Virtually Vitreous
February 2000

Borax IT goes glassmaking
Many glass properties can be treated as equal to the summed contributions of each of the indivdual oxides, assuming the way the glass is made has no effect. The following equation is used to represent this idea: where n is the number of oxides in the glass, ai is a constant representing the effect of each oxide on the property, and xi is the proportion of the given oxide in the glass.

The confidence in predicted values is good, and is similar to the precision of a single glass property measurement (see two properties of textile fiberglass below). This effectively means that virtual glassmaking is as reliable as real glassmaking, but it is much quicker, cheaper and more convenient.

	Fiberizing temperature*	Thermal expansion coefficient
Measured:	1270°C	50.0 x 10-7/°C
Calculated:	1265°C	53.2 x 10-7/°C

* for a viscosity of 103 poise

Vitreous applications account for more than half of global borate usage, and hence half of Borax output. They are divided into five main families – insulation fiberglass, textile fiberglass, borosilicate glass, ceramic frits and glazes, and porcelain enamels for steel and aluminium.

Each of these five main families of glass has significantly different thermal, chemical and mechanical properties, although their compositions all contain chemicals drawn from the same list of about ten oxides. The properties of each are quite different, and are determined by the precise balance of the individual oxides present in the glass. Borax is now harnessing information technology to help decide what mixture is best for which glass, which kind of glass-making process, and which application.

Glass properties are very important for two reasons. They are key in both the fabrication, and the end-use of the glass. For example, to make a glass resistant to heat it has to have a low thermal expansion coefficient to resist thermal shock, so it can be used for cookware and laboratory glassware.

The low thermal expansion required means that oxides of sodium and potassium cannot be used in the glass as they increase this. However, these oxides also reduce the viscosity of molten glass, which makes it easier to melt and form. So, to make heat resistant glass, an oxide is needed that reduces melt viscosity, but does not increase thermal expansion – this is boric oxide, B2O3; the resulting glass, borosilicate. In fact, without B2O3, it would be very difficult to fabricate this type of glass commercially.

Many glass properties can be treated as equal to the summed contributions (see opposite page) of each of the individual oxides, assuming the way the glass is made has no effect. By using a model like this, it is possible to predict glass properties from the oxide composition and this is very useful for several reasons.

Chiefly, the laboratory measurement of some parameters can be difficult, expensive, and time consuming. For example, measuring melt viscosity takes one to two days per composition, requires platinum equipment and a skilled operator, and the results from different laboratories can often vary. The calculation of properties helps to eliminate these problems, and only requires good property/composition models.

Property models have been constructed from statistical analysis of property/composition data. The problem is that with ten or more oxide variables to take into account, there is an enormous number of composition profiles. Until recently only hard copy versions have been available, but now glass databases are published on CD-ROM, giving a more than adequate supply of data.

The Sciglass database, for example, contains 105,000 property/composition relationships. Great for reference purposes as it stands, but to Borax it presented the opportunity to add an extra dimension – the ability to manipulate the glass batch on-screen, and see precisely what effect on the end product these adjustments would have. Borax's own tailor-made software has made computer-generated glass a reality.

The benefits of calculated 'virtual glassmaking' are many. It can remove the trial and error approach to product development by calculating new glass compositions which have the correct desired properties. This is particularly useful for developing ceramic frits and glazes, where the number of oxides is large (around ten), and there are several properties that are of interest— thermal expansion, viscosity, hardness, refractive index, color and so on.

For some glass properties, there is a shortage of compatible and consistent data. Chemical resistance, for one, is an example that cannot be modeled reliably, as test methods vary from laboratory to laboratory. For other properties this problem does not arise, as a given value is not laboratory, make of equipment, or operator dependent. Borax has already developed models for eleven commonly investigated properties, and assembled data to construct many more as they are required. The compositional ranges over which property predictions are valid are very wide, in most cases for B2O3 up to 100 percent.

Borax technology eliminates guesswork
At Borax, another application being studied is the addition of B2O3 to a glaze for sanitaryware, as it is known to greatly improve the resistance of the glaze to crystallization on firing. Glass property calculation is able to ensure that the addition of B2O3 does not change the glaze viscosity or thermal expansion, and the other oxide levels are adjusted to achieve this.

Virtual glassmaking also helps to show how each individual oxide affects each glass property – not just in which direction of change, but also by how much. The modeling process yields oxide 'effects', which are the change in the particular glass property caused by adding or subtracting percent by percent of each oxide.

A great benefit of computer-aided glass property design is the ability to answer 'what if...?' questions at the desk in minutes, rather than in the laboratory over hours or days. It enables Borax to predict what would happen to the properties of a glass if the composition were changed, either by reformulation or by changes in process conditions.

Virtual glassmaking is increasingly being called upon by Borax Technology. One example is the situation when Dehybor® anhydrous borax is being considered as a replacement for Neobor® borax pentahydrate in a glass raw material batch. This can significantly reduce the amount of borate volatilized or lost from the molten glass, and in turn increase the sodium and boron content of the melt, usefully resulting in lower viscosity, increased productivity, higher throughput, and reduced refractory corrosion.

Borax, thanks to its new program, is now able to use 'virtual glassmaking' to predict how, in customers' individual situations, any potential production difficulties can be avoided in a Neobor to Dehybor switch while reaping its rewards.

Dr. Simon Cook is a glass scientist with Borax Europe Limited.
The two graphs show the effects (on a weight basis) of common oxides on viscosity and thermal expansion, two important properties of borosilicate glass.

Figure 1 shows that the effects of Al2O3, B2O3, and SiO2 on thermal expansion are almost exactly the same, so introducing B2O3 to the composition will not change the thermal expansion. Adding Na2O on the other hand, which would lower the viscosity, gives a sharp increase in thermal expansion.

Figure 2 shows that, although B2O3 is less effective in lowering viscosity than Na2O, if it is substituted for Al2O3 or SiO2 (as in borosilicate) the viscosity will be less.

Dead heat to boronThey were both born in 1778. They both achieved the highest scientific and academic preeminence and social honor as a result of their talents. One was a friend of Coleridge and no mean poet himself, the other, something of an adventurer who took a 23,040 foot world record for a balloon ascent. One was the son of an English wood carver, the other the son of a French judge.

Both, independently and during their 30th year, for the first time isolated elemental boron. In the quest to discover elements and the honor of giving them their names, Humphrey Davy vied with Joseph Louis Gay-Lussac. The time was when the two great imperial, powers of Europe were struggling for supremacy with Napoleon of France pitted against George of England.

Perhaps it seems odd to us today that Napoleon Bonaparte, having been trounced in the Battle of Trafalgar, and with his arch-strategist Wellington beginning to chase him out of the Iberian peninsular, awarded a 3,000 franc prize medal for the best experiments that year in 'the galvanic field' – that is, electro-chemistry – to the Englishman Davy. Ironic too that his chemical rival Gay-Lussac was the adjudicating committee's secretary, and that the two were able, during a war, to visit each other and correspond freely.

But then, science transcended politics and wars.
Both men were attempting to isolate and name the element behind boric acid and borax, substances of great interest because of their remarkable properties. Some authorities held that the compositions contained copper (because of the green flame test) or even carbon. But Antoine Lavoisier was the first to suspect it was an unknown element. In 1789 he wrote:

"Borax is a neutral salt with excess of base consisting of soda partly saturated with boracic acid.... The boracic radical is hitherto unknown, no experiments having, as yet, been able to decompose the radical."

To chemists like Davy and Gay-Lussac this kind of statement was a magnet. There was a 'new' element to discover, and the cachet of being the first.

Davy had made his name several times over with the isolation of barium, calcium, strontium, sodium and potassium. With the last two, he had employed his 'galvanic' method to decompose salts by electricity, made possible by the new voltaic pile which was capable of generating continuous current.

In his address to the Royal Society (an organization of eminent scientists) about the electrical isolation of sodium and calcium on November 19, 1807, Davy also mentioned that he had tried it out on boric acid and found that:

"In the electrification of boracic acid moistened with water, I find that a dark coloured combustible matter is evolved at the negative surface; but the researches upon the alkalis have prevented me from pursuing this fact, which seems however to indicate a decomposition."

Gay-Lussac heard the information about Davy's success on November 25, and formally reported it to the Institute, the most prestigious learned body in France (later to become the Académie des Sciences), on January 4, 1808.

Perhaps spurred by the English advances using the new technique, Emperor Napoleon made a gift of a powerful voltaic pile to Gay-Lussac's École Polytechnique. Using this, Gay-Lussac and his assistant L. J. Thenard managed to replicate the isolation of sodium and potassium; they also devised another and particularly dangerous chemical method to isolate it.

In June 1808, Gay-Lussac and Thenard published their findings that boracic acid could be decomposed by being heated with potassium. Boracic acid was "composed of an olive-grey combustible substance and oxygen" (sic), but the report went no further. Later on November 14 they presented another paper, published a day or so later, to the Institute in which they revealed their success:

"The composition of boracic acid is no longer a problem... we decompose and recompose the acid at will."

They called the new element bore.
In the meantime, Humphrey Davy had been taken seriously ill. Unlike Gay-Lussac, he was working solo, so from the time of his first 'dark coloured matter' lecture until spring 1808, he was sidelined.

The Gay-Lussac 'bore' news stung Davy. He was able to present a paper to the Royal Society on June 30, 1808, published in its Philosophical Transactions the next month, saying he had repeated the French experiment, this time using a gold tube. The black substance he obtained was similar to that which he got from the electrical experiments, but:

"The quantities that I have operated upon, have been as yet too small to enable me to separate and examine the products, and till this is done no ultimate conclusion can be drawn."

By December, he had succeeded to his satisfaction and on the 15th told the Royal Society how he had decomposed boracic acid in several ways to a dark olive powder, and then recomposed it again. He also carried out many experiments on its properties and, prophetically, concluded that the dark olive substance was not a 'simple body', foreshadowing its extraordinary range of properties and applications.

He first called it boracium but changed his mind later, "...for I supposed that in its pure form it would be found to be metallic; subsequent experiments have not justified this conjecture. It is more analogous to carbon than to any other substance...." As a result he renamed boracium, boron.

So, who won the race? Undeniably, Davy produced the first ever sample of the element more than a year before anyone else, but did not establish its properties until later. As a matter of scientific record, Gay-Lussac and Thenard were, in mid-November 1808, a month ahead of Davy publicly to record the discovery and description of this unknown element. If, and we only have his later personal memoirs to support it, Davy knew some of boron's vital secrets before that November, he certainly lost the race by waiting, as was then the convention, to announce it first to the Royal Society at its next meeting. This happened to be a month after that of the Institute.

Let us call it an elemental dead heat. Nevertheless, it was a classic example of scientific rivalry and genius, working in parallel to unlock the secrets of chemistry for our benefit.

Science was then a risky and hazardous business. Gay-Lussac and Thenard's method was to heat a fused sodium or potassium compound with iron filings to red heat, inside an iron gun barrel coated with clay. Vaporized elemental metal was then collected and condensed. An explosion during one experiment nearly cost Gay-Lussac his eyesight. The voltaic pile was equally sensitive: Napoleon himself received a huge shock from it during his presentation ceremony at the École.