Continuing our discussion of the three configurations of capillary column backflush, we cover the coated pre-column configuration.
The use of pre-columns dates back to the early years of capillary chromatography. The primary purpose of using pre-columns was to protect the analytical column; it is cheaper to replace a pre-column than a complete analytical column, and retention times don’t change as much as when the analytical column is repeatedly trimmed. Press fit glass connectors or straight unions were typically used to connect a pre-column directly to a capillary column in these scenarios. By using a purged device between the pre-column and the analytical column instead of the usual straight connector, one can now backflush the pre-column, even while analysis is continuing in the second column. Figure 1 illustrates a typical pre-column backflush configuration.
One can use either an uncoated (but deactivated) pre-column, or a pre-column coated with stationary phase. The uncoated pre-column configuration will be covered next month. There are several advantages of using coated pre-columns over uncoated pre-columns.
Advantages of the coated pre-column configuration
- The coated pre-column backflush configuration is a more inert approach than the uncoated pre-column configuration that will be discussed next month. So, it is a more appropriate choice when analysing “active” compounds that might be lost on uncoated tubing.
- One has the ability to tune the onset of backflush fairly well between closely eluting components, although one can never do this as well as with the post-column configuration discussed last month.
- Backflushing can happen concurrent with the analysis. This means that backflush of unwanted components is initiated sometime during the analytical run; such that unwanted components are backflushing from the pre-column while analysis is proceeding in the second column. The shorter the pre-column, the sooner the last peak of interest has progressed passed the union, and the sooner backflush of the pre-column can commence.
- During concurrent backflush, the flow rate through the second column is a normal flow (as opposed to the elevated flow rates experienced with post-column backflush), so one can use flow limited detectors such as diffusion pump mass spectrometers.
- Coated pre-column configurations are simpler to implement than uncoated pre-columns. One can cut sections of existing columns for use as pre-columns or perhaps use shorter lengths of commercially available columns such that the total length equals the original column (e.g., use a 10 m length as the pre-column and a 20 m length as the analytical column instead of the original 30 m column)
- One can do multidimensional separations by using a pre-column of different stationary phase from the second column.
Disadvantages of coated pre-column configuration
- Of the three configurations, backflush timing is the most difficult to predict with the coated pre-column configuration. Theory has not been published that accurately describes the position of a peak in a capillary column during a temperature program.
- Coated pre-columns are more costly to replace than uncoated pre-columns
- Backflushing occurs later into the run than with uncoated columns, so there is less time for concurrent backflush
- It takes longer to backflush coated pre-columns than uncoated pre-columns. This is especially true if using longer pre-columns such as commercially available lengths (> 10 m).
- Contaminants have more influence on retention times of subsequent runs than with uncoated pre-columns (same issue as post-column configuration). Coated columns tend to focus samples into tighter zones just after the inlet than uncoated columns. So contaminants that cannot be backflushed are more concentrated at the head of the column and subsequent injections are in turn more tightly associated with the contaminants than when uncoated pre-columns are used.
- It is difficult to port retention time locked methods to coated pre-column configurations because both columns are operating under different flow rates.
Pre-column length
In the most typical embodiment, one would simply select a pre-column that is the same diameter, stationary phase type and film thickness as the original analytical column. I recommend a pre-column length that is 10–20% of the length of the original column. What I do is simply cut a 10% piece off of the original column and insert a purge tee between the two. For future replacement of a spent pre-column, I buy a second column and cut the same lengths off as needed (yielding 5 to 10 pre-columns).
Figure 1: A typical coated pre-column backflush configuration. A simple inert purged tee with programmable pressure controller is placed between the pre-column and the analytical column. Once the last peak of interest has passed the purged union, inlet pressure is decreased so that flow through the pre-column reverses and unwanted sample components are backflushed.
Although it is a tempting (and feasible) to simply cut a column in half or use two commercial columns of half the length of the original, there are several valid reasons to keep the pre-column short. First, the closer the length of the pre-column to the analytical column, the less suitable it is for concurrent backflush. In compressed systems (where the column inlet pressure is ≥ 1 atm), linear velocity in the first column is much slower than in the second, so by the time a peak passes the junction, it is barely retained in the second column and backflush time is very close to elution temperature/time from the second column. In addition, backflush flow through the first column is slow (there is less pressure at the purged tee and the pre-column is long, so flow is low), so backflushing takes longer. Finally, the pressure change during initiation of backflush takes longer to stabilize (the longer column provides a bigger buffer delay, so the transition from forward to reverse flow takes longer and this equilibration time increases the total backflush time further.
However, using a pre-column that is too short can also be problematic. It is trickier to determine the backflush timing and small changes in pre-column length (new pre-column not exactly the same length or diameter as old one) can ruin backflush timing. In a similar vein, if the pre-column is too short, sample components are poorly separated when they cross the junction, so it is more difficult to backflush between solutes of similar volatility (not a typical goal, but sometimes of interest). Finally, it is almost impossible to used pressure pulse injections with short pre-columns; whatever is driven onto the column during injection backflushes back into the inlet when the pressure pulse ends and returns to run pressure (the purged device pressure provides a source of the momentary reversed flow until pressures re-equilibrate).
Flow rates
My rule is that one should always add 10% flow through each purged connection in order to avoid peak tailing and losses up purge lines because of diffusion. As peaks in the pre-column and analytical column are experiencing standard partitioning processes, flow rate is an important parameter to set appropriately; one should avoid running the pre-column flow rate too far below optimum. I recommend taking the original column flow rate, running the pre-column at 95% of that and the analytical at 105% of that.
Always set pressures from the detector end toward the inlet and determine setpoints at initial oven temperature. So, if the original column flow was 1 mL/min at the start of the run, then set the junction pressure (the head pressure of the analytical column) so that its flow is 1.05 mL/min. Then, based on that junction pressure (now the outlet pressure of the pre-column) set inlet pressure such that the pre-column flow rate is 0.95 mL/min. In this manner, peak widths and retention times/temperatures will closely approximate those of the original method and my rule of 10% additional purge flow through the junction will be maintained. There are more sophisticated approaches to setting flows, but this one covers most circumstances well enough.
Determining backflush timing
It is extremely difficult to predict a priori the time/temperature at which the last component of interest passes the purged connection device (the junction) such that backflush can begin. So, it is most practical to simply use a sample or standard that contains the last eluting compound of interest plus a few other components that elute close to it and determine backflush time empirically. Depending on the length of the pre-column relative to the analytical column and the temperature program rate, the transit temperature of a component past the junction point can be 5 to 50 °C below its elution temperature from the analytical column. The closer the length of the pre-column to the length of the analytical column, the closer the temperature/time differential. If the pre-column is the same length as the analytical column (“split column configuration”) then the elution temperature differential is only a few column volumes (it is barely retained in the second column) so one might as well wait until the component elutes form the second column before initiating backflush (post-run backflush).
If the pre-column is 10% the length of the analytical column, then the elution temperature differential can be significant. If the pre-column is 10-20% the length of the original column, then I recommend trying a backflush that starts at a time corresponding to 25 °C below the elution temperature of the component of interest from the analytical column. Depending on the observed results, adjust the backflush time 10 °C earlier (if later eluting compounds still elute) or later (if the last compound of interest is not observed). Fine tune in 5 °C increments if more precision is necessary. In this manner within a couple experiments one can hone in on a suitable backflush time/temperature. Make sure to leave yourself a little buffer room to accommodate slight changes in column length, diameter, or film thickness as pre-column is replaced.
To accomplish a backflush, simply time program inlet pressure to reduce to a low, but controllable, pressure, such as 1-2 psig. Even though the inlet pressure goes down, the junction pressure is maintained such that the analysis can continue in the normal manner in the second column. In this way, backflush of the pre-column happens while the analysis continues. Remember to ensure that the split vent flow is at least 10% higher than the calculated backflush flow rate.
The length of time one should backflush follows the same guidelines as when post-column backflush is used: even though less column volumes might work, 10 void volumes is a safe benchmark. Calculate the reversed column flow through the pre-column based on its dimensions, temperature, and carrier gas type. If the requisite number of void volumes can not be accomplished before the elution of the last component of interest from the analytical column, then add the extra time needed as a post-run backflush event, or simply maintain the temperature and pressure conditions at the time of the elution of the last compound of interest. One always has the option to increase the junction pressure to speed up the flow in the pre-column and decrease post-run backflush time. Just be aware that this also increases the forward flow through the analytical column, so be careful to not exceed the recommended flow going to any flow sensitive detectors.
This blog article series is produced in collaboration with Dr Matthew S. Klee, internationally recognized for contributions to the theory and practice of gas chromatography. His experience in chemical, pharmaceutical and instrument companies spans over 30 years. During this time, Dr Klee’s work has focused on elucidation and practical demonstration of the many processes involved with GC analysis, with the ultimate goal of improving the ease of use of GC systems, ruggedness of methods and overall quality of results. If you have any questions about this article send them to techtips@sepscience.com
