cond-mat0312199/w2s.tex
1: \documentclass[aps,prl,twocolumn,groupedaddress]{revtex4}
2: 
3: \usepackage{graphicx}
4: 
5: \bibliographystyle{apsrev}
6: 
7: \begin{document}
8: 
9: \title{Weak- to strong pinning crossover}
10: 
11: \author{G.\ Blatter, V.B.\ Geshkenbein, and J.A.G.\ Koopmann}
12: 
13: \affiliation{Theoretische Physik, ETH-H\"onggerberg, CH-8093 Z\"urich,
14: Switzerland}
15: 
16: \date{\today}
17: 
18: \begin{abstract}
19:   Material defects in hard type II superconductors pin the
20:   flux lines and thus establish the dissipation-free current
21:   transport in the presence of a finite magnetic field.
22:   Depending on the density and pinning force of the defects
23:   and the vortex density, pinning is either weak-collective
24:   or strong. We analyze the weak- to strong pinning crossover
25:   of vortex matter in disordered superconductors and discuss
26:   the peak effect appearing naturally in this context.
27: \end{abstract}
28: 
29: \maketitle
30: 
31: Pinning of vortices by material defects is crucial in establishing
32: the defining property of a superconductor, its ability to
33: transport electrical current without dissipation. Collective
34: pinning theory \cite{LO_79}, describing the concerted action of
35: many weak pins on the vortex system, is playing a central role in
36: our understanding of this complex statistical mechanics problem
37: \cite{review}. On the other hand, first attempts describing flux
38: pinning go back to Labusch \cite{Lab_69}, who described the
39: interaction between vortices and strong pinning centers which
40: introduce large (plastic) deformations in the vortex system. In
41: this letter, we describe how these two theories relate to one
42: another; given the density $n_\mathrm{p}$ and force $f_\mathrm{p}$
43: of pinning centers, as well as the vortex density $n_\mathrm{v}=
44: 1/a_0^2$, we identify the regimes where individual vortex lines
45: and the bulk vortex lattice are pinned by the collective action of
46: many weak pins or by the independent action of strong pins, see
47: Fig.\ \ref{fig:w2s_1}. We naturally recover the peak effect
48: \cite{pippard_69} described in the work of Larkin and Ovchinnikov 
49: \cite{LO_79} and establish its formal relation to the Landau 
50: theory of phase transitions.
51: 
52: In a type II superconductor, the field (${\bf B}$) induced vortices
53: subject to a current flow ${\bf j}$ experience the Lorentz force
54: density ${\bf F}_{\rm \scriptscriptstyle L} = {\bf j}\wedge{\bf B}/c$
55: and the resulting vortex motion leads to dissipation. The
56: superconducting response is resurrected through material
57: inhomogeneities pinning the vortices at energetically favorable
58: locations. The pinning force density ${\bf F}_\mathrm{pin}$ defines a
59: critical current density $j_c = c F_\mathrm{pin}/B$ below which the
60: current can flow free of dissipation. Usually, this critical current
61: density is considerably reduced with respect to the depairing current
62: density $j_0 \sim c H_c / 4 \pi \lambda \sim c \varepsilon_0/
63: \Phi_0\xi$; here, $H_c = \Phi_0/2\sqrt{2} \pi\lambda\xi$ is the
64: critical magnetic field, $\lambda$ and $\xi$ denote the penetration
65: depth and the coherence length $\xi$, respectively, $\Phi_0 = h c/2e$
66: is the flux unit, and $\varepsilon_0 = (\Phi_0/4\pi \lambda)^2$ is the
67: (line) energy scale. Below, we focus on the most generic situation of
68: isotropic superconductors and ignore effects due to thermal
69: fluctuations.
70: %
71: \begin{figure}
72:    \includegraphics[scale=0.318]{w2s_1.eps}
73:    \caption[]{Pinning diagram delineating the 
74:    various pinning regimes involving collective-
75:    versus individual pinning and 1D-line- versus 3D-bulk
76:    pinning ($f_{\rm Lab}$ denotes the Labusch force):
77:    3D wcp -- bulk weak collective pinning, 1D cp --
78:    collective line pinning, 1D sp -- strong line pinning,
79:    3D sp -- bulk strong pinning. The insets illustrate
80:    the weak- and strong (plastic) distortions of the
81:    vortex system characterizing the pinning regimes. 
82:    Lines refer to crossovers.}
83:    \label{fig:w2s_1}
84: \end{figure}
85: %
86: 
87: When pinning is strong \cite{Lab_69,LO_79,LO_NSC_86}
88: defects act individually and the pinning force density
89: $F_\mathrm{p}$ is linear in the density $n_\mathrm{p}$ and
90: average pinning force $\langle f_\mathrm{pin} \rangle$ of
91: defects. The classic arguments characterizing strong pinning
92: go back to Labusch \cite{Lab_69}, see also \cite{LO_79,LO_NSC_86}:
93: A strong pinning defect induces plastic deformations in the
94: vortex lattice \cite{Brandt_86gen,OvIv_91,Schoenenberger_96})
95: and the energy landscape $e_\mathrm{pin}({\bf R})$ becomes
96: multi-valued in the vortex position ${\bf R}$, see 
97: Fig.\ \ref{fig:w2s_2}. The averaging over defect locations 
98: then has to account for the preparation of the system.
99: We concentrate on the critical current density and
100: thus search for the force against drag; the vortex 
101: position then is parametrized through the two-component
102: drag parameter ${\bf R}_\mathrm{d}$ fixing the position 
103: of the unperturbed lattice with respect to the defect. 
104: Dragging the system along the $x$-direction, we express 
105: the drag force $-\partial_x e_\mathrm{pin} (x,y)$ 
106: integrated along $x$ through the jump $\Delta 
107: e_\mathrm{pin}(y)>0$ in the pinning energy and 
108: average over `impact parameters'~$y$,
109: %
110: \[
111:    \langle f_\mathrm{pin} \rangle
112:    =-\int_0^{L_x}\!\!\!\! d x \int_0^{L_y} \!\!\!\!d y
113:    \frac{\partial_x e_\mathrm{pin}(x,y)}{L_x L_y}
114:    =-\int_0^{a_0}\!\!\! dy
115:    \frac{\Delta e_\mathrm{pin}(y)}{a_0 \tilde{a}(y)},
116: \]
117: %
118: where $\tilde{a}$ denotes the distance between periodic
119: jumps \cite{note}. For moderately strong pins with 
120: deformations not exceeding the lattice constant we 
121: have $\tilde{a} \approx a_0$ and assuming a maximal 
122: transverse trapping distance $t_\perp$ along the 
123: $y$-axis we obtain the mean pinning force
124: %
125: \begin{equation}
126:    \langle f_\mathrm{pin} \rangle
127:    \approx -\frac{t_\perp}{a_0^2} \Delta e_\mathrm{pin}(0)
128:    \approx -\frac{t_\perp t_\parallel}{a_0^2} f_\mathrm{p}
129:    \approx -\frac{S_\mathrm{trap}}{a_0^2} f_\mathrm{p},
130:    \label{fav}
131: \end{equation}
132: %
133: with the jump $\Delta e_\mathrm{pin}(0) \approx t_\parallel
134: f_\mathrm{p}$ expressed via the typical impurity force
135: $f_\mathrm{p}$ and the bistability range $t_\parallel$ of
136: $e_\mathrm{pin}(x,0)$; the product $t_\perp t_\parallel$
137: defines the trapping area $S_\mathrm{trap}$ associated with the
138: strong pin \cite{OvIv_91}. The low impurity concentration 
139: $n_\mathrm{p}$ implies non-interfering defects and we obtain 
140: a critical current density $j_c=-c n_\mathrm{p}\langle 
141: f_\mathrm{pin}\rangle/B$ linear in $n_\mathrm{p}$,
142: %
143: \begin{equation}
144:    j_c
145:    \approx ({c}/{B}) n_\mathrm{eff} f_\mathrm{p}
146:    \approx j_0 [n_\mathrm{p} \xi S_\mathrm{trap}]\,
147:    {f_\mathrm{p}}/{\varepsilon_0},
148:    \label{j_sp}
149: \end{equation}
150: %
151: with the effective impurity density $n_\mathrm{eff} = n_\mathrm{p}
152: (S_\mathrm{trap}/a_0^2)$.
153: %
154: \begin{figure}
155: \includegraphics[scale=0.383]{w2s_2e.eps}
156: \caption[]{Energy landscape $e_\mathrm{pin}$ and pinning force
157: $f_\mathrm{pin}$ versus displacement $R_\mathrm{d}$ of the vortex
158: lattice relative to the defect; for weak pinning these are
159: single-valued functions in $R_\mathrm{d}$ ({\itshape dashed
160: lines}), while strong pinning produces plastic deformations and
161: renders $e_\mathrm{pin}$, $f_\mathrm{pin}$ multi-valued ({\itshape
162: solid lines}; {\itshape dotted lines} indicate unstable branches).
163: Bottom right: Trapping geometry
164: (top view) for a circularly symmetric situation.}
165: \label{fig:w2s_2}
166: \end{figure}
167: %
168: 
169: In order to derive a quantitative criterion for the appearance of
170: strong pinning, we consider a single defect at the origin
171: with a pinning potential $e_\mathrm{p} ({\bf r})$. Such a defect
172: acts on the vortex system to produce a pinning energy density
173: $E_\mathrm{p}({\bf r},{\bf u}) = \sum_{\nu}e_\mathrm{p}
174: ({\bf r}) \delta^2({\bf R}-{\bf R}_\nu -{\bf u}({\bf R}_\nu,z))$,
175: with vortices positioned at ${\bf R}_\nu+{\bf u}({\bf R}_\nu,z)$,
176: ${\bf R}_\nu$ the equilibrium positions and ${\bf u}$ the
177: displacement field. The latter follows from the solution
178: of the implicit equation (${\bf r}_\nu=({\bf R}_\nu,z)$)
179: %
180: \begin{eqnarray}
181:    &&{u}_\alpha({\bf r}_\nu)
182:    = \!\! \int \! d^3 r^\prime\,
183:    G_{\alpha\beta}({\bf r}_\nu-{\bf r}^\prime)
184:    [-\partial_{u_\beta} {E}_\mathrm{p}]({\bf r}^\prime,{\bf u}^\prime)
185:    \nonumber \\
186:    &&\quad =
187:    \sum_{\nu^\prime} \int dz^\prime
188:    G_{\alpha\beta}({\bf r}_\nu-{\bf r}_{\nu}^\prime)
189:    {f}_{\mathrm{p},\beta}({\bf R}_{\nu}^\prime
190:    +{\bf u}({\bf r}_{\nu}^\prime),z^\prime)
191:    \nonumber \\
192:    &&\quad = G_{\alpha\beta}({\bf R}_\nu-{\bf R}_\mathrm{d},0)
193:    {f}_{\mathrm{p},\beta}({\bf R}_\mathrm{d}+
194:    {\bf u}({\bf R}_\mathrm{d},0),0),
195:    \label{u_int_f}
196: \end{eqnarray}
197: %
198: with $G_{\alpha\beta}(\bf r)$ the elastic Green's function and
199: ${\bf f}_\mathrm{p} = -\nabla_u e_\mathrm{p} ({\bf u})$ the
200: pinning force of the defect. In the last equation we have assumed
201: a moderately strong pinning potential (pinning one vortex at most)
202: of range much smaller than the lattice constant $a_0$ and have
203: chosen ${\bf R}_\mathrm{d}$ as the distance to the vortex closest
204: to the defect \footnote{Effects of very strong pinning 
205: (overdrag into the next cell and multi-vortex pinning) 
206: are not considered here.}.
207: Evaluating (\ref{u_int_f}) for ${\bf r}_\nu =
208: ({\bf R}_\mathrm{d},0)$, we arrive at the self-consistency equation
209: (note that $G_{\alpha\beta} ({\bf r}=0)$ is diagonal)
210: %
211: \begin{equation}
212:    {\bf u}({\bf R},0) \approx  \bar{C}^{-1} {\bf f}_\mathrm{p}
213:    ({\bf R}+{\bf u}({\bf R},0),0),
214:    \label{self}
215: \end{equation}
216: %
217: with the effective elastic constant $\bar{C}^{-1} = \int 
218: d^3 k/(2\pi)^3$ $G_{xx}({\bf k})$. 
219: For weak pinning the displacement ${\bf u}$ is small and the
220: solution ${\bf u}({\bf R},0) \approx {\bf f}_\mathrm{p} ({\bf
221: R})/\bar{C}$ is unique. Strong pinning, however, produces
222: multi-valued functions ${\bf u}({\bf R},0)$ and
223: ${e}_\mathrm{pin}({\bf R})$, cf.\ Fig.\ \ref{fig:w2s_2}. The
224: solution of (\ref{self}) turns multi-valued as the displacement
225: collapses when $\partial_R u \rightarrow \infty$. Assuming a
226: defect symmetric in the plane, $e_\mathrm{p}({\bf R},z) =
227: e_\mathrm{p} (R,z)$, and dragging the lattice through the defect
228: center along the $x$-axis, we find $u' = f'_\mathrm{p}(x+u)$
229: $[\bar{C}-f'_\mathrm{p}(x+u)]^{-1}$ (note that $x>0$ implies $u<0$) and
230: arrive at the (Labusch) criterion \cite{Lab_69} in the form
231: %
232: \begin{equation}
233:    \partial_x f_\mathrm{p} = -\partial_x^2 e_\mathrm{p}
234:    = \bar{C};
235:    \label{sp_crit}
236: \end{equation}
237: %
238: hence, in order to produce strong pinning the (negative) curvature
239: of the pinning energy $e_\mathrm{p}$ has to overcompensate the
240: lattice elasticity (the Labusch criterion involves the maximal
241: negative curvature above the inflection point). Note that the
242: Labusch criterion tests an individual pinning center and
243: classifies it as a weak or strong one.
244: 
245: When pinning is weak, the elastic forces dominate over the pinning
246: forces and the defects compete; we then are faced with the problem
247: of the statistical summation of individual pinning forces. For
248: weak pins the average $\langle f_\mathrm{pin}\rangle$ vanishes and
249: pinning is due to fluctuations in the pinning force density: the
250: forces of the competing pins (with pinning force $f_\mathrm{p}$,
251: density $n_\mathrm{p}$, and extension $r_\mathrm{p} \sim \xi$) add
252: up randomly and produce the pinning energy
253: %
254: \begin{equation}
255:    \langle {\cal E}_\mathrm{pin}^2 (V) \rangle^{1/2}
256:    \approx \bigl[f_\mathrm{p}^2 n_\mathrm{p}
257:    (\xi/a_0)^2 V\bigr]^{1/2}
258:    \, \xi;
259:    \label{E2pin_g}
260: \end{equation}
261: %
262: only vortex cores are pinned by the disorder, hence the factor
263: $(\xi/a_0)^2$. Within weak collective pinning theory the sublinear
264: growth of $\langle {\cal E}_\mathrm{pin}^2 (V) \rangle^{1/2}$ with
265: volume turns linear when the displacement $u$ increases beyond the
266: scale $\xi$ of the pinning potential, thus defining the collective
267: pinning volume $V_c$. Each volume of size $V_c$ is pinned
268: independently with a pinning energy $U_c = \langle {\cal
269: E}_\mathrm{pin}^2 (V_c) \rangle^{1/2}$ and we obtain a proper
270: pinning force density
271: %
272: \begin{equation}
273:    F_\mathrm{pin}
274:    \sim {U_c}/{V_c \,r_\mathrm{p}}
275:    \sim \bigl({f_\mathrm{p}^2
276:    n_\mathrm{p}(\xi/a_0)^2}/{V_c}\bigr)^{1/2};
277:    \label{f_pin_g}
278: \end{equation}
279: %
280: balancing this pinning force density against the Lorentz force
281: density $j B /c$ we find a finite critical current density $j_c
282: \sim c F_\mathrm{pin}/B$. The remaining task is the determination
283: of the collective pinning volume $V_c$; its calculation is
284: complicated by the dispersion and anisotropy of the vortex
285: lattice, see below and Ref.\ \cite{review} for a detailed
286: discussion.
287: 
288: It is instructive to compare the weak- and strong pinning schemes
289: and their dependence on dimensionality, particularly in the limit
290: of a small defect density $n_\mathrm{p}$ (in the following, we
291: assume pinning sites characterized by their force $f_\mathrm{p}$
292: and extension $\xi$). An isolated vortex line (1D) is always
293: subject to strong pinning forces as the effective elastic
294: coefficient $\bar{C}$ vanishes due to the diverging integral.
295: At the same time, the deformation of the line due to
296: the pins is large and we cannot ignore their mutual competition.
297: Comparing the elastic energy $\varepsilon_0 \xi^2/L_c$ and the
298: pinning energy $U_c = (f_\mathrm{p}^2 n_\mathrm{p} L_c
299: \xi^2)^{1/2} \xi$, we find the collective pinning length $L_c \sim
300: (\varepsilon_0^2 /f^2_\mathrm{p} n_\mathrm{p})^{1/3}$ and a
301: critical current density
302: %
303: \begin{equation}
304:    j_c \sim j_0 \bigl({n_\mathrm{p}\xi^3
305:    f^2_\mathrm{p}}/{\varepsilon_0^2}\bigr)^{2/3}.
306:    \label{jc_1d_cp}
307: \end{equation}
308: %
309: This result is valid as long as many pins compete within the
310: volume $\xi^2 L_c$; the condition $n_\mathrm{p} \xi^2 L_c > 1$
311: defines the lower limit $\bar{n}_{\rm \scriptscriptstyle 1D} \sim
312: f_\mathrm{p} /\varepsilon_0 \xi^3 < n_\mathrm{p}$ where the
313: critical current density assumes the value $\bar{j_c} \sim j_0
314: (f_\mathrm{p}/\varepsilon_0)^2$.
315: 
316: For small densities $n_\mathrm{p} < \bar{n}_{\rm
317: \scriptscriptstyle 1D}$ the pins acts individually and we
318: determine $j_c$ from the force balance $(\Phi_0/c) j_c l u \sim
319: \Delta e_\mathrm{pin} \sim f_\mathrm{p} u$, with $u\sim
320: t_\parallel$ the displacement directed along the force. The
321: displacement $u$ and the length $l$ between two subsequent pins
322: fixing the vortex derives from an analysis of the pinned vortex
323: geometry, see Fig.\ \ref{fig:w2s_1} inset: integrating the force
324: equation $\varepsilon_0 {\partial_z^2 u} = f(z)$ (with $f(z)$ the
325: force per unit length acting on the line) over one pinning center,
326: we find the distortion angle $\theta =\partial_z u \sim u/l \sim
327: f_\mathrm{p}/\varepsilon_0$ \cite{alternative}. A vortex segment
328: of length $l$ deformed by the angle $\theta$ in the direction of
329: the driving force encounters $\theta l^2 \xi n_\mathrm{p}$ defects
330: (with the trapping length $t_\perp \sim \xi$). At the distance
331: $l$, this number is unity, hence $l \sim
332: \sqrt{\varepsilon_0/f_\mathrm{p} n_\mathrm{p} \xi}$ and we obtain
333: the critical current density
334: %
335: \begin{equation}
336:    j_c \sim j_0
337:    \bigl({n_\mathrm{p}
338:    \xi^3 f^3_\mathrm{p}}/{\varepsilon_0^3}\bigr)^{1/2}.
339:    \label{jc_1d_sp}
340: \end{equation}
341: %
342: At the crossover density $\bar{n}_{\rm\scriptscriptstyle 1D} \sim
343: f_\mathrm{p}/ \varepsilon_0 \xi^3$ the critical current density
344: matches up with the weak pinning result; also, the displacement $u
345: \sim l f_\mathrm{p}/\varepsilon_0$ is of order $\xi$ at the
346: crossover density $\bar{n}_{\rm \scriptscriptstyle 1D}$ and hence
347: matches the displacement field relevant in the collective pinning
348: scenario. Note that collective pinning (\ref{jc_1d_cp}) dominates
349: over the strong pinning (\ref{jc_1d_sp}) at \textit{large}
350: densities $n_\mathrm{p} > \bar{n}_{\rm\scriptscriptstyle 1D}$.
351: 
352: For the vortex lattice (3D bulk pinning; we assume $a_0<\lambda$)
353: the Labusch criterion (\ref{sp_crit}) offers a distinction between
354: weak and strong pinning centers; using the Green's function for
355: the vortex lattice (see, e.g., \cite{review}) we find $\bar{C} 
356: \sim \varepsilon_0/a_0$.
357: According to (\ref{sp_crit}) a pinning center changes from weak to
358: strong at $f_\mathrm{p} \sim f_\mathrm{Lab} \equiv \varepsilon_0
359: \xi/a_0$. We first review the weak pinning situation with
360: $f_\mathrm{p} < f_\mathrm{Lab}$ (where necessary, we encode
361: quantities in this regime with a superscript
362: `$^{\scriptscriptstyle <}$'). The determination of the anisotropic
363: collective pinning volume $V_c = R_c^2 L_c^\mathrm{b}$ has to
364: account for the dispersion in the tilt modulus at intermediate
365: scales $a_0 < R_c < \lambda$, see Ref.\ \cite{review}, and
366: produces the results
367: %
368: \begin{eqnarray}
369:    &&j_c \sim j_0 \,\frac{\xi^2}{a_0^2}
370:    \left[\frac{a_0}{L_c}\right]^\nu
371:    e^{-2c\left[L_c/a_0\right]^3} , \quad R_c < \lambda,
372:    \label{sbp}\\
373:    &&j_c \sim j_0 \,\frac{\xi^2}{\lambda^2}
374:    \left[\frac{a_0}{L_c}\right]^6,
375:    \quad R_c > \lambda;
376:    \label{lbp}
377: \end{eqnarray}
378: %
379: we have made use of the single vortex pinning parameter $L_c/a_0
380: \sim (f_\mathrm{Lab}^2/f_\mathrm{p}^2 a_0 \xi^2
381: n_\mathrm{p})^{1/3}$. The numericals $c$ and $\nu$ follow from a
382: 2-loop renormalization group analysis \cite{wagner,book}. These
383: results are valid as long as many (competing) pins are active in
384: the volume $V_c$, $n_\mathrm{p}(\xi^2/ a_0^2) V_c > 1$. For
385: $f_\mathrm{p}< f_\mathrm{Lab}$ this condition is violated in the
386: large $n_\mathrm{p}$ limit. However, with increasing pinning
387: density $n_\mathrm{p}$, the collective pinning radius $R_c$
388: decreases, first below $\lambda$ at $\bar{n}_\lambda \sim
389: f_\mathrm{Lab}^2/f_\mathrm{p}^2 a_0\xi^2 \ln(\lambda/a_0)$
390: delineating the dispersive regime, and then below $a_0$ at
391: $\bar{n}_{a_0}^{\scriptscriptstyle <} \sim
392: f_\mathrm{Lab}^2/f_\mathrm{p}^2a_0\xi^2$ where the condition
393: $n_\mathrm{p}(\xi^2/ a_0^2) V_c > 1$ is still fulfilled. At the
394: crossover density $\bar{n}_{a_0}^{\scriptscriptstyle <}$ the 3D
395: weak collective pinning crosses over to the 1D weak collective
396: pinning result (\ref{jc_1d_cp}).
397: 
398: Turning to strong pinning $f_\mathrm{p} > f_\mathrm{Lab}$ (encoded
399: with a superscript `$^{\scriptscriptstyle >}$') we start at high
400: densities; as the Labusch criterion is not effective in 1D, the
401: system remains collectively pinned for $n_\mathrm{p}> \bar{n}_{\rm
402: \scriptscriptstyle 1D}$ and crosses over to 1D strong pinning as
403: $n_\mathrm{p}$ drops below $\bar{n}_{\rm \scriptscriptstyle 1D}$.
404: Decreasing $n_\mathrm{p}$ further, the pinning distance $l \sim
405: a_0\sqrt{(f_\mathrm{Lab}/ f_\mathrm{p})/n_\mathrm{p}a_0\xi^2}$
406: increases beyond $a_0$ as $n_\mathrm{p}$ decreases below
407: $\bar{n}_{a_0}^{\scriptscriptstyle >} \sim (f_\mathrm{Lab} /
408: f_\mathrm{p})/a_0\xi^2$ and the system enters the 3D strong
409: pinning regime, see Fig.\ \ref{fig:w2s_1}. The calculation of the
410: mean pinning force density $F_\mathrm{pin} \sim n_\mathrm{p}
411: \langle f_\mathrm{pin}\rangle$ proceeds along the lines 
412: discussed above and involves the trapping area $S_\mathrm{trap} 
413: \sim t_\perp t_\parallel$ with $t_\perp \sim \xi$ and
414: $t_\parallel \sim u \sim f_\mathrm{p}/\bar{C}$; we obtain the
415: force density $F_\mathrm{pin} \sim n_\mathrm{p} (\xi/a_0)
416: f^2_\mathrm{p} /\varepsilon_0$ and a critical current density
417: %
418: \begin{equation}
419:    j_c \sim j_0\,
420:    a_0 \xi^2 n_\mathrm{p}
421:    \frac{f^2_\mathrm{p}}{\varepsilon_0^2}
422:    \sim j_0\, \frac{\xi^2}{a_0^2}
423:    n_\mathrm{p} a_0 \xi^2 \,
424:    \frac{f^2_\mathrm{p}}{f_\mathrm{Lab}^2}.
425:    \label{jc_3d_sp}
426: \end{equation}
427: %
428: The bulk strong pinning result (\ref{jc_3d_sp}) smoothly
429: transforms into the 1D expression (\ref{jc_1d_sp}) at
430: $\bar{n}_{a_0}^{\scriptscriptstyle >}$ where $l \sim a_0$. On the
431: contrary, the strong pinning expression (\ref{jc_3d_sp}) does not
432: match up with the bulk weak collective pinning results (\ref{sbp})
433: and (\ref{lbp}) at $f_\mathrm{p} = f_\mathrm{Lab}$ (we concentrate
434: on low impurity densities with $n_\mathrm{p} a_0\xi^2 < 1$, cf.\
435: Fig.\ \ref{fig:w2s_1}). However, we have to keep in mind that our
436: rough derivation of the strong pinning result (\ref{jc_3d_sp})
437: breaks down on approaching the critical force $f_\mathrm{Lab}$.
438: Indeed, since the displacement field ${\bf u}({\bf r})$ turns
439: single valued below $f_\mathrm{Lab}$, strong pinning vanishes
440: altogether (with a power $[f_\mathrm{p}-f_\mathrm{Lab}]^2$, see
441: (\ref{jc_3d_sp_gen})) and pinning survives only in the form of
442: weak collective pinning due to fluctuations in the impurity
443: density. Within the approximative scheme adopted here the sharp
444: rise of the critical current density at $f_\mathrm{p} >
445: f_\mathrm{Lab}$ is encoded in a jump $j_c|_\mathrm{sp}/
446: j_c|_\mathrm{wcp} \sim (\lambda^2/a_0^2) /n_\mathrm{p} a_0 \xi^2 >
447: 1$ for $n_\mathrm{p} < \bar{n}_\lambda$ ($\sim \exp[2c/
448: n_\mathrm{p} a_0 \xi^2]$ for $n_\mathrm{p} > \bar{n}_\lambda$).
449: 
450: The crossover from strong to weak pinning at the Labusch condition
451: (\ref{sp_crit}) can be analyzed within a Landau type expansion: We
452: define the free energy functional $e_\mathrm{pin}({\bf u},{\bf
453: R}_\mathrm{d}) = \bar{C} u^2/2 + e_\mathrm{p}({\bf R}_\mathrm{d}
454: +{\bf u})$ from which the self-consistency equation (\ref{self})
455: follows by variation. Note that the derivative $-\partial_x
456: e_\mathrm{pin} = f_{\mathrm{p},x}({\bf R}_\mathrm{d}+{\bf u})$ 
457: provides the force along $x$ acting on a vortex separated from 
458: the defect by ${\bf R}_\mathrm{d}$ and deformed by ${\bf u}$, 
459: c.f.\ Fig.\ 1; it is this force which has to be averaged 
460: over in the definition of $\langle f_\mathrm{pin}\rangle$.
461: 
462: We first concentrate on the trajectory ${\bf
463: R}_\mathrm{d} = (x,0)$ with ${\bf u} = (u,0)$. The curvature
464: $e_\mathrm{p}^{\prime \prime}(u)$ relevant in (\ref{sp_crit})
465: assumes a maximal negative value; we denote the corresponding
466: location and value by $u_\kappa$ and $-\kappa$, respectively.
467: Next, we expand the curvature around $u_\kappa$,
468: $e_\mathrm{p}^{\prime\prime}(u) \approx -\kappa + \alpha
469: (u-u_\kappa)^2/2$; integrating in $u$ and combining with 
470: the elastic term $\bar{C} u^2/2$ we arrive at the expansion
471: %($u_\kappa = \xi$, $\kappa = e_0/2\xi^2$, and $\alpha = 3
472: %e_0/\xi^4$ for the Lorentzian potential $e_\mathrm{p,
473: %{\scriptscriptstyle L}}(u) = -e_0/(1+u^2/\xi^2)$)
474: %
475: %\begin{eqnarray}
476: %   e_\mathrm{pin}[\upsilon,x]
477: %   &\approx& \bar{C}\,(\upsilon-x)^2/2
478: %   -\epsilon+\nu\,(\upsilon-u_\kappa)\nonumber\\
479: %   &-&\kappa(\upsilon-u_\kappa)^2/2
480: %   +\alpha(\upsilon-u_\kappa)^4/24,
481: %   \label{L_expansion}
482: %\end{eqnarray}
483: %
484: %where we have shifted the coordinate $\upsilon = u+x$ for
485: %convenience.
486: %%($\epsilon = e_0/2$ and $\nu = e_0/2\xi$ for the
487: %%Lorentzian potential).
488: %In order to find the regime of bistability in the drag parameter
489: %$x$, we analyze the derivative $\partial_{\upsilon} e_\mathrm{pin}
490: %[\upsilon,x]$ and find the extreme points $x_\pm$ bounding the
491: %interval with three zeroes $\partial_{\upsilon}e_\mathrm{pin}
492: %(\upsilon) = 0$; they are defined by the joint conditions
493: %$\partial_{\upsilon} e_\mathrm{pin}(\hat{\upsilon}_\pm,x_\pm) = 0$
494: %and $\partial_{\upsilon}^2 e_\mathrm{pin} (\hat{\upsilon}_\pm
495: %,x_\pm) = 0$. From the latter condition we find the solution
496: %$\hat{\upsilon}_\pm - u_\kappa \approx \pm \sqrt{2\bar{C}/\alpha}
497: %\, (\kappa/\bar{C}-1)^{1/2}$, while the former provides us with
498: %the result $x_\pm - u_\kappa \approx \nu/\bar{C} \mp (2/3)
499: %\sqrt{2\bar{C}/\alpha}\, (\kappa/\bar{C} -1)^{3/2}$. The
500: %displacement $\upsilon_-(x)$ ($\upsilon_+(x)$) is associated with
501: %the pinned (unpinned) state of the vortex and extends from
502: %$\check{\upsilon}_-$ to $\hat{\upsilon}_-$ (from
503: %$\hat{\upsilon}_+$ to $\check{\upsilon}_+$) where the boundaries
504: %$\check{\upsilon}_\pm - u_\kappa = \pm 2 \sqrt{2\bar{C}/\alpha}\,
505: %(\kappa/\bar{C}-1)^{1/2}$ follow from the condition
506: %$\partial_{\upsilon} e_\mathrm{pin}(\check{\upsilon}_\pm,x_\mp) =
507: %0$. The total jump $\Delta e_\mathrm{pin}$ is obtained from the
508: %(equal) jumps in $e_\mathrm{pin}[\upsilon,x]$ at $x_\pm$, $\Delta
509: %e_\mathrm{pin} = 2[e_\mathrm{pin}[\hat{\upsilon}_-,x_-]
510: %-e_\mathrm{pin}[\check{\upsilon}_+,x_-]]$,
511: %%
512: %\begin{equation}
513: %   \Delta e_\mathrm{pin} = 9 (\bar{C}^2/\alpha)
514: %   \bigl(\kappa/\bar{C}-1\bigr)^2.
515: %   \label{d_e}
516: %\end{equation}
517: %
518: %
519: \begin{eqnarray}
520:    e_\mathrm{pin}[u,x]
521:    &\approx& \bar{C}\,u^2/2
522:    +\nu\,(x+u-u_\kappa)\label{L_expansion}\\
523:    &-&\kappa(x+u-u_\kappa)^2/2
524:    +\alpha(x+u-u_\kappa)^4/24.
525:    \nonumber
526: \end{eqnarray}
527: %
528: This pinning energy maps to the free energy $e_\mathrm{mag}
529: [\phi,h] = \tau \phi^2/2 +\alpha \phi^4 /24-h\phi$ of a 
530: one-component magnet in a magnetic field \cite{LL} 
531: if we define the `order parameter' $\phi = x+u-u_\kappa$, 
532: the `temperature difference' $\tau = \bar{C} -\kappa$, and 
533: the `magnetic field' $h = \bar{C}(x-u_\kappa -\nu/\bar{C})$.
534: The `high-temperature' phase $\tau>0$ describing the 
535: paramagnet corresponds to weak pinning, while the
536: two ferromagnetic phases at `low temperatures' $\tau < 0$ 
537: stand for the pinned ($\phi <0$) and unpinned ($\phi >0$) 
538: states; the transition between these states is discontinuous 
539: and the associated coexistence regime extends over the `field' 
540: domain $|h| < h^\ast = (2/3\bar{C})\sqrt{2/\alpha}|\tau|^{3/2}$.
541: At $h^\ast$ the trapping/detrapping of the vortex from the defect
542: produces jumps $\Delta\phi = 3\sqrt{2/\alpha}|\tau|^{1/2}$
543: in the `order parameter', leading to jumps $\Delta e = \Delta 
544: e_\mathrm{pin}/2=(9/2\alpha)\, \tau^2$ in the energy. 
545:  
546: For finite `impact parameters' $y$ we have to determine
547: the trapping distance $t_\perp$. Assuming rotational
548: symmetry, the bistable regime is bounded by a circle
549: of radius $R = x^\ast = u_\kappa +\nu/\bar{C}$ and 
550: hence $t_\perp \approx 2 x^\ast$ (note that at $\tau = 0$ 
551: we have $h^\ast = 0$ but the critical drag parameter 
552: $x^\ast$ does not vanish). The (uncompensated)
553: trapping area determining the average pinning force $\langle
554: f_\mathrm{pin}\rangle$ is shown in Fig.\ \ref{fig:w2s_2};
555: combining the above results for the jump in pinning energy 
556: and the transverse trapping distance we find the averaged 
557: pinning force (c.f.\ (\ref{fav}))
558: %
559: \begin{equation}
560:    \langle f_\mathrm{pin} \rangle 
561:    \approx 18(u_\kappa +\nu/\bar{C})[\bar{C}-\kappa]^2/\alpha a_0^2.
562:    \label{fpin_Lab}
563: \end{equation}
564: %
565: Defining the individual force of (equal) pinning centers 
566: via $f_\mathrm{p} = \max_u[\partial_u f_\mathrm{p}](u) \xi =
567: \kappa \xi$ (then $f_\mathrm{Lab} = \bar{C} \xi$)
568: we can translate (\ref{fpin_Lab}) into an expression
569: for the critical current density $j_c$ extending the strong 
570: pinning result (\ref{jc_3d_sp}) to the vicinity of the
571: Labusch point,
572: %
573: \begin{equation}
574:    j_c \sim j_0\,(\xi^2/a_0^2)
575:    n_\mathrm{p} a_0 \xi^2 \,
576:    [f_\mathrm{p}/f_\mathrm{Lab}-1]^2.
577:    \label{jc_3d_sp_gen}
578: \end{equation}
579: %
580: Comparing with the weak pinning result (\ref{lbp}),
581: we note a sharp rise in the critical current density $j_c$
582: once the strong pinning force overcomes the weak 
583: pinning result \cite{LO_79}. With the small parameter 
584: $\delta = (a_0/\lambda) \sqrt{n_\mathrm{p} a_0 \xi^2} < 1$, 
585: this crossover appears above but still close to the Labusch 
586: point as $f_\mathrm{Lab} \propto \bar{C}$ decreases 
587: below $f_\mathrm{p}/(1+\delta)$.
588: 
589: Another remarkable result is the interpretation of the collective
590: pinning scenario in terms of the strong pinning picture; indeed,
591: summing over competing pins within the collective pinning volume
592: $V_c$ produces the corresponding critical Labusch force.
593: Quantitatively, we compare the force gradient $f' \sim
594: [n_\mathrm{p}(\xi^2/a_0^2)V]^{1/2} f_\mathrm{p}/\xi$ accumulated
595: within the (anisotropic) volume $V = L R^2 = (\lambda/a_0)R^3$
596: with the elastic parameter $\bar{C}(R) =\varepsilon_0 \lambda
597: R/a_0^3$ for smooth distortions on the scale $R>\lambda$
598: (non-dispersive regime) and apply the Labusch criterion
599: (\ref{sp_crit}). We then find the scale $R_c = \lambda
600: f_\mathrm{Lab}^2/f_\mathrm{p}^2 n_\mathrm{p} a_0 \xi^2$, where the
601: accumulated pinning force overcompensates the elastic force; this
602: length agrees with the 3D collective pinning length in the
603: non-dispersive regime \cite{review}. The resulting bistable
604: solutions are the signature of the alternative pinning valleys
605: which the collective pinning volume can select beyond the scale
606: $R_c$.
607: 
608: The above discussion sheds light on the general concept of pinning.
609: Pinning is absent in the rigid limit. A finite but large elasticity
610: (with $f_\mathrm{Lab} > f_\mathrm{p}$) allows only for weak
611: deformations and individual pins cannot hold the lattice as the
612: averaging over individual pinning forces produces a null result. Hence,
613: pinning is due only to fluctuations in the pinning forces and thus
614: collective. Reducing the elasticity, strong pinning defects appear when
615: $f_\mathrm{Lab}$ drops below $f_\mathrm{p}$; they pin the lattice
616: individually and strong pinning, linear in the defect density
617: $n_\mathrm{p}$, outperforms collective pinning. The important role
618: played by the curvature $e_\mathrm{p}''<0$ in the pinning potential is
619: an interesting topic for  numerical studies. The crossover between weak
620: collective and strong pinning can be realized in  experiments:
621: increasing the magnetic field  towards its critical value $H_{c_2}$
622: leads to a marked softening of the elastic moduli. The reduction in the
623: elastic moduli entails a decrease of the Labusch force $f_\mathrm{Lab}$
624: and triggers the crossover from weak- to strong pinning, producing the
625: well known peak effect in the critical current density
626: \cite{LO_79,pippard_69}.
627: 
628: We acknowledge discussions with Anatoly Larkin and financial 
629: support from the Swiss National Foundation.
630: 
631: \begin{thebibliography}{99}
632: 
633: \bibitem{LO_79} A.I.\ Larkin and Yu.N.\ Ovchinnikov,
634:    J.\ Low Temp.\ Phys. \textbf{34}, 409 (1979).
635: 
636: \bibitem{review} G.\ Blatter \textit{et al.},
637:    Rev.\ Mod.\ Phys.\ \textbf{66}, 1125 (1994).
638: 
639: \bibitem{Lab_69} R.\ Labusch,
640:    Cryst.\ Latt.\ Defects \textbf{1}, 1 (1969).
641: 
642: \bibitem{pippard_69} A.B.\ Pippard, Phil.\ Mag.\ {\bf 19},
643: 217 (1969).
644: 
645: \bibitem{LO_NSC_86} A.I.\ Larkin and Yu.N.\ Ovchinnikov,
646:   in {\itshape Nonequilibrium Superconductivity},
647:   eds.\ D.N.\ Langenberg and A.I.\ Larkin
648:   (Elsevier, Amsterdam, 1986), p. 493.
649: 
650: \bibitem{Brandt_86gen} E.H.\ Brandt,
651:    Phys.\ Rev.\ B \textbf{34}, 6514 (1986).
652: 
653: \bibitem{OvIv_91} Yu.N.\ Ovchinnikov and B.I.\ Ivlev,
654:    Phys.\ Rev.\ B \textbf{43}, 8024 (1991).
655: 
656: \bibitem{Schoenenberger_96} A.\ Sch\"onenberger \textit{et al.},
657:    Phys.\ Rev.\ Lett.\ \textbf{77}, 4636 (1996).
658: 
659: \bibitem{note} The integration along $x$ can be interpreted
660:    as an average of pinning forces or as the drag work over 
661:    the period $\tilde{a}$.
662: 
663: \bibitem{alternative} Alternatively, we find $\bar{C}$ using $G_{xx} =
664: 1/\varepsilon_0 k_z^2$ and cut the integral on the scale
665: $l$, $\bar{C} \sim \varepsilon_0/l$; the relation $u \sim
666: f_\mathrm{p} /\bar{C}$ provides the angle $\theta \sim u/l \sim
667: f_\mathrm{p}/\varepsilon_0$.
668: 
669: \bibitem{wagner} O.S.\ Wagner, V.B.\ Geshkenbein, A.I.\ Larkin,
670:    and G.\ Blatter,
671:    Phys.\ Rev.\ B \textbf{59}, 11551 (1999).
672: 
673: \bibitem{book} G.\ Blatter and V.B.\ Geshkenbein, in
674: {\it The Physics of Superconductors}, eds.\ K.H.\ Bennemann and
675: J.B.\ Ketterson (Springer, Berlin, 2003).
676: 
677: \bibitem{LL}  L.D.\ Landau and E.M.\ Lifshitz, 
678:    Vol.\ \textbf{5}
679:    {\itshape Statistical Physics}
680:    (Pergamon Press, London/Paris, 1958). 
681: 
682: \end{thebibliography}
683: 
684: \end{document}
685: 
686: 
687: