Synthesis, structure and magnetic properties of Ti doped La2MnNiO6 double perovskite

We report sol-gel synthesis, structural characterization and magnetic properties of La2Mn1–xTixNiO6 (0 ≤ x ≤ 1.0). Ti doping removed the biphasic structure of La2MnNiO6 by suppression of rhombohedral structure and all the Ti containing samples crystallized in monoclinic P21 / n symmetry. La2MnNiO6 exhibits multiple magnetic transitions. The high temperature ferromagnetic transition of La2MnNiO6 gradually shifted to lower temperatures with increase in Ti doping. La2TiNiO6 (x = 1.0) does not show any long-range magnetic ordering. The suppression of magnetic transition by Ti doping is ascribed to the destruction of Mn – O – Ni superexchange interaction. However, the signature of ferromagnetic phase persists up to 70 % Ti doping, indicating the robustness of magnetic ordering in La2MnNiO6. These results suggest that the addition of Ti 4+ truncates the ferromagnetic Mn – O – Ni superexchange path and it likely promotes ferromagnetic cluster formation. The robustness of ferromagnetic state towards Ti substitution compared to the simple perovskite or spinel structure can be attributed to cationic ordering in double perovskite structure. Both the pure and Ti-doped samples exhibit magnetic frustration at lower temperatures due to partial cationic disordering. The absence of long-range ordering in La2TiNiO6, unlike La2TiCoO6 or Pr2TiCoO6, could be related to cationic disordering.


Introduction
Oxides of transition metals with perovskite structure LnMO 3 (where Ln is rare earth elements or alkaline earth and M is transition element) exhibit various exotic physical properties related to the correlation between spin, charge, lattice and orbital degrees of freedom [1][2][3]. These strongly correlated electronic systems show competing electronic and magnetic states. Accordingly, compounds with perovskite structure have most extensively been studied by the physics and chemistry communities due to their magnetic, transport and magnetotransport properties [4]. Presence of second transition metal in the perovskite structure further improves the characteristic features of the so-called double perovskites, Ln 2 MM'O 6 . The important features of double perovskite are cationic ordering, antiphase boundaries and multiple exchange interactions [5][6][7]. The rock salt ordering with alternate occupancy of octahedra by different metal ions takes place when there is a large size mismatch or the charge difference is greater than 2 [6,8,9]. All these are governed by the local electronic configuration of the transition metal ions. Extensive investigations on double perovskites reveal that by virtue of wide M / M' cationic range they exhibit multifunctional properties like insulating, metallic, ferromagnetic, magnetodielectric, multiferroic, etc., which are suitable for various technological applications [10][11][12][13].
Among the double perovskites, La 2 MnNiO 6 has attracted special attention due to its striking properties such as multiple structures, multiple magnetic ground states, spin frustration and ferromagnetic insulating behaviour near room temperature with giant magnetodielectricity and magnetoresistance [14][15][16][17][18][19]. La 2 MnNiO 6 is reported to crystallize in biphasic nature with rhombohedral (R3c) and monoclinic (P2 1 / n) [14,15,[20][21][22] or rhombohedral and orthorhombic (Pbnm) symmetry [17,23]. The high temperature synthesized sample is rhombohedral and low temperature one is monoclinic or orthorhombic, whereas for intermediate temperature range it is biphasic [23]. The phase fraction in the sample is also sensitive to the synthesis condition and post annealing treatment [14,15,18]. It was reported that the hightemperature phase transforms to P2 1 / n phase at low temperature [14]. The magnetic ordering temperatures are different for different phases. The R3c phase shows ferromagnetic ordering at relatively higher temperature (~280 K) compared to the P2 1 / n or Pbnm phase (~150 K) [16,17]. The high temperature magnetic phase is governed by Mn 4+ -O -Ni 2+ , where for low temperature it is Mn 3+ -O -Ni 3+ superexchange interactions [15-17, 23, 24]. There is also report on ferromagnetic transition at T C~1 00 K in partially disordered sample, which was attributed to the Mn 3+ -O -Ni 3+ interaction [15]. Furthermore, there is also appearance of spin glass behaviour in La 2 MnNiO 6 associated with the competing interaction between ferromagnetic superexchange and disorder induced antiferromagnetic Mn 4+ -O -Mn 4+ and Ni 2+ -O -Ni 2+ interactions [15,18,25]. This suggests that the cationic disordering suppressed the high temperature ferromagnetic Mn 4+ -O -Ni 2+ interaction of perfectly ordered phase. Another extensively studied compound is La 2 MnCoO 6 with ferromagnetic T C~2 30 K [26]. The double perovskites in which both metal ions are magnetic exhibit ferromagnetic ordering. However, for the ordered perovskite with single magnetic ion the long-range magnetic ordering is antiferromagnetic and the interaction is supersuperexchange with separation of magnetic centres larger than 5 Å [27,28]. Such interaction also becomes weaker as reflected from the ordering temperatures. The observed T N values for La 2 CoTiO 6 and Pr 2 CoTiO 6 are respectively, 15 K and 17 K [28,29]. Thus, it would be interesting to investigate the ferromagnetic to antiferromagnetic crossover starting from an ordered ferromagnetic system by gradual replacing one magnetic ion by a nonmagnetic one.
In the present studies we have gradually replaced Mn 4+ in La 2 MnNiO 6 by Ti 4+ to better understand the role of Mn 4+ in determining the structural and magnetic properties. To the best of our knowledge there is no report in the literature on in-vestigation of Ti doping in La 2 MnNiO 6 . The ionic radius of Ti 4+ (0.605 Å) is larger than that for Mn 4+ (0.530 Å) [30]. In addition, there are other aspects of this substitution. The replacement of magnetic Mn 4+ by nonmagnetic Ti 4+ will make magnetic dilution in system, which in turn will truncate the ferromagnetic exchange path. Thus, substitution of Ti 4+ in La 2 MnNiO 6 is expected to rapidly destroy the ferromagnetic state. Furthermore, by consid-ering the retention of cationic ordering it will bring antiferromagnetic supersuperexchange Ni 2+ -O -Ti 4+ -O -Ni 2+ in place of ferromagnetic Ni 2+ -O -Mn 4+ -O -Ni 2+ exchange path. We observed that the ferromagnetic ground state of La 2 MnNiO 6 is very robust and it persists up to 70 % doping of Ti 4+ . We did not observe any antiferromagnetic ordering in La 2 TiNiO 6 could be due to cationic disordering.

Experimental
Polycrystalline La 2 Mn 1-x Ti x NiO 6 (0 ≤ x ≤ 1) samples were prepared by modified sol-gel technique. At first, stoichiometric amounts of metal nitrates (La, Ni) and acetate (Mn) were dissolved in 100 ml double distilled water followed by the addition of about 5 ml concentrated nitric acid to prevent the hydrolysis of the aquated metal ions. In another small beaker stoichiometric amount (1:1) of titanium isopropoxide [Ti{OCH(CH 3 ) 2 } 4 ], and acetyl acetone were mixed and stirred for about several minutes. These two different solutions were mixed in one single beaker and were stirred for about one hour using a magnetic stirrer to get a clear solution.
Citric acid was then added to the solution at four times mole ratio of the total metal ions. The final pH of the solution was found to be ~2. The resulting solution was stirred overnight followed by evaporation of solvent at 100 °C to obtain the gel. The obtained gel was dried by increasing the temperature of hotplate to 150 °C to transform the gel into crude precursor. The crude powders were ground thoroughly by using an agate mortar-pestle and calcined at 500 °C for 6 h in air. The final calcination was performed at 700 °C for 6 h in air to achieve the pure phase samples.
The powder X-ray diffraction (PXRD) patterns were registered with a Bruker D8 Advance X-ray diffractometer using Cu Kα radiation (λ = 1.5418 Å) operating at 40 kV and 40 mA. The PXRD patterns were recorded in the 2θ range of 10-120° using Lynxeye detector (1D mode) with a step size of 0.02° and a dwell time of 1s per step. Iodometric titration of the samples confirms the oxygen stoichiometry fixed to ''O 6 " within the limit of accuracy ±0.05.
The dc magnetization measurements were performed using a superconducting quantum interference device (SQUID) magnetometer with a variable temperature cryostat (Quantum Design, San Diego, USA). The magnetic ac susceptibility, χ ac (T) was measured with a PPMS (Quantum Design, San Diego, USA) with the frequency ranging from 10 Hz to 10 kHz. All the magnetic measurements were performed on powder samples putting inside a Teflon capsule.

Results and discussion
The PXRD patterns for all the La 2 Mn 1-x Ti x NiO 6 (x = 0, 0.2, 0.3, 0.5, 0.7 and 1.0) samples recorded in the 2q-range 10-120° are shown in Fig. 1. All the pat-terns are refined by Rietveld method using FullProf suit program [31]. The pattern of the parent compound (x = 0) La 2 MnNiO 6 can only be refined properly by considering a mixed rhombohedral R3c and monoclinic P2 1 / n phases. The biphasic nature composed of rhombohedral and monoclinic phases for La 2 MnNiO 6 sample was reported in the literature by several researchers [14,15,18,32]. However, the phase fraction depends on synthesis condition as well as on post-synthesis annealing treatment. We observed predominant monoclinic phase (80 %) over rhombohedral (20 %) one for the sol-gel synthesized La 2 MnNiO 6 with final heat treatment in air at 700 °C for 6 h. However, all the Ti doped samples can be nicely indexed with monoclinic structure (sp. gr. P2 1 / n) (Fig. 1). Thus, the Ti doping suppressed the rhombohedral phase of La 2 MnNiO 6 . The detailed structural and refinement parameters are given in Table 1. From the Table, one can see the systematic increase in cell volume with the increase in Ti 4+ content as expected for its larger ionic radius compared to that of Mn 4+ [30]. We observed that the size of NiO 6 octahedra are slightly larger than that of MnO 6 as reported in the literature [33]. Fig. 2 shows the temperature dependent dc-magnetization measured in both the zero field-cooled (ZFC) and fieldcooled (FC) protocols under an applied magnetic field of 500 Oe for the temperature range 5-300 K. The magnetization data of x = 0 parent phase exhibit ferromagnetic transition (T C~2 80 K) just below room temperature (Fig. 2, a). One should notice the large thermomagnetic irreversibility between ZFC and FC data branch immediate below T C~2 80 K , where ZFC data show a hump. The ZFC data also  Table 1 Structural refinement parameters for Ln 2 Mn 1-x Ti x NiO 6 (0 ≤ x ≤ 1) ceramics (4) 5.516 (2) 5.483 (9) 5.490 (8) 5.511 (8) 5.524 (5) 5.542 (6) b (Å) 5.493 (3) 5.516 (1) 5.506 (4) 5.513 (6) 5.527 (2) 5.538 (6) 5.553 (6) c (Å) 7.742 (4) 13.237 (7) 7.760 (4) 7.767 (4) 7.783 (7) 7.802 (5)

Continuation of table
show a second broad hump at low temperature centred around 25 K. Below T C , FC data show a plateau followed by a definite slope change below 100 K (Fig. 2, a). Such magnetic behaviour of the parent phase is in good agreement with the reported data which supports the prevalence three magnetic phases in the temperature window 5-300 K [15-17, 23, 24]. The high temperature T C is associated with the ferromagnetic Mn 4+ -O -Ni 2+ superexchange interaction of cation ordered state. The transition below 100 K ascribed to the Mn 3+ -O -Ni 3+ superexchange interaction and lowest temperature anomaly was attributed to the magnetic frustration arising out of partial cationic disordering [15,18,25] For the Ti doped sample with x = 0.2, the high temperature magnetic transition largely shifted (~60 K) to lower temperature at T C~2 20 K (Fig. 2, a). However, the shape of the curve remains similar to that of the parent phase, though the thermomagnetic hysteresis loop shrinks compared to the x = 0 sample (Fig. 2, a). However, the upturn in magnetization below 100 K is not much shifted for x = 0.2 sample compared to x = 0. This result indicates that the high temperature magnetic transition associated with Mn 4+ -O -Ni 2+ superexchange interaction is largely hampered compared to the Mn 3+ -O -Ni 3+ superexchange interaction. This could be due to preferential isovalent substitution effect. Thus, it can be suggested that for 20 % Ti doped sample the cationic ordering persists. For x = 0.5 the T C value comes down to 150 K and still one can observe the double humps in ZFC branch data, upturn in FC data below 50 K as well as the thermomagnetic irreversibility (Fig. 2, b). It is worth mentioning that the parent phase sample with orthorhombic Pbnm or monoclinic P2 1 / n structures exhibits ferromagnetic transition at T C~1 50 K ascribed to the Mn 3+ -O -Ni 3+ superexchange interaction [15][16][17]24]. However, we believe that the T C~1 50 K for x = 0.5 sample is not associated with Mn 3+ -O -Ni 3+ superexchange interaction like in parent phase, rather it is related to the weakening of Mn 4+ -O -Ni 2+ superexchange interaction due to substitution of nonmagnetic Ti 4+ in place of Mn 4+ . The second magnetic phase further shifted below 50 K as indicated by the upturn in FC data in x = 0.5 sample (Fig. 2, b). This indicates that still there is possible cationic ordering up to 50 % Ti doping in La 2 MnNiO 6 . On further increase in Ti doping to x = 0.7, there is only one broad hump in ZFC data around 20 K and FC data show up turn below 100 K (Fig. 2, b). This indicates that the high and Inset in (b) shows the data for La 2 TiNiO 6 low temperature transitions as well as low temperature magnetic frustration merged together. For complete substitution of Mn by Ti i.e. x = 1.0 sample does not show any anomaly in the magnetization data and ZFC-FC superimposed as shown in inset of Fig. 2, b. This behaviour is typical for a paramagnetic material. This suggests that La 2 TiNiO 6 does not show any kind of long-range ordering. The gradual change in magnetic properties with replacement of magnetic Mn 4+ by nonmagnetic Ti 4+ in La 2 MnNiO 6 is also supported from the isothermal magnetization data measured at 5 K under an applied field of ±5 T (Fig. 3). There is also a systematic decrease in magnetization as expected for magnetic dilution, but all the samples up to x = 0.7 exhibit clear hysteresis loop suggesting the prevalence of ferromagnetic component in the system. However, the opening of the loop is small as reported for the parent phase [15,24]. There is no hysteresis loop for x = 1.0 sample (see inset in Fig. 3) which practically shows a linear increase in magnetization with field as expected for a paramagnet (see inset in Fig. 3).
Let us discuss the observed change in magnetization in La 2 Mn 1-x Ti x NiO 6 . The substitution of Mn 4+ by nonmagnetic Ti 4+ will truncate the ferromagnetic  [34]. Thus, with increase in Ti doping the ferromagnetism in La 2 MnNiO 6 becomes more and more weaker as reflected by the change in T C shown in Fig. 4. This result is very contrasting with the effect of Ti doping in single perovskite manganite Sm 0.55 Sr 0.45 MnO 3 , where just 4 % of Ti doping leads for disappearing of the ferromagnetic state [35]. The robustness of ferromagnetism towards magnetic dilution in double perovskite La 2 MnNiO 6 may be related to the cationic ordering. Most likely the disruption of Ni 2+ -O -Mn 4+ -O -Ni 2+ superexchange path leads to fragmented ferromagnetic clusters which are responsible for observed magnetic behaviour of the Ti doped samples even for 70 % doping. This is even higher than the effect of Ti doping in spinel structure. Mn 1+x Fe 2−2x Ti x O 4 has been reported to exhibit ferrimagnetic ordering up to 50 % Ti doping [36,37]. However, one cannot rule out the cationic disordering with the increase in Ti doping. La 2 CoTiO 6 and Pr 2 CoTiO 6 have been reported to show long-range antiferromagnetism at 15 K and 17 K, respectively, in cationic ordered state. The antiferromagnetic exchange interaction path is Co 2+ -O -Ti 4+ -O -Co 2+ in cationic ordered samples [28,29]. However, we did not observe any such long range antiferromagnetic ordering in La 2 TiNiO 6 . The absence of such long range ordering can be related to the cationic disordering.
To look at the magnetic features above T C we have fitted the high temperature data with Curie-Weiss law. The calculated effective paramagnetic moments also found to decrease with the increase in Ti doping. The µ eff value for the parent phase is 6.30 µ B / f.u. which is slightly larger than the calculated value 5.97 µ B / f.u., which has been attributed to the possible formation of ferromagnetic cluster above T C [38,39]. The µ eff value decreases from 4.8 µ B / f.u. for x = 0.2 to 1.72 µ B / f.u. for x = 1.0 sample revealing the effect of magnetic dilution. Finally, to confirm the association of low temperature magnetic anomaly with spin glass behaviour as reported for the parent phase we have measured the ac-susceptibility of x = 0 and x = 0.5 samples at different driving frequencies in the low temperature regions. Fig. 5 shows the real and imaginary parts of the ac-susceptibility data for these two samples. The χ'(T) for x = 0 sample is too broad to uniquely identify the glass transition temperature T g  Fig. 5, a). However, the χ"(T) revealed two frequency dependent peak around 20 K and 50 K (Fig. 5, c), respectively, suggesting the magnetic frustration. The multiglass behaviour of La 2 MnNiO 6 has been reported in the literature [18]. This magnetic frustration is associated with the competing interaction between the antisite disorder induced antiferromagnetic Mn 4+ -O -Mn 4+ and Ni 2+ -O -Ni 2+ interactions and ferromagnetic clusters [18]. The x = 0.5 sample exhibit only one peak around 20 K, which is frequency dependent as revealed from both χ'(T) and χ"(T) data (Figs. 5, b, d).
This indicates the presence of magnetic frustration in pure and Ti doped samples.

Conclusions
In the present study, we have synthesized La 2 Mn 1-x Ti x NiO 6 (0 ≤ x ≤ 1.0) by modified citrate-based sol-gel method. Rietveld analysis of the PXRD patterns revealed that the parent phase (x = 0) is biphasic in nature composed of rhombohedral R3c and monoclinic P2 1 / n structures, whereas Ti doped samples crystalized in single phase monoclinic P2 1 / n structure with the suppression of rhombohedral phase. The cell volume of the Ti doped samples increased due to larger size of Ti 4+ compared to Mn 4+ ion. The magnetic measurements suggest the multiple magnetic transition in La 2 MnNiO 6 . The high temperature ferromagnetic transition with T C~2 80 K associated with the cationic ordered ferromagnetic superexchange interaction Ni 2+ -O -Mn 4+ becomes weaker by replacement of Mn 4+ by nonmagnetic Ti 4+ . There is a gradual shift in T C with increase in Ti doping eventually leading to a paramagnetic state in La 2 TiNiO 6 . The ferromagnetic state of double perovskite exhibits robustness towards Ti substitution compared to the simple perovskite manganite as well as spinels, which may be related to the cationic ordering. Unlike other Ti containing double perovskite, La 2 TiNiO 6 fails to show long-range antiferromagnetic ordering probably due to cationic disordering. Both the pure and Ti doped samples show magnetic frustration at lower temperatures.