Silica as a Support Material for Reversed Phase HPLC :
Silica as a Support Material for Reversed Phase HPLC
Silica consists of siloxane bridges (silylether linkages), in which silicon atoms are three-dimensionally bridged by oxygen atoms, and terminate at the surface with silanol species - (Figure 1)
Figure 1 – Chemical structure of a silica particle
Bare silica, chemically untreated, can be used as the support material for normal phase (adsorption) chromatography and after chemical modification is the primary support material for reversed phase (partition) chromatography.
Silica is a support material with exceptional assets including:
High mechanical strength - required to cope with back pressures of 1000+ bar
High surface area - required for high efficiency separations (300m2/g is typical for a 5mm particle with 100Å pore diameter)
Available in a form pure enough to chromatograph polar and ionisable components when used in conjunction with bonded phases
Whilst we think of silica support materials as ‘spheres’, in reality there is less than 1% of the silica available surface on the outer surface of the sphere, the vast majority being found within the internal pore structure.
The major drawback to using silica is its susceptibility to hydrolysis at high pH (typically >pH7.5) especially in highly aqueous environments. Factors that influence the rate of silica hydrolysis include:
Aqueous content of the eluent (silica is more soluble in highly aqueous eluent systems)
Buffer type used
Temperature of the column / eluent
Pore volume (higher surface areas, achieved with smaller pore widths / volumes, will be more susceptible to faster hydrolysis especially in highly aqueous environments)
Key Physico-Chemical Properties of Silica as a Substrate for HPLC
Surface Area – to a first approximation, the larger the surface area, the greater the number of stationary phase ligands and hence the greater the efficiency and loading capacity that can be achieved. The surface area is inversely proportional to the both the particle diameter and the pore width.
Is directly related to the molecular weight (actually hydrodynamic volume) of the analytes for which the column is intended. In order for the analyte to permeate into the pores of the substrate material, they must be sufficiently wide and a useful rule of thumb is:
Analytes ≤3,000Da use a pore size of 100Å or less.
Analytes 3,000Da – 10,000Da use a pore size of 100Å - 130Å
Proteins and Peptides ≥10,000, use a pore width of 300Å
However, if the pore width is less than 10 times the size of the analyte of interest, considerable band broadening can be observed due to the hindered mass transfer kinetics associated with this excessive retention within the pore. Pore widths can be measured by either mercury porosimetry or nitrogen absorption. Nitrogen absorption is generally favoured as it is effective for pore widths from 0.2nm – 30nm and can be used for compressive and fragile materials such as polymeric packing materials.
Also be of interest and is a direct measure of the interstitial volume within the particle. The pore volume has a direct affect on the mechanical strength of the particle with large pore volumes yielding weaker particles. These larger pore volume particles are generally reserved for absolute necessity, such as size exclusion and low pressure HPLC work.
Carbon loading is a measure of the carbon contained within the column from the modifying ligand and is determined by elemental analysis. For a similar alky chain length, C18 for instance, the larger the carbon load usually the greater the number of modifying ligands and the greater the retentivity for hydrophobic analytes and loading capacity.
Perhaps a better measure of column retentivity / hydrophobicity. Carbon load from elemental analysis is compared to the surface area of the support material, the molecular weight of the modifying ligand and the number of carbon present on each modifying ligand. This normalised value will define the concentration of modifying ligands on the surface and is of great importance. Initially it appears that the greater the surface coverage the better, yielding larger numbers of modifying ligands and fewer residual silanols. However, on occasions, where polar interactions with the silica surface are desired in order to illicit retention of polar analytes or altered selectivity, the surface coverage can be reduced to expose the analyte to polar silica surface. In these instances, it is essential the silica surface is fully hydroxylated (inter-hydrated, low energy) prior to chemical modification. The phase will then offer both hydrophobic (via the bonded phase ligand) and hydrophilic (via the polar but low energy surface silanol groups) interactions, which give a desirable mix between hydrophobic retention and polar selectivity. These phases are also suitable for work with 100% aqueous eluents (required for the retention and separation of highly polar analytes) and are commonly referred to mixed mode columns.
The particle size of the support material is of primary importance when selecting a stationary phase. By reducing the particle size diameter, not only is the efficiency of the column increased (plate height (H) is reduced), but the optimum linear velocity at which minimum plate height achieved is also increased. More efficient peaks can be achieved at elevated flow rates, leading to a reduction in analysis time accompanied, potentially, with an increase in resolution. Shorter diffusion paths and hence increased mass transfer kinetics (C term from the Van Deemter equation) are one reason behind this reduction in plate height. Figure 2 shows the relationship between particle size, flow rate (linear velocity) and plate height for columns packed with various silica particle diameters.
Figure 2 – Van Deemter curves for various particle diameter particles
It may seem that the smallest diameter particles will be the obvious choice for high efficiency separation; however this benefit does come at a substantial price...that of increased back pressure. The pressure increase is inversely proportional to the square of the particle diameter as shown in Equation 1 oppisite
Equation 1 – Pressure
Where any of the numerators increase, flow (F) and Column length (L) for instance, a resultant increase in back pressure is observed. Where any of the denominators decrease, column radius (r) and particle size (dp)for instance, an increase in back pressure is also observed. Note that the Radius and Particle Size functions are raised to the second power, thus causing their impact on back pressure to be much more dramatic.
Particle Size Distribution
The particle size distribution is another key parameter but in practice as this is not a user controlled variable, little practical attention is given to it.
Any particle size stated for an HPLC column (e.g. 5mm) will actually be the mean value from a distribution of particle sizes, achieved practically by the manufacturer using automated sub-sieve sizing techniques. In reality a wide distribution of particles will lead to heterogeneity within the column and as such an increase in variable path lengths the analyte molecules can take through the column (Van Deemter A term, often called the Column Packing term) which will give rise to band broadening.
This phenomenon has contributed to the sudden upsurge in commercial availability of Superficially Porous Silica (SPS) particles with small pore volumes (80Å - 120Å) with their increased efficiency / reduced plate height for the analysis of low molecular weight analytes. SPS particles have been available in their 300Å guise for the analysis of high molecular weight analytes for a number of years because of the improved mass transfer kinetics (C term) which were debilitating for conventional porous microspheres. These particles consist of a solid silica core with a layer of porous silica deposited around the outside, effectively shortening the path length for diffusion within the internal pore structure and improving mass transfer kinetics (Figure 3).
Figure 3 – An example of a typical SPS particle
Figure 4 – Various particle size distributions
However, for lower molecular weight analytes (<600 Da) the effects of increased mass transfer are less important and the increase in efficiency / reduction in plate height attributed, in the main part, to the very narrow particle size distribution that can be achieved from creating a particle with a solid core and the reduction multiple path effects (A term). Figure 4 shows a typical plot of normalised particle size distributions for an SPS particle and various porous microsphere particle sizes.
Fully hydroxylated silica will have a Silanol surface concentration of ≈8µmol/m2. Following chemical modification > 4µmol/m2 of these silanols may remain even with optimum bonding conditions due to steric limitations of the modifying ligands. This indicates that on a molar basis there are more residual silanols remaining than actual modified ligand. In order to remove some of these residual silanols, an end-capping process may be undertaken. Short chain, less sterically hindered hydrophobic ligands, (commonly trimethyl / tri-iodo chlorosilanes or similar), are chemically reacted with the remaining unbounded silanol species, leading to improved peak shape with polar and ionisable analytes. This is only a partial solution, however, as not all of the surface silanol groups will be reacted even using sterically very small liagnds and optimised bonding conditions, also the end-capping ligand is prone to hydrolysis especially at low pH.
Silanol groups are present in numerous conformations, with some being more active than others at causing analyte peak tailing and / or irreversible retention.
Figure 5 – Various Silanol conformations
Acidic (lone) surface silanol groups give rise to the most pronounced secondary interactions with polar and ionisable analytes. Modern silica is designated as being Type I or Type II, which primarily describes the nature of the silanol surface. Type I silica is ‘high energy’ (non-homogenous) and contains a higher density of lone silanol groups, whereas Type II silica is much more homogenous (inter-hydrated) and therefore gives rise to much improved peak shapes. In order to create a more uniform (homogeneous) silica surface, manufacturers ensure the silica surface is fully hydroxylated prior to chemical modification. The incorporation of an acid wash step and avoidance of treatments at elevated temperature renders the majority of the surface in the lower energy geminal and bridged (vicinal) confirmation, creating Type II silica. Figure 6 shows how various basic and polar analytes are affected when analysed using Type I silica, Hypersil, as compared to Type II silica (Hypersil BDS).
Figure 6 – Basic, polar analytes analysed using Type I and Type II silica
Mobile phase pH will affect the degree of silanol ionisation and therefore the degree with which they interact with polar and ionisable analytes, causing peak tailing. Typically the pKa of surface silanol species lies in the range pH 3.8 – 4.5 and at eluent pH ≤3 all but the most acidic will be fully protonated and therefore peak tailing will be at a minimum (Figure 7).
As the eluent pH increases the degree of ionisation (through deprotonation) increases, and peak tailing becomes more pronounced as the silanol
groups interact with charged and polar species in solution (Figure 8)
Figure 8 A – Silanol – Base interactions and effect on peak shape at pH < 3.0
Figure 8 B – Silanol – Base interactions and effect on peak shape at pH > 3.0
Metal Ion Content
Where the silica contains metal ions from the manufacturing and particle sizing production steps, in particular iron and aluminium, the acidity of the lone silanols is further increased leading to an increase in peak tailing with polar and ionisable compounds. Further, analytes which are capable of chelating will also interact with the column in this way leading to additional unwanted secondary retention and more pronounced peak asymmetry. Manufacturers use various silica washing processes to remove metal ions from silica and also to avoid the electrolytic decomposition of metal ions from the stainless steel column tubing and metal column end frits used to retain the packing within the column.
Metal Concentration (ppm)
nd = none detected
Table 1 – Comparison of metal ion content in Type I silica (Hypersil) vs. Type II silica (Zorbax Rx-SIL)
Hydrolytic (pH) Stability
The chemical properties of silica are also of paramount importance, in particular are:
the stability of the silica support material to hydrolytic attack at elevated pH
the mode by which the surface has been chemically modified
the stability of the modifying ligand at low pH
secondary treatment of the remaining unreacted silica surface (endcapping)
The stability of silica as a support material at pH ≥ 8 and the approaches of various manufacturers to combat this have already been discussed.
There are many options available to enable work at elevated pH without any gross degradation of the support material.
The mode by which the modifying ligand is attached to the support material had been an area of considerable interest previously, but nowadays most manufacturers use a similar process and have been doing so for over two decades. All applications utilise the silanols on the silica surface onto which the ligand is attached through the formation of a siliyl ether linkage using a condensation reaction. Early work into esterification of the silanol groups with alcohols was discounted due the ease at which the reaction is reversed in the presence of water. The most chemically robust, in terms of hydrolytic attack, is the chlorinated of the silanols forming a stable silicon-carbon bond. The reaction scheme for octyldimethylchlorosilane, in the presence of a basic hydrochloric acid scavenger, is shown in Figure 9
Figure 9 – Chlorination, and subsequent hydrolysis, of silica with a C8 ligan
The reaction scheme above depicts a monofuntional chemical modification of the support material as the alkylsilane contains one group able to react with the surface. The two methyl side substituent groups are most commonly employed due to their small size, allowing maximum surface coverage / minimum steric effects during bonding. The side substituents could be any hydrocarbon and will be discussed in greater detail later. In order to improve the surface coverage, di- or trifunctional silanes are used in place of the conventional monofunctional silane. Figure 10 depicts a difunctional modification where two Chlorine groups are employed at the expense of one of the methyl side substituent.
On average, only around half the modifying ligands bind to the surface in a bidentate manner. Due to steric orientations, the remainder bind with a single siloxane bond with the second functional chlorine being readily hydrolysed to form an additional residual silanol group. Therefore the number of residual silanol groups remaining is very similar to the monofunctional approach. Unfortunately, the situation is no better when a trifunctional reagent is used, as it is sterically impossible for a tridentate reaction with the surface. There will however be a greater degree of difunctional siloxane bonds found but once again, following hydrolysis of the remaining functional chlorines, the number is remaining silanols is almost identical.
Figure 10 – Difunctional chlorination of the support material
When the reaction takes place in the presence of water, the hydrolysis of the additional functional groups results in the attachment of the additional functional group to the surface. Therefore greater surface coverage can be achieved using multifunctional, (or so called polymeric bonding), where a high surface coverage is required. Most manufacturers still operate monofunctional binding as the water content during the reaction and the state of the silica surface have to be very carefully controlled in order to produce batch to batch homogeneity. Small variations in the surface coverage can have quite dramatic variations on the retention of analytes that are reacting with silica surface via hydrogen bonding or dipole interactions. Multifunctional bound modifying ligands serve to impart some enhanced stability of the modifying ligand which is prone to hydrolysis at low pH.
Manufacturers have produced various alternatives in order to operate outside the conventional pH limit associated with silica phases (typically 2.5 < pH <7.5). Some employ a modified version of silica whilst other approaches involve using alternative materials other than silica.
Traditional di-methyl chlorosilanes can be replaced with a bidentate ligand which incorporates an alkyl bridge (Figure 11). This approach acts to sterically protect the silica surface from hydrolysis associated with high pH work
Figure 11 – An example of a
ligand for high pH work
Other approaches use a hybridsilica that is manufactured using organic (alkylsiloxane) and inorganic (silane) monomers (Figure 12). This approach is becoming increasingly popular and now includes the use of hybrid materials which incorporate bridged surface inorganic species
Figure 12 – Chemical structure of a typical hybridised silica particle
Polymers consisting of divinylbenzene, methacrylate and vinyl alcohol are also commercially available although their current use is mainly restricted to proteomics. Polymeric phases tend to be non-porous and hence exhibit limited efficiency due to their restricted surface area. Unlike silica, polymeric support materials imbibe large enough of amounts of solvent and demonstrate swelling and shrinkage of the substrate. These interstitial volumes that solvent molecules penetrate are also accessible to small molecular weight analytes and due to the selling of the packing material, they demonstrate sufficiently poor mass transfer kinetics to adversely affect the band broadening and hence reduce column efficiency.
It is worth noting that the larger the modifying ligand, the less it is prone to hydrolysis. It is this reason why C18 ligands demonstrate much greater hydrolytic stability as compared with C3 for example. In additional to monofunctional binding, enhanced stability at low pH’s can be achieved by sterically protecting the siloxane bond from hydrolytic attack. This is generally achieved by replacing the methyl substituent groups with much larger, bulky groups such as isopropyl or tert.-butyl groups (Figure 13). Where such groups are employed, the separations with eluent pH ≈1 are possible without noticeable phase loss due to hydrolytic attack. Please do note however, due to the size of these side groups, limited surface coverage can be obtained as compared to methyl substituted analogues.
Figure 13 – Larger side substituents groups are less prone to hydrolytic attack
Porous Graphitic Carbon (PGC) has been successfully used as a support material to extend the pH operating range. It is mechanically strong, stable from pH 1 – 14 and does not suffer from same shrinkage and swelling associated with polymeric materials. It has been commercially available since the late 1980’s and interest in this and continues to find new applications and provides unique retention and separation of highly polar compounds. The backbone structure is constructed of intertwined bands of graphite and demonstrates very comparable retention to a C18 packing for hydrophobic analytes. The retention mechanisms in action are a combination of dispersive interactions between the analyte-mobile phase and analyte-graphitic surface together with charge induced interactions of polar analytes with the polarisable surface of the graphite. Analyte interaction with the stationary phase is dependent upon the type and positioning of functional groups together with the molecular area in contact with the planar graphite surface given the quest for analysis to be carried out on more complex mixtures containing polar and hydrophobic analytes, at more inhospitable pHs, PGC seems to offer a platform for controlled mixed mode chromatography and has the added advantage of being able to operate at inhospitable temperatures.
A further alternative to conventional porous silica microspheres are the so-called monolithic stationary phases. The monoliths consist of a single rod of porous organic polymer or silica cast inside a column. This highly porous rod consists of a bi-modal structure containing mesopores (≈ 13nm diameter) and macropores (≈ 2µm diameter) (Figure 14). The surface area created by the mesopore structure is ≈ 300m2/g. Operating back pressures associated with monolithic columns are lower due to the larger porosities associated with monolithic columns, 80% compared with 68% for discrete particles. There are also substantial advantages in efficiency available with monolithic columns due to the reduced mass transfer kinetics. Resistance to the eluent flow is low with the monolith columns and hence column back pressures are lower than seen with porous microspheres – leading to the possibility of flow programming to effect faster analyte elution without compromising resolution. Flow programming can be used in both isocratic and solvent gradient modes.
Figure 14 – SEM images of bimodal monolith pore structure
Alternative Stationary Phases for the Retention and Separation of Polar Analytes
One factor that has driven HPLC stationary phase development has been the surge in popularity of LC-MS, principally following the introduction of Electrospray Ionisation (ESI). A prerequisite for high sensitivity ESI is that the analytes are ionised in solution and this is most conveniently done by adjusting the mobile phase pH. This can be problematic for reversed phase chromatography as in order to generate any retention of these highly polar / ionised analytes a highly aqueous (≥95%(Aq)) mobile phase needs to be employed. Historically this has not been an option for traditional reversed phase columns where such eluent systems caused the modifying ligands to undergo a phenomenon known as phase collapse or self-association. Here the hydrocarbon modifying ligands will fold in on themselves in order to elude the highly polar mobile phase system and reduce the energy within the system. This is often catastrophic and necessitates the replacement or lengthy regeneration of the column. There are various options commercially available whereby reversed phased separations can be run using silica as the support material at 100% aqueous mobile phase compositions.
There have been many instances when secondary interactions have actually been used to gain retention of very polar analytes or to separate a particular troublesome critical peak pairs. The influence of the silica surface on analyte separation can be promoted by reducing the ligand surface spacing i.e. the distance between the chemically modified ligands (controlled using the surface area to carbon load ratio), and by reducing the modifying ligand chain length, thus making the surface more accessible.
Use of high purity silica with a fully hydroxylated (energetically homogenous) surface prior in combination with a low surface coverage of bonded phase ensures large areas of silanols remain in the bridged (vicinal) conformation (Figure 15) that are ‘available’ to the analyte. These hydrophobic regions will bind a layer of water (via hydrogen bonding interactions) which ensures that the surface is fully hydrated (Figure 16). The tendency for the hydrophobic ligand to self-associate in the presence of highly aqueous eluent systems will be negated by the repulsion from the adsorbed water layer, ensuring the modifying ligand remains activated. These columns are generally given the suffix (Aq) denoting their ability to operate in highly aqueous eluent systems. These stationary phase types are also very useful for the retention and separation of complex mixtures containing both polar and non-polar analytes.
Figure 15 –Comparison of inter-ligand bonding distances
Figure 16 – A fully hydrated silanol surface
Figure 17 – A hydrated polar endcapped surface
An alternative approach to the retention and separation of highly polar analytes is the use of a polar endcapping reagent. This works in much the same fashion, whereby the polar surface ensures the hydrophobic ligands remain activated by lowering the energetic stability of self-associating (Figure 17). Once again, the polar end-capping group can be utilised for the increased retention of polar analytes. A low surface coverage of modifying ligand, as described above, is generally favoured in this approach as the polar embedded ligands still have the propensity to being hydrolysed at low pH.
A final approach to operating in fully aqueous mobile phase systems in the reversed phase mode is to incorporate a polar function group into an alkyl ligand, typically an amide or carbamate group. There are numerous benefits associated with operating polar embedded columns (Figure 18). It is worth pointing out that the enhanced retention of polar analytes by way of dipole interaction or hydrogen bonding has been called into question in recent years. In fact there has been some evidence that the retention of polar analytes actually decreases when employing a polar embedded group, attributed to the polar embedded group interacting with the silica surface in place of the polar analyte. The advantage of this, the reduction of peak tailing commonly observed for basic analytes, far outweighs any slight reduction in retention. Once again it is the adsorbed water layer which ensures that no self association takes place. Typically, the alkyl chain length is not given by the manufacturer but it is typically somewhere in the C13 – C14 region.
Figure 18 – Advantages of polar embedded phases
Here the analyte polar functional group is shown interacting with the amide ‘spacer’ – this is analogous to the interaction with the silanol surface to produce alternative selectivity
A layer of water attracted to the polar embedded moiety acts to stop phase collapse.
The polar moiety interacts with a lone silanol group to shield the surface and reduce secondary interactions (peak tailing)
An alternative mode of chromatography that has gained significant momentum in recent years is Hydrophilic Interaction Chromatography (HILIC). Although it can trace its origins back to the 1970’s it wasn’t until Alpert published its mechanism and coined the name HILIC in 1990 that it started gaining wider popularity. It is often easiest to view HILIC as an extension of normal phase chromatography, where analyte retention increases with analyte polarity and decreases with the polarity of the mobile phase. The subtle difference between HILIC and normal phase is the incorporation of aqueous mobile phase systems and has lead some commentators to referring to HILIC as ‘Reverse Reversed Phase Chromatography’. This has substantial impact on LC-MS ESI as previous normal phase solvents contained such poor electrolytic properties that they were incapable of maintaining ions in solutions – a pre-requisite for high sensitivity ESI. The primary function of a HILIC stationary phase is to absorb water and as such common stationary phases are bare silica or polar bonded phases such as amino, cyano or diol. Due to the very strong interactions of these types of columns with water, the absorbed layer at the stationary phase surface will be more highly aqueous than the mobile phase. The mobile phase must always contain a minimum of 3% water in order to achieve this water enriched layer and as water is the ‘strongly eluting’ solvent in HILIC mode, gradient separations typically operate from low to higher percentages of water in an organic solvent (typically acetonitrile). Polar analytes partition out of the stationary phase into the water enriched layer by way of hydrogen bonding and dipole interactions. All above examples of stationary phases that are capable of operating at 100% aqueous will have the ability to undergo HILIC interactions, however, as water is the strong eluting solvent in HILIC but the weak solvent in reversed phase, the likelihood is that polar analytes will undergo very little partitioning from the mobile phase under initial mobile phase conditions of predominately aqueous conditions. A very useful adjunct to conventional HILIC stationary phases is the inclusion charged groups to the modifying ligand such that ionised species can undergo ion-exchange reactions in a ‘mixed mode’ type separation. Columns are commercially available in cationic and anionic varieties but the most useful has been the zwitterionic stationary phases (Figure 19). The proximity of the opposing charges has led to much weaker electrostatic interactions than usually observed and as such analytes do not suffer peak tailing.
Figure 19 Example of a
zwitterionic stationary phase