Improving the properties of these devices requires piezoelectric materials capable of delivering

a large longitudinal strain on the application of an electric field.

A large longitudinal strain of more than 1% is generally anticipated in suitably oriented single
crystals of specific compositions of ferroelectric materials.

Polycrystalline piezoceramics typically show a longitudinal strain of approximately 0.2–0.4%.

PbZrxTixO3.

PZT52.

We demonstrate that when the thickness of a polycrystalline piezoceramic is reduced to such an

extent that a large fraction of the grains are in the triaxial–biaxial crossover regime, the domain-
switching fraction increases considerably.

We show that oxygen vacancies in piezoceramics cause asymmetrical switching at the positive
and negative surfaces, which causes thin piezoceramics to bend. Here positive and negative
means the electrode connections.

Electric field→ domain switching→ grain deformation→ total deformation of ceramic.

It is a common perception that polycrystalline piezoceramics cannot exhibit a large longitudinal

electro strain comparable to single crystals because of the orientational averaging of the
anisotropic strain responses of dipoles and the hindrance to a change in the shape of the grains
due to mutual clamping.

Polycrystals.

In contrast, a volume element in the bulk is triaxially constrained, which limits its ability to
switch domains and elongate in the field direction.

Our strain and polarization measurement unit (Radiant Technologies Inc.) comprises a large-
area flat bottom electrode and a spring-loaded moving top electrode with a tip diameter of 3
mm gently pressing the centre of the piezoceramic disc.

Longitudinal: disc gets thicker/thinner → top electrode moves because of thickness change.

Bending: disc curves → centre moves relative to the rest of the disc.

They also measured the strain of the crystal lattice.

Interestingly, the strain behaved exactly like the domain switching:

More domain switching→more strain

The biaxial effect was till 100microm.

With thinner sample the domain switching as well as strain increased more, because now more
sample was will be in the biaxial range.

Also tested by computer simulation got the same result, by surface effect and mechanical
coupling.

Apart from switching the tetragonal domains, the field also induced a tetragonal to
rhombohedral transformation in PZT52.

We estimated the tetragonal domain-switching fraction using in situ XRD (Cu Kα1) with electric
field measurements under bipolar cycling.

Because in the thin sample, a large fraction of grains are close enough to the surfaces to
experience the biaxial–triaxial crossover.

They compared three different quantities as the thickness changed:

1. How much the domains switch back (Δη)

2. How much the crystal changes between tetragonal and rhombohedral (Δf)

3. How much reversible strain occurs (Srev)

Pb(Zr0.60Ti0.40)O3 (PZT60).

K₀.₅Na₀.₅NbO₃, BiFeO₃, BaTiO₃ and Na₀.₅Bi₀.₅TiO₃ 2-mm-thick specimens invariably displayed

asymmetric bipolar S–E loops with a large strain in the positive half-cycle and a small strain in
the negative half-cycle. Then they directly saw all non-pzt samples and confirmed and stated the
phenomenon of bending.

This means there wasn't a significant internal bias electric field pushing the domains
preferentially in one direction.

So the asymmetry in strain was not simply because the material had a built-in electric field.

We confirmed the large strain to be a measurement artefact … by demonstrating that when the
diameter was reduced from 12 to 5 mm, the displayed strain reduced from 2.5% to 0.25%.

The center of the body moved up, detected as if it is straining.

This confirmed the increased population of oxygen vacancies at the negative surface in
comparison to the positive surface on application of the field.

Other oxide materials, such as SrTiO₃, also show oxygen vacancies moving over short distances
at room temperature.
Pinning means oxygen vacancies can act like obstacles that make it difficult for a ferroelectric
domain to change its orientation.

First measure true strain.

So making PZT thinner initially increases the strain, but after a certain point, bending starts
reducing the true longitudinal strain.

The blocking transition is a crossover [1,2], in an ensemble of bistable elements, between two
states that respond distinctly to external drive.

In fact, FE-like signatures on the surface of the oxygen-deficient STO have been found at room
temperature.

PUCKERED OXYGEN ION??

It becomes largest around:

This maximum is associated with the blocking crossover.

In simple terms, around this temperature the dipoles have the right balance between:

 being able to respond/switch

 being slow enough to retain a history-dependent state

So the hysteresis becomes strongest.

Raman spectrum = a fingerprint of how the atoms in a material vibrate.

STO is actually a paraelectric down to T = 0 K but with an inclination towards an FE transition as

temperature approaches 0 K.
A ferroelectric material is a material that has a permanent electric polarization that can be
switched by applying an external electric field.

Because the surface environment is different, the surface atoms may rearrange slightly.

This rearrangement could make the surface behave like a ferroelectric, even though the bulk
STO is not ferroelectric.

The gate electric field can change the displacement of a puckered oxygen ions and their dipole
moment along the field, like in an FE.

Note that an earlier study by Sachs et al. ruled out ferroelectricity in the bulk STO using the
absence of hysteresis in the capacitance of Au/STO/Au parallel-plate capacitors.

There are two competing timescales:

1. Dipole switching time

How quickly can the dipoles switch?

2. Gate-voltage sweep time

How quickly are we changing the gate voltage?

As temperature increases, dipoles first become mobile enough to respond to the gate field,
causing hysteresis to increase; around , their switching rate matches the measurement
timescale and hysteresis is maximum; at still higher temperature they switch too rapidly and
lose their memory, so hysteresis decreases.

V_h is defined to be the difference between values of the backward and forward sweeps at a
drain current, used for hysteresis size measurement.

The counterclockwise hysteresis, observed at room temperature, decreases and vanishes with
cooling.

They tried both:

and

and got essentially the same results if while changing temperature.

So the temperature itself wasn't creating a major additional history effect.

On the other hand, the contribution of the bulk STO polarization to the transconductance will
reduce with rising temperature due to a decline in its dielectric constant.

The hysteresis peaks at , where the dipole response time and the sweep-time match.
The results on sample MSTO-RH indicates that surface roughness did not alter the polarization
dynamics much at the MoS₂-STO interface.

The effect of bulk STO polarization will dominate over that of the surface dipoles at low
temperatures, due to other one freezing.

That frozen polarization changes the electrical behavior of MoS₂.

Because an oxygen vacancy removes a negatively charged ion from the STO surface. This
changes the local electric forces and makes the remaining surface oxygen ions easier to displace
(“pucker”).