Supercond. Sci. Technol. 9 (1996) 734–735. Printed in the UK An analytical solution for the extended critical state model V Meerovich†, M Sinder‡ and V Sokolovsky† † Physics Department, Ben-Gurion University of the Negev, POB 653, Beer-Sheva 84105, Israel ‡ Material Engineering Department, Ben-Gurion University of the Negev, POB 653, Beer-Sheva 84105, Israel Received 21 February 1996 Abstract. The equations of the extended critical state model are solved analytically for the√case when an external magnetic field is increased according to the law H = A t . Distributions of magnetic and electric fields in a superconductor are found to depend on the relation between characteristic times of two processes: the increase of applied external magnetic field and the diffusion of magnetic field into a superconductor. If the diffusion time is much smaller than the characteristic time of the magnetic field increase, the solution reduces to that of the classical critical state model. In the opposite case, the solution tends to the solution for normal metals. In this case, the discrepancy between the results of the expended and classical critical state models increases with time. The extended critical state model introduced by Bean [1] has explained some non-linear effects in high-Tc superconductors [2, 3]. Following Bean, we consider the penetration of magnetic field in a semi-infinite superconducting slab whose inward surface normal points in the positive x direction. The magnetic field H is assumed to be applied in the z direction, which is parallel to the slab surface, and the electric field E and currents j are induced along the y axis. For this special case, the equations of the extended critical state model become ∂E/∂x = −µ0 ∂H /∂t (1) ∂H /∂x = −jc − E/ρf (2) E = ρf [j − sign(j )jc ] (3) where jc is the critical current density, ρf is the flux flow resistivity, and µ0 is the magnetic constant. Assuming that the critical current density and resistivity are constants and eliminating E from equations (1)–(3), we obtain a onedimensional diffusion equation for the magnetic field. This equation in a dimensionless form becomes ∂ H̃ /∂ x̃ = ∂ H̃ /∂ t˜ 2 2 (4) where H̃ = H (x, t)/H0 , x̃ = x/xm , t˜ = t/τ , xm = H0 /jc and τ = µ0 H02 /(ρf jc2 ). Here we use an arbitrary normalizing constant H0 . xm is the penetration depth in the Bean critical state model [4] when the applied magnetic field is equal to H0 ; τ is the diffusion relaxation time introduced by Bean [1]. The diffusion equation (4) may be integrated analytically or numerically subject to appropriate boundary conditions. To obtain an analytical solution, we take the boundary condition for the magnetic field on the superconductor c 1996 IOP Publishing Ltd 0953-2048/96/090734+02$19.50 surface, i.e. at x = 0, in the form p H̃ (t˜, 0) = à t˜ (5) √ √ where à = τ/τ1 , τ1 = H0 /A, and τ√ 1 is the time to increase applied magnetic field H = A t up to the value H0 . In other words, we consider the case when the magnetic field applied to a superconductor increases with time according to a power law with the exponent√of onehalf. Introducing Boltzmann’s variable η = x̃/(2 t˜), we can seek a solution in the form p H̃ = t˜G(η) (6) where G(η) is a function of η. The boundary conditions inside the superconducting slab and the initial condition are H̃ (t˜, x̃ = x̃p ) = 0 ∂ H̃ (t˜, x̃)/∂ x̃|x̃=x̃p = −1 (7) H̃ (0, x̃) = 0. The second expression in (7) is needed to determine the penetration depth of magnetic field x̃p . The corresponding solution is p √ H̃ = −x̃ + à t˜{exp(−η2 ) − η π [erf(η) − erf(η0 )]} (8) √ where η0 = x̃p /2 t˜; erf(x) is the error integral; the penetration depth x̃p is determined from the equation p (9) à = (x̃p / t˜) exp(η02 ). Using (2), we find the electric field: √ Ẽ = à π [erf(η0 ) − erf(η)]/2. (10) An analytical solution for the extended critical state model p H̃ = à t˜ − x̃ p Ẽ = Ã2 [1 − x̃/(à t˜)]. (11) (12) This solution is coincident with the solution given by the critical state model: the current density is equal to the critical one. The second member of the Taylor series is proportional to Ã3 for H̃ and for Ẽ it is proportional to Ã4 . In the other limiting case à 1, the penetration depth is estimated as q x̃p ≈ 2 t˜ ln(Ã) and the electric field becomes √ Ẽ = (à π /2) erfc(η). Figure 1. The penetration rate as a function of the parameter Ã. Figure 2. The distribution of the electric field inside a superconductor for different values of Ã. As follows from (10), the electric field on the superconductor surface (η = 0) is time independent. The distributions of magnetic and electric fields are determined only by the parameter Ã. Figures 1 and 2 illustrate the influence of the parameter à on the solutions. Let us consider two limiting cases: (i) à 1 and (ii) à 1. To analyse the first, we will expand the solutions (8)–(10) in Taylor series in the parameter Ã. The first member in the expansion for H̃ contains à with the exponent of power equal to unity while the first member in the expansion √ for Ẽ is proportional to à squared. For x̃ ≤ x̃p = à t˜, the magnetic and electric fields become Note that, in contrast to the previous case, the electric field is proportional to Ã, as for normal metals. The electric field markedly differs from zero in the range where η < 2–3. In deeper points of the superconductor, j ≈ jc and the magnetic field decreases linearly as in the critical state model. In proximity to the surface (η 1), the electric field is close to a constant and the magnetic field is determined by the expression p p H̃ ≈ à t˜(1 − x̃/ t˜). In conclusion, we have shown that the distribution form of the magnetic and electric fields inside a superconductor depends on the relation of the diffusion time τ and the time τ1 , which characterizes an increase of the applied magnetic field. For τ τ1 (à 1), the diffusion process proceeds quickly and the solution differs little from that given by the critical state model. For τ τ1 (à 1), the magnetic field distribution is determined by the diffusion process as in the case of normal metals. In the course of time, the difference between solutions (8), (10) and the critical state model increases (see figure 1). This result differs critically from the results which have been obtained in [5]. Ampere’s law used in the work referred to has a form exactly equivalent to equation (2) but these treatments did not use Faraday’s law (1). References [1] Bean C P 1989 Superconductivity and Applications ed H S Kwok, Y-H Kao and D T Shaw (New York: Plenum) p 767 [2] Jiang H and Bean C P 1994 Appl. Supercond. 2 689 [3] Fisher L M, Il’in N V, Voloshin I F et al 1993 Physica C 206 195 [4] Bean C P 1962 Phys. Rev. Lett. 8 250; 1964 Rev. Mod. Phys. 36 31 [5] Liu Z and Bowden G J 1991 Supercond. Sci. Technol. 4 122 735