- Email: [email protected]
Curie temperature in InMnP and the mechanism of phase transition
ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 309 (2007) 183–187 www.elsevier.com/locate/jmmm
Curie temperature in InMnP and the mechanism of phase transition C.S. Park, T.W. Kang Physics Department and Quantum-functional Semiconductor Research Center, Dongguk University, 3-26 Pil-dong, Chung-ku, Seoul, South Korea Received 17 April 2006; received in revised form 27 June 2006 Available online 31 August 2006
Abstract We present the phase transition of high Curie temperature InMnP system grown by liquid phase epitaxy. InMnP system has a distribution of uniform islands on the surface layer which has a width of 40 nm and height of 30 nm. Two kinds of phase transition in the magnetization were observed. The origin of these structures and their influence for the magnetization are discussed from the view of twodimensional Ising model. r 2006 Elsevier B.V. All rights reserved. PACS: 71.20.Nr; 71.55.Eq; 73.43.Qt; 75.50.Pp keywords: Curie temperature; Magnetization; Phase Transition; InP
1. Introduction Considerable attention has been focused on the application for spin-sensitive electronics, and the possibility of fabricating devices which have magnetic semiconductors for novel spintronic devices, i.e., diluted magnetic semiconductor (DMS) alloy, as well as the fundamental physical properties of such a system [1–4]. In III–Vsemiconductors, substituted Mn2+ ions lead to both local moment formation and act as acceptors. The holes mediate interaction between the Mn moments, correlating their orientation and making ferromagnetism possible. The eligible performance of the device for such a system strongly depends on the high Curie temperature and room-temperature operation for the magnetic spin alignments. A suitable example of the argument for the roomtemperature transition is well reported in the digital alloy systems. In (Ga,Mn)Sb system, the ferromagnetism at low temperatures is responsible for some similarities between the GaSb/Mn digital alloys and GaMnSb random alloys, and the observed ferromagnetism of such a system at higher temperatures is associated with the 2D MnSb islands (precipitates), which may be related to the high Corresponding author. Tel.: +820222603489; fax: +820222784519.
E-mail address: [email protected] (C.S. Park). 0304-8853/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2006.06.031
TC in MnSb [5]. However, the origin of high TC is still ambiguous and a detailed theoretical study is needed to fully understand the mechanism. Up to now, manganese-doped InP has mainly discussed as a deep acceptor of impurity levels by some group [6,7]. However, InP is very useful material in the area of compound semiconductor and is known to have excellent physical properties and good characteristics of devices. In view of many DMS materials, InP is considered as a possible candidate according to a theoretical report [4]. We had reported the study of characteristics for unintentionally doped n-type (6.1–6.3 1015 cm3) InP implanted with various doses of Mn [8]. In the case of the Mn concentration of 5 1015 cm2(1%) and 5 1016 cm2 (10%), the ferromagnetic hysteresis loops at 10 K appeared very weakly because ferromagnetic coupling coexisted with antiferromagnetic coupling, and the ferromagnetic hysteresis loop did not appear at 300 K. In contrast, the temperature-dependent magnetization of samples grown by liquid phase epitaxy shows the two-step magnetization at each Curie temperature. Curie temperatures were obtained near 40 and 300 K, together with the hysteresis loops. In this paper, we report the results of the high Curie temperature of InMnP co-doped with Zn and clarify the mechanism of phase transition.
ARTICLE IN PRESS 184
C.S. Park, T.W. Kang / Journal of Magnetism and Magnetic Materials 309 (2007) 183–187
2. Experimental details InMnP epilayers were grown by liquid phase epitaxy (LPE). Mn and Zn-doped InP layers are grown on Zndoped (1 0 0) p-type InP substrate. The concentration of Zn is 3 1018 cm3 and that of Mn is 5 1016 cm2 (3%). Before the growth of InP epilayer, the substrate was cleaned by trichloroethylene (TCE), ethanol, methanol and rinsed in deionized water for 5 min. The solvents were 6 N In melt and InP bulk, and Mn powder was used for the growth of In1xMnxP epilayers. After prebaking the In and InP melt at 657 1C for 2 h, appropriate amount of Mn was added to the growth melt. The growth was performed at 637 1C with one-step cooling process. The composition was determined using the lattice constant and the band gap energy measured by X-ray diffractometer (XRD) and was compared with energy-dispersive X-ray spectroscopy (EDX). The measurements of photoluminescence (PL) and y/2y XRD were performed to characterize the optical and structural properties of InMnP, and superconducting quantum interference device (SQUID) was used to characterize the magnetic properties and the phase transition temperature of InMnP. The surfaces of films were investigated by high-resolution scanning electron microscopy (HRSEM). 3. Results and discussion The carrier concentrations of sample determined from Hall effect measurements at room temperature has a metallic behavior (1020 cm3) but is not the exact value because of the existence of the anomalous Hall effect. Fig. 1 shows the photoluminescence spectrum measured at 15 K for a In0.97Mn0.03P epilayer co-doped with Zn. The peak near 1.38 eV, which is usually referred to as the Al peak, frequently appears in InP grown by various methods and recently was attributed to carbon as the acceptor. The
Fig. 1. PL spectra measured for the (In, Mn)P film with x3% at 15 K and the inset shows the XRD patterns, (0 0 2) and (0 0 4) of (In, Mn)P. Upper spectrum means the film with surface islands and lower without.
peak near 1.38 eV is an unidentified donor-carbon acceptor pair transition. With regard to the above transition, the first phonon replica appeared near 1.33 eV. The transition related to Zn appeared near 1.29 eV. The intensity of band luminescence related to Mn from 0.9 to 1.20 eV is low and the energy positions of impurities shift to lower energy region due to the surface state well known in InP, and the emission related to Mn has a broad band. It has been reported that the intentionally doped InP:Mn has a very high Mn emission than band-related emission spectra [6]. In our result, it is considered that there is a similar mechanism of the extinction of free carriers through nonradiative recombination via deep levels for the reason of low intensity although Mn is intentionally doped [7]. The high density of carriers results in tunneling transport through thermal energy between two Mn acceptor levels. The tunneling probabilities increase as temperature increases. While the dominant process at low temperatures is hole trapping, that at relatively high temperatures is the transport of hole hopping through the tunneling process [8]. Furthermore, this variance of hole concentrations is consistent with the Hall resistance data at room temperature which shows a negative differential conductivity behaviors with two energy levels from two step saturation of resistance in Fig. 3a. InP with the Mn concentration of high impurity is known to have strong phonon replica, n ¼ 0, n ¼ 1 and n ¼ 2 ranging to 1.1–1.2 eV with the contribution of coupled-band luminescence related to Mn [9]. We confirmed that the transition is ascribed to a welllocalized transition metal level, that is, ferromagnetic type Mn centers that possibly form in InP [10]. This is given by a following mechanism, e+Mn3+(3d4)+hMn2+(3d5)+h, (bound hole binding energy ¼ 0.23 eV). Fig. 2a shows the surface morphology of distribution of islands in InMnP alloy through HRSEM and the images without islands (Fig. 2b) at 300 K. These images show the distribution of islands on the surface layer although the sample has been grown by liquid phase epitaxy. The diameter and the height of islands have been estimated approximately as 40 and 30 nm, respectively, from two images above. The island has been analyzed using EDX spectra to investigate the creation of InMnP islands on the homoepitaxial layer. The concentration of Mn of InMnP islands is 3 1015 cm2, 3%. y/2y XRD measurement have been performed in order to estimate the structural properties of InMnP layer, and x composition from the lattice constant to the band gap energy is approximately 3% same as the EDX results. 2y scan shows the only high intensity of InP (0 0 2) and (0 0 4) plane as seen in the inset of Fig. 1. The Hall resistance RHall of a magnetic thin film is conventionally observed to contain two distinct components. The first results from the normal Hall effect which is proportional to the applied magnetic field, H, and the second, called the anomalous Hall contribution, is proportional to the magnetization: RHall ¼ R0 H þ RS M,
(1)
ARTICLE IN PRESS C.S. Park, T.W. Kang / Journal of Magnetism and Magnetic Materials 309 (2007) 183–187
185
Fig. 3. (a) Anomalous Hall Effect measurements for the (In, Mn)P film with surface islands with Van der Pauw method at 10, 150 and 300 K. The solid lines are linear fits and (b) magnetoresistance measurements for the (In, Mn)P film with surface islands at low temperatures. Fig. 2. (a) HRSEM image of (In, Mn)P film with x3% having surface islands. (b) HRSEM images of (In, Mn)P film with x3%.
where M is the magnetization, and R0, Rs are the ordinary and anomalous Hall coefficients, respectively. In general, the anomalous Hall effect is a consequence of spin–orbit coupling in the system. Fig. 3 indicates that both InMnP and InMnP islands are fundamentally ferromagnetic from the anomalous Hall effects at 10, 150 and 300 K. The results of anomalous Hall effect at 150 and 300 K in Fig. 3a indicates that the magnetization is attributed to InMnP island. The magnetization of InMnP film saturates at 1 T with a half-slope of anomalous Hall effect at 10 K. By the way, the slopes of saturation from the fits imply that InMnP film have a stronger saturation response at high temperature. The magnitude of slopes is 2.4 at 150 K and 4.5 and 0.22 at 300 K, which demonstrate that the strength of the hysteresis more than low temperature as seen a good agreement with the magnitude of the temperature-dependent magnetization of Fig. 4. Fig. 3b shows the field
Fig. 4. Temperature dependence of magnetization for the (In, Mn)P films with x3%, obtained by SQUID with a magnetic field of 1000 Oe. Upper spectrum means the film with surface islands and lower without. The inset 1 shows the temperature-dependent magnetization of InMnP after thermal diffusion of Mn(3%) co-doped with Zn using MBE. The inset 2 shows the hysteresis curves for the (In, Mn)P film with x3% at 10 and 300 K.
ARTICLE IN PRESS 186
C.S. Park, T.W. Kang / Journal of Magnetism and Magnetic Materials 309 (2007) 183–187
dependence of the low-temperature magnetoresistance. The resistivity (r ¼ (r(B)r(0))/r(0)) of sample was measured with the magnetic field perpendicular to the sample. r(B) is a resistivity of sample in applied magnetic field and r(0) means the initial resistivity without magnetic fields. A negative magnetoresistance as a function of applied field was observed below 40 K. This is ascribed to spin scattering interactions between the localized magnetic moments of the Mn acceptors and the holes in an impurity band. The magnitude of the negative magnetoresistance decreases as the temperature increases. Hysteresis in the magnetoresistance was observed at 5 K, which shows the evidence that remnant magnetization within the sample continues to decrease the resistivity of the sample when the magnetic field is removed. This behavior was reported for InMnAs at 3.5 K by Ohno et al. [11]. Curie point at low-temperature range is found to be 40 K as determined from the maximum temperature at which the negative magnetoresistance is observable [12]. The reappearance of magnetoresistance at 50 K means that the ferromagnetic behavior having the anomalous Hall contribution is caused by InMnP islands to be confirmed in Fig. 4. Fig. 4 shows the results of zero-field cooled temperaturedependent magnetization using SQUID. The transition temperature appeared to be 300 K and another transition appears near 40 K, whereas InMnP without islands does not have hysteresis. Moreover, phase transition is not found in the temperature dependence of magnetization, which means that this film has no ferromagnetic property even due to MnP precipitates. It is reported that InMnP treated with various methods by the implantation of Mn and the thermal diffusion of Mn has the same tendency of temperature dependence. In addition, the study of p-type InMnP:Zn co-doped with Zn(2.1–2.2 1018 cm3) was investigated [13]. Mn was evaporated on the top of (1 0 0) InP:Zn co-doped with Zn using molecular beam epitaxy (MBE) system. After the evaporation of Mn, the thermal diffusion of Mn was carried out by heat treatment. The inset 1 of Fig. 4 shows the temperature-dependent magnetization of samples annealed at 500 1C for 60 s and 550 1C for 30 s after the evaporation of Mn. The concentration of Zn (2.1–2.2 1018 cm3) and Mn (3 1015 cm3, 3%) is similar to those of Zn (3 1018 cm3) and Mn(3 1015 cm3, 3%) for our present sample. Two Curie temperatures were obtained near 50 and 300 K. We showed additionally the results of temperature-dependent magnetization obtained by thermal diffusion of Mn in the inset 1 of Fig. 4 so as to clarify data. The inset 2 of Fig. 4 shows that the ferromagnetic hysteresis curve at 10 K has a weak loop(HC60 Oe) with slight saturation, and the hysteresis curve at 300 K has a stronger loop (HC100 Oe) and significant saturation. Compared with these results, the ferromagnetic behavior of InMnP at high temperature is much stronger than that of InMnP at low temperature. From these results of hysteresis curves and temperature dependence of magnetizations, it is found that the response of magnetization of InMnP sample occurs strongly at high-temperature. There
is an interesting phenomenon that the transition temperature is shown at high temperature region in the upper curve of Fig. 4 contrary to the lower case of sample without island. We believe that the main reason for the significant increase of Curie temperature originates to the distribution of islands or MnP [14]. However, it is very important that the film has the two-dimensional Ising chains, which is made of islands or Mn precipitates. Even though the codoping and the activation of Mn deep acceptor is considered, the increasing value of Curie temperature is at most expected to the extent of 30 K through interlayer coupling [15]. For this reason, we consider the twodimensional Ising model for the distribution of islands on the surface layer in order to generalize. We apply the theory of ferromagnetism on the two-dimensional Ising model for our result of temperature-dependent magnetization. Spontaneous magnetization was reduced by introducing an artificial limiting process [16], which may be charge of the formation of either a two-dimensional ferromagnetism or metamagnetism. Among most of semiconductors, it is caused by that sometimes phase transition happen, and sometimes phase transition does not happen irrespective of the ferromagnetic hystesresis. Lattice configuration of samples has two kinds of Ising lattices, which are made of the elementary lattice and the additional island sites on the surface or MnP distribution. We know that there are uniform islands formation in the Fig. 2a that can be considered as a giant Ising lattice with the topological chains. Two free energies are defined with a distinguished energy difference. We assume the free energy forming island is significantly larger than the general lattice. Spontaneous magnetization per spin of two-dimensional ferromagnetism is given by [16] 8 ðT4T C Þ; > <0 2 1=4 2 4 1=8 ð1 þ z Þ ð1 6z þ z Þ mI ð0; TÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffi ðToT C Þ; > :¼ 1 z2 (2) where the magnetic moment per spin has been taken to be unity and ze2be. The transition temperature TC correpffiffiffi sponds to the value zC ¼ 2 1. If e ¼ F ¼ FL+FD, then z ¼ e2bðF L þF D Þ ¼ e2bF L e2bF D . Here, FL and FD mean the free energy to form the lattice and island in thermodynamic system during the crystal growth. In this case, there are two situations with the change of temperature variable kBT. kB is the Boltzman’s constant. As seen in Fig. 4, Curie points appear in two transition temperatures in the application of this model. First, in case of kBT5eL, the first Curie point is a response of zC1 ¼ e2bF L as e2bF D goes to 1, corresponding to a low TC of 40 K, as seen in the Fig. 4. Likewise, in case of eLokBT5eL, zC2 ¼ e2bF D as e2bF L goes to 1. The second Curie temperature of 300 K in the Fig. 4 corresponds to a Curie point of MnP, island on the surface. Even in the MnP precipitates, two phases of metamagnetic in low temperature and ferromagnetism at high temperature works on this hypothesis.
ARTICLE IN PRESS C.S. Park, T.W. Kang / Journal of Magnetism and Magnetic Materials 309 (2007) 183–187
4. Summary and conclusion In conclusion, InMnP doped with Zn was grown by LPE and it has islands on the surface layer. The results of magnetoresistance and anomalous Hall effect demonstrates that InMnP film has properties of ferromagnetic semiconductor. Temperature dependence of magnetization shows that the phase transition occurs at high TC of 300 K. Two mechanisms of phase transition exist in the sample according to InMnP, and MnP or InMnP islands, respectively. We obtained the correlation between the free energy of the two-dimensional lattice formation of islands or precipitates, and the Curie point is determined on the competitive relation of two-dimensional lattices. Acknowledgement This work was supported by the Quantum-Functional Semiconductor Research Center (QSRC) through the Korean Science and Engineering Foundation and by a research fund of the Dongguk University, 2005. References [1] H. Munekata, H. Ohno, S. von Molnar, A. Segmuller, L.L. Chang, L. Esaki, Phys. Rev. Lett. 63 (1989) 1849; H. Ohno, H. Munekata, T. Penney, S. von Molnar, L.L. Chang, Phys. Rev. Lett. 68 (1992) 2664.
187
[2] J. De Boeck, R. Oesterholt, A. van Esch, H. Bender, C. Bruyseraede, C. Van Hoof, G. Borghs, Appl. Phys. Lett. 68 (1996) 2744. [3] H. Ohno, Science 281 (1998) 951; G.A. Prinz, Science 282 (1998) 1660; H. Shimizu, T. Hayashi, T. Nishinaga, M. Tanaka, Appl. Phys. Lett. 74 (1999) 398. [4] T. Dietl, H. Ohno, F. Matuskura, J. Cibert, D. Ferrand, Science 287 (2000) 1019. [5] X. Chen, M. Na, M. Cheon, S. Wang, H. Luo, B.D. McCombe, X. Liu, Y. Sasaki, T. Wojtowicz, J.K. Furdyna, S.J. Potashnik, P. Schiffer, Appl. Phys. Lett. 81 (2002) 511. [6] L. Eaves, A.W. Smith, M.S. Skolnick, B. Cockayne, J. Appl. Phys. 53 (1982) 4955. [7] Y. Ishitani, S. Minagawa, H. Hamada, T. Tanaka, J. Appl. Phys. 82 (1997) 1336. [8] A.A. Burkov, L. Balents, Phys. Rev. Lett. 91 (2003) 057202. [9] T. Takanohashi, T. Tanahashi, M. Sugawara, K. Kamite, K. Nakajima, J. Appl. Phys. 63 (1988) 1961. [10] Y. Shon, W.C. Lee, Y.S. Park, Y.H. Kwon, S.J. Lee, K.J. Chung, H.S. Kim, D.Y. Kim, D.J. Fu, T.W. Kang, X.J. Fan, Y.J. Park, H.T. Oh, Appl. Phys. Lett. 84 (2004) 2310. [11] H. Ohno, H. Munekata, S. von Molnar, L.L. Chang, J. Appl. Phys. 69 (1991) 6103. [12] R.P. Khoslar, J.R. Fischer, Phys. Lett. 32A (1970) 96. [13] Y. Shon, H.C. Jeon, Y.S. Park, W.C. Lee, S.J. Lee, D.Y. Kim, H.S. Kim, H.J. Kim, T.W. Kang, Y.J. Park, C.S. Yoon, K.S. Chung, Appl. Phys. Lett. 85 (2004) 1736. [14] E.E. Huber, D.H. Ridgley, Phys. Rev. 135 (1964) A1033. [15] T. Jungwirth, W.A. Atkinson, B.H. Lee, A.H. MacDonald, Phys. Rev. B 59 (1999) 9818. [16] C.N. Yang, Phys. Rev. 85 (1952) 808.
Journal of Magnetism and Magnetic Materials 309 (2007) 183–187 www.elsevier.com/locate/jmmm
Curie temperature in InMnP and the mechanism of phase transition C.S. Park, T.W. Kang Physics Department and Quantum-functional Semiconductor Research Center, Dongguk University, 3-26 Pil-dong, Chung-ku, Seoul, South Korea Received 17 April 2006; received in revised form 27 June 2006 Available online 31 August 2006
Abstract We present the phase transition of high Curie temperature InMnP system grown by liquid phase epitaxy. InMnP system has a distribution of uniform islands on the surface layer which has a width of 40 nm and height of 30 nm. Two kinds of phase transition in the magnetization were observed. The origin of these structures and their influence for the magnetization are discussed from the view of twodimensional Ising model. r 2006 Elsevier B.V. All rights reserved. PACS: 71.20.Nr; 71.55.Eq; 73.43.Qt; 75.50.Pp keywords: Curie temperature; Magnetization; Phase Transition; InP
1. Introduction Considerable attention has been focused on the application for spin-sensitive electronics, and the possibility of fabricating devices which have magnetic semiconductors for novel spintronic devices, i.e., diluted magnetic semiconductor (DMS) alloy, as well as the fundamental physical properties of such a system [1–4]. In III–Vsemiconductors, substituted Mn2+ ions lead to both local moment formation and act as acceptors. The holes mediate interaction between the Mn moments, correlating their orientation and making ferromagnetism possible. The eligible performance of the device for such a system strongly depends on the high Curie temperature and room-temperature operation for the magnetic spin alignments. A suitable example of the argument for the roomtemperature transition is well reported in the digital alloy systems. In (Ga,Mn)Sb system, the ferromagnetism at low temperatures is responsible for some similarities between the GaSb/Mn digital alloys and GaMnSb random alloys, and the observed ferromagnetism of such a system at higher temperatures is associated with the 2D MnSb islands (precipitates), which may be related to the high Corresponding author. Tel.: +820222603489; fax: +820222784519.
E-mail address: [email protected] (C.S. Park). 0304-8853/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2006.06.031
TC in MnSb [5]. However, the origin of high TC is still ambiguous and a detailed theoretical study is needed to fully understand the mechanism. Up to now, manganese-doped InP has mainly discussed as a deep acceptor of impurity levels by some group [6,7]. However, InP is very useful material in the area of compound semiconductor and is known to have excellent physical properties and good characteristics of devices. In view of many DMS materials, InP is considered as a possible candidate according to a theoretical report [4]. We had reported the study of characteristics for unintentionally doped n-type (6.1–6.3 1015 cm3) InP implanted with various doses of Mn [8]. In the case of the Mn concentration of 5 1015 cm2(1%) and 5 1016 cm2 (10%), the ferromagnetic hysteresis loops at 10 K appeared very weakly because ferromagnetic coupling coexisted with antiferromagnetic coupling, and the ferromagnetic hysteresis loop did not appear at 300 K. In contrast, the temperature-dependent magnetization of samples grown by liquid phase epitaxy shows the two-step magnetization at each Curie temperature. Curie temperatures were obtained near 40 and 300 K, together with the hysteresis loops. In this paper, we report the results of the high Curie temperature of InMnP co-doped with Zn and clarify the mechanism of phase transition.
ARTICLE IN PRESS 184
C.S. Park, T.W. Kang / Journal of Magnetism and Magnetic Materials 309 (2007) 183–187
2. Experimental details InMnP epilayers were grown by liquid phase epitaxy (LPE). Mn and Zn-doped InP layers are grown on Zndoped (1 0 0) p-type InP substrate. The concentration of Zn is 3 1018 cm3 and that of Mn is 5 1016 cm2 (3%). Before the growth of InP epilayer, the substrate was cleaned by trichloroethylene (TCE), ethanol, methanol and rinsed in deionized water for 5 min. The solvents were 6 N In melt and InP bulk, and Mn powder was used for the growth of In1xMnxP epilayers. After prebaking the In and InP melt at 657 1C for 2 h, appropriate amount of Mn was added to the growth melt. The growth was performed at 637 1C with one-step cooling process. The composition was determined using the lattice constant and the band gap energy measured by X-ray diffractometer (XRD) and was compared with energy-dispersive X-ray spectroscopy (EDX). The measurements of photoluminescence (PL) and y/2y XRD were performed to characterize the optical and structural properties of InMnP, and superconducting quantum interference device (SQUID) was used to characterize the magnetic properties and the phase transition temperature of InMnP. The surfaces of films were investigated by high-resolution scanning electron microscopy (HRSEM). 3. Results and discussion The carrier concentrations of sample determined from Hall effect measurements at room temperature has a metallic behavior (1020 cm3) but is not the exact value because of the existence of the anomalous Hall effect. Fig. 1 shows the photoluminescence spectrum measured at 15 K for a In0.97Mn0.03P epilayer co-doped with Zn. The peak near 1.38 eV, which is usually referred to as the Al peak, frequently appears in InP grown by various methods and recently was attributed to carbon as the acceptor. The
Fig. 1. PL spectra measured for the (In, Mn)P film with x3% at 15 K and the inset shows the XRD patterns, (0 0 2) and (0 0 4) of (In, Mn)P. Upper spectrum means the film with surface islands and lower without.
peak near 1.38 eV is an unidentified donor-carbon acceptor pair transition. With regard to the above transition, the first phonon replica appeared near 1.33 eV. The transition related to Zn appeared near 1.29 eV. The intensity of band luminescence related to Mn from 0.9 to 1.20 eV is low and the energy positions of impurities shift to lower energy region due to the surface state well known in InP, and the emission related to Mn has a broad band. It has been reported that the intentionally doped InP:Mn has a very high Mn emission than band-related emission spectra [6]. In our result, it is considered that there is a similar mechanism of the extinction of free carriers through nonradiative recombination via deep levels for the reason of low intensity although Mn is intentionally doped [7]. The high density of carriers results in tunneling transport through thermal energy between two Mn acceptor levels. The tunneling probabilities increase as temperature increases. While the dominant process at low temperatures is hole trapping, that at relatively high temperatures is the transport of hole hopping through the tunneling process [8]. Furthermore, this variance of hole concentrations is consistent with the Hall resistance data at room temperature which shows a negative differential conductivity behaviors with two energy levels from two step saturation of resistance in Fig. 3a. InP with the Mn concentration of high impurity is known to have strong phonon replica, n ¼ 0, n ¼ 1 and n ¼ 2 ranging to 1.1–1.2 eV with the contribution of coupled-band luminescence related to Mn [9]. We confirmed that the transition is ascribed to a welllocalized transition metal level, that is, ferromagnetic type Mn centers that possibly form in InP [10]. This is given by a following mechanism, e+Mn3+(3d4)+hMn2+(3d5)+h, (bound hole binding energy ¼ 0.23 eV). Fig. 2a shows the surface morphology of distribution of islands in InMnP alloy through HRSEM and the images without islands (Fig. 2b) at 300 K. These images show the distribution of islands on the surface layer although the sample has been grown by liquid phase epitaxy. The diameter and the height of islands have been estimated approximately as 40 and 30 nm, respectively, from two images above. The island has been analyzed using EDX spectra to investigate the creation of InMnP islands on the homoepitaxial layer. The concentration of Mn of InMnP islands is 3 1015 cm2, 3%. y/2y XRD measurement have been performed in order to estimate the structural properties of InMnP layer, and x composition from the lattice constant to the band gap energy is approximately 3% same as the EDX results. 2y scan shows the only high intensity of InP (0 0 2) and (0 0 4) plane as seen in the inset of Fig. 1. The Hall resistance RHall of a magnetic thin film is conventionally observed to contain two distinct components. The first results from the normal Hall effect which is proportional to the applied magnetic field, H, and the second, called the anomalous Hall contribution, is proportional to the magnetization: RHall ¼ R0 H þ RS M,
(1)
ARTICLE IN PRESS C.S. Park, T.W. Kang / Journal of Magnetism and Magnetic Materials 309 (2007) 183–187
185
Fig. 3. (a) Anomalous Hall Effect measurements for the (In, Mn)P film with surface islands with Van der Pauw method at 10, 150 and 300 K. The solid lines are linear fits and (b) magnetoresistance measurements for the (In, Mn)P film with surface islands at low temperatures. Fig. 2. (a) HRSEM image of (In, Mn)P film with x3% having surface islands. (b) HRSEM images of (In, Mn)P film with x3%.
where M is the magnetization, and R0, Rs are the ordinary and anomalous Hall coefficients, respectively. In general, the anomalous Hall effect is a consequence of spin–orbit coupling in the system. Fig. 3 indicates that both InMnP and InMnP islands are fundamentally ferromagnetic from the anomalous Hall effects at 10, 150 and 300 K. The results of anomalous Hall effect at 150 and 300 K in Fig. 3a indicates that the magnetization is attributed to InMnP island. The magnetization of InMnP film saturates at 1 T with a half-slope of anomalous Hall effect at 10 K. By the way, the slopes of saturation from the fits imply that InMnP film have a stronger saturation response at high temperature. The magnitude of slopes is 2.4 at 150 K and 4.5 and 0.22 at 300 K, which demonstrate that the strength of the hysteresis more than low temperature as seen a good agreement with the magnitude of the temperature-dependent magnetization of Fig. 4. Fig. 3b shows the field
Fig. 4. Temperature dependence of magnetization for the (In, Mn)P films with x3%, obtained by SQUID with a magnetic field of 1000 Oe. Upper spectrum means the film with surface islands and lower without. The inset 1 shows the temperature-dependent magnetization of InMnP after thermal diffusion of Mn(3%) co-doped with Zn using MBE. The inset 2 shows the hysteresis curves for the (In, Mn)P film with x3% at 10 and 300 K.
ARTICLE IN PRESS 186
C.S. Park, T.W. Kang / Journal of Magnetism and Magnetic Materials 309 (2007) 183–187
dependence of the low-temperature magnetoresistance. The resistivity (r ¼ (r(B)r(0))/r(0)) of sample was measured with the magnetic field perpendicular to the sample. r(B) is a resistivity of sample in applied magnetic field and r(0) means the initial resistivity without magnetic fields. A negative magnetoresistance as a function of applied field was observed below 40 K. This is ascribed to spin scattering interactions between the localized magnetic moments of the Mn acceptors and the holes in an impurity band. The magnitude of the negative magnetoresistance decreases as the temperature increases. Hysteresis in the magnetoresistance was observed at 5 K, which shows the evidence that remnant magnetization within the sample continues to decrease the resistivity of the sample when the magnetic field is removed. This behavior was reported for InMnAs at 3.5 K by Ohno et al. [11]. Curie point at low-temperature range is found to be 40 K as determined from the maximum temperature at which the negative magnetoresistance is observable [12]. The reappearance of magnetoresistance at 50 K means that the ferromagnetic behavior having the anomalous Hall contribution is caused by InMnP islands to be confirmed in Fig. 4. Fig. 4 shows the results of zero-field cooled temperaturedependent magnetization using SQUID. The transition temperature appeared to be 300 K and another transition appears near 40 K, whereas InMnP without islands does not have hysteresis. Moreover, phase transition is not found in the temperature dependence of magnetization, which means that this film has no ferromagnetic property even due to MnP precipitates. It is reported that InMnP treated with various methods by the implantation of Mn and the thermal diffusion of Mn has the same tendency of temperature dependence. In addition, the study of p-type InMnP:Zn co-doped with Zn(2.1–2.2 1018 cm3) was investigated [13]. Mn was evaporated on the top of (1 0 0) InP:Zn co-doped with Zn using molecular beam epitaxy (MBE) system. After the evaporation of Mn, the thermal diffusion of Mn was carried out by heat treatment. The inset 1 of Fig. 4 shows the temperature-dependent magnetization of samples annealed at 500 1C for 60 s and 550 1C for 30 s after the evaporation of Mn. The concentration of Zn (2.1–2.2 1018 cm3) and Mn (3 1015 cm3, 3%) is similar to those of Zn (3 1018 cm3) and Mn(3 1015 cm3, 3%) for our present sample. Two Curie temperatures were obtained near 50 and 300 K. We showed additionally the results of temperature-dependent magnetization obtained by thermal diffusion of Mn in the inset 1 of Fig. 4 so as to clarify data. The inset 2 of Fig. 4 shows that the ferromagnetic hysteresis curve at 10 K has a weak loop(HC60 Oe) with slight saturation, and the hysteresis curve at 300 K has a stronger loop (HC100 Oe) and significant saturation. Compared with these results, the ferromagnetic behavior of InMnP at high temperature is much stronger than that of InMnP at low temperature. From these results of hysteresis curves and temperature dependence of magnetizations, it is found that the response of magnetization of InMnP sample occurs strongly at high-temperature. There
is an interesting phenomenon that the transition temperature is shown at high temperature region in the upper curve of Fig. 4 contrary to the lower case of sample without island. We believe that the main reason for the significant increase of Curie temperature originates to the distribution of islands or MnP [14]. However, it is very important that the film has the two-dimensional Ising chains, which is made of islands or Mn precipitates. Even though the codoping and the activation of Mn deep acceptor is considered, the increasing value of Curie temperature is at most expected to the extent of 30 K through interlayer coupling [15]. For this reason, we consider the twodimensional Ising model for the distribution of islands on the surface layer in order to generalize. We apply the theory of ferromagnetism on the two-dimensional Ising model for our result of temperature-dependent magnetization. Spontaneous magnetization was reduced by introducing an artificial limiting process [16], which may be charge of the formation of either a two-dimensional ferromagnetism or metamagnetism. Among most of semiconductors, it is caused by that sometimes phase transition happen, and sometimes phase transition does not happen irrespective of the ferromagnetic hystesresis. Lattice configuration of samples has two kinds of Ising lattices, which are made of the elementary lattice and the additional island sites on the surface or MnP distribution. We know that there are uniform islands formation in the Fig. 2a that can be considered as a giant Ising lattice with the topological chains. Two free energies are defined with a distinguished energy difference. We assume the free energy forming island is significantly larger than the general lattice. Spontaneous magnetization per spin of two-dimensional ferromagnetism is given by [16] 8 ðT4T C Þ; > <0 2 1=4 2 4 1=8 ð1 þ z Þ ð1 6z þ z Þ mI ð0; TÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffi ðToT C Þ; > :¼ 1 z2 (2) where the magnetic moment per spin has been taken to be unity and ze2be. The transition temperature TC correpffiffiffi sponds to the value zC ¼ 2 1. If e ¼ F ¼ FL+FD, then z ¼ e2bðF L þF D Þ ¼ e2bF L e2bF D . Here, FL and FD mean the free energy to form the lattice and island in thermodynamic system during the crystal growth. In this case, there are two situations with the change of temperature variable kBT. kB is the Boltzman’s constant. As seen in Fig. 4, Curie points appear in two transition temperatures in the application of this model. First, in case of kBT5eL, the first Curie point is a response of zC1 ¼ e2bF L as e2bF D goes to 1, corresponding to a low TC of 40 K, as seen in the Fig. 4. Likewise, in case of eLokBT5eL, zC2 ¼ e2bF D as e2bF L goes to 1. The second Curie temperature of 300 K in the Fig. 4 corresponds to a Curie point of MnP, island on the surface. Even in the MnP precipitates, two phases of metamagnetic in low temperature and ferromagnetism at high temperature works on this hypothesis.
ARTICLE IN PRESS C.S. Park, T.W. Kang / Journal of Magnetism and Magnetic Materials 309 (2007) 183–187
4. Summary and conclusion In conclusion, InMnP doped with Zn was grown by LPE and it has islands on the surface layer. The results of magnetoresistance and anomalous Hall effect demonstrates that InMnP film has properties of ferromagnetic semiconductor. Temperature dependence of magnetization shows that the phase transition occurs at high TC of 300 K. Two mechanisms of phase transition exist in the sample according to InMnP, and MnP or InMnP islands, respectively. We obtained the correlation between the free energy of the two-dimensional lattice formation of islands or precipitates, and the Curie point is determined on the competitive relation of two-dimensional lattices. Acknowledgement This work was supported by the Quantum-Functional Semiconductor Research Center (QSRC) through the Korean Science and Engineering Foundation and by a research fund of the Dongguk University, 2005. References [1] H. Munekata, H. Ohno, S. von Molnar, A. Segmuller, L.L. Chang, L. Esaki, Phys. Rev. Lett. 63 (1989) 1849; H. Ohno, H. Munekata, T. Penney, S. von Molnar, L.L. Chang, Phys. Rev. Lett. 68 (1992) 2664.
187
[2] J. De Boeck, R. Oesterholt, A. van Esch, H. Bender, C. Bruyseraede, C. Van Hoof, G. Borghs, Appl. Phys. Lett. 68 (1996) 2744. [3] H. Ohno, Science 281 (1998) 951; G.A. Prinz, Science 282 (1998) 1660; H. Shimizu, T. Hayashi, T. Nishinaga, M. Tanaka, Appl. Phys. Lett. 74 (1999) 398. [4] T. Dietl, H. Ohno, F. Matuskura, J. Cibert, D. Ferrand, Science 287 (2000) 1019. [5] X. Chen, M. Na, M. Cheon, S. Wang, H. Luo, B.D. McCombe, X. Liu, Y. Sasaki, T. Wojtowicz, J.K. Furdyna, S.J. Potashnik, P. Schiffer, Appl. Phys. Lett. 81 (2002) 511. [6] L. Eaves, A.W. Smith, M.S. Skolnick, B. Cockayne, J. Appl. Phys. 53 (1982) 4955. [7] Y. Ishitani, S. Minagawa, H. Hamada, T. Tanaka, J. Appl. Phys. 82 (1997) 1336. [8] A.A. Burkov, L. Balents, Phys. Rev. Lett. 91 (2003) 057202. [9] T. Takanohashi, T. Tanahashi, M. Sugawara, K. Kamite, K. Nakajima, J. Appl. Phys. 63 (1988) 1961. [10] Y. Shon, W.C. Lee, Y.S. Park, Y.H. Kwon, S.J. Lee, K.J. Chung, H.S. Kim, D.Y. Kim, D.J. Fu, T.W. Kang, X.J. Fan, Y.J. Park, H.T. Oh, Appl. Phys. Lett. 84 (2004) 2310. [11] H. Ohno, H. Munekata, S. von Molnar, L.L. Chang, J. Appl. Phys. 69 (1991) 6103. [12] R.P. Khoslar, J.R. Fischer, Phys. Lett. 32A (1970) 96. [13] Y. Shon, H.C. Jeon, Y.S. Park, W.C. Lee, S.J. Lee, D.Y. Kim, H.S. Kim, H.J. Kim, T.W. Kang, Y.J. Park, C.S. Yoon, K.S. Chung, Appl. Phys. Lett. 85 (2004) 1736. [14] E.E. Huber, D.H. Ridgley, Phys. Rev. 135 (1964) A1033. [15] T. Jungwirth, W.A. Atkinson, B.H. Lee, A.H. MacDonald, Phys. Rev. B 59 (1999) 9818. [16] C.N. Yang, Phys. Rev. 85 (1952) 808.