As the generation of memory devices evolve, the successful fabrication of high aspect ratio (HAR) features becomes more and more challenging. Apart from the traditional patterning, deposition and etch related issues, conformal film deposition onto these HAR structures becomes a critical parameter in determining the overall device yields. It is well established now that compared to conventional deposition techniques such as chemical vapor deposition (CVD) and physical vapor deposition (PVD), atomic layer deposition (ALD) offers a pathway to highest conformality and step-coverage. On the other hand, film conformality, especially roughness property has trade-off relationship to step-coverage property. High-pressure process is appropriate to obtain high step-coverage performance in short ALD cycle time process, but roughness value becomes worse due to higher chance of crystalline film growth at high-pressure process condition. High roughness causes local variations in step-coverage and poses problems for deposition of subsequent layers and device performance.In the present study, we report that it is possible to obtain conformal and smooth TiN film by using low/high pressure two-step growth, initially low-pressure TiN growth, and secondly high-pressure TiN growth on the non-pattern and HAR pattern wafers. Eugenus’s versatile 300mm mini-batch system and HAR structures with approximately 120:1 (Dimension of bottom opening, and total height are approximately 17nm and 2μm, respectively) were used in this research, respectively. TiCl4 and NH3 were used as the precursors for TiN film deposition and grown TiN films were analyzed by using ellipsometry, atomic force microscopy (AFM) and transmission electron microscopy (TEM) to study the film properties and step-coverages. In addition, reactor-scale computational fluid dynamics (CFD) simulations were performed and correlated to on-wafer experimental results and utilized as a predictive tool.It was found that TiN film grown at low-pressure condition was more conformal and smoother than that grown at high-pressure condition, although step coverage value grown at low-pressure condition was worse than that grown at high-pressure condition. By optimizing low/high pressure two-step TiN growth condition, it was achieved that not only better roughness than that of pure high-pressure TiN, but also, better step coverage value than that of pure low-pressure TiN as well as than that of pure high-pressure TiN (Figure 1). Figure 1
Read full abstract