Electrochemical Characterization of Ruthenium Using Potassium Bromate As Oxidizer for Titania Based CMP Slurry

Introduction: Chemical mechanical planarization (CMP) uses chemical and mechanical action for removing the surface topography. It is widely used in integrated chip fabication. Due to the miniaturization of semiconductor chips, focus on the use of ruthenium (Ru) as barrier layer has increased as it allows direct electrodeposition of copper (Cu). The present work involved study of different parameters such as concentration and pH on Ru removal rate (RR) using a slurry consisting of titania as abrasive and potassium bromate (KBrO 3 ) as oxidizer. Electrochemical characterization of Ru was carried out in KBrO 3 solutions to investigate the dissolution phenomena. The corrosion current ( I corr ) values of Ru were obtained from the extrapolation of Tafel curves. An electrical equivalent circuit (EEC) was proposed to model the electrochemical impedance spectroscopy (EIS) results. Methods: A bench top polisher was used to perform the CMP experiments. A pressure of 3.5 psi was applied to the Ru disc and the platen speed was maintained at 150 rpm for all the CMP runs. The specimen holder speed was kept at 250 rpm. For studying the effect of pH on Ru RR, 1 wt% concentration of oxidizer and abrasive was used. For studies on the effect of concentration of abrasives, 1 wt% concentration of oxidizer was used while maintaining the pH at 2. Similarly, the effect of oxidizer concentation was carried out at pH 2, maintaining abrasive concentration at 1 wt%. Slurry pH was adjusted using HNO 3 or KOH. Electrochemical studies were performed to understand the chemical dissolution of Ru. All the electrochemical runs were conducted using 1 wt% oxidizer concentration. Three electrode configuration with platinum counter and Ag/AgCl reference electrodes were used for the electrochemical studies. A scan rate of 1 mV/s was used for potentiodynamic polarization studies and the EIS runs were performed at OCP between the frequency range 100 kHz and 0.01 Hz. Results: The effect of pH on Ru RR reveals that the Ru RR is maximum at pH 2, decreases at pH 3 and increases till pH 7 and decreases gradually in the alkaline region. The effect of abrasive loading on Ru RR in presence of 1 wt% KBrO 3 at pH 2 shows that the removal rate of Ru increases significantly till abrasive loading of 3 wt% and saturates after that. The effect of oxidizer concentration on Ru RR in presence of 1 wt% titania at pH 2 shows that the Ru RR increases with increase in oxidizer concentration of the slurry due to the formation of oxide layer on the surface which is removed with the help of polishing pad and/or the abrasive particles. The potentiodynamic polarization curves of Ru (as shown in Fig. 1) at various pH values in presence of 1 wt% KBrO 3 were studied. I corr value of ruthenium in KBrO 3 solution at various pH follows a trend similar to the CMP RR. Fig. 1 shows the Nyquist plot of Ru at various pH values in presence of 1 wt% KBrO 3 . The Nyquist plots of Ru exhibits capacitive loop at high frequency with depressed semicircle. Discussion: It has been reported that the isoelectric point of ruthenium oxide ranges from 4 to 6 and titanium (IV) oxide ranges from 5 to 6 [1-3]. Therefore, Ru and titania particles act as positively charged at low pH [4]. KBrO 3 dissociates to form K + and BrO 3 - ions. Bromate ions surround the titania particles. Thus the particle charge is negative in the presence of KBrO 3. Due to the opposite charge between the abrasive particle and the ruthenium surface, there exists a strong electrostatic interaction, which causes higher removal rate at lower pH. References: Jiang, Y. He, Y. Li, J. Luo, Appl . Surf. Sci. , 317 , 332 (2014). Zhang, L. Shi, S. Yuan,Y. Zhao, J. Fang, J. Colloid Interface Sci. 330(1), 113 (2009). N. Victoria, P. P. Sharma, I. I. Suni, S. Ramanathan, Electrochem. Solid-State Lett. 13(11) , H385 (2010). Asthana, A. Kumar, N. B. Dahotre, Materials Processing and Manufacturing Science , First ed. Butterworth-Heinemann, United Kingdom (2006). Figure 1