Sol-Gel Chemistry

SOL-GEL CHEMISTRY AND TECHNOLOGY

A process that has, in the past years, gained much notoriety in the glass and ceramic fields is the sol-gel reaction. This chemistry produces a variety of inorganic networks from silicon or metal alkoxide monomer precursors. Although first discovered in the late 1800s and extensively studied since the early 1930s, a renewed interest 11-12 surfaced in the early 1970s when monolithic inorganic gels were formed at low temperatures and converted to glasses without a high temperature melting process.13 Through this process, homogeneous inorganic oxide materials with desirable properties of hardness, optical transparency, chemical durability, tailored porosity, and thermal resistance, can be produced at room temperatures, as opposed to the much higher melting temperatures required in the production of conventional inorganic glasses. 13-15 The specific uses of these sol-gel produced glasses and ceramics are derived from the various material shapes generated in the gel state, i.e., monoliths, films, fibers, and monosized powders. Many specific applications include optics, protective and porous films, optical coatings, window insulators, dielectric and electronic coatings, high temperature superconductors, reinforcement fibers, fillers, and catalysts. 15

The sol-gel process, as the name implies, involves the evolution of inorganic networks through the formation of a colloidal suspension (sol) and gelation of the sol to form a network in a continuous liquid phase (gel). 12 The precursors for synthesizing these colloids consist of a metal or metalloid element surrounded by various reactive ligands. Metal alkoxides are most popular because they react readily with water. The most widely used metal alkoxides are the alkoxysilanes, such as tetramethoxysilane (TMOS) and tetraethoxysilane (TEOS). However, other alkoxides such as aluminates, titanates, and borates are also commonly used in the sol-gel process, often mixed with TEOS.

At the functional group level, three reactions are generally used to describe the sol-gel process: hydrolysis, alcohol condensation, and water condensation. This general reaction scheme can be seen in Figure 4. However, the characteristics and properties of a particular sol-gel inorganic network are related to a number of factors that affect the rate of hydrolysis and condensation reactions, such as, pH, temperature and time of reaction, reagent concentrations, catalyst nature and concentration, H2O/Si molar ratio (R), aging temperature and time, and drying. 16-17 Of the factors listed above, pH, nature and concentration of catalyst, H2O/Si molar ratio (R), and temperature have been identified as most important. Thus, by controlling these factors, it is possible to vary the structure and properties of the sol-gel-derived inorganic network over wide ranges. 17 For example, Sakka et al. observed that the hydrolysis of TEOS utilizing R values of 1-2 and 0.01 M HCl as a catalyst yields a viscous, spinnable solution. It was further shown, that these solutions exhibited a strong concentration dependence on the intrinsic viscosity and a power law dependence of the reduced viscosity on the number average molecular weight:

[n] = k(Mn)a(1)

Values for a ranged from 0.5 to 1.0, which indicates a linear or lightly branched molecule or chain.

In contrast, when R values greater than two (2) and/or base catalysts were utilized, solutions were produced which were not spinnable at equivalent viscosities. 18-19 Values of a in eq. 1 ranged from 0.1 to 0.5, indicating spherical or disk shaped particles. These results are consistent with the structures which emerge under the conditions employed by the Ströber process20 for preparing SiO2 powders. It was further shown that with hydrolysis under basic conditions and R values ranging from seven (7) to twenty-five (25), monodisperse, spherical particles could be produced.

Generally speaking, the hydrolysis reaction (Eq. 2), through the addition of water, replaces alkoxide groups (OR) with hydroxyl groups (OH). Subsequent condensation reactions (Eqs. 3 and 4) involving the silanol groups (Si-OH) produce siloxane bonds (Si-O-Si) plus the by-products water or alcohol. Under most conditions, condensation commences before hydrolysis is complete. However, conditions such as, pH, H2O/Si molar ratio (R), and catalyst can force completion of hydrolysis before condensation begins. 15 Additionally, because water and alkoxides are immiscible, a mutual solvent such as an alcohol is utilized. With the presence of this homogenizing agent, alcohol, hydrolysis is facilitated 16 due to the miscibility of the alkoxide and water. As the number of siloxane bonds increases, the individual molecules are bridged and jointly aggregate in the sol. When the sol particles aggregate, or inter-knit into a network, a gel is formed. Upon drying, trapped volatiles (water, alcohol, etc.) are driven off and the network shrinks as further condensation can occur. It should be emphasized, however, that the addition of solvents and certain reaction conditions may promote esterification and depolymerization reactions according to the reverse of equations (2), (3), and (4). 15, 17

In the following sections, specific factors that influence the hydrolysis and condensation reactions of the sol-gel process will be discussed. It has been established that certain reaction parameters are more important than others. 11-12, 14-24 Hereinafter, we will primarily focus on the following influences: pH, nature and concentration of catalyst, and H2O/Si molar ratio (R).

pH

Iler 25 divides this polymerization process into three pH domains: < pH 2, pH 2-7, and > pH 7. However, regardless of pH, hydrolysis occurs by the nucleophilic attack of the oxygen contained in water on the silicon atom as evidenced by the reaction of isotopically labeled water with TEOS that produces only unlabelled alcohol in both acid- and base-catalyzed systems: 17

It can be seen in Figure 5 how pH affects the hydrolysis rate.

Figure 5. pH Rate Profile for Hydrolysis in Aqueous Solution.

Although hydrolysis can occur without addition of an external catalyst, it is most rapid and complete when they are employed. Mineral acids (HCl) and ammonia are most generally used, however, other catalysts are acetic acid, KOH, amines, KF, and HF. 13 Additionally, it has been observed that the rate and extent of the hydrolysis reaction is most influenced by the strength and concentration of the acid- or base catalyst. 26

Aelion et al. found that all strong acids behave similarly, whereas weaker acids require longer reaction times to achieve the same extent of reaction. 26 From a plot of the logarithm of the hydrolysis rate constant versus acid concentration, a slope of one was obtained. They concluded that the reaction was first-order in acid concentration.

Under basic conditions, the hydrolysis reaction was found to be first-order in base concentration. However, as the TEOS concentration was increased the reaction deviated from a simple first-order to a more complicated second-order reaction. With weaker bases such as, ammonium hydroxide and pyridine, measurable speeds of reaction were produced only if large concentrations were present. Therefore, compared to acidic conditions, base hydrolysis kinetics is more strongly affected by the nature of the solvent. 26

Acid-Catalyzed Mechanism

Under acidic conditions, it is likely that an alkoxide group is protonated in a rapid first step. Electron density is withdrawn from the silicon atom, making it more electrophilic and thus more susceptible to attack from water. This results in the formation of a penta-coordinate transition state with significant SN2-type character. 13 The transition state decays by displacement of an alcohol and inversion of the silicon tetrahedron, as seen in Figure 6. Figure 6. Acid-Catalyzed Hydrolysis

Based-Catalyzed Mechanism

Base-catalyzed hydrolysis of silicon alkoxides proceeds much more slowly than acid-catalyzed hydrolysis at an equivalent catalyst concentration. 26 Basic alkoxide oxygens tend to repel the nucleophile, -OH. However, once an initial hydrolysis has occurred, following reactions proceed stepwise, with each subsequent alkoxide group more easily removed from the monomer then the previous one. 27 Therefore, more highly hydrolyzed silicones are more prone to attack. Additionally, hydrolysis of the forming polymer is more sterically hindered than the hydrolysis of a monomer. Although hydrolysis in alkaline environments is slow, it still tends to be complete and irreversible. 15

Thus, under basic conditions, it is likely that water dissociates to produce hydroxyl anions in a rapid first step. The hydroxyl anion then attacks the silicon atom. Again, an SN2-type mechanism has been proposed in which the -OH displaces -OR with inversion of the silicon tetrahedron. This is seen in Figure 7.

Figure 7. Base-Catalyzed Hydrolysis

H2O/Si Molar Ratio (R)

As stated previously, the hydrolysis reaction has been performed with R values ranging from less than 1 to over 50, depending on the desired polysilicate product. From equation 2, an increased value of R is expected to promote the hydrolysis reaction. Aelion et al. found the acid-catalyzed hydrolysis of TEOS to be first-order in water concentration; however, they observed an apparent zero-order dependence of the water concentration under basic conditions. 26 This is probably due to the production of monomers by siloxane bond hydrolysis and redistribution reactions (i.e., reverse reactions 3 and 4). Nonetheless, the most obvious effect of the increased value of R is the acceleration of the hydrolysis reaction. Additionally, higher values of R caused more complete hydrolysis of monomers before significant condensation occurs. Differing extents of monomer hydrolysis should affect the relative rates of the alcohol- or water-producing condensation reactions. Generally, with understoichiometric additions of water (R << 2), the alcohol producing-condensation mechanism is favored, whereas, the water-forming condensation reaction is favored when R ³ 2.28.

Although increased values of R generally promote hydrolysis, when R is increased while maintaining a constant solvent: silicate ratio, the silicate concentration is reduced. This in turn reduces the hydrolysis and condensation rates, resulting in longer gel times. This effect is evident in Figure 8 which shows gel times for acid-catalyzed TEOS systems as a function of R and the initial alcohol: TEOS molar ratio. 29

Finally, since water is the by-product of the condensation reaction (Eq, 3), large values of R promote siloxane bond hydrolysis (reverse of Eq. 3).

Figure 8. Gel Times as a function of H2O: TEOS Ratio, R.

B. Condensation

pH

Polymerization to form siloxane bonds occurs by either an alcohol-producing or a water-producing condensation reaction. It has been shown by Engelhardt et al. that a typical sequence of condensation products is monomer, dimer, linear trimer, cyclic trimer, cyclic tetramer, and higher order rings. 30 This sequence of condensation requires both depolymerization (ring opening) and the availability of monomers which are in solution equilibrium with the oligomeric species and/or are generated by depolymerization (reverse of Eqs. 3 and 4). 13

The rate of these ring opening polymerizations and monomer addition reactions is dependent upon the environmental pH. In polymerizations below pH 2, the condensation rates are proportional to the [H+] concentration. Because the solubility (see Figure 9) of silica is quite low below pH 2, formation and aggregation of primary silica particles occur together and ripening (i.e., growth of a network) contributes little to growth after particles exceed 2nm in diameter. Thus, developing gel networks are composed of exceedingly small primary particles. 13

It is generally agreed that between pH 2 and pH 6 condensation rates are proportional to [-OH] concentrations. Condensation preferentially occurs between more highly condensed species and those less highly condensed and somewhat neutral. This suggests that the rate of dimerization is low, however, once dimers form, they react preferentially with monomers to form trimers, which in turn react with monomers to form tetrameters. Cyclization occurs because of the proximity of the chain ends and the substantial depletion of the monomer population. Further growth occurs by addition of lower molecular weight species to more highly condensed species and aggregation of the condensed species to form chains and networks. The solubility of silica in this pH range is again low and particle growth stops when the particles reach 2-4 nm in diameter. 15

Above pH 7, polymerization occurs the same as in the pH 2 to pH 6 range. However, in this pH range, condensed species are ionized and therefore, mutually repulsive. Growth occurs primarily through the addition of monomers to the more highly condensed particles rather then by particle aggregation. Due to the greater solubility of silica and the greater size dependence of solubility above pH 7, particles grow in size and decrease in number as highly soluble small particles dissolve and reprecipitate on larger, less soluble particles. Growth stops when the difference in solubility between the smallest and largest particles becomes indistinguishable. This process is referred to as Ostwald ripening. Particle size, is therefore, mainly temperature dependent, in that higher temperatures produce larger particles. Additionally, in this pH range, the growth rate depends upon the particle size distribution. 13

Figure 9. Dissolution rate and relative gel time as a function of pH.

Nature and Concentration of Catalyst

As with hydrolysis, condensation can proceed without catalyst, however, their use in organosiloxanes is highly helpful. Furthermore, the same type catalysts are employed: generally those compounds which exhibit acidic or basic characteristics.

It has been shown that condensation reactions are acid and base specific. 31 In addition, Iler has shown that under more basic conditions, gel times are observed to increase. 25 Condensation reactions continue to proceed, however, gelation does not occur. Again, catalysts which dictate a specific pH, can and do drive the type of silica particle produced as seen in the previous discussion on pH.

It is generally believed that the acid-catalyzed condensation mechanism involves a protonated silanol species. Protonation of the silanol makes the silicon more electrophilic and thus susceptible to nucleophilic attack. The most basic silanol species (silanols contained in monomers or weakly branched oligomers) are the most likely to be protonated. Therefore, condensation reactions may occur preferentially between neutral species and protonated silanols situated on monomers, end groups of chains, etc. 13

Base-Catalyzed Mechanism

The most widely accepted mechanism for the base-catalyzed condensation reaction involves the attack of a nucleophilic deprotonated silanol on a neutral silicic acid: 25

Figure 10: Nucleophilic Attack to Form Siloxane Bond Furthermore, it is generally believed that the base-catalyzed condensation mechanism involves penta- or hexa-coordinated silicon intermediates or transition states , similar to that of a SN2 type mechanism. 13

C. Summary

According to Iler, Sol-gel polymerization occurs in three stages:
1. Polymerization of monomers to form particles
2. Growth of particles
3. Linking of particles into chains, then networks that extend throughout the liquid medium, thickening into a gel.

Within the context of these stages, many factors affect the resulting silica network, such as, pH, temperature and time of reaction, reagent concentrations, catalyst nature and concentration, H2O/Si molar ratio (R), aging temperature and time. However, it can generally be said that sol-gel derived silicon oxide networks, under acid-catalyzed conditions, yield primarily linear or randomly branched polymers which entangle and form additional branches resulting in gelation. On the other hand, silicon oxide networks derived under base-catalyzed conditions yield more highly branched clusters which do not interpenetrate prior to gelation and thus behave as discrete clusters (See Figure 11).

Figure 11: Summary of Acid/Base Sol-Gel Conditions

1. Butler, G.B.; O’Driscoll, K.F.; Wilkes, G.L. JMS-Rev. Macromol. Chem. Phys. 1994, C34(3), 325-373.
2. NafionÒ is a registered trademark of E.I. DuPont de Nemours & Co.
3. Robertson, M.A.F. Ph.D. Thesis, University of Calgary, 1994.
4. Yeager, H.J.; Eisenberg, A. in: Perfluorinated Ionomer Membranes; eds. A. Eisenberg and H.L Yeager, ACS Symp. Ser. No.180, (American Chemical Society: Washington, DC, 1982) pp. 1-6, 41-63.
5. Mauritz, K.A.; Storey, R.F.; Jones, C.K. in: Multiphase Polymer Materials: Blends, Ionomers, and Interpenetrating Networks; eds. L.A. Utracki and R.A. Weiss, ACS Symp. Ser. No. 395; (American Chemical Society: Washington, DC, 1989) p. 401-417.
6. Mauritz, K.A.; Hora, C.J.; Hopfinger, A.J. in: Ions in Polymers; ed. A. Eisenberg, ACS Advances in Chemistry Ser. No. 187 (American Chemical Society: Washington, DC, 1980) pp. 124-154.
7. Mauritz, K.A. JMS-Rev. Macromol. Chem. Phys. 1988, C28(1), 65-98.
8. A. Eisenberg and M. King, Ion-Containing Polymers: Physical Properties and Structure, vol. 2. (Academic Press: New York, 1977) pp. 163-169.
9. Sondheimer, S.J.; Bunce, N.J.; Fyfe, C.A. JMS-Rev. Macromol. Chem. Phys. 1986, C26(3), 353-413.
10. Brookman, P.J.; Nicholson, J.W. in: Developments in Ionic Polymers, vol. 2; eds. A.D. Wilson and H.J. Prosser, (Elsevier Applied Science Publishers: London, 1986) pp. 269-283.
11. Hench, L.L.; West, J.K. Chem. Rev. 1990, 90, 35-40.
12. Lev, O. et al. Analytical Chemistry. 1995, 67(1), 22A-30A.
13. C.J. Brinker and G.W. Scherer, Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing (Academic Press, Inc.: New York, 1990).
14. Brinker, C.J.; Scherer, G.W. J. Non-Crystalline Solids. 1985, 70, 301-322.
15. K.D. Keefer, in: Silicon Based Polymer Science: A Comprehensive Resource; eds. J.M. Zeigler and F.W.G. Fearon, ACS Advances in Chemistry Ser. No. 224, (American Chemical Society: Washington, DC, 1990) pp. 227-240.
16. M. Prassas and L.L. Hench, in: Ultrastructure Processing of Ceramics, Glasses, and Composites; eds. L.L Hench and D.R. Ulrich, (John Wiley & Sons: New York, 1984) pp. 100-125.
17. Brinker, C.J. J. Non-Crystalline Solids. 1988, 100, 31-50.
18. Sakka, S.; Kamiya, K. J. Non-Crystalline Solids. 1982, 48, 31.
19. S. Sakka, in: Better Ceramics Through Chemistry, eds. C.J. Brinker, D.E. Clark, and D.R. Ullrich, (Elsevier-North Holland: New York, 1984) p. 91.
20. Ströber, W.; Fink, A.; Bohn, E. J. Colloid Interface. Sci., 1968, 26, 62-69.
21. Jiang, C.Y.; Mark, J.E. Makromol. Chem. 1984, 185, 2609-2613.
22. Zerda, T.W.; Artaki, I.; Jonas, J. J. Non-Crystalline Solids. 1986, 81, 374-376.
23. Assink, R.A.; Kay, B.D. J. Non-Crystalline Solids. 1988, 107, 35-40.
24. Brinker, C.J. et al. J. Non-Crystalline Solids. 1982, 48, 47-64.
25. R.K. Iler, The Chemistry of Silica (Wiley: New York, 1979).
26. Aelion, R.; Loebel, A.; Eirich, F. J. Am. Chem. Soc., 1950, 72, 5705-5712.
27. Kelts, L.W.; Effinger, N.J.; Melpolder, S.M. J. Non-Crystalline Solids. 1986, 83, 353-374.
28. Assink, R.A.; Kay, B.D. J. Non-Crystalline Solids. 1988, 99, 359.
29. Klein, L.C. Ann. Rev. Mater. Sci., 1985, 15, 227-248.
30. Engelhardt, V.Q. et al. Anorg. Allg. Chem., 1977, 418, 43.
31. E.R Pohl and F.D. Osterholtz in: Molecular Characterization of Composite Interfaces; eds. H. Ishida and G. Kumar, (Plenum: New York, 1985) p. 157.
32. Mauritz, K.A.; Warren, R.M. Macromolecules 1989, 22, 1730.
33. Mauritz, K.A.; Stefanithis, I.D. Macromolecules 1990, 23, 1380.
34. Stefanithis, I.D.; Mauritz, K.A Macromolecules 1990, 23, 2397.
35. Stefanithis, I.D.; Mauritz, K.A.; Wilkes, G.L.; Huang, Hao-Hsin. Am. Chem. Soc. Polym. Prepr. 1991, 32(1), 236.
36. Davis, S.V.; Mauritz, K.A. Am. Chem. Soc. Polym. Prepr. 1992, 33(2), 363.
37. Davis, S.V.; Mauritz, K.A. Am. Chem. Soc. Polym. Prepr. 1993, 34(1), 608.
38. Bordayo, C.; et al. Am. Chem. Soc. Polym. Mat. Sci. Eng. Prepr. 1994, 70, 230.
39. Davis, S.V.; Mauritz, K.A.; Moore, R.B. Am. Chem. Soc. Polym. Prepr. 1994, 35(1), 419.
40. Deng, Q.; Mauritz, K.A.; Moore, R.B. Am. Chem. Soc. Polym. Prepr. 1995, 36(1), 433.
41. Brinker, C.J.; Scherer, G.W. J. Non-Crystalline Solids. 1985, 70, 304-309.
42. Davis, S.V. Ph.D. Thesis, University of Southern Mississippi, 1995.

	Go back to the site index page
	Go back to the current research page