DSpace at VNU: Tuning Magnetic Properties of BiFeO3 Thin Films by Controlling Rare-Earth Doping: Experimental and First-Principles Studies

DSpace at VNU: Tuning Magnetic Properties of BiFeO3 Thin Films by Controlling Rare-Earth Doping: Experimental and First-...

Rare-Earth Doping: Experimental and First-Principles Studies

Hoa Hong Nguyen, Ngo Thu Huong, Tae-Young Kim, Souraya Goumri-Said, and Mohammed Benali Kanoun

J Phys Chem C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.5b03834 • Publication Date (Web): 03 Jun 2015

Downloaded from http://pubs.acs.org on June 15, 2015

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication They are posted

online prior to technical editing, formatting for publication and author proofing The American Chemical

Society provides “Just Accepted” as a free service to the research community to expedite the

dissemination of scientific material as soon as possible after acceptance “Just Accepted” manuscripts

appear in full in PDF format accompanied by an HTML abstract “Just Accepted” manuscripts have been

fully peer reviewed, but should not be considered the official version of record They are accessible to all

readers and citable by the Digital Object Identifier (DOI®) “Just Accepted” is an optional service offered

to authors Therefore, the “Just Accepted” Web site may not include all articles that will be published

in the journal After a manuscript is technically edited and formatted, it will be removed from the “Just

Accepted” Web site and published as an ASAP article Note that technical editing may introduce minor

changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers

and ethical guidelines that apply to the journal pertain ACS cannot be held responsible for errors

or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Tuning Magnetic Properties of BiFeO 3 Thin Films by Controlling Rare-Earth Doping: Experimental and First-Principles Studies

Nguyen Hoa Hong 1*, Ngo Thu Huong 1,2, Tae-Young Kim 1, Souraya

Goumri-Said3, and Mohammed Benali Kanoun4, †

1

Nanomagnetism Laboratory, Department of Physics and Astronomy, Seoul National University,

Seoul 151-747, Korea 2

Hanoi University of Science, 334 Nguyen Trai, Thanh Xuan, Hanoi, Vietnam

3 School of Chemistry and Biochemistry and Center for Organic Photonics and Electronics, Georgia Institute of Technology, Atlanta, Georgia 30332-0400, United States

4 School of Physics, Georgia Institute of Technology, Atlanta, Georgia 30332-0400, United

2

ABSTRACT

Rare Earth (RE) - doped BiFeO3 (BFO) thin films were grown on LaAlO3 (LAO) substrates by

using pulsed laser deposition technique All of BFO films doped with 10% of RE exhibit a

rhombohedral single phase As for the Pr and Nd doping cases, the ferromagnetic phase is less

favored because Fe2+ amount is not dominant When dopant concentration was increased up to

20%, the RE-doped BFO films have gone through a structural transition from rhombohedral to

either pure orthorhombic phase (for Ho, Sm), or a mixed phase of orthorhombic and tetragonal

(for Pr, Nd), or pure tetragonal (for Eu) As an important consequence, magnetic properties of

RE-doped BFO films have drastically changed Our results give a guide for how to tailor the

ferromagnetism of BFO films by appropriate controlling the type of RE dopant as well as dopant

concentration The experimental findings are completed by performing density functional theory

calculations to explore the effect of RE doping in BFO for the considered three phases

INTRODUCTION

Multiferroics are an interesting group of materials that exhibit both ferroelectricity and ferromagnetism with coupled electric and magnetic order parameters [1-3] Multiferroism is currently the subject of intensive scientific investigation, since they potentially offer a wide range of interesting applications [2-6] BiFeO3 (BFO) is known to be the only ABO3-type simple perovskite that shows multiferroic at room-temperature and, thus, is considered to be the most promising candidate for practical applications among multiferroic materials [3, 5, 6] At room temperature, BiFeO3 exhibits a distorted perovskite structure with rhombohedral polar R3c

symmetry At higher temperatures (≈1100 K), the rhombohedral (R) phase undergoes a first order phase transition to a GdFeO3-like Pbnm structure [7-9] and a (probable) orthorhombic γ-

phase [10] Basically, BiFeO3 should be G-type antiferromagnetic due to the local spin ordering

of Fe3+, that forms a cycloidal spiral spin structure [11] There are several ways to stress the spiral magnetic ordering by applying a very high magnetic field, or reducing the dimensions of the samples, or by replacing Bi3+ or Fe3+ by other ions of comparable ionic sizes [12].Dimension reduction seems to be an effective method to enhance the magnetic moment in BFO thin films and in nanoparticles [6, 13] Some groups have reported about the increase of magnetization in the bulk, thin films, and nanoparticles of BFO, either by substituting on the Bi-site by trivalent rare-earth and divalent ions, or on the Fe-site by transition metal ions Thakuria and Joy had showed that the magnetic moment of the nanoparticles could be enhanced 3 times by substituting

Bi by Ho However, the reported saturated magnetization is still found only at a quite high field

as of 6 T [12, 14] Partial substitution of Bi by Rare-Earth (RE) ions is known to induce a

4

ferromagnetic response which has been attributed to suppression of the spiral modulation [15,

16] In particular, doping of the Bi3+ A site of the perovskite with rare earth cations (RE) has

received extensive attention, with a variety of symmetries, and magnetic and electric behaviors

reported with increasing values of x in Bi1-xRExFeO3 and/or decreasing ionic radii of the rare

earth cation [8, 9, 17–18] Recent studies of bulk and thin films of (Bi,RE) FeO3 (RE = Nd, Sm)

have revealed a formation of a stable antipolar, PbZrO3-like structure in a narrow rare-earth

concentration range [8, 18-19]

In this respect, ab initio calculations based on the density-functional theory (DFT) have played

an important role in the description, understanding, as well as prediction—via identification of

suitable material design rules—of magnetic, ferroelectric and magnetoelectric properties of

multiferroics, due to its ability to describe the many active degrees of freedom within a

comparable level of accuracy [3, 20-23] Theoretical modeling based DFT approach and

effective Hamiltonian scheme have been subject of few recent works on RE doped BFO [19, 23,

24] where they mainly reported the dependence of critical temperature on the RE compositions

In the present study, we attempt to tune the magnetic properties of BiFeO3 by several ways

such as: selecting the suitable RE ion for doping; screening the appropriate concentration,

targeting to obtain the largest magnetization possible at room temperature, and at a relatively low

field The aim of the present work is to understand the relationship between structural and

magnetic properties of RE-doped BFO films by combining ab initio calculations and thin-film

growth experiments in order to control magnetism of this family of compounds, with hope to

guide correctly the materials strategy for spintronic and magnetic applications

EXPERIMENTAL AND COMPUTATIONAL DETAILS

RExBi1-xFeO3 ceramic targets (where RE= Ho, Sm, Pr, Nd, and Eu; x = 0; 0.1, and 0.2) were

prepared by a sol-gel auto ignition method [14] RE-doped BiFeO3 (RBFO) thin films have been deposited by Pulsed Laser Deposition (PLD) technique (eximer KrF laser with λ= 248 nm; the repetition rate was 13 Hz and the energy density was 2.1 J/cm2), with a typical thickness as of

200 nm All the films were grown on (001) LaAlO3 (LAO) substrates During deposition, the substrate temperature was kept at 700°C and the oxygen partial pressure (PO2) was 1.4×10-3 Torr After deposition, the sample was kept in the chamber at 500°C with the same oxygen partial pressure as during deposition for 30 min, and then finally cooled down slowly to room temperature [see also Ref 14 for details] The structural analysis was carried out by High Resolution X-ray diffraction (HRXRD) with Cu Kα radiation The M-T and M-H curves were

collected by a Quantum Design Superconducting Quantum Interference Device (SQUID) system

with magnetic field (H) ranging from 0 up to 0.5 T and temperatures (T) ranging from 350 K

down to 5 K The oxidation states of RE-doped BFO thin films were characterized by X-ray photoelectron spectroscopy (XPS, KRATOS, AXIS-HSi) XPS measurements were performed with an Mg/Al X-ray source The energy calibrations were made against the C 1s peak and the Shirley background subtraction was used [as in Ref 14] The chemical elements’ content was also checked by Energy-dispersive X-ray diffraction spectroscopy (EDX) at room temperature

Our calculations were performed using all electron linearized augmented plane wave method with local orbitals basis set based on DFT as implemented in the WIEN2k computer program [25] Exchange and correlation were treated within the generalized gradient approximation

(GGA) of Perdew-Burke-Ernzerhof (PBE) [26] Onsite Hubbard interaction between the 5f

6

electrons was treated within the fully rotationally invariant version [27] Here, U (equal to what

is often called Ueff = U − J) is taken as the on-site interaction term as suggested in Ref 28 The

cutoff R mt *Kmax was set to 7.0 to determine the basis sets To obtain more information on RE

doped-BiFeO3, we were carried out DFT calculations by adopting the following supercell

approach: (i) a 2×2×2 supercell that contains 12 FeBiO3 formula cells for the hexagonal R3c

structure, and (ii) a 2×2×1 supercell that contains 8 formula cells for the orthorhombic Pbnm and

tetragonal P4mm structures In each supercell, a Bi atom was substituted by one RE impurity, in

order to obtain Bi1-xRExFeO3 with x = 0.0833 for hexagonal structure and 0.25 orthorhombic and

tetragonal structures For the integration over the Brillouin zone, a 4×4×1 Monkhorst–Pack

k-point mesh [29] was used for the rhombohedral (hexagonal) cell while a 5×5×3 Monkhorst–Pack

k-point grid was adopted for the orthorhombic and tetragonal cells The convergence of

self-consistent calculations was attained with a total energy convergence tolerance of 0.1 mRy

RESULTS AND DISCUSSION

As we know, the EDX method could not give a very precise evaluation of content of each

element in the case of thin films due to the fact that it is just most sensitive to the surface of the

film but not as the whole However, one can see a slight tendency of deviation in resulting

concentration if comparing to the starting doping concentration For example, if we substituted

20% of Bi by Ho, then the resulting Ho:Bi ratio is 22:78, if we substitute 20% of Bi by Sm, then

the resulting Sm :Bi ratio is 26.4: 73.6 Therefore, thoroughly in this report we keep naming the

compounds as their starting stoichiometry The XRD data show that for the case of doping of

10%, there is no significant change in structures in comparison with the pristine BiFeO3 All

RE0.1Bi0.9O3 films have a rhombohedral structure showing very strong peaks of the BFO phase

Doping with different RE elements only causes some certain shift of the peak positions in the spectra, showing some change in lattice parameters The out-of-plane lattice parameter is 3.925, 3.925, 3 9329, 3.933, and 4.585 Ǻ for Sm, Ho, Nd, Pr and Eu doped BFO, respectively (as discussed partially in Ref 14] When the RE doping concentration increases up to 20%, a drastic change in structure of the compound has appeared: As for doping of Ho and Sm , the structure is single phase orthorhombic, while for Pr and Nd, it has become a mixed phase of orthorhombic and tetragonal, and for Eu doping case, it is single phase tetragonal Some typical spectra are shown in Figure 1 for comparison between structures of 10% and 20% doping cases, showing changing from rhombohedral to orthorhombic for Ho doping case, and changing from rhombohedral to a mix of orthorhombic and tetragonal for Pr doping case In order to make it easier later for reference, we summarize the structure types for all RExFe1-xO3 thin films in Table

8

0 50

0 50 100

0 50 100

0 50 100

Figure 1: X-ray diffraction patterns for 200 nm-thick- films grown on LaAlO3 substrates of (a)

Ho0.1Bi0.9FeO3-showing Rhombohedral phase; (b) Ho0.2Bi0.8FeO3-showing Orthorhombic phase;

(c) Pr0.1Bi0.9FeO3-showing Rhombohedral phase; and (d) Pr0.2Bi0.9FeO3-showing a mix of

Orthorhombic and tetragonal phases

Table 1: List of structural phases and Fe2+:Fe3+ ratio calculated from XPS data for RExBi1-xFeO3

films (RE= Ho, Sm, Eu, Pr, and Nd)

element concentration

M s was obtained at a much greater field (as of 6 T) [12, 14], while we got much a larger Ms but at much lower field (as of 0.2 T) This makes a difference and it is quite meaningful for applications In comparison to those three dopants mentioned above, the Pr- and Nd-doped BFO films show much weaker magnetism: the curves in Figure 2 (a) show a much smaller magnitude

for magnetization, and the curves of M(H) at room temperature are almost linear indicating a

10

Figure 2: (a) Magnetization versus magnetic field taken at 300 K for (a) 200 nm-thick

RE0.1Bi0.9FeO3 films; (b) 200 nm-thick RE0.2Bi0.8FeO3 films (RE= Ho, Sm, Eu, Pr, and Nd), and

(c) Zoom for the low field region of the M-H curve of 200 nm-thick Sm0.2Bi0.8FeO3; The insets

in (b) shows the zoomed M-H curves of the two typical 10% doping cases (Pr and Sm) in order

to compare directly with the 20% doping case

In order to identify the origin of magnetism of our films, XPS measurements were performed

The peak of Fe 2p could be well observed at 711 eV for Fe3+ and at 709.5 eV for Fe2+ [14] From

the shape of the peaks, it is seen that there are Fe2+ and Fe3+ in all doping cases (Ho, Sm, Nd, Pr,

and Eu) The rough fitting analysis to the peaks reveals the oxidation state of Fe in our films

which is listed in Table 1 One can see from Table 1 that for the 10% doping case, Fe2+: Fe3+

ratio is about 50%: 50% for the cases of Ho and Sm, roughly so for Eu, but only about 40%: 60%

for the Pr and Nd cases [14] The coexistence of Fe2+ and Fe3+ is thought to be in favor of the

ferromagnetic phase in BFO films due to the double exchange between Fe2+ and Fe3+ via the role

of oxygen as intermediates [32, 33] When the amount of Fe3+ is more favored, the

ferromagnetism get weaker due to the fact the Fe3+- Fe3+ interaction is in favor of

antiferromagnetic ordering [33] This explains well the corresponding SQUID data shown in Fig

2 (a), we may understand why in Pr and Nd-doped BFO films, the ferromagnetic phase is not

favored [14]

When the doping concentration is increased up to 20%, the Fe2+:Fe3+ ratio has changed as the

results of changes in structure (recalling the difference in phase obtained that was discussed

earlier concerning Figure 1) In Table 1, one can see that as for Pr and Nd, there is a significant

increase of Fe2+ amount, leading to an enforcing of the ferromagnetic ordering On the contrary,

as for Ho, Sm, and Eu doping cases, Fe2+ seems to be decreased, relating to the weakening of

ferromagnetic phase This can be seen clearly from M-H curves for the 20% doping case shown

in Figure 2 (b) As for the Pr case, whose phase is a mix of orthorhombic and tetragonal, one can see that magnetic phase has changed from pure paramagnetic (see in the inset on the left for

the 10% doping case) into mixed paramagnetic and ferromagnetic (the M-H curve of

Pr0.2Bi0.8FeO3 film shows an accumulation of a part as of paramagnetic and another part as weak ferromagnetic phase) Differently, as for Sm case (similarly for Ho and Eu cases also), it has changed from very strong ferromagnetic into “almost” paramagnetic (compare the curve seen in Fig 2 (b) for the 20% Sm case with that of the 10% Sm case shown in the inset on the right hand side) It is totally in accord with the structural analysis shown above The zoom for the low field region of a typical ferromagnetic sample (in this case as of 20%Sm-doped BFO film) is shown in Figure 2(c) to see clearly the loop showing their ferromagnetic behavior Indeed our samples have coercivity HC ranging from 25-70 Oe (Pr-BFO: ~27 Oe; Nd-BFO~55Oe; Ho-BFO ~52 Oe, Sm-BFO~48 Oe, Eu-BFO~ 70 Oe) In fact doping RE does not change much the coercivity of the host compound (BFO target and film have almost the same magnitude of HC (61 Oe- see in

the M -H data show in Figure 3 (a) and (b)) However, doping RE has enhanced the

magnetization to about 2 orders in comparing to that of the pristine BiFeO3 (refer Figure 3 (a) and (b)) Previously we have tried to dope RE with lower concentration as of 5% Our results showed that 10% samples have larger magnetic moment One can see a typical example shown

in Figure 3 (c), and more details in Ref 14 From this work, one could see that starting from doping 20%, samples are not single phase Therefore, we assume that roughly for the sake of

applications, doping 10% seems to be most suitable Paudel et al have proposed some model to

suppose that ferromagnetic moment in BFO may be due to intrinsic defects [34] But no one has confirmed that defects should be the main source of magnetism in this type of compounds, either undoped BFO or Rare-Earth doped BFO For films, moreover, it is hard to determine oxygen

12

vacancies and from the M-H data, “oxygen vacancies” seem to be unlikely the reason for

magnetic moment here One can see that with the same environment, magnetic behaviors of

different compounds (pristine, RE-doped BFO with different RE element, or with different RE

concentration) are different And Fe2+:Fe3+ ratio is seen to be quite different It is more logical to

think that this ratio should be the main reason to alter the magnetic properties of RE-BFO

compounds From other aspect, the strains might influence properties of BFO somehow We also

noticed that surface magnetism could also play a role (for films, due to strains of substrates [14]

However in this work we conclude for films grown by PLD on LaAlO3 substrates only, not to

generalize for all cases including bulks It should be too much complicated to include all on one

plate

Figure 3: Magnetization versus magnetic field taken at 300 K for (a) undoped-BiFeO3 ceramic

target; (b) 130 nm-thick undoped-BiFeO3 film; and (c) for 200nm thick-Ho0.05Bi0.95FeO3 and

Ho0.1Bi0.9FeO3 films

From our theoretical calculation, we calculated the magnetic moments values for

orthorhombic, tetragonal and rhombohedral phases, as shown in Figure 4(a-d) In Fig 4 (a), the

rhombohedral phase shows the largest total magnetic moment for all the RE elements doping

BFO The magnetic moment per RE atom shows that the magnetic moment value of the Eu is

bigger than Pr, Nd, Sm, and Ho values, as shown in Figure 4 (b) The RE doped- BFO leads to a change of magnetic moment of Fe atom However, the magnetic moment per Fe atoms, as displayed in Figure 4 (c), shows no trend in its variation In fact, its value varies between 2.6 and 3.6 µB per Fe atom for all considered RE dopants in three phases This might be resulting from the difference of structure and symmetry In fact, the RE doping enhanced the magnetic moment

of BFO but it also induces the polarization of oxygen atoms situated in nearest site to RE and Fe The magnetic moment values of O are also presented in Figure 4 (d), where we see that the largest polarization of O is found in the tetragonal phase [35, 36]

Figure 4: (a) Total magnetic moment values for RExBi1-xFeO3 for three phases and the magnetic

moment per atom: (b) RE, (c) Fe, and (d) O The y-axis shows the magnetic moment values