How to choose the correct GC column – Part 3

In this post, I will explain how one can use column capacity (along with detector selection) to help narrow down GC column choices even further. Before reading this post, I suggest you review Parts 1 & 2, which can be found in the links below.

How to choose the correct GC column – Part 1

How to choose the correct GC column – Part 2

First, let’s talk about column capacity (also referred to as sample loading capacity).  A column’s internal diameter (ID) and film thickness (µmdf) dictates a column’s capacity.  Keep in mind that the capacities shown in the figure below are for compounds which are soluble in a column’s phase.  If the compound is not soluble, a column’s capacity will be less (in this case, capacity will be determined entirely by column ID).  For columns of the same internal diameter, capacity has a linear relationship with film thickness.  For example, a capillary column with a 1µmdf film thickness has approximately twice the capacity of a column with a 0.5µmdf film thickness.


To read more about identifying and preventing column overloading, please see the link below.

How to understand and deal with overloading in GC


Next, let’s look at the sensitivity of common GC detectors (excluding mass spectrometers). Generally speaking, I am going to consider all common GC detectors as being sensitive enough to use with most capillary columns (except for a TCD, which is discussed briefly below). However, one needs to remember that many detectors are only sensitive for a specific compound class, and not all compounds. For example:

FID’s are the most sensitive for compounds with carbon-hydrogen bonds. The more C-H bonds, the more sensitive.

ECD’s are the most sensitive for halogenated compounds. Order of decreasing sensitivity is I > Br > Cl > F

NPD’s are only sensitive for compounds containing nitrogen or phosphorus.

FPD’s are only sensitive for compounds containing sulfur or phosphorus.

PID’s can only detect compounds whose ionization potential is less than the output of the PID lamp.  The link below lists ionization potentials of many common compounds.

Photoionization Characteristics of Selected Compounds

TCD’s will (theoretically) detect all compounds except the carrier gas. However, they have the lowest sensitivity of all the common GC detectors. If using a TCD, consider using only a 0.53mmID capillary column (or larger ID micropacked or packed column) to ensure that there will be enough column capacity.

Finally, we need to consider the detector’s linearity range for the compounds of interest. After all, if your detector isn’t sensitive enough, or is overly sensitive, sample size and/or concentration will need to be modified to make sure you can obtain linearity and produce a successful calibration curve. Ideally, you want to obtain a compound calibration curve that looks like the one below.

Please note, for all curves, the Y-axis ↑ is increasing response, and the X-axis → is increasing concentration.


A calibration curve which lacks compound sensitivity (or may contain matrix interferences) at lower compound concentrations could look like this:

An overloaded detector at higher compound concentrations may have a compound calibration curve which looks like this:

By determining the linearity range of your compounds using a specific instrument/detector, you will be able to predict the proper internal diameter (and to a lesser degree, film-thickness) needed to make sure the column has adequate column capacity. You will then choose a film-thickness based upon the compounds you need to separate. But try not to overcompensate because larger ID columns (and thicker films) will produce broader compound peaks and higher column bleed than their smaller ID, thinner-film counterparts, and may actually decrease overall compound sensitivity.

One final thought before I end this post, remember the β values (also known as phase ratios) I listed in Part 1? Well, besides using them as a tool to help one determine if a column could be considered a volatile or semi-volatile column, their true purpose is to help the analyst with column selection if switching to a different ID column (of the same phase). Simply stated, you will try and match the closest numbers from each column dimension. Here is how it works; let’s assume you were currently using a 30m x 0.53mmID x 1.0µmdf, which has a β value of 128. If you wanted to switch to a 0.25mmID column, you would choose a 30m x 0.25mmID x 0.5µmdf, which has a β value of 125. Just remember to make sure that this 0.25mmID column will have enough column capacity before you switch. In addition, don’t forget that you will also need to modify carrier gas flows (and possibly detector gas flows, especially for the make-up gas) and even the GC oven temperature program to obtain the same compound elution order and retention times.

Hopefully these three posts are able to help everyone narrow down column choices to one, or only a few column catalog numbers. If not, there is still one last post that will cover a few topics not discussed so far.

Thanks for reading.

Leave a Reply