The Effect of Cu Distribution on Post-Patterning Grain Growth and Reliability of Al-1%Cu Interconnects

  • PDF / 1,953,642 Bytes
  • 6 Pages / 414.72 x 648 pts Page_size
  • 18 Downloads / 225 Views

DOWNLOAD

REPORT


EXPERIMENT Al-lwt.%Cu films were sputter deposited at room temperature using a Varian 3180 sputtering system. A film thickness of 0.75 prm was used, and the initial grain size was on the order of 361

Mat. Res. Soc. Symp. Proc. Vol. 391 0 1995 Materials Research Society

1000A. Three different types of post-deposition annealing sequences were used to vary the film microstructures. One set of samples was annealed at 400'C for 1 hour, air cooled to room temperature, heated at 50*C/min to 280'C, annealed at 280'C for 6 hours, and finally air cooled to room temperature (sample set 1). A second set of samples was annealed at 400'C for 1 hour, cooled to 280°C at 10 °C/min. aged at 280°C for 6 hours and then air cooled to room temperature (sample set 2). A third set of samples was simply aged for 30 min at 200'C (sample set 3). A fourth set of samples was cooled in air from 400'C (sample set 4). Table I summarizes the thermal histories used as well as the resulting film microstructures. TABLE I. Grain sizes and precipitate spacings in continuous films and lines for all annealing schedules

Thermal History

Grain Size

Precipitate Spacing in the Continuous Film

Precipitate Spacing in 1.0 pm Wide Lines

400-25-280 400-280 Aged for 30 min at 200 0C Air Cooled from 400'C

1.6 pm 1.6 pm 0.8 pm

2.6 pm 25. pm 0.1 Pm

6.7 pm 625 pm - 0.1 Pm

1.6 pm

film did not contain precipitates

film did not contain precipitates

Films from sample sets 1 and 2 were subsequently patterned into 1.0 pm wide lines, which were tested at a current density of 2.5x10 6 A/cm2 . The electromigration tests were performed over a range of temperatures from 250'C to 300°C. Films from sample set 3 were also patterned, but linewidths ranged from 0.5 pm to 2.0 pm. These lines were prepared for TEM observation and a hot stage TEM sample holder was used to heat the samples in situ for post-patterning grain growth observation.

RESULTS AND DISCUSSION Control of Continuous Film Microstructure In order to investigate the effect A12Cu precipitate spacing has on reliability, it is necessary to isolate this effect from any grain structure effects. It is therefore necessary to vary the precipitate structure while keeping the grain structure fixed. As shown earlier in Table I, we were able to do this with sample sets 1 and 2. The average grain size of both films is 1.6 pm, while the precipitate spacing is 2.6 pm and 25. pm, respectively. TEM micrographs of each film are shown in Figure 1. The final grain structure of a continuous film is determined by its initial grain size and by the amount of any subsequent grain growth 2. The driving force for grain growth is a strong function of temperature, such that the final grain size distribution is fixed by the highest temperature a film experiences during processing. Since both films in sample sets 1 and 2 had the same starting grain structure, the initial 400°C anneal served the purpose of maintaining the same grain structure characteristics in both films, regardless of subsequent low temperature steps.

362

Controlli