Insights into the magnetic dead layer in La0.7Sr0.3MnO3 thin films from temperature, magnetic field and thickness dependence of their magnetization

Experimental investigations of the magnetic dead layer in 7.6 nm thick film of La0.7Sr0.3MnO3 (LSMO) are reported. The dc magnetization (M) measurements for a sample cooled to T = 5 K in applied field H = 0 reveal the presence of negative remanent magnetization (NRM) in the M vs. H (magnetic field) measurements as well as in the M vs. T measurements in H = 50 Oe and 100 Oe. The M vs. T data in ZFC (zero-field-cooled) and FC (field-cooled) protocols are used to determine the blocking temperature TB in different H. Isothermal hysteresis loops at different T are used to determine the temperature dependence of saturation magnetization (MS), remanence (MR) and coercivity HC. The MS vs. T data are fit to the Bloch law, MS (T) = M0 (1 – BT 3/2), showing a good fit for T < 100 K and yielding the nearest-neighbor exchange constant J/kB ≅ 18 K. The variations of TB vs. H and HC vs. T are well described by the model often used for randomly oriented magnetic nanoparticles with magnetic domain diameter ≈ 9 nm present ...

Mixed valence manganese oxides with the formula RE 1-x A x MnO 3 (RE = trivalent rare earth element; A = divalent alkaline earth element) have a wide range of electronic and magnetic phases depending on doping level and cation radii, making them useful in spintronic and magnetic memory devices. 1 Studies of thin films of La 1-x A x MnO 3 showed above room temperature Curie temperature T C ∼ 370 K is achieved by doping Sr at the A-site with an optimum ratio of x ∼ 0.3, 2 making La 0.7 Sr 0.3 MnO 3 thin films good candidates for resistive switching memory devices, ferroelectric/ferromagnetic systems and heterostructures. [3][4][5] Studies on thin films of La 0.7 Sr 0.3 MnO 3 also showed that their magnetic properties are affected by film thickness and an inactive magnetic layer (dead layer) is present in thin films. Huijben et al. performed magnetic and transport measurements on LSMO thin films fabricated by pulsed laser deposition (PLD) with different thicknesses and reported that thin films remain ferromagnetic down to 3 unit cells. 6 Monsen et al. studied the thickness dependence of magnetic properties of LSMO thin films and determined a magnetic dead layer thickness d = 1.6 nm although the nature of magnetic state in the dead layer was not explored. 7 It has been suggested that the "dead layer" likely contains oxygen vacancies which disrupt exchange coupling and hence destroy long range order (LRO). 8 Here we report detailed magnetic studies on a 7.6 nm thick film of La 0.7 Sr 0.3 MnO 3 on SrTiO 3 (100) prepared by pulsed laser deposition (PLD). Measurements of DC magnetization (M) were carried out in different applied magnetic fields (H) and temperatures (T ) under the ZFC (zero-fieldcooled) and FC (field-cooled) protocols to show that the blocking temperature (T B ) usually observed in magnetic nanoparticles (NPs) is also present in thin films. In addition, isothermal hysteresis loops were measured to determine the temperature dependence of saturation magnetization (M S ), coercivity (H C ) and remanence (M R ). M vs. T data show that in low H, negative remanent magnetization (NRM) exists below 100 K. Although the observation of NRM in LSMO film has been attributed to negative magnetic field trapped in superconducting magnet or uncompensated spins between ferromagneticferroelectric layers, 9  The LSMO thin film with D = 7.6 nm was deposited on the TiO 2 -terminated SrTiO 3 (100) substrate from the stoichiometric La 0.7 Sr 0.3 MnO 3 target by using a KrF excimer laser with the repetition rate of 5 Hz by PLD. During growth, the sample was heated in 750 • C and was exposed to 100 mTorr O 2 pressure and cooled down to room temperature at the rate of 15 o C/min in 250 mTorr O 2 pressure. The growth was monitored in-situ by reflection high energy electron diffraction (RHEED) to provide precise control of thickness down to unit cell scale. A physical property measurement system (PPMS) from Quantum Design with maximum H = 90 kOe was used to measure M vs. T from 5 K to 400 K with H applied in the plane of the film. The measured magnetic moment was scaled to the volume of the film (thickness * area). Details of the procedures for magnetic measurements are given as a note in Ref. 10.
The plots of M vs. T for the ZFC and FC cases measured in H = 50, 100, 200, 500 Oe and 1 kOe are shown in Fig. 1. For the ZFC case, the sample was cooled to 5 K in H = 0 Oe and measuring H was then applied and M vs. T data taken up to 400 K. For the FC case, the sample is cooled to 5 K in non-zero H and M vs. T data taken similarly up to 400 K in the cooling H. Blocking temperature T B defined here by the bifurcation of the M (FC) from the M (ZFC) data represents the temperature above which all spins are unblocked 11 and T B decreases with increasing H. As expected, T B is less than T C , the latter defined by the inflexion point in the M vs. T data. Note the negative magnetization for the ZFC cases below 100 K for H = 50 Oe and 100 Oe. For H = 200 Oe and higher H, only positive values of M (ZFC) are observed. Any negative trapped residual magnetic field was practically eliminated by demagnetizing the magnet coil as described in Ref. 10. Note that the NRM has been reported in LSMO thin films. 9,12 and in some nanoparticle systems also. 13,14 Temperature dependence of saturation magnetization (M S ), remanent magnetization (M R ) and coercivity (H C ) determined from the isothermal hysteresis loops for the ZFC sample are shown in Fig. 2. These data were corrected for the diamagnetic contribution of the substrate which was evident from the negative slope of the M vs. H plots at higher H (not shown here) where the ferromagnetic component of LSMO gets saturated. In the insets of Fig. 2, the hysteresis loop at 5 K and its lowfield zoom are also shown. M R is measured for H = 0 and M S at H = 4 kOe. In Fig. 2, the initial M starts at negative values with the virgin loop requiring slightly larger field to switch sign. For H = 50 Oe, M (ZFC) is negative for T < 100 K becoming positive for T > 100 K. In the data of H C vs. T in Fig. 2(c), H C has dropped below 50 Oe. This explains why for H < H C , the M (ZFC) is negative, leading to the result that H < H C leads to the observation of NRM for the ZFC sample.
For very thick films, M S near 0 K was 583 (emu/cm 3 ) with T C ≈ 350 K. 7 For ferromagnetism M S = N V µ where N V is the number of spins per unit volume each with magnetic moment µ. The decrease of M S with increasing T shown in Fig. 2(a) for the 7.6 nm film is due to the excitation of spin waves (magnons). This temperature dependence of M S is fit to the Bloch's T 3/2 law given by 15 ∆M Here Q = 1, 2, or 4 for a sc, bcc or fcc structures and ∆M S = M S (0) − M S (T ). Bloch's law is derived using the Heisenberg Hamiltonian: where J is the exchange constant, and S i is the spin of atom "i". For LSMO, there is only one magnetic ion per unit cell and so Q = 1. The lines in Fig. 2(a) are fits to Eq. (1) using S = 3/2 and exchange constant J/k B = 18 K and 20 K. Thus the temperature variation of M S can be explained by the excitation of magnons represented by the Bloch's T 3/2 law reasonably well. Another estimate of J is determined from T C using molecular field theory, yielding J/k B = 3T c/2ZS(S + 1). With Z = 6 as the number of exchange-coupled nearest neighbors each with spin S =3/2 and T C = 305 K for the 7.6 nm LSMO film gives J/k B = 20 K close to J/k B = 18 K determined above.
The temperature variation of M R and H C in Fig. 2 shows that both become zero near 240 K coinciding with measured T B ≈ 230 K in H = 50 Oe. In Fig. 3(a), the dependence of T B on H is seen to fit very well to T B (H) = T B (0)[1-(H/H 0 ) 2 ] observed in NPs. 16 It is proposed that the observation of a bifurcation between ZFC and FC is due to nanoclusters of spins in the dead layer of thickness  18,19 In Fig. 2(c), the fit of the data to this Eq. for lower T appears to be valid but the fit for higher T fails and extrapolated T B = 120 K < T B =230 K with H = 50 Oe. This difference is likely due to the fact that to measure H C , H = 4 kOe was used and T B is strongly dependent on H.
To understand the issue of the 'dead layer" and its role in the magnetic properties of thin films, variation of T C and M S with respect to film thickness D are modeled by using the available data from literature on thin films prepared in different O 2 deposition pressures which is shown to significantly affect the measured properties. Experiments show that T C and M S decrease with decrease in thickness of the films which is similar to the observations reported in magnetic NPs. 6