1: %\documentclass[prb,preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}

2: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}

3: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}

4: %\documentclass[prb,twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}

5: 

6: \usepackage{graphicx}% Include figure files

7: \usepackage{dcolumn}% Align table columns on decimal point

8: \usepackage{bm}% bold math

9: \usepackage{graphicx}

10: \usepackage{epsfig}

11: \include{graphic}

12: 

13: \begin{document}

14: \preprint{APS/123-QED}

15: \title{Resonances in Ferromagnetic Gratings Detected by Microwave Photoconductivity}

16: 

17: \author{Y. S. Gui, S. Holland, N. Mecking, and C. -M. Hu\footnote{Electronic address: hu@physnet.uni-hamburg.de}}

18: 

19: \affiliation{Institut f\"ur Angewandte Physik und Zentrum f\"ur

20: Mikrostrukturforschung, Universit\"at Hamburg, Jungiusstra\ss e

21: 11, 20355 Hamburg, Germany}

22: 

23: \date{\today}

24: 

25: \begin{abstract}

26: 

27: We investigate the impact of microwave excited spin excitations on

28: the DC charge transport in a ferromagnetic (FM) grating. We

29: observe both resonant and nonresonant microwave photoresistance.

30: Resonant features are identified as the ferromagnetic resonance

31: (FMR) and ferromagnetic antiresonance (FMAR). A macroscopic model

32: based on Maxwell and Landau-Lifschitz equations reveals the

33: macroscopic nature of the FMAR. The experimental approach and

34: results provide new insight in the interplay between photonic,

35: spintronic, and charge effects in FM microstructures.

36: 

37: \end{abstract}

38: 

39: \pacs{73.50.Pz, 41.20.Jb, 76.50.+g, 42.79.Dj}

40: 

41: %41.20.Jb    Electromagnetic wave propagation; radiowave

42: %propagation (for light propagation, see 42.25.Bs; for

43: %electromagnetic waves in plasma, see 52.35.Hr; for ionospheric and

44: %magnetospheric propagation, see 94.20.Bb and 94.30.Tt)

45: %

46: %42.25.Bs    Wave propagation, transmission and absorption

47: %

48: %42.79.Dj    Gratings

49: %

50: %76.50.+g    Ferromagnetic, antiferromagnetic, and ferrimagnetic

51: %resonances; spin-wave resonance (see also 75.30.Ds Spin

52: %waves)

53: %

54: %73.50.Pz    Photoconduction and photovoltaic effects

55: %

56: %75.47.-m    Magnetotransport phenomena; materials for

57: %magnetotransport  (for spintronics, see 85.75.-d; see also

58: %72.15.Gd, 73.50.Jt, 73.43.Qt, and 72.25.-b in transport phenomena)

59: %

60: %84.40.-x    Radiowave and microwave (including millimeter wave)

61: %technology

62: 

63: \maketitle The connection between the DC and high frequency

64: response of the metal to external fields looks like a one-way

65: path. On the one hand, it is text book knowledge that due to the

66: ohmic dissipation, the DC conductivity $\sigma_{0}$ determines the

67: skin depth $\delta=\sqrt{2/\mu_{0}\sigma_{0}\omega}$ of the

68: electromagnetic radiation with the frequency $\omega = 2\pi f$,

69: where $\mu_0$ is the permeability of vacuum. On the other hand,

70: there is little knowledge about the inverse effect of the high

71: frequency response on the DC transport in metals, which is in

72: contrast to the case of semiconductors, where a whole zoo of

73: photoconductivity phenomena, ranging from the intrinsic,

74: extrinsic, to the bolometric effect, are all based on such an

75: influence.

76: 

77: Recently, a breakthrough has been achieved in ferromagnetic (FM)

78: metals. By combining the giant magnetoresistance effect of a FM

79: multilayer with the microwave absorption, high frequency

80: resonances were detected by measuring the DC resistance

81: \cite{Tsoi2000}. It was the first photoconductivity experiment on

82: FM multilayers, which bridged static and dynamic properties, and

83: paved the way for recent highlights of generating microwave

84: oscillations by a spin-polarized DC current \cite{Kiselev2003}.

85: Despite of broad interest in studying the interplay of static and

86: dynamic responses in FM multilayers, the basic question of the

87: impact of the high frequency response on the DC transport in a

88: single layer FM metal remains open.

89: 

90: In this paper, we answer this question by performing microwave

91: photoconductivity measurements directly on a single layer FM

92: microstrip. Our primary aim is to explore the bolometric effect

93: \cite{Hu2003} in the FM metal, which may bridge the high frequency

94: absorbance $A(\omega)$ with the DC resistance change $\Delta R$

95: via a simple relation

96: \begin{equation}

97: \Delta R = S\cdot A(\omega),

98: \end{equation}

99: where $S = \frac{\partial R}{\partial

100: T}\frac{P_{0}\tau_{e}}{C_{e}}$ is a sensitivity parameter that

101: depends on the specific heat $C_{e}$ of electrons, the incident

102: power $P_{0}$ of the radiation, and the energy relaxation time

103: $\tau_{e}$ of photo-excited spin/charges. The relation was

104: previously only known for semiconductors \cite{Neppl1979}. We

105: demonstrate that based on the interplay of the spin dynamics and

106: the DC charge transport, both the ferromagnetic resonance (FMR)

107: \cite{Kittel} and ferromagnetic antiresonance (FMAR)

108: \cite{Yager1949} can be detected by the photoconductivity

109: technique. Using a model based on Maxwell and Landau-Lifschitz

110: equations, we reveal the unique macroscopic nature of the FMAR in

111: the FM grating, which has the potential to integrate spintronic

112: and photonic features in FM microstructures.

113: 

114: Our experiments are performed on an array of Ni$_{80}$Fe$_{20}$

115: (Permalloy, Py) microstrip with a width $W$= 50 $\mu $m and a

116: thickness $d$ = 60 nm. As illustrated in insets of Fig. 1, the

117: strip has a total length $L \approx$ 10 cm and runs meandering in

118: a square of about 3$\times$3 mm$^{2}$, forming 30 periods of FM

119: grating with a period $a$ = 70 $\mu$m. The Py strip is deposited

120: on a semi-insulating GaAs substrate using photolithography and

121: lift-off techniques. The DC conductivity $\sigma_{0}$ of the Py

122: strip is determined to be 3.2 (5.0) $\times$ 10$^{4}$

123: $\Omega^{-1}$cm$^{-1}$ at 300 (4.2) K. A swept-signal generator is

124: connected with a circular oversized waveguide, which brings the

125: microwave radiations with $f$ between 17.5 - 20 GHz down to the

126: sample set in a cryostat.

127: 

128: \begin{figure} [t]

129: \begin{center}

130: \epsfig{file=fig1.eps,height=7.2 cm} \caption{(color online). (a)

131: Parallel (\textit{$\theta $} = 0$^{o}$) and (b) perpendicular

132: (\textit{$\theta $} = 90$^{o}$) AMR effect measured with the

133: applied magnetic field and current bias as shown in the

134: inset.}\label{Fig.1}

135: \end{center}

136: \end{figure}

137: 

138: Before discussing the photoconductivity of the Py strip, we show

139: in Fig. 1 the static property of our sample without microwave

140: radiations. By applying the external magnetic field $ H =

141: B/\mu_{0}$ along the easy axis parallel (\textit{$\theta $} =

142: 0$^{o}$) to the current flow in the strip, we measure the

143: anisotropic magnetoresistance (AMR) and plot it \cite{AMR} in Fig.

144: 1(a). The sharp minimum at $\pm$ 1.2 mT corresponds to the

145: coercive field of the strip \cite{Adeyeye1996}, which increases

146: with increasing the angle \textit{$\theta $} (not shown). At

147: \textit{$\theta $} = 90$^{o}$ when the applied B field is along

148: the hard axis perpendicular to the strip plane, perpendicular AMR

149: is measured and plotted in Fig. 1(b). The estimated saturation

150: magnetization ($M_0$) is about 1.2 T/$\mu_{0}$ and the normalized

151: AMR is about 3.2{\%}, both in agreement with earlier reports

152: \cite{Adeyeye1996}.

153: 

154: We perform the photoconductivity experiment at \textit{$\theta $}

155: = 90$^{o}$ in the Faraday configuration with the microwave wave

156: vector $\textbf{k}\parallel\textbf{B}$. Fig. 2 shows typical

157: photoresistance traces measured as a function of the B field at

158: 4.2 K for different microwave frequencies. The curves are

159: vertically offset for clarity. A DC current of $I$ = 90 $\mu $A is

160: applied. The radiation-induced voltage change $\bigtriangleup V$

161: is measured via lock-in technique by modulating the microwave

162: power with a frequency of 123 Hz. The photoresistance

163: $\bigtriangleup R =\bigtriangleup V/I$ measures the

164: microwave-induced DC magnetoresistance change of the Py strip. In

165: addition to a nonresonant background photoresistance at the order

166: of 10 m$\Omega$, which is about a few ppm of the DC

167: magnetoresistance $R$ of the Py strip, we observe clearly two

168: resonances. One appears as a peak and the other as a dip. The

169: resonant field for both shifts with $f$. We find that

170: $\bigtriangleup R$ increases with increasing power. The data shown

171: in Fig. 2 are measured by setting the output power of the

172: swept-signal generator at 24 dbm, however, the power that reaches

173: the sample via the long waveguide is significantly reduced. At

174: 17.75 GHz when $f$ approaches the cut off frequency of the

175: waveguide, $\bigtriangleup R$ is obviously reduced.

176: 

177: \begin{figure} [t]

178: \begin{center}

179: \epsfig{file=fig2.eps,height=7.2 cm} \caption{(color online).

180: Microwave photoresistance (vertically offset for clarity) of the

181: Py strip measured as a function of the magnetic field at 4.2 K and

182: at different microwave frequencies. The inset shows the

183: measurement configuration.} \label{Fig.2}

184: \end{center}

185: \end{figure}

186: 

187: To shed light onto the observed photoconductivity effect, we begin

188: by analyzing the magnetodynamic response function of our sample.

189: The dynamic susceptibility tensor $\widehat{\chi}$, which links

190: the dynamic magnetization $\textbf{m}$ and the dynamic magnetic

191: field $\textbf{h}$ via $\textbf{m}=\widehat{\chi}\cdot

192: \textbf{h}$, can be obtained by solving Landau-Lifschitz equation

193: \cite{Hubert}. For simplicity, we restrict our analysis to the

194: field range of $H

195: > M_{0}$ where resonances are observed. Since in our sample $d \ll L, W$,

196: we start by treating it as a 2D film, taking into account the

197: demagnetization field but neglecting the anisotropy and the

198: exchange field.  We get the dynamic permeability tensor

199: $\widehat{\mu}=\widehat{1}+\widehat{\chi} = \begin{pmatrix}

200: \mu_{L} & \mu_{T} & 0 \cr -\mu_{T}& \mu_{L} & 0 \cr 0 & 0 & 1

201: \end{pmatrix}$ with the longditudinal ($\mu_{L}$) and transversal

202: ($\mu_{T}$) complex permeability given by

203: \begin{eqnarray}

204: \mu_{L}=1+\frac{\omega_{M}(\omega_{r}-i\alpha\omega)}{(\omega_{r}-i\alpha\omega)^{2}-\omega^{2}},

205: \cr

206: \mu_{T}=\frac{i\omega_{M}\omega}{(\omega_{r}-i\alpha\omega)^{2}-\omega^{2}}.

207: \end{eqnarray}

208: Here, $\alpha$ is the dimensionless Gilbert damping parameter. We

209: define $\omega _M = \gamma M_0 $ and $\omega _r = \gamma (H - M_0

210: )$, with $\gamma = {g\mu _B \mu_{0}}/ \hbar$ the gyromagnetic

211: ratio which depends on the $g$ factor and the Bohr magneton $\mu

212: _B$.

213: 

214: The dynamic permeability tensor $\widehat{\mu}$ describes the

215: gyrotropic response of the FM metal. In the Faraday configuration,

216: its eigenvalues can be found by solving the equation

217: $\textbf{k}(\textbf{k}\cdot

218: \textbf{h})+(k_{0}^{2}\epsilon\widehat{\mu}-\textbf{k}^{2})\textbf{h}

219: = 0$ deduced from Maxwell equations \cite{Jackson}. We obtain

220: \begin{equation}

221: \mu_{\pm}=\mu_{L}\mp i\mu_{T}

222: =\frac{\omega_{r}+\omega_{M}\mp\omega-i\alpha\omega}{\omega_{r}\mp\omega-i\alpha\omega}

223: \end{equation}

224: which define two circular polarized electromagnetic eigenmodes

225: propagating in the FM film, whose wave vectors are given by

226: $k_{\pm}^{2}=\epsilon\mu_{\pm}k_{0}^{2}$. Here $\epsilon \approx

227: i\sigma_{0}/\epsilon_{0}\omega$ is the complex permittivity of the

228: FM film, $\epsilon_0$, $c$, and $k_{0} = \omega/c$ are the

229: permittivity, the velocity and the wave vector of light in vacuum.

230: The $k_{+}$ mode results from the strong coupling of the right

231: circular electromagnetic wave with the magnetization, which

232: excites the FMR at the resonant frequency $\omega _r$. The FMR is

233: inactive for the left circular electromagnetic wave and hence the

234: $k_{-}$ mode is only weakly influenced by the magnetization.

235: 

236: In Fig. 3(a), we plot the magnetic-field dispersion of the peak

237: (solid square) and dip (open circle) measured from

238: photoconductivity spectra. By fitting the dispersion of the peak

239: using the relation $\omega _r = \gamma (H - M_0 )$, we obtain

240: $\gamma$ = 183$\mu_{0}$ GHz/T (corresponding to $g$ = 2.08) which

241: agrees well with the published values \cite{Nibarger2003}, and

242: $M_{0}$ = 1.15 T/$\mu_{0}$ which is consistent with the value (1.2

243: T/$\mu_{0}$) estimated from the AMR effect. Therefore we identify

244: the resonant peak of the photoresistance as the FMR, which has the

245: microscopic origin of Larmor precession of spins in the FM metal

246: \cite{Kittel}.

247: 

248: \begin{figure} [t]

249: \begin{center}

250: \epsfig{file=fig3.eps,height=9 cm} \caption{(color online). (a)

251: The measured resonance positions for the photoresistance peak

252: (solid square) are fitted to the FMR dispersion (solid line). The

253: measured (open circle) FMAR dispersion is compared with that

254: calculated using $\mu_{+}$ (dashed line) and $\mu_{L}$ (dotted

255: curve). (b) $\mu_{+}$ and (c) $\mu_{L}$ are calculated at

256: $\omega/2\pi$ = 35 GHz, using the parameters $M_0$ = 1.15

257: T/$\mu_{0}$, $\alpha$ = 0.0075, and $\gamma$ = 183$\mu_{0}$ GHz/T.

258: Arrows indicate the condition for $Re(\mu)$ = 0.} \label{Fig.3}

259: \end{center}

260: \end{figure}

261: 

262: With fitted values for $\gamma$ and $M_{0}$, we calculate and plot

263: in Fig. 3(b) the B-field dependence of $\mu_{+}$ for $\omega/2\pi$

264: = 35 GHz. From a line shape fit that we will describe later, we

265: take $\alpha$ = 0.0075. The real part of $\mu_{+}$ has two zeros.

266: At the zero located at $\omega = \omega_{r} = \gamma (H - M_0 )$,

267: which is indicated by the upward arrow in Fig. 3(b), $Im(\mu_{+})$

268: shows a pole. This is the macroscopic definition of the FMR based

269: on the magnetodynamic response function. It corresponds to the

270: condition of resonantly enhanced absorption due to the FMR. For $H

271: > 0$, $\mu_{-}$ has neither pole nor zero (not shown), because the

272: FMR is inactive to the left circular polarized electromagnetic

273: wave.

274: 

275: Note that there is a second zero for $Re(\mu_{+})$ located at

276: $\omega = \omega_{+} = \gamma H$, which is indicated by the

277: downward arrow in Fig. 3(b). At this condition, $Im(\mu_{+})$ is

278: also nearly zero, hence the dynamic susceptibility $\chi_{+}

279: \simeq$ -1. This is the resonant condition for the FMAR of the FM

280: film at which the ohmic dissipation due to eddy currents is

281: suppressed. Early microwave transmission experiments performed on

282: thick ($d

283: > \delta$) FM films have confirmed enhanced transmission and reduced

284: absorption at the FMAR \cite{Yager1949}. One would therefore

285: attribute the resonant photoresistance dip in Fig. 2 to the FMAR.

286: However, as shown in Fig. 3(a), the measured resonances for the

287: dips (open circle) lie far away from the dashed line plotted for

288: the relation $\omega_{+} = \gamma H$.

289: 

290: The significant discrepancy reflects an intriguing macroscopic

291: nature of the FMAR in the microstructured FM layer. As shown in

292: insets of Figs. 1 \& 2, our grating has a large $L/W$ ratio with a

293: period ($a$ = 70 $\mu$m) much smaller than the wavelength of the

294: imposed microwave ($\lambda \approx$ 1.5 cm). Similar metallic

295: gratings with subwavelength period have long been investigated,

296: which display anomalous optical effects \cite{Neviere}. Recently,

297: they have got renewed interest due to exotic photonic effects

298: showing extraordinary optical transmission \cite{Porto}. We

299: demonstrate here that the FM grating has its own unique

300: macroscopic optical behavior based on the spin dynamics. In the

301: simplest approximation, we treat the grating as a linear polarizer

302: with a permittivity tensor $\widehat{\epsilon} =

303: \begin{pmatrix} 1 & 0 & 0 \cr 0 & \epsilon & 0 \cr 0 & 0 &

304: 1\end{pmatrix}$, in which we neglect the microscopic geometric

305: details of the patterned FM film, but focus instead on its

306: macroscopic characteristics of the anisotropic conductivity. By

307: using $\widehat{\epsilon}$ instead of $\epsilon$ in Maxwell

308: equations, we find that the eigenmode propagating in the FM

309: grating is nearly linear polarized with the wave vector given by

310: $k_{L}^{2}\approx\epsilon\mu_{L}k_{0}^{2}$.

311: 

312: The characteristics of $\mu_{L}$ plotted in Fig. 3(c) looks at

313: first glance similar to that of $\mu_{+}$. Indeed, both define the

314: same FMR since a linear polarized electromagnetic wave can be

315: split equally into a left and a right circular polarized wave,

316: with only the right one active for FMR. The characteristic

317: difference between $\mu_{L}$ and $\mu_{+}$ lies in the FMAR. From

318: $Re(\mu_{L}) = 0$, we get $\omega_{L} = \gamma \sqrt{H(H - M_0 )}$

319: for the FMAR, which we plot in Fig. 3(a) as the dotted curve. It

320: allows us to identify the photoresistance dip as the FMAR in the

321: FM grating. The small discrepancy left might be lifted if one

322: includes the details of the sample geometry \cite{geometric}.

323: 

324: \begin{figure} [t]

325: \begin{center}

326: \epsfig{file=fig4.eps,height=7 cm} \caption{(color online). The

327: microwave absorbance of the Py grating calculated for different

328: microwave frequencies. The curves are vertically offset for

329: clarity.} \label{Fig.4}

330: \end{center}

331: \end{figure}

332: 

333: Before addressing the technical and physical implication of our

334: results, we go a step further to calculate the absorbance

335: $A(\omega)$ of the Py grating on top of an insulating GaAs

336: substrate. The procedure is similar to that we derived recently

337: for a semiconductor multilayer system \cite{Bittkau}, except now

338: we include $\mu_{L}$ given in Eq. (2). The results of $A(\omega)$

339: plotted in Fig. 4 recover nicely the main feature \cite{FMAR} of

340: $\Delta R$ shown in Fig. 2. In particular, the agreement of the

341: calculated line shape for the FMR with the measured curve is

342: excellent, which allows us to fit accurately the dimensionless

343: Gilbert damping parameter $\alpha$ = 0.0075 \cite{alfa}. The

344: result also confirms Eq. (1), which demonstrates that the

345: bolometric effect in the FM metal bridges the spin dynamics and

346: the DC charge transport.

347: 

348: We summarize our work from both technical and physical point of

349: view. The technical difference between the photoconductivity and

350: transmission experiment is obvious. While a transmission

351: experiment measures $A(\omega)$ in Eq. (1) (or equivalently, the

352: \textit{high frequency surface impedance}) by monitoring the

353: absorption of photons, the photoconductivity experiment probes

354: $\Delta R$ via the change of the \textit{DC resistance} of

355: spin/charges. The parameter $S$ bridges both and opens free room

356: to enhance the sensitivity for the photoconductivity measurement.

357: We note that the FMAR in the FM thin film with $d < \delta$ was

358: unable to be detected by transmission measurements

359: \cite{Yager1949}. In our case where the skin depth ($\delta \sim$

360: 1 $\mu $m) is more than one order of magnitude larger than the

361: thickness ($d$ = 60 nm) of the Py, the FMAR is clearly observed as

362: a reduction of the nonresonant photoresistance. Our technique may

363: also provide a new alternative means to investigate spin

364: excitations such as quantized spin waves, which were used to be

365: measured by Brillouin light scattering spectroscopy

366: \cite{Mathieu1998}. In principle, the photoconductivity technique

367: can probe the spin dissipation via $\alpha$, as well as the energy

368: dissipation via $\tau_{e}$, both are currently of great interest

369: for investigating magnetodynamics.

370: 

371: From the physical point of view, we uncover an intrinsic different

372: nature of the FMR and FMAR. While both have the common microscopic

373: origin of the magnetodynamic excitation with Larmor precession of

374: spins, FMAR is sensitive to the macroscopic geometric pattern. We

375: demonstrate a characteristic frequency shift of the FMAR in the

376: periodic FM grating from that known for FM films. Similar gratings

377: made of normal metals are currently of great interest for their

378: enhanced transmission ability based on macroscopic optical effects

379: \cite{Porto}. Replacing normal metals with the FM metal, one may

380: bring in new optical effects utilizing the strong coupling of the

381: electromagnetic wave with magnetodynamic excitations. The

382: macroscopic nature of the FMAR, together with its intrinsic nature

383: for enhancing transmission through FM films, could pave the way

384: for integrating spintronic and photonic effects using FM

385: microstructures.

386: 

387: This work is partially supported by the EU 6th-Framework Programme

388: through project BMR-505282, the DFG through SFB 508 and BMBF

389: through project 01BM905. We thank D. Heitmann, F. Giessen, D.

390: Grundler, H.P. Oepen, and R.E. Camley for discussions, D.

391: G\"{o}rlitz and G. Meier for technical help.

392: 

393: 

394: \begin{thebibliography}{20}

395: 

396: \bibitem{Tsoi2000}

397: M. Tsoi, \textit{et al.}, Nature {\bf 406}, 46 (2000).

398: 

399: \bibitem{Kiselev2003}

400: S. I. Kiselev, \textit{et al.}, Nature {\bf 425} 380 (2003); W.H.

401: Rippard, \textit{et al.}, Phys. Rev. Lett. {\bf 92}, 027201

402: (2004).

403: 

404: \bibitem{Hu2003}

405: C. -M. Hu, \textit{et al.}, Phys. Rev. B {\bf 67}, 201302(R)

406: (2003); S. Holland, \textit{et al.}, Phys. Rev. Lett. {\bf 93},

407: 186804 (2004).

408: 

409: \bibitem{Neppl1979}

410: F. Neppl, \textit{et al.}, Phys. Rev. B {\bf 19}, 5240 (1979); K.

411: Hirakawa, \textit{et al.}, Phys. Rev. B {\bf 63}, 085320 (2001);

412: C. Zehnder, \textit{et al.}, Europhys. Lett. {\bf 63}, 576 (2003).

413: 

414: \bibitem{Kittel}

415: C. Kittel, \textit{Introduction to Solid State Physics, 6th

416: Edition} (John Wiley and Sons, Inc., New York, Chichester,

417: Brisbane, Tornonto, Singapore, 1986).

418: 

419: \bibitem{Yager1949}

420: W. A. Yager, Phys. Rev. {\bf 75}, 316 (1949); B. Heinrich and J.

421: F. Cochran, Phys. Rev. Lett. {\bf 29}, 1175 (1972); M. Scheffler,

422: M. S. thesis, University of Maryland, 1998; A. Schwartz,

423: \textit{et al.}, cond-mat/0010172 (2000).

424: 

425: \bibitem{AMR}

426: Due to the remanence of the superconducting coil of our magnet, a

427: magnetic field offset of -1.5 mT has been corrected in the plot.

428: 

429: \bibitem{Adeyeye1996}

430: A. O. Adeyeye, \textit{et al.}, J. Appl. Phys. {\bf 79}, 6120

431: (1996); M. Steiner, \textit{et al.}, J. Appl. Phys. {\bf 95}, 6759

432: (2004).

433: 

434: \bibitem{Hubert}

435: A. Hubert and R. Schaefer, \textit{Magnetic Domains - The Analysis

436: of Magnetic Microstructures} (Springer-Verlag, Berlin, Heidelberg,

437: New York, 1998).

438: 

439: \bibitem{Jackson}

440: J.D. Jackson, \textit{Classical Electrodynamics, 3rd Edition}

441: (John Wiley and Sons, Inc., New York, Chichester, Brisbane,

442: Tornonto, Singapore, 1998).

443: 

444: \bibitem{Nibarger2003}

445: J. P. Nibarger, \textit{et al.}, Appl. Phys. Lett. {\bf 83}, 93

446: (2003).

447: 

448: \bibitem{Neviere}

449: M. Nevi\`{e}re, in \textit{Electromagnetic Theory of Gratings},

450: edited by R. Petit, Springer Topics in Modern Physics  Vol. {\bf

451: 22} (Springer-Verlag, Heidelberg, 1980).

452: 

453: \bibitem{Porto}

454: J.A. Porto, \textit{et al.}, Phys. Rev. Lett. {\bf 83}, 2845

455: (1999); A.P. Hibbins, \textit{et al.}, Phys. Rev. Lett. {\bf 92},

456: 143904 (2004).

457: 

458: \bibitem{geometric}

459: To include geometric details, one needs using exact calculation as

460: have been nicely discussed by R.E. Camley, \textit{et al.}, Phys.

461: Rev. B {\bf 53}, 5481 (1996).

462: 

463: \bibitem{Bittkau}

464: K. Bittkau, \textit{et al.}, Phys. Rev. B {\bf 71}, 035337 (2005).

465: 

466: \bibitem{FMAR}

467: To reproduce the FMAR amplitude, we have phenomenologically

468: assumed a small energy loss ($<$ 1\%) for reflections at the

469: FM/GaAs interface in our calculation, which is not yet clear

470: whether it may reflect the influence of the surface leaky wave of

471: the grating as in the Wood's anomaly (Ref. [12]).

472: %

473: %To reproduce the FMAR amplitude, we have phenomenologically

474: %assumed a small energy loss ($<$ 1\%) for reflections at the

475: %FM/GaAs interface in our calculation, which is not clear yet

476: %whether it reflects the surface roughness of the FM layer, or the

477: %influence of the surface leaky wave of the grating as in the

478: %Wood's anomaly (Ref. [12]).

479: 

480: \bibitem{alfa}

481: Both the intrinsic and extrinsic contribution to $\alpha$ can be

482: obtained by analyzing its field dependence, which we leave to a

483: forthcoming longer paper.

484: 

485: \bibitem{Mathieu1998}

486: C. Mathieu, \textit{et al.}, Phys. Rev. Lett. {\bf 81}, 3968

487: (1998); Z.K. Wang, \textit{et al.}, Phys. Rev. Lett. {\bf 89},

488: 027201 (2002); K. Perzlmaier, \textit{et al.}, Phys. Rev. Lett.

489: {\bf 94}, 057202 (2005).

490: 

491: \end{thebibliography}

492: 

493: 

494: \end{document}

495: