Most foundries that get asked for a casting carbon footprint answer with a document instead of a number. The document usually has a certificate, a tree-planting photo and no declared unit. That is not an answer, and since the EU moved CBAM into its definitive phase it is now an answer that costs your importer real money.
This guide gives you the arithmetic instead. We build a cradle-to-gate casting carbon footprint line by line for a ductile iron housing, show which two inputs actually move it, then run the same part in A356 aluminium and get a number that swings 8× on a single line of the purchase spec. Every figure below comes from a stated emission factor, so you can swap in your own foundry’s data. The method follows the GHG Protocol corporate standard; process ranges are cross-checked against AFS practice notes.

What a Casting Carbon Footprint Actually Measures
A casting carbon footprint is the greenhouse gas burden allocated to one declared unit of product — normally one tonne of shipped castings — cradle-to-gate. “Cradle” means the ore, the scrap yard and the smelter, not your foundry gate. “Gate” means the moment the casting leaves the foundry; machining at a second plant, transport and end-of-life sit outside it unless you say otherwise.
Three boundary choices decide whether two quotes are even comparable:
- Declared unit. Per tonne of shipped castings, per tonne of metal poured, or per part. These differ by the yield factor, and quoting the wrong one is the most common way a number gets quietly halved.
- Allocation of returns. Gates, risers and rejected castings are remelted. Under the cut-off method they carry zero burden of their own; under recycled-content accounting they carry part of it. Ask which one was used.
- Direct vs indirect. Direct means fuel burned in the process. Indirect means the emissions behind the electricity you bought. For steel and aluminium under CBAM, only direct counts — which surprises most people.
Get those three in writing before you compare anything: a casting carbon footprint is only comparable when the unit, the allocation rule and the direct/indirect split all match. A “2.0 t CO2e per tonne” figure is meaningless until you know it is 2.0 per shipped tonne, cut-off basis, direct-only, for a specific part number.
Four Lines That Move Your Casting Carbon Footprint
For induction-melted ferrous castings the whole calculation collapses into four lines plus a mass balance that most people get wrong. Get the mass balance right and your casting carbon footprint stops moving every time somebody re-counts it. Take a ductile iron pump housing: 18 kg finished, 55% yield, 12,000 per year, melted in a medium-frequency induction furnace.
Here is the mass balance that trips everyone up. Per tonne of shipped castings at 55% yield, the foundry pours 1 / 0.55 = 1.818 t of metal. Of that, 1.0 t leaves as castings and 0.818 t comes back as internal returns and is remelted. So the purchased metal per shipped tonne is 1.0 t — always, whatever the yield. Yield does not change how much metal you buy. It changes how much metal you melt.
| Input line | Basis | Emission factor | t CO2e / t shipped |
|---|---|---|---|
| Pig iron | 0.550 t purchased | 1.80 t CO2e / t | 0.990 |
| Purchased steel scrap | 0.450 t purchased | 0.35 t CO2e / t | 0.158 |
| Internal returns | 0.818 t remelted | 0.00 (cut-off) | 0.000 |
| Ferrosilicon + carburiser | 1.8% of metal melted | 4.0 / 3.0 t CO2e / t | 0.120 |
| Melting electricity | 580 kWh/t × 1.818 t = 1,055 kWh | 0.570 kg CO2e / kWh | 0.601 |
| Moulding, coring, fettling | per tonne melted | 0.090 t CO2e / t | 0.164 |
| Total cradle-to-gate | per tonne shipped | — | 2.032 |
That gives 2.03 t CO2e per tonne of shipped castings, or 36.6 kg CO2e per 18 kg housing. At 12,000 parts a year that is 216 t of castings and roughly 439 t CO2e. Note where the weight sits in that casting carbon footprint: charge material — mostly the pig iron — is 62% of the total, and electricity is 30%. Process efficiency is not the main lever. What you put in the furnace is. Material choice per alloy family works the same way; see gray iron vs ductile iron for where the metallurgy forces your hand.
A Casting Carbon Footprint Is Not Yield × Everything
Because purchased metal per shipped tonne is fixed at 1.0 t, improving yield only removes melting energy and melting consumables. It removes no upstream metal burden at all. This is where most published estimates go wrong, and it changes the ranking of what is worth paying for.
| Change from the 2.032 t baseline | New total (t CO2e / t shipped) | Saving | Rank |
|---|---|---|---|
| Charge to 100% purchased scrap, no pig iron | 1.235 | −0.797 t (−39%) | 1 |
| Electricity grid 0.570 → 0.250 kg/kWh | 1.695 | −0.337 t (−17%) | 2 |
| Yield 55% → 65% (better rigging / risers) | 1.896 | −0.136 t (−6.7%) | 3 |
| Melting 580 → 520 kWh/t (furnace upgrades) | 1.970 | −0.062 t (−3.1%) | 4 |
Read that order again. Cutting yield loss — the thing every foundry improvement project targets, and the thing riser design e solidification simulation both help you do — is worth about one sixth of what switching the charge mix is worth. Not because yield does not matter, but because the metal you did not ship is recycled rather than thrown away, and returns carry no burden of their own under cut-off accounting.
There is a hard engineering limit on lever 1 though. A 100% scrap charge in ductile iron drags in tramp elements — copper, chromium, tin, lead — that push pearlite and carbides where you wanted ferrite. Most production foundries land somewhere between 20% and 60% pig iron and buy the rest as certified low-residual scrap. Knowing that range is what lets you sanity-check a supplier’s casting carbon footprint claim.
Same Housing in Aluminium: One Line in the Spec, an 8× Swing
Redesign the housing in A356-T6 and it comes out at about 6.8 kg. Yield on a permanent-mould or low-pressure aluminium casting runs near 50%, so the foundry melts 2.0 t per shipped tonne and buys 1.0 t of ingot. Everything except that ingot line is small, which is why the casting carbon footprint for aluminium is really a purchasing specification rather than a process metric.
| Aluminium charge route | Ingot factor | Melting + aux | t CO2e / t shipped | Per 6.8 kg part |
|---|---|---|---|---|
| Secondary (recycled) A356 ingot | 0.60 t CO2e / t | 0.947 | 1.547 | 10.5 kg |
| Primary smelter A356 ingot | 11.5 (mid of the 8–16 published band) | 0.947 | 12.447 | 84.6 kg |
Both are true. Same drawing, same foundry, same furnace — and the casting carbon footprint moves from 10.5 kg to 84.6 kg per part on one line of your material spec. For comparison, the ductile iron version was 36.6 kg.
- Aluminium on recycled ingot: 10.5 kg — 3.5× lower than the iron casting.
- Aluminium on primary ingot: 84.6 kg — 2.3× higher than the iron casting.
“Aluminium is the low-carbon choice” is not a fact. It is a purchasing decision you have to write down. Specify recycled content with a mill certificate and a minimum percentage, and make the foundry quote both routes — the price difference is small and the casting carbon footprint difference is not. The mechanical side of the same choice is covered in A356 vs A380.

CBAM: The CN Code Decides, Not the Material
The EU Carbon Border Adjustment Mechanism entered its definitive phase on 1 January 2026. From that date only an authorised CBAM declarant can release covered goods into free circulation, and the declarant — your importer, or their indirect customs representative — carries the bill. Full detail sits on the European Commission CBAM portal.
What catches buyers out is that CBAM does not cover “steel” and “aluminium” as materials. It covers a list of CN codes in Annex I. Two castings on the same pallet can fall on opposite sides of it.
| CN heading | What it covers | CBAM status | Practical note |
|---|---|---|---|
| 7325 | Cast articles of iron or steel | Not listed — outside scope | The heading most grey and ductile iron castings clear under |
| 7326 | Other articles of iron or steel | Listed — in scope | Where machined or further-worked iron articles often land |
| 7616 | Other articles of aluminium | Listed — in scope | Aluminium castings and machined aluminium parts |
| 7301–7311, 7318 | Sheet piling, tubes, structures, fasteners | Listed — in scope | 13 Chapter 73 headings in total |
| 8400–8700 series | Pumps, valves, vehicle parts | Outside scope | A casting classified as a machine part leaves CBAM entirely |
So an iron housing under 7325 carries no CBAM obligation, while the aluminium version of the same housing under 7616 does. Meanwhile 7325 is exactly the heading that takes the hardest hit on US Section 232 derivative duties. One classification decision sets your duty rate and your casting carbon footprint bill in opposite directions depending on destination — which is why the HS code and import duty walkthrough insists you settle the code before you model any landed cost.
Two more rules that matter commercially. First, there is a 50 tonne de minimis: below 50 t of CBAM goods per legal entity per calendar year there is no obligation, and the threshold is cumulative across all covered goods, not per shipment. Second, non-preferential origin still governs — a casting poured in one country and machined in another keeps the origin the pouring gave it if that operation gave it its essential character. Facing a flange does not change origin.
The aluminium list is published heading by heading; the UK government’s CBAM goods-in-scope list mirrors the EU Annex I codes and is the quickest place to check a specific one. Confirm yours against TARIC, and where the code is genuinely arguable, buy a Binding Tariff Information decision. A few weeks of waiting buys three years of certainty.
What CBAM Counts, and What the Bill Looks Like
Under the definitive-period methodology, steel and aluminium are assessed on direct emissions only. Indirect emissions from purchased electricity are excluded for these two sectors, although they are included for cement and fertilisers. That has an awkward consequence: the number your foundry can measure most cheaply — its kWh meter — is the number CBAM does not want. What it wants is the direct emissions embedded in the metal that went into the furnace, supported by a verifier’s report covering the right period.
If you cannot produce verified actuals, your declarant falls back to published default values, and those are deliberately punitive. Defaults are set conservatively and carry a mark-up that steps up each year: +10% for 2026, +20% for 2027, +30% from 2028. Certificate prices in the €75–82 per tonne CO2e range tracked the EU ETS through 2026.
| Scenario — 400 t/yr of aluminium castings under 7616 | Embedded emissions | Certificate cost at €80/t CO2e |
|---|---|---|
| Verified actuals, secondary-ingot route | 480 t CO2e | €38,400 / yr |
| Default values at 2× the actual figure | 960 t CO2e | €76,800 / yr |
| Gap paid for having no data | 480 t CO2e | €38,400 / yr |
That gap is margin destroyed by an administrative omission — a casting carbon footprint nobody measured still gets billed, at the default rate. and it lands on your importer — which means it lands in your next price negotiation. Certificates go on sale on 1 February 2027 and the first annual declaration, covering all of 2026, is due 30 September 2027, so the liability is already accruing. Fold it into your total cost of ownership model rather than treating it as a compliance footnote.
One forward note: the Commission has proposed extending CBAM to roughly 180 additional downstream steel and aluminium product lines. It is not law. If it passes, the earliest realistic effect is 2028, and the 7325 exclusion above is exactly the kind of gap it is aimed at.
Twelve Data Fields to Put in the RFQ
Ask for these casting carbon footprint fields in the first RFQ, not after the award. A supplier who can produce them in a week has a measurement system; one who needs three months does not. Either way you learn something useful. This slots directly into the casting RFQ package you are already sending.
- Declared unit and boundary. Per tonne of shipped castings, cradle-to-gate, for this part number.
- Yield. Shipped weight divided by poured weight, per part number.
- Melting route and intensity. Induction, cupola or arc; kWh per tonne melted, and what the meter actually covers.
- Electricity factor. Grid emission factor used, plus any on-site renewable or PPA share.
- Charge mix. Percentage pig iron, purchased scrap, internal returns.
- Purchased metal factors. Emission factor per input, with mill certificate or EPD.
- Returns treatment. Cut-off or recycled-content, stated explicitly.
- Consumables. Resin and binder kilograms per tonne of metal.
- Heat treatment. Cycle and kWh per tonne treated, if treated — see tratamento térmico de peças fundidas e heat treatment services.
- Machining and finishing. kWh and consumables per part, if in scope — see secondary operations e CNC machining services.
- Verification status. Actual values verified by an EU-accredited verifier, or defaults — and which.
- Data period and revision rule. Which 12 months, and what triggers a restatement.
Six Answers That Should Make You Push Back
- “We are carbon neutral.” Ask for the number in t CO2e per shipped tonne. Offset purchases do not change it.
- Per tonne poured instead of per tonne shipped. At 55% yield that understates by 1.82×.
- Electricity only. For steel and aluminium CBAM excludes indirect emissions, so an electricity-only report is the wrong number for the wrong purpose.
- No charge-mix breakdown. The pig iron share is the single largest line. Refusing to disclose it is refusing to disclose the answer.
- “Industry average.” CBAM runs on installation-level data. An average is a default value with extra steps.
- No verifier named. Unverified actuals cannot be used, so your declarant pays the default mark-up anyway.
Treat this exactly like the rest of supplier qualification: a credible casting carbon footprint is four numbers, not a brochure, and the question is not whether they hold a certificate, it is whether they can produce the four numbers that generate it.

Five Levers, Ranked by What They Actually Move
| Lever | Typical reduction | Cost and constraint |
|---|---|---|
| 1. Charge mix — cut pig iron, buy certified low-residual scrap | −0.80 t/t (−39%) | Little capital cost; limited by tramp-element control and scrap supply |
| 2. Lower-carbon electricity — PPA, on-site solar, green tariff | −0.34 t/t (−17% at 0.250 kg/kWh) | Real contract cost; excluded from CBAM for steel and aluminium |
| 3. Yield improvement — rigging, risers, simulation | −0.14 t/t (−6.7% at 55→65%) | Also cuts melt cost directly; strongest combined lever |
| 4. Melting efficiency — lining, preheating, scheduling | −0.06 t/t (−3.1% at 580→520 kWh/t) | Capital cost; see foundry automation |
| 5. Cut rework and scrap castings | Scales with your current defect rate | Zero cost at the margin; every remelt is a second full melt cycle |
Lever 5 is the one that gets forgotten in carbon accounting and never forgotten in cost breakdown: a casting that fails final inspection has already been melted, moulded, poured, shaken out, heat treated and machined. Remelting it means paying the whole melting line again. Cutting remelt by one percentage point on a 216 t programme takes roughly 2 t of metal out of the furnace and about 1.1 t CO2e with it — for free. The arithmetic is identical across fundição em areia, fundição sob pressão e fundição à cera perdida programmes: every remelt is a second full cycle.
Then close the loop commercially. Put the data fields into the RFQ, put the revision trigger into the quality agreement, and put the casting carbon footprint next to unit price in your sourcing sheet. A casting carbon footprint you can audit is a negotiating position. One you cannot audit is a surcharge waiting to arrive.
Perguntas frequentes
Is a lower casting carbon footprint always more expensive?
Usually not, and often the opposite. Scrap-based charging and higher yield both cut metal and energy purchases, so the cheapest route and the lowest-carbon route coincide more often than not. The two that genuinely cost money are certified low-residual scrap and low-carbon electricity contracts. Everything else on this list lowers unit cost and casting carbon footprint at the same time.
Can I use one number for my whole casting programme?
No. Yield, charge mix and heat treatment all vary by part and by alloy, and the spread is large — roughly 1.5 to 2.5 t CO2e per shipped tonne across a typical ferrous programme. CBAM wants the casting carbon footprint per CN code and per installation anyway. Ask per part family at minimum.
Who actually pays the CBAM bill?
The authorised declarant — normally your EU importer, or their indirect customs representative if they use one. But an importer facing a charge they cannot control will come back to you at the next price review, which is why the number belongs in your quote rather than in their compliance file.
Do castings shipped to the US face anything similar?
Not a carbon mechanism. The comparable exposure is Section 232 derivative duty on the metal content, which is a classification question rather than a data one — covered in casting import duty and HS codes.
How often should the data be refreshed?
Annually, aligned to the reporting period. A casting carbon footprint is a maintenance item, not a one-off deliverable. Grid factors, charge mix and yield all drift, and a number that is two years old will not survive a verifier’s review or your customer’s own Scope 3 audit.
À procura de uma fundição de metal fiável na China?
- Podemos ajudá-lo a otimizar o design de produtos e a poupar custos.
- Podemos ajudá-lo com peças fundidas de alta qualidade e em grandes quantidades.
- Podemos entregar a tempo e obter mais oportunidades no mercado de vendas.
- Irá beneficiar do serviço de fundição de metais da Supro MFG.
*Emission factors are stated inline so they can be substituted with installation data. Method follows the GHG Protocol corporate standard; CBAM scope references Regulation (EU) 2023/956 Annex I and Commission guidance. Process ranges cross-checked against AFS publications and Supro MFG production records for tratamento de superfície, iron casting, aluminium casting e steel casting programmes.
