Supercond. Sci. Technol. 10 (1997) 523–542. Printed in the UK PII: S0953-2048(97)78196-X TOPICAL REVIEW Characterization of high-temperature superconductors by AC susceptibility measurements Fedor Gömöry† Institute of Electrical Engineering, Slovak Academy of Sciences, Dúbravská 9, 84239 Bratislava, Slovakia Received 16 December 1996, in final form 2 May 1997 Abstract. A magnetic field harmonically varying in time (to probe the sample) and a lock-in technique (to register the sample response sensed by a pick-up coil) are widely used for characterizing superconductors. Measuring the temperature dependence of the complex AC susceptibility is the most common procedure of this type. This paper reviews these techniques, introducing in addition the complex AC susceptibility, the so-called ‘wide-band AC susceptibility’. The latter quantity refers to the magnetic flux and often offers an easier meeting between theory and experiment. Starting from models for linear flux diffusion, reversible screening, volume and surface flux pinning and the intermediate regime in a type II superconductor, the expressions for the complex AC susceptibility in different cases are presented and compared with those derived for the wide-band AC susceptibility. Derivation of the basic physical properties of high-Tc superconducting materials from the AC data (resistivity, critical temperatures and fields, London and Campbell penetration depths, critical current density, granularity and content of superconducting phase, irreversibility line, pinning potential) is then thoroughly discussed. 1. Introduction A magnetic field harmonically varying in time (to probe the sample) and the lock-in technique (to register the sample’s induced magnetic response sensed by a pick-up coil) are often used to study the electromagnetic properties of superconductors. Measurement of the temperature dependence of the complex AC susceptibility is the most common experiment of this type. As we will see, a common theoretical background exists for the experimental techniques that can be treated as members of the same family: complex AC susceptibility, AC losses, inductive measurement of Tc and critical magnetic fields, AC methods to study pinning. This paper is an attempt to systematize the knowledge gathered in the field, with particular attention paid to the modifications introduced after the discovery of high-Tc superconductors (HTSCs). A simple registration of the AC susceptibility is enough to characterize the type I superconductor (Shoenberg 1937, Hein and Falge 1961). The development of the type II superconductors required experimental methods to study the volume pinning force and the surface barriers † Present address: Pirelli Cavi SpA, c. 2714, Viale Sarca 222, 20126 Milano, Italy. c 1997 IOP Publishing Ltd 0953-2048/97/080523+20$19.50 (Campbell 1969, Rollins et al 1974, Griffiths et al 1976). AC susceptibility was used to reveal the filamentary superconductivity (Maxwell and Strongin 1963) and effects of granularity (Ishida and Mazaki 1979) and the importance of correlating the AC susceptibility data with the structure was pointed out (van der Klein 1972). The AC losses are probably one of the most important topics related to the AC susceptibility (Clem 1979a, Campbell 1995, Müller et al 1996). Recently the wide-band AC susceptibility technique was developed (Dubots and Cave 1988, Gömöry 1991b, Gugan 1994, Stoppard and Gugan 1995) that could serve as an integrating concept for all these methods. It will be used throughout the paper, exploiting such favourable features as simple derivation of formulas and easy interpretation of results. The phenomenology of the macroscopic magnetic properties of type II superconductors is nicely mapped in the studies centred around the critical state model and the concept of the critical current density (Campbell and Evetts 1972, Ullmaier 1975). After the discovery of HTSCs, all the previous ‘low-Tc ’ experience was confronted with new problems (Malozemoff 1989), in particular an extremely rich phase diagram of the flux line lattice (Blatter et al 1994). Higher operational temperatures and strong intrinsic anisotropy emphasize the importance of dynamic effects (Brandt 1992a). 523 F Gömöry Physical models used to explain the AC susceptibility are identical with those utilized in DC magnetic experiments (Senoussi 1992, Pérez et al 1996) and comparison of both methods gives valuable information (Frischhertz et al 1995). Other experimental techniques that allow us to study similar properties are the vibrating reed experiments (Köber et al 1991, Parvin et al 1993, D’Anna et al 1994, Ziese et al 1994, Rogacki et al 1996) and the AC transport measurements (Doyle R A et al 1994). The basic arrangement of the experiment (a detailed description can be found e.g. in the papers of Ramakrishnan et al (1985), Rillo et al (1991), Couach and Khoder (1991) and Nikolo and Hermann (1991)) is given in figure 1. The sample in the form of a slab (large and high enough for these dimensions to be considered infinite compared with the sample width) is placed in the magnetic field Bext = Bd +Bac having the static component Bd and the AC component Bac characterized by the frequency f = ω/2π and the amplitude Ba given by Bac = Ba cos(ωt) (1.1) parallel to the surface of the slab. We suppose that the DC and AC fields are parallel in space, although sometimes the perpendicular arrangement of the fields is used (de la Cruz et al 1994, Filippi et al 1994, Waldmann et al 1996). The sample’s magnetic response is sensed by a pick-up coil surrounding the sample. The temperature control is another subsystem of the apparatus, allowing the dependence of the sample’s parameters on temperature to be studied. Various methods exist to analyse the pick-up coil voltage: the whole waveform can be analysed (e.g. Cave et al 1991, Gjomesli and Fossheim 1994b, Kerchner et al 1995), the waveform analyser can give the Fourier spectrum of the signal (e.g. Fabbricatore et al 1993a, 1994, Mazaki et al 1995) or, as the most simple way, only the effective value is detected by an AC voltmeter. In this paper we consider the phase-sensitive detection (PSD) that allows us to measure AC susceptibility and seems to be a good compromise between complexity of the data and simplicity of the experiment. The organization of the paper is as follows: in section 2 the basic terms are introduced. Together with the complex AC susceptibility components, the wide-band AC susceptibilities will be defined. To interpret the experimental data correctly we need to recognize what mechanisms participate in the observed behaviour. I gathered the formulas for both types of AC susceptibilities derived according to the models of linear flux diffusion (including flux flow and thermally assisted flux flow), reversible screening (covering both the Meissner–Ochsenfeld effect and Campbell’s reversible motion of flux lines), critical state, surface barrier pinning and flux creep in section 3. Then, in section 4 the procedures that allow the critical temperature, critical magnetic fields, linear resistivity, critical current density, surface barrier and granularity together with the content of the superconducting fraction, London’s and Campbell’s penetration depths, irreversibility line and so-called pinning potential (the height of barrier for thermal activation), to be determined are recommended. 524 2. Definition of basic terms The magnetic flux crossing the sample area can be expressed with the help of the mean value B of the flux density through the sample cross-section A: Z φm = AB = B · dA. (2.1) A In an AC experiment B is a function of time and controls the voltage in one turn of the pick-up coil: um (t) = − dB(t) dφm (t) = −A . dt dt (2.2) It is important to bear in mind that the pick-up coil measures an integrated value of the flux density in the sample. Recently experiments have been performed using miniature Hall probes to record the local magnetic response of the sample (Prozorov et al 1995, van der Beek et al 1996, Morozov et al 1996), but these are outside the scope of the present review. Using equation (2.1) one can define the sample magnetization M(t) = B(t) − Bext (t) = φm (t) − Bext (t). A (2.3) The complex AC susceptibility components are defined as (Maxwell and Strongin 1963) 1 χ = π Ba 0 χ 00 = 1 π Ba Z 2π M(ωt) cos(ωt) d(ωt) 0 Z 2π M(ωt) sin(ωt) d(ωt). 0 The physical meaning of χ 0 and χ 00 is the following. The energy converted into heat during one cycle of the AC field is (Clem 1988) Wq = −2π χ 00 Ba2 . 2µ0 (2.4) This expression explains why the lock-in can be used to determine AC losses (Sekula 1971, Bozec et al 1991, LeBlanc and LeBlanc 1992, Jiang and Bean 1994). Because Wq is always negative, χ 00 in a correctly designed experiment must take the positive sign. The time average of the magnetic energy stored in the volume occupied by the sample is (Gömöry 1991a) Wm = χ 0 Ba2 2µ0 (2.5) where the normal-state value was taken as the reference level, i.e. Wm (T > Tc ) = 0. Diamagnetic behaviour leads to reduction of the magnetic field compared with the normal state, reflected in a negative value of Wm . Thus we expect in the case of a superconductor χ 0 ≤ 0. The wide-band susceptibilities are defined as (Gömöry 1991b). Ma Mr χr = (2.6) χa = Ba Ba Characterization by AC susceptibility Figure 1. The set-up for AC susceptibility measurement on superconductors (schematic). where Ma is the sample magnetization at the moment when the external field reaches the maximum: we can call it the ‘amplitude magnetization’. Mr is the magnetization remaining in the sample at zero instantaneous value of the AC field: we can call it the ‘remanent magnetization’ (see figure 2). Therefore in the following χa and χr will be called the remanent and the amplitude AC susceptibility, respectively. The diamagnetic nature of the superconductor leads to a negative magnetization of the sample at the moment when the external field reaches extremal (both negative and positive) values −Ba and Ba , respectively. The absolute value of χa ∈ (−1, 0) is a measure of the external flux density that is not allowed to penetrate the sample. If the flux density inside the sample has some delay with respect to the external field, a non-zero remanent magnetization (i.e. left in the sample at zero Bac ) appears. It is seen as a hysteresis of the AC magnetization loop M(Bac ) and leads to χr > 0. Regarding the instrumentation, complex AC susceptibility requires a lock-in with a harmonic reference waveform or, in the case of an instrument with a rectangular reference waveform, to filter the input voltage. To record the amplitude and the remanent AC susceptibility, we must pass the entire spectrum of the pick-up coil signal to the PSD with the rectangular reference waveform. If the pickup coil signal does not contain higher harmonics, exact equivalence is reached: χ 0 = χa , χ 00 = χr . Both χ 0 and χa reflect the shielding ability, while χ 00 as well as χr are measures of the magnetic irreversibility. The main difference from the practical point of view is in the complexity of calculations: the complex AC susceptibility is related to the whole magnetization cycle, while χa and χr are determined at only two significant instants of the cycle. This sometimes results in a simpler calculation of χa and χr in comparison with χ 0 and χ 00 and consequently easier comparison of the experimental results with theory. 3. Superconductor in a harmonic external field The interaction between superconductor and magnetic field consists of various processes. Those treated here are the expulsion of flux due to Meissner–Ochsenfeld effect, the flux pinning by a surface barrier, the pinning of the flux entering the bulk of superconductor in the form of flux lines, the diffusion of these flux lines across the sample and the reversible motion of pinned flux lines in potential wells. Analysing the physical models of these mechanisms one can see that they can be grouped in the following way (Brandt 1990). The diffusion equation (we denote by x the transverse coordinate) 1 ∂B µ0 ∂B ∂ 2B = = 2 ∂x ρ(B, E) ∂t D(B, E) ∂t (3.1) 525 F Gömöry Figure 2. One cycle of the magnetization of a YBaCuO melt-grown sample. Left: the time dependences of the driving field and the sample magnetization. Right: the magnetization loop. The significant values of magnetization recorded when Bac ± Ba and Bac = 0 will be used in the definition of the wide-band susceptibilities. Note the obvious anharmonicity of M (ωt ). together with the approximate current–voltage characteristics (Vinokur et al 1991, Gilchrist 1994, Gilchrist and van der Beek 1994) j (B, E) = jc (B) E |E| |E| Ec 1/n (3.2) allow us to model the linear diffusion of flux lines (n = 1), critical state (n → ∞) and intermediate regimes (n > 1). The resistivity ρ(B, E) from equation (3.2) defines the relation between the electrical field and the current density: E(x, t) . j (B, E, x, t) = ρ(B, E) δ= 2ρlin µ0 ω 1/2 . (3.3) where 3 is a material constant. This equation describes well the reversible screening either by Meissner currents or by Campbell’s reversible motion of flux lines in potential wells. Finally, the irreversible surface barrier must be treated by a model derived from a simplified form of the magnetization loop. 3.1. Linear diffusion Setting n = 1 in equation (3.2) gives j linearly proportional to E, i.e. E (3.1.1) j (B, E) = ρlin (B) and only one material parameter ρlin (B) = Ec (B)/jc (B) remains. Substituting equation (3.1.1) into equation (3.3) converts equation (3.1) into a linear differential equation, which leads to an AC flux profile with the local AC field amplitude B0 (x) decaying roughly exponentially (Brandt 1991a): (3.1.2) B0 (x) ≈ Ba exp(−x/δ). (3.1.3) Important mechanisms of the flux dynamics belonging to this case are normal-state eddy currents, linear flux flow and thermally assisted flux flow. They differ in the value of the linear resistivity. For eddy currents this is equal to the normal resistivity: ρlin = ρn . For some mechanisms one has to use the London (1961) equation ∇ × (3j) = −B (3.4) 526 The characteristic space scale of the AC field decay is (3.1.4) When the Lorentz force fL = j B exceeds the pinning strength of the material (characterized by the volume pinning force fP ), the flux line lattice (FLL) starts to move as a whole. Kim and Stephen (1969) found this movement to be viscous, with the viscosity coefficient ηF F = BBc2 ρn (3.1.5) where Bc2 is the upper critical magnetic field. The flux lines move with velocity vL proportional to the gradient in flux density, and this movement is accompanied by the electrical field E = B × vL (Josephson 1965). Thus, the resistivity ρF F that appears in equation (3.1.3) is now called the ‘flux flow resistivity’ (Kim and Stephen 1969): ρF F = ρn B/Bc2 . (3.1.6) This becomes field independent when we use a large DC field Bd superimposed on the AC field because then ρF F = ρn Bd /Bc2 . Then, for linear flux flow, ρlin = ρn Bd /Bc2 . (3.1.7) The FLL deformed by local pinning forces tends to reach states with lower energy. This process could by activated by several mechanisms, thermal activation being Characterization by AC susceptibility χ 00 and χr will be reached when δ = 0.887R in the case of a cylinder (radius R) or δ = 0.556R in the case of a slab (width 2R). In section 4 we will see how difficult it is to cope with a combination of several mechanisms appearing simultaneously in the flux dynamics. One possible approach is to work out linearized models that can be superimposed. Brandt (1992b) succeeded in deriving a complex resistivity that combines the effects of Meissner screening, thermally assisted flux flow, flux flow and reversible motion of flux lines (Campbell 1971): ρAC (ω) = iωµ0 λ2L + ρT AF F Figure 3. AC susceptibility in the regime of linear flux diffusion. the most important. At low density gradients (Kes et al 1989), for thermally assisted flux flow, ρlin = ρF F exp(−U/kB T ). (3.1.8) Here U is the typical height of the pinning barrier. The linear diffusion is treated in textbooks of classical electrodynamics (e.g. Smythe 1950, p 390). Because of the linear character of the diffusion, higher harmonics are not generated in the pick-up signal and χ 0 = χa as well as χ 00 = χr . For a slab of width 2R in a parallel magnetic field one finds (Kes et al 1989) χ 0 = χa = δ sinh(2R/δ) + sin(2R/δ) −1 2R cosh(2R/δ) + cos(2R/δ) δ sinh(2R/δ) + sin(2R/δ) . χ = χr = 2R cosh(2R/δ) + cos(2R/δ) 00 1 + iωτ . 1 + iωτ0 (3.1.11) Here λL is the London penetration depth and the time constants are as follows: τ = αL /ηF F is given by the Labusch parameter αL (Labusch 1969, Brandt 1991b, Zeisberger et al 1994) and the flux flow viscosity ηF F , and τ0 = τ exp(−U/kB T ). Linear AC response theory provides elegant and compact solutions (Campbell 1991) and can account for the anisotropy (Wu and Tseng 1996). The linear AC susceptibility could be used to reflect the conductivity of the samples (e.g. Kötzler et al 1994, Ando et al 1994, Li et al 1994), but one should always test whether the conditions of linearity are fulfilled (Mehdaoui et al 1993). However, many practical applications require the determination of the superconducting properties in the conditions of a strongly non-linear regime (Takács and Gömöry 1995). 3.2. Critical state (3.1.9a) The behaviour of a superconducting material able to pin the flux line lattice is well described by the so-called criticalstate model (Bean 1964). The j (B, E) dependence is given by the expression (3.2) with n → ∞. The flux density gradient in the critical state will be (3.1.9b) ∂B = ∓µ0 jc (B). ∂x The expressions for a spherical sample (Khoder and Couach 1991) are of similar form. Another important geometry is the case of a cylinder with radius R in a parallel magnetic field. The corresponding formulas are (Clem et al 1976): 2J1 (kR) 0 −1 (3.1.10a) χ = χa = Re kRJ0 (kR) 2J1 (kR) χ 00 = χa = Im . (3.1.10b) kRJ0 (kR) The penetration depth δ enters the expressions (3.1.9) and (3.1.10) through an auxiliary variable k = (1 + i)/δ, i is the imaginary constant and J0 , J1 are the Bessel functions of the first kind (e.g. Spanier and Oldham 1987, p 509). Re means the real part while Im is the imaginary part. The graphical form of (3.1.9) and (3.1.10) is presented in figure 3. The linear regime is dissipative, which means that a non-zero loss of magnetic energy appears and the peak in (3.2.1) Accordingly, the material is characterized by the critical current density jc . Suppose the dependence of jc on the AC magnetic field can be omitted. This approximation is valid if for example Bd Ba because jc (B) ≈ jc (Bd + Ba ). At the moment when the AC magnetic field (1.1) reaches a maximum value (i.e. at ωt = 0, 2π , 4π, . . . , 2nπ ), the AC flux density profile takes the form B(x, 0) = Ba − µ0 jc x. (3.2.2) Field penetration is stopped in the depth xc = Ba . µ0 jc (3.2.3) Because of the strong non-linearity of the current– voltage relation, higher harmonics will appear in the sample magnetization causing anharmonicity in the pick-up coil voltage (Fabbricatore et al 1994). As a consequence, χr is no longer equal to χ 00 or χa to χ 0 . In the following formulae 527 F Gömöry we use an auxiliary variable y = xc /R = Ba /µ0 jc R. For a slab (width 2R) in a parallel magnetic field (Ji et al 1989, Matsushita and Ni 1989, Goldfarb et al 1991) y/2 − 1 for 0 ≤ y ≤ 1 1 y − 1 cos−1 (1 − 2y) χ0 = π 2 4 4 + −1+ − 2 (y −1)1/2 for 1 ≤ y 3y 3y (3.2.4a) 2y/3π for 0 ≤ y ≤ 1 (3.2.4b) χ 00 = 4 1 6 − 2 for 1 ≤ y 3π y y and the wide-band susceptibilities take a form that can be easily derived from the flux density profiles (Bean 1964): ( y/2 − 1 for 0 ≤ y ≤ 1 (3.2.5a) χa = −1/2y for 1 ≤ y y/4 χr = 1 − y/4 − 1/2y 1/2y for 0 ≤ y ≤ 1 for 1 ≤ y ≤ 2 (3.2.5b) for 2 ≤ y. If the sample has the form of a cylinder with the radius R, then in a similar way as for the slab we find 5y y 1 − −1 for 0 ≤ y ≤ 1 16 2 5y 2 χ0 = −1+y − cos−1 (1−2y) π 16 2 19 5y 1 1/2 +− + + − (y −1) for 1 ≤ y 12 8 y 3y 2 (3.2.6a) 4 y(1 − y/2) for 0 ≤ y ≤ 1 3π (3.2.6b) χ 00 = 1 4 1 1− for 1 ≤ y 3π y 2y ( y − y 2 /3 − 1 for 0 ≤ y ≤ 1 χa = (3.2.7a) −1/3y for 1 ≤ y 2 for 0 ≤ y ≤ 1 y/2 − y /4 2 χr = −1/3y + 1 − y/2 + y /12 for 1 ≤ y ≤ 2 1/3y for 2 ≤ y. (3.2.7b) These dependences are presented in figure 4. It is worth mentioning some important features. (1) When xc R (i.e. y → 0), χa ∼ = χ and ∼ χr = (3π/8)χ 00 for both geometries. (2) The penetration of the AC field is dissipative and the persistent currents cause the appearance of the remanent 00 magnetization. The peak √ in χ as well as in χr is reached at xc = R and xc = 2R in the case of cylinder and slab, respectively. (3) There is a notable difference in the heights of the maxima between χ 00 and χr for both geometries. (4) The substantial difference between χa and χ 0 for xc R indicates that the higher harmonics significantly 528 Figure 4. AC susceptibility in the regime of flux pinning in the sample bulk, described by the critical-state model. contribute to the pick-up coil signal. This can be used to detect the onset of the critical state by the appearance of the third harmonics in the susceptibility signal (Gilchrist and Konczykowski 1990, Yang et al 1994b, van der Beek et al 1995, 1996, Dubois et al 1996). The portion of the χr (xc /R) and χa (xc /R) curves when xc > 2R can be used very simply for the determination of the critical current density. In the case of a slab-like sample (width 2R) we find by combining equation (3.2.3) with equations (3.2.5) that jc = 2χa Ba 2χr Ba =− µ0 R µ0 R for jc < Ba 2µ0 R (3.2.8) and, in a similar way, combining equation (3.2.3) with equations (3.2.7) we find that for a cylindrical sample with radius R jc = 3χa Ba 3χr Ba =− µ0 R µ0 R for jc < Ba . 2µ0 R (3.2.9) These expressions can be used to transform directly the measured χ(T ) curves into the jc (T ) plot. Only wideband susceptibilities provide this opportunity because of the linear form of the relations (3.2.8) and (3.2.9) in the interval xc > 2R. The use of χ 00 (T ) for this purpose (Senoussi 1992) requires the more restrictive condition xc R to be fulfilled. The critical-state model was found to explain the behaviour of many types of HTSC samples (Berling et al 1996b). 3.3. Intermediate regimes This is the case of equation (3.2) when n is greater than 1 but not enough to be considered infinitely large. Therefore it is called the intermediate regime (Civale et al 1991, Karkut et al 1993, Polichetti et al 1994, DiGioacchino et al 1996). Typical examples of this behaviour are the extreme Characterization by AC susceptibility flux flow (Takács and Gömöry 1990) and the flux creep (Anderson and Kim 1964) regimes. The extreme flux flow model was worked out to treat the flux flow at Bd = 0, when because of the flux flow resistivity dependence on the magnetic field (3.1.6) the diffusion equation (3.1) is no longer linear. An approximate solution for the penetration depth was found, i.e. 3πBa ρn 1/2 δEF F = (3.3.1) 8µ0 ωBc2 that nicely demonstrates the characteristic feature of the intermediate regimes: the penetration depth depends on both the AC field amplitude as well as the frequency. Careful experiments we performed (Takács and Gömöry 1995) demonstrated that the regime of this type with n = 2 can be found in melt-grown YBaCuO samples near Tc . In contrast, the flux creep covers a wide range of temperatures and fields in the case of HTSCs. This regime is met when in the sample, driven originally into the critical state, separate pieces of the FLL (flux line segments, bundles of flux lines, etc.) are hopping over pinning barriers to reach metastable states with lower energy, and the movement against the density gradient can be neglected. Symptomatic of thermally activated creep is the appearance of a drift velocity of the flux lines vL that obeys the Arrhenius law: Ueff vL = v0 exp − . (3.3.2) kB T Here, Ueff is the effective height of the barrier that the hopping segment of the FLL must overcome, kB = 1.38 × 10−23 J K−1 is the Boltzmann constant, and v0 is the line velocity at Ueff → 0 that can be estimated considering the transition from flux creep to flux flow (van der Beek et al 1992). The flux creep is often taken as a small perturbation in the pinning mechanism (e.g. Fedorov and Stepanov 1996). Then, in equation (3.2), jc is the unperturbed value of the current density given by the flux pinning strength while the exponent ∂ ln E n= (3.3.3) ∂ ln j expresses the significance of relaxation mechanisms. On decreasing n we can model situations with increasing importance of the flux creep (Fabbricatore et al 1996). The dynamic penetration depth for the combination of pinning and creep x0 is found for example by replacing jc in equation (3.2.3) by j given by equation (3.2): x0 = Ba Ba = . µ0 j µ0 jc (E/Ec )1/n (3.3.4) The characteristic electrical field the sample experiences in our experiment is (Campbell 1991) E≈ R ωBa . 2 (3.3.5) Inserting this quantity into equation (3.3.4) one finds 1−1/n x0 = Ba . µ0 jc (Rω/2Ec )1/n (3.3.6) Figure 5. AC susceptibility in the regime of reversible screening. We expect that the maximum of χ 00 and χr will be reached when x0 ≈ R (Geshkenlein et al 1991). At high values of n, the penetration depth (3.3.6) is only weakly frequency dependent while for n = 1 the amplitude dependence vanishes. Measurements at different frequencies (x0 ∼ ω−1/n ) could be used to find n. 3.4. Reversible screening In this regime the sample passes through the equilibrium states. This leads to a reversible character of the sample magnetization, yielding χ 00 = χr = 0. The pick-up coil signal does not contain higher harmonics and χ 0 = χa . From equation (3.4) it follows that the magnetic field decreases exponentially with the characteristic length λ = (3/µ0 )1/2 . The susceptibilities are (Campbell et al 1991) R λ 0 χ = χa = − 1 − tanh (3.4.1) R λ for a slab of width 2R in a parallel field and 2λ I1 (R/λ) 0 χ = χa = − 1 − R I0 (R/λ) (3.4.2) for a cylinder of radius R in a parallel field. In the formula for the cylinder, the modified Bessel functions I0 and I1 , sometimes called the Neumann functions (Spanier and Oldham 1987, p 533), are used. The graphical form of the susceptibilities is given in figure 5. Two important points are to be noted. (1) The temperature dependence of the penetration depth can be studied using AC susceptibility data measured on an object with well-defined geometry. The highest sensitivity of the data on the changes in λ is reached for R ≈ λ. 529 F Gömöry (2) A material of thickness comparable with λ is far from completely expelling the magnetic flux in spite of the fact that it is perfectly superconducting. This effect, called ‘magnetic invisibility’ (Clem and Kogan 1987), could lead to severe underestimation of the superconducting phase content particularly in the case of BiSrCaCuO compounds (Plecháček and Gömöry 1990). There are two typical examples of reversible mechanisms entering the flux dynamics: screening by Meissner currents and Campbell’s reversible screening. Meissner currents try to expel the magnetic flux from the superconducting volume to reach thermodynamical equilibrium. They are maximal on the interface with the external magnetic field; the characteristic space scale is the London penetration depth λL . The maximum drop of the field across the sample they can produce is equal to Bc1 , the lower critical magnetic field. At higher fields these currents only weaken the external field Bext to the actual value entering the sample bulk B = Bext − Bc1 . Both Bc1 as well as λL are temperature dependent, the values at T → 0 being among the basic physical characteristics of the material. In the experiment we can neglect the Meissner currents either when the magnetic fields are much larger than Bc1 or when there is no structure in the sample with dimensions comparable with λL (Ishida and Mazaki 1987). Taking into account that the values of λL are in the range of 1000 Å one can expect that the Meissner currents could be important for micrometre-sized samples. In the mixed state, we can imagine the individual flux lines to be positioned in potential wells. At small driving force, the displacement of the flux line from the bottom of the potential well could be reversible. Because of mutual repulsion between flux lines, compression of the flux line lattice appears starting from the surface. The deformation vanishes in the depth 1/2 c (3.4.3) λC = αL given by the elastic modulus of the lattice c and the Labusch parameter αL . In the case of Bd kBac the compressional modulus c11 should be applied, which for B Bc1 is material independent; c11 ≈ B 2 /µ0 where B 2 ≈ Bd2 + Ba2 . The Labusch parameter characterizes the steepness of the potential wells for the flux lines (Seow et al 1995). It is defined as the mean value of the second derivative of the potential energy U l ascribed to the flux line of unit length in the field of pinning forces: Z 2 l ∂ U 1 dA. (3.4.4) α= A A ∂x 2 Indications of this effect have been found (Matsushita et al 1992) and force–displacement curves derived from the AC magnetization data (Campbell 1971, Johnson et al 1994, Johnson and Campbell 1996). 3.5. Surface pinning In the previous considerations we supposed that there is no barrier to prevent the flux lines entering or leaving the 530 Figure 6. AC susceptibility in the regime controlled by the surface pinning. The magnetization loop of a sample with the irreversible surface barrier Bb for the flux entry and exit is plotted in the insert. sample. However, the existence of such a barrier was necessary to explain certain experiments on low-Tc samples (Bean and Livingston 1964, Dunn and Hlawiczka 1968, Melville 1972). Here we present a simple model (Cesnak et al 1984) that is equivalent to the existence of a layer with thickness δb and extremely high critical current density jb on the surface of the sample. The relevant quantity is the irreversible surface barrier Bb = µ0 jb δb . (3.5.1) The magnetization loop of a sample with the irreversible surface barrier is given in figure 6 (Clem 1991). The susceptibilities do not depend on the sample geometry, and one obtains (Clem 1979b) χ0 = 1 (2 − sin 2 cos 2) π (3.5.2a) 4u π(1 − u) (3.5.2b) χ 00 = where u = Bb /Ba and cos 2 = 1 − 2u. The wide-band susceptibilities can be calculated directly from the magnetization loop with the results ( −1 for 0 ≤ Ba ≤ Bb (3.5.3a) χa = 1 − Bb /Ba for Bb ≤ Ba 0 χr = 1 − Bb /Ba Bb /Ba for 0 ≤ Ba ≤ Bb for Bb ≤ Ba ≤ 2Bb for 2Bb ≤ Ba . (3.5.3b) Characterization by AC susceptibility These formulae are presented in graphical form in figure 6. Both χ 00 and χr reach a maximum at Ba = 2Bb . Notable differences between the complex AC susceptibility components χ 0 , χ 00 and the wide-band susceptibilities χa , χr indicate a rich content of higher harmonics in the sample magnetization. We see that Ba /Bb plays in this mechanism the role that is equivalent to for example the xc /R ratio in the case of bulk pinning. In order to achieve a formal analogy we can define the penetration depth as xb = R Ba . Bb (3.5.4) 3.6. Identifying the regime that controls the flux dynamics A common feature of the mechanisms discussed in this section was the existence of a characteristic space length xp . When xp < R (1) the time variation of fields and currents takes place in the surface layer with thickness xp and (2) the contribution to the magnetization of the sample due to the currents circulating deeper than xp is negligible. As far as χp depends on at least one of the external parameters (steady magnetic field Bd , the AC field amplitude Ba , the AC field frequency f = ω/2π and the temperature T ), we have the possibility to change xp or, more precisely, the ratio xp /R during the experiment. In the neighbourhood of the value xp ≈ R, we expect to observe a steep change of χ 0 or χa that will in the case of a dissipative mechanism be accompanied by a peak in χ 00 or χr . In a multiphase sample, several steps in χ 0 or χa corresponding to several peaks in χ 00 or χr should be found (e.g. Couach et al 1988, Cesnak 1992, Xia et al 1993). Changes of the sample quality due to aging, contamination etc lead to changes of the observed transition (Mehdaoui et al 1989) To distinguish between different mechanisms, we must check how this pattern is influenced by the controllable parameters of the experiment (Yamaguchi et al 1994). A summary of the essential features exhibited by the mechanisms considered in this paper is given in tables 1 and 2. 4. Determination of a superconductor’s parameters The scope of this paper is to review the procedures allowing us to extract as much information about the sample material as possible from the AC data (summarized in table 2). With the background information given in the preceding sections we can deal with this task. 4.1. Linear resistivity (ρlin ) and critical temperature (Tc ) In section 3.2 we have seen that in the case of linear diffusion the resistivity ρlin is the only material parameter controlling the flux dynamics. Then it can be found from Figure 7. Dependence of AC susceptibilities (both wide-band and complex AC susceptibility components) on the superimposed DC field can be used to determine the critical magnetic fields. Perfect reversible screening is expected for Bdc < Bc 1 while the superconducting response completely vanishes at Bdc > Bc 2 . Uncertainty in determining the critical fields is roughly given by the AC field amplitude. equation (3.1.9) or (3.1.10), e.g. numerically (Polichetti et al 1996). To test whether the data used to determine ρlin were taken in the linear regime we can compare the susceptibilities measured at different amplitudes (to be sure that they are amplitude independent) or various frequencies (the resulting ρlin should be frequency independent) or check the coincidence between the complex AC susceptibility and the wide-band susceptibilities (to confirm the non-existence of higher harmonics). In the case of known Bc2 and normal-state resistivity, one can derive the value of the typical height of the pinning barrier, U , from the susceptibility measured in the TAFF regime. The transition from a normal to a superconducting state is always accompanied by a resistivity drop. Therefore an AC susceptibility measurement can be used to determine the critical temperature, Tc (e.g. Khan et al 1994, Masini et al 1994, Das and Suryanarayanan 1995). Sometimes the onset temperature of transition, Ton , is used to characterize the transition temperature and thus the quality of the sample. In granular samples, however, Ton is often connected with the Meissner shielding, resulting in a decrease of χ and χa , while χ 00 and χr remain zero (Opagiste et al 1994). 4.2. Critical magnetic fields Bc1 , Bc2 The upper critical magnetic field Bc2 plays, on the flux density scale, a role similar to that of Tc on the temperature scale: crossing of Bc2 should be accompanied by a sharp change in the linear resistivity. Therefore the AC susceptibility data can be used to find Bc2 by analysing how the superimposed DC field shifts the onset temperature in the linear regime (Küpfer et al 1987). This is schematically illustrated in figure 7. Nevertheless, probably because of the existence of other more exact methods, the AC data are rarely used to determine Bc2 (Vlakhov et al 1994). In the case when the lower critical field Bc1 is to be determined, the situation is different. At low fields and/or low temperatures the Meissner currents secure the complete screening of a bulk sample without any magnetic 531 F Gömöry Table 1. Qualitative characteristics of the mechanisms controlling the flux dynamics. xp dependence on parameter Mechanism Formula for xp T Ba ω Bd Higher harmonics Eddy currents Linear flux flow Thermally assisted flux flow Bulk pinning Flux creep Meissner current Campbell’s reversible screening Surface pinning (3.1.3) + (3.1.4) (3.1.3) + (3.1.7) Weak Strong No No Strong Strong Weak Strong No No (3.1.3) + (3.1.8) (3.2.3) (3.3.6) λL (London) (3.4.3) (3.5.4) Strong Strong Strong Weaka Strong Strong No Strong Weak No No Strong Strong No Weak No No No Strong Weak Weak No Strong Weak No Yes Yes No No Yes a Strong near Tc . Table 2. Significant points on the AC susceptibility curves. xp /R at peak of χ 00 and χr Model Figure Linear diffusion Critical state Flux creep Reversible screening Surface barrier 3 4 — 5 6 Cylinder, radius R 0.556 1 ∼1 — 2 irreversibility: χ 0 = χa = −1, χ 00 = χr = 0. Crossing the Bc1 (T ) line towards higher temperatures or fields destroys this perfect screening (Goldfarb and Clark 1985). It is easy to find that χa = −1 χa = − Bc1 Ba for Ba < Bc1 for Ba ≥ Bc1 (4.2.1) χr = 0. A more sophisticated method based on a dynamic analysis of the flux line lattice in the reversible regime was proposed by Khoder et al (1991). When the flux lines enter the sample bulk they interact strongly with the material. One can expect the appearance of dissipation as well as incomplete diamagnetism when crossing the Bc1 (T ) line and an upper limit for Bc1 (T ) can be determined in this way (Goldfarb et al 1987b, Babic et al 1987, Loegel et al 1990, Loughran and Goldfarb 1991). Gömöry and Takács (1997) found that in the presence of flux pinning (characterized by the critical current density jc ) it is still possible to find the conditions allowing the determination of Bc1 , because at large enough AC field amplitudes χ a + χr = − Bc1 Ba for Ba ≥ 2µ0 jc R + Bc1 . (4.2.2) In this way the temperature dependences χa (T ), χr (T ) can be used to find Bc1 . 532 Slab, width 2R 0.887 1.414 ∼ 1.414 — 2 xp /R at χ 0 ≈ χa ≈ −0.5 Cylinder, radius R 0.556 0.634 ∼ 0.634 0.301 2 Slab, width 2R 0.887 1 ∼1 0.522 2 4.3. Critical current density jc Let us first analyse the simplest case when the flux pinning is the only mechanism controlling the flux dynamics, and the current-carrying capability is supposed to remain constant throughout the sample. A rough but simple estimation of jc can be performed by realizing that the maximum of χ 00 or χr is reached at a particular value of the AC field penetration depth: in the case of a cylinder (Clem 1988) Ba = µ0 jc R at max{χ 00 } or max{χr } (4.3.1a) and in the case of a slab with thickness 2R (Gömöry 1989) √ at max{χ 00 } or max{χr }. (4.3.1b) Ba = 2µ0 jc R Measuring at different AC field amplitudes Ba leads to finding the peaks at different temperatures T and with the help of equations (4.3.1) the jc (T ) dependence can be found (Gömöry and Lobotka 1988, Mezzetti et al 1994, Widder et al 1995). The drawback of this approach is that it uses only one point from the whole susceptibility curve. It is more effective to use the expressions (3.2.5) or (3.2.7) and to process a larger portion of the susceptibility data. Both χa and χr must give the same jc , and the parts of the jc (T ) curve determined from the data taken at different AC field amplitudes must overlap (Lera et al 1992, Fábrega et al 1993). Figures 8 and 9, respectively, illustrate how wide-band AC susceptibility data can be used to plot directly jc (T ) and jc (B) dependences. A slight difference Characterization by AC susceptibility Figure 9. (a ) Wide-band AC susceptibility of a YBaCuO melt-grown sample as a function of the DC background field, recorded at T = 84 K. The selected points were used to construct the jc (B ) curve in (b ). Figure 8. (a ) Wide-band AC susceptibility of a YBaCuO melt-grown sample. The marked portion of the curves was utilized to construct the jc (T ) curve in (b ). To allow the comparison, the AC field amplitudes and frequencies were chosen to induce the same electrical field on the sample’s surface. is observed between the jc values derived from χa and χr . I suppose that this is due to periodic changes of the electrical field that reaches a maximum during the cycle just at the moment when the remanent magnetization is determined, while the determination of the amplitude susceptibility happens when the actual E ∼ dB/dt is zero. Therefore the jc derived from χr are higher and represent the sample’s current-carrying capability better. Until now we have supposed that the magnitude of jc does not change during the cycle. A possible check of validity of this condition is the comparison of the experimentally found peak height of χ 00 or χr with the theoretical values given in figure 4. However, deviations from these values can have other origins, the most important being: reversible motion of flux lines (Matsushita et al 1991) and material granularity (Clem 1988). The value of the critical current density can vary for various reasons. Here we discuss two cases: jc changing as a result of variations in the local value of magnetic field and jc changing as a result of sample spatial inhomogeneity. Suppose first that the properties of the sample are the same everywhere and no DC field is applied. Then the value of the magnetic field that affects the sample ranges from 0 to Ba . Chen et al (1989) derived the formulae for both components of the complex AC susceptibility by considering Kim’s type of jc (B) dependence (Kim et al 1964): jc ≈ (B + B0 )−1 where B0 is a constant. They succeeded in finding in this way the jc (T ) dependence for a polycrystalline YBaCuO sample from the AC susceptibility 00 ) was data. A simpler analytical solution found for T (χpeak used by Lee and Kao (1995) to interpret a large set of data. Numerical calculations were performed to calculate the complex AC susceptibility components for different types of jc (B) dependences (Chen and Sanchez 1991a, Forsthuber and Hilscher 1992, Sanchez 1994). For the dependence B −β (4.3.2) jc (B) = jc0 1 + B0 having three parameters jc0 , B0 and β (where in particular β = 0 gives constant jc and β = 1 corresponds to Kim’s 533 F Gömöry model), the complex AC susceptibilities χ 0 and χ 00 were found analytically in the approximation Bd Ba (Sun et al 1995b) as well as the wide-band susceptibilities χa and χr (Gömöry et al 1993a). For β > 0 the height of the χ 00 or χr maximum exceeds the values illustrated in figure 4, calculated for constant jc . The necessity to consider jc (B) can be always avoided by application of a large DC field. In contrast, Bd comparable with Ba causes the effects created rather by a non-linear interaction between both of the fields, masking the intrinsic properties of the material (Gömöry et al 1989, Gjomesli and Fossheim 1994a, Vuong 1996). Another case is a macroscopically inhomogeneous sample. Here I do not mean the granularity of polycrystalline samples which will be analysed in section 4.5. The consequences of a broad distribution of the critical temperatures in an Ag–BiSrCaCuO composite were explained by Larrea et al (1994); sometimes the FWHM of χ 00 peak is reported as 1Tc (Nafidi and Suryanarayanan 1995). A powerful AC magnetic method to study the spatial inhomogeneity modelled by a sample consisting of cylindrical shells with different jc was developed by Campbell (1971). It is based on the evaluation of the mean value of the flux density inside the sample at the moment when the external field reaches maximum, B amp , and its change with AC field amplitude. As the amplitude Ba increases, the AC field probes successively more and more internal shells of the sample. The depth δAC , to which the AC field has penetrated, was found to be in the case of a slab (thickness 2R) δAC = R dB amp dBa (4.3.3) and in the case of a cylinder (radius R) dB amp 1/2 . δAC = R 1 − 1 − dBa (4.3.4) Taking into account the definition of the amplitude AC susceptibility (2.6), the χa data can be used for the determination of δAC by replacing dB amp /dBa by d[(1 + χa )Ba ]/dBa because B amp = Ba + Ma = Ba (1 + χa ). An example of a graph obtained when applying Campbell’s method is given in figure 10. The penetration depth δAC increases with Ba , and the local critical current density controls this proportionality: jc (δAC ) = 1 dBa . µ0 dδAC (4.3.5) This means that the local critical current density in the shell at a distance δ0 from the sample surface is proportional to the slope of the Ba (δAC ) curve in figure 10 at δAC = δ0 . In the case of a homogeneous sample with constant jc the Ba (δAC ) curve should be a straight line. 4.4. Surface barrier Bb In the case when the surface pinning is the only mechanism controlling the flux dynamics, values of the surface barrier Bb can be found by fitting the experimental data to 534 Figure 10. Campbell’s plot (‘flux density profile’) for a YBaCuO single crystal with higher jc on the well-oxygenated surface than in the interior. the theoretical expressions (3.5.2). However, one must normally consider at least the existence of volume pinning, characterized by the critical current density jc . Clem (1979b) derived the flux density profiles that can be used for calculation of the wide-band susceptibilities in this case (Gömöry and Takács 1997). Complicated expressions are expected if one includes the dependence of jc and Bb on the magnetic field. However, this is very desirable because the role of surface pinning seems to be more important than generally supposed yet (Chen and Sanchez 1992, Sun et al 1994b, Mamsurova et al 1994, Ding et al 1995). 4.5. Granularity and content of superconducting phase (vs ) In the initial period of HTSC development, all the prepared samples were polycrystalline, exhibiting behaviour explainable by the model of a ‘better’ phase embedded in a ‘worse’ matrix (Goldfarb et al 1987a). A typical manifestation of such behaviour is the AC susceptibility transition with two steps in χ 0 (T ) accompanied by two peaks in χ 00 (T ) as illustrated in figure 11 as well as the Campbell graph composed of two lines with different slopes (figure 12). It was early recognized that such behaviour is expected in the case of a sample consisting of superconducting grains representing elementary volumes of the phase coherence of the order parameter, interconnected by a system of weak links (Ishida and Mazaki 1987, Küpfer et al 1988, Müller 1990, Senoussi et al 1991, Levy et al 1994). When cooled under Tc the grains become superconducting first, and shielding is performed by the so-called intragrain currents with density jg . On lowering the temperature further, locking of the order parameter phases in different grains is reached (Clem 1988) and the whole sample behaves like a material able to carry the macroscopic intergrain current density jm , defined as P Atot II G,i (4.5.1) jM = Atot where the intergrain currents II G are summed flowing through a plane parallel to Bext and containing the principal symmetry axis: see figure 13. Such a definition does not give the actual current density carried by the intergrain contacts locally, but is equivalent to that determined in Characterization by AC susceptibility Figure 11. AC susceptibility curve of a polycrystalline YBaCuO sample with pronounced granularity. Figure 13. (a ) Schematic drawing of the paths of the intragrain current density jg and the (macroscopic) intergrain current density jm . (b ) The intergrain current density is related to the macroscopic area Atot . Figure 12. Campbell’s plot (‘flux density profile’) for the polycrystalline YBaCuO sample with pronounced granularity. the contact transport measurements. Then one expects to observe a dependence of the susceptibility on the magnetic history of the sample (Gömöry et al 1989, Puig et al 1992, Dhingra et al 1995). Probably the most complete explanation of this effect was given by Saha and Das (1993). By enhancing the quality of the intergrain contacts, the difference between inter- and intragrain currents vanishes, and the two steps (and peaks) merge (Sauv et al 1993, Planinić et al 1994, Vo et al 1996). However, very small fractions (∼1%) of a poorer quality superconductor or the weak links in single crystals could be detected (Moreira et al 1994, Ren et al 1994) by AC susceptibility. Let us demonstrate the way of extracting essential information from the AC susceptibility data taken on a granular sample. We start with the simplest model: inside the grains the flux dynamics is governed by the intragrain pinning force characterized by the intragrain critical current density jgc while the system of intergrain contacts forms an effective pinning medium (Clem 1989, Müller et al 1991, Ravi and Seshu Bai 1994, Doyle T B et al 1994, Feigelman and Ioffe 1995, Chen et al 1995), characterized by the intergrain critical current density jmc . Because of the smaller dimensions of grains with respect to the whole sample and because generally jgc is larger than jmc , the χ 00 (T ) peak appearing at higher temperature (T1 in figure 11) has to be ascribed to the situation when the AC field penetrates just to the centres of grains. The critical state model (section 3.2) allows us to determine jgc |T =T1 = Ba µ0 Rg (4.5.2) where Rg is the grain radius. For simplicity we suppose that all the grains are identical and the sample is dense enough to exhibit strong magnetic coupling between individual grains (Senoussi et al 1990). Then the system of grains can be handled as a set of cylinders parallel to the external field (Müller 1989). Questionable is here the estimation of the grain size. Early studies revealed that the elementary superconducting coherent volume can differ from the grain size observed for example by optical methods (Däumling et al 1988, Chen et al 1990, Nicolas et al 1995) and the 535 F Gömöry existence of intragrain contacts was suggested (Küpfer et al 1989). An analogous problem does not exist in the determination of jmc from the χ 00 (T ) peak observed at lower temperature (T2 in figure 11). This peak is reached when the intergrain currents arrive at the value allowing the AC field to penetrate through the intergrain space just to the sample centre. Then, for a cylinder (radius R) and a slab (thickness 2R) respectively jmc |T =T2 = Ba µ0 R Ba jmc |T =T2 = √ . 2µ0 R (4.5.3) Using a series of AC field amplitudes, a set of peak temperatures T2 is obtained and the jmc (T ) dependence can be constructed. When the temperatures T1 , T2 are sufficiently distant the peaks can be handled separately. Then the models valid for a homogeneous superconductor, i.e. involving more mechanisms, can be applied to the portions of data manifesting the inter- and the intragrain currents (Chen and Sanchez 1991b, Sanchez and Chen 1991). When trying to reveal the existence of links controlling the intergrain currents, Campbell’s method is very useful. At low amplitudes Ba the field is supposed to penetrate first the intergrain volume, and therefore the initial part of the Ba (δAC /R) curve corresponds to the shielding performed by the intergrain currents. At a certain value Ba∗ the AC field is just sufficient to occupy the whole intergrain space and the AC flux starts to enter the grains. The flux penetration into grains at B > Ba∗ is opposed by the intragrain current density and the part of the Campbell curve at B > Ba∗ can be used to determine jgc . Thus, analysing the graph from figure 12 one finds Ba∗ 1 dBa jgc = . (4.5.4) jmc = µ0 R µ0 Rg d(δAC /R) B>Ba∗ At B = Ba∗ the whole intergrain space is occupied by the AC field. Because the total volume of the sample Vtot is the sum of the intergrain space Vi and the space occupied by the superconducting grains Vg , one finds from simple geometrical considerations that the ‘porosity’ of the sample can be defined for a slab (thickness 2R) and a cylinder (radius 2R) as respectively δAC ∗ R Vi = (4.5.5) vi = 2 Vtot δAC ∗ . R The same result is obtained when we suppose that the grains are perfectly shielded by the intragrain Meissner currents at Ba ≈ Ba∗ . Let us derive the wide-band AC susceptibilities in the following model case (Opagiste et al 1993): the grains are perfectly diamagnetic in the considered interval of temperatures and AC field amplitudes owing to Meissner currents, while the intergrain currents shield the sample as a whole, creating the critical state described by the critical current density jmc (Bekeris et al 1994). For simplicity 536 we neglect the possible coupling between the inter- and intragrain currents (Kasatkin et al 1994). The diamagnetic action of the intragrain currents excludes the sample volume fraction fs from the penetration as well as capturing the AC magnetic flux (Calzona et al 1989, Marohnić and Babic 1991). This leads to the modification of formulae (3.2.5) or (3.2.7). Then, for a cylinder (radius R) at Ba > 2µ0 jmc R χa = − µ0 jmc R − fs 3Ba χr = µ0 jmc R (1 − fs ) (4.5.6a) 3Ba and in the case of a slab (thickness 2R) χa = − µ0 jmc R −fs 2Ba χr = µ0 jmc R (1−fs ). (4.5.6b) 2Ba This means that the Meissner shielding of grains will depress the height of the intragrain χr peak and in the same time sharpen the χa step. Data of this type were found on a Bi-2212 melt-grown sample (Gencer et al 1996a) and a Bi-2223 polycrystal (Gencer et al 1996b). The case of large AC amplitudes or low intergrain currents is important, when χa |Ba µ0 jcm R = −fs χr |Ba µ0 jcm R = 0. (4.5.7) This means that if a ‘valley’ in χr exists between T1 and T2 that goes down to zero, the corresponding value of χa gives directly the diamagnetic fraction of the sample. Too weak an AC field amplitude would disable this type of determination because a minor superconducting fraction could also shield the whole sample against a low AC field (Dubois et al 1995). Alternatively, fs can be derived from the height of the intergrain maximum (i.e. that found at T2 in figure 11), if it is supposed that the diamagnetic portion cannot hold any magnetic flux (Ravinder Reddy et al 1995). Then we find that for the cylinder √ χr (T2 ) = (1 − fs )/4 and for the slab χr (T2 ) = (1 − 1/ 2)(1 − fs ), and fs can be found: √ 2[1 − χr (T2 ) − 1] fs = fS = 1 − 4χr (T2 ) (4.5.8) √ 2−1 for a cylinder (radius R) and a slab (thickness 2R) respectively. Ghivelder et al (1994) studied the susceptibility of a cylindrical sample consisting of YBaCuO grains in an Ag matrix and explained the observed behaviour with the help of the effective-medium theory (McLachlan et al 1990), which for the incompletely penetrated intergrain volume (i.e. low AC fields) yields χ 0 = −1 + 2 1 Ba Ba − (1 − fs )1.5 (4.5.9a) µ0 jc R 3 µ0 jc R 2 Ba 4Ba (1 − fs )1.5 −2 . χ = µ0 jc R µ0 jc R 3π 00 (4.5.9b) A difference in the estimation of fs using different formulae (4.5.7) and (4.5.8) is found when for example the model of jmc independent of the magnetic field is no longer valid. Another reason for the discrepancy could be that the diamagnetic fraction of the sample volume has changed its value owing to the temperature dependence of λL or Bc1 . Characterization by AC susceptibility So far we have supposed that the grains are perfectly shielded. Thus, the diamagnetic fraction of the sample volume, fs , was identical with the volume fraction occupied by the superconducting grains, vg . This condition requires the grain dimension Rg to be much larger than the London penetration depth, λL . Otherwise we must account for the effect of so-called magnetic invisibility (Clem and Kogan 1987) mentioned in section 3.4. Then, the fraction of the sample volume screened by the intragrain Meissner currents, fs , will differ from the volume fraction occupied by the superconducting grains, vg (Chen et al 1990). Replacing again the individual grains by a system of identical cylinders (radius Rg ) and taking into account the formula (3.4.2) we find that for the fields under Bc1 vg = 1− fs 2λL I1 (Rg /λL ) Rg I0 (Rg /λL ) . (4.5.10) For large grains (Rg λL ) the second term in the denominator goes to zero. Then vg = fs and the content of superconducting phase, given by vg , can be determined in a straightforward way. Otherwise, vg is larger than fs . In the limiting case of tiny grains with Rg λL , the denominator of equation (4.5.10) goes to zero and one will observe a negligible diamagnetic fraction of the sample, fs , also in the case of a dense, single-phase sample with vg → 1. Therefore the knowledge of λL as well as the grain shape and dimension is important (Yang et al 1994a). In particular for Bi-2223 superconductors with plate-like grains, the susceptibility depends strongly on the field orientation (Berling et al 1996a) and this difference can be used to estimate the degree of texturing (Gömöry et al 1996). For the grains supposed to have the form of slab-like columns instead of cylinders the formula (4.5.10) should be modified using tanh(Rg /λL ) in place of the Bessel functions (Angurel et al 1994). The previous considerations apply well to dense samples with the grains oriented parallel to the external field. In the opposite case (isolated grains in the form of plates perpendicular to the field) the magnetic moments of individual grains increase the strength of the effective field, leading to overestimation of vg . The existence of two peaks does not necessarily signify sample granularity. For example in untwinned YBaCuO single crystals this was explained by a first-order melting transition in the flux line lattice (Giapintzakis et al 1994, Billon et al 1996, Tatara et al 1996). 4.6. London’s (λL ) and Campbell’s (λC ) penetration depths Determination of the reversible screening depth λ by the AC magnetization technique requires the preparation of particular samples that contain grains comparable in dimensions with expected values of λ (Sargánková et al 1996), preferably in an insulating matrix (Guerrin et al 1994). Small AC amplitudes should be used to make this effect dominant (Shoenberg 1937). By fitting the experimental temperature-dependent data to the theoretical expressions (3.4.1) and (3.4.2) one can find the appropriate λL (T ) dependence and compare it with predictions of various microscopic theories of superconductivity (Mühlschlegel 1974, Riseman et al 1994, Guerrin et al 1995). Because of the intrinsic anisotropy of HTSCs, grains should be aligned, e.g. magnetically (Boolchand et al 1992, Porch et al 1993), and the values of the penetration depth in two principal directions determined separately (Athanassopoulou et al 1994). The first attempt (Takács and Campbell 1988) to perform the electromagnetic calculations for the case when the bulk pinning (characterized by the critical current density jc ) is combined with the reversible motion of the flux lines (characterized by the Campbell penetration depth λC ) was followed by Matsushita et al (1991). They obtained for the slab with thickness 2R the numerical solutions and approximated then as 1 1 + 3(λC /R)2 + y (4.6.1a) y 2 π 3[1 + 2(λC /R)2 ]2 + y 2 (4.6.1b) χ0 = χ 00 = where y = Ba /µ0 jc R. The height of the imaginary peak, 00 , is depressed significantly (Campbell et al 1991) and χmax allows λC to be estimated regardless of the pinning strength: λC = R . √ 00 (2π χmax 3)1/2 (4.6.2) 4.7. Irreversibility line We define the irreversibility line (IL) as the line of zero critical current density: jC (Birr ) = 0 (Matsushita 1993, Dubois et al 1996). Remember that in the formulae for the determination of jc (see section 4.3), the critical current density is always proportional to the AC field amplitude Ba . Thus, the direct determination of the jc = 0 line would require the measurement at Ba = 0, which is impossible. We can extrapolate to this line by using the data taken at low amplitudes or finding the point where the upper part of the χ 00 peak goes to zero (Dhingra and Das 1993, Noetzel et al 1996). Another possibility is to select a certain level of critical current density as that approaching closely enough, from the point of view of the studied problem, the zero level. Then, instead of determining the hardly accessible true IL, we can study a ‘practical’ IL defined by jc = jt (Sun et al 1994a, Mosbah et al 1994) where jt is commonly chosen in the interval 1–100 A cm−2 (Gömöry and Takács 1993, Fabbricatore et al 1993b, Gömöry et al 1993b). ILs determined from the susceptibility peaks are exactly the practical ILs (Loegel et al 1993). This is because, for the conditions of the χ 00 peak found at different DC fields superimposed on the AC field with constant amplitude, the critical current density is always the same, given by equations (4.3.1). A quantitative comparison of the ILs determined separately from the AC susceptibility and the DC magnetization is also possible when we compare the same extent of irreversibility expressed by identical values of jt (Gömöry et al 1994). 537 F Gömöry The difficulty in employing a dynamic AC method in the determination of a static critical current density clearly increased when ILs were revealed to be frequency dependent (Nikolo and Goldfarb 1989, Samarappuli et al 1992). The flux density gradient created in the sample during the change of the external magnetic field can no longer be assumed to persist in time, especially in granular samples (Nikolo and Missey 1994, Pérez et al 1994). Its relaxation towards equilibrium is more pronounced at lower frequencies because of the longer cycle duration. The macroscopic current density determined for example from the AC susceptibility peak in that case decreases with decreasing frequency. When the frequency is lowered, the ILs shift towards lower temperatures because the gain in the unrelaxed value of current density must compensate for its loss due to relaxation. The problem of handling strongly relaxing flux density gradients was discussed in section 3.3. The approach used there can be judged as a modification of the original model of Bean (1962): we assume that the actual current density j flowing in the sample slightly differs from jc and depends on the electrical field (Yeshurun and Malozemoff 1988). In the general treatment of the flux dynamics in section 3 we used for the j (E) dependence a phenomenological model (3.2). This model together with the electrical field estimation (3.3.6) is quite sufficient to predict the shift of the IL with frequency. Thorough calculations of the electrical fields can be found in the work of Brandt (1995). The shift of the ILs when the frequency is changed is illustrated in figure 14. Experimental ILs are determined from the AC susceptibility peak and from the quasistatic magnetization loops recorded on a SQUID susceptometer. Theoretical lines are calculated using the hypothetical j (E) characteristics of the form (3.2). The only parameter that had to be fitted was the exponent n (Gömöry et al 1995). 4.8. Pinning potential U0 Another interesting parameter that can be found from the AC susceptibility data taken at different frequencies is the height of the barrier that the moving element of the flux line lattice must overcome when hopping to the next pinning site (Savvides et al 1992, Kusevic et al 1994, Fábrega et al 1994, Devos et al 1994, Sun et al 1995a, Mehdaoui et al 1995). Generally, the flux line can be considered as a particle moving in a force field created by the potential wells at pinning sites. The effective depth, Ueff , of the pinning well in the presence of a flux density gradient can be separated into two terms, i.e. Ueff = U0 (T , B)f [j/jc (T , B)] (4.8.1) the first of them being a material parameter while the second reflects the modification of the barrier height by the Lorentz force. For the models of Anderson and Kim (1964), Feigelman et al (1989), Zeldov et al (1990) and Fisher et al (1991) one can find U0 from the slope of the j (E) curve at j = jc (Welch et al 1990, Cesnak et al 1996): ∂Ueff ∂ ln E = k T . (4.8.2) U0 = −jc B ∂j j =jc ∂ ln j j =jc 538 Figure 14. The irreversibility lines measured at frequencies of 0.1 (F), 2, 12.3, 123, 1230, 12 300 ( ) Hz on a YBaCuO melt-grown sample, compared with the lines calculated on the basis of the thermally activated flux creep model. • As we can see from the last expression, the determination of U0 is affected by the definition of jc except the case when ∂ ln E/∂ ln j = constant. Exactly this happens for the logarithmically shaped potential Ueff = U0 ln(jc /j ) analysed by Zeldov et al (1990) and leading to the j (E) curve of the type j ∼ E n where n = ∂ ln E/∂ ln j . We can collect the susceptibility data at different AC field amplitudes much larger than Bp = µ0 j R and process them by supposing that the thermal activation is only a small perturbation of the pinning-controlled behaviour (Shi et al 1995). Then the critical-state model formulae (3.2.8) or (3.2.9) can be still used but we must replace jc in these formulae by the electrical field dependent j from equation (3.2). Including the expression (3.3.5) estimating the electrical field, we can construct the E(j ) curve and determine U0 with the help of relation (4.8.2): see figure 15. There is still a need for discussion if the pinning potential found in the creep regime should be identical with the activation energy derived from the Arrhenius plots of the resistivity in the transition region: Ua . (4.8.3) ρ = ρ0 exp kB T The logarithmic U (j ) dependence found for melt-textured YBCO by Quin et al (1996) seems to support this idea. 5. Conclusions The goal of this review was to get together the basic information necessary for an experimentalist to extract from the AC susceptibility data of an HTSC sample as much information as possible. In addition to the complex Characterization by AC susceptibility of various models. I would like to mention at least some of them: L Cesnak, P Fabbricatore, M Forsthuber, K Fröhlich, G Hilscher, T Holubar, P Lobotka, I Madera, V Plecháček, I Pochaba, M Polichetti, A Rosová, P Kováč, S Takács, Yu P Tarenkov. The financial support from the Slovak Grant Agency under Contracts GAV 2/1087/94 and GAV 108/96 is also acknowledged. References Figure 15. Construction of the E (j ) curve (insert) from the wide-band AC susceptibility data measured at 85 K, 0.1 T on a YBaCuO thin film. The portion of the data used for this purpose is indicated by the arrows. From the logarithmic slope of the E (j ) curve, the activation energy was found as U0 /kB = nT = 1445 K. AC susceptibility, the wide-band AC susceptibilities were analysed. The formulae for susceptibilities corresponding to the models of linear diffusion, reversible screening, critical state, surface barrier pinning and flux creep can be used in two ways: for the recognition of the mechanism controlling the flux dynamics and in the determination of material properties. We have seen that such important parameters as the critical temperature, critical magnetic field, linear resistivity, critical current density, surface barrier, granularity, volume filling by superconducting grains, London and Campbell penetration depths, and pinning barrier height can be derived under favourable circumstances from the AC magnetization data. To discuss clearly the principal physical features, the models used in this paper are the simplest possible. Therefore I did not present here for example the results for field-dependent critical current density or surface barrier, which can be found elsewhere (see references in corresponding sections), or a discussion about the influence of the sample shape (Chen and Goldfarb 1989, Sun J Z et al 1991, Forsthuber and Hilscher 1992, Clem and Sanchez 1994, Brandt 1994a, 1994b). Acknowledgments What could not be referred to in the text of this review is the support of many colleagues who helped me substantially either in constructing the apparatus and performing the experiments or in discussing the results and applications Anderson P W and Kim Y B 1964 Rev. Mod. Phys. 36 39 Ando Y, Kubota H, Sato Y and Terasaki I 1994 Phys. Rev. B 50 9680 Angurel L A, Lera F, Rillo C and Navarro R 1994 Physica C 230 361 Athanassopoulou N, Cooper J R and Chrosch J 1994 Physica C 235–240 1835 Babic E, Marohnić Ž, Drobac D and Prester M 1987 Int. J. Mod. Phys. B 1 973 Baldini A, Borchi E, Garré R, Lascialfari A, Masi L and Peruzzi A 1996 Physica C 262 68 Bean C P 1962 Phys. Rev. Lett. 8 250 —— 1964 Rev. Mod. Phys. 36 31 Bean C P and Livingston J D 1964 Phys. Rev. Lett. 12 14 Bekeris V, Ferrari H, Acha C, Levy P, Polla G and Leyva G 1994 Physica C 234 49 Berling D, Beille J, Loegel B, Mehdaoui A, Noudem J G and Tournier R 1996a Supercond. Sci. Technol. 9 161 Berling D, Bolmont D, Loegel B and Mehdauoi A 1996b Solid State Commun. 97 731 Billon B, Charalambous M, Chaussy J, Lejay P, Koch R H and Hardy W N 1996 Czech. J. Phys. 46 (suppl S3) 1581 Blatter G, Feigelman M V, Geshkenbein V B, Larkin A I and Vinokur V M 1994 Rev. Mod. Phys. 66 1125 Boolchand P, Pradhan S, Wu Y, Abdelgadir M, Huff W, Farrell D, Coussement R and McDaniel D 1992 Phys. Rev. B 45 921 Bozec X, Maury R, Fert A R, Redoulis J P, Bertrand D, Potel M, Levet J C, Gougeon P and Noël H 1991 Physica C 179 22 Brandt E H 1990 Z. Phys. B 80 167 —— 1991a Int. J. Mod. Phys. B 5 751 —— 1991b Phys. Rev. Lett. 67 2219 —— 1992a Physica C 195 1 —— 1992b Supercond. Sci. Technol. 5 S25 —— 1994a Phys. Rev. B 49 9024 —— 1994b Phys. Rev. B 50 4034 —— 1995 Phys. Rev. B 52 15 442 Calzona V, Cimberle M R, Ferdeghini C, Putti M and Siri A S 1989 Physica C 157 425 Campbell A M 1969 J. Phys. C: Solid State Phys. 2 1492 —— 1971 J. Phys. C: Solid State Phys. 4 3186 —— 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 129 —— 1995 IEEE Trans. Appl. Supercond. 5 682 Campbell A M, Blunt F J, Johnson J D and Freeman P A 1991 Cryogenics 31 732 Campbell A M and Evetts J E 1972 Adv. Phys. 21 199 Cave J R, Critchlow P R, Lambert P and Champagne B 1991 IEEE Trans. Magn. 27 1379 Cesnak L 1992 Czech. J. Phys. 42 1025 Cesnak L, Gömöry F, Kokavec J and Takács S 1984 Cryogenics 24 119 Cesnak L, Gömöry F and Kováč P 1996 Supercond. Sci. Technol. 9 184 Chen D X and Goldfarb R B 1989 J. Appl. Phys. 66 2489 Chen D X, Mei Y and Luo H L 1990 Physica C 167 317 Chen D X, Nogues J and Rao K V 1989 Cryogenics 29 800 Chen D X and Sanchez A 1991a J. Appl. Phys. 70 5463 539 F Gömöry Chen D X and Sanchez A 1991b Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 259 —— 1992 Phys. Rev. B 45 10 793 Chen D X, Sanchez A and Hernando A 1995 Physica C 243 123 Civale L, Worthington T K, Krusin-Elbaum L and Holtzberg F 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 313 Clem J R 1979a AC Losses in Type II Superconductors (Ames, IA: Iowa State University) —— 1979b J. Appl. Phys. 50 3518 —— 1988 Physica C 153–155 50 —— 1989 Physica C 159 717 —— 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 177 Clem J R, Kerchner H R and Sekula T S 1976 Phys. Rev. B 14 1893 Clem J R and Kogan V R 1987 Japan. J. Appl. Phys. 26 (suppl 26-3) 1161 Clem J R and Sanchez A 1994 Phys. Rev. B 50 9355 Couach M and Khoder A F 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 35 Couach M, Khoder A F, Monnier F, Henry J Y and Barbara B 1988 Phys. Rev. B 38 748 D’Anna G, André M O, Indenbom M V and Benoit W 1994 Physica C 235–240 3197 Das A and Suryanarayanan R 1995 J. Physique I 5 623 Däumling M, Seuntjens J and Larbalestier D C 1988 Appl. Phys. Lett. 52 590 de la Cruz F, Pastoriza H, Lopez D, Goffman M F, Arribére A and Nieva G 1994 Physica C 235–240 83 Devos P, Buekenhoudt A, D’Anna G D, André M O, Indenbom M V, Benoit W, DeBatist R and Cornelis J 1994 Physica C 235–240 3223 Dhingra I and Das B K 1993 Supercond. Sci. Technol. 6 765 Dhingra I, Moorthy V N and Das B K 1995 Supercond. Sci. Technol. 8 252 DiGioacchino D, Pace S, Testa A M, Tripodi P, Celani F and Polichetti M 1996 Czech. J. Phys. 46 (suppl S3) 1801 Ding S Y, Wang G Q, Yao X X, Peng H T, Peng Q Y and Zhou S H 1995 Phys. Rev. B 51 9107 Doyle R A, Seow W S, Johnson J D, Walker D W, Campbell A M, Doyle T B, Somekh R E and Evetts J E 1994 Proc. 7th Int. Workshop on Critical Currents in Superconductors ed H W Weber (Singapore: World Scientific) p 181 Doyle T B, Doyle R A, Minkov D, Stepankin V N and Yakovets U P 1994 Physica C 233 253 Dubois S, Carmona F and Flandrois S 1995 Physica C 245 26 —— 1996 Physica C 260 19 Dubots P and Cave J 1988 Cryogenics 28 661 Dunn W J and Hlawiczka P 1968 J. Phys. D: Appl. Phys. 1 1469 Fabbricatore P, Farinon S, Gemme G, Musenich R, Parodi R and Zhang B 1993a Cryogenics 33 1170 —— 1993b Physica C 213 200 —— 1994 Phys. Rev. B 50 3189 Fabbricatore P, Priano C, Sciutti A, Gemme G, Musenich R, Parodi R, Gömöry F and Thompson J R 1996 Phys. Rev. B 54 12 543 Fábrega L, Fontcuberta J, Civale L and Piňol S 1994 Phys. Rev. B 50 1199 Fábrega L, Fontcuberta J, Piňol S, nam der Beek C J and Kes P H 1993 Phys. Rev. B 47 15 250 Fedorov Y A and Stepanov Y P 1996 Czech. J. Phys. 46 (suppl S3) 1641 Feigelman M V, Geshkenbein V B, Larkin A I and Vinokur V M 1989 Phys. Rev. Lett. 63 2303 540 Feigelman M V and Ioffe L B 1995 Phys. Rev. Lett. 74 3447 Filippi J, Pugnat P, Barbara B, Ingold M, Bourgault D and Tournier R 1994 Physica C 235–240 3138 Fisher D S, Fisher M P A and Huse D A 1991 Phys. Rev. B 43 130 Forsthuber M and Hilscher G 1992 Phys. Rev. B 45 7996 Frischherz M C, Sauerzopf F M, Weber H W, Murakami M and Emel’chenko G A 1995 Supercond. Sci. Technol. 8 485 Gencer A, Muthu R H, Belenli I, Özogul Ö and Aydinuraz A 1996a Supercond. Sci. Technol. 9 284 Gencer A, Nezir S, Altunbas M and Aydinuraz A 1996b Supercond. Sci. Technol. 9 467 Geshkenbein V B, Vinokur V M and Fehrenbacher R 1991 Phys. Rev. B 43 3748 Ghivelder L, Abrego Castillo I, Pimentel L M, Pureur P and Schilling O F 1994 Physica C 235–240 3221 Giapintzakis J, Neiman R L, Ginsberg D M and Kirk M A 1994 Phys. Rev. B 50 16 001 Gilchrist J 1994 Physica C 219 67 Gilchrist J and Konczykowski M 1990 Physica C 168 123 Gilchrist J and van der Beek C J 1994 Physica C 231 147 Gjomesli S and Fossheim K 1994a Physica C 220 33 —— 1994b Physica C 235–240 3203 Goldfarb R B and Clark A F 1985 IEEE Trans. Magn. 21 332 Goldfarb R B, Clark A F, Braginski A I and Panson A J 1987a Cryogenics 27 475 Goldfarb R B, Clark A F, Panson A J and Braginski A I 1987b High Temperature Superconductors vol EA-11 ed D V Gubser and M Schluter (Pittsburgh, PA: Materials Research Society) p 261 Goldfarb R B, Lelental M and Thompson C A 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 49 Gömöry F 1989 Solid State Commun. 70 879 —— 1991a Thermochim. Acta 174 299 —— 1991b Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 289 Gömöry F, Forsthuber M and Hilscher G 1993a Cryogenics 33 277 Gömöry F, Fröhlich K, Šouc J, Kováč P and Tarenkov Y P 1993b Applied Superconductivity ed H C Freyhardt (Oberursel: DGM Informationsgesellscheft) p 395 Gömöry F and Lobotka P 1988 Solid State Commun. 66 645 Gömöry F, Pachla W, Kováč P, Pochaba I and Hušek I 1996 Czech. J. Phys. 46 (suppl S3) 1483 Gömöry F and Takács S 1993 Physica C 217 297 —— 1997 Superlattices Microstruct. 21 Suppl. A 219 Gömöry F, Takács S, Holubar T and Hilscher G 1994 Physica C 235–240 2753 —— 1995 Proc. CEC/ICMC ’95 Adv. Cryog. Eng. Materials vol 42, ed L T Summers (New York: Plenum) p 587 Gömöry F, Takács S, Lobotka P, Fröhlich K and Plecháček V 1989 Physica C 160 1 Griffiths D J, Koch C C and Charlesworth J P 1976 Phil. Mag. 33 505 Guerrin L, Alloul H and Collin G 1994 Physica C 235–240 1797 —— 1995 Physica C 251 219 Gugan D 1994 Physica C 233 165 Hein R A and Falge R L Jr 1961 Phys. Rev. 123 407 Ishida T and Mazaki H 1979 Phys. Rev. B 20 131 —— 1987 Japan. J. Appl. Phys. 26 L2003 Ji L, Sohn H R, Spalding G C, Lobb C J and Tinkham M 1989 Phys. Rev. B 40 10 936 Jiang H and Bean C P 1994 Appl. Supercond. 2 689 Johnson J D and Campbell A M 1996 Czech. J. Phys. 46 (suppl S3) 1673 Johnson J D, Campbell A M, Doyle R A, Athanassopoulou N and Cooper J R 1994 Proc. 7th Int. Workshop on Critical Currents in Superconductors ed H W Weber (Singapore: World Scientific) p 185 Characterization by AC susceptibility Josephson B D 1965 Phys. Lett. 16 242 Karkut M G, Slaski M, Heill L K, Sagdahl L T and Fossheim K 1993 Physica C 215 19 Kasatkin A L, Pan Y M, Petrukhno Y A, Prokhorov V G and Tretiachenko C G 1994 Proc. 7th Int. Workshop on Critical Currents in Superconductors ed H W Weber (Singapore: World Scientific) p 335 Kerchner H R, Sun Y R, Thompson J R, Khan H R, Thompson J O and Christen D K 1995 Physica C 250 331 Kes P H, Aarts J, van den Berg J, van der Beek C J and Mydosh J A 1989 Supercond. Sci. Technol. 1 242 Khan H R, Loebich O, Fabbricatore P, Sciutti A and Zhang B 1994 Physica C 229 165 Khoder A F and Couach M 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 213 Khoder A F, Couach M, Monnier F and Henry J Y 1991 Europhys. Lett. 15 337 Kim Y B, Hempstead C F and Strnad A R 1964 Rev. Mod. Phys. 36 43 Kim Y B and Stephen M J 1969 Superconductivity vol 2, ed R D Parks (New York: Marcel Dekker) p 1107 Köber J, Gupta A, Esquinazi P, Braun H F and Brandt E H 1991 Phys. Rev. Lett. 66 2507 Kötzler J, Nakielski G, Baumann M, Behr R, Goerke F and Brandt E H 1994 Phys. Rev. B 50 3384 Küpfer H, Apfelstedt I, Flükiger R, Keller C, Meier-Hirmer R, Runtsch B, Turowski A, Wiech U and Wolf T 1989 Cryogenics 29 268 Küpfer H, Apfelstedt I, Schauer W, Flükiger R, Meier-Hirmer R and Wühl H 1987 Z. Phys. B 69 159 Küpfer H, Apfelstedt I, Schauer W, Wolf T and Wühl H 1988 Physica C 153–155 367 Kusevic I, Babic E, Ivkov J, Marohnic Z, Liu H K, Hu Q Y and Dou S X 1994 Solid State Commun. 92 735 Labusch R 1969 Cryst. Lattice Defects 1 1 Larrea A, Snoeck E, Badia A, de la Fuente G F and Navarro R 1994 Physica C 220 21 LeBlanc D and LeBlanc M A R 1992 Phys. Rev. B 45 5443 Lee C Y and Kao Y H 1995 Physica C 242 167 Lera F, Navarro R, Rillo C, Angurel L A, Badı́a A and Bartolomé J 1992 J. Magn. Magn. Mater. 104–107 615 Levy P, Ferrari H, Acha C and Bekeris V 1994 Physica C 222 212 Li J, Ding S Y, Shao H M, Zhu J S and Wang Y N 1994 Physica C 232 10 Loegel B, Mehdaoui A and Bolmont D 1990 Supercond. Sci. Technol. 3 504 Loegel B, Mehdaoui A, Bolmont D, Danesi P, Bourgault D and Tournier R 1993 Physica C 210 432 London F 1961 Superfluids vol 1 (New York: Dover) Loughran R J and Goldfarb R B 1991 Physica C 181 138 Malozemoff A P 1989 Physical Properties of High Temperature Superconductors ed D M Ginsberg (Singapore: World Scientific) p 71 Mamsurova L G, Pigalskiy K S, Sakun V P and Scherbakova L G 1994 Proc. 7th Int. Workshop on Critical Currents in Superconductors ed H W Weber (Singapore: World Scientific) p 153 Marohnić Ż and Babic E 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 267 Masini R, Dimesso L, Migliori A, Francesconi M G and Calestani G 1994 Physica C 223 189 Matsushita T 1993 Physica C 214 100 Matsushita T and Ni B 1989 Japan. J. Appl. Phys. 28 L419 Matsushita T, Otabe E S, Kimura K, Hashimoto M and Küpfer H 1992 Physica C 196 98 Matsushita T, Otabe E S and Ni B 1991 Physica C 182 95 Maxwell E and Strongin M 1963 Phys. Rev. Lett. 10 212 Mazaki H, Yasuoka H, Kakihana M, Fujimori H, Yashima M and Yoshimura M 1995 Physica C 252 275 McLachlan D S, Blaszkiewicz M and Newnham R E 1990 J. Am. Ceram. Soc. 73 2187 Mehdaoui A, Berling D, Bolmont D and Loegel B 1995 Mater. Sci. Eng. B 34 132 Mehdaoui A, Bolmont D, Danesi P, Bourgault D and Loegel B 1993 Supercond. Sci. Technol. 6 322 Mehdaoui A, Bolmont D and Loegel B 1989 Phys. Status Solidi a 116 777 Melville P H 1972 J. Phys. D: Appl. Phys. 5 613 Mezzetti E, Colombo S, Gerbaldo R, Ghigo G, Gozzelino L, Minetti B and Cherubini R 1994 Physica C 235–240 2977 Moreira F M A, de Carvalho J S, Ortiz W A and de Lima O F 1994 Physica C 235–240 2305 Morozov N V, Zeldov E, Majer D and Khaykovich B 1996 Czech. J. Phys. 46 (suppl S3) 1543 Mosbah M F, Senoussi S, Frikach K, Taoufik A, Fert A and Odier P 1994 Physica C 235–240 2769 Müller K H 1989 Physica C 159 717 —— 1990 Physica C 168 585 Müller K H, Nikolo M and Driver R 1991 Phys. Rev. B 43 7976 Müller K H, Savvides N, Katsaros A, Guo Y C and Dou S X 1996 Czech. J. Phys. 46 (suppl S3) 1297 Mühlschlegel B 1974 Phys. Rev. B 10 1127 Nafidi A and Suryanarayanan R 1995 Physica C 249 262 Nicolas M, Mettout B and Sauv K 1995 Physica C 249 377 Nikolo M and Goldfarb R B 1989 Phys. Rev. B 39 6615 Nikolo M and Hermann A M 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 553 Nikolo M and Missey M J 1994 Physica C 235–240 2899 Noetzel R, Brinkmann M, Rex T, Bach H and Westerholt K 1996 Czech. J. Phys. 46 (suppl S3) 1629 Opagiste C et al 1993 Physica C 205 247 Opagiste C, Couach M, Khoder A, Triscone G, Jorda J L and Cohen-Adad M T 1994 Physica B 194–196 1809 Parvin E M, Singfield A, Vinen M F and Cox G F 1993 Supercond. Sci. Technol. 6 525 Pérez F, Obradors X, Fontcuberta J, Bozec X and Fert A 1994 Physica C 235–240 2941 —— 1996 Supercond. Sci. Technol. 9 161 Planinić P, Bašic I, Tonković M, Brničević N and Drobac D 1994 Physica C 235–240 919 Plecháček V and Gömöry F 1990 Solid State Commun. 73 349 Polichetti M, Pace S, Reggiani L and Testa A M 1996 4th Euro Ceramics vol 7, ed A Barone, D Fiorani and A Tampieri (Rome: Faenza) p 85 Polichetti M, Pace S, Saggese A, Testa A M, Federico C, Celani F, Bontet M and DiGioacchino D 1994 Physica C 235–240 3217 Porch A, Cooper J R, Zheng D N, Waldram J R, Campbell A M and Freeman P A 1993 Physica C 214 350 Prozorov R, Shaulov A, Wolfus Y and Yeshurun Y 1995 Phys. Rev. B 52 12 541 Puig T, Pont M and Muňoz J S 1992 Supercond. Sci. Technol. 5 S188 Qin M J, Ding S Y, Ren C, Yao X X, Fu Y X, Cai C B, Shi T S and Wang G Y 1996 Physica C 262 127 Ramakrishnan S, Sundaram S, Pandit R S and Girish Chandra 1985 J. Phys. E: Sci. Instrum. 18 650 Ravi S and Seshu Bai V 1994 Physica C 230 51 Ravinder Reddy R, Murchidhar M, Hari Babu V and Venugopal Reddy P 1995 Supercond. Sci. Technol. 8 101 Ren H T, Xia J A and Taylor K N R 1994 Physica C 219 221 Rillo C, Lera F, Badı́a A, Angurel L A, Bartolomé J, Palacio F, Navarro R and van Duyneveldt A J 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 1 Riseman T M et al 1994 Hyperfine Interact. 86 481 541 F Gömöry Rogacki K, Zeng X, Esquinazi P and Dabrowski B 1996 Czech. J. Phys. 46 (suppl S3) 1615 Rollins R W, Küpfer H and Gey W 1974 J. Appl. Phys. 45 5392 Saha S and Das B K 1993 Supercond. Sci. Technol. 6 840 Samarappuli S, Schilling A, Chernikov M A and Ott H R 1992 Physica C 201 159 Sanchez A 1994 Physica C 225 136 Sanchez A and Chen D X 1991 Magnetic Susceptibility of Superconductors and Other Spin Systems ed R A Hein, T L Francavilla and D H Liebenberg (New York: Plenum) p 251 Sargánková I, Diko P, Tweed J D, Anderson C A and Brown N M D 1996 Supercond. Sci. Technol. 9 688 Sauv K, Nicolas M, Nguyen van Hong C, Dubon A, Legeay P and Kandel L 1993 Supercond. Sci. Technol. 6 327 Savvides N, Katsaros A, Andrikidis C and Müller K H 1992 Physica C 197 267 Sekula S T 1971 J. Appl. Phys. 42 16 Senoussi S 1992 J. Physique. III 2 R1041 Senoussi S, Aguillon C and Manuel P 1991 Physica C 175 202 Senoussi S, Hadjoudj S, Maury R and Fert A 1990 Physica C 165 364 Seow W S, Doyle R A, Johnson J D, Kumar D, Somekh R, Walker D J C and Campbell A M 1995 Physica C 241 71 Shi K X, Zeng Z Y, Ding S Y, Ren C, Yao X X, Fu Y X and Cai C B 1995 Physica C 254 318 Shoenberg D 1937 Proc. Cambridge Philos. Soc. 33 559 Smythe W R 1950 Static and Dynamic Electricity (New York: McGraw-Hill) Spanier J and Oldham K B 1987 Atlas of Functions (Washington, DC: Hemisphere) Stoppard O and Gugan D 1995 Physica C 241 375 Sun H B, Russel G J and Taylor K N R 1995a Physica C 241 219 —— 1995b Physica C 243 139 Sun H B, Taylor K N R and Russel G J 1994a Physica C 227 55 Sun J Z, Sharen M J, Boutne L C and Schrieffer J R 1991 Phys. Rev. B 44 5275 Sun Y R, Thompson J R, Kerchner H R, Christen D K, Parathaman M and Brynestad J 1994b Phys. Rev. B 50 3330 Takács S and Campbell A M 1988 Supercond. Sci. Technol. 1 53 Takács S and Gömöry F 1990 Supercond. Sci. Technol. 3 94 —— 1995 Studies of High Temperature Superconductors vol 14, ed A V Narlikar (New York: Nova) p 407 542 Tatara G, Matsuda M, Katsumata K, Kambara T, Awaya Y, Mitamura T, Terasawa M, Endoh Y, Yamada K and Hosoya S 1996 Czech. J. Phys. 46 (suppl S3) 1555 Ullmaier H 1975 Irreversible Properties of Type II Superconductors (Berlin: Springer) van der Beek C J, Indenbom M, D’Anna G and Benoit W 1996 Physica C 258 105 van der Beek C J, Konczykowski M, Vinokur V M, Li T W, Kes P and Crabtree G W 1995 Phys. Rev. Lett. 74 1214 van der Beek C J, Nieuwenhuys G J, Kes P H, Schnack H G and Griessen R 1992 Physica C 197 320 van der Klein C A M 1972 PhD Thesis Leiden University, Leiden Vinokur V M, Feigelman M V and Geshkenbein V B 1991 Phys. Rev. Lett. 67 915 Vlakhov E S, Nenkov K A, Ciszek M, Zaleski A and Dimitriev Y B 1994 Physica C 225 149 Vo N V, Liu H K and Dou S X 1996 Supercond. Sci. Technol. 9 104 Vuong N V 1996 Physica C 258 77 Waldmann O, Lichtschlag G, Steinmeyer F, Talalaevski A and Müller P 1966 Czech. J. Phys. 46 (suppl S3) 1589 Welch D O, Suenaga M, Xu Y and Ghosh A R 1990 Advances in Superconductivity II ed T Ishiguro and K Kajimura (Tokyo: Springer) p 655 Widder W, Bauernfeind L, Stebani M and Braun H F 1995 Physica C 249 78 Wu C J and Tseng T Y 1996 Physica C 259 61 Xia J A, Ren H T, Zhao Y, Andrikidis C, Liu H K and Dou S X 1993 Supercond. Sci. Technol. 6 315 Yamaguchi Y, Ha D H, Iga F and Nishiharu Y 1994 Physica C 228 141 Yang M, Kao Y H, Xiu Y and Wong K W 1994a Phys. Rev. B 50 13 653 Yang Y, Yi Z, Beduz C and Scurlock R G 1994b Physica C 235–240 3039 Yeshurun Y and Malozemoff A P 1988 Phys. Rev. Lett. 60 2202 Zeisberger M, Klupsch T, Gawalek W and Habisreuther T 1994 Proc. 7th Int. Workshop on Critical Currents in Superconductors ed H W Weber (Singapore: World Scientific) p 145 Zeldov E, Amer N M, Koren G, Gupta A 1990 Appl. Phys. Lett. 56 1700 Ziese M, Esquinazi P, Braun H F and Sherman A B 1994 Proc. 7th Int. Workshop on Critical Currents in Superconductors ed H W Weber (Singapore: World Scientific) p 189