none of these sources of extra column dispersion can be completely eliminated but, by careful design, they can be reduced to a level where they no longer significantly impair column performance. Before extra column dispersion can be considered in detail, however, it is necessary to determine the volume of the peaks produced in contemporary high efficiency columns by the normal column dispersive processes. This is necessary in order to place a specific limit on the acceptable level of extra column dispersion. It will be seen that extra column dispersion can be far more serious in liquid chromatography than in gas chromatography and so the two types of column system will be considered separately. Extra column dispersion in LC systems will be considered first

in chromatographic columns was taken by Van Deemter (8) who developed the dispersion equation for a packed GC column. Van Deemter's development did not take into account the compressibility of the mobile phase which was dealt with later by Katz, Ogan and Scott (9). A simple form of this theory will be given that does not accommodate the compressibility of the mobile phase but a more detailed and comprehensive treatment is given in Dispersion in Chromatography Columns and Extra Column Dispersion. Van Deemter et al. assumed that there were four band spreading processes responsible for peak dispersion, namely, multi-path dispersion, longitudinal diffusion, resistance to mass transfer in the mobile phase and resistance to mass transfer in the stationary phase. Van Deemter derived an expression for the variance contribution of each process to the overall variance per unit length of the column. Furthermore, as the individual dispersion processes can be assumed to be

10% to the column variance, is determined by the type of column, dimensions of the column and the chromatographic characteristics of the solute. In practice, the majority of the permitted extra-column dispersion should be allotted to the sample volume, as a large sample volume may be necessary to handle a particular sample type for successful analysis. In any event, the chromatograph should always be designed so that the dispersion from other parts of the system is kept to the absolute minimum. Dispersion in sample valves can be minimized by mechanical design (internal loop valves tend to provide the minimum dispersion). Dispersion from unions can be minimized by using drilled-out unions or low dead volume unions. Stainless steel frits provide very little dispersion and can be employed without great concern for their contribution to the overall dispersion of the system. Connecting tubes are one of the major sources of extra-column dispersion and should be kept as short as possible

conditions are not always achieved and channeling often occurs, under which circumstances lack of radial equilibrium could result in the column efficiency being reduced with consequent loss in resolution. To ensure radial equilibrium, it must either be achieved on injection (using sample distribution device) or by employing narrow bore columns where radial equilibrium is more quickly reached. The latter alternative, however, will depend on the resolution required and the nature of the sample. Dispersion Processes that take Place in an LC Column There are four basic dispersion processes that can occur in a packed column that will account for the final band variance. They are, Multipath dispersion, dispersion from Longitudinal Diffusion, dispersion from the Resistance to Mass Transfer in the Mobile Phase and dispersion from the Resistance to Mass Transfer in the Stationary Phase. All these processes are random and essentially non-interacting and, therefore, provide individual

Permissible Extra Column Dispersion The total variance of an eluted peak ()will be the sum of the variance due to column dispersion () and the variance from extra column dispersion ().                Thus,                    Now, the maximum increase in peak variance from extra column dispersion that can be tolerated while not significantly effecting the resolution is 10 %. This value was suggested by Klinkenberg (3) in 1960 and has been accepted as the criteria for extra column dispersion since that time.               Thus,                              or,                      Substituting for () from equation (2), then for an LC column,                                            (3)         and,                               (4)   Values of  and  are included in table 1. It is seen that the permissible peak widths (taken as (4sv(k'=0))) are very smalland range from about 1.5 ml for the peak from the microbore column to about 4.8 ml

and better performance has been achieved, so has more attention been given to apparatus design as opposed to merely optimizing the distribution system that produces the separation. This also implies that if older chromatographs are employed with modern LC columns (particularly microbore columns) then the expected efficient performance from such columns may not be realized, and, in fact, may not even be attainable. Sources of Extra Column Dispersion There are a number major sources of extra column dispersion in any given chromatographic apparatus, and these are depicted diagramatically in figure 1. Figure 1 Sources of Extra Column Dispersion