SOME CHEMICAL AND ELECTROCHEMICAL ASPECTS OF THE CHEMICAL MECHANICAL POLISHING OF COPPER

Open Access
Author:
Al-Hinai, Ashraf Talib
Graduate Program:
Materials Science and Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
August 21, 2002
Committee Members:
  • Barbara Shaw, Committee Member
  • Howard W Pickering, Committee Member
  • Kwadwo Osseo Asare, Committee Chair
  • Tarasankar Debroy, Committee Member
Keywords:
  • hydrogen peroxide
  • bare metal
  • hydroxylamine
  • chemical mechanical polishing
  • cmp
  • bta
  • benzotriazole
  • copper
Abstract:
Since the early 1990s, the growing interest in copper as an interconnect metal has led to a large increase in the number of studies covering various aspects of copper chip processing. One of the important topics studied is the chemical mechanical polishing (CMP) of copper, the main method used for material removal. Several copper CMP observations have gone without adequate explanation despite their importance to the CMP process. This study investigates three of these observations. The first observation concerns the discrepancy between the observed polishing rate and the electrochemically determined corrosion rate. In general, the electrochemically determined corrosion rates are found to be much lower than the polishing rate. In this study, copper CMP is studied as a corrosive wear process, where the enhanced dissolution is due to passive film rupture. The most common CMP slurry ingredients were selected: benzotriazole (BTAH) as a corrosion inhibitor and ferric ion and hydrogen peroxide as oxidizers. The first step was to study the kinetics of the Cu(I)BTA film growth. The rate of growth of this film was monitored as a function of potential, BTAH concentration, and pH. The growth of the film followed a 3D growth model, with progressive nucleation at low pH and instantaneous nucleation at higher pH. By modeling the surface coverage of Cu(I)BTA as a function of time, it was shown, for the first time, that if the Cu(I)BTA film is ruptured during a CMP process, there is not enough time for it to grow back. This leaves the copper metal in the bare metal state exposed to the oxidizing slurry. Using a continuous scratching setup, the electrochemical behavior of bare copper metal was studied. In the presence of BTAH, scratching away the Cu(I)BTA film increased the corrosion current. The results also indicated that, even in the absence of the Cu(I)BTA passive film, BTAH will still inhibit the corrosion of copper. However, the inhibitive effect was more prominent for the cathodic than the anodic reaction, which classifies BTAH -under these conditions- as a cathodic inhibitor. The corrosion current of bare copper metal was determined in ferric ion and in hydrogen peroxide solutions and was found to be close to the CMP rate reported in the literature. This leads to the new conclusion that copper CMP is mainly a corrosive wear process that occurs at the abraded spots on the wafer surface, the high spots. Cu(I)BTA film forms only at the low areas which are not being abraded. This finding calls for the revision of the conventional view that CMP involves a successive sequence of film formation and removal. The second topic studied deals with the dissolution of copper in hydrogen peroxide solutions. The literature reports that the polishing rate of copper increase with hydrogen peroxide concentration then drops at higher concentrations. Copper oxide has been detected on the metal surface although the bulk solution pH and metal ion concentration are too low for precipitation to occur. In this study, the near surface pH of copper dissolving in hydrogen peroxide was found to be ~2.3 pH units higher than the bulk value. At high metal dissolution rates, the surface concentration of the metal ions increases rapidly, exceeding the solubility limit and precipitating an oxide. Aqueous stability diagrams confirmed the possibility of oxide formation under the identified surface conditions. The effect of hydrogen peroxide concentration on oxide formation was modeled. Another effect of pH was investigated, this time for the dissolution of copper in hydroxylamine solutions; the third topic of this study. The literature shows that the polishing rate of copper in hydroxylamine solutions exhibits a peak at pH ~6. Aqueous stability diagrams and electrochemical measurements showed that the increase in the dissolution rate of copper coincided with the regions of stability of the copper hydroxylamine complexes. The peak observed around pH 6 was explained as a result of two competing processes; one is the increase in the dissolution rate of the metal due to the increase in the concentration of the unprotonated ligand, and the other is the precipitation of the oxide.