The GC Separation Process

There may be moments when you sit in front of a gas chromatograph, waiting for a peak and ask yourself how the peak travels through the system. You may have studied textbooks to answer this question and probably have found some mathematical concepts. If they satisfied you, please stop reading here; you might find the following text too dilettante.

I am not satisfied with mathematical descriptions. They usually start with theoretical plates in the explanation. But when you try to imagine a column as a distillation tower with tens of thousands of evaporation steps, you are quickly frustrated. You should not take theory as literally as that. Since I cannot see the process with my own eyes, I try to imagine what it is like.

In my simple model, every molecule goes through a three-step cycle, thousands of times. Each cycle is different, but averages of these many times are similar to those of other molecules of the same species.

The Three Steps

For the description of Step 1, let us start with a molecule that just evaporated from the column wall into the gas phase. The molecule flies a small fraction of a millimeter until it hits a particle of the carrier gas, changes direction and either picks up or loses energy. It has no eyes and no intention where to go. It flies back and forth, as well as towards the center and towards the wall of the tubing. The carrier gas moves it forwards, but gas flow is not like swimming in a river. More than 99% of the space in the gas phase is empty and does not move. Flow in the gas phase merely means that more of the particles flying past are directed toward the detector, not in the other direction. After a time, which is short in one instance and long in another, our solute molecule hits the stationary phase, where it is likely to remain attached like a fly on a flypaper.

The solute molecule is attracted to the surface of the stationary phase by intermolecular forces, but since it continues moving, it still has some chance to free itself and take off again. If success is not immediate, however, it is grabbed and dives into the liquid (Step 2). It goes up and down, maybe even several times to the support surface. Sooner or later it returns to the surface of the stationary phase film, from where it may dive back into the flood - or take off into the gas phase.

Step 3: At the surface of the liquid, the solute molecule finds itself in a cavity formed by stationary phase that pulls it back into the liquid. The molecule has kinetic (thermal) energy and vibrates: it tries to tear itself free and to escape into the gas phase (to evaporate). Its energy may be high or low, depending on the last collisions with the stationary phase molecules. If its movements are sufficiently violent, the molecule will take off into the gas phase. Otherwise, the liquid takes it again and Step 2 is repeated (maybe many times).

The average kinetic energy depends on temperature: The higher the temperature, the more violently the molecule moves and the more often it is able to take off, i.e. Step 2 is repeated fewer times until evaporation succeeds. Hence, the solute will arrive at the end of the column in a shorter retention time. Increase in temperature accelerates the GC process. Higher temperature also accelerates diffusion (Steps 1 and 2), but it is primarily Step 3 that renders GC so temperature-dependent.

Where does separation take place?

Diffusion in the gas phase (Step 1) does not significantly contribute to separation, since all types of molecules behave similarly. Nor does it occur in the layer of the stationary phase (Step 2). Separation occurs through Step 3, and is related to the probability of take off: Because intermolecular forces differ, different molecule types have different chances to pull themselves free. As an average, one molecule may take off once out of 5 times it is at the surface. If another does once out of 5.1 times only, the two may end up being separated if the process is repeated sufficiently many times.

Selectivity

Selectivity of the stationary phase works through its influence on the probability of take off. For instance, if the stationary phase contains cyano groups and if two solutes differ by adouble bond, the additional interaction with the double bond hinders take off and adds cycles of Step 2 until the molecule is able to make the next jump through the column. According to this model, selectivity of a column is determined by the properties of a thin surface layer of the stationary phase only; the bulk could consist of any other liquid. If, for example, a few nanometers of a polar stationary phase could be deposited onto a normal film of an apolarphase, the column should show high polarity. Maybe one day, this concept can be used to produce good columns with a stationary phase of poor diffusivity.

Column diameter

If you were to re-invent capillary GC, what would you optimize? You recognized that every time a molecule takes off from the stationary phase surface, it contributes to the separation. You will, therefore, try to obtain a maximum number of these events. This has to do with Step 1: jumps between two contacts with the stationary phase should be short in order to make maximum use of the column length available. Your first idea is probably to slow down the carrier gas in order to provide the molecule more time to find its way back to the stationary phase surface.That is correct, but there is a limit to this, because the molecule also moves longitudinally, spreading the band as long as the molecule is in the gas phase. In fact, this is why there is an optimum gas velocity (see van Deemter plots). You can gain more if you reduce the capillary diameter. If the distance to the stationary phase surface is shorter, the molecule will touch the latter more frequently, i.e. there are more contacts with the stationary phase per unit time and there is less time for longitudinal diffusion until the molecule returns to the flypaper. In fact, narrow bore capillary columns have always been more efficient. Why then did the wide or megabore columns become so popular?

Column capacity

Steps 1 and 3 have obvious purposes: transport of and differentiation between substances. What is the usefulness of Step 2? Diffusion in the ocean of stationary liquid is the most time-consuming step of GC, as shown by the following estimation: if a peak is isothermally eluted after 10 min. and the gas hold-up time is 30 seconds, molecules travel 30 seconds in the carrier gas and dive in the stationary phase for 9½ minutes (minus the few seconds they spend altogether at the surface trying to take off). 9½ minutes are spent to periodically give them a chance to perform Step 3. If the film were 10 times thinner, every Step 2 cycle would be 10 times shorter and the same number of opportunities for Step 3 could be obtained in less than 1 min; including the gas hold-up time, the run time would be less than 1½ min. If the film were just thick enough to embed the molecules for take off, GC would be much faster. However, this creates a practical problem:insufficient column capacity. Step 3 assumes that during take off the solute molecule is surrounded by stationary phase only. If solute/solute interaction interfered, the probability for take off would be altered. Since the solute concentrations in the center and at the borders of the solute band differ, take off would occur under inhomogeneous conditions, resulting in the well known broadened and asymmetric overloaded peaks. The ocean of liquid has the effect of diluting the solute. It removes solute molecules from the surface layer. Indeed, if the film is ten times thicker, ten times more can be injected to achieve the same solute concentration in the liquid, holding back the molecule in Step 3.

Retention time and elution temperature

Increase of film thickness proportionally prolongs retention times in isothermal runs because each cycle of Step 2 takes longer. Hence, columns with a ten times thicker film should be used with ten times longer retention times (lower program rates) in order to achieve the same separation processresulting from Step 3. In reality, however, users tend to select conditions resulting in similar retention times and increase elution temperatures by roughly 15 degrees for a factor of two in film thickness. This involves a trade between Steps 2 and 3. Temperature increase slightly accelerates diffusion speeds (reducing the duration of Step 2), but the main effect is an increased probability for take off (Step 3). If diving into the stationary phase takes twice as long, this is compensated by a doubled probability of take off. It is paid for by reduced selectivity: if the probability to evaporate from the stationary phase surface becomes high, small differences in the structure of two solutes have less influence on the retention time and the two peaks will get closer together. In fact, thick film columns provide lower resolution, not because separation efficiency in terms of theoretical plates is lower, but because relative retention times (alpha values) are smaller (J. Chromatogr. 207 (1981) 291).

Conclusions

I find it exciting that a model as simple as that described is capable of correctly describing the principal phenomena observed in GC separation. This helps us to understand our daily observations, making the process taking place in the tiny capillary column behind the oven door much more vivid. The above considerations have not been brought to an end, and I would not be surprised if such simple models would turn out to be fertile ground for developing new insight and techniques.

Originally published in the Restek Advantage 1996, Volume 4