Microstructure and Nonohmic Properties of SnO2-Ta2O5-ZnO System Doped with ZrO2

The microstructure and nonohmic properties of SnO2-Ta2O5-ZnO varistor system doped with different amounts of ZrO2 (0–2.0 mol%) were investigated. The proposed samples were sintered at 1400°C for 2 h with conventional ceramic processing method. By X-ray diffraction, SnO2 cassiterite phase was found in all the samples, and no extra phases were identified in the detection limit. The doping of ZrO2 would degrade the densification of the varistor ceramics but inhibit the growth of SnO2 grains. In the designed range, varistors with 1.0 mol% ZrO2 presented the maximum nonlinear exponent of 15.9 and lowest leakage current of 110 μA/cm2, but the varistor voltage increased monotonously with the doping amount of ZrO2.


Introduction
SnO 2 varistors are semiconducting ceramic devices, which possess nonlinear voltage-current characteristics due to their grain boundary effects formed commonly by sintering SnO 2 powder with minor additives (impurity). Due to their excellent energy handling capabilities, they can be applied extensively to protect electronic circuits, various semiconductor devices, and electric power systems from dangerous abnormal transient overload [1,2].
Moreover, during modern ceramics processing, high energy attrition milling and ZrO 2 grinding media were often applied. As a result, Zr 4+ contamination in ceramic samples is a common phenomenon. However, up to now, no literature about the role of Zr 4+ ion (ZrO 2 ) in SnO 2 -based varistors has been reported.
Recently, we optimized a SnO 2 -Ta 2 O 5 -ZnO varistor system, which presents varistors of good nonlinear properties but very low varistor voltage [17]. Based on it, in the present study, SnO 2 -Ta 2 O 5 -ZnO-based varistor system was doped with ZrO 2 (0-2.0 mol%), and the effect of ZrO 2 doping on the microstructure and nonohmic properties of SnO 2 -Ta 2 O 5 based varistors was investigated. To our surprise, varistors with fully dense structure and high breakdown voltage could be obtained. The Scientific World Journal composition of (99.45-) mol% SnO 2 + 0.05 mol% Ta 2 O 5 + 0.5 mol% ZnO + mol% ZrO 2 ( = 0, 0.25, 0.5, 1.0, 2.0). All the oxides were raw powders of analytical grade. At beginning, the raw powders were mixed in deionized water and ball-milled in polyethylene bottle for 24 h with 0.5 wt% of PVA as binder and highly wear-resistant ZrO 2 balls as grinding media. Subsequently, the obtained slurries were dried at 110 ∘ C in an open oven. After drying, the powder chunks were crushed into fine powders, sieved, and pressed into pellets of 6 mm in diameter and 1.5 mm in thickness under a pressure of 40 MPa. Then, the pressed pellets were sintered at 1400 ∘ C for 2 h in a Muffle oven by heating at a rate of 300 ∘ C/h and cooling naturally. To measure the electrical properties, silver electrodes were prepared on both surfaces of the sintered disks by heat treatment at 500 ∘ C for half an hour.

Materials
Characterization. The density of the samples was measured by Archimedes method according to international standard (ISO18754). Their crystalline phases were identified by X-ray diffractometer (XRD, D/ max2550HB+/PC, Cu K , and = 1.5418Å) through a continuous scan mode with speed of 8 ∘ /min. The microstructure was examined on the fresh fracture surfaces of the samples via a scanning electron microscope (SEM, Tescan XM5136). And the average size of SnO 2 grains in the samples was determined using linear intercept method from the SEM images.
A high-voltage source measurement unit (Model: CJ1001) was used to record the characteristics of the applied electrical field versus current density ( -) of the samples. The varistor voltage ( ) was determined at 1 mA/cm 2 and the leakage current ( ) was the current density at 0.75 . Then, the nonlinear coefficient ( ) was obtained by the following equation: where 1 and 2 are the electric fields corresponding to 1 = 1 mA/cm 2 and 2 = 10 mA/cm 2 , respectively. Figure 1 illustrates the XRD patterns of the as-prepared SnO 2 - Ta  No extra phases were identified, possibly because the doping levels of the additives were too low to be detected in XRD limits. And, because of the same ionic valence and almost no radius difference between Sn 4+ (0.071 nm) and Zr 4+ (0.072 nm) ions, the doped ZrO 2 is fully soluble in SnO 2 lattice, which can be seen from almost the same positions of XRD diffraction peaks of the prepared samples as shown in Figure 1(b) in a close view to the patterns in 2 from 50 to 55 ∘ . As for the splitting of the XRD peaks in the figure, it might be due to the peak doublet of K-alpha 1 and K-alpha 2. SEM images of the as-prepared SnO 2 -Ta 2 O 5 -ZnO based varistor ceramics also confirmed the solubility of ZrO 2 into SnO 2 lattice (please see Figure 2). The images reveal that, although doped with different amounts of ZrO 2 , the typical microstructure of the samples almost has no change: almost fully dense structure of SnO 2 grains without any obvious second phases. The detailed microstructural parameters are also summarized in Table 1. With increasing doping amount of ZrO 2 , the density of samples decreases in a very narrow range from 6.93 to 6.80 g/cm 3 partly because the density of ZrO 2 (5.89 g/cm 3 ) is lower than that of the matrix SnO 2 (6.95 g/cm 3 ), but the relative density of the samples also decreases although also in a very narrow range from 99.8% to 98.2%, which indicates a decreased sample densification and could be attributed to the lower diffusion ability of solid ZrO 2 particles in SnO 2 matrix at the designed sintering temperature because the melting point of ZrO 2 (2680 ∘ C) is much higher than that of SnO 2 (1630 ∘ C). Moreover, from these SEM images, it can be clearly seen that, with increasing ZrO 2 contents in the ceramics, the average size of SnO 2 grains decreases, which might be owing to the inhibited transportation of SnO 2 during sintering by the doped ZrO 2 with lower diffusion ability.

Electrical Properties.
The -characteristics of the asprepared SnO 2 -Ta 2 O 5 -ZnO-based ceramic varistors doped with different contents of ZrO 2 are illustrated in Figure 3, and their corresponding detailed electrical parameters calculated from the -curves are listed in Table 1.
The results indicate that, with increasing doping content of ZrO 2 up to 1.0 mol%, the nonlinear coefficient of the samples increased up to 15.9, possibly owing to the increased carrier concentration in the varistors, decreased electrical resistivity of SnO 2 grains and thus enhanced barrier height by doping, and higher number of voltage barriers due to the decrease in grain size, but it would drop down with more ZrO 2 doped, due to the corresponding less dense sample structure, degraded effective grain boundary, destroyed depletion layer structure, and thus decreased barrier height. The leakage current of the samples presented an opposite trend to that of nonlinear coefficient with ZrO 2 doping, and the varistors with 1 mol% ZrO 2 presented the lowest leakage current, 110 A/cm 2 , which is completely consistent with classic theory on their relationship [18]. Thus, it can be concluded that the optimum doping amount of ZrO 2 in the proposed SnO 2 -Ta 2 O 5 -ZnO-based ceramic varistor system was 1 mol%. The varistor voltage of the samples increased monotonously with the doping amount of ZrO 2 , which could be mainly attributed to the decreased SnO 2 grain size, thus increasing the number of grain boundary in unit thickness after doping.   for 2 h with conventional ceramic processing method. The doping of ZrO 2 would degrade the densification of the varistor ceramics, but inhibit the growth of SnO 2 grains. In the designed range, varistors with 1.0 mol% ZrO 2 presented the maximum nonlinear exponent of 15.9 and lowest leakage current of 110 A/cm 2 ; but the varistor voltage increased monotonously with the doping amount of ZrO 2 .