High pressure induced superconductivity

High pressure induced superconductivity

Physica C 392–396 (2003) 17–21 www.elsevier.com/locate/physc High pressure induced superconductivity K. Amaya *, K. Shimizu Graduate School of Engine...

144KB Sizes 0 Downloads 55 Views

Physica C 392–396 (2003) 17–21 www.elsevier.com/locate/physc

High pressure induced superconductivity K. Amaya *, K. Shimizu Graduate School of Engineering Science, Osaka University, Machikaneyama 1-3, Toyonaka, Osaka 560-8531, Japan Received 13 November 2002; accepted 28 February 2003

Abstract We have developed complex extreme condition of very low temperature down to 30 mK and ultra high pressure exceeding 200 GPa by assembling compact diamond anvil cell (DAC) on a powerful 3 He/4 He dilution refrigerator. We have also developed measuring techniques of electrical resistance, magnetization and optical measurement for the sample confined in the sample space of the DAC. Using the newly developed apparatus and techniques, we have searched for superconductivity in various materials under pressure. In this paper, we will shortly review our newly developed experimental apparatus and techniques and discuss a few examples of pressure induced superconductivity which were observed recently. Ó 2003 Elsevier B.V. All rights reserved. PACS: 74.25.Dw; 74.62.Fj Keywords: Pressure induced superconductivity; Iron; Lithium

1. Production of very low temperature and ultra high pressure In the past 15 years, we have developed the apparatus for complex extreme condition of very low temperature and ultra high pressure as well as the measuring techniques to be used in the extreme condition. We noticed that a compact diamond anvil cell (DAC) could be easily assembled on a powerful 3 He/4 He dilution refrigerator and that the combined system enabled us to attain the very low temperature of 30 mK and ultra high pressure

exceeding 200 GPa at the sample position. For this purpose, we employed good thermal conductors and non-magnetic materials for all parts of the pressure cell to perform the very sensitive measurement such as Meissner effect detection at very low temperature. Also, DAC was designed so as to produce close pressure values both at room temperature and liq. He temperature to avoid the breaking of diamonds during the heat cycles. Our typical DAC is shown in Fig. 1.

2. Measuring techniques for electrical resistance and magnetization * Corresponding author. Address: Graduate School of Engineering Science, Osaka University, 14-2 Hamakazecho Ashiya, Hyogo 659-0032, Japan. Tel./fax: +81-797-22-1720. E-mail address: [email protected] (K. Amaya).

For the purpose of detecting superconducting transition, we employed sensitive ac-4 terminal method for electrical resistance measurement [1].

0921-4534/$ - see front matter Ó 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0921-4534(03)01072-4


K. Amaya, K. Shimizu / Physica C 392–396 (2003) 17–21

Fig. 3. Schematic drawing of the detecting coil wound around a diamond for magnetic measurements using a SQUID magnetometer. Fig. 1. Photograph of the DAC made of non-magnetic Be–Cu which used for measurements at pressures up to 250 GPa and temperatures down to 30 mK. (A) upper diamond holder, (B) main body, (C) cylinder with lower diamond, (D) loading nut, d––diamond, s––plastic ring.

In Fig. 2, configuration of Au or Pt electrodes on the pressure surface of diamond is shown. In the case of Fig. 2a, the sample is pressed directly by one of the diamond. However, to improve the quality of pressure, proper pressure media such as mixed alcohol or NaCl is used as shown in Fig. 2b. In both cases, hard metallic gasket such as rhenium is covered with thin layer of aluminum oxide for electrical insulation between the gasket and the electrodes. The setting of the electrodes is made by hand and the extremely small size of the sample of the order of 10 8 cm3 makes this procedure very difficult. We tried to use the lithography technique

to make fine pattern on the pressure surface of diamond. But the contact between deposited electrodes and diamond was not so strong as to stand against ultra high pressure and further improvement is still going on. On the other hand, magnetization measurement is rather simple. A pair of pick up coils is connected in series and the out put signal is introduced into the probe of sensitive magnetometer SQUID as shown in Fig. 3. One of the coil is for the sample in DAC and the other is for the reference sample as well as for compensation against uniform external noise. This dc method is available at temperatures below several K, where the pick up coil is at superconducting state. To extend the sensitive magnetization measurement using SQUID magnetometer up to liquid nitrogen temperature region, vibrating coil method was employed and has

Fig. 2. Schematic drawing of the arrangement of sample and electrodes on the gasket. (a) Conventional configuration, (b) hydrostatic configuration; d––diamond, m––gasket, a––aluminum insulating layer, e(p)––platinum electrode, g––gold wire, s––sample.

K. Amaya, K. Shimizu / Physica C 392–396 (2003) 17–21

been developed successfully [2]. Typical example of application of this ac method is seen in the detection of superconducting transition of vanadium at 17 K, which was the maximum TC value among elemental materials [3]. As well as electrical and magnetic measurement, optical measurement is very important to perform the material study of the sample in DAC. Fortunately, we obtain optically transparent synthetic diamond for the various optical studies such as infra-red absorption, Raman scattering and so on. However, in the present study we performed optical measurement only for determination of pressure value by ruby method. Several small ruby chips on the sample are irradiated by argon laser and empirical relation between observed wave length of ruby fluorescent line and pressure value is employed for determination of pressure up to about 200 GPa. There are considerable developments in other field of measuring techniques. For example, structural analysis of the sample in DAC has been performed recently even at temperatures below 1 K or high temperature by laser heating in the facility of the synchrotron orbital radiation but details are to be discussed elsewhere.

3. Examples of pressure induced superconductivity 3.1. Iron under pressure It is well known that a ferromagnetic metal does not show superconductivity down to the lowest temperature and even a small amount of paramagnetic impurities could suppress considerably the superconducting transition temperature TC of non-magnetic metals because of the pair-breaking magnetic interaction between the Cooper pairs and the local magnetic moments. So long as we look at the periodic table of elements, the above idea of absence of TC of magnetic metals seems to be accepted quite in general. On the contrary, we may expect the appearance of TC of magnetic metals in its non-magnetic state under certain high pressure and at low temperature. In the case of iron, there appears non-magnetic state after the crystal phase transition from bcc to


hcp phase at around 10 GPa. The non-magnetism of hcp iron (eFe) is confirmed by Mossbauer experiments [4]. Also, there are several theoretical predictions on TC of iron in its non-magnetic state, giving relatively low TC below 1 K [5]. We have searched for superconductivity of iron under pressure in the past 10 years. After searching in the wide range of pressure up to 100 GPa and temperature down to 50 mK, we noticed that purification of the sample and application of quasi hydrostatic pressure using proper pressure media should be the most important. In fact, pressurization of sample in pressure media such as NaCl was found to be very effective in the present case of iron, avoiding excess crystal distortion and uniaxial pressure distribution in the course of pressurization. According to the Mossbauer effect experiment by Nasu of Osaka University, [4] the phase transition from bcc to hcp phase is found to take place most successfully under hydrostatic pressure and also to show that the amount of the remaining bcc Fe in the hcp Fe is pressure dependent and could be reduced down to the amount within the resolution of Mossbauer experiment after application of pressure up to 100 GPa. On the other hand, by heating the iron sample rod in a high vacuum chamber, we tried purify the starting sample. This was done successfully by Onuki of Osaka University. We performed electrical resistance and magnetization measurement of purified samples under quasi hydrostatic pressure. At relatively high temperature region of 100 < T < 300 K, we observed linearly decreasing resistance with decreasing temperature and at low temperature below 5 K, the temperature independent residual resistance was observed. By decreasing temperature further, small but sudden drop of resistance was observed at just below 2 K under pressure of 23 GPa. This drop is also confirmed to be magnetic field dependent and disappears under the external field exceeding 0.2 T as shown in Fig. 4. Therefore, the observed resistance drop is expected to be due to superconducting transition. We studied the pressure dependence of TC and found that TC appears at 15 GPa and disappears at 30 GPa, giving the maximum value of 2 K at 21 GPa as shown in Fig. 5. Further, we could confirm the superconductivity


K. Amaya, K. Shimizu / Physica C 392–396 (2003) 17–21

3.2. Lithium under pressure

Fig. 4. The resistance drop of superconducting transition in the applied magnetic field.

For long time, mono-valent metals like alkali metals are believed not to show superconductivity even at very low temperature. But, since Cs metal is found to show superconductivity under pressure, there arises possibility of pressure induced superconductivity also in other alkali metals. In the case of lithium, Ashcroft et al. [8] predicted the new structural transition of Li towards the formation of pairing of Li atoms under higher pressure. In fact, experimental studies show the increase of electrical resistance by increasing applied pressure, suggesting the tendency towards insulator in consistent with the theoretical prediction. Quite recently, Christensen et al. [9] calculated the pressure dependence of TC of lithium and give considerable high TC values reaching up to 80 K. We started search for superconductivity of lithium by the electrical resistance measurement under pressure. There are difficulties on confinement of Li sample in our DAC because of the strong chemical reaction with surrounding material. However, we made a sample hole in a diamond anvil and confined there lithium sample at low temperature. This procedure worked well and enabled us to perform measurements up to 50 GPa. We could observe the resistance drop showing the appearance of superconductive state of lithium under pressure [10]. As shown in Fig. 6, super-

Fig. 5. T –P phase diagram for iron.

by detection of the diamagnetic signal due to Meissner effect at the corresponding pressure and temperature to the resistance measurements [6]. Recently, a theoretical calculation was performed by Jarlborg [7], in which the spin fluctuation coupling, especially of ferromagnetic nature, is estimated to be sufficiently large to explain superconductivity of hcp iron. Now the superconductivity was observed in the case of typical ferromagnet of iron in its nonmagnetic state and we can expect analogously the similar behavior in the case of ferromagnetic cobalt, nickel and antiferromagnetic manganese and chromium as well.

Fig. 6. TC –P phase diagram for Li.

K. Amaya, K. Shimizu / Physica C 392–396 (2003) 17–21

conductivity is observed at pressures above 25 GPa and the TC reached 20 K at about 50 GPa. The observed TC value is the highest in elements. With increasing pressure, suppression of the spin fluctuation and increase of the electron–phonon interaction are expected, which causes appearance of superconducting of Li at high pressure in spite of the decrease of carrier density there. We are now studying further on the pressure dependence of TC of Li and also expecting the possibility in the case of other alkali metals. 4. Concluding remarks We have reviewed recent developments of complex extreme condition of low temperature and high pressure and also recent results of pressure induced superconductivity observed in our laboratory. Following the case of iron, new type of superconductivity may be discovered in varieties of magnetic metals. Now we reached lithium, which is the light element close to hydrogen and showing the highest TC among elements. Indeed, our final target is production of metallic hydrogen and detection of its possible high TC superconductive


state which are expected to be realized under ultra high but almost attainable pressure of 400 GPa.

References [1] K. Shimizu, M.I. Eremets, T.C. Kobayashi, K. Amaya, Proceedings of AIRAPT-17, Hawaii, 1999, p. 77. [2] M. Ishizuka, K. Amaya, Rev. Sci. Instrum. 66 (1995) 3307. [3] M. Ishizuka, M. Iketani, S. Endo, Proceedings of AIRAPT-17, Hawaii, 1999, p. 709. [4] G. Cort, R.D. Taylor, J. Willis, J. Appl. Phys. 53 (1982) 2064; R.D. Taylor, M. Pasternak, R. Jeanloz, J. Appl. Phys. 69 (1991) 6126; S. Nasu, Private communication. [5] E.P. Wohlfarth, Phys. Lett. A 75 (1979) 141; A.J. Freeman, A. Continenza, S. Massida, J.C. Grossman, Physica C 166 (1990) 317; N. Suzuki, T. Souraku, I. Hamad, Proceedings of EHPRGÕ39, Santander, 2001. [6] K. Shimizu, T. Kimura, S. Furomoto, K. Takeda, K. Kontani, Y. Onuki, K. Amaya, Nature 412 (2001) 316. [7] T. Jarlborg, Phys. Lett. A 300 (2002) 518. [8] J.B. Neaton, N.W. Ashcroft, Nature 400 (1999) 141. [9] N.E. Christensen, D.L. Novikov, Phys. Rev. Lett. 86 (2001) 1861. [10] K. Shimizu, H. Ishikawa, D. Takao, T. Yagi, K. Amaya, Nature 419 (2002) 597.