I just finished my Master’s Thesis on reducing embodied carbon for residential construction. One focus was to simplify material selection for residential builders, to do this I created charts that make it pretty easy (I hope) to identify the materials that should be selected if you’re prioritizing embodied carbon reduction. Take a look at the attached files to see the charts; read on if you’re interested in the data used and how the charts are structured. Please send me any feedback, I’d love to add more material categories and would be happy to work in any updates.
The charts are broken down into the various material subcategories from the EC3 Tool. A scale was created ranging from 0 to the CLF baseline of the subcategory, then the conservative value of the individual material type was mapped to the scale (each 10% has a different color ranging from green to red). For example, in the Board Insulation Chart, the range is from 0 to 9.29kgCO2e. XPS has the highest conservative value at 6.91kgCO2e and scores an 8 on the scale.
Since biogenic carbon hasn’t been standardized, I decided to have it as a separate column. If more than 50% of the EPDs within a certain material type listed biogenic carbon, then I put a checkmark under the “stores carbon” column.
Lastly, I converted the framing member from weight or volume to Linear Foot (I don’t know any builders who buy studs by the pound). If anyone is interested in the conversions let me know and I’ll send you the numbers.
Feel free to use these charts any way you see fit, the more knowledge out there, the better.
The insulation chart is awesome, but the structural ones aren’t comparing equivalent quantities. The insulation compares Co2/R which is functionally equivalent, but the structural members are equivalent sizes but not equivalent capacities, it would be interesting if there were a way to compare Co2/ a value of structural capacity.
That’s a great point! I’ll try to do a comparison based on span tables for the rafters and joists. Regardless, four 2x10s have lower emissions than the LVLs or Glulam beams, which is pretty surprising. Although harvesting more lumber just to store it doesn’t seem like a great idea.
Great work Josh and thanks for sharing. To Melissa’s comment I would add that it would be helpful to display the RSI values used in your calculations for each of the insulation materials being compared. These values differ between manufacturers and databases. It would also be helpful if the table included the GWP of the blowing agents being used in the materials being compared. XPS, for example, can be manufactured using HFCs or HFOs.
I agree with the earlier comment that more detail around the values that were used for insulation materials would be helpful. There is significant variability within a product grouping, so it might be better shared as a range rather than showing so many significant figures which implies a precise value. For XPS, there are a number of variations that have been coming out this year, even from the same manufacturer and the numbers vary greatly based on blowing agent used.
I’ll upload the excel spreadsheet containing the data for the material categories, once I clean it up.
The charts are meant for quick reference to guide material selection (I’m hoping that builders will pull it up on their phones to help decide between pink rigid insulation or white rigid insulation when at the local supply house), not to specify a certain manufacturer. I know this oversimplifies embodied carbon, but I thought the charts would be useful in steering people away from materials produced with HFCs or petroleum-based products and towards natural, sustainably harvested materials.
As for the blowing agents, as of April 2021, the only blowing agents listed for XPS were HFCs, but this should be made clear in the chart and I’ll add it to my list of updates. I’ll add XPS with HFOs once the EPDs are shown in the EC3 database.
@cheryl.smith@philippe.st-jean I just confirmed with the authors of the EC3 tool that the RSI is 1.0 per square foot or square meter, depending on your settings. I believe this means the unknown variable in my charts is the thickness needed to achieve a specified insulation value. However, for a builder, if a wall assembly is specified at RSI 3.5 (R-20), they could easily calculate the GWP of various materials, then check that a certain sized wall cavity can support the specified amount of insulation value (or vice versa). I know two steps is a barrier, I’ll try to think of a method for combining those variables while still keeping the charts easy to read.
Apologies, what I meant to write is the lambda value and not RSI value of the various insulation materials. The RSI value is held constant across materials in the comparison and is therefore inconsequential. The lambda value, however, can vary significantly between manufacturers or installation detail for the same material (e.g., mineral wool, dense pack cellulose of varying densities, etc.). Thus, knowing both the lambda value and density of the materials used would help better contextualize the results. It’s also important to note that in Canada insulation materials that contain refrigerants with a GWP greater than 150 can no longer be sold as of January 2021. Therein the importance of stating the GWP of the blowing agent to allow for informed decisions to be made across markets.
Hi Josh - thanks so much for sharing this interesting work!
I am curious if you (or others) have any insight on the dramatic differences between EPS and XPS. It seems that LCA studies in the literature deviate a bit on this, with some indicating similar values for both materials and others indicating more significant differences. [See e.g., Grazieschi et al.'s 2021 review https://www.sciencedirect.com/science/article/pii/S2666789421000246.]
I guess I don’t know enough about the processing differences between EPS and XPS to understand the source of the large (order-of-magnitude) embodied carbon difference. Is it just the HFC contribution along with the CO2-eq multipliers? I might have thought these materials might have been more similar based on an assumption that most of the embodied carbon is in the feedstock - but that’s just me making relatively uneducated guesses. Maybe @jharmony, @cheryl.smith or others have some better insight on this.
Other questions: How many data points were available in EC3 for each material? Are you pretty confident in the ranking? I’m not sure I’m clear on the methodology you used to establish the “conservative” value for each material.
The polystyrene in both EPS and XPS will tend to have higher embodied carbon from a material standpoint than something like fiberglass for the same functional unit. The blowing agent impact, however, can potentially dwarf the material impact because the multipliers can be so high.
When the literature and websites tend to generalize values for EPS and XPS, they can be misleading since the values can vary pretty dramatically between manufacturers based on the blowing agent blend they are using.
Thanks @cheryl.smith. Really interesting. Glad I stumbled on this post and discussion! Looks like the new product had a significant use phase benefit too. Do you know why that would be? Offgassing?
Yes, the blowing agent emissions happen over time. Since use phase is a significant length of time (75 years per the PCR), there are significant savings.
Hopefully someone can correct me if I’m wrong, but what I’ve found is that EPS uses air as a filler and XPS uses a blowing agent. The blowing agent is what creates the vast difference in embodied carbon between the materials. My charts focused on XPS produced with HFCs since there were very few EPDs with HFOs at the time. Looking at the EC3 tool now there are quite a few more EPDs for XPS with HFO. In a future iteration I would break apart those two subcategories of XPS.
For the data points, I’m not sure how many there were in April, but currently there are 15 for XPS (8 HFO average conservative value of 1.4kgCO2e, 7 HFC average conservative value 8.63kcCO2e). For EPS there are 3 EPDs covering 7 manufacturing plants with a conservative average value of .317 kgCO2e.
I’m confident in the ranking based on the EC3 tool. I systematically went through every EPD within the insulation category to make sure the product listed was in the correct subcategory so the averages were correct. That being said, every EPD has a margin for error - I’ve seen as high as 40% - based on the parameters and omissions of the LCA.
The conservative value was developed by Building Transparency (The operators of the EC3 Tool). The image below is a boxplot diagram that you see when do a material search in the tool, the conservative value is the value that 80% of the EPDs fall below. I chose this value instead of the highest value to eliminate outliers, but I acknowledge that this adds some uncertainty into the charts.
Lots of interesting stuff here! Thank you for sharing. It brings up for me the tangential question about how we sub-categorize product types when setting the CLF Material Baselines, where we’ve wrestled a fair bit with the insulation category.
Comparing EC of like products is super useful. But how ‘like’ do the products have to be before the comparison loses meaning? XPS and EPS and mineral wool and spray foam all do the same fundamental job of insulating, but they have so many differences in terms of performance and application.
One very small note: Open-cell spray foam uses water as the blowing agent. (At least typically…I’m sure there’s another blowing agent buff out there who can let us know if there are exceptions.) So I believe it should not say ‘HFO’ after ‘open-cell’ in your charts.
“One very small note: Open-cell spray foam uses water as the blowing agent. (At least typically…I’m sure there’s another blowing agent buff out there who can let us know if there are exceptions.) So I believe it should not say ‘HFO’ after ‘open-cell’ in your charts.”
Thanks Brook! I’ll make the correction on the next iteration.
Maybe 