Material Substitution for Embodied Carbon
Material substitution is where embodied carbon is actually reduced. The EN 15978 method tells you how to count a building’s carbon and the RICS WLCA standard tells you how to report it — but neither tells you whether swapping Portland cement for a slag blend, primary aluminium for recycled, or a concrete frame for timber actually delivers the saving the headline factor implies.
It often does not — or not by as much as the cradle-to-gate numbers suggest. A substitution saving is only real when the alternative does the same job (functional equivalence), is counted across the same boundary, and uses a declared allocation method. Break any of those three and the saving is an artefact of accounting, not a reduction in atmospheric carbon. This page is the methodology for getting it right: how to compute a defensible substitution delta, where the carbon hides rather than disappears, and which swaps survive scrutiny.
A material-substitution saving is the embodied-carbon delta between a baseline and a functionally equivalent alternative, on a consistent unit basis, boundary and allocation method. Replacing CEM I with a CEM III slag blend cuts binder carbon by about 35%.
Put this into practice with the Material Substitution Savings Calculator, which implements the method described on this page.
Provider directory
Now it has to survive the next design change.
See who does this work. Every listing names the standards it works to, and paid placements are labelled. Including Circular Ecology and Enviropass Consulting.
Browse 4 embodied carbon & LCA providers →Do this work? A listing is US$390 a year. Get listed →
What Material Substitution Accounting Is
Material substitution accounting quantifies the embodied-carbon consequence of replacing one material with another that performs the same function. It is the analytical core of low-carbon design: every value-engineering decision, every specification swap, every “concrete or timber?” question is a substitution problem, and each has a right and a wrong way to be counted.
Substitution vs specification change
Not every material change is a substitution in the accounting sense. A specification change alters the building’s function — a thicker slab, a different fire rating, a larger glazing ratio. A substitution holds the function constant and changes only the material that delivers it. Only the latter yields a clean carbon delta, because only the latter compares like with like. Confusing the two is the first error: a “saving” that is really a reduction in performance is not a saving at all.
Where it sits relative to EN 15978 and RICS WLCA
Substitution analysis is a layer above the counting and reporting standards. EN 15978 supplies the module factors and the summation rules; RICS WLCA fixes the conventions and the element scope. Substitution accounting consumes both — it takes two compliant inventories, one baseline and one alternative, and isolates the difference attributable to the material change. The discipline that makes the difference meaningful is the subject of this page.
The substitution delta, defined
The carbon saving from a substitution is ΔC = (Qbase × EFbase) − (Qalt × EFalt), summed across every module in the boundary. The trap is in the quantities: Qbase and Qalt are the amounts of each material needed to deliver the same function — they are rarely equal, and they are rarely in the same unit. A saving computed with Qbase = Qalt and a single module is the number most people quote, and it is usually wrong.
The Three Validity Conditions
Before any saving is claimed, three conditions must hold. Each is a gate: fail it and the delta is not defensible, however precise the factors.
1 — Functional equivalence
The alternative must deliver the same function to the same performance — structural capacity, fire rating, acoustic and thermal performance, durability over the reference study period. You cannot compare 1 m³ of concrete with 1 m³ of timber: they carry different loads. The comparison must be on the functional unit (e.g. “a floor spanning 6 m at 5 kN/m²”), not the material unit.
2 — Consistent system boundary
Both options must be counted across the same modules. Comparing a baseline on A1–A3 against an alternative on A1–C4 — or quoting one material’s cradle-to-gate against another’s cradle-to-grave — manufactures a saving from the boundary mismatch. Biogenic materials make this acute: timber’s negative A1–A3 reverses at C3, so an A1–A3-only comparison flatters it dramatically.
3 — Declared allocation method
Recycled content and end-of-life credits can be accounted by cut-off or by avoided-burden (system expansion) — and the two give different, sometimes opposite, answers. The method must be declared and applied identically to both options. An undeclared or mixed allocation method is the most common way a substitution claim becomes unauditable.
The single most common substitution error is the m³-for-m³ swap across materials of different structural capacity. Concrete and timber, steel and timber, dense and lightweight masonry — none are interchangeable by volume. Substitute on the functional unit (the load carried, the span achieved, the U-value met), derive the quantity of each material that meets it, then compute the delta. Skipping this step can invert the result entirely.
The Allocation Decision (ISO 14044 Hierarchy)
When a material carries recycled content, or yields a recyclable output, its carbon must be split between the lives it serves. ISO 14044 sets a ranked hierarchy for resolving this, and the rank chosen materially changes a substitution result.
The ranked hierarchy
| Rank | Method | Principle |
|---|---|---|
1 |
Subdivision | Split the process into sub-processes; avoid allocation entirely where possible |
2 |
System expansion | Expand the boundary to include the avoided/displaced function (avoided-burden) |
3 |
Physical allocation — mass | Allocate by a physical relationship, typically mass |
4 |
Physical allocation — energy | Allocate by energy content where mass is not the driver |
5 |
Economic allocation | Allocate by market value — the fallback of last resort |
6 |
Recycling — substitution | Credit avoided primary production (avoided-burden / Module D logic) |
7 |
Recycling — cut-off | Recycled input enters burden-free; no end-of-life credit |
Ranks are live from the MasterBrain (ISO 14044:2006 §4.3.4 hierarchy). Lower rank number = higher preference. The two “special” recycling rows are the cut-off vs avoided-burden choice that drives recycled-content substitution.
Cut-off vs avoided-burden — why the method changes the answer
The two recycling methods are not interchangeable. Cut-off treats recycled input as burden-free (someone else carried the primary-production carbon) and assigns no end-of-life credit. Avoided-burden (system expansion) credits the material for displacing future primary production — the logic behind EN 15978 Module D. The same recycled product can look better or worse depending solely on which method is applied.
Aluminium shows this vividly with live factors. On the product stage (A1–A3), recycled secondary aluminium at 0.26 kg CO₂e/kg beats primary ingot at 10.1 kg CO₂e/kg — a 27% cut-off saving. But the Module D signs flip: primary aluminium earns a recovery credit at end-of-life while recycled secondary, having already consumed the scrap, carries a Module D burden. Add Module D under the avoided-burden method and the ranking can reverse.
Never mix allocation methods across a comparison. If the baseline uses cut-off and the alternative uses avoided-burden, the delta is meaningless — you have measured the difference between two accounting conventions, not two materials. Declare one method, apply it to both options, and report the recycled-content lines and Module D separately so a verifier can see the choice.
The Calculation Logic
The substitution-delta formula
For each option, sum quantity × factor across every in-scope module; the delta is the difference. Expressed per functional unit:
ΔC = Σmodules (Qbase · EFbase,m) − Σmodules (Qalt · EFalt,m)
where Q is the quantity of each material required to deliver one functional unit (not one material unit), EFm is the module-m emission factor, and the summation spans every module in the agreed boundary — A1–A3 at minimum, A1–C4 for any biogenic or recycled material. A positive ΔC is a saving; a negative ΔC means the “alternative” is worse.
The unit-basis bridge
Substitution factors arrive in different units, and a delta computed across mismatched units is silently wrong. Cement factors are per kg; ready-mix concrete is per m³; sheet products are per m². To combine a per-kg binder factor with a per-m³ concrete volume you need a density bridge — ready-mix concrete declares roughly 2,400 kg/m³, so a binder saving expressed per kg must be multiplied through the binder content per m³ before it touches the concrete total.
Three unit families coexist in the building-material factor set and must never be netted directly: per kg (cement, steel, primary/recycled aluminium, glass wool, sand-lime brick), per m³ (ready-mix concrete, all timber, most insulation, clay and aerated masonry), and per m² (aluminium sheet). Within insulation, stone wool is per m³ while glass wool is per kg; within brick, clay and aerated are per m³ while sand-lime is per kg. Confirm factor.unit on both sides of every substitution before computing the delta.
Marginal vs average factors
A substitution that scales — switching a portfolio to recycled steel, or a region’s concrete to slag blends — eventually runs into supply limits. Average factors assume the alternative is freely available; marginal analysis recognises that displacing primary production at scale changes the supply mix and the factor itself. For a single project, average factors are appropriate. For a corporate target or a national policy, the marginal availability of the low-carbon input is a real constraint that average-factor substitution maths ignores.
Worked Substitution Families
Four substitution families cover most building decisions. Each endpoint factor below is live from the MasterBrain (ÖKOBAUDAT 2024); each delta is computed from those endpoints and shown as an audit record.
Family 1 — Binder substitution (cement)
The highest-leverage building substitution: replacing Portland clinker with supplementary cementitious materials. At the cement level the factors are per kg and the pair is live.
| Binder | A1–A3 factor (live) | Unit | Saving vs CEM I |
|---|---|---|---|
| CEM I (Portland) — baseline | 0.665 |
kg CO₂e/kg | — |
| CEM III (blast-furnace / GGBS blend) | 0.433 |
kg CO₂e/kg | −34.9% (0.232 kg/kg) |
| CEM IV (pozzolanic / fly-ash composite) | 0.70414 |
kg CO₂e/kg | +5.9% (worse) |
| CEM II (Portland composite) | 0.89498 |
kg CO₂e/kg | +34.6% (worse) |
| LC3 (limestone calcined clay) — no MB row | ~40% clinker reduction vs CEM I (cited) | kg CO₂e/kg | ≈ −40% (hardcoded) |
CEM factors live from ÖKOBAUDAT 2024. LC3 has no MasterBrain row — the ~40% figure is hardcoded from the limestone-calcined-clay cement literature (MPA, IStructE/ICE low-carbon concrete guidance) and must be verified against a project EPD before use.
The CEM III slag blend cuts binder carbon by 35%, but in this dataset CEM IV (a pozzolanic/fly-ash composite) is 5.9% higher than CEM I, and CEM II is 35% higher. “Blended cement” is not a synonym for “low carbon” — the saving depends entirely on the supplementary material and its replacement level. Always check the specific binder’s factor; never assume a composite cement beats Portland. The binder is applied to a ready-mix volume through a density bridge — the per-kg cement factor must be combined with the binder content per m³, not netted directly against the per-m³ ready-mix concrete factor.
Family 2 — Metals (primary vs recycled route)
Aluminium offers a clean live A1–A3 pair — and the textbook illustration of the allocation problem.
| Option | A1–A3 (live) | Module D (live) | Net A1–A3 + D |
|---|---|---|---|
| Primary ingot — baseline | 10.1 |
[gc_factor key="materials.aluminium.ingot_primary.module_d"] (credit) |
0.929 |
| Recycled secondary | 0.26 |
[gc_factor key="materials.aluminium.recycled_secondary.module_d"] (burden) |
2.610 |
Factors live from ÖKOBAUDAT 2024 (per kg). On A1–A3 alone, recycled wins by 27%. With Module D under avoided-burden, the ranking reverses — the cut-off vs avoided-burden choice, made visible.
On cradle-to-gate (A1–A3), recycled secondary aluminium beats primary by 27%. But primary aluminium carries a Module D recovery credit (−0.967) because it can be recycled at end-of-life, while recycled secondary carries a Module D burden (+1.227) because its scrap input was already consumed. Add Module D and primary’s net (0.929) is lower than recycled’s (2.610). Neither answer is “wrong” — they are different allocation methods. The lesson: declare the method, report A1–A3 and Module D separately, and never claim a recycled saving without stating whether Module D is in scope.
Steel has no primary-vs-recycled split in the building-material EPDs (the ÖKOBAUDAT steel rows are blended-route averages). The route difference can be shown only at the process level: the basic-oxygen-furnace route emits 1.46 tonne CO₂/tonne against the electric-arc-furnace route at 0.08 tonne CO₂/tonne — but these are process-only figures in CO₂ (not CO₂e), and the EAF value excludes the electricity that dominates its real footprint. Use them to illustrate the route logic, not to substitute a full A1–A3 steel factor.
Family 3 — Structural frame (concrete → timber)
The marquee substitution — and the one where functional equivalence and biogenic reversal both bite hardest.
| Material (per m³) | A1–A3 (live) | C3 (live) | Module D (live) | A1–A3 + C3 |
|---|---|---|---|---|
| Concrete C30/37 — baseline | 196 |
5.05 |
-12.1 |
201.1 |
| CLT | -660.1 |
753.4 |
-409.9 |
93.3 |
| Glulam | -608.4 |
753.4 |
-409.9 |
145.0 |
Factors live from ÖKOBAUDAT 2024. The negative A1–A3 is sequestered biogenic carbon; the large positive C3 is its release at end-of-life. On A1–A3 alone CLT appears ~856 kgCO₂e/m³ better than concrete; carried through to A1–A3 + C3 the gap narrows to ~108 kgCO₂e/m³.
Concrete-to-timber concentrates two errors at once. Functional equivalence: you cannot swap 1 m³ of concrete for 1 m³ of CLT — they carry different loads, so the substitution must be derived per structural functional unit, not per cubic metre. Biogenic reversal: timber’s negative A1–A3 is largely cancelled by its positive C3 at end-of-life, so an A1–A3-only comparison overstates the saving by roughly a factor of eight here. Carry both materials through A1–A3 + C3 (and report storage and Module D separately), on a like-load functional unit, before claiming any frame-substitution saving.
Family 4 — Envelope (insulation and masonry)
Lower-stakes but frequent substitutions, both with same-unit pairs that must be respected.
| Substitution (per m³) | Baseline A1–A3 | Alternative A1–A3 | A1–A3 delta |
|---|---|---|---|
| XPS → wood-fibre insulation | 93.84 |
-253.4 |
−347 kgCO₂e/m³ (alternative is biogenic-negative) |
| Perforated → solid clay brick | 146 |
113 |
−33 kgCO₂e/m³ (−22.6%) |
Factors live from ÖKOBAUDAT 2024 (per m³). Wood-fibre carries the same biogenic-reversal caveat as timber — carry it through C3 before claiming the full negative figure. Use stone mineral wool (per m³) for an m³ insulation comparison; glass wool is declared per kg and is not directly comparable.
Edge Cases and Where the Burden Hides
Substitution rarely deletes carbon outright; more often it moves it — to another module, another life-cycle stage, or another part of the value chain. A credible methodology tracks where it goes.
- Biogenic reversal. Bio-based substitutes (timber, wood-fibre, hemp) store carbon at A1–A3 and release it at C3/C4 unless permanent storage is demonstrated — at least
100years under the standard convention, with biogenic CO₂ defaulting to a GWP of0. An upfront-only saving that ignores the reversal is overstated. - Burden-shifting between modules. A material with low A1–A3 but a short service life shifts carbon into B4 replacement over the reference study period. The product-stage saving can be erased by repeated replacement — the comparison must span the whole life, not the gate.
- Recycled-content double counting. If both the supplier of recycled input and the user of the recyclable output claim the benefit, the saving is counted twice. The cut-off vs avoided-burden choice exists precisely to prevent this — pick one, apply it consistently, and never let both ends of the loop claim credit.
- Marginal supply. Slag and fly ash are finite by-products of declining heavy industry; a substitution that works for one project may not scale to a portfolio without displacing the supplementary material from another user. The saving is real locally but bounded globally.
- Performance penalties. Some low-carbon binders cure slower or need higher replacement levels to match strength, which can raise the total binder quantity — eroding the per-kg saving at the functional-unit level.
Before reporting any substitution saving, ask three questions: did the carbon move to another module (B4 replacement, C3 release)? Did it move to another actor (recycled-content double counting)? Could the alternative scale to the whole programme, or only this project (marginal supply)? A saving that survives all three is defensible; one that fails any is an accounting artefact.
Error Traps with Calculable Magnitudes
Each error below produces a specific, quantifiable distortion, shown against the worked families above.
| Error | What happens | Magnitude | How to avoid |
|---|---|---|---|
| m³-for-m³ swap across materials | Concrete substituted for timber by volume, ignoring different structural capacity. | Can invert the result — the quantity of each material per functional unit differs, sometimes by 2× or more. | Substitute on the functional unit (load, span, U-value), derive Q for each, then compute ΔC. |
| A1–A3-only biogenic comparison | Timber’s negative A1–A3 quoted; the positive C3 reversal dropped. | Overstates the CLT-vs-concrete saving ~8× (856 vs 108 kgCO₂e/m³ on A1–A3 vs A1–A3+C3). | Carry biogenic materials through C3/C4. Report storage separately. |
| Mixed allocation method | Baseline on cut-off, alternative on avoided-burden. | For aluminium, flips the ranking entirely (recycled wins by 27% on cut-off; loses on avoided-burden). | Declare one method; apply to both; report Module D separately. |
| Assume “blended cement” = low carbon | CEM II or CEM IV specified as the low-carbon option. | CEM IV is +5.9% and CEM II is +34.6% vs CEM I — a “saving” that adds carbon. | Check the specific binder factor. Only CEM III (GGBS) delivers the −35% here. |
| Unit-basis mismatch | Per-kg cement factor netted against per-m³ concrete, or stone wool against glass wool. | Silent and unbounded — a missing 2,400 kg/m³ density bridge corrupts the line entirely. | Confirm factor.unit both sides; bridge per-kg ↔ per-m³ by declared density. |
| Ignore B4 replacement | Low-A1–A3 short-life substitute compared on product stage only. | Product-stage saving erased by repeated replacement over the 60-year RSP. | Span the whole life; include B4 for any element with an ESL below the RSP. |
Governance, Verification and Reporting
A substitution claim that influences a design decision, a planning submission, or a corporate disclosure must be auditable. Three governance disciplines make it so.
- Declare the functional unit. State the function held constant (the structural element, the thermal performance, the durability) and the quantity of each material that delivers it. A delta without a declared functional unit cannot be verified.
- Declare the boundary and allocation method. State the modules in scope and the recycling-allocation method (cut-off or avoided-burden), applied identically to both options, with Module D and biogenic storage reported separately.
- Document the factor tier. Note whether each endpoint factor is a product-specific EN 15804 EPD, a generic database value (ÖKOBAUDAT, the ICE database), or a proxy — and recompute against project EPDs before a saving is locked into a contract or a disclosure.
For corporate reporting, a substitution saving reduces a Scope 3 Category 1 or 2 figure — but only if the baseline was itself reported. A saving claimed against an unstated or inflated baseline is unauditable; the baseline inventory must be on record before the reduction can be credited.
What the Calculator Handles vs What You Decide
The material-substitution calculator automates the delta arithmetic across live factors. The validity judgements remain yours.
Calculator handles
Per-module quantity × live ÖKOBAUDAT factor for both options; the substitution delta and percentage saving; the per-kg ↔ per-m³ unit bridge by declared density; biogenic A1–A3 / C3 reversal; separate Module D and storage lines; the allocation-method toggle (cut-off vs avoided-burden) applied consistently to both sides.
You decide
The functional unit and the quantity of each material that meets it; whether the swap is a true substitution or a specification change; the allocation method; the boundary (and whether biogenic materials are carried through C3/C4); the factor tier for each endpoint; whether the alternative scales beyond the project (marginal supply); whether a performance penalty raises the alternative’s quantity.
Frequently Asked Questions
Compute ΔC = (quantity of baseline × baseline factor) − (quantity of alternative × alternative factor), summed across every module in the boundary. The critical detail is the quantities: they are the amounts of each material needed to deliver the same function — the functional unit — not equal material units. Replacing CEM I with a CEM III slag blend, for example, cuts binder carbon from 0.665 to 0.433 kgCO₂e/kg, a 35% saving, but only if the slag blend meets the same strength and durability at the same binder content. A delta computed with equal quantities and a single module is the number most people quote, and it is usually wrong.
It depends on the allocation method. On cradle-to-gate (A1–A3), recycled secondary aluminium at 1.383 kgCO₂e/kg beats primary ingot at 1.896 — a 27% saving. But under the avoided-burden method, primary aluminium earns a Module D recovery credit (it can be recycled at end-of-life) while recycled secondary carries a Module D burden (its scrap was already consumed). Including Module D reverses the ranking. Neither answer is wrong — they are different conventions. The rule is to declare one allocation method, apply it identically to both options, and report Module D separately so the choice is visible.
Because a cubic metre of concrete and a cubic metre of CLT do not carry the same load — they are not functionally equivalent. A valid comparison fixes the function (a floor spanning a given distance at a given load) and derives the quantity of each material that meets it, which is rarely equal by volume. There is also a second trap: timber’s A1–A3 factor is negative (sequestered carbon) but its C3 factor is strongly positive (release at end-of-life), so an A1–A3-only comparison overstates the timber saving by roughly eight times here. Compare on the functional unit, across A1–A3 + C3, with storage reported separately.
No. “Blended” is not a synonym for “low carbon” — the saving depends on the supplementary material and its replacement level. A CEM III blast-furnace (GGBS) cement cuts binder carbon by about 35% versus CEM I Portland. But in the ÖKOBAUDAT dataset, CEM IV (a pozzolanic/fly-ash composite) is about 6% higher than CEM I, and CEM II (Portland composite) is about 35% higher. Always check the specific binder’s factor rather than assuming any composite beats Portland. And remember the cement factor is per kg while ready-mix concrete is per m³, so the binder saving must be applied through a density bridge, not netted directly against the concrete volume.
They are two ways to account for recycling, ranked in the ISO 14044 hierarchy. Cut-off treats recycled input as burden-free — the primary-production carbon stays with whoever first made the material — and assigns no end-of-life credit. Avoided-burden (system expansion) credits a material for displacing future primary production, which is the logic behind EN 15978 Module D. The same recycled product can look better or worse depending solely on which method applies, so the method must be declared and applied identically to both sides of a comparison. Mixing them measures the difference between two conventions, not two materials.
Usually to another module, stage, or actor. Biogenic substitutes move carbon from A1–A3 (storage) to C3/C4 (release). Short-life substitutes move it into B4 replacement over the reference study period — a low product-stage figure erased by repeated replacement. Recycled-content claims can move the benefit to another actor and be counted twice if both ends of the loop claim it. And at scale, a finite supplementary material (slag, fly ash) simply displaces the saving from one user to another rather than creating new capacity. A defensible saving is one that survives all four checks: it didn’t move to another module, another stage, another actor, or rely on supply that can’t scale.