Artificial design for new ferroelectrics using nanosheet-architectonics concept
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2015 Nanotechnology 26 244001 (http://iopscience.iop.org/0957-4484/26/24/244001) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 22.214.171.124 This content was downloaded on 28/05/2015 at 01:43
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 26 (2015) 244001 (6pp)
Artiﬁcial design for new ferroelectrics using nanosheet-architectonics concept Yoon-Hyun Kim1,2, Lei Dong1,2, Minoru Osada1,2, Bao-Wen Li1, Yasuo Ebina1 and Takayoshi Sasaki1 1
International Center for Materials Nanoarchitectonics (MANA), National Institute for Materials Science, Tsukuba, Ibaraki 305-0044, Japan 2 Grduate School of Advanced Science and Engineering, Waseda University, Shinjyu-ku, Tokyo 169-8555, Japan E-mail: [email protected]
Received 1 December 2014, revised 6 March 2015 Accepted for publication 9 April 2015 Published 27 May 2015 Abstract
Control over the emergence of ferroelectric order remains a fundamental challenge for the rational design of artiﬁcial materials with novel properties. Here we report a new strategy for artiﬁcial design of layered perovskite ferroelectrics by using oxide nanosheets (high-k dielectric Ca2Nb3O10 and insulating Ti0.87O2) as a building block. We approached the preparation of superlattice ﬁlms by a layer-by-layer assembly involving Langmuir–Blodgett deposition. The artiﬁcially fabricated (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattices are structurally unique, which is not feasible to create in the bulk form. By such an artiﬁcial structuring, we found that (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattices possess a new form of interface coupling, which gives rise to ferroelectricity with a good fatigue-free characteristic. Considering the ﬂexibility of self-assembled nanosheet interfaces, this technique provides a route to synthesize a new kind of layered ferroelectric oxides. Keywords: oxide nanosheets, nanoarchitectonics, artiﬁcial superlattice, layer-by-layer assembly, ferroelectricity (Some ﬁgures may appear in colour only in the online journal) 1. Introduction
been two well-known approaches to control ferroelectric properties in these systems: strain engineering [5–7] and tailoring electrostatic boundary conditions . Besides these approaches, a more complete picture of the engineered interfaces in these systems also needs to incorporate the possibility of additional structural distortions, electronic redistributions, and interface polarization. Such protocols relying on interface engineering open up pathways to create new artiﬁcial ferroelectrics with tailored properties. Here, we report a new strategy for artiﬁcial design of layered perovskite ferroelectrics by the assembly of oxide nanosheets in a unit-cell-upon-unit-cell manner (ﬁgure 1). Ferroelectricity is ubiquitous in the ABO3 perovskite titanates, niobates, and tantalates. However, the only layered perovskites to display a ferroelectric transition has been Aurivillius compounds such as SrBi2Ta2O9, Bi4−xLaxTi3O12, and a few others [9, 10]. These layered perovskites consist of perovskite ‘building blocks’ of TiO6, NbO6, or TaO6
Over the past decades, scientists have aspired to exploit bottom-up approaches to create new artiﬁcial materials with hierarchical structures and tailored properties. As a direction for the bottom-up assembly, dimension-reduced approaches such as layering 2D nanostructures become an important target for a rational design of precisely controlled lamellar nanomaterials with tailor-made properties. State-of-the-art thin-ﬁlm technology now allows a layer-by-layer epitaxial growth of multicomponent thin ﬁlms with atomic level control . This makes possible a wide range of novel artiﬁcial materials having physical phenomena that are not observable in bulk materials. One of the fundamental challenges for such a rational design is control over the emergence of ferroelectric order [2–4]. Complex oxide heterostructures, layered thin ﬁlms, and other low-dimensional systems provide an important platform to address this ongoing challenge. There have 0957-4484/15/244001+06$33.00
© 2015 IOP Publishing Ltd Printed in the UK
Y-H Kim et al
Nanotechnology 26 (2015) 244001
Figure 1. Artiﬁcial design for new ferroelectrics using nanosheet architectonics.
octahedra, separated at intervals by bismuth oxide planes. Ferroelectricity arises from a highly polarized nature of perovskite-type layers and the fatigue-free property is mainly attributed to the insulating (Bi2O2)2+ layers. (Bi2O2)2+ layers may minimize the interlayer stress upon polarization. Our principal concern is artiﬁcial design of layered perovskite ferroelectrics by oxide nanosheets as a building block. Oxide nanosheets obtained via exfoliation of layered compounds possess a high 2D anisotropy with a molecular thickness (typically ∼1 nm) and, therefore, can be regarded as the thinnest self-standing 2D nanostructures having functionalities inherent from the parent compounds [11–15]. We focus on two different oxide nanosheets [16, 17]: high-k dielectric Ca2Nb3O10 and insulating Ti0.87O2. An important aspect is that Ca2Nb3O10 nanosheet consist only of three NbO6 octahedral nanoblocks susceptible to a large molecular polarizability, which makes the nanosheet an ideal base for high-k dielectrics and ferroelectrics with a critical thickness. In addition, due to their polyelectrolytic nature, oxide nanosheets can be employed as a building block for the electrostatic layer-by-layer self-assembly [18–21]. Using Ca2Nb3O10 as a precursor, we prepared a layered perovskite-type oxide sandwiched with insulating Ti0.87O2. This artiﬁcial oxide has a structure like Bi4−xLaxTi3O12 with three TiO6 octahedral nanoblocks. Also, considering the ﬂexibility of self-assembled nanosheet interfaces, our nanosheet technique provides a route to synthesize a new kind of layered ferroelectric oxide. Here, we describe the results of the preparation, characterization, and ferroelectric properties of artiﬁcial layered ferroelectric based on oxide nanosheets.
Figure 2. AFM images of (a) Ca2Nb3O10 and (b) Ti0.87O2
nanosheets. A tapping-mode AFM in vacuum condition was used to evaluate the morphology of the nanosheets on Si substrate.
were 1.85 ± 0.10 nm, 3–10 μm for Ca2Nb3O10 and 1.0 ± 0.10 nm, 10–20 μm for Ti0.87O2, respectively. The thickness values obtained are nearly comparable to the crystallographic thickness of the host layer in the corresponding parent compounds, supporting the formation of unilamellar nanosheets [22, 24]. We approached the preparation of superlattice ﬁlms by a layer-by-layer assembly involving the Langmuir–Blodgett (LB) process [20, 21] (ﬁgure 3). A diluted colloidal suspension of the nanosheets was used as a subphase to form LB ﬁlms without amphiphilic molecules. During LB deposition, the packing density of the nanosheets in the ﬁlm could be controlled by the surface pressure of the air–water interface. The best results were obtained for the surface pressure at 15 mN m−1, yielding >95% coverage with only occasional overlaps and gaps. The procedure for the LB depositions involving two different nanosheets was repeated an appropriate number of times to synthesize a superlattice assembly. The use of an atomically ﬂat SrRuO3 substrate is effective for the fabrication of an atomically uniform superlattice ﬁlm with a highly dense characteristic. These ﬁlms were irradiated by UV white light from a Xe lamp (4 mW cm−2) for 48 h in order to decompose tetrabuthylammonium (TBA) hydroxide ions used in the exfoliation process. The ﬁnal product was identiﬁed as an inorganic multilayer assembly accommodating NH 4+ ions as a consequence of total photocatalytic removal of TBA ions.
2. Artiﬁcial superlattices fabricated by nanosheet architectonics Oxide nanosheets (Ca2Nb3O10, Ti0.87O2) were prepared by delaminating layered perovskites (KCa2Nb3O10, KTi0.87O2) according to previously described procedures [22, 23]. This exfoliation process deduced the colloidal suspension of monodispersed and unilamellar nanosheets, which are appropriate for the layer-by-layer assembly. Atomic force microscopy (AFM) images (ﬁgure 2) revealed a sheet-like morphology, which is inherent to the host layer in the parent compounds. The average thicknesses and lateral dimension 2
Y-H Kim et al
Nanotechnology 26 (2015) 244001
Figure 3. Fabrication procedure for superlattice ﬁlms using the LB method.
Figure 5. Interface characterization by HRTEM and EELS. (a)
HRTEM image of (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattice on SrRuO3 substrate. (b) The integrated intensity proﬁles of the Ti-L2,3 and Ca-L2,3 edges across the interface.
Figure 4. UV–visible absorption spectra in the superlattice
assemblies: (a) (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) and (b) (Ti0.87O2) (Ca2Nb3O10)2(Ti0.87O2). (c) UV–visible absorption spectra taken from the monolayer ﬁlms of Ti0.87O2 and Ca2Nb3O10 nanosheets.
235 nm is characteristic of Ca2Nb3O10, whereas the absorption peak at 275 nm is attributable to Ti0.87O2. These features are clear from the data for the monolayer ﬁlms composed of each nanosheet (ﬁgure 4(c)). In the superlattice ﬁlms, each deposition cycle causes progressive enhancements for superlattice-type buildup, and the spectral proﬁle observed can be understood in terms of the superimposition of the proﬁle of each nanosheet in the designed sequences.
The layer-by-layer assembly process of nanosheets could be monitored by the UV–visible absorption spectra. Figures 4(a) and (b) show the buildup process monitored with UV–visible spectroscopy in two different superlattices on the quartz glass substrate. The broad absorption centered at 3
Y-H Kim et al
Nanotechnology 26 (2015) 244001
substrate surface is covered with the superlattice nanoﬁlm, and the regular growth structure composed of Ti0.87O2 and Ca2Nb3O10 layers is clearly seen. The thicknesses of the constituent layers were approximately 1.0 and 1.6 nm, which are in good agreement with the crystallographic thicknesses of Ti0.87O2 and Ca2Nb3O10 nanosheets, respectively. Electron energy-loss spectroscopy (EELS) in TEM was used to probe compositional changes in the (Ti0.87O2/ Ca2Nb3O10)2(Ti0.87O2) superlattice (ﬁgure 5(b)). The integrated intensity proﬁles of the Ti-L2,3 and Ca-L2,3 edges clearly resolve a compositional modulation in the lamellar parts; the Ti-L2,3 and Ca-L2,3 peaks are alternatively appeared at 1.0 and 1.6 nm steps. We also note that there are no detectable interdiffusion and strains at the interface. Ti-L2,3 and Ca-L2,3 edge spectra showed a compositional abruptness between SrRuO3 substrate and subsequently deposited nanosheets. These results indicate the production of a deadlayer-free perovskite superlattice directly assembled on the SrRuO3 substrate. The superlattice ﬁlms of a similar quality were achieved in different stacking sequences and different substrates such as SrTiO3:Nb or Si. Therefore, our superlattice approach is quite general; the exact control of interface engineering can be realized in the self-assembled superlattices. The artiﬁcially fabricated (Ti0.87O2/ Ca2Nb3O10)2(Ti0.87O2) superlattices are structurally unique, which is not feasible to create in the bulk form. In the case of typical superlattices, in which the in-plane lattice parameter of all the constituents is constrained to that of the underlying substrate. The principal consideration is the minimization of polarization mismatch between layers, any mismatch giving rise to very high electrostatic energy penalties. Our nanosheet approach is based on soft-landing process of oxide layers, offering a route to synthesize a new kind of superlattices without elastic constraint imposed by the substrates.
Figure 6. (a) P−E curves taken from two superlattices: (Ti0.87O2/ Ca2Nb3O10)2(Ti0.87O2) [SL-1] and (Ti0.87O2) (Ca2Nb3O10)2(Ti0.87O2) [SL-2]. (b) Fatigue test (at 1 kHz) of (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattice [SL-1].
3. Engineered dielectric/ferroelectric responses in nanosheet superlattices We now turn to dielectric/ferroelectric properties of nanosheet superlattices. The constituent nanosheets are not ferroelectric, but paraelectric with a highly insulating nature [16, 17]. Here, question is whether engineered interface can control over the emergence of ferroelectric order in the self-assembled perovskite superlattice. We note that the artiﬁcially fabricated (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattice has an analogous structure as Bi4−xLaxTi3O12 . Also, considering the ﬂexibility of self-assembled nanosheet interfaces, it is hopeful to synthesize a new kind of layered ferroelectric oxide with good fatigue-free property in similar to Aurivillius compounds. Figure 6(a) shows the P−E curves of two superlattices. The (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattices showed a clear hysteresis behavior inherent to ferroelectricity. Ferroelectric properties are dependent on the stacking order of nanosheets: (Ti0.87O2)(Ca2Nb3O10)2(Ti0.87O2) with thicker perovskite layer exhibited a reduced polarization compared to
Figure 7. Dielectric properties of (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattice and multilayer ﬁlms of (Ti0.87O2)10 and (Ca2Nb3O10)10. The estimated εr values from the parallel capacitor model are also indicated by a dotted line.
Cross-sectional high-resolution transmission electron microscopy (HRTEM) observations (ﬁgure 5) provide direct information on the superlattice formation. An image of the (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattice revealed that the 4
Y-H Kim et al
Nanotechnology 26 (2015) 244001
for the ferroelectric state. We expect a similar ferroelectric instability in the present case; the weakly bonding of the adjoined nanosheets can facilitate a zone-boundary distortion involving tilts of the oxygen octahedral at the interface. In such a case, the ferroelectric/antiferrodistortive coupling is very strong at the interface, and such an interface effect produces a net coupling between the ferroelectric motion and the oxygen in-plane anti-ferrodistortive motion, a situation being similar to the improper ferroelectricity observed in ferroelectric/paraelectric superlattices . The ﬂexibility of self-assembled nanosheet interfaces facilitates a ferroelectric distortion: Ti0.87O2 layers may minimize the interlayer stress upon polarization, causing a fatigue-free nature. In this context, reduced polarization of (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) is possibly due to a less distorted nature of the perovskite interface. Although full understanding on ferroelectric character awaits further studies, our approach provides a new route to realize artiﬁcial lead-free ferroelectric, which enables it to comply with the RoHS regulation.
Figure 8. Schematic view of the (Ti0.87O2/Ca2Nb3O10/Ti0.87O2) interface and ferroelectric distortions.
the (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) case. In this paper, we focus on the (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) case. The hysteresis loop of the (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattice showed a well-saturated P−E loop with the remanent polarization (Pr) of 7.1 μC cm−2. The polarization property exhibited a fatigue-free characteristic with no signiﬁcant degradation up to 1010 switching cycles (ﬁgure 6(b)). These ferroelectric responses are an intrinsic signature due to the interfacial effect and not a consequence of the leakage current; the (Ti0.87O2/Ca2Nb3O10)2(Ti0.87O2) superlattice exhibits a highly insulating nature with leakage current density of