Premium
Detection and Diagnosis of Various Contaminants in an Electroplating Bath by a Voltammetric Sensor: a Case Study
Author(s) -
Jaworski Aleksander,
Wikiel Hanna,
Holverson Peter,
Nelson Andrew,
Wikiel Kazimierz
Publication year - 2015
Publication title -
electroanalysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.574
H-Index - 128
eISSN - 1521-4109
pISSN - 1040-0397
DOI - 10.1002/elan.201400390
Subject(s) - mahalanobis distance , computer science , contamination , sliding window protocol , biological system , pattern recognition (psychology) , projection (relational algebra) , artificial intelligence , mathematics , analytical chemistry (journal) , chemistry , algorithm , chromatography , window (computing) , ecology , biology , operating system
Chemical contamination cases encountered during semiconductor manufacturing which were resolved with the automated analytical system are described. The system is capable of collecting on‐line voltammetric data and investigating similarities between measurements at proper conditions and measurements with upset behaviour. The analytical system uses an in‐situ sensor employing voltammetric techniques providing a current response strongly affected by the presence of investigated disturbances (contamination with iron ions and with hydrogen peroxide). Soft Independent Modeling of Class Analogy (SIMCA) is a pattern recognition method which describes each class separately in eigenvector space. In this supervised classification technique the projected new measurements are evaluated aiming to determine whether they belong to certain class of disturbance or not. An assessment of the membership in a class of specific disturbance is a method of diagnosis whether such a disturbance affects the plating solution. The shape differences between deformed and reference set voltammograms are quantified by Mahalanobis Distance (MD)‐SIMCA. In addition to the numerical approach, a graphical projection is utilized to diagnose the root cause of the detected disturbances.