Advances in GC Leading to Present-Day Instrumentation

A number of additional advances in GC instrumentation were necessary before the advantages of capillary columns over packed columns could be widely employed  

The actual amount of stationary phase per unit column length on a capillary column is much less than that supported on the particles of the typical packed column. Therefore, the capacity for analytes of the capillary is much less than that of the packed column. When the capacity is exceeded, the analytes will spread out over the front of the column after injection, and the improved resoln-tion of the capillary will be lost. This results in the phenomenon of leading or fronting peak shape 

The typical GC syringe injection volume of 0.5-2 pL of liquid solvent typically expands to 0.5-1.5 mL vapor volume, depending on injector temperature. This is a volume increase on the order of lOOOx. A capillary GC column is operated at a much lower flow rate than a packed column (typically on the order of 1 mL/min instead of 20-40 mL/min). The handling of the volume of vaporized solvent and analyte(s) in capillary GC required a more complex control of gas flows in and out of the injector volume. At the typical capillary column flow rates, it may require a minute or more for aU this vapor volume to be transferred to the column. In order that the analytes (if not the solvent) should not spread out after entering the column, the column oven must be maintained at a temperature much lower than that used to vaporize the sample in the injector space. This requires that a separate injector port volume be independently heated to a different and higher temperature than the initial oven temperature. 

The development of the flexible fused silica capillary GC column described in Section 11.5 resulted in the domination of the GC market by open tubular columns and instruments tailored to their use. Packed columns are relegated to a few special applications. Fused silica capillaries were easy to install. They were more inert (i.e., contained less active sites causing tailing) than glass or metal capillaries or the particle supports used in packed column GC. Even when the resolution of packed columns was sufficient for simple separations, a capillary coated with the same stationary phase could 

If the dimensions of stationary phase coating thickness and diffusion distance to the film from the gas phase have been optimized, for further improvement of GC resolution, it becomes necessary to increase the length of the column. This is seen from the simple relation of Eq. (11.7), which indicates that for an optimal minimized value of H (the height equivalent to a theoretical plate) the number of plates, N, is proportional to the length of the column. From Eqs. (11.8) and (11.9) we note that the resolution is proportional to the square root of N. For columns packed with particles of optimal size, and operated at the optimal linear flow rate at the minimum of the Van Deempter curve (Fig. 11.3), the typical maximum pressure of 100 psi achievable from a regulated gas cylinder requires that most packed columns be less than 4-7 m long. More typically they are only 1 -2 m in length. These considerations hmit the resolution achievable in packed column GC. 

In 1958 Marcel Golay demonstrated that coating the stationary phase on the inner walls of narrow bore tubing of capillary dimension (generally in the range of 

5 mm diameter) could allow columns up to 50 or even 100 m to be operated within these pressure ranges. When the flow rate was optimized to a minimum of the open tubular column relation, similar in form to the Van Deempter equation [Eq. (11.10)], the H values were similar to or even better than those obtainable from the best packed columns, so the much longer open mbular or capillary columns could have much larger values of N, and consequently much greater resolution. Since they have no packing, the H values for capillary columns can be smaller because there is no contribution from a multipath A term as in Eq. (11.10). 

---