Acoustomigration in SAW Devices

We investigate the stress induced acoustomigration effects in SAW structures under microscopic observation

The experimental study of the stress induced material transport phenomena in finger electrodes of Surface Acoustic Wave (SAW) structures, so called acoustomigration, is of special interest in our group [1]. Using in situ experiments under microscopic observation we can estimate some important effects in microscopic range to describe the acoustomigration mechanism that is theoretically not understood yet.

By applying an optical microscope (Axiotech / ZEISS) as shown in Fig. 1 we can not only compare two different metallizations at the same load conditions using a special power SAW test structure, but also study the microstructural damage over time to calculate power durability and lifetime.

Inside a scanning electron microscope (Ultra Plus 55 / ZEISS or high-pressure SN3500 / Hitachi) the material flux can be studied more in detail using some advantages of the SEM as its higher depth of focus, higher magnifications beyond that of light microscopy, and a simultaneous use of analytical methods like EDX or EBSD. Therefore, SAW test devices were mounted inside the specimen chamber of the SEM and an RF signal was applied to it through a drive unit to initiate a SAW. This type of investigations allow us to answer some important questions in dependence, especially, of the layer system, the power load, and time:

  1. Where are locations in microstructure preferred for either void or hillock formation during acoustomigration damage
  2. How void and hillocks are formed over time
  3. What kind of transport paths occur (grain volume, grain boundaries, interfaces) in correlation to their activation energies
  4. Why does the material migrate.

In last years such in situ experiments were carried out in Al-based and Cu-based finger electrodes [1-7].

The following examples in Fig. 2 and Fig. 3 show two video sequences of such investigations. Here, Fig. 2 demonstrates the different power durability of Al/Ti and Cu fingers loaded under the same conditions with respect to temperature, ambient and RF power load using optical microscopy. Observations using the SEM, as shown in Fig. 3, indicate that voids and hillocks in columnar grown Al fingers are generated at grain boundary triple junctions with high misorientation angle, and the material transport often occurs along high angle grain boundaries >30°.

Fig. 1: Experimental test setup using an optical microscope. Left: electrical measuring circuit; right: arrangement under the optical microscope.
Fig. 1: Experimental test setup using an optical microscope. Left: electrical measuring circuit; right: arrangement under the optical microscope.
Fig. 2: Example of results to compare power durability of Al- and Cu-based finger material during acoustomigration tests.
Fig. 3: In situ SEM experiment to study the correlation of the microstructure and the void and hillock formation during acoustomigration in Al-fingers.
Fig. 2: Example of results to compare power durability of Al- and Cu-based finger material during acoustomigration tests.
Fig. 3: In situ SEM experiment to study the correlation of the microstructure and the void and hillock formation during acoustomigration in Al-fingers.

References

[1] T. Knoth: Diploma thesis,Westsächsische Hochschule (FH) Zwickau(2001).

[2]  H. Schmidt et al. : Proc. of the IEEE Ultrasonics Symp., Atlanta, USA (2001) pp. 97-100.

[3]  S. Menzel et al.: in: Proc. 6th Internat. Workshop on Stress Induced Phenomena in Metallizations, Ithaca, AIP Conf. Proc. 612 (2002) 133-137 (doi:10.1063/1.1469898).

[4] S. Menzel and K. Wetzig, in K. Wetzig and C.M. Schneider (Eds.): Metal Based Thin Films for Electronics. Whiley-VCH, Weinheim (2003), p. 235 (ISBN-13:978-3-527-40650-0).

[5] M. Pekarciková: PhD thesis, TU Dresden (2005).

[6] M. Pekarcikova et al.: Microelectron. Engineering 82 [3-4] (2005) 607-612 (doi:10.1016/j.mee.2005.07.064).

[7] M. Pekarcíkova et al.: IEEE Trans. Ultrason. Ferroelectrics Frequency Control 52 [5] (2005) 911-917 (doi:10.1109/TUFFC.2005.1503977).