Electrochemical reliability of plasma-polymerized cyclohexane films deposited on copper in microelectronic devices

Abstract

As the circuit density of electric equipment steadily increases, metallization processes demand new conductors with higher electrical conductivities and interlayer dielectrics with lower dielectric constants, in order to overcome the limitation of conventional aluminum and SiO2 metallization schemes. Recently, copper is in a stage of replacing aluminum as the interconnecting metal in advanced microelectronic devices due to its higher conductivity [1–3]. Plasma-polymerized thin films with low dielectric constant have been developed to lower the signal delay in ultra-large-scaleintegrated circuits (ULSIs) [4, 5]. However, in view of electrochemical properties, copper does not form a self-passivating oxide relative to Al2O3 on aluminum and the resistance to corrosion decreases in humid and oxygen-containing environments [6]. Also, plasmapolymerized thin films as interlayer dielectrics have a high tendency to absorb water in the presence of moisture which leads to an increase in its dielectric constant. Furthermore, absorbed water in the film can act as an electrolyte between two dissimilar metals, forming a galvanic cell in the microelectronic devices and electromigration (EM) [7]. To inhibit effectively the multilevel interconnections from corrosion-related malfunction, it is necessary to examine the corrosion problem resulting from a newly-developed interlayer dielectric in microelectronic devices. In this paper plasma-polymerized cyclohexane films were considered as a possible candidate for a interlayer dielectric for multilevel metallization of ULSI semiconductor devices. It has been known that the dielectric constant of cyclohexane was very low (k < 3) [8]. Plasma polymerization was carried out in a vacuum chamber made of stainless steel. Polymerlike thin films were deposited on disk-shaped copper specimens (99.99% copper) by the plasma-enhanced chemical vapor deposition (PECVD). After cleaning the sample using acetone, isopropanol and distilled water, the substrates were in-situ pre-treated with Ar plasma to give an oxygen-free surface and a buffer layer for enhancing film adhesion. The deposition lasted up to 2 h, depending on the RF power and deposition temperature. The general deposition pressure was 2–4 × 10−1 Torr and the deposition temperature was 298 K and 373 K, respectively. The typical conditions of the PECVD process applied in this study are 30 W and 50 W of RF power, 20 sccm of Ar carrier gas, and 20 sccm of H2 bubbler gas. Cyclohexane monomer was used as an organic precursor. The thickness of the coating was fixed at 1 μm layer. Polarization measurements of the electrodes, coated and uncoated with cyclohexane films, were carried out potentiodynamically in a 3.5 wt% NaCl solution open to air at room temperature. For each materials and electrolyte combination, the corrosion sample was allowed to absorb water in the electrolyte for 3 h. After immersion of the specimen in the solution, the potential of the electrode was swept at a rate of 0.166 mV/s from the initial potential of −250 mV vs. Ecorr to the final potential of 1000 mV vs. Ecorr. For a better understanding of the effect of cyclohexane films on corrosion behavior, impedance measurements for the copper electrodes coated and uncoated with the films were performed at Ecorr after immersion for 3 h in 3.5 wt% NaCl solution open to air. Impedance measurements were performed in the frequency range between 100 kHz and 1 mHz. Sinusoidal voltage of ±10 mV was supplied. All potentials recorded in this paper were referred to the saturated calomel electrode (SCE) and the counter electrode was a high-purity graphite rod. The as-grown plasma-polymerized thin films were ex-situ