cond-mat0506269/m4.tex
1: %
2: \documentclass[prl,aps,twocolumn,showpacs]{revtex4}
3: %  \documentclass[prl,aps,preprint]{revtex4}
4: \usepackage[dvips]{graphicx}
5: \usepackage{dcolumn}
6: \newcommand{\be}{\begin{equation}}
7: \newcommand{\ee}{\end{equation}}
8: \newcommand{\bea}{\begin{eqnarray}}
9: \newcommand{\eea}{\end{eqnarray}}
10: \newcommand{\nn}{ \nonumber}
11: \newcommand{\ds}{\displaystyle}
12: \mathsurround=2pt
13: \begin{document}
14: \topmargin=-20mm
15: %    \Large
16: 
17: 
18: 
19: \title{ Local Features of the Fermi Surface Curvature and the Anomalous Skin Effect in Metals}
20: 
21: 
22: 
23: \author{Natalya A. Zimbovskaya}
24: 
25: 
26: \affiliation
27: {Department of Physics and Electronics, University of Puerto Rico at Humacao, Humacao, PR 00791}
28: 
29: \begin{abstract}
30:  In this paper we present a theoretical analysis of the effect of local geometrical
31:  structure of the Fermi surface on the surface impedance of a metal at the anomalous
32:  skin effect. We show that when the Fermi surface includes nearly cylindrical and/or
33:  flattened segments it may significantly change both magnitude and frequency dependence
34:  of the surface impedance. Being observed in experiments these unusual frequency
35:  dependencies could bring additional information concerning fine geometrical
36:  features of the Fermi surfaces of metals.
37:    \end{abstract}
38: 
39: \pacs{ 73.21 Cd; 73.40. --C}
40: %\vspace{2mm}
41: \date{\today}
42: \maketitle
43: 
44: 
45: 
46: 
47: \section{I. Introduction}
48: 
49:  It is well known that electromagnetic waves incident at the
50: surface of a metal cannot penetrate deeply inside. Actually, the
51: field inside the metal vanishes at the distances of the order of $
52: \delta $ from the surface. This effect is called the skin effect,
53: and the characteristic depth $ \delta $ is called the skin depth.
54: The suppression of the electromagnetic field inside the metal
55: originates from the response of conduction electrons, and it occurs
56: when the frequency $ \omega $ of the incident wave is smaller than
57: the electrons plasma frequency $ \omega_p. $ The latter is the
58: characteristic frequency for the response of the conduction
59: electrons system to an external disturbance. When $ \omega >
60: \omega_p $ the electrons are too slow to respond, and the
61: electromagnetic field penetrates into the metal without decay. Due
62: to the skin effect the incident electromagnetic field could affect
63: condition electrons only when they move inside the layer of the
64: thickness $ \delta $ near the metal surface. The skin depth depends
65: on the electric conductivity of the metal $ \sigma $ and on the
66: frequency $ \omega $ of the incident wave as well. Increase in $
67: \sigma $ and/or $ \omega $ leads to the decrease in the skin depth.
68: At high frequencies $ \tau^{-1} \ll \omega \ll \omega_p\ (\tau $ is
69: the scattering time for conduction electrons) and low temperatures,
70: $ \delta $ may become smaller than the electrons mean free path $ l.
71: $ When the condition $ \delta < l $ is satisfied the effect is
72: referred to as the anomalous skin effect. At the anomalous skin
73: effect the response of a metal to an incident electromagnetic wave
74: is determined with the electrons moving in the skin layer nearly in
75: parallel with the surface of the metal sample. These ``efficient"
76: electrons are associated with a few small ``effective segments" on
77: the Fermi surface (FS).  The remaining electrons stay in the skin
78: layer only for a very short while which prevents them from
79: responding to the electromagnetic field.
80: 
81: A theory of the anomalous skin effect in metals was first proposed
82: more than five decades ago by A. B. Pippard \cite{1} and G. E.
83: Reuter and E. H. Zondheimer \cite{2}, and R. B. Dingle \cite{3}
84: using an isotropic model for a metal. The main results of these
85: studies were presented in some books where high frequency phenomena
86: in metals were discussed \cite{4}. Then the theory was further
87: developed to make it applicable to realistic metals with anisotropic
88: Fermi surfaces \cite{5,6,7,8,9}. It became clear that the response
89: of conduction electrons to an external electromagnetic field under
90: the anomalous skin effect depends on the Fermi surface (FS)
91: geometry, especially its Gaussian curvature
92:  $\ K({\bf p}) = 1/R_1 ({\bf p}) R_2 ({\bf p}), \ $
93:  where $ R_{1,2}\bf (p) $ are the principal radii of curvature.
94: For the most of real metals FSs are complex in shape, and their
95: curvature turns zero at some points. These points could be partitioned
96: in two classes. First, these exist zero curvature points where only one
97: of the principal radii has a singularity, whereas another one remains
98: finite. Usually, such points are combined in lines of zero curvature.
99: The latter are either inflection lines or they label positions of
100: nearly cylindrical strips on the FSs. Also, some points could be
101: found where both principal radii tend to infinity. These points are
102: set out separately, and the FSs are flattened in their vicinities.
103: 
104: When a FS includes points of zero curvature it leads to an enhancement of
105: the contribution from the neighborhoods of these points to the electron
106: density of states (DOS) on the FS. Normally, this enhanced contribution is
107: small compared to the main term of the DOS which originates from the major
108: part of the FS. Therefore it cannot produce noticeable changes in the
109: response of the metal when all segments of the FS contribute  essentially
110: equally. However, when the curvature turns zero at some points on an
111: ``effective" part of the FS, it can give a sensible enhancement in the
112: number of efficient electrons and, in consequence, a pronounced change
113: in the response of the metal to the disturbance.
114: 
115: It has been shown that when the FS includes nearly cylindrical
116: and/or flattened segments, noticeable changes may be observed in the
117: frequency and temperature dependencies of sound dispersion and
118: absorption \cite{10,11,12,13}. Also, the shape and amplitude of
119: quantum oscillations in various characteristics of a metal could be
120: affected by  of the FS local geometry in the vicinities of the
121: extremal cross sections. Qualitative anomalies in the de Haas-van
122: Alphen oscillations associated with cylindrical pieces of the FSs
123: were considered in Refs. \cite{14,15}. Similar anomalies in quantum
124: oscillations in the static elastic constants and the velocity of
125: sound were analyzed in \cite{16,17}.
126: 
127: 
128: 
129: Here, we concentrate on the analysis of possible manifestations of
130: the FS local geometry in the surface impedance of a metal at the
131: anomalous skin effect. In this case the main contribution to the
132: surface impedance of a metal originates from electrons moving nearly
133: in parallel with the surface of the metal. These electrons are
134: efficient quasiparticles, and they belong to the ``effective" part
135: of the FS. The effects of the FS geometry on the metal response at
136: the anomalous skin effect  were analyzed before \cite{9,18} adopting
137: some simplified models for the FS. The purpose of the present work
138: is to carry out a general analysis whose results are independent on
139: particularities in energy-momentum relations and could be applied to
140: a broad class of metals.
141: 
142: \section{II. results and discussion}
143: 
144: We consider a metal filling the half-space $ z < 0 $. A plane
145: electromagnetic wave is incident on the metal surface making a right
146: angle with the latter.  To analyze the response of the metal to the
147: wave we calculate the surface impedance:
148:    \be % f1
149:   Z_{\alpha\beta} = E_\alpha (0) \bigg/ \int_0^{-\infty} J_\beta (z)
150:   d z
151:   \ee
152:    Here, $ \alpha,\beta = x,y;\ E_\alpha(z) $ and $ J_\beta (z) $
153:    are the components of the electric field $ \bf E $ and electric
154:    current density $ \bf J, $ respectively.
155:  Considering the anomalous skin effect we can limit our analysis to the case
156:  of specular reflection of electrons from the surface. Then the surface
157:  impedance tensor has the form:
158:   \be % f2,3
159: Z_{\alpha \beta} = \frac{8 i\omega}{c^2} \int_0^\infty
160: \bigg(\frac{4 \pi i \omega}{c^2} \sigma - q^2 I\bigg )_{\alpha \beta}^{-1} d q.
161:  \ee
162:  Here, $ \omega $ and $ \bf q $ are the frequency and the wave vector of the
163:  incident wave, respectively $ ({\bf q} = (0,0,q)); \ \sigma $ is the electron
164:  conductivity tensor, and $ I_{\alpha \beta} = \delta_{\alpha \beta}. $
165: 
166: To proceed we assume that the FS has a mirror symmetry in a momentum space
167: relative to a plane $ p_z = 0.$ To simplify calculations of the electron
168: conductivity we divide each sheet of the FS in segments in such a way that
169: the momentum $ \bf p $ is a one-to-one function of the electron velocity
170: $ \bf v$ over a segment. The segments may coincide with the FS sheets.
171: Also, it could happen that  some sheets include a few segments. This
172: depends on the FS shape. In calculation of the conductivity we carry out
173: integration over each segment using spherical coordinates in the velocity
174: space, namely, the velocity magnitude at the $j$-th segment $ v_j, $ and
175: the spherical angles $\theta, \varphi. $ So, the element of the surface
176: area is given by the expression:
177: $ d A_j = \sin \theta d \theta d \varphi/|K_j(\theta,\varphi)| $
178:  where $ K_j (\theta,\varphi)$ is the Gaussian curvature of the $j$-th FS
179:  segment. Summing up contributions from all these segments we obtain:
180:     \bea %f3,4
181: &&\sigma_{\alpha\beta} (\omega, q)= \frac{i e^2}{4 \pi^3 \hbar^3 q}
182:  \sum_j \int d \varphi
183:   \nn \\ &&
184: \times
185: \int \frac{ n_\alpha n_\beta   \sin \theta d \theta}
186: {\big |K_j(\theta, \varphi)\big | \big [(\omega + i/\tau) /q v_j - \cos
187: \theta \big ]}.
188:           \eea
189:    Here, $ n_{\alpha,\beta} = v_{j\alpha,\beta}/v_j, $ and $ \tau $ is the
190:    electron scattering time. The limits in the integrals over $ \theta, \varphi $
191:    are determined with the shape of the segments. We remark, however, that the
192:    effective strips on the FS are determined by the condition $ v_z \approx 0 $ for
193:    efficient electrons move in parallel with the metal surface at $ z=0. $
194:    Therefore, the upper limit in the integral over $ \theta $ in the terms
195:    corresponding to the segments including the effective strips must equal $ \pi/2. $
196:    In the following calculations we omit the term $ i/\tau $ in the Eq. (3)
197:    assuming $ \omega \tau \gg 1$ which is typical for the anomalous skin
198:    effect in good metals. Using Eq. (3) we can easily write out the
199:    expressions for the conductivity components. We have:
200:   \bea %f4,3,4
201: && \sigma_{xx} (\omega, q)= \frac{i e^2}{4 \pi^3 \hbar^3 q}
202: \sum_j \int d \varphi
203:    \nn \\ &&\times
204: \int \frac{d \theta \cos^2 \varphi \sin^3 \theta}
205: {\big |K_j(\theta, \varphi)\big | \big [(\omega + i/\tau) /q v_j - \cos
206: \theta \big ]}.
207:           \eea
208:   Another conductivity component $ \sigma_{yy} $ is described with the similar
209:   expression where $ \cos^2 \varphi $ in the integrand numerator is replaced
210:   by $ \sin^2 \varphi. $ In further calculations we assume for simplicity that
211:   the chosen $ z $ axis coincides with a high symmetry axis for the FS, so that
212:   both conductivity and surface impedance tensors are diagonalized.
213: 
214: The main contribution to the surface impedance under the anomalous skin effect
215: comes from the region of large $ q $ where $ \omega/q v \ll 1 .$ To  calculate
216: the corresponding asymptotic expressions for the conductivity components we
217: expand the integrand in the Eq. (4) in powers of $ \omega/qv $.
218: Then we can write the well known result for the principal
219: term in the expansion of the conductivity component $\sigma_{xx}(\omega, q) $:
220:   \be %f5,4,5
221:  \sigma_0 (q) = \frac{e^2}{4 \pi^3 \hbar^3 q} \sum \limits_l
222:  \int d \varphi \frac{\cos^2 \varphi}{\big |K_l (\pi/2, \varphi) \big |}
223:  \equiv \frac{e^2}{4 \pi \hbar^3 q} p_0^2.
224:            \ee
225:  The same asymptotics could be obtained for $ \sigma_{yy}, $ so
226: the indices are omitted for simplicity here and in following
227: expressions. Summation over $ l $ is carried out over all segments of
228: the FS containing effective strips which correspond to $ \theta = \pi/2 \ (v_z = 0) $
229: and the curvature $ K_l (\pi/2, \varphi ) $ is supposed to take finite
230: and nonzero value at any point of any effective strip. For a spherical
231: FS $ p_0 $ equals the Fermi momentum $ p_F.$ In realistic metals the two
232: are not equal but have the same order of magnitude. In general, $ p_0 $
233: is determined by the Eq. (5). Then we can calculate the next term in the
234: expansion of conductivity in powers of $ \omega/qv. $ For a FS whose
235: curvature everywhere is finite and nonzero we arrive at the result
236:    \be %f6,5,6
237:  \sigma_1 (\omega, q) = \sigma_0 (q) \frac{i \omega}{q v_0}.
238:            \ee
239:   Here, the velocity $v_0 $ has the order of the Fermi velocity $ v_F: $
240:   \bea %f7
241:   \frac{1}{v_0} &=& \frac{2}{\pi^2p_0^2} \sum_l \int d \varphi
242: \int_{\alpha_l}^{\pi/2} \frac{d\theta \cos^2 \varphi \sin\varphi}{\cos^2\theta}
243:    \nn \\ \nn \\ &\times &
244:  \left[\frac{1 + \cos^2 \theta/\cos^2 \alpha_l}{|K_l (\pi/2,\varphi)|
245:  v_l (\pi/2,\varphi)} - \frac{\sin^2 \theta}{|K_l (\theta,\varphi)|
246:  v_l (\theta, \varphi)}    \right]
247:    \nn \\ \nn \\ &-&
248:  \frac{2}{\pi^2 p_0^2} \sum_{j\neq l} \int d \varphi \int_{\theta < \pi/2}
249: \frac{d \theta \cos^2 \varphi \sin^3 \theta}{|K_j (\theta,\varphi)|
250: v_j (\theta,\varphi) \cos^2 \theta}. \nn\\
251:    \eea
252:    Here, the lower limit $ \alpha_l $ in the integral over $ \theta $
253:    in the first term takes on values determined by the FS shape
254:    $ (\alpha_l < \pi/2). $ The second term corresponds to the
255:    contribution from the FS segments which do not include effective
256:    lines. For the spherical FS we have $v_0=\pi v_F/4.$
257: 
258: When the curvature at any effective line turns zero, it changes the
259:  asymptotics for the conductivity, as we show below. First, we assume
260:   that the curvature becomes zero at a whole effective line passing
261:   through one of the segments of the Fermi surface. Keeping in mind
262:   that $ z $ axis in the chosen reference system runs along a high  order
263:   symmetry axis we can present the relevant energy-momentum relation in the form:
264:   \be %f8,6
265:   E{\bf(p)} = E_1 (p_x,p_y) + E_2(p_z).
266:   \ee
267:   Near the effective line where $ v_z \equiv \partial E_2/\partial p_z = 0 $
268:   we can approximate $ E_2 (p_z) $ as follows:
269:    \be %f9,7
270:   E_2 (p_z) \approx E_0 \left(\frac{p_z - p^*}{p^*} \right)^{2l};\qquad l\geq 1.
271:   \ee
272:   Here, $ E_0, p^* $ have the dimensions of the energy and momentum,
273:   respectively; $ p_z = p^* $ corresponds to the effective line.
274: 
275: In the vicinity of the effective line we can write the following
276: expression for the curvature of the FS corresponding to the Eq. (8):
277:   \be %f10,8
278:   K{\bf (p)} = \frac{1}{v^4} \frac{\partial v_z}{\partial p_z}
279: \left (v_y^2 \frac{\partial v_x}{\partial p_x} +
280: v_x^2 \frac{\partial v_y}{\partial p_y} -
281: 2 v_x v_y \frac{\partial v_x}{\partial p_y} \right ).
282:    \ee
283:   The value of the curvature at $ v_z = 0 $  is determined by the factor
284:   $\ \partial v_z/\partial p_z \sim [(p_z - p^*)/p^*]^{2l-2}. $ The curvature
285:   becomes zero at the effective line when $ l > 1. $ Expressing this factor as a
286:   function of velocity (which is necessary to carry out integration over a region
287:   in the velocity space) we get $ \partial v_z/\partial p_z \sim (v_z)^{-\beta} $
288:   where $ \beta = -1 +1/(2l -1). $
289: 
290: So, we can use the following approximation for the curvature
291: $K (\theta, \varphi) $ at $ \theta \le \pi/2: $
292:      \be %f11,6, 7
293: K(\theta,\varphi) = W (\theta,\varphi) (\cos \theta)^{-\beta},
294:           \ee
295:  In this expression, the function $ W(\theta,\varphi) $ everywhere
296:  assumes finite and nonzero values, and the exponent $ \beta $ takes on
297:  negative values which correspond to the line of zero curvature  at
298:  $\theta =\pi/2. $  In the close vicinity of this line the FS is nearly
299:  cylindrical in shape. The closer $ \beta $ to $-1,$ the closer to
300:  a cylinder is the effective strip on the FS. The contribution to
301:  the conductivity from the nearly cylindrical segment on the FS is given by:
302:      \bea % f7,8
303: \!\!\! & &
304: \sigma_a (\omega, q) = \frac{i e^2 \omega}{2 \pi^3 \hbar^3 q}
305: \Bigg [ \int d \varphi \int_\alpha^{\pi/2} \! d \theta
306: \Bigg (\frac{\sin^2 \theta}{\big |W(\theta, \varphi)\big |v(\theta,\varphi)}
307:  \nn \\ \nn \\  \! \!\!\! & - &
308: \frac{1}{\big |W(\pi/2, \varphi)\big
309: |v(\theta, \varphi)} \Bigg) \frac{\cos^2 \varphi \sin \theta
310: (\cos \theta)^\beta}{\big (\omega /q v (\theta, \varphi)\big)^2 - \cos^2 \theta}
311: \nn \\  \nn \\ \! \!\!\! &+& \!\!\!
312:  \int
313: \frac{ d \varphi \cos^2 \varphi }{\big|W(\pi/2, \varphi)
314: \big |v(\pi/2, \varphi)} \int_{\alpha}^{\pi/2} \!\!\!
315:   \frac{d \theta \sin \theta (\cos \theta)^\beta}
316: {\big(\omega /q v (\theta, \varphi)\big)^2 - \cos^2 \theta} \Bigg ] . \nn \\
317:           \eea
318: 
319:  Using this asymptotic expression we can calculate the ``anomalous"
320:  contribution to the conductivity $ \sigma_a (\omega, q) $ for small
321:  $ \omega/q v. $ Introducing the largest magnitude of the velocity
322:  on the effective line $ v_a $ we have:
323:     \bea %f13,8,9
324:  \sigma_a (\omega, q) &=& \rho \sigma_0 (q)
325: \left (\frac{\omega}{q v_a} \right )^\beta \left [1 - i \tan
326: \bigg (\frac{\pi \beta}{2} \bigg) \right ] ,
327:    \\ \nn \\
328:  \rho &\approx& \frac{1}{\pi p_0^2} \int \frac{d \varphi
329:  \cos^2 \varphi}{|W (\pi/2, \varphi)|},
330:       \eea
331:   Comparing Eq. (14) with the definition for $ p_0^2 $ introduced
332:   earlier by Eq. (5) we see that $ \rho $ is a dimensionless factor
333:   whose value is determined with the relative number of the ``effective"
334:   electrons concentrated at the nearly cylindrical effective segment.
335: 
336: The value of the contribution to the conductivity from the
337: "anomalous" effective strip depends on the character of the
338: curvature anomaly at given strip, and on the relative number of
339: effective electrons concentrated here as shown in the Fig. 1. In
340: this figure we display plots of $ |\sigma (\omega,q)/\sigma_0 (q)|
341:  \equiv |1 + \sigma_a(\omega,q)/\sigma_0 (q)| $ versus $ \omega/q v. $
342:  When the parameter $ \rho $ takes on values of the order or greater
343:  than 0.1 (the number of effective electrons associated with the
344:  anomalous sections on the FS is comparable to the total number
345:  of the effective electrons), the term $ \sigma_a (\omega,q) $
346:  can predominate over $ \sigma_0 (q) $ and determine the conductivity
347:  value at large $ q. $ This occurs when the shape parameter $ \beta $
348:  accepts values not too close to zero, and the curvature anomaly at
349:  the effective line is well pronounced. When either $ \rho $ or
350:  $ \beta $ or both are very small in magnitude, the main approximation
351:  to the conductivity is described with Eq. 5 as well as for a metal
352:  whose FS curvature is everywhere nonzero. Nevertheless, in such
353:  cases the term $ \sigma_a (\omega, q) $ also is important for it
354:  gives the first correction to the principal term in the expression
355:  for the conductivity.
356: 
357:  \begin{figure}[t]
358: \begin{center}
359: \includegraphics[width=8.6cm,height=4.4cm]{m4_1.eps}
360: %[width=5.0cm,height=6.6cm]{M1.eps}%{g2.eps}
361: \caption{
362: Conductivity component $ \sigma_{xx} (\omega,q ) $ including the
363: contribution from a zero-curvature segment on the FS at large $ q \
364: (\omega/qv \ll 1). $ Left panel: the curves are plotted at
365: $ \rho = 0.2 $ and $ \beta = -0.8,-0.7, -0.6,-0.4,-0.2 $
366: from the top to the bottom. Right panel: the curves are
367: plotted at $ \beta=- 0.6 $ and $ \rho = 0.2, 0.1,0.05,0.03, 0.01 $
368: from the top to the bottom.
369: } \label{rateI}
370: \end{center}
371: \end{figure}
372: 
373: Also, the anomalous contribution to the conductivity could appear
374: when the FS is flattened at some points belonging to an effective
375: segment. To avoid lengthy calculations we illustrate the effect of
376: such points on the conductivity using a simple expression
377: representing the energy momentum relation near the point of
378: flattening $ {\cal M}_0 (p_1,0,0): $
379:    \be % f15
380:   E {\bf (p)} = \frac{p_1^2}{2m_1} \left(\frac{p_x^2}{p_1^2}\right) + \frac{p_2^2}{m_2}
381:   \left(\frac{p_y^2 + p_z^2}{p_2^2}\right)^l
382:    \ee
383:    where $p_1,p_2 $ have dimensions of  momentum. When $ l = 1$ this expression
384:     corresponds to the ellipsoidal FS, and $ m_1, m_2 $ are the principal values
385:     of the effective mass tensor. The FS curvature equals:
386:    \bea % f16
387: \!\!\!\!&&\!\! K{\bf(p)}= \frac{l}{m_2v^4} \left(\frac{p_y^2 +
388: p_z^2}{p_2^2}\right)^{l-1}
389:   \nn\\ \!\! \!\!\!& &\!\!\!\! \times\!
390: \left[\frac{1}{m_1}(v_y^2 + v_z^2) + v_x^2 \frac{l(2l -1)}{m_2}
391: \left(\frac{p_y^2 +p_z^2}{p_2^2}\right)^{l-1}\right].
392:     \eea
393:    For  $ l > 1$ the curvatures of both principal cross sections of the FS
394:     become zero at the point $ (p_1,0,0) $ indicating the FS local flattening.
395: 
396: Turning to the spherical coordinates in the velocity space we can
397: rewrite the expression (16) in the form:
398:   \be % f17
399:  K(\theta,\varphi) = W (\theta, \varphi) (\cos^2\theta + \sin^2 \varphi)^{(1-\beta)/2}
400:     \ee
401:       where the shape parameter $ \beta = -1 +2/(2l - 1). $ When $ l > 1, $ the
402:       FS curvature becomes zero at $ \theta = \pi/2,\ \varphi = 0 $ which correspond
403:       to the point $ {\cal M}_0 . $ The parameter $ \beta $ takes on values from the
404:       interval $ (-1,1) $, and the more pronounced is the FS flattening near the point
405:       $ {\cal M}$ ( the greater is the value of $l) $ the closer is $ \beta $ to $ -1. $
406:       The ``anomalous"  contribution to the conductivity originating from the flattened
407:       segment of the FS has the form similar to Eq. (13), namely:
408:  \be %f18
409:  \sigma_a (\omega,q) = \mu \sigma_0 (q) \left (\frac{\omega}{q
410:  v(\pi/2,0)}    \right)^\beta
411:  \left[1 -i\tan \left(\frac{\pi\beta}{2}\right)\right].
412:   \ee
413:   Here, $ \mu $ is a small dimensionless factor proportional to the relative number
414:   of conduction electrons associated with the flattened part of the FS. Due to the
415:   smallness of $ \mu $ the term (17) may be significant only when $ \beta \leq 0\
416:   (l \geq 1.5). $ Otherwise, it could be neglected.
417: 
418: Now, we proceed in calculations of the surface impedance given by the expression (2).
419: Under anomalous skin effect conditions the impedance can be represented as an
420: expansion in inverse powers of the anomaly parameter $ (\xi \gg 1). $ Representing
421: the conductivity as the sum of terms (5) and (6), we can calculate two first terms
422: in the expansion of the surface impedance in inverse powers of the anomaly parameter:
423:      \bea % f19,16,10,12
424:  Z &\equiv &  R - i H = - \frac{8 i \omega}{c^2} \delta \int_0^\infty d t
425: \frac{1}{1 - i t^3 (1 + i t/\xi)}  \nn \\ \nn \\ & \approx &
426:  Z_0 \left(\frac{\omega}{\omega_0}\right)^{2/3}
427: \left[1 -i\sqrt3 - \frac{2}{3} \left (\frac{\omega}{\omega_0}\right)^{2/3}
428: \big(1 + \sqrt 3 \big) \right ] \nn \\
429:            \eea
430:  where $ \delta = (c^2 \hbar^3/ e^2 p_0^2 \omega)^{1/3} $ is the skin depth,
431:   $$
432:   \ds Z_0 = \frac{8 \pi}{3 \sqrt 3} \frac{v_0}{c^2}; \qquad \xi =
433:   \frac{v_0}{\omega \delta} \equiv \left(\frac{\omega}{\omega_0}\right)^{2/3}
434:   $$
435:   is the anomaly parameter, and frequency $ \omega_0 $ equals:
436:   \be % f20,17,11
437: \omega_0 = \left (\frac{v_0}{\hbar} \right)^{3/2}
438: \frac{e p_0}{c} .
439:    \ee
440:  Keeping in mind that $ v_0 \sim v_F $ and $ p_0 \sim p_F $ we can roughly
441:  estimate the characteristic frequency $ \omega_0. $ In good metals the
442:  electron density has the order of $ 10^{21}-10^{22}$ cm$^{-3}$, so
443:  $ \omega_0 \sim 10^{12}-10^{13} s^{-1}. $ This is significantly smaller
444:  that the plasma frequency $ \omega_p $ which in good metals is of the
445:  order of $ 10^{15}-10^{16} s^{-1}. $ As one would expect, the inequality
446:  $ \omega \ll \omega_0 \ (\xi \gg 1) $ agrees with the general requirement
447:  on frequencies $ \omega \ll \omega_p, $ and could be satisfied at
448:  $ \omega \sim 10^{10}-10^{11} s^{-1}. $
449: 
450: The expression of the form (19) was first obtained by R. B. Dingle
451: (see \cite{3}) within the isotropic model of metal. Later it was
452: generalized to be applied to realistic metals, assuming that their
453: FSs do not include segments of zero curvature \cite{8}. For such FSs
454: the frequency dependence of the surface impedance has the same
455: character, as for a Fermi sphere. The main approximation of the
456: surface impedance is proportional to $ \omega^{2/3} $ and the first
457: correction to it is proportional to $ \omega^{4/3}. $
458:  \begin{figure}[t]
459: \begin{center}
460: \includegraphics[width=8.6cm,height=4.4cm]{m4_2.eps}
461: %[width=5.0cm,height=6.6cm]{M1.eps}%{g2.eps}
462: \caption{Frequency dependence of the real $ ( R) $ and imaginary $
463: (H) $ parts of the surface impedance of a metal under the anomalous
464: skin effect. Dashed-dotted lines are plotted using Eq.(12) at $ \rho
465: = 0.2 $ and $ \beta = -0.8,-0.7,-0.6,-0.4, $ from the top to the
466: bottom. Solid lines represent real and imaginary parts of the
467: surface impedance of a metal whose FS does not include nearly
468: cylindrical and/or flattened segments. } \label{rateI}
469: \end{center}
470: \end{figure}
471: 
472: When the FS includes locally flattened or
473: nearly cylindrical segments the asymptotic expression for the
474: surface impedance changes. The effect of this anomalous
475: local geometry of the FS on the impedance is especially strong for
476: $ \beta < -0.5, \ \rho \stackrel{ >}{\sim} 0.1. $ Under these
477: conditions the ``anomalous" contribution dominates over the other
478: terms in the expression for conductivity and determines the principal
479: term of the surface impedance. As a result we have:
480:   \be %f21,18,12
481:  Z \approx Z_0 \zeta (\beta) \left (\frac{\omega}{\omega_0} \right)^{2/(3+ \beta)}
482:     \ee
483:  where
484:   \bea % f22,19,13
485: \zeta (\beta) &=& \frac{3 \sqrt 3 \, \rho^{-1/3 + \beta}}{3 + \beta}
486: \left( \cos\frac{\pi \beta}{2} \right)^{1/(3+\beta)}
487:    \nn\\\nn\\ &\times&
488: \left[ \cot \left(\frac{\pi}{3 + \beta}\right) - i \right].
489:     \eea
490: 
491:  The surface impedance described with (21) differs in magnitude from that
492:  of a conventional metal whose FS does not include zero curvature segments.
493:  Frequency dependence of the surface impedance also changes as shown in
494:  the Fig. 2. Now it is proportional to $\omega^{2/(\beta+3)}. $ For a
495:  nearly cylindrical effective strip the exponent $ 2/(\beta + 3) $ varies
496:  in the interval $ (0.8,1) $ where the value $1$ corresponds to the
497:  precisely cylindrical strip. So, when the significant part of effective
498:  electrons is associated with a nearly cylindrical effective strip, the
499:  impedance should slower increase with increase of frequency than in a
500:  ``conventional" case.
501: \begin{figure}[t]
502: \begin{center}
503: \includegraphics[width=8.6cm,height=4.4cm]{m4_3.eps}
504: %[width=5.0cm,height=6.6cm]{M1.eps}%{g2.eps}
505: \caption{ Frequency dependence of the real $ (\Delta R) $ and
506: imaginary $ (\Delta H) $ parts of the first correction to the main
507: term in the surface impedance expansion in the inverse powers of the
508: anomaly parameter. The curves are plotted at $ \rho= 0.01, \ \beta =
509: -0.9; \ \rho = 0.02, \ \beta = - 0.8 $ (dashed lines); $ \rho = 0.1,
510: \ \beta = -0.5; \ \rho = 0.1, \ \beta = - 0.4 $ (dotted lines).
511: Solid lines present the real and imaginary parts of the first
512: correction to the main approximation
513:  for the impedance of a metal whose FS does not include nearly cylindrical
514:  and/or flattened segments.}
515:  \label{rateI}
516: \end{center}
517: \end{figure}
518: 
519: 
520: 
521:  Now we consider more realistic case when either $ \rho $ or $ \beta $ or
522:  both take on values close to zero (a zero curvature segment on the effective
523:  part of the FS is narrow and/or the curvature anomaly is only moderately
524:  pronounced). In this case the anomalous contribution (13) is the first
525:  correction to the main approximation for the conductivity, and it
526:  determines the first correction to the approximation for the surface impedance:
527:   \be %f23,20,14
528: Z = Z_0 \left (\frac{\omega}{\omega_0} \right)^{2/3} (1 - i \sqrt 3) + \Delta Z.
529:     \ee
530:  Here,
531:   \bea %f24-25,21-22,15-16
532:  \Delta Z & \equiv & \Delta R - i \Delta H \approx - Z_0 \eta (\beta)
533:  \left (\frac{\omega}{\omega_0} \right)^{2(\beta + 1)/3};
534:     \\ \nn \\
535: \eta(\beta) & = &  \frac{\rho(\beta+1)}{\sqrt 3}
536: \frac{1}{\cos (\pi \beta/2)}
537: \bigg \{\cot \bigg( \frac{ \pi (\beta + 1)}{3} \bigg) + i \bigg \} . \nn \\
538:    \eea
539: 
540:  We can apply this result (24) to describe the contribution to the
541:  surface impedance from narrow or weakly developed nearly cylindrical
542:  strip and also from a point of flattening located on the effective segment
543:  of the FS. The correction to the main approximation of the surface
544:  impedance now is proportional to $ \omega^{2(\beta+1)/3}$ as we show
545:  in the Fig. 3. The presented analysis shows that the surface impedance
546:  of a semiinfinite metal whose FS has points or lines of zero  curvature
547:  can be described by the formulas (21)--(24). The obtained asymptotic
548:  expressions indicate that an anomaly of curvature on an effective line
549:  changes frequency dependence of the surface impedance, and under certain
550:  conditions it can essentially change its magnitude. This follows from
551:  the discussed above relation between the curvature of the FS and the
552:  number of effective electrons.
553: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
554: 
555: 
556: 
557: 
558: \section{III. conclusion}
559: 
560:  The concept of Fermi surface is recognized as one of the most
561: meaningful concepts in condensed matter physics, providing  an
562: excellent insight in the main electronic properties of conventional
563: metals and other materials with metallic-like conductivity.
564: Extensive studies based on experimental data concerning effects
565: responsive to the structure of electronic spectra in metals and
566: using advanced computational methods were carried out to restore the
567: FS geometries. These efforts were resulted in the impressive mapping
568: of the FSs of conventional metals. However, in the course of these
569: studies a comparatively little attention was paid to fine local
570: features in the FS geometries including zero curvature lines and/or
571: points of flattening. These local curvature anomalies do not
572: significantly affect main geometrical characteristics of FSs (such
573: as connectivity, locations of open orbits, sizes and arrangements of
574: sheets) which are usually determined from the standard experiments.
575: Therefore these local features could be easily missed when a FS is
576: restored, if one does not expect them to be present, and does not
577: pay special attention to keep them in the resulting FS image. So, it
578: is important to explore possible experimental manifestations of the
579: FS curvature anomalies. Adoption of the phenomenological models is
580: justified in these studies as far as these models are based on
581: reasonable assumptions concerning the FS geometry. Actually,
582: phenomenological models were commonly used to develop the theory of
583: ``standard" effects such as de Haas-van Alphen effect, which were
584: (and are) employed as tools to obtain informations concerning FSs
585: shapes \cite{19}. This approach to fermiology does not contradict
586: that one based on electron band structure calculations. It
587: supplements the latter. Such supplementing analysis could bring new
588: insight in the physical nature and origin of some physical effects
589: including these considered in the present work, and show their
590: usefulness in studies of the FSs geometries.
591: 
592: As for the particular models  adopted in the present work they may
593: be reasonably justified within a nearly free electron approximation.
594: Adopting the nearly free-electrons approach we arrive at the
595: energy-momentum relation for conduction electrons:
596:   \be % f26
597:  E = \frac{\bf k^2}{2m} + \frac{\bf g^2}{2m} - \frac{1}{m} \sqrt{{\bf
598:  (k\cdot g)^2} +  m^2V^2 },
599:   \ee
600: where $ m $ is the effective mass, $ {\bf k= g - p;\ g = \hbar G}/2;
601: \ \bf G $ is a reciprocal lattice wave vector; $ V $ is Fourier
602: component of the potential energy of electron in the lattice field
603: which corresponds to the vector $ \bf G. $
604:   Within the nearly free-electron model the energy $ V $ is assumed
605:   to be small compared  to the Fermi energy $ E_F ,$ so we introduce
606:   a small parameter $ \epsilon = \sqrt{V/E_F} $.
607: 
608: The corresponding FS looks like a sphere with ``knobs" located at
609: those segments which are close to the boundaries of the Brillouin
610: zone. Inflection lines of zero curvature pass along the boundaries
611: between the knobs and the main body of the FS. A FS segment
612: including a knob and its vicinity is axially symmetric, and the
613: symmetry axis is directed along the corresponding reciprocal lattice
614: vector. In further analysis we single out such segment to consider
615: it separately. For certainty we choose the coordinate system whose
616: $``x"$ axis is directed along the reciprocal lattice vector. Within
617: the chosen segment the FS curvature is described with the
618: expression:
619:     \be % f27
620:  K = \frac{m^2 v_x^2 + p_\perp^2 dv_x/dp_x}{(p_\perp^2 + m^2 v_x^2)^2}
621:   \ee
622:  where $ p_\perp^2 = p_y^2 + p_z^2. $
623: 
624: Equating the FS curvature to zero, and using the energy-momentum
625: relation (26) we find the values $p_{x0} $ and $p_{\perp 0} $
626: corresponding to the inflection line.
627:  We get:
628:    \be % f28
629:  p_{x0} = p_F  (1 - \epsilon^2/\sqrt 2), \qquad p_{\perp 0} \approx p_F \epsilon,
630:   \ee
631:   where $p_F $ is the radius of the original Fermi sphere. Now, we can expand
632:   the variable $ p_x $ in powers of $(p_\perp - p_{\perp 0}) $ near the zero
633:   curvature line. Taking into account that $ d^2 p_x/ dp_\perp^2 $ turns zero
634:   at points belonging to the inflection line, and keeping the lowest-order
635:   terms in the expansion, we obtain:
636:    \be %f29
637:         p_x \approx p_{x0} - \frac{\epsilon}{\sqrt 2} (p_\perp - p_{\perp 0}) -
638:   \frac{p_F}{\sqrt 2\, \epsilon} \left(\frac{p_\perp - p_{\perp 0}}{p_F}\right)^3.
639:    \ee
640:     Substituting this approximation into Eq. (26) we arrive at the following
641:     energy-momentum relation:
642:   \be % f30
643:   E{\bf (p)} = \frac{p_x^2}{2m} + \frac{2}{\epsilon} \frac{p_F^2}{2m}
644:   \left(\frac{p_\perp - p_{\perp 0}}{p_F} \right)^3.
645:    \ee
646:   The latter could be employed near the zero curvature line where
647:   $ p_{\perp 0} \ll p_F. $ Omitting $ p_{\perp 0} $ we arrive at the
648:   energy-momentum relation of the form (15) where $ l = 3/2 . $ Also,
649:   we can compare the equations describing cross-sections $ p_y = 0 $
650:   of the FS corresponding to the Eq. (30) and our phenomenological
651:   model (8), (9). Again, we see that these equations are in
652:   agreement with each other provided that $ l = 3/2. $
653: 
654: 
655: There is an experimental evidence that "necks" connecting
656: quasispherical pieces of the FS of copper include nearly cylindrical
657: belts \cite{19}. It is also likely that the FS of gold possesses the
658: same geometrical features for it closely resembles that of copper.
659: As for possible flattening of the FS, experiments of \cite{20,21,22}
660: on the cyclotron resonance in a magnetic field normal to the metal
661: surface give grounds to conjecture that such anomalies could be
662: found on the FSs of cadmium, zinc and even potassium.
663:  Another group of materials where we can expect the FS curvature anomalies
664:  to be manifested includes layered structures with metallic-type conductivity
665:  (e.g. $\alpha - (BEDT - TTF)_2 Mhg(SCH)_4 $ group of organic metals). Fermi
666:  surfaces of these materials are sets of rippled cylinders, isolated or
667:  connected by links. There exists experimental evidence that the
668:  quasi-two-dimensional FSs of some organic metals include segments
669:  with zero curvature \cite{23}.
670:  Also, recent investigations give grounds to expect the FSs of some new
671: conducting materials include flattened segments \cite{24,25,26}.
672: 
673: 
674: 
675: The most important result of the present work is that it shows how
676: such fine geometric features as points of flattening and/or zero
677: curvature lines could be manifested in experiments on the anomalous
678: skin effect. It is shown that when the FS includes nearly
679: cylindrical segments or it is flattened at some points, qualitative
680: changes may occur in frequency dependencies of the surface impedance
681: under the anomalous skin effect. Being observed in experiments, such
682: unusual frequency dependencies would indicate the presence of
683: zero-curvature lines and points on the FS, and display their
684: location. Also, analyzing these frequency dependencies, the shape
685: parameter $ \beta $ could be found giving additional information on
686: the FSs local structure. This information may be used in further
687: studies of the FSs geometries.
688: % \vspace{2mm}
689: 
690: 
691: 
692: \section{ Acknowledgments} The author thanks G. M. Zimbovsky for
693: help with the manuscript. This work was supported  by NSF Advance
694: program SBE-0123654, NSF-PREM 0353730, and PR Space Grant
695: NGTS/40091.
696: 
697: \begin{thebibliography}{99}
698: 
699: 
700: \bibitem{1} A. B. Pippard, Proc. Roy. Soc. {\bf A 191}, 385 (1947).
701: 
702: \bibitem{2} G. E. Reuter and E. H. Zondheimer, Proc. Roy. Soc. {\bf A 195}, 336 (1949).
703: 
704: \bibitem{3} R. B. Dingle, Physics {\bf 19}, 311 (1953).
705: 
706: \bibitem{4} See e.g.: A. B. Pippard, {\it The Dynamic of Conduction Electrons}
707: (Gordon and Breach, New York, 1965); P. M. Platzman and P. A. Wolff,
708: {\it Waves and Interactions in Solid State Plasmas} (Academic Press,
709: new York, 1973); A. A. Abrikosov, {\it Introduction to the Theory of
710: Normal Metals} (Academic Press, New York, 1972); I. M. Lifshitz, M.
711: Ya. Azbel, and M. I. Kaganov, {\it Electron Theory of Metals}
712: (Consultants Bureau, New York, 1973); N. A. Zimbovskaya, {\it Local
713: Geometry of the Fermi Surface and High-Frequency Phenomena in
714: Metals} (Springer-Verlag, New York, 2001).
715: 
716: \bibitem{5} A. B. Pippard, Proc. Roy. Soc. {\bf A224}, 273 (1954).
717: 
718: \bibitem{6} E. H. Sondheimer, Proc. Roy. Soc. {\bf A224}, 160 (1954).
719: 
720: \bibitem{7} M. L. Glasser, Phys. Rev. {\bf 176}, 1110 (1968).
721: 
722: \bibitem{8} V. V. Ustinov and D. T. Khusainov, Sov. J. Low. Temp. Phys. {\bf 11}, 617 (1985).
723: 
724: \bibitem{9} M. I. Kaganov and P. Kontreras, JETP {\bf 79}, 360 (1994).
725: 
726: \bibitem{10} G. T. Avanesjan, M. I. Kaganov and T. Yu. Lisovskaya, JETP, {\bf 48}, 900 (1978).
727: 
728: \bibitem{11} V. M. Kontorovich and N. A. Stepanova, JETP {\bf 49}, 321 (1978).
729: 
730: \bibitem{12} I. M. Suslov, Sov. Phys. Solid State {\bf 23}, 1114 (1981).
731: 
732: \bibitem{13} N. A. Zimbovskaya, JETP {\bf 80}, 932 (1995).
733: 
734: \bibitem{14} G. Lacueva and A. W. Overhauser, Phys. Rev. Lett. {\bf 46}, 1273 (1992).
735: 
736: \bibitem{15} Ya. M. Blanter, M. I. Kaganov and D. B. Posvyanskii, Phys-Usp. {\bf 165}, 178 (1995).
737: 
738: \bibitem{16} N. A. Zimbovskaya, J. Low Temp. Phys. {\bf 21}, 217 (1995).
739: 
740: \bibitem{17} N. A. Zimbovskaya, Phys. Rev. B {\bf 71}, 024109 (2005).
741: 
742: \bibitem{18} M. I. Kaganov and Yu. V. Gribkova, J. Low Temp. Phys. {\bf 17}, 473 (1991).
743: 
744: \bibitem{19} D. Shoenberg, {\it Magnetic Oscillations in Metals}
745: (Cambridge University Press, New York, 1984).
746: 
747: \bibitem{20} V. P. Naberezhnykh and N. K. Dan'shin, JETP {\bf 29}, 658 (1969).
748: 
749: \bibitem{21} V. P. Naberezhnykh and V. L. Mel'nik, Sov. Phys. Solid State {\bf 7}, 197 (1965).
750: 
751: \bibitem{22} G. A. Baraff, C. C. Grimes and P. M. Platzman, Phys. Rev. Lett. {\bf 22}, 590 (1969).
752: 
753: \bibitem{23} J. Wosnitza, {\it Fermi Surface of Low-Dimensional Organic Metals and
754: Superconductors}  (Springer-Verlag, Berlin, 1996).
755: 
756: \bibitem{24} C. Bergemann, S. R. Julian, A. P. Mackenzie,
757: S. NishiZaki, and Y. Maeno, Phys. Rev. Lett. {\bf 84}, 2662 (2000).
758: 
759: \bibitem{25} Y.-D. Chuang, A. D. Gromko, D. S. Dessau, T. Kimura and Y. Tokura,
760:   Science, {\bf 292}, 1509 (2001).
761: 
762: \bibitem{26} T. Dahm and N. Schopohli, Phys. Rev. Lett. {\bf 91}, 017001 (2003).
763: 
764: 
765: 
766: 
767: 
768: \end{thebibliography}
769: \end{document}
770: