PRODUCTION OF HOLLOW GLASS MICROSPHERES
FROM AMBER GLASS FRIT

Krista Carlson

Faculty Mentor: Dr. Matt Hall

Summary

A unique way of dealing with the new demands of recycling amber container glass is the production of hollow glass amber spheres. The proposed research studied the feasibility of this and found it to be fairly successful. Traditionally, the manufacture of HGM relies on the intentional addition of a sulfur-containing compound to a powdered borosilicate glass, but in the case of amber glass, sulfur is a pre-existing constituent added to achieve the distinctive brown coloration. HGM were produced from frit by a flame spraying process in which the sulfur already present in the amber glass decomposes to form small internal bubbles. Frit size that accomplished this task the best was 75-106 mm using an oxy-methane flame. This range gave spheres with the most internal bubbles hence, the greatest density change. With more monitoring equipment to regulate the process, HGM technology could be used as an alternate method for the recycling of amber glass.

Introduction

There are many potential applications for hollow glass microspheres (HGM). Sodium borosilicate HGM are often used as light-weight fillers of composite plastics for ship-building, aviation and car-making industries, sensitizing additives in manufacture of industrial explosives, varnishes, and paint fillers. In contrast to mineral and organic fillers, HGM are unique because they have low density but high strength.

Another application of HGM is in hydrogen storage. Hydrogen gas is a viable, non-polluting automotive fuel which could be used to replace gasoline and diesel fuels used in internal combustion engines.

In addition, HGM have proven to be very useful in the field of biotechnology. One common application of HGM is for drug delivery. The hollow cavities of HGM may be loaded with a pharmaceutical agent of choice and then injected or applied to a target treatment area.

The production of HGM is a well-established technology. There are several methods available to produce HGM, but every approach depends on the decomposition of a substance known as a “blowing agent” to form a gas within in a liquid. The rapid expansion of this gaseous product causes the formation of a bubble. One of the most common methods for producing HGM is to intentionally mix a trace amounts of a sulfur-containing compound such as sodium sulfate with a sodium borosilicate glass that is similar in composition to traditional Pyrex^® glassware.^1-2 This mixture is then dropped into a hot flame that melts the powdered glass and sodium sulfate. The melting of sodium sulfate results in a decomposition reaction that releases minute amounts of sulfur gas that form bubbles within the molten glass droplets. The hollow droplets are then rapidly cooled from the liquid state to form HGM. As previously mentioned, such an approach relies on the intentional addition of a sulfur-containing compound to the glass. However, the amber container glass that is produced in large quantities for the beverage and chemical industries already contains approximately 0.04 weight percent sulfur to achieve the signature brown color. The origin of the amber color is commonly attributed to a ferric ion (Fe³⁺) that is complexed by three oxygen ions (O^2-) and one sulfide ion (S^2-), as depicted in Figure 1. The most

Figure 1. The complex that forms the amber chromophore in silicate glasses.

typical application of amber glass is for the production of containers to store light-sensitive fluids, but the fact that sulfur is a natural constituent presented an interesting opportunity to explore amber glass as a potential precursor material for the formation of HGM. The decomposition of sulfur from amber glass melts, also known as sulfur reboil, is a well known phenomenon that has come under intense investigation in recent years. Part of this renewed interest in sulfur re-boil is related to the increased usage of recycled glass cullet in manufacturing due to both legislative and cost considerations. From a manufacturing perspective, uncontrolled sulfur decomposition is something generally to be avoided since the violent foaming that may occur during re-boil can lead to catastrophic effects in an industrial glass tank.³ The type of HGM produced in this research could be suitable for thermal insulation, filers and extenders, and light-weight additives for weight reduction and buoyancy.

This experiment found that through the use of an oxygen methane mixed flame, internal bubbles could be formed in the spheres. Even though very few fully hollow spheres were produced, the internal bubbles in a lot of spheres did produce a significant decrease in density. With more regulation on gas flow and flame temperature, HGM would most likely have been produced.

Methods

Different amber glass containers were made into frit by placing pieces in a ball milling device. Different brands of beer bottles and chemical container glass were used to simulate a realistic environment. Sieving was preformed to sort the frit into discrete particle size intervals. Some problems arose in the process with the finer particles sizes, mainly below 106mm. Clumping of the powder due to moisture and particles caught in the screens hindered correct distribution. Placing the powder in a drying oven before sieving and only using small amounts at a time reduced these problems. Glass microspheres were then produced from each size fraction using a set up as seen in Figure 2.

Torch

Irregular Glass Frit

Collection Tube

Glass

Microspheres

Figure 2. Schematic diagram of the apparatus used to produce glass microspheres.

The flame was produced by using a mixture of oxygen and propane, but when only solid spheres were produced, it was changed to an oxygen methane flame. This produced spheres with internal bubbles.

Density measurements were made on the different samples using helium pycnometry to estimate the amount of spheres that contained internal bubbles.

Results and Discussion

First attempts at producing spheres were unsuccessful because the frit was dropped into the flame too far towards end of the tip, where the flame temperature was not hot enough. This produced partially melted glass chunks. Spheres were produced when the frit was dropped in the beginning or middle of the flame, but were not hollow.

When internal bubbles could not be produced with the oxygen propane mix, an oxygen methane mix was used instead. The oxygen propane flame was not producing bubbles because the flame temperature was too high and the spheres collapsed on themselves before they could solidify.

Although purely hollow spheres could not be produced using the oxy-methane flame, a reasonable amount of spheres with internal bubbles were produced. Bubbles could be seen in different levels in the spheres by adjusting the focus. This was done in every picture taken to make sure that the spheres that appeared solid actually were.

The 75- 106 mm frit size produced spheres with high amounts of small and large internal bubbles. Density change between the oxy-propane and oxy-methane mixes suggested that this frit size had the greatest amount of internal bubbles as seen in Figure 3.

Figure 3. Spheres made from 75-106 mm frit. Picture taken on optical microscope with 10x objective.

Density was 2.4350 g/cm³ with the oxy-propane set up and 2.1326 g/cm³ with the oxy-methane flame. Very few partially melted chunks were observed indicating that the flame temperature was adequate enough to melt the glass but not so high as to get solid spheres. More single bubble spheres were seen in this frit size than any other.

The 106-150 mm frit size produced fewer internal bubbles than the 75- 106 mm, as seen in Figure 4.

Figure 4. Spheres made from 106-150 mm frit. Picture taken on optical microscope with 10x objective.

In spheres where bubbles were present, more small internal bubbles were produced than a single large one as seen in the 75-106 mm. Density changed from 2.5007 g/cm³ from the oxy-propane set up to 2.4407 g/cm³ with the oxy-methane flame.

The 63-75 mm frit size produced more internal bubbles than the 106-150 mm size. It also tended to produced quite a few single bubble spheres as seen in Figure 5.

Figure 5. Spheres made from 63-75 mm frit. Picture taken on optical microscope with 10x objective.

Density measurements showed a small change from 2.5661 g/cm3 to 2.5061g/cm3.

Bubbled and solid spheres were difficult to obtain with the <63 mm frit sphere size. The frit was so light that it was blown off the top of the flame as it was dropped in. This caused half melted agglomerates and partially melted particles to be produced as seen in Figures 6 and 7.

Figures 6 and 7. Spheres and agglomerates from the <63mm frit size. Pictures were taken on the optical microscope using the 10x objective.

Density was not measured on these spheres because of the large amount of agglomeration. Better results could have been obtained with a modified vibrating spatula, where a small tube coming off the end could direct the frit right into the flame.

Reproduction and improving on positive results was difficult because of the torch/tank setup. No flow meters were attached so markings on the torch knobs were made to try to get the same flame settings each time. This helped some but since pressure from the tanks themselves varied, the flame was not reliably reproduced. An optical pyrometer would have also helped to see if the flame temperature was staying constant.

Conclusions

Even though semi-hollow microspheres were produced, the experiment would have been more successful with more monitoring equipment. Flow meters and a thermal pyrometer would have enabled a more successful and reproducible flame temperature. This would have allowed more accurate comparing of the different size microspheres, since all samples were not made at the same time. Further research taking into these considerations would give a reliable way to help in the recycling of amber glass containers.

References

1. V.V. Budov and V.Y. Stetsenko, “Choice of Glass Composition for Producing Hollow Microspheres,” Steklo i Keramika, [8] 15-16 (1988).

2. V.V. Budov, “Physicochemical Processes in Producing Hollow Glass Microspheres,” Steklo i Keramika, [3] 9-10 (1990).

3. J.E. Shelby, “Optical Properties,” Chapter 10 in Introduction to Glass Science and Technology, The Royal Society of Chemistry, Cambridge, UK, pg. 204 (1997).