Supporting Evidence for Strongly Bound Water

 

Scott and Traiman also attempted to identify the source of the water that arose from each desorption process. Likewise, the particular interest of Scott and Traiman lay in the identification of the source of the water from the second desorption process over which there was some uncertainty, and even controversy. Consequently, they embarked on some further thermal analytical procedures. They heated samples of silica for 2 hours over a range of temperatures from 100^oC to 700^oC, each consecutive treatment temperature differing by 50^oC. Immediately after treatment, each sample was reacted exhaustively with dimethlyoctylchlorosilane in a dry xylene solvent. After washing and drying, the carbon content of each sample of silanated silica was determined. The purpose of this experiment was to determine whether any silanol groups were eliminated when the silica was heated between 120^oC and 400^oC. The results obtained are shown in figure 33 as a curve relating carbon content of sample to treatment temperature. It is clearly seen that there is little or no decrease in carbon content of the gel samples between 120^oC and 400^oC and the curve for the first seven points is, within experimental error, a horizontal straight line, indicating that no silanol groups had been converted to siloxane bonds. After 400^oC, however, the carbon content of the samples does begin to fall indicating that most, or all, of the strongly bound water has been lost and heating is now progressively eliminating silanol groups in the production of siloxane bonds. The results support the concept that the second desorption step is still involved in the elimination of strongly held water and that silanol groups are not beginning to be converted to siloxane bonds until the temperature is in excess of 400^oC. This experiment is, however, subject to some criticism as it has already been established that only about 58% of the silanol groups react with the silanizing reagent as a result of stearic hindrance and reagent exclusion. Nevertheless, one might expect that if any silanol groups were removed, they would be randomly distributed and not just those groups that were stearically hindered from reaction availability. Consequently, the results given in figure 10 gives significant support to the conclusions of Odlyha et al.

Figure 33. Graph of Carbon Content of Silanated Silica gel against Treatment Temperature

 

Scott and Traiman also confirmed their findings by infrared measurements taken on the derivatized silica. An IR spectrum of derivatized silica gel can be obtained by pressing a well-mixed sample of the bonded derivative into a potassium bromide disc. An example of such a spectrum is shown in figure 34.

Figure 34. The IR Spectrum Showing CH₂ and CH₃ Stretching Bands for the Dimethyloctyl Bonded Phase.

 

The peaks between 2800Å and 3000Å are the CH₂ and CH₃ stretching bands contained in the dimethyloctyl chain attached to the silica. These adsorption peaks can be used to determine the methyl/methylene content of the derivatized silica.

 

Figure 35. Graph of Area of CH₂/CH₃ Group Adsorption Peaks from Silanized Silica gel (between 2700Å and 3000Å) against Treatment Temperature

 

A spectrum was obtained for each of the derivatized samples and the peaks in each spectrum, situated between 2800Å and 3000Å, were cut out and weighed. The weight, which was proportional to the concentration of methyl/methylene groups attached to the silica and curves relating peak area to temperature are shown in figure 35.

 

It is seen that the results derived from IR spectra measurements, confirm those shown in figure 33, that were obtained by microanalysis, albeit with somewhat less precision. The degree of derivatization appears to remain constant until the silica is heated to a temperature in excess of 400^oC, which again indicates that no silanol groups are lost over this temperature range. Nevertheless, the same caveat holds regarding the complete reaction of all the silanol groups with the reagent.

 

Scott and Traiman carried out one further set of experiments in order to determine whether the water lost from silica gel between 200^oC and 400^oC was derived from strongly bound or hydrogen bonded water, or from the condensation of silanol groups. They noted that in the IR spectrum of silica gel contained in a KBr disc, the strong adsorption peaks between 3000 and 4000 wave numbers could be allocated to free water. An example of the IR spectrum of silica gel is shown in figure 36.

 

3mg of silica gel, previously equilibrated by exposure to air having 50% humidity was pressed it into a 300mg KBr disc. The disc was then heated to 400^oC in 25^oC intervals and after each heating cooled under anhydrous conditions and the spectrum taken. Unfortunately, at temperatures above 400^oC, partial fusion of the KBr occurred and satisfactory spectra could not be obtained.

 

 

Figure 36. The IR Spectrum of Silica Gel

 

The areas of the adsorption peaks between 3000 and 4000 wave numbers (which were proportional to the mass of free water present) were measured and a curve relating peak area to temperature is shown in figure 37.  

 

Figure 37. Graph of Peak Area of IR Adsorption between 3000 and 4000 Wave Numbers against Temperature to which the Silica Gel was Heated.

 

It is clearly seen that free water is continually being lost up to a temperature of 400^oC, which supports the concept of strongly bound or hydrogen bonded water persisting on the silica gel up to at least 400^oC. This also supports the conclusions of Odlyha et al based on their TGA results and their mass spectrometric identification of  water on the silica gel surface after having been heated to 180^oC