Effects of Substitution and Substrate Strain on the Structure and Properties of Orthorhombic Eu1−xYxMnO3 (0 ≤ x ≤ 0.5) Thin Films

Highlights Orthorhombic Eu1−xYxMnO3 (0.0 ≤ x ≤ 0.5) thin films were successfully synthesized by spin-coating. The film phase evidences an additional contribution to lattice strain with increasing Y-content. Appreciable modification of magnetic response from bulks to films due to structural distortions. Abstract The effects on the structure and magnetic properties of Eu1−xYxMnO3 (0.0 ≤ x ≤ 0.5) thin films due to lattice strain were investigated and compared with those obtained in equivalent composition ceramics. The films were deposited by spin-coating chemical solution onto Pt\TiO2\SiO2\Si (100) standard substrates. X-ray diffraction and Raman spectroscopy measurements revealed that all films crystallize in orthorhombic structure with space group Pnma, observing an added contraction of the unit cell with increasing Y-substitution ou Eu, corresponding to a broadening of the Mn-O1-Mn angle and a gradual decrease in magnetic order response.


Introduction
Transition metal oxides have many fascinating physical properties, such as high temperature, superconductivity, piezoelectricity, ferroelectricity, magnetism, and multiferroicity. Hence, these materials are still objects of intense scientific, engineering, and economic interest, considering the extensive technological applications already in use and further yet to be discovered. Rare-earth manganites RMnO 3 are among some of the most relevant functional oxide materials, mainly due to intrinsic strong correlated electron mechanisms, leading to exhibiting structural-orbital-magnetic-electric coupling effects. In particular, it can be used for barriers and interfaces in controlled spin transport devices [1][2][3][4][5][6]. According to the ionic radius of rare earth elements, the manganites can exhibit two main distinct structural symmetries. The perovskite-like orthorhombic phase, comprising a network of octahedral MnO 6 building blocks, surrounds relatively larger R 3+ cations from La to Dy. In particular, systems like GdMnO 3 , TbMnO 3 , and DyMnO 3 reveal magnetoelectric properties owed to competitive ferromagnetic (FM) and antiferromagnetic (AFM) exchanges. In such cases, the Mn-O1-Mn angle is considered the primary order parameter to understand the modulated spin-lattice coupling mediated ferroelectric (FE) ground states [7,8]. The hexagonal space group can be found in the RMnO 3 systems of smaller R 3+ cations, including Lu, Y, and Sc [9][10][11][12][13][14]. These compounds are based on MnO 5 layers and present only AFM order at low temperatures due to weak spin-orbit coupling (<90 K). Nonetheless, the FE order prevails up to substantial high temperatures. In addition to ionic displacements in these systems, other complex mechanisms related to exchange effects of electron orbital polarization contribute to the ferroelectric state [15][16][17][18]. In this particular case, the Eu 1−x Y x MnO 3 system can exhibit both the orthorhombic (x ≤ 0.5) and hexagonal phase (x > 0.5), allowing the possibility of tuning the magnetic and polar phases through manipulation of the A-site size, the partial substitution of larger Eu 3+ ions for smaller isovalent Y 3+ ions without increasing the R-site magnetic complexity. The variation of the Mn-O1-Mn bond angle in these systems is also associated with the development of complex magnetic states and ferroelectric phases, leading to the magnetoelectric properties exhibited by this system [19][20][21]. The Eu 1−x Y x MnO 3 compound shows the characteristics of a paramagnetic to an incommensurate antiferromagnetic phase (AFM-1) at Néel temperature (T N ), varying monotonously from 52−45 K (x = 0) to 45 K (x = 0.50) [7,22]. For x < 0.1, a transition from the AFM-1 to a canted A-type antiferromagnetic phase occurs at 43 K (x = 0) to 33 K (x = 0.1) [22]; next while it emerges for x < 0.25 [23]. For the compositions 0.25 < x < 0.35, a cycloidal modulated antiferromagnetic and ferroelectric phase (AFM-2) becomes apparent. For 0.35 < x < 0.55, a re-entrant magnetic phase, corresponding to two successive magnetic transitions, is associated with the cycloidal plane's rotation from the ab to the bc plane [7,22,23]. This rotation implies that the spontaneous electric polarization changes from the c-axis to the a-axis [23]. Other authors consider, for 0.15 < x < 0.35, a further low temperature modulated conical magnetic and ferroelectric phase (AFM-3) [7,[22][23][24]. However, the magnetic structure of low-temperature phases continues to be intensely studied. Besides the studies of Eu 1−x Y x MnO 3 as single crystals and conventional ceramics [22], it becomes relevant to investigate this system in the form of thin films since the properties may differ from those of bulk ceramics due to the influence of dimensional strain and the substrate interface. Even with their granular nature, thin films have attracted great interest and require a systematic comparison between the behavior of single crystals and ceramics [22,25,26]. In addition to contributing to extend the understanding and enhancing the underlying multiferroic mechanisms, thin film materials are crucial for developing potential applications and further fabrication of functional electro-optical devices. Chemical solution deposition methods like the sol-gel spin-coating technique provide higher composition control, lower processing temperatures, shorter fabrication time, and relatively low cost, compared to other chemical and physical vapor deposition thin-film-forming processes such as evaporation, pulsed laser, and sputtering [27].
As thin films can be subjected to strong elastic strains due to structural mismatch with respect to substrates, their intrinsic physical properties can be actively modified. These effects are under study on other strained perovskite thin films, which in bulk form exhibit orthorhombic structures: EuMnO 3 , GdMnO 3 , LaMnO 3 , DyMnO 3 , YMnO 3 , etc. [6,27]. From this perspective, the main objective of this work is to analyze and determine the effects of Y substitution of Eu and substrate strain on the structural and magnetic properties of orthorhombic Eu 1−x Y x MnO 3 thin films and, in particular, thin film oxides prepared by the sol-gel spin-coating method, taking into consideration some practical technical factors, including low production cost, equipment versatility, simplicity of preparation, composition control, and low sintering temperature. The experimental results encompassing crystal structure, dynamics, and the magnetic properties of these thin films are further compared with bulk ceramic samples of equivalent nominal composition (x = 0; 0.1; 0.2; 0.3; 0.4; 0.5).

Experimental Procedure
The studied Eu 1−x Y x MnO 3 thin films were prepared by a sol-gel method using standard Pt\TiO 2 \SiO 2 \Si (100) wafers as substrates. Details of precursor solutions and film preparation are available in refs. [27,28]. Thermogravimetric (TG) and differential thermogravimetric analyses (DTA) were conducted in the Eu 1−x Y x MnO 3 precursor solutions using a Seteram Labsys TG-DTA/DSC analyzer; each solution was previously dried at 100 • C for 48 h, forming the so-called dry solution. Afterward, it was heated in an air atmosphere at 10 • C/min from room temperature up to 1000 • C. The structural characterization of the films was carried out with a PANalytical MRD diffractometer, equipped with an X'Celerator detector in Bragg-Brentano geometry, using a Cu Kα radiation (λ = 1.5418 Å) source, a step of 0.017 • /100 s from 20 • up to 80 • (2θ). Unpolarized micro-Raman spectroscopy was performed at ambient conditions for the 200-800 cm −1 spectral range, using a 514.5 nm Argon laser line as the excitation source. The incident power was kept below 10 mW to avoid local sample heating. The scattered light was analyzed by a T64000 Jobin-Yvon triple spectrometer, operating in triple subtractive mode and equipped with a liquid-nitrogencooled charge-coupled device. The spectral slit width was ∼1.5 cm −1 . The parameters of the observed Raman modes (frequency, line width, and amplitude) were calculated using IgorPro software 6.0 from the best fit of a sum of damped oscillator functions [25]. Samples magnetization was measured using a Quantum Design SQUID magnetometer in reciprocating sample option (RSO) mode with a sensitivity of ∼10 −7 emu. Measurements were carried out in a heating run from 5 to 300 K following protocols of zero field-cooled (ZFC) and field-cooled (FC) at 100 Oe. The experimental and calculated data were processed according to the error analysis procedures, accounting the intrinsic precision of the methods and corrections derived from the formulas used [29].  The decomposition process includes three main steps: (i) solvent evaporation and (ii) calcination and pre-oxidation of precursor cations; these two preparation steps are similar to the ones reported for other oxide films [27,28]. The third step (iii) is the most important since it evolves the formation of the intended crystalline phase. The TG-DTA curves profiles, steps, and temperature intervals, obtained during the decomposition of each precursor solution exhibit slight variations depending on the ratio of substitution elements used. Overall, the formation of the Eu 1−x Y x MnO 3 perovskite crystalline phase is indicated by the precursor's loss weight rate stabilizing beyond 770-880 • C. As seen in Figure 1c, the temperature threshold for the phase formation increases gradually with Y substitution (x) in relation to the EuMnO 3 film. Consequently, the final sintering treatment was set at 800 • C for the x = 0 up to 850 • C for x = 0.5.

C.2. Surface Morphology
Representative SEM images of the Eu 1−x Y x MnO 3 films surface view and respective crosssection backscattering are shown in Figure 2a  As the content of Y 3+ substitution increases, the average grain (crystallite) size and average area enclosed by cracks (agglomerates) decreases, whereas the film thickness gradually increases. A summary of the results extracted from these images is displayed in Table 1.  Figure 3 presents the grazing angle X-ray diffraction (GA-XRD) patterns obtained at ambient conditions from the series of Eu 1−x Y x MnO 3 thin films, where the diffraction peaks resultant from the substrate are labeled with ( * ) symbols. Thin films with composition 0 ≤ x ≤ 0.5 exhibit a typical XRD diffractogram profile associated with the Pnma (62) symmetry group, as observed in analogous bulk samples [22,25,26]. Likewise, no relevant secondary phases were detected. On the other hand, films with x ≥ 0.6 revealed concomitance of hexagonal and perovskite phases and are not the object of the present study since it is not viable to evaluate a direct correlation to each phase and its composition. A control film of YMnO 3 (x = 1.0), synthesized by the same spin-coating sol-gel method and sintered at 850 • C, confirms the dominance of P6 3 cm symmetry [28,30,31]. Overall comparison of the graphs enables us to observe a shift of reflection peaks towards higher angles as x increases, suggesting a relative contraction of the unit cell volume. In the diffractogram of film EuMnO 3 (x = 0.0) and to a lesser degree in film Eu 0.9 Y 0.1 MnO 3 (x = 0.1), it is possible to discern that the intensity of the reflection peak at 2θ∼24 • , indexed to planes (020) of orthorhombic phase, are over-proportional to the spectra expected from a random polycrystalline film. This specific feature, not detected on the other film diffractograms, is suggestive that the films with x = 0.0 and 0.1 can have partially grown with a preferred orientation along the b-axis. For x > 0.1, the films XRD do not show signs of oriented growth and are assumed to be dominated by polycrystalline form.

C.3. Crystal Structure
Detailed (Le Bail) analysis of the diffractograms enables calculating the respective thin films lattice parameters. The fitting model used in the indexation to the Pnma space group and conventional pseudo-Voigt function to describe the peak shape [25]. Figure 4a plots the dependence of the pseudo-cubic ( pc ) lattice parameters defined as a pc = a/ √ 2, b pc = b/2, and c pc = c/ √ 2 with x-composition, where a, b, and c are the conventional orthorhombic Pnma lattice parameters.
The relation a pc > c pc > b pc is fulfilled for all compounds characteristic of the designated O' structure, typically found in rare-earth manganites. In this O' structure, the octahedral MnO 6 frames are associated with a strong Jahn-Teller distortion and orbital ordering [32]. While in bulk ceramics, a linear decrease of lattice volume is observed in direct correlation with the cube of the effective A-site radius [25,26], it is evident that the series of films' pseudo-cubic lattice parameters gradually decrease with increasing Y-substitution content in the range 0.0 to 0.5.  Table 2 resumes the values of quadratic |∆ 2 y/∆x 2 | and linear |∆y/∆x| slopes obtained from best fitted quadratic degree polynomial functions (y = |∆ 2 y/∆x 2 |·x 2 + |∆y/∆x|·x + y 0 ). Analogously to Eu 1−x Y x MnO 3 ceramics, the lattice compression, due to Y 3+ substitution, occurs in all axes, more manifested in cand band less in adirections. The quadratic terms |∆ 2 y/∆x 2 | are residual, and the pseudo-cubic lattice parameters follow a straight linear slope with the relation |∆a/∆x| < |∆b/∆x| ≤ |∆c/∆x| [33], which can be traced directly to the system tolerance factor (t = <(1 − x)·r Eu + x·r Y + r O >/ √ 2·<r Mn + r O >). Quite differently from the described linear behavior reported for the ceramic series, the contraction of lattice parameters in the series of Eu 1−x Y x MnO 3 thin films follow a non-linear correlation to t. The results indicate an additional compression mechanism at play, patently due to the strain in the film phase induced by the substrate effect. Figure 4a,b expose the film lattice strain by separating it from the corresponding linear behavior found on bulks. The film lattice contractions are more relevant and evenly distributed in the basal plane aand cdirections (< −0.5% to −1.7%) than in the b-direction (~−0.5%). This conjugation can explain why it is possible to advocate partial preferential growth of the film phase with x = 0 compatible with a less strained b-axis. On films, the relations become |∆a/∆x| ≈ |∆c/∆x| > |∆b/∆x|, respectively, ∼10, ∼3.5, and ∼2 times larger than reported for ceramics. Moreover, the quadratic terms have a similar magnitude to the linear ones and follow the same distribution |∆ 2 a/∆x 2 | ≈ |∆c 2 /∆x 2 | > |∆b 2 /∆x 2 |. Subsequently, the continuous reduction of the unit cell parameters and volume attests that the systems remain in the original Pnma phase, suggesting a supplementary adjustment of MnO 6 octahedra distortions and tilting [34]. The lattice volume contraction is near −1.7% for x = 0 and reaches a maximum of −3.7% for x ≥ 0.3. Figure 5 shows the unpolarized Raman spectra obtained from the series of Eu 1−x Y x MnO 3 thin films recorded at ambient conditions.

C.4. Raman Spectroscopy
Due to the main contribution of a polycrystalline phase in the series of samples, the experimental spectra simultaneously exhibited all Raman-active modes A g , B 1g , B 2g, and B 3g ; the assignment [13,35,36] assumes the band at ∼359 cm −1 is resulting from the bending or tilt (T) mode of the MnO 6 octahedra, which is activated by the (101) rotations (symmetry A g ); the band at ∼478 cm −1 is assigned to a Jahn-Teller type asymmetric stretching (AS) mode also involving the O(II) atoms (symmetry A g ); the band at ∼498 cm −1 is assigned to a bending mode (B) (symmetry B 2g ); and the band at ∼612 cm −1 is associated to a symmetric stretching mode (SS) involving the O(II) atoms (symmetry B 2g ). These modes can be detected in the range from 350 to 650 cm −1 and are also signaled in ceramic samples [25,35,36], corroborating the Pnma symmetry of the films already ascertained from the XRD data.
In the Raman spectrum of the Eu 1−x Y x MnO 3, thin films with x = 0 and 0.1 are observed as additional modes at ∼220 and ∼685 cm −1 ; these can be interpreted as a Raman activation of active infrared modes-a direct consequence of the lattice distortions occurring in thin films due to the lattice mismatch between film and substrate and the film-substrate interaction [32], consistent with the partial presence of an orientated layer as mentioned from the XRD analysis. In samples with 0.2 ≤ x ≤ 0.5, the contribution of an orientated layer is not so significant. Hence, the mode at ∼220 cm −1 vanished, while the broad shoulder band above ∼620 cm −1 can still be attributed to distortions in the film lattice induced by the substrate effect. Still, due to the dominant polycrystalline structure, these distortions are disordered and relaxing through the film thickness, contributing to the convolution of peaks deriving from a distribution of modes. Figure 5 shows the dependence in x of the peak center position wave number (ω) of the four most intense Raman modes assigned T, AS, B, and SS, measured in the films and compared to the ceramics with matching composition [25]. These modes are particularly interesting because they indicate the structural deformations in the Eu 1−x Y x MnO 3 thin films. For the thin films, the wave number of the tilt mode (T) increases linearly with Y 3+ content with a slope ∆ω T /∆x films = 50 ± 2 cm −1 , a similar linear behavior ∆ω T /∆x bulks = 29 ± 2 cm −1 is found for the ceramic's samples [25]. Iliev et al. [35] demonstrated that such out-of-phase rotation mode and the tilt angle (θ T ) exhibit a linear behavior with slope ∆ω T /∆x = 23.5 cm −1 . By combining the above results, a variation of the rotation mode and the tilt angle with x can be obtained, for the series of Eu 1−x Y x MnO 3 thin films, of ∆θ T /∆x = 2.1 • . This value indicates that the volume reduction, anticipated from X-ray diffraction analysis, corresponds to an increase in the tilt angle with up to x = 0.5, close to 1 • . While in Y-doped ceramics, the variation was ∆θ/∆x = 1.3 • , representing an increase in the tilt angle of only 0.05 • [25]. The symmetric stretching mode SS follows a near analogous constant (∆ω SS ∼2 cm −1 ) behavior for films and ceramics, suggesting minor variations of Mn-O 2 bond lengths with increasing x. However, in the same concentration range, the wave numbers measured for the thin films of the Jahn-Teller type asymmetric stretching mode (AS) and bending mode (B) exhibit a step-like discrepancy relative to the near linear ∆ω/∆x behavior identified in ceramics. In effect, specifically between x = 0.2 and 0.3, where the four mode defined curves intersect, the pseudo-perovskite lattice c pc (films) and b pc (bulks) lengths coincide, as can be seen in Figure 4a. Mode assignment: T-tilt mode of the MnO 6 octahedra (symmetry A g ); SS-symmetric stretching mode (symmetry B 2g ); B-bending mode (symmetry B 2g ), and AS-Jahn-Teller type asymmetric stretching mode (symmetry A g ). Dashed red lines are eyes-guides for the indicated modes identification. Insets: detailed graphs of the T; AS; B; and SS modes dependence with x in the Eu 1−x Y x MnO 3 thin films and ceramics [24].

C.5. Magnetic Properties
The structural distortions associated with the film phase and Y 3+ substitution are expected to significantly influence the magnetic properties of the studied samples, particularly the competitive ferromagnetic (FM) and antiferromagnetic (AFM) exchange interactions. Figure 6 presents the ZFC/FC magnetization measurements performed on the Eu 1−x Y x MnO 3 thin films from x = 0.0 to 0.4.
In overview, the experimental data follow a similar behavior to the ones reported in single crystal and ceramic samples of the equivalent compounds [22,25]. The difference between FC and ZFC magnetization ∆M (5 K) = M FC (5 K)-M ZFC (5 K) is 27.9 emu/cm 3 for x = 0.0, but ensuing films 0.1 ≤ x ≤ 0.4 ∆M (5 K) yield, respectively, only 0.67; 0.36, 0.07; and 0.08 emu/cm 3 . In EuMnO 3 and Eu 0.9 Y 0.1 MnO 3 films, an expected irreversibility behavior is observed for temperatures up to T N ∼50 K. Below this temperature, the characteristic weak FM arising from the canted A-type AFM was exhibited similar to EuMnO 3 single crystals and ceramics samples [22,37]. However, in films with Y 3+ concentrations ranging from 0.2 to 0.4, the irreversibility behavior persists at higher temperatures, making the precise determination of T N difficult. The relative difference between FC and ZFC magnetization curves decreases with increasing Y 3+ concentration, and the concavity of the FC curves inflect downward for x ≥ 0.3, as seen in Figure 6. It is reasonable to interpret the reduction of the induced magnetic moments in Eu 1−x Y x MnO 3 thin films with increasing x as a consequence of the volume contraction of the unit cell and the increase in the Mn-O1-Mn bond angle. While in films with x = 0.0 and 0.1, the weak FM stems from the canted A-type AFM, the incommensurate AFM for Y 3+ concentration in the range from 0.2 to 0.4 may not be assessed as the cause of the irreversibility at temperatures above T N . In this case, the film-substrate interface may emulate an apparently emergent FM response caused by pinning local magnetic moments induced by the excitation magnetic field [38]. This weak FM character, prolonging above 100 K, hinders the application of a Curie-Weiss model [6]. The previous results agree with those reported in Lu-doped EuMnO 3 thin films [17,39]; however, irreversibility behavior seems more marked in these samples.

Conclusions
Thin films of Eu 1−x Y x MnO 3 (0 ≤ x ≤ 0.5) were successfully prepared from the precursor solutions and deposited onto Pt\TiO 2 \SiO 2 \Si (100) substrates. The X-ray diffraction analysis reveals that the films stabilize in a single Pnma orthorhombic phase up to x = 0.5, while further substitution results in phases concomitance or full crystallization for x = 1.0 on P6 3 cm hexagonal structure. Compared to equivalent orthorhombic Eu 1−x Y x MnO 3 bulk ceramics, the thin films revealed an enhanced decrease in unit cell volume with increasing substitution concentration, and a further increase in the inclination of the MnO 6 octahedra was ascertained by Raman spectroscopy. In contrast to ceramics samples, the films significantly smeared out the paramagnetic-canted A-type AFM transition. A weak ferromagnetic character was detected even above 100 K and attributed to the pinning of local magnetic moments at the film/substrate interface. These results corroborate the pertinence of the substrate-induced strain to modify and tune the distortions of the compound crystal lattice and ferroic properties.

Data Availability Statement:
The data used to support the findings of this study are available from the corresponding author upon request.