years, one of the rods was released from the strain: within 48 hours it returned to its original shape.

R.C. Plumb offers an excellent theory as to why old windows are sometimes thicker on the bottom than on top. He reports that the old technique of manufacturing windows involved collecting a large amount of melted glass at the end of a metal blowpipe, blowing a vase, and attaching the vase bottom to a solid metal rod called a ponty. The end that was blown into is now removed, leaving an open end pointing away from the glassblower. By reheating and then rapidly spinning the hot (soft) vase, the glassblower would use centrifugal force to make the open end flair out, thus transforming the vase into a flat circular pane up to five feet in diameter. From this pane (or "table," as it was called), the glassblower would cut square sections. The sections would have varying thicknesses depending on how far from the center of the "table" they were cut.

Plumb does not offer a strong reason as to why he believes the thicker sections were placed on the bottom. He states, "It would certainly make good sense to install the glass with the thick edge down!" I am unaware of anyone acknowledging if any windows have ever been found to be thicker on the top than on the bottom. It is conceivable that if any such windows were ever noticed, they were disregarded because they did not fit into the pre-expected pattern of being thicker on the bottom.

Because of the belief that glass may sag under its own weight, there has been concern about the storage of glass tubing and rods at an angle. The only danger to glass being stored at an angle is fear of damage to the ends of the glass. Otherwise, there is no problem with storing glass vertically, at an angle, or on its side.

1.1.3 Phase Separation

Pure quartz glass, lead glass, borosilicate glass, or any other type of glass that is clear is in phase. In-phase glass is completely homogeneous throughout. Glass that has any cloudy nature to its appearance can easily be discerned as being out of phase or has phase separation. The cloudy nature is due to inseparable phase (or materials) from the glass phase. As mentioned, glass is glass because it cools too fast for the molecules to align themselves into their crystalline structure. If there are nucleating agents that can enhance the growth of crystals or if the glass is held at too hot a temperature for too long, some crystallization will occur.

Sometimes phase separation can be visually desirable such as that which occurs in opaline glass. By placing an earth alkali fluoride or phosphate material on the surface of the glass, the quickly generated fine-crystalline surface disperses light so efficiently that an opal glow is created. Photosensitive glass is an excellent example of a more practical/commercial use of phase separation. This phase separation is activated by ultraviolet light; and once the ultraviolet light is removed, the glass rephases to the glassy state.

Phase separation is not always a surface phenomenon. The glass that is eventually changed into a pyroceramic material has a nucleating agent mixed throughout

6 Materials in the Lab

the original vitreous material. After the object has been formed and examined, it is slowly baked through its phase separation in an oven.

Vycor®, a high-temperature glass that often can be substituted for quartz glass, is also made by a phase separation process (see page 16).

The phase separation producing opalescence and photosensitivity are produc- tion-created. That is, during the production of the glass, the phase separation occurs. The phase separation that occurs with pyroceramic material and Vycor requires baking the glass at high temperatures for an extended time. This elevated temperature provides the time for the molecules to align and/or separate themselves in a crystalline pattern.

Unfortunately, not all phase separation is desirable. When borosilicate glass is heated for too long near its annealing temperature, a phase separation will occur. This tends to exhibit itself throughout the glass, but can only be observed with an electron microscope. Despite it not being observable to the naked eye, the ramifications of this separation are considerable. The glass separates itself into two phases: One is rich in silicic acid, while the other is rich in alkali borate.5 The result of this change is that the glass has much greater sensitivity to chemical attack.

The significance of the chemical attack sensitivity can best be demonstrated by heat exchangers that must deal with high-temperature water. There are several issues and conditions that come together for this effect:

1.Heat exchangers aremade of thick glass.

Because thick glass requires a longer annealing process, there is a greater opportunity for phase separation to occur.6

2. Due to the manufacturing process,they must go throughthe annealing process several times.

Phase separation is a result of the total length of time the glass is held to high temperatures, not the length of time at any one setting.

3. Glasses with high alkali content are more susceptible to chemical attack.

Water is not generally thought of as a caustic material, but it can be to less chemically resistant glass (e.g., lead and soda-lime glass). Even soda-lime glass that has too great a percentage of soda is more chemically vulnerable than a soda-lime glass with a lower percentage of soda. Generally, borosilicate glass is generally very resistant to water. However, if the alkali concentration is too high (due to phase separation) and this glass is subjected to high-temperature water (more corrosive than room temperature water), greater glass erosion can be expected.

8 Materials in the Lab

weaker and is more vulnerable to chemical attack. If a glass is held within its crystallization temperature* for a sufficiently long time, phase separation occurs and the atoms have time to align themselves into a crystalline structure. Once the temperature is allowed to drop, the glass becomes increasingly more viscous, until it cannot further devitrify.

There are several ways to force glass to devitrify (whether devitrification is desired or not). One technique is to heat the glass until it begins to soften, then mechanically work, or flex, the glass while it cools. Eventually a whitish frost will appear on the surface in the region of compression (see Fig. 1.2).

The risk of devitrification rises the longer a glass is kept in a softened or melted state, and it is also linked to how dirty the glass is. Devitrification typically begins as a surface phenomenon, using either dirt or some other surface defect as a nucleation point.7 The devitrification process may be assisted by variations in the exterior composition (which is typically different from the interior) of a glass object.8 These variations may be the result of flame-working the glass, surface contamination, or chemical attack.

Devitrification can often be removed by reheating a glass up to its melting temperature and avoiding any mechanical action while it recools to a rigid state. However, if a glass is overworked or dirty when originally flame-heated, removal of the devitrification may be impossible. If glass has been held too long at an annealing temperature or has been repaired too often, it may not be possible to remove the devitrification.

Mechanical stress is not a requirement for devitrification. The phenomenon is also common in quartz glass furnace tubes maintained at high temperatures for extended periods of time. Devitrification of silica occurs at increasing rates from 1000°C to 1710°C, which is the crystallization temperature and melting point range of B-cristobalite.1^ Insufficient surface cleaning and very slow cool-down times typically facilitate devitrification on these tubes. Early-stage devitrification on a quartz glass furnace tube may be removed by a hydrofluoric acid dip. This cleaning procedure can remove only surface cristobalite. Devitrification deeper than surface level cannot be removed.

The best way to limit or prevent devitrification on quartz glass is to ensure that it is maintained scrupulously clean: No fingerprints, oils, dirt, or chemicals of any

*The specific crystallization temperature is not commonly identified, but typically is between the annealing and softening points.

^-Cristobalite is transparent. We do not normally consider devitrified glass as a transparent material. However, once fused silica has cooled below 250°C, 6-cristobalite is transformed into a-cristo- balite. This substance is the white opaque material we usually associate with devitrified silica. When fused silica is reheated into the devitrification range, the a-cristobalite turns back into 6-cristobalite. However, because a-cristobalite has many fissures and cracks, the opacity remains when it is reheated back into fi-cristobalite.