Diffusion Barrier Integrity

A simple method to directly probe the continuity of barrier layers on open-pored films is illustrated in Figure 1. Positrons are injected into a dielectric film and form Ps, which diffuses rapidly throughout the interconnected pores. Despite many collisions with the barrier, a continuous overlayer should prevent Ps from escaping into vacuum. Hence, detection of Ps in vacuum through its telltale 140 ns vacuum lifetime clearly indicates leakage through pin-holes or discontinuities in the barrier.

Figure 1. Illustration of Ps diffusion barrier testing in blanket (upper) and patterned (lower) samples. The patterned samples have trenches, but they are not filled with Cu.

Barrier continuity: Pinholes/defects in blanket samples allow Ps to escape.
Side wall integrity: Comparison between blanket and patterned/trenched samples can highlight the side wall integrity.
Thermal stability: Heat treatment and repeat test will reveal any structural change.

Barrier continuity

We have studied a series of Ta- and TaN-capped A10C blanket films (with no trenches or vias) deposited on Si substrates. The thickness of the capping layer is 25 nm, 35 nm or 45 nm for each of the two barrier materials. The intensities of Ps in vacuum are plotted in Figure 2.

If both barrier materials at all three thicknesses were completely continuous all six curves would be the same.
The 25 nm curves have far too much Ps has escaped into vacuum: 25 nm barriers on A10C do not form continuous layers,
35 nm and 45 nm Ta- and TaN-capped samples have no escaped Ps in vacuum: these thick barriers are continuous.
The minimum critical thickness must be greater than 25 nm using the current deposition processes to perform effectively as a continuous barriers on A10C.
The results are substantiated by TEM results.

Figure 2.Intensities of Ps vacuum component in Ta and TaN-capped A10C silica films.

Barrier thermal Stability

An effective barrier is certainly required to be continuous at room temperature. Moreover, it must be able to maintain its integrity at high temperatures (at least 350 °C) encountered in the subsequent device processing. TaN barriers with thickness of 35 nm and 45 nm on A10C, continuous at room temperature, were selected for demonstrating PALS capability of thermal stability test. These samples, as well as the oxide-capped control sample, were progressively annealed in vacuum up to 500 °C for 1 hour at each temperature.

Figure 3. The lifetimes of Ps annihilating gin the pores of annealed TaN-capped A10C films.

PALS spectra were collected at room temperature in between each heating cycle.
All the diffusion barriers remain continuous even after annealing to 500 °C with no Ps leakage into vacuum through the barriers.
The Ps lifetime in the mesopores started to drop sharply at 350 °C for 35 nm TaN and 400°C for 45 nm TaN: some contaminants are thermally activated to diffuse into the mesopores.
PALS results are consistent with sputtering depth-profiled SIMS results.

Summary

The study demonstrates that PALS is capable of detecting two failure modes of diffusion barriers on low-k films with open porosity.

It can readily detect pin-holes or discontinuities in the barrier layers signaled by the Ps that escapes into vacuum.
It is sensitive to barrier diffusion into the porous structure, indicated by the reduction of Ps lifetime in mesopores.

References:

J. N. Sun, D. W. Gidley, T. L. Dull, W. E. Frieze, A. F. Yee, E. T. Ryan, S. Lin and J. Wetzel, Probing Diffusion Barrier Integrity on Porous Silica Low-K Thin Films Using Positronium Annihilation Lifetime Spectroscopy, Journal of Applied Physics, 89, 5138 (2001)