Giant energy storage density in lead-free dielectric thin films deposited on Si wafers with an artificial dead-layer

Highlights

• High quality lead-free thin films deposited on Si wafers were obtained.

• An ultrahigh energy storage density (W) on Si wafers ~ 89 J/cm³ was achieved.

• Artificial Dead-layer implantation benefits the enhancement of W and DBS.

• Artificial Dead-layer implantation benefits the improvement of thermal stability and fatigue endurance.

• High electrons injection barrier can suppress Schottky Emission.

Abstract

High-performance lead-free thin-film capacitors deposited on the silicon (Si) wafers with large energy storage density (W) and high reliability are strongly attractive in the modern electrical and electronic devices. Here, an ultrahigh W was achieved in the Ba_0.3Sr_0.7Zr_0.18Ti_0.82O₃ (BSZT) relaxor ferroelectric thin films deposited on the Si wafers with the help of an ultrathin Ca_0.2Zr_0.8O_1.8 (CSZ) artificial “dead-layer” simultaneously possessing high resistivity, wide band gap and high permittivity among linear dielectrics. As the CSZ was implanted, the W of the Ba_0.3Sr_0.7Zr_0.18Ti_0.82O₃ (BSZT) thin films was greatly increased from 64.9 J/cm³ to 89.4 J/cm³, which is comparable to the best W of thin film deposited on expensive single crystal substrates, and is the largest one reported so far than those of lead-free thin films deposited on the Si wafers, and even for lead thin films. Due to the formation of ultrahigh electrons injection barrier (3.92 eV) between the interface of the CSZ dead layer and the Au top electrode, the Schottky emission of the BSZT thin films under high electric field and at high temperatures was effective suppressed, which is responsible for the greatly improved dielectric breakdown strength and thermal stability. Moreover, the fatigue endurance was also enhanced. It is concluded that the implantation of the CSZ artificial dead-layer could be used as a universal-simple-effective strategy to improve the electrical performances of ferroelectric materials working in the harsh environment of high electric field.

Introduction

Energy storage techniques are being pushed toward high efficiencies, high storage capability and environmentally friendly by the ever-increasing energy requirement and the sustainable development of the society [1,2]. Dielectric capacitors show superior advantages over other energy storage devices due to their high-power density and voltage endurance, good reliability, long cycling lifetime and simple device structure. Different kinds of dielectric capacitors are also critical components widely used in electronic devices and electrical power systems [[3], [4], [5]]. Recently, much attentions have been paid on how to improve the energy storage capability of dielectric thin films to meet the novel and advanced technologies applications. Many efforts have been developed to improve the energy storage density (W) of dielectric capacitor. It is agreed that the dielectrics with a medium dielectric constant (ε_r), high dielectric breakdown strength (DBS), and a slim hysteresis loop (low remnant polarization P_r and high maximum polarization P_m) are highly desired to realize a high energy density and high energy efficiency in the dielectric capacitor [6,7]. Relaxor ferroelectrics (RFE) and anti-ferroelectrics (AFE) have attracted much attention due to their unique polarization-electric field (P-E) hysteresis loops.

In the past few decades, the excellent energy storage properties were achieved in lead-based RFE and AFE films [[8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18]]. Ma et al. reported an ultrahigh W of 85 J/cm³ at room temperature in the Pb_0.92La_0.08(Zr_0.52Ti_0.48)O₃ (PLZT) thin films on LaNiO₃-buffered Ni (LNO/Ni) foils [18]. Zhongqiang et al. reported that an ultrahigh W of 61 J/cm³ was realized in the PLZT thin films deposited on Pt coated silicon wafers [11]. Recently, a lot of breakthrough progresses have been made on the energy storage performance in the lead-free thin films [6,7,[19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32]]. For instance, a giant W of 166 J/cm³ at a high electric field of 5.7 MV/cm was realized in the epitaxial Ba(Zr_0.2Ti_0.8)O₃ thin films by the domain engineering [19]. A markedly enhanced W of 137 J/cm³ to 154 J/cm³ was obtained in the (111)- and (100)-oriented La/Zr co-doped (1-x)(Bi_1/2Na_1/2)Ti–xBaTiO₃ epitaxial lead-free RFE thin films by the FE/AFE composition design, respectively [6]. The W of 112 J/cm³ was reported in the lead-free BiFeO₃-BaTiO₃-SrTiO₃ solid-solution RFE films by the polymorphic nanodomain design [7]. The W of 86 J/cm³ was reported in the 0.88Ba_0.55Sr_0.45TiO₃–0.12BiMg_2/3Nb_1/3O₃ RFE thin films [20]. By using pulsed laser deposition, the energy storage density of 86 J/cm³ was realized in the 0.25BiFeO₃-0.75BaTiO₃ thin films [21]. Though controlling the number of interfaces, a high W of 83.9J/cm³ was obtained on the BaZr_0.15Ti_0.85O₃/BaZr_0.35Ti_0.65O₃ multilayer thin films [22]. A high W of 78.7 J/cm³ was obtained in the BaZr_0.35Ti_0.65O₃ relaxor thin films [23]. However, these films with ultrahigh W were fabricated on the SrTiO₃ or Nb-doped SrTiO₃ single crystal substrate, which leads to high fabrication cost and were difficult to realize a large-area fabrication or mass production due to the high cost of SrTiO₃ substrate and lack of large substrate sizes [9]. For industrial applications, high-performance lead-free dielectric thin films should be integrated on Si wafers with low cost by using standard integrated circuit (IC) process. However, the W of these reported lead or lead-free RFE and AFE thin films deposited on the Si wafers still cannot be competitive with these reported films grown on the SrTiO₃ single crystal substrate. Thus, there is an urgent requirement to develop the high-performance lead-free RFE or AFE thin films with high W on the Si wafers to meet the demands of advanced electronic and electrical systems for integration, compactness, miniaturization and environmentally friendly [5,7,33].

The energy storage density W can be calculated from P-E loops by this equation as follows [6]:

$$W=\underset{P_r}{\overset{P_m}{\int }}EdP$$

where E is the external electric field, and P the polarization.

According to the equation, improving the discrepancy ΔP between P_m and P_r or DBS can be both effective approaches to improve W. Due to the trade-off between polarization and DBS [5], increasing ΔP was often realized at relatively low electric field. Some effective strategies, like multiple polar structures design [34], multilayer heterostructure design [35], ect, was proposed to improve the ΔP for a high W at low electric field. Besides, A large number of studies have also indicated that improving the DBS could be a positive solution for an ultrahigh W in the lead-free dielectric thin films. Therefore, in recent years, some proposed familiar strategies of enhancing DBS can be concluded as follows: a) reducing the internal defects; b) optimizing the films growth process. c) suppressing or regulating the movement of internal defects; d) stopping the development of electric trees in the films; e) “dead-layers” engineering. The cyclic cooling heating process was offered to restore the internal defects in the STO films to improve breakdown voltage from 40 V to 68.4 V [2]. By means of oxygen annealing, an enhanced DBS of 2000 kV/cm was obtained in Ba_0.85Ca_0.15Ti_0.90Zr_0.10O₃ thin films, which is twice that of the thin film annealed in the air [24]. The optimizing the films growth process mainly include utilization of perovskite oxide electrodes (ABO₃, such as LaNiO₃ [36], SrRuO₃ [9], etc.), and the uses of single crystal substrates [19], like SrTiO₃, LaAlO₃, etc. By using these substrate or electrode, a high-quality epitaxial or textured thin film can be obtained with less structure defects than the polycrystalline film, and then the DBS was improved. Recently, suppressing or regulating the movement of internal defects in the films starts to become an effective approach to enhance the DBS. The vertical interface in the vertically aligned heteroepitaxial (BaTiO₃)_0.5:(Sm₂O₃)_0.5 nanocomposite thin films can become a sink to attract the oxygen vacancies, and consequently the DBS was enhanced [37]. The low-temperature-poling method was firstly proposed to make the defect dipoles highly ordered at low temperature under the high electric field [38]. After this pretreatment, the DBS and W in the Pb_0.97La_0.02(Zr_0.905Sn_0.015Ti_0.08)O₃ thin films at room temperature are greatly enhanced from 1286 kV/cm to 2000 kV/cm, and from 16.6 J/cm³ to 31.2 J/cm³, respectively. Besides, stopping the development of electric trees by using multi-interface blocking effect in the films can be also proved as an effective approach to improve the DBS [26]. When the periods n increases from 2 to 8, the DBS and W in the (Ba_0.7Ca_0.3TiO₃/BaZr_0.2Ti_0.8O₃)_n thin films at room temperature are greatly enhanced from 3 MV/cm to 4.5 MV/cm, and from 26.4 J/cm³ to 52.4 J/cm³, respectively. Moreover, the concept of “dead-layers” engineering was firstly proposed by McMillen et al. [39]. A parasitic “dead-layer” can be used to increase the DBS and W though using high DBS ultrathin layer to share the load of the voltage applied on the dielectric films. In this study, a 6 nm Al₂O₃ ultrathin layer was obtained as this “dead-layer” by oxidizing the ultrathin alumina layer sputtered on the films. As a result, the W is increased in the (BiFeO₃)_0.6-(SrTiO₃)_0.4 thin films from 13 J/cm³ to 17 J/cm³. The Al₂O₃ or SiO₂ may be a good candidate material as a parasitic dead-layer for the ferroelectric films due to its ultrahigh DBS. However, the polarization of the ferroelectric thin films usually degraded sharply as implanted with the dead-layer mentioned above due to their low dielectric constant (4–6). As a result, little improvement will be obtained for the W. Therefore, a new “dead-layer” will be expected. Besides, to our knowledge, the function of artificial “dead-layers” is more than sharing the voltage applied on the films, and how to design a proper and effective artificial “dead-layer” need further investigation. Nevertheless, this artificial “dead-layers” design could be a universal-simple-effective strategy to significantly enhance the DBS for the most dielectric films.

As one of the most studied dielectric material systems, Ba_1-xSr_xTiO₃ (BST) thin films show a low dielectric loss and strong relaxor behavior with the characteristics of field-induced reversible transitions, low hysteresis, and large polarization, which make them a potential candidate dielectric material in the energy storage fields [20]. However, its inferior DBS (hardly exceed 1 MV/cm) becomes one of main obstacles that hinder its application [5]. Researches have shown that the substitution of Ti⁴⁺ with Zr⁴⁺ in BST can improve the DBS obviously by reducing the dielectric loss or leakage current [[40], [41], [42]]. Ricketts et al. present a high W of 4 J/cm³ under the electric field of 0.5 MV/cm in the Ba_xSr_1-xZr_yTi_1-yO₃ (BSZT) ceramics when x ≤ 0.3, y ~ 0.2 [43]. It suggests that BSZT thin film could be a potential high energy storage density dielectric thin film.

In this paper, Ba_0.3Sr_0.7Zr_0.18Ti_0.82O₃ thin films were fabricated on the Pt coated Si wafers by radio-frequency (RF) magnetron sputtering. In order to obtain higher DBS and W, an artificial “dead-layer” composed of the Ca_0.2Zr_0.8O_1.8 (CSZ) linear dielectric material was inserted between Ba_0.3Sr_0.7Zr_0.18Ti_0.82O₃ and top Au electrode to limit the infusion of space charge from the Au electrode into Ba_0.3Sr_0.7Zr_0.18Ti_0.82O₃ thin film under high electric field. After this interface modulation, the dielectric properties, electric breakdown strength and energy storage characteristics as a function of electric field, temperature and fatigue endurance, were investigated. For a more insightful understanding of this modulation, we conducted measurements and analysis of the leakage current mechanism of Pt/BSZT/Au and Pt/BSZT/CSZ/Au films, respectively.