Effect of negative pressure on superconducting transition temperature of MgB2
Chengguo Zhang,X. Zhang
DOI: https://doi.org/10.1016/j.commatsci.2010.11.007
IF: 3.572
2011-01-01
Computational Materials Science
Abstract:Atomistic simulation has been performed with interatomic potentials to investigate the effect of negative hydrostatic pressure on the superconducting transition temperature ( T c ) of MgB 2 . The calculation reveals that T c can be greatly enhanced by applying negative pressure and it can reach up to 52.2 K at −13 GPa, about 13 K higher than that at ambient pressure. The mechanism for the enhancement of T c is attributed to that negative pressure reduces high-frequency phonon vibrations of B and thus dramatically enhances the electron–phonon coupling. Our result seems to open up a possible way for the enhancement of T c in MgB 2 . Keywords Superconducting transition temperature MgB 2 Negative pressure 1 Introduction The discovery of non-copper-oxide MgB 2 , which has relatively high T c (∼39 K) and simple hexagonal structure, has aroused enormous interest [1] . Researchers have made great efforts to try to enhance the T c further by employing various methods such as isotopic substitution, neutron irradiation, and chemical doping. However, Mg isotopic substitution had almost no effect on the T c [2] . Neutron irradiation resulted in a significant reduction of T c due to the radiation-induced disorder, although it could be healed after post high-temperature annealing [3] . For chemical doping in MgB 2 , dopants like Al, Fe and Cu were tried to substitute at the Mg site [4,5] , and for B site C was the most tried dopant [6] . Nevertheless, the results found that T c was suppressed for most of the dopants due to the formation of fine secondary phase, disorders and crystal lattice distortions. High-pressure experiments on MgB 2 have been carried out with pressure up to 40 GPa in anticipation of T c enhancement. On the contrary, the results revealed that T c was depressed at positive pressure [7–9] . For even higher positive pressure, density-functional-based calculation found that the superconductivity in MgB 2 vanished by 100 GPa, and then reappeared at a pressure of 137 GPa with T c = 2 K [10] . However, the recent first-principles investigation demonstrated that this was not the case, instead, T c reduced monotonically and finally vanished at 200 GPa [11] . The recent report by Ma et al. predicted a KHg 2 -type polymorph of MgB 2 above ultrahigh pressure of 190 GPa through ab initio simulation, and this high-pressure phase was absent of superconductivity [12] . Up to date, most of the attempts to enhance T c of MgB 2 give disappointing results, although there are some exceptional cases such as in strained MgB 2 films and 10 B isotope substituted samples. MgB 2 thin film grown epitaxially on (0 0 0 1) SiC substrate [13] or boron crystal [14] exhibited an enhanced T c above 41 K and the T c enhancement was attributed to tensile strain. In the case of boron isotopic substitution, T c was increased by ∼1 K for Mg 10 B 2 compared to Mg 11 B 2 [2] . Although in these samples the enhancement of T c has been observed, the extent is quite small and only about 1∼2.5 K. Negative pressure is allowed for condensed matters and can be a few hundred MPa for liquids and a few GPa for solids [15] . Presently, modern experimental methods can generate negative pressure with a variety of techniques such as Berthelot-method and shock waves [16,17] , which make it possible to investigate the substance properties at negative pressures. Researchers have studied the effect of negative pressure on structure and properties of materials such as Si [18] , TiO 2 [19,20] , Sb 2 Te 3 [21] , ice [22] , and even the potato starch (biological molecules) [23] theoretically and experimentally, and found it could result in phase transition or lead to remarkable change of properties. Some researchers have also speculated that T c may be changed by negative pressure in MgB 2 [3,24,25] . Islam et al. found negative pressure could influence the band structure and Fermi level of MgB 2 by first-principles density functional calculations [26] . In this work we investigate the effect of negative hydrostatic pressure on structure and superconductivity of MgB 2 in order to search for higher T c . It is found that T c can be greatly enhanced up to 52.2 K at negative pressure of −13 GPa. The mechanism of this effect is also discussed briefly. 2 Simulation method Free energy minimization methods can identify the crystal structure of lowest energy of a material at a given temperature and pressure by adjusting the cell volume and atomic positions until the net pressure or stress is zero. The calculation is based on the widely used shell model generalization of the Born model of a solid. Details of these techniques are available in Ref. [27] . With this model, the lattice energy E can be expressed as (1) E = 1 2 ∑ i , j q i q j r ij + V ( r ij ) , where the first item is Coulombic energy introduced by long range interactions of effective charges, the second item denotes the short range interactions. Short range interaction used in this work is represented by a Buckingham potential (2) V ( r ) = A exp ( - r / ρ ) - Cr - 6 , where A , ρ , and C are fitting parameters. A , ρ is related to the hardness and size of ions, while C represents Van der waals’ forces between ions. To describe the polarization of ions, shell model [28] is employed embodied in the GULP code [29] , in which the interaction between the core and shell of any ion is treated as harmonic with a spring constant k and the interaction is represented by (3) E v ( d i ) = 1 2 kd i 2 , where d i is the relative displacement of core and shell of ion i . The polarization of a massless shell with Y charge and a core with X charge ( X + Y is the charge of the ion) can be calculated as (4) α = Y 2 k , where Y is related to dielectric constant, and k is the force constant between core and shell, related to the phonon frequency. Parameters Y and k are fitting parameters. This method has been successfully used to the simulation of colossal magnetoresistance material [30,31] , A 4 B 3 O 12 δ -phase [32] , thermal-transport properties of SrTiO 3 [33] , surface behaviour of dolomite CaMg(CO 3 ) 2 [34] , pressure effect on the properties in 410/[0 0 1] tilt grain boundary of MgO [35] , YBa 2 Cu 3 O 8 [36] and MgB 2 [37] . It should be stressed that the reliability of the simulation strongly depends on the validity of the potential model used, and the latter is assessed primarily by its ability to reproduce experimental properties. All the potential parameters of MgB 2 used in this calculation, were developed in our group [37] . The difference in lattice constants and bond lengths of MgB 2 between calculated and experimental data is less than 0.0005 Å. This potential model has been used successfully for simulating the effect of positive pressure on MgB 2 up to 240 GPa. We also compared our structural data of MgB 2 under negative pressure with the available structural results from first-principles density functional calculations [26] and found that they are in good agreement ( Fig. 1 ). This good agreement suggests that our potential can be applied to negative pressure. 3 Results and discussion The pressure effect on lattice parameters of MgB 2 over the range from 20 GPa to −13.4 GPa was shown in Fig. 1 . The lattice parameters a and c increase monotonically over the studied pressure range, and c changes more rapidly than lattice parameter a , which means that the interplane Mg–B bond is weaker than the intraplane Mg–Mg and B–B bonds. Accordingly, the c / a ratio increases almost linearly to −13.4 GPa. The calculation also reveals that intraplane bond angles (B–B–B and Mg–Mg–Mg) remain 120° in the studied pressure range (not shown here). Our calculated results are in excellent agreement with the experimental data at positive pressures [8,38,39] and the calculated data at negative pressures by first-principles density functional calculations [26] . This lends a support for the validity of the potential we used under both positive pressure and negative pressure. The structural stability of MgB 2 under pressure has also been examined through lattice dynamics calculation. Fig. 2 illustrates the pressure dependence of the calculated phonon frequencies at zone-center Γ -point. There are six optical bands in MgB 2 and they can be classified as four distinct modes: A 2u , B 1g , E 1u , E 2g . The A 2u and B 1g singly degenerate modes involve vibrations along c axis; for A 2u mode Mg and B atoms both move along c , while for B 1g the B atoms move along c with Mg atoms stationary. The E 1u and E 2g doubly degenerate modes involve only in-plane motions. For E 1u mode Mg and B atoms all vibrate along x or y axis. For E 2g mode, B atoms vibrate along x or y axis with Mg atoms stationary [40] . Among the four modes, E 2g seems to be much more active. Under positive pressure, E 2g grows stiff notably, while the other three maintain rather stable. Upon applying negative pressure, the out-plane A 2u mode remains extremely stable while B 1g mode softens slightly. However, the in-plane E 2g and E 1u modes behave quite differently. The frequency of E 2g decreases noticeably to −13.4 GPa. For E 1u mode, one branch sustains rather stable, whereas the other branch softens moderately to −13.4 GPa and becomes imaginary at around −13 GPa, as the arrow pointed ( Fig. 2 ). The emergence of imaginary frequency of E 1u mode indicates that MgB 2 structure becomes unstable above this pressure, which corresponds well with the abnormity of structural changes for negative pressure being larger than −13 GPa (not shown here). The superconducting transition temperature of MgB 2 is a result of both the raising effect of anisotropy in electron–phonon couplings and the lowering effect of anharmonicity in relevant phonon modes [41] . If we want to study the superconducting transition temperature, fully anisotropic Eliashberg equation is more suitable for analyzing the pressure effect on T c . Nevertheless, there is as yet quite few reports on such a solution due to its complexity, although some preliminary results have been achieved by Choi et al. [42] . Under the circumstances, the McMillan formula [43] , although which was deduced for the isotropic structures, has been employed to calculate T c of MgB 2 in many cases, such as bulk [44] , under pressure [45] , doped by carbon at B site [6] , doped by aluminum at Mg site [4,6] , or the strained MgB 2 [13] , and it gives the T c reasonably well. Therefore, in this work we use the McMillan expression to analyze the pressure effect on T c . McMillan expression takes the form as (5) T c = 〈 ω 〉 1.2 exp - λ + 1 × 1.04 λ - μ ∗ ( 1 + λ × 0.62 ) which connects the value of T c with the electron–phonon coupling parameter ( λ ), the logarithmically averaged phonon frequency 〈 ω 〉, and Coulomb repulsion μ * . Following the analysis of Chen et al. [45] , we set 〈 ω 〉 = 〈 ω 0 〉( V / V 0 ) − γ G , λ = λ 0 ( V / V 0 ) φ and μ * = μ 0 ∗ ( V / V 0 ) Φ , where φ = − γ N − 0.67 + 2 γ G and Φ = −0.1 γ G − 0.09 γ N + 6.7 × 10 −3 . The parameters 〈 ω 0 〉 = 670 K, T c = 39.25 K and λ 0 = 0.9 were obtained from inelastic neutron studies [46] . The value μ 0 ∗ = 0.1 obtained from theoretical calculation [47] was the most commonly used. The variations of the parameters 〈 ω 〉 and λ with volume are decided by two parameters: γ N and γ G . The former is defined as γ N = ∂ ln N ( E f )/ ∂ ln V , which represents an effect of the density of electronic states at the Fermi level, and the latter is defined as γ G = − ∂ ln ( ω 2 ) 1/2 / ∂ ln V , which represents an effect of the lattice stiffening. Chen et al. suggested the values of γ N as 0.46 and γ G as 2.30 [45] . Shao et al. have studied the variation of lattice stiffening with pressure and determined γ G as 2.55 [37] , which is close to the calculated value of 2.36 [48] and experimental value 2.90 ± 0.3 [38] . Thus, in this work, using γ N = 0.46, we adopt γ G = 2.30 and 2.55 to calculate T c of MgB 2 under different pressures varying from 20 GPa to −13 GPa. Fig. 3 shows the variation of T c with pressure. Our calculated T c is in good agreement with the available measured data at positive pressure [7,48,49] , indicating that our adopted γ G is reasonable and our model can describe the behaviour of T c under positive pressure. So we speculate naturally that our model can be also applied to study T c of MgB 2 under negative pressure. In contrast to the depression of T c in MgB 2 at positive pressure, T c can be greatly enhanced by applying negative pressure ( Fig. 3 ). At −13 GPa, T c can reach to its maximum value of 52.2 K (calculated with γ G = 2.55) and 50.6 K (calculated with γ G = 2.30) respectively, which increases by 29.1∼33.2% in comparison with the T c at ambient pressure. Such a remarkable enhancement is close to the effect of positive pressure on T c of the HgBa 2 Ca 2 Cu 3 O 8+ δ superconductor [50] . We also obtain dT c / dP ≈ −1.0 K/GPa under both negative and positive hydrostatic pressure, which is in quantitative agreement with the suggested intrinsic value of −1.11 K/GPa measured under hydrostatic pressure [9] . MgB 2 is an electron–phonon mediated superconductor and electron–phonon interactions play a significant role in its superconductivity. To obtain further insight into the effect of negative pressure on superconductivity of MgB 2 , we calculated the electron–phonon coupling with γ G = 2.30 and 2.55, respectively. Pressure dependence of λ is illustrated in Fig. 4 . In contrast to the decline of electron–phonon interactions at positive pressure, λ becomes much stronger with increasing negative pressure and reaches 1.28 at −13 GPa, which is about twice as large as that at positive pressure of 20 GPa, but it is still in the valid range for the McMillan formula ( λ ⩽ 1.5) [43] . As the weakening of electron–phonon coupling was responsible for the suppression of T c under positive pressure [10,45,46] , it is reasonable to speculate the strengthening of λ plays a significant role in the enhancement of T c under negative pressure. We find that the λ calculated with γ G = 2.55 agrees much better with the reported data by Chen et al. [45] than that calculated using γ G = 2.30 ( Fig. 4 ), indicating that the value of γ G = 2.55 is more reasonable than 2.30 in our work. Further calculation of the fine electron–phonon interaction by first principles calculations is under work. To clarify the role of phonon vibration in the T c enhancement under negative pressure, we take into account the pressure dependence of total and partial phonon density of states (DOS) which provides an understanding of the contribution from Mg and B atoms. The calculated phonon DOS at 10 GPa, 0 GPa and −10 GPa are shown in Fig. 5 . In general, the vibration of B phonon contributes much more than Mg to the total phonon DOS, which is consistent with the result obtained from isotopic substitution studies [2] . Fig. 5 shows that at the three different pressures, the vibrational frequency of Mg phonon changes quite a little except that the last peak takes a slightly shift. For B, it seems that the low-frequency phonon (left of the dotted line) are insensitive to pressure, however, the high-frequency phonon shift strikingly in their frequencies with a dramatic decrease at −10 GPa and a remarkable increase at 10 GPa. The overlapped part with B phonon DOS in Fig. 5 indicates that only B atoms vibrate while Mg atoms remain stationary, and this corresponds to the B 1g and E 2g modes. As indicated in Fig. 2 and Ref. [11] , B 1g mode seems quite less sensitive to pressure than E 2g . Besides, the superconductivity of MgB 2 occurred only on the x – y plane. It seems the out-plane B 1g mode makes little contribution to the variation of T c and thus E 2g mode plays crucial roles in the enhancement of T c under negative pressure. This speculation is compatible with the previous experimental and theoretical results that the decrease of T c under positive pressure was ascribed to the increased E 2g phonon frequencies of B [38] and consistent with the result that lowering the E 2g phonon frequency can increase T c [13] . As a consequence, the remarkable decrease of phonon vibration frequencies by applying negative pressure dramatically enhance the electron–phonon coupling, and at last the significant strengthening of electron–phonon coupling plays a dominant role for the enhancement of T c . Our calculation found that the in-plane B–B bond length increased from 1.7393 Å at 20 GPa to 1.7805 Å at 0 GPa and then raised to 1.8263 Å at −13 GPa, with corresponding T c increasing from 19.2 K to 39.25 K then to 52.2 K. This interesting result indicates that the B–B bond has a close correlation with T c . The stretched B–B bond at negative pressure corresponds to a higher T c while the shortened B–B bond leads to a lower T c . This result opens up a possible way to enhance T c of MgB 2 through enlarging the B–B bond. Such an enlargement by applying negative pressure is qualitatively analogous to an effective strain engineering by growing MgB 2 film on appropriate substrates [13,14] or applying tensile stress with a mechanical load on the crystal. Experimentally in tensile strained MgB 2 thin films a slight increase in T c (∼2.5 K) is actually observed with a stretching B–B bond compared to that in the bulk [13] . Seeking for other appropriate means to enlarge the B–B bond enough should also be considered. 4 Conclusion In conclusion, atomistic simulation has been performed to investigate the effect of negative hydrostatic pressure on T c for the MgB 2 . We find that the structure turns unstable when negative pressure becomes larger than −13 GPa. The T c of MgB 2 can be greatly enhanced by applying negative pressure and it can reach up to 52.2 K at −13 GPa, which is about 13 K higher than that at ambient pressure. The mechanism for T c enhancement by negative pressure is discussed using McMillan formula: Application of negative pressure reduces high-frequency phonon vibrations of B and thus dramatically enhances the electron–phonon coupling. This remarkable strengthening of electron–phonon coupling plays a dominant role for the T c enhancement. Our result seems to open up a possible way for the enhancement of T c and indicates there is still plenty of room to raise T c through elongating the in-plane B–B bond and increasing the electron–phonon coupling strength using various techniques. Acknowledgments We are grateful to Dr. F.L. Tang for his helpful discussions. We would like to acknowledge the financial support by The National Science Foundation of China (U0734001 and 50772054) and The Ministry of Science and Technology of China (2008CB617601 and 2009CB929202). References [1] J. Nagamatsu N. Nakagawa T. Muranaka Y. Zenitani J. Akimitsu Nature 410 2001 63 [2] D.G. Hinks H. Claus J.D. Jorgensen Nature 411 2001 457 [3] R.H.T. Wilke S.L. Bud’ko P.C. Canfield J. Farmer S.T. Hannahs Phys. Rev. B 73 2006 134512 [4] G. Profeta A. Continenza S. Massidda Phys. Rev. B 68 2003 144508 [5] C.S. Lue T.H. Su B.X. Xie S.K. Chen J.L. MacManus-Driscoll Y.K. Kuo H.D. Yang Phys. Rev. B 73 2006 214505 [6] W.X. Li Y. Li R.H. Chen R. Zeng S.X. Dou M.Y. Zhu H.M. Jin Phys. Rev. B 77 2008 094517 [7] M. Monteverde M. Nunez-Regueiro N. Rogado K.A. Regan M.A. Hayward T. He S.M. Loureiro R.J. Cava Science 292 2001 75 [8] P. Bordet M. Mezouar M. Núñez-Regueiro M. Monteverde M.D. Núñez-Regueiro N. Rogado K.A. Regan M.A. Hayward T. He S.M. Loureiro R.J. Cava Phys. Rev. B 64 2001 172502 [9] T. Tomita J.J. Hamlin J.S. Schilling D.G. Hinks J.D. Jorgensen Phys. Rev. B 64 2001 092505 [10] P.P. Singh Phys. Rev. Lett. 97 2006 247002 [11] Y.C. Wang J. Lv Y.M. Ma T. Cui G.T. Zou Phys. Rev. B 80 2009 092505 [12] Y.M. Ma Y.C. Wang A.R. Oganov Phys. Rev. B 79 2009 054101 [13] A.V. Pogrebnyakov J.M. Redwing S. Raghavan V. Vaithyanathan D.G. Schlom S.Y. Xu Q. Li D.A. Tenne A. Soukiassian X.X. Xi M.D. Johannes D. Kasinathan W.E. Pickett J.S. Wu J.C.H. Spence Phys. Rev. Lett. 93 2004 147006 [14] N. Hur P.A. Sharma S. Guha M.Z. Cieplak D.J. Werder Y. Horibe C.H. Chen S.W. Cheong Appl. Phys. Lett. 79 2001 4180 [15] A.R. Imre Phys. Stat. Sol. (b) 244 2007 893 [16] Q. Zheng D.J. Durben G.H. Wolf C.A. Angell Science 254 1991 829 [17] (a) F. Albergamo J. Bossy P. Averbuch H. Schober H.R. Glyde Phys. Rev. Lett. 92 2004 235301 (b) T. Antoun L. Seaman D.R. Curran G.I. Kanel Spall Fract. 2003 Springer New York [18] M. Wilson P.F. McMillan Phys. Rev. Lett. 90 2003 135703 [19] B. Montanari N.M. Harrison J. Phys.: Condens. Matter 16 2004 273 [20] Y. Liu L.H. Ni Z.H. Ren G. Xu C.H. Song G.R. Han J. Phys.: Condens. Matter 21 2009 275901 [21] T. Thonhauser Solid State Commun. 129 2004 249 [22] S.J. Henderson R.J. Speedy J. Phys. Chem. 91 1987 3069 [23] N.W.A. Van Uden H. Hubel D.A. Faux A.C. Tanczos B. Howlin D.J. Dunstan J. Phys.: Condens. Matter 15 2003 1577 [24] K. Prassides Y. Iwasa T. Ito D.H. Chi K. Uehara E. Nishibori M. Takata M. Sakata Y. Ohishi O. Shimomura T. Muranaka J. Akimitsu Phys. Rev. B 64 2001 012509 [25] J. Tang L.C. Qin A. Matsushita Y. Takano K. Togano H. Kito H. Ihara Phys. Rev. B 64 2001 132509 [26] A.K.M.A. Islam F.N. Islam Int. J Mod. Phys. B 17 2003 3785 [27] S.C. Parker, G.D. Price, in: C.R.A. Catlow, S.C. Parker, M.P. Allen (Eds.), Computer Modeling of Fluids, Polymers and Solid, NATO ASI Series, 293 (1988) 405. [28] B.G. Dick A.W. Overhauser Phys. Rev. 112 1958 90 [29] J. Gale A.L. Rohl Mol. Simul. 29 2003 291 [30] F.L. Tang X. Zhang Phys. Rev. B 73 2006 144401 [31] F.L. Tang X. Zhang Comput. Mater. Sci. 40 2007 434 [32] A. Chernatynskiy R.W. Grimes M.A. Zurbuchen D.R. Clarke S.R. Phillpot Appl. Phys. Lett. 95 2009 161906 [33] C.R. Stanek C. Jiang B.P. Uberuaga K.E. Sickafus A.R. Cleave R.W. Grimes Phys. Rev. B 80 2009 174101 [34] K.F. Austen K. Wright B. Slater J.D. Gale Phys. Chem. Chem. Phys. 7 2005 4150 [35] D.J. Harris G.W. Watson S.C. Parker Phys. Rev. B 64 2001 134101 [36] X. Zhang C.R.A. Catlow Physica C 193 1992 221 [37] Y. Shao X. Zhang J. Phys.: Condens. Matter 16 2004 1103 [38] A.F. Goncharov V.V. Struzhkin E. Gregoryanz J.Z. Hu R.J. Hemley H.K. Mao G. Lapertot S.L. Bud’ko P.C. Canfield Phys. Rev. B 64 2001 100509 [39] T. Vogt G. Schneider J.A. Hriljac G. Yang J.S. Abell Phys. Rev. B 63 2001 220505 [40] T. Yildirim O. Gülseren J. Phys. Chem. Solids 63 2002 2201 [41] H.J. Choi D. Roundy H. Sun M.L. Cohen S.G. Louie Phys. Rev. B 66 2002 020513 [42] H.J. Choi S.G. Louie M.L. Cohen Phys. Rev. B 79 2009 094518 [43] P.B. Allen R.C. Dynes Phys. Rev. B 12 1975 905 [44] J. Kortus I.I. Mazin K.D. Belashchenko V.P. Antropov L.L. Boyer Phys. Rev. Lett. 86 2001 4656 [45] X.J. Chen H. Zhang H.U. Habermeier Phys. Rev. B 65 2002 144514 [46] R. Osborn E.A. Goremychkin A.I. Kolenikov D.G. Hinks Phys. Rev. Lett. 87 2001 017005 [47] Y. Kong O.V. Dolgov O. Jepsen O.K. Andersen Phys. Rev. B 64 2001 020501 [48] S. Deemyad J.S. Schilling J.D. Jorgensen D.G. Hinks Physica C 361 2001 227 [49] F.S. Razavi S.K. Bose H. Ploczek Physica C 366 2002 73 [50] J.D. Tristan R.J. Wijngaarden H. Wilhelm R. Griessen S.M. Loureiro J.J. Capponi A. Schilling H.R. Ott Phys. Rev. B 54 1996 4265