387 Effect of Oxygen Expnsurc and Deposition Environment on Thermal Stability of Ta Barriers To Cu Penetration. N.A. Bojarczuk, L.A. Clevenger, K.Holloway, J.M.E. Harper, C. Cabral, R.G. Schad and L.Stolt, I.B.M. T.J. Watson Research Center, Yorktown Heights, N.Y. 10598. TPresent Address: Institute for Microelectronics 16428 Kista, Sweden. Abstract The effect of deposition pressure and controlled oxygen dosing on the diffusion barrier perform- ance of thin film Ta to Cu penetration was investigated. hi-situ resistivity, Auger compositional profiling, scanning electron microscopy and cross-sectional transmission electron microscopy were used to determine the electrical, chemical and structural changes that occur in Cu/Ta bilayers on Si upon heating. A 20 iun Ta barrier allowed the penetration of Cu at temperatures ranging from 320 to 630°C depending on processing conditions. Barrier failure temperature is dependent upon the deposition pressure and oxygen contamination at the Ta/Cu interface. This indicates the im- portance of understanding how deposition conditions affect diffusion barrier performance. In troduction Thin film wiring inctallurgies are an important part of semiconductor technology. In particular, good barrier materials for Cu interconnects are needed. Vaiious possibilities have been proposed with Ta being one of the more promising barrier materials.'-l Hu et al. demonstrated that sput- tered Ta was resistant ýo Cu penetration up to 750 *C. 4 Spi-c et al. have used radiotracer analysis to investigate the diffurivity of Cu through Ta from 400 to 750 °C. 5 Recently Holloway and Fryer have shown that sputtcred deposited Ta and 7a 2 N prevent the diffusion of Cu through the barrier until 600 and 700 °C respectively. 6 In this work, we have chosen to study the effects of two vari- ables onl the resistance of a Ta thin film to 1::1u penetration. These variables are the amount of residual impurities pre,;t-nt during Ta deposition and the presence of oxygen dosing at the Cu/Ta interface. Experimental Procedures Cu/'ra bilayer structures were deposited onto 125mm lI.-cleaned < 100 > p-type 2.0-4.0 le-.cm Si wafers by electron-hcam evaporation in two different systems. The first reaches ultra-high vac- uum (IJIIV) base pre,.sures ranging from 2x111-10 to 5xl0-1 Torr. The second is a high vacuum (HV) system with 5x19.-1 to lxl0" 7 Torr base pressures. In both cases, Van der Ilauw resistivity patterns were created by depositing through . 304 stainless steel mask. These devices allow for 4-point probe resistivity measurements. Twenrty nim Ta films were deposited at a rate of 3 A/see, then 150 nm of Cu (4 A/sec) was deposited vwithout breaking vacuum. The pressure during the deposition of the Ta layer ranged from 5x10 1' to 5x10 I Torr for the UIIV system and 5xl0- 7 to Ixl0- 6 Torr for the IIV7 system. Shutters were used in both systems to close half of the wafers with 1.8xi0 5 l..angmuirs of oxygen at 100 'C, after thie deposition o-f the Ta layer followed by cooling to 30 *C before the depo, ition of the top Cu laycr. This allowed both unclosed and dosed samples to be prepared on the same wafer in the same process cycle. Individual Van der Pauw devices cleaved from the wafers were an•iealed in flowing lie purified over a Ti getter. The films were analyzed by resistivity measurermients, Auger compositi::.nal profiling, scanning electron microscopy (SEM) and cross-sectional transmission electron microscopy (TENI) to determine the temperature and mechanism of Cu penctration through Ta. Results and Disen•ion Figure I is a plot of the resistance versus -annealing temperature of an undosed Van der Pauw device deposited in the UIIV system. The pressure during deposition of the Ta layer was 3xl0 6 - Mat. Res. Soc. Symp. Proc. Vol. 203. 01991 Materials Research Society