Скачать 41.91 Kb.
|The Micro-structure Studies of Ni-BaTiO3 Nanocomposite Films by TEM and EELS|
CHEN Shijuan1,2, GE Fangfang2, MA Yongjun2, WANG Xuemin2, CHEN Liying2, HAN Shangjun2, ZHANG Hong1, WANG Hongbin2, TANG Yongjian2，WU Weidong 2,＊
(1 Institute of Atomic and Molecular Physics, Sichuan University, Chengdu 610065 China;
2 Research Center of Laser Fusion, CAEP, P.O. Box 919-983, Mianyang 621900, China)
Abstract: Epitaxial BaTiO3 films with embedded metallic Ni nanocrystal (Ni-BaTiO3) were successfully fabricated on SrTiO3 (001) single-crystalline substrate through the laser molecular beam epitaxial (L-MBE) technique. High resolution transmission electron microscopy (HRTEM) and electron energy loss spectrum (EELS) with Kramers-Kronig analysis methods were employed to characterize the microstructures, elementary distribution and the electron structure of these films. HRTEM results suggested that the structure of BaTiO3 was tetragonal with lattice parameters of a=0.399 nm and c=0.403 nm. Energy dispersive X-Ray spectroscopy (EDX) confirmed metallic Ni nanocrystal embedded successfully in BaTiO3 epitaxial films. The Ni-BaTiO3 composite films were compound of the epitaxial BaTiO3 (110) layers alternating with Ni NCs array (111) layers. Furthermore, the existence of the misfit dislocations induced by the embedding of Ni nanoparticles was also clearly demonstrated by the HRTEM images. The Ni L2, 3 edges of EELS revealed that Ni NCs in their metallic state were embedded uniformly in the BaTiO3 matrix. A chemical shift of about 7 eV regarding L3 edges in the Ni EELS was also observed. The optical band gap of BaTiO3 in these films was about 3.84 eV, bigger than 3.55 eV for pure BaTiO3 films at room temperature.
Keywords: Laser molecular beam epitaxial; transmission electron microscopy; electron energy loss spectrum; Kramers-Kronig analysis
Ferroelectric thin films have attracted considerable attention because of their potential use in nonvolatile memory devices, [1-3] nonlinearity optical devices,  thin-film capacitors, and pyroelectric detectors. The perovskite-type semiconductor Barium titanate (BaTiO3) is a remarkable ferroelectric material with a high dielectric constant, large ferroelectric response, high electron-optic coefficient, and a second- and third-order nonlinear susceptibility . Nanocomposite thin films are fabricated by embedding metallic nanocrystals (MNCs) in BaTiO3 matrix which can dramatically improve the intrinsic properties of pure BaTiO3 and make it a much more attractive material in several applications. For instance, magnetic MNCs (such as Fe, Co, Ni etc.) embedded in BaTiO3 matrix can greatly enhance the third-order nonlinear susceptibility (3) compared with the pure dielectric materials [6,7]. A large third-order optical nonlinearity is very useful for optical switching devices so it is critical to fabricate high-quality MNCs-BaTiO3 composite films.
It was observed by Chen and her co-worker [8-12] that the third-order nonlinear susceptibility (3) and the ratio (Re) (3)/Im(3) were obviously enhanced when geometric anisotropy nanoparticles were embedded in the BaTiO3 matrix. These results were consistent with the theoretical calculations of Sheng et al [13, 14]. Recently, the growth mechanism by reflection high-energy electron diffraction (RHEED), the structure characterized by atomic force microscopy (AFM) and X-ray diffractometry (XRD), and the optical properties by absorption spectra (AS) and photoluminescence (PL) spectra of metal-BTO nano-composition films have been studied by Wu and his co-worker [15-18]. However, there are few reports studying the micro-structure, electron-structure and misfit dislocation of these nanocomposite films using transmission electron microscopy (TEM) and electron energy loss spectrum (EELS). The properties of these films depend mainly on the size and shape of the nanoparticles, the embedding environment, and many other factors [19, 20] so it is necessary to study the properties and electron-structure of Ni nanocrystals (NCs).
In this study, the L-MBE method was employed to fabricate the Ni-BaTiO3 composite films, and TEM and EELS were used to analyze them. The aim of this study is to characterize the micro-structure of the as-prepared films and the spatial distribution as well as electronic structure of the embedded Ni particles in detail. The optical band gap of BaTiO3 matrices was investigated and the embedded Ni particles did introduce chemical shifts to Ni-BaTiO3 films
Ni-BaTiO3 nanocomposite films were deposited on the SrTiO3 substrate (001) layer by layer using L-MBE. A KrF excimer laser (λ=248 nm) was focused at a 45° angle onto a ceramic stoichiometric BaTiO3 target. The experimental methods were discussed in details previously. 
Grain size, grain boundary conditions, interfacial microstructure and the element distribution of the Ni-BaTiO3 thin films were investigated using high resolution transmission electron microscopy (HRTEM) with a field-emission gun of 200 keV. Cross-section TEM samples were prepared by focused ion beam (FIB) using Ga gas. The experimental HRTEM images were recorded with a charge-coupled device (CCD) camera. EELS were acquired in diffraction mode with 5 eV/channel dispersion, an aperture of 2 nm and a collection semi-angle of 1.195 mard along with a convergence semi-angle of 3.826 mard. The spectrum resolution was determined by measuring the full width at half maximum (FWHM) of the elastic peak and this was close to 1.5 eV.
Gatan DigitalMicrograph TM software was used to analyze HRTEM images and EEL spectra. The data processing of the low energy (0-100 eV) EELS was used to calibrate the zero-loss peak. Then the 1st Log-polynomial mode was used for removal background and the plural scattering was removed by the Fourier-log deconvolution method .
3 Results and Discussion
3.1 HRTEM and the selected area diffraction (SAD) of the Ni: BaTiO3 films
Fig.1 (a) is the energy dispersive X-ray spectroscopy (EDX) analysis of the Ni-BaTiO3 films, which confirms the existence of Ni NCs with a relatively low concentration in BaTiO3 matrix. It can be seen that the additional substances such as Ga, Cu, and C are introduced into the TEM measurement. The Cu element comes from the Cu net during the TEM experiment; Ga and C elements exist because of the focused ion beam (FIB) . The existence of these extra elements does not affect the analysis of the Ni-BaTiO3 films. Some degree of surface damage is observed in FIB-milled samples, as shown in Fig.1 (b). The black spots may be Ni NCs since the background is a BaTiO3 matrix. There is no significant difference in grain size across the films' thicknesses, as clearly shown in the undamaged area in Fig .1 (b). Fig.1(c) is the energy-filtered transmission electron microscopy (EFTEM) of the Ni element. It is obviously seen that the bright places are Ni NCs. It is well in agreement with Fig.1 (b) and proves that the uniform black spots in a line are corresponding to Ni NCs in the Fig.1 (b). All the figures confirm that Ni has been successful incorporated in the BaTiO3 matrix, and the measured size of Ni NCs are around 6 nm, as depicted by Fig1. (d).
Fig. 1(a) The EDX of the Ni-BaTiO3 films; Fig. 1(b) the cross-sectional TEM image of Ni-BaTiO3 nanocomposite films -- it consists of eight Ni NCs layers alternating with BaTiO3; Fig. 1(c) the Energy filter image of Ni element; and Fig. 1(d) HRTEM of the Ni-BaTiO3 films.
Fig.2. (a) is the HRTEM micrograph which shows the microstructure of the Ni and BaTiO3 coexisting system. The Ni nanoparticles are determined to be endowed with a face-center cubic (fcc) structure along the axis of (111). The BaTiO3 is along the (110) crystal plane. The perfect electron-diffraction pattern is shown in Fig.2 (a). It can be deduced that BaTiO3 is tetragonal with a=0.399 nm and c=0.403 nm. This is the same as the bulk BaTiO3 at room temperature, implying that the BaTiO3 matrix is a perfect single phase and it grows very well in Ni-BaTiO3 films.
Fig. 2(a) The HRTEM micrographs of the BaTiO3 matrix, the Ni nanoparticles, and the Ni-BaTiO3 interfaces are indicated in the Figure; Fig. 2(b) the SAED pattern from Figure (a)
Fig. 3(a) The HRTEM of the Ni-BaTiO3 films; and Fig. 3(b) the Fourier masking from mask of (a) the boundary of Ni and BaTiO3 matrix, A and B areas are the BTO but B area is Ni Nanoparticles.
Fig.3 (a) is HRTEM of Ni and BaTiO3 coexistence area. Fig.3 (b) is a Fourier masking image for the mask place of (a). It has been divided into three areas. A and C areas are both BaTiO3 matrix. B is the Ni NCs. It is observed that the lattices are highly organized and grow very well in the A area. In the proximity of B area however, the lattices appear to have some dislocations such as the atom planes missing or other defects. But when getting farther from the Ni NCs into the C area, the films grow well again. The above 11% mismatch between BaTiO3 and Ni NCs is deduced from the formulation:
where, af, as are lattice parameters of Ni and BaTiO3 respectively. The big lattice mismatch probably leads to the misfit dislocations.
3.2 Electron Energy Loss Spectrum (EELS) and Kramers-Kronig(K-K) transformation
Fig.4 is the smoothing spectrum of the EELS of the Ni element. It has been stripped of the pre-edge background by fitting it to a power law AE-r . The Ni L2, 3 edges confirm that Ni is in its metallic state, not as an oxide [23, 24].
Fig. 4 The EELS of the Ni element.
Actually the EELS edge energies may vary about 7eV due to the bonding energy during the procedure of the experiment. But after being taken into consideration of the bonding energy, the Ni L3 edge still has a chemical shift of about 7 eV, which should be caused by its chemical environment. BaTiO3 is a ferroelectric material, so there are many reasons for the chemical environment changing. For example, the strong electron beam of the TEM and the embedding of metallic Ni NCs both can cause very high local electron fields around the BaTiO3 matrix. Furthermore, the outer electrons have bigger influences than inner ones, that is why L3 has a 7 eV left chemical shift but L2 is almost normal. More reasons and mechanisms of this phenomenon will be researched deeper in the future.
The observed energy-loss spectrum is closely related to a quality referred to as the energy-loss function, [-1/(E)] via this relationship 
where, S(E) is the single scattering distribution, K is a proportionality constant and and E are the effective collection and characteristic scattering angles respectively. Hence, after suitable corrections are applied and scaling using the proportionality constant, the energy-loss function can be retrieved from the low-loss spectrum with no plural scattering. It is related to the energy-loss function via the Kramers-Kronig transformation 
where, P indicates the Cauchy principal part of the integral should be taken to avoid singularities at E' = E. Once Re[1/(E)] has been retrieved, the real and imaginary parts of the dielectric function can be calculated from [-1/(E)] and Re[1/(E)] via the relationship 
Then the optical absorption co-efficient α can be deduced from the following formulation 
where, c is the speed of light in a vacuum.
Putting the parameters into Kramers-Kronig analysis — where the number of iterations is 20 and optical refractive index set to be 100 — the optical absorption α can be calculated.
Because BaTiO3 is a direct gap material, it can deduced that the optical band gap from the following formulation 
where, Eg is the optical band gap of BaTiO3.
Fig. 5(a) The EELS (0-100 eV) after stripping off the background using 1st Log-polynomial mode; Fig. 5(b) the optical absorption α from KKT of the EELS (0-100 eV); Fig. 5(c) the band gap of the films, and Fig. 5(d) the magnifying image of the band gap.
Fig.5 (a) is the spectrum after removal of the background of low energy (0-100 eV) EELS using 1st Log-polynomial mode .The optical absorption coefficient is shown in Fig.5 (b). Fig.5(c) is the optical band gap image. Fig.5 (d) is a magnifying image of Fig.5(c) during 3-7eV. The optical band gap of Ni-BaTiO3 films is around 3.84 eV, bigger than the bulk BaTiO3 (3.55 eV) at room temperature. Embedding the Ni NCs probably gives rise to this phenomenon. Because the above 11 % lattice mismatch between BaTiO3 and Ni NCs will cause strain, there are many misfit dislocations appearing in the interface to decrease this strain. The deformation areas nearby the misfit dislocations of the interface between BaTiO3 and Ni NCs can give rise to deformation energy. And they can introduce a localized state of energy. Furthermore, the strain caused by mismatching can at least change the valence band or the conduction band, thus making the band structure change . This is probably the reason why there are chemical shifts and optical band gap transference.
In summary, we successfully prepared epitaxial BaTiO3 films embedded with metallic Ni nanocrystal using the L-MBE technique. And the final thickness of the films was around 370 nm. Combined methods of TEM, EDS, and EELS were applied to characterize the structures of these films. SAED pattern revealed that pure BaTiO3 matrix grows very well in these films with lattice parameters of a=0.399 nm and c=0.403 nm (which is very close to the value of tetragonal BaTiO3). The embedded Ni particles were verified to be metallic, and a chemical shift of 7 eV for Ni L3 EELS edges was observed. The HRTEM indicates that the Ni-BaTiO3 composite films were a compound of the epitaxial BaTiO3 (110) layers alternating with Ni NCs array (111) layers. The big lattice mismatch of 11 % between Ni and BaTiO3 was potentially responsible for this remarkable chemical shift. The band gap of BaTiO3 in these films is around 3.84 eV, which is larger than 3.55 eV for the bulk pure BaTiO3 at room temperature. Embedding Ni NCs and the big mismatch probably gave rise to this phenomenon.
作者简介：CHEN Shijuan(陈士娟): 1987-, 河南省（信阳市）淮滨县人，硕士研究生。