Effect of pad groove width on slurry mean residence time and slurry utilization efficiency in CMP Yan Mu a, ⁎, Yun Zhuang a,b , Yasa Sampurno a,b , Xiaomin Wei a , Toranosuke Ashizawa c , Hiroyuki Morishima c , Ara Philipossian a,b a Department of Chemical and Environmental Engineering, University of Arizona, Tucson, AZ 85721, USA b Araca, Inc., Tucson, AZ 85718, USA c Hitachi Chemical Company Limited, Hitachi-shi, lbaraki, Japan abstract article info Article history: Received 27 August 2015 Received in revised form 5 February 2016 Accepted 21 February 2016 Available online 23 February 2016 This paper studies the effect of pad groove width on slurry mean residence time (MRT) in the pad–wafer interface as well as slurry utilization efficiency (η) during chemical mechanical planarization. Three concentrically grooved pads with different groove widths were tested at different polishing pressures to experimentally determine the corresponding MRT using the residence time distribution (RTD) technique. Results showed that MRT and η in- creased significantly when the groove width increased from 300 to 600 μm. On the other hand, when the groove width increased further to 900 μm, MRT continued to increase while η remained constant. Results also indicated that MRT was reduced at a higher polishing pressure while η did not change significantly with pressure for all three pads. © 2016 Elsevier B.V. All rights reserved. Keywords: Mean residence time Slurry utilization efficiency Chemical mechanical planarization 1. Introduction It is well known that the presence of slurry in the pad–wafer inter- face is critical to the chemical mechanical planarization (CMP) process [1–3]. Various factors such as slurry mixing and transport, slurry film thickness and the tribological mechanism in the pad–wafer interface can affect material removal rate and planarization efficiency. Different pad groove designs are used to transport fresh slurry into the pad– wafer interface [4,5]. In addition, pad grooves discharge polish debris, heat and spent slurry from the pad–wafer interface and also prevent wafer hydroplaning [4,5]. The effect of different pad groove designs on coefficient of friction (COF), pad surface temperature, and material re- moval rate for interlayer dielectric (ILD) and copper CMP has been in- vestigated extensively [5–9]. Additionally, Muldowney introduced a 3D fluid flow model for simulating the influence of pad groove pitch, width and depth on slurry flow in the pad–wafer gap which revealed that it took longer to renew the slurry in the pad–wafer gap for a pad with larger groove pitch and wider and deeper groove design [10]. While pad groove width is an important factor that impacts slurry flow during wafer polishing, no experimental study has been performed to illustrate the effect of groove width on slurry mixing and transport in the pad–wafer interface. In previous studies, classical residence time distribution (RTD) tech- nique was used to investigate slurry mean residence time (MRT) in the pad–wafer interface [11,12]. MRT represents the average time it takes for fresh incoming slurry to replace the existing slurry in the region bound between the pad and the wafer. As the used slurry contains polishing by-products and pad conditioning debris that have been shown to cause polishing defects [13], a shorter slurry MRT is preferred to reduce polishing defects and increase process yield. In this study, MRT was obtained for three concentrically grooved pads with different groove widths at different polishing pressures using the RTD technique. Results illustrated how groove width affects slurry mixing and transport in the pad–wafer interface. In addition, slur- ry utilization efficiency was calculated to show that the pad groove width can be optimized to increase slurry utilization and minimize slur- ry usage for CMP processes. 2. Theoretical approach For a typical CMP process, MRT can be extracted from the corre- sponding RTD curve by employing classical reactor design principles to CMP as described by Levenspiel [11]. According to Levenspiel, an imaginary reactor can be assumed to form between the pad–wafer as shown in Fig. 1, with its volume defined as the space bound in that inter- facial region. The slurry may enter or exit the reactor anywhere along its circumference. The slurry remains within the reactor for a finite amount of time, and the average period that the fluid remains in the system can be quantified using the reactor design theory. With silica nano-particles, the extremely low values of Stokes number (calculated to be less than 0.1) suggest that we can assume to have creeping flow in the pad– wafer interface where the abrasive nano-particles present in the slurry ought to follow along the same flow fields as the bulk fluid. Based on Microelectronic Engineering 157 (2016) 60–63 ⁎ Corresponding author. E-mail address: muy@email.arizona.edu (Y. Mu). http://dx.doi.org/10.1016/j.mee.2016.02.035 0167-9317/© 2016 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Microelectronic Engineering journal homepage: www.elsevier.com/locate/mee