Page 1813 – Physics World

Home » Page 1813

Thermopower in a spin

The thermoelectric effect relies on the fact that the flow of electrons in a solid produces an entropy current as well as a charge current. The spin of the electrons could also be a source of entropy, although this does not happen in ordinary metals and semiconductors. However, in certain complex materials – such as the layered sodium cobalt oxide material discovered by the Japanese group – the spin might be important.

To prove that spin entropy effects are responsible for the enhanced thermopower of this cobalt oxide, the Princeton team applied a magnetic field in the plane of the material. “We observed that a magnetic field of 10 Tesla suppresses the thermopower completely to zero at temperatures below 4 K,” Ong told PhysicsWeb. This suppression occurs because the magnetic field couples to the spins and traps them in one direction so they can no longer move. “This finding is exceptional because it is rare to see a field alter the magnitude of the thermopower, let alone suppress it to zero.”

The results agree with theory and hold true for a wide range of temperatures. At 2 K the spin entropy term accounts for almost all of the thermopower, and for about two-thirds of it at 300 K.

Superconductivity and magnetism in harmony

According to the Bardeen-Cooper-Schreiffer theory of superconductivity, electrons with opposite spins form pairs that can move through a material without resistance. A magnetic field can destroy superconductivity in two ways: by breaking up the electron pair, or by trying to make both of the electron spins point in the same direction. These effects also limit how much current can flow through the superconductor because of the disruptive effect of the magnetic field produced by the current itself.

Until now, only a few compounds remained superconducting under the influence of an applied magnetic field. Moreover, the number of materials in which an applied field could actually induce superconductivity – by the so-called magnetic field induced superconductivity effect – were very few.

Lange and co-workers placed a layer of cobalt-palladium ferromagnetic dots, each 800 nanometres in diameter and separated by 1.5 micrometres, on top of a superconducting thin film made of lead. Each dot produces a stray magnetic field that destroys the superconductivity in the thin film. The researchers then applied an external magnetic field, which enhanced the destructive effect of the dots’ magnetic field in the area directly beneath the dots and, to compensate, reduced it everywhere else in the film. The overall effect was an increase in the current carried by the superconductor.

“In fact, the technique can be compared with an array of small sponges, which are already a bit wet, put on a wet floor: the sponges will absorb the water,” said team leader Victor Moshchalkov. “The floor between the sponges will become dry at the expense of having more water in and under the sponges.”

This new ‘field compensation effect’ is not restricted to specific superconductors, the researchers say, so magnetic field induced superconductivity could be achieved in any superconducting thin film. The team believes that using nanodots and nanopillars, which have larger stray fields, could allow superconducting materials to remain in higher magnetic fields. The nano-dot array could also be used to design logical devices for use in quantum computers.

Cobalt breaks magnetism record

Magnetic anisotropy is one of the most important properties of a magnetic material, and the MAE controls the alignment of the atomic spins that give rise to magnetism in a material. The larger the MAE, the more stable the magnet.

Gambardella and co-workers deposited single cobalt atoms onto a platinum substrate using molecular beam epitaxy, applied a magnetic field of up to 7 tesla, and then measured the magnetisation of the cobalt atoms both parallel and perpendicular to the field. They calculated a MAE value of 9.3 +/- 1.6 meV per cobalt atom, which is about 200 times larger than that of cobalt atoms in a bulk crystal. In comparison, samarium cobalt, a widely used permanent magnet, has a MAE of just 1.8 meV per cobalt atom.

At present, over 100 000 atoms are needed to make a stable magnetic bit for use in a hard disk. As the MAE of cobalt is so high, only a few hundred atoms would be needed for one bit, the researchers say. This would allow the information storage density to be greatly increased.

Entanglement reaches new lengths

A quantum computer could, in principle, outperform a classical computer by exploiting the ability of a quantum system to be in two states – often called 0 and 1 – at the same time. When two qubits are entangled, they behave as one system: this means that the quantum state of one qubit directly depends on the state of the other. It was once thought that entanglement was only possible with individual quantum particles – such as photons – but recent experiments have shown that macroscopic objects can also be entangled. Superconducting materials are good for making qubits because decoherence effects – which wipe out quantum behaviour – can be limited.

Berkley and co-workers made their qubits from a Josephson Junction – a type of superconducting ‘reservoir’ – and coupled two qubits together using a capacitor. Under certain conditions, the qubits can exist in one of two states: a ground state or an excited state. When the two qubits are entangled, if qubit 1 is in the ground state then qubit 2 is in the excited state, and vice versa.

The researchers measured the entangled states by applying microwaves to the system and recording transitions from the ground state to higher energy states. “Such evidence for entanglement over a macroscopic length is particularly promising for the construction of a quantum computer, as this will require many spatially separated qubits,” said Berkley.

Route one to the centre of the Earth

The centre of the Earth is thought to consist of a solid inner core and a molten outer core. Both are made up mainly of iron but also contain small amounts of nickel and possibly oxygen. Our knowledge of the core comes from a variety of indirect measurements that include the study of seismic waves and earthquakes, the analysis of meteorites, and high-pressure laboratory experiments.

Stevenson proposes sending down a small “grapefruit-sized” probe, placed in a large volume of liquid-iron alloy. The mass of iron contained in the volume would need to be about 10⁸ kg, which is equivalent to the amount of iron produced by the world’s foundries in about an hour. The probe would travel at about 5 metres per second along a crack under the influence of gravity, and would reach the centre in about a week. It would be made of an alloy with a high melting point and would include instruments to measure temperature, electrical conductivity and the abundance of various elements.

The signal sent out by the probe could, in principle, be detected by the Laser Interferometer Gravitational Wave Observatory (LIGO) – a facility built to detect gravitational waves. However, LIGO is currently isolated from vibrations in the ground and could not be sensitive to both gravitational waves and signals from the probe at the same time. Stevenson believes that his proposal is “modest” compared to the space programme and may appear unrealistic only because so little effort has been devoted to such direct measurements.

Could neutrinos destroy nuclear weapons?

Neutrinos are one of the fundamental particles of matter and come in three ‘flavours’ – electron, muon and tau neutrino. They are electrically neutral and only interact weakly with matter, which means that they can pass through thousands of kilometres of matter without being absorbed.

In 1999 the first so-called long-baseline neutrino oscillation experiment, K2K, involved sending a neutrino beam from KEK to the Superkamiokande detector 250 km away. There are plans underway to send a neutrino beam from Fermilab to the Soudan lab in Minnesota, 710 km away and from CERN to Gran Sasso in Italy, 730 km away. The new method for destroying nuclear weapons proposed by Hirotaka Sugawara, Hiroyuki Hagura and Toshiya Sanami is a “vast extrapolation” of such experiments.

The researchers suggest sending a neutrino beam with an energy of 1000 TeV through the Earth to wherever the nuclear weapon was located (see figure). The beam would produce neutrons in a ‘hadron shower’ and would cause fission reactions in the plutonium or uranium in the bomb. These reactions would either melt or vaporize the bomb.

Such a high energy neutrino beam would be difficult to produce, the physicists admit. The storage ring would have to be 1000 km across – hundreds of times larger than the biggest present day accelerators. The magnets in the specially built muon storage ring would need to be one to two orders of magnitude stronger than those currently available to construct a realistically sized machine. Moreover, the cost of building such a device could be over $100 billion and it would consume 50 GW of energy – the entire power consumption of the United Kingdom.

Finally there is the risk, the authors point out, that the interaction of the neutrino beam with the bomb “could lead to a full explosion” instead of eliminating it.

First light for pure silicon

Silicon dominates the microelectronics industry but it is not used in optoelectronics applications because it is believed that it does not emit light efficiently. Previous research on improving silicon light emission has focused on porous silicon, nanocrystalline silicon and silicon doped with rare-earth metals such as erbium and cerium.

Green and co-workers excited a variety of commercially available pure silicon wafers using light with a wavelength of 780 nm from a laser diode. They calculated the ‘external’ light emission efficiency by dividing the number of photons emitted by a sample by the number of photons used to excite it. Photons are emitted when electrons that have been photo-excited into the conduction band recombine with ‘holes’in the valence band.

The team calculated that the external light emission efficiency was up to 6.1% at room temperature and up to 10.2 % at 130 K. The researchers then used a theoretical model to predict that the ‘internal’ light efficiency is larger than 20% at room temperature under optimal excitation conditions. The internal light efficiency is the ratio of the number of photons that are spontaneously emitted inside the sample and the total internal recombination rate per unit area.

“The result means that ‘radiative recombination’ can be one of the dominant recombination channels in pure crystalline silicon,” team member Thorsten Trupke told PhysicsWeb. “This is in stark contrast to a widely accepted perception that silicon is an inherently poor light emitter.” The group is now investigating ways to modulate the light emitted by the silicon. It also hopes to improve the surface texture of the samples to further increase efficiency.

Tinnitus treatment tones up

If you were exposed to ‘white noise’ – a broadband sound signal – that contained a frequency gap, your brain would ‘hear’ a single-frequency sound for several seconds after the real signal had ended. This auditory illusion is known as the Zwicker tone and is surprising because the note ‘heard’ by the brain has a frequency that falls in the ‘gap’ in the original signal.

Scientists agree that the Zwicker tone does not originate in the cochlea or the auditory nerve, but until now they have also struggled to explain the effect in terms of neuron activity. The unusual frequency characteristics rule out the possibility that stimulated neurons are active for a short time after the signal is switched off, which makes the Zwicker tone fundamentally different to the ‘habituation’ of neurons that produces after-images in the visual system.

Now Franosch and co-workers say that anti-noise neurons in the brain could produce the Zwicker tone, which occurs for many sound configurations that contain broadband noise. Imagine a pure tone superimposed onto white noise. According to the researchers, the neurons that respond to pure tones suppress the activity of noise-reducing neurons over a narrow frequency range around the frequency that the particular pure-tone neuron is sensitive to. But this ‘hole-burning’ effect is asymmetrical and the anti-noise neurons on the low-frequency side of the pure tone are still inhibited after the signal is switched off – so the brain ‘hears’ a tone of this lower frequency.

Franosch and colleagues have performed computer simulations that they say can explain all evidence of the Zwicker tone to date. They also hope their model will provide new insights into how the auditory system handles noise. This could help some sufferers of tinnitus, who persistently ‘hear’ noises with frequencies at which they have hearing loss.

Potassium-40 heats up Earth’s core

Potassium-40, which has a radioactive half-life of about 1.2 billion years, could be an important source of heat in the Earth’s core but this has never been unambiguously confirmed in an experiment. Murthy and colleagues used an iron and iron- sulphur mixture to represent the Earth’s core and potassium silicate glass to represent the shell. They measured the partition coefficient – the concentration of potassium-40 in the sulphur mix divided by its concentration in the silicate – at temperatures and pressures approaching those found deep in the Earth’s mantle.

The researchers found that the logarithm of the partition coefficient is inversely proportional to temperature. The results suggest that potassium-40 can move from the silicate ‘shell’ to the iron-sulphur ‘core’ and that it would be possible for a high enough concentration of potassium-40 to build up in the core.

The team calculated a core potassium-40 content of between 60 and 130 ppm, which produces between 0.4 and 0.8 TW of heat. Estimates of the core-mantle boundary heat flux are between 8 and 10 TW, so the heat produced by potassium-40 could significantly contribute to the heat flux at the boundary. Recent studies have shown that the present level of heat flux would have been insufficient to sustain the Earth’s magnetic field for the past 3.5 billion years. This ‘extra’ radioactive heat could thus have allowed the field to exist.

“We now plan to expand these measurements to much higher pressures and temperatures,” Murthy told PhysicsWeb. “We shall also extend the experiments to the other major radioactive heat sources in the Earth, uranium and thorium.”

Carbon nanotubes light up

Light-emitting devices rely on charge carriers – electrons and holes – being brought together so that they can recombine to emit photons. Single-walled carbon nanotubes have been used as field-effect transistors before and now the IBM team has succeeded in obtaining light from them. Previous nanotubes have only emitted light when excited by another light source such as a laser.

Avouris and colleagues used single nanotubes to make a three-terminal FET device. They randomly dispersed the nanotubes – each about 1.4 nanometres in diameter – onto a silicon substrate that contained a 150 nanometre silicon dioxide layer. ‘Source’ and ‘drain’ contacts were then added at either end of the device so that electrons and holes could be injected (figure 1). The electrons and holes recombined in the nanotube to emit infrared radiation at wavelengths longer than about 0.8 microns. This included light at a wavelength of 1.5 micrometres, which is widely used in fibre-optic communications (figure 2).

Figure 2

The device does not rely on doping to create charge carriers, as silicon transistors do, but is ‘biased’ so that one part of the nanotube conducts electrons while the other conducts holes. This is achieved by the formation of Schottky barriers – potential barriers that electrons can tunnel through – at the source and drain.

The wavelength of the emission is determined by the band gap of the nanotube, which depends on the diameter of the nanotube. Changing the thickness of the silicon dioxide layer and using other materials to construct the device might improve the overall efficiency according to the IBM team.

Browse articles by content type