Thin-film reactions of Au with Ti, Zr, V, and Nb
- PDF / 722,609 Bytes
- 7 Pages / 595.44 x 841.68 pts Page_size
- 60 Downloads / 155 Views
I. INTRODUCTION As the dimensions of integrated circuits are further reduced, the material requirements for interconnects and contacts become increasingly stringent. The individual interactions of the metallization components with each other need to be understood to understand the system as a whole. Metal-silicon interactions have been extensively studied.1"3 Two reviews of metal-metal reactions have been published.4'5 Although aluminum is the metal used most frequently for interconnects, Au is considered as a possible candidate due to its low resistivity. We have previously investigated a number of aluminide systems^ 8 and now seek to extend this work to Au/metal reactions. In contrast to Al, which forms compounds with all transition metals, Au forms compounds with only a few. Columns IVA and VA elements do form intermetallic phases with Au. The reactions of Au/Ti 9 ' 10 and Au/ Ta n ~ 13 have been previously investigated. In the work of Tisone and Drobek,9 transmission electron microscopy (TEM) and glancing-angle electron diffraction were used to study thin Au/Ti bilayers. They found that Au2 Ti formed first, between 200° and 300 °C, and when the Au was depleted, AuTi formed between the Ti and Au2 Ti layers at 400°-500 °C. As the reaction continued Au 2 Ti was consumed and AuTi 3 formed. They speculated that Au 4 Ti could have formed initially, but was not detectable. The phase Au 4 Ti was found in thick films (Au 10 kA, Ti 1 kA). They concluded that Ti diffused into Au more rapidly than Au into Ti, with Ti diffusing mainly along the Au grain boundaries. A private communication from R. Sun was cited that found the opposite result for bulk diffusion, i.e., that Au diffuses into Ti faster than Ti into Au. Backscattering and x-ray diffraction (XRD) were used by Poate et al.10 to study the Au/Ti system. They found no evidence for layered growth of the phases; however, a 389 °C anneal produced Au, Au4Ti, and Au2 Ti. No phase diagram has been published for the Au/Ta system, but the crystal structures of three phases 28
J. Mater. Res. 2 (1), Jan/Feb 1987
http://journals.cambridge.org
(AuTa, AuTa 2 , and AuTa 3 ) have been reported.14 Christou and Day11 observed AuTa formation at temperatures above 450 °C using XRD, electrical resistivity, and various microscopy techniques. The growth of the AuTa layer at the Au/Ta interface followed a parabolic rate law. Using similar techniques, Tisone and Lau12 observed AuTa formation above 511 °C for Au/ Ta, and 347 °C for Au//?-Ta. Again, the growth followed a parabolic rate law. More recent work by Liaw et al.13 found AuTa and AuTa 2 after a 400 °C anneal using TEM. All of these papers discussed grain boundary diffusion at temperatures below those where AuTa formed. II. EXPERIMENTAL Thin-film samples of Au/Ti (1350/5650 A), Au/ Ti/Mo/Ti (2200/100/20/700 A), Au/Zr (700/2450 A) and Au/V (1650/7250 A) were evaporated in an oil-free vacuum at pressures of ~ 5 X l O ~ 7 Torr on (100) Si covered with SiO 2 . The Au/Nb (750/2600 A) film was dc sputtered in 2 mTorr of Ar, in a system with
Data Loading...
