Eastern Analytical Symposium (EAS) 2005

Somerset, NJ, USA
November 14 - 17, 2005

ACD/Labs' Talk Schedule

Title:	Automated Chromatographic Peak Matching Using UV-Vis and MS Detection
Authors:	Michael McBrien, Mark Bayliss, Vitaly Lashin, Alexy Galin, and Andrey Bogomolov
Session:	HPLC Applications - Part II
Date:	Tuesday, November 15 @ 4:00 PM
Abstract:	One of the key challenges in efficient chromatographic method development is the tracking of component peaks between runs as experimental conditions are changed. Following peak movement with changes in parameters can prove to be a burden on both instrument and operator time. The advent of hyphenated detection in the chromatographic laboratory has created an opportunity to automate the process of peak tracking through the use of components; spectral signatures. This paper will describe the effective application of chemometric treatment of the hyphenated signals to extract retention times and other peak information for components across two or more sets of chromatographic conditions. The effectiveness of MS versus UV-Vis detection will be discussed. The use of both MS and UV-Vis detectors has been found to be particularly powerful; while mass spectrometry provides sensitivity and selectivity, UV-Vis can provide key information when components do not show signal in MS! Baseline and signal-to-noise ratio effects will be highlighted as they pertain to effectiveness of the matching algorithms.
Title:	Group Baseline Corrections on Spectroscopic and Chromatographic Data
Authors:	Michel Hachey Michael McBrien, Michael Boruta, Andrey Bogomolov, and Alexey Gallin
Session:	PAT: Chemometrics and Modeling with Spectroscopic Techniques
Date:	Wednesday, November 16, 2005 @ 2:20 PM
Abstract:	Baseline corrections are a double-edge sword. On the one hand, baseline slope and offset can be removed, greatly simplifying the analysis. For example, in the case of diode array detectors (DAD) it is difficult to obtain a good spectral match or to extract peak purity information in the presence of the non-routine background resulting from gradient experiments. On the other hand, baseline corrections run the risk of distorting the spectrum and adding unwanted variance. This situation can result in meaningless responses complicating subsequent chemometric analyses. There are many methods for describing a function that is parallel to the slope and curvature in the baseline. In the case of multivariate data, correcting each spectrum individually can result in an increase in rank of the set, i.e., it introduces additional components. In this paper, we will show a group baseline correction, which was implemented in collaboration with Windig and co-workers that is meant to minimize the rank increase in multivariate data. Essentially, this means that the algorithm is designed to remove detrimental background variations while minimizing the risks of introducing unwanted artifacts. Applications to hyphenated chromatography data sets will be given which display better spectral matches and extracted peak purity.
Title:	The Coblentz Society Infrared Spectral Collection - A Resource Renewed
Authors:	John Hellgeth and Michael Boruta
Session:	Applications of Spectral Databases and Searching
Date:	Thursday, November 17, 2005 @ 2:00 PM
Abstract:	Comprised over 9500 spectra and offered selectively in a set of five books, The Coblentz Society Infrared Spectra Collection was recently updated to a digitized and computer searchable format. Although there are some advantages to paging through a book of infrared spectra, use of this digitized spectral database with search software has definite and significant advantages. Search algorithms coupled with an appropriately structured database permits rapid examination of a large and diverse spectral data set and increases the utility of this database. Further, the ability to conduct searches based on queries such as chemical structure, name, or spectral peak table affords additional specificity in search results. This paper presents an overview of the development of the Coblentz Society Infrared Spectral Collection from a paper-based atlas of spectra into a modern digital database available for today's computers.
Title:	Raman Spectral Interpretation and Database
Authors:	Michael Boruta and Eugene Karpushkin
Session:	Applications of Spectral Databases and Searching
Date:	Thursday, November 17, 2005 @ 2:40 PM
Abstract:	With the continued growth of Raman spectroscopy usage and applications, the need for databases and interpretation tools grows with it. As with infrared, the interpretation of Raman spectra can be a difficult task. This is often made more difficult because of the reduced number of resources available to the spectroscopists for interpreting their data. Traditionally, interpretation of spectra relies on experience, text books, reference spectra and any prior knowledge of the chemistry of the sample. Software tools that can assist the spectroscopists in the verification, interpretation or classification of these spectra can be an invaluable and cost-effective asset, both in terms of efficiency and accuracy. This paper will briefly look at the resources available today to the Raman spectroscopists and examine some recent progress in software assisted structure verification and spectral interpretation. We will describe the use of software combined with a knowledgebase of spectra-structure correlations to aid in the verification of suspected materials and the interpretation of unknown materials. The organization of the knowledgebase along with some limitations of the knowledgebase and several examples to explore the knowledgebase will be discussed.
Title:	Using Analytical Data Management Systems to Accelerate Salt and Polymorph Screening
Authors:	Michel Hachey, Michael Boruta, and Karim Kassam
Session:	Pharmaceutical Analysis, Part II
Date:	Thursday, November 17, 2005 @ 3:40 PM
Abstract:	Salt and polymorph screening are critical steps that occur between the late discovery and early development phases. These steps have far-reaching effects on the ultimate efficacy, patent protection, ease of manufacturing, and profitability of the final drug. A delicate balance must be maintained between speed and risks in choosing which form to develop. This decision process surounding preformulation is quite complex, involving large quantities of chemical data from disparate sources that need to be evaluated in a coordinated way across multiple criteria. With the advent of high throughput data acquisition, preformulation scientists are being asked to explore even greater numbers of potentially developable forms than before. However, the data evaluation and final decision must still be made in the same amount of time (or faster). In this work, we discuss the acceleration of the preformulation process through the use of modern software systems. We will demonstrate a data system that reads, processes, stores, retrieves, visualizes, and reports a wide variety of data types XRD, Raman, HPLC, NMR, TGA, DSC, chemical structures, properties, and images in a homogeneous fashion. Time savings between 20% to 80% are obtained at various points in the workflow.