1: \documentstyle[multicol,aps,prl,epsf]{revtex}

2: 

3: \pagestyle{empty}

4: \begin{document}

5: 

6: \draft

7: 

8: \title{Atomic and Electronic Structures of Unreconstructed Polar

9: MgO(111) Thin Film on Ag(111)}

10: 

11: \author{Manabu Kiguchi, Shiro Entani, Koichiro Saiki}

12: 

13: \address{Department of Complexity Science $\&$ Engineering, Graduate

14: School of Frontier Sciences, The University of Tokyo, \\7-3-1 Hongo,

15: Bunkyo-ku,Tokyo 113-0033, Japan}

16: 

17: \author{Takayuki Goto,  Atsushi Koma}

18: 

19: \address{Department of Chemistry, Graduate School of Science, The 

20: University of Tokyo, \\7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan}

21: 

22: \date{\today}

23: 

24: \maketitle

25: 

26: \begin{abstract}

27: 

28: Atomic and electronic structures of a polar surface of MgO formed on

29:  Ag(111) was investigated by using reflection high energy electron

30:  diffraction (RHEED), Auger electron spectroscopy, electron energy loss

31:  spectroscopy (EELS), and ultraviolet photoemission spectroscopy

32:  (UPS). A rather flat unreconstructed polar MgO(111) 1$\times$1 surface

33:  could be grown by alternate adsorption of Mg and O$_{2}$ on

34:  Ag(111). The stability of the MgO(111) surface was discussed in terms

35:  of interaction between Ag and Mg atoms at the interface, and charge

36:  state of the surface atoms. EELS of this surface did not show a band

37:  gap region, and finite density of states appeared at the Fermi level in

38:  UPS. These results suggest that a polar MgO(111) surface was not an

39:  insulating surface but a semiconducting or metallic surface.

40: 

41: \end{abstract}

42: 

43: \medskip

44: 

45: \pacs{PACS numbers: 79.60.Jv, 61.14.Hg, 68.55.-a}

46: 

47: \begin{multicols}{2}

48: \narrowtext

49: 

50: \section{INTRODUCTION}

51: \label{sec1}

52: 

53: Polar surfaces have attracted wide attention not only for fundamental

54: science but also for technological applications, because several

55: interesting properties, such as novel catalytic activity, two

56: dimensional electron system, are expected for the polar

57: surface. Figure~\ref{fig1} shows the atomic geometry of non-polar (100)

58: and polar (111) faces of rocksalt crystals schematically. The (111)

59: surface consists of only one atomic species, either cations or anions,

60: while the (100) surface consists of the same number of cations and

61: anions. In a rocksalt structure crystal, each atom is surrounded by six

62: atoms of different species. The coordination number decreases to five

63: for atoms on a (100) surface, while it changes drastically to three for

64: atoms on a (111) surface. Therefore, surface energy of the (111) face is

65: much higher than that of the (100) face, which makes the (111) surface

66: unstable\cite{1}.  In other words, alternate stacking of cation and

67: anion layers forms a dipole layer along the [111] direction

68: (Fig.~\ref{fig2}-a). Accumulation of the dipole layers would produce a

69: macroscopic electric field, which causes appearance of a flat (111)

70: surface to be quite unpreferable in the rocksalt structure compounds and

71: the flat (111) surface does not occur in nature.

72: 

73: From a theoretical viewpoint, the macroscopic electric field in the

74: polar surface can be cancelled either by surface reconstruction or by

75: reduction of effective charge of the surface layer. According to Wolf,

76: the polar (111) surface of rocksalt crystals would be stable under an

77: octapole termination, which leads to formation of a

78: 2$\times$2 reconstruction on the (111) surface\cite{1}

79: (Fig.~\ref{fig2}-b). On the other hand, Tsukada and Hoshino pointed out

80: that the change in charge state of surface atoms could stabilize the

81: (111) surface of rocksalt structure compounds\cite{2}. If the charge of

82: top atoms is reduced to half of the bulk atoms, the macroscopic electric

83: field would not appear.

84:  

85: Under these backgrounds, several attempts to grow the polar surface of

86: rocksalt structure compounds have been performed so far\cite{3}. Metal

87: oxides have been extensively studied to create crystals or thin films

88: having (111) surfaces. For a NiO(111) surface, p(2$\times$2)

89: reconstruction was found, and the surface was determined to be a single

90: Ni termination with double steps by the measurement of grazing-incidence

91: X-ray diffraction\cite{4}. The unrecostrcuted surface is stabilized by

92: some adsorbates. Langell and Berrie reported that the NiO(111) surface

93: is stabilized by OH adsorbates and that desorption of the hydroxyl

94: changes the surface to a thermodynamically more stable

95: NiO(100)\cite{5}. Stabilization by impurities was found also for Pb and

96: Si on NiO(111)\cite{6}.

97:   

98: In contrast with the surface of single crystals, the electrostatic

99: energy does not reach exorbitant values for the small thickness of a

100: film (Fig.~\ref{fig2}-c). When the thin film is grown on a metal

101: substrate, further reduction of energy could come from the image charge

102: induced in the supporting metal surface. 

103: Ventrice {\it et al.} have studied

104: the growth of NiO on Au(111) by LEED and STM, and obtained a 6 ML thick

105: stable p(2$\times$2) reconstructed NiO(111) film\cite{7}. For a FeO(111)

106: thin film, interesting results were reported by Koike and

107: Furukawa\cite{8}. They obtained the FeO (111) p(2$\times$2) surface by

108: oxidizing Fe(110) and measured the polarization of secondary electrons

109: emitted from the FeO (111) surface. Ferromagnetic ordering was found for

110: the FeO (111) p(2$\times$2) surface, although FeO itself was an

111: antiferromagnetic material with the Neel temperature of 198 K. They

112: thought that the ferromagnetic ordering comes from the reconstruction of

113: a polar FeO (111) surface to reduce the large electrostatic surface

114: energy. On the other hand, the relaxation, that is, the decrease in

115: interlayer spacing, was observed for the FeO (111) grown on Pt

116: (111)\cite{9}. In this case, the layer by layer growth was limited to a

117: maximum of 2.5 ML.

118: 

119: Thus, although the (111) surface of rocksalt structure oxides has been

120: studied by many groups, there have been few studies on unreconstructed

121: adsorbate-free (111) surfaces of a rocksalt structure until

122: now. Furthermore, the electronic structure of the polar surface has not

123: been clear. In the preset study, we have examined growth of a MgO thin

124: film on Ag(111) by supplying Mg and O, alternately. MgO has a rocksalt

125: structure with a lattice constant of 4.21 \AA, while Ag has a fcc

126: structure with a lattice constant of 4.09 \AA. As the lattice misfit of

127: MgO to Ag is only -2.9 \%, the first Mg layer is expected to become a

128: template for the growth of the MgO film along the [111] direction. The

129: commensurate bonding between Mg and Ag atoms at the interface might help

130: alternate stacking of Mg and O layers, leading to a flat (111)

131: surface. Actually, we have observed rather streak RHEED pattern,

132: implying formation of a flat (111) MgO surface. We have characterized

133: the grown surface by using EELS and ultraviolet photoemission

134: spectroscopy (UPS) and discussed stabilization mechanism of the polar

135: surface.

136:  

137:   

138: \section{EXPERIMENTAL}

139: \label{sec2}

140: 

141: The experiments were performed in a custom-designed ultrahigh-vacuum

142: (UHV) system with a base pressure of 2$\times$10$^{-8}$ Pa. A

143: mechanically and electrochemically polished Ag(111) surface was cleaned

144: by repeated cycles of Ar$^{+}$ sputtering and annealing at 900 K. After

145: repeated preparation cycles, a sharp RHEED pattern was observed, and no

146: contamination was detected by AES. A MgO film was grown by alternate

147: adsorption of Mg and O$_{2}$ on Ag(111) at substrate temperature of 300

148: K. First, 1 ML (2 \AA) Mg was deposited by evaporating high purity (99.98

149: \%) Mg onto Ag(111). The growth rate was monitored using a quartz

150: crystal oscillator. The Mg film was dosed with 10 L O$_{2}$. Real-time

151: observation of the crystallinity and orientation of films was done by

152: RHEED. Surface compositions and the electronic structure of the grown

153: films were investigated {\it in situ} by AES, 

154: EELS, and UPS. The analyses were

155: performed with a double pass CMA (PHI 15-255G) in the analysis

156: chamber. Adoption of a pulse counting detector reduced the probing

157: electron current down to 1 nA, which reduced surface damage as much as

158: possible.

159: 

160: \section{RESULTS}

161: \label{sec3}

162: \subsection{Heteroepitaxial growth of MgO on Ag(111)}

163: \label{sec3a}

164: 

165: Figure~\ref{fig3} shows the Auger spectrum of the 10 ML thick MgO film

166: grown on Ag(111). Disappearance of the Ag LMM Auger peak (351 eV)

167: indicated that the grown MgO film covered the Ag(111) substrate

168: completely because the inelastic mean free path of 351 eV electron was

169: around 1 nm. The chemical state of Mg atoms of the MgO film was known by

170: the energy of the Mg LMM Auger peak, because the energy of the Mg LMM

171: Auger peak is 45 eV and 32 eV for metallic Mg and MgO,

172: respectively\cite{10}. The energy of the Mg LMM Auger peak was 32 eV for

173: the grown MgO film, showing that Mg was completely oxidized. The

174: stoichiometry of the MgO film was also examined by the ratio of the Mg

175: KLL Auger peak intensity to the O KLL Auger peak intensity. Considering

176: the Auger electron emission probabilities\cite{11}, the ratio of the

177: amount of Mg to that of O was 1 : 0.92$\pm$0.18\cite{12}, supporting the

178: idea that a stoichiometric MgO film was grown on Ag(111).

179: 

180: Figure~\ref{fig4} shows the typical sequence of RHEED pattern during the

181: growth at a substrate temperature of 300 K. The incident electron beam

182: was parallel to the [1$\bar{1}$0] azimuth of the substrate. The result

183: of RHEED patterns indicated that the MgO film grew heteroepitaxially on

184: Ag(111). The epitaxial orientation of the MgO film was determined to be

185: (111)$_{MgO}$//(111)$_{Ag}$ and 

186: [1$\bar{1}$0]$_{MgO}$//[1$\bar{1}$0]$_{Ag}$. The half-order streaks did

187: not appear during the growth, showing that the (1$\times$1)

188: unreconstructed MgO(111) film was grown on Ag(111). Streaks in RHEED

189: patterns indicated that a rather flat (111) surface could be

190: obtained. The RHEED pattern became blurred with increasing film

191: thickness, suggesting that a thick MgO(111) film was unstable.

192: 

193: The in-plane lattice constant of the MgO(111) film was calculated from

194: the spacing between streaks in the RHEED pattern. For the 10 ML thick

195: MgO(111)/Ag(111), the in-plane lattice constant was determined to be

196: 3.28$\pm$0.03 \AA \cite{12}, which was +10 \% larger than that 

197: of bulk one (2.97 \AA). The in-plane lattice constant

198: was 3.25$\pm$0.03 \AA {} for the 2

199: ML thick film and did not change with film thickness, indicating that

200: the expansion was uniform throughout the epitaxial layer. The increase

201: of the in-plane lattice constant leads to reduction of the surface

202: electron density, which might help a decrease in the electrostatic

203: energy of the MgO(111) film.

204: 

205: 

206: \subsection{Stability of the (1$\times$1) MgO(111) film on Ag(111)}

207: \label{sec3b}

208: 

209: Having established the existence of the (1$\times$1) unreconstructed

210: MgO(111) film on Ag(111), we should discuss why this structure is stable

211: in spite of the previously mentioned arguments on the instability of a

212: polar surface of ionic crystals\cite{1}. We think that the stabilization

213: of the (1$\times$1) unreconstructed MgO(111) film can be explained in

214: terms of interaction between Ag and Mg atoms at the interface and charge

215: state of surface atoms.

216:  

217: In the previous study, a single crystalline MgO(100) film grew

218: heteroepitaxially on Ag(100) by evaporating MgO congruently from an

219: electron beam evaporator(EB)\cite{13}. On Ag(111), however, a MgO film

220: could not grow heteroepitaxially by EB. Heteroepitaxial growth of a

221: single crystalline MgO film was achieved by alternate adsorption of Mg

222: and O$_{2}$ on Ag(111). These facts indicate that the strong interaction

223: between Mg and Ag atoms plays a decisive role in stabilizing the

224: (1$\times$1) MgO(111) film. The free energy of the interface

225: ($E_{inter}$) is smaller for the case with coherent Mg-Ag bonds as

226: compared with the case, in which fine particles with thermodynamically

227: more stable \{100\} faces grow on Ag(111) without coherent Mg-Ag

228: bonds. The structure of the grown film is determined to minimize the sum

229: of the electrostatic energy of the film ($E_{electro}$) and

230: $E_{inter}$. Here, we should notice that $E_{electro}$ of a (111) film

231: does not reach exorbitant values for ultrathin films (see

232: Fig.~\ref{fig2}-c), because $E_{electro}$ is proportional to the film

233: thickness. Therefore, the gain in interaction between Mg and Ag

234: overcomes the disadvantage of the electrostatic energy, and the

235: (1$\times$1) unreconstructed MgO(111) film is grown on Ag(111). In

236: addition to the strong interaction with Mg atoms, the Ag substrate plays

237: another important role for stabilization of the MgO(111) film. The MgO

238: film was prepared by alternate adsorption of Mg and O$_{2}$. Therefore,

239: the film is unstable due to breakdown of charge neutrality

240: (Fig.~\ref{fig2}-d), when the film thickness is odd number (Mg top). On

241: the metal substrate, however, the image charge is induced in a metal,

242: helping stabilization of the film with odd layers.

243:  

244: Besides the strong interaction between Mg and Ag atoms at the interface,

245: the stabilization of MgO(111) can be explained in terms of charge

246: state. Due to the reduction of the coordination number, Madelung

247: potential largely decreases for the surface O$^{2-}$ on the (111)

248: surface. The binding energy of the 2p band, thus, decreases for the

249: surface O$^{2-}$, and the Fermi level might be located in the 2p band. In

250: such a case, electrons flow out of the upper valence states of the

251: surface O$^{2-}$, and the charge of the surface O$^{2-}$ is reduced. 

252: The charge of the Mg$^{2+}$ at the interface would 

253: also decrease to keep the film neutral

254: (Fig.~\ref{fig2}-e). The decrease in the charge reduces a macroscopic electric

255: field, and the instability of the film should be remedied. That is the

256: second point to stabilize the polar MgO(111) film on Ag(111). Here we

257: should notice that the reduction in charge of surface O$^{2-}$ leads to

258: reduction of Madelung potential. Therefore, the amount of the charge

259: transfer is determined under the condition, in which both Madelung

260: potential and the charge of surface O$^{2-}$ are self consistent. Tsukada and

261: Hoshino studied the O-terminated MgO(111) surface by DV-X$\alpha$ calculations,

262: and revealed that the charge of the top and bottom atoms is reduced to

263: half of the bulk atoms\cite{2}. When the charge of the surface atom is

264: reduced to half of the bulk atoms, the macroscopic electric field would

265: not appear, irrespective of the film thickness. In the present study, a

266: thick MgO(111) film could not be grown on Ag(111), implying that the

267: charge of the topmost O atoms did not reduce to half of the bulk one.

268: 

269: Finally, we would discuss the surface reconstruction of the MgO(111)

270: film. Since the surface energy of a \{100\} face is lower than that of a

271: \{111\} face in rocksalt structure compounds, \{100\} faces are expected

272: to appear for growth normal to the (111) face. In case of

273: NaCl/GaAs(111)A, triangular pyramids with three 

274: exposed \{100\} faces grow

275: on the substrate\cite{3,14}. As discussed in Introduction, a

276: p(2$\times$2) reconstructed NiO(111) film is grown on Au(111)\cite{7},

277: and the p(2$\times$2) structure corresponds to the smallest triangular

278: pyramids surrounded by \{100\} faces. Here, we should notice that both

279: films were grown by congruent evaporation of NiO or NaCl at the

280: substrate temperatures higher than 420 K. Therefore, the grown film

281: forms in a thermodynamically stable structure. In the present case,

282: however, we have grown a metastable flat MgO(111) film by alternate

283: adsorption of Mg and O$_{2}$ at 300 K. Once a flat (111) film is formed

284: on Ag(111), activation energy is needed to change the metastable (111)

285: film into stable triangular pyramids surrounded by \{100\}

286: faces. Therefore, the flat unreconstructed (111) MgO film remains in a

287: metastable form on Ag(111).

288: 

289: \subsection{Electronic structure of MgO(111)}

290: \label{sec3c}

291: 

292: As we could prepare the unreconstructed polar MgO(111) surface, the

293: electronic structure of this surface was investigated by EELS and UPS,

294: comparing with that of a non-polar MgO(100) surface\cite{13,15}. Absence

295: of Ag LMM Auger peak in the 10 ML MgO(111)/Ag(111) assures that EELS and

296: UPS probe the MgO film with influences of the substrate negligible,

297: because their probing depth is smaller than or equal to that of AES.

298: 

299: Figure~\ref{fig5} shows UPS spectrum for the MgO(111) surface measured

300: with a He I (21.2 eV) source. For comparison, the spectrum of the 10 ML

301: thick MgO(100) film on Ag(100) is included in the same figure. The main

302: features for MgO(111) were almost similar to those for MgO(100) except

303: around the Fermi level. For MgO(100), there were no density of states

304: (DOS) near the Fermi level, while an appreciable DOS appeared in the

305: region from 2 eV to the Fermi level for MgO(111). These finite DOS at

306: the Fermi level did not originate from metallic Mg, since component of

307: metallic Mg was not observed in AES. Furthermore, we have revealed that

308: the finite DOS at the Fermi level disappeared even for the incompletely

309: oxidized film in the previous study\cite{16}. Therefore, the finite DOS

310: at Fermi level did not originate from metallic Mg or the incompletely

311: oxidized Mg, the surface states of MgO(111). The UPS results suggested

312: that the polar MgO(111) surface was a metallic surface.

313: 

314: Figure~\ref{fig6} shows the EELS of the MgO(111) surface measured with

315: primary electron energy of 60 eV. For comparison, spectrum of the

316: MgO(100) surface is also shown in the figure. The structure above 7 eV

317: for MgO(111) was almost similar to that of MgO(100), showing that a

318: stoichiometric MgO film grew on Ag(111)\cite{13,15}. On the other hand,

319: difference appears in the structure below 5 eV between MgO(100) and

320: (111). In contrast with MgO(100), the clear band gap region was not

321: observed for MgO(111). In addition, the tail of the elastic peak became

322: larger and a new peak (p1) appeared at 1.2 eV, indicating existence of

323: the inelastically scattered electrons with small

324: loss-energy\cite{17}. Since the AES signal indicative of Mg metal was

325: not observed at all, the tail or the new peak did not originate from Mg

326: aggregates in the MgO film.

327: 

328: Due to the low growth temperature and the significant mismatch between

329: the substrate and the oxide lattice, the film may contain a significant

330: concentration of defects. Therefore, the spectroscopic features could be

331: due to emission from defect states in the MgO(111) film. Defective MgO

332: surfaces have been studied by a variety of techniques such as UPS, EELS,

333: theoretical calculation, etc. In EELS of defective MgO surfaces, sharp

334: peaks are observed at 2.3 eV and 5 eV, which are attributed to Mg

335: vacancy (V center) and O vacancy (F center), respectively. Sharp

336: structures are also observed for MgO powders by diffuse reflectance

337: spectra, and these structures at 5.8 and 4.6 eV have been attributed to

338: ions with four and three-fold ligand coordination\cite{18}. On the other

339: hand, such sharp peaks were not observed in the EELS of MgO(111) below 5

340: eV. Recently, we have revealed that metal induced gas states(MIGS) were

341: formed at the insulator-metal interface\cite{19}. Since the new states

342: of the MgO(111) film were observed independently of the film thickness,

343: the states did not originate from MIGS. Therefore, the structure below 5

344: eV could be assigned to the excitation of electrons in the unfilled band

345: derived from O 2p, or the excitations of newly appearing electronic

346: states on the (111) surface.

347: 

348: The UPS and EELS results suggested that the MgO(111) surface was not an

349: insulating surface but a semiconducting or metallic surface. This

350: electronic structure of the MgO(111) surface can be explained in terms

351: of Madelung potential as discussed in the previous section. Compared to

352: the binding energy of isolated ions, the binding energy of Mg$^{2+}$ 3s

353: orbital decreases, and the binding energy of O$^{2-}$ 2p orbital

354: increases by the Madelung potential in a crystal phase. Because of the

355: decrease in the coordination number, the Madelung potential largely

356: decreases for the (111) surface, compared with the (100)

357: surface. Therefore, the band gap is reduced and the MgO(111) surface

358: changes into a semiconductor or metallic surface. However, a possibility

359: of certain new electronic state characteristic of the (111) surface

360: cannot be excluded.

361:   

362: \section{CONCLUSIONS}

363: \label{sec4}

364: Heteroepitaxial growth of MgO on Ag(111) has been investigated by RHEED,

365: AES, UPS and EELS. The flat unreconstructed polar MgO(111) surface was

366: obtained on Ag(111) by alternate adsorption of Mg and O$_{2}$ at

367: substrate temperature of 300 K. The EELS and UPS of the (111) surface

368: were different from those of the (100) surface. For MgO(111), a clear

369: band gap region was not observed in EELS, while finite density of states

370: appeared in UPS, implying that the MgO(111) surface was a semiconducting

371: or metallic surface.

372: 

373: 

374: \acknowledgments{

375: The authors are grateful for the financial support of the Grant-in-Aid

376: from Ministry of Education, Culture, Sports, Science and Technology 

377: of Japan.}

378: 

379: \begin{references}

380: 

381: 

382: \bibitem{1} D. Wolf, Phys. Rev. Lett. {\bf 68} (1992) 3315.

383: \bibitem{2} M. Tsukada and T. Hoshino, J. Phys. Soc. Jpn. {\bf 51}

384:  (1982) 2562.

385: \bibitem{3} K. Saiki, A. Goda, and A. Koma, Jpn. J. Appl. Phys. {\bf 36} (1997) L55.

386: \bibitem{4} A. Barbier, C. Mocuta, H. Kuhlenbeck, K.F. Peters,

387:  B. Richter and G. Renaud, Phys. Rev. Lett. {\bf 84} (2000) 2897.

388: 

389: \bibitem{5} M.A. Langell and C.L. Berrie, Surf. Sci. {\bf 320} (1994)25.

390: 

391: \bibitem{6} P.A. Cox and A.A. Williams, Surf. Sci. {\bf 152/153} (1985) 791.

392: \bibitem{7} C.A. Ventrice, T. Bertrams, H. Hannemann, A. Brodde, and

393:  H. Neddermeyer, Phys. Rev. B {\bf 49} (1994) 5773.

394: \bibitem{8} K. Koike and T. Furukawa, Phys. Rev. Lett. {\bf 77} (1996) 3921.

395: \bibitem{9} W. Ranke, M. Ritter and W. Weiss, Phys. Rev. B {\bf 60} (1999) 1527.

396: 

397: \bibitem{10} S.A. Flodstrom, C.W.B. Martinsson, Surf. Sci. {\bf 118} (1982) 513.

398: \bibitem{11} Handbook of Auger electron spectroscopy, 2nd Ed.,

399:  Perkin-Elmer Corporation, 1976.

400: \bibitem{12} Error is estimated from the assumed precision of the

401:  technique (atomic ratio determined by AES: 20 \%, lattice constant

402:  determined by RHEED: 1\%).

403: \bibitem{13} M. Kiguchi, T. Goto, K. Saiki, T. Sasaki, Y. Iwasawa, and

404:  A. Koma, Surf. Sci. {\bf 512} (2002) 97.

405: \bibitem{14} K. Saiki, Y. Nakamura, and A. Koma, Surf. Sci. {\bf

406:  269/270} (1992) 790.

407: \bibitem{15} V. E. Henrich, G. Dresselhaus, and H. J. Zeiger,

408:  Phys. Rev. Lett. {\bf 36} (1976) 158.

409: \bibitem{16} K. Nishita, K. Saiki, and A. Koma, Appl. Surf. Sci. {\bf

410:  169/170} (2001) 180.

411: \bibitem{17} K. Ueno, Y. Uchino, K. Iizumi, K. Saiki, and A. Koma,

412:  Appl. Surf. Sci. {\bf 169/170} (2001) 184.

413: \bibitem{18} The Surface Science of Metal Oxides, edited by V.E.Henrich and P.A.Cox, (Cambridge University Press, 2000), p136

414: \bibitem{19} M. Kiguchi, R. Arita, G. Yoshikawa, Y. Tanida, M. Katayama,

415:  K. Saiki, A. Koma and H. Aoki, Phys. Rev. Lett. {\bf 90}, (2003)

416:  196803.

417: \end{references}

418: 

419: \begin{figure}

420: \begin{center}

421: \leavevmode\epsfysize=30mm \epsfbox{fig1.eps}

422: \caption{Atomic geometry of the non-polar(100) and polar(111) faces of a

423:  rocksalt crystal.}

424: \label{fig1}

425: \end{center}

426: \end{figure}

427: 

428: \begin{figure}

429: \begin{center}

430: \leavevmode\epsfysize=20mm \epsfbox{fig2.eps}

431: \caption{Charge distribution and resulting dipole moments (p) for (111)

432:  surfaces discussed. The numbers in the layers refer to the typical

433:  charge per atom. (a) Thick (111) film, (b) Reconstructed surface, only

434:  1/4 and 3/4 of the lattice positions are occupied for the first and the

435:  second layer, respectively, (c) Neutral 4 ML thick (111) film, and (d)

436:  3 ML thick (111) film on a metal substrate neutralized by an image

437:  charge, (e) Neutral 4 ML thick (111) film, the charge of the top and

438:  bottom atoms reduces to x of the bulk atoms (1/$\leq$x$\leq$1). The

439:  dipole moment of (e) case is smaller than that of (c) case.}

440: \label{fig2}

441: \end{center}

442: \end{figure}

443: 

444: \begin{figure}

445: \begin{center}

446: \leavevmode\epsfysize=50mm \epsfbox{fig3.eps}

447: \caption{Auger electron spectra for the 10 ML thick MgO film on

448:  Ag(111). The primary-electron energy is 3 keV.}

449: \label{fig3}

450: \end{center}

451: \end{figure}

452: 

453: \begin{figure}

454: \begin{center}

455: \leavevmode\epsfysize=50mm \epsfbox{fig4.eps}

456: \caption{A typical sequence of RHEED patterns and schematic models

457:  during the growth of the MgO film on Ag(111) at a substrate temperature

458:  of 300 K. a) Ag(111), b) 1 ML Mg/Ag(111), c) 2 ML MgO/Ag(111), d) 10 ML

459:  MgO/Ag(111). The incident beam was parallel to the [1$\bar{1}$0]

460:  azimuth of the Ag(111).}

461: \label{fig4}

462: \end{center}

463: \end{figure}

464: 

465: \begin{figure}

466: \begin{center}

467: \leavevmode\epsfysize=50mm \epsfbox{fig5.eps}

468: \caption{UPS of the polar MgO(111) surface taken with He I

469:  source. Spectra of the MgO(100) surface is shown for comparison.}

470: \label{fig5}

471: \end{center}

472: \end{figure}

473: 

474: \begin{figure}

475: \begin{center}

476: \leavevmode\epsfysize=50mm \epsfbox{fig6.eps}

477: \caption{EEL spectra of the polar MgO(111) surface (line). The primary

478:  electron energy was 60 eV. Spectra of MgO(100) surface is shown for

479:  comparison (dot line).}

480: \label{fig6}

481: \end{center}

482: \end{figure}

483: 

484: 

485: 

486: 

487: 

488: \end{multicols}

489: \end{document} 

490: 

491: 

492: