1: %\documentclass[prb]{revtex4}
2: \documentclass[prb,twocolumn,showpacs]{revtex4}
3: \usepackage{graphicx}
4: %\documentstyle[prb,aps]{revtex}
5: \begin{document}
6: \title{Simultaneous existence of two spin-wave modes
7: in ultrathin Fe/GaAs(001) films studied by Brillouin Light Scattering:
8: experiment and theory}
9: \author{M. G. Pini$^{1,2}$}
10: \email{mgpini@ifac.cnr.it}
11: \author{P. Politi$^{1,2}$}
12: \author{A. Rettori$^{2,3}$}
13: \author{G. Carlotti$^{4,5}$}
14: \author{G. Gubbiotti$^{5}$}
15: \author{M. Madami$^{4,5}$}
16: \author{S. Tacchi$^{4,5}$}
17: %\homepage[]{Your web page}
18: %\thanks{}
19: %\altaffiliation{}
20: \affiliation{
21: $^1$ Istituto di Fisica Applicata ``N. Carrara",
22: CNR, Via Madonna del Piano, I-50019 Sesto Fiorentino, Italy
23: \\
24: $^2$ INFM, Unit\`a di Firenze, Via G. Sansone 1, I-50019 Sesto Fiorentino, Italy
25: \\
26: $^3$ Dipartimento di Fisica, Universit\`a di Firenze,
27: Via G. Sansone 1, I-50019 Sesto Fiorentino, Italy
28: \\
29: $^4$ Dipartimento di Fisica, Universit\`a di Perugia,
30: Via Pascoli, I-06123 Perugia, Italy
31: \\
32: $^5$ INFM, Unit\`a di Perugia, Via Pascoli, I-06123 Perugia, Italy
33: }
34: \date{\today}
35: \begin{abstract}
36: A double-peaked structure was observed in the {\it in-situ} Brillouin
37: Light Scattering (BLS) spectra of a 6 \AA$~$ thick epitaxial
38: Fe/GaAs(001) film for values of an external magnetic field $H$, applied
39: along the hard in plane direction,
40: lower than a critical value $H_c\simeq 0.9$ kOe. This experimental finding is theoretically
41: interpreted in terms of a model which assumes a non-homogeneous magnetic ground state
42: characterized by the presence of perperpendicular up/down stripe domains.
43: For such a ground state, two spin-wave modes, namely an acoustic and an optic mode, can exist.
44: Upon increasing the field the magnetization tilts in the film plane,
45: and for $H \ge H_{c}$ the ground state is homogeneous,
46: thus allowing the existence of just a single spin-wave
47: mode.
48: %The value of $H_{c}$ is determined by the balance between
49: %the uniaxial out-of-plane anisotropy energy and the demagnetization
50: %energy in the domain structure, and depends also on the orientation
51: %of the in-plane field with respect to the crystallographic axes
52: %owing to a uniaxial in-plane anisotropy.
53: The frequencies of the two spin-wave modes
54: were calculated and successfully compared with the experimental data.
55: The field dependence of the intensities of the corresponding two peaks
56: that are present in the BLS spectra was also estimated, providing further
57: support to the above-mentioned interpretation.
58: \end{abstract}
59: \pacs{75.70.-i}
60: %75.70.-i Magnetic properties of thin films, surfaces, and interfaces
61: \maketitle
62:
63: \section{Introduction}
64:
65: Ferromagnetic-semiconductor heterostructures, like ultrathin Fe/GaAs(001) films,
66: have received considerable attention\cite{Brockmann,Gester,Steinmueller,Monchesky}
67: for their potential technological applications in new magnetoelectronic
68: devices.\cite{Prinz,libro} A sharp and well ordered interface,
69: without any dead magnetic layer, can be obtained, as bcc Fe
70: grows epitaxially on GaAs thanks to the small lattice
71: mismatch (1.4\%). Together with the expected cubic anisotropy of bulk bcc
72: Fe, a strong in-plane uniaxial anisotropy has been found in ultrathin
73: Fe/GaAs(001) films, resulting from the atoming bonding at the
74: interface.\cite{Bland}
75: The evolution of the latter anisotropy with film thickness has been
76: quantitatively analyzed by some of us\cite{esteso}
77: in a thorough {\it in-situ} Brillouin scattering study of the dynamical
78: magnetic properties of such films.
79: %In Ref. \onlinecite{esteso}, an external magnetic field $H$ was
80: %applied in the film plane and the frequency of the
81: %spin-wave excitations was measured as a function of the field
82: %intensity, of the incidence angle of light and of the angle between
83: %the field direction and an in-plane crystallographic axis.
84: %Upon decreasing the iron thickness, it was observed that
85: %both the out-of-plane and the in-plane uniaxial anisotropies
86: %strongly increase, while the dipolar field, favouring in-plane
87: %magnetization, decreases.
88: %It was also shown that, surprisingly, the presence
89: %of a Cu overlayer completely suppresses the in-plane
90: %anisotropy in the thinner films.
91:
92:
93: % Quello che segue dovrebbe sostituire, nella introduzione,
94: % il pezzo che va da "In this paper we present a careful .."
95: % fino a ".. but the field dependence of the frequencies could
96: % not be quantitatively reproduced."
97:
98:
99: The same system Fe/GaAs(001) has now been found to display a very
100: interesting phenomenon, for small iron thickness ($t_{Fe}=6$\AA):
101: below a critical field $H_c\simeq 0.9$ kOe {\it in-situ} BLS
102: spectra show a ``double-peak" structure, therefore revealing the
103: existence of two spin-wave modes for $H<H_c$. This feature is not
104: completely new: it was already observed\cite{Falco} in Co/Au(111)
105: films, for $t_{Co} \ge 6$ML and $H < H_c \sim 3$ kOe. However, the
106: novelty of our contribution is twofold. (i)~The experimental
107: observation of the double-peak structure has been done in two
108: different samples and both upon increasing and decreasing the
109: magnetic field. This confirms that the phenomenon can be well
110: reproduced and rules out the possibility it may be due to
111: metastability effects. (ii)~We develop a theory which explains the
112: field dependence of the observed spin-wave frequencies as well as
113: the intensities of the corresponding BLS peaks.
114:
115: The starting point of our theory is that the observed splitting of
116: the spin-wave spectrum into two modes is not compatible with a
117: homogeneous ground state. For this kind of system, the simplest
118: and most natural explanation is to assume a perpendicularly
119: magnetized up/down domain structure (for such low values of film
120: thickness, in-plane magnetized domains are {\it not} energetically
121: favoured\cite{rassegna,EPJB,Pescia}). Another possibility, a two
122: sublattices spin arrangement, can be readily disregarded because
123: the splitting is only observable below a critical value of the
124: field and because it is hardly applicable to an epitaxial iron
125: film. Our assumption is therefore an up/down domain ground state.
126: With increasing the magnetic field $H$, the magnetization of each
127: domain gradually tilts in the plane and, for $H$ greater than a
128: critical value $H_c$, the ground state is homogeneously magnetized
129: in plane. Two branches are found for $H<H_{c}$ and a single
130: (uniform) mode for $H \ge H_{c}$.
131:
132: At present an {\it in-situ} high resolution mapping of the
133: magnetization is outside the reach of conventional microscopic
134: techniques. In the absence of detailed
135: information on the actual domain structure and for the sake of
136: generality, we are going to assume a stripe domains structure.
137: Such a one-dimensional model has the advantage that the
138: frequencies of the spin-wave excitations can be easily evaluated
139: and their field dependence can be reproduced for different
140: in-plane directions.
141:
142:
143: The format of the paper is as follows. In Section II the
144: experimental details concerning the sample preparation and the BLS
145: technique are summarized. In Section III the frequencies of the
146: spin-wave excitations with respect to a non-homogeneous ground
147: state with up/down stripe domains are calculated using the
148: Landau-Lifshitz equations of motion. We also estimate the field
149: dependence of the spin-wave intensity of the two modes in the
150: framework of a classical macroscopic model which relates the
151: scattered intensity to the magnetization-dependent permittivity
152: tensor. In Section IV the experimental results are presented and
153: compared with the predictions of the theoretical model. Finally,
154: in Section V the conclusions are drawn. Details about the
155: calculation of the spin-wave frequencies can be found in Appendix
156: A.
157:
158: \section{Experiment}
159:
160: \begin{figure}
161: \label{fig1}
162: \includegraphics[width=8cm,angle=0,bbllx=40pt,bblly=60pt,%
163: bburx=500pt,bbury=760pt,clip=true]{Fig1new.eps}
164: \caption{
165: {\it In-situ} Brillouin spectra taken from a $t_{Fe}=6$\AA$~$thick
166: epitaxial Fe/GaAs(001) film for different values of the external magnetic
167: field $H$, applied within the film plane along $[-110]$, the hard
168: in-plane direction. Two peaks, indicated by arrows, are simultaneously
169: observed for $H=0.6$ and 0.7kOe.
170: }
171: \end{figure}
172:
173: Following the previous investigation of Fe/GaAs(100) films with
174: different thickness,\cite{esteso} in this work we focus our
175: attention on 6\AA$~$ thick Fe films, because for this particular
176: thickness BLS spectra exhibit the double-peak structure described
177: in the following. The reproducibility of the results was checked
178: studying two different samples with the same nominal thickness.
179: The two iron films were grown on a Si-doped GaAs(001) single
180: crystal in an ultra-high-vacuum (UHV) chamber, specially designed
181: to allow {\it in-situ} BLS measurements\cite{apparatus,
182: apparatus2} (background pressure $3 \times 10^{-10}$ mbar) at
183: GHOST laboratory, Perugia University.\cite{ghost} Details about
184: sample preparation and structural characterization can be found in
185: Ref.~ \onlinecite{esteso}. About 200 mW of monochromatic
186: p-polarized light, from a solid state laser (532 nm line), was
187: focused onto the sample surface using a camera objective of
188: numerical aperture 2 and focal length 50 mm. The back-scattered
189: light was analyzed by a Sandercock type (3+3)-pass tandem
190: Fabry-P\'erot interferometer.\cite{Sander} The external dc
191: magnetic field was applied parallel to the surface of the film and
192: perpendicular to the plane of incidence of light {\it i.e.} in the
193: so-called Damon-Eshbach geometry. BLS measurements of the
194: spin-wave frequency were performed {\it in situ} at room
195: temperature as a function of both the intensity and the in-plane
196: direction of the applied magnetic field. Typical BLS spectra for
197: such a film, taken with the magnetic field applied along the hard
198: in-plane direction [-110], are shown in Fig.~1. The presence of a
199: double-peak structure is evident for applied field values
200: $0.4<H<0.9$ kOe. In contrast, a single peak was observed in the
201: BLS spectra measured for magnetic field applied along both the
202: [100] and the [110] directions (not shown).
203:
204:
205: \section{Theory}
206:
207: As discussed in the Introduction, the observed splitting of the
208: spin-wave spectrum into two modes is not compatible with a
209: homogeneous ground state. In the absence of experimental data
210: about the actual spin configuration, we assume a simplified model
211: with perpendicularly magnetized up/down domains in the shape of
212: parallel stripes of infinite length along the in-plane field
213: direction and with vanishing thickness of the wall between
214: opposite domains: see Fig. 2. While more complicated up/down (or
215: possibly canted) domain patterns cannot in principle be ruled out,
216: the former one-dimensional model for the non-homogeneous ground
217: state presents the advantage that the frequencies of the spin-wave
218: excitations can be easily evaluated using the theory of domain
219: mode ferromagnetic resonance
220: (DMFMR),\cite{RW,Ramesh,Ebels,notaBLS} and analytical results can
221: be obtained for field applied in plane along high symmetry
222: directions. Such a simple model turns out to be a useful tool to
223: reproduce the spin-wave behavior; however, one should not expect
224: it to be entirely realistic or able to account for other
225: properties of the system, like the domain wall structure, the
226: spatial dependence of the demagnetization factors, as well as the
227: domain width and its field dependence.
228:
229: \begin{figure}
230: \label{draw}
231: \includegraphics[width=8cm,angle=90]{Stripe.eps}
232: \caption{One-dimensional up/down stripe domain structure
233: and coordinate system used to define the domain variables:
234: $z$ denotes the normal to the film plane.}
235: \end{figure}
236:
237: The free energy per unit volume of the system in Fig.~2
238: is given by the sum of various contributions
239: \begin{equation}
240: \label{G}
241: G=G_{H}+G_{dip}+G_{2\perp}+G_{2\Vert}
242: \end{equation}
243: where $G_H$ is the Zeeman energy term due to the external field,
244: $G_{dip}$ is the term due to demagnetization, while $G_{2\perp}$
245: and $G_{2\Vert}$ are the terms due to uniaxial out-of-plane and
246: in-plane anisotropies, respectively. One has\cite{RW}
247: \begin{eqnarray}
248: \label{free}
249: G_H&=&-{{HM_s}\over 2} [\sin\theta_1 \cos\phi_1
250: +\sin\theta_2\cos\phi_2]
251: \cr
252: G_{dip}&=&{{\pi}\over 2}M_s^2[\cos\theta_1+\cos\theta_2]^2
253: \cr
254: &+&{{\pi}\over 2}M_s^2 N_{zz}[\cos\theta_1-\cos\theta_2]^2
255: \cr
256: &+&{{\pi}\over 2}M_s^2 N_{yy}[\sin\theta_1 \sin\phi_1-\sin\theta_2 \sin\phi_2]^2
257: \cr
258: G_{2\perp}&=&-{{K_{2\perp}}\over 2}[\cos^2 \theta_1+\cos^2\theta_2]
259: \end{eqnarray}
260: where $M_s$ is the saturation magnetization;
261: $H$ is the external field applied within the film plane ($xy$)
262: along the $x$ direction;
263: $K_{2\perp}>0$ is a uniaxial out-of-plane anisotropy
264: that favours the up/down spin alignment along $z$,
265: the normal to the film plane;
266: $N_{zz}$ is the demagnetization factor associated to the up/down
267: domain structure: since the stripes are assumed to be parallel to the field direction
268: ($x$ axis), one has $N_{xx}=0$ and $N_{yy}=1-N_{zz}$.
269: In general, the static demagnetization factor
270: $N_{zz}$ is a function of the domain aspect ratio $L/t$ (where $L$ is the domain
271: width and $t$ the film thickness) and of the rotational permeability
272: $\mu$.\cite{RW,Ramesh,KooyEnz}
273: For magnetic field applied along the hard in-plane direction,
274: the in-plane anisotropy contribution to the free energy is written as
275: \begin{equation}
276: \label{ahard}
277: G_{2\Vert}^{(i)}=-{{K_{2\Vert}}\over 2}
278: [\sin^2\theta_1 \sin^2\phi_1+\sin^2\theta_2 \sin^2 \phi_2]
279: \end{equation}
280: while for field applied along the easy in-plane direction
281: \begin{equation}
282: \label{beasy}
283: G_{2\Vert}^{(ii)}=-{{K_{2\Vert}}\over 2}
284: [\sin^2\theta_1 \cos^2\phi_1+\sin^2\theta_2 \cos^2 \phi_2].
285: \end{equation}
286: In the former case the uniaxial in-plane anisotropy
287: $K_{2\Vert}>0$ favours the $y$ direction (perpendicular to the field and to
288: the stripes), and in the latter case it favours the $x$ direction
289: (parallel to the field and to the stripes).
290:
291: In terms of the free energy parameters, we define the out-of-plane anisotropy
292: field $H_{2\perp}=2K_{2\perp}/M_s$, the in-plane field $H_{2\Vert}=2K_{2\Vert}/M_s$,
293: and the dipolar field $H_{dip}=4 \pi M_s$.
294: It is customary to introduce the quality factor $Q=H_{2\perp}/H_{dip}$
295: as the ratio between the out-of-plane anisotropy field $H_{2\perp}$,
296: favouring the perpendicular direction ($z$), and the dipolar field $H_{dip}$,
297: favouring the film plane ($xy$). In the case under study, we have $Q<1$.
298: Finally, the saturation field is defined as
299: $H_{sat}=H_{2\perp}-H_{dip}N_{zz}$.\cite{Ramesh,RW}
300:
301: In the following, the equilibrium values of the polar and azimuthal
302: angles, obtained by minimizing the free energy Eq.~(\ref{G}),
303: will be denoted by the suffix ``{\it e}".
304: The frequencies of the spin-wave excitations
305: are evaluated\cite{RW,Ramesh} by the Landau-Lifshitz
306: equations of motion (see Appendix A for details).
307: Two modes, denoted by $\omega^+$ (acoustic mode)
308: and $\omega^-$ (optic mode),
309: are found for $H<H_{c}$ and a single (uniform\cite{Kittel})
310: mode for $H \ge H_{c}$.
311: It is worth noticing that, despite their names, neither of
312: the two modes is fully in-phase or fully out-of-phase:
313: their peculiar character is disclosed by the analysis of the
314: eigenvectors associated to the two modes.\cite{Ebels}
315: In fact, assuming the external magnetic field to be parallel to
316: the $x$ direction (along which the domains are infinitely long),
317: one finds that the acoustic mode with frequency $\omega^+$ is
318: characterized by dynamic fluctuations
319: such that $m_1^x(t)+m_2^x(t)=0$ and $m_1^y(t)+m_2^y(t) \ne 0$.
320: This corresponds to an out-of-phase precession of the dynamic moments
321: ${\bf M}_1$ and ${\bf M}_2$ in the direction parallel to the domain wall
322: and an in-phase precession perpendicular to the domain wall.
323: In contrast, for the optic mode the precession parallel to
324: the domain wall is in-phase and the precession
325: perpendicular to the domain wall is out-of-phase.\cite{Ebels}
326:
327: Hereafter we give the expressions of the spin-wave
328: frequencies\cite{limite} when the field is applied in plane along
329: the hard axis, see Eq.~(\ref{ahard}), or the easy axis, see
330: Eq.~(\ref{beasy}). In both cases one has
331: $\phi_{1e}=\phi_{2e}=0$.\cite{nota2}
332: %(upon variation of the field
333: %intensity the domain magnetizations ${\bf M}_1$ and ${\bf M}_2$ reorient themselves
334: %within the $xz$ plane).\cite{nota2}
335:
336: $\bullet$ Case ({\it i}): $H$ is along the hard in-plane axis.
337: %, see Eq.~(\ref{ahard}).
338: %The in-plane anisotropy favours the $y$ direction,
339: %perpendicular to the field and to the stripes.
340: For $H<H_c=H_{sat}$, the minimum free energy is obtained
341: for $\theta_{2e}=\pi-\theta_{1e}$
342: where $\sin \theta_{1e}=H/H_{sat}$. The frequencies of the acoustic
343: and optic modes are
344: \begin{eqnarray}
345: \label{omegapma}
346: (\omega^+/\gamma)^2&=&
347: \left\lbrack
348: H_{sat}^2-H^2
349: \Big(
350: 1- {{H_{dip}N_{yy}}\over {H_{sat}}}
351: \Big)
352: \right\rbrack
353: \left\lbrack
354: 1-{{H_{2\Vert}}\over {H_{sat}}}
355: \right\rbrack
356: \cr
357: \cr
358: (\omega^-/\gamma)^2&=& \left\lbrack H_{sat}^2-H^2 \right\rbrack
359: \left\lbrack 1+ {{ H_{dip}N_{yy}-H_{2\Vert} }\over {H_{sat}}}
360: \right\rbrack
361: \end{eqnarray}
362: respectively ($\gamma$ is the gyromagnetic factor).
363: For $H\ge H_c=H_{sat}$, one has $\theta_{1e}=\theta_{2e}=\pi/2$:
364: the stripe domain structure is wiped out, and the film is
365: homogeneously in-plane magnetized. The optic mode ($\omega^-$)
366: disappears and the acoustic one ($\omega^+$)
367: becomes the saturated, in-plane, uniform\cite{Kittel} mode with frequency
368: \begin{equation}
369: \label{omegaua}
370: (\omega/\gamma )^2=
371: \left\lbrack H - H_{2\Vert} \right\rbrack
372: \left\lbrack H - (H_{2\perp}-H_{dip}) \right\rbrack
373: \end{equation}
374:
375: $\bullet$ Case ({\it ii}): $H$ is along the easy in-plane axis.
376: %, see Eq.~(\ref{beasy}).
377: %The in-plane anisotropy favours the $x$ direction,
378: %parallel to the field and to the stripes.
379: For $H<H_c=H_{sat}-H_{2\Vert}$, the ground state has
380: $\theta_{2e}=\pi-\theta_{1e}$ where $\sin \theta_{1e}=H/H_c$
381: and the frequencies of the acoustic
382: and optic spin-wave excitations are
383: \begin{eqnarray}
384: \label{omegapmb}
385: (\omega^+/\gamma)^2&=&
386: H_{sat} H_c \cr
387: &\times& \left\lbrack
388: 1- \Big({{H}\over {H_c}}\Big)^2 \Big(1-{{H_{dip}N_{yy}}\over {H_c}}\Big)
389: \right\rbrack
390: \cr
391: \cr
392: (\omega^-/\gamma)^2&=& H_{sat} H_c \cr
393: &\times& \left\lbrack 1- \Big({{H}\over {H_c}}\Big)^2 \right\rbrack
394: \left\lbrack 1+ {{ H_{dip}N_{yy} }\over {H_{c}}}
395: \right\rbrack
396: \end{eqnarray}
397: For $H\ge H_c=H_{sat}-H_{2\Vert}$, one has $\theta_{1e}=\theta_{2e}=\pi/2$
398: and the frequency of the uniform mode is
399: \begin{equation}
400: \label{omegaub}
401: (\omega/\gamma )^2=
402: \left\lbrack H +H_{2\Vert} \right\rbrack
403: \left\lbrack H -(H_{2\perp}- H_{dip})+H_{2\Vert} \right\rbrack
404: \end{equation}
405:
406: \section{Results and discussion}
407:
408: \subsection{Spin-wave frequencies}
409:
410: \begin{figure}
411: \includegraphics[width=8cm,angle=0,bbllx=140pt,bblly=85pt,%
412: bburx=481,bbury=720pt,clip=true]{Fig3new.eps}
413: \caption{Full-line curves: frequencies of the acoustic ($\omega^{+}$),
414: optic ($\omega^{-}$) and uniform ($\omega$) mode, as a
415: function of the in-plane field $H$,
416: calculated using $H_{2\Vert}=0.9$kOe,
417: $H_{2\perp}=13.9$kOe, $H_{dip}=17$kOe, $N_{zz}=0.76$
418: ($H_{sat}=0.98$kOe), for three different cases:
419: (a) field applied
420: parallel to the stripes and along [-110], the hard in-plane
421: direction (see Eqs. (\protect\ref{omegapma}) and
422: (\protect\ref{omegaua}));
423: (b) field applied parallel to the stripes and
424: along [100], the intermediate in-plane direction;
425: (c) field applied
426: parallel to the stripes and along [110], the easy in-plane
427: direction (see Eqs. (\protect\ref{omegapmb}) and
428: (\protect\ref{omegaub})).
429: Open circles: experimental data.
430: }
431: \end{figure}
432:
433: In Fig.~3 the measured spin-wave frequencies are plotted as
434: a function of the intensity of the in-plane applied
435: magnetic field $H$. When the field is parallel to the hard in-plane
436: direction [-110] (Fig.~3a), two spin-wave modes are observed
437: for $0.4<H<0.9$ kOe. No differences in frequency, within experimental error, are
438: found upon increasing or decreasing the field intensity, thus
439: ruling out the possibility of metastability effects.
440: In contrast, a single spin-wave mode is observed for $H \ge 0.9$ kOe.
441: When the field is applied along either the intermediate [100] or the easy [110]
442: in-plane directions (see Fig.~3b and 3c, respectively),
443: a single spin-wave mode is experimentally observed
444: at any investigated field intensity.
445: The full-line curves in Fig. 3 are the theoretical spin-wave
446: frequencies.
447: %, calculated either analytically
448: %for field along the high symmetry directions [-110] and [110],
449: %or numerically for H along the intermediate in-plane axis [100].
450: %In the latter case (see Fig. 3b), a hybridization and repulsion
451: %between the acoustic and the optic mode is found.
452: The Hamiltonian parameters for the calculations were deduced
453: from previous experimental work on Fe/GaAs(001) films,\cite{esteso} where
454: a detailed fit of the BLS data was made for a whole set of samples with
455: different values of the iron thickness, ranging between $t_{Fe}=4$ \AA$~$
456: and 100 \AA. As $t_{Fe}$ increases, the out-of-plane
457: single-ion anisotropy $K_{2\perp}$ strongly decreases, while the dipolar field
458: $H_{dip}=4\pi M_s$, favouring in-plane magnetization, increases;
459: moreover, a biaxial in-plane anisotropy, favouring the [100] and [010]
460: crystallographic axes, gradually develops.
461: For the sample with $t_{Fe}=6$ \AA, only the data at fields high enough
462: for the magnetization to be homogeneous and in plane, were used
463: to obtain the fit. The dipolar field was estimated to be $H_{dip}=17$kOe,\cite{Trieste}
464: the out-of-plane anisotropy field $H_{2\perp}=13.9$kOe and the in-plane anisotropy
465: field $H_{2\Vert}=0.9$kOe.\cite{neglect}
466: Using these parameters, the demagnetization factor $N_{zz}$ was self-consistently
467: calculated\cite{KooyEnz} for different values of the domain aspect ratio $L/t$.
468: The variation of the domain size with the applied field intensity,
469: and thus the corrections to the frequency due to such variations,
470: were neglected as being of a higher order.\cite{RW,Ramesh}
471:
472: The best overall agreement between theory and experiment was found
473: assuming for the static demagnetization factor the value $N_{zz}=0.76$.
474: Such an assumption, although it corresponds to a probably too low
475: aspect ratio,\cite{nota3} is nevertheless able to justify the
476: presence of domains in the system in spite of the fact that,
477: for the considered iron thickness $t_{Fe}=6$\AA, one has
478: $H_{2\perp}<H_{dip}$ (quality factor $Q<1$).
479: In fact, for the onset of up/down stripe domains,
480: the condition $H_{2\perp}-H_{dip}N_{zz}>0$ has to be satisfied.
481: As a further support to the domain hypothesis, it is worth observing that
482: in epitaxial Co/Pt multilayers a perpendicular (up/down) stripe domain structure
483: was indeed experimentally observed at remanence by magnetic force microscopy,
484: while torque magnetometry measurements, providing $Q<1$,
485: %negative values for the
486: %effective out-of-plane anisotropy,
487: had suggested a preference for
488: in-plane orientation of the magnetization.\cite{Stamps}
489: %In spite of this, magnetic force microscopy images, taken at remanence,
490: %clearly showed a fine perpendicular domain structure consisting of regularly spaced
491: %stripes with domain period of the order of 100nm.\cite{Stamps}
492:
493: Note that the two modes are well observable only for the case of
494: field applied along the hard in-plane direction (Fig.~3a).
495: Otherwise (see Fig.~3b,c) one has a considerable shrinking of the
496: coexistence region of the two modes and moreover metastability
497: phenomena are likely to occur since the energy of the stripe
498: domain ground state is close to that of a homogeneous in-plane
499: configuration.
500:
501:
502: \subsection{Light scattering intensities}
503:
504: %In order to obtain further support for our interpretation of the
505: %coexistence of two spin-wave modes as a consequence of the
506: %onset, below a critical field $H_c$, of a perpendicularly
507: %magnetized up/down stripe domain structure, we also estimated
508: %the field dependence of the intensities of the two modes.
509:
510: For the experimental backscattering geometry
511: ($H$ in plane parallel to the $x$ axis and scattering plane perpendicular to $H$)
512: the incident light has p-polarization
513: (the optical wavevector and the optical incident electric field {\bf E}$_I$
514: have only $y$ and $z$ components)
515: while the scattered light has s-polarization. Then its intensity is
516: proportional to the square modulus of the $x$ component of the
517: polarization {\bf P} induced in the
518: film\cite{Cottam,Mills,DutcherJMMM,TesiDutcher,Zivieri}
519: \begin{eqnarray}
520: \label{tesi}
521: 4\pi P_x&=&m^x(t)~~ \Big\lbrack
522: -K E^y_I
523: \sin\theta_{e}\cos\theta_{e}
524: \cr
525: &-&2G_{44}M_s E^z_I
526: \cos\theta_{e}(\sin^2\theta_e-\cos^2\theta_e)
527: \Big\rbrack
528: \cr
529: \cr
530: &+& m^y(t)~~ \Big\lbrack
531: 2G_{44}M_s E^y_I\sin\theta_e -KE^z_I
532: \Big\rbrack
533: \cr
534: \cr
535: &+& m^z(t)~~\Big\lbrack
536: K E^y_I\sin^2\theta_e
537: \cr
538: &+& 2G_{44}M_s E^z_I
539: \sin\theta_e(\sin^2\theta_e-\cos^2\theta_e)
540: \Big\rbrack
541: \end{eqnarray}
542: \noindent where $K$ and $G_{44}$ denote the first- and second-order
543: (complex) magneto-optic coupling coefficients, respectively.
544: For the film with up/down domains, we assume that
545: $m^{\alpha}(t)=m_1^{\alpha}(t)+m_2^{\alpha}(t)$
546: ($\alpha=x,y,x$) since the size of the laser spot is much greater
547: than the lateral size of the domains.\cite{Zivieri}
548: Taking into account that $\phi_{1e}=0$ and $\sin\theta_{1e}=H/H_c=h$
549: and expressing the dynamic fluctuations of the magnetization in terms of the
550: fluctuations of the angle coordinates $\Delta \theta^{\pm}(t)$, $\Delta \phi^{\pm}(t)$
551: defined in Appendix A, we obtain
552: \begin{eqnarray}
553: \label{amplitude}
554: 4\pi P_x &=& \Delta \theta^-(t)
555: ~~\Big\lbrack
556: -KE^y_I ~h (1-h^2)
557: \cr&-&2G_{44}M_s E^z_I ~(1-h^2)(2h^2-1)
558: \Big\rbrack
559: \cr
560: &+& \Delta \phi^+(t)~~ \Big\lbrack
561: 2G_{44}M_s E^y_I~ h^2 -KE^z_I ~h
562: \Big\rbrack
563: \cr
564: &+&\Delta \theta^+(t)~~\Big\lbrack
565: -K E^y_I ~ h^3 \cr
566: &-& 2G_{44}M_s E^z_I ~ h^2 (2h^2-1)
567: \Big\rbrack
568: \end{eqnarray}
569: The field dependence of the intensities of the two modes can now be estimated
570: taking into account that the eigenvector associated with the acoustic mode
571: %with frequency $\omega^+$,
572: is characterized by $\Delta \theta^-(t)=\Delta \phi^-(t)= 0$
573: %and nonzero $\Delta \theta^+(t)$, $\Delta \phi^+(t)$ angle fluctuations,
574: while the optic mode
575: %, with frequency $\omega^-$
576: has $\Delta \theta^+(t)= \Delta \phi^+(t)= 0$.
577: %and nonzero $\Delta \theta^-(t)$, $\Delta \phi^-(t)$.
578: %Moreover, regardless of the direction of propagation of the incident or scattered
579: %light, one shoud consider that the electric field in the medium has a
580: %component parallel to the surface, $E^y$, that is
581: %greater than the perpendicular one, $E^z$.\cite{Mills}
582:
583: $\bullet$ Acoustic mode with frequency $\omega^+$.
584: The intensity $I^+(H)$ of the light scattered by the
585: acoustic mode has a maximum for $H \to H_c$ since the
586: fluctuations become very large. For zero field
587: the intensity is zero, $I^+(0)= 0$,
588: since $\Delta \theta^-(t)= 0$ and the coefficients
589: of the ``$^+$" angle fluctuations are zero for $h=0$.
590: %see Eq.~(\ref{amplitude}).
591: For $H\to + \infty$, the intensity vanishes, $I^+(H)\to 0$, since,
592: upon increasing $H$ above $H_c$, the $\omega^+$ mode
593: evolves into the uniform mode $\omega$ and
594: the fluctuations progressively decrease.
595: This behavior for $I^+(H)$ is similar to that of a perpendicularly
596: magnetized uniform film.\cite{Dutcher}
597:
598: $\bullet$ Optic mode with frequency $\omega^-$.
599: For $H=H_c$, the intensity of the light scattered by the optic mode
600: is zero $I^-(H_c)= 0$, since $\Delta \theta^+(t)=
601: \Delta\phi^+(t)=0$ and the coefficient of the $\Delta \theta^-(t)$
602: angle fluctuation is zero for $h=1$.
603: % see Eq.~(\ref{amplitude}).
604: For zero field, the intensity $I^-(0)$ can be finite provided
605: that the second order magneto-optic coupling coefficient is
606: nonzero, $G_{44}\ne 0$.
607:
608: The field dependence of the intensity of both the acoustic and the
609: optic modes, as deduced from Eq.~(\ref{amplitude}), is
610: qualitatively confirmed by the experimental spectra in Fig.~1.
611: The intensity of the former mode exhibits a neat maximum for field
612: values sligtly lower than $H_{c} \approx 0.9$ kOe and then it vanishes
613: as the field is reduced below about $0.5$ kOe.
614: The optic mode intensity, instead, shows a minimum approaching
615: $H_c$, in agreement with the theoretical predictions.
616:
617: \section{Conclusions}
618:
619: In conclusion, we have shown that a double-peaked structure is displayed
620: by the Brillouin Light Scattering spectra of Fe/GaAs(001) films with
621: $t_{Fe}=6$\AA$~$ when the field is applied in plane along the hard axis
622: and is smaller than a critical value $H_c=0.9$ kOe.
623: The existence of two peaks in the BLS spectrum should be the
624: general feature of a film with a perpendicular domain
625: structure (it is irrelevant whether the magnetization is canted or not).
626: This feature disappears when $H \ge H_c$ and the magnetization lies
627: in the film plane.
628: The reason why the unravelling of such a two-peaked structure
629: in the Brillouin light scattering spectra of ultrathin magnetic
630: films is so rare might well be that many conditions have to be
631: simultaneously satisfied.
632: In fact, the optic mode, with frequency $\omega^-$,
633: has enough intensity in an appreciable
634: range of fields only if $G_{44}\ne 0$
635: and $H_c$ are not too small.
636: %since for $G_{44}=0$ its intensity
637: %vanishes in both the limits $H \to 0^+$ and $H \to H_c^-$.
638: In contrast, the acoustic mode, with frequency $\omega^+$,
639: has more chances to be observed since its intensity,
640: though always vanishing in the $H\to 0^+$ limit, is expected
641: to increase as $H$ increases and to reach a maximum just at $H_c$.
642: Another stringent requirement for the simultaneous observation
643: of two modes is that the competing out-of-plane anisotropy field $H_{2\perp}$
644: and easy-plane dipolar anisotropy field $H_{dip}$ are of comparable
645: magnitude and that $H_{2\perp}-H_{dip}N_{zz}>0$,
646: so that a perpendicularly magnetized up/down
647: domain structure is energetically favoured for $H \to 0^+$.
648: This seems just to be the case of the Fe/GaAs films
649: with $t_{Fe}=6$\AA. In fact, for higher Fe thickness, one has
650: $H_{dip} \gg H_{2\perp}$, so that a homogeneous in-plane magnetized
651: ground state is realized, while, upon reducing the Fe thickness,
652: one would expect $H_{dip} \ll H_{2\perp}$ and
653: a single spin-wave mode to be excited with respect to a homogeneous,
654: perpendicularly magnetized metastable state.
655: %This mode would be expected to display the peculiar field dependence
656: %of the frequency and intensity, firstly observed by Dutcher {\it et al.}
657: %in ultrathin Fe/Cu films,\cite{Dutcher}
658: %and associated with the reorientation of the magnetization from an out-of-plane
659: %to an in-plane direction. However, the observation of such a single,
660: %well-defined spin-wave mode in Fe/GaAs films is prevented by the fact that the system
661: %becomes superparamagnetic for $t_{Fe}<6$\AA.\cite{esteso} This opens the
662: %way to further experimental and theoretical study of the magnetic
663: %excitations in Fe/GaAs(001) ultrathin magnetic films in the absence of true
664: %long range order.
665:
666: We hope that the results of this paper can stimulate other
667: experimental groups to directly visualize the domain pattern, {\it
668: e.g.} using magnetic microscopy techniques as a function of the
669: external magnetic field intensity.\cite{Meyer} This should be done
670: {\it in situ}, because the magnetic anisotropy is strongly
671: affected by the presence of a protective overlayer, so that
672: formation of magnetic domains can be prevented.\cite{esteso}
673: %In any case, we think that the present experimental and theoretical
674: %study of spin-wave occurrence in ultrathin Fe/GaAs films can
675: %be useful in order to stimulate a more systematic search for
676: %other magnetic film systems where the simultaneous presence
677: %of an acoustic and an optic mode should be possible, reproducible and easily
678: %detectable using high resolution Brillouin Light Scattering
679: %techniques.\cite{rassegnaCarlotti}
680:
681: \begin{acknowledgments}
682: The authors gratefully acknowledge financial support from the Italian
683: Ministery for the Instruction, University and Research (MIUR), under projects
684: PRIN 2003025857 and FIRB RBNE017XSW. This work was performed in the
685: framework of the joint CNR-MIUR programme (Legge 16/10/2000, Fondo FISR).
686: \end{acknowledgments}
687:
688: \appendix
689: \section{Spin-wave frequencies}
690:
691: The static equilibrium configuration of the system is obtained
692: by minimizing the free energy $G$, Eq.~(\ref{G}), with respect
693: to the polar and azimuthal variables
694: %$\theta_1$, $\phi_1$, $\theta_2$, $\phi_2$,
695: while the frequencies of the spin-wave excitations
696: are evaluated \cite{RW} by the Landau-Lifshitz
697: equations of motion
698: \begin{equation}
699: {{d \theta_{1,2}}\over {dt}}= - {\gamma\over {\sin\theta_{1,2}}}
700: {{\partial {\cal G}}\over {\partial \phi_{1,2}}},
701: ~~~~
702: {{d \phi_{1,2}}\over {dt}}= {\gamma\over {\sin\theta_{1,2}}}
703: {{\partial {\cal G}}\over {\partial \theta_{1,2}}}
704: \end{equation}
705: where ${\cal G}=2G/M_s$. The small oscillations of the
706: system in response to an external perturbation
707: are obtained by expanding ${\cal G}$ in a Taylor series about its equilibrium value
708: ${\cal G}_e$ up to the second order.
709: Next, assuming for the set of variables
710: a harmonic time dependence with frequency $\omega$ and
711: introducing the normal coordinates $\Delta \theta^{\pm}=\Delta \theta_1 \pm \Delta
712: \theta_2$ and $\Delta \phi^{\pm}=\Delta \phi_1 \pm \Delta \phi_2$
713: (where $\Delta$ denotes a small variation), the
714: equations of motion can be rewritten in matrix form as
715: \begin{equation}
716: \label{matricial}
717: \left\lbrack
718: \begin{array}{l}
719: ~A^+~ \; -iz \; ~~0~ \; ~~B^+ \;
720: \cr
721: ~iz \; ~~~~~C^+~ \; B^-~\; 0~\;
722: \cr
723: ~~0~~ \; ~~~B^-~\; A^-\; -i z\;
724: \cr
725: ~B^+~ \; ~~0\;~~~~i z \;~~~ C^-
726: \end{array}
727: \right\rbrack
728: \left\lbrack
729: \begin{array}{l}
730: \Delta \theta^+ \;
731: \cr
732: \Delta \phi^+
733: \cr
734: \Delta \theta^- \;
735: \cr
736: \Delta \phi^- \;
737: \end{array}
738: \right\rbrack =0
739: \end{equation}
740: where $z=\omega ~\sin \theta_{1e}$ and $A^{\pm}={\cal G}_{11} \pm
741: {\cal G}_{13}$, $B^{\pm}={\cal G}_{12} \pm {\cal G}_{23}$,
742: $C^{\pm}={\cal G}_{22} \pm {\cal G}_{24} $. By ${\cal G}_{ij} = {
743: {\partial^2 {\cal G} }\over {\partial X_i \partial X_j}
744: }\Big\vert_e$ we denote the second partial derivatives of the free
745: energy with respect to the angular variables
746: ($X_1=\theta_1,X_2=\phi_1,X_3=\theta_2,X_4=\phi_2$). The
747: frequencies of the normal modes are obtained by imposing the
748: condition for nontriviality of the solutions of Eq.
749: (\ref{matricial}), {\it i.e.} the vanishing of the matrix
750: determinant.
751:
752: For $H \ge H_c$ the ground state is homogeneously in-plane
753: magnetized, ($\theta_{1e}=\theta_{2e}=\pi/2$ and $N_{zz}=1$) and
754: it results that $A^+=A^-={\cal G}_{11}$, $B^+=B^-=0$, and
755: $C^+=C^-={\cal G}_{22}$, so that a single, uniform
756: mode\cite{Kittel} is obtained, with frequency $\big( \omega/\gamma
757: \big)^2 = {\cal G}_{11} {\cal G}_{22}$.
758:
759: \begin{thebibliography}{99}
760:
761: \bibitem{Brockmann}
762: M. Brockmann, M. Zolfl, S. Miethaner, G. Bayreuther, J. Magn. Magn. Mater.
763: {\bf 198}, 384 (1999).
764:
765: \bibitem{Gester}
766: M. Gester, C. Daboo, R. J. Hicken, S. J. Gray, A. Ercole, J. A. C. Bland,
767: J. Appl. Phys. {\bf 80}, 347 (1996).
768:
769: \bibitem{Steinmueller}
770: S. J. Steinm\"uller, M. Tselepi, V. Strom, J. A. C. Bland, J. Appl. Phys. {\bf 91},
771: 8679 (2002).
772:
773: \bibitem{Monchesky}
774: T. L. Monchesky, B. Heinrich, R. Urban, K. Myrtle, M. Klaua, J. Kirschner,
775: Phys. Rev. B {\bf 60}, 10242 (1999).
776:
777:
778: \bibitem{Prinz}
779: G. A. Prinz, ``Spin-polarized Transport", Physics Today {\bf 48}, 58 (1995).
780:
781: \bibitem{libro}
782: {\it Spin Dynamics in Confined Magnetic Structures I},
783: Eds. B. Hillebrands and K. Ounadjela, Springer
784: Topics in Applied Physics vol. 83 (Springer, Berlin, 2002).
785:
786: \bibitem{Bland}
787: S. Mc Fail, C. M. Gurtler, F. Montaigne, Y. B. Xu, M. Tselepi and J. A. C. Bland,
788: Phys. Rev. {\bf 67}, 24409 (2003).
789:
790: \bibitem{esteso}
791: M. Madami, S. Tacchi, G. Carlotti, G. Gubbiotti, R. L. Stamps,
792: Phys. Rev. B {\bf 69}, 144408 (2004).
793:
794: \bibitem{Falco}
795: A. Murayama, K. Hyomi, J. Eickmann, C. M. Falco, J. Magn. Magn.
796: Mater. {\bf 198-199}, 372 (1999).
797:
798: \bibitem{rassegna}
799: P. Politi, Comments Cond. Mat. Phys. {\bf 18}, 191 (1998),
800: and references therein.
801:
802: \bibitem{EPJB}
803: P. Politi, M. G. Pini, Eur. Phys. J. B {\bf 2}, 475 (1998).
804:
805: \bibitem{Pescia}
806: C. Stamm, F. Marty, A. Vaterlaus, V. Weich, S. Egger, U. Maier,
807: U. Ramsperger, H. Fuhrmann, D. Pescia,
808: Science {\bf 282}, 449 (1998); J. Harris and D. Awschalom,
809: ``Thin films squeeze out domains", Physics World, p. 19
810: (January 1999).
811:
812: \bibitem{apparatus}
813: L. Albini, G. Carlotti, G. Gubbiotti, M. Madami, S. Tacchi,
814: J. Appl. Phys. {\bf 89}, 7383 (2001).
815:
816: \bibitem{apparatus2}
817: G. Gubbiotti, G. Carlotti, M. Madami, S. Tacchi, L. Verdini, Phys.
818: Stat. Sol. 189, 403 (2002).
819:
820: \bibitem{ghost}
821: http://ghost.fisica.unipg.it/
822:
823: \bibitem{Sander}
824: J. R. Sandercock, in: Light Scattering in Solids III, Eds. M. Cardona and
825: G. Guntherodt, Springer Ser. Topics Appl. Phys., Vol. 51, Springer-Verlag
826: 1982 (p.173)
827:
828: \bibitem{RW}
829: M. Ramesh and P. E. Wigen, J. Magn. Magn. Mater. {\bf 74}, 123
830: (1988).
831:
832: \bibitem{Ramesh}
833: M. Ramesh, E. W. Ren, J. O. Artman, and M. H. Kryder,
834: J. Appl. Phys. {\bf 64}, 5485 (1988).
835:
836: \bibitem{Ebels}
837: U. Ebels, L. D. Buda, K. Ounadjela, and P. E. Wigen,
838: ``Small amplitude dynamics of nonhomogeneous magnetization
839: distributions: the excitation spectrum of stripe domains",
840: in Ref.~\onlinecite{libro}, pp.~167-216.
841:
842: \bibitem{notaBLS}
843: Clearly within DMFMR theory one neglects the wavevector
844: of the spin waves ($q \approx 10^5$ cm$^{-1}$ for
845: a spin wave revealed in a BLS) with respect to the zone
846: boundary wavevector ($q_{ZB} \approx 10^8$ cm$^{-1}$).
847:
848: \bibitem{KooyEnz}
849: C. Kooy and U. Enz, Philips Res. Rep. {\bf 15}, 7 (1960).
850:
851: \bibitem{Kittel}
852: C. Kittel, Phys. Rev. {\bf 73}, 155 (1948).
853:
854: \bibitem{limite}
855: It is worth noticing that in the absence of in-plane anisotropy, $H_{2\Vert} \to 0$,
856: the results of Ref. \onlinecite{RW} for the mode frequencies (Eqs.~31, 32, 35)
857: are correctly recovered.
858:
859: \bibitem{nota2}
860: This is obvious for field applied along the
861: easy axis. For field along the hard axis,
862: this is verified provided that $H_{2\Vert}<H_{sat}$,
863: as is the case for the system under study;
864: when $H_{2\Vert}>H_{sat}$,
865: the excitation of two modes would not be allowed since the ground
866: state is characterized by a homogeneous magnetization lying within
867: the film plane.
868: %(and possibly subject to an in-plane reorientation for
869: %external field applied along a hard in-plane axis).
870:
871: \bibitem{Trieste}
872: L. Giovannelli, C. -S. Tian, P. L. Gastelois, G. Panaccione, M.
873: Fabrizioli, M. Hochstrasser, M. Galaktionov, C. H. Back and G.
874: Rossi, Physica B {\bf 345}, 177 (2004).
875:
876: \bibitem{neglect}
877: For the sake of simplicity we neglect the
878: anisotropy field $H_{4\Vert} \approx 0.1$kOe associated to a biaxial in-plane
879: anisotropy that favours the [110] and [-110] axes, since it is much smaller than
880: the uniaxial in-plane anisotropy fied $H_{2\Vert}$.\cite{esteso}
881:
882: \bibitem{nota3}
883: A self-consistent calculation \protect{\cite{KooyEnz}} gives
884: $L/t \approx 15$ in correspondence to $N_{zz}=0.76$.
885: %OK ho ricontrollato
886:
887: \bibitem{Stamps}
888: L. Louail, K. Ounadjela, M. Hehn, K. Khodjaoui, M. Gester, H. Danan,
889: R. L. Stamps, J. Magn. Magn. Mat. {\bf 165}, 387 (1997);
890: L. Louail, K. Ounadjela, R. L. Stamps, J. Magn. Magn. Mat. {\bf 167},
891: L189 (1997); R. L. Stamps, L. Louail, M. Hehn, M. Gester, K. Ounadjela,
892: J. Appl. Phys. {\bf 81}, 4751 (1997).
893:
894: \bibitem{Cottam}
895: W. Wettling, M. G. Cottam, J. R. Sandercock, J. Phys. C: Solid St. Phys.
896: {\bf 8}, 211 (1975).
897:
898:
899: \bibitem{Mills}
900: R. E. Camley and D. L. Mills, Phys. Rev. B {\bf 18}, 4821 (1978).
901:
902: \bibitem{DutcherJMMM}
903: J. F. Cochran, J. R. Dutcher, J. Magn. Magn. Mat. {\bf 73}, 299 (1988).
904:
905: \bibitem{TesiDutcher}
906: J. R. Dutcher, Ph. D. Dissertation, Simon Fraser University
907: (1988). Available on microfiche from the National Library of
908: Canada, Ottawa, Ontario, Canada, K1A0N4. Notice that
909: $\theta_e=\pi/2-\alpha_{Dutcher}$.
910:
911: \bibitem{Zivieri}
912: R. Zivieri, L. Giovannini, P. Vavassori, ``Theory of Brillouin
913: cross section from magnetic nanostructured multilayers", in
914: {\it Magnetic Nanostructures}, Editor H. S. Nalwa, Volume 1,
915: (American Scientific, 2002), and references therein.
916:
917: \bibitem{Dutcher}
918: %intensita`
919: J. R. Dutcher, J. F. Cochran, I. Jacob, W. F. Egelhoff, Jr.,
920: Phys. Rev. B {\bf 39}, 10430 (1989).
921:
922: \bibitem{Meyer}
923: G. Meyer, A. Bauer, T. Crecelius, I. Mauch, G. Kaindl,
924: Phys. Rev. B {\bf 68}, 212404 (2003), and
925: references therein.
926:
927: %\bibitem[34]{rassegnaCarlotti}
928: %G. Carlotti and G. Gubbiotti, J. Phys. C {\bf 14}, 8199-8233 (2002)
929: %and references therein.
930:
931: %\end{references}
932: \end{thebibliography}
933: \end{document}
934:
935: