Monitoring the composition of electroless copper baths using quantitative infrared spectroscopy

This paper discusses electroless copper plating baths used in the manufacture of printed circuit boards. The quality, morphology, and thickness of the copper deposit is directly affected by bath composition which changes as a function of time and usage. No technique currently exists for on-line monitoring of these baths. The plating solutions are known to contain many chemical species. These species either participate in the deposition reactions or perform some other function such as bath stabilization. Due to proprietary concerns, though, the exact composition of the baths is not known. However, three chemicals have been identified that are critical to the deposition reactions. They are copper sulfate, sodium formate, and sodium tartrate. Therefore, as a first step toward on-line monitoring of these baths, partial least squares (PLS) multivariate calibration methods were used to analyze infrared attenuated total reflection spectral data taken from model bath solutions containing the above chemicals. The PLS calibrations were performed on chemical component at a time on 40 calibration solutions. It was found that optimum modeling could be achieved effectively by narrowing the spectral range to spectral regions that contained strong correlations associated with each chemical