【作者】 卢振亚；

【导师】 吴建青；

【作者基本信息】 华南理工大学 ， 材料学， 2012， 博士

【摘要】 本文研究了氧化物压敏陶瓷晶界特性与宏观电性能的关系，取得了如下成果：1、测试分析了电负荷历史对氧化锌压敏陶瓷样品的端电容量的影响，并据此研究了经受电负荷时压敏陶瓷内部晶体缺陷结构的动态变化过程。电负荷对样品端电容量影响很大，测试得到的端电容值会随电负荷大小而变化，端电容量测试值的变化反映了材料内部晶界势垒两侧耗尽层厚度的变化。本文对此提出了两种物理机构加以解释，一是导电电子被界面态俘获和发射（释放），二是耗尽区深施主缺陷的电离和复合（重新俘获电子）。当对压敏陶瓷样品施加弱电测试信号时，一些电子被界面态俘获，而撤去测试信号后，被俘获的电子会逐步通过热发射释放。晶界俘获电子使得界面负电荷积累增加，并且因此会使势垒两侧耗尽层厚度增加，而测试信号过后的电子释放过程则会使界面电荷密度和耗尽层厚度逐渐恢复。当对压敏陶瓷样品施加足够大的脉冲电流时，除了材料晶界界面态发生上述电子俘获现象外，在晶界两侧耗尽层会发生碰撞电离现象，耗尽层的深施主杂质被碰撞电离出电子，使得耗尽层正电荷密度增加、耗尽层厚度减小。脉冲电流试验后，晶界界面态俘获电子的释放和耗尽层电离施主重新俘获电子是对缺陷结构动态变化和耗尽层厚度恢复过程的两个相反作用因素，并且这两个因素作用过程的时间常数不一致，因此能观察到强脉冲电流试验后瞬间端电容量比测试前还高以及恢复过程中的非单调变化现象。2、对比研究了几种测试计算压敏陶瓷晶界肖特基势垒高度的方法，结果认为：基于修正的Mott-Schottky方程的C-V特性测试计算法物理意义清晰合理，但是由于压敏陶瓷的内部微观结构的复杂性及深施主杂质缺陷的影响，采用C-V特性法难以获得真实可信的势垒高度测试计算结果。采用基于Schottky热发射方程的J-E-T法计算晶界势垒高度时可以不必考虑压敏材料的微观结构和缺陷结构，但是测试得到的ln（J0）¹/T曲线通常不是一条直线，选取的温度范围不同，势垒高度测试计算结果不一样。采用基于阿伦尼乌斯方程的低电场的lnρ-1/T特性曲线法估算压敏陶瓷材料的晶界势垒高度方法相对简单，由于所测试材料在低电场下的导电机理单一，即导电电子通过热发射越过晶界势垒，所以通过lnρ-1/T特性曲线计算的激活能可以近似认为是晶界势垒高度。3、与以往一些文献报道的研究结果不同，本工作发现初始氧化锌粉体颗粒大小对氧化锌基压敏陶瓷晶粒尺寸和压敏电压梯度影响有限。分别采用等离子体气相法制备的纳米级氧化锌和间接法生产的亚微米级氧化锌制作了压敏陶瓷样品，与后者相比，采用前者制备的样品烧结活性较好，但在相同烧结温度下，两种样品的陶瓷晶体颗粒大小相差不大。采用等离子气相法纳米氧化锌制作的压敏陶瓷材料样品的大电流非线性特性较好，大电流脉冲残压比较低，过压保护特性较好。通过分析样品的介电特性和晶界势垒高度等参数发现，采用等离子气相法纳米氧化锌制作的样品内部氧化锌晶粒的施主浓度比采用普通间接法氧化锌制作的样品内部氧化锌晶粒的施主浓度高，施主浓度高意味着晶粒电导率大，晶粒电导率大有利于提高压敏材料在大电流区的非线性特性。4、研究了钛酸铜钙CaCu₃Ti₄O₁₂（CCTO）陶瓷的电压电流特性。在制备的无掺杂CCTO陶瓷样品上测试到了异常大的电压非线性系数，此结果与一些在该领域有重大影响的文献报道类似。但是，进一步深入研究发现，所测得的这种异常大的非线性特性来源于测试方法的偏差。与ZnO压敏陶瓷样品相比，CCTO陶瓷样品的热发射电流高出6个数量级，当对样品施加测试电压或电流信号时，由于强烈的电阻负温度系数（NTCR）效应，热发射电流发生反馈放大现象导致测试计算得到的电压非线性系数异常增大，采用适当方法降低或消除这种反馈放大现象，测试得到材料的实际电压非线性系数小得多。测试CCTO陶瓷样品（包括添加各种成分的样品）的电压电流特性时应设法避免热发射电流的反馈放大现象。更多还原

【Abstract】 Relationship between electrical performances and grain boundary characteristics ofoxide varistor ceramics was investigated. The main achievements are as follows:1. The effect of the electrical loading history on the teriminal capacitance of ZnOvaristor samples was tested and analyzed. Dynamic change of the defect structure in the grainboundaries of ZnO varistor materials was investigated. The teriminal capacitance is stronglyinfluenced by the electrical loading. The changing of the measured terminal capacitance is areflection of the changing of the depletion layer thickness in the grain boundaries. Twomechanisms have been proposed. One is the capture and emission of electrons by the interfacestates, and the other is impact ionization and recombination (trapping of electrons) of the deeplevel donor defects in the depletion layer. Some electrons are captured by the interface stateswhen the varistor sample is tested with a electric field, and the captured electrons are emittedafter disconnecting of the electric field. The electron capturing in the interfaces increase thenegative charge and also the depletion layer thickness, and opposite process will happen withthe captured electrons emitting. When the varistor sample is subjected to a strong electricalimpulse, in addition to the electron capturing in the interfaces, impact ionization of deep leveldonor defects will happen in the depletion layer. The ionizing of the deep level donor defectswill increase the positive charge concentration in the depletion layer, the depletion layer willthus decrease. The captured electrons emitting in the interface and the electrons trapping inthe depletion layer are two opposite factors influencing the dynamic change of the defectstructure and the recovering process of the depletion layer thickness after the surge currenttest. The non-monotonic changing process of the terminal capacitance is attributed to thedifferent time constants of the captured electrons emitting in the interfaces and the electronstrapping by the ionized deep level donor defects in the depletion layers.2．Different approaches for estimation of the Schottdy barrier height in the grain boundaries of oxide varistor ceramic materials were investigated. The C-V method (based onthe modified Mott-Schottky equation) is reasonable in physical principle. But it is difficult toobtain a true value of the barrier height because microstructure and defect structure of varistormaterial is much more complicated than a metal-semiconductor contact junction. TheSchottdy barrier’s parameters including the the barrier height can be calculated according tothe J-E-T method (based on the Schottky emission equation) without concern for themicrosturcture or defect structure. But the ln(J0)~1/T plots aren’t usually straight lines. Thecalculted barrier heights are different with different temperature range choosing. The lowelectrical field lnρ-1/T plot method (based on the Arrhenius equation) is easy and feasible toestimated the Schottdy barrier height. The barrier height can be reasonbly approcimated as theactivation energy calculated according to the low electrical field lnρ-1/T plot because theconduction mechanism in low field is solely electrons thermal activation acrossing theSchottdy barriers.3. Different from the results presented in literature, the starting ZnO particle sizes have alimited impact on the grain sizes and the varistor voltage gradients of the ZnO varistorceramics. Varistor samples were fabricated by using the nano-ZnO powder synthesized byplasma vapor-phase reaction (PVPR) process and submicron-ZnO powder manufactured byFrench process, respectively. The sintering activity of the samples prepared with thenano-ZnO powder is higher, but there is only a little difference between the mean grain sizesof the samples sintered at the same temperature. The current–voltage (I–V) characteristics inthe low-current region are similar. However, when subjected to surge currents, the residualvoltage ratio of the samples prepared using the PVPR ZnO powder is much lower.Comparative analysis of the dielectric property and the grain-boundary barrier height revealsthat the donor concen-tration of the ZnO crystal grains in the sample prepared using thePVPR ZnO powder is higher than that prepared using French process ZnO powder. A higher donor concentration gives rise to a higher grain conductivity, a higher nonlinear coefficient inthe high-current region and a better protection level in applications.4. Ceramic samples of CaCu3Ti4O12(CCTO) without additive were prepared and thevoltage–current characteristics were investigated. Extremely strong nonlinear behavior,similar to the reported results in literature [Nature Materials3[11]774–8(2004); Appl. Phys.Lett.,89[19]191907(2006); Appl. Phys. Lett.,89[21]212102(2006); Appl. Phys. Lett.,91[9]091912(2007)] was observed. However, we found that this is a measurement illusion due to athermionic emission current feedback amplification effect. Compared to that of ZnO varistor,the thermionic emission current of CCTO sample is about six orders of magnitude higher andit is feedback amplified due to strong negative temperature coefficient resistance (NTCR)behavior when the CCTO sample is tested with a voltage (or current) source. Eliminating thethermionic emission current feedback amplification effect, the true value of nonlinearcoefficient of the CCTO sample is much lower. The nonlinearity of all variants of CCTO(CCTO with different dopants) should be measured carefully to avoid the thermionic emissioncurrent amplification effect.更多还原