|Monthly Tech-Tip |
A transcript of a presentation at the 3rd Whitewares conference at Alfred University in the spring of 2000 by Richard Eppler.
Formulating a glaze for high gloss is a challenging assignment, particularly when cost is a consideration, as it almost always is these days. As frits are among the most expensive glaze ingredients, the easy solution of an all-fritted glaze is not an answer. While partially fritted satin and matte glazes have long been used, partially fritted gloss glazes have only been attempted at cone 6 or higher. One reason is the completeness of reaction required of the firing process in making a gloss glaze.
In the first place, a gloss glaze must have a very smooth, mirror-like surface1. When the surface is less than smooth, some of the light reflected specularly from the surface is reflected at angles other than the incident angle, because the surface is not flat and parallel at the point of interaction. Hence, the apparent specular reflection is reduced, and with it the gloss.
Thus, a textured glaze can never be a high gloss glaze! Attaining a high gloss surface requires developing a smooth surface. On the other hand, an isolated defect, however large, that does not disturb the human response to the rest of the surface will not lower the gloss the way that texture will.
Second, if there are internal surfaces within the glaze, due for example to bubbles or crystals or phase separation, those surfaces can themselves cause reflection1. Reflection from these internal surfaces produces the phenomenon called scattering. The effect of internal scattering on gloss depends on the refractive index difference between the glaze and the dispersed particle. When the refractive index difference is large, as with an opacifying crystal, the gloss may be enhanced by additional specular reflectance from crystals near the surface. But, when the refractive index difference is small, but not zero, scattering from the internal particles causes the diffuse reflectance to become overwhelming, leading to a satin or a matte. Hence, the presence of bubbles, phase separation, or low index crystals must be minimized or eliminated.
As we all know, the cost of partially fritted glazes is lower than all fritted glazes. Hence, there is a strong incentive to develop partially fritted gloss glazes. This paper will discuss optimization of both the oxide formula and the raw materials chosen, to make it possible to prepare partially fritted high gloss leadless glazes at most firing conditions above cone 02, or ll00C in a fast fire.
In optimizing the oxide formula of the glaze, there are two issues to be considered1. The first issue affects the clarity of the glaze, and the smoothness of its surface. To achieve good clarity and a smooth surface, the glaze must be free from phase separation, low refractive index crystals, and from gaseous defects.
The study of phase equilibria2 has shown that there is a large incompatibility gap between silica, the principal component of glazes, and both boron oxide and the alkaline earth oxides. This immiscibility gap covers the concentrations normally found in glazes. The result of phase separation is microspheres of the other oxides in a matrix of increased silica content3. The visual effect is haze from scattering of light from the boundaries between the two phases. Hence, phase separation must be suppressed to produce a high gloss glaze.
It is well known in the glass industry that inclusion of alumina in a glass formulation serves to suppress phase separation4. Thus, most commercial glass contains small amounts of alumina. For glazes, which usually contain higher concentrations of the phase separable elements, the alumina content needs to be at least 4-5 mole percent to fully suppress phase separation5.
Alkalis also help to suppress phase separation5. However, their use is limited by thermal expansion requirements. Thus, for high gloss glazes, the recommendation is to use as much alkali as possible, consistent with thermal expansion limitation.
Another source of surface roughness and gloss reduction is crystallization of low index phases. This is the technique used to produce satin and matte glazes. The same technique in bulk glass yields glass-ceramics. In high gloss glazes, therefore, crystallization of low index phases must be prevented.
In most glazes, the most likely low index crystal to crystallize out of a glaze melt is anorthite - CaOAl2O32SiO2. The most effective way to prevent anorthite crystallization is to limit the alumina content of the glaze. The maximum alumina concentrations to suppress crystallization vary with the firing conditions1 from 7-8 mole percent at cone 06 to over 12 mole percent at cone 8.
Hence, we see that optimization of the alumina concentration is important to the formulation of a high gloss glaze5. The optimum gloss is achieved when the alumina content is:
0.055 to 0.06 mole ratio at cone 06; 0.07 to 0.08 mole ratio at cone 1; 0.075 to 0.085 mole ratio at cone 4; 0.08 to 0.09 mole ratio at cone 6; and 0.09 to 0.10 mole ratio at cone 8.
The final concern in formulating a glaze is to see that it does not create gaseous defects that disturb the glaze surface, or create scattering centers. We will have more to say about this when we discuss raw material selection to achieve a selected oxide formula, but there is one aspect that relates to the oxide formula itself. For all applications where the body has not been pre-fired to maturity, a major source of gas is elimination of the air between the particles of the body. This very large quantity of gas must be eliminated before the glaze seals over. This places a premium on maximizing the seal-over temperature. In terms of the oxide formula, seal-over temperature is increased when the concentrations of boron oxide and the alkalis in a frit are minimized, while the concentrations of calcia, magnesia, and zinc oxide when used, are increased.
The other issue in designing the oxide formula is the refractive index of the glaze itself. Fresnel's Law shows that the specular reflectance of a glaze is a strong function of the refractive index of the glaze. In turn, the refractive index is a function of the atomic number of the cation molecules comprising the glaze1. Unfortunately, there are not many opportunities for adding high atomic number, and hence high refractive index materials to a glaze. One important opportunity is to substitute strontium oxide for some of the calcium oxide or the alkalis. This has been shown to increase gloss1. Strontium (atomic number 38) has a higher atomic number than calcium (20), potassium (19), or sodium (11). Including some zinc (atomic number 30) oxide in the formulation is also beneficial. Zirconium (atomic number 40) oxide, below the concentration which produces opacification, will also help.
Bismuth oxide (atomic number 83) has also been recommended for improvement of gloss, but it is a high cost material with relatively high volatility. Hence, it is suitable only for very low fire applications such as glass colors.
Raw material selection is not a trivial consideration in the formulation of a high gloss tile glaze for a number of reasons6. The first reason is melting and/or dissolution rate. During the time when the glaze is at the high temperature during the firing process, all of the various raw materials must either melt or dissolve. That time may be as short as two to ten minutes in a fast-fire operation. Even for slow-fire the time is only one to six hours. Moreover, as kinetic processes, melting and dissolution are strongly temperature dependent. Thus, while these processes have modest effects on a sanitary operation firing at cones 7 to 10, they have major effects on tile operations with a cone 1 fast-fire.
This consideration affects the choice of raw materials to supply the refractory ingredients silica, alumina, and zirconia. Flint (quartz sand) does not dissolve readily below cone 5; alumina dissolves slowly at even higher firing. Other sources are needed.
For supplying the silica requirements, there are several options. Normally, two or more are needed, in order to not add too much of some other element. The options include frits, feldspar, wollastonite, various clays, mica, and pyrophyllite. For supplying the alumina requirements, the options include frits, feldspar, various clays, mica, and pyrophyllite. For the portion of the zirconia requirement that is below the solubility, addition by means of a frit is recommended.
Second, the glaze melting process releases large quantities of gas, which must be eliminated during the firing process7. The space between the particles in the dried-but-not-fired glaze alone amounts to over 40% of the total volume, and is the largest single source of gas. This places a premium on obtaining a dense laydown of the slip during the application process, so as to minimize the air space that must be eliminated during firing. The use of deflocculants can help reduce the air space7. Deflocculated slips settle in an efficient manner to give a dense coating.
A deflocculated slip is produced by addition of certain electrolytes called deflocculants. These materials include several complex salts of sodium and phosphoric acid - sodium tripolyphosphate, tetrasodium pyrophosphate, and sodium metaphosphate. They also include several monovalent salts such as sodium nitrite, borax, sodium aluminate, NH4OH, Na2CO3 or K2CO3, and sodium silicate.
The use of one or more of these deflocculants, at the one quarter to one half percent level, improves the laydown of glaze slips, by reducing the amount of water needed to obtain a free flowing slip. With less water to remove, there is less air-space to be displaced in fusing the glaze.
In addition, while gas escapes readily during the initial stages of firing, once the glaze seals over, the only way to eliminate gas is by diffusing it to the glaze surface, where a bubble of gas may burst and the surface heal over. This diffusion process is inherently too slow for current manufacturing processes. Hence, there is need to maximize the temperature at which seal-over occurs8.
Table 1 lists some of the principal glaze raw materials and the temperatures at which they melt or decompose. The seal-over temperature is increased by reducing the amounts of low melting raw materials. This Table indicates that proper frit selection is very important here, as the frits are the first ingredients to melt in partially fritted glazes. Commercial frits are available with a range of softening temperatures. It is important to maximize the softening temperature, consistent with obtaining a smooth surface.
Another limitation is that some raw materials release substantial amounts of gas on heating. If the temperature of decomposition is not several hundred degrees below the firing temperature, substantial gas may be trapped, producing bubble defects7.
Table 2 lists the decomposition temperatures of selected glaze raw materials, and the extent of weight loss accompanying the decomposition.
Melting of Glaze Raw Materials and Decomposition Products (from ref. 11).
Raw Material Melting Temperature alumina D BaO D CaO D Feldspar 117OC Frits 700-1000C MgO D MgSiO3 1200C metakaolin D nepheline syenite 1100C silica D wollastonite D ZnO D zircon D
D = the melting point is above 1300C. The material, therefore, dissolves in glazes rather than melts.
Material Decomposition Product Weight % Loss Temperature alumina hydrate 250C alumina 35 % clay 500-650C metakaolin 14 % dolomite 800C CaO, MgO 48 % whiting 850-900C CaO 44 % talc 1000C MgSiO3 7 % strontium carbonate l200-1300C SrO 30 % barium carbonate l300-l400C BaO 22 %
When possible, a raw material substitution can eliminate this major source of bubble defects.
The closer the temperature at which the gas comes off is to the firing temperature, the more important it is to eliminate that raw material. Hence, the first recommendation is to not use strontium or barium carbonates directly as raw materials. Rather, obtain strontium or barium values from appropriate frits. A more important recommendation here is to eliminate talc from the glaze formula. Talc loses its water of hydration at approximately 1000C, well above the temperature at which the glaze melts. If more magnesia is required in the glaze formula than is provided as impurity in other raw materials, there are several suitable frits available.
Another recommendation is to replace calcium carbonate (whiting) with wollastonite. Calcium carbonate loses 44 percent of its weight at about 900C, which is near the glaze melting temperature in many cases. This requires a suitable adjustment of the silica content. At cone 1 and below, the silica content in the wollastonite must be compensated molecularly. However, at high temperatures (cone 6 and above) the greater melting power of wollastonite makes it unnecessary to compensate for the silica.
In addition, silica and other refractory materials can serve as an anchor for gas, promoting its retention in the glaze. This can leave bubble defects in the glaze even when the particle does finally dissolve before the end of the firing process, and is another reason why slowly melting refractory materials cannot be used in formulating high gloss fast-fire glazes, as was discussed above.
The glaze processing can also affect the gloss. For example, we made up a glaze slip with varying amounts of water, thereby adjusting the viscosity over the range from 7 minutes to 8 seconds on a #5 Zahn cup. As the viscosity was reduced, the surface smoothness improved, and the gloss increased.
The application process can also affect the gloss and the surface smoothness. The tile industry uses the waterfall or Bell technique, because it gives the smoothest surface of all application methods. The reason is that leadless glazes do not move that much during firing. You get pretty much what you apply.
Let us now consider an example where all these ideas are put into practice14. The oxide formula of this glaze is:
Oxide Mole Ratio Weight Percent Na2O 0.04389 4.11 K2O 0.01541 2.19 CaO 0.12762 10.81 MgO 0.00089 0.05 SrO 0.00552 0.86 ZnO 0.01478 1.82 Fe2O3 0.00100 0.24 B2O3 0.01901 2.00 Al2O3 0.08020 12.35 SiO2 0.66243 60.12 ZrO2 0.02828 5.26 TiO2 0.00031 0.04 P2O5 0.00067 0.14
In this formula, the MgO, Fe2O3, and TiO2 are due to impurities in the raw materials used. The P2O5 is incidental to the deflocculant used.
The alkali oxides, at 0.05880 mole ratio, are near to the maximum concentration consistent with a thermal expansion that will fit tile bodies. Part of the alkaline earth requirement is provided by SrO and ZnO, in order to raise the refractive index. The high alkaline earth to alkali ratio serves to maximize the seal-over temperature. The alumina level is at the edge of the zone of maximum clarity. The B2O3 and ZnO levels are kept low in order to prevent excessive gassing.
To prepare this glaze formula, the following materials were weighed out and blended.
Material Amount Frit P-4K05 11.88 Frit P-4K47 9.75 NC-4 feldspar 42.54 NYAD-400 wollastonite 17.98 Zircoplus Zircon 6.81 Zinc Oxide 1.67 Bentonite 1.17 EPK 3.44 Pyrophyllite 2253 4.50 TSPP 0.25 CMC 0.25
You will observe that no materials with a substantial volatile content are used. Neither is flint or calcined alumina used. Even though the SiO2 requirement is over 0.66 mole ratio, all of it is provided by more readily dissolvable or meltable materials than flint. The frits, feldspar, wollastonite, zircon, clays, and pyrophyllite all contribute to the SiO2 content. Similarly, the 0.08 mole ratio Al2O3 is provided by several materials - the frits, feldspar, clays, and pyrophyllite.
The frit selections are made to minimize the use of low melting products. P-4K05 is a moderate melting frit that serves as a zirconia source sufficient to equal the zirconia solubility in the glaze. P-4K47 is a hard frit that serves as the SrO source. To provide for a dense laydown, the TSPP and the CMC are first dispersed in water before adding the other ingredients, and milling for one hour. The milled slip was applied by a waterfall technique. After drying the glazes were fired to cone 1.
The fired glaze is a smooth, glossy, white opaque glaze. As the zircon content is modest, the glaze is only approximately white. The color data will be found in Table 3. A variety of colors can be achieved by appropriate additions of pigment to the base glaze. A true white is achieved by the addition of 5% additional zircon. As shown in Table 3, this raises the L value and lowers the b value, indicating a whiter glaze. A pink is achieved by the addition of 0.5 % of iron coral pigment K-1868 to the same glaze base.
Note that the pink glaze is made by addition to a base made with 6.81 weight percent zircon, and the true white by adding extra zircon to the same base. If we had tried to make the pink with the true white glaze, substantial more pigment would have been required.
Color Data on Tiles (Specular excluded)
Sample L a b Base Glaze 91.3 - 0.6 + 4.4 Base Glaze + 5% Zircon 92.4 - 0.6 + 3.9 Base Glaze + 0.5% iron coral 75.3 +10.9 + 8.9 Clear Glaze + 5% Pr yellow 84.9 - 4.2 +69.9
To make a strong, pure color, we prepared the following variation on the base glaze:
Material Amount Frit P-4K05 11.88 Frit P-4K47 9.75 NC-4 feldspar 42.54 NYAD-400 wollastonite 17.98 Zircoplus Zircon 1.81 Zinc Oxide 1.67 Bentonite 1.17 EPK 3.44 Pyrophyllite 2253 4.50 Pigment 5.00 TSPP 0.25 CMC 0.25
Here we replaced 5% of the zircon with pigment. Otherwise, it is the same glaze as above. The zircon reduction increases the effectiveness of the pigment in producing color, and results in a pure yellow color.
These samples illustrate that it is now possible to prepare partially fritted high gloss glazes.
G2926B - Cone 6 Whiteware/Porcelain Transparent Base Glaze
A base transparent glaze recipe created by Tony Hansen for Plainsman Clays, it fires high gloss and ultra clear with low melt mobility.
Phase separation is a phenomenon that occurs in transparent ceramic glazes. Discontinuities in the internal glass matrix affect clarity and color.
Every glossy ceramic glaze is actually a base transparent with added opacifiers and colorants. So understand how to make a good transparent, then build other glazes on it.