Questions raised about giant piezoresistance

To enjoy free access to all high-quality "In depth" content, including topical features, reviews and opinion sign up

Dec 1, 2010

Four years after scientists in the US reported seeing "giant piezoresistance" in silicon nanowires, a team of researchers in France and Switzerland claims that this phenomenon may not exist after all.

Giant piezoresistance is a large change in electrical resistance that occurs when a material is stretched. After it was first reported in tiny silicon wires, claims were made that it could significantly improve nanoelectronic devices, such as nanoscale transistors, and help make ultrasensitive nanosensors.

Now, new work by Jason Milne and Alistair Rowe at the Ecole Polytechnique, Steve Arscott of IEMN-CNRS and Christoph Renner at the University of Geneva calls such applications into question.

Physicists have known about piezoresistance (PZR) – whereby the electrical resistance of a semiconductor changes when a small mechanical stress is applied to it – for many years. Giant piezoresistance occurs when there is a much larger change in resistance for the same applied strain. For example, the change in resistance per unit of strain (the "gauge factor") typically ranges up to 100 in bulk silicon but in giant PZR, this value can reach several thousand.

Giant PZR would find many practical applications. For example, it might be used to detect motion in nanomechanical systems (NEMS) because traditional detectors lose their sensitivity at these length scales. Furthermore, because mechanical stress is currently employed to enhance the performance of electronic devices (in so-called "strain engineering"), it might also help enhance nanoscale transistors too.

The very act of measuring the resistance changes its value Alistair Rowe, Ecole Polytechnique

Four years ago Peidong Yang's team at the University of California at Berkeley first observed giant PZR in silicon nanowires and the discovery created a flurry of interest in labs worldwide. Indeed, the researchers measured gauge factors up to almost 6000. The effect was thought to be a new phenomenon occurring in an otherwise well-characterized material resulting from the sample's reduced size and characteristic surface states.

In a paper just published in Physical Review Letters, the France-Switzerland team claims that these observations were probably artefacts in no way related to the mechanical stress applied to the silicon nanowires. They were, instead, caused by surface trapping of charges induced by the voltage applied to measure the resistance. "In other words, the very act of measuring the resistance changes its value," explained Rowe.

PZR is usually measured by performing a standard resistance measurement on a sample while gradually changing the applied mechanical stress on it. The trouble is that any non-stress-related drift in the value of the resistance cannot be separated from that caused by the applied stress.

The France-Switzerland team says it overcame this problem by applying an oscillating stress to its samples. In this way, stress repeatedly increases and then decreases as function of time. "This is a fairly standard technique (called heterodyne detection) in physics and engineering and is used to separate two or more signals and give artefact-free measurements," said Rowe.

According to Rowe, scientists had never applied heterodyne techniques to PZR measurements before, so previous measurements revealed large (but not stress-related) resistance changes in the silicon nanowires. "This meant that the resistance drift due to charge trapping (also known as dielectric relaxation) was assumed to be the result of the applied stress", he added. "This now appears to have been an incorrect assumption."

Yang himself disagrees: "They are reporting PZR measurements on a collection of top-down micro- and nano-wires while our measurements were on bottom-up grown nanowires. Their results might not actually be that surprising as we now all know that bottom-up synthetic bridging nanowires have quite different strain levels, surface states and dopant profiles from those of top-down fabricated ones. In fact, the lack of giant PZR effect in such nanowires was already reported back in 2003. However, the lack of giant PZR effect in these new fabricated samples should not automatically imply the same in our synthetic bridging nanowires.

The observed PZR effect in our nanowires, whether it is intrinsic or from the surface states effect, has already proven to be useful Peidong Yang University of California at Berkeley

"After all, the observed PZR effect in our nanowires, whether it is intrinsic or from the surface states effect, has already proven to be useful," he added. "For example, we recently demonstrated the first piezoresistively transduced very high frequency silicon nanowire resonators with on-chip electronic actuation at room temperature. We clearly showed that, for very thin silicon nanowires, their time-varying strain can be exploited for self-transducing the devices' resonant motions at frequencies as high as 100 MHz. This simply would not be possible without the enhanced PZR effect."

The debate looks set to continue.

View the original article here

Giant Faraday rotation spotted in graphene

To enjoy free access to all high-quality "In depth" content, including topical features, reviews and opinion sign up

Nov 15, 2010

The polarization of light can be rotated by almost 6° as it passes through a single sheet of graphene in a magnetic field, according to an international team of physicists. This latest property of graphene – a sheet of carbon just one atom thick – was unexpected because large rotations normally occur only in much thicker materials. The scientists believe that this newly discovered property of graphene could be exploited in new devices that switch light using electric and magnetic fields.

The fact that the polarization of light can rotate as it travels through a material exposed to a magnetic field is, of course, nothing new. Physicists have long known that it is to do with that fact that right- and left-circularly polarized light can propagate at different speeds. It means that when linearly polarized light passes through such a material, the right and left components of the light interfere such that the polarization is rotated by a certain angle when it emerges.

But because the size of this "Faraday angle" is proportional to the thickness of the material, graphene – being just one atomic layer thick – was not expected to generate a large rotation. However, Alexey Kuzmenko and colleagues at the University of Geneva have found that the material can twist the polarization of light by 0.1 radians, or about 6°. Researchers at the Fritz Haber Institute in Berlin and the University of Erlangen-Nueremberg – both in Germany – and the Lawrence Berkeley Laboratory in the US were also involved in the work.

According to Kuzmenko, the team made its discovery while using infrared light to study aspects of the quantum Hall effect in graphene. "We didn't expect to see a large [rotation] in graphene," he says "We expected to see a rotation of about 0.01 radians and instead we saw 0.1 radians." The result means that graphene has a bigger Faraday rotation per atomic layer than any other material – beating out its nearest semiconductor rivals in the infrared by a factor of 10.

The team measured the Faraday rotation by passing infrared light through a polarizing filter to create a linearly polarized beam. This beam was then sent through a graphene sample with a magnetic field perpendicular to its surface. After the light emerged, it was passed through a second polarizing filter and on to a detector. If the polarizations of the two filters are exactly 90° apart, no light should be detected. But if the polarization of the light is rotated as it passes through the graphene, the angle at which no light is detected will be shifted by the Faraday angle.

The physicists believe that the large rotation is a result of graphene's electrons behaving as if they have no mass. When subjected to a magnetic field, the electrons occupy a spectrum of circular "cyclotron" orbits that is very different to that found in other materials. Transitions between these orbits affect the circular polarization of the transmitted light and result in a much enhanced Faraday angle.

According to Kuzmenko, the effect could be used to create switches in which light can travel in one direction, but not in the opposite direction. These optical diodes, known as "Faraday isolators", are not currently available for infrared light.

One important benefit of making such magneto-optical devices from graphene is that the direction of the Faraday rotation can be reversed by simply applying an electric field to the graphene. In other materials, in contrast, this is only possible by reversing the applied magnetic field, which is a slower and more complicated process. The reason, according to Kuzmenko, is graphene's unique ability to change the sign of its charge carriers from negative to positive by simply applying an electric field.

Andrea Ferrari of the University of Cambridge in the UK believes that this newly discovered optical property of graphene is yet more evidence that the material's future lies in photonics and optoelectronics. "The Faraday effect and the associated magneto-optical Kerr effect are widely used in optical communications, data storage and computing," he told physicsworld.com. "These, combined with the [other known] properties of graphene, could lead to uniquely performing devices."

There are, however, several challenges involved in making practical devices. One is that about 10 independent layers of graphene would be needed to achieve a rotation of about 45° – which would be required in practical devices. Another problem is that graphene absorbs infrared light, which would lead to significant signal loss in devices.

The research is published in Nature Physics doi: 10.1038/NPHYS1816.

View the original article here

Do giant spiral galaxies thwart clusters of young stars?

To enjoy free access to all high-quality "In depth" content, including topical features, reviews and opinion sign up

Nov 2, 2010

Astronomers in Scotland and Germany say simple physics may explain a long-standing paradox: why large clusters of young stars tend to reside in relatively small galaxies and not in giants like the Milky Way. The reason, according to the astronomers, is that giant spiral galaxies, like the Milky Way, spin fast, shearing star clusters before they grow into monsters.

The stunning 30 Doradus complex is the most luminous nursery of young stars in the Local Group – a collection of several dozen nearby galaxies that includes the Milky Way. It stands to reason, therefore, that 30 Doradus would inhabit an equally impressive galaxy, either Andromeda or the Milky Way, the two largest galaxies in the Local Group.

But instead the stunning 30 Doradus complex lies in the Large Magellanic Cloud, a satellite galaxy of our own that emits only one tenth as much light. The newborn stars of 30 Doradus have set gas aglow over an area 700 light-years wide, 30 times the diameter of the well known Orion nebula.

Now Carsten Weidner and Ian Bonnell of the University of St Andrews in Fife and Hans Zinnecker of the Astrophysical Institute of Potsdam have conducted computer simulations that model interstellar clouds of molecular gas which collapse to form star clusters. Says Weidner, "It seems that rotation inhibits the formation of very massive star clusters."

Giant spiral galaxies spin fast. For example, the Milky Way rotates at about 230 kilometres per second, and the even larger Andromeda galaxy spins faster still. By contrast, smaller galaxies, such as the Large Magellanic Cloud, rotate slowly.

Weidner's team ran four computer simulations, each with a different spin speed. "Each model took about a month to compute," Weidner says. In the fast-spinning models, stars and clusters formed over a wide area, because the spin prevented the gas from collapsing into one gigantic cluster. By contrast, in the slowest-spinning model, the gas collapsed and gave birth to a single huge star cluster at the centre. That model might explain why the huge 30 Doradus complex arose in a galaxy much smaller than our own.

This work also applies to colliding galaxies. Says Weidner, "In the collision region, you have less rotational support, so you would also expect more massive clusters." In fact, the famous Antennae galaxies – two large spiral galaxies that are smashing together in the constellation Corvus – have created young star clusters far greater than any young clusters in either the Milky Way or Andromeda.

Bruce Elmegreen, a star-formation expert at the IBM Research Division in Yorktown Heights, New York, says the study is interesting, but he's sceptical of the result. "The connection between galaxy spin and molecular cloud spin is vague," he says. "Does galaxy spin correlate with the spin of molecular clouds? I'm not aware of an answer to that." Weidner responds that fast-spinning galaxies should indeed have faster-spinning clouds, because the clouds interact with one another.

Elmegreen also says that 30 Doradus may owe its great size to factors other than its home galaxy's slow rotation. The Large Magellanic Cloud – which is only 160,000 light-years from Earth – is plowing through the Milky Way's halo. Gas in the halo compresses gas in the Large Magellanic Cloud, a process called "ram pressure" that may have sparked the star formation in 30 Doradus.

Weidner acknowledges that ram pressure may have played a role. "30 Doradus is a complex object," he says, "and we do not claim that we can explain every detail of it. We just say there might be a trend with rotation."

Weidner and his colleagues will publish their work in The Astrophysical Journal and a preprint (arXiv: 1009.1618) is available.

Ken Croswell is the author of eight astronomy books, including Magnificent Universe.

View the original article here