Electron paramagnetic resonance in transition metal-doped ZnO nanowires

JOURNAL OF APPLIED PHYSICS 101, 024324 共2007兲

Electron paramagnetic resonance in transition metal-doped ZnO nanowires A. O. Ankiewicz,a兲 M. C. Carmo, and N. A. Sobolev Departamento de Física, Universidade de Aveiro, 3810-193 Aveiro, Portugal

W. Gehlhoff Institut für Festkörperphysik, Technische Universität Berlin, D-10623 Berlin, Germany

E. M. Kaidashev,b兲 A. Rahm, M. Lorenz, and M. Grundmann Institut für Experimentelle Physik II, Universität Leipzig, D-04103 Leipzig, Germany

共Received 31 July 2006; accepted 9 October 2006; published online 29 January 2007兲 The wide-band-gap zinc oxide-based diluted magnetic semiconductors currently attract considerable attention due to their possible use in spintronic devices. In this work, we studied ZnO nanowire samples synthesized on 10⫻ 10 mm2 a-plane sapphire substrates by high-pressure pulsed laser deposition. The samples were characterized by scanning electron microscopy 共SEM兲 and electron paramagnetic resonance 共EPR兲 in the X-band 共⯝9.3 GHz兲 from T = 4 to 300 K. According to the SEM pictures, the nanowires exhibit a length of about 1 ␮m and are aligned perpendicular to the substrate surface. The structures have a hexagonal cross section and their diameter ranges from 60 nm up to 150 nm. For the lowest nominal concentrations of xMn = 3 at. % and xCo = 5 at. %, we detect the anisotropic EPR spectra of isolated Mn2+ 共3d5, 6S兲 and Co2+ 共3d7, 4F兲, respectively, on Zn sites. The detection of the well-resolved anisotropic spectra proves a coherent crystallographic orientation of the nanowires. The linewidth was larger than the best values reported in the literature. Nevertheless, it was possible to identify two different components, A and B, of the reported spectra. From the temperature dependence of the EPR intensity, we found that both components exhibit paramagnetic behavior and are present in a concentration ratio of NB / NA = 1.4. In the case of the Mn-doped ZnO wires, the linewidth increases with increasing Mn concentration due to the dipole-dipole interaction of the paramagnetic ions. At the highest used nominal concentration, xMn = 10 at. %, an additional broad single line is observed. © 2007 American Institute of Physics. 关DOI: 10.1063/1.2402095兴 I. INTRODUCTION

Developments in spin-transport electronics depend in large part on the elaboration of ferromagnetic semiconductors with Curie temperatures above room temperature. The diluted magnetic semiconductors 共DMSs兲 in which a fraction of nonmagnetic elements is substituted by magnetic transition metal ions are ideal candidates. The wide-band-gap zinc oxide-based DMSs attract currently considerable attention due to their possible applications in spintronic and UV devices.1,2 An important step forward in the field was the prediction by Dietl et al.3 of high-temperature ferromagnetism in some magnetically doped wide-band-gap p-type semiconductors. Even though the ab initio calculations predict that the incorporation of V, Cr, Fe, Co, or Ni in ZnO in the 5%–25% concentration range should give rise to metallic behavior and a ferromagnetic state without need of additional doping,4,5 the experimental results are quite controversial even for bulk materials. For Mn- and Co-doped ZnO films, usually only weak ferromagnetism has been found.6–8 Very recently, a giant magnetic moment of 6.1 ␮B / Co and a high Curie temperature of 790 K have been observed in isolating ZnO films doped with 4 at. % Co grown at low temperature 共200 °C兲.9 As to nanostructures, their magnetic a兲

Electronic mail: [email protected] On leave from Rostov-on-Don State University, Mechanics and Applied Mathematics Research Institute, 344090 Rostov-on-Don, Russia.

b兲

0021-8979/2007/101共2兲/024324/6/$23.00

properties can be completely different from their bulk counterparts.10–13 The advantage of the electron paramagnetic resonance 共EPR兲 spectroscopy is its extreme sensitivity to the microscopic environment of the paramagnetic center. Therefore, it has been applied by us to the study of the incorporation of Co and Mn ions into the crystalline lattice of ZnO nanowires grown by pulsed laser deposition 共PLD兲. II. EXPERIMENTAL DETAILS

The ZnO nanowire samples were synthesized on 10 ⫻ 10 mm2 a-plane sapphire substrates by high-pressure PLD.14,15 A gold catalyst was applied prior to the growth. The growth temperature varied between 780 and 880 °C. Argon was used as a carrier gas at a background pressure of 100 mbar and a constant flow of 100 sccm. The PLD targets were made from pressed and sintered 5N powders for 12 h at 1150 °C in air, nominally containing 3 at. % Mn, 10 at. % Mn, or 5 at. % Co. The target-to-substrate distance was varied between 10 and 20 mm, and 4800–12 000 pulses were used to ablate the targets. In a previous work, Rahm et al.15 carried out elemental analysis of the nanowire samples grown from the 5 at. % Co and from the 3 at. % Mn PLD targets by Rutherford backscattering 共RBS兲. For the first sample, the Co content measured at seven different spots on the sample varied between 0.15 and 0.3 at. %. In the second sample, the Mn content was more inhomogeneous varying between 0.20 and 0.75 at. %. In both cases, the RBS mea-

101, 024324-1

© 2007 American Institute of Physics

Downloaded 30 Jan 2007 to 193.136.173.43. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp

024324-2

J. Appl. Phys. 101, 024324 共2007兲

Ankiewicz et al.

FIG. 1. Typical SEM images of ZnO:5 at. % Co 共a兲; ZnO:3 at. % Mn 共b兲, and ZnO:10 at. % Mn. All pictures were taken under a 45° viewing angle.

surements indicate that the doping content in the nanowires is much lower than that expected from the targets compositions. The scanning electron microscopy 共SEM兲 measurements were performed with a CamScan CS 44 Microscope using 10-keV electrons. Figure 1 shows three SEM pictures that evidence the good quality and alignment of the nanowires. The nanowires exhibit a length of about 1 ␮m and are aligned perpendicular to the substrate. The structures have a hexagonal cross section and their diameters range from 60 to 150 nm. The EPR spectra were measured in the X-band 共⬇9.3 GHz兲 at temperatures between 4 and 300 K using a Bruker ESP 300E spectrometer equipped with an Oxford Instruments continuous flow helium cryostat. III. EPR ANALYSIS

ZnO crystallizes in the wurtzite structure P63mc with a C3v point symmetry for the substitutional sites. As with all 3d transition metal ions, Co and Mn are expected to substitute Zn atoms. In their neutral charge state 共as referred to the charge of the Zn ion兲, the ions have 3d7 and 3d5 electron valence configurations, yielding 4A2 and 6A1 ground states for the free ions, respectively. A. Co incorporation

The atomic 4F ground state of Co2+ splits under the influence of the tetrahedral component of the crystal field 共CF兲 into a 4A2 orbital singlet state and two orbital triplets, 4T2 and 4T1. The first excited state 4T2 is separated from the lowest level by the amount 10Dq ⯝ 4000 cm−1 共Ref. 16兲. The value of 10Dq is much larger than the thermal energy at room temperature; thus, the occupations of all excited states are much smaller than of the ground-state levels. The EPR and the magnetic susceptibility are therefore determined almost entirely by the properties of the singlet ground state 4 A2, with only a small admixture from the higher-lying excited states. Under the action of the trigonal component of the CF and the spin-orbit coupling both triplets and the singlet undergo further splitting. The fourfold degenerated 共S = 3 / 2兲 4A2 ground state divides into two Kramers doublets E±1/2 and E±3/2 with a zero-field splitting equal to 2D, where in agreement with the optical16 and EPR results,17 the doublet E±1/2 共SZ = ± 1 / 2兲 is the lowest. The EPR data can be described by the following spin Hamiltonian 共SH兲:18 ˆ = ␮ B共g cos ␪S + g sin ␪S 兲 + A S I H S B 储 Z ⬜ X 储 Z Z

冋

册

1 + A⬜共SXIX + SY IY 兲 + D SZ2 − S共S + 1兲 , 3

共1兲

where ␮B is the Bohr magneton, B is the magnetic field, and

S and I are the electronic and nuclear spin operators, respectively, D is the axial fine-structure 共FS兲 parameter, and g and A are the g-tensor and the hyperfine structure 共HFS兲 tensor, respectively. The label Z 共or 储兲 applies for the c axis 共hexagonal 关0001兴 axis of ZnO兲 and X, Y 共or ⬜兲 apply for all axes perpendicular to it. In the present case, I = 7 / 2. In this limiting case of the zero-field splitting being much larger than the Zeeman energy, only the electron spin transitions mS = −1 / 2 ↔ + 1 / 2 within the S = 3 / 2 manifold can be observed in the available magnetic field range. The corresponding spectrum can be, in good approach, described by an effective SH given by ˆ = ␮ B · g⬘ · S ⬘ + S ⬘ · A⬘ · I , H S B

共2兲

with an effective spin of S⬘ = 1 / 2 and taking into account the hyperfine 共HF兲 interaction in the doublet. The angular dependence of the line positions Bm of the allowed HF transitions 共⌬m = 0兲 within the 兩±1 / 2典 spin doublet is given by the resonance condition h␯ = ␮Bg⬘共␪兲Bm共␪兲 + A⬘共␪兲m,

共3兲

where ␪ is the angle between the c axis and the applied external magnetic field B. The effective g-value g⬘共␪兲 is connected with the g-values used in Eq. 共1兲 in the S = 3 / 2 manifold in first order by19 g⬘共␪兲 = 冑g储2 cos2共␪兲 + 共2g⬜兲2 sin2共␪兲,

共4兲

and the apparent hyperfine constant A⬘共␪兲 by

19

A ⬘共 ␪ 兲 =

冑g储2A储2 cos2共␪兲 + 16g⬜2 A⬜2 sin2共␪兲 g ⬘共 ␪ 兲

.

共5兲

In this approach with B 储 c 共␪ = 0°兲, the apparent g-values g⬘ and A⬘ are equal to g储 and A储, respectively, while for B ⬜ c 共␪ = 90°兲, one obtains

⬘ = 2g⬜ g⬜

and

⬘ = 2A⬜ . A⬜

共6兲

Whereas for ␪ = 0°, the parameters g⬘ and A⬘ are identical to the corresponding g- and A-values, the exact diagonalization of the SH 关Eq. 共1兲兴 gives very small corrections for ␪ ⫽ 0ⴰ. These small contributions are given in good approach by the perturbation theory using the correction up to the third order in the Zeeman energy.19 The deviation for B ⬜ c 共␪ = 90°兲 is given by g⬜ ⬘ = 2g⬜关1 − 共3 / 16兲共g⬜␮BB / D兲2兴, which raises the g⬜-value calculated with Eq. 共4兲 only by 2.2⫻ 10−3 using the D-value of 2.75 cm−1 given by Estle et al.17 The magnetic field was applied in the 共1210兲 plane of ZnO, the angle ␪ being varied between 0° and 180°. The SH parameters were obtained by fitting the experimental spectra with the Minirock program.20 The analysis of the spectra revealed that they consist of two components, A and B, as illustrated in Fig. 2 for B 储 c. The intensity ratio between the two components allows us to determine the concentration ratio of the centers NB / NA = 1.4. The parameters A储, A⬜, g储, and g⬜ determined for both components of the spectra are given in Table I. The angular dependence of the experimental Co2+ EPR spectrum together with the HF splitting calculated using parameters given in Table I is shown in Fig. 3. In Fig. 4, the angular dependencies of the experimental line positions for

Downloaded 30 Jan 2007 to 193.136.173.43. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp

024324-3

J. Appl. Phys. 101, 024324 共2007兲

Ankiewicz et al.

FIG. 2. 共a兲 Experimental EPR spectrum of Co2+ in ZnO nanowires 共nominal content of 5 at. %兲, measured in the X-band at 4.2 K for B 储 c 共␪ = 0 ° 兲. 共c兲 Result of the fitting of the experimental spectrum by the sum of components A and B shown in 共d兲 and 共b兲, respectively.

both components A and B are compared with the calculated ones using Eq. 共4兲 supplemented with the small third-order corrections in the Zeeman energy.19 The linewidths 共for B 储 c, ⌬B pp = 0.7 and 1.9 mT for components A and B, respectively兲 were larger than that reported in the literature 共⌬B pp = 0.04 mT for B 储 c兲.21 Contributions to the broadening may occur due to the presence of a high defect concentration and random strains in the wires. However, the main part is probably caused by some irregularity of the wires arrangement observed by SEM. All determined SH parameters, except g储, agree fairly well with those reported in the literature.17,21 The g储-values are slightly too large, lowering the line position by about 0.6 mT for component A. This suggests that the sample was slightly misoriented during the measurement 共in this case we did not measure the exact values for B 储 c兲. Using the data from literature,17 one finds that a misalignment of only ⬃2.0ⴰ can be responsible for this deviation. Thus, the misalignment probably comes from the misorientation of the sample holder’s surface in the cavity. Nevertheless, it was shown by x-ray diffraction that the substrate was cut correctly and the wires were well aligned.22 Additional information about the magnetic state of the nanowires can be obtained from the variation of the EPR spectrum intensity I with the measurement temperature T. In the case of cations without exchange interaction, I is proportional to the difference in the population of the four lowest levels, whose energies can be calculated using the SH given by Eq. 共1兲. In Fig. 5, we show the EPR spectrum for B 储 c, which was the best resolved one, for temperatures between 5 and 40 K. The spectrum could be measured up to 90 K, although for T ⬎ 40 K we could not resolve the HFS any

FIG. 3. Angular dependence of the EPR spectra of Co2+ in ZnO nanowires 共nominal content of 5 at. %兲, measured in the X-band at 4.2 K for the rotation of the magnetic field B in the 共1210兲 plane of ZnO. ␪ = 0° corresponds to B 储 c. The result of the fitting of the angular dependence of the HF line positions is plotted in solid lines on top of the spectra.

FIG. 4. Experimental values of the centers of gravity of the A 共dots兲 and B 共squares兲 EPR spectra, as extracted from the fittings, and calculated 共solid lines兲 angular variations of the line positions of the 兩±1 / 2典 transitions for Co2+ in ZnO nanowires 共nominal content of 5 at. %兲. The spectra were measured in the X-band at 4.2 K.

TABLE I. SH parameters for Co2+ in Co-doped ZnO nanowires, at T = 4.2 K. Except for g, all values are given in 10−4 cm−1. For Co2+, Estle et al. 共Ref. 17兲 determined 2D = 共5.5± 0.3兲 cm−1. x

兩A储兩 兩2A⬜兩 g储 g⬜

5 at. % A

B

Ref. 17

16.2± 0.6 3.2± 1.0 2.247± 0.001 2.276± 0.001

16.8± 2.0 3.2± 1.0 2.245± 0.002 2.276± 0.007

16.11± 0.05 3.00± 0.03 2.243± 0.001 2.2791± 0.0002

FIG. 5. Experimental temperature dependence of the EPR spectra for the Co2+ in ZnO nanowires 共nominal content of 5 at. %兲, measured for B 储 c in the X-band.

Downloaded 30 Jan 2007 to 193.136.173.43. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp

024324-4

J. Appl. Phys. 101, 024324 共2007兲

Ankiewicz et al.

FIG. 7. Temperature dependence of the spectral widths for the Co2+ lines in ZnO nanowires 共nominal content of 5 at. %兲, measured for B 储 c in the X-band. Dots represent the A component and squares the B component of the spectra.

FIG. 6. Temperature dependence of the EPR intensity 共top兲 and inverse intensity 共bottom兲 of the A 共dots兲 and B 共squares兲 components of the 兩±1 / 2典 transition of the Co2+ spectrum in ZnO nanowires 共nominal Co content of 5 at. %兲, measured in the X-band for B 储 c. The dashed lines are the calculated curves for the 兩±1 / 2典 transition within the S = 3 / 2 manifold, using the zerofield splitting of D = 2.75 cm−1 given by Estle et al. 共Ref. 17兲. The solid lines represent the same calculated lines but including a correction that takes into account a small error that scales linearly with temperature.

longer. Figure 6 shows the temperature dependence of the EPR intensity and of the inverse intensity of the 兩±1 / 2典 transition for Co2+ in ZnO nanowires 共nominal content of 5 at. %兲, measured in the X-band, from 5 to 40 K. Furthermore, Fig. 6 also shows the curves of the temperature dependence for the 兩±1 / 2典 transition of both components A and B within the S = 3 / 2 manifold calculated using the zero-field splitting of D = 2.75 cm−1 given by Estle et al.17 We find that both components exhibit a paramagnetic behavior. Moreover, in the same figure, we plot a simulation including a correction that takes into account a possible small intensity error linearly dependent on the temperature. The simulation now perfectly agrees with the experimental data. This error might be caused by any of the following factors or a combination of them: 共i兲 miscalibration of the thermoelement, 共ii兲 variation of the quality factor of the microwave cavity with temperature, and 共iii兲 setting in of a saturation of the EPR transition with decreasing temperature due to increasing spin-lattice relaxation time 共see below the discussion of the temperature behavior of the linewidth兲. With the correction necessary to perfectly fit the measured values being very small, none of the above reasons can be excluded. On the other hand, such small effects are difficult to be checked with certainty. We find that the curvature of the 1 / I versus T dependence caused by a large zero-field splitting can be misleading, giving the impression of the existence of a finite Curie-Weiss temperature. Thus, we do not believe to observe any ferromagnetic coupling between the Co2+ spins in our sample. Figure 7 shows the temperature dependence of the linewidth. The increase of the linewidth with temperature due to

the increasing spin-lattice relaxation explains why the HFS is no longer resolved for T ⬎ 40 K and the signal cannot be measured, at least with the same microwave power, for T ⬎ 90 K. Taking into account that both components, A and B, follow the same angular and temperature dependencies, and that the essencial difference between them is the linewidth, we suggest that these components of the spectrum describe two different environments of the Co2+ ions. These different environments may be due to different local concentration of Co ions and/or different local strains. B. Mn incorporation

In the case of manganese, the trigonal component of the crystal field 共CF兲 and the spin-orbit interaction split the S = 5 / 2, 6A1 ground state into three Kramers doublets E±1/2, E±3/2, and E±5/2 with zero-field splittings equal to 2D and 4D, respectively. The SH for S = 5 / 2 is given by23

冋

ˆ = ␮ B · g · S + S · A · I + D S2 − 35 H S B Z 12 −

冋

册 冋

册

册

7F 4 95 2 81 707 a SZ − SZ + + 共S␰4 + S␩4 + S␨4兲 − . 14 16 16 36 6 共7兲

The axially symmetric component of the CF, characterized by the terms proportional to D and F, lies on the Z axis; i.e., the c axis of the hexagonal crystal. In the wurtzite structure, this corresponds to the 关111兴 axis of the cubic tetrahedral field component defined by the 共␰ , ␩ , ␨兲 coordinate frame, which is proportional to a. From the SH given by Eq. 共7兲, we obtain the following energy eigenvalues for B 储 c:23 W±5/2 = ± g␮BB0 − D/3 + 共a − F兲/2 ±

冑冋

3D +

a−F 3 ⫿ g ␮ BB 0 2 6

3 2 W±3/2 = ± g␮BB0 + D − 共a − F兲, 2 3

册

2

+

20 2 a , 9 共8兲

Downloaded 30 Jan 2007 to 193.136.173.43. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp

024324-5

J. Appl. Phys. 101, 024324 共2007兲

Ankiewicz et al.

W±1/2 = ⫿ g␮BB0 − D/3 + 共a − F兲/2 +

冑冋

3D +

a−F 3 ± g ␮ BB 0 2 6

册

2

+

20 2 a . 9

The experimental angular dependencies of the EPR spectra of the ZnO nanowires containing 3 at. % and 10 at. % of Mn are shown in Figs. 8 and 9. Because the parameter a is small as compared to the Zeeman energy and to D, the terms proportional to 20a2 / 9 can be neglected in a good approximation, and the line positions for the FS transitions are given by 4 B+5/2↔+3/2 = B0 − 4D⬘ + 共a⬘ − F⬘兲, 3

FIG. 8. Experimental angular dependence of the EPR spectra of Mn2+ in ZnO nanowires 共nominal content of 3 at. %兲, measured in the X-band at 4.2 K.

5 B+3/2↔+1/2 = B0 − 2D⬘ − 共a⬘ − F⬘兲, 3 B+1/2↔−1/2 = B0 ,

共9兲

5 B−1/2↔−3/2 = B0 + 2D⬘ + 共a⬘ − F⬘兲, 3 4 B−3/2↔−5/2 = B0 + 4D⬘ − 共a⬘ − F⬘兲 3 with B0 = h␯ / 共␮Bg兲. The prime indicates that the values are in magnetic field units 共mT兲. From the fittings of the experimental spectra we extracted the line positions for B 储 c 共␪ = 0°兲. This way we have determined g储, D, and 兩a − F兩 given in Table II. The HF splitting parameter A储 was determined using Eq. 共3兲. The values of the SH parameters do not vary significantly with the Mn content, in the studied Mn concentration range, as observed by Diaconu et al. in Mndoped ZnO thin films,25 and are very near to the ones observed for Mn-doped single crystals.24 Figure 10 shows the EPR spectra measured at 4.2 K and ␪ = 0° for the two Mn concentrations. With increasing Mn content, the linewidth increases due to the dipole-dipole interaction of the paramagnetic ions. At the highest nominal concentration, xMn = 10 at. %, an additional broad line appears, probably stemming from local regions with higher concentrations of magnetic ions, where the HFS of the spectra is obscured by the dipole-dipole broadening and breaks down due to the exchange interaction.

FIG. 9. Experimental angular dependence of the EPR spectra of Mn2+ in ZnO nanowires 共nominal content of 10 at. %兲, measured in the X-band at 4.2 K. TABLE II. SH parameters for Mn2+ in Mn-doped ZnO nanowires, at T = 4.2 K. Except for g储, all values are given in 10−4 cm−1. x 兩A储兩 g储 D 兩a − F兩

3 at. % 76± 1 2.003± 0.001 −231± 1 6±1

10 at. % 78± 3 2.000± 0.002 −230± 3 7±3

⬍0.01 at. %a 74.1 1.9984± 0.0002 −235.5 5.4

a

Reference 24.

IV. CONCLUDING REMARKS

We investigated the incorporation of Mn and Co into nanometric ZnO wires grown by the PLD technique. For the lowest nominal concentrations, xMn = 3 at. % and xCo = 5 at. %, we detect the anisotropic EPR spectra of isolated Mn2+ 共3d5, 6S兲 and Co2+ 共3d7, 4F兲, respectively, on Zn sites. The ions could be unambiguously identified by their HFSs due to a nonzero nuclear spin 共I = 5 / 2 for Mn55 and I = 7 / 2 for Co59 with 100% natural abundance each兲 and their valence states by the observed FSs. For the Co-doped ZnO wires, only the FS transition 兩±1 / 2典 could be observed in the X-band, as the zero-field splitting 2D of the orbital singlet

FIG. 10. Experimental EPR spectra for B 储 c of Mn2+ in ZnO nanowires with two different nominal contents 共3 and 10 at. %兲, measured in the X-band at 4.2 K.

Downloaded 30 Jan 2007 to 193.136.173.43. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp

024324-6

ground state 4A2 is much larger than the Zeeman energy at the available magnetic fields up to 1.5 T. The detection of well-resolved anisotropic spectra proved a coherent crystallographic orientation of the nanowires. Nevertheless, it was possible to identify two different components, A and B, of the spectra that probably describe two different environments of the Co ions. From the temperature dependence of the EPR intensity we found that both components exhibit paramagnetic behavior and we do not believe to observe any ferromagnetic coupling between the Co2+ spins in our sample. The linewidths were larger than the best values reported in the literature21 and increase with increasing temperature. Contributions to the broadening may occur due to the presence of a high defect concentration and random strains in the wires. However, the main part is probably caused by some irregularity of the wires arrangement observed by SEM. In the case of the Mn-doped ZnO wires, the linewidth increases with increasing Mn concentration due to the dipole-dipole interaction of the paramagnetic ions. For the highest used nominal concentration, i.e., xMn = 10 at. %, an additional broad single line is observed. ACKNOWLEDGMENTS

This work has been supported by the SANDiE Network of Excellence of the EU, by the Fundação para a Ciência e a Tecnologia of Portugal 共project POCI/FIS/61462/2004 and bursary SFRH/BD/21659/2005兲, and by the DFG within FOR 522 共Project Gr 1011/11–1兲. We are very thankful to Gert Denninger for making his fitting program 共Minirock V3.4 05.04.2006 ©GWAD兲 available to us, with which we were able to fit all of our spectra. 1

J. Appl. Phys. 101, 024324 共2007兲

Ankiewicz et al.

Y. W. Heo, D. P. Norton, L. C. Tien, Y. Kwon, B. S. Kang, F. Ren, S. J. Pearton, and J. R. LaRoche, Mater. Sci. Eng., R. 47, 1 共2004兲. 2 Z. L. Wang, J. Phys. Condens. Matter 16, R829 共2004兲.

3

T. Dietl, H. Ohno, M. Matsukura, J. Cibert, and D. Ferrand, Science 287, 1019 共2000兲. 4 K. Sato and H. Katayama-Yoshida, Jpn. J. Appl. Phys., Part 2 39, L555 共2000兲. 5 H. Katayama-Yoshida, K. Sato, and T. Yamamoto, JSAP Int. 6, 20 共2002兲. 6 P. Sharma, A. Gupta, K. V. Rao, F. J. Owens, R. Sharma, R. Ahuja, J. M. Osorio Guillen, B. Johansson, and G. A. Gehring, Nat. Mater. 2, 673 共2003兲. 7 M. Diaconu, H. Schmidt, H. Hochmuth, M. Lorenz, G. Benndorf, J. Lenzner, D. Spemann, A. Setzer, K. W. Nielsen, P. Esquinazi, and M. Grundmann, Thin Solid Films 486, 117 共2005兲. 8 K. Ueda, H. Tabat, and T. Kawai, Appl. Phys. Lett. 79, 988 共2001兲. 9 C. Song, K. W. Geng, F. Zeng, X. B. Wang, Y. X. Shen, and F. Pan, Phys. Rev. B 73, 024405 共2006兲. 10 T. Ono, H. Miyajima, K. Shigeto, K. Mibu, N. Hosoito, and T. Shinjo, Science 284, 468 共1999兲. 11 D. J. Sellmyer, M. Zheng, and R. Skomski, J. Phys. Condens. Matter 13, R433 共2001兲. 12 C. Yu, D. Li, J. Pearson, and S. D. Bader, Appl. Phys. Lett. 79, 3848 共2001兲. 13 K. Liu, C. L. Chien, P. C. Searson, and K. Yu-Zhang, Appl. Phys. Lett. 73, 1436 共1998兲. 14 M. Lorenz, E. M. Kaidashev, A. Rahm, Th. Nobis, J. Lenzner, G. Wagner, D. Spemann, H. Hochmuth, and M. Grundmann, Appl. Phys. Lett. 86, 143113 共2005兲. 15 A. Rahm, E. M. Kaidashev, H. Schmidt, M. Diaconu, A. Pöppl, R. Böttcher, C. Meinecke, T. Butz, M. Lorenz, and M. Grundmann, Microchim. Acta 156, 21 共2006兲. 16 P. Koidl, Phys. Rev. B 15, 2493 共1977兲. 17 T. Estle and M. De Witt, Bull. Am. Phys. Soc. 6, 445 共1961兲. 18 A. Abragam and B. Bleaney, Electron Paramagnetic Resonance of Transition Ions 共Clarendon, Oxford, 1970兲. 19 J. Kreissl, W. Gehlhoff, P. Omling, and P. Emanuelsson, Phys. Rev. B 42, 1731 共1990兲. 20 G. Denninger, Minirock V3.4 共05.04.2006兲 共private communication兲. 21 N. Jedrecy, H. J. von Bardeleben, Y. Zheng, and J.-L. Cantin, Phys. Rev. B 69, 041308 共2004兲. 22 E. Alves 共private communication兲. 23 J. Schneider, Z. Naturforsch. A 17, 189 共1962兲. 24 A. Hausmann and H. Huppertz, J. Phys. Chem. Solids 29, 1369 共1968兲. 25 M. Diaconu, H. Schmidt, A. Pöppl, R. Böttcher, J. Hoentsch, A. Rahm, H. Hochmuth, M. Lorentz, and M. Grundmann, Phys. Rev. B 72, 085214 共2005兲.

Downloaded 30 Jan 2007 to 193.136.173.43. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp