Superconductivity from magnetic elements under high pressure

Superconductivity from magnetic elements under high pressure

ARTICLE IN PRESS Physica B 378–380 (2006) 632–635 Superconductivity from magnetic elements under high pressure ¯ nukid...

118KB Sizes 2 Downloads 43 Views


Physica B 378–380 (2006) 632–635

Superconductivity from magnetic elements under high pressure ¯ nukid Katsuya Shimizua,, Kiichi Amayab, Naoshi Suzukic, Yoshichika O a

KYOKUGEN, Research Center for Materials Science at Extreme Conditions, Osaka University, Osaka 560-8531, Japan b Toyota Physical and Chemical Research Institute, Aichi 480-1192, Japan c Graduate School of Engineering Science, Osaka University, Osaka 560-8531, Japan d Graduate School of Science, Osaka University, Osaka 560-0043, Japan

Abstract Can we expect the appearance of superconductivity from magnetic elements? In general, superconductivity occurs in nonmagnetic metal at low temperature and magnetic impurities destroy superconductivity; magnetism and superconductivity are as incompatible as oil and water. Here, we present our experimental example of superconducting elements, iron and oxygen. They are magnetic at ambient pressure, however, they become nonmagnetic under high pressure, then superconductor at low temperature. What is the driving force of the superconductivity? Our understanding in the early stages was a simple scenario that the superconductive state was obtained as a consequence of an emergence of the nonmagnetic states. In both cases, we may consider another scenario for the appearance of superconductivity; the magnetic fluctuation mechanism in the same way as unconventional superconductors. r 2006 Elsevier B.V. All rights reserved. PACS: 74.62.Fj; 74.25.Ha Keywords: Superconductivity; High pressure; Magnetism; Iron; Oxygen

1. Introduction

2. Iron

We have investigated high-pressure electrical properties of simple elemental materials which mainly focus on their superconductivity for the past several years [1]. Simple systems formed by a small number of elements were pressurized up to mega-bar pressure (1,000,000 atmospheric pressure  100 GPa) using diamond-anvil cell (DAC) which could be easily assembled on a powerful 3 He/4He dilution refrigerator. One of our aggressive challenges in the investigation is aiming to find a universality of superconductivity, in other words, demonstrating that every element shows superconductivity in a certain pressure and temperature condition. According to the challenging scenario above, it can be expected that not only nonmetal elements but ferromagnetic metals will be a superconductor at the certain P–T condition.

Iron is the most typical example for magnetic elements among various elemental metals that does not show superconductivity at ambient pressure. It is well known that there exists a pressure-induced crystal phase transition under pressure around 14 GPa from the ferromagnetic BCC phase (a-Fe) to the nonmagnetic HCP phase (e-Fe). Therefore, we could expect the superconducting transition at a certain low temperature and high pressure roughly above 14 GPa, where the iron is nonmagnetic [2]. The nonmagnetism of e-Fe was confirmed by Mo¨ssbauer experiments [3,4] and the superconductivity of e-Fe was predicted by theory to be around 0.25 K [5]. We have searched superconductivity in the high-pressure nonmagnetic phase of iron and could find the superconductivity experimentally [6]. The superconductivity exists at pressure range only between 15 and 30 GPa with the maximum T c of 2 K. The structural transition was known to be sluggish, and it was considered that remaining small mount of a-Fe may suppress the appropriate superconductivity. The superconductivity could hardly be

Corresponding author. Tel.: +81 66850 6675; fax: +81 66850 6662.

E-mail address: [email protected] (K. Shimizu). 0921-4526/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.physb.2006.01.182

ARTICLE IN PRESS K. Shimizu et al. / Physica B 378–380 (2006) 632–635

detected with sample of low residual resistance ratio (RRRo10) or low hydrostatic condition. It seemed that a complete transition is needed to detect superconducting transition. The low quality of sample and pressure obstruct the complete formation of nonmagnetic e-Fe. Only one sample with no pressure medium showed very week evidence of superconductivity while reducing the pressure from 90 GPa. Then, we had concluded, at that time, that the superconductivity occurred with the appearance of the nonmagnetic state by the structural phase transition. 2.1. Unconventional behavior Soon after the discovery of superconductivity of iron, a possibility of the unconventionality was pointed out. First, the superconductivity of iron appears close to the magnetic phase which has a similarity to that of heavy-fermion superconductors, although the phase transition is a firstorder structural transition. A first principle calculation [7] claims that the superconductivity is not a simple nonmagnetic-metal superconductivity from electron–phonon interaction but an unconventional one such as a spin fluctuation-mediated superconductivity. Our obtained superconducting region of iron in T–P phase diagram showed a dome shape as shown in Fig. 1. The shape of the dome is independent to the experimental sequence. A usual pressure dependence of T c in conventional superconductor is mostly monotonic. This suggests an existence of some characteristic condition for superconductivity at the center of the dome. Recently, we have also calculated T c of nonmagnetic e-Fe and found that T c decreases monotonically with increasing pressure, which contradicts the experimental results. This may suggest that the origin of the superconductivity is by other mechanisms than the phonon-mediated one, for example, the pairing mechanism to magnetic fluctuations.


Other experimental investigations [8,9] on the superconductivity of iron were performed. It was found that the resistivity for the normal state of e-Fe shows T5=3 dependence which was predicted by the nearly ferromagnetic Fermi-liquid model which supports the scenario of the first principle calculation [7]. 2.2. Structural and electrical phase boundary Although many structural studies have been done for iron at temperature above room temperature, we could find no experimental determination for the phase boundary between a and e at low temperature. We performed X-ray diffraction measurements under high pressure and the resistivity was measured at the same time. The sample from the same lot with the same pressure condition was pressurized. At room temperature the diffraction peaks from e phase appeared at around 15 GPa with preservation of a phase, then a collapsed until 18 GPa. In the course of pressure decrease, e phase remained until 9 GPa and a phase appeared at the same pressure. Pressure width of coexistence of a and e phases was narrower than previous report [4] and it is considered that the superconductivity occurs with the appearance of e phase. However, as strictly considered, the structure of superconductive iron is not clear because the measurement at low pressure has not been done yet. The transition can be also detected by electrical resistance changes. For determining the boundary at low temperature, the resistance is measured as a function of pressure at fixed temperature of 10 and 290 K with the same arrangement as that of the superconductivity of iron was observed. Resistance curves show the broad step at the structural boundary both in the course of increasing and decreasing pressures. At 290 K, the coexistence of a and e phase was 13–17 and 5–10 GPa in the course of increasing and decreasing pressures, respectively, which is in good agreement with the structural results. In the course of increasing pressure at 10 K, a clear step in resistance was not observed but the coexistence was 5–10 GPa in the course of decreasing pressure. It seems that the a phase collapses at lower pressure than the region of superconductivity and the structural boundary does not lead to the dome. Recently, X-ray magnetic circular dichroism of iron under high pressure was performed at room temperature and it was found that the magnetic moment was suddenly reduced to zero [10]. These results support that iron is a conventional superconductor. The origin of the pairing mechanism is still unclear and will be solved by a further investigation using a neutron scattering experiments under pressure and low temperature. 3. Oxygen

Fig. 1. T c as a function of pressure. The numbers indicate the experimental order. Open circle and square correspond to different runs.

Solid oxygen is a unique among diatomic molecular crystal which has antiferromagnetic properties. The mag-


K. Shimizu et al. / Physica B 378–380 (2006) 632–635

netism considered to be an important role of the crystal variety of this molecule. At ambient pressure, oxygen solidified into three different crystal structures of a, b and g [11]. Under very high pressure, optical measurements showed that reflectivity of solid oxygen starts to increase at around 90 GPa [12] and the metallization was confirmed by temperature dependence of the resistivity [13]. The structural change keeping diatomic molecular bonds was reported [14] at pressure of the metallization. The new structural phase was named as z-phase. From above experimental results, oxygen undergoes metallization at molecular phase, and then a question arises on its magnetism in the metallic oxygen. It was found that solid oxygen becomes superconductor above 98 GPa [15] with T c of 0.6 K as in the case of pressure-induced metallic crystals of such as iodine [16] and bromine [17], which are diatomic molecular crystals at ambient pressure. It was considered that the superconductivity occurs subsequent to the disappearance of the magnetism in z-O2 . However, there is evidence of remaining of the magnetic signature. When we compare the observed T c of 0.6 K with other VI group elements, it is unexpectedly low, where lower-Z element has a tendency of higher T c in this group as shown in Fig. 2. For example, a high T c exceeding 17 K was observed in sulfur [18,19]. The large difference in the value of T c may be due to residual magnetism of the metallic oxygen which keeps the molecular form, where other VI elements dissociate to monatomic form at superconducting phase. Taking into account the molecularity of oxygen, the magnetism at e phase might remain and suppress the proper superconductivity of z phase. No experiment has ever been done in such high pressure and low temperature, however, there is a theoretical report using a first principle calculation [20]. According to the report, at around 100 GPa the pressure-induced metallization occurs whereas the molecular dissociation does not occur and the spin polarization still remains. The most exciting scenario is, in analogy, with some heavy fermion

Fig. 2. Pressure dependence of T c for VI elements; S [18,19], Te [21], Se [22] and O [15].

systems, where the small residual magnetization in z phase itself mediates the superconductivity. On the other hand, magnetic neutron diffraction experiments were performed for solid oxygen. It was reported that d phase which exists between 6 and 8 GPa at low temperature is found to be also antiferromagnet and the magnetic order collapses at 8 GPa [23]. At pressure much lower than z phase, oxygen molecule already loses its magnetism. Magnetism of metallic z phase is still unknown, however, the magnetism of lower phases may keep out of the superconductivity. In any case, from analogy with VI elements, a possible high T c of metallic oxygen can be expected after monatomic phase, however, pressure dependence of T c was negligible in the pressure range between 98 and 130 GPa. Possible high-T c superconductivity in the monatomic metallic oxygen is left for future study.

4. Other magnetic elements We have also studied other antiferromagnetic elements such as a-Mn and Cr over wide range of temperatures. T N was defined at the temperature with the anomaly in the R–T curves. In both cases, their antiferromagnetic ordering temperature, T N , was suppressed by applying pressure as reported by several authors (Fig. 3). By simple extrapolation of the guide lines for eyes, T N of a-Mn and Cr will be zero at roughly 2.5 and 15 GPa, respectively. From analogy with iron, superconductivity was expected for both elements at the pressure range, however, we have observed no evidence of superconductivity down to 50 mK. We believe that further purification of sample and improvement of pressure quality will drive the superconductivity of them.

Fig. 3. Pressure dependence of T N for a-Mn and Cr obtained by electrical resistance measurements except solid circles from thermal expansion measurements. Ref. [24] Dashed lines are guide for eyes.

ARTICLE IN PRESS K. Shimizu et al. / Physica B 378–380 (2006) 632–635

5. Summary In both case of iron and oxygen, there are two different scenarios for occurrence of superconductivity, however, the pairing mechanism is still an open question. Even if it is a phonon or a spin fluctuation mechanism or others, superconductivity in elemental systems such as simple elements will demonstrate an important key for solving the mechanism of ‘‘superconductivity’’. In future, other new superconductor from magnetic elements and compounds will appear and they might cast a new superconductor. Acknowledgements This work was supported by the grant-in-aid for the 21st Century COE Research Program and JSPS KAKENHI (Nos. 15204032 and 15GS0123). References [1] K. Shimizu, K. Amaya, N. Suzuki, J. Phys. Soc. Japan 74 (2005) 13454. [2] R.D. Taylor, G. Cort, J.O. Willis, J. Appl. Phys. 53 (1982) 8199. [3] G. Cort, R.D. Taylor, J. Willis, J. Appl. Phys. 53 (1982) 2064. [4] R.D. Taylor, M. Pasternak, R. Jeanloz, J. Appl. Phys. 69 (1991) 6126. [5] E.P. Wohlfarth, Phys. Lett. A 75 (1979) 141. [6] K. Shimizu, T. Kimura, S. Furomoto, K. Takeda, K. Kontani, ¯ nuki, K. Amaya, Nature 412 (2001) 316. Y. O [7] T. Jarlborg, Phys. Lett. A 300 (2002) 518.


¯ nuki, Phys. Lett. [8] D. Jaccard, A.T. Holmes, G. Behr, Y. Inada, Y. O A 299 (2002) 282. ¯ nuki, J. Phys.: [9] A.T. Holmes, D. Jaccard, G. Behr, Y. Inada, Y. O Condens. Matter 16 (2004) S1121. [10] F. Baudelet, S. Pascarelli, O. Mathon, J.P. Itie, A. Polian, M. d’Astuto, J.C. Chervin, J. Phys.: Condens. Matter 17 (2005) S957. [11] Yu.A. Freiman, Sov. J. Low Temp. Phys. 16 (1990) 559. [12] S. Desgreniers, Y.K. Vohra, A. Ruoff, J. Phys. Chem. 94 (1990) 1117. [13] K. Shimizu, K. Suhara, K. Amaya, S. Endo, J. Phys. Soc. Japan 65 (1996) 1527. [14] Y. Akahama, H. Kawamura, D. H usermann, M. Hanfland, O. Shimomura, Phys. Rev. Lett. 74 (1995) 4690. [15] K. Shimizu, K. Suhara, M. Ikumo, M.I. Eremts, K. Amaya, Nature (London) 393 (1998) 767. [16] K. Shimizu, N. Tamitani, N. Takeshita, M. Ishizuka, K. Amaya, S. Endo, J. Phys. Soc. Japan 61 (1992) 3853. [17] K. Shimizu, K. Amaya, S. Endo, Proceedings of the International Conference on High Pressure Science, Warsaw, 1995. [18] S. Kometani, K. Shimizu, M.I. Eremets, M. Kobayashi, K. Amaya, J. Phys. Soc. Japan 66 (1997) 2564. [19] V.V. Struzhkin, R.J. Hemley, H.K. Mao, Y.A. Timofeev, Nature 390 (1997) 382. [20] K. Kususe, Y. Hori, S. Suzuki, K. Nakao, J. Phys. Soc. Japan 68 (1999) 2692. [21] Y. Akahama, M. Kobayashi, H. Kawamura, Solid State Commun. 84 (1992) 803. [22] E. Gregoryanz, V.V. Struzhkin, R.J. Hemley, M.I. Eremets, H.K. Mao, Y.A. Timofeev, Phys. Rev. B 65 (2002) 064504. [23] I.N. Goncharenko, J. Phys.: Condens. Matter 17 (2005) S947. [24] T. Kagayama, K. Honda, G. Oomi, J. Magn. Soc. Japan 23 (1999) 406.