Introduction
A ZnO-based varistor is a ceramic device used to prevent electric surges. The behavior of the current-voltage (I-V) curve in varistors is non-linear and can be explained by the presence of Schottky barriers at the grain boundaries [1-3]. Varistors can work with either alternate or direct current in a wide range of power or voltage values. Additionally, varistors have high capacity to absorb energy, presenting good stability in time. The versatility of varistors enables its use in high voltage applications as well as in microelectronic circuitry.
The conventional fabrication of ZnO-based varistors consists of the mixture of ZnO powders (> 95 wt %) with another oxide powders such as Bi2O3, CoO, MnO, Sb2O3 and Cr2O3, followed by consolidation and conventional sintering [4]. From this process a polycrystalline microstructure results consisting of semiconductor ZnO grains surrounded by a thin intergranular phase [2]. In addition to forming the intergranular layer, the functions of various additives in ZnO-based varistors have been investigated. For example, the densification and grain growth kinetics of ZnO can be controlled by the occurrence of a Bi2O3-rich eutectic liquid at the grain boundaries and grain junctions [5]. The addition of CoO or MnO can prevent Bi2O3 evaporation at the sintering temperature, and the addition of Cr2O3 and Sb2O3 can control ZnO grain growth [2].
The Reaction Bonding Process was first developed in Germany by Claussen et. al. [6-7] with the name of Reaction Bonding Aluminum Oxide (RBAO), and it was then adapted to produce silicon nitride (RBSN: Reaction Bonding Silicon Nitride) and silicon carbide (RBSC: Reaction Bonding Silicon Carbide). Materials produced by reaction bonding (RB) process displays better physical and mechanical characteristics as well as a homogeneous microstructure consisting of very small grains. These are desirable characteristics in the case of ZnO based varistors since the breakdown voltage is proportional to the number of grain boundaries between electrodes and a very well distributed resistive phase in the microstructure is required. Hence, the objectives of this work were (i) to explore the possibility to produce ZnO-based varistors by the Reaction Bonding Zinc Oxide (RBZO) process, (ii) to improve the homogeneity of the oxide mixture by taking advantage of the milling action, (iii) to improve the sintering behavior of the oxide composition since small particles are produced by milling and (iv) to investigate the effect of RB process on the varistor properties such as the Breakdown Voltage and the non-linear coefficient.
Experimental
Powders of Zn, Co, Mn, Cr and Sb, ZnO and Bi2O3 were used as raw material. The powder mixture was milled in a conventional ball mill with ZrO2 media, the rotation speed of the mill was 200 rpm during 16 h of milling. The ball-to-powder volume ratio was of 10:1. With the milled powder mixture, cylindrical samples of 2 cm diameter and 0.2 cm thickness were fabricated by uniaxial pressing at 270 MPa pressure. The pressed samples were sintered in air following a heating cycle that allows the Zn oxidation to produce more ZnO. In order to fabricate devices for electrical tests, circular golden spots (0.196 cm2) were deposited on both sides of the disks through a mask by sputtering. Then, the leads were cold welded using silver paste. A scheme of the experimental procedure is presented in Figure 1. The reason for using aggregate ZnO powder in the starting mixture rather than use pure metallic Zn, is because the original ZnO particles act as seeds or nuclei in order to obtain a complete oxidation of pure Zn, through RB.
Figure 1. Scheme of the experimental procedure.
The phases and microstructure of the sintered samples was analyzed by X-ray diffraction (XRD, Siemens D-5000, Germany) and scanning electron microscopy (SEM, Jeol-6300, Japan) respectively, with samples polished with standard ceramographic techniques. Reaction behavior was investigated by differential thermal analysis (DTA, Setaram TG/DTA92, France) and thermogravimetric analysis (TGA, Setaram TG/DTA92, France) following the same sintering cycle of the samples.
Results and Discussion
Thermal Analysis
The typical weight change (thermogravimetric analysis, TG) observed in samples pressed at 270 MPa is presented in Figure 2 as a function of the temperature. In the range from room temperature to 330°C, evaporation of either organic species or water adsorbed during milling leads to a weight loss. At higher temperatures there is a weight gain from 330 to 470°C which corresponds to the Zn oxidation. It is worth noting that milled powder oxidizes faster when Zn is still in the solid phase (melting point of Zn is at 419°C) [6]. Therefore, in this process (RBZO) a slow heating rate before melting temperature has to be used in order to achieve the best characteristics of the new ZnO grains. The differential thermal analysis (DTA) curve presented in Figure 2, shows there is a small increment in the heat flow between 300 and 470°C; this exothermic behavior has been associated with the oxidation of Zn. At 419°C an endothermic peak corresponding to the melting temperature can be observed.
Figure 2. Curves of the differential thermal and thermogravimetric analysis as a function of the temperature.
X- Ray Diffraction Analysis (XRD)
In the XRD patterns corresponding to the unmilled and milled sample powders only the peaks of Zn and ZnO are clearly observed (Figure 3) suggesting that no contamination or other chemical reaction has occurred during milling. On the other hand, XRD peaks associated with the additives such as; Cr, Mn, Co, Sb, or Bi2O3 are not observed, probably because the amount is very small and they could not be detected. Thus we do not have any experimental evidence indicating that the starting metal exists as oxides. Nevertheless, it is most likely that the metals were oxidized under the experimental condition of this study, in which specimens were heated in air with slow heating rate so as to oxidize Zn completely, because all the metals we used have negative values of the standard free energy of formation as shown in Table 1.
The XRD pattern of sintered sample at 1200°C for 1 h is shown in Figure 3c and large peaks of ZnO can be seen, whereas the Zn peaks have disappeared suggesting that oxidation of Zn was complete during the heating. Species related to the additives do not appear due to their low content. However, the free energy formation of Cr, Mn, Co and Sb are more negative than that for the Zinc, so it is considered that the additives were also oxidized during the heating cycle [8].
Table 1. Free energy of formation of the oxides [8].
|
|
|
-287086 |
|
-190585 |
|
-347969 |
|
-969119 |
2[Sb]+ 3/2O2 → [Sb2O3] |
- 264126 |
Figure 3. X-ray diffraction patterns of a) unmilled powders, b) powders milled during 16 hrs and c) sintered sample after the RBZO cycle.
Microstructure
Figure 4 shows a typical microstructure after RBZO cycle. The microstructure is homogeneous and it is formed by small ZnO grains with angular shapes typical of this type of ceramics. The material shows some porosity. Some new ZnO grains produced during the RBZO process are observed. These new ZnO grains are localized principally at intergranular zones, however, also are observed some ZnO grains at intragranular positions. The presence of particles other than ZnO could not be observed.
Figure 4. Microstructure of the ZnO-based varistor after the reaction bonding cycle.
Current-Voltage (I-V) Curve
It is observed in Figure 5 that the threshold voltage of the ZnO-based varistors fabricated by the RBZO is very high. This effect has been attributed to the small grains as well as a highly resistive character of the grain boundaries. The presence of very fine ZnO grains is a result of the RBZO process that lets the formation of new ZnO at grain boundaries preventing the agglomeration of the old ZnO grains into large particles. The insulating phases homogeneously distributed throughout the microstructure build the varistor effect observed in Figure 5.
Figure 5. Voltage versus current for ZnO-based varistors. Sintered at 1200°C for 1h.
Conclusions
Varistor devices were fabricated using a novel method based on the reaction bonding process. Slow heating rates were used during the oxidizing process at temperatures lower than the melting temperature of Zinc. The sintered pieces showed fine grain microstructure formed by both original ZnO particles, added as raw material, and new ZnO grains produced during the oxidizing process. The varistor could operate at high voltages and the I-V curve displayed the non-linear behavior typical in varistors which suggests that the insulating phases (Bi rich) are homogenously distributed in the microstructure.
References
1. F. Gruter and G. Blatter, “Electrical Properties of Grain Boundaries in Polycrystalline Compound Semiconductors”, Semiconductor Science and Technology, 5, 110-113 (1990).
2. M. Matsuoka, “Nonohmic Propierties of Zinc Oxide”, Jpn. J. Appl. Phys., 10, 736-741 (1971).
3. L. M. Levinson and H. R. Philipp, “Zinc Oxide Varistors – A review”, American Ceramic Society Bulletin, 65, 639-641 (1986).
4. J. P. Caffin. “Zinc Oxide-Based Composition for Low and Medium Voltage Varistors”, US patent No. 5, 143-146 (1992).
5. M. Tao, B. Ai, O. Dorlanne and A. Loubiere, “Different Single Grain Junctions within a ZnO Varistor”, J. Appl. Phys., 5, 1562-1565 (1987).
6. N. Claussen, N. Travitzky and S. Wu, “Tailoring of Reaction Bonded Al2O3 (BBAO) Ceramics”, Ceram. Eng. Sci. Proc., 11, 806-820. (1990)
7. N. Claussen, “Processing Reaction Mechanisms and Properties of Oxidation-Formed Al2O3-Matrix Composites”, Journal de Physique IV, 1327-1330 (1993).
8. J. F. Shackelford and W. Alexander, “Materials Science and Engineering Handbook”, CRS Press, Boca Raton Florida (2001).
Contact Details |