ESD in this context is a high-voltage transient with fast rise time and fast decay time. Several thousand volts of ESD with a high rise time (dv/dt) could break through the junction layer of protective devices easily and cause damage. ESD surge energy, however, is very small, and it does not require much energy-handling capability from a protective device. Electrical overstress (EOS) is a much slower phenomenon than ESD. Therefore, the following factors should be considered when designing for EOS and ESD protection:

Some design engineers still prefer to use SMD ceramic capacitors for ESD protection because of their simplicity and low cost. Capacitors, however, can barely withstand high-voltage ESD surges. For example, 10 pieces out of 100 were damaged at 5 kV of ESD, and all 100 pieces were damaged at 15 kV (see Figure 1). In addition, the surge energy through a capacitor does not dissipate as heat, but rather filters through a device to the ground plane. This means that the filtered surge current can wander around the circuit via the ground plane unless it is dissipated at the ground dc resistance. Figure 1 compares device survival rates at various ESD voltage levels.

Figure 1. Comparative survival rates at ESD for devices protected by capacitors, zener diodes, and multilayer varistors.

Zener Diodes

Zener diodes are designed for voltage regulation, not for protection against surge impulses. However, these devices are widely used by design engineers worldwide because of their low cost. A zener diode is more effective than a ceramic capacitor because it provides a stronger defense against surge. Zener diodes have a higher clamping ratio (the ratio between impulse clamping voltage and dc breakdown voltage), which makes it difficult to lower the impulse clamping to a level safe enough for the device being protected. However, these devices are too slow to protect against nanosecond ESD events. Some devices, including microprocessor chips, are sensitive to ESD even at 200 V. Heat dissipation of ESD at the p-n junction is slower, which increases the clamping voltage level. For example, 10 pieces using zener diodes were shorted at 10 kV, and all failed at 20 kV (see Figure 1).

Also known as avalanche breakdown diodes, TVS devices have several advantages in ESD suppression—such as lower clamping ratio and stronger resistance to surges—over ceramic capacitors and zener diodes. TVS diodes are also too slow to protect against nanosecond ESD events. The structure and characteristic curves are shown in Figure 2. For ESD protection, a 500-W TVS diode is typically adequate. The wattage rating is based on the maximum clamping voltage and peak surge current at that moment, such as a 500-W, 5-V device with a peak surge current of 52.3 A (10 x 1000 microseconds) and maximum clamping voltage of 9.6 V. The device wattage would be calculated as 52.3 x 9.6 = 502 W. This number does not represent the conventional meaning of wattage (i.e., the energy during a 1-second period). Therefore, the energy of a 500-W TVS diode could be estimated by multiplying the device wattage by surge duration (1000 microseconds in this case). The result is approximately 0.5 J.

Figure 2. TVS diode structure and characteristic curve.

The question that arises is whether the joule rating is necessary for a TVS diode in ESD protection. As described earlier, although ESD has tens of thousands of volts of amplitude, it lasts only several nanoseconds, and the joule rating is almost negligible. Because of its high voltage, ESD is still a great threat even though the energy level is small. Tens of thousands of volts can cause dielectric breakdown of insulation, puncture a wafer junction, or burn off a tiny trace of a microprocessor circuit. Therefore, protection devices should be strong enough to meet high-voltage surges.

When a TVS diode is placed in a high-speed signal line, the capacitance of the diode could upset the line impedance or attenuate the line signal considerably. Connecting a low-capacitance diode forward in series with the device to be protected could reduce the capacitance of the TVS diode (see Figure 3).

Figure 3. Low-capacitance configuration with a TVS diode.

TVS diodes are available in small packages for ESD protection, in both axial leaded and SMD (TO-92, DO-215AA, DO-214AA, etc.). Array packages and hybrid packages with diodes are also available.

Multilayer varistors are relatively new devices for ESD protection. They come in a surface-mount package ranging in size from 0402 to 1206. Single-layer devices are available with the same package sizes. The main body substance is constructed of a ceramic material, which is rugged against ESD surges. These devices will not fail even at the highest ESD voltage level. The multilayer structure consists of very thin layers that provide reasonable mechanical strength (see Figure 4). That means the dc breakdown of this device can go as low as <5 V dc.

Figure 4. Multilayer varistor structure.

Multilayer varistors lower device capacitance by adjusting the electrode sizes while still functioning in a low-voltage circuit protection mode for ESD. The design of these structures enables them to reduce device capacitance and still function as an ESD protector. These devices are surge absorbers with rated voltages from 100 V dc to several hundred volts with capacitance of about 1 pF.

Termination. Although the termination doesn't affect ESD protection, the termination process for these devices is pertinent because it differs greatly from that for the other devices. The termination of these multilayer varistors and surge absorbers requires very sophisticated process control compared with other SMD devices such as capacitors or inductors. Metal oxide varistors contain a bonding material that could react easily with the electroplating solution, forming a conductive layer.

Only a few manufacturers have been able to resolve this problem, with rather sophisticated processes. Other manufacturers simply create the termination by dipping the varistor into silver alloy paste and then drying it. Unfortunately, silver amalgam reacts with the solder paste during the solder process, which makes silver alloy paste termination difficult. The silver content of the terminal is drawn down during the amalgam, so nothing is left to hold down the terminal. This phenomenon is called the tombstone effect.

To minimize this tombstone effect, the maximum flow solder temperature is typically set at lower than 240°C for this type of termination.

These devices are also used for ESD protection in limited parts of the PCB. Two diodes can be forwardly connected in series between the positive and negative power supply lines, with the center point connected to the data and input/output ports where the protection is desired. Schottky diodes provide low clamping voltage. They are typically forward voltage biased.

When using these diodes, it is important to consider that the diode-forward characteristic rating must meet the expected ESD voltage and current ratings. Like zener and TVS diodes, these devices are too slow to protect against nanosecond ESD events. Also, when a fast surge current is conducted through these diodes to the ground plane, radiated magnetic flux could induce noise on nearby circuitry. Schottky diodes are also available in surface-mount package sizes from 0603 to 0805.

All of the devices described in this article are suitable only for small energy surges, not for lightning strikes or heavy-duty inductive surges. Therefore, the location for the protection device—no matter which one is used—should be chosen carefully. The best location is as close as possible to the circuitry to be protected.