---
title: Sliding Ferroelectricity in Layered Materials
url: https://www.emergentmind.com/topics/sliding-ferroelectricity
type: topic
---

# Sliding Ferroelectricity in Layered Materials

Sliding ferroelectricity is a form of ferroelectricity in which spontaneous polarization is generated and reversed by relative sliding of weakly coupled structural units—most commonly van der Waals layers, but also chain-like and cluster-assembled motifs—rather than by conventional ionic off-centering within a fixed unit cell. Its defining feature is that the ferroelectric order parameter is tied to stacking registry: two or more distinct registries break inversion symmetry in different ways, produce opposite or otherwise distinct polar states, and are connected by low-energy shear-like pathways. In contemporary literature, this mechanism has been established in bilayer and multilayer van der Waals systems, generalized to cluster-assembled and quasi-one-dimensional crystals, directly observed electrically in an amphidynamic bulk crystal, and coupled to topological, spintronic, electro-optic, and photovoltaic responses [2410.00776] [2302.03414] [2407.12235].

## 1. Symmetry origin and defining mechanism

The central symmetry principle is that sliding ferroelectricity arises when relative translation changes the registry between weakly bonded units and thereby changes the global symmetry from nonpolar to polar, or between two polar configurations with opposite polarization. In 3R-stacked bilayer MoS\(_2\), two stackings related by a mirror through the midplane, or equivalently by a specific in-plane translation, have equal \(|P_z|\) and opposite sign; polarization depends continuously on the interlayer shift vector \(\mathbf{u}\) along a path such as \(\mathbf{u}=t\,\tfrac{1}{3}(\mathbf{a}-\mathbf{b})\) [2410.00776]. In bilayer MBi\(_2\)Te\(_4\), the relaxed polar states AB′\(_1\) and AB′\(_2\) are related by lateral sliding from the high-energy AA′ configuration and carry opposite out-of-plane polarization \(P_\downarrow\) and \(P_\uparrow\) [2506.09317].

A key distinction from conventional displacive ferroelectrics is that the layers or subunits remain nearly rigid and the polarization is controlled by interfacial charge redistribution induced by registry, not by large ionic displacements along the polarization axis. This distinction is explicit in systems built from nonpolar monolayers such as 3R bilayer MoS\(_2\), where the out-of-plane dipole is almost entirely electronic and originates from asymmetric interlayer charge transfer [2410.00776]. It also appears in systems where the individual building block is not a simple 2D monolayer. In trilayer quasi-hexagonal C\(_{60}\), “cluster sliding ferroelectricity” is generated by the relative sliding and orientational arrangement of fullerene cluster layers, with different stackings realizing nonpolar, in-plane-polar, or simultaneously in-plane- and out-of-plane-polar phases [2407.13985]. In bulk NbI\(_4\), the same concept is generalized to a quasi-one-dimensional geometry: each NbI\(_4\) chain is nonpolar, but a particular relative binding arrangement of two van der Waals chains creates a polarization perpendicular to the chains, and reversal is achieved mainly by interchain sliding along the chain direction [2407.12235].

Sliding ferroelectricity is therefore not restricted to bilayers of centrosymmetric monolayers. Across-layer sliding ferroelectricity in graphene-based heterolayers shows that a centrosymmetric graphene bilayer can acquire switchable vertical polarization when embedded in a multilayer environment, because asymmetry of next-neighbor interlayer couplings makes the two graphene layers electronically inequivalent [2212.05182]. The broader implication is structural rather than chemical: whenever registry-dependent symmetry breaking and interfacial charge transfer generate a switchable dipole, sliding ferroelectricity becomes possible.

## 2. Polarization, switching coordinates, and dynamical behavior

The standard formal description of polarization in insulating sliding ferroelectrics follows the modern theory of polarization. In the Berry-phase form used for MBi\(_2\)Te\(_4\) bilayers,
\[
\mathbf{P}=\frac{ie}{(2\pi)^3}\sum_n^{\mathrm{occ}}\int_{\mathrm{BZ}} d\mathbf{k}\,\langle u_{n\mathbf{k}}|\nabla_{\mathbf{k}}|u_{n\mathbf{k}}\rangle,
\]
with the out-of-plane component \(P_z\) being the relevant quantity in most 2D cases [2506.09317]. For photoexcited sliding ferroelectrics, where partial occupation of valence and conduction bands invalidates the standard insulating Berry-phase treatment, direct integration of the total charge density along \(z\) is used instead, as in photoexcited 3R bilayer MoS\(_2\) [2410.00776].

Because these are effectively 2D systems, polarization is often reported in pC/m. Representative magnitudes span a wide range. Bilayer MBi\(_2\)Te\(_4\) gives \(|P_z|=0.571\) pC/m for GeBi\(_2\)Te\(_4\), \(0.615\) pC/m for SnBi\(_2\)Te\(_4\), and \(0.623\) pC/m for PbBi\(_2\)Te\(_4\) [2506.09317]. Bilayer CuF\(_2\) yields \(P_z=\pm 1.23\) pC/m [2603.10907]. Trilayer qHP C\(_{60}\) has out-of-plane polarization \(0.22\)–\(0.25\) pC/m in its fully polar stackings and in-plane polarization \(0.17\)–\(0.18\) pC/m in its in-plane-only ferroelectric stackings [2407.13985]. Janus monolayer In\(_2\)S\(_2\)Se supports two distinct non-degenerate sliding ferroelectric phases with \(P_{WZ′}=14.16\) pC m\(^{-1}\) and \(P_{ZB′}=-11.51\) pC m\(^{-1}\) [2507.22722]. In the quasi-one-dimensional bulk limit, NbI\(_4\) exhibits \(0.11\) \(\mu\)C/cm\(^2\) perpendicular to the Nb chains [2407.12235].

The switching coordinate is a shear or slip coordinate in real space, and the corresponding energy landscape is typically a low-barrier double well. In MBi\(_2\)Te\(_4\), the AB′\(_1\)\(\leftrightarrow\)AB′\(_2\) barriers are 51.30 meV/f.u. for GeBi\(_2\)Te\(_4\), 64.15 meV/f.u. for SnBi\(_2\)Te\(_4\), and 62.66 meV/f.u. for PbBi\(_2\)Te\(_4\) [2506.09317]. In photoexcited 3R bilayer MoS\(_2\), the dark and illuminated sliding barriers remain small, about \(17\)–\(21\) meV/u.c. [2410.00776]. In bilayer CuF\(_2\), the sliding barrier is \(3.83\) meV/atom [2603.10907]. In trilayer qHP C\(_{60}\), NEB barriers range from \(1\) to \(7\) meV/atom depending on the switching path [2407.13985].

The dynamics of switching are unconventional. In bilayer h-BN, the microscopic coupling between an out-of-plane electric field and an in-plane sliding coordinate is governed by off-diagonal Born effective charges, so that a vertical field produces lateral forces and switching follows an avalanche-like rather than climbing-like intrinsic-coercive-field rule [2505.09084]. In bilayer 3R-MoS\(_2\), domain walls are broad, with width \(\sim 100\) Å, and are described by a sine-Gordon field theory in the sliding coordinate; machine-learning-assisted molecular dynamics predicts uniformly accelerated motion under field, a relativistic-like velocity limit set by the in-plane transverse acoustic speed, and constant-velocity motion after field removal, i.e. undamped soliton-like domain wall propagation [2502.02137]. These features are unusual in the context of conventional ferroelectrics and follow directly from weak interlayer corrugation combined with strong intralayer stiffness.

## 3. Materials landscape and representative realizations

The known materials landscape spans symmetric and Janus 2D semiconductors, topological chalcogenides, altermagnetic fluorides, fullerene cluster solids, amphidynamic molecular crystals, and quasi-one-dimensional chain compounds. The table lists representative systems and the quantities explicitly reported in current work.

| System | Representative polar states or magnitude | Distinctive feature |
|---|---|---|
| 3R bilayer MoS\(_2\) | \(P_z \approx 0.82\) pC/m in the dark | Prototype sliding ferroelectric; photo-tunable [2410.00776] |
| Bilayer MBi\(_2\)Te\(_4\) | \(0.571\)–\(0.623\) pC/m | Ferroelectric QSH bilayers [2506.09317] |
| Bilayer CuF\(_2\) | \(P_z=\pm 1.23\) pC/m | Sliding-ferroelectric altermagnet [2603.10907] |
| Trilayer qHP C\(_{60}\) | OP \(0.22\)–\(0.25\) pC/m; IP \(0.17\)–\(0.18\) pC/m | Cluster sliding ferroelectricity [2407.13985] |
| Janus In\(_2\)S\(_2\)Se | \(14.16\) and \(-11.51\) pC m\(^{-1}\) | Non-degenerate sliding ferroelectric phases [2507.22722] |
| Bulk NbI\(_4\) | \(0.11\) \(\mu\)C/cm\(^2\) | Quasi-one-dimensional sliding ferroelectricity [2407.12235] |
| (15-Crown-5)Cd\(_3\)Cl\(_6\) | \(P_S \approx 0.3\)–\(0.4\) \(\mu\)C/cm\(^2\) | Direct electrical observation of sliding contribution [2302.03414] |

This diversity also clarifies several recurring misconceptions. Sliding ferroelectricity is not confined to bilayer van der Waals semiconductors, since it has been identified in cluster-assembled carbon allotropes, a bulk amphidynamic crystal with coupled geometric and sliding polarization, and a quasi-one-dimensional chain solid [2407.13985] [2302.03414] [2407.12235]. It is likewise not restricted to two exactly degenerate \(\pm P\) states. Janus In\(_2\)S\(_2\)Se hosts two strengthened and distinct non-degenerate sliding ferroelectric phases, WZ′ and ZB′, with different polarization magnitudes and different photovoltaic behavior [2507.22722]. Rhombohedral-stacked \(\gamma\)-InSe shows multiple sliding ferroelectricity, with \(n-1\) distinct sliding barriers in an \(n\)-layer stack and \(2^{n-1}\) thermodynamically stable polarization states along the full switching route [2407.21270].

## 4. Coupling to topology, spin transport, metallicity, and optics

One of the defining developments in the field is that sliding ferroelectricity is not an isolated structural phenomenon but a control parameter for other quantum responses. In bilayer MBi\(_2\)Te\(_4\) (\(M=\) Ge, Sn, Pb), spin-orbit coupling drives band inversion between Bi-\(p_z\) and Te-\(p_z\) states, and both polar states are quantum spin Hall phases. The spin-orbit-coupled bandgaps reach 31 meV for GeBi\(_2\)Te\(_4\), 36 meV for SnBi\(_2\)Te\(_4\), and 35 meV for PbBi\(_2\)Te\(_4\); the topological character is verified by \(Z_2=1\) from Wannier charge center flow and by gapless helical edge states [2506.09317].

Sliding ferroelectricity also couples strongly to spin transport. In bilayer 1T′-WTe\(_2\), it reversibly switches the signs and magnitudes of both conventional and anomalous spin Hall conductivities by shifting spin Berry curvature contributions near the \(\Gamma\)-X path. Reported anomalous spin Hall conductivities include \(\sigma_{xy}^{y}=45.62\) \((\hbar/e)\)S/cm and \(\sigma_{yx}^{y}=56.84\) \((\hbar/e)\)S/cm in monolayer 1T′-WTe\(_2\), enhanced in the bilayer to \(\sigma_{xy}^{y}=-96.77\) \((\hbar/e)\)S/cm and \(\sigma_{yx}^{y}=104.03\) \((\hbar/e)\)S/cm, with sign reversal under ferroelectric switching [2606.22015]. In bilayer CuF\(_2\), interlayer translation produces a switchable \(P_z\) that directly couples to and reverses d-wave altermagnetic spin splitting; the bilayer states FE-I and FE-II are \((+1,+1,+1)\) and \((-1,-1,-1)\) in the notation \((P,S,L)\), and quadrilayer CuF\(_2\) supports four such polarization–spin–layer states [2603.10907].

The electronic-structure consequences can extend to metallicity and excitonic order. In Janus sliding ferroelectric TMD bilayers and trilayers, the intrinsic electric field of each Janus monolayer modulates interlayer polarization and the electronic band structure; decreasing interlayer distance is identified as a major contributor to increasing polarization and reducing the band gap, and TeMoS trilayer reaches \(P_{CD}\approx 2.54\) pC/m with \(E_g=0.18\) eV [2505.22207]. In bilayer WTe\(_2\), a theoretical treatment that includes excitonic effects finds that exciton condensation contributes significantly to stabilizing the ferroelectric state upon sliding, increasing the effective energy difference between the ferroelectric and glide-mirror-symmetric structures from the \(\sim 0.3\)–\(0.4\) meV DFT scale to \(\approx 4\) meV at the equilibrium displacement and \(\approx 6\) meV at \(d=0.4\) Å [2510.01465].

Optical functionality is another major branch. In \(\beta\)-ZrI\(_2\), a prototype sliding ferroelectric for electro-optics, the clamped linear electro-optic response is dominated by the electronic term rather than the ionic term, with \(r_{33}^{\eta}=28.7\) pm/V and strain enhancement to \(43.2\) pm/V under \(+2\%\) biaxial strain [2510.03738]. This is explicitly opposite to the usual phonon-dominated picture of conventional ferroelectrics and implies ultrafast electro-optic response. The same work identifies a nearly linear anti-correlation between \(r_{33}^{\eta}\) and the band gap under both biaxial and uniaxial strain and reports a large elasto-optic coefficient \(p_{33}\approx 0.7\) that is independent of biaxial strain [2510.03738].

## 5. Experimental observation and characterization

Experimental confirmation has progressed from indirect nanoscale signatures to direct electrical measurements. The most explicit macroscopic demonstration is in the amphidynamic crystal (15-Crown-5)Cd\(_3\)Cl\(_6\), a van der Waals coordination polymer in which sliding ferroelectricity coexists with geometric ferroelectricity. The high-temperature phase is centrosymmetric \(P2_1/n\), the low-temperature phase is polar \(P2_1\), and a transition occurs at \(T_C \approx 320\) K. Direct electrical measurements show \(P\)–\(E\) hysteresis loops with saturated polarization \(P_S \approx 0.3\)–\(0.4\) \(\mu\)C/cm\(^2\) along \(b\), coercive field \(\sim 40\) kV/cm at 293 K increasing to \(\sim 60\) kV/cm at 273 K, and a DFT band gap \(>4\) eV that suppresses leakage and enables direct loop measurement [2302.03414]. DFT decomposition gives a geometric dipole \((0,0.112,0)\) eÅ, a bulk dipole \((0,0.263,0)\) eÅ, and a sliding contribution \((0,0.151,0)\) eÅ, i.e. 57.4% of the total [2302.03414].

In rhombohedral \(\gamma\)-InSe, multiple sliding ferroelectricity has been observed by dual-frequency resonance tracking PFM, scanning Kelvin probe microscopy, and conductive AFM. PFM resolves switchable out-of-plane polarization states; KPFM detects a retained surface-potential contrast of \(\sim 100\) mV after field removal; conductive AFM shows multiple discrete current steps consistent with multiple sliding states [2407.21270]. A graphene/\(\gamma\)-InSe/graphene tunneling device further demonstrates a tunable bulk photovoltaic effect, with photovoltaic current density of \(\sim 15\) mA/cm\(^2\), photo-responsivity of \(\sim 255\) A/W, and fast response suitable for real-time imaging [2407.21270].

Device demonstrations in 3R bilayer MoS\(_2\) have established the functional viability of sliding ferroelectric memories. Sliding ferroelectric memories on rigid and flexible substrates exhibit a memory window of \(>8\) V, a conductance ratio above \(10^6\), retention time of \(>10\) years, and programming endurance greater than \(10^4\) cycles; flexible devices maintain performance after bending over \(10^3\) cycles, support synapse-specific Hebbian forms of plasticity, and enable image recognition with 97.81% accuracy [2401.16150]. These results are particularly significant because they connect the structural switching coordinate directly to reproducible device metrics.

## 6. Modulation routes, non-degenerate phases, and future directions

External perturbations modify sliding ferroelectricity unusually strongly because the order parameter is tied to weak interlayer or inter-subunit registry. Optical excitation is one such route. In 3R bilayer MoS\(_2\), constrained-DFT calculations show that photoexcitation tunes \(P_z\) over a large range for a given sliding coordinate: from the dark value \(P_z \approx 0.82\) pC/m, to \(\sim 0.47\) pC/m at \(n_{\mathrm{ph}}=0.01\) e/u.c., down to \(\sim 0.03\) pC/m at \(n_{\mathrm{ph}}=0.2\) e/u.c., and back to \(\sim 0.4\) pC/m at \(n_{\mathrm{ph}}=0.8\) e/u.c.; at \(n_{\mathrm{ph}}\gtrsim 0.9\) e/u.c. a \(2\times 2\) reconstructed \(Cm\) phase appears with \(P_z\approx -4.8\) and \(+5.4\) pC/m in the two distorted structures [2410.00776]. The modulation is non-monotonic because photoexcited carriers and photoinduced structural distortion contribute differently at different carrier densities.

A second route is superlubricity engineering. In “superlubric sliding ferroelectricity,” inserting an intermediate layer such as graphene or BN between a sliding-ferroelectric homobilayer produces incommensurate interfaces and dramatically flattens the sliding potential. For 3R bilayer MoS\(_2\), the switching barrier drops from \(\approx 2.38\) meV/atom in the bare bilayer to \(\approx 0.0267\) meV/atom in MoS\(_2\)/graphene/MoS\(_2\) and \(\approx 0.167\) meV/atom in MoS\(_2\)/BN/MoS\(_2\), while the required switching voltage is reduced by about one order of magnitude [2501.16118]. The same work emphasizes a tradeoff already implicit in the field: ultralow barriers improve write energy and speed, but very small polarization or strong metallic screening can reduce readout contrast and retention.

A third route is phase engineering away from degenerate \(\pm P\) switching. Janus In\(_2\)S\(_2\)Se demonstrates that sliding ferroelectric phases can be non-degenerate and functionally distinct: the WZ′ phase, with enhanced polarization, gives superior photoelectric conversion efficiency in the visible region, while the ZB′ phase, with higher carrier mobility, a moderate band gap, and an indirect-to-direct transition, yields a marked red-shift and enhancement of the photocurrent peak in the infrared [2507.22722]. This establishes a broader design principle: sliding need not merely reverse an existing dipole but can connect different polar phases with different optical and transport functions.

Current research therefore points in several convergent directions. One is the systematic exploration of non-bilayer geometries, where cluster, chain, and multilayer degrees of freedom create many-state switching landscapes rather than simple Ising-like polarity [2407.13985] [2407.12235] [2407.21270]. Another is the use of sliding ferroelectricity as a nonvolatile control parameter for quantum responses, including topology, altermagnetism, spin Hall transport, and excitonic order [2506.09317] [2603.10907] [2606.22015] [2510.01465]. A third is the transition from proof-of-principle materials to robust platforms for memories, photovoltaics, electro-optics, and neuromorphic devices, where the distinctive advantages are atomically thin form factors, low sliding barriers, broad tunability, and the possibility of electrically controlling functionalities that in conventional ferroelectrics are not directly tied to a simple shear coordinate [2401.16150] [2510.03738].

Source: https://www.emergentmind.com/topics/sliding-ferroelectricity