Photoluminescence study of magnetic spin clusters

0 downloads 0 Views 509KB Size Report
Nov 9, 2012 - In contrast, the L3-component is due to the LEMPs formed ..... 31G. Mackh, W. Ossau, D. R. Yakovlev, A. Waag, and G. Landwehr, Phys. Rev.

Photoluminescence study of magnetic spin clusters and their temperature evolution in Cd0.70Mn0.30Te spin-glass compound Yu. P. Gnatenko and P. M. Bukivskij Citation: J. Appl. Phys. 112, 093715 (2012); doi: 10.1063/1.4764922 View online: http://dx.doi.org/10.1063/1.4764922 View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v112/i9 Published by the American Institute of Physics.

Related Articles Structural and optical properties of ZnCdO/ZnO multiple quantum wells grown on sapphire substrates using pulsed laser deposition J. Appl. Phys. 112, 083513 (2012) Pronounced Purcell enhancement of spontaneous emission in CdTe/ZnTe quantum dots embedded in micropillar cavities Appl. Phys. Lett. 101, 132105 (2012) Effects of Co content on the structural, luminescence, and ferromagnetic properties of Zn1−xCoxSy films J. Appl. Phys. 112, 063712 (2012) Mechanisms of infrared photoluminescence in HgTe/HgCdTe superlattice J. Appl. Phys. 112, 063512 (2012) Leaky mode analysis of luminescent thin films: The case of ZnO on sapphire J. Appl. Phys. 112, 063112 (2012)

Additional information on J. Appl. Phys. Journal Homepage: http://jap.aip.org/ Journal Information: http://jap.aip.org/about/about_the_journal Top downloads: http://jap.aip.org/features/most_downloaded Information for Authors: http://jap.aip.org/authors

Downloaded 09 Nov 2012 to 188.231.181.233. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

JOURNAL OF APPLIED PHYSICS 112, 093715 (2012)

Photoluminescence study of magnetic spin clusters and their temperature evolution in Cd0.70Mn0.30Te spin-glass compound Yu. P. Gnatenkoa) and P. M. Bukivskij Institute of Physics of NASU, Prospect Nauky 46, Kyiv 03028, Ukraine

(Received 5 October 2012; accepted 12 October 2012; published online 9 November 2012) We have investigated microscopic magnetic spin states (MMSSs) (“loose spins, finite superparamagnetic, ‘locked’ and infinite clusters”) both above and below the freezing temperature in Cd0.70Mn0.30Te spin glass (SG). We used the localized exciton magnetic polarons, which we observed in the photoluminescence spectra, as a probe. This makes it possible to estimate the MMSS’s relative concentrations and to study their temperature evolution and thus to elucidate one of the most important issues in this field of research. Furthermore, the findings described here open new prospects for further studies of spin freezing in the different SGs, C 2012 American Institute of Physics. especially, in dilute magnetic semiconductors. V [http://dx.doi.org/10.1063/1.4764922]

I. INTRODUCTION

Spin glass (SG) formation is one of the most complex and exciting problems in the condensed matter physics and material science.1–12 SG behavior has been found in many materials: crystalline and amorphous metal alloys,13,14 diluted magnetic semiconductors (DMSs),1–5,15 high temperature superconductors,16 insulating alloys,17 magnetic nanoparticles,18 etc. In order to understand more deeply the nature of SG freezing and the frozen state, it is necessary to carry out systematic and careful experimental studies of SG systems. Such research is mainly based on magnetic1–3,19 or magnetooptical4,5 measurements. In particular, it was shown that the transition from paramagnetic to SG phase is characterized by a sharp cusp in the low-field ac susceptibility at the freezing temperature (Tf).1,6,7 The thermoremanent magnetization studies indicate that there is a very slow dynamics of spins when the temperature is lowered towards Tf.2 Usually, a spin freezing process can be analyzed in terms of clustering.6–10 Far above Tf SG properties are explained by a collection of independent finite superparamagnetic clusters with local order and fluctuations.6,7 According to Morgownik and Mydosh’s theory,6,7 there are local magnetic inhomogeneities in the paramagnetic spin distribution, resulting in superparamagnetic clustering appeared at T < 5 Tf which is larger than it can be explained by random statistics and Mn-Mn near-neighbour interactions (small clusters). The clusters are in thermodynamical equilibrium with the phase of free spins at T > Tf.20 These finite dynamic superparamagnetic clusters (or magnetization fluctuations) vary in magnitude and direction from one Mn site to another, and are a function of time.21 A typical size of the superparamagnetic clusters for different SG compounds is about 3 nm.21–25 For temperatures reaching Tf (Tf < T < 2Tf), the cooperative behavior of spins (or “locked” clusters) becomes more important6–8,32 than behavior of independent clusters. At Tf, an infinite cluster appears a)

Author to whom correspondence should be addressed. Electronic mail: [email protected]

0021-8979/2012/112(9)/093715/6/$30.00

as a result of the interlocking of different clusters.12 The temperature range 0.7Tf  T  2Tf, we used to carry out our measurements, corresponds to strongly nonuniform distribution of Mn ions, because different types of clusters (finite, “locked” or infinite) coexist in this range. In spite of the intensive theoretical and experimental investigations of SG systems, a number of issues still remain open. In particular, the relative concentrations (RCs) of “loose” spins, i.e., the fraction of the spins which do not belong to any clusters (single spins),26 and various magnetic spin clusters in SGs are one of the important unanswered questions. Another problem is lack of detailed quantitative information on how these microscopic magnetic spin states (MMSSs) evolve with temperature, as well as how temperature affects the redistribution between the spin states. The alloy we investigated in this research, Cd0.70Mn0.30Te DMS, is a typical three-dimensional Heisenberg semi-insulating SG system.3,12,26 Only a short range anti-ferromagnetic interaction between Mn2þ ions takes place. The SG freezing in this material is a result of the spins’ frustration on a fcc sublattice and occurs both above and below the percolation threshold of XC  0.136.27 One of the brightest effects taking place in DMSs is the formation of exciton magnetic polarons.28–31 The effect occurs due to the strong exchange interaction between the spins of carriers associated with excitons (bound electronhole pairs) and spins of magnetic ions. In this case, the primary-carrier localization is necessary. Such localization occurs on component content alloy fluctuations or on magnetization fluctuations. As a result of the exchange interaction, the localized exciton magnetic polarons (LEMPs) are formed. For Cd1XMnXTe with large Mn-component content (X  0.10), the magnetization fluctuations at low temperatures (near Tf) are created by local magnetic fields due to the formation of different magnetic spin clusters. The exchange constant and effective mass of holes in Cd1XMnXTe are considerably greater than those in the case of the electrons. Thus, the energy of the LEMPs is determined by the exchange interaction of the heavy-hole spins of excitons

112, 093715-1

C 2012 American Institute of Physics V

Downloaded 09 Nov 2012 to 188.231.181.233. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

093715-2

Yu. P. Gnatenko and P. M. Bukivskij

with the total magnetic moment of clusters. The exchange field splits the levels of excitons localized as a result of their primary localization and, thus, induces additional decreasing of their energy due to the formation of the LEMPs in the regions of different MMSSs. The value of the decreasing energy depends on the value of the exchange fields. Thus, the magnetic polaron effect is larger for clusters with large total magnetic moment, i.e., the magnetic polaron effect for the LEMPs localized in the region of the “locked” clusters is greater than that in the region of the finite clusters. The low-temperature PL spectra of DMSs are dominated by a broad band. The shape of this band is defined by the recombination of LEMPs, which are formed in different regions of MMSSs. At low temperatures, the energy of LEMPs is determined by the radiative recombination from the lowest excitonic states which arise due to both the primary localization and the splitting of the localized exciton states by the exchange field. The Stokes shift between the PL band maximum and the position of the free exciton reflects the characteristic energy of localization. This energy includes both the primary localization, mainly caused by the compositional fluctuations, and the magnetic contribution due to the formation of the exciton magnetic polarons for the localized exciton states. Consequently, the PL bands of the LEMPs possess important information about the MMSSs, where excitons are localized. Therefore, they can be used to probe these magnetic states. In this paper, we report the first estimation of the RCs of various MMSSs: (1)“loose” spins, (2) finite, (3)“locked,” and (4) infinite clusters in Cd0.70Mn0.30Te (Tf ¼ 6.45 K32) SG in the vicinity of its freezing temperature (0.7Tf  T  2Tf). These results are based on the analysis of lowtemperature PL spectra structure caused by the appearance of various LEMPs. This made it possible to study the MMSSs temperature evolution and correlation between these magnetic states. II. EXPERIMENTAL DETAILS

The Cd1XMnXTe bulk crystals were grown by the vertical Bridgman technique. The initial concentration of Mn atoms was X ¼ 0.30. The actual concentration of Mn atoms for the investigated sample was measured using the following equation: Eex (X) ¼ 1.596 eV þ 1.61 eV,33 where Eex (X) determines the energy position of the free exciton reflection band, and equals to 29.9 6 0.1%. The measurement of the PL and reflection spectra were carried out from freshly cleaved sample surface. The PL measurements were carried out using a SDL-1 grating spectrometer with a high aperture and a variable temperature liquid-helium cryostat. The temperature stabilization by the UTREKS system was 0.01 K. The PL measurements were performed with 1000 s delay after the temperature stabilization, because the spins relaxation in Cd0.70Mn0.30Te near Tf is very slow (100 s). The duration of the PL spectrum measurement was about 1 h. In order to avoid the heating of the samples, we used a low-power diode-pumped solid state (DPSS) laser (k ¼ 532 nm, P ¼ 2.6 mW) produced by Orion Telescopes & Binoculars Com.

J. Appl. Phys. 112, 093715 (2012)

FIG. 1. Temperature dependence of the PL spectra of Cd0.70Mn0.30Te crystal obtained using DPSS laser excitation (k ¼ 532 nm, P ¼ 2.6 mW) after cooling of the crystal sample in the dark.

III. RESULTS AND DISCUSSION

The temperature dependence of the PL bands in Cd0.70Mn0.30Te is shown in Fig. 1. Below Tf, the appearance of a strong long-wavelength (LW) asymmetry of PL band (L1 component) is caused by the exciton localization in the crystal region where an infinite cluster is formed. As illustrated in Fig. 1, two other components, indicated by arrows, are also clearly observed at the temperatures of 16.0 K and 20 K. Their nature is similar to that obtained for Cd1XMnXTe (X ¼ 0.127) in Refs. 34 and 35. This means that the L2-component is mainly caused by the emission of LEMPs formed in the crystal region of the finite superparamagnetic clusters due to exchange interaction between carriers spins and the total magnetic moment of these clusters. In contrast, the L3-component is due to the LEMPs formed in the crystal region of “loose” spins as a result of exchange interaction between the hole and electron spins and “loose” spins. As can be seen from Fig. 1, the relative intensity of L2-component decreases with increasing temperature. The decreasing intensity of this component is caused by temperature destruction of the local magnetic fields of the finite clusters. The intensity of L3-component increases with temperature because their concentration grows with increasing temperature at the expense of the finite and “locked” clusters due to their temperature destruction. In addition to above mentioned components, the background on the LW wing of bands shows up. This background is caused by the emission associated with intracenter transitions between the excited 4T1- and the main 6A1-states of Mn2þ ions. Taking into account above circumstances, we decomposed the PL bands into four components using OriginLab (OriginPro 8) software. It should be noted that the peak fitting of the observed PL bands has produced a single unique combination of different peaks (or components) including their location, intensity, and width. Then, the RCs, namely, integrated intensities, of three excitonic components were determined. Fig. 2(a) shows PL band for Cd0.70Mn0.30Te crystal at T ¼ 4.5 K and its components. Fig. 2(b) presents the exciton reflection (curve 1) and the PL spectra (curve 2) at T ¼ 4.5 K obtained using incandescent lamp excitation of Cd0.70Mn0.30Te crystal. The excitonic reflection spectrum is due to the excitation of free excitons with the energy of EFex,

Downloaded 09 Nov 2012 to 188.231.181.233. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

093715-3

Yu. P. Gnatenko and P. M. Bukivskij

FIG. 2. (a) PL spectrum of Cd0.70Mn0.30Te crystal at T ¼ 4.5 K. The PL data were obtained using a DPSS laser excitation (k ¼ 532 nm, P ¼ 2.6 mW) after cooling of the crystal sample in the dark. Curves 1–3 correspond to the L1-, L2-, and L3-components, respectively. Curve 4 is associated with the emission of Mn2þ ions caused by the intracenter transitions. (b) Reflection and photoluminescence spectra of Cd0.70Mn0.30Te crystal at T ¼ 4.5 K, where a schematic representation of the optical transitions with the formation of LEMPs is also shown.

while the PL spectrum is caused by the PL of LEMPs. Thus, a dispersion-like structure in the reflection spectrum observed at ˚ is a result of the photogeneration of free excitons. k1 ¼ 5970 A ˚ is caused by PL spectrum with the maximum at k2 ¼ 6120 A the emission of various types of localized excitons. The simultaneous presence of both free excitonic reflection band and the PL band of the LEMPs under excitation with incandescent lamp indicates excellent optical quality of the investigated crystals. In Fig. 2(b), where PL spectrum is observed, the schemes of the optical transitions are shown by dotted arrows directed down. These optical transitions are caused by recombination of different LEMPs’ components. The solid and dashed arrows directed upwards show the location of different PLcomponents. At low temperatures, recombination of the LEMPs occurs from the lowest localized exciton states. Unlike for regular semiconductors (not alloys), for DMSs the low-energy tails in the density of excitonic states appear, as shown in the schemes (solid lines). This is due to both the primary localization of excitons caused by the random distribution of alloy components (non-magnetic contribution) and the result of the magnetic polaron formation (magnetic contribution) for these excitonic states. The total localization energy for various LEMPs is shown as DElocex1, DElocex2, and DElocex3 values. The dashed curves indicate the threshold

J. Appl. Phys. 112, 093715 (2012)

energy between the delocalized and the localized excitons for alloys. The transition from delocalized to localized exciton states is not sharp. The energy position of LEMPs formed in the region of “locked” clusters (L1-component) will be smaller than that for LEMPs appeared in the region of finite clusters (L2-component), since the local magnetic fields of “locked” clusters is larger than that of finite clusters. It should be noted that the observation of different components of LEMPs, as follows from our work, is possible under the following experimental conditions: (1) measuring the PL spectra with use of low power laser excitation (less than 70 mW), (2) cooling of the crystal sample from 2Tf to 0.7Tf must be carried out in the dark; (3) using only high optical quality crystal samples for the measurements. These conditions let us to observe the different LEMPs in the PL spectra. Figure 3(a) shows the temperature dependence of the RCs for different components reflecting the contributions of the “loose” spins, finite, and infinite (for T  Tf) or “locked” (for Tf < T  2Tf) clusters and their temperature evolution obtained using a low-power DPSS laser after cooling the crystal sample in the dark. It is assumed that the sum of the excitonic components is equal to 100%. At Tf, the proportions of these parts are equal to 5:20:75. It should be noted that the RC value for the “loose” spins (5%) correlates with the effective concentration X* of Mn atoms in the Cd1XMnXTe SGs. In this case, it is necessary to compare the RC value with the relative effective concentration X*R obtained from X*R ¼ X*/X (100)%, where X* ¼ XM(X)/ M(X ¼ 5 at. %) is the effective concentration of Mn atoms (for X ¼ 0.30 X*  0.015 (Ref. 3). The increase of the RC for infinite cluster with decreasing temperature from Tf to 4.5 K indicates that the cluster continues to grow at the expense of the finite clusters and its contribution equals to 78.5% at T ¼ 0.7Tf. The RC for the “locked” clusters comes down (2.5%) with decreasing temperature from 1.15Tf to Tf. This may be related to the fact that not all “locked” clusters are involved in the formation of the infinite cluster at Tf. It is possible that some of them break down and, thus, the number of finite clusters (1.5%) and “loose” spins (1%) increases. At T > 1.15Tf, the RC of “locked” clusters decreases with increasing temperature and at T > 1.5Tf this dependence is linear and its extrapolation indicates that these clusters disappear at T  3Tf, i.e., the intercluster interaction in Cd0.70Mn0.30Te SG is effective below this temperature. Thus, the increase of temperature up to 14.0 K leads to the significant redistribution of the spins between various magnetic spin clusters as well as the “loose” spins. It was found that the values of RCs do not depend on the laser excitation power up to 70 mW. It means that the excitons at such power excitation do not affect the distribution of different MMSSs in Cd0.70Mn0.30Te SG and consequently can be used to probe these states. The half-widths (HW) of LEMP’s bands are caused by the compositional fluctuations (non-magnetic contribution) and the MMSS’s size fluctuations (magnetic contribution). First contribution to the inhomogeneous broadening of the bands is practically invariable at low temperatures. Consequently, the temperature dependence of the HWs of LEMP’s bands is due to the inhomogeneous distribution of local

Downloaded 09 Nov 2012 to 188.231.181.233. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

093715-4

Yu. P. Gnatenko and P. M. Bukivskij

J. Appl. Phys. 112, 093715 (2012)

(curve 3) occurs. Such dependence correlates with the corresponding dependence for the “locked” clusters. This indicates that ID for “loose” spins is mainly defined by the dissociation of “locked” clusters. It should be noted that the distribution of “loose” spins is strongly inhomogeneous since they are also located at the places where the internal fields coming from different clusters cancel out or at least do not contribute more than the thermal energy kBT.25 IV. CONCLUSIONS

FIG. 3. (a) Temperature dependence of the RCs of the PL band excitonic components obtained using a DPSS laser (P ¼ 2.6 mW). (b) Temperature dependence of the half-widths of the PL band excitonic components obtained using a DPSS laser (P ¼ 2.6 mW). Curves 1–3 correspond to the L1-, L2-, and L3-components, respectively.

magnetic fields for different MMSSs. Figure 3(b) shows the temperature dependence of the HWs for different PL band components. Analysis of these dependences gave us information about inhomogeneity distribution (ID) of various MMSSs and their temperature evolution in Cd0.70Mn0.30Te SG. This shows that the HW values at 0.7Tf  T  1.15Tf do not change, which obviously indicates the independence of ID for the magnetic clusters and “loose” spins in this temperature range. At 1.15Tf < T < 2Tf, the decrease of HW for the L1-component (curve 1) reflects the improvement of the “locked” clusters ID that may be caused by reduction of their size. The ID for finite clusters (curve 2) is weakly dependent on the temperature showing a minimum at T ¼ 1.5Tf, which coincides with the minimum for RC. It can be assumed that at 1.15Tf  T  1.5Tf decrease RC and improvement ID for finite clusters may be caused by the destruction of larger finite clusters, while the destruction of “locked” clusters leads only to emergence of additional “loose” spins. The increasing of the RC and ID values for finite clusters at 1.5Tf  T  1.75Tf can be explained by the increase of finite clusters at the expense of the destruction of “locked” clusters. At 1.15Tf < T  2Tf, an increase in HW of the “loose” spins

We have carried out a detailed PL study of high optical quality Cd0.70Mn0.30Te SG compound using LEMP as nano-size probe of the various MMSSs in this material. This enables us to estimate the RCs of “loose” spins, finite superparamagnetic, and infinite or “locked” clusters. This is the first such quantitative evaluation for any SG compounds. In addition, we have also investigated the temperature evolution of various MMSSs both above and below Tf. Surprisingly, there is a direct correlation between the RCs and the IDs of “locked” clusters and “loose” spins. Our studies are crucial for better understanding of the MMSSs temperature evolution in the dynamical freezing range 0.7Tf < T < 2Tf, since the available concepts of SG treat the temperature evolution of various spin clusters only qualitatively. The results obtained in this work may be used both for better understanding of the nature of aging, memory, and chaos effects36,37 as well as of the intrinsic photoinduced magnetic responses10 in SGs and for more effective using of these effects in various practical applications. Thus, this work provides a better insight into challenging issues of the freezing process of spins and also opens doors for a broad range of research opportunities to future investigations of the MMSSs in semi-insulating SGs. Moreover, our findings can be applied to other SGs since the present results show that at Tf < T < 2Tf the “locked” clusters are primarily responsible for the temperature dependence of magnetic properties both the semi-insulating and metallic6,7 as well as amorphous SGs.8 In particular, the two-fold increase of the magnetic susceptibility value for these SGs when temperature is lowered from 2Tf to Tf is correlated with the increase in the RC of “locked” clusters. The data concerning the evolution of various MMSSs obtained in this work may be especially useful for DMSs, both bulk and nanostructured magnetic systems. These objects are very promising materials for nanomagnetism and spintronics, because they show ferromagnetic properties near room temperature.38–41 In particular, magnetization measurements and a scanning photoelectron microscopy analysis of Ge1XMnX (X ¼ 0.04) indicate that ferromagnetism in Mndoped Ge is not of intrinsic nature.23,42 This is due to the presence of two different magnetic phases: (1) superparamagnetic Mn5Ge3 clusters undergoing a blocking transition around Tb ¼ 210 K and (2) superparamagnetic Mn-rich nanoclusters performing a blocking (or freezing) transition around Tf ¼ 12 K. These clusters are embedded in a crystalline Ge matrix with low Mn concentration and, in contrast to the spin glasses, are not dynamic superparamagnetic clusters having intrinsic nature. Below Tb, the magnetization of

Downloaded 09 Nov 2012 to 188.231.181.233. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

093715-5

Yu. P. Gnatenko and P. M. Bukivskij

Mn5Ge3 clusters aligns along the magnetic easy direction of magnetic anisotropy. In order to reorient the magnetization by 180 , an energy barrier EB has to be overcome. The possible presence of weak interparticle dipolar interactions results in a spread of the individual EB’s, resulting in inhomogeneous freezing. For strong interactions, homogeneous freezing should be expected.23 In Ref. 42, it is indicated that the superparamagnetic response below 150 K is attributed to the presence of Mn-rich nanoclusters. Alternatively, the magnetization below 100 K may be explained43 by the formation of bound magnetic polarons (BMPs).44 The formation of BMPs is the result of the exchange interaction of localized holes with some surrounding magnetic atoms. These local ferromagnetic regions grow in size, finally leading to a disordered glassy state at low temperatures. It should be noted that, regardless of the real physical nature of the low-temperature magnetic state, a complex magnetic behavior of Ge1XMnX alloys occurs which is inherent to magnetically inhomogeneous systems showing glassy behavior. The origin of high-temperature ferromagnetism in another DMS material, namely, Mn-doped ZnO system, is controversial. It was suggested40 that the high-temperature ferromagnetism is rather caused by an oxygen-vacancy-stabilized metastable Mn2XZnXO3d phase than by carrier-induced interaction between separated Mn atoms in ZnO.38 This phase arises from diffusion of Zn into the Mn-oxide grains. In our opinion, various microscopic magnetic states inherent to above mentioned materials may also be studied by measurements of their PL spectra. It is expected that LEMPs with different localization energies will be formed in magnetically inhomogeneous semiconductor systems. The temperature dependence of LEMPs will reflect the temperature evolution of different magnetic inhomogeneities. Furthermore, information on the possible interparticle interaction near the freezing temperature may also be obtained. These studies together with the magnetic, structural, and transport measurements of magnetically inhomogeneous semiconductor systems will help one to better understand the nature of their magnetic inhomogeneity as well as the freezing process in these materials. It should be noted that ZnO-Mn DMS is a very suitable system for PL studies since the exciton binding energy (60 meV) for this material exceeds thermal energy at room temperature (25 meV). We think that as a result of these comprehensive researches, the technology of materials with desired and controllable physical properties which will be suitable for spintronics could be developed. Thus, the present results may encourage researchers for more detailed studies of freezing process in various inhomogeneous magnetic glassy systems. This also opens intriguing prospects for further studies of spin freezing and frozen states in these systems, especially under influence of extrinsic factors (magnetic field, pressure, ultrasound, etc) as well as using time-resolved PL measurements. ACKNOWLEDGMENTS

The research has been supported by the National Academy of Sciences of Ukraine (Grants Nos. BC-139-15 and B-146-15) and Ministry of Education and Science of Ukraine

J. Appl. Phys. 112, 093715 (2012)

(Grant No. 01 10U001151). We wish to thank Professor Sergej Lyuksyutov for his kind help with a low-power DPSS laser and to Dr. O. A. Shigiltchoff and Dr. A. P. Bukivsky for technical assistance. 1

R. R. Galazka, S. Nagata, and P. N. Keesom, Phys. Rev. B 22, 3344 (1980). A. Mauger, J. Ferre, M. Ayadi, and P. Nordblad, Phys. Rev. B 37, 9022 (1988). 3 S. B. Oseroff, Phys. Rev. B 25, 6584 (1982). 4 B. Leclereq, C. Rigaux, A. Mycielski, and M. Menant, Phys. Rev. B 47, 6169 (1993). 5 S. Geschwind, A. T. Ogielski, G. Devlin, and J. Hegarty, J. Appl. Phys. 63, 3291 (1988). 6 A. F. J. Morgownik and J. A. Mydosh, Phys. Rev. B 24, 5277 (1981). 7 C. A. Mulder, A. J. van Duyneveldt, and J. A. Mydosh, Phys. Rev. B 23, 1384 (1981). 8 K. V. Rao, M. F€ahnle, E. Figueroa, O. Beckman, and L. Hedman, Phys. Rev. B 27, 3104 (1983). 9 A. E. Jacobs, Phys. Rev. B 32, 7607 (1985); 32, 7430 (1985). 10 H. Kawai and T. Sato, J. Appl. Phys. 85, 7310 (1999). 11 E. A. Goremychkin, R. Osborn, B. D. Rainford, R. T. Macaluso, D. T. Adroja, and M. Koza, Nat. Phys. 4, 766 (2008). 12 J. K. Furdyna, J. Appl. Phys. 61, 3526 (1987); 64, R29 (1988). 13 F. Wang, J. Shen, J. Zhang, J. Sun, and B. Shen, J. Appl. Phys. 105, 07A944 (2009). 14 R. M. Pickup, R. Cywinski, C. Pappas, B. Farago, and P. Fouquet, Phys. Rev. Lett. 102, 097202 (2009). 15 A. L. Chudnovskiy, R. Oppermann, B. Rosenow, D. R. Yakovlev, U. Zehnder, and W. Ossau, Phys. Rev. B 55, 10519 (1997). 16 L. Leuzzi, G. Parisi, F. Ricci-Tersenghi, and J. J. Ruiz-Lorenzo, Phys. Rev. Lett. 103, 267201 (2009). 17 V. M. Galitski and A. I. Larkin, Phys. Rev. B 66, 064526 (2002). 18 J. A. Quillian, S. Meng, C. G. A. Mugford, and J. B. Kycia, Phys. Rev. Lett. 101, 187204 (2008). 19 A. A. Timopheev, S. M. Ryabchenko, V. M. Kalita, A. F. Lozenko, P. A. Trotsenko, V. A. Stephanovich, and M. Munokata, J. Appl. Phys. 108, 033919 (2010). 20 M. Cyrot, Solid State Commun. 39, 1009 (1981). 21 S. M. Bhagat and H. A. Sayadian, J. Magn. Magn. Mater. 61, 151 (1986). 22 T. Giebultowicsz, B. Lebech, B. Buras, W. Minor, H. Kepa, and R. R. Galashka, J. Appl. Phys. 55, 2305 (1984). 23 C. Jaeger, C. Bihler, T. Vallaitis, S. T. B. Goennenwein, M. Opel, R. Gross, and M. S. Brandt, Phys. Rev. B 74, 045330 (2006). 24 R. B. Goldfarb and C. E. Patton, Phys. Rev. B 24, 1360 (1981). 25 R. Nigam, A. V. Pan, and S. X. Dou, Eur. Phys. J. B 74, 429 (2010). 26 M. Escone and A. Mauger, Phys. Rev. B 25, 4674 (1982). 27 N. A. Novak, O. G. Symko, and D. J. Zheng, Phys. Rev. B 33, 6391 (1986). 28 R. Fiederling, D. R. Yakovlev, W. Ossau, G. Landwehr, I. A. Merkulov, K. V. Kavokin, T. Wojtowicz, M. Kutrowski, K. Grasza, G. Karczewski, and J. Kossut, Phys. Rev. B 58, 4785 (1998). 29 K. V. Kavokin, I. A. Merculov, D. R. Yakovlev, W. Ossau, and G. Landwehr, Phys. Rev. B 60, 16499 (1999). 30 S. Takeyama, S. Adachi, Y. Takagi, and V. F. Aguekian, Phys. Rev. B 51, 4858 (1995). 31 G. Mackh, W. Ossau, D. R. Yakovlev, A. Waag, and G. Landwehr, Phys. Rev. B 49, 10248 (1994). 32 Y. Zhou, C. Rigaux, A. Mycielski, M. Menant, and N. Bontemps, Phys. Rev. B 40, 8111 (1989). 33 P. I. Babij, P. N. Bukivskij, and Yu. P. Gnatenko, Fiz. Tverd. Tela (Leningrad) 28, 3358 (1986). 34 P. N. Bukivskij, Yu. P. Gnatenko, and A. Kh. Rozhko, Pis’ma Zh. Eksp. Teor. Fiz. 61, 380 (1995) [JETP Lett. 61, 386 (1995)]. 35 P. N. Bukivskij, Yu. P. Gnatenko, and A. Kh. Rozhko, Fiz. Tverd. Tela (St. Petersburg) 37, 2916 (1995) [Phys. Solid State 37, 1610 (1995)]. 36 H. G. Katzgraber and F. Krzakala, Phys. Rev. Lett. 98, 017201 (2007). 37 K. Jonason, E. Vincent, J. Hammann, J. P. Bouchaud, and P. Nordblad, Phys. Rev. Lett. 81, 3243 (1998). 38 P. Sharma, A. Gupta, K. V. Rao, F. J. Owens, R. Sharma, R. Ahuja, J. M. Osorio Guillen, B. Johansson, and G. A. Gehring, Nature Mater. 2, 673 (2003). 39 Y. D. Park, A. T. Hanbicki, S. C. Erwin, C. S. Hellberg, J. M. Sullivan, J. E. Mattson, T. F. Ambrose, A. Wilson, G. Spanos, and B. T. Jonker, Science 295, 651 (2002). 2

Downloaded 09 Nov 2012 to 188.231.181.233. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

093715-6 40

Yu. P. Gnatenko and P. M. Bukivskij

D. C. Kundaliya, S. B. Ogale, S. E. Lofland, S. Dhar, C. J. Metting, S. R. Shinde, Z. Ma, B. Varughese, K. V. Ramanujachary, L. Salamanca-Riba, and T. Venkatesan, Nature Mater. 3, 709 (2004). 41 L. Chen, J. H. He, Y. Mei, Y. Z. Cao, B. Q. Liu, Z. W. Zhu, and Z. A. Xu, J. Appl. Phys. 105, 123910-1 (2009).

J. Appl. Phys. 112, 093715 (2012) 42

M. Passacantando, L. Ottaviano, F. D’Orazio, F. Lucari, and M. De Biase, Phys Rev. B 73, 195207 (2006). 43 A. P. Li, J. Shen, J. R. Thompson, and H. H. Weitering, Appl. Phys. Lett. 86, 152507 (2005). 44 A. Kaminski and S. Das Sarma, Phys. Rev. Lett. 88, 247202 (2002).

Downloaded 09 Nov 2012 to 188.231.181.233. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

Suggest Documents