ARTICLE IN PRESS Journal of Magnetism and Magnetic Materials 322 (2010) 1311–1314
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm
Magnetotransport properties of ferromagnetic LaMnO3+d nano-sized crystals V. Markovich a,, G. Jung a, I. Fita b,c, D. Mogilyansky d, X. Wu a,e, A. Wisniewski b, R. Puzniak b, L. Titelman f, L. Vradman f, M. Herskowitz f, G. Gorodetsky a a
Department of Physics, Ben-Gurion University of the Negev, 84105 Beer-Sheva, Israel Institute of Physics, Polish Academy of Sciences, 02-668 Warsaw, Poland c Donetsk Institute for Physics and Technology, National Academy of Sciences, 83114 Donetsk, Ukraine d Materials Engineering Department, Ben-Gurion University of the Negev, 84105 Beer-Sheva, Israel e Department of Materials Engineering, Monash University, Clayton 3800, Australia f Blechner Center for Industrial Catalysis & Process Development, Department of Chemical Engineering, Ben-Gurion University of the Negev, 84105 Beer-Sheva, Israel b
a r t i c l e in fo
abstract
Available online 16 April 2009
Transport and magnetic properties of LaMnO3+d nanoparticles with average size of 18 nm have been investigated. The ensemble of nanoparticles exhibits a paramagnetic to ferromagnetic (FM) transition at TC246 K, while the spontaneous magnetization disappears at TE270 K. It was found that the blocking temperature lies slightly below TC. The temperature dependence of the resistivity shows a metal–insulator transition at TE192 K and low-temperature upturn at To50 K. The transport at low temperatures is controlled by the charging energy and spin-dependent tunnelling through grain boundaries. The low temperature I–V characteristics are well described by indirect tunnelling model while at higher temperatures both direct and resonant tunnelling dominates. & 2009 Elsevier B.V. All rights reserved.
Keywords: Manganite Nanoparticle Resistivity Tunnelling Magnetization Curie temperature
1. Introduction Doped perovskite manganites (R1xAx)MnO3 (R is a rare-earth ion and A is a divalent ion such as Ca, Sr, Ba, etc.) exhibiting colossal magnetoresistance (CMR) effect have been studied intensively in the last decade [1,2]. An important feature of CMR manganites is the possibility of inducing ferromagnetism by cationic substitutions at different sites in the antiferromagnetic (AFM) parent compound LaMnO3. The parent compound, characterized by a Ne´el temperature TNE140 K, may exhibit various magnetic ground states such as AFM, canted AFM, spin glass, ferromagnetic (FM) insulating, or even FM metal with high Curie temperature TC, depending on the level of nonstoichiometry on the lanthanum and manganese sites [3]. It is expected that magnetic and transport properties of manganite particles change significantly with reduction of the particle size. When the size of magnetic nanoparticles is reduced to the nanometric scale, some of their basic magnetic properties, like spontaneous magnetization, Curie temperature, coercivity, etc., may differ significantly from the bulk values. The size effects become more pronounced with decreasing size of the particles. Recent studies of nano-sized crystalline samples of nonstoichiometric LaMnO3+d (LMO) revealed the presence of a FM state with TC4200 K [4,5]. Magneto-transport properties of hole-doped
Corresponding author. Tel.: +972 8 647 7127; fax: +972 8 647 2903.
E-mail address:
[email protected] (V. Markovich). 0304-8853/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2009.04.012
nano-sized FM metallic manganites, in particular the temperature behavior of the magnetoresistance and its magnetic field dependence as well as the nature of low-temperature resistivity minima, were intensively investigated experimentally [6]. In this work, we discuss magnetic and transport properties of ensembles of LMO nanoparticles with an average grain size of 18 nm.
2. Experimental The nanocrystalline LaMnO3 particles were prepared by the citrate method described in detail elsewhere [5,7]. The crystalline properties and the distribution of grain sizes were determined using high-resolution transmission electron microscopy and X-ray diffraction techniques. For magnetic measurements the nanocrystalline powders were pressed into cylinder-shaped samples by applying a pressure of 2 kbar at room temperature. Magnetization (M) was measured using PAR (Model 4500) vibrating sample magnetometer in magnetic fields up to 16 kOe at temperatures ranging from 10 to 300 K. For electrical transport measurements, the nano-powders were pressed into pellet-shaped samples by applying pressure of 5 kbar at room temperature. To weld the grains in cold-pressed pellets the samples were sintered for 30 min at the calcination temperature of 700 1C. Sintering was performed at the temperature not exceeding the calcination temperature in order not to increase the grain size. Electrical transport properties of sintered
ARTICLE IN PRESS 1312
V. Markovich et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 1311–1314
pellets were measured in a standard four-point arrangement with separation between the voltage contacts of 0.4–0.5 mm. During measurements of temperature dependence of the resistivity, bias current was provided by a pulsed current source to avoid excessive Joule heating. Dynamic resistance Rd ¼ dV/dI was measured using a lock-in detector with 5 mA ac current modulation at 390 Hz.
3. Results and discussion
M (emu/g)
9
M 0 (emu/g)
The average crystalline size /DS was calculated using Debye–Scherrer equation /DS ¼ 0.9l/b cos y, where b is the fullwidth at half-maximum of the X-ray diffraction peak, y is the corresponding Bragg angle and l of (0 2 4) reflection is 1.54059 A˚ for CuKa1 line. The average grain size was determined to be 1871 nm in both powder and pellet samples. The temperature dependence of the magnetization at 100 Oe, recorded in zero-field cooled (ZFC) and field cooled (FC) conditions, MZFC and MFC, is shown in Fig. 1(a). The Curie temperature TCE246 K of an ensemble of LMO nanoparticles was determined as the temperature of a minimum in the dMFC(T)/dT curve. The maximum in MZFC(T) dependence usually appears at the blocking temperature TB. The ZFC curve in Fig. 1 peaks at TmaxE195 K. The temperature dependence of MFC exhibits local weak maximum at TE60 K. FC and ZFC magnetization become significantly different at temperatures below some 220 K. The field dependence of magnetization shown in Fig. 1(b) was measured by cooling samples in the zero-field conditions and applying magnetic field at the required temperature. The nonlinearity in M(H) curves, which indicates the onset of the ferromagnetism in the system, is clearly seen already at temperatures below 270 K, while hysteresis loops (not shown in the graph), with the coercivity and small remnant magnetization appear only below 200 K. The coercive field increases with decreasing temperature and reaches HCE270 Oe at T ¼ 10 K. Spontaneous magnetization M0, determined from the linear
40
FC
20
6
Tmax
3
100 200 T (K)
TC
ZFC
0
0 0
0
100
200
300
10 K 50 K 150 K 200 K 230 K 250 K 270 K ZFC 280 K
40 20 0 0
4
8
12
H (kOe)
16
MZFC/M0 , MFC /M0
M (emu/g)
Temperature (K)
0.3
M /M FC
0
0.2 M
/M
ZFC
0.1 0
TB = 242 K
0
50 100 150 200 250
T (K)
Fig. 1. (a) Temperature dependence of ZFC and FC magnetization of LMO sample, recorded in H ¼ 100 Oe. Inset: temperature dependence of the spontaneous magnetization M0. (b) Field dependence of the magnetization for a LMO sample at various temperatures after ZFC and (c) temperature dependences of normalized MZFC(T)/M0(T) and MFC(T)/M0(T) magnetization.
extrapolation of the M(H) dependence at high fields to zero field, approaches M0E45 emu/g at T ¼ 10 K. The temperature evolution of M0 is shown in the inset to Fig. 1(a). The relative volume of FM phase calculated using thus determined M0 values is close to 50% at T ¼ 10 K. In the frame of the core-shell model of manganite nanoparticles [5,6], the inner part of a particle, the core, has the same properties as the bulk material, while the outer shell contains majority of oxygen faults and vacancies and acts as a magnetically dead layer. Assuming that the spontaneous magnetization of a nonmagnetic shell is zero, we evaluate the diameter of the FM core and the thickness of the nonmagnetic shell to be /DSE14 nm and tE2 nm, respectively. It has been recently demonstrated that to establish the correct value of the blocking temperature in an ensemble of nanoparticles, one should take into account the temperature dependence of the spontaneous magnetization [8]. The results of such analysis are shown in Fig. 1(c) illustrating the temperature dependence of MZFC(T) normalized to M0(T). The normalized magnetization exhibits a maximum at the blocking temperature TBE242 K, much higher than the temperature Tmax ¼ 195 K at which a maximum in the temperature dependence of ZFC magnetization in Fig. 1(a) is seen. It appears that the real blocking temperature in the LMO nanoparticle system at H ¼ 100 Oe is in fact much closer to the Curie temperature, determined from dMFC(T)/dT, than it would follow from a simplified magnetization analysis. This result agrees well with recent experiments concerning 12-nm-size La0.7Sr0.3MnO3 particle ensembles [8]. Fig. 1(c) shows that the normalized MZFC(T)/M0(T) and MFC(T)/M0(T) fully coincide at high temperatures and start to diverge at TETB. The magnetic transition observed in Fig. 1(a) is very broad. This is due to the distribution of particle sizes and shapes in the assembly, as well as to inter- and intra-particle magnetic interactions. Additional broadening of the transition as well as a significant decrease of the normalized MZFC(T)/M0(T) and MFC(T)/ M0(T) in Fig. 1(c) with decreasing temperature may be caused by dispersion and temperature dependence of anisotropy fields [8]. Therefore, the Curie temperature TCE246 K should be seen as an averaged value of dispersed TC’s of individual nanoparticles in the investigated LMO sample. Fig. 2(a) shows the temperature dependence of the resistance R(T), recorded under different current flow. One can see a clear metal-to-insulator (M–I) transition around TPE192 K, a quasimetallic (dR/dT40) behavior in a wide temperature range 55 KoTo190 K, and a sharp resistivity upturn below TE50 K. The increase of the low-temperature resistivity is commonly considered as a hallmark of the presence of insulating tunnel barriers in the system. The low-temperature resistivity increase may however also result from a breakdown of double exchange interactions caused by broken Mn–O–Mn bonds in the disordered interfacial region at nanoparticles surfaces. Ferromagnetic double exchange interactions in nanogranular manganites at low temperatures can also be disrupted by the Coulomb blockade [6]. Resistivity of granular metals in the Coulomb blockade regime follows the relation r ¼ A exp (D/T)1/2, where D is proportional to the charging energy EC [6,9]. By fitting experimental R(T) dependence in the range 10–40 K to the above expression we have found that D ¼ 14.2 K, consistently with the literature data for other nanogranular manganite systems [10]. The most striking feature of Fig. 2 is a significant decrease of the resistance with increasing current in the entire investigated temperature range. Similar effects have been already reported for other CMR systems [11,12]. The decrease of resistivity with increasing current at temperature range between the M–I transition and room temperatures is likely related to anomalous thermal effects associated with small heating and cooling steps caused by turning on and off the pulsed dc current source, as
ARTICLE IN PRESS V. Markovich et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 1311–1314
1313
-3
10
T=10 K
T=120 K
I =0.001 mA
T=192 K
300
-4
10
Current (A)
Resistance (Ω)
400
200
I =1 mA 100
7/3
I = (G0+G1)V+G2V +G3V
-5
10
7/2
G3=0
0 0
50
100
150
200
250
Temperature (K)
-6
10
0.00
295 K
1.0
0.10
0.15
0.20
Voltage (V) 120 K
192 K
G0+G1, G2
50 K dV/dI/dV/dI (0)
0.05
10 K 0.9
G0+G1
0.008
G2
0.004
0.000 -0.3
0.0
0
0.3
50
Current (mA)
100
150
200
250
Temperature (K)
Fig. 2. (a) Temperature dependence of the resistance measured under different dc current flow I ¼ 0.001mA, 0.01mA, 0.1 and 1 mA. Arrow shows the direction of the current increase and (b) normalized Rd as a function of bias current at various temperatures.
Fig. 3. (a) Experimental I–V characteristics (open symbols) at various temperatures fitted (solid lines) to the GM tunnelling theory I ¼ (G0+G1) V+G2V7/3 and (b) temperature variation of parameters G0+G1 and G2.
discussed recently in Ref. [13]. For more detailed inspection of the current influence upon the resistivity, we have measured the dynamic resistance Rd ¼ dV/dI using a lock-in ac technique and plotted the results of a function of the dc current at various temperatures in Fig. 2(b). Since we have determined that sample temperature starts to increase for dc bias currents above 1 mA, we have limited dc current range in this experiment to Imax ¼ 0.4 mA to prevent Joule heating from influencing the data. Fig. 2(c) shows that the nonlinearity in transport characteristics appears already in the vicinity of M–I transition and starts to increase significantly at low temperatures. We have analyzed voltage–current (I–V) characteristics using the Glazman–Matveev (GM) tunneling model [14]. In the GM model, the nonlinearity of I–V characteristics is due to processes of multi-step indirect tunneling via n localized states. For eVbkBT one has
which contains only the linear and the lowest nonlinear term from Eq. (1),
I ¼ ðG0 þ G1 ÞV þ
1 X
Gn V nþ12=ðnþ1Þ ,
I ¼ ðG0 þ G1 ÞV þ G2 V 7=3 .
(2)
Results of the fitting of 10, 120 and 192 K experimental I–V curves to Eq. (2) are shown in Fig. 3(a) with solid lines. The temperature variation of the weight of linear (G0+G1) and nonlinear G2 terms is shown in Fig. 3(b). The term (G0+G1) has a minimum in the vicinity of the metal–insulator transition, while G2 increases monotonically with decreasing temperature. Nevertheless, one can see that the nonlinear term becomes really significant only at low temperatures below some 100 K [15]. The fact that the temperature dependence of the conductivity can be described by the lowest-order GM nonlinear term indicates that in the considered temperature range the indirect tunneling through two impurity states in the barrier dominates transport properties of the system.
(1)
n¼2
where coefficients Gn are exponentially decreasing functions of the barrier thickness. Coefficient G0 accounts for the direct tunneling, G1 for the resonant tunneling via one impurity, while the nonlinear terms describe inelastic multi-step tunneling via n localized states. We found that within entire experimental temperatures our I–V curves can be well fitted to the GM formula using Eq. (2)
4. Conclusions In conclusion, we have found that the ensemble of LaMnO3+d nanocrystals with 18 nm average grain size exhibits broad transition to the FM state with average Curie temperature, determined by the temperature of a minimum in the dMFC(T)/dT curve, TCE246 K. In the same time the experiments unambiguously demonstrate that the spontaneous magnetization in the
ARTICLE IN PRESS 1314
V. Markovich et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 1311–1314
system appears already at To270 K. At low temperatures each nanoparticle consists of a FM core and a paramagnetic insulating shell. FM phase occupies about 50% of the nanoparticles volume at temperatures around 10 K. The blocking temperature determined from normalized MZFC(T)/M0(T) dependence is only slightly lower than TC and is much higher than TB which would follow from the position of a maximum in MZFC(T). The resistivity of LaMnO3+d nanoparticles exhibits metal–insulator transition at T190 K, followed by a minimum at T50 K and a sharp upturn at low temperatures. The low-temperature increase is attributed to tunnel transport mechanism and additionally to an action of Coulomb blockade. Electrical transport mechanism at low temperatures is dominated by inelastic tunneling of charge carriers via two localized states in the insulating barriers separating adjacent FM cores. At higher temperatures direct and/or resonant tunneling prevails.
Acknowledgments This work was supported in part by the Israeli Science Foundation, Grant 391/07 and by the Polish State Committee for Scientific Research under a research project No. 1 P03B 123 30.
References [1] E. Dagotto, Nanoscale Phase Separation and Colossal Magnetoresistance, Springer Series in Solid State Physics, Springer-Verlag, Berlin, 2003. [2] M. Ziese, Rep. Prog. Phys. 65 (2002) 143. [3] J. Topfer, J.B. Goodenough, J. Solid State Chem. 130 (1997) 117. [4] G. Dezanneau, M. Audier, H. Vincent, C. Meneghini, E. Djurado, Phys. Rev. B 69 (2004) 014412. [5] V. Markovich, I. Fita, D. Mogilyansky, A. Wisniewski, R. Puzniak, L. Titelman, L. Vradman, M. Herskowitz, G. Gorodetsky, J. Phys.: Condens. Matter 19 (2007) 346210. [6] L1. Balcells, J. Fontcuberta, B. Martinez, X. Obradors, Phys. Rev. B 58 (1998) R14697. [7] M.S.C. Baythoun, F.R.J. Sale, Mater. Sci. 17 (1982) 2757. [8] V.M. Kalita, A.F. Lozenko, S.M. Ryabchenko, A.A. Timopheev, P.A. Trotsenko, I.A. Danilenko, T.E. Konstantinova, J. Low Temp. Phys. 34 (2008) 436. [9] P. Sheng, B. Abeles, Y. Arie, Phys. Rev. Lett. 31 (1973) 44. [10] B. Roy, A. Poddar, S. Das, J. Appl. Phys. 100 (2006) 104318. [11] J. Gao, S.Q. Shen, T.K. Li, J.R. Sun, Appl. Phys. Lett. 82 (2003) 4732. [12] V. Markovich, E. Rozenberg, Y. Yuzhelevski, G. Jung, G. Gorodetsky, D.A. Shulyatev, Y.M. Mukovskii, Appl. Phys. Lett. 78 (2001) 3499. [13] J. Sacanell, A.G. Leyva, P. Levy, J. Appl. Phys. 98 (2005) 113708. [14] L.I. Glazman, K.A. Matveev, JETP 67 (1988) 1276. [15] V. Markovich, G. Jung, I. Fita, D. Mogilyansky, X. Wu, A. Wisniewski, R. Puzniak, N. Froumin, L. Titelman, L. Vradman, M. Herskowitz, G. Gorodetsky, J. Phys. D: Appl. Phys. 41 (2008) 185001.
