Deconvolution and background subtraction by least-squares fitting with prefiltering of spectra

Journal Article · Thu Dec 01 04:00:00 EST 1977 · Anal. Chem.; (United States)

DOI:https://doi.org/10.1021/ac50022a014· OSTI ID:5220120

Statham, P J

Deconvolution of overlapped peaks in a spectrum is complicated by the presence of a high background component. A method of background subtraction, which involves suppressing the background in both the specimen data and peak models with a digital filter before proceeding to a conventional least-squares fit, is analyzed. The dimensions of a ''top-hat'' filter are found which give a suitable compromise with regard to statistical accuracy and sensitivity to both background curvature and possible errors in the peak models. The major advantages over conventional techniques are that the shape of the background need not be known explicitly, there is no need to find suitable points away from peaks for background scaling, and any background that is approximately linear over the range covered by a single peak will be effectively removed.

• Research Organization: Univ. of California, Berkeley
• OSTI ID: 5220120
• Journal Information: Anal. Chem.; (United States), Journal Name: Anal. Chem.; (United States) Vol. 49:14; ISSN ANCHA
• Country of Publication: United States
• Language: English

Related Subjects

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY
400104* -- Spectral Procedures-- (-1987)
BACKGROUND NOISE
DATA PROCESSING
LEAST SQUARE FIT
MAXIMUM-LIKELIHOOD FIT
NOISE
NUMERICAL SOLUTION
PROCESSING
SPECTRA
SPECTRA UNFOLDING
X-RAY SPECTRA