Magnetic resonance study of nanocrystalline 0.10MnO/0.90ZnO

Transcription

1 Cent. Eur. J. Phys. 11(2) DOI: /s Central European Journal of Physics Magnetic resonance study of nanocrystalline 0.10MnO/0.90ZnO Research Article Grzegorz Żołnierkiewicz 1, Janusz Typek 1, Niko Guskos 1,2, Urszula Narkiewicz 3, Daniel Sibera 3 1 Institute of Physics, Faculty of Mechanical Engineering and Mechatronics, West Pomeranian University of Technology, Al. Piastów 48, Szczecin, Poland 2 Department of Solid State Physics, Faculty of Physics, University of Athens, Panepistimiopolis, Zografou Athens, Greece 3 Institute of Chemical and Environmental Engineering, West Pomeranian University of Technology, Al. Piastów 17, Szczecin, Poland Received 30 May 2012; accepted 14 August 2012 Abstract: Magnetic properties of nanosize ZnO powders doped with MnO magnetic dopand have been studied. Sample designated as 0.10MnO/0.90ZnO was characterized by XRD that revealed the presence of ZnO and ZnMnO 3 phases. An average size of magnetic ZnMnO 3 nanocrystallites was 9 nm. Magnetic resonance study has been carried out in the K temperature range. The spectrum at each temperature was analyzed in terms of three components. The temperature dependences of resonance field, linewidth and integrated intensity of these components have been determined. Magnetic centers responsible for producing the observed spectra have been proposed. PACS (2008): Lf, Fc, Wx Keywords: magnetic resonance magnetic nanoparticles Versita sp. z o.o. 1. Introduction Spintronics is an emerging field of nanoscale electronics involving the detection and manipulation of electron spin [1]. The incorporation of spin into semiconductor electronics requires fabrication of a ferromagnetic system with the Curie temperature above room temperature (RT). Theoretical prediction of RT ferromagnetism in diluted magnetic semiconductors brought wide attention to this class gzolnierkiewicz@zut.edu.pl of materials [2, 3]. Nanostructures made of a wide-gap (E g = 3.36 ev) ZnO semiconductor doped with transition metals (Fe, Ni, Mn, Co) are of particular interest in spintronics. Despite a large volume of experimental results the conclusions are not convincing [4, 5]. Often contradictory reports by different groups lead to a debate on whether the system shows RT ferromagnetism at all. In the case that it does, the ferromagnetism might not be intrinsic to the material. Among all transition metals, the doping with Mn in ZnO is the most promising because Mn 2+ has the highest possible magnetic moment. So far, many Mn-doped ZnO systems have been synthesized by different methods (ion implanta- 226

2 Grzegorz Żołnierkiewicz et al. tion, RF magnetron sputtering, solid state reaction, pulsed laser deposition, molecular beam epitaxy, sol-gel process, hydrothermal or solvothermal) and extensively studied but controversial results are reported [6 14]. The magnetic properties of Mn-doped ZnO have been widely studied and different magnetic behaviors, like spin-glass, paramagnetic, antiferromagnetic and ferromagnetic, have been found [15 18]. These results show that the magnetization in ZnO:Mn is strongly dependent on the preparation method and is sensitive to the crystalline structure and the concentration of oxygen vacancies and defects. In this work, we present study of magnetic properties of sample containing nanosize ZnO powders doped with MnO magnetic additive, by using the magnetic resonance method. The magnetic resonance spectra registered at temperatures in the K range will be analyzed in term of three different components and attributed to specific magnetic centers. 2. Experimental The investigated sample was synthesized by the wet chemical method. Initially, from an aqueous solution of nitrides the mixture of manganese hydroxides and zinc hydroxides was obtained. Then the obtained hydroxides were filtered, dried at the temperature of 70 C and calcined 1 hour at 300 C. The final sample contained 10 wt.% MnO and 90 wt.% ZnO. The phase composition was determined by X-ray diffraction (CoK α radiation, X Pert Philips). It was found that the sample contains only two phases: ZnO and ZnMnO 3. The mean crystallite size of magnetic ZnMnO 3 was calculated by using Scherrer s formula and was found to be 9 nm [19]. Magnetic resonance spectra were obtained on a conventional X-band (ν = 9.4 GHz) Bruker E 500 spectrometer with the 100 khz magnetic field modulation. The measurements were carried in the K temperature range using an Oxford Instrument helium-flow cryostat. The registered spectra are the first derivative of the absorption curve with respect to the sweeping external magnetic field. 3. Results and discussion In Fig. 1 a selection of the registered magnetic resonance spectra of 0.10MnO/0.90ZnO are presented. In the upper panel in Fig. 1 three spectra obtained at T<30 K are shown; in the bottom panel a few spectra taken at higher temperatures are presented. At low temperatures the spectra of two components are easy to recognize. One component forms a collection of many narrow lines that are Figure 1. Examples of registered spectra of nanocrystalline 0.10MnO/0.90ZnO at low (top panel) and high (bottom panel) temperatures. apparently the result of a hyperfine structure (hfs) of manganese ion and will be further designated as component S1. The other component constitutes a single, broader line that is superimposed on the S1 spectrum. That component will be designated as S2. Both components are visible in all spectra registered in K range. At temperatures T>30 K yet another component is visible a very broad line that spans almost the whole range of magnetic field. This component will be designated as S3. To obtain more information on the involved magnetic centers responsible for the observed spectra the fitting and simulation has been carried out. The single lines S2 and S3 were each fitted by a Lorentzian-shaped line that also included the part of the line in negative magnetic fields. This is especially important for very broad lines (our S3 line). An example of simulation is presented in Fig. 2 for a spectrum registered at T = 290 K. The simulation includes all three components and in general the result is satisfactory. The S1 spectrum with hfs lines was simulated by using Simfonia program. As an example, the result of simulation at T = 10 K is presented in Fig. 3. In simulation of the S1 spectrum it was assumed that the spectrum could be attributed to isolated Mn 2+ ions. For Mn 2+ ion (3d 5, S = 5/2) the spectrum should consist of five fine transitions lines. Each transition is further split into six hyperfine lines due to 55 Mn nuclear spin (I = 5/2). Assuming that the g-factor and the hyperfine coupling constant A are isotropic, the spin-hamiltonian can be written as H = gβhs +D [S 2z 13 ] S (S + 1) +E ( Sx 2 + Sy) 2 +ASI. (1) 227

3 Magnetic resonance study of nanocrystalline 0.10MnO/0.90ZnO Figure 2. Experimental (top, black) and simulated (bottom, red) magnetic resonance spectra of nanocrystalline 0.10MnO/0.90ZnO at T = 290 K. The simulation includes all three components. Figure 3. Experimental (black) and simulated (red) magnetic resonance specta of nanocrystalline 0.10MnO/0.90ZnO at T = 10 K. Only the S1 and S2 components were included in the simulation. The first term describes the Zeeman interaction, the second and third terms the axial zero-field splitting caused by the hexagonal symmetry of the wurtzite ZnO having small rhombic distortion. The last term accounts for the electron nuclear hyperfine interaction. Good agreement with experimental spectrum was obtained using the following spin Hamiltonian parameters: g = , A = 80 G, D = 225 G and E = 10 G. In simulation, Gaussian-shaped lines with an internal peak-to-peak linewidth B = 8 G were used. The calculated values of the spin Hamiltonian parameters agree reasonably well with the values reported in the literature for Mn 2+ doped ZnO single crystal [20]. It was also observed that the obtained values of the spin Hamiltonian parameters did not Figure 4. Temperature dependence of the resonance field (open squares, right axis) and the peak-to-peak linewidth (filled squares, left axis) of the S2 component. change much with temperature in the K range, indicating on a rather weak contact with the surrounding lattice. The S2 spectrum consisting of a single broad line was fitted by a Lorentzian-shaped line at every applied temperature and thus the relevant parameters (amplitude, resonance field, peak-to-peak linewidth) could be calculated. Temperature dependence of the resonance field and linewidth of the S2 spectrum is presented in Fig. 4. On cooling the sample from RT (g = 2.071) the resonance field increases (g-factor decreases) and reaches a maximum value at 75 K (g = 1.985). On further cooling the resonance field quickly decreases to a value corresponding to g = Thermal behavior of the peak-to-peak linewidth, B pp, is different the line becomes narrower on cooling from RT down to 4 K. It is interesting to notice the alteration in the rate of linewidth change with temperature, d( B pp )/dt, which coincides with the temperature T = 75 K at which the g-factor is the smallest. For the S2 spectrum the integrated intensity I int has been calculated. Integrated intensity was defined as the product of the signal amplitude A and the square of the peakto-peak linewidth, I int = A ( B pp ) 2. This quantity is proportional to the magnetic susceptibility of the spin system. Temperature dependence of the integrated intensity for the S2 spectrum is presented in Fig. 5. On cooling from RT the integrated intensity decreases, but below 30 K it increases sharply. To check whether the low temperature rise could be described by the Curie-Weiss law, I int = const/(t T CW ), the temperature dependence of the reciprocal integrated intensity was studied. The inset in Fig. 5 shows the temperature dependence of the inverse integrated intensity. The straight line in the inset is the least-squares fit to the Curie-Weiss law with 228

4 Grzegorz Żołnierkiewicz et al. Figure 5. Temperature dependence of the integrated intensity of the S2 component. The inset shows the temperature dependence of the inverse integrated intensity. The straight line in the inset is the least-squares fit to the Curie-Weiss law. Figure 6. Temperature dependence of the resonance field (open squares, right axis) and the peak-to-peak linewidth (filled squares, left axis) of the S3 component. T CW = 0.8 K. A negative sign of T CW indicates on a weak antiferromagnetic interaction among spins involved in formation of the S2 spectrum. The S3 component was investigated in a similar manner as S2 component. The simulation of the S3 component by a single Lorentzian line could be only seen as a very rough approximation of the FMR spectrum of magnetic nanoparticles. In our sample the easy axes of nanoparticles are oriented randomly with respect to the eternal magnetic field and the resulting FMR spectrum is a sum of many individual lines with different resonance fields, linewidts and amplitudes. Nevertheless, the obtained values of the spectral parameters in such simulation should generally reflect the real temparature changes of the true parameters. Fig. 6 shows the temperature dependence of the resonance field and the peak-to-peak linewidth for that component. This line is visible only above 25 K due to an excessive broadening in the low temperature range. The resonance field increases slightly on cooling the sample from RT, but below 75 K it starts to decrease significantly, reaching almost zero at 25 K. The linewidth increases with temperature decrease in the whole investigated temperature range, but the broadening is especially large at low temperatures (T < 75 K). Such behavior of resonance field and linewidth closely resembles behavior of magnetic nanoparticles [21]. In Fig. 7 the temperature dependence of the integrated intensity of the S3 spectrum component is presented. Above 100 K the temperature changes of I int are erratic, but below 100 K a clear increase is observed. The inset shows the temperature dependence of the inverse integrated intensity at low temperatures. Below 100 K the experimental points align along the straight line, indicating that the Curie- Figure 7. Temperature dependence of the integrated intensity of the S3 component. The inset shows the temperature dependence of the inverse of the integrated intensity. The straight line in the inset is the least-squares fit to the Curie- Weiss law. Weiss law is in effect. The straight line in the inset in Fig. 7 is the least-squares fit to the Curie-Weiss law with T CW = 2.3 K. This value and the sign of T CW points to ferromagnetic interaction in the spin system responsible for the S3 line. To attribute an appropriate spin system to the S1, S2 and S3 components the phase composition of the investigated sample should be considered. According to XRD study the only magnetic phase is ZnMnO 3 and it appears as nanoparticles with an average crystallite size of 9 nm. ZnMnO 3 has a T C higher than 300 K and exhibits a spinglass behavior with the freezing temperature T f = 15 K [4]. The other important information is that most of manganese ions in ZnMnO 3 are in 4+ oxidation state [14]. The relative weights of the three components could be calculated 229

5 Magnetic resonance study of nanocrystalline 0.10MnO/0.90ZnO as the area under the appropriate absorption curves (and not their derivative signal registered by the spectrometer). As expected the most intense spectrum was obtained for the S3 component (90.7% of the total intensity), followed by much weaker S2 component (9.1%) and a very weak S1 component (0.2%). Bearing this in mind we propose the following attribution of the spin system to the S1, S2 and S3 spectra. The S1 spectrum is ascribed to isolated Mn 2+ ions which appear as defects in the ZnMnO 3 structure. They might appear on the surface of the nanograins or inside the grains. Simulation of the hfs of S1 lines assuming it was produced by the Mn 4+ (S = 3/2) ion was not successful. For the broad S2 spectrum clusters of the exchange-coupled Mn 4+ ions are responsible. The hfs structure is not visible in S2 spectrum because it was washed out by a stronger interaction between neighboring ions. A small value of Curie-Weiss temperature obtained for that spectrum indicates a very weak interaction (of the dipole-dipole type) between clusters so they must be relatively well isolated from each other. The very broad line of the S3 spectrum is most likely attributed to the magnetic ZnMnO 3 nanoparticles. Thus the S3 spectrum is the ferromagnetic resonance (FMR) line originating in a system of magnetic nanoparticles. FMR line vanishes from the spectrum before the magnetic structure of nanoparticles freezes at T f. A small value of the Curie-Weiss temperature indicates on a weak dipole-dipole interaction between magnetic nanoparticles that is not suprising by considering that only 10% (by weight) of the sample is magnetic. Thus ZnMnO 3 nanoparticles are well isolated by the ZnO nonmagnetic phase. Because the S2 and S3 spectra display similar changes at the same temperature 75 K thus the magnetic cluster of Mn 4+ ions must be influenced by the magnetic field of a nanoparticle. A possible connection will be that the S2 clusters are formed on the surface of nanoparticle thus ensuring a close interaction of both systems. 4. Conclusions Temperature study of the magnetic spectrum of 0.10MnO/0.90ZnO nanocomposite revealed the existence of three kinds of magnetic centers. Two components of the observed spectra, designated as S1 and S2, are connected with defects in magnetic structure of the ZnMnO 3 nanoparticles: the S1 spectrum displaying hfs is formed by isolated Mn 2+ ions, while the S2 line is attributed to the clusters of the exchange-coupled Mn 4+. Both paramagnetic centers are probably located on the surfaces of ZnMnO 3 nanoparticles. The FMR line S3 is produced by the ZnMnO 3 nanoparticles. References [1] I. Zutic, J. Fabian, S.D. Sarma, Rev. Mod. Phys. 76, 322 (2004) [2] T. Dietl, Semicond. Sci. Technol. 17, 377 (2002) [3] J. Fabian, A. Matos-Abiague, C. Ertler, P. Stano, I. Zutic, Acta Phys. Slovaca 57, 565 (2007) [4] S. Kolesnik, B. Dabrowski, J. Mais, Phys. Stat. Sol. C 1, 900 (2004) [5] O.D. Jayakumar et al., Nanotechnology 17, 1278 (2006) [6] D.P. Shoemaker, E.E. Rodriguez, R. Seshadri, I.S. Abumohor, T. Proffen, Phys. Rev. B 80, (2009) [7] N. Guskos et al., Acta Phys. Pol. A 120, 1074 (2011) [8] R. Sanz et al., Nanoscale Res. Lett. 4, 878 (2009) [9] J. Blasco, J. Garcia, J. Solid State Chem. 179, 2199 (2006) [10] H.B. Wang et al., Mater. Chem. Phys. 113, 884 (2009) [11] L. Duan et al., J. Magn. Magn. Mater. 323, 2374 (2011) [12] E. Schlenker et al., Appl. Phys. A 91, 375 (2008) [13] C.J. Cong, L. Liao, Q.Y. Liu, J.C. Li, K.L. Zhang, Nanotechnology 17, 1520 (2006) [14] L.V. Saraf, P. Nachimuthu, M.H. Engelhard, D.R. Baer, J. Sol-Gel Sci. Techn. 53, 141 (2010) [15] T. Fukumura et al., Appl. Phys. Lett. 78, 958 (2001) [16] J. Alaria et al., J. Appl. Phys. 99, 08M118 (2005) [17] W.Y. Shim et al., J. Electron. Mater. 35, 635 (2005) [18] A.M. Abdel Hakeem, J. Magn. Magn. Mater. 322, 709 (2010) [19] I. Kuryliszyn-Kudelska et al., J. Phys.: Conf. Ser. 200, (2010) [20] L.W. Yang, X.L. Wu, G.S. Huang, T. Qiu, Y.M. Yang, J. Appl. Phys. 97, (2005) [21] M.M. Yulikov, P.A. Purtov, Appl. Magn. Reson. 29, 231 (2005) 230