pss standard-reprint webversionstability of 71-stripe domains in epitaxial bifeo 3 films upon...

reprint

Phys. Status Solidi B 249, No. 11, 2278–2286 (2012) / DOI 10.1002/pssb.201248329 p s sb

statu

s

soli

di

www.pss-b.comph

ysi

ca

basic solid state physics

Stability of 71- stripe domains inepitaxial BiFeO3 films upon repeatedelectrical switching

Florian Johann**,1, Alessio Morelli1, and Ionela Vrejoiu*,1,2

1Max Planck Institute of Microstructure Physics, Weinberg 2, 06120 Halle, Germany2Max Planck Institute for Solid State Research, Heisenbergstr. 1, 70569 Stuttgart, Germany

Received 22 June 2012, accepted 20 July 2012

Published online 21 August 2012

Keywords BiFeO3, domain walls, ferroelectric domains, multiferroics, polarization switching

*Corresponding author: e-mail [email protected], Phone: þ49 711 689 1574, Fax: þ49 711 689 1796** e-mail [email protected]

The 718 stripe domain patterns of epitaxial BiFeO3 thin films

are frequently being explored to achieve new functional

properties, dissimilar from the BiFeO3 bulk properties. We

show that in-plane switching and out-of-plane switching of

these domains behave very differently. In the in-plane

configuration the domains are very stable, whereas in the

out-of-plane configuration the domains change their size and

patterns, depending on the applied switching voltage frequency.

� 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1 Introduction Recently particular multiferroicdomains patterns and multiferroic domain wall (DW)properties have moved into the focus of intense research. Ithas been observed that these domain patterns and DWs canhave their own functional properties, dissimilar to the bulkdomain properties [1]. In thin films the fraction of DWscompared to the filmvolume can become relatively high, andtherefore the overall functional properties of the filmmay bedominated by the DW properties. Taking into account thatDWs are in general mobile and can be manipulated forinstance by an electric field, particular domain patterns maybecome promising for active memory devices [2, 3]. As anexample 1098 and 718 stripe domain patterns of multiferroicBiFeO3 (BFO) have been investigated intensively, dueto their unique properties. BFO has a rhombohedralstructure with the direction of the ferroelectric polarizationalong the pseudocubic [111]C direction. As a result of the

rhombohedral symmetry, the energetically favorabledomain configurations in defect-free epitaxial BFO filmsconsist of stripes, which are made of either 718 or 1098DWs [4]. In the last years there have been several reportsabout novel functionalities exhibited by 718 and 1098 DWs.Seidel et al. [5, 6] demonstrated that native 1098 DWs and1808 DWs are conductive, although the bulk of thesedomains is insulating. Although the 718DWswhere found tobe insulating by Seidel et al., conductivity has been recentlyreported for 718 DWs of BFO films grown under differentgrowth parameters [7]. Recently it has even been shownthat 1098 DWs exhibit a magnetoresistance effect [8].Another peculiar property of a particular 718 stripe domainconfiguration in BFO films is related to an above band gapphotovoltaic effect [9]. One of themain driving forces for theintensive research on multiferroics in the last decades is thatthey hold promise for magnetoelectric effects that are


Phys. Status Solidi B 249, No. 11 (2012) 2279

Original

Paper

interesting for device applications. Heron et al. demonstratedthat room temperature electrical switching of the polariz-ation of 718stripe domains of epitaxial BFO films canbe used to reverse the magnetization of a CoFe layer in aCoFe/BFO heterostructure [10]. In this case, on one hand,not the value of the polarization of individual ferroelectricdomains is relevant, but the yielded net in-plane valueacross the 718stripe domain pattern is utilized. Onthe other hand, the switching of the polarization ofthe individual domains is coupled to the rotation of theantiferromagnetic order parameter of the BFO in thesame and this is further coupled to the magnetization ofthe CoFe top layer, leading ultimately to the switching of theCoFe magnetization.

The reliability of switching of theses stripes is animportant issue if BFO is considered to be used in devicefabrication. However, not many studies have been reportedso far regarding domain pattern stability. Shafer et al. [11]reported on the possibility of studying the polarizationswitching between SrRuO3 (SRO) bottom in-plane electro-des on a BFO/DyScO3(110)O sample whose BFO film had718 stripe domains. Balke et al. [12] reported on the in-planeswitching of [110]C-oriented BFO on SrTiO3(110)C, whereswitching between only two domain states was investigated.Recently, Lee et al. [13] investigated the in-plane switchingof electrically leaky BFO films deposited on SrTiO3(001)and correlated the domains with local photoconductivemeasurements. Folkman et al. [14] investigated theelectrical switching characteristics of 718stripe domainsand the effects of defect-dipoles on a BFO film grown onTbScO3(110)O, using interdigitated gold electrodes on topof the BFO film.

However, there are no reports about the stability ofthese domains upon many electrical switching cyclesand at different frequency of the applied voltage pulses.Here, we investigate the stability of 718 stripe domains inepitaxial thin BFO films for two configurations: afterin-plane switching by applying an electrical field betweenin-plane electrodes aligned parallel to the 718 stripes(Fig. 1b) and after out-of-plane switching by applying thefield perpendicular to the BFO film in plan-parallelcapacitors (Fig. 3a). The latter configuration is preferablefor applications in high density data storage devices.We show that the two electrical switching configurationshave different behavior. For the in-plane configurationthe 718 stripe domains can be switched repeatedly formany cycles at frequencies up to 100 kHz, resulting inonly minor changes of the stripe domain pattern. Forthe out-of-plane configuration, however, a marked depen-dence on the applied field frequency is observed. Atfrequencies below 1 kHz the 718 stripe domains aremaintained, nevertheless the domain pattern of the filmchanges to stripes with larger widths compared to theas-grown state. For frequencies above 1 kHz the stripescan either be completely destroyed or the areas with samenet-polarization break into smaller areas depending on thestrain state and defect density.

www.pss-b.com

2 Experimental SrTiO3(100)C (STO) substrates werepretreated by etching in a bufferedHF solution and annealingat 950 8C for 2 h. DyScO3(110)O (DSO) substrates wereprepared in two different ways: either annealed in O2

atmosphere at 1250 8C for 3 h and used for growing BiFeO3

films for the in-plane switching configuration, or annealed inair at 1200 8C for 2 h and used for the out-of-plane switchingconfiguration. All substrates were atomically flat afterannealing. SRO and BFO were deposited by pulsed laserdeposition (PLD) at 650 8C in 0.14mbar O2 with 5Hz laserrepetition rate and a growth rate of 5 and 1.5 nmmin�1,respectively. After deposition the films were cooled to roomtemperature in 200mbar O2. Copper and gold top electrodeswere deposited ex situ by thermal evaporation through ashadow mask: for the in-plane configuration two 1mm longelectrodes were evaporated parallel to the stripe directionwith a gap of 20mm and for the out-of-plane configurationsquare shaped electrodes of size 50� 50mm2. Switchingcurrents were measured with an Aixxact TF2000 Analyzer.The voltage pulses for cycling the films were applied by aTektronix AFG310 pulse generator, amplified if needed bythe TF Analyzer amplifier. The rise time of the square pulseswithout amplifier were 75 ns, with amplifier about 200 ns.For the out-of-plane switching configuration the copperelectrodes were etched with a diluted (NH4)2S2O8 solutionand the gold electrodes were etched with gold etchingsolution from Alfa Aesar. Piezoresponse force microscopy(PFM) measurements were performed on a XE-100 Parkscanning probe microscope in ambient conditions. Through-out the paper a bright (dark) contrast corresponds in verticalPFM to a polarization pointing downwards (upwards) and inlateral PFM to a polarization pointing to the right hand side(left hand side) of the image. XRDmeasurements were donewith a Philips X’Pert diffractometer.

3 Results and discussion For the in-plane configur-ation, 50 nm thick BFO films were grown directly onDyScO3(110)O (DSO) substrates that were annealed inoxygen atmosphere before the deposition. The as-growndomain structure consists of 718 stripe domains along theorthorhombic [001]O direction of the substrate, with a netpolarization pointing out-of-plane towards the film surfaceand in-plane along [110]O substrate direction (see sche-matics in Fig. 1a and b). The direction of the polarizationcould be revealed by vertical PFM (VPFM) and lateralPFM (LPFM) imaging (see Fig. 1d and e, respectively). Ithas been reported that annealing of the DSO substrates in O2

atmosphere at elevated temperatures prior to depositingBFOcan lead to formation of two ferroelastic domain variantswith 718 stripe domain patterns in epitaxial BFO films [9].However, the formation has not been studied much so far. Itshould be noted that the surface morphology (Fig. 1c) is verysimilar to BFO films with 1098 stripe domains, which areobtained for BFO films grown on DSO substrates annealedin air (compare with Ref. [15]). Here, for our BFO film with718 stripe domains, the VPFM image in Fig. 1d reveals still afew small 1098 stripes domains, which are aligned with the


2280 F. Johann et al.: Stability of 718 stripe domains in epitaxial BiFeO3 filmsp

hys

ica ssp st

atu

s

solid

i b

(a)

(c) (d) (e)

(b)

Figure 1 (online color at: www.pss-b.com)(a) Schematics of the orthorhombic DyScO3

unit cell with two of the four monoclinic dis-torted perovskite cubes inside. (b) Top-viewschematics of the BiFeO3/DyScO3(110) sam-ple with 718 stripe domain orientation andelectrode alignment. (c) Topography, (d)VPFM phase, and (e) LPFM signal imagesof the as-grown state of the BiFeO3/DyScO3(110) sample. All images are8� 8mm2. Inset in (e) shows a scheme ofthe in-plane domain configuration.

features in the morphology and perpendicular to the 718stripe domains (short segments along [110]O with whitecontrast). Furthermore, for our BFO film with only 1098stripe domains on DSO substrate annealed in air [15], thestripes are as well aligned along the [110]O axes, exactly asthe few 1098 stripe domains in the film with predominantly718 stripe domains discussed here. From XRD measure-ments it can be deduced that for both samples the samestructural variants are present (see Fig. S1 in the SupportingInformation, online at www.pss-b.com). This suggests thatthe DSO(110)O substrate imposes in both cases the samestructural/ferroelastic variants. If we consider one pseudo-cube of the DSO substrate, with its structure beingmonoclinic(Fig. 1a), the ferroelastic BFO domains are arranged in such away as to adapt to the monoclinic distortion of the DSO [15].The perpendicular orientation between the 1098 stripedomains and the 718 stripe domains, which are built upfrom the same structural variants, is in agreement with thetheoretical consideration by Streiffer et al. (see Figs. 2 and 4in Ref. [4]). Whether the 718 or 1098 stripe domains areformedmight be imposed by the DSO termination. It has beenreported that the termination of the DSO substrate can beinfluenced by the atmosphere during annealing [16], althoughother reports claim that for perfect single terminated DSOsurfaces also a selective wet etching has to be applied [17].Moreover, the resulting single-terminated DyScO3(110)Osurfaces are polar, being negatively charged if ScO�

2

terminated, and positively charged if DyOþ terminated [17].In epitaxial rhombohedral ferroelectric films in general and


in BFO films in particular, 718 and 1098 stripes compete andthe energetically more favorable configuration will form,dependingonseveral parameters suchas thefilm thickness andthe electrostatic boundary conditions [18, 19]. It is reasonableto assume that the different DSO terminations given by thedifferent annealing procedures may affect the electrostaticenergy terms, in particular the depolarization energy, and as aresult stabilize different stripe domain patterns. Consideringthat the polarization of our 718 stripe domain BFO film onDSO substrate is pointing upwards, it might be inferred thatour DSO substrate after O2 annealing is mostly DyOþ

terminated.The net in-plane polarization can be switched by

applying a voltage to the in-plane electrodes. To investigatethe domain development after switching, PFM wasperformed in an area between the electrodes on the as-grown BFO film and in the same area after repeatedswitching cycles. Figure 2a shows the LPFM image of theas-grown state and Fig. 2b the LPFM image after applyingone unipolar rectangular pulse of þ200V for 5ms. Thepolarization of each stripe domain switched by 718, resultingin a total 1808 switching of the net in-plane polarization, asalready reported [10]. After 5� 105 switching cycles, whichwere performed by applying square pulses of �200V at100 kHz, the domain pattern hardly changed. Only at a fewplaces, where the stripes were interrupted in the as-grownstate of the BFO film, some small changes appeared, leadingto partial removal of the initial interruptions, as can be seen inthe comparison of the domain images in Fig. 2c.

www.pss-b.com


Original

Paper

(a)(d)

(e)

(b)

(c)

Figure 2 (online color at: www.pss-b.com) (a) LPFM image of as-grown state between the in-plane electrodes, (b) LPFM image afterswitching the net polarization with a unipolar voltage pulse, and (c) LPFM image after 5� 105 complete switching cycles with the domainwalls of the as-grown state superimposed as black lines. The green circle indicates one of the few changes of the domain pattern. All imagesare 5� 2.5mm2. (d) Switching current and integrated polarization of an in-plane hysteresismeasured at 1 kHz andRTafter 1.7� 106 cycleson a second electrode. The capacitive charging contribution was subtracted from the polarization curve. (e) Development of the coercivevoltages and imprint with number of cycles measured on the same electrode.

The switching of the net in-plane polarization can beinvestigated by measuring also macroscopic in-planeswitching currents. Figure 2d shows the switching currentscollected for a hysteresis measurement performed at 1 kHz atroom temperature after the BFO film was switched for1.7� 106 cycles. Assuming an effective polarization for thisconfiguration of Peff¼Q/(lh), with Q being the measuredcharge, l the electrode length (1mm) and h the film thickness(50 nm) [14], values of�50mCcm�2 for the polarization ofBFO along [110]O are measured. The coercive voltages arevery asymmetric and the resulting hysteresis loop for thepolarization is thus strongly imprinted, indicating a built-infield aligned parallel to the direction of the as-grown netpolarization. Imprint effects were attributed to defect dipolesfor BFO films having 718 stripes domains, deposited onTbScO3(110)O [14]. Figure 2e shows the development of thecoercive voltages and imprint with increasing cycle number.For the in-plane configuration, the imprint decreases after1.7� 106 cycles to �66% of the original value. In addition,the coercive voltage for both polarities drops with increasingnumber of cycles as well. This improvement with switchingcycles is an advantageous behavior with respect to deviceapplications.

So far the conclusion is that in case of the in-planeswitching configuration the domain pattern is very stablewith improving electrical switching properties. However,in case of the out-of-plane electrical switching a steadytransition of the as-grown stripe domain pattern to a new

www.pss-b.com

pattern occurs upon several switching cycles. For thisconfiguration, 150 nm thick BFO films were grown onSRO-coated DSO(110)O and SRO-coated STO(100)C.Copper top electrodes were thermally evaporated on topof the BFO films to form plan-parallel capacitors. Weperformed similar experiments with gold top electrodes andthe results are summarized in Fig. S2 in the SupportingInformation. For BFO films on both types of SRO-coatedsubstrates, the as-grown domain structure consists of 718stripe domains. The schematic in Fig. 3a describes theorientation of the 718 domains with respect to the DSOsubstrate and its surface terrace orientation. The VPFMimages reveal that the polarization is pointing downwards tothe SRO interface for both type of substrates (Figs. 3band 4a), as often reported for BFO films deposited on SRO-coated substrates [15, 20, 21]. There are minor exceptionswhere some dark curved lines with polarization pointingupwards exist. These seem to have charged domain walls, asdetailed in the Supporting Information (Fig. S4). Thesedomains were seen by us before in BFO films grown onSRO/DSO and they seemed to be quite stable [22]. Theyappeared again after annealing the samples in vacuum at350 8C, subsequent to a chemical switching of the filmpolarization by oxygen plasma treatment. This points to apossible role played by oxygen vacancies in the stabilizationof these domains. The LPFM image for the BFO film onSRO/DSO (Fig. 3c) shows that, unlike the 718 stripedomains of the BFO film grown directly on DSO, the 718



hys

ica ssp st

atu

s

solid

i b

(a)

(d)

(e)

(f)

(b) (c)

Figure 3 (online color at: www.pss-b.com) (a) Schematics of the Cu/BiFeO3/SrRuO3/DyScO3(110) sample. (b) VPFM phase and(c) LPFM signal of as-grown state of the BiFeO3/SrRuO3/DyScO3 sample. The arrows indicate the net in-plane polarization in the darkand bright regions. (d) LPFM images of capacitors which have been cycled 5000 times at different frequencies, and (e) capacitors cycledat 0.1 kHz with different number of cycles. All images are 8� 8mm2. (f) Average stripe domain width as a function of the number ofswitching cycles at 0.1 kHz, measured for a BiFeO3 film on SrRuO3/DyScO3(110).

stripes are nowaligned along the [110]O direction, so they are908 rotated. Additionally, there are regions with different netin-plane polarization directions, which are very oftendivided by the dark curved lines seen in the VPFM image.It can be deduced from LPFM images that at these lines thefilm often forms charged domain walls (see the domainanalysis given in Fig. S4), however the PFM-measured


domain pattern is too complex to allow an unambiguousassignment. TEM studies revealed as well that chargeddomain walls can form in epitaxial BFO films, either in theas-grown state [23] or after electrical switching [24].

In the case of the BFO film on SRO/STO substratethere is predominately a single stripe direction with one netin-plane polarization direction. The direction of the stripes

www.pss-b.com


Original

Paper

Figure 4 (online color at: www.pss-b.com) (a) VPFM phase and(b) LPFM signal of as-grown state of the BiFeO3/SrRuO3/SrTiO3(001) sample. The arrow in (b) indicates the direction ofthe net in-plane polarization. (c) LPFM images of Cu/BiFeO3/SrRuO3/SrTiO3 capacitors that have been cycled 5000 times atdifferent frequencies and (d) capacitors cycled at 0.1 kHz withdifferent number of cycles. After 10,000 cycles the direction ofnet in-plane polarization is inverted compared to the as-grown state.All images are 6� 6mm2.

of BFO films grown on both substrates coated with SROmight be due to the epitaxial orientation and structure ofthe SRO layer, which is induced by the terraces of thesubstrate [25–27].

Because lateral domain imaging was not possiblethrough the top metal electrode, the electrode had to beetched away after the switching cycles. As a consequence,only the state after the cycling and removal of the electrodecan be observed by PFM. The development of the domains inthe same area of the BFO film after successive switchingcycles cannot be directly observed. We had to investigateseveral electrodes that were used to switch the film withincremental number of cycles, subsequently etching theelectrode away and then investigating the area underneath byPFM. The polarization switching of both Cu/BFO/SRO/DSO and Cu/BFO/SRO/STO capacitors shows importantfrequency dependence. The BFO film on SRO/STO hasmuch stronger frequency dependence than the BFO film onSRO/DSO. Figure 3d shows LPFM images taken in the areaswhere Cu/BFO/SRO/DSO capacitors have been cycled 5000times at different cycle frequencies. At all frequencies therestill exist 718 boundaries, but the domain pattern has changedcompared to the as-grown state. First, in the correspondingVPFM images (not shown here) a completely uniform

www.pss-b.com

contrast is seen with the out-of-plane polarization pointingto the SRO interface. The narrow dark lines belonging todomains of opposite out-of-plane polarization (Fig. 3b)disappeared after the capacitor cycling. One reason for theirdisappearance might be that defects, for example, oxygenvacancies, cation defects [13, 24], or cation impurities,which were produced during growth and were probablyresponsible for the formation of the curved narrow domainswith different polarization in the first place, are mobileand redistribute during the switching cycles [24]. Secondly,for the lower frequencies of 0.1–1 kHz, the areas withsame net in-plane polarization grew laterally, whereas forhigher switching frequencies these areas broke up into tinyareas.

Figure 3e shows the development of the stripe domainswith the number of cycles at 0.1 kHz and Fig. 3f the averagestripe width which has been extracted from these LPFMimages. It can be seen that the average stripe width increasedto the double of the as-grown value within the first 1500–2000 cycles and remained stable afterwards. However, theareas with same net in-plane polarization kept on increasinguntil the maximum performed switching cycles.

Figure 4 summarizes the investigations performed onCu/BFO/SRO/STO capacitors for which the frequencydependence of the stripe domains upon repeated switchingwas more dramatic. Figure 4a and b display VPFM andLPFM images of the as-grown state of the BFO film grownon SRO/STO. Figure 4c shows the LPFM images ofCu/BFO/SRO/STO capacitors that were cycled 5000 timesat different frequencies. At 0.1 kHz the 718 stripe domainpattern was maintained. For the low frequency, the resultingstripe domain width also doubled, very similar to theCu/BFO/SRO/DSO capacitors. In addition, for Cu/BFO/SRO/STO capacitors the direction of the net in-planepolarization was reversed after cycling, with respect to theas-grown state. From Fig. 4d it can be seen that for 0.1 kHzswitching a similar number of pulses as for theDSO substratewas needed for the transition to the new domain state.Moreover, already for switching at 1 kHz the stripes began tobreak up, whereas at 10 kHz small disordered mosaic-likelateral domains formed.

Figure 5 shows the electrical switching current measure-ments for the out-of-plane configuration performed on BFOfilms on both DSO and STO substrates. The voltage wasapplied to the top Cu electrode while the bottom SROelectrode was grounded. In the as-grown state (green curvesin Fig. 5) the coercive fields were imprinted, indicating abuilt-in field pointing to the bottom electrode, parallel to theas-grown out-of-plane component of the net polarization. Inaddition, for negative voltages the leakage current increasedwith cycling. After 1.1� 104 cycles, the coercive voltageincreased at both switching frequencies (0.1 and 10 kHz), butwas more pronounced for the film on SRO/STO substrate.The magnitude of the leakage current improved for 0.1 kHzcycling, while for 10 kHz it worsened.

Figure 5b and d show the development of the coercivevoltage and the imprint of BFO films grown on SRO-coated



hys

ica ssp st

atu

s

solid

i b

Figure 5 (online color at: www.pss-b.com) Measurement of the switching current at 1 kHz and room temperature for the sample(a) Cu/BiFeO3/SrRuO3/DyScO3(110) and (c) Cu/BiFeO3/SrRuO3/SrTiO3(100) with the as-grown hysteresis (green) and after1.1� 104 cycles with 0.1 kHz (red) and 10 kHz (black). The development of the positive coercive voltage VCþ and the negativecoercive voltage VC� as well as the imprint is plotted for the film on (b) DyScO3(110) substrate and (d) SrTiO3(100) substrate.

DSO and STO substrates, switched at 0.1 and 10 kHz. It canbe seen that, although the coercive voltages increased, theimprint stayed unchanged and their dependence on voltagepulse frequency was marginal.

It was established for other ferroelectric thin films, suchas Pb(Zr,Ti)O3 (PZT), that the alignment of defect-dipolescan strongly impact several material properties. Forexample, it can lead to enhanced voltage shifts, that is,imprint. It was proposed that the net polarization determinesthe spatial location of the asymmetrically trapped chargesthat are the cause for the voltage shifts [28].

The in-plane and the out-of-plane switching configur-ations show clear differences in the stability upon repeatedswitching at all frequencies. To understand the differencesit is worth to have a closer look at the two configurations.The ratio between the electrode area, which should beproportional to the amount of nucleation centers, andthe switched volume is more than two orders of magnitudehigher for the out-of-plane electrode configuration. There-fore, in the out-of-plane configuration the BFO film will beswitched in different areas of the electrode independentlyand then these areas have to merge together. If unfavorable


domain configurations meet, the system needs to relax to amore favorable state via ferroelastic switching, which needsa certain relaxation time [29], and thus leads to the strongerfrequency dependence of the out-of-plane configuration.

Defects, especially the oxygen vacancies, may play amajor role in the frequency dependence of the switching.Nelson et al. [24] showed by in situ TEM investigations ofswitching of a BFO lamella by applying voltage between aneedle and a planar bottom electrode that charged domainwalls formed during the switching process and orderedplanes of oxygen vacancies formed and acted as pinningcenters. Moreover, cation defects, such as Bi substitutions ofFe were detected in the vicinity of the 1808 domain wallsformed during the in situ TEM switching experiment.Lee et al. [13] showed that the oxygen vacancies moveand redistribute during in-plane switching of epitaxial BFO.The effects of the oxygen vacancy motion and the formationof defect-dipoles in ferroelectric perovskites were exten-sively studied for other ferroelectrics, such as BaTiO3 andPZT [28, 30]. Oxygen vacancies and the defect-dipolesthey may form were correlated with the degradation of theelectrical resistance and also, indirectly, with the imprint of

www.pss-b.com


Original

Paper

the ferroelectric capacitors, by affecting the potential wellsfor trapped electrons. It was also shown by Kim et al. [22]that oxygen plasma exposure of as-grown BFO films led toswitching the out-of-plane direction of the polarization. Asecond step of annealing at 350 8C in vacuum could recoverthe initial direction of the polarization, thus indicating theimportant role of oxygen content in the stabilization ofpreferential orientation of polarization in BFO films grownon SRO/DSO(110). Additional evidence that defects play animportant role is that another BFO film grown on SRO/DSO,which hasmore lines of opposite polarization in the as-grownstate, that is most likely more defects, exhibited strongerfrequency dependence and degradation of the stripe domains(see Fig. S3 in the Supporting Information).

The substrate on which the BFO film is deposited islikely to influence the switching behavior. Although forCu/BFO/SRO/STO a similar amount of dark domains withopposite polarization exists in the as-grown state comparedto the Cu/BFO/SRO/DSO sample, the frequency depen-dence is more dramatic for the former. The STO substratesubjects the BFO film to a higher compressive strain than theDSO substrate, due to the larger in-plane lattice mismatch.From X-ray diffraction reciprocal space mapping it wasseen that the BFO film on SRO/STO is relaxed, whereasthe BFO film on SRO/DSO is still almost completelycoherently strained (see Fig. S1c and S1d in the SupportingInformation). The relaxation is usually accomplished byformation of misfit dislocations. We expect therefore alarger density of misfit dislocations to form in the BFO filmgrown on SRO/STO substrate, due to the larger in-planemismatch.

4 Conclusions Summarizing, we investigated theswitching of 718 stripe domains in capacitor structures, bothin in-plane configuration and in out-of-plane configuration.The two configurations show very different behavior forboth the macroscopic electrical switching characteristics aswell as the 718 stripe domain pattern stability. For thein-plane configuration the stripes are very stable for manyswitching cycles and independent of the investigatedswitching frequencies, up to 100 kHz. The coercive voltagesand the imprint decreased with increasing number ofswitching cycles. For the out-of-plane configuration changesof the domain pattern occurred upon repeated switchingwithin the first 2000–5000 switching cycles. For frequencieshigher than 1 kHz the stripes were either completelydestroyed or broke into smaller areas, depending on thestrain state, defect type and defect density present in the film.For cycling at 0.1 kHz the 718 stripes were maintained,however the stripe width doubled with respect to the as-grown state. The coercive voltage increased with increasingnumber of cycles and the imprint of the hysteresis loops,however, stayed constant. The electrical resistance degra-dation improved for cycling at 0.1 kHz. The switchingbehavior and the limitations of the ferroelectric/ferroelasticBiFeO3 domains unraveled here are very important in view

www.pss-b.com

of employing BiFeO3 films with 718 stripe domain patternsin magneto-electric devices [10].

Acknowledgements Financial support by the GermanScience Foundation in the framework of SFB 762 is gratefullyacknowledged.We thankDietrichHesse for a careful reading of thismanuscript.

References

[1] G. Catalan, J. Seidel, R. Ramesh, and J. F. Scott, Rev. Mod.Phys. 84, 119 (2012).

[2] E. K. H. Salje and H. Zhang, Phase Transit. 82, 452 (2009).[3] E. K. H. Salje, ChemPhysChem 11, 940 (2010).[4] S.K. Streiffer, J. Appl. Phys. 83, 2742 (1998).[5] J. Seidel, L.W. Martin, Q. He, Q. Zhan, Y.H. Chu, A. Rother,

M. E. Hawkridge, P. Maksymovych, P. Yu, M. Gajek,N. Balke, S. V. Kalinin, S. Gemming, F. Wang, G. Catalan,J.F. Scott, N. A. Spaldin, J. Orenstein, and R. Ramesh, NatureMater. 8, 229 (2009).

[6] J. Seidel, P. Maksymovych, Y. Batra, A. Katan, S.-Y. Yang,Q. He, A. P. Baddorf, S. V. Kalinin, C.-H. Yang, J.-C. Yang,Y.-H. Chu, E. K. H. Salje, H. Wormeester, M. Salmeron, andR. Ramesh, Phys. Rev. Lett. 105, 197603 (2010).

[7] S. Farokhipoor and B. Noheda, Phys. Rev. Lett. 107, 127601(2011).

[8] Q. He, C.-H. Yeh, J.-C. Yang, G. Singh-Bhalla, C.-W.Liang, P.-W. Chiu, G. Catalan, L. W. Martin, Y.-H. Chu,J. F. Scott, and R. Ramesh, Phys. Rev. Lett. 108, 067203(2012).

[9] S. Y. Yang, J. Seidel, S. J. Byrnes, P. Shafer, C. H. Yang,M. D. Rossell, P. Yu, Y.-H. Chu, J. F. Scott, J. W. Ager,L. W. Martin, and R. Ramesh, Nature Nanotechnol. 5, 143(2010).

[10] J. T. Heron, M. Trassin, K. Ashraf, M. Gajek, Q. He, S. Y.Yang, D. E. Nikonov, Y.-H. Chu, S. Salahuddin, andR. Ramesh, Phys. Rev. Lett. 107, 217202 (2011).

[11] P. Shafer, F. Zavaliche, Y. H. Chu, P. L. Yang, M. P. Cruz,and R. Ramesh, Appl. Phys. Lett. 90, 202909 (2007).

[12] N. Balke, M. Gajek, A. K. Tagantsev, L. W. Martin, Y.-H.Chu, R. Ramesh, and S. V. Kalinin, Adv. Funct. Mater. 20,3466 (2010).

[13] W.-M. Lee, J. H. Sung, K. Chu, X. Moya, D. Lee, C.-J. Kim,N. D. Mathur, S.-W. Cheong, C.-H. Yang, andM.-H. Jo, Adv.Mater. 24, OP49 (2012).

[14] C. M. Folkman, S. H. Baek, C. T. Nelson, H. W. Jang,T. Tybell, X. Q. Pan, and C. B. Eom, Appl. Phys. Lett.96, 052903 (2010).

[15] F. Johann, A. Morelli, D. Biggemann, M. Arredondo, andI. Vrejoiu, Phys. Rev. B 84, 094105 (2011).

[16] R. Dirsyte, J. Schwarzkopf, G. Wagner, R. Fornari,J. Lienemann, M. Busch, and H. Winter, Surf. Sci. 604,L55 (2010).

[17] J. E. Kleibeuker, B. Kuiper, S. Harkema, D. H. A. Blank,G. Koster, G. Rijnders, P. Tinnemans, E. Vlieg, P. B. Rossen,W. Siemons, G. Portale, J. Ravichandran, J. M. Szepieniec,and R. Ramesh, Phys. Rev. B 85, 165413 (2012).

[18] C. W. Huang, L. Chen, J. Wang, Q. He, S. Y. Yang, Y. H.Chu, and R. Ramesh, Phys. Rev. B 80, 140101 (2009).

[19] C.W. Huang, Z. H. Chen, J. Wang, T. Sritharan, and L. Chen,J. Appl. Phys. 110, 014110 (2011).



hys

ica ssp st

atu

s

solid

i b

[20] Y.-H. Chu, Q. Zhan, L. Martin, M. Cruz, P.-L. Yang,G. Pabst, F. Zavaliche, S.-Y. Yang, J.-X. Zhang, L.-Q. Chen,D. Schlom, I.-N. Lin, T.-B. Wu, and R. Ramesh, Adv. Mater.18, 2307 (2006).

[21] Y. H. Chu, Q. He, C. H. Yang, P. Yu, L. W. Martin, P. Shafer,and R. Ramesh, Nano Lett. 9, 1726 (2009).

[22] Y. Kim, I. Vrejoiu, D. Hesse, and M. Alexe, Appl. Phys. Lett.96, 202902 (2010).

[23] Y. Qi, Z. Chen, C. Huang, L. Wang, X. Han, J. Wang,P. Yang, T. Sritharan, and L. Chen, J. Appl. Phys. 111,104117 (2012).

[24] C. T. Nelson, P. Gao, J. R. Jokisaari, C. Heikes, C. Adamo,A. Melville, S.-H. Baek, C. M. Folkman, B. Winchester,Y. Gu, Y. Liu, K. Zhang, E. Wang, J. Li, L.-Q. Chen,C.-B. Eom, D. G. Schlom, and X. Pan, Science 334, 968(2011).


[25] Q. Gan, R. A. Rao, C. B. Eom, L. Wu, and F. Tsui, J. Appl.Phys. 85, 5297 (1999).

[26] A. Vailionis, W. Siemons, and G. Koster, Appl. Phys. Lett.93, 051909 (2008).

[27] C. J. M. Daumont, S. Farokhipoor, A. Ferri, J. C. Wojdeł,J. Iniguez, B. J. Kooi, and B. Noheda, Phys. Rev. B 81,144115 (2010).

[28] W. Warren, G. Pike, D. Dimos, K. Vanheusden, H. Al-Shareef,B. Tuttle, R. Ramesh, and J. Evans, Mater. Res. Soc. Symp.Proc. 433, 257 (1996).

[29] S. H. Baek, H. W. Jang, C. M. Folkman, Y. L. Li,B. Winchester, J. X. Zhang, Q. He, Y. H. Chu, C. T. Nelson,M. S. Rzchowski, X. Q. Pan, R. Ramesh, L. Q. Chen, andC. B. Eom, Nature Mater. 9, 309 (2010).

[30] W. L. Warren, K. Vanheusden, D. Dimos, G. E. Pike, andB. A. Tuttle, J. Am. Ceram. Soc. 79, 536 (1996).

www.pss-b.com

pss standard-reprint webversionstability of 71-stripe domains in epitaxial bifeo 3 films upon...

Documents