Do all silica flash columns provide the same results?

Silica flash columns are the most frequently used to purify reaction mixtures. Prepacked columns are available from a vast number of vendors as is the silica itself for those wishing to pack their own.

The silica used for chromatography is made using a synthetic process. Depending on that process and the ratio and purity of the reagents used, the final silica product can vary significantly, especially in surface area and porosity.

The most commonly used silica has an irregular or granular shape, Figure 1. However, when initially synthesized, this silica starts out as millimeter-sized spherical or spheroidal particles which is then ground and sieved to achieve the proper particle size distribution (no silicas are monodisperse). Typical flash chromatography particle size distribution is 40-63 µm with a median of ~50 µm.

Figure 1. SEM image of granular silica.

This media typically has a surface area around 500 m²/g, a pore volume of 0.75 mL/g, and a pore diameter of 60Å (they are all related parameters). Pore diameter is calculated from the measured surface area and pore volume, Equation 1.

Equation 1.

Granular silica has a mix of larger and smaller particles which can pose solvent and sample flow issues when packed into a column.

Though granular silica is still the most commonly used, spherical silica is now being packed into commercially available columns at an increasing rate. Spherical silica is all that we pack into our Biotage® Sfär columns.

Spherical silica also varies in surface area, pore volume, pore diameter, and particle size depending on who manufactured it. These silicas can also vary surface chemistry.

How? Well silica’s surface chemistry is made of silanols (SiOH). There are different types of silanols including single, geminal, vicinal (bridged)[1], Figure 2. It is the ratio of these three silanol types that influences the chromatographic separation results (selectivity).

Figure 2. Various silanol types comprising chromatographic silica.

So, how do these differences manifest themselves chromatographically? Well, the geminal silanol is the most polar and therefore is most influential in polar compound retention while the vicinal silanol is the least polar. More vicinal silanols means less polar compound retention.

As a chemist using silica columns for flash chromatography, it is not important to know the ratio of these silanols, just that the three types of silanols exist and can influence your purification. Within a flash column manufacturer’s product line, only one type or source of silica is used, though different particle sizes may be available.

Working at Biotage, I believe that understanding how our silica columns compare to our competition is important and so I decided to do some testing. I compared similar sized flash columns from two well-known vendors, which I will call brand G and brand P, to a Biotage® Sfär HC column. All column brands used spherical silica with a published particle size between 20 and 30 µm. From these vendors' literature I was able to find the other physical parameters – surface area, column volume, and pore diameter (few publish the pore volume as pore diameter seems to be more relevant), Table 1.

Table 1. Silica physical property comparison

As can be seen, the physical properties vary considerably. Unfortunately, flash column vendors either do not know (highly likely) or do not publish data on the chemical make-up (silanol types and ratio). Again, the actual values are not something you need to know, just that each silica brand will likely behave differently when the same sample is purified.

My experimentation included an amide reaction mixture and a lipid mixture consisting of α-tocopherol and cholesterol (1:1), both of which were able to highlight differences in the silicas.

Reaction mixture

I used a 7-50% EtOAc in hexane gradient in the same number of column volumes (10) using the vendor’s suggested flow rates. Sample size was equal (415 mg) and dry loaded (1:4 ratio of sample to silica). The results showed some noticeable differences, Figure 3.

Figure 3. Reaction mixture flash chromatography comparison showing performance differences. Top – Biotage® Sfär HC. Middle - Brand G. Bottom - Brand P.

There are clear differences in the chromatographic results. While all three columns show a trailing shoulder on the product peak, brands G and P also show a shoulder on the front of the product peak and reduced resolution between the product peak (green) and the major trailing impurity (yellow). These are indicated by colored circles.

Reversed-phase purification of the product peak shows how much by-product contamination was present, Figure 4.

Figure 4. Reversed-phase chromatography of isolated product peak from normal-phase chromatography shows different by-product levels. Top - Biotage Sfär analysis. Middle brand G analysis. Bottom brand P analysis.

While all show a large by-product 1 contamination, as expected, it is by-product 2 that highlights the differences in silicas. By-product 2 is the shoulder on the front of the product peak in the normal-phase purification with brands G and P in Figure 3. The reversed-phase data shows a large amount of this contaminant coming from brand G, and far less from brand P. Very little by-product 2 was found in the Biotage® Sfär HC product fraction. What this data tells me is that the silica used by brand G does not have the right chemical and/or physical parameters to remove this by-product. The data also shows the Biotage silica was able to remove by-product 2 from the product most effectively.

Lipid mixture

Fatty compounds such as cholesterol and tocopherol often pose separation challenges for chemists. These two compounds, however, are easily separated by normal-phase chromatography. What I leaned however, is that impurities or degradants are not so easy to resolve, Figure 5.

For this comparison, a hexane/MTBE gradient of 5-60% in 10 CV was used. Vendor-recommended flow rates were again used and sample size was identical (200 mg).

Figure 5. Chromatographic comparison of a 1:1 mixture of α-tocopherol and cholesterol. Top - Biotage® Sfär HC. Middle - brand G. Bottom - brand P.

This comparison again shows the impact that silicas from different sources have on chromatography results. While the Biotage column does not separate impurity 1 from the peak preceding it as well as the other two column brands, it does the best job of separating a tocopherol isomer from α-tocopherol and impurity 2 from cholesterol. Unlike with the reaction mixture purification, brand G provided a better separation than brand P, which does not resolve either the tocopherol isomer or impurity 2.

These two examples help highlight differences in chromatographic silicas. Hopefully, this data has provided you with some awareness that not all silica is the same and variability between brands will be seen. Make your column buying decision based on your scientific results.

[1] Neue, Uwe D. HPLC Columns Theory, Technology, and Practice; Wiley-VCH: New York, 1997; p. 109.