Journal of Physics D: Applied Physics

ACCEPTED MANUSCRIPT

Current-induced skyrmion motion on magnetic nanotubes

To cite this article before publication: Xiaofan Wang et al 2019 J. Phys. D: Appl. Phys. in press https://doi.org/10.1088/1361-6463/ab0c64

Manuscript version: Accepted Manuscript

Accepted Manuscript is “the version of the article accepted for publication including all changes made as a result of the peer review process,
and which may also include the addition to the article by IOP Publishing of a header, an article ID, a cover sheet and/or an ‘Accepted
Manuscript’ watermark, but excluding any other editing, typesetting or other changes made by IOP Publishing and/or its licensors”

This Accepted Manuscript is © 2019 IOP Publishing Ltd.

During the embargo period (the 12 month period from the publication of the Version of Record of this article), the Accepted Manuscript is fully
protected by copyright and cannot be reused or reposted elsewhere.
As the Version of Record of this article is going to be / has been published on a subscription basis, this Accepted Manuscript is available for reuse
under a CC BY-NC-ND 3.0 licence after the 12 month embargo period.

After the embargo period, everyone is permitted to use copy and redistribute this article for non-commercial purposes only, provided that they
adhere to all the terms of the licence https://creativecommons.org/licences/by-nc-nd/3.0

Although reasonable endeavours have been taken to obtain all necessary permissions from third parties to include their copyrighted content
within this article, their full citation and copyright line may not be present in this Accepted Manuscript version. Before using any content from this
article, please refer to the Version of Record on IOPscience once published for full citation and copyright details, as permissions will likely be
required. All third party content is fully copyright protected, unless specifically stated otherwise in the figure caption in the Version of Record.

View the article online for updates and enhancements.

This content was downloaded from IP address 150.216.68.200 on 07/03/2019 at 12:59

Page 1 of 5 AUTHOR SUBMITTED MANUSCRIPT - JPhysD-119501.R1

1
2
Current-induced skyrmion motion on magnetic nanotubes
3 Xiaofan Wang,1 X. S. Wang,1, a) C. Wang,1 Huanhuan Yang,1 Yunshan Cao,1 and Peng Yan1, b)
4 School of Electronic Science and Engineering and State Key Laboratory of Electronic Thin Films

pt
and Integrated Devices, University of Electronic Science and Technology of China, Chengdu 610054,
5
China
6
7 Magnetic skyrmions are believed to be the promising candidate of information carriers in spintronics. However,
8 the skyrmion Hall effect due to the nontrivial topology of skyrmions can induce a skyrmion accumulation or
9 even annihilation at the edge of the devices, which hinders the real-world applications of skyrmions. In this

cri
10 work, we theoretically investigate the current-driven skyrmion motion on magnetic nanotubes which can be
11 regarded as “edgeless” in the tangential direction. By performing micromagnetic simulations, we find that the
12 skyrmion motion exhibits a helical trajectory on the nanotube, with its axial propagation velocity proportional
to the current density. Interestingly, the skyrmion’s angular speed increases with the increase of the thickness
13
of the nanotube. A simple explanation is presented. Since the tube is edgeless for the tangential skyrmion
14
motion, a stable skyrmion propagation can survive in the presence of a very large current density without
15
any annihilation or accumulation. Our results provide a new route to overcome the edge effect in planar
16

us
geometries.
17
18
19 Ever since its experimental discovery,1 the magnetic two antiparallel magnetic domains coupled antiferromag-
20 skyrmion, a chiral quasiparticle,2,3 has been an active re- netically, through which skyrmions with opposite polari-
21 search area in condensed matter physics because of not ties can transform mutually.36 Under proper conditions,
22 only the potential for future spintronic applications such the domain boundary can also act as a reconfigurable
23
24
25
26
27
28
as skyrmion racetrack memories4–6 and logic devices,7,8

interaction (DMI) of two types:1,14–22 the bulk DMI and

an
but also the fundamental interests.9–13 In chiral magnets,
skyrmions can be stabilized by the Dzyaloshinskii-Moriya

the interfacial one. The bulk DMI typically exists in non-

centrosymmetric magnets, and can support the forma-
channel for skyrmion propagations.36 All these proposals
were aiming to eliminate the skyrmion Hall effect in a
planar geometry. Different from a planar strip with two
edges, a closed curved geometry, e.g., magnetic spheres
and/or cylinders, can be edgeless. In such geometries,
the skyrmion cannot vanish at edges any more, even
dM
tion of Bloch-type (vortex-like) skyrmions,1,15–18 while in the presence of the skyrmion Hall effect. Of course,
29
the latter one usually exists in inversion-symmetry break- the curvature of the non-planar geometry may induce
30
ing thin films, and can give rise to Néel-type (hedgehog- other interesting consequences like those in domain wall
31
like) skyrmions.19–22 dynamics.37,38 These facts motivate us to consider the
32 Several methods have been proposed to drive the skyrmion motion on a nanotube that a planar strip is
33 skyrmion motion, such as spin-polarized currents via spin rolled up, as shown in Fig. 1.
34 transfer torque or spin orbit torque,23,24 microwaves,25 In this work, we show, via micromagnetic simulations,
35 and thermal gradients,26 to name a few. However, when that the skyrmion can be created on magnetic nanotubes
36 the skyrmion is driven by an in-plane current via the and the skyrmion motion exhibits a helical trajectory
37 spin transfer torque, the trajectory of its motion deviates when it is driven by an electric current along the tube.
38 from the current direction due to the intrinsic skyrmion Further, we demonstrate that the skyrmion can travel
pte

39 Hall effect.2,3,27–30 Furthermore, there exists a threshold over arbitrarily long distances in the presence of a very
40 current density above which skyrmions can annihilate at large current density since the nanotube geometry are
41 the film edge.31 This edge effect strongly limits the speed edgeless. The skyrmion’s angular speed increases with
42 of skyrmion propagation which is of vital importance for the increase of the thickness of the nanotube, which is
43 real applications. Several solutions have been proposed different from the case in planar geometry.
44 to overcome this problem. Upadhyaya et al. showed that We consider the magnetic energy density in a nan-
45 the skyrmion can be guided in a desired trajectory by otube,
46 applying electric fields in a certain pattern.32 Barker et
ce

47 al. reported an antiferromagnetic skyrmion in antiferro- 2

magnetic materials.33 Zhang et al. proposed an antiferro- E = Aex |∇m| + Dm · (∇ × m) − K(m · ρ̂)2 + EDDI , (1)
48
49 magnetically exchange-coupled bilayer system, where the
skyrmions move straightly along the current direction.34 where m is the unit magnetization vector with a satu-
50 ration magnetization Ms , Aex is the ferromagnetic ex-
Zhang et al. reported the skyrmion Hall effect can be
51 change constant, D is the bulk Dzyaloshinskii-Moriya
depressed by imposing an additional pinning layer at the
52 interaction (DMI) strength, K > 0 is the easy-normal
edge of the track.35 More recently, Yang et al. discov-
Ac

53 ered a novel twisted skyrmion state at the boundary of anisotropy constant along ρ̂ direction, and EDDI is the en-
54 2
ergy density of dipole-dipole interaction. |∇m| is short
55 2 2
for |∇mx | + |∇my | + |∇mz | .
2
56 To study the current-driven magnetization dynamics,
57 a) Corresponding author: justicewxs@connect.ust.hk we solve the Landau-Lifshitz-Gilbert equation with the
58 b) Corresponding author: yan@uestc.edu.cn spin transfer torque τstt associated with the electric cur-
59
60
 AUTHOR SUBMITTED MANUSCRIPT - JPhysD-119501.R1 Page 2 of 5
2
1
rent flowing along the tube,39,40 mρ = 1 in the rest part, and then relaxed the system from
2
the initial state by minimizing the total energy. Numeri-
3 ∂t m = −γm × Heff + αm × ∂t m + τstt . (2) cal results are shown in Fig. 2, in which three phases are
4 identified: single domain of mρ = 1 (rhombuses), ordi-

pt
5 Here γ is the gyromagnetic ratio, α is the Gilbert damp- nary isolated skyrmion (circles), and stretched skyrmion
6 ing constant, and Heff = − MδE s δm
is the effective field. (squares) where the skyrmion is elongated like a spiral
7 The spin transfer torque τstt can be written as reaching the ends of nanotube. The typical magnetiza-
8 tion profiles are shown next to the phase diagram. When
τstt = −(vs · ∇)m + βm × (vs · ∇)m, (3)
9 D is small, the stable state is a single domain. The phase

cri
10 where vs = −µB pje /[eMs (1+β 2 )] is a vector with dimen- boundary between the single-domain phase and the ordi-
11 sion of velocity and parallel to the spin-polarized current nary skyrmion phase is mesh-size-dependent. That is be-
12 density je , e is the electronic charge, p is the polariza- cause when D is small, the skyrmion size is also small42,43
13 tion rate of the current, µB is the Bohr magneton, and so that the 2-nm mesh is not small enough to mimic the
14 β is the dimensionless parameter describing the degree of continuous model. To justify this, we have tested that
15 non-adiabaticity.39 when the mesh size is 1 nm, the stable state becomes
16 an ordinary isolated skyrmion for D = 1 mJ/m2 . This

us
17 rolling mesh-size-related annihilation of skyrmions is well-known
18 (a) (b) d and has been studied in detail.44 For an intermediate D,
19 an isolated skyrmion can exist. The sectional view cut
20 in xy plane (upper) and the front view after expanding
21 the tube into a plane (lower) are shown for D = 1.2
mJ/m2 and d = 40 nm. From the inner surface to the
22
23

an outer surface, the skyrmion size is getting larger. The

l

skyrmion size (or radius) is almost linearly dependent on

24
ρ. It can also be observed that the skyrmion is tilted to
25 the right [see also Fig. 4(a) below]. This is because of
26 the effective DMI induced by the curvature of the tube,
27 which will be explained later. When D is large enough,
28
dM
the stable state is a stretched skyrmion. The stretched
29 skyrmion forms a right-handed spiral on the tube when
30 D > 0 as shown in the figure, or a left-handed spiral
31 FIG. 1. (a) Schematic illustration of a Bloch-type skyrmion in
when D < 0 (not shown). The phase boundary between
a planar film. Green arrows refer to the local spin directions.
32 (b) Skyrmion on a nanotube by rolling up (a). Colors refer the ordinary skyrmion phase and the stretched skyrmion
33 to the ρ-component of the magnetization. The coordinate phase depends on the diameter of the tube, similar to the
34 system is defined in the inset. well known fact that the confinement effect of the sample
35 boundary is important in stabilization of the skyrmion
36 To visualize the skyrmion motion on magnetic nan- in planar films.45 In planar films, the upper limit of D
37 otubes, we performed micromagnetic simulations by em- for the existence of an isolated skyrmion is larger in a
38 ploying the MuMax3 package.40 The nanotube for nu- small sample than that in an infinite film.43,45 However,
the upper limit of D here in the nanotube is smaller than
pte

39 merical study is defined by fixed outer radius R = 50 nm,

40 various thickness d = 10 − 40 nm, and length l = 600 nm. that in an infinite film, probably because the skyrmion
41 The mesh size of 2 × 2 × 2 nm3 is used in our simula- size is larger than the diameter of the nanotube. The
42 tions. The magnetic nanotubes are assumed to be made skyrmion size increases with d due to the demagnetizing
of FeGe and the following material parameters are used:41 field, similar to the case in the planar geometry.42,43
43
44 exchange stiffness Aex = 8.78 pJ/m, saturation magne- To investigate the current-driven skyrmion dynamics in
45 tization Ms = 1.1 × 105 A/m, bulk DMI parameter infinite long nanotubes, we employ the periodic boundary
46 D varying from 0.8 mJ/m2 to 1.5 mJ/m2 , easy-normal condition in z-direction. The DMI is fixed to be D = 1.2
anisotropy parameter K = 2 × 105 J/m3 , and Gilbert mJ/m2 . A typical current-driven skyrmion motion in the
ce

47
48 damping constant α = 0.1. For the spin transfer torque, nanotube is plotted in Fig. 3(a) for d = 20 nm. We have
49 we assume p = 0.5 and β = 0.5. Figure 1 schematically compared the results obtained by using 2 nm and 1 nm
50 shows a Bloch-type skyrmion (a) in a planar film and mesh sizes and found no observable difference. So in the
(b) in a nanotube. The coordinate system is shown at calculations below, we use 2 nm mesh size. A skyrmion
51
the lower right corner of Fig. 1(b). ρ, φ and z represent is initially created at one end of the nanotube. Then we
52
Ac

the radial, tangential, and axial coordinates, respectively. inject an electric current along the axis of the nanotube.
53 The origin is set to be the center of the tube. The current exerts spin transfer torques on the magneti-
54 Firstly, in order to see how a skyrmion can exist on zation texture. Similar to the skyrmion motion in planar
55 magnetic nanotubes, we numerically calculate the phase film,2,3,27 the skyrmion moves not only along the current
56 diagram by tuning parameters D and d. We initially set direction, but also in the tangential direction at the same
57 mρ = −1 on the intersection of a 20-nm-diameter cylinder time because of the skyrmion Hall effect. As a result, the
58 along the half-line φ = 0, z = 0 and the nanotube, and set skyrmion trajectory follows a helical curve, as shown in
59
60
Page 3 of 5 AUTHOR SUBMITTED MANUSCRIPT - JPhysD-119501.R1
3
1
(a) (b)
2
3 mρ
4 +1
(c) (d)

pt
5
0
6
7 -1
8
9 single normal stretched

cri
10 domain skyrmion skyrmion

11
(e) (f)
12 FIG. 2. The magnetic phase diagram as a function of D and
d. Three phases are identified: single domain (rhombuses),
13
ordinary isolated skyrmion (circles), and stretched skyrmion
14 (squares). Representative magnetization profiles are shown at
15 the right panel. From left to right: single domain, ordinary
16 isolated skyrmion with both the sectional view (upper) and

us
17 the side view (lower), and stretched skyrmion.
18
19
20 the green line of Fig. 3(a). As a comparison, the tra-
21 FIG. 3. The trajectory of current-driven skyrmion motion
jectory of skyrmion motion in a planar film of the same in the nanotube (a) and the planar film (b). The current-
22
23
24
25
26
27
28
thickness is shown in Fig. 3(b). Due to the skyrmion
Hall effect, the skyrmion will annihilate at the edge when
the current density is large enough. To see more details,
we calculate the instantaneous velocity of the skyrmions
from the simulation results. Considering the outer sur-
face only, the skyrmion velocity v has two components:
the component parallel to the direction of the electric cur-
an dependence of skyrmion velocity, vk (c) and v⊥ (d), in which
d is fixed to 20 nm and the vertical grey line denotes the
annihilation of skyrmion at edges for the case of planar geom-
etry, where the velocity is calculated before the annihilation of
skyrmions. (e)-(f) The d-dependence of vk and v⊥ in both pla-
nar and tube geometries, in which j is fixed to 5 × 1012 A/m2 .
Symbols are numerical results. Solid lines are analytical re-
dM
29 sults obtained from Eqs. (6) and (7) for planar films and
rent, vk = v· ẑ, and a perpendicular one, v⊥ = v· φ̂ = Rω,
dashed curves are from Eq. (8).
30 where ω is the skyrmion’s “angular speed”. Figures 3(c)
31 and (d) show the current dependence of vk and v⊥ , re-
32 spectively. The two velocity components of the skyrmion
Fig. 3(f)]. See supplementary material MOVIE 1 for
33 in planar film are also plotted as comparisons (It is noted
skyrmion motions in different thicknesses.
34 that in planar film the velocity is measured before the
To better understand the numerical findings, we first
35 skyrmion annihilation at the edge). As the injected cur-
consider the effect of the curvature of the nanotube. Sup-
36 rent density j varies from 1×1010 A/m2 to 1×1013 A/m2 ,
pose the nanotube is constructed by many coaxial thin
37 both vk and v⊥ are proportional to j. In planar film, the
layers of tubes, with their thicknesses much smaller than
38 skyrmion annihilates when j is larger than 2×1011 A/m2 , their radii ρ. For each layer, we can express the energy
as shown by the vertical grey lines in Figs. 3(c) and (d).
pte

39 density in local coordinates on the outer surface of the

40 The corresponding longitudinal speed is no more than nanotube constructed by basis vectors (ρ̂, φ̂, ẑ).46,47 Intu-
41 100 m/s. However, because the tube is closed in the tan- itively, this means to expand the tube into a planar film.
gential direction, the skyrmion does not vanish even at a
42 To be more clear in comparison with a planar film, we
very large current density. The vk can reach ∼ 2000 m/s
43 rename the local basis vectors as ρ̂ → x̂0 , φ̂ → yˆ0 , and
when we inject a very large j = 1 × 1013 A/m2 . So the ẑ → zˆ0 . In the local coordinates, the energy density is
44
skyrmion motion in nanotube geometries possesses the
45
advantages that a stable skyrmion propagation can sur-
46 2
E = Aex |∇0 m| + Dm · (∇0 × m) − Km2x0 + EDDI
vive in the presence of a very large current density and
ce

47 the propagation speed can be very fast.


1 2

∂my0 ∂mx0

48 + Aex 2 + mx0 − my 0
We then fix the current density j = 5×1012 A/m2 , and ρ ρ ∂y 0 ∂y 0
49
50 investigate the d-dependence of skyrmion velocity at the Aex D
outer surface of the nanotube. The outer radius of the − 2 m2z0 + my0 mz0 , (4)
51 ρ ρ
nanotube is still R = 50 nm as above. Figures 3(e) and
52
(f) show the d-dependence of vk and v⊥ , respectively. For where ∇0 denotes the derivatives in primed coor-
Ac

53 different thickness d, vk is almost a constant [blue circles dinates. Compared to a planar film, the curva-
54 in Fig. 3(e)], and shows no apparent difference in com- ture induces three extra terms. The first term
55 parison with that in planar films [green squares in Fig. 2Aex (mx0 ∂y0 my0 − my0 ∂y0 mx0 ) /ρ comes from the ex-
56 3(e)]. However, v⊥ increases with d [blue circles in Fig. change interaction. For a fixed ρ, this term has the
57 3(f)], which is different from that in planar films, where same mathematical form as an interfacial DMI along ẑ
58 v⊥ stays unchanged when d increases [green squares in direction48 (the constant term Aex /ρ2 does not affect the
59
60
 AUTHOR SUBMITTED MANUSCRIPT - JPhysD-119501.R1 Page 4 of 5
4
1
2
magnetization profile). A left-handed Néel wall49,50 is (a)
thus preferred along ẑ direction. However, it is known
3
that the bulk DMI prefers a Bloch skyrmion. The su-
4

pt
perposition of this two effects therefore gives a skyrmion
5 tilting to the right, as schematically plotted in the Fig. d
6
7
4(a), which is consistent with the numerical observation (b) v⊥ R
O d

shown in Fig. 2. When the sign of D is reversed, the ro- R ω

8 tation direction of the Bloch skyrmion is also switched, so

d
9

d
the orientation of the tilting is flipped consequently. The

cri
10 stretched skyrmion also grows with a certain tilting direc-
v⊥ ω

11 tion for the same reason. The second term is an effective

12 easy-axis anisotropy along z (0) direction, due to which
13 the skyrmion is stretched along the axial direction of the FIG. 4. (a) Schematic diagram to explain the tilting of
14 tube. Note that Aex /2ρ2 is smaller than K for the param- the skyrmion as well as the rotation sense of the elongated
skyrmion. (b) Schematic picture to interpret why the linear
15 eters we used so that the easy-normal anisotropy K still
velocities are different in the planar and nanotube geometries.
16 dominates. The third term Dmy0 mz0 /ρ comes from the

us
The left panel shows the sectional view of the skyrmion motion
17 DMI, and it prefers that mz0 and my0 have opposite signs, in the planar geometry, in which v⊥ stays unchanged when the
18 which competes with the first term. We note that the thickness of planar geometry varies from d to d0 . The right
19 first Aex -term is approximately 2Aex my0 mz0 /(ρw) where panel shows the sectional view of the skyrmion motion in the
20 w ≈ πD/(4K) is the skyrmion wall width.42,43 For pa- nanotube geometry, in which the skyrmion’s angular velocity
21 rameters used in the simulations, 2Aex /w is much larger inside the nanotube is faster than the one outside the nan-
than D. Thus, the first Aex -term dominates over the D- otube, if layers are assumed to be decoupled.
22
23
24
25
26
27
28
term, and the contribution from the latter term can be
safely ignored.
Below, we analytically understand the motion of the
skyrmion on the nanotube. Let’s first consider a pla-
nar film, in which the skyrmion follows the Thiele’s
equation,2,23,29,31,34,51,52
an (e), which show that the parallel velocity components are
nearly the same in a planar film and in a tube. Numeri-
cally, the planar film gives Dxx = Dyy = 15.8, while the
flattened tube at d/2 gives Dxx = 17.4, Dxy = −2.2 and
Dyy = 16.4 . The longitudinal speed calculated with this
dM
29 G × (v − vs ) + D(αv − βvs ) = 0, (5) tensor D is hardly distinguishable from the one with the
30 scalar D. However, the perpendicular component v⊥ is
31 where G = 4πQx̂ is the gyrovector where Q = ±1 is the significantly different due to the curving nature of the
32 tube. This can be understood following the schematic
R
skyrmion number, Dij = ∂i m · ∂j mdydz is the dissipa-
33 tion tensor, and v is the skyrmion velocity. For a rota- diagrams in Fig. 4(b): Supposing that each layer is in-
34 tionally symmetric skyrmion, D degenerates to a scalar dependent to each other, the v⊥ is the same as that in
35 that can be calculated from the numerical results.29 In the planar film. However, for the layer at different ρ,
36 our simulations, the electric current is applied only along the angular speed ω(ρ) = v⊥ /ρ is larger (smaller) for
ẑ direction, and the Thiele’s equation can be easily solved. smaller (larger) ρ. Since the layers are strongly coupled
37
The parallel component (z-component) and the perpen- by exchange interactions and the skyrmion in each layer
38
dicular component (y-component) are is closely bounded together, the smaller angular speed in
pte

39 outer layers is dragged to become faster so that all layers

40 share the same angular speed in the end. As a result, v⊥
G2
 
β α−β
41 vk = + vs , (6) at the outer surface is faster than that in a planar film,
42 α α G2 + (αD)2
and it increases with d in a nanotube. The angular speed
43 (α − β)GD ω = v⊥ /R also increases with d for fixed outer radius R.
v⊥ = 2 vs . (7)
44 G + (αD)2 By fitting with the numerical data, we find that v⊥ as
45 well as ω can be estimated by considering the skyrmion
46 As shown by solid lines in Figs. 3(c)-(f), our analytical motion at the layer with radius R − d/2, i.e., the half-
ce

47 results obtained from Eqs. (6) and (7) are consistent with thickness of the tube wall. At ρ = R − d/2, the Thiele’s
48 the micromagnetic simulations.
equation gives v⊥ |ρ=R−d/2 = G(α−β)GD
2 +(αD)2 vs = ω(R − d/2).
49 In the nanotube, for each layer of radius ρ expanded to
a planar film, the skyrmion still follows the Thiele’s equa- Thus, we obtain
50
tion. The gyrovector G does not depend on ρ. Strictly (α − β)GDR
51 v⊥ |ρ=R = vs . (8)
speaking, the dissipation tensor D is no longer a scalar
52 [G2 + (αD)2 ] (R − d/2)
Ac

because the skyrmion is tilted, and its value also depends

53 on the skyrmion size. However, as long as the skyrmion In Figs. 3(d) and (f), the above expression for v⊥ is
54 stays in the isolated phase as shown in Fig. 2, this tilting plotted in dashed curves, which show excellent agreement
55 is small enough. The dependence of D on the skyrmion with the numerical results.
56 size is also weak. For the parameter we used in Fig. 3, it Magnetic nanotube is the key to test our theoretical
57 is good enough to adopt the scalar D in the correspond- predictions. Experimentally, there are several methods
58 ing planar film, as shown by solid lines in Figs. 3(c) and to produce the nanotube geometry. Sui et al. reported
59
60
Page 5 of 5 AUTHOR SUBMITTED MANUSCRIPT - JPhysD-119501.R1
5
1
that the nanotubes can be generated by hydrogen re- Li R, Jin C, Chen X, Zang J, Zhang Y and Tian M 2015 Nat.
2
duction in nanochannels of porous alumina templates.53 Commun. 6, 8504
3 19 Heinze S, von Bergmann K, Menzel M, Brede J, Kubetzka A,
Nielsch et al. proposed that the nanotubes can be made
4 by electrodeposition.54 Daub et al. reported that the
Wiesendanger R, Bihlmayer G and Blügel S 2011 Nat. Phys. 7,

pt
5 magnetic nanotubes can be synthesized by atomic layer
713
20 Romming N, Hanneken C, Menzel M, Bickel J E, Wolter B, von
6 deposition into porous membranes.55 Although it is still Bergmann K, Kubetzka A and Wiesendanger R 2013 Science
7 challenging to manufacture patterned nanotubes nowa- 341, 636
8 days, our proposal provides another paradigm to elim-
21 Jiang W, Upadhyaya P, Zhang W, Yu G, Jungfleisch M B, Fradin

F Y, Pearson J E, Tserkovnyak Y, Wang K L, Heinonen O, te

9 inate the drawbacks of skyrmion Hall effect and design

cri
Velthuis S G E and Hoffmann A 2015 Science 349, 283
10 the skyrmionic devices. 22 Krause S and Wiesendanger R 2016 Nat. Mater 15, 493

11 To conclude, we investigate the static and dynamic 23 Tomasello R, Martinez E, Zivieri R, Torres L, Carpentieri M and

12 properties of skyrmions on magnetic nanotubes. Through Finocchio G 2014 Sci. Rep. 4, 6784
24 Ado I A, Tretiakov O A and Titov M 2017 Phys. Rev. B 95,
13 micromagnetic simulations, we show that the electric cur- 094401
14 rent can drive a skyrmion propagation with a helical tra- 25 Wang W, Beg M, Zhang B, Kuch W and Fangohr H 2015 Phys.

15 jectory on the tube because of the skyrmion Hall effect. Rev. B 92, 020403
26 Kong L and Zang J 2013 Phys. Rev. Lett. 111, 067203
16 The skyrmion velocity is proportional to the injected cur-

us
27 Stone M 1996 Phys. Rev. B 53, 16573
17 rent without conventional upper limit. The skyrmion’s 28 Iwasaki J, Mochizuki M and Nagaosa N 2013 Nat. Nanotechnol.
18 angular velocity increases with the thickness of the nan- 8, 742
19 otube, which is different from the fact in the planar geom- 29 Iwasaki J, Mochizuki M and Nagaosa N 2013 Nat. Commun. 4,

20 etry. Our proposal of transporting the skyrmion in nan- 1463

30 Jiang W, Zhang X, Yu G, Zhang W, Wang X, Jungfleisch M
21 otube geometry will stimulate future design of skyrmionic
devices. B, Pearson J E, Cheng X, Heinonen O, Wang K L, Zhou Y,
22
23
24
25
26
27
28
ACKNOWLEDGMENTS

This work is supported by the National Natural Sci-

ence Foundation of China (NSFC) (Grants No. 11604041
an Hoffmann A and te Velthuis S G E 2017 Nat. Phys. 13, 162
31 Yoo M W, Cros V and Kim J V 2017 Phys. Rev. B 95, 184423
32 Upadhyaya P, Yu G, Amiri P and Wang K L 2015 Phys. Rev. B

92, 134411
33 Barker J and Tretiakov O A 2016 Phys. Rev. Lett. 116, 147203
34 Zhang X, Zhou Y and Ezawa M 2016 Nat. Commun. 7, 1023
35 Zhang Y, Luo S, Yan B, Ou-Yang J, Yang X, Chen S, Zhu B and

You L 2017 Nanoscale 9, 10212

dM
and 11704060), the National Key Research Development 36 Yang H, Wang C, Wang X, Wang X S, Cao Y and Yan P 2018
29 Program under Contract No. 2016YFA0300801, and Phys. Rev. B 98, 014433
30 the National Thousand-Young-Talent Program of China.
37 Otálora J A, López-López J A, Vargas P and Landeros P 2012

31 XSW acknowledges the support from NSFC (Grant No.

Appl. Phys. Lett. 100, 072407
38 Goussev A, Robbins J M, Slastikov V and Tretiakov O A 2016
32 11804045) and China Postdoctoral Science Foundation Phys. Rev. B 93, 054418
33 (Grant No. 2017M612932 and 2018T110957). 39 Zhang S and Li Z 2004 Phys. Rev. Lett. 93, 127204

34 1 Mühlbauer
40 Vansteenkiste A, Leliaert J, Dvornik M, Helsen M, Garcia-
S, Binz B, Jonietz F, Pfleiderer C, Rosch A, Neubauer Sanchez F and Van Waeyenberge B, 2014 AIP Adv. 4, 107133
35 A, Georgii R and Böni P Science 323, 915 (2009). 41 Wang W, Albert M, Beg M, Bisotti M -A, Chernyshenko D,
36 2 Sampaio J, Cros V, Rohart S, Thiaville A and Fert A 2013 Nat.
Cortés-Ortuño D, Hawke I and Fangohr H 2015 Phys. Rev. Lett.
37 Nanotechnol. 8, 839
3 Nagaosa N and Tokura Y 2013 Nat. Nanotechnol. 8, 899
114, 087203
38
42 Büttner F, Lemesh I and Beach G S D 2018 Sci. Rep. 8, 4464
4 Fert A, Cros V and Sampaio J 2013 Nat. Nanotechnol. 8, 152 43 Wang X S, Yuan H Y and Wang X R 2018 Commun. Phys. 1, 31
pte

39 5 Kang W, Huang Y, Zheng C, Lv W, Lei N, Zhang Y, Zhang X, 44 Lucia A D, Litzius K, Krüger B, Tretiakov O A and Kläui M
40 Zhou Y and Zhao W 2016 Sci. Rep. 6, 23164 2017 Phys. Rev. B 96, 020405(R)
6 Müller J 2017 New J. Phys. 19, 025002
41 7 Zhang X, Ezawa M and Zhou Y 2015 Sci. Rep. 5, 9400
45 Rohart S and Thiaville A 2013 Phys. Rev. B 88, 184422
46 Gaididei Y, Kravchuk V P and Sheka D D 2014 Phys. Rev. Lett.
42 8 Zhang X, Zhou Y, Ezawa M, Zhao G P and Zhao W. 2015 Sci.
112, 257203
43 Rep. 5, 11369 47 Kravchuk V P, Rößler U K, Volkov O M, Sheka D D, van den
44 9 Mruczkiewicz M, Gruszecki P, Zelent M and Krawczyk M 2016
Brink J, Makarov D, Fuchs H, Fangohr H and Gaididei Y 2016
45 Phys. Rev. B 93, 174429
10 Kim J, Yang J, Cho Y -J, Kim B, and Kim S -K 2017 Sci. Rep.
Phys. Rev. B 94, 144402
48 Huang S, Zhou C, Chen G, Shen H, Schmid A K, Liu K and Wu
46 7, 45185 Y 2017 Phys. Rev. B 96, 144412
ce

47 11 Yang W, Yang H, Cao Y and Yan P 2018 Opt. Express 26, 8778 49 Heide M, Bihlmayer G and Blügel S 2008 Phys. Rev. B 78,
48 12 Yang H, Wang C, Yu T, Cao Y and Yan P 2018 Phys. Rev. Lett.
140403(R)
121, 197201
49 13 Li Z -X, Wang C, Cao Y and Yan P 2018 Phys. Rev. B 98,
50 Emori S, Bauer U, Ahn S -M, Martinez E and Beach G S D 2013

Nat. Mater. 12
50 180407(R) 51 Thiele A A 1973 Phys. Rev. Lett. 30, 230
51 14 Rößler U K , Bogdanov A N and Pfleiderer C 2006 Nature (Lon- 52 Tretiakov O A, Clarke D, Chern G -W, Bazaliy Y B and Tch-

52 don) 442, 797 ernyshyov O 2008 Phys. Rev. Lett. 100, 127204
Ac

15 Yu X Z, Onose Y, Kanazawa N, Park J H, Han J H, Matsui Y,

53
53 Sui Y C, Skomski R, Sorge K D and Sellmyer D J 2004 J. Appl.
Nagaosa N and Tokura Y 2010 Nature (London) 465, 901 Phys. 95, 7151
54 16 Onose Y, Okamura Y, Seki S, Ishiwata S and Tokura Y 2012 54 Nielsch K, Castãno F J, Ross C A and Krishnan R 2005 J. Appl.
55 Phys. Rev. Lett. 109, 037603 Phys. 98, 034318
17 Park H S, Yu X, Aizawa S, Tanigaki T, Akashi T, Takahashi Y,
56 55 Daub M, Knez M, Goesele U, and Nielsch K 2007 J. Appl. Phys.
Matsuda T, Kanazawa N, Onose Y, Shindo D, Tonomura A and
57 Tokura Y 2014 Nat. Nanotechnol. 9, 337
101, 09J111
58 18 Du H, Che R, Kong L, Zhao X, Jin C, Wang C, Yang J, Ning W,
59
60