It is difficult to etch Cu with a conventional plasma etching process because it does not form volatile products at room temperature. Additional energy has to be used to vaporize the Cu compound [1]. A room temperature plasma based Cu etch process was reported by the Kuo group [2-4]. The Cu thin film was converted into CuCl_x or CuBr_x in a RIE reactor, which was subsequently removed by dipping in a diluted HCl solution. This method has been used in preparing ICs and TFT LCDs [5]. A barrier layer is commonly used to enhance Cu adhesion and to prevent diffusion into the adjacent film. A capping layer can prevent Cu from oxidation in air and diffusion to the surrounding dielectric. The lifetime of the Cu line is usually estimated using the electromigration (EM) method [6]. Previously, the TiW capping layer effect on the Cu line broken time was studied [7]. In this study, the line width effect on the Mo capped Cu line is studied. Two types of samples, i.e., Mo (25 nm)/Cu (280 nm) and Mo (25 nm)/Cu (280 nm)/Mo (25 nm), were prepared on the Corning glass substrate. All films were sputter deposited. The Mo layer was etched under the condition of CF₄/O₂ 10/10 sccm, 70 mTorr, and 600W for 2 minutes. The Cu layer was etched by exposing it in the HCl/CF₄ (20/5 sccm) plasma at 70 mTorr, 400 W for 1 minute followed by dissolving the reaction product CuCl_x in a 8:1 diluted HCl solution. The plasma reaction and etching steps were done in the PlasmaTherm 700C system. The EM test was done under the constant current stress at room temperature. Figure 1 shows the relationship between the current density and the line broken time of the Mo capped and uncapped lines. All lines have the same length of 800 µm but different width of 10 µm or 30 µm. First, the breakdown time decreases exponentially with the increase of the current density, which is consistent with the case of EM of aluminum lines [8]. The line broken could be explained by the momentum transfer between the conduction electrons and the metal ions [6] Second, when stressed with the same current density, the 10 µm wide line has a longer lifetime than the 30 µm wide line independent of the existence or absence of the capping layer. Since the wider line has more grain boundary intersection points than the narrow line has, it is easier to form and grow voids in the former than in the latter, which is the cause of the quick line broken of the former [9]. Third, based on the same line width, Cu with the Mo capping layer has a shorter lifetime than that without the capping layer. The high line temperature at the EM condition might cause the diffusion of Cu form the bulk film to the Mo capping layer. This process can facilitate voids formation in the Cu film and therefore, the early broken of the line [7]. Figure 2 shows the relationship between the stress time and the line resistance under various current densities. The resistance of the line increases slightly with the increase of time at the beginning. It increases drastically when approaching the broken point. In addition, the Mo capped line has a larger resistance than the uncapped line has. The formation of the Mo-Cu interface may contribute to the high resistance [10]. More results will be reported on the line broken mechanism. Authors acknowledge the financial support of this work through NSF CMMI project 1633580. [1] P.A. Tamirisa, G. Levitin, N.S. Kulkarni, and D.W. Hess, Microelectronic Engineering, 84, 105-108 (2007). [2] Y. Kuo and S. Lee, Jpn. J. Appl. Phys., 39(3AB), L188-L190 (2000). [3] Y. Kuo and S. Lee, Appl. Phys. Lett., 78(7), 1002-1004 (2001). [4] Y. Kuo and S. Lee, Jpn. J. Appl. Phys., 41(41), 7345-7352 (2014). [5] Y. Kuo, Proc. 16^th Intl. Workshop AMFPD, 211-214 (2009). [6] J. R. Black, IEEE Rel. Phys. Symp., 142-149 (1974). [7] M. Li, J. Q. Su and Y. Kuo, Abst. #120271, 235^th ECS National Meeting, Dallas, TX, May 26-31 (2019). [8] A. Buerke, H. Wendrock and K. Wetzig, Crystal Res. Technol., 35(6-7), 721-730 (2010). [9] M. Li and Y. Kuo, ECS . Trans., 86(8), 41-47 (2018). [10] M. Seiss, T. Mrotzek, N. Dreer, S. Knippscheer and W. Knabl, Mat. Sci. Forum, 825, 297-304 (2015). Figure 1