Technical Article

Combining Creepage with Miniaturization in High-Voltage Automotive Drives

May 18, 2016 by Reggie Phillips

This article highlights the concerns for the miniaturization in high voltage automotive drives which is arcing and introduces KEMET's ArchShield MLCCs.

The inverters and charging systems in hybrid vehicles (HV) or fully electric vehicles (EV) provide a typical example of a high-voltage application that faces extreme space constraints. Where multilayer ceramic capacitors (MLCCs) are used as filters across high-voltage lines, the pressure to miniaturize can direct designers to select devices in the smallest available case sizes, such as 0603. A 0603 chip-size device saves 75% of the board space occupied by a 1206-size MLCC, for example. However, these smaller case sizes challenge device manufacturers to maximize capacitance within the reduced package volume, and to ensure reliability.

As far as reliability is concerned, the shorter distance between the device terminals brings a greater risk that creepage – the natural tendency of an electric field to spread out over a dielectric surface – may allow arcing between the terminals (figure 1) when the full working voltage is applied across the device. This can result in failure of the capacitor and may cause thermal damage to other components nearby. Factors such as high atmospheric humidity or contamination on the component surface further increase the likelihood of arcing.

 

Surface arcing between MLCC terminations
Figure 1: Surface arcing between MLCC terminations

 

Analysis of Arcing Phenomenon

When a high-voltage DC bias is applied to a high voltage MLCC, a potential difference is established between the opposing terminations and the opposing electrode structure. Simultaneously, an electric field concentration is localized in the termination area and a respective first counter electrode within the MLCC, as illustrated in figure 2. This difference in potential begins to build along the surface of the chip, ionizing the air above it once the electrical breakdown of air is reached.

 

Electrical conditions around the capacitor surface that can allow arcing to occur
Figure 2: Electrical conditions around the capacitor surface that can allow arcing to occur

 

Once the inception voltage of the ionized air is reached, a conductive path is created allowing the energy in the concentrated electric field of the termination area to discharge. This discharge of energy travels through the air, along with the surface of the capacitor and onto an area of lower potential, rather than through the capacitor. During discharge, there is a visible and audible electric arc across the surface of the chip.

This type of arcing can occur at applied voltages of about 300V or more. For some high-voltage capacitors, this may be lower than the rated voltage of the device.

If the arcing occurs between a termination surface and through the dielectric material of the ceramic body to the first internal counter electrode, this usually causes dielectric breakdown of the capacitor resulting in a short-circuit condition that leads to catastrophic failure.

 

Prevention of Arcing

Capacitor vendors have tried a number of approaches to prevent arcing. One of these is to apply a polymer or glass coating along the surface of the chip to fill any voids and provide a smooth surface that has a naturally lower susceptibility to creepage.

Filling these voids with the insulating material also helps exclude contaminants and improves the dielectric stability across the surface of the chip.  Improving this stability reduces the ionization of the air and increases the inception voltage along the surface, thereby reducing the potential for arcing and improving the voltage performance of the capacitor.

Designers have used surface coatings on PCBs in high-voltage applications for decades. This technology has been proven to increase performance, but its primary disadvantage is the cost of applying the coating. Many designers choose to avoid such cost unless it is deemed absolutely necessary to meet specific electrical safety standards.

Another hazard is that surface coatings can be damaged during handling and assembly processes. A breach in the coating effectively reduces the creepage distance capability along the surface, leaving the capacitor susceptible to contamination and arc-over concerns (figure 3). In addition, when choosing a device that has a pre-applied coating, it is important to ensure that the coating material is compatible with all applicable assembly materials, processes and conditions. Incompatibility could result in damage or premature failure of the surface coating.

There are also concerns with air gaps under mounted components, and voids in and under the epoxy coating. These gaps and voids allow for the same arcing potential as an uncoated device.

 

Imperfections in the coating can leave the device vulnerable to arcing
Figure 3: Imperfections in the coating can leave the device vulnerable to arcing

 

Series Electrode

An alternative technique, illustrated in figure 4, is a “series electrode” construction. The first part of the diagram illustrates how five individual 1000V 1000µF capacitors can be connected in series to form an array that effectively raises the breakdown capability to 5000V, even though the total electric field experienced is the same as that for a single capacitor. One disadvantage, however, is that the total capacitance is reduced to 200µF. The second part of the diagram shows the entire block of capacitors placed into a single monolithic structure with the same characteristics as the five series devices.

 

Top: Five individual capacitors in series. Bottom: Monolithic series-electrode construction raises the breakdown voltage but reduces capacitance
Figure 4: Top: Five individual capacitors in series. Bottom: Monolithic series-electrode construction raises the breakdown voltage but reduces capacitance.

 

KEMET has implemented floating-electrode or serial-capacitor technology in a number of device families covering low-to-mid capacitance values. These devices feature a cascading internal electrode design that effectively forms multiple capacitors in series within the device. While certainly reducing susceptibility to surface arcing, this type of series connection is also highly effective as a flex-crack mitigation technology that reduces the risk of capacitor short-circuit failure. A flex crack cannot cross electrodes originating from both ends of the capacitor. It can only cross electrodes that originate from one end of the capacitor and those floating between the active areas. Even if a crack propagates through one of the active areas, the device may lose capacitance but will not typically fail shortly since there is no conductive pathway between the electrodes connected to the opposed terminations. For this reason, the floating-electrode fails open.

 

ArcShield

An additional internal shield electrode, as shown in figure 5, opposes the effects that can cause surface arcing, without the known disadvantages of a coating or serial-electrode construction. The shield electrodes form a barrier to the terminal-to-terminal arcing seen in standard designs.  In a standard design, the electric field at the surface is very close to the terminal, which reduces the energy barrier for arcing to occur across the surface.  The ArcShield design has a larger energy barrier because of the presence of the shield electrode of similar polarity to the termination.

When a high-voltage bias is applied to an ArcShield MLCC, a potential difference is established between the opposing terminations and the opposing electrode structure, but the electric field concentration is localized in the shield electrodes rather than the termination surface and respective first counter electrode. This minimizes the difference in potential along the surface of the chip and drastically improves the creepage distance capability even in smaller case size devices and when there is high porosity in the dielectric surface.

 

The shield electrode reduces field strength in the region of the capacitor surface and first counter electrode
Figure 5: The shield electrode reduces field strength in the region of the capacitor surface and first counter electrode

 

Review of Shield Effects

A standard overlap X7R MLCC is vulnerable to three basic high-voltage failure mechanisms:

  1. arcing between a terminal and the nearest electrode of opposite polarity
  2. arcing between terminals
  3. internal breakdown

KEMET ArcShield MLCCs address these failure mechanisms by introducing a shield electrode that prevents arcing between terminals and any nearby opposing electrode. The devices also incorporate thicker active areas that effectively increase the breakdown voltage.

 

Voltage breakdown in the air (50pcs), comparing standard 1206 MLCC to ArcShield
Figure 6: Voltage breakdown in the air (50pcs), comparing standard 1206 MLCC to ArcShield

 

Surface arcing can occur at voltages as low as 300V, especially with small case sizes. Applying ArcShield technology to smaller case sizes such as 1206 (figure 6) and 0805 or 0603 (table 1) results in high voltage breakdown and reliable life test performance.

 

Performance data for smaller case size ArcShield MLCC
Table 1: Performance data for smaller case size ArcShield MLCC

 

The results show the capacitors can withstand exposure to voltages much higher than typical hybrid or EV inverter or battery-charging voltages, indicating that X7R high-voltage MLCCs in case sizes as small as 0603 can be used safely.

 

About the Author

Reggie Phillips works as the Senior Global Product Manager of KEMET Electronics Corporation where his responsibility focuses on high voltage, high temperature, and high-frequency products as well as application of specific products. He is also responsible for power electronics applications and new business development, quality, delivery, pricing and profitability. He is particularly skilled in the field of product development, research and development as well as manufacturing. He holds a Bachelor's Degree in Ceramic Engineering from which he earned from Clemson University located in South Carolina, USA.

 

This article originally appeared in the Bodo’s Power Systems magazine.