1: \documentclass[aps,prb,twocolumn,showpacs]{revtex4}
2: %\documentclass[preprint,aps,draft]{revtex4}
3: \usepackage{graphicx}
4: \begin{document}
5: %=======================================================================
6: %Zn$_{1-x}$Co$_x$O
7: %Zn$_{0.9}$Co$_{0.1}$O
8: %
9: \title{Electronic structures of Zn$_{1-x}$Co$_x$O using photoemission \\
10: and x-ray absorption spectroscopy}
11:
12: \author {S. C. Wi, J.-S. Kang$^{*}$, J. H. Kim, S.-B. Cho, B. J. Kim,
13: S. Yoon, and B. J. Suh}
14:
15: \affiliation{Department of Physics, The Catholic University of Korea,
16: Puchon 420-743, Korea}
17:
18: \author {S. W. Han and K. H. Kim}
19:
20: \affiliation{Department of Physics, Gyeongsang National University,
21: Chinju 660-701, Korea}
22:
23: \author {K. J. Kim, B. S. Kim, H. J. Song, and H. J. Shin}
24:
25: \affiliation{Pohang Accelerator Laboratory (PAL),
26: POSTECH, Pohang 790-784, Korea}
27:
28: \author {J. H. Shim and B. I. Min}
29:
30: \affiliation{Department of Physics, POSTECH, Pohang 790-784, Korea}
31:
32: \date{\today}
33:
34: \begin{abstract}
35:
36: Electronic structures of Zn$_{1-x}$Co$_x$O have been investigated
37: using photoemission spectroscopy (PES) and x-ray absorption
38: spectroscopy (XAS).
39: The Co $3d$ states are found to lie near the top of the O $2p$
40: valence band, with a peak around $\sim 3$ eV binding energy.
41: The Co $2p$ XAS spectrum provides evidence that the Co ions
42: in Zn$_{1-x}$Co$_{x}$O are in the divalent Co$^{2+}$ ($d^7$) states
43: under the tetrahedral symmetry.
44: Our finding indicates that the properly substituted Co ions
45: for Zn sites will not produce the diluted ferromagnetic semiconductor
46: property.
47:
48: \end{abstract}
49:
50: %\pacs{79.60.-i, 75.47.Lx, 71.30.+h}
51:
52: \maketitle
53:
54: %\narrowtext
55: %==============================================================================
56: %\section{Introduction}
57: %\label{sec:intro}
58:
59: %Spintronics, which utilizes both the spin and charge degrees of
60: %freedom, becomes an important field in magnetism.
61: Diluted magnetic semiconductors (DMSs) are considered as good
62: candidates for spintronics materials because of the possibility of
63: incorporating the spin degree of freedom in traditional semiconductors
64: \cite{Furd88,Wolf01}.
65: Motivated by the recent theoretical calculations by Dietl {\it et al.}
66: \cite{Dietl00} which predicted room-temperature ferromagnetism in
67: Mn-doped ZnO, Zn$_{1-x}$T$_x$O systems (T$=3d$ transition-metal atom)
68: have been investigated extensively.
69: Indeed there were reports that Zn$_{1-x}$Mn$_x$O epitaxial thin films
70: exhibited ferromagnetic properties \cite{Fuku99,Jung02} and that
71: Zn$_{1-x}$Co$_x$O films showed ferromagnetism
72: with the Curie temperature $\rm T_C \sim 300$ K \cite{Ueda01}.
73: However, the reproducibility is somewhat questionable,
74: and there are contradicting reports on Co-doped ZnO
75: \cite{Ueda01,Kim02,Lee02}.
76: Technologically ZnO-based DMSs of Zn$_{1-x}$T$_x$O attract much
77: attention due to the cheapness and abundance of ZnO,
78: and the easy solubility of T atoms in ZnO
79: up to several ten $\%$ \cite{Fuku99,Jin00}.
80: In addition, since ZnO has a wide band gap energy of $\sim 3.4$ eV,
81: ZnO-based DMSs would be useful for short wavelength magneto-optical
82: applications.
83:
84: %In spite of extensive studies on the magnetic properties of
85: %Zn$_{1-x}$T$_x$O-type DMSs, the origin of the observed ferromagnetism
86: %in Zn$_{1-x}$T$_x$O has not been clarified yet.
87: In order to understand the magnetic interaction in Zn$_{1-x}$T$_x$O,
88: it is essential to understand the electronic structure of the doped
89: $3d$ T impurities in Zn$_{1-x}$T$_x$O.
90: %Two competing factors are expected, such as
91: %the strong on-site Coulomb interaction between $3d$ electrons
92: %($U_{dd}$), and the hybridization between $3d$ and the host
93: %valence-band electrons ($V_{pd}$) \cite{Mizo02}.
94: Photoemission spectroscopy (PES) is one of the powerful experimental
95: methods for providing direct information on the electronic structures
96: of solids.
97: To our knowledge,
98: only a few PES studies have been reported on Zn$_{1-x}$T$_x$O:
99: the PES study on Zn$_{1-x}$Mn$_x$O \cite{Mizo02}
100: and the Co $2p$ core-level PES spectra for Zn$_{1-x}$Co$_x$O
101: \cite{Lee02}.
102:
103: In this study we have investigated the electronic structure of
104: homogeneous bulk samples of Zn$_{1-x}$Co$_x$O using
105: PES and x-ray absorption spectroscopy (XAS).
106: Differently from previously reported film samples\cite{Ueda01},
107: magnetic properties of our bulk Zn$_{1-x}$Co$_x$O samples revealed
108: that the Co-Co magnetic
109: interaction is dominated by the antiferromagnetic coupling
110: \cite{Yoon03}.
111: Magnetic susceptibility at high temperature exhibits a typical Curie-Weiss
112: behavior with negative $\rm T_C$ and the magnetization of Co ion
113: is reduced with increasing the Co ion concentration reflecting
114: an increase in average antiferromagnetic interaction between Co ions.
115: Hence the present bulk sample is not a diluted ferromagnetic semiconductor.
116:
117: %\section{Experimental and Calculational Details}
118: %\label{sec:exp}
119:
120: Polycrystalline Zn$_{1-x}$Co$_x$O samples ($x=0, 0.1$) were synthesized
121: using the standard solid-state reaction method.
122: Zn$_{1-x}$Co$_x$O samples were obtained by sintering a mixed powder
123: of ($1-x$)ZnO$+x$CoO at T $\simeq 1400 ^{\circ}$ C for $2$ hours
124: followed by the furnace cooling in air.
125: The x-ray diffraction (XRD) analysis showed that all the samples have
126: the single phase of the Wurtzite structure with no impurity phase.
127: Valence-band PES, XAS and Co $2p \rightarrow 3d$ resonant photoemission
128: spectroscopy (RPES) measurements were performed at
129: the 2B1 and 8A1 beamlines of the PAL.
130: Samples were cleaned {\it in situ} by repeated scraping with a diamond
131: file and the data were obtained at room temperature with the pressure
132: better than $4 \times 10^{-10}$ Torr.
133: The Fermi level $\rm E_F$ \cite{calib} and the overall instrumental
134: resolution (FWHM) of the system were determined from the valence-band
135: spectrum of a sputtered Au foil in electrical contact with a sample.
136: The FWHM was about $0.1-0.6$ eV between a photon energy $h\nu\sim 30$
137: eV and $h\nu\approx 600$ eV.
138: All the spectra were normalized to the mesh current.
139: The XAS spectra were obtained by employing the total
140: electron yield method.
141: The experimental energy resolution for the XAS data was set
142: to $\sim 0.2$ eV at the Co $2p$ absorption threshold
143: ($h\nu \approx$ 700 eV).
144: Core-level spectra and the valence-band spectrum with $h\nu=1486.6$
145: eV were obtained by using the monochromatized Al $K\alpha$ radiation
146: with a FWHM of $\sim 0.6$ eV.
147: %For both PES and XAS measurements, samples were cleaned
148: %{\it in situ} by repeated scraping with a diamond file.
149:
150: %\section{Results and Discussion}
151: %\label{sec:results}
152:
153: %-------------------------------------
154: \begin{figure}[t]
155: \includegraphics[scale = 0.5]{fig1.eps}
156: \caption{Left: Comparison of the Co $2p_{3/2} \rightarrow 3d$
157: on-resonance spectrum ($h\nu
158: \approx 777.1$ eV) and the off-resonance spectrum
159: ($h\nu\approx 776.4$ eV) for Zn$_{0.9}$Co$_{0.1}$O,
160: scaled at $\sim 7$ eV BE.
161: Right: Comparison of the valence-band spectrum of ZnO
162: and that of Zn$_{0.9}$Co$_{0.1}$O, obtained with
163: $h\nu=1486.6$ eV.
164: For both cases, the difference curves are shown at the bottom.}
165: \label{3d}
166: \end{figure}
167: %-------------------------------------
168: \begin{figure}[t]
169: \includegraphics[scale = 0.65]{fig2.eps}
170: \caption{
171: The Co $2p$ XAS spectrum of Zn$_{0.9}$Co$_{0.1}$O
172: in comparison to those of
173: CoO (Co$^{2+}$) [Ref.~\cite{Regan01}],
174: LiCoO$_2$ (Co$^{3+}$) [Ref.~\cite{Elp91}],
175: and Co metal [Ref.~\cite{Regan01}].
176: The data for CoO, LiCoO$_2$, and Co metal were shifted
177: by $+0.1$ eV, $-3.64$ eV and $-0.1$ eV, respectively.
178: The calculated Co $2p$ XAS spectrum
179: for low-spin Co$^{2+}$ under the $T_d$ symmetry is presented
180: at the bottom. }
181: \label{2pxas}
182: \end{figure}
183: %-------------------------------------
184:
185: We have measured the valence-band PES spectra of
186: Zn$_{1-x}$Co$_x$O ($x=0,~0.1$) over a wide several photon energy
187: range 25 eV $\le h\nu \le$ 1486.6 eV, which includes
188: the Co $2p \rightarrow 3d$ RPES energies (see Fig.~\ref{2pxas}).
189: Our valence-band PES spectrum of ZnO agrees with the existing
190: data in the literature \cite{Ruckh94}.
191: Valence-band spectra for ZnO and Zn$_{0.9}$Co$_{0.1}$O
192: were rather similar, showing two large structures, one sharp peak
193: around 11 eV binding energy (BE),
194: and the other broad band between $\sim 3$ eV and $\sim 7$ eV BE.
195: The former and latter peaks are assigned as the Zn $3d$
196: and O $2p$ emissions, respectively.
197: They also reveal negligible emission near $\rm E_F$, reflecting
198: that both ZnO and Zn$_{0.9}$Co$_{0.1}$O are insulating.
199:
200: The left panel of Fig.~\ref{3d} compares
201: the on-resonance ($h\nu \approx 777.1$ eV) and off-resonance
202: ($h\nu\approx 776.4$ eV) spectra for Zn$_{0.9}$Co$_{0.1}$O
203: in the Co $2p_{3/2} \rightarrow 3d$ RPES,
204: which are scaled to each other at $\sim 7$ eV BE.
205: The difference between on-resonance and off-resonance
206: can be considered as representing the Co $3d$ partial spectral
207: weight distribution (PSW)
208: \cite{Kang02}.
209: The right panel of Fig.~\ref{3d} compares the valence-band spectrum
210: of Zn$_{0.9}$Co$_{0.1}$O to that of ZnO, obtained with $h\nu=1486.6$ eV.
211: The differences between
212: $x=0$ and $x=0.1$ also reflect the effect of the Co $3d$ states.
213: In both cases the differences are shown at the bottom, which
214: are very similar to each other, showing a peak around $\sim 3$ eV BE
215: and negligible emission near $\rm E_F$ \cite{scale}.
216: These results provide evidence that the Co $3d$ states in
217: Zn$_{1-x}$Co$_x$O are located near the top of the O $2p$ valence band.
218:
219: The extracted Co $3d$ PSW reveals almost no emission between $\rm E_F$
220: and 2 eV BE.
221: This feature does not agree with the calculated
222: Co $3d$ PDOS (partial density of states),
223: obtained from the LSDA (local spin-density approximation) band
224: structure calculation \cite{Sato02}.
225: The calculated Co $3d$ PDOS for Zn$_{0.75}$Co$_{0.25}$O
226: shows a sharp $e_{g}^2\downarrow$ peak near $\rm E_F$.
227: To resolve this discrepancy,
228: more elaborate calculations such as LSDA$+U$ incorporating
229: the Coulomb correlation $U$ between Co $3d$ electrons
230: will be required because it is well known that the LSDA
231: calculation does not properly describe the electronic structures
232: of insulators with strongly correlated electrons \cite{Park00}.
233:
234: Figure~\ref{2pxas} compares the Co $2p$ XAS spectrum of
235: Zn$_{0.9}$Co$_{0.1}$O with those of reference Co oxides, having
236: the formal Co valences of $2+$ (CoO, Ref.~\cite{Regan01})
237: and $3+$ (LiCoO$_2$, Ref.~\cite{Elp91}),
238: and that of Co metal (Ref.~\cite{Regan01}).
239: Note that Co ions both in CoO and LiCoO$_2$ are located at
240: the octahedron ($O_h$) centers.
241: The peak positions and the line shape of the T $2p$ XAS spectrum
242: depend on the local electronic structure of the T ion,
243: providing the information about the valence state
244: and the ground state symmetry of the T ion \cite{Groot90,Laan92,Kim03}.
245: The Co $2p_{3/2}$ and $2p_{1/2}$ spectral parts are clearly separated
246: by the large $2p$ core-hole spin-orbit interaction, and the core-hole
247: lifetime broadening is small, resulting in the sharp multiplet
248: structures.
249: The Co $2p$ XAS spectrum of Zn$_{0.9}$Co$_{0.1}$O looks similar to
250: that of CoO except for the absence of the low-$h\nu$ shoulder
251: ($h\nu\sim 776$ eV), but quite different from those of LiCoO$_2$ and Co metal.
252: This observation
253: indicates that Co ions in Zn$_{1-x}$Co$_x$O might be
254: in the divalent Co$^{2+}$ valence states, but not in the trivalent
255: Co$^{3+}$ valence states, and that the formation of the Co metal
256: cluster in our Zn$_{1-x}$Co$_x$O samples can be ruled out.
257:
258: In order to estimate the valence state of Co ions in the ground states,
259: we have analyzed the Co $2p$ XAS spectrum of Zn$_{0.9}$Co$_{0.1}$O
260: within the cluster model where the effects of the multiplet
261: interaction, the crystal field, and the hybridization
262: with the O $p$ ligands are included \cite{Laan92}.
263: The calculated XAS spectrum is shown at the bottom of Fig.~\ref{2pxas}.
264: We have found that including only one configuration, corresponding to
265: Co$^{2+}$ ($d^7$), and the tetrahedral ($T_d$) crystal field energy
266: of $10Dq=0.6$ eV yields
267: a good fit for the measured $2p$ XAS spectrum \cite{cluster}.
268: The analysis shows that it is not necessary to include
269: the charge-transfer configuration, $d^{n+1}\underbar{L}^{1}$
270: ($\underbar{L}^{}$: a ligand hole), indicating that the hybridization
271: between Co $3d$ and O $p$ orbitals is small, in contrast
272: to Mn-doped ZnO case \cite{Mizo02}.
273: We have checked that small changes in the parameters do not affect
274: the overall spectral shape including the prominent peak positions.
275: Based on the calculated XAS spectrum, one can conclude that
276: the doped-Co ions in Zn$_{1-x}$Co$_{x}$O are divalent in the ground
277: states, i.e., Co$^{2+}$ ($S=3$)
278: under the $T_d$ symmetry with a small crystal field energy \cite{LS}.
279: Our finding suggests that, in our samples,
280: Co ions substitute properly
281: for the Zn sites with Co$^{2+}$ valence states. In view of no long
282: range ferromagnetic order in our samples,
283: the reported DMS properties
284: in some Co-doped ZnO \cite{Ueda01,Lee02} are likely to be extrinsic,
285: that is, to originate from Co cluster or other impurity phases.
286:
287:
288: %\section{Conclusions}
289: %\label{sec:conc}
290:
291: In conclusion, the electronic structures of bulk Zn$_{1-x}$Co$_x$O
292: samples have been investigated by employing PES and XAS.
293: According to the Co $2p \rightarrow 3d$ RPES and the careful comparison
294: between $x=0$ and $x=0.1$, the Co $3d$ states in Zn$_{1-x}$Co$_x$O are
295: found to lie near the top of the O $2p$ valence band with a peak
296: around $\sim 3$ eV BE. The measured Co $2p$ XAS spectrum shows that
297: the ground states of Co ions in Zn$_{1-x}$Co$_{x}$O are the divalent
298: Co$^{2+}$ states under the $T_d$ symmetry corresponding to the total
299: spin of $S=3$ per Co ion.
300: Our finding suggests that the properly substituted Co ions
301: in Co-doped ZnO do not produce the DMS property with the long range
302: ferromagnetic order.
303:
304: Acknowledgments$-$
305: This work was supported by the KRF (KRF-2002-070-C00038) and
306: by the KOSEF through the CSCMR at SNU and the eSSC at POSTECH.
307: The PAL is supported by the MOST and POSCO in Korea.
308:
309: \begin{thebibliography}{99}
310:
311: \bibitem[*]{AAAuth} Author to whom all correspondence should be
312: addressed: kangjs@catholic.ac.kr
313:
314: \bibitem{Furd88} J. K. Furdyna,
315: J. Appl. Phys. {\bf 64}, R29 (1988).
316: \bibitem{Wolf01} S. A. Wolf, D. D. Awschalom, R. A. Buhrman,
317: J. M. Daughton, S. von Molnar, M. L. Roukes,
318: A. Y. Chtchelkanova, and D. M. Treger,
319: Science {\bf 294}, 1488 (2001).
320: \bibitem{Dietl00} T. Dietl, H. Ohno, F. Matsukura, J. Cibert,
321: and D. Ferrand,
322: Science {\bf 287}, 1019 (2000).
323: \bibitem{Fuku99} T. Fukumura, Z. Jin, A. Ohtomo, H. Koinuma,
324: M. Kawasaki,
325: Appl. Phys. Lett. {\bf 75}, 3366 (1999).
326: \bibitem{Jung02} S. W. Jung, S.-J. An, G. C. Yi, C. U. Jung,
327: S. I. Lee, and S. Cho,
328: Appl. Phys. Lett. {\bf 80}, 4561 (2002).
329: \bibitem{Ueda01} K. Ueda, H. Tabata, and T. Kawai,
330: Appl. Phys. Lett. {\bf 79}, 988 (2001).
331: \bibitem{Kim02} J. H. Kim, H. Kim, D. Kim, Y. E. Ihm,
332: and W. K. Choo,
333: J. Appl. Phys. {\bf 92}, 6066 (2002).
334: \bibitem{Lee02} H. J. Lee, S. Y. Jeong, C. R. Cho, and C. H. Park,
335: Appl. Phys. Lett. {\bf 81}, 4020 (2002).
336: % \bibitem{Sato01} K. Sato and H. K. Yoshida,
337: % Physica E {\bf 10}, 251 (2001).
338: \bibitem{Jin00} Z. Jin, M. Murakami, T. Fukumura, Y. Matsumoto,
339: A. Ohtomo, M. Kawasaki, and H. Koinuma,
340: J. Crystal Growth {\bf 214/215}, 55 (2000).
341: \bibitem{Mizo02} T. Mizokawa, T. Nambu, A. Fujimori, T. Fukumura,
342: and M. Kawasaki,
343: Phys. Rev. B {\bf 65}, 085209 (2002).
344: \bibitem{Yoon03} S. W. Yoon, S.-B. Cho, S. C. Wi, S. Yoon,
345: B. J. Suh, H. K. Song, and Y. J. Shin,
346: J. Appl. Phys. {\bf 93}, 7879 (2003).
347: \bibitem{calib} Since Zn$_{1-x}$Co$_x$O samples are insulating,
348: there is some ambiguity in determining $\rm E_F$.
349: However, the charging did not distort the valence-band
350: lineshape with time.
351: % \bibitem{Yeh85} J. J. Yeh and I. Lindau,
352: % {\it Atom. Data Nucl. Data Tables} {\bf 32}, 1 (1985).
353: % \bibitem{Kang03} J. -S. Kang, unpublished.
354: \bibitem{Ruckh94} M. Ruckh, D. Schmid, and H. W. Schock,
355: J. Appl. Phys. {\bf 76}, 5945 (1994).
356: \bibitem{Kang02} J.-S. Kang, J. H. Kim, A. Sekiyama, S. Kasai,
357: S. Suga, S. W. Han, K. H. Kim, T. Muro, Y. Saitoh, C. Hwang,
358: C. G. Olson, B. J. Park, B. W. Lee,
359: J. H. Shim, J. H. Park, and B. I. Min
360: Phys. Rev. B {\bf 66}, 113105 (2002).
361: \bibitem{scale} When the difference is obtained with the
362: on-resonance and off-resonance spectra being scaled at
363: the Zn $3d$ peak, we get another broad feature between
364: $\sim 4$ eV and $\sim 8$ eV BE. This feature is ascribed
365: to the Co $3d$-O $2p$ hybridized states.
366: Here we focus on the resonating feature which has rather
367: pure Co $3d$ electron character.
368: \bibitem{Sato02} K. Sato and H. K. Yoshida,
369: Semicond. Sci. Technol. {\bf 17}, 367 (2002).
370: \bibitem{Park00} J. H. Park, S. K. Kwon, and B. I. Min,
371: Physica B {\bf 281-282}, 703 (2000).
372: \bibitem{Regan01} T. J. Regan, H. Ohldag, C. Stamm,
373: F. Nolting, J. L\"{u}ning, J. St\"{o}hr,
374: and R. L. White,
375: Phys. Rev. B {\bf 64}, 214422 (2001).
376: \bibitem{Elp91} J. van Elp, J. L. Wieland, H. Eskes,
377: P. Kuiper, G. A. Sawatzky, F. M. F. de Groot,
378: and T. S. Turner,
379: Phys. Rev. B {\bf 44}, 6090 (1991).
380: \bibitem{Groot90} F. M. F. de Groot, J. C. Fuggle, B. T. Thole,
381: and G. A. Sawatzky,
382: Phys. Rev. B {\bf 42}, 5459 (1990).
383: \bibitem{Laan92} G. van der Laan and I. W. Kirkman,
384: J. Phys.: Condens. Matter {\bf 4}, 4189 (1992).
385: \bibitem{Kim03} J.-Y. Kim, J.-H. Park, B.-G. Park, H.-J. Noh,
386: S.-J. Oh, J. S. Yang, D.-H. Kim, S. D. bu, T.-W. Noh,
387: H.-J. Lin, H.-H. Hsieh, and C. T. Chen,
388: Phys. Rev. Lett. {\bf 90}, 017401 (2003).
389: \bibitem{cluster}
390: The fitting parameters are scaled by 80\% of the atomic
391: Slater integrals ($F(dd),F(pd)$, and $G(pd)$).
392: \bibitem{LS} In the $T_d$ symmetry for Co ions,
393: one cannot distinguish between the low-spin (LS) and
394: high-spin (HS)
395: divalent Co$^{2+}$ states because the $e_g$ states are located
396: below the $t_{2g}$ states.
397: This results in the same total spin $S=3$ for both
398: the small (the so-called LS state) and
399: high (the so-called HS state) values of $10Dq$.
400:
401: \end{thebibliography}
402:
403: \end{document}
404:
405: