The Future of Glass Melting

Reducing carbon emissions

Glass plays an important role in our society. Its usage in housing, transportation, communication, food storage, etc. is crucial to enjoying a high quality of life. To produce glass, we need raw materials and energy. We can reduce the need for materials by recycling more. Indeed, a significant advantage of glass is that it can be endlessly recycled without loss in quality or purity although glass waste needs to be purified, cleaned, and color separated before use.

Sustainable glass recycling (UN Goal 11)

Using more cullet for melting means not only considerable savings in raw materials costs and energy usage, but CO2 emissions are also lower. Clean cullet needs to be reheated and homogenized, but melting reaction energy is not required and every 10% cullet addition reduces the energy consumption of glass melting by 2-3%. To melt soda-lime glass from raw materials requires energy of about 2.6 MJ/ kg. As pure cullet, this is reduced to 1.9 MJ/ kg. More importantly, re-melting cullet avoids CO2 emissions from soda ash (Na2CO3) and lime (CaCO3) in the batch. Every metric ton of waste glass recycling saves about 315 kg of CO2 that would be released manufacturing a new glass product [3]. The most common, efficient, end-fired, container glass furnaces, melting with an average of 50% cullet, consume about 3.5 MJ/kg.

Sustainable responsible glass production & climate action (UN Goal 12&13)

Melting glass requires considerable energy to reach the necessary high temperatures (>1500°C). Glass production used to take place in “glass houses” where people had local resources – sand and wood ash as raw materials and wood from the forest for energy. Old glass houses can still be found in forested areas. As much as 150-200 kg of wood was needed then to melt a kg of glass [4]. Assuming wood burning generates about 19 MJ/kg, this equates to >2850 MJ for a kg of glass. Today’s result of 3.5 MJ/kg is astonishingly 800 times more efficient.

Over the last century, the main energy source has shifted to fossil fuels such as oil and natural gas. Modern glass melting uses about 1% of all industrial energy [5] much less than for example steel production. Nevertheless, it is energy intensive and massive improvements have been made over the years. Asahi Glass Company have plotted these downward trends, and the reduction in pollutants such as NOx, SOx and dust emissions for flat glass production (Fig. 1).

Figure 1. Energy efficiency gains over 150 years and NOx, SOx and dust emissions for the last quarter century. (Source http://www.agc-glass.eu/sustainability/environmental-achievements/air)

Figure 1 also shows that since 2000 the relative specific energy line has flattened, suggesting little improvement in recent furnace designs for flat glass. Furnace efficiency had increased because new refractories allowed higher combustion and crown temperatures, and increased melting temperatures. Furnaces became larger, producing more glass per m2 of heat loss surface. Some flat glass furnaces now produce a remarkable 1200-1500 tons/ day while container glass furnaces can melt a high 800 tons/day. But furnace size is limited by the maximum crown span (width), the size of equipment, flame length, and other factors. Larger regenerators have increased heat regeneration from 50% to 70%, close to the theoretical maximum of 75%. This maximum arises from the difference in heat flow in the waste gas (greater mass and specific heat) than the air being preheated.

Figure 2. A 350 TPD container glass melter. Courtesy of Glass Service a.s. (www.gsl.cz)

Figure 2 shows the design of the most common U flame (end-fired) container glass melting furnace, producing about 350-380 TPD (tonnes/day). Cold air enters the base of the regenerator at the right and is preheated to 1200-1300°C, before leaving at the top and entering the combustion chamber. Gas (or oil) is injected into the hot air at the base of the port. This example has four injectors. The iso-temperature surfaces indicate the flame shape that develops. The hot gases radiate heat to the glass melt surface, the furnace walls and the crown, the latter two re-radiating energy to the glass. The waste gases then circulate round the furnace and exit via the left exhaust port, entering the opposite regenerator, and preheating it until the process is reversed after 20-30 minutes. Raw materials enter into the melting basin from two sides. First the batch under the flames is melted. Some designs have a barrier wall (0.8 m high) on the bottom of the furnace to bring the glass from a typical depth of about 1.3 m to the melt surface to aid the removal of small bubbles, the so-called fining process. The glass then dives down into the sunken throat to be delivered into the distributor which connects to the forehearth which takes the glass to the forming machines. The small rods protruding from the bottom of the glass basin are molybdenum electrodes that assist in melting the glass by electrical Joule heating, often called electric boosting. Such a melter is typically about 15 m long by 6 m wide.

The second most common glass melter is the cross-fired regenerative float glass furnace. Flat glass is formed after leaving the melter by floating the melt on a molten tin bath. This glass is mainly used for window glass or automotive windshields also solar panels or sometimes LCD products can be produced. The furnaces can be 35-40 m long and 10-12 m wide. The most typical pull rate is 600-800 TPD, but some furnaces produce 1200 or even 1500 TPD. These cross-fired regenerative furnaces alternate firing from opposite sides. They have five to nine burner ports on each side and the preheated air comes from brick regenerators on each side. Injectors introduce gas into preheated air to create flames crossing the glass melt surface, the hot waste gases exiting to the opposite regenerators. This process is reversed about every 30 minutes.

Figure 3. A cross-fired regenerative 600 TPD float glass melting furnace. Courtesy of Glass Service a.s. (www.gsl.cz)

Figure 3 shows a 600 TPD float furnace with 5 ports with 2 gas injectors on each side. Raw materials are introduced by batch chargers. After melting, the glass is cooled in the working end and leaves by the canal onto the molten tin, where it spreads out to form a flat sheet.

Other furnace designs

Other technologies include the recuperative and the oxy-gas furnace. Oxy-gas furnaces use pure oxygen, extracted from air and may seem more energy efficient than the best regenerative furnaces. A correct analysis though requires the energy and cost of separating the oxygen be considered and usually favors a regenerative furnace. However, oxy-gas furnaces can bring other benefits – NOx reductions and a smaller footprint. Recently, two industrial gas suppliers have reduced energy consumption by preheating the fuel and oxygen.

Linde (Praxair) developed the OptiMeltTM technology to save another 20% of energy by preheating the natural gas with waste gas from the oxymelter to create a syngas (CO + H2) formed by cracking CH4 with CO2 in the waste gas [6]. An interesting side benefit is that CO tends to reduce foam on the glass surface, increasing heat transfer and lowering seed counts.

Air Liquide designed HeatOx technology with heat exchanging recuperators using furnace waste heat to preheat the natural gas and oxygen indirectly to 400-500°C, giving 9-10% additional energy savings. [7-9]. Should this technology be installed in a conventional regenerative float furnace converted to oxy-gas firing, a total of 20-25% energy savings may be achieved. A side effect would be a major NOx reduction.

Finally, an oxy-gas furnace is apparently converted to burn hydrogen more easily than an air-fired furnace. Burning hydrogen with air gives higher flame temperatures typically equating to higher NOx emissions. Oxy-gas furnaces may therefore be more attractive when hydrogen is affordable.

Electric melting

The first continuous regenerative glass melting furnace was invented by Charles William Siemens of Westminster England between 1872 and 1880 and modern regenerative furnaces have changed little since.

Many do not realize though that continuous all electric melting (AEM) is almost as old as gas-fired regenerative melting. The first electric furnace was built in 1905 following French Sauvageon’s design and was for window production. The specific energy consumption was even then only 0.73 kWh/kg. Many designs have been implemented since but recently electric melting has fallen in popularity due to its high cost compared to widely available fossil fuels.

Global warming and pressure on carbon footprints, has rekindled interest in full or partial (hybrid) electric melting. Alternative energy sources for electricity have helped to lower costs and production is essentially CO2 free; for example, in Germany, 40% of electricity is generated using renewable resources such as wind, solar, hydro, and bio. The question for the future is not if more electricity will be used for glass melting but what will be the balance between fully electric and hybrid furnaces (substituting bio fuel for fossil fuel).

Glass is important in generating green renewable energy, or “green electricity.” Most wind turbine blades are composed of reinforced glass fiber. And most solar panels use large quantities of flat glass. In the future photovoltaics will probably be widely integrated into windows. These applications mean that glass is not only a consumer of renewable energy but also has an important role in generating it.

For larger furnaces with higher pull rates, the higher volumes and lower wall losses make recuperators or regenerators sensible. Gas-fired furnaces can be cheaper than the efficient electric melter. This was historically so in most countries because electricity was generated from fossil fuels, and typically 2.5 to 3x more costly per kWh than the fuel alone.

Even small electric furnaces are 70-85% thermally efficient. While a fuel-fired furnace without a recuperator at a low pull is only 10% efficient, adding a regenerator improves efficiency to 45% and an oxy-gas fired furnace, can achieve 50% efficiency.

Figure 4. An 80 TPD cold top rectangular all electric melter using top, side and bottom molybdenum electrodes. Courtesy of IWG Wagenbauer and Glass Service

Most common all-electric melters produced 10-30 TPD, sometimes up to 80 TPD. They were round or hexagonal to avoid heat losses via the furnace walls and to allow more easily distributed batch charging and electric connections. Figure 4 shows a larger rectangular melter at 80 TPD. These cold top electric melters used the batch cover as a heat insulating blanket, conserving heat inside the melt. They were called vertical melters, as the glass melts on the surface near the batch, refines at lower levels and flows out via a bottom throat into a working end/distributor. To maintain batch coverage and hence an insulating blanket, the cullet content was usually below 50%. Electric melters were mostly used for high-quality clear glasses and crystal (lead) glasses, as the redox (color) control is best managed with this process.

During the 1970 global oil crisis, some glass producers, especially in the United States converted their regenerative furnaces to all electric melters. They retained the infrastructure and horizontal configuration because other shapes were difficult to incorporate into their existing space; sidewall losses are less important at higher pull rates.

The future of carbon free melting – electric, hydrogen or hybrid?

Currently, 95% of all glass melting uses fossil fuels, mostly natural gas or heavy oil; but industries are now strongly encouraged to follow the Paris Climate Agreement guidelines and are seeking to minimize CO2 emissions. Many but not all countries are enforcing rules, with penalties for carbon emissions and benefits for reductions. Either way, the glass industry knows its consumers expect low-carbon or carbon-free production, so are working to achieve this while remaining competitive amongst themselves and with other packaging materials.

Four key technologies for carbon reduction exist, in addition to those already discussed. They are:

Cold top all-electric vertical melting (AEM)
Hydrogen combustion (replacing natural gas in regenerative or oxy-gas furnaces)
Horizontal hot top electric melting (H2EM) also referred to as hybrid melting
Horizontal hot top hydrogen-electric melting (H3EM)

The question is: What is the best solution – not just now – but for 2030? 2050? After 2050?

Hydrogen

Currently, truly green hydrogen produced by electrolysis using renewable electric energy is the first choice, but there is simply insufficient available. Even with low electric pricing, hydrogen at 6€/kg is three times too costly to compete with natural gas. So, in most regions, it would be uneconomical, without state subsidy. More research on hydrogen combustion is needed, specifically the effect on the molten glass and refractories of water concentrations approaching 100% in the combustion atmosphere. Certainly, concentrations near 50% in the combustion atmosphere of oxy-gas furnaces created problems. Using electricity to break water into H2 and O2 by electrolysis is expensive and is only now reaching 70% efficiency levels. However, expectations are that investment costs should decline while efficiency continues to increase so that, as more renewable electricity becomes available, hydrogen will become affordable.

But why consider hydrogen? If electricity is used directly, the furnace melting efficiency is much higher than via the hydrogen route. An advantage of hydrogen is the possibility of storage for long periods, allowing long-distance transportation and the creation of a buffer against supply hiccoughs. Storing electricity for similar times is simply not efficient. Unused batteries slowly lose power while storing sufficient energy would require huge batteries. Different storage options are shown in Figure 5; some, such as hydro power have been created but are not universally applicable, mountains and water reservoirs, as in Norway or Austria being necessary. Energy storage today is facilitated by methane which can be stored for millennia in caves with appropriate geology [10].

Figure 5. Showing the capacity and discharge times for different storage technologies. Source RMIT

FlammaTec (part of GS Group) developed THE first Hydrogen burner for glass melting in 2018.

All-Electric Melting

Electric melting has been a proven technology for over a century so why not convert all furnaces to all-electric melting? Mainly because electricity typically costs three times that for natural gas /kWh. While electric melters are twice as thermally efficiency, they are more expensive to operate. Another obstacle remains. Most electric melters are producing less than 80 TPD. Only a handful in the entire world melt more than 100 TPD; and only two have produced 200 TPD – both were stopped due to production issues. All-electric melters greater than 200 TPD, have diameters so large that maintaining a well distributed insulating batch blanket across the melt surface is difficult although a key requirement for keeping the furnace operational. Should the batch cover disappear, the furnace loses heat from the top, the glass cools, melt quality and pull rate fall and production deteriorates. There is also limited long-term experience at that size of producing reduced colored glasses or melting with high cullet levels.

Hybrid melting

Hybrid melting entered the glass dictionary in 2017 being mentioned by companies such as Glass Service, FIC, BDF Industries, Fives, Teco, Horn and Sorg. Previously discussion was limited, though hybrid melting simply means more than one heat source and has a long history. It is analogous to hybrid cars where the engine is the main power source, while the battery-driven electric motors can move the car short distances and add extra power during acceleration. Previously, electric boosting in glass production was often for 15-30% of the total energy input. Combustion is also used in hybrid melters (H2EM) but 50% or more energy comes from electric heating. The thermal efficiency of the electricity is 85-90%, while combustion is about 50%.

A smaller all-electric furnace (<4 TPD/m2) has the following advantages:

No emissions (NOx, SOx) or particulate dust, so no filter or cleaning costs for waste gas
No chimney stack and therefore fewer complaints from neighbors
Lower investment: no crown, regenerator or flue gas channels
No regenerators to clean
Lower raw materials costs, because volatilization reduced
Lower repair costs and shorter repair times
Efficiency is less impacted by furnace size and capacity

Common disadvantages are:

Less pull rate flexibility
Shorter furnace lifetime (8 years for smaller furnaces 50-80 TPD)
Limited experience of operators
Dependent on electrical power stability)
Proven melting only up to 55% cullet
Limited experience with producing reduced colored container glass (hybrid melting helps)

Hybrid melting removes most of the disadvantages.

Glass Service and FIC in cooperation with BDF Industries developed THE first Hybrid design in 2017. A flexible design independent of energy source, melting at times with 80% fossil fuel/ H2 and 20% electric boost (at 3 MJ/kg), or conversely 80% boost and 20% combustion (at 2.5 MJ/kg). This should reduce the risks of adopting a new technology. Figure 6 shows the concept design of such a horizontal hybrid electric melter for container glass.

Fig. 6. shows 3D view of the combustion space and glass melt in a Horizontal Hybrid Electric Melter at 80% electric mode and 20% firing mode. Courtesy of Glass Service a.s. (www.gsl.cz)

Hybrid electric melting and oxy-gas furnace such as this can break the magic energy barrier undercutting a specific energy consumption of 3 GJ/ton of glass (with 70-80% cullet).

Table 1 shows that using electric energy directly in the glass melt is much more efficient than hydrogen whether by combustion or via the fuel cell. Direct efficiency is estimated to be 79%, whereas hydrogen reduces efficiency below 30%.

Table 1. Comparison of electric melting efficiency versus hydrogen route

Also, for producing Float glass it will be possible to use much more electric heating or super boosting then was common till now. A design made by FIC UK to make some first steps into this direction is shown in figure 7 with a 6 MW bottom melter boosting installed in a conventional regenerative float furnace. To make the complete transition it maybe more interesting to combine this also with oxy combustion and then at some point replace the natural gas to more Hydrogen and waste heat recovery. But the efficiency route using Electricity directly will be always higher.

Figure 7. Float furnace with 6 MW Super boosting, going Hybrid.

Conclusions and outlook

With the help of Industry 4.0 automation & renewable energy sources, the required 55% reduction of carbon emissions should be possible before 2030 through:

improved glass recycling (in both amount and quality)
improved Model Based Predictive Furnace control (Dynamic balancing of Electric vs Combustion firing)
greater use of low-cost green electricity, in hybrid or all electric furnaces
the use hydrogen for combustion or electricity generation

Generating hydrogen using green electricity will become important post-2030. The 2050 goal of an 80% CO2 reduction, will require large amounts of green electricity and a functioning hydrogen economy to replace fossil fuels for glass production, and transportation to and from the factory.

Industry 4.0 automation will continue its forward progression. A dark glass factory may be difficult to imagine by 2030, but not by 2050 when the light from hot gobs falling from the forehearth spout will be all that illuminates the factory hall.

This article was originally published in IGS Magazine’s Spring 2022 Issue – Decarbonising the Glass Industry: Read the full Magazine here for more thought-leadership from those spearheading the industry

Author: Ir. H.P.H. Erik Muijsenberg, GLASS SERVICE, a.s.

Erik Muijsenberg is a Mechanical Engineering graduate of the University of Eindhoven from the Class of 1990. For the eight years following graduation, he was employed by TNO Glass group in Eindhoven focusing his efforts in furnace modeling and glass melt technology. In 1997 he became the Glass Department leader.

In 1998 he became a GLASS SERVICE B.V. Managing Director, first GLASS SERVICE subsidiary office in Maastricht, the Netherlands. After eleven years he moved to GLASS SERVICE headquarters in Czech Republic to become group Vice President. GLASS SERVICE employs over 100 engineers with offices worldwide including Czechia, Slovakia, Netherlands, Germany, UK, France, USA, China and Japan. In 1997 he was awarded, together with his former colleagues, with the Otto Schott Award.

In 2012 he received the Adolf Dietzel Industry Award from the German Glass Society for his contribution to the development and acceptance of glass furnace modeling & optimization in the German glass industry. He was chosen as a Fellow member by the British Glass Society in 2014.

Erik is also active Vice Chairman and past Chairman of the Technical Committee 21 – Furnace Design & Operations – of the International Commission on Glass (ICG). As of 2016 Erik became an ICG Steering Committee member.

In 2017 he became a Phoenix Award Committee member. Erik has been selected to become the next Vice Chair and future Chair of the Phoenix Award Committee.

Erik has actively promoted Industry 4.0 smarter model-based furnace and forehearth control and CO2 emission reductions to the Glass Industry for over twenty years.

References