| CHROMacademy Resolver Issue 3: Column Choice for Capillary GC |
 |
In this month’s Resolver, CHROMacademy Technical Director Tony Taylor discusses key factors in GC Column Choice, including Stationary Phase Chemistry and the Rules we can use to choose the optimum phase for our separations. |
 |
| Stationary Phase Interactions: |
|
jump to section » |
| Stationary Phase Chemistry : |
|
jump to section » |
| Golden Rules of Column Selection : |
|
jump to section » |
|
|
| Stationary Phase Interactions: |
|
|
|
| |
Mechanisms of Analyte – Stationary Phase Interaction
In order to understand the basis of the choice of stationary phase for capillary GC columns, you should know something about the intermolecular forces which govern the analyte – stationary phase interactions (i.e. how well retained an analyte is by the stationary phase).
The intermolecular forces responsible for adsorption interactions between analyte molecules and GC stationary phases can be broadly categorized as:
- (Table 1) |
 |
Interaction |
Energy (kJ/mol) |
Dispersive |
<<1 |
| Dipole-Induced Dipole |
1 |
| Dipole-Dipole |
3.3 |
| Hydrogen bonding |
19 |
Table 1 – Typical energies for various intermolecular interactions |
|
|
|
| |
The Concept of Polarity
All covalently bonded molecules will share electrons between the bonded atoms. A ‘non-polar’ covalent bond has a uniform distribution of electron charges between the atoms –the simplest non-polar covalent bonds exist in homonuclear diatomic species such as Cl2 or H2. In this type of molecule there is no permanent localised electrical charge build up, electrons are shared uniformly within the molecule.
Alternatively, a ‘polar’ bond displays a non-uniform electron distribution cloud. This typically occurs when two non-metal atoms which are more than two positions apart in the periodic table are involved in the bond –some examples are shown below. Here an electrical ‘dipole’ is established where the atoms within the bond are permanently charged due to the delocalization of the electron cloud.
The blue arrow beneath the HCl molecule indicates a permanent charge associated with the bond called a ‘Dipole Moment’.
The chlorine atom carries a small negative charge and the hydrogen a small positive charge |
 |

Figure 1: Graphic representation of a Dipole Moment |
|
| |
| Such dipole moments occur when the electronegativity of the atoms (essentially their power to attract electrons), differ substantially. You can find a great link to more information and an excellent table of values at this link:
http://en.wikipedia.org/wiki/Electronegativity |
| |
|
| |
| One can predict the strength of a dipole within a molecule by comparing the relative electronegativity values of the atoms involved in the bond. |
| |
| C-S = 2.5 – 2.5 = 0.0 |
| C-H = 2.5 – 2.1 = 0.4 |
| C-N = 3.0 - 2.5 = 0.5 |
| C-O = 3.2 - 2.5 = 1.0 |
| O-H = 3.5 – 2.1 = 1.4 |
| C-F = 4.0 - 2.5 = 1.5 |
|
| EN units difference |
| EN units difference |
| EN units difference |
| EN units difference |
| EN units difference |
| EN units difference |
|
Increasing electronegativity
 |
|
|
Table 2: Electronegativity differences for some typical covalent bonds |
| |
|
| |
The C-F bond has a dipole moment so high that it borders on becoming an ionic bond.
You will notice above that some of the functional groups typical of our own analytes have relatively high dipole moments.
When assessing the polarity of a molecule one should look out for common functional groups such as: Acids, Alcohols, Amines, Amides, Thiols etc. |
 |

Figure 2: Dipole moments within common organic molecules |
|
| |
|
| |
Dispersive Interactions
All substances contain small dipoles (because electronegativities of all atoms differ). These dipoles fluctuate throughout the molecule and as two molecules approach each other their dipoles can induce the opposite dipole in the other molecule and a small attractive effect is seen – this is often called a dispersive interaction and occurs between compounds which are predominantly non-polar. Dispersive interactions occur with all substances, regardless if there is another overriding interaction (such as a dipole interaction).
Subscribe to CHROMacademy for Multi-Media explanations of Dispersive Interactions |
| |
|
| |
Dipole Interactions
Dipole interactions come in two sorts – Dipole-Dipole and Dipole-Induced Dipole.
Dipole interactions occur between substances whose permanent dipoles come into close contact with each other. Dipole-Induced dipole interactions occur when a polar substance meets a less polar or non-polar compound and the stronger dipole induces a more permanent dipole in the other substance and an intermolecular attraction occurs.
Subscribe to CHROMacademy for Multi-Media explanations of Dipole Interactions |
| |
|
| |
Hydrogen Bonding Interactions
Hydrogen bonding is a special case of a dipole-dipole interaction in which the dipoles
associated with the functional groups of the two molecules come into close proximity. Hydrogen bonding interactions
are very strong compared to dispersive interactions and in the extreme (e.g., the association of water with methanol)
the dipole-dipole interaction energy can approach that of a chemical bond.
It should be noted that even when molecules are undergoing hydrogen bonding –there is
still an underlying weak dispersive interaction occurring simultaneously. |
 |

Figure 3: Typical Hydrogen Bonding Interactions |
|
| |
| Subscribe to CHROMacademy for Multi-Media explanations of Hydrogen Bonding |
| |
|
|
|
| |
We tend to chose GC stationary phases on the basis of the three interaction types outlined above.
For example, a non-polar analyte will undergo predominantly dispersive types of interaction. For good retention, we would therefore choose a
dispersive type stationary phase.
We should begin then to classify stationary phase types according to their predominant interactions with analytes to enable
us to make good stationary phase choice for retention, and hopefully resolution!
Most GC stationary phases are highly viscous liquid polymers whose chemical composition depends upon the monomeric units used
to derive the polymeric form. The careful choice of these monomers and their ratio in the polymeric form will dictate the chemical
nature of the stationary phase and its application area.
Figure 4 shows the broad classification of phases currently available and their predominant interactions and Table 3 summarises
the information
Figure 4: Various Stationary Phase Chemistries |
|
|
|
|
|
|
|
| |
| Functional Group |
|
Dispersion |
|
Dipole |
|
Hydrogen Bonding |
| Methyl |
|
Strong |
|
None |
|
None |
| Phenyl |
|
Very Strong |
|
Weak |
|
None |
| Cyanopropyl |
|
Strong |
|
Very Strong |
|
Moderate |
| Triflouropropyl |
|
Strong |
|
Moderate |
|
Weak |
| PEG |
|
Strong |
|
Strong |
|
Moderate |
|
|
|
| |
Table 3: Summary of Various Stationary Phase Types and their predominant interactions |
| |
| Subscribe to CHROMacademy for a fantastic tool to compare Capillary GC columns and their manufacturers Trade Names |
| |
| |
|
|
| |
Dispersive Interactions
Using the ‘like dissolves analogy’ dispersive phases are typically used to analyse non-polar analytes. When a series of n-alkanes is chromatographed with a diemthylpolysiloxane column, they elute in boiling point order with the lowest boiling analyte eluting first.
However, the boiling point analogy is lost when a series of aromatic compounds are also chromatographed at the same time. Whilst they remain in boiling point order the relative spacing of the analyte bands is not as predicted and this is a reflection that the two ‘species’ of analyte undergo different degrees of dispersive interaction with the stationary phase. |
 |
 |
| |
Figure 5: Boiling point separation of an n-alkane and aromatic series on a 100% polydimethylsiloxane stationay phase |
|
|
|
|
 |
 |
As a general rule of thumb, if the boiling point of two compounds differs by 30oC or more, then they may be separated by most stationary phases. This is due to the fact that dispersion is the dominant interaction for a wide range of stationary phase types. If the compounds boiling point differs by less than 10oC (and the compounds do not belong to a homologous series), then the boiling point analogy is more likely to be in error.
When dealing with analytes with mixed functionality –the boiling point analogy is often confounded as can be seen. Note that s-butylbenzene and n-decane have the same boiling point and are separated via the differing strengths of their dispersive interaction. |
|
Column: DB1 (30m x 0.25mm, 0.25μm) : Carrier: Helium @ 32 cm/sec. : Oven: 45 – 115 oC @ 5 oC/min. |
Figure 6: Failure of the boiling point analogy when analytes of mixed functional groups are chromatographed simultaneously |
| |
|
| |
Dispersive Interactions / Polarity
Another stationary phase characteristic that may affect retention in a predictable manner is the phenyl
content. The higher the phenyl content of the stationary phase, the higher the retention of aromatic solutes RELATIVE to
aliphatic solutes. This does not necessarily mean that the aromatics are more retained, but that they shift relative to
aliphatic solutes. This is demonstrated in the top example opposite. Increasing the phenyl content of the phase causes
earlier elution of the n-alkanes and later elution of the alkyl-benzenes. |
| |
 |
 |
1. Toluene, 2. n-Octane, 3. Ethylbenzene,
4. n-Nonane, 5. Propylbenzene,
6. sec-Butylbenzene, 7. n-Decane,
8. 1,2-Dichlorobenzene, 9. Butylbenzene,
10. n-Undecance, 11. n-Dodecane |
|
Figure 7: Predictable retention pattern change when moving to a more aromatic phase –
aliphatic compounds move to shorter retention time relative to the aromatics |
| |
|
| |
| Altering the phenyl content of the stationary phase affects the polarity of the column.
By increasing the phenyl content the polarity increases (due to the presence of p -electrons in the aromatic groups)
which renders the stationary phase ‘less non-polar’. This is well described in Figure 8 where the retention of the
polar analytes (alcohols) is increased relative to the non-polar analytes (n-alkanes) where the phenyl content of the
phase (and hence the polarity) is increased. |
| |
 |
Figure 8: Polar compounds better retained when
increasing the phenyl content of the phase – increasing an increase in the overall stationary phase polarity |
| |
|
| |
Dipole Interactions / Hydrogen Bonding
If the stationary phase is capable of dipole interaction, it enhances its power to separate solutes
whose dipole moments are different. Cynopropyl, trifluoropropyl and Poly Ethylene Glycol (PEG) phases all show good dipole
interaction properties.
The degree of peak separation for solutes with different dipoles often changes if a stationary
phase with a different amount of the dipole interaction is used. If the dipole difference between compounds is small, a greater amount of the appropriate phase is required (i.e. 50% cyanopropylphenyl-methyl instead of 14% cyanopropylphenyl-methyl). It is difficult to be predictive about the magnitude of the separation change for all peaks. Empirical studies show that this type of phase are well suited for compounds which have a base or central structure to which different groups are attached in various positions. Examples include substituted aromatics, halocarbons, pesticides and drugs.
The same stationary phases that undergo dipole interactions also undergo hydrogen bonding interactions
–with the PEG phases showing the strongest interaction. Again where the analyte hydrogen bonding potential differs only slightly,
a stationary phase with a greater amount of the appropriate group is required. This is demonstrated in Figure 9.
It could be argued that by using a 50% Trifluoropropyl phase the para, meta substituted forms could be even further resolved. |
| |
 |
Figure 9: Increased isomeric separation using a TFP phase which shows strong dipole interactions |
| |
|
| |
| Hydrogen bonding interactions also show similar trends, again relying on increased amounts of the hydrogen bonding monomer within the phase when the dipole moment differences between analytes is small. |
| |
 |
Figure 10: Increased Creol isomar separation due to hydrogen bonding with the TFP phase |
| |
|
| |

Figure 11: Hydrogen Bonding interaction between m-Cresol and the trifluoropropyl stationary phase |
 |
Strength |
Compounds |
Strong |
Alcohols, carboxylic acids, amines. |
| Moderate |
Aldehydes, esters, ketones |
| Weak to None |
Hydrocarbons, halocarbons, ethers |
Table 4 – Hydrogen bonding capability of various analyte species |
|
| |
|
| |
Porous Layer Open Tubular (PLOT) Columns
Porous Layer Open Tubular (PLOT) columns are intended for Gas Solid Chromatographic (GSC) applications. PLOT columns are capillaries in the conventional sense but the inner wall of the capillary is coated with small, solid porous particles using a binder. The particles are usually Alumina or Molecular sieve and solutes are separated on differences in their adsorption properties, with size and shape differentiation also occurring.
PLOT columns are used primarily for the separation of highly volatile liquids and permanent gases without the need for cryogenic or subambient cooling of the GC oven. Separations that would require column temperatures well below ambient temperatures, even with thick film capillary columns, can be obtained at ambient temperatures or above using PLOT column technology.
Alumina columns are well suited to the analysis of C1 – C10 hydrocarbons and small aromatics, whilst the KCl derivatised version of the column produces altered selectivity for the same compound group.
The Q designated columns show better selectivity for C1-C3 hydrocarbons, but give very long retention and broadened peaks for anything heavier than C6. These columns are also able to separate sulphur gases and most light hydrocarbons.
Molecular sieve columns are used to separate many noble and permanent gas samples and are also good for the separation of solvents. |
 |

Figure 12: Permanent gas analysis using a Molecular Sieve coated PLOT column |
|
| |
|
| |
Stationary Phase Selection Summary |
| 1. |
|
If no information or ideas about which stationary phase to use is available, start with a 100% polydimethylsiloxane or 5% phenlmethyl polysiloxane phase. |
| 2. |
|
Low bleed ("MS type") columns are usually more inert and have higher temperature limits. |
| 3. |
|
Use the least polar stationary phase that provides satisfactory resolution and analysis times. Non-polar stationary phases have superior lifetimes to polar phases. |
| 4. |
|
Use a stationary phase with a polarity similar to that of the solutes. This approach works more times than not; however, the best stationary phase is not always found using this technique. |
| 5. |
|
If poorly separated solutes possess different dipoles or hydrogen bonding strengths, change to a stationary phase with a different amount (not necessarily more) of the dipole or hydrogen bonding interaction. Other co-elutions may occur upon changing the stationary phase, thus the new stationary phase may not provide better overall resolution. |
| 6. |
|
If possible, avoid using a stationary phase that contains a functionality that generates a large response with a selective detector. For example, cyanopropyl containing stationary phases exhibit a disproportionatly large baseline rise (due to column bleed) with Nitrogen Phosphorous detectors (NPD). |
| 7. |
|
100% Methyl (or 5% Phenyl), 50% Phenyl, 14% Cyanopropylphenyl and WAX (PEG) cover the widest range of selectivities with the smallest number of columns. |
| 8. |
|
PLOT columns are best used for the analysis of gaseous samples at above ambient column temperatures. |
|
| |
|
| |