www.advmat.dewww.MaterialsViews.com

CO

MM

UN

ICATIO

Yingbang Yao , QingXiao Wang , Hongtao Wang , Bei Zhang , Chao Zhao , Zhihong Wang , Zhengkui Xu , Ying Wu , Wei Huang , Pei-Yuan Qian , and Xi Xiang Zhang *

Bio-Assembled Nanocomposites in Conch Shells Exhibit Giant Electret Hysteresis

N

Novel physics and properties have been discovered in different assembled nanomaterials synthesized by the bottom-up tech-nique, which enables us to make and build new electronic, photonic, and magnetic devices based on those properties. [ 1–7 ] Nature has long used bottom-up synthesis to fabricate nanoma-terials that exhibit much better physical properties than their man-made counterparts. [ 8–13 ] For example, superior optical properties are observed in the nanometer-scale architectures of brittle stars, butterfl ies, and many insects; [ 14 ] super-hydrophobic effects are evident in lotus plants and water bugs. [ 15 , 16 ] Superior mechanical properties are found in seashells, which are com-posed of well-aligned and highly packed aragonite nanolaminas glued together by biopolymers. [ 8–12 ] Although the strength and toughness of the nanocomposites in seashells have been exten-sively studied, [ 8–12 ] their other extraordinary physical proper-ties have not been described. Here, we show that conch shells, with bio-assembled, hierarchical architectures of nano-CaCO 3 laminas sandwiched between biopolymers, exhibit ferroelectret behavior: ferroelectric-like hysteresis loops. [ 17–20 ] Their rema-nent electrical polarization (2000–4000 μ C cm − 2 ) is one order of magnitude higher than the largest electrical polarization reported in man-made ferroelectric materials (146 μ C cm − 2 ) [ 21 ]

© 2012 WILEY-VCH Verlag Gm

DOI: 10.1002/adma.201202079

Dr. Y. B. Yao, Q. X. Wang, H. T. Wang, [+] B. Zhang,C. Zhao, Z. H. Wang, Prof. X. X. ZhangNanofab, Imaging & Characterization Core LabKing Abdullah University of Science and TechnologyThuwal 23955-6900, Saudi Arabia E-mail: xixiang.zhang@kaust.edu.sa Z. K. XuDepartment of Physics and Materials ScienceCity University of Hong KongKowloon, Hong Kong, P. R. of China Dr. Y. WuComputer, Electrical and Mathematical ScienceKing Abdullah University of Science and TechnologyThuwal 23955-6900, Saudi Arabia W. HuangInstitute of Advanced MaterialsNanjing University of Posts and TelecommunicationsNanjing 210046, P. R. of China P.-Y. QianKAUST Global Collaborative ResearchDivision of Life ScienceHong Kong University of Science and TechnologyHong Kong, P. R. of China [ + ] Present address: Zhejiang University, Hangzhou 310027,P. R. of China

Adv. Mater. 2012, DOI: 10.1002/adma.201202079

and several orders of magnitude higher than that in elec-trets, [ 20 , 22 ] as obtained from square polarization–electric-fi eld hysteresis loops. This novel property suggests the possibility of developing nanocomposites with high electric polariza-tion using the bottom-up technique for applications requiring high-performance electret motors/generators [ 22 ] or high-density energy storage.

We cut slices of a conch shell that were less than 1 mm thick along the shell’s length. The shell from which the sam-ples were cut is shown in Figure 1 a. We used scanning elec-tron microscopy (SEM) and transmission electron microscopy (TEM) to confi rm the microarchitecture of the shell as previ-ously reported [ 9 ] (Figures 1 b–d). The basic building blocks of the conch were found to be nanolaminas that have cross sec-tions of 50 nm × 200 nm and are 10-20 μ m in length. We con-ducted X-ray diffraction (XRD) experiments and found that only a few peaks appeared in the spectrum in comparison with the standard powder XRD spectrum of aragonite (Figure S1, Sup-porting Information), indicating that the nanolaminas have an orderly assembly with preferential crystalline orientations, sim-ilar to self-assembled Fe 3 O 4 nanoparticles. [ 23 ]

We milled the samples to a thickness of 0.50 ± 0.01 mm for all electric measurements. The dielectric properties of the sam-ples were then measured in the frequency range from 100 Hz to 10 MHz at room temperature. Figure 2 shows representative curves of the frequency dependence of the dielectric properties measured on different samples. The relative dielectric constant lay in the range of 80–300 at the low-frequency end (100 Hz) and decreased to ca. 13–19 at 1 MHz. The dielectric loss (tan δ ) decreased from 0.3–1.0 at 100 Hz to 0.05–0.3 at 1 MHz (inset in Figure 2 ). These properties differ signifi cantly from those of single-crystal slices (0.5 mm thick) of CaCO 3 , for which the dielectric constant is about 9 and independent of frequency. In addition, the dielectric loss of a single crystal of CaCO 3 is very small ( < 0.01) across the entire frequency range. The unusually large dielectric constant of the shell indicates that the electrical properties of the bio-assembled nanocomposite differ from those of single crystals of CaCO 3 . The large dielectric loss of the conch is indicative of signifi cant electrical leakage.

To understand this electrical leakage in the shell materials, we performed standard current–electric-fi eld measurements by sweeping the electric fi eld in the range ± 40 kV cm − 1 with a ram-ping rate of 500 V cm − 1 s − 1 , corresponding to a fi eld frequency of 3.1 mHz. Shown in Figure 3 a are the fi rst and third current density–electric-fi eld ( J – E ) loops. The fi rst J – E loop showed weak hysteresis. With fi eld cycling, peaks developed gradually in both directions. After a number of cycles, the J – E behavior became nearly saturated and reproducible, and the peaks in the

bH & Co. KGaA, Weinheim 1wileyonlinelibrary.com

www.advmat.dewww.MaterialsViews.com

CO

MM

UN

ICATI

ON

Figure 1 . a)Photograph of the conch shell used in this study. How the samples were cut is indicated. b) SEM image showing the microarchitecture of the conch shell. c) SEM image showing how the nanoaragonite laminas are assembled. d) TEM image of a cross section of the nanolaminas and the bio-organic matrix between the nanolaminas.

Figure 2 . Frequency dependence of dielectric constant and loss for sev-eral representative samples from a conch shell.

J – E loop grew higher and sharper. Figure 3 c shows an example of a nearly saturated J – E loop (eighth), which is very similar to the behavior of J – E loops of conventional ferroelectric mate-rials, such as lead zirconate titanate (PZT) (Figure S2, Sup-porting Information). After each J – E loop, a polarization–elec-tric-fi eld ( P – E ) loop was measured by sweeping the fi eld from zero without pre-poling, as shown in Figure 3 b. In these P – E measurements, the fi eld frequency is 0.01 Hz, the lowest fre-quency of the instrument for P – E loop measurements. Actually, the measured P – E hysteresis is the integration of current den-sity with time using the data of the J – E loop according to

P (t) =∫ t

0J (t) dt (1)

where J ( t ) is the measured total current density, including current induced by dipole reversal, leakage, and capacitive charging. The P – E loop shown in Figure 3 c is the integration of the eighth J – E loop. For an ideal ferroelectric material, that is, a perfect insulator, both the leakage and capacitive charging currents are negligible in comparison with the current induced by dipole reversal, that is, the peak in the increasing-fi eld process in the J – E loop (Figure S2, Supporting Information). Since the leakage is much larger than the capacitive charging

2 wileyonlinelibrary.com © 2012 WILEY-VCH Verlag G

contribution in our samples, the current density value of the decreasing-fi eld curves in the J – E loops ( J d ) can be safely taken as the leakage. The leakage-corrected hysteresis loop can then be easily calculated as

mbH & Co. KGaA, Weinheim Adv. Mater. 2012, DOI: 10.1002/adma.201202079

www.advmat.dewww.MaterialsViews.com

CO

MM

UN

ICATIO

N

Figure 3 . a–f) J – E loops and P – E loops measured at room temperature between –40 and + 40 kV cm − 1 . a) The fi rst and third J – E loops measured with a fi eld-frequency of 0.0031 Hz, corresponding to a fi eld ramping rate of 500 V cm − 1 s − 1 . b) P – E loops measured after the fi rst and third J – E cycles with a fi eld frequency of 0.01 Hz. c)The eighth J – E loop and the corresponding integrated P – E loop with a fi eld frequency of 0.0031 Hz. d) The leakage-corrected J – E and P – E loops of (c). e) P – E loop measured with pre-poling protocol at 0.01 Hz, the lowest frequency of the instrument, and the corresponding J – E loop. f) The leakage-corrected J – E and P – E loops of (e).

-40 -30 -20 -10 0 10 20 30 40-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0 After 3rd J-E after 1st J-E

P (

mC

/cm

2 )

E (kV/cm)

(b)

-40 -30 -20 -10 0 10 20 30 40-6

-4

-2

0

2

4

6

P(m

C/c

m2 )

E (kV/cm)

-400

-200

0

200

400

J (

A/c

m2 )

0.0031 Hz (c)

-40 -20 0 20 40

-4

-2

0

2

4

P (

mA

/cm

2 )E (kV/cm)

0.0031Hz (d)

-400

-200

0

200

400

J (

A/c

m2 )

-40 -20 0 20 40

-4

-2

0

2

4

P(m

C/c

m2 )

E (kV/cm)

0.01Hz

-800

-400

0

400

800

J (

A/c

m2 )

(e)

-40 -20 0 20 40

-4

-2

0

2

4 0.01Hz

P(m

C/c

m2 )

E (kV/cm)

(f)

-800

-400

0

400

800

J (

A/c

m2 )

P (t) =∫ t

0

[J (t) − Jd (t)] dt (2)

Figure 3 d shows the leakage-corrected hysteresis loop and leakage-corrected current density that was used to generate the hysteresis loop.

We also measured the P – E loop using the standard protocol (by averaging two P – E loops, one with negative pre-poling and one with positive) after the eighth J – E loop with a fi eld fre-quency of 0.01 Hz (Figure 3 e). Similarly, we used Equation 2 to calculate the leakage-corrected hysteresis loop, as shown in Figure 3 f. It is clear that the leakage-corrected P – E loops are saturated, square hysteresis loops with all the features of the P – E loops of typical ferroelectret polymers [ 17–20 ] or ferroelectric materials, [ 24 ] including nonzero remanent polarization, satura-tion of polarization, and a coercive fi eld. It is clear that as the fi eld was cycled, the P – E loops also evolved from an olive-shaped

© 2012 WILEY-VCH Verlag GAdv. Mater. 2012, DOI: 10.1002/adma.201202079

P – E loop (typical for leakage) to a square P – E loop, typical for a ferroelectric/ferroelectret.

The most striking difference between the shell and standard ferroelectret and ferroelectric materials is the huge remanent polarization, ∼ 3100 μ C cm − 2 at 3.1 mHz ( ∼ 2150 μ C cm − 2 at 0.01 Hz), which is more than 20 times larger than the largest remanent polarization in PZT thin fi lms (146 μ C cm − 2 ) [ 21 ] and 2–4 orders of magnitude larger than that in ferroelectret foams and other electrets. [ 17 , 19 , 20 , 22 ] We measured tens of samples and found that they all exhibited similar behaviors, but the values of the parameters, such as the remanent polarization and the coercive fi eld, fl uctuated, likely owing to the non-uniformity/non-homoge-neity of all bio-materials. The shape of the hysteresis loops shown in Figure 3 d and f are very similar to those observed in ferroelec-tric polymers in which the remanent polarization and reproduc-ible hysteresis loops were induced by fi eld cycling. [ 25 ] We refer to the samples with sharp/reproducible peaks as poled samples.

3wileyonlinelibrary.commbH & Co. KGaA, Weinheim

www.advmat.dewww.MaterialsViews.com

CO

MM

UN

ICATI

ON

Figure 4 . The unipolar J – E loops measured with positive fi eld on a poled, 0.5 mm thick sample with a fi eld ramping rate of 500 V cm − 1 s − 1 . Inset a: The fi rst J – E loop and the fi fth J – E curve measured with a fi eld ramping rate of 500 V cm − 1 s − 1 . Inset b: The leakage-corrected P – E curve measured with pre-poling after the fi fth J – E curve with a frequency of 0.01 Hz and corre-sponding J – E loop. Inset c: Decrease of the current density at maximum fi eld with the number of cycles n of the fi eld.

-40 -20 0 20 40

-500-250

0250500

J (µ

A/c

m2 )

E(kV/cm)

(a)

-40 -20 0 20 40-6

-3

0

3

6

P(m

C/c

m2 )

E (kV/cm)

(b)

0 20 400

100

200

300

400

J (µ

A/c

m2 )

6th

11th

E (kV/cm)

J (µ

A/c

m2 )

1st

2nd

3rd

-400

-200

0

200

400

J(µA

/cm

2 )

0 5 100

100200 (c)

n

Figure 5 . Leakage-corrected P – E loops measured with a pre-poling pro-tocol at different frequencies on a poled 0.5 mm thick sample.

-40 -20 0 20 40

-1.2

-0.6

0.0

0.6

1.2

P (

mC

/cm

2 )

E (kV/cm)

Frequency (Hz)

0.01

0.02

0.04

0.08

1.0

To show that these peaks in J – E curves were due to current induced by the reversal of polarization, we performed unipolar J – E loop measurements, [ 25 ] that is, the loops were in the fi rst or third quadrant only, although the origin of the polarization or dipoles is not clear at this stage. Insets a and b of Figure 4 show, respectively, the (fi rst and fi fth) J – E loops and the leakage-corrected P – E loop using the fi fth J – E loop. Before the unipolar loop measurement, the sample was poled negatively. The fi rst unipolar J – E curve (0 → 40 kV cm − 1 ) in Figure 4 exhibits a sharp peak at about 24 kV cm − 1 . In the following J – E loops (second to eleventh), no peak was observed in either the increasing-fi eld or decreasing-fi eld curves, because most “dipoles” had been aligned in the positive direction during the fi rst increasing-fi eld process. The peak in the fi rst increasing-fi eld curve must therefore be due to the current induced by the reversal in polarization from negative to positive. [ 25 ]

It is worth noting that when the fi eld was cycled, the cur-rent decreased and gradually approached a constant. To show clearly the trend, we plotted the current density at 40 kV cm − 2 as a function of the number of cycles (loops) (Figure 4 , inset c) This observation can be interpreted as follows. During the uni-polar J – E loops, the applied electric fi eld was always positive, which always forced the “dipoles” to line up in its direction. Consequently, the alignment/reversal of the dipoles always induced a current in the same direction in both increasing-fi eld and decreasing-fi eld processes. With more and more “dipoles” aligned in the fi eld direction, fewer and fewer “dipoles” were left for the next loop. Consequently, the current density caused by the “dipole” alignment/reversal decreased with fi eld cycling and gradually vanished. The saturated current density will

4 wileyonlinelibrary.com © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinh

fi nally refl ect the real leakage of the sample. This again supported our conclusion that the peaks in the J – E curves are due to the cur-rent induced by the “dipole” reversal. Con-sequently, the leakage-corrected hysteresis loops originate purely from the reversal of electric “dipoles”. It is also clear from the data in inset c (Figure 4 ) that we have over-corrected the data and that the alignment of the “dipoles” in the shell is a very slow process.

In order to understand the characteristics of the polarization reversal, P – E loops were measured using the standard protocol at dif-ferent frequencies (0.01–200 Hz) on a poled sample, as shown in Figure 5 . To better present the data, all the loops were leakage-corrected. It is evident that the f = 0.01 Hz loop has all the features of a standard fer-roelectret/ferroelectric-like P – E loop with a remanent polarization of P r ∼ 1013 μ C cm − 2 and a coercive fi eld of about E c = 23 kV cm − 1 . When the frequency increased from 0.01 Hz to 0.02, 0.04, 0.08, and 1 Hz, P r decreased from 1013 to 0.007 μ C cm − 2 . We also found that when f > 0.2 Hz, the shape of the uncor-rected P – E loop changed to one in which the leakage current is dominant. Another important feature in these P – E loops is that

the coercive fi eld increased with frequency. The frequency dependence of the remanent polarization and the coercive fi eld is similar to characteristics of ferroelectric materials and sug-gests that the reversed polarization is strongly associated with the activation process [ 26 , 27 ] (Figure S3, Supporting Information). However, the reversal of the polarization in the conch shell is extremely slow in comparison with that in conventional ferro-electrics and ferroelectret polymers. [ 17–20 , 24–29 ]

eim Adv. Mater. 2012, DOI: 10.1002/adma.201202079

www.advmat.dewww.MaterialsViews.com

CO

MM

UN

ICATIO

N

Figure 6 . a) The leakage-corrected J – E loops measured at different times after the sample was well-poled negatively with a fi eld ramping rate of 500 V cm − 1 s − 1 . b) The reversed polarization calculated using the J – E curves in (a). The polarization values in (b) are proportional to the remanent polarization.

0 20 400

50

100J (

A/cm

2 )

E (kV/cm)

Virgin 20min 40min 1h 2h 4h 16h

(a)

0.1 1.0 10.0

0.5

1.0

1.5

2.0

2.5

3.0

P (m

C/cm

2 )

Retention time (h)

(b)

To explore the stability of the remanent polarization of the samples, we performed a retention experiment at room tempera-ture on a well-poled sample with the following procedure. a) Just after the sample was well-poled negatively, the positive unipolar J – E loop was measured immediately; the fi rst unipolar loop was denoted as the virgin state. b) To guarantee that the sample reached the same saturated state before the next unipolar loop, the sample was well-poled negatively again. After waiting for a time t , we measured the second unipolar loop. c) Step b was repeated, but with different waiting time t . The reversed polari-zation during the fi eld increase process can then be calculated using the data of the unipolar J – E loop. During the waiting time t , the negatively aligned dipoles begin to randomize to reduce the electrostatic energy, which leads to a decrease of the remanent polarization. Consequently, the current induced by the reversal of the remanent polarization diminished with time t . Figure 6 a shows the unipolar curves at different waiting times. Here, we have corrected the unipolar loop by subtracting the decreasing-fi eld curve. The relaxation of the polarization with time is shown in Figure 6 b. As we discussed for the unipolar loops in Figure 4 , only part of the remanent polarization was reversed. However, the reversed polarization should be proportional to the remanent polarization before the fi eld is switched. It is clear that the rema-nent polarization decreased from 2.6 mC cm − 2 to 0.7 mC cm − 2 within 2 h then remained nearly constant for the following 14 h.

To understand the origin of the giant polarization in bio-assembled nanocomposites, we measured the same J – E and

© 2012 WILEY-VCH Verlag GAdv. Mater. 2012, DOI: 10.1002/adma.201202079

P – E loops of a single crystal slice of calcite (CaCO 3 ) that was 0.5 mm thick and of a sample taken from a pearl oyster shell, a different bio-assembled aragonite composite. [ 8 , 10–12 ] We did not observe peaks in the J – E loops nor did we fi nd ferroelec-tric/ferroelectret-like P – E hysteresis loops in either sample (Figures S4–S6, Supporting Information), which indicates that both calcite and aragonite CaCO 3 are nonpolar. The varied behaviors of the conch shell and the pearl oyster shell materials must come from their different microstructures/microarchitectures.

In pearl oyster shells, the basic building blocks — aragonite platelets — stack in the same direction. These platelets are about 0.5 μ m thick and several hundred micrometers in length. The thickness of the organic matrix between the platelets is a few tens of nanometers [ 8 , 10–12 ] (Figures S7 and S8, Supporting Information). The basic building blocks of conch shells are much ( ∼ 250 times) smaller than those of pearl oyster shells. More importantly, the hierarchical architecture in conch shells is much more complex than the simple stacking in pearl oyster shells. [ 8–12 ] The biopolymer layers between the nanolaminas in the conch shells form a complex three-dimensional (3D) network, which must play an important role in the extreme polarization.

Under a strong applied electric fi eld (up to 40 kV cm − 1 ), the molecules in the conch shell’s biopolymer layers could be gradually ionized, allowing the electrons to hop between the molecules. In addition to the fi eld-induced conduction, the biopolymer itself might also have had weak electrical con-duction. The biopolymer layers became conducting paths with high resistivity for electrons. More importantly, these layers will also act as electrodes for the “nanocapacitors”, the CaCO 3 laminas sandwiched by conducting biopolymer layers. The bio-assembled nanocomposite is essentially a complex 3D network of resistors and capacitors in which the resistors and capacitors should be electric-fi eld dependent. In the poled samples, the nanocapacitors should be well charged.

These charged nanocapacitors should be analogue to the charged voids in ferroelectret cellular polymers. When an applied electric fi eld is large enough, it triggers the breakdown of gases or creates a microplasma discharge in the voids of the polymer. The positive and negative charges are then sepa-rated and trapped at opposite void surfaces under the electric fi eld. [ 17–20 ] The large number of charges and distance between the positive and negative charges in a void of typical size of 100 μ m × 100 μ m × 10 μ m form a macroscopic dipole or giant dipole that is huge in comparison with the electrical dipoles in conventional ferroelectric materials and electrets. [ 20 , 22 ] The direction of these giant dipoles can be switched back and forth by cycling a electric fi eld if the fi eld is stronger than the break-down fi eld of the gases in the voids. The switching of the giant dipoles has been ascribed to the internal microplasma dis-charge in the voids, which is analogous to the reversal process between thermally stable polarization states in ferroelectric materials and is the origin of the displayed ferroelectric-like hysteresis loops. [ 18 ] The corresponding coercive fi eld of the hysteresis loop in the ferroelectret polymers should be equal to the breakdown fi eld.

Therefore, the reversal of the polarity of the charged capacitors likely came from the change in the positions of the

5wileyonlinelibrary.commbH & Co. KGaA, Weinheim

6

www.advmat.dewww.MaterialsViews.com

CO

MM

UN

ICATI

ON

Figure 7 . a) The experimental data of the current induced by dipole reversal and the fi tted data. b)The fi tted distribution ( D ) of U 0 (or size of the dipoles).

0 20 40

0

100

200

J (m

A/c

m2 )

E (kV/cm)

(a)

0.0 1.0 2.0 3.0 4.0 5.0 6.0

D (

a.u.

)

U0 (eV)

(b)

electrons and the positive charges in the polymer layer following the application of the electric fi eld. This process may be similar to what happens in cellular ferroelectrets. [ 17–20 ] However, the position exchange of the electrons and positively charged mol-ecules in the biopolymer layers is very slow. This slow reversal process allows us to observe a ferroelectret-like P – E loop only with a very low fi eld frequency (Figures 3 and 5 ). In this model, the charges are distributed into millions of nanocapacitors in the body of the sample and each nanocapacitor behaves like a “giant dipole” similar to the microscopic dipoles in conven-tional ferroelectret foams.

This process can be correlated to the equivalent-circuit models developed for understanding the switching kinetics and hysteresis loops of ferroelectric materials. [ 30 , 31 ] In these models, a unit cell of a ferroelectric material can be described as a fi eld-dependent capacitor and a fi eld-dependent resistor in series. It is therefore reasonable to expect to observe ferroelectric-like behaviors in our samples.

We assume that the reversal of a giant dipole has similar characteristics to those of a conventional ferromagnetic dipole. For example, ferromagnetic dipoles have a double-well potential and two stable states, “up” and “down”, and their reversal can be caused by thermal activation. The reversal rate between these two states can be then described by the Arrhenius law

� = �0 exp(

U(E )

kBT

) (3)

where Γ 0 is 10 12 –10 13 Hz, [ 32 , 33 ] k B is the Boltzmann constant, T is the absolute temperature, and U is the energy barrier, which depends on the applied electric fi eld E [ 32 ] as

U(E ) = V∗(WB − ps E )2 = V∗WB

(1 − ps E

WB

)2

= U0

(1 − E

E0

)2

(4)

where V* is the activation volume in which the dipole is reversed coherently, p s is the polarization, and W B is the energy barrier between the two states for p s at E = 0, which is a material property. Therefore, E 0 = W B / p s should be the characteristic fi eld that removes the energy barrier completely.

Owing to the inhomogeneity of the biopolymer layers, the size of the nanolaminas and the distance between the nanolam-inas as shown in Figure 1 , there must be a activation volume distribution, f ( V* ). The time-dependent polarization change can be written as [ 33 ]

U(E ) = V∗(WB − ps E )2 = V∗WB

(1 − ps E

WB

)2

= U0

(1 − E

E0

)2 (5)

The corresponding current induced by Δ P ( t ) is then given by

I(t) = d[�P(t)

]dt

= d[�P(t)

]dE

dE

dt= r

d[�P(t)

]dE

(6)

where dE/dt is the ramping rate of the applied electric fi eld. We then fi tted the leakage-corrected J – E curve in Figure 3 d

using Equation 6 , because it is purely due to the reversal of

wileyonlinelibrary.com © 2012 WILEY-VCH Verlag G

the giant dipoles. To fi t the J – E curve, we have to fi nd a suit-able distribution of V *. In practice, we simply need to fi nd the distribution of U 0 ( = V*W B ). Figure 7 a shows the leakage-corrected J – E curve and fi tting data. The fi tted distribution ( D ) of U 0 is shown in Figure 7 b; it is a non-symmetric distribution with a peak value of U 0 = 1.6 eV. The other parameters are E 0 = 40 kV cm − 1 and Γ 0 = 1 × 10 12 Hz. From U 0 = 1.6 eV, the reversal frequency for the dipoles at E = 0 and room temperature can be easily calculated to be about 10 − 5 Hz.

To explore the potential applications of this giant polariza-tion (or stored charges), we measured measured thermally stimulated currents in Conch shells. The currents obtained on the well-poled and nonpoled samples at temperatures ranging from 40 ° C to 300 ° C are shown in Figure 8 a. In the measure-ments, the discharge currents were recorded while the temper-ature was increased linearly with time (temperature ramping technique [ 34–36 ] ). Two similar samples (cut from the same loca-tion and of the same thickness) were measured, one unpoled and one poled, and it is interesting to note that below 240 ° C similar current profi les were observed in the two samples. The sharp increase of the current at 200 ° C for both samples could be caused by the decomposition and evaporation of the polymer as evidenced by the thermogravimetric analysis (TGA) meas-urements (Figure S9, Supporting Information), because such evaporated polymer fragments, monomers, molecular groups, etc. may carry positive or negative charges, depending on their compositions. [ 37,38 ] Such a discharge current caused by the evaporating polymer fragments reached a maximum at about 240 ° C and then decreased at temperatures higher than 248 ° C, forming a broad peak between 200 and 265 ° C. The two small

mbH & Co. KGaA, Weinheim Adv. Mater. 2012, DOI: 10.1002/adma.201202079

www.advmat.dewww.MaterialsViews.com

CO

MM

UN

ICATIO

N

Figure 8 . a) The pyroelectric currents measured on an unpoled and a poled sample with the same thickness and cut from the same position on the Conch shell. b)The calculated pyroelectric coeffi cients from the measured pyroelectric currents in the temperature range 238–264 ° C.

100 200 300

0

50

100

Unpoled PoledI (

nA)

T (oC)

(a)

240 250 260

0

60

120

p (m

C/m

2 K)

T (oC)

(b)

peaks observed at temperatures of approximately 140 ° C and 190 ° C could be due to the evaporation of the polymers around the surfaces or the evaporation of polymers with different com-positions. The difference in the currents at low temperatures for these two samples could be due to the inhomogeneity of the bio-materials, or a change in the state of the polymer after poling. As the temperature increased above 265 ° C, the meas-ured currents in both samples began to fl uctuate, which might be due to the severe decomposition of the polymer layers. The most intriguing observation is the high and narrow discharge current peak formed in the temperature range of 240–265 ° C in the poled sample, which is found on the top of the broad peak for the unpoled sample. This huge peak must be a con-sequence of the softening and decomposition of the polymer, leading to the collapse of the nanocapacitors, that is, the dis-appearance of the giant dipoles. We suggest, based on results found by subtracting the currents of the unpoled sample from the currents of the poled samples, that pyroelectricity is observ-able in conch shells. We then calculated the pyroelectric coef-fi cients using

Ipoled − Iunpoled = A · dP

dt= A · dP

dt

dT

dt= A · p

dT

dt (7)

that is

Ipoled − Iunpoled

p=

(A · dT

dt

) (8)

where A is the area of the electrodes, P the polarization, p the pyroelectric coeffi cient, T the temperature, and t the time.

© 2012 WILEY-VCH Verlag GAdv. Mater. 2012, DOI: 10.1002/adma.201202079

Figure 8 b shows the temperature-dependent “pyroelectric” coeffi cients at 8–115 mC m − 2 K − 1 . Without having removed irre-versible contributions from the currents, we note a coeffi cient 2–3 orders of magnitude larger than those of conventional fer-roelectric materials (100–1000 μ C m − 2 K − 1 ). [ 34–36 , 39 ] This agreed with our expectations considering their giant polarization, and materials with pyroelectric coeffi cients of such great magni-tudes in a specifi c temperature range may fi nd applications in thermal sensors/detectors.

In summary, the nanolaminas and the biopolymer layers of conch shells exhibit ferroelectret behaviors, including square P – E hysteresis loops with giant “polarization” and indications of pyroelectricity. These behaviors originate from the charged nanocapacitors that are created by the hierarchical microar-chitecture of the CaCO 3 nanocomposite, in which biopolymer layers form a complex 3D network. The results strongly sug-gest that by varying the properties of the polymers and the sizes of the nanolaminas, the polarization, coercive fi eld, and reversal frequency of the giant dipoles can be tailored for varied applications utilizing ferroelectret materials, including electret motors/generators [ 22 ] and of high-density energy storage. Inter-estingly, the fabrication of such composites has already seen great progress. [ 10 , 11 ]

Experimental Section Sample preparation: For the dielectric ferroelectric measurements,

circular, 200 nm thick electrodes of Pt or Au were deposited through a mask onto both sides of 0.5 mm slices of the conch shell using pulsed laser deposition.

Materials characterization : XRD patterns were collected on a Bruker D8 Advanced X-Ray Diffractometer. Field emission SEM (FESEM) images were obtained on a FEI Quanta 600 microscope. TEM images were taken on FEI Titan ST microscopes.

Electrical measurements : The dielectric properties were measured fi rst for all the samples in the frequency range from 100 Hz to 10 MHz at room temperature using an Agilent 4294A impedance analyzer. All the J – E loops and the P – E loops were measured using an aixACCT ferroelectric tester (TFA 2000) with the electric fi eld varying linearly as 0 → + 40 kV cm − 1 → –40 kV cm − 1 → 0. To reduce arcing, the measurements were performed with the sample immersed in silicon oil. For the measurements of pyroelectric currents, the currents were recorded while the temperature of the sample was linearly ramped. Details of the system are described in the literature. [ 36 ]

Supporting Information Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgements The authors are grateful to Prof. P. Sheng and Prof. Z. Q. Zhang at HKUST for useful discussions and Dr. Yang Yang, and Mr. Xiaodong Xu for technical assistance. Dr. Virginia Unkefer and Ms. Kan Zhang improved the English.

Received: May 24, 2012 Revised: September 14, 2012

Published online:

7wileyonlinelibrary.commbH & Co. KGaA, Weinheim

8

www.advmat.dewww.MaterialsViews.com

CO

MM

UN

ICATI

ON

[ 1 ] V. P. Colvin , M. C. Schlamp , A. P. Alivisatos , Nature 1994 , 370 , 354 .

[ 2 ] Y. D. Yin , A. P. Alivisatos , Nature 2005 , 437 , 664 . [ 3 ] J. A. Fan , C. Wu , K. Bao , J. Bao , R. Bardhan , N. J. Halas ,

V. N. Manoharan , P. Nordlander , G. Shvets , F. Capasso , Science 2010 , 328 , 1135 .

[ 4 ] Y. Huang , X. F. Duan , Y. Cui , L. J. Lauhon , K. H. Kim , C. M. Lieber , Science 2001 , 294 , 1313 .

[ 5 ] M. S. Gudiksen , L. J. Lauhon , J. F. Wang , D. C. Smith , C. M. Lieber , Nature 2002 , 415 , 617 .

[ 6 ] S. H. Park , D. Qin , Y. Xia , Adv Mater 1998 , 10 , 1028 . [ 7 ] V. F. Puntes , P. Gorostiza , D. M. Aruguete , N. G. Bastus ,

A. P. Alivisatos , Nat. Mater. 2004 , 3 , 263 . [ 8 ] B. L. Smith , T. E. Schaffer , M. Viani , J. B. Thompson ,

N. A. Frederick , J. Kindt , A. Belcher , G. D. Stucky , D. E. Morese , P. K. Hansma , Nature 1999 , 399 , 761 .

[ 9 ] S. Kamat , X. Su , R. Ballarini , A. H. Heuer , Nature 2000 , 405 , 1036 . [ 10 ] Z. Y. Tang , N. A. Kotov , S. Magonov , B. Ozturk , Nat. Mater. 2003 , 2 ,

413 . [ 11 ] M. Rubner , Nature 2003 , 423 , 925 . [ 12 ] X. D. Li , W. C. Chang , Y. J. Chao , R. Z. Wang , M. Chang , Nano Lett.

2004 , 4 , 613 . [ 13 ] J. B. Thompson , J. H. Kindt , B. Drake , H. G. Hansam , D. E. Morse ,

P. K. Hansma , Nature 2001 , 414 , 773 . [ 14 ] P. Vukusic , J. R. Sambles , Nature 2003 , 424 , 852 . [ 15 ] L. Feng , S. H. Li , Y. S. Li , H. J. Li , L. J. Zhang , J. Zai , Y. L. Song ,

B. Q. Liu , L. Jiang , D. B. Zhu , Adv. Mater. 2002 , 14 , 1857 . [ 16 ] R. Blossey , Nat. Mater. 2003 , 2 , 302 . [ 17 ] S. Bauer , R. Gerhard-Multhaupt , G. M. Sessler , Phys. Today 2004 ,

57 (2), 37 . [ 18 ] M. Wegener , S. Bauer , ChemPhysChem 2005 , 6 , 1014 . [ 19 ] R. Gerhard-Multhaupt , IEEE Trans. Dielectr. Electr. Insul. 2002 , 9

850 .

wileyonlinelibrary.com © 2012 WILEY-VCH Verlag

[ 20 ] M. Lindner , H. Hoislbauer , R. Schwodiauer , S. Bauer-Gogonea , S. Bauer , IEEE Trans. Dielectr. Electr. Insul. 2004 , 11 , 255 .

[ 21 ] K. Y. Yun , D. Ricinschi , T. Kanashima , M. Noda , M. Okuyama , Jpn. J. Appl. Phys. 2004 , 43 , L647 .

[ 22 ] Electrets , 2nd ed. (Ed: G. M. Sessler ), Springer , Berlin 1987 , pp. 52 , 299 , 368 .

[ 23 ] R. K. Zheng , H. W. Gu , B. Xu , K. K. Fung , X. X. Zhang , S. P. Ringer , Adv. Mater. 2006 , 18 , 2418 .

[ 24 ] M. Dawber , K. M. Rabe , J. F. Scott , Rev. Mod. Phys. 2005 , 77 , 1083 . [ 25 ] B. Dickens , E. Balizer , A. S. DeReggi , S. C. Roth , J. Appl. Phys. 1992 ,

72 , 4258 . [ 26 ] S. Horiuchi , Y. Tokunaga , G. Giovannetti , S. Picozzi , H. Itoh ,

R. Shimano , R. Kumai , Y. Tokura , Nature 2010 , 463 , 789 . [ 27 ] R. E. Cohen , Nature 1992 , 358 , 136 . [ 28 ] J. F. Scott , Science 2007 , 315 , 954 . [ 29 ] W. Eerenstein , N. D. Mathur , J. F. Scott , Nature 2006 , 442 , 759 . [ 30 ] S. Sivasubramanian , A. Widom , Y. Srivastava , IEEE Trans. Ultrason.

Ferroelectr. Freq. Control 2003 , 50 , 950 . [ 31 ] D. B. A. Rep , M. W. J. Prins , J. Appl. Phys. 1999 , 85 , 7923 . [ 32 ] A. K. Tagantsev , I. Stolichnov , N. S. Jeffrey , S. Cross , M. Tsukada ,

Phys. Rev. B 2002 , 66 , 214109 . [ 33 ] M. Vopsaroiu , J. Blackburn , M. G. Cain , P. M. Weaver , Phys. Rev. B

2010 , 82 , 024109 . [ 34 ] I. Lubomirsky , O. Stafsudd , Rev. Sci. Instrum. 2012 , 83 , 051101 . [ 35 ] S. B. Lang , F. Steckel , Rev. Sci. Instrum. 1965 , 36 , 929 . [ 36 ] Z. K. Xu , J. W. Zhai , W.-Hu. Chan , H. Chen , Appl. Phys. Lett. 2006 ,

88 , 132908 . [ 37 ] B. Y. Lee , J. X. Zhang , C. Zueger , W.-J. Chung , S. Y. Yoo , E. Wang ,

J. Meyer , R. Ramesh , S.-W. Lee , Nat. Nanotechnol. 2012 , 7 , 351 . [ 38 ] I. Gitlin , J. D. Carbeck , G. M. Whitesides , Angew. Chem. Int. Ed.

2006 , 45 , 3022 . [ 39 ] R. W. Whatmore , Rep. Prog. Phys. 1986 , 49 , 1335 .

GmbH & Co. KGaA, Weinheim Adv. Mater. 2012, DOI: 10.1002/adma.201202079