Bloch-Point Domain Wall in Nanowires

• Bloch-point domain walls are 3D magnetic textures featuring a singular Bloch point where magnetization vanishes, enforcing a full azimuthal flux closure in cylindrical nanowires.
• They are experimentally identified via XMCD-PEEM and micromagnetic simulations, which distinguish them from transverse or vortex walls through their topologically distinct winding.
• Understanding their lattice pinning and dynamic behavior under applied fields and currents informs the development of high-speed, 3D magnetic devices.

Searching arXiv for recent and foundational papers on Bloch-point domain walls. A Bloch-point domain wall is a three-dimensional magnetic domain wall that contains a Bloch point, a point singularity in the magnetization field at which the local magnetization vanishes or, in the unit-vector description, becomes undefined as the texture covers all directions on a surrounding sphere. In cylindrical soft-magnetic nanowires this wall type emerges for sufficiently large diameters and is topologically distinct from transverse or transverse–vortex walls. Its defining features are full azimuthal flux closure, a nontrivial three-dimensional winding, and dynamics that differ qualitatively from planar-wall motion because the wall can be intrinsically pinned by lattice discreteness yet can also avoid the usual Walker breakdown under appropriate driving conditions (Kim et al., 2013, Col et al., 2013).

1. Topological definition and magnetic structure

A Bloch point is a point-like topological defect in a three-dimensional ferromagnet whose magnetization covers every direction on the unit sphere once as one goes around the defect. In the simplest hedgehog form near the origin,

$M(r)=\pm M_s \hat r,$

so that the unit vector $\hat m(r)=M(r)/M_s$ carries skyrmion charge

$$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$

(Kim et al., 2013). A complementary formulation describes the Bloch point as a singular point where the continuous magnetization vector field vanishes locally, $|m|\to0$, in order to accommodate all directions of magnetization on a closed surface around it; topological protection follows because the mapping $S^2\to S^2$ has non-zero degree and cannot be unwound without passing through $|m|=0$ (Col et al., 2013).

Within a cylindrical nanowire, a Bloch-point wall is a head-to-head or tail-to-tail $180^\circ$ wall in which the magnetization curls azimuthally around the wire axis and continuity requires a singular point on the axis. In contrast, a transverse–vortex wall has a core magnetization approximately transverse to the wire axis and no internal point singularity (Fruchart et al., 2018, Riz et al., 2020). The distinction is therefore not merely geometric but topological: the Bloch-point wall contains a true three-dimensional defect, whereas the transverse or transverse–vortex wall remains a smooth texture.

Several topological descriptors are used in the literature. On a two-dimensional cut one may define

$$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$

while for an isolated Bloch point one may assign a three-dimensional winding or degree, and in one short-nanowire formulation a quantity denoted $Q_{3D}$ is written as

$$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$

with value $\hat m(r)=M(r)/M_s$0 for an isolated Bloch point (Col et al., 2013, Caso et al., 2023). The same body of work distinguishes polarity, vorticity, and helicity as additional discrete or quasi-discrete characteristics of Bloch-point textures (Hermosa-Muñoz et al., 2023).

2. Stability in cylindrical nanowires

Bloch-point domain walls are characteristic of sufficiently thick cylindrical nanowires. In one formulation, a head-to-head wall contains a Bloch point when the wire radius satisfies $\hat m(r)=M(r)/M_s$1 a few exchange lengths $\hat m(r)=M(r)/M_s$2 (Kim et al., 2013). In another, the energy crossover occurs for diameters $\hat m(r)=M(r)/M_s$3, with $\hat m(r)=M(r)/M_s$4 for Permalloy (Col et al., 2013). A closely related statement is that for wire diameters larger than roughly seven times the dipolar exchange length, the Bloch-point wall is the ground-state wall because full three-dimensional flux closure strongly reduces its dipolar energy (Riz et al., 2020).

The micromagnetic energy functional used for soft materials is typically written as exchange plus magnetostatic terms, with anisotropy neglected for Permalloy-like systems:

$\hat m(r)=M(r)/M_s$5

$\hat m(r)=M(r)/M_s$6

(Col et al., 2013). A more general continuum form used for field-driven dynamics includes exchange, Zeeman, and magnetostatic self-energy:

$\hat m(r)=M(r)/M_s$7

(Fruchart et al., 2018).

For experimentally relevant Permalloy parameters, the exchange stiffness is reported as either $\hat m(r)=M(r)/M_s$8 with $\hat m(r)=M(r)/M_s$9 and $$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$0 (Kim et al., 2013), or $$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$1 with $$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$2 and $$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$3 (Col et al., 2013). These values set the characteristic scale for the Bloch-point core, described as exchange-limited and of order a few $$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$4, i.e. $$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$5 (Col et al., 2013).

Experiment and simulation identify two wall families in cylindrical nanowires: transverse walls in smaller-diameter wires and Bloch-point walls in larger-diameter wires. For example, transverse walls were imaged in wires of about $$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$6 diameter, whereas Bloch-point walls were identified in wires of about $$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$7 diameter (Col et al., 2013). This suggests that geometry-driven energy balance between exchange and magnetostatics is the primary selector of wall topology in soft cylindrical systems.

3. Experimental observation and identification

Direct experimental observation of Bloch-point domain walls in cylindrical nanowires was reported using combined surface and transmission XMCD-PEEM in released Fe$$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$8Ni$$q=\frac{1}{8\pi}\oint dA_i\,\epsilon_{ijk}\,\hat m\cdot(\partial_j \hat m\times \partial_k \hat m)=\pm1$$9 nanowires of diameter between about $|m|\to0$0 and $|m|\to0$1 and lengths of a few micrometers (Col et al., 2013). The imaging geometry used circular polarization at the Fe $|m|\to0$2 edge, grazing incidence $|m|\to0$3, and a contrast

$|m|\to0$4

with spatial resolution about $|m|\to0$5 in surface-sensitive PEEM mode and effective resolution about $|m|\to0$6 in transmission shadow mode because the shadow is inflated by $|m|\to0$7 (Col et al., 2013).

The Bloch-point wall was identified through a characteristic combination of signals. Surface XMCD showed a two-fold symmetric pattern corresponding to orthoradial curling of magnetization around the wire axis, while shadow XMCD showed a clear curling signature and zero contrast exactly on the wire axis, interpreted as the location of the Bloch-point singularity (Col et al., 2013). By contrast, the transverse wall displayed monopolar black/white contrast at the wall center in surface XMCD and a weaker, more asymmetric shadow signal without full orthoradial curling (Col et al., 2013).

Micromagnetic simulations post-processed by ray tracing reproduced both the surface and shadow contrasts. In those simulations the Bloch-point wall showed the same two-fold symmetry and zero-contrast streak at the axis, while the transverse wall reproduced the monopolar signature (Col et al., 2013). Quantitative line scans across the wall matched experiment within the $|m|\to0$8 resolution. The experimental result therefore established not only the existence of Bloch points in domain walls, but also an imaging protocol capable of distinguishing Bloch-point walls from topologically trivial cylindrical wall textures.

A related experimental study on diameter-modulated FeNi and CoNi wires used shadow XMCD-PEEM to identify both Bloch-point walls and transverse–vortex walls at rest, and then followed field-induced transformations between them above threshold fields (Fruchart et al., 2018). This broadened the experimental picture from static identification to topology-changing dynamics.

4. Lattice pinning and depinning physics

A central property of the Bloch-point domain wall is that the Bloch point has no intrinsic core size below the exchange length and therefore feels the discreteness of the underlying atomic lattice. For a simple cubic array of spins with lattice constant $|m|\to0$9, the exchange energy oscillates as the Bloch point moves along the axis and is minimized at the center of a cubic cell and maximized on a cell face. To leading order in $S^2\to S^2$0, the energy is

$S^2\to S^2$1

with amplitude

$S^2\to S^2$2

where $S^2\to S^2$3 is a numerical constant of order unity (Kim et al., 2013).

Differentiation gives the maximal pinning force

$S^2\to S^2$4

In compact form,

$S^2\to S^2$5

(Kim et al., 2013). Because the factor of $S^2\to S^2$6 cancels, the pinning force remains finite even as the computational mesh size becomes much smaller than the exchange length. OOMMF simulations explicitly confirmed that a finite depinning field remains even when the mesh size satisfies $S^2\to S^2$7, precisely because the Bloch point has no intrinsic core width (Kim et al., 2013).

For Permalloy with $S^2\to S^2$8, $S^2\to S^2$9, and $|m|=0$0, taking $|m|=0$1 gives

$|m|=0$2

(Kim et al., 2013). Equating this to the Zeeman force on a wall of magnetic charge $|m|=0$3,

$|m|=0$4

yields

$|m|=0$5

For $|m|=0$6 this gives

$|m|=0$7

that is, on the order of tens of oersted (Kim et al., 2013).

The same paper reports OOMMF depinning fields in the $|m|=0$8–$|m|=0$9 range for cylinders of radius $180^\circ$0–$180^\circ$1 and realistic $180^\circ$2 and $180^\circ$3, with instantaneous wall velocity oscillating with period $180^\circ$4 as the Bloch point traverses the mesh (Kim et al., 2013). This is direct numerical evidence for a periodic lattice-pinning potential. A common misconception is that refining the discretization eliminates the barrier; for Bloch points the reported result is the opposite, because the singularity remains sensitive to lattice discreteness even in the fine-mesh limit.

5. Field-driven and current-driven dynamics

Bloch-point walls are notable because their fully curling structure strongly suppresses the internal dipolar restoring torque that drives Walker breakdown in planar strips. Accordingly, under magnetic field or spin-polarized current a Bloch-point wall is expected to propagate in a steady below-Walker regime, and for spin-transfer motion the velocity is written

$180^\circ$5

(Riz et al., 2020). This is the basis for the long-standing expectation of very high domain-wall speeds in cylindrical conduits.

Experiment and simulation have examined the interplay between topology and dynamics in several regimes. In one study, field-driven motion above a threshold transformed a transverse–vortex wall into a Bloch-point wall through precessional motion of a surface vortex–antivortex pair, which merged and annihilated as an $180^\circ$6 winding number jumped from about $180^\circ$7 to about $180^\circ$8, signaling Bloch-point injection (Fruchart et al., 2018). In a $180^\circ$9 diameter simulated wire the threshold was $$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$0, while experiments gave about $$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$1–$$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$2 in FeNi and about $$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$3 in CoNi (Fruchart et al., 2018). The reverse transformation, Bloch-point wall to transverse–vortex wall, was observed experimentally but not reproduced in continuum micromagnetics, which instead left the Bloch-point wall stable or produced spiraling instabilities at high fields (Fruchart et al., 2018). This discrepancy has been attributed to the difficulty of treating point singularities, thermal activation, disorder, and atomistic exchange variations within standard micromagnetics.

Current-driven dynamics in cylindrical nanowires are strongly influenced by the Oersted field

$$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$4

inside a solid wire (Riz et al., 2020). This azimuthal field lowers the energy of the Bloch-point wall circulation aligned with it and can convert either a transverse–vortex wall or a Bloch-point wall of the wrong circulation into a Bloch-point wall of the correct circulation. The switching threshold obeys

$$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$5

with $$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$6, and for Permalloy wires of radius $$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$7 one finds $$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$8 (Riz et al., 2020). Once the appropriate circulation is established, simulated tubes of outer diameter $$Q_{2D}=\frac{1}{4\pi}\iint m\cdot(\partial_x m\times \partial_y m)\,dx\,dy,$$9 and inner diameter $Q_{3D}$0 show a perfectly linear $Q_{3D}$1 with speeds exceeding $Q_{3D}$2 for $Q_{3D}$3, in agreement with experiment (Riz et al., 2020).

A more extreme dynamical regime was reported in field-driven simulations of Fe nanowires, where three regimes appear: linear low-field motion for $Q_{3D}$4, velocity stagnation with spin-Cherenkov emission for $Q_{3D}$5, and re-acceleration for $Q_{3D}$6 (Tejo et al., 2023). The minimal magnon phase velocity was estimated at $Q_{3D}$7–$Q_{3D}$8, but velocities up to $Q_{3D}$9 were obtained for $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$0 through a reported jet-propulsion effect involving cone elongation, Bloch-point pair nucleation, and ejection of texture carrying momentum (Tejo et al., 2023). This suggests that in cylindrical geometries the absence of Walker breakdown does not by itself exhaust the available high-speed regimes.

6. Extensions, stochasticity, and controlled transformations

Although the canonical Bloch-point wall is a cylindrical-nanowire object, related Bloch-point-mediated walls appear in other confined geometries. In sub-$$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$1 perpendicularly magnetized disks driven by spin-transfer torque, finite-temperature micromagnetics identified a regime in which a propagating wall slows near the disk center and thermal fluctuations nucleate a Bloch point or Bloch line inside the wall (Bouquin et al., 2020). This event pins the wall for a long time and produces rare long-tail multiple-swing switching events with transition times $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$2 (Bouquin et al., 2020). The same work reports an incubation-delay mean $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$3 at $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$4 bias and an estimated relative width $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$5 (Bouquin et al., 2020). By choosing a disk diameter such that Bloch and Néel walls are degenerate at the center, specifically $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$6 for the stated parameters, the “retention pond” in phase space vanishes and Bloch-point nucleation is suppressed, yielding a narrow transition-time distribution with standard deviation $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$7 of the mean (Bouquin et al., 2020).

Short cylindrical nanowires provide another setting where Bloch-point walls coexist with competing topological states. For wires of diameter about $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$8 and length about $$Q_{3D}=\frac{1}{8\pi}\iiint \epsilon_{ijk}\,m_i\,\partial_x m_j\,\partial_y m_k\,dV$$9–$\hat m(r)=M(r)/M_s$00, simulations found two metastable configurations: a single-vortex state and a vortex-domain-wall state containing a Bloch point (Caso et al., 2023). Microwave excitation tuned to the lowest spin-wave mode at about $\hat m(r)=M(r)/M_s$01 for $\hat m(r)=M(r)/M_s$02 can induce the single-vortex to Bloch-point-domain-wall transition with threshold amplitude $\hat m(r)=M(r)/M_s$03–$\hat m(r)=M(r)/M_s$04 and pulse duration about $\hat m(r)=M(r)/M_s$05–$\hat m(r)=M(r)/M_s$06 (Caso et al., 2023). The reverse transition is driven by opposite spin currents with critical current density

$\hat m(r)=M(r)/M_s$07

for $\hat m(r)=M(r)/M_s$08–$\hat m(r)=M(r)/M_s$09 and $\hat m(r)=M(r)/M_s$10–$\hat m(r)=M(r)/M_s$11, with transition probability above $\hat m(r)=M(r)/M_s$12 in simulation (Caso et al., 2023). A Thiele-like model then gives a gyrotropic frequency $\hat m(r)=M(r)/M_s$13 without pinning and about $\hat m(r)=M(r)/M_s$14 with a modest pinning well, while micromagnetics finds a resonance near $\hat m(r)=M(r)/M_s$15 (Caso et al., 2023).

Bloch points in domain-wall-like environments have also been engineered in ferrimagnetic trilayers. X-ray vector magnetic tomography identified standard circulating Bloch points in a bottom Gd$\hat m(r)=M(r)/M_s$16Co$\hat m(r)=M(r)/M_s$17 layer, with $\hat m(r)=M(r)/M_s$18 and average $\hat m(r)=M(r)/M_s$19, and hyperbolic or twisted-hyperbolic Bloch points in a top Gd$\hat m(r)=M(r)/M_s$20Co$\hat m(r)=M(r)/M_s$21 layer, with $\hat m(r)=M(r)/M_s$22 between $\hat m(r)=M(r)/M_s$23 and $\hat m(r)=M(r)/M_s$24 and average $\hat m(r)=M(r)/M_s$25 (Hermosa-Muñoz et al., 2023). In that system the upper-layer Bloch points nucleate within a Néel-like exchange-spring domain wall about $\hat m(r)=M(r)/M_s$26 above an antiferromagnetically coupled interface, and the larger helicity is attributed to the combination of low saturation magnetization and long exchange length (Hermosa-Muñoz et al., 2023). This suggests that the internal character of a Bloch-point wall can be materially engineered through $\hat m(r)=M(r)/M_s$27, $\hat m(r)=M(r)/M_s$28, $\hat m(r)=M(r)/M_s$29, thickness, and interlayer coupling.

7. Significance and unresolved issues

Bloch-point domain walls occupy a distinctive position in nanomagnetism because they combine topological protection, singular micromagnetic structure, intrinsic lattice sensitivity, and unusual transport properties. Several implications follow directly from the cited results.

First, Bloch-point walls are not “ultra-soft” in the sense sometimes associated with conventional transverse walls. Because the atomic lattice generates a periodic potential with amplitude $\hat m(r)=M(r)/M_s$30, Bloch-point walls require a finite threshold field, typically about $\hat m(r)=M(r)/M_s$31–$\hat m(r)=M(r)/M_s$32 in Permalloy nanowires, to move (Kim et al., 2013). This matters for any domain-wall device based on cylindrical conduits.

Second, their dynamics can be more robust than those of planar walls once the correct topology and circulation are established. The absence of Walker breakdown and the linear law $\hat m(r)=M(r)/M_s$33 under spin transfer make them suitable for high-speed operation, while experimentally relevant current densities can stabilize the desired circulation through the Oersted field (Riz et al., 2020). This is consistent with proposals for three-dimensional racetrack-like architectures using cylindrical nanowires (Col et al., 2013).

Third, Bloch-point-mediated processes remain a challenge for continuum modeling. The experimentally observed BPW $\hat m(r)=M(r)/M_s$34 TVW conversion under field was not reproduced by finite-element micromagnetics, which points to unresolved issues in modeling singularity expulsion, atomistic barriers, finite temperature, and disorder (Fruchart et al., 2018). A plausible implication is that multiscale approaches combining atomistic spin dynamics near the Bloch point with micromagnetics elsewhere will remain important for quantitatively reliable predictions in this area.

Finally, the topic has expanded beyond simple ferromagnetic nanowires. Bloch-point domain walls or Bloch-point-mediated wall states now appear in reversal stochasticity studies, microwave-controlled short wires, and engineered ferrimagnetic exchange springs (Bouquin et al., 2020, Caso et al., 2023, Hermosa-Muñoz et al., 2023). This suggests that the Bloch-point domain wall is best understood not as a single static texture but as a broader topological motif governing wall stability, transformation pathways, and singular dynamics in three-dimensional magnetic systems.