1 Introduction
Pressure sensors are gaining tremendous attention due to their applications in wearable sensing devices [
1,
2], medical diagnosis [
3,
4], electronic skins (e-skin) [
5,
6], and soft touch displays [
7,
8]. On the basis of the transduction mechanism, pressure sensors can be divided into the piezoresistive type [
9], capacitance type [
10], and piezoelectric type [
11]. Among them, the piezoelectric-type pressure sensors, which rely on the generation of electrical charges in solid materials in response to an applied mechanical pressure, have received special attention because of their high sensitivity and fast response [
12,
13]. Till now, various piezoelectric materials, including polymeric materials [
14], inorganic semiconductors [
15] and ceramics [
16], were employed for the fabrication of piezoelectric devices. Compared with inorganic semiconductors and piezoceramic, piezoelectric polymer materials are flexible and can sustain larger mechanical strains but with the lower piezoelectric coefficient [
17-
19].
As a kind of piezoelectric ceramic materials, BiFeO
3 has attracted increasing attention since the discovery of BiFeO
3 thin film with a remnant ferroelectric polarization [
20]. BiFeO
3 has a rhombohedral twisted crystal structure, in which the hybridization between Bi
3+ lone pairs (6s orbital) and O
2− (2p orbital) results in a displacement of Bi
3+ from a centrosymmetric position, resulting in high spontaneous polarization [
21]. It was demonstrated that BiFeO
3 exhibited high multiferroic properties over a wide temperature range. However, mostly reported bulk BiFeO
3 ceramics showed limited piezoelectric properties due to leakage problems [
22]. Piezoelectricity is the ability of a material to generate an internal electric field when subjected to mechanical stress, which is largely depended on the thickness [
23] and crystal structure of the material [
24], as well as the magnitude of epitaxial strain [
25,
26]. It was reported that piezoelectric materials with a multilayer structure exhibited an improved piezoelectric output, attributing to the electric capacity enhancement of the double-layered structure [
27]. Moreover, DFT calculation of the piezoelectric coefficients of monolayer transition metal dichalcogenides indicated that some of them exhibited stronger piezoelectric coupling than traditionally employed bulk structures [
28]. Therefore, there is a reason to believe that the piezoelectric properties of bulk BiFeO
3 crystals could be improved by breaking them into thinner layers. BiFeO
3 has a layered crystal structure, where the bismuth ions are distributed above and below the ferrite octahedron and share a common edge. A previous theoretical work predicted that the vertical repulsion of bismuth atoms between the layers might make the exfoliation of bulk BiFeO
3 become possible [
29]. Unfortunately, unlike with van der Waals layered materials whose interlayer interaction can be easily overcome, direct liquid phase exfoliation of BiFeO
3 into thin nanoplates/nanosheets remains challenging.
In this study, thin-layer BiFeO3 nanoplates were prepared by a hydrothermal method followed by a freeze-drying assisted liquid exfoliation process for the first time. The liquid exfoliation mechanism was detailed by changing the exfoliation solvent and ultrasonication time. The as-exfoliated BiFeO3 nanoplates with optimized conditions showed lateral sizes of 400−500 nm with thicknesses of 10−20 nm. As a proof of concept demonstration, pressure sensors based on the BiFeO3 nanoplates exhibited largely enhanced pressure response, which is nearly 5 times higher than that of bulk BiFeO3 crystals under an applied pressure of 693 Pa, attributing to interface effect enhanced piezoelectric property.
2 Experiments
2.1 Materials
Bismuth nitrate pentahydrate (Bi(NO3)3.5H2O, > 99%) was purchased from Adamas (Shanghai, China). Ferric chloride hexahydrate (FeCl3·6H2O) and sodium hydroxide (NaOH, > 99%) were purchased from Innochem (Beijing, China). Ethanol (anhydrous, > 99.7%), methanol (anhydrous, > 99.9%), and acetone (anhydrous, > 99.7%) were purchased from Shanghai Chemical Reagent Co., Ltd (Shanghai, China). The main agent and curing agent of polydimethylsiloxane (PDMS) were purchased from Sylgard (Suzhou, China). All the materials were used without further purification.
2.2 Synthesis of bulk BiFeO3 crystals
The raw materials, [Bi(NO3)3·5H2O] and [FeCl3·6H2O] with a stoichiometric molar ratio of 1:1 were dissolved in acetone and distilled water with a volume ratio of 3:10 by continuous stirring until the formation of a clear solution. Then, 13 M ammonia hydroxide solution was added until the pH of the mixed solution reached 11−12. The obtained sediment by centrifugation was washed with distilled water for several times until the pH of the liquid supernatant was 7.0. Afterwards, the sediment was re-dispersed into 5 M NaOH solution and stirred for 30 min. Then, the mixed solution was transferred to a sealed Teflon-lined steel autoclave and heated at 140 °C for 12 h. The as-prepared black powder (BiFeO3) was transferred to 3 mL 4 M HNO3 and stirred for 8 h to remove impurities.
2.3 Freeze-drying assisted liquid exfoliation of bulk BiFeO3
The BiFeO3 powder was dispersed in 2 mL methanol and freeze dried in liquid nitrogen. Afterwards, 0.046 mg freeze-dried powder was re-dispersed in 3 mL methanol and ultrasonicated in a water bath for 8 h at room temperature. Subsequently, the extracted upper suspension was washed by methanol for three times to obtain thin BiFeO3 nanoplates.
2.4 Fabrication of BiFeO3 nanoplate-based pressure sensors
First, the elastomer and curing agent of PDMS (mass ratio = 10:1) were mixed thoroughly and then degassed under vacuum to remove air bubbles. This mixture was then poured into a mold with a thickness of 1 millimeter and cured at 80 °C for 1 h to prepare a PDMS thin film. A copper wire was fixed on the surface of the PDMS film by using conductive silver glue. Two of such films were used as the top and bottom electrodes of a pressure sensor. Finally, pressed BiFeO3 nanoplates was placed between the two parallel electrodes.
2.5 Characterization
The samples were characterized with a scanning electron microscopy (SEM JEOL JSM-7800F, Japan), a transmission electron microscope (TEM JEOL 2100 F, Japan), a high-resolution transmission electron microscope (HRTEM JEOL 2100 Plus, Japan), an atomic force microscope (AFM, MultiMode8, Bruker), an X-ray diffractometer (XRD, SmartLab Rigaku, Japan) with Cu Kα radiation at λ = 1.54 Å, a LabRAM HR Evolution Raman spectrometer (Horiba Jobin Yvon) with a 532 nm laser focused through a 100× objective lens, a polar device (PZT-JH10/4, China) and Hydrostatic electric coefficient (d33) measuring instrument (ZJ-3, China).
3 Results and discussion
3.1 The morphology and microstructure of BiFeO3 nanoplates
Fig.1(a) shows the schematic diagram of the preparation process for BiFeO
3 nanoplates. Bulk BiFeO
3 was prepared by a hydrothermal coprecipitation method with Bi(NO
3)
3·5H
2O and FeCl
3·6H
2O as precursors. BiFeO
3 nanoplates were obtained by combining a freeze-drying process and a following liquid exfoliation process. As shown in the scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images in Fig.1(b) and (c), the bulk BiFeO
3 crystals showed large sizes of 1−2 µm, while the as-exfoliated BiFeO
3 nanoplates show lateral sizes of 400 to 500 nm with thicknesses of 10−20 nm (Fig. S1). The phase structures of the BiFeO
3 crystals before and after exfoliation were determined by XRD. As shown in Fig.1(d), well-defined XRD peaks assignable to (012), (110), (006), (202), and (211) planes of BiFeO
3 (JCPDS No.71-2494) can be clearly observed in both samples. The high resolution TEM (HRTEM) image [Fig.1(e)] of an as-exfoliated BiFeO
3 nanoplate shows interplanar distances of 0.28 nm and 0.23 nm which can be assigned to (110) and (006) planes of BiFeO
3, respectively, suggesting the successful exfoliation of bulk BiFeO
3 by freeze-drying assisted liquid exfoliation method. Moreover, the Raman spectra of bulk BiFeO
3 and BiFeO
3 nanoplates in Fig. S2 both show characteristic modes of E-1, A
1-1 and A
1-2 at 75 cm
−1, 152.6 cm
−1 and 177.5 cm
−1, respectively, further confirming the phase purity of the as-prepared samples [
30].
3.2 The formation mechanism of thin-layer BiFeO3 nanoplates by exfoliation
As described above, the BiFeO
3 nanoplates were successfully prepared by a freeze-drying assisted liquid exfoliation method. Here, to detail the whole formation mechanism, the effect of freeze-drying process during exfoliation was explored. Without the freeze-drying process, BiFeO
3 crystals treated under the same ultrasonication condition showed much thicker morphology as shown in Fig. S3, indicating that the freeze-drying process is an important step in the whole exfoliation process. We first examined the morphologies of the BiFeO
3 crystals before freeze-drying and after freeze-drying with different solvents, and found no obvious morphological changes (Fig. S4). We therefore compared their XRD patterns [Fig.2(a), Fig. S5]. Interestingly, their XRD peaks shift slightly to lower angle after freeze-drying, and the amount of shift is the largest when methanol was used as the solvent for freeze-drying. This suggests that the liquid to solid and then to gas transition of the solvent during freeze-drying can pose stress on the dispersed crystals, and cause lattice strain in them. This agrees with previous reports that freeze-drying was employed to weaken the van der Waals force in layered crystals [
31-
33]. As shown in Fig.2(b), our work demonstrates that the freeze-drying can also induce strain in non-van der Waals crystals such as BiFeO
3. The reason why the freeze-drying process with methanol as the solvent led to the largest lattice strain might be due to the larger volume change of methanol during liquid-to-solid transition as compared to water and ethanol (Tab.1). Such a large strain helps further break up the crystals into thinner nanoplates under subsequent ultrasonication.
After confirming the important role of the freeze-drying process, the effect of the subsequent liquid exfoliation conditions, i.e., the ultrasonication time and solvents were then analyzed. In these control experiments, BiFeO3 crystals treated by the optimized freeze-drying condition with methanol as the solvent were used as the starting material. Fig.3 shows the XRD patterns and TEM images of products obtained after ultrasonication in water, ethanol and methanol for 8 h. Interestingly, the XRD peaks all shift to lower angles (Fig. S6), suggesting further lattice expansion after ultrasonication. The amount of shift is in the order of methanol > water > ethanol. This observation is in line with their morphological difference characterized with TEM as shown in Fig.3(b)−(d). Obviously, the thickness of exfoliated BiFeO3 nanoplates by using methanol is smaller than that of nanoplates exfoliated in water and ethanol. According to the cross-section TEM images in Fig. S7, the thicknesses of typical exfoliated BiFeO3 nanoplates with water and ethanol are 37 and 39 nm, respectively, we further investigated the effect of DMF solvent on the exfoliation and found that the thickness of exfoliated BiFeO3 nanoplates by using DMF (33 nm) as exfoliation solvent in Fig. S8, which indicate the more effective exfoliation with methanol as the solvent. As for the effect of ultrasonication time, it is evident that the exfoliated nanoplates became thinner as the time increased from 4 to 8 h (Figs. S9 and S10).
The reason why methanol is the most effective in facilitating the exfoliation can be explained based on the surface tension and molecular size of the solvent. Compared with water molecules, the surface tension of methanol and ethanol match better with bulk BiFeO
3 (Tab.1) [
34,
35]. However, ethanol has a relatively larger molecule size than methanol, and therefore is more difficult to be inserted inbetween crystal planes of BiFeO
3. This is why methanol as the liquid exfoliation solvent owns the better exfoliation ability compared to ethanol with the larger molecule size or water with the unmatchable surface tension [
36].
3.3 The sensing performance of BiFeO3 nanoplates/PDMS based pressure sensor
As a demonstration, the exfoliated BiFeO
3 nanoplates were fabricated into a pressure sensor to measure pressures from 34.8 Pa to 9.9 kPa [Fig.4(a)]. For comparison, the bulk BiFeO
3 based pressure sensor was also fabricated, but showed a narrower working range that the sensing film was fractured when the pressure exceeded 693 Pa. The fracture of sensing film is closely associated with the intrinsic stress during the fabrication process. In analogy, with the flexible characteristics of two-dimensional materials, exfoliated thin-layer BiFeO
3 nanoplate tend to have the decreased intrinsic stress, which can withstand the larger external pressures compared to bulk BiFeO
3 crystal [
37]. Fig.4(b) shows their responses to varying pressures over the range from 34.8 Pa and 693 Pa. The sensor response is defined by Δ
C/
C0, where
C0 is the initial capacitance and Δ
C is the capacitance variation under an external pressure. The sensitivity was defined by (Δ
C/
C0 × 100%)/Δ
P, where Δ
P is the variation of pressure, The BiFeO
3 nanoplate based sensor showed up to five times of increased responses as compared to the sensor based on bulk BiFeO
3 in the lower pressure range. The sensitivity of the two types of sensors were further calculated by fitting their response-pressure curves, as shown in Fig.4(c). the bulk BiFeO
3 sensor showed a sensitivity of 0.09 kPa
−1 in the range of 34.8−693 Pa, while the sensor based on BiFeO
3 nanoplates presented an improved sensitivity of 0.24 kPa
−1 in the range of 34.8 Pa−9.9 kPa, and the sensor response reached 250% at 9.9 kPa. Besides, the durability of our sensor based on BiFeO
3 nanoplates was evaluated by sensing tests under repeated loading of 34.8 Pa and 251.5 Pa pressure. As shown in Fig.4(d), the sensor exhibited stable sensing responses without obvious degradation after more than 40 cycles, suggesting a good stability.
The enhanced pressure sensing property of BiFeO
3 nanoplates can be explained based on the interface effect promoted piezoelectricity. The piezoelectric effect is the ability of a piezoelectric material to generate an electric field in response to an applied mechanical stress [
38]. A material’s piezoelectric coefficient (
d33) is defined by d
Q/d
F = (
Ud
C +
Cd
U)/d
F, where d
Q is the variation of electric charge, d
F is the variation of external force, d
C is the variation of capacitance and d
U is the variation of voltage [
39]. When the voltage and applied force are given, the
d33 value is directly proportional to changes in capacitance. Therefore, the sensitivity of the capacitance-based pressure sensor increases with increasing piezoelectric effect. The piezoelectric property of a material depends on its atomic orientation, size and edge types. Generally, the bond breaking of edge parts can lead to the relaxation of bond identities, including bond length and strength, which significantly influence the related piezoelectric properties of the system [
40]. Based on the XRD patterns of BiFeO
3 crystals before and after exfoliation, the bond breaking process along with lattice expansion was also involved during the exfoliation process. From their TEM images we can see that the exfoliated nanoplates showed irregular edges, suggesting that the exposed facet/edge types are random and some of them are probably different from those of bulk crystals. These changes of exposed facet/edge types and the increased amount of exposed surfaces can lead to the change of piezoelectric coefficient (
d33) [
41,
42]. We further measured
d33 of bulk BiFeO
3 crystals and exfoliated BiFeO
3 nanoplates as shown in Table S1. It was found that the coefficient of exfoliated BiFeO
3 nanoplates (5.6 pC/N) was two times higher than that of bulk BiFeO
3 crystals (1.77 pC/N). Besides, as illustrated in Fig.5(a), under the equal mass of pressure sensing materials, the bulk BiFeO
3 based pressure sensor possessed less inter-particle interface as compared to the sensor based on exfoliated thinner nanoplates. Therefore, the electric charge induced by external pressure in the sensor based on bulk BiFeO
3 crystals is mostly stored at the interface between the film and the electrodes, while the sensor based on BiFeO
3 nanoplates could have additional electric charges accumulated at the extra interfaces between thin layers, contributing to the enhanced capacitance [
27,
43]. This is further demonstrated in Fig.5(b) that the exfoliated BiFeO
3 exhibited the higher output capacitance than that of bulk BiFeO
3 on the basis of equal mass.
4 Conclusion
In summary, different from the commonly synthesized bulk BiFeO3 crystals, thin-layered BiFeO3 nanoplates were successfully prepared by a hydrothermal method followed by freeze-drying assisted liquid exfoliation process. The as-prepared BiFeO3 nanoplates showed lateral sizes of 400−500 nm. The vital role of freeze-drying and the effect of liquid exfoliation solvent were also detailed. It is believed that the freeze-drying is the precondition of the whole fabrication process, which can generate internal stress to facilitate the subsequent solvent exfoliation. Besides, the methanol solvent with the suitable surface tension and relatively smaller molecule size can be more easily driven into the layers for maximizing the exfoliation effect. Compared with bulk BiFeO3, the as-fabricated pressure sensor based on the BiFeO3 nanoplates exhibited largely enhanced sensitivity with a wider working range of pressure, attributing to enhanced piezoelectric effect induced by the extra electrical charges generated at interlayer interfaces. We hope the adopted method for the fabrication of thin-layer BiFeO3 could extend to the preparation of other non-van der Waals layered crystals for wide applications.