Colloids and Surfaces A: Physicochem. Eng. Aspects 366 (2010) 68–73 Contents lists available at ScienceDirect Colloids and Surfaces A: Physicochemical and Engineering Aspects journal homepage: www.elsevier.com/locate/colsurfa Role of amines and amino acids in enhancing the removal rates of undoped and P-doped polysilicon films during chemical mechanical polishing P.R. Veera Dandu, B.C. Peethala, Naresh K. Penta, S.V. Babu ∗ Department of Chemical & Biomolecular Engineering, Center for Advanced Materials Processing, Clarkson University, Potsdam, NY 13699, United States article info Article history: Received 15 March 2010 Received in revised form 28 April 2010 Accepted 14 May 2010 Available online 24 May 2010 Keywords: Polysilicon Removal rate Chemical mechanical polishing Potentiodynamic and open circuit potential measurements Amines Doping abstract During chemical mechanical polishing, removal rates of undoped and P-doped polysilicon films as high as 200 and 250 nm/min, respectively, have been achieved using several abrasive-free solutions, each consisting of an amine or amino acid. It was observed that only -amine(s) solutions enhance the removal rates of both undoped and P-doped polysilicon films. Potentiodynamic, zeta potential, contact angle, thermo gravimetric and EDS measurements were performed to examine the role of these -amines in achieving high polysilicon removal rate. Possible removal mechanism of both undoped and P-doped polysilicon film in the presence and absence of the different additives is also proposed. © 2010 Elsevier B.V. All rights reserved. 1. Introduction During the fabrication of microelectromechanical systems (MEMS) and integrated circuit devices, polysilicon has to be pol- ished to planarize the surface in order to avoid the issues associated with depth of focus non-uniformity and photoresist step coverage, etc. [1,2]. Since the polysilicon layers are quite thick, exceeding several micrometers, especially in MEMS fabrication [1–5], they need to be polished at a high rate to maintain throughput. Sev- eral commercial slurries [3–10] containing various combinations of abrasives and additives have been developed to meet these require- ments. In several emerging applications [1,2,10], preferential removal of polysilicon over silicon dioxide and silicon nitride is required. We recently reported several silica- and ceria-based dispersions that enhanced the undoped polysilicon RRs to >500 nm/min in the presence of arginine and lysine HCl [8,9] and guanidine carbon- ate (GC) [10] at pH 9 and 10, while simultaneously suppressing silicon dioxide and silicon nitride RRs to <2 nm/min [8,9]. The sil- icon dioxide and silicon nitride RR behavior in the presence of these additives was discussed in our earlier publication [8], but not the polysilicon removal mechanism. Now, while investigating the polysilicon RR in the presence of large number of different amines ∗ Corresponding author. E-mail address: babu@clarkson.edu (S.V. Babu). and amino acids (Fig. 1), we identified several abrasive-free slurries which polish both undoped and P-doped polysilicon films at >200 and >250 nm/min, respectively, with ∼0 nm/min dissolution rates. The absence of abrasives and dissolution offer a compelling com- bination since they can facilitate planarization with minimal defect formation. Since the P-doped films (prepared with an implantation dose of ∼10 15 ions/cm 2 at 80 keV for ∼15 s) have adequate electrical conductivity, it was possible to perform several potentiodynamic measurements with them. Even though doping specifics like the level, type and depth of doping have a significant effect on polysil- icon RRs [11–13], our focus here is only on the role played by the different additives, solution pH and doping (not its specifics) on the RRs. Since our earlier results [8–10] showed that arginine, lysine and picolinic acid enhanced undoped polysilicon RRs in silica- and ceria-based dispersions at pH 10, we investigated a large number of amines and amino acids as potential additives. The different amines and amino acids investigated here are shown in Fig. 1. Several of these choices are based on the molecular structure of arginine, which can be considered to comprise of glycine and guani- dine, and that of lysine to comprise of glycine and ethyl amine. We used GC since it is several orders of magnitude cheaper than guani- dine. Interestingly, with 1 wt% glycine or GC abrasive-free solutions, the polysilicon (both doped and undoped) RRs were enhanced to >200 nm/min from >70 nm/min obtained with just pH adjusted water, while there was no increase in the case of 1 wt% ethyl amine abrasive-free solution, all at pH 10 (all as shown later in Fig. 2). Since 0927-7757/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2010.05.026