1/f noise as a diagnostic tool to investigate the quality of isotropic conductive adhesive bonds

Abstract

Reliability assessment of conductive adhesive bonds by thermo-cycling up to 830 cycles is time consuming, and does not give much information about the details of the onset of degradation. There is a need for faster tests giving more details about degradation. In this paper, low frequency noise of such contacts is investigated. 1/f Noise stems from conductance fluctuations. The observed voltage noise is enhanced due to current crowding in the electrical contacts on a microscopic scale. In this research contact bonds were made and compared of isotropic conductive adhesives from three suppliers. The 1/f noise of the contact resistance can be interpreted in terms of a multispot contact behavior. We investigated the relative noise C versus contact resistance R in two ways: (1) after an increasing number of thermo-cycles; (2) after increasing mechanical stress. The results often show an increase in relative noise of three orders of magnitude for poor quality polymer bonds. A maximum increase of one order of magnitude is observed for the best quality conductive adhesive. The contact resistance increases by a factor 1.7 and not more than 1.14 for the poor and best quality bonds, respectively. From the analysis based on a noise model for multispot contact, the onset of delamination can be characterized as a reduction in electrical contact area A/sub e/. The relative noise is proportional to A/sub e//sup -5/2/. The surprising result is that samples submitted to a mechanical stress show pictures similar to thermocycled samples. Thermo-cycling with less than 200 cycles leads to less noise, an increase in electrical contact area, and hence a contact improvement. This behavior is understood. Noise analysis under mechanical stress on nondegraded or slightly cycled bonds is a fast diagnostic tool for reliability characterization. The degree of delamination is expressed quantitatively by the D-factor D=A/sub emax//A/sub emin//spl cong/(C/sub max//C/sub min/)/sup 2/5/.