Professional Documents
Culture Documents
BDS Column Article
BDS Column Article
Technical Article
Base-Deactivated and High-Purity Silica Particles for Improved Peak Shape in High-Speed, High-Efficiency LC
mance are improved dramatically when the appropriate competing agent is used. The effect of the additive is to compete with any silanol interactions that interfere with analyte retention and peak shape. Consequently, the additive must be present in fairly high concentrations, often as much as 1% volume fraction of the mobile phase. The use of additives in such a concentration can often have a deleterious effect on the column lifetime and also on the reproducibility of the method. These difficulties provided the impetus for the development of improved chromatographic silicas that facilitate the analysis of both basic and acidic compounds without the requirement for a competing acid or base in the mobile phase to achieve acceptable peak shapes for problematic analytes.
by Dafydd Milton
The use of covalently bonded silica stationary phases in HPLC allows the analysis of a broad range of analytes. Along with rapid equilibration times and significantly improved mass transfer characteristics over liquidliquid partition chromatography, this has resulted in the highly successful advancement of HPLC as a modern-day analytical technique. However, covalently bonded silica stationary phases often have specific limitations. Many chemical properties associated with derivatized silicas used in HPLC have a strong effect on analyte interactions. These properties are specific to either the derivatized ligand itself or the remaining underivatized silanol groups on the silica surface. In particular, the number and acidity of these remaining silanol groups are significant. It is the silanol groups that are responsible for the acidbase properties of the base silica, contributing to the overall polarity of the surface even when the surface method is derivatized. Their type and acidity play an important role in determining resolution and peak shape for various classes of compounds being analyzed. Peak tailing and low efficiencies of both basic and acidic compounds can occur due to unwanted silanol interactions. The effect is most apparent with some of the earlier silicas developed for HPLC in which silanol groups are quite acidic. For these types of silicas, the observed effect on peak shape requires the mobile phase to include either a competing base such as triethylamine or a competing acid such as acetic acid. Both peak shape and column perfor-
Discussion
This article will consider two means to improving peak shapebase deactivation and the use of highly pure silica supports.
Base deactivation
The popularity of columns packed with C18 derivatized silica is due to their wide breadth of application, including nonpolar and neutral, acidic, and basic analytes. The selectivity of a given C18 phase can depend on the type of silane used and the synthetic conditions, since both of these factors will affect the density of the bonded phase on the surface. This density of the bonded phase is important since the greater the access of an analyte to the underlying silica support, the greater the opportunity for secondary interactions such as hydrogen bonding. There are approximately five silanol (SiOH) groups per nm 2 of surface on the silica, corresponding to 89 mM/m2. It is stereochemically impossible to react more than ~50% of the silanol groups even with ligands as small as trimethylsilane (C1). The surface composition of silica prior to derivatization is very important. As illustrated in Figure 1a, at most silica surfaces it is customary to have a variety of silanol groups: 1) lone silanols, 2) siloxanes, and 3) geminal silanols and vicinal silanols. The presence of these silanols in
Table 1 Typical metal impurity content of silica used for Hypersil ODS, Hypersil BDS, and Hypersil GOLD columns
Hypersil ODS Hypersil BDS Hypersil GOLD Mg 30 10 <1 Fe 240 90 <20 K 10 10 <1 Na 2000 1350 <5 Ca 30 10 <5 Ti 60 30 <1 Al 250 130 <1
Figure 1
a derivatized silica can result in unwanted silanol interactions with the analytes, which can give rise to peak tailing and changes in retention and selectivity on a typical alkyl C18 packing, such as: SiOHNH2-R (hydrogen bonding with base) SiOHO=RCOH (hydrogen bonding with acid) SiO+NH3-R (ion exchange with base). Using a base deactivation process, the surface of the silica is made much more homogeneous, so that all silanols are of the same type (vicinal silanols), as illustrated in Figure 1b. The resultant silica surface is more uniform and ready for surface derivatization. Special care is taken to ensure a high density of coverage followed by thorough end-capping in order to further reduce the possibility of any silanol interactions. As a consequence of the base deactivation process, the silanols that are still present after surface derivatization become much more friendly toward basic and acidic compounds, and the packing material becomes an excellent choice to develop highly reproducible methods. Silanols are less acidic and less likely to be available for ion exchange interaction with ionized basic analytes, and are also less likely to hydrogen bond with polar analyte. With these reduced silanol interactions, Thermo Scientific Hypersil BDS columns (Thermo Fisher Scientific, Runcorn, U.K.) are well suited for the analysis of a wide range of analytes, including both acids and bases, with peak shape and column performance significantly improved.
ering the separation of polar and basic compounds. Older, less pure silica supports (often termed Type A silicas) contain a greater number of metallic impurities. These metallic impurities can complex with chelating solutes, resulting in asymmetrical or tailing peaks. In extreme cases, these interactions may be strong enough to result in complete retention of the solute; consequently, elution of the solute does not occur. The presence of certain metallic impurities (particularly aluminum) in silica can also activate the silanols so that they become highly acidic, which can also lead to peak tailing for basic solutes. The use of a highly pure silica particle is therefore desirable for the separation of basic or polar compounds. Table 1 compares typical metallic impurities for the base silica used in three generations of Thermo Scientific chromatography columns. As can be seen from the data shown in Table 1, the base deactivation process used in Hypersil BDS columns also has the effect of reducing metallic impurities. The most dramatic reduction in metallic impurities, however, comes with the use of highly pure silica support as used in Hypersil GOLD columns; these columns can be expected to give the best peak shape. This assumption has been tested using a method based on the USP method for the analysis of diazepam tablets. The resulting peak shape is shown in Figure 2. Not surprisingly, as the chromatography is transferred from an older silica (Hypersil ODS) to basedeactivated (Hypersil BDS) to highly pure silica (Hypersil GOLD), there are improvements in peak shape, efficiency, and sensitivity.
High-purity silica
Further improvements to chromatographic peak shape can be achieved by utilizing a highly pure silica support for the column packing material. The purity of the silica support is of particular importance when consid-
Figure 2 Comparison of peak shape for Hypersil ODS, Hypersil BDS, and Hypersil GOLD. Column: 150 4.6 mm, 5 m Hypersil ODS, Hypersil BDS, and Hypersil GOLD. Mobile phase: methanol/water 65%/35% v/v. Flow rate: 1.0 mL/min. Detection: UV at 254 nm.
stant between the original and new method. 2. Small particle-based methods are most often transferred to smaller-volume columns; thus the same injection volume will take up a larger proportion of the new column, possibly leading to column saturation or band broadening. Figure 3 Column backpressure as a funcIt is therefore important tion of flow rate for 2.4-, 3-, and 5-m particle to scale down the injec- packed columns. tion volume to match the change in column volume. 3. Transfer of the gradient to maintain an equivalent separation requires calculating the number of column volumes of mobile phase in each segment (time interval) of the gradient in the original method to ensure that the new calculated gradient takes place over the same number of column volumes, for the new column. One advantage of using 2.4-m particle size columns is that highspeed, high-efficiency separations are achievable using conventional HPLC systems, even for narrow columns. For example, the backpressure for a 100 2.1 mm i.d. column packed with 2.4-m particles is shown in Figure 3. The optimum flow rate for this column is 0.40.5 mL/min, which will give a backpressure within the limits for a conventional HPLC system. Figure 4 illustrates method transfer to 2.4-m Hypersil BDS particles using omeprazole as an example. For the separation using the column packed with 2.4-m particles, omeprazole elutes in 3.3 min, compared with 7.5 min for the column packed with 5-m particles, and the separation efficiency is improved by 20%. With 1.9-m high-purity silica particles, even greater time savings can be realized. This is illustrated in Figure 5, which shows the chromatographic profiles obtained for the original HPLC method for
is available with a particle size of 2.4 m and Hypersil GOLD with 1.9 m. The smaller the particle size, the higher the efficiency. Sub-3-m particles give higher efficiency than 5-m particles, and this efficiency is delivered over a greater range of optimum linear velocity. This makes it possible to operate at higher flow rates without losing performance. Because shorter columns packed with 1.9- or 2.4-m particles provide equivalent efficiency to longer columns packed with 5-m particles, faster analysis and solvent savings for the chromatographer become a reality. When transferring methods to columns packed with sub-3-m particles, the following should be considered: 1. To maintain an equivalent separation when transferring a method, it is important to keep the reduced linear velocity con-
Figure 4 Columns packed with 2.4-m particles give faster, more efficient chromatography than columns packed with 5-m particles.
Figure 5 Chromatograms obtained with the original HPLC and geometrically scaled UHPLC methods. Comparison of run time, USP resolution, and pressure drop across the column.
ibuprofen and impurities and the geometrically scaled ultrahighperformance liquid chromatography (UHPLC) methods on smaller columns packed with 1.9-m particles. When the method is transferred to the 50 2.1 mm column, the resolution of peaks 5 and 6 is maintained, while analysis time is reduced by approximately sevenfold. The resolution between peaks 5 and 6 increases from 1.6 with the HPLC method to 2.8 with the UHPLC method on the 100 2.1 mm column. This also highlights the flexibility of using 1.9-m particles for high-speed or high-resolution separations.
Second, because peak widths are narrower with fast chromatography, the detector time constant and sampling rate need to be selected carefully. If fast gradients are being used, then a pump with a low dwell volume is desirable to transfer the gradient to the column quickly. These system considerations are addressed when using modern UHPLC instruments, such as the Thermo Scientific Accela High Speed LC.
Summary
Base-deactivated silica and highly pure silica can be used to obtain improvements in chromatographic peak shape. This facilitates the use of faster, more efficient chromatography using smaller particles.
Dr. Milton is Product Manager, Thermo Fisher Scientific, 112 Chadwick Rd., Runcorn WA7 1PR, U.K.; tel.: +44 1928 588089; e-mail: [email protected].
System considerations
There are some considerations when using short columns packed with 1.9-m or 2.4-m particles in order to maximize their performance. First, to limit dispersion, which can lead to peak broadening, the system volume (connecting tubing i.d. and length, injection volume, and UV detector flow cell volume) should be minimized.