Investigation on the validity of holding voltage in high-voltage devices measured by transmission-line-pulsing (TLP)

Latch-up is one of the most critical issues in high-voltage (HV) ICs due to the high power-supply voltages. Because the breakdown junction of an HV device is easily damaged by the huge power generated from a dc curve tracer, the device immunity against latch-up is often referred to the transmission-line-pulsing (TLP)-measured holding voltage. An n-channel lateral DMOS (LDMOS) was fabricated in a 0.25-μm 18-V bipolar CMOS DMOS process to evaluate the validity of latch-up susceptibility by referring to the holding voltage measured by 100- and 1000-ns TLP systems and curve tracer. Long-pulse TLP measurement reveals the self-heating effect and self-heating speed of the n-channel LDMOS. The self-heating effect results in the TLP system to overestimate the holding voltage of HV n-channel LDMOS. Transient latch-up test is further used to investigate the susceptibility of HV devices to latch-up issue in field applications. As a result, to judge the latch-up susceptibility of HV devices by holding voltage measured from TLP is insufficient.