12 June 2012
Nanoscale wires could yield next gen quantum computers
Making transistors ever smaller has been a hallmark of the semiconductor industry. Now, making wires ever thinner – down to the nanoscale – is holding the prospect of many new potential applications.
Such nanowires can now be made with a diameter of just 1nm, though researchers tend to work with nanowires that are between 30 and 60nm wide. At these dimensions, materials can acquire properties very different to those they exhibit at larger scales. That is partly because at such tiny scales, quantum confinement effects alter the behaviour of fundamental particles like electrons within the material. Such effects can change how materials conduct electricity and heat, or interact with light.
Paul Warburton, a Reader in Nanoelectronics at the London Centre for Nanotechnology at University College London (UCL), has worked extensively on nanowire techniques.
"There are a host of niche applications within the electronics sector, such as solar cells and sensors, where nanowires could have a significant impact in the medium term and rapid advances are being made. You could say nanowires have taken on some of the attention that a few years ago was being spent on nanotubes."
A key question about nanowires concerns how best to manufacture them: building them from the bottom up, or from the top down? A top down approach involves taking the material that will form the nanowire and reducing it until you reach nanoscale dimensions. As its name suggests, the bottom up approach is an assembly process where the nanowire is 'grown', by adding particles gradually.
An example of a top down approach is fibre optic nanowires, which can be created by heating a sapphire rod, wrapping the cable around the rod, and stretching it to create a nanowire. Another method uses a tiny furnace made from a small cylinder of sapphire. The fibre optic cable is drawn through the furnace and stretched. A third procedure, called flame brushing, uses a flame under the cable while it is stretched.
For the other approach, bottom up, the essential material science process used to grow nanowires, called the vapour liquid solid technique, has been around since the 1950s. This uses the chemical vapour deposition (CVD) process well known in the semiconductor industry. CVD refers to a group of processes where solids are formed out of a gaseous phase. Catalysts, such as gold nanoparticles, are deposited on a substrate where they act as an attraction site for nanowire formation. The substrate is placed in a chamber with a gas containing the appropriate element, such as silicon, and the atoms attach themselves to the catalyst. Then additional gas atoms attach to those atoms, and so on, creating a chain or wire.
There are still major challenges with either technique to achieve mass production, but Warburton is clear about which is the more promising. "My strong feeling is that it is bottom up techniques that will enable us to achieve mass production and control the functionality of the growing nanowire."
Mass production is only one manufacturing challenge. Another is arranging the nanowires properly once they are built. The small scales make it difficult to build devices automatically – today, nanowires are typically manipulated into place with tools while being observed through a powerful microscope.
But there is no shortage of potential applications for nanowires. One project underway at UCL is to develop nanowire based electronic noses. This aims to exploit the fact that the transport of an electrical signal through the nanowire changes if a molecule lands on the outside of the wire. This works because with a nanowire you have a very high surface area to volume ratio.
Warburton's team has already had a paper published in the Journal of Applied Physics, describing the theory underlying the electronic nose and he is now working on the development of the materials needed to implement the theory.
"We are starting to look at nanowire heterostructures (essentially, the combination of two layers or regions of dissimilar crystalline semiconductors) where, as you grow your nanowire, you can change the material that you are growing along its length. Hopefully, you can do that with atomically sharp interfaces and that gives you some quantum confined structures that you can then exploit in applications."
Another potential nanowire application is solar cells – and there are two ways this could happen. The first is to enhance the efficiency of the conventional solar cell by incorporating nanowires.
There are two aspects to how a conventional solar cell works. One is the conversion of the incoming light into electronic charges; the second is to get those charges to an electrode, where they can be coupled to a circuit to create useful power. These processes are rather inefficient and are currently done by a single material – amorphous silicon.
"One potential way of improving efficiency is to embed nanowires in a polymer material; the polymer does the light absorption and the nanowires carry the charge," Warburton says, "This allows you to tailor these two functions independently to maximise the light to electron conversion efficiency by tweaking the polymer parameters, and do the same for the charge transport by tweaking those of the nanowires."
The second way in which nanowires could improve solar technology is in 'solar concentrate' applications. These involve the use of huge mirrors located in suitable areas like deserts, where sunlight is focused to a tiny spot. Here, a huge amount of energy is collected in a small area and, as a result, the economics are radically different.
For this kind of application, UCL is collaborating with Danish company SunFlake, a spin out from the Nanoscience Center at Copenhagen University. SunFlake says its nanowire based solar cells combine III-V semiconducting nanotechnology with standard silicon solar cells. Material costs are similar to or less than crystalline silicon solar cells, but with potentially 50% greater conversion efficiency.
"Nanostructures provide an ideal light trapping geometry," says SunFlake. "They provide a large surface to volume ratio and offer a high absorption coefficient of the incoming sunlight without the use of an antireflective layer. In addition, the nanostructures reduce the number of processing steps required, and therefore costs, compared with traditional manufacturing."
As well as collecting light, nanowires could generate it, in the form of LEDs. Recent work at MIT has resulted in new techniques for growing nanowires with greater precision – and LED light bulbs are a potential application. The most important colours of light to produce from LEDs are in the blue and ultraviolet range and the ZnO and GaN nanowires produced by the MIT group can, potentially, produce these colours efficiently and at low cost.
An advantage of this approach is that it could enable the use of cheaper substrate materials, replacing today's sapphire or silicon carbide substrates. The nanowire devices also have the potential to be more efficient.
Energy scavenging is a further possible application. Most nanowire researchers have worked with silicon or other compound semiconductors. But other work involves ZnO, which UCL is starting to investigate. A group at Georgia Tech in the US has pioneered this, exploiting the fact that ZnO is piezoelectric – when it is bent or stretched, it generates an electrical current. The idea is that you could wear a nanowire based material that could power devices simply by being shaken, for example. A key challenge here is to increase the conversion efficiency.
Other challenges nanowires face are the economics and control of the material, Warburton says.
"Silicon does what it does phenomenally successfully and, in practice, phenomenally cheaply. To achieve anything like this with nanowires is a major challenge. Growing commercially useful nanowires is going to require complete control over their length, cross sectional area, the mobility of the electrons within them and the level of defects, all of which can affect how fast switching can occur. These are all being worked on and great advances are being made, but there is still a way to go."
One interesting phenomenon nanowires can exhibit is ballistic conduction. In normal conductors, electrons collide with the atoms in the material, slowing them and generating heat as a byproduct. In ballistic conduction, electrons can travel through the conductor without collisions. Nanowires could conduct electricity efficiently without producing intense heat, but there are challenges.
"How do you get the current into the nanowire in the first place?," Warburton asks. "If you use a metallic electrode to feed the electrons into the nanowire, a lot of heat will be generated as that happens. Although you have removed heat from the conductor, there will still be plenty of heat in the total circuit. This is of course a problem faced by conventional silicon technologies."
Ballistic conductance is different to superconductivity; with the former. there is still resistance. This 'quantum of resistance' means that, for example, a nanowire with a single ballistic conduction channel will have a resistance of 6kohm, irrespective of length. A superconducting nanowire will have zero resistance.
In fact, superconducting nanowires are another area that Warburton is addressing, in collaboration with the National Physical Laboratory. Strangely, the aim is to destroy the superconductivity.
"There is an open question in physics as to how small a superconductor can be. There are predictions that a superconducting material measuring 10 to 20nm in cross section becomes very resistive. This may sound like bad news, but there is an application in which the NPL is interested; creating a new current standard and making it possible to define the ampere extremely precisely." (For more, click here.)
Superconductivity has also featured in nanowire research at Delft University of Technology. Here, a team made nanowires out of indium arsenide, each about 100nm in diameter and 100 times as long. By attaching aluminium electrodes to the wires, they could test the wires' behaviour at temperatures close to absolute zero. At these temperatures, aluminium starts to superconduct, offering zero resistance.
The team discovered the superconducting electrodes prompted similar behaviour in the nanowires because the superconducting properties 'leaked' into the wire – something known as the 'proximity effect'. By applying a voltage to the substrate on which the nanowires rested, they could alter the strength of the proximity effect and turn the superconductivity on and off at will.
And that's an interesting prospect. Because superconductivity is an inherently quantum effect, superconducting nanowires could, in principle, form part of the circuitry necessary for a future quantum computer.