The nucleation and growth of copper nanoclusters on silicon surfaces from deoxygenated ultra pure water solutions.

摘要

Due to the recent adoption of copper interconnect technology by the semiconductor industry, there has been great interest in understanding the kinetics and mechanisms of copper metal deposition on silicon wafer surfaces in ultra pure water (UPW) solutions. To study copper deposition mechanisms on silicon surfaces, silicon [100] samples were immersed in deoxygenated and non-deoxygenated UPW solutions contaminated with a copper concentration ranging from 0.01 ppb to 1000 ppb while holding the dipping time constant at 300 seconds. By using synchrotron radiation total reflection x-ray fluorescence (TXRF) and x-ray absorption near edge spectroscopy (XANES) in the TXRF geometry, the surface concentration as well as the chemical state of the deposited copper for ultra-low surface concentrations in the range of 4E9 atoms/cm2 (∼10 -6 ML) were determined. From these measurements, it was seen that metallic Cu is deposited in deoxygenated UPW solutions while a mixture of metallic Cu and Cu oxides are deposited in non-deoxygenated UPW solutions. In addition, the copper fluorescence signal was measured as a function of the angle of incidence of the incoming x-rays to determine whether the deposited copper was atomically dispersed or particle-like in nature. It was revealed that samples prepared in nondeoxygenated UPW had Cu that was atomically dispersed near the silicon surface, while the samples immersed in deoxygenated UPW had Cu particles that rose above the silicon surface. To further explore the growth of Cu particles in deoxygenated UPW solutions, silicon samples were immersed in deoxygenated UPW solutions copper concentrations of 100 ppb with dipping times ranging from 5 to 300 seconds. The size and surface density of the metallic copper nanoparticles deposited in deoxygenated UPW solutions was determined for the whole range of dipping times, by using AFM as well as measuring the fluorescence signal as a function of angle. Mathematical models were developed to describe the particle size and density as a function of time. These models had similar results to classical Ostwald ripening mechanisms, where large particles grow at the expense of smaller less thermodynamically stable clusters.