Spatially Resolved Polarization Manipulation of Ferroelectricity in Twisted hBN

Abstract

two layers of hexagonal boron nitride (hBN) together with a small rotation [5][6][7][8] and detected using piezo force microscopy (PFM), [9][10][11] electrostatic force micro scopy, [12,13] or nanoinfrared microscopy. [11] Besides hBN, other 2D transitional metal dichalcogenides with a small twist are also being actively explored owing to their ferroelectricity. [9,[14][15][16][17][18][19] Unlike the afore mentioned phenomena that require cryo genic temperature to protect the fragile quantum states, twisted ferroelectricity is robust at room temperature, which opens up the possibilities of facile utiliza tions. One of the most important applica tions is ferroelectric tunneling junction (FTJ), [20,21] which fits perfectly with the atomically thin twisted hBN (ThBN) with outofplane polarization. In this work, we used a conductive atomic force micro scope (CAFM) tip to probe the tunneling through different domains of ThBN, and found that the polarization dependence of the tunneling resistance contrasts with that in conventional FTJs. In addition, we demonstrated two different mechanisms for the manipulation of polarization in ThBN. First, the stress applied using the CAFM tip or stretching of the ThBN device can be utilized to move the domains, indicating the "sliding" origin of the ferroelectricity in ThBN. Second, a sufficiently large voltage applied at the tip can switch polarization, similar to the mechanism in classical ferroelectric materials. However, Robust room-temperature interfacial ferroelectricity has been formed in the 2D limit by simply twisting two atomic layers of non-ferroelectric hexagonal boron nitride (hBN). A thorough understanding of this newly discovered ferroelectric system is required. Here, twisted hBN is used as a tunneling junction and it is studied at the nanometer scale using conductive atomic force microscopy. Three properties unique to this system are discovered. First, the polarization dependence of the tunneling resistance contrasts with the conventional theory. Second, the ferroelectric domains can be controlled using mechanical stress, highlighting the original meaning of the emergent "slidetronics". Third, ferroelectric hysteresis is highly spatially dependent. The hysteresis is symmetric at the domain walls. A few nanometers away, the hysteresis shifts completely to the positive or negative side, depending on the original polarization. These findings reveal the unconventional ferroelectricity in this 2D system.

Theoretical designs of low‐barrier ferroelectricity

“…Such symmetry breaking leads to so-called sliding ferroelectricity, where the vertical polarizations are switchable via in-plane translation, that is, interlayer sliding. 6 Over the next few years, such ferroelectricity have been experimentally detected in a series of bilayer or multilayers including BN, [7][8][9] InSe, [10][11][12] transition-metal dichalcogenides (TMDs) like WTe 2 , [13][14][15][16] MoTe 2 , 17 MoS 2 , [18][19][20][21][22][23] ReS 2 , 24 and even multiwall nanotubes 25 and amphidynamic crystal. 26 Taking the AB stacking BN bilayer displayed in Figure 1a as an example, the N atom of the upper layer is over the hexagon center of the down layer, while the B atom of the upper layer is right over the N atom of the down layer (staggered with B over N), giving rise to inequality between two layers and interlayer charge transfer.…”

“…44 The low barriers far blow $k B T at room temperature may invite the question about their thermal stability. However, almost all the Curie temperature in the previous experimental reports on sliding ferroelectricity are above 300 K. [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23]25,26 In the ferroelectric bilayer configurations, the uniformity of local stacking translation can be ensured by the intralayer rigidity, enforcing all the local dipole moments aligned to the same directions. Taking WTe 2 bilayer as a paradigmatic case, as shown in Figure 2a, the artificial antiferroelectric state can be formed by in-plane inhomogeneous strain, where the translation is positive in one unitcell but negative in the other with opposite vertical polarizations.…”

Across‐Layer Sliding Ferroelectricity in 2D Heterolayers

“…In 2017, we proposed that for most non-ferroelectric 2D materials, certain stacking of bilayer or multilayer may break the symmetry and give rise to so-called sliding ferroelectricity, where the vertical polarizations are switchable via in-plane translation, i.e., interlayer sliding. [2] Over the next few years, such ferroelectricity have been experimentally detected in bilayer or multilayer BN, [3][4][5] configuration, the inversion symmetry  I is preserved and forbids the formation of vertical polarization, which will be broken if the two interfacial BN layers are parallel. We have compared the energies of different possible stacking configurations, and the ground state of graphene/BN interface turns out to be staggered with C over B atoms.…”

“…In 2017, we proposed that for most non‐ferroelectric 2D materials, certain stacking of bilayer or multilayer may break the symmetry and give rise to so‐called sliding ferroelectricity, where the vertical polarizations are switchable via in‐plane translation, i.e., interlayer sliding. [ 2 ] Over the next few years, such ferroelectricity have been experimentally detected in bilayer or multilayer BN, [ 3–5 ] transition‐metal dichalcogenides (TMDs) like WTe 2 , [ 6–9 ] MoS 2 , [ 10–16 ] MoTe 2 , [ 17 ] ReS 2 , [ 18 ] InSe [ 19 ] and even multiwall nanotubes, [ 20 ] and amphidynamic crystal. [ 21 ] This mechanism is applicable to most 2D materials except for mono‐element systems like graphene and phosphorene, as the inversion symmetry of their bilayers can always be maintained at any interlayer‐sliding vector.…”

Epitaxially ferroelectric hexagonal boron nitride on graphene

“…The realization of ferroelectricity at the atomic-limit thickness, i.e., two-dimensional (2D) ferroelectrics, has long shown great promise for device innovation in the post-Moore era, where efficient energy storage and dense digital information retrieval at reduced dimensions are crucial [1][2][3][4][5][6][7][8][9][10][11][12][13] .…”