With the emergence of new and sometimes difficult to measure materials being integrated into advanced technology node integrated circuit film stacks, alternate methods for determining thickness and removal rate during Chemical Mechanical Planarization (CMP) may be required. Measurement of Coefficient of Friction (COF) is one such alternate method that can be used to overcome several technical obstacles, such as films with unknown composition, films with high a refractive index, very thinly deposited films, or films that oxidize heavily during CMP. This work will demonstrate how COF was used during CMP slurry development. The inherent difference in frictional force measured between an active film and the sublayer film upon which it is deposited can be exploited as a means to develop a selective CMP slurry. Coefficient of Friction in CMP is primarily determined by process conditions and the lubrication regime. Mathematically, the interaction of these two is defined by following equation: COF=ωμ/P [1] Where ω is the rotational speed, µ is the slurry viscosity, and P is the polish pressure. To achieve high removal rates on a given film, a CMP process must interact with that film in the boundary lubrication regime. The process parameters of downforce, platen speed, head speed, slurry flow and the slurry chemistry interact with the film to create COF values that are usually > 0.25 (1). With these process parameters fixed to generate boundary lubrication during CMP, one can determine the change in COF as the active film is removed and the sublayer is exposed, as shown in Figure 1. From the change in COF, the removal rate of the active film can be determined. Figure 1: COF Comparison between active film and sublayer The capability of determining removal rate on a difficult to measure film by COF can now be used as a method to observe changes in removal rate when altering a Control formulation with a chemical additive. Two different film types with the same sublayer were polished to clear, in order to expose the sublayer, while in-situ coefficient of friction (COF) measurements were taken. The goal was to chemically modify a Control formulation with a near 1 to 1 removal rate selectivity between the two polished films to increase the selectivity to >4 to 1. This work used a COF methodology that consisted of a fixed CMP process that enabled polishing in the boundary lubrication regime to determine the difference in COF observed between Film 1 and Film 2 and their respective sublayers. With this COF baseline established, the removal rate selectivity could be increased, as determined by COF, between Film 1 and Film 2 by altering the composition of the Control formulation with the addition of two chemical additives. Additive A was used to increase the removal rate of Film 1 and Additive B was used to suppress the removal rate of Film 2. Additive A increased Film 1 removal rate by 64%, but was only marginally selective to Film 1 because it also increased Film 2 removal rate by 25%. Mechanistically, Additive A functions by increasing the overall viscosity of the Control formulation. Additive B reduced Film 2 removal rate by 46% and was completely selective to Film 2, as there was no impact on Film 1 removal rate. Mechanistically, Additive B adsorbs to Film 2 by way of hydrophobic-hydrophobic interaction. Overall, the Film 1 : Film 2 selectivity was increased from 1.39 to 4.63 with the incorporation of Additive A and Additive B to the Control Formulation, demonstrating that the inherent difference in COF observed between two different material types can be used to measure and improve the selectivity performance of a CMP slurry on multiple films. This work demonstrated that coefficient of friction data can be implemented as an alternate methodology to determine removal rates of difficult to measure film types and can thus be used as a CMP slurry development tool. Figure 1
Read full abstract