When I ask what stationary phase is used in an HPLC method, the answer is almost always the manufacturers name (Agilent, Phenomenex, Thermo, Waters, YMC etc.)  and the phase name (Zorbax, Luna, Hypersil, BEH, Triart etc.) and sometimes the phase chemistry (C18, C8 etc.), but usually only when the bonded phase is an alkyl silica.  When the phase is anything other than a conventional alkyl silica, the fashion is to know the phase by it’s trade name – for example Agilent Zorbax Bonus-RP.  But what is the chemistry of this phase?  What are the mechanisms of interaction with the analyte and hence how is retention and selectivity gained from this phase?  How can we troubleshoot separation problems or develop suitable methods without a good knowledge of the bonded phase chemistry.

It also very important to realise that, even if the bonded phase chemistry is known, a significant contribution to retention and selectivity will be made by the underlying silica and any surface treatments which are carried out after the stationary phase is bonded.

With this in mind, let’s take a whistle stop tour around the surface of some reversed phase stationary phases to assess what factors actually influence the retention and separation of our analytes. I can’t include the vagaries of the stationary phase ligands, silica and post bonding treatments of every manufacturer, but the information that follows will give a reasonably comprehensive summary of the chemistry inside the column on your HPLC!

Let’s start with a look at the nature and treatments of the silica used for stationary phase particles.

Section 1 - various types of silica which are used to construct stationary phase particles.  The traditional polysiloxane is shown in the box region and this fully inorganic polymer has reasonable pH and mechanical stability.  Some manufacturers use organic / inorganic silanes (methyl - polysiloxane or ethyl-polysiloxane) to impart better pH stability.

Section 2 – shows the surface silane (Si-OH) groups of a polysiloxane matrix which are used to chemically attach the bonded phase ligand.  The chemical nature and environment of these siloxanes can have a big influence of the retention and especially the selectivity of a stationary phase.  Silanes or ‘silanol’ groups are inherently acidic (pKa typically 3.5 – 4.5) and are therefore either polar (eluent pH > 5.5) or anionic (eluent pH < 2.5) and can interact with our analytes to alter the analyte / stationary phase interaction.  If the analyte contains polar functional groups then, the desired analyte / stationary phase interactions are supplemented by secondary polar-polar interactions, which can significantly affect the selectivity of the stationary phase.  At low eluent pH the silanes will be ionised (Si-O) and the secondary interactions will be stronger (electrostatic) and can produce significant secondary interactions with ionised basic analytes, which is sometimes seen as peak tailing in the chromatogram.  Section 2 shows a special case in which adjacent silanol groups are hydrogen bonded with each other, reducing their acidity and resulting in lower peak tailing but retaining the ability to provide polar interactions with analytes.  Manufacturers can ensure a high degree of silanol inter-hydrogen bonding and this is a typical feature of modern ‘Type B’ silica.

Section 3 - shows the inclusion of a metal ion close to the surface of the silica matrix.  These metal ions can make the silica surface more acidic (show as a co-ordination effect with a proton producing a Bronsted acid) and again can lead to secondary interactions with polar or ionogenic analytes, altering selectivity and potentially leading to peak tailing.  Most modern Type B silica’s are washed to remove the majority of metal ions which are known to give rise to the most pronounced secondary effects (Fe and Al are known to be significant in this respect).

Section 4 – shows an alkyl bonded phase ligand (C8) and the siliyl ether linkage (Si-O-Si) which is formed to anchor the stationary phase to the silica surface.  At low pH it is this ether linkage which is susceptible to hydrolysis and some manufacturers replace the methyl group shown in the highlight with bulkier isopropyl or isobutyl groups to provide increased steric protection to hydrolysing agents.  The alkyl ligand shown is monomerically bonded, however it is possible to use dimeric or trimeric ligands (di- or tri-chloro alkylsilanes) which increases the numbers of points of attachment to the silica surface.  , as  illustrated in Section 7. If hydroxyl groups are used in place of the methyl substituents, it is possible to attain some degree of chemical cross-linking post bonding.  Phases which use more than one point of attachment or cross-linking are often known as ‘polymeric’.

Section 5 – shows a ‘lone’ or ‘acidic’ silanol group.  This is a silanol which is isolated after the bonding of the stationary phase ligand and will cause a high degree of secondary interaction with polar and ionogenic analytes.  It is inevitable that residual silanol groups will remain after the stationary phase is bonded and depending upon the nature of the silica and the bonded phase, as much as 40% of the original silanols may remain after bonding, and therefore significantly influence the selectivity of the stationary phase.  As was noted above, low energy silanols are often very useful in adjusting the selectivity of the phase, however these lone / acidic silanols are undesirable due to their detrimental effect on peak shape.

Section 6 – shows a lone silanol group which has been ‘deactivated’ using an end-capping reagent, in this case a trimethyl silyl moiety.  This treatment reduces the ability of the silanol group to participate in secondary retention effects with polar / ionogenic analytes and therefore will help to reduce peak tailing effects.  There are innumerable approaches and reagents used to improve the degree of end capping, however, even the most reactive reagents will still leave a significant number of unreacted silanols post treatment.  It is also possible to use more polar, non-ionogenic, reagents to alter the selectivity of the phase and in some cases these polar end-capping reagents will allow the phase to be used with 100% aqueous eluents, which are sometimes necessary for the retention of more highly polar analytes in reversed phase mode.

In many HPLC selectivity classification models, these polar and electrostatic interactions are identified as being particularly significant in defining the stationary phase selectivity.  Therefore the type of silica used by the manufacturer, the bonding chemistry, post bonding chemical derivatisation (end-capping etc), washing and heat treatment will all affect the selectivity of the phase.  This may go some way to explaining why there are almost 200 C18 stationary phases available to chromatographers, and that defining the column used as simply a C18 or even L1 (as per the USP classification) is too little information in order to reproduce a separation!

Let’s now look at some chemical variants of bonded phases and their typical applications.

Alkyl phases typically separate analytes based on their degree of hydrophobicity and differences in LogP values can be used as a very crude measure to assess the possibility of separation.  As the alkyl ligand length is decreased (C18 to C8 to C4 etc.) the hydrophobic retentivity of the phase decreases as does the degree of shape selectivity.  This latter feature describes the ability of the phase to discriminate between analytes of different size (typically length) or whose sphere of hydration (associated water layer) is significantly different, for example when comparing non-polar analytes to those containing a number  of polar functional groups.  The ratio of the amount of stationary phase ligands bonded to the stationary phase surface and the silica surface area will control the spacing between the ligands and hence the ability of the analyte to access to the various moieties on the silica surface.  Adjusting this ratio is also a popular technique to adjust the selectivity of the stationary phase.  Perhaps counter intuitively, for a given ligand density (stationary phase loading per unit silica surface area) the longer the alkyl ligand the greater will be the effect of secondary interactions with silica surface silanol species.

Phenyl phases are used to alter separation selectivity by exploiting the π-basic (electron donating) nature of the phase.  Analytes which show altered selectivity include bases, nitroaromatics, polar compounds with stronger dipole moments and heterocycles.  It is important to note that the π-π interactions are often enhanced when methanol is used as the organic modifier.  Variants of the phenyl phase include a heteroatom in the alkyl spacer (Phenyl-X) which is reported to enhance the π-donating nature of the phase and the inclusion of an alkyl spacer (such as phenyl-hexyl phases) to improve the hydrophobicity of the phase, providing greater retention of non-polar analytes.

Embedded polar group (EPG) phases are unique in that they contain a polar functional moiety within the alkyl chain.  This moiety is typically a carbamate, carbimide or other amine containing group and this functional group serves several purposes, including association with surface silanol groups to significantly reduce their secondary effects on retention and selectivity, thus improving peak shape.  The amine group is also capable of more pronounced hydrogen bonding interaction with acidic analytes and evidence exists for a degree of ‘shielding’ of basic analytes, thus the selectivity of these phases can be very different from alkyl bonded ligands.  These phases can also be used with 100% aqueous eluents, again for the retention of more polar analytes in reversed phase mode.

Pentafluoro phenyl (PFP) phases are capable of a wide variety of interactions with analytes including hydrophobic, π-π, dipole and hydrogen bonding interactions and are known to be particularly ‘shape selective’.  The highly polar C-F bonds give rise to strong interaction with highly polar analytes and are also popular for the analysis of unsaturated and aromatic compounds, phenols, halogenated compounds and positional isomers.

Alkyl cyano phases show altered selectivity when analysing mixtures of polar and non-polar analytes and exhibit both dipole and hydrophobic interactions.  This makes them particularly amenable for the separation of analytes with differing functional group chemistry and to reduce the retention of highly hydrophobic analytes which exhibit long retention times with alkyl ligands.  Earlier cyano phases had a reputation for high phase bleed, but most manufacturers now offer low bleed versions of their alkyl cyano phases.

‘Aq’ type phases are designed for the retention of polar analytes using highly aqueous eluents in reversed phase mode.  Typically they may contain polar endcapping groups with an alkyl bonded ligand or the ligand spacing may be increased to allow analyte access to the silanol groups on the silica surface.  It is proposed that a water enriched layer is developed at the silica surface due to the presence of more polar species (encapping or surface silanols) and that this acts to prevent phase collapse (self-association).  

It may well be the case that you are unaware of the chemistry of the phase you are using, referring to it merely by its trade name.  However, with a little research and using manufacturers literature, you should be able to discover much more about the surface chemistry which yields the selectivity of your separation, enabling you to make much more informed choices for method development and be better informed when troubleshooting separations.  Remember also that stationary phase selectivity in reversed phase HPLC is not just about the chemistry of the bonded phase, but that the silica used, the nature of any post bonding treatments and the conditions used can play a major part in defining the selectivity of the stationary phase.