Development of Dielectric Materials for Monolithic Ceramic Capacitors for Compact High-Capacitance Devices and Power Electronics

Tomoyuki Nakamura, Takayuki Yao, Jun Ikeda, Harunobu Sano

Original Papers:

T. Nakamura, Takayuki Yao, Jun Ikeda, Noriyuki Kubodera, Hiroshi Takagi IOP Conf. Series: Materials Science and Engineering 18 (2011)
T. Nakamura, H. Sano, T. Konoike, K. Tomono: Key Engineering Materials (Volumes 169-170) Electroceramics in JapanII (1999) pp.19-22
T. Nakamura, H. Sano, T. Konoike and K. Tomono: Japanese Journal of Applied Physics vol.38 (1999) pp.5457-5460 Part1 No.9B September 1999

This paper provides a review of the accomplishments that were evaluated in the process of awarding the technology prize of the Ceramic Society of Japan, and the content of this paper includes some results from several years ago. The names of the representative members who were involved in the research are listed here.

Resistors, capacitors, inductors, and other passive components are essential components in today's cutting-edge semiconductor devices. Among these, monolithic ceramic capacitors (MLCC) are an extremely important component. Without this capacitor, semiconductor devices cannot be expected to operate normally. MLCCs serve an important role by supporting the required power supply to semiconductor devices and removing noise, which otherwise can cause malfunctions and reduced performance. This paper describes the design of ceramic materials for developing MLCCs capable of use in compact high-capacitance applications and power electronics applications.

Compact High-Capacitance Device Applications

The major trend for development of MLCCs with smaller sizes and larger capacities has led to the development of thinner layers of dielectric elements. Today, products are being mass-produced having dielectric elements of less than 1μm that are stacked in several hundred layers. The major issue in the development of thinner layers is the maintaining of insulation performance, and particularly the maintaining of insulation performance over long periods of time under high temperatures and high electric fields (referred to as "reliability" below) is needed.

Figure 1 shows a schematic diagram of an MLCC and an explanatory diagram of the parts of the ceramic structure. To achieve high reliability for MLCCs, it is important to understand the contribution to reliability and the role of each part of the ceramic structure. This will enable further strengthening of the strong parts or improvement of the weak parts.

Fig. 1 MLCC Schematic Diagram and Explanatory Diagram of the Parts of the Ceramic Structure

Figure 2 shows a comparison of the reliability for samples where the number of grain boundaries per element was changed by adjusting the number of grains per element. In this example, the number of grain boundaries refers to the quantity in the element thickness direction. This figure shows that samples with a large number of grain boundaries last a longer time until the insulation resistance is reduced, and thus have a higher reliability. This suggests that the grain boundary has an extremely high contribution to reliability, and it is important to obtain a certain number of grain boundaries per element in order to achieve sCompact High-Capacitance Device Applicationsufficient reliability.

This means that the development of thinner layers for dielectric elements requires smaller grain diameters for BaTiO3, which is the main raw material. However, as the grain diameter of BaTiO3 becomes smaller, the dielectric constant is reduced, and the desired capacitance will not be obtained. We then decided to investigate the reliability of the BaTiO3 grain interior. To do this, a fabrication process was devised so that samples were fabricated that had only a single grain in the thickness direction of the dielectric element. We confirmed that samples without grain boundaries in this element thickness direction had an extremely low reliability.

Therefore, to further improve reliability without relying solely on the grain boundary, we thought that it would be effective to modify the grain interior. In actuality, as shown in Fig. 3, using (Ba,Ca) TiO3 as the main raw material where a portion of the Ba in BaTiO3 was replaced with Ca enabled a significant improvement in reliability. To confirm the effect of modifying the grain interior, as described above, a (Ba,Ca) TiO3 sample without a grain boundary in the element thickness direction was fabricated for evaluation, and it was found that the reliability was improved. This means that, as we had intended, the modified grain interior resulted in higher reliability. Based on the changes in the lattice constant of various samples after the Ca replacement amount was changed and the changes in the activation energy resulting in insulation deterioration, it is surmised that this improvement in reliability by replacing with Ca is due to its suppressing of electromigration of oxygen vacancies, which is a cause of breakdowns, by lattice shrinkage due to the Ca replacement.From this standpoint, we were successful in developing high-reliability materials that support thin layers for dielectric elements.

Fig. 2 Change in IR (Insulation Resistance) Over Time Under High Temperatures and High Electric Fields for MLCCs with Different Numbers of Grain Boundaries

Fig. 2 Change in IR (Insulation Resistance) Over Time Under High Temperatures and High Electric Fields for MLCCs with Different Numbers of Grain Boundaries
Number of grain boundaries: a) 5.4 b) 6.6 c) 7.6 d) 8.1 Testing conditions: 150°C, 10kV/mm Composition: BaTiO3-Dy2O3-MgO-MnO-SiO2

Fig. 3 Change in IR (Insulation Resistance) Over Time Under High Temperature and High Electric Field Conditions for (Ba,Ca) TiO3-Based Material and BaTiO3-Based Material

Fig. 3 Change in IR (Insulation Resistance) Over Time Under High Temperature and High Electric Field Conditions for (Ba,Ca) TiO3-Based Material and BaTiO3-Based Material
Testing conditions: 150°C, 20kV/mm Composition: (Ba1-xCax) TiO3-Dy2O3-MnO-SiO2

Power Electronics Applications

In motor driving inverter applications and other applications, capacitor performance must provide both an effective capacitance when applying a direct current voltage (DC bias) and a large allowable ripple current. However, MLCCs that use BaTiO3 have difficulties providing the right balance between these two requirements. For these types of applications, SrTiO3 conventionally has been used, but with the shift from Pd to Ni for the internal electrode for reducing costs, use of SrTiO3 has run into problems.

Because Ni easily oxidizes in air-fired environments and does not act as a metal, firing must be performed under low oxygen partial pressure conditions. However, SrTiO3 is converted to a semiconductor under these conditions. For these reasons, we developed the material design concept shown in Fig. 4 for BaTiO3-based materials that can be fired even in low oxygen partial pressure conditions. In other words, large amounts of rare earth elements and other additive components were added (Conventionally: 1, 2at% added, This time: approx. 20at% added) to shift the Curie point of the original BaTiO3 to the low-temperature side and use the paraelectric phase, which has a small loss and low dielectric constant at room temperature. In this way, it was thought, both an effective capacitance when applying a DC bias and a large allowable ripple current should be obtained. To realize this concept, the results when using different element types were investigated, and Gd was found to have excellent characteristics for low loss in BaTiO3.

Fig. 4 Design Concept of Dielectric Materials Suitable for Power Electronics Applications: Description based on temperature dependence of dielectric constant

Fig. 4 Design Concept of Dielectric Materials Suitable for Power Electronics Applications: Description based on temperature dependence of dielectric constant
Composition: BaTiO3-Rare earth oxide-MgO-MnO-Li/Al/Ti/Si/O-glass

Compared to the same rare earth elements of Dy and Y, Gd easily forms a solid solution in BaTiO3 , and the Curie point was shifted to the low-temperature side. This was thought to be due to the difference in the ion radii. Because this material type contains large amounts of effective rare earth metals for improving reliability, it results in a ceramic with an extremely high reliability and is suitable for applications exposed to high temperature and high voltage conditions.

For materials containing large amounts of Gd added to BaTiO3 , an examination was made for improvement of the temperature characteristics of the dielectric constant, which was an issue for MLCCs, to find the additive effect of BaZrO3. By adding BaZrO3, we found that the dissolution of Gd in BaTiO3 was no longer uniform, and the BaTiO3 ferroelectric phase region remained so that the temperature characteristics were improved.

This enabled the material with the basic characteristics shown in Table 1 to be obtained. The dielectric constant is close to the conventionally-used SrTiO3-based material (X7R-B) , and the temperature dependence of the dielectric constant is also identical to this material.

Dielectric Ceramics	BTL	X7R-A	X7R-B
Dielectric Constant	300	2800	200
Fired Grain Size (µm)	0.7	0.4	1.0
Capacitance (nF)	10	10	10
DF (%)	0.20	1.50	0.02
log IR (Ω)	11.0	11.0	11.5
TCC (EIA standard)	X7R^*	X7R	X7R
Internal Electrode	Ni	Ni	Pd
Chip Size (mm)	4.5 x 3.2 x 2.0

Table 1 Comparison of Basic Characteristics of Various Materials

BTL: BaTiO3-Based Low-loss Dielectrics (Developed material)
X7R-A: Conventional BaTiO3-Based X7R Dielectrics
X7R-B: SrTiO3-Based Dielectrics
*EIA's X7R specification, which demands that the dielectric constant should not change
by more than+15% or-15% from the 25°C value over the temperature range -55°C to 125°C

As shown in Fig. 5, in terms of loss under high frequency and high voltage conditions, this material has better characteristics than BaTiO3-based material (X7R-A) and identical characteristics to SrTiO3 -based material.

Fig. 5 Comparison of Loss Characteristics of Various Materials (Capacitance: 10nF)

Fig. 5 Comparison of Loss Characteristics of Various Materials (Capacitance: 10nF)
a) Comparison of BTL (developed material) and X7R-A AC: 20 kHz
b) Comparison of BTL and X7R-B AC: 300 kHz

In this way, the realizing of the material design concept resulted in successful development of material suitable for power electronics applications.

Looking Ahead

With BaTiO3 as the base material, we tried to understand the ceramic structure and improve the characteristics by focusing on aspects of the composition. To meet even higher market needs in the future, further innovations will be required. In particular, breakthroughs will be needed not only for aspects of the composition, but also for processes. In addition, to better understand this phenomena, more advanced analysis technology and simulation technology will be required. By building up these types of technology, we expect to develop monolithic ceramic capacitors with even higher performance.

Paper Review