高温多层陶瓷电容器 ( MLCC ) 的材料要求

Recent trends and progresses in power electronics rely on fast, miniaturized, high current and high  temperature switching semiconductors. However, in electronics system design, passive components technologies,  in particular capacitors, limit their full exploitation. The goal of the present paper is, firstly, to analyse the capacitor  technologies, in particular ceramic technologies for the limits set by high current, high voltages, high temperatures,  high frequencies, and by cost items. Secondly, for each of the identified limits, to define the tools and technologies  that are capable to overcome it. The new technologies include a new PLZT based antiferroelectric ceramic material  with high switching field, copper inner electrode for the multilayer construction, firing in reducing atmosphere,  applying a sputter method for the outer connecting electrode combined with silver sintering of a flexible termination  or of a pressfit connector. The third goal is to show how the new technologies are applied such that the cost forecast  in production scale is giving a good economic perspective for future use in power electronic systems. Finally, specific  applications in power electronics are selected and described, where the benefits resulting from the new technologies  are made evident.

The main components of power electronic systems  are the switching semiconductors, the control  electronics, the inductive and capacitive passive  components, sensors, chassis, cooling, control  electronics, and connector elements. All these need  to be robust, highly efficient, and have a good relation  of cost to performance. For optimum and low system  size, low losses, and for lowest possible cost, a  careful balancing of the semiconductor switching  speed, frequency of switching, current density, and  ripple and turn-off overshoot voltage is necessary. The recent trend of the development of switching  semiconductors is towards higher switching  frequencies, steeper switching slopes, higher junction  temperatures, higher current density, and less size  while keeping high robustness. The progresses in the  semiconductor technology, there are interactions with  the supporting passive components, in particular the  DC-link capacitors.

In the present paper these shortcomings of the  classical MLCC with respect to their application in  power electronics will be dealt. At first, the relation  between the MLCC specifications and the basic  material properties for ferroelectric and for  antiferroelectric ceramics are discussed. From the  power electronics requirements, it is shown, that the  ferroelectric MLCC cannot comply with the needs of  high currents at high temperaures simultaneously. A  change in development trend direction is needed, and  it is recommended to develope and improve antiferroelectric dielectrics in the future, which are able to  be processed with base metal electrodes. These  materials can comply with the high current / high  temperature needs, because of their material  properties.

In marketplace, there is common understanding that  MLCC specification is such that the capacitance is  measured at 25°C, 10kHz, Vrms = 1 Volt, no bias  voltage applied and the same holds for the dissipation  factor, and comparisions are made on the basis of this  condition. However, in power electronics, e.g. when  using the capacitor as a DC - link, or a snubber there  is the need for specification at high link - voltage (e.g.  400 Volt DC- link voltage for 650 Volt  semiconductors), at higher ripple voltages, and at  much higher frequencies according the switching  features of the semiconductors.

Fig1

It is important to realize, that all ferroelectrics show a  decrease of dielectric constant with bias voltage, with  no exception. The high K values noted above always  go down to 400 ~ 500, when the voltage has induced  full ferroelectric polarization. The corresponding antiferroelectric behaviour of the electronic polarisation,  is schematically depicted in Fig.1 (b), in contrast to  the ferroelectric case the dielectric constant i slow for  low field, and increases then. Fig. 2 shows examples. The field dependence of K is shown for an  antiferroelectric capacitor (with Cu inner electrodes,  commercially available as « CeraLink » from  EPCOS), in comparision with a commercial BaTiO₃ -  based MLCC. For the latter it is important to  emphasize, that the zero field value of K does not  matter, the high field value of K will go down to 400 to  500 anyway. A further note is, that in (BaTiO₃) - based  MLCCs designed for power electronics use the field  strength at reated voltage reaches values >20  Volt/µm, otherwise the capacitance density will go  down considerably.