Dzyaloshinskii-Moriya domain walls in magnetic nanotubes
Arseni Goussev, J. M. Robbins, Valeriy Slastikov, Oleg A. Tretiakov
DDzyaloshinskii-Moriya domain walls in magnetic nanotubes
Arseni Goussev, J. M. Robbins, Valeriy Slastikov, and Oleg A. Tretiakov
3, 4, ∗ Department of Mathematics and Information Sciences,Northumbria University, Newcastle Upon Tyne, NE1 8ST, UK School of Mathematics, University of Bristol, Bristol, BS8 1TW, UK Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan School of Natural Sciences, Far Eastern Federal University, Vladivostok 690950, Russia
We present an analytic study of domain-wall statics and dynamics in ferromagnetic nanotubes with spin-orbit-induced Dzyaloshinskii-Moriya interaction (DMI). Even at the level of statics, dramatic effects arise from theinterplay of space curvature and DMI: the domains become chirally twisted leading to more compact domainwalls. The dynamics of these chiral structures exhibits several interesting features. Under weak applied currents,they propagate without distortion. The dynamical response is further enriched by the application of an externalmagnetic field: the domain wall velocity becomes chirality-dependent and can be significantly increased byvarying the DMI. These characteristics allow for enhanced control of domain wall motion in nanotubes withDMI, increasing their potential as information carriers in future logic and storage devices.
PACS numbers: 75.78.Fg, 75.60.Ch, 75.70.Tj
I. INTRODUCTION
In recent years, ferromagnetic nanostructures featuring nar-row and stable domain walls (DWs) have been in the spotlightof experimental and theoretical research, with an overarchingaim to achieve more compact spintronic logic and memorydevices.
In particular, numerous efforts have been focusedon DWs in ferromagnetic nanowires, nanotubes, andthin films with perpendicular magnetic anisotropy featur-ing the Dzyaloshinskii-Moriya interaction (DMI).
Herewe report striking effects arising from the interplay betweenspace-curvature and DMI in ferromagnetic nanostructures,leading to narrow and stable DWs controllable, efficiently andreliably, by means of electric current and magnetic field.Curvature effects play a significant role in various fieldsof physics, and are attracting increasing attention in con-densed matter, particularly in nanomagnetism. The simplestsystem with curvature where DW dynamics can be consid-ered is a magnetic nanotube. Furthermore, thin ferromag-netic nanotubes have attracted recent attention from experi-mentalists owing to a number of technologically advantageousproperties, including enhanced DW stability understrong external fields, allowing for higher DW velocities com-pared to flat geometries; increased DW velocities underelectric current pulses; and the possibility of switching chi-rality in vortex DWs through magnetic field pulses. We show that in thin ferromagnetic nanotubes, theDMI induces qualitatively different effects to those foundin flat nanostructures, such as thin films and rectangularnanowires. In nanotubes, DMI causes the domains to be-come twisted, with helical lines of magnetization as in Fig. 1,forcing the DWs to become narrower. In contrast, in rectangu-lar nanowires with DMI, the magnetization far from the DWremains parallel to the wire axis while DWs become broader.This sharpening effect of DMI in nanotubes can enable sub-stantial downscaling in future nanodevices.We further demonstrate that in a certain thin-nanotuberegime specified below, DWs exhibit perfectly stable motionunder an applied electric current, propagating without any dis- z e φ e ρ e FIG. 1. (Color online) Domain wall profile in a thin nanotube withDzyaloshinskii-Moriya interaction. The magnetization lies tangentto the nanotube surface. tortion. The adiabatic spin-transfer torque is absent in thespin dynamics equations, and the non-adiabatic term takes theform of the adiabatic one.Complimentarily to the current, a magnetic field along thenanotube triggers a rich dynamical response in the magnetiza-tion texture. We show that the DW velocity becomes stronglydependent on polarity and chirality, and can be significantlyenhanced by DMI, which is favorable for memory applica-tions. Moreover, the onset of magnonic breakdown, imped-ing DW transport at high fields, can be efficiently suppressedby DMI. II. STATICS
We consider a ferromagnetic nanotube with inner radius R and thickness w . In the thin-nanotube regime w (cid:28) R , the a r X i v : . [ c ond - m a t . m e s - h a ll ] F e b micromagnetic energy with DMI takes the form E ( m ) = (cid:90) d r (cid:26) A |∇ m | + K (cid:2) − ( m · e z ) (cid:3) + DM s m · ( ∇ × m ) + µ M s m · e ρ ) (cid:27) , (1)where the integral runs over the volume of the nanotube, A is the exchange constant, K is the easy-axis crystallineanisotropy, D is the DMI constant, M s is the saturation mag-netization, and µ is the magnetic permeability of vacuum.The cylindrical-coordinate unit vectors e z , e ρ , and e φ areshown in Fig. 1.The last term on the right-hand side of Eq. (1) representsthe shape anisotropy stemming from the thinness of the nan-otube. In nanotubes with radius R much larger than the mag-netostatic exchange length (cid:112) A/ ( µ M s ) , this term forces themagnetization to lie nearly tangent to the surface. In this case,the unit vector of magnetization may be described by its ori-entation Θ( z, ρ, φ ; t ) in the ( z, φ ) -tangent plane: m = e z cos Θ + e φ sin Θ . (2)Substituting Eq. (2) into Eq. (1) and introducing the dimen-sionless coordinates s = r /R , ζ = z/R , ξ = ρ/R , anisotropy κ = KR /A , and DMI constant η = DM s R/ (2 A ) , wethereby obtain the expression for the dimensionless energy E = E/ (2 AR ) : E = (cid:90) d s ( ε (Θ) + ε (Θ)) , (3)where the energy densities ε and ε are given by ε = | ∂ ζ Θ | − V (Θ) + (1 + κ ) , (4) ε = | ∂ ξ Θ + η | + | ∂ φ Θ | . (5)The “potential” V , which appears in ε , is given by V ( θ ) = a cos 2( θ − δ ) , (6) a = (cid:2) (1 + κ ) + 4 η (cid:3) , tan 2 δ = − η κ , (7)where δ is taken between − π/ and π/ . Below we shall seethat δ determines the orientation of the twisted domains, while /a is the DW width.Next we look for a magnetization profile Θ which mini-mizes the energy E . The ε term vanishes (and is thus min-imized) by taking ∂ ξ Θ = − η and ∂ φ Θ = 0 . Then Θ is ofthe form θ ( ζ ) − η ( ξ − . In the thin-nanotube limit (andtaking κ, η = O (1) ), the ε term can be simultaneously min-imized by taking θ ( ζ ) to satisfy the Euler-Lagrange equa-tion θ (cid:48)(cid:48) = − V (cid:48) ( θ ) , subject to the boundary conditions that θ ( ±∞ ) correspond to maxima of V (not minima). Thesemaxima, given by θ = δ + nπ , describe the orientations ofmagnetization in domains. In the case of zero DMI, ie, η = 0 ,the magnetization far from the DW center is parallel to thenanotube axis. However, for η (cid:54) = 0 the magnetization profilebecomes helical, as shown in Fig. 1. Domain walls correspond to boundary conditions θ ( ±∞ ) describing oppositely oriented domains. There are four dis-tinct DW profiles, characterized by polarity σ and chirality χ (one is shown in Fig. 1), for more details see Appendix A. Po-larity determines whether the DW is head-to-head ( σ = 1 ) ortail-to-tail ( σ = − ), while chirality determines the sense ofrotation of m with increasing ζ , so that χ = σ sgn θ (cid:48) . TheEuler-Lagrange equation may be solved exactly to obtain θ = 2 χ arctan (cid:0) e σaζ (cid:1) + δ. (8)The DW profiles may be understood qualitatively in terms of amechanical analogy – see Fig. 2. We regard θ ( ζ ) as the trajec-tory of a particle moving in a potential V ( θ ) with ζ playing therole of time. In the static case, the DW boundary conditionscorrespond to the particle approaching consecutive maximaof V (located at δ mod π ) as ζ → ±∞ . At times in between,the particle traverses the intervening potential well (this is anexample of a so-called instanton orbit). θ / π -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 V − h e c o s θ -2.5-1.5-0.50.51.52.5 η = 0, h e = 0 η = 1, h e = 0 η = 1, h e = 1 η = 1, h e = 2 FIG. 2. (Color online) Mechanical analogy: The profile θ ( ζ ) maybe regarded as the instanton orbit of a particle θ taking infinite time η to move between consecutive local maxima of V ( θ ) − h e cos θ ( κ = 1 throughout). Dashed curve: With no DMI or applied field,the local maxima are θ = 0 and θ = π , corresponding to domainsaligned along the nanotube axis. The instanton orbit, indicated by thearrows, describes a tail-to-tail DW with negative chirality. Dotted-dashed curve: With DMI parameter η = 1 but no applied field, thelocal maxima are shifted by δ = − π/ , corresponding to twisted do-mains. The instanton orbit corresponds to a head-to-head DW withpositive chirality. Solid curve: With η = 1 and applied field h e = 1 ,the values of V − h e cos θ at consecutive maxima are no longer equal;a specific value j + v , the coefficient of linear damping in Eq. (15),is required to ensure that particle reaches the second maximum with-out overshooting. Hashed curve: For large enough h e , a maximumand minimum of V − h e cos θ coalesce, and the instanton orbit isdestroyed. An important characteristic of the DW is its width ∆ givenby /a , or in physical units, ∆ = √ A [( K + A/R ) + ( DM s /R ) ] / . (9)As is clear from this expression, DWs in thin nanotubes be-come sharper in the presence of DMI, in marked contrast tothe case of rectangular nanowires. DWs also become sharperin nanotubes with higher curvature /R . III. DYNAMICS
Under an applied current, the magnetization dynamics in aferromagnet far below the Curie temperature is described bythe Landau-Lifshitz-Gilbert (LLG) equation: ∂ m ∂t = γ H × m + α m × ∂ m ∂t − J ∂ m ∂z + βJ m × ∂ m ∂z , (10)where H = − ( M s ) − δE/δ m is the effective magnetic field, γ is the gyromagnetic ratio, α is the Gilbert damping constant, J is the current along the nanotube in units of velocity, and β is the nonadiabatic spin-transfer torque parameter. In theregime w, (cid:115) Aµ M s (cid:28) R (11)(as considered in the static case) and for currents J satisfying J (cid:46) γAM s R , (12)it can be shown that m lies nearly tangent to the nanotube,and Eq. (10) reduces to (see Appendix B for the details of thiscalculation): ∂ m ∂τ = − m × ( m × h t ) − j ∂ m ∂ζ . (13)Here τ = γAαM s R t is the dimensionless time, h t is the tan-gential component of the dimensionless effective field h = M s R A H , and j = M s R γA βJ is the dimensionless current. Notethat j is proportional to nonadiabatic torque parameter β ,whereas the current term itself assumes the adiabatic-torqueform. This has an important consequence, as described be-low.Proceeding as in Eq. (2), we write m = e z cos Θ+ e φ sin Θ with Θ( ζ, ξ, τ ) = θ ( ζ, τ ) + η ( ξ − to obtain ∂θ∂τ = ∂ θ∂ζ − j ∂θ∂ζ + V (cid:48) ( θ ) . (14)We look for traveling-wave solutions of the form θ ( ζ, τ ) = ϑ ( ζ − vτ ) describing axially symmetric DWs propagating withvelocity v . From Eq. (14), the profile ϑ satisfies ϑ (cid:48)(cid:48) = ( j − v ) ϑ (cid:48) − V (cid:48) ( ϑ ) (15)subject to the same boundary conditions as in the static case.It is easy to see that the moving profile ϑ coincides with thestatic profile θ in Eq. (8) with velocity v = j . In physicalunits, the DW velocity is given by V = βJ/α . (16)From Eqs. (11) and (12), Eq. (16) holds for velocities in theregime V (cid:28) γ βα (cid:112) µ A. (17) Thus, under an applied current, the DW propagates withoutdistortion and with velocity independent of polarity, chiralityand DMI. As the velocity approaches γ βα √ µ A , the magne-tization acquires a non-negligible radial component, and newbehavior can be expected to appear .Next we study the effect of an external magnetic field onDW dynamics. We take the field to be uniform along thenanotube axis, H e = H e e z . In the thin-nanotube limit, theapplied field generates an additional term in Eq. (15): ϑ (cid:48)(cid:48) = ( j − v ) ϑ (cid:48) − V (cid:48) ( ϑ ) − h e sin ϑ , (18)where h e = M s R A H e . The boundary conditions are modi-fied so that ϑ ( ±∞ ) correspond to consecutive maxima of amodified potential, V ( ϑ ) − h e cos ϑ . In terms of the mechan-ical analogy (Fig. 2), ϑ ( ξ ) again describes the trajectory of aparticle moving from one potential maximum to another, asabove. However, the potential difference at consecutive max-ima induced by the applied field is compensated now by theadditional (anti)damping term ( j − v ) ϑ (cid:48) . η -3 -2 -1 0 1 2 3 v -0.8-0.6-0.4-0.200.20.40.60.8 χ , σ = (+,+) χ , σ = (+,-) χ , σ = (-,+) χ , σ = (-,-) FIG. 3. (Color online) DW velocity v vs DMI parameter η for differ-ent chiralities χ and polarities σ ( h e = κ = 1 and j = 0 ). Changingthe sign of χ and η leaves v unchanged, while changing the sign of σ and η alters the sign of v . Numerical solutions of Eq. (18) show that the applied fieldcauses the DW velocity to depend strongly on DMI, chirality,and polarity, see Fig. 3. For a given field strength, the veloc-ity achieves a maximum for a nonzero value of η , and varieswith DMI through η by a factor exceeding 2. While Eq. (18)cannot be solved analytically, one can develop an expansionin powers of h e (see Appendix C for the details): v = j + σ √ κ + a a h e + χ πη a h e . (19)As shown in Fig. 4, this quadratic approximation is in goodagreement with the numerical results. In the limit of no cur-rent and DMI, j = η = 0 , it yields v = σh e / ( √ a ) in accordwith Ref. 55, or in physical units the DW velocity due to mag-netic field reads V = σγ √ α R (cid:112) KR /A H e . (20)In the limit of R → ∞ this expression reduces to a well knownresult for the velocities of transverse DWs in flat nanostrips. η -3 -2 -1 0 1 2 3 v h e = 1 . h e = 0 . h e = 0 . FIG. 4. (Color online) DW velocity v vs DMI parameter η for dif-ferent values of the external field h e ( σ = + , χ = + , κ = 1 , and j = 0 throughout). The solid curves are obtained from the numericalsolution of Eq. (18); the dotted curves are given by the approximateanalytical formula (19). h e v η = 0 η = 1 , χ = + η = 1 , χ = − η = 3 , χ = + η = 3 , χ = − FIG. 5. (Color online) DW velocity v vs. applied field h e for differ-ent values of the DMI parameter η and chirality χ ( σ = + , κ = 1 ,and j = 0 ). For η = 0 , v is given by the exact linear relation v = σh e /a . For η (cid:54) = 0 , v is obtained by solving Eq. (18) numeri-cally. Curves are computed up to the critical field h c . The dependence of the DW velocity on applied field andchirality is shown in Fig. 5. It follows from Eqs. (19) and(7) that for small fields, the DW velocity is suppressed by theDMI. However, for larger fields and chirality χ = σ sgn( ηh e ) ,the velocity may be enhanced by DMI (for the opposite chi-rality, v is always reduced).At a certain critical applied field h c , a bifurcation occurs,beyond which the DW velocity is suppressed. In terms of themechanical analogy of Fig. 2, as h e approaches h c , a maxi-mum and minimum of the potential V − h e cos ϑ coalesce, theinstanton orbit is destroyed, and the character of the travelingDW changes. This phenomenon has been discussed in termsof the spin-Cherenkov effect, and more recently in termsof pulled wavefronts of the KPP equation. It is straightfor-ward to obtain an analytic expression for h c in terms of η ,shown in Fig. 6. For η (cid:28) κ the leading-order behavior isgiven by h c = 1 + κ − cη ) / , with c = √ κ/ (2 √ ,while for η (cid:29) κ the leading-order behavior is h c = η . Theimportant conclusion is that the critical field can be enhancedby increasing DMI, thus allowing for faster DW propagation. η -3 -2 -1 0 1 2 3 h c (1 / , √ FIG. 6. (Color online) Critical applied field h c vs. DMI parameter η . IV. DISCUSSION AND CONCLUSIONS
In recent years there have been ongoing efforts to use fer-romagnetic materials with perpendicular anisotropy toproduce sharp and stable domain walls with a view to makepotential spintronic logic and memory devices more com-pact and faster. Here we have described an alternative ap-proach to this goal via domain walls in thin nanotubes withDzyaloshinskii-Moriya interaction. These DWs are foundto have novel properties: The domains themselves becometwisted about the nanotube, forcing the DWs to becomesharper with increasing DMI, the opposite of what is seen inthin nanowires. Under applied currents in the regime specified by Eqs. (11)and (12), these DWs propagate without distortion with a ve-locity proportional to the current. Applying a magnetic field,we find a rich dependence of DW velocity on polarity, chi-rality, DMI and field strength, which may provide enhancedcontrol in future spintronic devices. The DW velocity can besignificantly increased by DMI, and the onset of the magnonicregime suppressed.This work provides the favorable material trends for engi-neering nanotubes with DMI for faster and more robust DWoperation. Using the DMI parameter from Ref. 37 ( DM s =0 . · − J/m ), A = 10 − J/m, and taking the nanotuberadius R ≈ nm, we estimate the dimensionless DMI pa-rameter η ≈ . For the same material parameters κ = 1 isreached for K = 10 J/m . These estimates show that theregime where DMI has visible effects is experimentally feasi-ble.In the thin nanotube limit, we are able to treat leading con-tributions of dipolar interactions exactly. We have derived ex-plicit analytic expressions for the DW profiles and their veloc-ities under applied currents and fields that are in good agree-ment with numerical solutions of the LLG equation. Theseresults are robust and potentially applicable even beyond thethin-nanotube limit, which is hinted at by recent micromag-netic studies. ACKNOWLEDGMENTS
We thank G.E.W. Bauer, J. Barker, G.S.D. Beach, M. Kl¨aui,S.S.P. Parkin, and J. Sinova for helpful discussions. A.G.acknowledges the support of EPSRC Grant EP/K024116/1.J.M.R. and V.S. acknowledge the support of EPSRC GrantEP/K02390X/1. O.A.T. thanks KITP at University of Cali-fornia, Santa Barbara for hospitality, and acknowledges sup-port by the Grants-in-Aid for Scientific Research (Grants No.25800184, No. 25247056 and No. 15H01009) from MEXT,Japan; the NSF under Grant No. NSF PHY11-25915; andSpinNet.
Appendix A: Domain walls of different chirality and polarity
There are four distinct DW profiles, characterized by po-larity σ = ± and chirality χ = ± , see Fig. 7. The polaritydetermines whether the DW is head-to-head ( σ = 1 ) or tail-to-tail ( σ = − ), while chirality determines the sense of rotationof m with increasing ζ , so that sgn θ (cid:48) = σχ . Appendix B: Derivation of Eq. (13) for current drivendomain-wall motion
We start with rewriting the LLG equation, Eq. (10), in theLandau-Lifshitz (LL) form. Taking the vector product of m with both sides of the LLG equation, we obtain m × ∂ m ∂t = − γ m × ( m × H ) − α ∂ m ∂t − J m × ∂ m ∂z − βJ ∂ m ∂z . (B1)Then, combining LLG equation and Eq. (B1) to eliminate the m × ∂ m /∂t term, we find (1 + α ) ∂ m ∂t = − γ m × H − αγ m × ( m × H ) − (1 + αβ ) J ∂ m ∂z − ( α − β ) J m × ∂ m ∂z , (B2)which leads to the LL equation: ∂ m ∂t = − ˜ γ m × H − ˜ α m × ( m × H )+ ˜ J ∂ m ∂z + ˜ β ˜ J m × ∂ m ∂z , (B3)where ˜ γ = γ α , (B4) ˜ α = α ˜ γ = αγ α , (B5) ˜ J = − αβ α J , (B6) ˜ β = α − β αβ . (B7) Introducing dimensionless time τ (cid:48) through t = M s R γA τ (cid:48) = (1 + α ) M s R γA τ (cid:48) , (B8)the LL equation takes the form ∂ m ∂τ (cid:48) = − m × h − α m × ( m × h )+ j ∂ m ∂ζ + j m × ∂ m ∂ζ , (B9)where j = M s R A ˜ J ˜ γ = − (1 + αβ ) M s R γA J , (B10) j = M s R A ˜ β ˜ J ˜ γ = − ( α − β ) M s R γA J , (B11)and h = M s R A H , ζ = z/R .Taking (cid:15) = AR µ M s (cid:28) and j , j = O (1) , we have thatthe radial component of m is O ( (cid:15) ) , and m lies nearly tangentto the nanotube. Let us decompose h into its tangential andnormal components, h = h t + h n , where h t = h − ( h · e ρ ) e ρ and h n = ( h · e ρ ) e ρ . The LL equation (B9) is now equivalentto a system of two equations: ∂ m ∂τ (cid:48) = − m × h n − α m × ( m × h t ) + j ∂ m ∂ζ , (B12) − m × h t − α m × ( m × h n ) + j m × ∂ m ∂ζ . (B13)Equation (B12) is the projection of Eq. (B9) on the tangentspace of the cylinder, and Eq. (B13) is the projection ofEq. (B9) on e ρ direction. From Eq. (B13) we obtain α h n = m × h t − j m × ∂ m ∂ζ (B14)and consequently, m × h n = 1 α m × ( m × h t ) + j α ∂ m ∂ζ . (B15)The substitution of Eq. (B15) into Eq. (B12) yields ∂ m ∂τ (cid:48) = − (cid:18) α + 1 α (cid:19) m × ( m × h t ) + (cid:18) j − j α (cid:19) ∂ m ∂ζ = − (cid:18) α + 1 α (cid:19) (cid:20) m × ( m × h t ) + j ∂ m ∂ζ (cid:21) , (B16)where j = − (cid:18) α + 1 α (cid:19) − (cid:18) j − j α (cid:19) = β M s R γA J . (B17)Rescaling time once again, τ (cid:48) = (cid:18) α + 1 α (cid:19) − τ , (B18) z e φ e ρ e σ = +1, χ = +1 σ = +1, χ = −1σ = −1, χ = +1 σ = −1, χ = −1 (a) (b)(c) (d) FIG. 7. (Color online) Domain wall profiles in a thin nanotube with Dzyaloshinskii-Moriya interaction and polarity σ and chirality χ . Themagnetization lies tangent to the nanotube surface. so that t = α M s R γA τ , (B19)we obtain ∂ m ∂τ = − m × ( m × h t ) − j ∂ m ∂ζ . (B20)This is the central equation, Eq. (13), that we analyze through-out the rest of the paper. Appendix C: Expansion of domain wall velocity in magneticfield
While Eq. (18) cannot be solved analytically, it is straight-forward to develop an expansion in powers of h e . It turns outthat quadratic order is sufficient to capture the leading depen-dence on polarity and chirality. Letting p ( ϑ ( ζ )) = ϑ (cid:48) ( ζ ) , we may write Eq. (18) equivalently as ddϑ (cid:0) p + V − h e cos ϑ (cid:1) = ( j − v ) p. (C1)In terms of the mechanical analogy of Fig. 2, this correspondsto energy balance. Letting (cid:15) = h e , we expand p = p + (cid:15)p + (cid:15) p and v = v + (cid:15)v + (cid:15) v . At zeroth order we deducethat p ( ϑ ( ζ )) = θ (cid:48) ( ζ ) , where θ is the static profile, i.e. givenby Eq. (8). Then it follows that v = j . The equations forthe next two corrections p and p can be readily solved, and v and v are then obtained by integrating Eq. (C1) over theinterval δ < ϑ < π + δ and noting that p vanishes at theendpoints. Up to terms of order h e /a , we obtain v = j + σ √ κ + a a h e + χ πη a h e . (C2) ∗ [email protected] S. S. P. Parkin, M. Hayashi, and L. Thomas, Science , 190(2008); M. Hayashi et al.
Science , 209 (2008). D. A. Allwood, G. Xiong, M. D. Cooke, C. C. Faulkner, D. Atkin-son, N. Vernier, and R. P. Cowburn, Science , 2003 (2002). D. A. Allwood, G. Xiong, C. C. Faulkner, D. Atkinson, D. Petit,and R. P. Cowburn, Science , 1688 (2005). A. Yamaguchi, T. Ono, S. Nasu, K. Miyake, K. Mibu, andT. Shinjo, Phys. Rev. Lett. , 077205 (2004). G. S. D. Beach, C. Nistor, C. Knutson, M. Tsoi, and J. L. Erskine,Nat. Mater. , 741 (2005). M. Kl¨aui, C. A. F. Vaz, J. A. C. Bland, W. Wernsdorfer, G. Faini,E. Cambril, L. J. Heyderman, F. Nolting, and U. R¨udiger, Phys.Rev. Lett. , 106601 (2005). M. Hayashi, L. Thomas, C. Rettner, R. Moriya, X. Jiang, andS. S. P. Parkin, Phys. Rev. Lett. , 207205 (2006). L. Thomas, M. Hayashi, X. Jiang, R. Moriya, C. Rettner, andS. S. P. Parkin, Nature , 197 (2006). G. Meier, M. Bolte, R. Eiselt, B. Kr¨uger, D.-H. Kim, and P. Fis-cher, Phys. Rev. Lett. , 187202 (2007). R. Moriya, L. Thomas, M. Hayashi, Y. B. Bazaliy, C. Rettner, andS. S. P. Parkin, Nature Phys. , 368 (2008). S. A. Yang, G. S. D. Beach, C. Knutson, D. Xiao, Q. Niu, M. Tsoi,and J. L. Erskine, Phys. Rev. Lett. , 067201 (2009). O. A. Tretiakov, D. Clarke, G.-W. Chern, Y. B. Bazaliy, andO. Tchernyshyov, Phys. Rev. Lett. , 127204 (2008). J. Rhensius, L. Heyne, D. Backes, S. Krzyk, L. J. Heyderman,L. Joly, F. Nolting, and M. Kl¨aui, Phys. Rev. Lett. , 067201(2010). D. Ilgaz, J. Nievendick, L. Heyne, D. Backes, J. Rhensius,T. Moore, M. Ni˜no, A. Locatelli, T. Mentes¸, A. v. Schmidsfeld,A. v. Bieren, S. Krzyk, L. Heyderman, and M. Kl¨aui, Phys. Rev.Lett. , 076601 (2010). A. Singh, S. Mukhopadhyay, and A. Ghosh, Phys. Rev. Lett. ,067206 (2010). C. T. Boone, J. A. Katine, M. Carey, J. R. Childress, X. Cheng,and I. N. Krivorotov, Phys. Rev. Lett. , 097203 (2010). L. Thomas, R. Moriya, C. Rettner, and S. S. Parkin, Science ,1810 (2010). A. Goussev, J. M. Robbins, and V. Slastikov, Phys. Rev. Lett. ,147202 (2010). R. A. Duine, A. S. N´u˜nez, and A. H. MacDonald, Phys. Rev. Lett. , 056605 (2007). Y. Tserkovnyak and M. Mecklenburg, Phys. Rev. B , 134407(2008). M. E. Lucassen, H. J. van Driel, C. M. Smith, and R. A. Duine,Phys. Rev. B , 224411 (2009). Y. Nakatani, A. Thiaville, and J. Miltat, Nat. Mater. , 521 (2003). K. M. D. Hals, A. Brataas, and Y. Tserkovnyak, Europhys. Lett. , 47002 (2010). O. A. Tretiakov, Y. Liu, and A. Abanov, Phys. Rev. Lett. ,247201 (2012). J. Linder, Phys. Rev. B , 041412 (2014). M. Yan, A. K´akay, S. Gliga, and R. Hertel, Phys. Rev. Lett. ,057201 (2010). H. Min, R. D. McMichael, M. J. Donahue, J. Miltat, and M. D.Stiles, Phys. Rev. Lett. , 217201 (2010). M. C. Depassier, EPL (Europhys. Lett.) , 27005 (2015). R. D. Benguria and M. C. Depassier, “Reaction diffusion dy-namics and the schryer-walker solution for domain walls of thelandau-lifshitz-gilbert equation,” Arxiv: 1510.04927. M. Yan, C. Andreas, A. K´akay, F. Garc´ıa-S´anchez, and R. Hertel,Appl. Phys. Lett. , 122505 (2011). M. Yan, C. Andreas, A. K´akay, F. Garc´ıa-S´anchez, and R. Hertel,Appl. Phys. Lett. , 252401 (2012). J. A. Ot´alora, J. A. L´opez-L´opez, A. S. N´u˜nez, and P. Landeros,J. Phys. C , 436007 (2012). D. Ruffer, R. Huber, P. Berberich, S. Albert, E. Russo-Averchi,M. Heiss, J. Arbiol, A. Fontcuberta i Morral, and D. Grundler,Nanoscale , 4989 (2012). A. Buchter, J. Nagel, D. R¨uffer, F. Xue, D. P. Weber, O. F.Kieler, T. Weimann, J. Kohlmann, A. B. Zorin, E. Russo-Averchi,R. Huber, P. Berberich, A. Fontcuberta i Morral, M. Kemmler,R. Kleiner, D. Koelle, D. Grundler, and M. Poggio, Phys. Rev.Lett. , 067202 (2013). S. Da Col, S. Jamet, N. Rougemaille, A. Locatelli, T. O. Mentes,B. S. Burgos, R. Afid, M. Darques, L. Cagnon, J. C. Toussaint,and O. Fruchart, Phys. Rev. B , 180405 (2014). A. Thiaville, S. Rohart, ´E. Ju´e, V. Cros, and A. Fert, Europhys.Lett. , 57002 (2012). S. Emori, U. Bauer, S.-M. Ahn, E. Martinez, and G. S. D. Beach,Nat. Mater. , 611 (2013). K.-S. Ryu, L. Thomas, S.-H. Yang, and S. S. P. Parkin, NatureNanotech. , 527 (2013). A. Brataas, Nat. Nano , 485 (2013). J. Torrejon, J. Kim, J. Sinha, S. Mitani, M. Hayashi, M. Ya-manouchi, and H. Ohno, Nat. Commun. , 4655 (2014). I. Dzyaloshinsky, J. Phys. Chem. Solids , 241 (1958). T. Moriya, Phys. Rev. , 91 (1960). R. Streubel, J. Lee, D. Makarov, M.-Y. Im, D. Karnaushenko,L. Han, R. Sch¨afer, P. Fischer, S.-K. Kim, and O. G. Schmidt,Advanced Materials , 316 (2014). R. Streubel, L. Han, F. Kronast, A. A. ¨Unal, O. G. Schmidt, andD. Makarov, Nano Letters , 3981 (2014). J. A. Ot´alora, J. A. L´opez-L´opez, P. Vargas, and P. Landeros,Appl. Phys. Lett. , 072407 (2012). O. A. Tretiakov and Ar. Abanov, Phys. Rev. Lett. , 157201(2010). G. Tatara and H. Kohno, Phys. Rev. Lett. , 086601 (2004). Z. Li and S. Zhang, Phys. Rev. Lett. , 207203 (2004). G. Chen, T. Ma, A. T. N’Diaye, H. Kwon, C. Won, Y. Wu, andA. K. Schmid, Nat. Commun. , 2671 (2013). G. Gioia and R. D. James, Proc. R. Soc. Lond. A , 213 (1997). T. L. Gilbert, Phys. Rev. , 1243 (1955). T. L. Gilbert, IEEE Trans. Mag. , 3443 (2004). A. Thiaville, Y. Nakatani, J. Miltat, and Y. Suzuki, Europhys.Lett. , 990 (2005). See Supplementary Material for the movie of current driven DW. A. Goussev, J. M. Robbins, and V. Slastikov, Europhys. Lett. ,67006 (2014). A. L. Gonz´alez, P. Landeros, and A. S. N´u˜nez, J. Magn. Magn.Mater. , 530 (2010). M. Depassier, Europhys. Lett.108