VIEWS: 12 PAGES: 14 POSTED ON: 2/2/2011
Topics to cover: 1. Resistance, magnetic ﬁeld and heat capacity observations. Statistical and Low Temperature Physics (PHYS393) 2. Explanation using macroscopic wavefunction. 7. Superconductivity 3. Quantised vortices. 4. Cooper pairs. Kai Hock 2010 - 2011 5. Applications: trains, accelerators, ... University of Liverpool Superconductivity 1 Zero resistance. Examples of superconductors. Metals conduct electricity. Normally, there is always some resistance, however small. In some materials, this resistance suddenly falls to zero below a certain temperature. In 1911, Kamerlingh Onnes discovered that this happened with mercury below 4.2 K http://hyperphysics.phy-astr.gsu.edu/hbase/solids/scond.html Notice that these are metals, and that the transition temperatures are close to liquid helium temperature. Superconductivity 2 Superconductivity 3 Meissner eﬀect. Measuring magnetic ﬁeld. When the resistance drops to zero, the superconductor all This graph shows the magnetisation of lead in liquid helium, expels all magnetic ﬁeld from its body. plotted against the applied ﬁeld. Livingston, Physical Review, vol. 129 (1963), p. 1943 http://www.materia.coppe.ufrj.br/sarra/artigos/artigo10114/index.html Below a certain critical ﬁeld, the magnetisation is equal and The ﬁeld inside the body of a superconductor can be obtained opposite to the applied ﬁeld. So the resultant ﬁeld inside the by inserting it in a coil and measuring the induced voltage. superconductor is zero. Superconductivity 4 Superconductivity 5 Levitation. Heat capacity. The expulsion of magnetic ﬁeld from a superconductor is called Recall the heat capacity of a normal metal: is Meissner eﬀect. Cv = γT + AT 3. A striking demonstration is the levitation of a superconductor Measurements show that for supercondctors, this changes above a magnet. completely below the transition temperature. This graph is the result of measurement for the niobium metal. Brown, et al, Physical Review, vol. 92 (1953), p. 52 Superconductivity 6 Superconductivity 7 If we select the normal conducting state of niobium by applying Niobium become superconducting below 9.5 K. It is possible to a strong magnetic ﬁeld, we would measure the curve labelled prevent it from becoming superconducting by applying a “normal.” suﬃciently large enough magnetic ﬁeld. We know from the Meissner eﬀect that a niobium expels all magnetic ﬁeld. However, if the ﬁeld is strong enough, it can “force” its way into the superconductor. This destroys the superconductivity and returns the niobium to a normal conducting state - even if temperature is below 9.5 K. Using this property, it is possible to select between the normal and the superconducting state. This follows the “normal” behaviour of Cv = γT + AT 3. Superconductivity 8 Superconductivity 9 The Macroscopic Wavefunction. If we do not apply any magnetic ﬁeld, we get the superconducting state. Then we would get the curve labelled In 1937, Fritz London suggested that if the electrons in a “superconducting.” superconductor somehow forms a macroscopic wavefunction. Using this assumption, London was able to explain it expels all magnetic ﬁeld. To understand this, we ﬁrst need to appreciate why expulsion of the magnetic ﬁeld is strange. Suppose the resistance going to zero is the only change in a metal. Consider what happens if we now bring a magnetic to the metal. If we subtract the phonon contribution of AT 3, we would ﬁnd that the curve is closer to the exponential form: The change in magnetic ﬂux through the metal induces an electric current, according to Faraday’s law. C = a exp(−b/T ) for some connstants a and b. This looks like the Boltzmann distribution. Superconductivity 10 Superconductivity 11 Lenz’s law. It looks like we have just “explained” the Meissner eﬀect. However, let us now look at what happens if the magnet is According to Lenz’s law, the current would ﬂow in such a way already there before cooling. as to produce a magnetic ﬁeld of its own that opposes the incoming ﬁeld. We start with a normal metal with a magnetic ﬁeld going through the body. Then we cool this down and the resistance In a metal with resistance, this induced current would quickly falls to zero. slow down to zero. The induced ﬁeld becomes zero, and only the incoming ﬁeld remains in the body of the metal. According to Faraday’s law, since there is no change in magnetic ﬂux, no current is induced. So the original ﬁeld from If the metal has no resistance, the induced current continues to the magnet remains in the body. ﬂow. The induced ﬂux has to be opposite to the incoming ﬂux. Therefore they cancel, and the ﬁeld in the body becomes zero. In a real superconductor, we know from the Meissner eﬀect that, even in this case, the magnetic ﬁeld must be expelled. In this way, the ﬁeld is “expelled.” This shows that there is something diﬀerent about a superconductor that the familiar laws of electromagnetism cannot explain. Superconductivity 12 Superconductivity 13 Macroscopic wavefunction. Vector potential We shall now see how a macroscopic wavefunction, ψ, can explain the Meissner eﬀect. In order to use the vector potential, lets review its meaning. It is deﬁned by Recall the operator in quantum mechanics for momentum: ∇ × A = B, dψ where B is the magnetic ﬁeld. This a bit similar to relation −i = pxψ dx between the electric ﬁeld and electric potential. where p is the momentum mv. http://en.wikipedia.org/wiki/Magnetic_potential In the presence of an electromagnetic ﬁeld, this is changed to dψ For a qualitative understanding, the integral form of this −i = (mv − qA)ψ equation is suﬃcient: dx where A is the vector potential and q the charge of the particle. A.dl = B.dS, C S http://quantummechanics.ucsd.edu/ph130a/130_notes/node29.html where the left integral is along any loop C, and and the right integral is over any surface S enclosed by the loop. Both equations are quantum mechanical postulates that have been shown to give correct results in physics. Superconductivity 14 Superconductivity 15 Ampere’s law. Phase. The right side of this equation is the magnetic ﬂux Φ, Let us now return to the quantum mechanical equation: dψ A.dl = B.dS −i = (mv − qA)ψ. C S dx and the left side is the line integral for magnetic potential. Recall the wavefunction we used for superﬂuids: ψ = e−iφ(x) If we make the following replacements: A → B and B → J, where J is the current density, we get Ampere’s law. where φ(x) is the phase. Substituting into the equation, we get dφ = mv − qA. In the more familiar Ampere’s law, the electric current is related dx to the integral of magnetic ﬁeld over a loop round the current. This relation along a straight line in x can be extended in a simple way to any path or loop in 3D. In the same way, the equation Consider a loop in a superconductor of length L enclosing an A.dl = Φ C area S. Integrating along this loop, we get tells us that magnetic ﬂux is equal to the integral of vector ∆φ = m v.dl − q A.dl. potential over a loop round the ﬂux. L L Superconductivity 16 Superconductivity 17 Let us now see how this equation m ∆φ = J.dl − qΦ. ∆φ = m v.dl − q A.dl. ρq L L L can help us understand Meissner’s eﬀect. The phase change ∆φ is zero or a multiple of 2π, because the wavefunction returns to the same value after one loop. For a simple lump of metal, the wavefunction would be continuous through the whole volume, so the phase change The integral over A gives the magnetic ﬂux Φ. would be zero. The equation then simpliﬁes to m The velocity v is related to the current density J by J.dl = qΦ. ρq L J = ρq v, This means that: where ρ is the number density of the electrons. The above equation then becomes if there is a magnetic ﬁeld in the macroscopic m wavefunction, then is a there is an electric current. ∆φ = J.dl − qΦ. ρq L To see why this is special, consider Faraday’s law again. Superconductivity 18 Superconductivity 19 Meissner eﬀect. London’s penetration depth. According to Faraday’s law, a change in magnetic ﬂux is Flux from the wavefunction, or superconducting, current would required before a current can be induced. cancel some of the incoming ﬂux. For a macroscopic wavefunction, the very presence of the ﬂux The amount cancelled depends on the density of the electrons produces the current. No change in ﬂux is needed! in the wavefunction. The higher the density, the larger the superconducting current, and more of the incoming ﬂux would Let us look at the case of transition to the superconducting be cancelled. state again. Previously, we have not been able to explain the expulsion of the ﬁeld using Faraday’s law. For a uniform external ﬁeld, this superconducting current would typically be circulating the metal. So it produces the greatest We can now explain this assuming that a macroscopic ﬁeld at the centre, where more cancellation takes place. wavefunction appears when the metal becomes superconducting, If there is a magnetic ﬁeld in the metal, it For larger electron density, the region of cancellation is also would produce a current. This current would in turn produce a larger. In a typical superconductor, there is suﬃcent density to ﬂux. expel the incoming ﬁeld from most of the volume. A more detailed reasoning would show that this wavefunction In practice, some ﬁeld would penetrate to a depth of about 100 ﬂux is in the opposite direction to the incoming ﬂux. nm on the surface. Superconductivity 20 Superconductivity 21 Vortices. The reason for the penetration depth is that a current is In the lectures on superﬂuid helium, we have seen that a needed to keep the ﬁeld expelled. macroscopic wavefunction can give rise to vortices that quantised. Recall that a ﬁeld must be present in the macroscopic wavefunction in order to produce the current. As the ﬁeld gets If the electrons in a superconductor also forms a macroscopic expelled from the center of the superconductor, the current at wavefunction, quantised vortices should also be possible in the the center would also stop. electrons. This is indeed observed: If the ﬁeld is completely expelled from the metal, there would be no current at all in the metal. Then there would be no opposing ﬂux to cancel the incoming ﬂux. The external ﬂux would come in again and start producing current. For this reason, a balance would to be reached. The ﬁeld would penetrate until a depth when there is suﬃcient current to keep the rest of the volume ﬁeld free. Essmann and Trauble, Physics Letters 24A, 526 (1967) Superconductivity 22 Superconductivity 23 Observing vortices. Likewise, Essmann and Trauble sprinkled some cobalt powder The method used to observe vortices is similar to the method on a Lead-Indium alloy. This is what they saw under an for observing magnetic ﬁeld lines in school. electron microscope: Sprinkle some iron ﬁlings on a piece of paper, place a magnet underneath, tap the paper gently, and this is what you would see: The cobalt powder collected at the centres of the vortices, where magnetic ﬁelds are strongest. A nice gallery of superconducting vortices can be found here: http://www.fys.uio.no/super/vortex/index.html Superconductivity 24 Superconductivity 25 Type II superconductors. Type II superconductors. The existence of vortices is in fact not consistent with This shows the magnetisation of Lead alloy with diﬀerent Meissner’s eﬀect. amount of Indium: We have learnt that when a metal becomes superconducting, it expels all magnetic ﬁeld (except for some near its surface). A vortex in a superconductor is a circulating current. This must produce a magnetic ﬁeld in the superconductor. This contradicts the Meissner’s eﬀect. Below a certain critical ﬁeld, the magnetisation (e.g. OB) is It turns out that the Meissner’s eﬀect is only true for some strong enough to cancel the applied ﬁeld. metals - mainly pure metals. These are called Type I superconductors. For higher ﬁeld, the magnetisation decreases (e.g. curve to the right of B). It is not enough to cancel the applied ﬁeld, which For alloys and other materials, it is possible for magnetic ﬁeld then penetrates the superconductor. These are called type II to penetrate the body of the superconductor to some extent. superconductors. Superconductivity 26 Superconductivity 27 Flux quantisation. Flux measurement. We have seen that vortices of electrons do exist in a Deaver and Fairbank measured the ﬂux through a long, thin superconductor. Lets now look at whether they are quantised. tube made of Tin: If the current around a vortex is quantised, so is the magnetic 1. Apply a magnetic ﬁeld to the tube. ﬂux produced. This can be measured, the has indeed been found to be quantised. 2. Cool below the 3.7 K transition temperature. 3. Move the tube up and down rapidly. 4. Place a coil near the end of the tube. 5. Measure the voltage induced in the coil. 6. Obtain the ﬂux from the voltage. Deaver and Fairbank, Physical Review Letters, vol. 7 (1961) p. 43 Superconductivity 28 Superconductivity 29 The measured ﬂux is plotted against applied ﬁeld: In the case of the superﬂuid, no magnetic ﬁeld is involved. So we still need to understand how vortex arise in the superconductor. Recall the relation between ﬂux and current in a macroscopic wavefunction: m ∆φ = J.dl − qΦ. ρq L The Tin tube is a solid with a hole through it. The wavefunction is no longer continuous over the whole volume, so phase change around the tube does not have to be zero: m The steps show that the possible ﬂux through the tube is 2nπ = J.dl − qΦ. ρq L indeed quantised. The magnitude of each step is The equation is true for any loop L in the wavefunction. It is h possible to choose the loop in such that the integral over Φ= . 2e current J is zero. Superconductivity 30 Superconductivity 31 This ﬁgure shows the cross-section of the Tin tube. We are interested in the ﬂux through the hollow. So if we choose the loop L away from either surfaces, then the current density along L would be zero. The equation m 2nπ = J.dl − qΦ. ρq L then becomes 2nπ = qΦ. The “-” sign can be left out if we are only interested in the magnitudes. Since q is the charge of an electron, the ﬂux is nh Φ= . e This means that one quantum step is h/e. When this is superconducting, current is only possible very near the surfaces A and B, within the penetration depth. Further in We have just found a problem. the bulk, there is no current because there is no ﬁeld, since all ﬁeld is expelled. Superconductivity 32 Superconductivity 33 Cooper pair. Using the macroscopic wavefunction, we have found that the ﬂux is quantised in steps of h/e. Notice the diﬀerence: The measurement results tell us that the ﬂux is quantised in Theory predicts h/e. Measurement gives h/2e. steps of h/2e. This means that something must be wrong with the theory. It seems to suggest that, instead of a charge of e, the particle should have a charge of 2e. This is one of the evidence to suggest that the electrons might somehow be moving in pairs. Superconductivity 34 Superconductivity 35 The Isotope Eﬀect. Lattice Vibration. If electrons repel each other, how can they form a pair? The clue: In 1950, the superconducting temperature of Why do the neutrons change the superconducting temperature? Mercury was found to be diﬀerent for diﬀerent isotopes of Mercury. One possible reason is that the movement of the atoms are somehow involved in causing the superconductivity. More neutrons means more mass. This would result in slower movement of atoms. This provides an important clue: Lattice vibration is known to scatter electrons and cause resistance. Reynolds, et al, Physical Review, vol. 78 (1950) p. 487 The only diﬀerence between isotopes is the number of neutrons in the nuclei. This should not aﬀect the conduction electrons! Superconductivity 36 Superconductivity 37 How electrons “attract” When a electron moves in a metal, it can attract the positive ions and bring them closer. The attractive potential between electrons is much smaller than the kinetic energy of the two electrons. So it should not normally be able to bind the electrons together. However, in this case, the two electrons are not in free space. They are in a Fermi sea - electrons stacked up to the Fermi energy. Another electron may then get attracted to the displaced ions. In the 1950s, Leon Cooper showed that two electrons near the Fermi energy is is able to form a bound pair. Bardeen, Cooper and Schrieﬀer (BCS) then developed a complete theory to that is able to explain the Meissner’s eﬀect, the zero resistance, the heat capacity behaviour, and other phenomena of superconductors. http://hyperphysics.phy-astr.gsu.edu/hbase/solids/coop.html Superconductivity 38 Superconductivity 39 Cooper pair in real space BCS versus BEC The wavefunction of an electron in a Cooper pair in real space The electrons in the pair have opposite spin, so that resultant is not unlike that of an electron around an atom. spin of the Cooper pair is zero - it is a boson. So, like the Bose-Einstein condensate, the Cooper pairs can condense into the ground state and form a condensate. Kadin, Spatial Structure of the Cooper Pair (2005) Ketterle and Zwierlein, Making, probing and understanding ultracold Fermi gases (2006) The size of the Cooper pair is a few hundred times the spacing between atoms, so there is a lot of overlap between Cooper However, because of the considerable overlap, it is normally pairs. called a BCS condensate instead. Superconductivity 40 Superconductivity 41 Energy gap. As an example of a prediction by the BCS theory, recall the Rearranging the relation gives this ratio: behaviour of heat capacity in a superconductor, 2∆ C = a exp(−b/T ). This can be written in the form: = 3.52. kB Tc ∆ The ratio for measured values are shown here: Cv = D exp − kB T This looks like the Boltzmann factor, in which ∆ is the energy between two levels. In the BCS theory, ∆ is the energy needed to excite one electron from the BCS condensate. This energy is now called the energy gap. It can be obtained directly from a heat capacity measurement by ﬁtting the above formula. Meservey and Schwarz, in Parks (1969) Superconductivity BCS theory predicts that the energy gap and the transition temperature are related by: The ratios are all fairly close to 3.52. This is another evidence 2∆ = 3.52kB Tc. that supports the BCS theory. Superconductivity 42 Superconductivity 43 BCS superﬂuid. Applications of superconductors. The most obvious property about a superconductor is the zero resistance. Unfortunately, there does not appear to be a simple Existing applications of superconductivity include: way to explain this. 1. Maglev train. The Cooper pairs can carry electric current, but why does it not get scattered by phonons and experience resistance? 2. Magnetic Resonance Imaging (MRI) Victor Weisskopf suggested that the Cooper pairs are packed like atoms in the helium-4 superﬂuid, and has zero resistance for similar reasons. So the diﬃculty in scattering a Cooper pair 3. Particle accelerators (e.g. LHC) is a result of interaction with other Cooper pairs. http://cdsweb.cern.ch/record/880131/files/p1.pdf 4. Detecting weak magnetic ﬁeld (SQUIDS) The Cooper pairs would ﬂow like a superﬂuid, unless there is http://www.superconductors.org/uses.htm enough energy to break all of them. This would happen at the transition temperature: kB Tc ≈ ∆. Superconductivity 44 Superconductivity 45 Maglev train. Magnetic Resonance Imaging A train can be levitated above its track using powerful, MRI requires a very strong magnetic ﬁeld. This is produced superconducting magnets, so that there is little friction. using supercondctors. One, built in Japan in 2005, travelled at half the speed of sound. http://www.magnet.fsu.edu/education/tutorials/magnetacademy/mri/ (and Wikipedia) http://en.wikipedia.org/wiki/Maglev_(transport) Superconductivity 46 Superconductivity 47 Particle accelerators Detecting weak magnetic ﬁeld A superconducting device called SQUID can detect very weak Particle accelerators use superconducting magnets and rf magnetic ﬁelds. cavities to accelerate particles to high energies. The Large Hadron Collider: (Wikipedia) It is useful for: - detecting brainwave, - diagnosing problems in various parts of the human body, - as an MRI detector, - oil prospecting, - earthquake prediction, - submarine detection, etc. Superconductivity 48 Superconductivity 49 High Temperature Superconductors Characteristics In 1986, materials that become superconducting above liquid nitrogen tempratures are discovered. This generated a lot of The ﬁrst of the high temperature superconductors discovered is excitement about possible applications, because liquid nitrogen YBCO (Yttrium-Barium-Copper-Oxide), is much cheaper than liquid helium. Being copper oxides, these materials are very poor conductors of electricity at room temperature. When they do become superconducting at liquid nitrogen temperatures, there are fewer Cooper pairs compared to metallic superconductors. As a result, they are strongly type II when superconducting - ﬂux lines can penetrate. Notice that the examples above 77 K are copper oxides. Superconductivity 50 Superconductivity 51 Applications? Using these high temperature superconductors for MRI, trains, LHC and SQUID would remove the need for expensive liquid helium. Unfortunately, being copper oxides again, they are brittle and very diﬃcult to make into electrical wires. Today, scientists are still trying to solve these engineering problems. Superconductivity 52