Application of Theory

 

The above theory was applied to the data obtained by Odlyha et al from the TGA curve shown in figure 27 using the above equation. Employing a simple iterative curve fitting computer program, values for the constants a_r, A_r, T_x(r) and T_S(r) were identified for each desorption process that gave the minimum error between the calculated and measured values given by the TGA analysis. To help in the mathematical iteration process, the differential form of the TGA curve provided useful reference temperatures. Optimum values for T_x(r) and T_S(r) were obtained by searching over a range of values that embraced those suggested by the minima shown in the differential TGA curves. The number of minima also helped to decide the number of desorption processes to consider. Using the values for the constants identified from the iterative procedure the theoretical TGA curve was constructed and is shown in figure 29.

 

Figure 29. Theoretical TGA Curve (Including Experimental Data Points

 

Included in the figure are the experimental points from the TGA experiment. It is seen from figure 29 that excellent agreement was obtained between the theoretical curve and the experimental points. It must be emphasized that the curve was obtained on the assumption that there were three distinct, and separate sources from which water was desorbed and that the distribution of energies of adsorption for each group was random about a mean energy. From the data obtained by the iterative curve fitting process the individual energy distribution of the three sources could be calculated and the results obtained are shown in figure 30.

 

Figure 30. Graph of Site Distribution against Temperature

 

It is seen that the three sources of water desorption are unambiguously identified. The first source, probably that of physically adsorbed water, is mostly lost below 200^oC whereas the third source, resulting from the condensation of silanol groups to siloxane bonds is not complete until a temperature in excess of 900^oC is reached.

 

The three different sources of weight loss are unambiguously confirmed but the nature of the material lost between 200^oC and 400^oC needs to be identified, which required further thermoanalytical examination.

 

Odlyha et al heated a sample of the silica gel to 180^oC for an hour to remove all physically bound water and then under carefully controlled anhydrous conditions reacted the silica with excess dimethyloctylchlorosilane.

 

The excess reagent was removed by reaction with dry phenol and the reacted silica removed by filtration. The reaction liquid was examined by high-resolution mass spectroscopy using FAB ionization. The spectrum obtained is shown in figure 31 together with that obtained from a blank run with no silica present. 

Figure 31. The Mass Spectra of the Contents of the Reaction Liquid and those of the Blank

The presence of a significant amount of the ion mass 359.3 which is the molecular weight + 1 of the ether confirmed the presence of dimethyloctyl silyl ether, ((CH₃)₂C₈H₁₇Si)₂O, which could only have formed if there were significant amounts of water remaining on the silica after heating to 180^oC.

 

                    2(CH₃)₂C₈H₁₇SiCl + H₂O = ((CH₃)₂C₈H₁₇Si)₂O + 2HCl

 

The blank, which was treated in exactly the same manner, shows no significant quantity of the ether even at an amplification of 10. It is obvious that, despite the results from the TGA curves heating to180^oC over an extended period of time does not remove all the molecular water from the silica gel. Thus, the second source of desorbable water that is lost between 200^oC and 400^oC appears to be strongly held (probably hydrogen bonded) water. The presence of water on the silica surface after being heated to 180^oC gives some concern as to the validity of the accepted value for the number of silanol groups per square meter of surface as determined by the methyl lithium method. It would appear that not all the water is removed from the silica surface, even on heating to 180^oC, thus, some of the methane collected from the methyl lithium reaction is likely to have come from strongly held water as well the silanol groups. If this is so, a false high value for the silanol content will be obtained. The amount of water lost during the three different desorption processes can also be calculated and the desorption curves are shown in figure 32.

Figure 32. Graph of calculated Weight Lost against Temperature for Each Desorption Stage

 

From the data given in figure 32, Odlyha et al (7) calculated the number of silanols per square meter of silica surface and estimated the error that might be involved in the methyl lithium method of measurement. The data they used was as follows,

 

Calculation of the True Silanol Concentration from the TGA Results

 

Weight of Silica taken                                                     21.0 mg

Weight of water lost in third desorption stage                     0.65

Weight of water lost in second desorption stage                  0.086

Surface area of silica gel                                               250m²g^-1

Now the concentration of silanol groups on the surface (C) in mmol.m^-2 of surface can be calculated from the following equation:-

 

 

where (g) is the mass of water lost during silanol condensation,

          (m) is the mass of silica gel sample,

          (M_(OH)) is the molecular weight of the hydroxyl group,

   and  (A) is the surface area per gram of the silica gel.

 

In a similar way, the concentration of silanol groups (c) equivalent to the mass of strongly bound water lost in the second description stage can be calculated from,

 

 

where (w) is the mass of water lost during silanol condensation,

          (m) is the mass of silica gel sample,

          (M_(H2O)) is the molecular weight of the hydroxyl group,

   and  (A) is the surface area per gram of the silica gel.

 

Thus, using the above eqations

 

Concentration of Silanol Groups on the surface                  7.28m.mol.m^-2

 

Concentration of Silanol Groups on the surface

equivalent to hydrogen bonded water lost                            1.93m.mol.m^-2

 

Therefore, silanol content if determined by the

methyl lithium procedure would be                                      9.21m.mol.m^-2

 

The actual value for the silanol concentration as determined by the methyl lithium procedure on silica gel preheated to 200^oC has been reported to be (8) 9.0m.mol.m^-2±1.0 m.mol.m^-2. It is seen that the calculated value of 9.21m.mol.m^-2 agrees well with the experimentally determined value. The lower value for the silanol content of the silica gel obtained in this way is important to bear in mind when calculating the necessary reagent concentration in any silica surface chemistry.