J Materiomics Vol 2, 2016

Graphical Abstracts

REVIEWARTICLES

J Materiomics 2016, 2, 213e224DyMnO3: A model system of type-II multiferroics

Chengliang Lua,*, Jun-Ming Liub,**aSchool of Physics & Wuhan National High Magnetic Field Center,

Huazhong University of Science and Technology, Wuhan 430074, ChinabLaboratory of Solid State Microstructures and Innovation Center of Advanced Microstructures,

Nanjing University, Nanjing 210093, China

In this short review, some recent experimental results on tailoring the multiferroicityand magnetoelectric coupling in DyMnO3 were represented. The multiferroicity andmagnetoelectric coupling of DyMnO3 were described based on various approachesincluding chemical substitution and strain/domain engineering. These results mayindicate some alternative strategies in designing superior multiferroicity andunconventional magnetoelectric controls for multiferroic applications.

J Materiomics 2016, 2, 225e236Recent development of n-type perovskitethermoelectrics

Hongchao Wang*, Wenbin Su, Jian Liu,Chunlei Wang**

School of Physics, State Key Laboratory of Crystal Materials,

Shandong University, Jinan, China

SrTiO3 and CaMnO3-based bulk thermoelectric perovskite oxides have been systematically reviewed. The element doping as traditionalmodification is generally used for improving the figure of merit of oxide. Nanostructuring, nanocomposite and defect chemistry engi-neering are beginning to significantly enhance the thermoelectric performance, especially for SrTiO3 based oxide and more state-of-the-artvalues of figure of merit have been achieved. So nanostructuring, nanocomposite and combining doping and defect chemistry engineeringcan be developed to further optimize the thermoelectric performance of oxides as the effective methods. And then more significantbreakthrough on high-temperature thermoelectric oxides will be achieved in future.

Available online at www.sciencedirect.com

ScienceDirect

J Materiomics 2 (2016) iveviwww.ceramsoc.com/en/ www.journals.elsevier.com/journal-of-materiomics/

Available online at www.sciencedirect.com

ScienceDirect

J Materiomics 2 (2016) 213e224www.ceramsoc.com/en/ www.journals.elsevier.com/journal-of-materiomics/

DyMnO3: A model system of type-II multiferroics

Chengliang Lu a,*, Jun-Ming Liu b,**a School of Physics & Wuhan National High Magnetic Field Center, Huazhong University of Science and Technology, Wuhan 430074, Chinab Laboratory of Solid State Microstructures and Innovation Center of Advanced Microstructures, Nanjing University, Nanjing 210093, China

Received 14 January 2016; revised 14 March 2016; accepted 7 April 2016

Available online 3 May 2016

Abstract

One of the main motivations for multiferroic researches is to explore distinctive magnetoelectric controls for conceptually novel electronicdevices and information storage technologies. DyMnO3 possesses two major microscopic mechanisms for ferroelectricity generation, i.e., theinverse DzyaloshinskiieMoriya interaction between Mn spin pairs and the exchange striction between DyeMn spin pairs, making it anaggressive model to address various opportunities for magnetoelectric controls. In this short review, some recent experimental results on tailoringthe multiferroicity and magnetoelectric coupling in DyMnO3 were represented. By means of various approaches including chemical substitutionand strain/domain engineering, we have demonstrated the multiferroicity and magnetoelectric coupling of DyMnO3, which can be manipulatedover a broad scale. These results may bring alternative strategies in designing superior multiferroicity and unconventional magnetoelectriccontrols for multiferroic applications.© 2016 The Chinese Ceramic Society. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Multiferroic; Ferroelectricity; Exchange striction; Thin film; Domain

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2132. Dual nature of multiferroicity in DyMnO3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2143. Ferroelectricity by tuning the Dy3þeMn3þ spin coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2164. “Copy” bulk multiferroicity to DyMnO3 thin films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2175. Continuous ME control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2206. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

1. Introduction

The terminology “multiferroics” refers to those materials inwhich two or more ferroic orders such as ferroelectricity and

* Corresponding author.

** Corresponding author.

E-mail addresses: cllu@mail.hust.edu.cn (C. Lu), liujm@nju.edu.cn

(J.-M. Liu).

Peer review under responsibility of The Chinese Ceramic Society.

http://dx.doi.org/10.1016/j.jmat.2016.04.004

2352-8478/© 2016 The Chinese Ceramic Society. Production and hosting by Elsevi

creativecommons.org/licenses/by-nc-nd/4.0/).

(anti-)ferromagnetism coexist [1,2]. Among all the availablemultiferroic materials, type-II multiferroics of magnetic originstand out due to the unique one-to-one correspondence be-tween ferroelectric (FE) and magnetic orders [3e6]. Theintimate coupling between the two orders gives rises to giantmagnetoelectric (ME) effect, such as polarization (P) switch-ing in multiferroics hosting a spiral spin order (SSO) [7]. Sofar, dozens of type-II multiferroic materials have beenexplored, and our understanding of the ferroelectricity andmagnetism coexistence which otherwise would have been

er B.V. This is an open access article under the CC BY-NC-ND license (http://

214 C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

impossible within the framework of ferroelectricity-magnetism exclusion [8].

In general, these type-II multiferroics with magnetic originfor ferroelectricity can be further classified according to thedetails of microscopic mechanism. The first group includesthose having a noncollinear spiral spin order, making abroken space inversion symmetry and thus a spontaneousFE polarization via the well-known inverse Dzya-loshinskiieMoriya interaction (DMI). The as-generated po-larization is formulated as P ~ eij � (Si � Sj), where eij is thevector connecting two neighboring spins Si and Sj [9]. Aunique property of these multiferroics in this group is themagnetic field (H ) driven polarization switching, leading to agiant ME effect which has been repeatedly revealed inTbMnO3 (TMO) and DyMnO3 (DMO) [10]. The secondgroup has a specific collinear spin configuration, and conse-quently an improper ferroelectricity may be driven by the so-called exchange striction mechanism. The polarization can beexpressed as P ~ Si,Sj, as demonstrated in Ca3Co1�xMnxO6

and orthorhombic HoMnO3 etc [11e15]. The FE polarizationgenerated via this exchange striction mechanism is usuallymuch larger than that from the DMI mechanism, and actuallyamong the largest for all the type-II multiferroics. In additionto these two mechanisms, there is still one more worthymentioning, namely the spin dependent p-d hybridization, andthe polarization is written as P ~ (Si ,eil)

2,eil where eil is thevector connecting Si and the neighbor ligand ions [16]. Onesymbolic material featured with this p-d hybridization isBa2CoGe2O7, in which the polarization can be rotated inharmonization with the spin Si rotation driven by magneticfield [16]. However, the as-generated polarization P in suchmultiferroics is usually small (<100 mC/m2) despite thismanipulation is impressive.

Indeed, a bunch of novel ME control functionalities havebeen demonstrated in these type-II multiferroics. Importantly,it is found that the magnetic manipulation of P can befundamentally distinct in different multiferroics because ofdifferent microscopic origins. For instance, the magnetic fieldcontrol of polarization P in the first group of multiferroics isrelatively easy but the polarization magnitude is insufficientlylarge. The opposite situation is often observed for the secondgroup of multiferroics where the polarization is large but itscontrollability using magnetic field is far from expectation[10,12].

Besides, the physics of multiferroics is evidenced withmuch more fascinating phenomena, and one of them is themore than one mechanism has been found to be involved insome multiferroics, coined as the duality or triplicity of mul-tiferroicity. This property provides an avenue for unconven-tional ME functionalities. As a typical example, one considersDyMnO3 (DMO) as one member of the RMnO3 family whereR is the rare-earth element, which was classified into the firstgroup and the ferroelectricity was thought to originate thenoncollinear spiral spin order until recent experiments [17].Now it is well accepted that the total polarization Pc along thec-axis in DMO includes two components, one from the sym-metric exchange striction between the R3þeMn3þ spin pairs

(R is the rare earth element), i.e., PR ¼ PDy ~ SDy$SMn and theother from the inverse DMI mechanism between the MneMnspin pairs, i.e., PMn ~ (Si � Sj)Mn. The total polarizationreaches up to Pc ~ 2800 mC/m2. Furthermore, a colossalmagneto-capacitance up to ~500% was reported, owing to thesubsequent response of the two components to magnetic field.In this sense, DMO can be viewed as a model system toinvestigate a series of unusual issues like multiferroic dualityand multi-fold ME manipulation, not only benefiting to ourunderstanding of unconventional phenomena in type-II mul-tiferroics but also allowing more opportunities for function-ality realization and optimization. We shall present abreakthrough point at which one sees how DMO opens upthese aspects. For example, a specific domain structure inDMO thin films properly deposited on some substrates hasbeen found, which as an additional degree of freedom addsmore functionality [18e20]. Some research aspects regardingcation substitution, strain and domain engineering, andmagnetoelectric control in both bulk and thin film DMO wererepresented.

2. Dual nature of multiferroicity in DyMnO3

We start from the duality of DMO, which is the basis forthe present work. A seminal work among a huge number ofearlier investigations on multiferroics was the discovery ofmultiferroicity in TbMnO3 (TMO) in 2003 and the ferroelec-tricity appears as a consequence of spiral spin ordering [21].Subsequently, several more materials in this category werefound [10,22], while DMO has been receiving more attentionthan TMO because of its intriguingly large FE polarization andcolossal magento-capacitance [23,24]. However, these supe-rior properties were not well understood until 2009 when thedetails of magnetic structure were unveiled.

In fact, DMO has an orthorhombic perovskite structurewith lattice parameters a ¼ 5.8337(1), b ¼ 7.3778(1), andc ¼ 5.2785(1) Å at room temperature [25]. A significantstructural characteristic, owing to the small ionic size of Dy3þ,is the GdFeO3-type distortion which drives the MnO6 octa-hedra tilting in the ab-plane and rotation with respect to the c-axis. Such a distortion reconstructs the multifold exchangeinteractions among Mn3þ, and thus stabilizes the Mn3þ spiralspin order if the Jahn-Teller effect is taken into account [26].As revealed by neutron diffraction, the Dy3þ spins do aligncoherently with the Mn3þ spins below a certain temperaturedue to the strong Dy3þeMn3þ coupling [27], making the Dy3þ

spins to align noncollinearly and incommensurately incoherence with the Mn3þ spiral spin order.

Indeed, the spin configurations of DMO are illustrated inFig. 1. In the cooling sequence, DMO experiences the Mn3þ

spin ordering to a sinusoidal antiferromagnetic (AFM) struc-ture at temperature T ¼ TN ~ 39 K, and then to an incom-mensurate (ICM) spiral structure at T ¼ TC ~ 19 K. Anemergent FE polarization along the c-axis (PMn) is induced viathe inverse DMI mechanism. Further cooling causes aremarkable enhancement of total polarization Pc, which can betwice that of TMO in which PMn is the dominant contribution

Fig. 1. (a) The T-dependence of out-of-plane FE polarization Pc for DyMnO3 (red and purple lines) and TbMnO3 (dashed line). Three regions can be identified

below TC, and the corresponding spin structures projected onto the bc-plane are shown in the top of figure. Reprinted by permission from Ref. [27]. (b) The T-

dependence of polarization Pc under various magnetic field for DyMnO3, in which the enhancement of Pc is due to the CM-ICM transition of Dy3þ spins.

Reprinted by permission from Ref. [23].

215C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

to Pc [10,27]. When the Dy3þ spins evolve coherently with theMn3þ spins below TC, as shown in the upper of Fig. 1, aspecific magnetic structure is developed, in which the Mn3þ

spins on each Mn-chain align parallel to the Dy3þ spins on oneof the nearest neighboring Dy chain but antiparallel to thoseon the other nearest neighboring Dy-chain. This configuration,called the synchronization between the Dy3þ spin sublatticeand Mn3þ spin sublattice, fits well the requirement for theexchange striction mechanism, giving rise to an additional netcomponent PR ¼ PDy and thus one has Pc ¼ PMn þ PR [17].At T ¼ TR ~ 7 K, the Dy3þeDy3þ interaction goes beyond theDy3þeMn3þ coupling, and an independent commensurate(CM) ordering of Dy3þ spins appears. This ICM-CM transi-tion of Dy3þ spins eliminates the polarization component PR,leaving the polarization component PMn alone and thus lead-ing to a rapid drop of Pc below TR ~ 7 K. As revealed byneutron diffraction, this ICM-CM transition exhibits anevident thermal hysteresis, well illustrating the bifurcation ofPc(T ) curves obtained through different cooling procedures[23,27]. The Pc(T ) dependence in bulk DMO is shown inFig. 1(a). On the contrary, this ICM-CM transition can besuppressed by a magnetic field H, resulting in a reentrancy ofcomponent PR, and thus a remarkable enhancement of Pc, asshown in Fig. 1(b) [24].

Interestingly, Schierle et al. observed the sizable b-axis andc-axis components of Dy3þ spins using the soft X-raydiffraction at the Dy-M5 resonance on DMO [28]. This sug-gests a cycloidal Dy3þ spin order at TR < T < TC, which mayalso contribute to the ferroelectricity further via the inverseDMI mechanism between Dy3þ spin pairs. This issue remainsto be addressed in future. Even though, such a possiblecycloidal Dy3þ spin order is still coherent with the Mn3þ spin

structure, allowing the Dy3þeMn3þ exchange strictionmechanism active.

To this stage, the above results do allow one to claim thatDMO could not be better to act as a platform for illustratingthe rich physics of multiferroicity. In fact, the strongR3þeMn3þ coupling is not an unusual phenomenon for mul-tiferroic RMnO3. TMO is another system in which the Tbspins develop an order with the same propagation vector (tR)as that (tMn) of the Mn spin order, i.e., t1

Tb ¼ tMn ¼ 0.274,right below TC ¼ 27 K [29]. At T ¼ TR ~ 7 K, vector t1

Tb shiftsto a higher value of 0.286, and simultaneously a second wavevector t2

Tb ¼ 0.4285 for the Tb3þ sublattice appears, leading toan evident kink in the Pc(T ) curve. The orthorhombicHoMnO3 with an E-type AFM order was theoretically pre-dicted to show a remarkable FE polarization along the a-axisvia the exchange striction of Mn3þ spins [11]. However,experimentally only a FE phase with Pc ~ 1500 mC/m2 whichcontinuously increases with decreasing T down to ~2 K, wasidentified [12]. Neutron diffraction on HoMnO3 revealed thatthe Ho spins also evolve coherently with the E-type AFMorder of Mn3þ spins, responsible for the emergent Pc via theHo3þeMn3þ exchange striction. As compared with DMO, asurprising feature for HoMnO3 is that the HoeMn couplingcontributes to the FE phase significantly till a very low T farbelow T ¼ TR, suggesting no independent ordering of theHo3þ spins [12,30]. The situation in GdMnO3 is a bit morecomplicated since this material doesn't show apparent ferro-electricity [10], unless a magnetic field is applied, whichrigidly tracks the field-induced modification of Gd3þ spinsublattice. While the underlying mechanism for the ferro-electricity in GdMnO3 has still not been well understoodbecause of the absence of sufficient magnetic data, the Mn3þ

Fig. 2. T-dependence of FE polarization in Dy1�xHoxMnO3 (0 � x � 0.3). The

inset shows the polarization at T ¼ 2 K as a function of Ho3þ substitution level

x. Reprinted by permission from Ref. [34].

Fig. 3. T-dependence of FE polarization in Dy1�xYxMnO3 (0 � x � 0.2).

Reprinted by permission from Ref. [35].

216 C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

cycloidal spin order modulated by the Gd3þ AFM phase at lowT was proposed [31,32].

While the above discussed R3þeMn3þ coupling contrib-utes to the multiferroicity, Mochizuki et al. discussed the na-ture of multiferroicity in RMnO3 which is of multifold even ifR3þ is nonmagnetic. In this case, both the symmetric(P ~ (Si�Sj)Mn) and antisymmetric (P ~ (Si � Sj)Mn) compo-nents from the Mn spiral spin order are possible [33]. Thisexplains why the spontaneous FE polarization associated withthe ab-plane spiral spin structure is larger than that associatedwith the bc-plane spiral spin structure for RMnO3. Neverthe-less, such multifold multiferroicity solely from the Mn3þ

spiral spin order remains to be identified experimentally.

3. Ferroelectricity by tuning the Dy3þeMn3þ spincoupling

As discussed, the duality of multiferroicity in DMO pro-vides an ideal platform to integrate the multiferroicity fromvarious mechanisms, i.e., the giant ME effect associated withmagnetic field driven polarization switching. Keeping in mindthis motivation, Zhang et al. carried out systematical experi-ments on tuning of the duality of multiferroicity, which in turndemonstrates the significant role of the Dy3þeMn3þ exchangestriction [17,34,35].

In details, the Dy3þ spin order undergoes the ICM-CMtransition at TR, which destabilizes the coherence of Dy3þ

spins with Mn3þ spins. Consequently, polarization componentPR is no longer available, making the total polarization Pc tofall down drastically from ~2800 mC/m2 to 600 mC/m2 [23].An alternative to save this component PR is to stabilize theDy3þ ICM order against the ICM-CM transition at TR, whichcan be accessed by applying a magnetic field, as demonstratedalready. Apart from this way, cation substitution is shown to beanother approach to manipulate PR. Similar to DMO,HoMnO3 also accommodates the strong Ho3þeMn3þ

coupling to preserve the coherence between the Ho3þ spinsand Mn3þ spins down to the lowest T [30], while no suchICM-CM transition occurs. In this case, the polarizationcomponent (PR) arising from the Ho3þeMn3þ exchangestriction can be saved.

In proceeding, a set of experiments on substitution of Dy3þ

in DMO by Ho3þ were performed. Two effects upon the Ho3þ

substitution can be expected. First, the substitution is expectedto preserve the R3þeMn3þ coupling, essential for ferroelec-tricity generation via the exchange striction. Second, thesubstitution would dilute the Dy3þ sublattice, which certainlyhinders the independent Dy3þ spin ordering below TR sincethe 4fe4f electron coupling is relatively weak. Nevertheless,such a substitution does not affect much the coherency of theDy3þ spins with the neighboring Mn3þ spins, noting that mostlikely the substituting Ho3þ spins may also align coherentlywith the neighboring Mn3þ spins. In this sense, the substitu-tion does not qualitatively change the ICM R3þ alignmentinduced by the Mn3þ spin order via the R3þeMn3þ coupling.In short, one is expected to observe a remarkable enhancementof the total polarization Pc in the extremely low-T end

(T < TR), crediting with the reentrancy of component PR.Fig. 2 presents the Pc(T ) curves of Dy1�xHoxMnO3

(0 � x � 0.3) [34]. Indeed, a drastic enhancement of Pc in thelow-T range is observed, suggesting a successful preservationof the synchronization of the Ho3þ spins with the Mn3þ spinsin the spiral alignment.

Regarding the Ho3þ substitution in DMO, the second effectis a natural consequence of chemical substitution, but the firsteffect can still be different if a nonmagnetic/weak magneticR3þ or other trivalent species. Along this line, Y3þ could be anideal candidate since it is nonmagnetic and has nearly the sameionic radius as Ho3þ. In this case, the diluted Dy3þ spin sub-lattice on one hand would prevent the onset of the independentDy3þ spin ordering, and consequently prevent the drop in thePc(T ) curve below TR. On the other hand, the damaged Dy3þ

spin network will certainly suppress the overall Dy3þeMn3þ

coupling, leading to a suppression of Pc(T ) within TR< T< TC.As shown in Fig. 3, one sees that the polarization Pc falls downover the whole T-range first upon increasing Y3þ up to x ~ 0.1,owing to the global breakage of the Dy3þeMn3þ coupling.

Fig. 4. Microstructural characterizations on the (001) DyMnO3 thin film on (001) NbeSrTiO3 substrates. (a) X-ray qe2q scan of the film. The inset shows a sketch

of the epitaxial growth of the film. (b) The rocking curve around the (002) reflection of the film. The reciprocal space mapping data around the (103)c and (113)creflections of the substrate are shown in (c) and (d), respectively. (e) The atomic force microscopy of the film.

Fig. 5. The plane-view transmission electron microscopy image of the film, in

which the twin-like domain structure can be identified.

217C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

Further increasing of x makes the dilution of the Dy3þ sub-lattice strong enough so that the independent Dy3þCMorder isreplaced, leading to a reversal of the low-T P(T ) curve [35].This is a good example showing the tunable duality of multi-ferroicity in DMO. A detailed discussion on this duality and itsmodification by chemical substitution in DMO can be found ina previous mini-review paper [17].

4. “Copy” bulk multiferroicity to DyMnO3 thin films

While enthusiastic endeavors have been motivated by thefascinating physics of type-II multiferroics, the inherent AFMnature does not give rise to any macroscopic magnetization,which seriously hampers the potential applications of multi-ferroics. Given this unfortunate situation, several approaches toenhance the ferromagnetism and to achieve the sizable sponta-neous magnetization have been reported. The favorable strategyincludes the strain and domain engineering in thin films andnanostructures [19,20,36e40]. In fact, multiferroic thin filmsbelong to an inter-discipline bridging the novel cross-control ofmultiple ferroic orders and multifunctional devices.

The TMO thin films deposited on SrTiO3 single crystalsubstrates have been intensively investigated during the pastfew years, and interesting emergent ferromagnetism has been

evidenced. Two possible origins for ferromagnetism wereproposed. One is the release of the MnO6 octahedron distor-tion induced by the substrate transferred strain [36]. The otherone is the difference of electronic structure at domain walls,which may give rise to apparent ferromagnetism [18]. Inparticular, a twin-like domain structure was observed in thesethin films [19,20], which may lead to unconventional ME

218 C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

controls to be discussed below. Although exciting weakferromagnetism in TMO thin films has been reported, themagnetically induced ferroelectricity known in the bulkcounterpart seems to be suppressed simultaneously, raising achallenge. According to the phase diagram of RMnO3, thenovel spiral spin structure is rather sensitive to the structuredistortion and can only exist within a very narrow region [23].

Fig. 6. (a) Measured out-of-plane dielectric constant εc as a function of T, in which t

data upon different cooling sequences. (c) Measured FE hysteresis loop and the cor

with the opposite chirality (Q�a and Qa) are presented.

Fig. 7. Measured Pc(T ) data under various magnetic fields

It is physically reasonable to expect a destabilization of thespiral spin structure in strained TMO thin films, since themagnetism of bulk TMO locates quite close to the phaseboundary between the A-type AFM and spiral spin order in thephase diagram. However, a clear signature of the ferroelec-tricity in RMnO3 thin films, which is associated with the spiralspin structure, can be found. For instance, a recent ultrafast

he multiple phase transitions are indicated by dashed lines. (b) Measured Pc(T )

responding switching current. In the upper of (c), a sketch of spiral spin orders

applied (a) along the c-axis, and (b) in the ab-plane.

Fig. 8. Measured magnetization (M ) as a function of magnetic field H applied (a) in the ab-plane and (b) along the c-axis at various temperatures.

219C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

optical spectroscopy study on TMO thin films revealed ananomaly around T ~ 30 K, which is very close to the FE Curietemperature TC ¼ 27 K in TMO bulk crystals. A coexistenceof Mn3þ spiral spin order and ferroelectricity in the thin filmswas thus argued [41]. The bulk orthorhombic YMnO3 has anE-type AFM ground state, and the observed FE state wasassociated with the exchange striction mechanism. However,for YMO thin films grown on SrTiO3 (001) substrates, apossible spiral spin structure and switchable FE polarizationwere identified [42,43]. These phenomena encourage us toexplore more emergent phenomena of multiferroicity inRMnO3 thin films, in particular the DMO thin films.

We then prepared DMO thin films on STO (001) singlecrystal substrates and a set of characterizations on variousaspects of microstructures and multiferroic functionalitieshave been performed [44]. Two motivations drive us toinvestigate the DMO thin films. The first is the superiormultiferroicity of DMO, and the second one is the location ofDMO in the multiferroic phase diagram of RMnO3, whereDMO seems to be more proper than TMO in terms of themultiferroic manipulation [23], and the two issues allow moreflexibility in DMO thin films. For example, a relatively small

Fig. 9. A sketch of spin spiral plane flopping from the bc-plane to the ab-plane

driven by magnetic field along the b-axis or a-axis, giving rise to different

critical fields Hc1 or Hc2, respectively. Here Q denotes the chirality of Mn3þ

spiral spin order.

compressive strain in the DMO thin films was evidenced bystructural characterizations (Fig. 4), although the lattice misfitbetween STO and DMO can be huge. Such a huge misfit maybring into some more variants such as twin-like domainstructure, which could release the strain energy largely. Anumber of techniques have been employed to characterize thetwin-like structure, including the X-ray reciprocal spacemapping and the plane-view image from transmission electronmicroscopy (TEM) which presents clear evidences, as shownin Fig. 5.

We look at the experimental results. First, the dielectric andpolarization data of the as-prepared DMO thin films shown inFig. 6 evidence clearly the ferroelectricity and its relevancewiththe spiral spin ordering ofMn3þ spins. The dielectric constant asa function of T, εc(T ), marks the multiple phase transitions at

Fig. 10. Top: 4 is defined as the angle between the applied magnetic field and

[100] of the SrTiO3 substrate. Bottom: Sketch of the twin-like domain struc-

ture comprising a- and b-domains, in which arrows indicate the local domain

[010] directions of the film.

Fig. 11. A sketch of concurrent and distinct polarization switching events as H

is applied at (a) 4 ¼ 45� and (b) 4 ¼ 0 in the DyMnO3 thin films. The

macroscopic polarization, including Pc, Pa1 associated with the polarization

switching of a-domains, and Pa2 associated with the polarization switching of

b-domains, are also illustrated.

220 C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

TN ~ 40 K, TC ~ 21 K, and TR ~ 10 K, demonstrating the one-to-one correspondence between the bulk crystals and thin films.Secondly, such magnetically induced ferroelectricity can befurther demonstrated by the non-zero polarization Pc belowTC ~ 21 K, as shown in Fig. 6(b). The measured Pc(T ) hysteresisis believed to be associated with the ICM ↔ CM transition ofDy3þ spins and coupled Mn3þ spins via the Dy3þeMn3þ ex-change striction. Third, the positive-up-negative-down (PUND)method was used to attest the FE polarization, as shown inFig. 6(c). A well-defined polarization-electric field (Pc ~ Ec)hysteresis is obtained too. According to the scenario of DMinteraction associated with the spiral spin order ofMn3þ, the FEpolarization is determined by the chirality of the spiral spinorder. A reversal of polarization Pc marks an electric control ofthis chirality, one aspect of the ME coupling. On the other hand,tuning polarization Pc by magnetic field is also attractive. Asshown in Fig. 7, a variation ofPc upon amagnetic field in the a-bplane can be as high as ~800%.

The multiferroicity of DMO thin films highly resemblesthat of the DMO bulk counterpart. First, the polarization of theDMO thin films is about ten times smaller than that of thebulk. This difference is reasonable considering the modifiedmicrostructure (twin-domains) in the films. The magneticproperties of the as-prepared thin films, noting that the ferro-electricity originates from the specific magnetic structure, areshown in Fig. 8 where the measured magnetization M(H )curves with H//ab plane show quick enhancement aroundH ~ 1.5 T, owing to the reorientation of Dy3þ spins. Thisbehavior is similar to the observation of DMO bulk crystals.However, a close checking of the M(H ) curves shows a big in-plane to out-of-plane magnetization ratio (Mab/Mc) of ~10,which is much bigger than the ratio of bulk DMO. Thisdiscrepancy evidences the spin order modulation in thin filmsby strain and microstructure [44].

In addition, we have investigated the ferroelectricity andME coupling in TMO and GdMnO3 thin films deposited onSTO (001) substrates. For TMO films [45], the spiral spinordering induced ferroelectricity was identified too, highlyresembling the observation in bulk TMO. However, it wasidentified that the ME effect in TMO thin films finds sub-stantial contribution from the Tb3þ spin ordering far above TR,which is different from the bulk counterpart. This contributionshould be associated with the strain modified microstructure.For GdMnO3 thin films [46], the FE phase with giant polari-zation (~5000 mC/m2) and relatively high TC (~75 K) wasidentified although the bulk GdMnO3 doesn't show suchferroelectricity induced magnetically. This was discussedbased on the twin-like domain structure and the Gd3þeMn3þ

exchange striction.

5. Continuous ME control

A unique feature of the ferroelectricity associated with theMn3þ spiral spin ordering is the polarization switching fromthe c-axis to the a-axis (i.e. Pc / Pa switching) as a conse-quence of the Mn3þ spiral plane flop from the bc-plane to theab-plane. This spiral plane flop can be triggered by magnetic

field or other stimuli, i.e., done by a magnetic field above acritical value Hc1 along the b-axis. This is also the origin forthe giant ME effect. Interestingly, for bulk DMO, the polari-zation switching can also be triggered when magnetic field isapplied along the a-axis but with a higher critical fieldHc2 > Hc1, as shown in Fig. 9. The difference (Hc2�Hc1) canbe as big as ~6.0 T at T ¼ 5 K in bulk DMO [10].

For the DMO thin films, a twin-like domain structure isevidenced by a series of structural characterizations, whichallows sufficient space to explore additional strategies towardsthe magnetic control of polarization. For instance, one canintentionally make use of the distinct polarization switchingdriven by magnetic field H along certain direction. Onerepresentative example is shown in Fig. 10, where the mi-crostructures are denoted as a- and b-domains in which thearrows represent the [010] direction of the DMO orthorhombiclattice, and the angle between the in-plane H orientation andthe <110> direction of DMO thin film is denoted as 4. Based

221C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

on such a twin-like domain structure, two characteristicpathways can be expected for the polarization switchingsequence [47]. For 4 ¼ 0, magnetic field H aligns along thediagonal direction of both the a- and b-domains, and hence thepolarization P within the two types of domains can beswitched concurrently as H > √2Hc1. However, for 4 ¼ 45�,magnetic field H is parallel to the b-axis of the a-domains butperpendicular to that of the b-domains, and then two separatedpolarization switching events can be anticipated, i.e., the po-larization switching in the a-domains first occurs as H > Hc1

and then in the b-domains as H > Hc2, which are schematicallyshown in Fig. 11.

To identify the above arguments, various measurements onthe ME effect in the DMO thin films have been carried out. InFig. 12, the measured out-of-plane dielectric constant εc andpolarization Pc are plotted as a function of H. We first look atthe εc(H ) curves for both 4 ¼ 0 and 4 ¼ 45� by sweeping Hup and down in a continuous manner, noting that dielectricconstant in comparison with polarization Pc(H ) seems to bemore sensitive in detecting the phase transitions. It is revealedthat three peaked anomalies in the εc(H ) curve for 4 ¼ 45� at

Fig. 12. The H dependences of Pc and εc measured at (a) 4 ¼ 45� and (b) 4 ¼ 0 a

figure. The short bars indicate the fields HDy, Hc1, and Hc2. (d) The Hc1 as a function

(PMn2 e PDy

2 ) is also plotted for a comparison.

H ~ 1.5 T, ~2.6 T, and ~6.0 T, can be seen. However, only twoanomalies in the εc(H ) curve for 4 ¼ 0 appear at H ~ 1.5 T and~4.0 T. Obviously, the first anomaly for the two cases isassociated with the Dy3þ spin CM ↔ ICM transition driven bymagnetic field H, since the critical field coincides with that ofthe metamagnetic transition as identified in the M(H) curves(see Fig. 8). For the case of 4 ¼ 45�, the second peak aroundH ¼ Hc1 ~ 2.6 T features the Pc ↔ Pa switching of the a-domains, leading to a gradual decrease of Pc. The third peakaround H ¼ Hc1 ~ 6.0 T reflects the Pc ↔ Pa switching of theb-domains. As expected, the polarization switching wouldhappen in the a- and b-domains concurrently for 4 ¼ 0, andthe critical field should be H ¼ √2Hc1. This is well repre-sented in the experiments shown in Fig. 12(b), in which thecritical field of the second anomaly is about 4.0 T, close to√2Hc1 (i.e., ¼3.7 T).

To obtain a comprehensive understanding of the MEcoupling in DMO thin films, a series of εc(H ) curves weremeasured at various T. The data obtained at T ¼ 4 Ke18 K for4 ¼ 45� are shown in Fig. 12(c), in which the critical fieldsHR, Hc1, and Hc2 are marked as olive, red, and blue bars,

t T ¼ 5 K. (c) Measured εc(H ) curves at various T labeled at the right side of

of T evaluated from the measured dielectric data. The predicted difference term

Fig. 13. The multiferroic phase diagram of the DyMnO3 thin films with H at

4 ¼ 45�. The values of Tc were obtained from the εc(T ) data. Open and closed

symbols denote the data obtained with increasing H and decreasing H,

respectively. For a comparison, the critical fields of anomalous transitions with

H applied at 4 ¼ 0 are also shown as circles (increasing H ) and dots

(decreasing H ) in the phase diagram. The olive crosses denote the effective

field H,cos45� along the a- or b-axis.

222 C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

respectively, where HR is the field for suppressing the inde-pendent Dy3þ spin ordering. As usual, HR shifts graduallytowards the low-H side upon increasing T until TR ~9 K,beyond which no more Dy3þ CM spin order can be perse-vered. The critical field Hc1(T ) first shifts towards the low-Tside but starts to turn to the high-T side as T ~ TR. The criticalfield Hc2(T ) dependence is trivial by a gradual increasing ofHc2 over the entire T-range.

To explain the dependences Hc1(T ) and Hc2(T ), one maydiscuss the stability of the bc-plane spiral spin state against H.An increasing of H beyond HR(T ) drives the CM-ICM tran-sition of Dy3þ spins via the Dy3þeMn3þ coupling. Thistransition is compensated by the stability penalty of the Mn3þ

spiral spin state. The ME energy terms for the bc-plane Mn3þ

spiral spin state and Dy3þ ICM state, and the H-inducedmagneto-static energy can be described respectively by:

DX

Mn�SSO

¼�X< i;j>

DMn$�SMn;i � SMn;j

�¼� 1

2pεMn

P2Mn

DX

Dy�ICM

¼X< i;k>

aMn�Dy$�SMn;i � SDy;k

�¼� 1

2pεDyP2Dy

Xstatic

¼�H$X< i>

SMn ¼�H$Seff

ð1Þ

where DSMn-SSO is the variation of the ME energy associatedwith the DM mechanism for PMn, which can be expressed bythe corresponding electric static energy, DMn is the DMfactor, and εMn is the dielectric permittivity of Mn “ferro-electric sublattice” corresponding to PMn. DSDy-ICM is thevariation of the ME energy associated with the Mn3þeDy3þ

exchange striction mechanism for PR ¼ PDy, which can bereplaced by the corresponding electric static energy, aMn-Dy isthe exchange striction factor and εDy is the dielectricpermittivity of Dy “ferroelectric sublattice” corresponding toPDy. Sstatic is the magneto-static energy and Seff is theeffective moment accounting for the Mn3þ spins SMn. In azero-order approximation, the instability of the bc-planeMn3þ spiral spin state and the Pc / Pa will occur at thepoint defined by:Xstatic

¼ DX

Mn�SSO

þDX

Dy�ICM

; ð2Þ

which yields

Hc1 ¼ 1

2pSeff

�1

εMn

P2Mn �

1

εDy

P2Dy

�; ð3Þ

Equation (3) explains Hc1(T ) and Hc2(T ) in a qualitativesense, clarifying the physics of magnetic control of polariza-tion. The evaluated Hc1 values from the data in Fig. 12(c) as afunction of T are plotted in Fig. 12(d) where the predicteddifference term (PMn

2 e PDy2 ) is plotted too, revealing the

qualitative consistency between Hc1 and (PMn2 e PDy

2 ) as afunction of T. A detailed discussion about the evaluation of thePMn(T ) and PDy(T ) can be found in a previous paper [17].

A phase diagram for the DMO thin films is summarized inFig. 13, based on the above discussion. It is seen that the phasediagram of the DMO thin films shows almost identical featuresto an integration of the two phase diagrams for bulk DMO,given H//a-axis and H//b-axis respectively [10]. The diagramis divided into five regions. The region at T > Tc1 is occupiedby the paraelectric phase. In 10 K < T < 20 K, there appear inorder the FE phase with P//c-axis (Pc), the FE phase withmixed components P//c-axis and P//a-axis (Pc þ Pa), and theFE phase with P//a-axis (Pa), with increasing H. BelowT ¼ 10 K, the increasing H drives the transitions from the FEphase with P//c-axis (PMn), to the FE phase with P//c-axis(Pc ¼ PMn þ PDy), then to the FE phase with mixed compo-nents P//c-axis and P//a-axis (Pc þ Pa), and eventually to theFE phase with P//a-axis (Pa).

The above described experiments have demonstrated thatthe polarization switching in the thin films can be one step orseparated events, corresponding to 4 ¼ 0 and 4 ¼ 45�,respectively. A natural question is what would happen if H isapplied along a direction in-between 4 ¼ 0 and 4 ¼ 45�? Inthe framework of Landau theory on ferroelectricity, the freeenergy F(Pc, H ) can be schematically drawn in Fig. 14(a) for aguide of eyes. The polarization Pc ¼ 0 state corresponds to theab-plane spiral spin state where the polarization Pa reaches themaximum. Intentionally, one expects that a highly sensitivemodulation of polarization by magnetic field H should bereachable around H ~ Hc1 to √2Hc1, at which there are threecomparable free energy minimal states with relatively lowbarriers between them, so that the inter-state transitions arekinetically easy. In the experiments, the in-plane H (i.e., ¼3.0T), which is higher than Hc1 and lower than √2Hc1, is rotated,triggering the Pc ↔ Pa switching of the a-domains and b-

Fig. 14. (a) Energy diagram on ferroelectric flopping under various magnetic fields. (b) Angular dependence of Pc measured under H ¼ 3 T at T ¼ 5 K.

223C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

domains, alternatively, so that the measured Pc(4) oscillateswith the two-fold rotation symmetry. One set of Pc(4) data areplotted in Fig. 14(b). It is seen that the minimal and maximalare ~125 mC/m2 and ~250 mC/m2, consistent with the modelprediction in Fig. 11. More interesting is the continuous andperiodic modulation of macroscopic Pc. This modulation ishighly robust and can be well reproduced upon an electric fieldat different swamping frequencies. So far, no such a contin-uous modulation of polarization has been reported in eitherbulk DMO or other multiferroics of spiral spin structure [7].One is for Eu0.55Y0.45MnO3 where the rotation of a developedconical spin structure from an ab-plane spiral phase wasobserved, driven by a magnetic field perpendicular to the spinpropagation vector [48]. The present continuous response ofpolarization Pc to the in-plane rotating H is essentially deter-mined by the in-plane twin-like domain structure of the DMOthin films, while the generation of polarization PDy in DMOalso makes contributions.

6. Conclusion

In this mini-review, we have thoroughly discussed themultiferroicity and magnetoelectric coupling tailored bychemical substitution and strain/domain engineering inDyMnO3 (bulk and thin film states), which can be a modelsystem for type-II multiferroic materials. The experimentsdemonstrated the duality of multiferroicity in DyMnO3, i.e.,the ferroelectricity is generated from the symmetricDy3þeMn3þ exchange striction and antisymmetric interaction(DM interaction) among Mn3þ spins. The multiferroicitycould be modulated intentionally by means of various stimuliroutes. In particular, designable magnetoelectric controls havebeen identified in DyMnO3 thin films grown on SrTiO3 sub-strates, owing to the cooperation between the twin-like nano-domain structure and the duality of multiferroicity. Theexperimental results outlined in this review would openadditional possibilities for exploring additional multiferroic

materials with superior properties and distinctive magneto-electric control roadmaps.

Acknowledgments

This work was supported by the National Natural ScienceFoundation of China (Grant Nos. 11374112, 51332006,11374147) and the National Basic Research Program of China(973 Program) (Grant No. 2015CB654602).

References

[1] Cheong S-W, Mostovoy M. Multiferroics: a magnetic twist for ferro-

electricity. Nat Mater 2007;6:13e20.

[2] Huang W, Yang S, Li XG. Multiferroic heterostructures and tunneling

junctions. J Materiomics 2015;1:263e84.

[3] Dong S, Liu J-M, Cheong S-W, Ren ZF. Multiferroic materials and

magnetoelectric physics: symmetry, entanglement, excitation, and to-

pology. Adv Phys 2015;64:519e626.

[4] Wang KF, Liu JM, Ren ZF. Multiferroicity: the coupling between mag-

netic and polarization orders. Adv Phys 2009;58:321e448.

[5] Lu CL, Hu W, Tian Y, Wu T. Multiferroic oxide thin films and hetero-

structures. Appl Phys Rev 2015;2:021304.

[6] Khomskii D. Classifying multiferroics: mechanisms and effects. Physics

2009;2:20.

[7] Tokura Y, Seki S, Nagaosa N. Multiferroics of spin origin. Rep Prog Phys

2014;77:076501.

[8] Hill NA. Why are there so few magnetic ferroelectrics? J Phys Chem B

2000;104:6694.

[9] Sergienko IA, Dagotto E. Role of the DzyaloshinskiieMoriya interaction

in multiferroic perovskites. Phys Rev B 2006;73:094434.

[10] Kimura T, Lawes G, Goto T, Tokura Y, Ramirez A. Magnetoelectric

phase diagrams of orthorhombic RMnO3 (R¼Gd, Tb, and Dy). Phys Rev

B 2005;71:224425.

[11] Sergienko I, S‚en C, Dagotto E. Ferroelectricity in the magnetic E-phase

of orthorhombic perovskites. Phys Rev Lett 2006;97:227204.

[12] Lee N, Choi YJ, Ramazanoglu M, Ratcliff W, Kiryukhin V,

Cheong SW. Mechanism of exchange striction of ferroelectricity in

multiferroic orthorhombic HoMnO3 single crystals. Phys Rev B

2011;84:020101R.

[13] Choi YJ, Yi HT, Lee S, Huang Q, Kiryukhin V, Cheong SW. Ferro-

electricity in an ising chain magnet. Phys Rev Lett 2008;100:047601.

224 C. Lu, J.-M. Liu / J Materiomics 2 (2016) 213e224

[14] Guo YJ, Dong S, Wang KF, Liu JM. Mean-field theory for ferroelec-

tricity in Ca3CoMnO6. Phys Rev B 2009;79:245107.

[15] Lu CL, Fan J, Liu HM, Xia K, Wang KF, Wang PW, et al. An investi-

gation on magnetism, spinephonon coupling, and ferroelectricity in

multiferroic GdMn2O5. Appl Phys A 2009;96:991e6.[16] Murakawa H, Onose Y, Miyahara S, Furukawa N, Tokura Y. Ferroelec-

tricity induced by spin-dependent metal-ligand hybridization in Ba2Co-

Ge2O7. Phys Rev Lett 2010;105:137202.

[17] ZhangN,Dong S, Liu JM. Ferroelectricity generated by spin-orbit and spin-

lattice couplings in multiferroic DyMnO3. Front Phys 2011;7:408e17.

[18] Farokhipoor S, Magen C, Venkatesan S, Iniguez J, Daumont CJ, Rubi D,

et al. Artificial chemical and magnetic structure at the domain walls of an

epitaxial oxide. Nature 2014;515:379e83.

[19] Venkatesan S, Daumont C, Kooi BJ, Noheda B, Hosson JTMD. Nano-

scale domain evolution in thin films of multiferroic TbMnO3. Phys Rev B

2009;80:214111.

[20] Daumont CJ, Mannix D, Venkatesan S, Catalan G, Rubi D, Kooi BJ,

et al. Epitaxial TbMnO3 thin films on SrTiO3 substrates: a structural

study. J Phys Conden Matter 2009;21:182001.

[21] Kimura T, Goto T, Shintani H, Ishizaka K, Arima T, Tokura Y. Magnetic

control of ferroelectric polarization. Nature 2003;426:55e8.

[22] Lu CL, Dong S, Wang KF, Liu JM. Enhanced polarization and magne-

toelectric response in Tb1�xHoxMnO3. Appl Phys A 2009;99:323e31.[23] Goto T, Kimura T, Lawes G, Ramirez AP, Tokura Y. Ferroelectricity and

giant magnetocapacitance in perovskite rare-earth manganites. Phys Rev

Lett 2004;92:257201.

[24] Kimura T. Spiral magnets as magnetoelectrics. Annu Rev Mater Res

2007;37:387e413.

[25] Alonso JA, Martinez-Lope MJ, Casais MT, Fernandez-Daz MT. Evolu-

tion of the Jahn-Teller distortion of MnO6 octahedra in RMnO3 perov-

skites (R ¼ Pr, Nd, Dy, Tb, Ho, Er, Y): a neutron diffraction study. Inorg

Chem 2000;39:917e23.

[26] Dong S, Yu R, Yunoki S, Liu JM, Dagotto E. Origin of multiferroic

spiral spin order in the RMnO3 perovskites. Phys Rev B 2008;78:

155121.

[27] Prokhnenko O, Feyerherm R, Dudzik E, Landsgesell S, Aliouane N,

Chapon L, et al. Enhanced ferroelectric polarization by induced Dy spin

order in multiferroic DyMnO3. Phys Rev Lett 2007;98:057206.

[28] Schierle E, Soltwisch V, Schmitz D, Feyerherm R, Maljuk A,

Yokaichiya F, et al. Cycloidal order of 4f moments as a probe of chiral

domains in DyMnO3. Phys Rev Lett 2010;105:167207.

[29] Prokhnenko O, Feyerherm R, Mostovoy M, Aliouane N, Dudzik E,

Wolter AUB, et al. Coupling of frustrated Ising spins to the magnetic

cycloid in multiferroicTbMnO3. Phys Rev Lett 2007;99:177206.

[30] Munoz A, Casais MT, Alonso JA, Martinez-Lope MJ, Martinez JL,

Fernandez-Diaz MT. Complex magnetism and magnetic structures of the

metastable HoMnO3 perovskite. Inorg Chem 2001;40:1020.

[31] Feyerherm R, Dudzik E, Wolter AUB, Valencia S, Prokhnenko O,

Maljuk A, et al. Magnetic-field induced effects on the electric polari-

zation in RMnO3 (R¼Dy,Gd). Phys Rev B 2009;79:134426.

[32] Lin L, Li L, Yan ZB, Tao YM, Dong S, Liu JM. Ferroelectricity of

polycrystalline GdMnO3 and multifold magnetoelectric responses. Appl

Phys A 2012;112:947e54.[33] Mochizuki M, Furukawa N, Nagaosa N. Spin model of magnetostrictions

in multiferroic Mn perovskites. Phys Rev Lett 2010;105:037205.

[34] Zhang N, Guo YY, Lin L, Dong S, Yan ZB, Li XG, et al. Ho substitution

suppresses collinear Dy spin order and enhances polarization in

DyMnO3. Appl Phys Lett 2011;99:102509.

[35] Zhang N, Dong S, Zhang GQ, Lin L, Guo YY, Liu JM, et al. Multiferroic

phase diagram of Y partially substituted Dy1�xYxMnO3. Appl Phys Lett

2011;98:012510.

[36] Cui Y, Tian Y, Shan A, Chen C, Wang R. Magnetic anisotropy and

anomalous transitions in TbMnO3 thin films. Appl Phys Lett 2012;101:

122406.

[37] Marti X, Skumryev V, Ferrater C, Garcia-Cuenca MV, Varela M,

Sanchez F, et al. Emergence of ferromagnetism in antiferromagnetic

TbMnO3 by epitaxial strain. Appl Phys Lett 2010;96:222505.

[38] Rubi D, de Graaf C, Daumont C, Mannix D, Broer R, Noheda B.

Ferromagnetism and increased ionicity in epitaxially grown TbMnO3

films. Phys Rev B 2009;79:014416.

[39] Kirby BJ, Kan D, Luykx A, Murakami M, Kundaliya D, Takeuchi I.

Anomalous ferromagnetism in TbMnO3 thin films. J Appl Phys 2009;

105:07D917.

[40] Marti X, Skumryev V, Cattoni A, Bertacco R, Laukhin V, Ferrater C,

et al. Ferromagnetism in epitaxial orthorhombic YMnO3 thin films.

J Magn Magn Mater 2009;321:1719e22.[41] Qi J, Yan L, Zhou HD, Zhu JX, Trugman SA, Taylor AJ, et al. Coex-

istence of coupled magnetic phases in epitaxial TbMnO3 films revealed

by ultrafast optical spectroscopy. Appl Phys Lett 2012;101:122904.

[42] Fina I, F�abrega L, Martí X, S�anchez F, Fontcuberta J. Chiral domains in

cycloidal multiferroic thin films: switching and memory effects. Phys

Rev Lett 2011;107:257601.

[43] Fontcuberta J, Fina I, F�abrega L, S�anchez F, Martí X, Skumryev V.

Ferroelectricity and strain effects in orthorhombic YMnO3 thin films.

Phase Transit 2011;84:555e68.

[44] Lu CL, Dong S, Xia Z, Luo H, Yan Z, Wang H, et al. Polarization

enhancement and ferroelectric switching enabled by interacting magnetic

structures in DyMnO3 thin films. Sci Rep 2013;3:3374.

[45] Hu N, Lu CL, Xia Z, Xiong R, Fang P, Shi J, et al. Multiferroicity and

magnetoelectric coupling in TbMnO3 thin films. ACS Appl Mater In-

terfaces 2015;7:26603e7.[46] Li X, Lu CL, Dai J, Dong S, Chen Y, Hu N, et al. Novel multiferroicity in

GdMnO3 thin films with self-assembled nano-twinned domains. Sci Rep

2014;4:7019.

[47] Lu CL, Deniz H, Li X, Liu JM, Cheong SW. Continuous magnetoelectric

control in multiferroic DyMnO3 films with twin-like domains. Sci Rep

2016;6:20175.

[48] Murakawa H, Onose Y, Kagawa F, Ishiwata S, Kaneko Y, Tokura Y.

Rotation of an electric polarization vector by rotating magnetic field

in cycloidal magnet Eu0.55Y0.45MnO3. Phys Rev Lett 2008;101:

197207.

Dr. Chengliang Lu is an associate professor of

material science of School of Physics, Huazhong

University of Science and Technology, Wuhan,

China. He received Ph.D. in 2009 from the Nanjing

University. He worked in Nanyang Technological

University as a postdoc from 2009 to 2010, and in

Max-Planck Institute of Microstructure Physics as a

Humboldt Research Fellow from 2013 to 2015. His

recent research focuses on multiferroic materials

and 5d iridates. Email: cllu@mail.hust.edu.cn

Dr. Jun-Ming Liu is a Professor of Physics with

Department of Physics and Laboratory of Solid State

Microstructures, Nanjing University. He obtained his

Pd.D degree from Northwestern Polytechnical Uni-

versity in 1989 before joining the Department of

Physics, Nanjing University. He specializes in

correlated quantum materials and computational

materials sciences. He has co-authored more than

600 technical papers in international referred jour-

nals. More information can be found at http://pld.

nju.edu.cn. Email: liujm@nju.edu.cn.