Published Chris Cook on September 8, 2020

2 responses to “Retention Time Issues in VUV Analyze Software Applications”

  1. Thanks for this outline !
    I don’t grasp how Chi^2 works exactly, this should result in zero for exact correspondence between experimental and database spectrum ? (apparently not).
    I am more a chromatographer. Could you please consider to update te Kovats calculations in VUV ?
    Based on the VUV database time table I get an iC5 Kovats index of 467.1, with a linear Kovats equation I get 465.7 but the iC5 Kovats is 470 according to Fig7. Facing the problems of gasoline separation on a single column, please use the Wikipedia TPGC definition of the Kovats index for measured and database Kovats index ! That is one problem less, solved forever. I would be happy to join you to keep your chromatography sharp and analyse faster while staying comparable to analytical standards !

  2. Christopher Cook says:

    Hi Walter, thank you for your comment!

    The Chi^2 is a goodness of fit (https://en.wikipedia.org/wiki/Goodness_of_fit) metric calculated as Chi^2 = ∑[((Raw-Fit)^2 x weighting function )/# of data points]. This would result in a zero value for a fit perfectly matching a library spectrum. Looking at the Log Chi^2 makes for easier viewing of the plot. Normal Log Chi^2 values in an analysis are around -5 to -4.5, with increases where there are saturated peaks or bad fits.

    We source our RI values from a variety of places, primarily from the NIST WebBook (https://webbook.nist.gov/chemistry/). In VUV Analyze applications the RI values are not used for identification, but are rather used to select a subset of the library to search for fitting the time slice data. Because of this, and because our applications tolerate such a large RI windows (+/-25 for PIONA and Jet), these small RI value variations do not affect the analyses.

Leave a Reply

Your email address will not be published. Required fields are marked *