Microscopic Measurements of Electromigration Damage Using Electrical Measurements
- PDF / 574,718 Bytes
- 12 Pages / 414.72 x 648 pts Page_size
- 41 Downloads / 214 Views
School of Physics and Chemistry, Lancaster University, Lancaster, LA I 4YB, United Kingdom
ABSTRACT
Continuous measurements have been made of the changes in several electrical quantities of Al 4%Cu tracks during the lifetime under electromigration stress. The changes in resistance, second harmonic and the second harmonic delay have been used to determine the changes in the track cross-section, resistivity and temperature. During most of the lifetime the results are consistent with a linear increase in the microscopic damage occurring uniformly throughout the track. During the final phase of the life there is evidence of large void formation with subsequent healing. Considerable information can be obtained about these processes and the extent to which they are reversible. The observed events are compared with the changes which have been predicted by specific mechanisms.
INTRODUCTION
Electromigration, and the related stress migration phenomena can be significant failure mechanisms in the metallisation of electronic devices. These are complex phenomena involving the distribution of alloy species and impurities, the creation and motion of defects and the changes in morphology. The general processes involved may not be easy to discover since each sample may be very individual in the propagation to failure. The general methods of study are time-to-failure accelerated tests, microscopy of various types including chemical and metallurgical aspects, modelling and electrical measurements. With such a complex process a range of well characterised samples are needed and a
combination of several experimental techniques are likely to reveal most information.
We
review here the electrical measurements and show that they can reveal considerable microscopic information. Electrical measurements are appropriate since the samples are used as conductors and electromigration is an electrical phenomenon.
There is probably no such thing as a typical specimen. The stress conditions that are used are high, to give a sample life of about 10 days. Thus the degradation mechanism may not be typical of the processes involved in real use. However the even higher acceleration conditions of wafer level tests do not produce radically different results so that perhaps the general effects seen are not untypical. If we take the change in resistance as a simple indicator of the sample properties, then there are three general time period 111 121. Initially there is a 489
Mat. Res. Soc. Symp. Proc. Vol. 391 @1995 Materials Research Society
fast change which is probably due to annealing effects after the changes brought about by
bringing the sample up to a high ambient temperature and then applying the current with further Joule heating. We will not discuss this here. In the main part of the life, from approximately 5% to 80% of the total life, there is an approximately linear rise in resistance. In this, Phase II, region we assume that the sample structure is slowly evolving under the stress with changes in the structure, impurity distribution
Data Loading...
