Premium
Accumulating resampling (modulation) in comprehensive two‐dimensional capillary GC (GC×GC)
Author(s) -
Blumberg Leonid M.
Publication year - 2008
Publication title -
journal of separation science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.72
H-Index - 102
eISSN - 1615-9314
pISSN - 1615-9306
DOI - 10.1002/jssc.200800424
Subject(s) - undersampling , resampling , chemistry , chromatography , oversampling , dimension (graph theory) , sampling (signal processing) , gas chromatography , interpolation (computer graphics) , standard deviation , analytical chemistry (journal) , two dimensional gas , mathematics , statistics , physics , detector , optics , combinatorics , motion (physics) , optoelectronics , cmos , classical mechanics , artificial intelligence , computer science
During each sampling period, an accumulating resampler (modulator) in comprehensive 2‐D chromatography accumulates all eluite from the first‐dimension column and reinjects the whole or a portion of the accumulated material into the second‐dimension column. The detrimental effect of the resampling on peak capacity of a 2‐D separation comes from the broadening of the peaks along the first‐dimension due to the resampling itself and due to the subsequent peak reconstruction. Sampling density (ρ S ) of resampling is the number of sampling periods per standard deviation of a peak at the outlet of the first‐dimension column. It is shown that a simple formula describes the peak broadening as a function of ρ S at any (even practically too low or too high) ρ S , for the peaks of any (not necessarily Gaussian) shape, for a wide class of peak reconstruction techniques, and for any 2‐D separation (GC×GC, LC×LC, etc .). In capillary GC×GC, optimal ρ S (ρ S,Opt ) depends on the type of the peak reconstruction and on the degree of the gas decompression along the second‐dimension column. When reconstructing using linear interpolation, ρ S,Opt = 0.7 at large and ρ S,Opt = 0.5 at small gas decompression. The choice of exact optimal conditions is not critical. Thus, two‐fold departure of actual ρ S from ρ S,Opt in either direction (under‐ or oversampling) causes only 10% drop in the net peak capacity of GC×GC. The quantitative effect of a much greater undersampling is also evaluated.