Flexoelectricity

Big Picture

Flexoelectricity matters because it broadens the electromechanical design space beyond conventional piezoelectric materials. A non-uniform strain field — a strain gradient — generates electrical polarization even in materials that are piezoelectrically inactive. That removes the symmetry constraint that makes piezoelectricity available in only 20 of the 32 crystal classes. Every dielectric, regardless of crystal symmetry, can in principle exhibit flexoelectric response.

In the broader phonon landscape, flexoelectricity is a reminder that geometry itself creates new effective behavior. Bending is not just a mechanical action — at the right scale, it is also an electrical one. As devices shrink to nanometer scales and biological systems become engineering targets, this effect transitions from a correction term to a primary design lever.

This Document Covers

This document covers flexoelectricity as polarization from strain gradients: the physical mechanism and how it differs from piezoelectricity, the key coefficient and its measured values across material classes, the dramatic size-scaling behavior, biological flexoelectricity including the 2025 cell-membrane results, device applications and emerging commercial directions, current research frontiers, and the main reasons the effect remains underexploited in device design.

What The Effect Does

Flexoelectricity couples non-uniform deformation to electrical polarization. The critical distinction from ordinary piezoelectricity is that the relevant quantity is the spatial gradient of strain, not strain itself.

In piezoelectricity: P = d × ε (polarization proportional to uniform strain)

In flexoelectricity: P = μ × (∂ε/∂x) (polarization proportional to strain gradient)

Because the strain gradient, not the strain, drives the response, two immediate consequences follow:

1. Symmetry is no longer limiting — even centrosymmetric crystals (which cannot be piezoelectric) have a non-zero flexoelectric response, because strain gradients inherently break the local inversion symmetry the material globally possesses
2. Scale dependence is intrinsic — for a material bent to a given curvature, the strain gradient scales as 1/thickness. As a device gets thinner, the same bending produces larger strain gradients, and therefore larger polarization per unit deformation

This scale dependence is not a quirk — it is built into the definition of the effect and makes flexoelectricity increasingly important as device dimensions shrink.

The Flexoelectric Coefficient

The flexoelectric coefficient μ has units of C/m (coulombs per meter), or equivalently μC/m. It quantifies how much polarization is generated per unit strain gradient.

Measured values span a wide range:

The SrTiO3 result deserves emphasis: reducing a cantilever from millimeter to nanometer thickness increases the effective flexoelectric coupling coefficient by nine orders of magnitude. This is the size scaling made concrete — at the nanoscale, flexoelectric coupling can approach that of conventional piezoelectrics even in materials that have no piezoelectric response at all.

Near-Curie-temperature enhancement (as seen in BST) arises because polarization fluctuations diverge near the phase transition, amplifying the response far above the theoretical intrinsic value. The BST coefficients of 115–124 μC/m significantly exceed the values predicted from first principles alone.

Why Scale Matters

Flexoelectric response grows in relative importance as thickness decreases because strain gradient magnitudes scale inversely with thickness for a given curvature. This means:

• A 1 mm thick film bent over a 1 cm radius has a strain gradient of ~0.1 m⁻¹
• A 10 nm thick film bent over the same radius has a strain gradient of 10⁴ m⁻¹ — a hundred-thousand-fold increase

At nanoscale, flexoelectricity can produce polarizations comparable to or exceeding those from piezoelectricity in conventional materials. Structures previously treated as mechanically passive may not be electrically passive at all. This means much of the thin-film and nanoscale literature has been modeled incorrectly — flexoelectric contributions have been ignored or treated as noise when they may have been the dominant electromechanical mechanism.

Practically, this has two immediate engineering implications:

1. Reinterpret existing data — thin-film ferroelectric and dielectric measurements may have significant flexoelectric artifacts or contributions that have been attributed to other sources
2. Design for it deliberately — devices that exploit curvature, cantilever geometry, or intentional strain gradients can harness flexoelectricity without any symmetry restriction on material choice

Biological Flexoelectricity

Biological cell membranes are thin (~5 nm), highly curved, and constantly fluctuating. They are also made of lipid bilayers — centrosymmetric in their average structure and therefore not piezoelectric. But they are flexoelectric.

This has been recognized for decades in the context of hearing (the cochlea): outer hair cells in the inner ear are thought to convert sound-driven membrane curvature changes into electrical signals partly through flexoelectric coupling. The physical dimensions and curvature magnitudes are squarely in the flexoelectric-dominant regime.

More recently, the significance has been extended dramatically. A December 2025 study published in PNAS Nexus (Khandagale, Liu, Sharma) showed that thermally driven fluctuations in cell membrane curvature — the spontaneous molecular restlessness of a living membrane — can generate transmembrane voltages of up to 90 mV. This is comparable to the amplitude of action potentials in neurons (~70–100 mV). The voltage shifts occur within milliseconds, matching the timescale of neuronal firing.

The implications are significant:

• Flexoelectricity may be a physical mechanism for neuronal action potential generation or modulation, operating in parallel with or as a driver for the ion-channel-based picture
• Cells may generate electricity from passive thermal fluctuations — a biological energy harvesting mechanism at the cellular level
• Ultrasound-driven membrane flexoelectricity is being studied as a mechanism for non-invasive neuromodulation in treating Alzheimer’s and Parkinson’s disease — mechanical waves bend membranes, generating electrical signals that modulate neuronal activity without drug delivery

Industrial and Device Applications

Thin-film energy harvesters — bending thin dielectric films generates flexoelectric voltage. Films of BaTiO3, SrTiO3, and BST on flexible substrates can harvest mechanical energy without the symmetry constraints of piezoelectrics, opening a wider material design space.

MEMS sensors — cantilever-based accelerometers, pressure sensors, and flow sensors in MEMS devices may already have significant flexoelectric contributions that are currently unmodeled. Explicitly designing for flexoelectricity enables sensors from materials (silicon, alumina, common dielectrics) that don’t require piezoelectric coatings.

Flexoelectric actuators — applying a voltage gradient across a film produces a strain gradient (the converse flexoelectric effect), bending the film without the material needing to be piezoelectric. This enables actuation from a much wider class of materials.

Soft robotics — polymer-based flexoelectric films are intrinsically flexible and can be integrated into soft robotic architectures for actuation and sensing without the brittle ceramic layers required by piezoelectric designs.

Drug delivery — the ability to generate controlled electric fields from mechanical deformation, in a biocompatible centrosymmetric material, enables stimuli-responsive drug release triggered by deformation of an implanted device.

Neuromorphic and bioelectronic interfaces — the proximity of flexoelectric voltage scales to biological signaling voltages (90 mV from membrane fluctuations) raises the possibility of direct mechano-electrical interfaces with cells, using mechanical inputs to drive or record from neurons and muscle.

Nanocatalysis — a 2024 review in Small identified strain-driven flexoelectric nanocatalysts as an emerging application: the local electric fields generated by flexoelectricity at nanoscale surface features can modulate catalytic activity by influencing adsorbate binding and charge transfer at reaction sites.

Why The Effect Is Underexploited

The document treats flexoelectricity as one of the least fully integrated effects in device thinking, and this assessment remains accurate for several reasons:

Historical framing — flexoelectricity was for decades considered a second-order correction to piezoelectric response, worth noting in textbooks but not designing around. The recognition that it dominates at nanoscale required a community-wide recalibration that is still in progress.

Absence from design tools — finite element simulation tools for electromechanical devices (COMSOL, ANSYS) do not include flexoelectric coupling as a standard material property. Designers can only account for it through custom implementations, creating a practical barrier.

Measurement difficulty — separating flexoelectric contributions from piezoelectric, pyroelectric, and electrostatic artifacts requires carefully designed experiments. Many published thin-film measurements do not adequately control for this.

No dedicated commercial ecosystem — unlike piezoelectricity (commercial ceramics, datasheets, application notes), flexoelectricity has no commercial materials supply chain or device engineering community organized around it.

This underexploitation creates a clear opportunity: any team that correctly incorporates flexoelectricity into thin-film and nanoscale device design is working with a more complete physical model than the mainstream.

Current Research Frontiers

First-principles coefficient calculation — ab initio computation of flexoelectric coefficients is maturing, enabling screening of candidate materials without synthesis. This is accelerating the search for high-μ materials beyond the BaTiO3/BST family.

2D materials — graphene, MoS2, h-BN, and MXenes are atomically thin and exhibit curvature at essentially every scale. Their flexoelectric responses are being studied both as a probe of fundamental coupling and as a design resource for atomically thin electromechanical devices.

Converse flexoelectricity for actuation — applying a voltage gradient (not a uniform field) across a thin film produces controllable bending. This enables actuation from centrosymmetric materials, including amorphous dielectrics, which is entirely inaccessible to piezoelectric design.

Biological systems as design templates — the 2025 cell-membrane results suggest that biological evolution has been using flexoelectricity for electromechanical transduction in membranes for billions of years. Reverse-engineering these architectures for synthetic sensors and actuators is an emerging research direction.

Flexoelectric enhancement of piezoelectrics — in piezoelectric materials, flexoelectric contributions add to (or subtract from, depending on geometry) the piezoelectric response. Designing thin-film geometry to have both effects work cooperatively can produce higher effective coupling than either mechanism alone.

Connections to the Larger Landscape

• Piezoelectricity is the nearest comparison: flexoelectricity extends the electromechanical story to any dielectric and to any scale. At the nanoscale, the distinction between the two effects can blur, requiring careful experimental separation.
• Part I provides the most direct conceptual bridge: this effect is almost a textbook example of geometry acting like force, and of symmetry breaking being a design variable rather than a fixed material property.
• Thin-film and interface work in Part VII could benefit substantially from treating flexoelectricity as a design variable rather than a correction term — particularly in any system where films are thinner than ~100 nm.
• Biological and soft-material systems (relevant throughout the landscape) are where flexoelectricity transitions from an engineering option to a fundamental operating principle. The cell membrane result suggests it may be essential to the electrophysiology of living systems.