Methods for Measuring the Concentration of Silanol Groups on the Surface of Silica Gel.

 

The surface silanol groups on silica gel (-Si-O-H groups) are the sites with which solvent and/or solute molecules interact. Furthermore, it is the silanol groups that react with the chlorosilane reagents to produce silica gel derivatives. It follows that knowledge of the concentration of silanol groups on the silica surface can be very important.

 

There are a number of methods described in the literature that purport to provide values for the hydroxyl group concentration-some more useful than others. One of the methods reported, that is claimed to give more accurate results, involves the reaction of methylithium with the silica gel (dried at 120^oC) to produce methane.

 

                   

 

A known weight of dried silica is reacted with excess of methylithium and the methane evolved determined by a gas chromatographic procedure. The amount of methane produced is then converted to the number of hydroxyl groups per sq.m of the original silica. Unfortunately, this method assumes that all the water has been removed by heating to 120^oC, which as will be seen later, does not necessarily appear to be the case. If there is any strongly bound, or hydrogen bonded water remaining on the silica gel after drying, this will be measured and determined as hydroxyl groups and the results obtained by this procedure are likely to be in significant error. The accepted value for the concentration of silanol groups on the surface determined by this procedure is 9.0mmol.m^-2±1.0mmol.m^-2.

 

Another method for measuring silanol groups that has been suggested uses diborane to react with the silica to release hydrogen, which is also determined by a gas chromatographic procedure.

 

           

 

Unfortunately, this method suffers from the same disadvantages as methylithium, it would also reacts with any water present as well as hydroxyl groups. Thus, any strongly adsorbed water remaining on the silica surface, for example, by hydrogen bonding, would also be determined as hydroxyl groups.

 

The use of an organic chlorosilane, as previously discussed, however, might appear to be a far more reliable method for measuring solely the hydroxyl content of silica. For example, reacting dimethyloctylchlorosilane with silica would attach the dimethyloctyl group to the silica by a siloxane bond, viz.,

 

            

 

As the presence of water would not cause any organic moiety to be bonded to the silica, it follows, that the carbon content would be solely related to the reacted hydroxyl groups. Up to this point the argument for the use of silane reagents seems attractive. Unfortunately, the chlorosilane reagents are stearically hindered, due to the close proximity of the silanol groups on the surface, and consequently, reaction is not complete. Furthermore, the reagent itself cannot enter some of the silica gel pores due to its size and thus, is not available to all the silanol groups due to size exclusion. As a result, the use of silane regents to measure the silanol group concentration is not recommended and if such methods are employed then the values obtained will be significantly in error.

 

Infra red spectroscopy has also been used to estimate the hydroxyl groups on the silica surface.  The dried silica is formed into pellets and the IR adsorption at 3747cm^-1, 3680cm^-1 and 3535cm^-1 measured. These frequencies are considered to be characteristic of isolated hydroxyl groups, internal hydroxyl groups and vicinal hydroxyl groups respectively. After heating silica to 600^oC the adsorption bands at 3680cm^-1 and 3535cm^-1 disappear. In view of the results from the TGA experiments, however, this might indicate that the adsorption at 3747cm^-1 corresponds solely to the hydroxyl groups where adsorption at the other wavelengths might be attributed to strongly adsorbed, or hydrogen bonded, water.

 

Nuclear Magnetic resonance Spectroscopy (NMR) of ²⁹Si with cross-polarization and magic angle spinning (MAS) has been used to try to identify free and geminal silanol groups (1,10). NMR bands observed at 91,100 and 109 ppm are considered to correspond to geminal silanol groups, isolated silanol groups and siloxane bonds without hydroxyl groups respectively. Broad bands at 1 and 2.2 ppm are considered to arise from isolated and geminal hydroxyl groups. It should be pointed out that the interpretation of the source of band identities has assumed that there is no strongly bound or hydrogen bonded water on the silica after heating to 130^oC which in view of more recent work employing thermogravimetric analysis may not be true. Nevertheless, extension of this work is likely to provide more information regarding the nature of the silica surface.

 

Probably the most informative experimental procedure to help in the elucidation of the structure of the silica surface is thermogravimetric analysis (TGA). In the early days this amounted to heating the silica for known times at known temperatures and noting the loss in weight. With the advent of the programmed TGA apparatus this procedure has been simplified a great deal and the basic thermogram can now provide considerable information on the nature of the surface hydroxyl groups and adsorbed water. 

The Thermogravimetric Analysis of Silica Gel

 

Briefly, the sample is suspended from the arm of a continuously recording microbalance in a temperature-controlled furnace. The sample is heated from a defined starting temperature to a specified final temperature at a designated heating rate usually given as temperature change per unit time.  As discussed above in the modern TGA instrument both the temperature and the sample weight are continuously digitized and the data stored. Depending on the software provided with the instrument, the results can then be printed out or an appropriate graph constructed relating sample mass to temperature. To help identify the desorption of different species, derivative curves can also be produced.

 

The results obtained by Odlyha et al from a sample of Matrex 20m LC silica gel taken directly from the Perkin Elmer TGA instrument is shown in figure 27.

Figure 27.  Thermogram of Silica Gel

 

It is seen from the derivative curve that there appears to be three distinctly different desorption processes. The first takes place from about 30^oC to 130^oC; the second between about 200^oC and 450^oC and the third between about 400^oC and 900^oC. The three different desorption processes (probably those discussed by Lange) are distinct and unambiguous and are similar to those previously identified by Scott and Traiman. The total loss from the sample was about 5%w/w but it would appear from the TGA curve that the condensation to siloxyl groups  was not entirely complete even at the temperature of 900^oC.

 

Odlyha et al considered that, the curve relating mass of water lost (obtained by subtracting each data point from the initial total mass of sample) against temperature in a TGA analysis, was a type of desorption isotherm that could be described by assuming three distinct and separate desorbable species on the silica surface, each evolving water over a specific temperature range. From the two TGA curves the authors proposed the following theoretical explanation.