No abstract

Split injection problems even at ratios 1 : 100 and 0.06 ml Band broadening effects, leading tail caused by reverse solvent injection effect for early peaks rel. to late solvent peak It is generally observed that minor peaks immediately preceding major peaks exhibit asymmetry in the form of a "leading tail"; this is especially evident in small peaks that precede the solvent peak in splitless injection. This phenomenon can be explained by, and provides further support for, Grob's hypothesis on the profile assumed by the solvent plug during splitless injection [l]. The solvent effect has also been rationalized in terms of the basic parameters of the phase ratio, the distribution coefficient and the partition ratio [2,3].Peaks following the solvent encounter an ever-decreasing phase ratio in the direction of migration, i. e. a range of partition ratios exists across each of these solute bands. Viewed from any point in a solute band, the partition ratio in the immediately preceding portion of the band is larger, and the partition ratio in the immediately following portion of the band is smaller; hence band narrowing occurs. Plate numbers calculated on these peaks would be expected to reflect this phenomenon, and should be a function of the magnitude and duration of this continual increase in the partition ratio; in addition, the range of peaks exhibiting these changes in plate numbers should reflect the thickness and extent of the solvent plug.Peaks preceding the solvent peak (or any other very major peak) encounterjust the oppositesituation.The partition ratioof asolute band immediately preceding any point in that band is smaller, and the partition ratio in that portion immediately following is larger; hence the band broadens atypically in this zone of condensed solvent. Because each portion of such a band travels at afaster rate than that portion of the band immediatelyfollowing, distortion in the form of a "leading tail" results; in extreme cases, complete degradation of peak shape is evident. These compounds have encountered a "reverse solvent effect", and the plate numbers reflected by these peaks would be expected to decrease as the amount of solvent is increased.Grob and Grob [I] mention that a solvent effect can also be demonstrated in split mode, but only if the injected sample is at least 10 pl at a split ratio of 1 : 20. It is our experience that both of these solvent-effect phenomena are readily demonstrable on relatively small peaks that reside within the "solvent envelope" of major sample components, even with an injection volume as small as 0.06 pl and a split ratio of 1 : 100.

No abstract