ARTICLE
Received 26 Oct 2016 | Accepted 28 Mar 2017 | Published 5 Jun 2017
T. Honda1,2,*, J.S. White3,*, A.B. Harris4, L.C. Chapon5, A. Fennell3, B. Roessli3, O. Zaharko3, Y. Murakami2,M. Kenzelmann6 & T. Kimura1,w
Despite remarkable progress in developing multifunctional materials, spin-driven ferroelectrics featuring both spontaneous magnetization and electric polarization are still rare. Among such ferromagnetic ferroelectrics are conical spin spiral magnets with a simultaneous reversal of magnetization and electric polarization that is still little understood. Such materials can feature various multiferroic domains that complicates their study. Here we study the multiferroic domains in ferromagnetic ferroelectric Mn2GeO4 using neutron diffraction, and show that it features a double-Q conical magnetic structure that, apart from trivial 180o commensurate magnetic domains, can be described by ferromagnetic and ferroelectric domains only. We show unconventional magnetoelectric couplings such as the magnetic-eld-driven reversal of ferroelectric polarization with no change of spin-helicity, and present a phenomenological theory that successfully explains the magnetoelectric coupling. Our measurements establish Mn2GeO4 as a conceptually simple multiferroic in which the magnetic-eld-driven op of conical spin spirals leads to the simultaneous reversal of magnetization and electric polarization.
1 Division of Materials Physics, Graduate School of Engineering Science, Osaka University, Toyonaka, Osaka 560-8531, Japan. 2 Condensed Matter Research Center, Institute of Materials Structure Science, High Energy Accelerator Research Organization, Tsukuba 305-0801, Japan. 3 Laboratory for Neutron Scattering and Imaging (LNS), Paul Scherrer Institut (PSI), Villigen CH-5232, Switzerland. 4 Department of Physics and Astronomy, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA. 5 Institut Laue-Langevin, BP 156X, Grenoble F-38042, France. 6 Laboratory for Scientic Developments and Novel Materials (LDM), Paul Scherrer Institut (PSI), Villigen CH-5232, Switzerland. * These authors contributed equally to this work. w Present address:
Department of Advanced Materials Science, University of Tokyo, Kashiwa 277-8561, Japan. Correspondence and requests for materials should be addressed to M.K. (email: mailto:[email protected]
Web End [email protected] ) or to T.K. (email: mailto:[email protected]
Web End [email protected] ).
NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 1
DOI: 10.1038/ncomms15457 OPEN
Coupled multiferroic domain switching in the canted conical spin spiral system Mn2GeO4
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15457
The nomenclature multiferroics has been originally coined for materials showing the coexistence of two or all three ferroic orders (ferroelectric, ferromagnetic and ferroelas
tic)1,2, and is expanded nowadays to comprise materials showing ferroelectric and antiferromagnetic orders3. Extensive studies in the past decade have led to the discoveries of various new multiferroics in which their magnetic order breaks inversion symmetry and gives rise to ferroelectricity37. Among such multiferroics, however, materials showing both spontaneous magnetization M and electric polarization P simultaneously, that is, multiferroics in the original denition, are not common. One of the few systems exhibiting both spontaneous M and P is a spiral magnet with the so-called transverse conical spin structure, which consists of a cycloidal spiral spin component and a ferromagnetic component along the spin rotation axis of the cycloid3,6. This is an example of a so-called multi-Q structure in which two distinct magnetic propagation vectors Q coexist. In the framework of the spincurrent or inverse DzyaloshinskiiMoriya (DM) mechanism8, the cycloidal component induces P in the direction perpendicular to both the spin rotation axis and the modulation wave vector. This means that P in transverse conical systems develops in the direction perpendicular to M. CoCr2O4 (refs 9,10) and some hexaferrites1117 are examples of such conicalspiral multiferroics in which simultaneous reversals of M and P are often observed9,10,12,16. However, the microscopic mechanism for the simultaneous reversal has not been fully understood to date. An important result of the present paper is to show that the phenomenological model we introduce here explains the switching mechanisms we observe in the double-Q conical-spiral multiferroic, Mn2GeO4.
Recently, an olivine-type compound Mn2GeO4 was reported to be a rare multiferroic below 5.5 K, where both spontaneous M and P develop in the same direction18,19. The crystal structure of Mn2GeO4 (Fig. 1a) is described by the orthorhombic
Pnma space group. The S 5/2 Mn2 ions occupy two distinct
crystallographic sites (Mn1 and Mn2) octahedrally coordinated by O2 anions, and form sawtooth-like chains along the b axis20.
The structure contains competing magnetic interactions, which lead to various magnetic phases and a complex magnetic order in the ground state multiferroic phase18,19,21. A previous neutron scattering study13 proposed a multi-Q magnetic order with both commensurate (C) and incommensurate (IC) components for the multiferroic state. The C component characterized by the propagation vector of (0 0 0) is described by a combination of two irreducible representations (irreps) G1 and G3 and allows nite
M along the c axis (Mc). The IC component forms a spin spiral with the propagation vector Qic (qh qk 0), where qh 0.136 and
qk 0.211. This structure is described by two irreps D1 and D2,
and contains a cycloidal component which can be responsible for P along the c axis (Pc). To the best of our knowledge, Mn2GeO4 is the only multiferroic with the M||P conguration where P can be understood as generated by a spiral spin order. Phenomenological symmetry analysis on the respective C and IC components well explained not only the simultaneous appearance of nite |M| and |P| but also the M||P conguration18. However, no clear accessible explanation for the MP coupling has been provided to date. In addition, as shown in Fig. 1b, a reversal of M accompanies that of P in the multiferroic state of Mn2GeO4, indicating a synchronized switching of both ferromagnetic and ferroelectric domains. This coupled switching mechanism still remains an unsolved issue, both microscopically and phenomenologically.
Here we report on the evolution of these domains in Mn2GeO4 by the application of electric and/or magnetic elds, and reveal one-to-one phenomenological correspondence relations between certain C and Q domains, which arise from the C and IC
components in the multi-Q magnetic order, respectively. From this we show that the multiferroic state of Mn2GeO4 can be regarded as canted conical spin spirals, as illustrated in Fig. 1c,d. We demonstrate experimentally that the effective spiral handedness, which is usually controlled by an electric eld in magnetic spiral ferroelectrics, couples not only with P but also with M via the C,Q domain switching, and its sign can be reversed by a magnetic eld in Mn2GeO4. We also clarify the mechanism of both the M||P conguration and the simultaneous reversal of M and P observed in this unique multiferroic compound.
ResultsConsideration of multiferroic domains. To elucidate the coupling mechanism between the ferromagnetism and ferroelectricity in Mn2GeO4, we investigated the effects of electric and magnetic elds on domains by means of unpolarized and polarized neutron scattering measurements on single crystal samples. Both the ferromagnetic and ferroelectric domains in this compound involve several types of magnetic domains, which are ascribed to the complex multi-Q magnetic order with both C and IC components. As illustrated in Supplementary Fig. 1a, the C component can host two pairs of ferromagnetic domains (dened as CA domain for G1"G3 and CB domain for G1"8G3). Each C domain possesses its time-reversal counterpart with the opposite sign of Mc (compare upper and lower panels of
Supplementary Fig. 1a). On the other hand, as illustrated in Supplementary Fig. 1b, the IC component can host two pairs of ferroelectric domains (dened as QA and QB domains for the propagation vectors (qh qk 0) and (qh qk 0), respectively). Each
Q domain possesses its space-inversion counterparts with the opposite sign of Pc (or the opposite sign of spin-helicity h)
(compare upper and lower panels of Supplementary Fig. 1b). Thus, the multiferroic state in Mn2GeO4 includes all four types of magnetic domains dened in ref. 22: orientation domains (CA2CB), time-reversed domains [C( Mc)2C( Mc)],
conguration (or Q) domains (QA2QB) and so-called chirality
domains [Q( h)2Q( h)].
Phenomenological coupling theory. First, we present a phenomenological theory that describes the possible coupling terms between the ferroelectric and magnetic order parameters in the multiferroic state of Mn2GeO4 (see Methods). To describe the couplings seven order parameters are identied: ferroelectric order Pz, two complex order parameters MD1Q and MD2Q for irreps
D1 and D2, respectively, for each IC Q domain, Q QA or QB (in
total four order parameters), and two C order parameters X1 and X3, belonging to irreps G1 and G3, respectively. MD1Q and MD2Q represent the amplitude of the magnetization distribution which transforms according to the irreps D1 and D2 for the space group Pnma and Qic (qh qk 0) (ref. 18), for example, MD1Q is even
under the glide operation perpendicular to the z axis (mz), and MD2Q is odd under mz. X1 describes 180 domains of the dominant
C antiferromagnetic component while X3 corresponds to the ferromagnetic component with magnetization along the z axis (Mz). Here the z axis corresponds to the c axis in Mn2GeO4.
Supplementary Table 1 and Supplementary Note 1 describe the symmetry operations of the space group and the transformation properties of these order parameters, respectively. We construct coupling terms that the reader can check are invariant under these transformation properties by using Supplementary Eqs. 1 7 in Supplementary Note 1, and this leads to three
2 NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15457 ARTICLE
a b
Ge
10
0.02
5
c( B per Mn)
0
0
5
M 0.02
10
P c(C m2)
c
O
Mn2
b
Mn1
0.4
0
0.4
a
Bc (T)
c
d
Conical chain Conical chain
c
b
a
Mc Mc
Pc Pc
Figure 1 | Structures and multiferroicity of Mn2GeO4. (a) Crystal structure of Mn2GeO4. (b) Magnetization along the c axis (Mc) and electric polarization along c (Pc) as a function of magnetic eld applied along c (Bc) at 4.5 K for a single crystal of Mn2GeO4. Before the measurement of Pc, the specimen was cooled from high temperature into the multiferroic phase in positive electric and magnetic elds. The gure is adapted with permission from ref. 18. (c,d) Magnetic structures in the multiferroic state of Mn2GeO4 (the most probable magnetic point group: 2 with 21 axis along c). Two of the possible magnetic domains are illustrated: (c) (CA, QA) domain [ G1" G3, (qh qk 0)] and (d) (CB, QB) domain [ G1" G3, (qh qk 0)]. Red and blue arrows
on cones denote Mn1 and Mn2 moments, respectively. In the magnetic structure, two types of conical spin chains extend along the b axis (conical chains a and b) and are placed alternately in a staggered arrangement.
such terms given by
U ir0Pz MD1QA MD2QA MD1QAMD2QA
MD1QB MD2QB MD1QBMD2QB
h i ;
1
V sX1X3 MD1QA
2
MD1QB
2
;
2
W irX1X3Pz MD1QA MD2QA MD1QAMD2QA MD1QB MD2QB MD1QBMD2QB
s0X1X3 MD2QA
2
MD2QB
2
:
3
Here s, s0, r, and r0 are real numbers. U describes the coupling between the spin-helicity and the electric polarization, and is equivalent to the trilinear coupling term developed for Ni3V2O8
(ref. 23) and TbMnO3 (ref. 24). The spin-helicity h is dened as h / i MD1QMD2Q MD1Q MD2Q
= MD1QMD2Q
, and is of opposite
sign for the QA and QB domains. Nevertheless, each Q domain hasthe same handedness of spin cycloidal component projected onto the bc plane, which induces Pz with the same sign. V describes the coupling between the C domains and the Q domains, and W describes the coupling between the magnetization and the electric polarization. These three coupling terms (U, V and W) dene strict relationships between the order parameters. Table 1 lists the relationships between the signs of the respective order parameters achieved after various eld cooling conditions. From these couplings, one can anticipate the effects of ferromagnetic and ferroelectric domain switching on the respective order parameters. Importantly, a reversal of X3 ( Mz) allows not only
a concomitant reversal of Pz but also a switching between the QA and QB domains. We will provide experimental evidence for the presence of all three coupling terms.
Electric-eld-cooling effect on spin-helicity. To prove the existence of the U coupling term (equation (1)), we studied the effect of a cooling electric eld Ecool on spin-helicity in the Q domains using polarized neutron diffraction (spherical neutron polarimetry (SNP)) at TASP, Paul Scherrer Institut (PSI). Details of the experiments are fully described in Methods. The intensity proles of the (28qh 1 qk 0) peaks were measured in both the non-
spin-ip (z,x) and spin-ip (z, x) polarization channels. From
the difference in the intensities observed in each polarization channel, the relative proportions of the two helicity domains within each Q domain can be determined directly (see Methods). These measurements were done in zero electric eld at 2 K, after having cooled the specimen from 10 to 2 K under various Ecool
along the c axis. In Fig. 2ad we show the wave-vector dependence of the scattering from the QA domain peak (2 qh 1 qk 0)
after cooling under Ecool 0.5 and 1.5 MV m 1. The results
show a strong sensitivity to both the neutron polarization state and size of Ecool, with the difference between the non-spin-ip and spin-ip scattering being larger in the data for Ecool 1.5
MV m 1. Furthermore, by switching the sign of Ecool the polarization-dependent magnitude of the spin-ip and non-spin-ip intensities becomes reversed. Similar data are also obtained for the QB domain. Importantly, these results show that the Pzx elements of the polarization matrix (see Methods) can be made nite, with both the magnitude and sign tunable by the electric-eld-cooling condition. This conrms the magnetic structure in the multiferroic state of Mn2GeO4 to include a spin spiral
NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 3
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15457
Table 1 | Relationships between the signs of order parameters.
Condition before measurement h[QA, QB] Pz X1 (C1) X3 Mz (C3) X1X3 (CA or CB) Q (QA or QB) E B cool [ , ] ( ) ( ) QA(QB) B sweep [ , ] ( ) ( ) QB(QA)
E B cool [ , ] ( ) ( ) QA(QB) B sweep [ , ] ( ) ( ) QB(QA)
E B cool [ , ] ( ) ( ) QA(QB) B sweep [ , ] ( ) ( ) QB(QA)
E B cool [ , ] ( ) ( ) QA(QB) B sweep [ , ] ( ) ( ) QB(QA)
The signs are determined by considering that an electric eld E and a magnetic eld B directly act on P and M , respectively, and by the U, V and W invariants in equations (1)(3). In the table, square brackets show the sign of h in each Q domain. The sign ( ) denotes that either or domain is possible. However, within each experimental condition, the signs in and out of parentheses
couple with those in and out of parentheses for other order parameters. Here, E B cool means the eld-cooled condition with positive E and B that are large enough to pole the system. B-sweep
denotes that B was swept to an opposite value once after the cooling procedure of the upper row. One-to-one correspondence between h and P is ascribed to the U term (equation (1)) and that between X X and Q originates from the V term (equations (2)). In the column of X X , and refer to C and C commensurate domains, respectively. The relationships between the respective order
parameters satisfy all experimental results (Table 2 and Supplementary Table 2).
component with a spin-helicity that can be selected by an electric eld.
We measured the spin polarization dependence of the magnetic scattering in more detail by examining the full polarization matrices for various peaks in the QA and QB domains (Methods). Using both the MuFIT program25 and the magnetic structure models reported previously18, we modelled the polarization matrix data to quantitatively determine the fractions of the spin-helicity domains (h domains) within each of the QA and QB domains, and for various Ecool. Figure 2e shows the h domain fractions (p( h) and p( h) in the QA and QB
domains, respectively), as a function of Ecool. For comparison, the Ecool dependence of Pc measured in the same condition is also plotted in Fig. 2e. The h domain fraction in each Q domain behaves in the same way, namely increasing in proportion with positive Ecool up to 2 MV m 1 and saturating above 2 MV m 1.
The Ecool dependence of the domain fraction matches up nicely with that of Pc. Thus, the QA and QB domains coexist even after the electric-eld cooling and contribute to the ferroelectric polarization, with their spin-helicity (or h domain fraction) equivalently tuned by Ecool. These observations, in particular the opposite sign of h in the two domains, are entirely consistent with the presence of the U coupling term. It is also worth pointing out that large Ecool (42 MV m 1) is required to obtain single spin-helicity in each Q domain. In most spin spiral ferroelectrics in single crystal form, an order of magnitude smaller Ecool is enough
to pole specimens (for example, o0.2 MV m 1 in Ni3V2O8 (ref. 26)).
Correlation between the C and Q domains. Next, we show the results of unpolarized neutron scattering measurements done at TriCS, PSI. Note that the unpolarized neutron scattering technique cannot distinguish a pair of time-reversed 180 domains or a pair of helicity domains. From this technique, however, we obtain quantitative information about the orientation domains (CA2CB) and the conguration domains (QA2QB)27. In addition, by carefully examining the eld-response of the respective domain populations, we also elucidate the correlation between the C and Q domains, that is, the coupling mechanism between ferromagnetism and ferroelectricity. Before the measurements, the specimen was cooled from 10 to 2 K at Ecool 3 MV m 1 and Bcool 1.5 T (magnetoelectric (ME)
cooling). These cooling elds are large enough to fully pole both the ferromagnetic and ferroelectric domains, which is conrmed by measurements of M and P or by checking the chiral scattering terms of the polarization matrix obtained from SNP experiments (see Figs 1b and 2e). After the ME cooling, both the electric eld E
Ecool = +0.5 MV m1
b
Ecool = 0.5 MV m1
Intensity (103counts per 30 s)
6
3
0
6
3
0
a
c
Ecool = +1.5 MV m1
d
Ecool = 1.5 MV m1
0.76 0.78 0.80 0.76 0.78 0.80 0.82
0.82
k (r.l.u.) k (r.l.u.)
20
e
Fraction of h domains
1.0
0.5
0.0
10
0
P c (C m2)
10
2.0 2.0
20
0
Ecool (MV m1)
Figure 2 | Electric-eld-cooling effect on spin-helicity. (ad) Polarized neutron scattering as a function of wave-vector transfer along the (0,k,0) wave-vector direction of the (2 qh 1 qk 0) magnetic reection in the QA
domain. The measurements were done at 2 K in the absence of electric eld after cooling the specimen from 10 to 2 K at the cooling electric eld Ecool 0.5 (a), 0.5 (b), 1.5 (c) and 1.5 MVm 1 (d) along the c axis. Red circles and blue triangles denote the diffracted intensities of non-spin-ip (z,x) and spin-ip (z, x) channels, respectively. (e) Fractions of
the spin-helicity h domains as a function of Ecool. Red open circles and blue
open squares show the fraction of h domain [p( h)] in the QA domain
and that of h domain [p( h)] in the QB domain, respectively. For
comparison, the Ecool dependence of Pc is also plotted (grey triangles). Error bars in (e) are dened as the difference between positive and negative polarization matrices.
4 NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15457 ARTICLE
Table 2 | Spin-helicity and G parameters in polarization matrix.
Q domain Reection Cooled at Ecool and Bcool After B-sweeph Pyx Pzx h Pyx Pzx
QA (2 qh 1 qk 0) 0.66(1) 0.73(1) 0.79(1) 0.82(1)
QB ( 2 qh 1 qk 0) 0.68(1) 0.64(1) 0.82(2) 0.80(2)
The spin-helicity h and the G parameters (Pyx, Pzx) of the Q and Q domains are determined from the intensities at the incommensurate magnetic reections of (2 q 1 q 0) and ( 2 q 1 q 0),
respectively. The data of Cooled at E and B were obtained after the magnetoelectric cooling condition with E 3 MVm
1 and B 1.2 T, and those of After B-sweep were obtained
after B was swept to 1.2 T once after the above cooling procedure. Both measurements were done at zero elds.
1
a (3 1 0)
2.0
1.5
1.0
0.5
Intensity (103counts per 30 s)
Integrated intensity
(arb. units)
b (3 1 0)
30.5 29.5
30 90 89.5 89 88.5
(2+qh 1+qk 0)
(2+qh 1qk 0)
4
c
d
3
2
99.5
(deg) (deg)
99 28 27
27.5
98.5
e
f
1 0 1 1 0 1
Bc (T) Bc (T)
Figure 3 | Coupled domain switching induced by a magnetic eld. Data obtained by unpolarized neutron scattering measurements. The y-scan proles of (a,b) the commensurate ( 3 1 0) and (c,d) the
incommensurate ( 2 qh 1qk 0) magnetic reections. The
measurements were done at 4.5 K. Black and red closed symbols denote the data measured after the ME cooling and after the reversal of M and P, respectively (see text). For comparison, the data measured after zero-eld cooling are also plotted (grey symbols). Error bars are dened as the square root of counts. (e,f) Integrated intensity of the commensurate (1 1 0) and the incommensurate (qh 1 qk 0) magnetic reections as a function of
magnetic eld along the c axis, Bc. Pink and blue marks in (e) denote the data of the (1 1 0) and the (qh 1 qk 0) magnetic reections, respectively.
Pink and blue marks in (f) represent the data of the (1 1 0) and the( qh 1 qk 0) magnetic reections, respectively. These data were taken
while sweeping a magnetic eld from positive to negative Bc.
and magnetic eld B were swept to zero at 2 K, and then we measured integrated intensities of 50 C and 70 IC magnetic peaks. For instance, y-scan proles of the ( 3 1 0), ( 3 1 0),
( 2 qh 1 qk 0) and ( 2 qh 1 qk 0) reections are
shown in Fig. 3ad, respectively (black symbols). Subsequently, B was swept to 1.5 T and then set to zero at 4.5 K. This
B-sweeping procedure causes the reversal of M and P. Then the same measurements were done and the integrated intensity data of the same magnetic peaks were obtained again. Red symbols in
Fig. 3ad show the y-scan proles of the above-mentioned reections after the reversal of M and P. By comparing the data before and after the reversal of M and P, it is apparent that the intensities of the ( 3 1 0) and ( 2 qh 1 qk 0) reections
are suppressed after the reversal while those of the ( 3 1 0)
and ( 2 qh 1 qk 0) reections are enhanced. In fact, as
shown in Fig. 3e,f, a stepwise change in the integrated intensity of both the C (1 1 0) and the IC (qh 1 qk 0) magnetic
reections is observed at the magnetic eld where the reversal of M and P occurs.
As described in Methods, renement of the integrated intensity data and determination of the respective domain fractions were done using both the FullProf program28 and the models proposed in ref. 18. From the analysis, the domain fractions were estimated to be p(CA) : p(CB) 0.60(1) : 0.40(1) and p(QA) :
p(QB) 0.600(3) : 0.400(3) before the reversal while those after
the reversal were p(CA) : p(CB) 0.33(1) : 0.67(1) and p(QA) :
p(QB) 0.371(2) : 0.629(2). Thus, the domain fractions of the CA
and CB domains are nearly the same with those of the QA and QB domains, respectively. This result suggests that the C and IC domains are strongly entangled, and that the CA (CB) domain is always coupled with the QA (QB) domain. Furthermore, the domain fraction relationships between the domains A and B are reversed in both the C and Q domains after the reversal of M and P. These coupling features were observed in all the data for various eld cooling conditions, as summarized in Supplementary Table 2. This indicates that a domain switching occurs not only between the CA and CB domains but also between QA and QB domains when the simultaneous reversal of M and P takes place. For a given sign of X1, CA and CB domains are distinguished by the sign of X3, and the presence of coupling term V (equation (2)) dictates a coupling of the Q domains with the C domains consistent with our observations.
Magnetic-eld effect on spin-helicity and domain switching. We now show the results obtained by polarized neutron diffraction (PND) measurements at D3, ILL (see Methods). As listed in Table 2, the signs of the G polarization matrix parameters ( Pyx, Pzx) for magnetic peaks in the QA and QB domains are
opposite to each other after the same ME cooling procedure, and are reversed after B-sweeping, that is, the M reversal accompanied by the P reversal. This switching of the average spin-helicity with magnetic eld arises due to a switching of the Q domain populations and not due to the switching of the helicity within a Q domain: it is this change of the Q domains that leads to the observed change of the spin-helicity because the QA and QB domains have opposite helicity according to coupling term U. This leaves the helicity-dependent term in the coupling term W overall invariant, (equation (3)), so that W describes the coupling between ferromagnetic and ferroelectric orders in Mn2GeO4.
DiscussionBased on these experimental results, we conclude that the full magnetic structure in the multiferroic state of Mn2GeO4 is of the
NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 5
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15457
double-Q type, being namely a superposition of the CA and QA domains (or the CB and QB domains). This means that only the following eight-types of C, Q domain combinations are possible: [CA(Mc), QA(Pc)], [CA(Mc), QA(8Pc)], [CB(Mc),
QB(Pc)], and [CB(Mc), QB(8Pc)] (the double sign applies in the same order as written). The sign of X1 determines whether M is parallel or antiparallel to P. In Fig. 1c,d we show two of the eight possible double-Q domains. The resulting magnetic structures can be simplied by considering them to include two types of conical spin chain (a and b) formed by Mn1 moments.
The simplied structure is illustrated in Fig. 4a,b. In each conical chain, the Mn moments form a cone with the cone axis (|| vector sum of the Mn moments in each chain) making a nite angle with the a, b, and c axes. The coupling between the conical chains a and b is canted antiferromagnetic. Though the net moments along the a and b axes are cancelled out, the c axis component remains nite. As a result, a net magnetization Mnet develops along the c axis, as schematically illustrated in Fig. 4b. Furthermore, each conical chain comprises a cycloidal
component which allows a local P in the direction perpendicular to both the component of the propagation vector ||b, and the spin rotation axis through the inverse DM mechanism8. The vector sum of the local P in the conical chains a and b is aligned along the c axis, and a net electric polarization Pnet develops along the c axis (Fig. 4b). Thus, the multiferroic state of Mn2GeO4 can be regarded simply as canted conical spin spirals in which magnetically-induced electric polarization develops in the direction parallel to magnetization.
Now, we turn to the mechanism of the simultaneous reversal of magnetization and electric polarization by applying B along c (Bc)
whose sign is opposite to that of Mc. As mentioned above, the results of unpolarized neutron scattering experiments suggest that domain switching occurs between the (CA, QA) and (CB, QB)
domains when the reversal of M and P takes place. Schematics of the change in a cone axis upon the reversal of M and P (or domain switching) are illustrated in Fig. 4c. Starting from the (CA( Mc), QA( Pc)) domain with positive Mc and Pc
(Top left panel of Fig. 4c), a magnetic eld antiparallel to Mc will lead to the domain switching into the (CB( Mc), QB( Pc))
domains with negative Mc and Pc (Top right panel of Fig. 4c). As displayed in Fig. 4c, each of the eight possible domains can be transformed into only one of the other domains by applying Bc.
Such a B-induced domain switching can be viewed as a op of the cone axis. Note that the cone axis ops across not only the ab plane but also the bc plane during the switching. The cone-axis op across the ab plane leads to the reversal of M along the c axis while the cone-axis op across the bc plane gives rise to the reversal of P along the c axis. This is because the effective spiral handedness of the cycloidal component in the cone projected onto the bc plane reverses after the op. The reversal of the handedness of the spin cycloid component with propagation vector along the b axis and the rotation axis along the a axis causes the reversal of magnetically-induced P along the c axis in the framework of the inverse DM mechanism8. This scheme is also consistent with no sign change in the spin-helicity h after the reversal of M and P, which was revealed by our unpolarized and polarized neutron scattering measurements. Thus, the simultaneous reversal of M and P observed in Mn2GeO4 is well explained in the light of a op of the cone axis in the conical spin chains.
Phenomenologically, the M-P coupling can be described by the coupling term W (equation (3)), as we will show now: from the coupling term V (equation (2)) we know that a reversal of X3 (or the magnetization) leads to switching of the Q domains (from QA to QB or vice versa). The coupling term W dictates that the two different Q domains couple with the same helicity to both magnetization and ferroelectricity. Experimentally, we established that a reversal of the magnetization leads to a conservation of the spin-helicity and to a switch of the Q domains, which leaves the term in W describing the magnetic structure invariant under a reversal of the magnetization. The term W describes thus a direct coupling between the ferroelectric polarization and the magnetizationconsistent with the microscopic scenario explained above.
In summary, unpolarized and polarized neutron scattering experiments have been carried out on single crystals of the olivine-type Mn2GeO4 to elucidate a unique coupling mechanism between ferromagnetism and ferroelectricity observed in this multiferroic compound. We investigated the double-Q magnetic structure of the multiferroic state by selectively preparing and controlling the populations of various types of magnetic domains. We have shown that the combination of canted commensurate and spin spiral incommensurate magnetic components offer unconventional magnetoelectric couplings such as the magnetic-eld control of Q domain switching and electric polarization
a
b
M M
cone cone
Pnet (=P +P ) || c
Mnet (=M +M ) || c
Conical chain
Conical chain
c
(CA,QA) domains (qh qk 0)
(CB,QB) domains (qh qk 0)
Bc
+Bc
Bc
Bc
+1
+1 +3
+Mc
1 3
Mc
Mc
3
+Bc
+Bc
Bc
1
+Bc
+Mc
+3
Figure 4 | Schematics of coupled multiferroic domain switching.(a) Simplied conical magnetic structure viewed along the c axis.
Red arrows denote conical axes. The structure consists of two-types of conical spin chains extended along the b axis (conical chains a and b).
(b) The schematic congurations of local magnetization Ma,b (red arrows) and magnetically-induced electric polarization Pa,b (blue arrows) for each conical chain are illustrated. In each conical chain, Ma,b and Pa,b are aligned in directions parallel and perpendicular to the cone axis, respectively. The coupling between the two types of chains is canted antiferromagnetic, which results in both net magnetization Mnet and electric polarization
Pnet along the c axis. (c) Schematic illustrations of coupled multiferroic domain switching by a magnetic eld along the c axis, Bc. Blue circular arrows represent the sense of spin spiral, which corresponds to the sign of the spin-helicity h. The application of Bc causes a op of the cone axis, which leads to coupled multiferroic domain switching.
6 NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15457 ARTICLE
reversal without a change of spin-helicity. The double-Q magnetic structure of multiferroic Mn2GeO4 can be regarded as canted conical spin spirals, which well explains the parallel conguration of spontaneous magnetization and electric polarization. Furthermore, we revealed that the simultaneous reversal of magnetization and electric polarization observed in this multiferroic can be understood in terms of a op of the cone axis in the conical spin chains. The present study describes how magnetization and ferroelectric polarization in spin-driven ferroelectrics with conical magnetic order can be coupled on a phenomenological level in the case where they are parallel to each other. Furthermore, it stands out as a text-book case of how complex and seemingly unexpected coupling terms can be engineered in such spin-driven ferroelectrics. The essence of the present phenomenological discussion can be also applied to other conical-spin-driven ferroelectrics.
Methods
Sample preparation and measurement of electric polarization. Single crystals of Mn2GeO4 were grown by the oating zone technique, as reported previously14. For measurements of neutron diffraction and electric polarization P, the crystals were cut into thin plate-shaped specimens. The widest faces of these specimens were perpendicular to the c axis in order to apply an electric eld along the ferroelectric c axis. For the measurements of P, a plate-shaped specimen with the dimension approximately 1 mm2 45 mm was used, and silver electrodes were
vacuum-deposited on the surfaces of the specimen. To obtain P as a function of cooling electric eld, pyroelectric current was measured with an electrometer after cooling the specimen from high temperature into the multiferroic phase in various cooling electric elds applied along the c axis.
Phenomenological coupling theory. To develop the phenomenological coupling theory, we rst determined the symmetry properties of the magnetic and ferro-electric order parameters found experimentally. The symmetry properties of the magnetic order parameters are determined by their irreducible representation, and lead directly to the transformation properties of the order parameters given in Supplementary Eqs 2 7 of Supplementary Note 1. The transformation property of
the ferroelectric polarization is given in Supplementary Eq. 1 of Supplementary Note 1, similar to the procedure used in ref. 29. Using these transformation properties, we construct coupling terms that are invariant under the space group symmetry elements, and those relevant to Mn2GeO4 are listed in equations (1)(3).
A full account of the phenomenological coupling theory is described in ref. 30.
Polarized neutron diffraction. Polarized neutron diffraction (PND) measurements were carried out using both the cold neutron triple-axis spectrometer TASP (l 3.189 ) equipped with the MuPAD device31 at the PSI, Switzerland, and the
hot diffractometer D3 (l 0.825 ) equipped with the CryoPAD device32 at the
Institut Laue-Langevin, France. All experiments were carried out on single crystal samples of approximate masses 50100 mg, and cut into thin plates of approximate cross-section 10 20 1 mm3, with thin dimension parallel to the ferroelectric
c axis. The samples were always mounted with a [100] [010] horizontal scattering
plane, which allowed access for SNP measurements over a range of commensurate(C) and incommensurate (IC) magnetic peak positions.
PND measurements at TASP. For the TASP experiment, a 10 nm layer of Cr and 50 nm layer of Au were sequentially evaporated onto the surfaces of the plate-like sample for making electrical connections. The sample was then oriented and mounted on an aluminium spacer positioned at the end of a bespoke sample stick designed for the application of high voltages (up to 5 kV) at cryogenic temperatures33. The high-voltage lines were connected directly to the CrAu sample electrodes using silver paint, so that electric elds could be applied along the ferroelectric c axis. The sample space at the end of the stick was indium-sealed and covered by an Al vacuum can. The sample space was evacuated in order to avoid electric-eld breakdown across the sample during the measurements. The high voltage stick with pumped sample space was then installed into a standard Orange cryostat (base temperature of 2 K), which itself was installed into the MuPAD device at the TASP beamline. Polarization matrix measurements were carried out at 2 K, after either zero-eld cooling or electric-eld cooling through the multiferroic transition at 5.5 K. When at the base temperature, the electric eld could be removed for the SNP measurements. We observed no effect on the resulting data due to the electric eld removal in the multiferroic phase.
PND measurements at D3. For the D3 experiments, a plate-like sample (without pre-deposited electrodes) was installed onto a different bespoke 5 kV high voltage sample stick. On this stick the application of electric elds along the c axis is
achieved by a potential difference applied across two parallel Al plates. The sample was xed to one plate by a silver epoxy, and the other plate positioned close to, but not touching the other sample surface. Thus the actual electric eld across the sample was not exactly the same as the applied voltage. By comparing between similar electric-eld cooling measurements as done at TASP we could estimate the absolute electric eld scale in the D3 experiments. Similarly as for the TASP experiments, the sample space of the sample stick used at D3 was also indium-sealed by an Al can and the sample space evacuated. The pumped sample stick was then installed into a thin tail Orange cryostat (base temperature of 3.5 K), which could then be installed into the CryoPAD device at the D3 beamline. SNP measurements were done after either zero-eld cooling or electric-eld cooling through the multiferroic transition, similarly as at TASP.
In addition at D3, an electromagnet (maximum applied eld 1.2 T) was available. By installing the thin-tail cryostat into the electromagnet, vertical magnetic elds could be applied along the c axis, in addition to electric elds. Thus at D3 we could investigate the effects of magnetoelectric cooling (simultaneous magnetic eld- and electric eld-cooling) into the multiferroic phase, and also magnetic eld sweeping from positive to negative magnetic elds within the multiferroic phase. Since the electromagnet cannot be integrated with the CryoPAD device, each sample state preparation involving the magnetic eld was done ofine; by using the beamline crane we could seamlessly interchange the high voltage stick and cryostat ensemble between the CryoPAD device and the electromagnet stationed nearby. Thus, once the sample had been either poled to3.5 K through the multiferroic transition under simultaneously applied electric-and magnetic-elds, or the magnetic eld was swept between 1.2 T at the slightly elevated temperature of 4.5 K, all elds were then removed before the cryostat and sample stick were installed back into the CryoPAD device for the SNP measurements of the newly prepared state.
Spherical neutron polarimetry technique. Both of the MuPAD and CryoPAD devices allowed for SNP measurements of magnetic Bragg scattering at TASP and D3, respectively. SNP is a technique that, given an incident neutron beam polarized along an arbitrary direction, enables the full measurement of the direction of neutron spin polarization after the scattering process. The results are expressed as a polarization matrix
Pij
r
j jj
h i
j j2 i M? q
r
j j
i M? q
D E
j
2
r
j jj
h i
j j2 i M? q
r
j j
i M? q
D E
j
2 ; 4
where the ket i
h j describes the initial neutron spin state, the bra j
j i describes the
nal neutron spin state (or j
j i the nal spin-ip state), r is the Pauli operator for
the neutron spin, and M? q
is the component of M q
that is perpendicular to the
neutron scattering vector q. Here M(q) is the magnetic structure factor. Full details of the calculation of M(q) for the different magnetic structures in Mn2GeO4, are given in the Supplement of ref. 18. The quantity Iij i M? q
r
j jj
h i
j j2 is the
intensity of the magnetic Bragg peak observed for the incoming i and outgoing
j spin polarizations.
The SNP technique is a particularly powerful probe of complex magnetic structures and magnetic domain distributions. For the IC spiral structure in the multiferroic phase of Mn2GeO4, and for a perfectly efcient polarized neutron scattering setup, a general polarization matrix can be written for a magnetic Bragg peak probed at q:
P
1 0 0
G F 0
G 0 F
0
@
1
A: 5
Here F Mj j M
j j , and G 2I M M
j j
f g
M
j j . In this work we use the standard
coordinate system where x is dened as always parallel to q, z perpendicular to the scattering plane, and y orthogonal to both x and z. Consequently at eachq, M? 0; My; Mz
.
The F and G parameters of P are measured experimentally and compared with
the values expected according to a model for magnetic structure and domain populations. The terms G are the so-called chiral scattering elements of P, and by
denition can only be nite when the magnetic structure has multiple modes that are out-of-phase with one another, such as is the case for non-collinear structures like cycloids, helices, and the IC spiral order in Mn2GeO4. The measured value of
G not only depends on the magnetic structure in the sample, but it is also sensitive to the populations of the possible domains of the magnetic structure. In the multiferroic phase of Mn2GeO4, the spin spiral magnetic structure of each Q domain has a helicity degree-of-freedom (h) the sign of which is intimately related to the direction of ferroelectric polarization along either c or c. After
zero-eld cooling, equal populations of opposite helicity domains are realized, which will give either Go0 or G40. Thus for equal helicity domain populations the so-called chiral scattering averages out and the measured value of G 0, as
would be observed for a collinear structure. On the other hand, as we demonstrate in our experiments at TASP, by applying an electric eld along the ferroelectricc axis while cooling, the helicity domain population within each Q domain can be controlled, and nite values of G are observed. Thus, not only does measuring a
M
NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 7
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15457
nite value of G allow the determination of the relative fraction of spiral domains of opposite handedness (which generate ferroelectricity along either c or c),
but it further conrms the IC phase of Mn2GeO4 to be symmetry-breaking
complex magnetic order with out-of-phase modes.
The full calculation of the expected F and G parameters further depends on the exact orientation of the magnetic scattering vector q with respect to the real-space spin distribution. If the measured q is exactly perpendicular to the spin rotation plane of the non-collinear spiral structure, then one obtains G 1 for a mono
handedness spiral state and F 0. Typically however, most of our polarization
matrix measurements were done at IC q vectors which lie at angles away from perpendicular to the spin spiral rotation plane. Consequently, even for a Q domain containing mono-handed spiral order, the magnitude of G becomes o1, and the magnitude of F larger than 0. Nevertheless, when armed with a full model for the magnetic structure, the measured values for F and G at any q allow for a full determination of the spiral-handedness domain populations. We used the MuFIT program25 for both analysing and tting the SNP data to determine the magnetic domain populations.
We also briey mention that in our SNP experiments we measured some polarization matrices of Bragg peaks due to the C magnetic orders in the various phases of Mn2GeO4. As expected for these simple q 0 antiferromagnetically
modulated structures, the measured matrices only had nite diagonal elements, and the G parameters were always measured to be 0 within uncertainty.
Unpolarized neutron diffraction. Unpolarized neutron diffraction measurements were carried out using the TriCS diffractometer at PSI (l 1.178 ). Similarly as
for the SNP experiments described above at TASP, a plate-like Mn2GeO4 sample was mounted on the same 5 kV high voltage stick33 so that electric elds could be applied parallel to the crystal c axis. The horizontal scattering plane at TriCS contained the [100] [010] directions. The high voltage stick was installed into the
variable temperature insert (VTI) of a 6 T vertical eld cryomagnet (base temperature 2 K), so that both electric and magnetic elds could be applied simultaneously along the crystal c axis. Unlike at D3, the cryomagnet apparatus could be installed on the beamline, and the electric- and/or magnetic-eld cooling from high temperature into the multiferroic phase done online. After eld-cooling the sample through the multiferroic transition, all elds were then removed before data collection at 2 K. In the case of magnetic-eld sweeping an already poled sample, the magnetic eld sweeps were done after carefully elevating the temperature to 4.5 K. We found that at this slightly higher temperature the effect of the magnetic eld sweeping on the magnetic domain populations was more pronounced than when sweeping was done at 2 K. After the sweeps the eld was removed and the temperature lowered back to 2 K for the data collection.
After each multiferroic state preparation, either by electric- and/or magnetic-eld cooling, or by magnetic eld sweeping, we studied both the C and IC magnetic structure components that exist in this state. Typical data sets constituted integrated intensity measurements for 50 C magnetic peaks and 70 IC magnetic peaks. Renements of the integrated intensity data and magnetic domain populations were done using the FullProf program28 and the models proposed in ref. 18. From these renements we determined the relative populations of the CA and CB C domain types, and the QA and QB IC domain types for each prepared state (with these domain types introduced in the main text), and consequently the relationships between them based on the response of their relative populations to the applied electric and magnetic elds.
Data availability. The data supporting the ndings of this study are available from the corresponding authors on request.
References
1. Schmid, H. Multi-ferroic magnetoelectrics. Ferroelectrics 162, 317338 (1994).2. Schmid, H. Some symmetry aspects of ferroics and single phase multiferroics.J. Phys.: Condens. Matter 20, 434201 (2008).3. Tokura, Y., Seki, S. & Nagaosa, N. Multiferroics of spin origin. Rep. Prog. Phys 77, 076501 (2014).
4. Jia, C., Onoda, S., Nagaosa, N. & Han, J. H. Microscopic theory of spin-polarization coupling in multiferroic transition metal oxides. Phys. Rev. B 76, 144424 (2007).
5. Cheong, S.-W. & Mostovoy, M. Multiferroics: a magnetic twist for ferroelectricity. Nat. Mater. 6, 1320 (2007).
6. Kimura, T. Spiral magnets as magnetoelectrics. Annu. Rev. Mater. Res. 37, 387413 (2007).
7. Johnson, R. D. & Radaelli, P. G. Diffraction studies of multiferroics. Annu. Rev. Mater. Res. 44, 269298 (2014).
8. Katsura, H., Nagaosa, N. & Balatsky, A. V. Spin current and magnetoelectric effect in noncollinear magnets. Phys. Rev. Lett. 95, 057205 (2005).
9. Yamasaki, Y. et al. Magnetic reversal of the ferroelectric polarization in a multiferroic spinel oxide. Phys. Rev. Lett. 96, 207204 (2006).
10. Choi, Y. J. et al. Thermally or magnetically induced polarization reversal in the multiferroic CoCr2O4. Phys. Rev. Lett. 102, 067601 (2009).
11. Ishiwata, S., Taguchi, Y., Murakawa, H., Onose, Y. & Tokura, Y. Low-magnetic-eld control of electric polarization vector in a helimagnet. Science 319, 16431646 (2008).
12. Taniguchi, K., Abe, N., Ohtani, S., Umetsu, H. & Arima, T. Ferroelectric polarization reversal by a magnetic eld in multiferroic Y-type hexaferrite Ba2Mg2Fe12O22. Appl. Phys. Express 1, 031301 (2008).
13. Chun, S. H. et al. Realization of giant magnetoelectricity in helimagnets. Phys. Rev. Lett. 104, 037204 (2010).
14. Tokunaga, Y. et al. Multiferroic M-type hexaferrite with a room-temperature conical state and magnetically controllable spin helicity. Phys. Rev. Lett. 105, 257201 (2010).
15. Soda, M., Ishikura, T., Nakamura, H., Wakabayashi, Y. & Kimura, T. Magnetic ordering in relation to the room-temperature magnetoelectric effect of Sr3Co2Fe24O41. Phys. Rev. Lett. 105, 257201 (2010).
16. Wang, F., Zou, T., Yan, L.-Q., Liu, Y. & Sun, Y. Low magnetic eld reversal of electric polarization in a Y-type hexaferrite. Appl. Phys. Lett. 100, 122901 (2012).
17. Kimura, T. Magnetoelectric hexaferrites. Annu. Rev. Condens. Matter Phys. 3, 93110 (2012).
18. White, J. S. et al. Coupling of magnetic and ferroelectric hysteresis by a multi-component magnetic structure in Mn2GeO4. Phys. Rev. Lett. 108, 077204 (2012).
19. Honda, T., Ishiguro, Y., Nakamura, H., Wakabayashi, Y. & Kimura, T. Structure and magnetic phase diagrams of multiferroic Mn2GeO4. J. Phys. Soc.
Jpn. 81, 103703 (2012).20. Creer, J. G. & Troup, G. J. The crystal and magnetic structures of Mn2GeO4. Solid State Commun. 8, 11831188 (1970).
21. Volkov, N. V. et al. Magnetic phase diagram of the olivine-type Mn2GeO4 single crystal estimated from magnetic, resonance and thermodynamic properties. J. Phys.: Condens. Matter 25, 136003 (2013).
22. Brown, P. J. Magnetic structure studied with zero-eld polarimetry. Physica B 192, 1424 (1993).
23. Lawes, G. et al. Magnetically driven ferroelectric order in Ni3V2O8. Phys. Rev. Lett. 95, 087205 (2005).
24. Kenzelmann, M. et al. Magnetic inversion symmetry breaking and ferroelectricity in TbMnO3. Phys. Rev. Lett. 95, 087206 (2005).
25. Poole, A. & Roessli, B. Analysis of neutron polarimetry data using MuFit.J. Phys.: Conf. Ser. 340, 012017 (2012).26. Cabrera, I. et al. Coupled magnetic and ferroelectric domains in multiferroic Ni3V2O8. Phys. Rev. Lett. 103, 087201 (2009).
27. White, J. S. et al. Stress-induced magnetic domain selection reveals a conical ground state for the multiferroic phase of Mn2GeO4. Phys. Rev. B 94, 024439 (2016).
28. Rodrguez-Carvajal, J. Recent advances in magnetic structure determination by neutron powder diffraction. Physica B 192, 5569 (1993).
29. Harris, A. B., Kenzelmann, M., Aharony, A. & Entin-Wohlman, O. Effect of inversion symmetry on the incommensurate order in multiferroic RMn2O5
(R rare earth). Phys. Rev. B 78, 014407 (2008).
30. Harris, A. B. Landau Potential for Multiferroic Mn2GeO4. Preprint at https://arxiv.org/abs/1701.04976
Web End =https://arxiv.org/abs/1701.04976 (2017).
31. Janoschek, M., Klimko, S., Gahler, R., Roessli, B. & Bni, P. Spherical neutron polarimetry with MuPAD. Physica B 397, 125130 (2007).
32. Tasset, F., Brown, P. J. & Forsyth, J. B. Determination of the absolute magnetic moment direction in Cr2O3 using generalized polarization analysis. J. Appl.
Phys. 63, 36063608 (1988).33. Bartkowiak, M., White, J. S., Rnnow, H. M. & Pra, K. Note: Versatile sample stick for neutron scattering experiments in high electric elds. Rev. Sci. Inst. 85, 026112 (2014).
Acknowledgements
We thank M. Fiebig for valuable discussions. We acknowledge nancial support from Grants-in-Aid for Scientic Research (No. 24244058), Grant-in-Aid for the Japan Society for the Promotion of Science Fellows (No. 24 1989), the Swiss National Centre of
Competence in Research program MaNEP and the SNF under Grant Nos. 200021_126687, 200021_138018, and 20021_153451, Switzerland. This work is based on experiments performed at SINQ, PSI, Villigen, Switzerland and ILL, Grenoble, France. The work was also supported by Condensed Matter Research Center (CMRC) in KEK, Japan.
Author contributions
T.H., J.S.W., M.K. and T.K. designed this work. T.H. and T.K. prepared single crystals of Mn2GeO4. T.H., J.S.W. and M.K. carried out the neutron scattering experiments with
L.C.C. (D3, ILL), A.F. and B.R. (MuPAD TASP, PSI) and O.Z. (TriCS, PSI). A.B.H.
developed the phenomenological coupling theory. T.H., J.S.W., M.K. and T.K. co-wrote the paper. All authors reviewed the manuscript and agree with the results and conclusions.
8 NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15457 ARTICLE
Additional information
Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications
Web End =http://www.nature.com/ http://www.nature.com/naturecommunications
Web End =naturecommunications
Competing interests: The authors declare no competing nancial interests.
Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/
Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/
Web End =reprintsandpermissions/
How to cite this article: Honda, T. et al. Coupled multiferroic domain switching in the canted conical spin spiral system Mn2GeO4. Nat. Commun. 8, 15457doi: 10.1038/ncomms15457 (2017).
Publishers note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
Web End =http://creativecommons.org/licenses/by/4.0/
r The Author(s) 2017
NATURE COMMUNICATIONS | 8:15457 | DOI: 10.1038/ncomms15457 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 9
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Copyright Nature Publishing Group Jun 2017
Abstract
Despite remarkable progress in developing multifunctional materials, spin-driven ferroelectrics featuring both spontaneous magnetization and electric polarization are still rare. Among such ferromagnetic ferroelectrics are conical spin spiral magnets with a simultaneous reversal of magnetization and electric polarization that is still little understood. Such materials can feature various multiferroic domains that complicates their study. Here we study the multiferroic domains in ferromagnetic ferroelectric Mn2 GeO4 using neutron diffraction, and show that it features a double-Q conical magnetic structure that, apart from trivial 180o commensurate magnetic domains, can be described by ferromagnetic and ferroelectric domains only. We show unconventional magnetoelectric couplings such as the magnetic-field-driven reversal of ferroelectric polarization with no change of spin-helicity, and present a phenomenological theory that successfully explains the magnetoelectric coupling. Our measurements establish Mn2 GeO4 as a conceptually simple multiferroic in which the magnetic-field-driven flop of conical spin spirals leads to the simultaneous reversal of magnetization and electric polarization.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer