Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.

  

Model Based Optimization of Analytical Instruments

Back
Abstract:
The development and optimization of sensor systems and analytical methods is best done when the whole system is considered simultaneously. Optimization of the individual parts of the system does not guarantee that the system as a whole will be optimal. The holistic approach discussed here requires understanding in depth the environment of the application, the physics/chemistry governing the response of the analytical system, and the data preprocessing and processing (i.e. chemometrics). This typically requires a (generally statistical) model of the environment (means, distributions and covariances of analytes in the environment that would produce a response on the sensor system, generally these are scenario specific), a model of how the sensor works, including a noise model, and possibly a drift model (due to things like imperfect temperature control, etc.), and of course a method for pretreating the data (methods used to eliminate extraneous variance and/or linearize the response) and ultimately for classification or quantification (pattern recognition or calibration). Once all of the models of the system are complete they can be used for optimizing the system configuration (e.g use pre-concentration or not, which wavelengths to measure, or what resolution), how the response is generated (e.g. integration time, amount of time used to collect sample on a sorbent bed). The models can also be used to generate performance estimates such as Receiver Operator Curves. An example is given of a system used to detect Gagents (aka nerve gas) based on Laser-Photoacoustic Spectroscopy (L-PAS). Barry M. Wise, Neal B. Gallagher, Eigenvector Research, Inc. Richard M. Ozanich and Bret Cannon, Battelle Pacific Northwest National Laboratories

Type of Seminar:
Public Seminar
Speaker:
Dr. Barry Wise
Eigenvector Research, Inc., Wenatchee, WA, USA
Date/Time:
Nov 04, 2008   16:00 /
Location:

ETH Zentrum, Gloriastrasse 35, Building ETZ , Room E6
Contact Person:

Prof. Manfred Morari
File Download:

Request a copy of this publication.
Biographical Sketch:
Barry Wise received B.S. degrees in Chemistry and Chemical Engineering from the University of Washington in 1982. After 3 years with Battelle Pacific Northwest National Laboratories (PNNL), he returned to University of Washington where he received Master of Science (1987) and Doctor of Philosophy (1991) degrees in Chemical Engineering. During that time he became interested in chemometrics, the application of mathematical and statistical methods to problems in chemistry. Barry Wise then returned to PNNL, working in the Environmental and Molecular Science Laboratory, until co-founding Eigenvector Research with Dr. Neal Gallagher in 1995.