The week the target arrives and the grid does not move
The frame is designed. Column grid fixed, slab depths on the general arrangement, piling priced off a load schedule that went out a month ago, and now a number has appeared — from a planning condition, a funder's requirement, or a sustainability consultant who joined the team late — that the frame as drawn does not meet. Nobody is offering to reopen the grid. The lead-in on the batching plant is what it is, and the first structural pour wants ordering inside three weeks.
So the useful question is not which frame has the lowest carbon in the abstract. It is which of the decisions between here and the chute has not yet been spent, and who has to agree before each one can move. Three parties own different pieces of it and none of them owns the whole thing: the exposure classification, the strength class and the specified test age belong to the engineer; under EN 206 and, in North America, ACI 301, the mixture proportions belong to the producer, who is being asked to demonstrate performance rather than follow a recipe; and the pour cycle, the striking times and the back-propping belong to the contractor, who will be the one paying if the mix arrives slower than the programme assumed.
There is a category error worth killing early too. Carbon is not a property you can write into a strength class; there is no C30/37 low-carbon. What you can write is a limit on the declared global warming potential per cubic metre alongside the exposure classes, then leave the producer free to meet it — which only works on a performance-specified concrete. EN 206 separates designed concrete, where the specifier states the required performance, from prescribed concrete, where the specifier names the constituents and owns the result. Write a prescriptive mix and then complain about its carbon and you have specified the problem away from the only party equipped to solve it.
Three levers, and they are not the same size
Rank the levers by what it costs to pull them today, not by how large they look in a case study. Volume is nearly always the biggest and usually the first one closed: every cubic metre deleted from an over-thick transfer slab or an over-sized pile cap takes its cement, its aggregate, its reinforcement and its delivery with it, at any binder chemistry. Binder substitution stays open longest, because it lives in the supplier's proposal rather than in the drawings. Changing the structural material is the largest single move available and, where the piling is priced, has usually been closed for months.
The table below includes the column people leave out — who else has to sign. A saving that costs the contractor a day per floor is not free, and pretending otherwise is how a target ends up abandoned at the third pour instead of designed for at the first.
| Lever | Where the saving comes from | What it costs, and who signs |
|---|---|---|
| Less concrete | Volume deleted from slabs, transfer structures and over-sized foundations, with its reinforcement | A design revision and re-analysis; closes the day the reinforcement drawings are issued |
| Lower-clinker binder | Portland clinker replaced by slag cement, fly ash or a calcined pozzolan | Slower early strength, so the contractor and the temporary works designer sign alongside the engineer |
| A later specified test age | Replacement levels that a 28-day conformity age will not permit | A specification change agreed before tender, plus a testing regime that matches it |
| Lower-carbon reinforcement | Electric-arc-furnace bar in place of steel from the blast-furnace route | A sourcing question, settled by the mill's declaration and the bar bender's supply chain |
| A different structural material | A steel, timber or hybrid frame doing the same job for less mass or a lower factor | A scheme redesign, different foundations, a different fire strategy; already closed on most jobs |
Measure the element you have before arguing about the one you want
Every argument that follows needs a baseline that both sides accept, and the baseline has to declare its boundary out loud. The figure a producer's Environmental Product Declaration gives you is modules A1 to A3 under EN 15804 — raw material supply, transport into the plant, and manufacturing — verified under the programme rules of ISO 14025 and, for construction products, ISO 21930. That is a cradle-to-gate number. It does not include the delivery to your site, the pumping, the formwork, the striking, or anything that happens after handover, and comparing it against a whole-life benchmark is the single most common way these conversations go wrong.
Run the element twice: once as designed, once as proposed. What matters in the output is less the headline total than the split. The reinforcement line on a normally reinforced suspended slab is routinely a fifth or more of the element, and it is invisible in any figure quoted per cubic metre of concrete. The per-cubic-metre figure is what you will be asked for, because it is what benchmarks are published in; the element total is what the building is actually made of.
Then treat the published range you started with as scaffolding and take it down. The spread between two ready-mixed suppliers offering the same strength class, drawing on different cement works with different fuel mixes and haul distances, is wider than several of the savings people go chasing — an assessment built on a generic factor can be beaten by the choice of plant alone. For anything reportable the figure has to come from the supplier's declaration for that mix from that plant, and the methods say so: the RICS Whole Life Carbon Assessment for the Built Environment and the Institution of Structural Engineers' How to Calculate Embodied Carbon both put product-specific data at the top of their data hierarchies, with generic figures such as the Inventory of Carbon and Energy standing in until the real one exists.
Put the element's volume, the mix you are being offered and the reinforcement density through it, and the reinforcement share in the breakdown answers the question the meeting is really about — whether your attention belongs on the binder at all, or on the steel that nobody put in the target.
Placed volume of the element.
Replacement level is the biggest lever available.
Used only when the mix above is set to Custom.
Steel content by volume of concrete.
Depends heavily on how the steel was made.
Embodied carbon (A1-A3)
19,000 kgCO₂e
Emission factors here are typical published ranges for orientation, not product data. A reportable assessment uses the supplier's own Environmental Product Declaration — the spread between two suppliers of the same nominal product routinely exceeds 30%.
- Tonnes CO₂e
- 18.83 tCO₂e
- Concrete
- 15,023.5 kgCO₂e
- Reinforcement
- 3,805.95 kgCO₂e
- Reinforcement share
- 20.21 %
- Per m³ of element
- 376 kgCO₂e/m³
- Saving against 100% Portland
- 0 kgCO₂e
- Reinforcement mass
- 11,040.38 lb
They open the calculator with your figures already in it
Embodied Carbon in Concrete Calculator: 18,829 kgCO₂e — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 19,000 kgCO₂e — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- Cradle-to-gate only (A1-A3). Transport to site (A4), site operations (A5), the in-use stage and end of life are excluded, and A4 alone can be significant for a remote site.
- Excludes formwork, which is reused a variable number of times and is genuinely difficult to allocate.
- Does not credit carbonation over the structure's life, which reabsorbs a modest fraction of the process emissions and is treated inconsistently between assessment methods.
What the exposure class will let you buy
Replacement level is not a dial that runs from zero to seventy at the engineer's discretion. The durability classification assigned to each element decides which cements and combinations are permitted at all, and a combination that is not listed against the class is unavailable at any percentage, however good it looks on a carbon sheet. That assignment is the engineer's to make and it belongs in the specification, so it is worth confirming it is written down before anybody proposes a mix against it.
In the European system the composition itself is standardised. EN 197-1 defines the common cements by their non-clinker fraction: CEM II/A-V and CEM II/B-V carry roughly six to twenty and twenty-one to thirty-five per cent siliceous fly ash, while the blastfurnace cements CEM III/A, III/B and III/C step up through approximately thirty-six to sixty-five, sixty-six to eighty and eighty-one to ninety-five per cent granulated slag. EN 197-5 extended the family further with the CEM II/C-M and CEM VI composite cements, which take the clinker fraction lower than the CEM II types of Part 1 allowed and let a producer get there on a ternary blend rather than on slag alone. BS 8500-1, the British complement to BS EN 206, is where those types meet the exposure classes: it tabulates which cement or combination may be used for each class, and against each it sets a minimum strength class, a minimum cementitious content and a nominal cover, all tied to an intended working life. Change the cover on the drawing and you can change the binder allowed underneath it.
In the North American system the same constraints arrive through different documents. Supplementary cementitious materials come in as separate constituents against ASTM C618 for fly ash and natural or calcined pozzolans, ASTM C989 for slag cement with its Grade 80, 100 and 120 activity classes, and ASTM C1240 for silica fume; or the blend arrives pre-made as a blended hydraulic cement to ASTM C595, or as a performance-specified cement to ASTM C1157 — the route that says least about composition and so leaves the producer most room. ACI 318 Chapter 19 does the constraining, assigning each element to freezing-and-thawing, sulfate, water and corrosion classes and attaching limits to each, including the cementitious material options permitted for the sulfate classes. ACI 201.2R sits behind those provisions, with ACI 232.2R, ACI 233R and ACI 234R as the committee reports on the three materials.
The reason the standards will not simply rank binders once is that replacement helps and hurts different mechanisms in opposite directions. High slag or fly ash contents markedly reduce chloride ingress, which is why marine and de-icing exposures are the classes where these combinations are actively favoured; they suppress alkali-silica reaction, which is why they appear as a mitigation route alongside the testing regimes of ASTM C1293 and ASTM C1567; and they cut peak heat of hydration, which is a genuine benefit on anything thick enough to be treated as mass concrete. Carbonation runs the other way, because less portland clinker means a smaller reserve of calcium hydroxide buffering the pH at the bar. Freeze-thaw exposure brings its own interaction with air entrainment and with the finishing method. So the mix that is best for a basement wall in chloride-bearing ground is not the mix that is best for an exposed soffit in a carbonation class, and no single replacement level is right across a frame.
The last constraint is in no standard at all: supply. Ground granulated blastfurnace slag is a by-product of making iron in a blast furnace, so the quantity that exists is set by how much iron is made that way, not by how much concrete would like to use it — and as primary steelmaking shifts towards the electric-arc route, that supply falls. Fly ash follows coal-fired generation and is going the same way, which is why reclaimed ash recovered from ponds has become a real route, still required to conform to ASTM C618 or its European equivalent before it goes near a mix. Calcined clays are what the composite cements in EN 197-5 were written to accommodate, and they are the constituent whose availability is not tied to a declining industry. One argument is worth knowing before somebody else raises it: with slag supply essentially fixed, using more of it here can displace it from another job, so a saving that is real at project level is contested at national level. That is an accounting dispute rather than a reason to specify clinker, but it is a reason to say which of the two you are claiming.
Replacement is paid for in programme
A high-replacement mix does not cost money in the way a carbon target implies. It costs time, at the point in the cycle where the contractor can least absorb it. Strength gain in the first days is slower, so striking arrives later, back-propping stays in longer, post-tensioning cannot be stressed on the same day, and a frame priced on a five-day floor cycle becomes a six-day one — twelve days of preliminaries on a twelve-storey building that somebody has to have agreed to. On most jobs this trade is discovered rather than negotiated, and discovering it at the third pour is how a specification quietly reverts to CEM I.
The lever that buys the replacement back is the conformity age, and it is under-used because twenty-eight days feels like a law rather than a convention. It is a convention. Both ACI 301 and EN 206 work from a specified compressive strength assessed at an age stated in the contract documents; twenty-eight days is the default everyone reaches for, not a requirement of the physics. Slag and fly ash mixes continue gaining strength well beyond it, so specifying conformity at fifty-six days on the elements that are not governing the cycle — foundations, cores poured well ahead of the frame, substructure walls — permits replacement levels that a twenty-eight-day assessment would rule out, at no cost to the programme. It has to be decided before tender, because it changes what the producer bids and it changes the cube testing regime, and it has to be stated in the same breath as the strength class so nobody reads one without the other.
Temperature is the multiplier on all of it. A high-slag mix placed in a cold October is slower than the same mix in June by a margin that matters to a striking decision, and ACI 306R is the guide for what cold-weather concreting demands in protection and in minimum placement temperature. The same chemistry that slows a winter pour is a benefit on a thick raft, where the reduced peak heat of hydration lowers the temperature differential that drives early thermal cracking — the one place where the low-carbon mix and the durable mix are the same mix and no argument is needed.
Which is why the striking criterion should be written as an in-place strength rather than as a number of days. The maturity method under ASTM C1074 is the practical way to prove one: log the concrete temperature, accumulate a maturity index, and read the strength off a calibration curve built by breaking cylinders of that mix at several ages. The trap is that the calibration belongs to the mixture. A curve developed for the CEM I mix the job started with is worthless the moment the binder changes, and using it anyway strips a deck on the strength of a concrete no longer in it. New mix, new calibration, developed early rather than argued about at seven in the morning.
Take the logged average temperature and the hours elapsed and this gives the maturity index the pour has actually accumulated, which is the number a striking decision on a slow mix should turn on instead of a day count borrowed from the mix the job was priced with.
The average concrete temperature over the elapsed curing period, in °C.
The total elapsed curing time, in hours.
The temperature below which the mix gains no useful strength. ASTM C1074 recommends 0°C.
Maturity index
3,360 °C-hours
This uses a single average temperature for simplicity — a real maturity study integrates continuously logged temperature data over time, which is more accurate when temperature varies significantly during curing (e.g. cold nights, hot days). The datum belongs to your mixture: 0°C is ASTM C1074's recommendation for plain Type I cement, and Annex A1 sets out how to measure it for anything else.
- Elapsed time
- 7 days
- Datum temperature used (°C)
- 0 °C
- Temperature above datum (Δ°C)
- 20 Δ°C
They open the calculator with your figures already in it
Concrete Maturity Method Calculator (Nurse-Saul): 3,360 °C-hours — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- A maturity index is not a strength. Until a strength-maturity relationship has been developed for the mix actually supplied — cylinders from that mix, broken at several ages against their own logged maturity — this number cannot be read as an in-place compressive strength, and on its own it does not release formwork, tendons, sawcutting or traffic.
- Nurse-Saul sums temperature over successive intervals; this takes one average for the whole period instead. Cold nights, the hydration peak and any spell the concrete spent below the datum are all flattened into a single figure, so the wider the temperature swing over the period the further this sits from what an embedded logger would report.
- It covers one temperature, not a position in the pour. Corners, edges, thin sections and the top surface run colder than the core and reach a given maturity later, and a striking decision belongs to the coldest critical location rather than to an average of the element.
- The datum is taken exactly as entered — nothing here checks it against your cement, admixtures or supplementary cementitious materials. The Nurse-Saul function is also linear in temperature, so it does not capture the reduced ultimate strength of concrete cured hot: a warm pour can accumulate maturity quickly and still fall short at later ages, which is what the Arrhenius equivalent-age methods exist to address.
- At or below the datum the index reports zero. That is the method's bookkeeping convention, not a verdict on the concrete: it says nothing about whether a pour that got cold or froze has been damaged, and it does not cover cold-weather protection requirements.
- The index is reported in °C-hours whichever measurement system you are reading the site in, and the two temperature rows beneath it are shown in Celsius terms for the same reason: the figure is only meaningful against the strength-maturity curve it is compared to, and that curve carries the units its own cylinders were logged in. Enter your temperatures in whichever units you measured — those fields convert — but if your calibration curve is in °F-hours, multiply this figure by 1.8 before reading a strength off it.
Changing the material is a different argument, made earlier
If whether the frame should be concrete at all is genuinely still open, the project is at scheme stage and this guide has arrived early rather than late. Past that point the comparison is still worth running as evidence for the next building, because the foundations have been designed for a mass a steel or timber frame would not impose.
The comparison itself is easy to rig, and it is usually rigged by accident. A fair one sets two designs against each other that both satisfy the same brief: same spans, same imposed loading, same fire resistance period, same acoustic separation, same service life. Compare lumps of material instead and timber wins every time, because its factor per kilogramme is a fraction of steel's, and the comparison silently omits the deeper sections, the acoustic topping, the encapsulation or sacrificial charring depth and the connection hardware a timber floor needs to do what the concrete one was doing. On the steel side the largest variable is not the sections at all but the route the steel came from.
The consequence people leave out is underground. A timber frame can weigh a fraction of a concrete one for the same floor plate, and that mass difference propagates straight into pile lengths, pile counts and cap volumes. On a soft site the foundation saving can rival the superstructure saving, and it is systematically missing from comparisons that stop at the floor plate — which is an argument in favour of running the comparison over the whole load path or being explicit that you did not.
Then there is biogenic carbon, which is where most published timber comparisons are made to say what their author wanted. EN 15804+A2 requires the carbon stored in the timber to be reported separately from the fossil figure, because whether it can be counted as a saving depends entirely on an assumption about end of life: timber reused or kept in service holds it, timber burned or landfilled does not. Netting the storage off without stating the assumption produces a number that cannot be checked. Keep the two apart, and if the storage is being claimed, say what happens to the material in eighty years and who is guaranteeing it.
The outcome, more often than not, is neither a pure swap nor a defence of the status quo. It is a hybrid — a timber or composite floor on a concrete core, or a steel frame with a lightweight deck — or it is the same concrete frame with less concrete in it, achieved by rationalising spans, thinning slabs that were sized by a rule rather than by a check, and deleting the transfer structure that the architectural plan created. That last route has no supply chain, no early-strength penalty and no argument with the contractor, which is why it should be exhausted before the binder discussion starts rather than after it.
Set the two options side by side on their masses and their declared factors, and the comparison stays honest only for as long as both columns describe designs that satisfy the same brief - so put the fire protection, the acoustic topping and the connection steel into the mass before reading the difference.
Total mass of option A.
A1-A3 GWP per kilogram.
Total mass of option B.
A1-A3 GWP per kilogram.
Carbon difference
11,000 kgCO₂e
Emission factors here are typical published ranges for orientation, not product data. A reportable assessment uses the supplier's own Environmental Product Declaration — the spread between two suppliers of the same nominal product routinely exceeds 30%.
- Option A total
- 23,879.57 kgCO₂e
- Option B total
- 12,600.12 kgCO₂e
- Lower option
- 2 (B is lower)
- Difference in tonnes
- 11.28 tCO₂e
- B as a percentage of A
- 52.77 %
They open the calculator with your figures already in it
Structural Material Carbon Comparison Calculator: 11,279 kgCO₂e — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 11,000 kgCO₂e — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- A fair comparison requires that both options actually do the same job — same span, same loading, same fire rating, same durability. Comparing masses that are not structurally equivalent produces a meaningless answer.
- Biogenic carbon stored in timber is excluded here. Whether it may be counted depends on the assessment method and on end-of-life assumptions, and netting it off silently is one of the more common ways these comparisons mislead.
- Cradle-to-gate only. Transport, erection, maintenance, service life and end-of-life all differ between materials and can reverse a close result.
The reinforcement nobody put in the target
Carbon targets for frames are written in cubic metres of concrete and satisfied in tonnes of steel. A suspended slab at eighty to a hundred and twenty kilogrammes of reinforcement per cubic metre, a transfer structure at two hundred and upwards, a ground slab at forty to sixty — at any of those densities the bar is a large minority of the element and it is entirely absent from a figure quoted per cubic metre of concrete. The first useful thing to do with a target expressed that way is to ask whether it includes the steel, because the answer determines whether the sensible response is a binder change or a sourcing change.
The evidence is the mill's own declaration, produced to EN 15804 or ISO 21930 in the same way as the concrete's, and it is worth insisting on rather than accepting a national average. Reinforcing bar made in an electric-arc furnace from scrap sits far below bar made through the blast-furnace and basic-oxygen route, and where a project's steel comes from is a question the bar bender can answer if asked at tender and cannot answer at all once the schedules are being fabricated. None of this touches product conformity, which continues to be governed by its own specification — BS 4449 for weldable reinforcing steel in the United Kingdom, ASTM A615 and ASTM A706 in North America, with A706 being the low-alloy grade specified where weldability and ductility are required — and by whatever third-party certification scheme the project accepts. A declaration is not a substitute for conformity, and a low-carbon bar that fails its bend test is not a saving.
Detailing has a smaller but real say: rationalising bar sizes and lengths cuts the offcut waste that is paid for in full and delivered in full, and a congested zone redesigned late tends to gain steel rather than lose it. Keep the claim proportionate, though — that is a few per cent, where the mill route is a multiple. And specify the declaration rather than a recycled-content percentage, because recycled content is an input and the declaration is the output an auditor will ask to see.
What actually goes on the page
A specification that produces lower-carbon concrete is mostly a specification that stops blocking it. The clauses below are the ones that do work: they state the performance, name the boundary of the number being limited, leave the composition to the party who can vary it, and remove the two default assumptions — a twenty-eight-day conformity age and a day-count striking rule — that quietly rule out the mixes the target needs.
Order them the way they close rather than the way they read. Exposure classification and conformity age are settled before tender because they change what is bid; a GWP limit added after a price is agreed is worth far less; the reinforcement route becomes unanswerable once the schedules are with the fabricator.
- Assign the exposure classes element by element, with the intended working life and the nominal cover they were derived against — no mix can be proposed against a classification that exists only in the designer's head.
- State the strength class and the age at which conformity is assessed in the same sentence, and use fifty-six days on elements that are not governing the pour cycle.
- Set a maximum declared A1-A3 global warming potential per cubic metre for each strength class, name EN 15804 as the boundary, and require the product-specific declaration it is demonstrated from.
- Permit any cement or combination type allowed for the assigned exposure class rather than naming one, so the producer can meet the limit with what its plant can actually get.
- Write the striking and stressing criteria as in-place strengths with the verification method named, and require the maturity calibration to be developed for the mixture actually supplied.
- Require the reinforcement supplier's declaration and its production route at tender, while leaving the product conformity specification untouched.
- Require the delivery documentation to state the cement or combination type actually supplied, so the mix that was accepted is the mix that can be shown to have arrived.
Claims that will not survive being audited
Three claims recur in these conversations and each will be taken apart by anyone competent reviewing the assessment. The first is carbonation credit: concrete does reabsorb a fraction of the carbon dioxide released in calcining its limestone, over decades and mostly at exposed surfaces, but the methods treat it inconsistently and it is not available as a deduction from a cradle-to-gate figure. The second is a product marketed as carbon neutral where the neutrality was bought rather than made — an offset attached to a mix does not change the declaration, and a GWP limit is met by the declaration or not at all. The third is a comparison whose baseline moved: a proposed mix reported at A1-A3 against an existing frame reported at A1-A5 shows a saving that is entirely an artefact of the boundary.
The unit deserves the same suspicion. Kilogrammes per cubic metre compares mixes; kilogrammes per square metre of floor area compares buildings; a scheme hitting a good per-cubic-metre figure with twenty per cent more concrete in it has gone backwards while appearing to improve. Whatever the target is written in, report the element total beside it so the two cannot drift apart.
What survives audit is dull and short: the boundary you assessed to, the source of every factor, the quantities they were applied to, and an honest statement of what is excluded. EN 15978 gives the modules to declare against, and the RICS and IStructE methodologies give the reporting conventions. Say plainly that the figure is cradle-to-gate, that formwork is not in it, that transport to site is not in it, and that the reinforcement is or is not. A defensible number with a stated boundary is worth more to the project than a better one nobody can reproduce.
On the desk before the mix is agreed
Half of this comes from the design file and half from the supply chain, and the specification cannot be written until both halves are on the table.
- Exposure classes per element, with working life and nominal cover — The durability classification decides which combinations are permitted at all, so it is the constraint every carbon proposal is checked against first.
- Placed volumes by element, taken off the design and not the order — Waste and over-order carry their own carbon but belong on a separate line; the element's declared figure is the design quantity.
- Reinforcement density for each element type — Eighty to a hundred and twenty kilogrammes per cubic metre for a normal suspended slab, far more through a transfer structure, and none of it visible in a per-cubic-metre concrete figure.
- Product-specific EPDs for the mixes being offered, by plant — Cradle-to-gate A1-A3 to EN 15804 from the works that will actually batch it, because the spread between suppliers exceeds several of the savings on offer.
- The mill route and declaration for the reinforcing steel — Electric-arc-furnace bar and blast-furnace bar sit several-fold apart per kilogramme, and the question is answerable at tender and not afterwards.
- The contractor's assumed pour cycle and striking regime — Any replacement level is bought with early strength, so the cycle time it costs has to be agreed by the party who pays for it.
Opens the calculators above on one screen with the dimensions from this article already filled in. Quantities only — this site publishes no price list, because local prices vary too much to publish honestly.