Twisted Charged Interfaces in Ferroelectrics

We have been seeking to explore the nature of interfaces in ferroelectrics, that not only support discontinuities in polarisation (and hence are "charged" and likely to show interfacial conduction), but also have lattice twists. We note that such interesting interfaces are not accessible through conventional epitaxial growth. Neither can naturally forming charged ferroelectric domain walls be twisted, since domains arise from a common parent lattice.
To create such new twisted, charged, conducting interfaces, we have therefore resorted to wafer bonding; z-cut single crystal wafers of lithium niobate have been selected for this purpose. There are several reasons for this choice of system: firstly, and most importantly, the melting temperature of lithium niobate is only slightly above its Curie temperature, so that effective thermomechanical bonding (sintering) can occur entirely within the ferroelectric state; secondly, lithium niobate is uniaxial and so, when wafers are stacked with opposing polarisations, reductions in electrostatic energy cannot occur through dipolar rotation (polarisation divergence should be maintained); thirdly, charged domain walls, which naturally form in lithium niobate, have already been seen to conduct many orders of magnitude more than the bulk and hence wafer-bonded charged interfaces should be conducting.
In this talk we will discuss the bonding process, and how, in general, it successfully allows clean charged interfaces, with sharp polar discontinuities, to be generated. Emergent interfacial conductivity, in head-to-head bonded interfaces, in particular, and the characteristics of the associated charge transport, will be discussed. The influence of relative twist on interfacial conductivity and the apparent twist-induced change from a conducting to an insulating state will also be presented, as will subtle details in dipolar configurations.