As the need to reduce carbon emissions becomes more acute, design teams must focus on reducing embodied carbon in their buildings. In last month’s blog, we identified the biggest drivers, including the decision to build in and of itself, as well as building size, extending service life and then reducing the largest impacts due to structural concrete and steel. The drive to reduce embodied carbon is also leading architects to request environmental product declarations (EPDs) for insulating glass. But are insulating glass EPDs important for design decision making?
This month, I am joined again by my colleague, Alexandra Blakeslee, to explore how to approach carbon emissions-focused design decisions for glass, and especially how to use (or not) EPDs in this product category. We conclude that using EPDs for insulating glass sourcing decisions is largely irrelevant. The most important factors are in the architectural design decisions, such as insulating glass make-up and lifespan, not where the glass comes from.
The Big Picture
A recent lifecycle assessment of an office building conducted by Anthony Pak of Priopta Innovations Inc. reveals that glass makes up less than 2% of the embodied carbon of a building. Structural elements, such as concrete and steel/iron, comprised 50% and 35%, respectively. The majority of the 6.9% contributed by aluminum is attributed to the curtainwall. Pak states, “Glass is not particularly carbon intensive, but the materials used for framing it can drive emissions up.”
Of course, the actual impact in each building depends on the window-to-wall ratio, and the building size and massing, but this gives us a sense its magnitude.
Pak reinforced what we have said previously: Aluminum production, while energy intensive, doesn’t have to be as carbon intensive, if it is made from renewable electricity sources or if recycled aluminum is used.
Insulating Glass: Dissecting the Embodied Carbon
Vitro Architectural Glass’ processed glass EPD aggregates data over multiple float glass manufacturing lines and glass processing locations across North America to give emissions data for an average insulating glass unit (IGU). It reports that the manufacturing of float glass is the largest contributor to the global warming potential (GWP) for both double- and triple-pane IGUs at ~75% and 72% respectively (see figure below).
The EPD from Vitro also confirms that triple-pane IGUs have approximately 1.5 times the GWP as dual-pane IGUs, which is driven by the impact of adding a third glass pane. Heat treatment contributes approximately 13%, and low-E coating approximately 2% to the total GWP for an IGU. The IGU manufacturing process itself contributes 10% of the GWP of a dual-pane IGU.
If we assume that insulating glass contributes less than 2% to the building’s embodied carbon based on Pak’s study, that means only 0.2% of the overall building’s GWP is due to the IGU manufacturing process itself (not including the impact of flat glass production, coating and heat treatment). This is a very small impact relative to the rest of the structure.
Vitro’s EPD also reveals that approximately 50% of the GWP of the IGU process (5% of the IGU’s total impact) is due to electricity consumption, and approximately 50% is due to IGU materials (sealants, desiccants, spacer, etc.). The emissions from electricity usage in the IGU process is mostly driven by where the fabricator is located, i.e. how “green” the local electrical grid is, with some contribution from throughput efficiencies.
Many of the impacts due to the IGU materials are somewhat fixed across fabricators, too. In commercial construction, similar materials are used by all fabricators—silicone, polyisobutylene, molecular sieve desiccant, etc., and need to be selected and used in the correct quantities to deliver critical durability performance.
Realistically, by the time the fabricator produces an IGU, the major impacts have already been determined by the architectural specification: The number of lites in the IGU, the thickness of these lites and whether they are heat treated, the glass area, the number of different unique sizes (more uniformity in size and shape means better fabrication efficiencies), etc.
The Implications
Firstly, consider that the range of uncertainty in the reported GWP for the flat glass portion of the impacts of an IGU is higher than the contribution to the total GWP of the IGU manufacturing process (10%). Typically, the uncertainty in EPD data ranges from 20% to above 40%.
Secondly, based on the EPD information discussed above, the extent to which a typical IGU fabricator can meaningfully impact this 10% through items reported by an EPD is small. It also is related typically to factory location (hard to change), throughput and glass scrap rates. But remember, one of the largest and most meaningful impacts an IGU manufacturer can make, which is not captured by an EPD, is with the durability of the delivered unit.
Coupled with the dominance of the flat glass processing in the total GWP of insulating glass, we conclude that asking for, and comparing, EPDs for IGUs from different IGU fabricators is not meaningful for project-level carbon reduction decision making. This is especially true given that any EPD data provided will be averaged over many different glass thicknesses, sizes, shapes, coating types and glass supply chains.
The question remains as to whether it makes sense to make glass choices based on EPDs for the flat glass supply used in the chosen IGU construction. At 75% of the GWP of an IGU, and glass contributing just less than 2% of the overall carbon impacts of a building, the flat glass embodied carbon contribution is estimated to be ~1.5%.
Currently both Guardian and Vitro have published EPDs for flat glass products averaged over their North American operations. The GWP data for both, given the known uncertainties, are basically the same at 1.45 x103 and 1.37 x103 Kg CO2 eq per metric tonne of glass respectively (<4% difference).
The industry-wide flat glass EPD through the National Glass Association reports a very similar number for the industry average GWP of 1.43 x 103 Kg CO2 eq per metric tonne. It includes data from NSG Pilkington’s and AGC Glass North America’s float lines in addition to Guardian and Vitro, and suggests that within the margin of error, the embodied carbon of float glass made in North America is basically the same, independent of manufacturer.
Therefore, design decisions related to IGU size and configuration (number of lites, thickness, heat treatment, number of unique sizes), and IGU service life have more implications for reducing embodied carbon than the choice of either the IG fabricator or the float glass manufacturer based on compared EPD data.
Unintended Consequences
Requiring EPDs from IGU fabricators can also disadvantage small, local manufacturers because of their high cost. Each EPD costs at least $30,000 and must be renewed every five years. This additional cost could unintentionally push high-quality, smaller, regional, fabricators out of the bidding process.
If you wanted to find the “lowest-embodied carbon” glass, you would probably find it from the newest, most energy-efficient glass furnaces. However, the installed float line stock comprises a range of furnace vintages and thus, efficiencies. While sourcing from the most efficient float lines sounds great on the surface, consider a scenario where demand is pushed primarily to newer, more efficient float lines. Will that force the manufacturers to shutter older lines and build new furnaces (creating yet more emissions), rather than maintain and run their older stock, maximizing its useful life, and leveraging its existing embodied carbon? And remember, we are talking about making a small difference in ~1.5% of a building’s embodied carbon.
What Can We Do?
While product EPDs can be valuable tools, it is very important to understand how to use their data to make sound carbon-based decisions. In the case of glazing, do not compare EPDs from different IGU or glass fabricators to make supplier selections. Instead, focus on the IGU design and specification details to reduce the embodied carbon of the glazing and enhance lifespan. Suggested considerations for glazing specifications for low embodied carbon be found here.
Ted Kesik of the University of Toronto points to the carbon triangle (see image above). He says, “The DNA that is imparted during the initial schematic design’s moment of conception largely determines the life cycle performance of the building asset. Getting the building DNA right is everything when it comes to lifecycle performance, including embodied carbon.” He goes on to say that “architects should pay attention to the big-ticket items (structure, interior finishes and fixtures, etc.) and ensure these are durable.”