Here we discuss the relationship between resolution, time of analysis and sample capacity and how they impact GC performance. We’ll keep it soft and simple and in layman’s terms. Perhaps we will delve deeper into more detail and specific equations in a future blog.
Not quite as mysterious as it’s Bermuda counterpart, the Chromatographic Triangle has confused many analysts over the years. Lost within its realm are countless man hours devoted to optimizing separation performance. Some have equated the triangle to the task of squeezing a balloon; the tighter you squeeze the more it squirts out between your hands until…. Bang!
The Chromatographic Triangle states that whenever an improvement is made in terms of either resolution, run time or sample capacity a sacrifice is imposed on one or both of the remaining factors. Let’s look at the relationship between the three points of the triangle and how they impact GC analysis.
- Need more resolution? Try using a longer column or slow down the temperature ramp. Of course that will cost you in terms of run time. Another way to increase resolution is to use a thinner stationary phase; however, your sample capacity will suffer.
- Want more sample capacity? Use a longer column or increase the stationary phase film thickness. The former will result in longer run times while the latter will negatively impact the resolution.
- Let’s see how we can save time with Fast GC methods. First, use hydrogen for the carrier gas. Then you can try to shorten the column and/or ramp the oven faster. Both will reduce the resolution. If you couple a shorter column with a thinner stationary phase you may retain resolution but you’ll lose sample capacity.
How can these concepts be applied to everyday methods? First and foremost the method requirements must be well defined. Is it imperative to achieve baseline separation for each and every peak? This is certainly not the case in many applications such as petroleum methods. Other times there are only a few peaks of interest that must be resolved from the remaining components. Is there a need to increase speed? Maybe not if you only running a few samples per shift or if the bottleneck lies in data reduction. Is sample capacity important? Real world samples with an uncontrolled matrix place a higher burden on sample capacity. Environmental labs often rely on thick film megabore columns to avoid sample capacity issues. On the other hand, a quality control lab running routine samples with a known matrix needn’t be as concerned with overloading the column. Once you fix the method requirements the conditions can be optimized to achieve the best results for a given application.
Let’s take a hypothetical example. We’re monitoring a production run of decaffeinated tea. The production conditions must be adjusted to keep caffeine levels with 0.5% of the target level. Samples are continually analyzed during the production run and results must be generated immediately. The tea sample is very complex, however, it can be cleaned up and the concentration reduced by a quick extraction into 1 mL hexane or methylene chloride. We consider an isothermal run, however, some sample components elute much later than caffeine and might show up as ghost peaks in subsequent runs. A standard 15 meter 320um X 0.25um 5% phenyl methyl silicone column is selected because it has the required sample capacity and selectivity for the peak of interest (caffeine). It is also rugged, relatively inexpensive and available from many vendors. The temperature ramp is set to start at the highest temperature that will isolate the caffeine from the solvent front and early peaks. High initial temperatures save time during oven cool-down. The temp program is adjusted to quickly ramp up after the caffeine has eluted to clean the column up for the next run.
That’s the fun part of chromatography. Each analysis has it’s own goals and challenges. In future blogs we can pull out some of the equations that describe column performance and see how they apply to real world applications. Do you have a specific application that you would like discussed? If so, send your request in to stevemac@zipsci.com