggle to resolve devices that are smaller than the wavelengths of the detected radiation, which typically are several thousand nanometers. And bringing a thermometer into contact with a small device generally disturbs the device's temperature. In addition, neither has demonstrated the resolution necessary to "see" the active features in modern transistors, which are typically 22 nanometers across or smaller. Without a way to measure the temperature of extremely small circuitry, manufacturers have worked blindly, relying on simulations to estimate the devices' temperatures. Now, PEET mapping will enable them to heat a transistor and accurately map which parts of it heat up and track how the heat is transported away -- knowledge that could help engineers revolutionize the design of the nanoscale electronics inside the next generation of computing devices. Led by Chris Regan, a member of UCLA's California NanoSystems Institute and first author Matthew Mecklenburg, a senior staff scientist at USC's Center for Electron Microscopy and Microanalysis, the research team built its technique on the same physical principles behind the glass-bulb thermometer that was invented by Daniel Gabriel Fahrenheit in 1724. Fahrenheit's thermometer gauges temperature from changes in the density of mercury. As mercury is warmed or cooled, it expands or contracts, causing it to move up or down inside a graduated glass cylinder. PEET determines temperature in the same way, by monitoring changes in density. However, the UCLA-USC team's key advance was to measure changes in the density of the microelectronic device itself rather than using a separate thermometer. In effect, the technique turns the device into its own thermometer. PEET maps density using a transmission electron microscope. For the Science research, the team demonstrated the technique on tiny aluminum wires that were heated on one end. They focused the microscope's electron beam to a point, scanned it across the wire and measured the energy of the beam electrons as they came out the other end. Passing through the wire, some of the beam's electrons create charge waves in the wire, called plasmons. Electrons lose energy making those waves, just as a motorboat burns gasoline to make a wake on a pond. Because that energy loss is sensitive to the wire's density, measuring it accurately determines the wire's density, and therefore its temperature; warmer parts of the wire have a slightly lower density. Repeating this measurement thousands of times as they moved the tightly-focused beam over the wire, the team was able to map the wire's temperature with nanometer-scale spatial resolution. "With the old techniques, measuring the thermal conductivity of a nanowire returns one number. Mapping temperature with PEET, we get 10,000 numbers as we go down the wire. It's the difference between seeing the score and watching the game -- one gives you much better knowledge of the players," said Regan, an associate professor of physics and astronomy at UCLA. According to Mecklenburg, the technique could be adopted easily by electronics manufacturers. "What's especially important is that the transmission electron microscope is already the primary tool used by manufacturers for examining individual microelectronic devices," he said. "We have developed a way to measure thermal gradients with that same microscope -- it is a perfect fit. Suddenly, manufacturers can see a new dimension in their devices with the tools they already have."
read more "Scientists devise breakthrough technique for mapping temperature in tiny devices"

Research shows benefits of nanocrystalline silicon carbide for sensors in harsh environments

The use of silicon carbide as a semiconductor for mechanical and electrical sensor devices is showing promise for improved operations and safety in harsh working environments, according to new research from Griffith University. Experiments with silicon carbide grown at the Queensland Micro- and Nanotechnology Centre (QMNC) at Griffith University have demonstrated the compound's superiority as a semiconductor for high performance sensors. The research has identified advantages for fields including mining, aerospace, aviation and the automotive, electrochemical and biomedical industries. The findings appear in the specialist publication ("The effect of strain on the electrical conductance of p-type nanocrystalline silicon carbide thin films") and for the first time present the effect of mechanical strain on the electrical conductivity of silicon carbide deposited on silicon wafer. "Over the past 50 years, silicon has been the dominant material used as a semiconductor for sensing devices and that continues today in computers, mobile phones, automobiles and more," says Dr Dzung Dao, from Griffith's School of Engineering and one of the lead researchers. "However, silicon is not suitable for electronic devices at high temperatures above 200°C due to the generation of thermal carriers and junction leakage. "Silicon carbide, on the other hand, possesses excellent mechanical strength, chemical inertness, thermal durability and electrical stability due to its unique electronic structure. "Thus it holds promise as the material for high performance sensors in, for example, deep-oil and coal mining, combustion engines, energy conversion devices and so on. "In areas where the temperature can reach well above 200°C, chemical corrosion and mechanical shock are extreme. That's where silicon carbide comes in. "Silicon carbide is already used in power electronics and these results are very encouraging for sensor technology, particularly in harsh working environments." The device-grade silicon carbide for this research was grown on six inches of silicon wafer at low temperature by Professor Sima Dimitrijev's team at QMNC.
read more "Research shows benefits of nanocrystalline silicon carbide for sensors in harsh environments"

Breakthrough may lead to industrial production of graphene devices

With properties that promise faster computers, better sensors and much more, graphene has been dubbed the 'miracle material'. But progress in producing it on an industrial scale without compromising its properties has proved elusive. University of Groningen scientists may now have made a breakthrough. Their results will be published in the journal ("Comparing Graphene Growth on Cu(111) versus Oxidized Cu(111)"). Graphene is a special material with crystals that are just one atom thick. Electrons pass through it with hardly any resistance at all, and despite being very flexible, it is stronger than any metal. The discoverers of graphene, Andre Geim and Konstantin Novoselov, famously made it by peeling graphite with Scotch tape until they managed to isolate a single atomic layer: graphene. It won them the 2010 Nobel Prize in Physics. 'The challenge is to find a substrate that not only preserves the properties of graphene, but also enables scalable production.', says Stefano Gottardi, PhD student at the University of Groningen Zernike Institute for Advanced Materials. A good candidate is chemical vapour deposition. Here heat is used to vaporize a carbon precursor like methane, which then reacts with a catalytically active substrate to form graphene on its surface. A transition metal is normally used as the substrate. However, not only does the transition metal act as a support, but it also tends to interact with the graphene and modify - or even deteriorate - its outstanding properties. Cumbersome To restore these properties after growth on the metal, the graphene has to be transferred to a non-interacting substrate, but this transfer process is cumbersome and often introduces defects. Nevertheless, many scientists are trying to improve graphene growth on transition metals, mostly using copper foil as the substrate. This is what the Surfaces and Thin Films group of Gottardi's supervisors Meike Stöhr and Petra Rudolf did too. 'When we analyzed a sample of graphene on copper, we made some strange observations', Stöhr recalls. The observations suggested that alongside the copper some copper oxide was also present. Indeed, a nice graphene film appeared to have formed on the copper oxide, and as oxidized metals might leave the properties of graphene unaltered, this was a potentially important observation. Achievement The Groningen team began to study this possibility in more detail. That was three years ago. Since then, Gottardi and his colleagues have managed to successfully grow graphene on copper oxide. This achievement together with an in-depth characterization of graphene's properties will be published in . The team also reports the remarkable finding that graphene on copper oxide is decoupled from the substrate, which means that it preserves its peculiar electronic properties. The results could be far-reaching. Stöhr: 'Other labs need to reproduce our findings, and quite a bit of work needs to be done to optimize growth conditions.' The best case scenario would be that large single-domain crystals of graphene could be grown on copper oxide. If this proves to be the case, it should then be possible to use lithographic techniques to make all sorts of electronic devices from graphene in a commercially viable manner. An unexpected observation three years ago may thus prove to be the start of a new era of graphene electronics.
read more "Breakthrough may lead to industrial production of graphene devices"