Electrical diagnostics of passive components failure during reliability testing

While the major reliability concerns in electronics packaging target often new types of semiconductor packages, passive components such as chip resistors (CR) or chip capacitors (CC) can be still responsible for a significant portion of failures that occur in electronic devices. Although passive components show generally good reliability figures, component failures may result from either a very large number of such components on a board or from unfavourable mounting positions they are placed in, e.g. close to large THT-connectors. The main challenge in performing reliability tests on passive components results from their superior robustness in comparison to semiconductor packages. Therefore efficient ways need to be used to test large quantities of such components. Thus, the focus in failure detection in passive components is on electrical diagnostics because of its speed and cost effectiveness.In order to address the specific problems of failure detection using electrical measurements on chip resistors one need to consider not only the problems of low resistance measurements, like lead resistance, thermoelectric voltages, triboelectric voltages, non ohmic contacts and device self-heating, but also the particular issues that come with the two-pole nature of the resistor. While a solder joint on a dummy semiconductor package can be simply connected in a four-wire-"Kelvin"-scheme, any chip resistor has a current-voltage path that connects two solder joints and the resistive element in series. Thus, it is difficult to figure out, which of the three elements contributes to small changes in the measured resistance.Test measurements were carried out using the Keithley Models 6221/2182A SMU/nanovoltmeter combination, the Keysight Model 34980A and the Keithley Model 3700A multifunction switch/measure mainframes with appropriate expansion cards to provide multichannel configurations. Measurements were executed using either the internal DMMs of the two mainframe systems or combining the switch units with external measurement devices such as the Keithley Model DAQ6510 or the Keysight Model 34461A. The DUTs were provided by special test boards that allow not only to electrically connect the chip resistors or chip capacitors to the measurement devices but also provide the possibility to separate individual components during the reliability test for further metallographic analysis. The results of the measurements on larger sets of zero-ohm resistors show that the precision and repeatability of the measurements depends fundamentally on measurement parameters such as the chosen test current as well as the measurement time. The paper will give a comparison of the different test parameters with respect to the achieved precision of the measurements. The limitations of electrical diagnostics for failure detection on passive components will be discussed.