In todays integrated circuit (IC) industry, the main cost consideration is the time-to-market and to this end adequate quality control is fundamental during the development phase, allowing design and process engineers to evaluate the functionality of a new device before it reaches the final testing stage.
Integrated Circuit Bonding Pads
IC bonding pads are of paramount importance in serving two functions, the first of which is to provide a defined contact with a test probe in order to test the continuity and basic functionality of a device, the second to provide a platform onto which connecting wires can be attached by thermosonic bonding. The latter is used to establish electrical contacts from a chip to the outer leads via very fine gold wires having a diameter down to 20μm.
In order to acertain whether a bonding pad is satisfactory or not requires a quantitative measurement of its Vickers hardness. The pad itself is produced by sputtering aluminium onto the SiO2 substrate creating a film of approximately 1μm thickness. Insufficient hardness of this film results in deep scrub marks (during probe testing) which then prevent a good bonding between the pad and the gold connecting wire. In addition, if the pad is too soft then substantial debris may be produced when the probe tip comes into contact with it, this being a very important consideration in such a particle sensitive environment.
For this comparative study a selection of four wafers were chosen from different manufacturers and the Vickers hardness of each measured using the Nano Hardness Tester (NHT) from CSM Instruments. Owing to the 1μm thickness of the Aluminium film, all indentations were performed with a maximum depth of 100nm in order not to have any substrate effects.
Figure 1 shows one of the bonding pads after having made a series of eight nanoindentations. The scrub mark is also clearly visible. With the NHT’s accurate positioning stages it was possible to make indentations within specific grains (Figure 3) and thus prevent hardness variations due to grain boundary effects.
Figure 1. SEM micrograph of one of the measured bonding pads showing the scrub mark (right) and a line of nanoindentations (left).
Figure 2. Vickers hardness of bonding pads from four different manufacturers.
The Vickers hardness of each bonding pad type is tabulated in Figure 2, each value being averaged from several measurements. It can be seen that type A has the highest hardness and can therefore be considered as being ‘good’, whereas types B and C have much lower hardness and so are considered as ‘bad’. Type D has an intermediate hardness value.
The NHT is a powerful analytical tool which is ideally suited to industrial quality control applications where short measurement times and automated operation are particularly important. In the case of IC bonding pads it has been shown that such a method is easily able to distinguish between different pad hardnesses and thus allow the choice of manufacturer to be made at an early stage in the development phase.
Localised measurements at specific sites make this instrument an attractive solution to many other IC and other related industrial problems. However, the instrument is also of great benefit in a smaller scale research environment for the development of better materials and processes.
Figure 3. SEM micrograph showing the sputtered aluminium grain structure and the ability of the NHT to make a highly localised indentation within a specified grain.