Giant soft-memory in liquid crystal nanocomposites

Ravindra Kempaiah,1 Yijing Liu,1 Zhihong Nie,1,a) and Rajratan Basu2,b)

1Department of Chemistry and Biochemistry, University of Maryland, College Park, Maryland 20740, USA2Department of Physics, The United States Naval Academy, Annapolis, Maryland 21402, USA

(Received 21 December 2015; accepted 5 February 2016; published online 23 February 2016)

A hybrid nanocomposite comprising 5CB liquid crystal (LC) and block copolymer (BCP)

functionalized barium titanate ferroelectric nanoparticles was prepared. This hybrid system

exhibits a giant soft-memory effect that was detected by dielectric hysteresis. Spontaneous polar-

ization of ferroelectric nanoparticles couples synergistically with the radially aligned BCP chains

to create nanoscopic domains where LC mesogens can align directionally. Such domains can be

rotated electromechanically and locked in space even after the removal of the applied electric field.

The resulting non-volatile memory is several times larger than the non-functionalized sample and

provides an insight into the role of non-covalent polymer functionalization on enhancing the size of

the nanoscopic domains. VC 2016 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4942593]

The use of liquid crystals (LCs) for electronic data stor-

age has attracted interest in both fundamental and applied

research communities.1–6 Inorganic nanoparticles (NPs) in

LC matrices bear a great potential for this application due to

the synergistic interactions between NPs and LCs.7–10

Creating programmable memory devices holds great promise

in multi-level data storage systems.11 Non-volatile memory

effects in the nematic phase of ferroelectric LCs doped with

bare and polymer capped gold nanoparticles (Au NPs) have

been studied using low-frequency dielectric spectros-

copy.12,13 The presence of ferroelectric NPs (FNPs) like bar-

ium titanate (BaTiO3) enhances electro-optical properties of

LC composites without distorting the global nematic direc-

tor.14–19 In the nematic phase of nanocomposites, the global

director trumps individual molecular orientation and the

effect of FNPs on the molecular orientation is hard to distin-

guish.16,20 However, non-volatile memory effect has been

reported in the isotropic phase of nanocomposites made of

5CB (4-cyano-40-pentylbiphenylcarbonitrile) LC and FNPs

of BaTiO3.21 When polymers are attached on the surface of

FNPs like BaTiO3, the system is found to exhibit interesting

physical22,23 and chemical properties.24–26

Here, we report a giant soft-memory effect, studied by

the dielectric hysteresis effect, found in 5CB LC doped with

block copolymer (BCP) functionalized BaTiO3 FNPs. The

soft-memory effect we are referring to in this report is a non-

volatile electromechanical effect at the interface of LC mole-

cules and FNPs’ surface.

The 5CB LC, obtained from Sigma Aldrich, has a nematic

to isotropic transition (NI) temperature TNI of 35 �C. BaTiO3

FNPs with a diameter of 50 6 5 nm were purchased from U.S.

Research Nanomaterials Inc. BCP of polyethylene oxide-b-

polystyrene (PEO45-b-PS670-SH) containing a small polyethyl-

ene oxide block (45 repeating units) and a long polystyrene

(670 repeating units) terminated with a thiol functional group

was synthesised following the reversible addition-

fragmentation chain transfer (RAFT) polymerization procedure

as described in the previous report. Another type of BCPs of

polystyrene-b-poly (acrylic acid) (PS-b-PAA) with varied poly-

styrene block length was also produced using the same syn-

thetic approach.27

Surface functionalization of BaTiO3 FNPs was per-

formed in several stages. First, 5 mg BaTiO3 powder was dis-

persed in N,N-dimethylformamide (DMF) and sonicated for

3 h. Sonicated solution was centrifuged at 6000 rpm for

20 min to collect the FNPs and washed with tetrahydrofuran

(THF) for three more cycles. Finally, the FNPs were redis-

persed in 5 ml of THF. This sample was used a control for

measuring the soft-memory effect arising from pristine, non-

functionalized FNPs. Note that this functionalization process

of BaTiO3 for our experiment was carried out at room tem-

perature. Therefore, it is unlikely that BaTiO3 FNPs’ surface

crystal structure would be modified by this process without

reaching above its Curie temperature, 130 �C.28

To attach the polymer ligands, 2.5 mg of amphiphilic

PEO45-b-PS670-SH BCP was dissolved in 5 ml of THF and it

was mixed with 5 ml premade THF solution containing non-

functionalized BaTiO3 FNPs. The mixture was sonicated for

30 min and left for 1 h to promote attachment of the ligands

on FNPs’ surface. The result is a homogeneous colloidal so-

lution containing BaTiO3 FNPs with BCPs tethered to their

surface. Block copolymer chains attach to the FNPs surface

via intermolecular interaction and/or van der Wall’s forces.

The solution was thermally annealed and degassed to remove

any residual organic solvent and sonicated again for an hour.

Three concentrations for 5CB/BaTiO3 and 5CB/BCP-modi-

fied-BaTiO3 were prepared for investigation, i.e.,

c1¼ 0.175 wt. %, c2¼ 0.275 wt. %, and c3¼ 0.375 wt. %. The

surface topology of these polymer-modified FNPs can be

affected by the external stimuli, and hence, both kinetics and

thermodynamic pathway play a role in alignment of meso-

gens around these particles.29

Dielectric measurements were carried out using an auto-

matic liquid crystal tester (Instec, Inc.) that was attached to a

precision temperature controller. Commercially available LC

cells (SA100A200uG180, planar rubbed) with 1 cm2 semi-

transparent indium tin oxide electrode area, 1� pretilt angle,

a)znie@umd.edub)basu@usna.edu

0003-6951/2016/108(8)/083105/5/$30.00 VC 2016 AIP Publishing LLC108, 083105-1

APPLIED PHYSICS LETTERS 108, 083105 (2016)

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 136.160.90.131 On: Tue, 23 Feb 2016 14:38:28

and a spacing d¼ 20 lm were used for our experiments.

Scanning electron microscopy (SEM) imaging was done

using Hitachi SU-70 field emission scope.

The 5CB LC shows positive dielectric anisotropy,

De¼ ejj�e?, where ejj and e? are the dielectric components

parallel and perpendicular to the nematic director, respec-

tively. In a parallel-plate planar cell, 5CB exhibits

Fr�eedericksz transition, where under an applied electric field

(E) the nematic director orients from a planar configuration

(e?) to a homeotropic configuration (ek ). This director reor-

ientation occurs because the LC experiences a torque propor-

tional to DeE2 in the presence of E.30 When the E is switched

off, the director reorients back to planar configuration due to

the LC’s long range splay elastic interaction and the surface

anchoring mechanism in the cell. When the temperature is

raised above TNI, the global nematic order is destroyed (i.e.,

De¼ 0) and the Fr�eedericksz transition is no longer observed.

Figure 1 shows the dielectric constant e as a function of

E (f¼ 1 kHz) in nematic phase for pure 5CB, 5CB doped

with bare BaTiO3 (5CB/BaTiO3) and 5CB doped with poly-

mer functionalized BaTiO3 (5CB/BCP-modified-BaTiO3), at

a doping concentration of 0.275 wt. % of FNPs. A typical

Fr�eedericksz transition with a threshold field of 0.04 V/lm�1

is observed for all three samples. No dielectric hysteresis

was observed on turning E down to zero.

Since BaTiO3 FNPs carry a spontaneous polarization

P¼ 0.26 Cm2,31 they enable very strong local electric fields

�1010 V m�1 near the surface. This local field, EFNP, attenu-

ates as 1/r3. The presence of EFNP creates nanoscopic domains

where LC mesogens orient along EFNP surrounding the FNPs,

and we call these regions—pseudonematic domains. In the ne-

matic phase, these short-range domains align with the global

nematic director to reduce the free energy of the nematic ma-

trix. These domains collectively increase the nematic orienta-

tional order, increasing De in the nematic phase of the

LCþFNP systems, as shown in Figure 1.

The dipole moment magnitude for a 5CB molecule is

p¼ 6.5D¼ 2.15� 10�29 C m.32 When the LC molecules

align with EFNP close to the FNP, the associated energy can

be written as UFNP¼�~p:~EFNP��10�19 J. The thermal

energy in the isotropic phase of 5CB at T¼ 42 �C¼ 315 K

(well above the transition temperature) is Uthermal

� kBT� 10�21 J. Apparently, the thermal energy is too small

to eliminate the FNP-induced LC-order in the isotropic

phase. Due to the presence of these domains, the isotropic

phase of 5CB/BaTiO3 maintains a net De and is expected to

interact with the external E.

Figure 2 shows e as a function of E in the isotropic phase

(T¼ 42 �C) for pure 5CB, 5CB/bare BaTiO3, and two differ-

ent 5CB/BCP-modified-BaTiO3 samples as listed in the

legend. Pure 5CB shows a featureless behaviour in the

isotropic phase as expected. The hybrid systems show an

increase in e above a threshold field, exhibiting a

Fr�eedericksz-like transition. The dielectric constant for these

nanocomposite systems does not relax back to its original

value on turning the E down to zero, manifesting a dielectric

hysteresis effect. The hysteresis area correlates to the soft-

memory effect in the hybrid samples. The memory effect for

5CB/BaTiO3 can be seen from the (� red) curve in Figure 2;

the De holds its value (Deiso5CB =FNP ¼ 0:4) even when the E is

switched off. The hysteresis curve for 5CB/PEO45-b-PS670-

SH polymer functionalized BaTiO3 FNPs (� blue curve)

clearly shows a giant increase in the magnitude of De(Deiso

5CB = ðPEO45�b�PS670ÞFNP ¼ 2:4) and a six-fold increase in

the hysteresis area compared to 5CB/BaTiO3 at the same

wt. % concentration. If we compare this value to De of ne-

matic 5CB, (Denem5CB ¼ 12Þ, we can estimate that even in iso-

tropic phase, 20% of all the mesogens have directional

orientation compared to fully nematic 5CB. This indicates an

interesting way of aligning the mesogens in the isotropic

phase.

Even though the hybrid system contains short-range

pseudonematic domains, the global nematic order is still

FIG. 1. Dielectric constant of pure 5CB, 5CB doped with BaTiO3 FNPs, and

polymer modified BaTiO3 FNPs at 0.275 wt. %. Dielectric constant is plotted

against applied RMS field (f¼ 1 kHz) in the nematic phase of LCs (25 �C).

FIG. 2. Dielectric hysteresis of pure 5CB (black), pure PEO45-b-PS670-SH

polymer (purple), 5CB þ bare BaTiO3 composite (red), polymer functional-

ized BaTiO3 FNPs (blue and green) as a function of applied electric field in

isotropic phase (T¼ 42 �C). Inset shows non-volatile effect as measured by a

capacitance bridge over 72 h.

083105-2 Kempaiah et al. Appl. Phys. Lett. 108, 083105 (2016)

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 136.160.90.131 On: Tue, 23 Feb 2016 14:38:28

absent in the isotropic phase. Therefore, there is no long-

range elastic interaction present in the isotropic phase.

Accordingly, these isolated pseudonematic domains do not

interact with the aligning layers of the LC cell. Therefore,

when the field is turned off, there is no restoring force to

mechanically torque these domains back into original orien-

tation in the isotropic phase and the domains stay oriented,

as schematically shown in Figure 3. In addition, the absence

of the back flow in the thin LC cell also allows the domains

to stay oriented. The thermal diffusion mechanism in this

case takes days to randomize the domains in the cell, as

can be seen from the inset in Figure 2. An electric field of

1.3 V/lm–1 was first applied to a hybrid system for 30 s to

initiate the reorientation of pseudonematic domains and then

switched off. The dielectric constant was monitored as a

function of time for the next 72 h and plotted in the inset in

Figure 2. No significant change was observed in this time

period. This is the essence of a non-volatile nano-electrome-

chanical memory effect; i.e., the domains at the nanoscale

are rotated mechanically and locked in thermodynamically.

When the FNPs are functionalized with a BCP of PS-b-

PAA, the PAA block is expected to bind strongly with

BaTiO3 FNPs’ surface via a combination of chemical and

van der Wall’s interaction.33 Even with a shorter PS chain in

the PS-b-PAA BCP, the hysteresis effect is found to be com-

parable to that of the PEO45-b-PS670-SH polymer (see Figure

2). This can be explained by the stronger binding of PS-b-

PAA to BaTiO3 than PEO-b-PS-SH BCP, because of the

multiple binding sites on PAA block. However, our measure-

ments were limited by the availability of polymer that had

longer PS block and higher affinity PAA block.

To understand this huge increase in dielectric hysteresis

in detail, we studied (i) the effect of only pure PEO45-b-

PS670-SH in 5CB, and (ii) the dispersibility of BaTiO3 when

functionalized with PEO45-b-PS670-SH. The repeating units

of hydrophobic PS block of BCPs have electron-rich benzene

ring that can interact with biphenyl group of 5CB via p-pstacking interaction. Therefore, a polymer that has more

units of hydrophobic PS block will favour aligning with 5CB

molecules because of this p-p stacking interaction. We then

doped 5CB with pure PEO45-b-PS670-SH such that the wt. %

of the dopant is approximately equal to the weight of the

polymer chains tethered on the BaTiO3 FNPs’ surface. This

normalized weight was essential as not all polymers we use

for functionalization go onto the surface of the FNPs. In this

case, we also noticed a hysteresis curve (purple) as shown in

Figure 2. This result clearly suggests that the presence of

pure PEO45-b-PS670-SH in 5CB also forms pseudonematicdomains due to the p-p stacking interaction between the LCs

and the polymer chains. These domains also interact with the

external E, showing a hysteresis effect. However, as clearly

seen in Figure 2, the hysteresis effect of 5CB/PEO45-b-

PS670-SH polymer functionalized BaTiO3 is significantly

larger than the sum of individual contributing components,

i.e., the pure polymer and the pure BaTiO3 FNPs.

Due to the dipole-dipole interaction between the FNPs,

it is expected that some FNPs are present in aggregates form

resulting in antiparallel-dipole configuration. An

antiparallel-dipole can be approximately treated as a quadru-

pole, whose field magnitude drops as �1/r4 and hence the

pseudonematic domains formed by the FNP-clusters would

have smaller De. On the other hand, the presence of the poly-

mer ligands on the FNP surface prevents the aggregation of

FNPs to a large extent. To visually examine the effects of

polymer ligands on aggregation and dispersibility, we pre-

pared two samples: (i) 5CB and bare BaTiO3 FNPs and (ii)

5CB and BaTiO3 FNPs functionalized with PEO45-b-PS670-

SH polymer. After preparation, they were allowed to stabi-

lize for 10 h and then a drop of each of sample in THF solu-

tion was deposited on a silicon substrate for SEM imaging.

Figure 4 shows representative SEM images of these samples.

It was apparent that bare 50 nm BaTiO3 sample formed sev-

eral micron-sized aggregates (Figures 4(a) and 4(c)),

whereas PEO45-b-PS670-SH polymer functionalized BaTiO3

FNPs showed a much better dispersion (Figures 4(b) and

4(d)). Without large clusters, therefore, the polymer func-

tionalized BaTiO3 FNPs retain their strong dipolar electric

field that drops as �1/r3 and create pseudonematic domains

with a higher De and a larger size, as illustrated in the sche-

matic in Figure 3. Due to this enhancement in the pseudone-

matic domains, the polymer-functionalized-BaTiO3 samples

exhibit a giant hysteresis effect.

FIG. 3. Schematic diagrams. (a) Schematics for each component of the sys-

tem. (b) Isotropic phase of 5CB with randomly oriented mesogens, (c) The

effect of introduction of FNP’s into 5CB. BaTiO3 molecules carries inherent

polarization, and as a result, 5CB molecules orient themselves along the

dipole field near the interface. A small arrow represents the direction of

polarization. (d) An illustration of the role of polymer functionalization of

BaTiO3 FNPs with amphiphilic block copolymers. Polymer tethers expand

the pseudonematic domains and act as scaffold for the radial alignment of

5CB mesogens. (e) Pseudonematic domains of mesogens (created by BCP

functionalized BaTiO3) mechanically rotate in the direction of the field. The

density of mesogens is greater along the poles. (f) The domains stay oriented

when the field is turned off.

083105-3 Kempaiah et al. Appl. Phys. Lett. 108, 083105 (2016)

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 136.160.90.131 On: Tue, 23 Feb 2016 14:38:28

It is reported that the size of the inorganic NP’s, the con-

centration of the dopant, the chemical nature, and the length

of polymer chain all play a major role in affecting the

dispersibility, long-range ordering, and electro-optic proper-

ties.34–36 To further understand the roles of these factors on

this soft-memory, especially the role of polymer chemistry

and FNPs concentration, we have studied three different con-

centrations of 5CB/PEO45-b-PS670-SH modified BaTiO3:

0.175 wt. %, 0.275 wt. %, and 0.375 wt. %. The schematics in

Figures 5(a) and 5(b) explain the p-p stacking interaction.

Figure 5(c) shows the effect of the dopant concentration on

the hysteresis effect. With increasing concentration of the

dopant, we see larger dielectric hysteresis, indicating the pres-

ence of more number of pseudonematic domains at higher

concentrations. The inset in Figure 5(c) explains the role of

different polymer chemistry on the dielectric hysteresis area.

Homopolymer of PAA600 alone (without the PS block) cannot

produce the same effect as PS-b-PAA which contains PS

block, due to absence of p-p stacking interaction (� red

curve). However, an introduction of even small PS block

(PS260) with PAA100 can produce much larger dielectric hys-

teresis (� green curve).36 The end-to-end length of each

hydrophobic PS670 block in a polymer chain is calculated to

be around 6–7 nm.27 Increasing the thickness of polymer

brushes, depending on the conformation, results in bigger

pseudonematic domains, enhancing the soft-memory effect.

In conclusion, we have shown that 5CB/BCP functional-

ized BaTiO3 samples yield giant enhancement of soft-

memory effect in the isotropic phase. Here, De is enhanced

multi-fold by simple attachment of BCP chains on the

surface of NPs. Both polymer chemistry and concentration

dependent studies have been conducted to understand these

physical phenomena. The results presented here are expected

to advance the conceptions about, and methodology toward,

nanoscale manipulation of nanomaterials via LCs and LC

orientation control using their interactions with functional-

ized nanomaterials. These studies will also lead directly to

the central role played by LC-influenced organizational

themes in the development of nanoscale directed self-

assembled systems, such as the pseudo-nematic domains—

which can be used as a nano-electromechanical memory

function.

FIG. 4. Scanning electron microscopy

images of the pure BaTiO3 and BCP

(PEO45-b-PS670-SH) functionalized

BaTiO3 FNPs. From (a) and (c), we

can infer that clusters and random

aggregates are prevalent.

Functionalized samples as shown in

(b) and (d) provide evidence of uni-

form colloidal dispersion.

FIG. 5. (a) Molecular structure of PEO45-b-PS670-SH BCP. (b) Schematic

illustrating the p-p stacking interaction between 5CB mesogens and benzene

rings of polystyrene units. (c) Dielectric hysteresis area of PEO45-b-PS670-SH

BCP functionalized BaTiO3 nanocomposite as a function of concentration of

dopants. The inset shows the dielectric hysteresis for various polymers that

were used to functionalize BaTiO3 doped 5CB liquid crystal; solid circle:

5CB/BaTiO3, solid square: 5CB/(PAA600b-PS260)-BaTiO3, solid triangle:

5CB/(PAA100b-PS260)-BaTiO3, solid star: 5CB/(PEO45-b-PS670)-BaTiO3.

083105-4 Kempaiah et al. Appl. Phys. Lett. 108, 083105 (2016)

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 136.160.90.131 On: Tue, 23 Feb 2016 14:38:28

Zhihong Nie is grateful for the support from the

National Science Foundation grant (CHE-1505839) and

startup fund from University of Maryland.

Rajratan Basu is grateful for the support from the Office

of Naval Research grant (Award No. N0001415WX01534)

and investment fund from the U.S. Naval Academy.

1Y. Ouchi, M. B. Feller, T. Moses, and Y. R. Shen, Phys. Rev. Lett. 68,

3040 (1992).2I. Mu�sevic and S. Zumer, Nat. Mater. 10, 266 (2011).3A. S. Matharu, S. Jeeva, and P. S. Ramanujam, Chem. Soc. Rev. 36, 1868

(2007).4T. Ikeda and O. Tsutsumi, Science 268, 1873 (1995).5H. K. Bisoyi and S. Kumar, Chem. Soc. Rev. 40, 306 (2011).6T. Hegmann, H. Qi, and V. M. Marx, J. Inorg. Organomet. Polym. Mater.

17, 483 (2007).7Q. Liu, Y. Cui, D. Gardner, X. Li, S. He, and I. I. Smalyukh, Nano Lett.

10, 1347 (2010).8Y. Zhang, Q. Liu, H. Mundoor, Y. Yuan, and I. I. Smalyukh, ACS Nano 9,

3097 (2015).9A. L. Rodarte, R. J. Pandolfi, S. Ghosh, and L. S. Hirst, J. Mater. Chem. C

1, 5527 (2013).10E. Pikina and C. Rosenblatt, Eur. Phys. J. E 35, 87 (2012).11J. Ouyang, C. W. Chu, C. R. Szmanda, L. Ma, and Y. Yang, Nat. Mater. 3,

918 (2004).12J. Prakash, A. Choudhary, A. Kumar, D. S. Mehta, and A. M. Biradar,

Appl. Phys. Lett. 93, 11294 (2008).13L. Marino, S. Marino, D. Wang, E. Bruno, and N. Scaramuzza, Soft

Matter 10, 3842 (2014).14R. Basu and A. Garvey, Appl. Phys. Lett. 105, 151905 (2014).15Y. Reznikov, O. Buchnev, O. Tereshchenko, V. Reshetnyak, A.

Glushchenko, and J. West, Appl. Phys. Lett. 82, 1917 (2003).16P. Ganguly, A. Kumar, S. Tripathi, D. Haranath, and A. M. Biradar, Appl.

Phys. Lett. 102, 222902 (2013).17A. Lorenz, N. Zimmermann, S. Kumar, D. R. Evans, G. Cook, M. F.

Martinez, and H.-S. Kitzerow, J. Phys. Chem. B 117, 937 (2013).

18A. Mikułko, P. Arora, A. Glushchenko, A. Lapanik, and W. Haase,

Europhys. Lett. 87, 27009 (2009).19N. Podoliak, O. Buchnev, M. Herrington, E. Mavrona, M. Kaczmarek, A.

G. Kanaras, E. Stratakis, J. F. Blach, J. F. Henninot, and M. Warenghem,

RSC Adv. 4, 46068 (2014).20A. Rudzki, D. R. Evans, G. Cook, and W. Haase, Appl. Opt. 52, E6

(2013).21R. Basu, Phys. Rev. E 89, 022508 (2014).22P. Kim, S. C. Jones, P. J. Hotchkiss, J. N. Haddock, B. Kippelen, S. R.

Marder, and J. W. Perry, Adv. Mater. 19, 1001 (2007).23S. A. Paniagua, Y. Kim, K. Henry, R. Kumar, J. W. Perry, and S. R.

Marder, ACS Appl. Mater. Interfaces 6, 3477 (2014).24B. Jiang, X. Pang, B. Li, and Z. Lin, J. Am. Chem. Soc. 137, 11760

(2015).25X. Pang, Y. He, B. Jiang, J. Iocozzia, L. Zhao, H. Guo, J. Liu, M. Akinc,

X. Tan, N. Bowler, and Z. Lin, Nanoscale 5, 8695 (2013).26H. Z. Guo, Y. Mudryk, M. I. Ahmad, X. C. Pang, L. Zhao, M. Akinc, V.

K. Pecharsky, N. Bowler, Z. Lin, and X. Tan, J. Mater. Chem. 22, 23944

(2012).27J. He, Y. Liu, T. Babu, Z. Wei, and Z. Nie, J. Am. Chem. Soc. 134, 11342

(2012).28Y. Li, Z. Liao, F. Fang, X. Wang, L. Li, and J. Zhu, Appl. Phys. Lett. 105,

182901 (2014).29M. Draper, I. M. Saez, S. J. Cowling, P. Gai, B. Heinrich, B. Donnio, D.

Guillon, and J. W. Goodby, Adv. Funct. Mater. 21, 1260 (2011).30P. G. De Gennes and J. Prost, The Physics of Liquid Crystals (Oxford

University Press, New York, 1994).31B. Jaffe, W. R. Cook, H. Ludwig, and C. Jaff�e, Piezoelectric Ceramics

(Academic Press, London, New York, 1971).32C. J. Adam, S. J. Clark, G. J. Ackland, and J. Crain, Phys. Rev. E 55, 5641

(1997).33P. Kim, N. M. Doss, J. P. Tillotson, P. J. Hotchkiss, M. J. Pan, S. R.

Marder, J. Y. Li, J. P. Calame, and J. W. Perry, ACS Nano 3, 2581 (2009).34A. Lapanik, A. Rudzki, B. Kinkead, H. Qi, T. Hegmann, and W. Haase,

Soft Matter 8, 8722 (2012).35Q. K. Liu, Y. Yuan, and I. I. Smalyukh, Nano Lett. 14, 4071 (2014).36G. H. Kirby, D. J. Harris, Q. Li, and J. A. Lewis, J. Am. Ceram. Soc. 87,

181 (2004).

083105-5 Kempaiah et al. Appl. Phys. Lett. 108, 083105 (2016)

Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 136.160.90.131 On: Tue, 23 Feb 2016 14:38:28