Cutting carbon from Rotterdam's chemical cluster works. It can also create higher critical material demand.

New modelling for the TRANSIENCE project traces the environmental impacts of a deep decarbonisation pathway for Europe's largest petrochemical cluster through its entire global supply chain. Climate and other environmental impacts fall, while critical material demand rises, especially for lithium and iridium.

TRANSIENCE project · 31 July 2026 · ~7 min read

Status: preliminary results from Deliverable D8.1, not yet approved by the European Commission and subject to revision. This case study is a proof of concept for a method still under development — see the limitations section.

Industrial decarbonisation is often judged on one number: how much carbon dioxide comes out of the chimney. It is also, on its own, an incomplete guide. Electrifying a chemical plant means building generating capacity, transmission networks and battery storage. Producing hydrogen means electrolysers. Switching feedstocks means new supply chains for biomass or captured carbon. Every one of those requires metals, land and energy that come from somewhere else — often from outside Europe, and always with environmental consequences of their own.

This case study asks what happens when you follow those consequences all the way through. We took a deep decarbonisation pathway for the chemical sector at the Port of Rotterdam and ran it through a full environmental life cycle assessment, tracking not just greenhouse gases but toxicity, air pollution, ecosystem damage and the extraction of eight critical metals — across the whole global supply chain, not just within the port.

A word of caution before the results. This is a proof of concept. It tests one industrial pathway against one background scenario, using a model coupling that the teams involved are still refining. The numbers below demonstrate what the method can see; they are not yet a basis for policy decisions. That distinction matters, and we return to it at the end.

What we modelled

The Port of Rotterdam is one of Europe's largest industrial and petrochemical clusters, central to the production, import, conversion and distribution of fuels, chemicals and materials. At a stakeholder workshop on 26 February 2026, the port authority and its partners identified a specific gap: plenty of studies examine the cluster's carbon emissions, but far less is known about the wider environmental consequences of decarbonising it.

The pathway assessed is called Carbon Looping. It replaces fossil naphtha and natural gas with a portfolio of alternative carbon sources — biomass-derived feedstocks, recycled plastics, captured carbon dioxide, and synthetic intermediates made from renewable electricity and hydrogen. Methanol becomes a central platform chemical, converted into olefins and aromatics. Carbon capture handles the emissions that remain, with the captured carbon either stored permanently or fed back in as feedstock. The objective is to close the carbon cycle within the sector.

Assessing this required linking several models. An industrial transformation model (ITOM) describes how the cluster's technologies, production volumes and energy demand evolve. Energy system and industrial models (OPEN-PROM, FORECAST) supply the changing electricity, hydrogen, ammonia and methanol mixes across Europe and the world. A global integrated assessment scenario (REMIND, following SSP2-NDC — a moderate pathway based on countries' current climate pledges) sets the background. All of this is translated into a life cycle inventory database that represents not today's global economy but a plausible future one.

That last step is what makes the assessment prospective rather than retrospective. A standard life cycle database is a snapshot of the economy as it is now; using it to assess a 2050 technology would assume the world's electricity, steel and transport stay frozen at present-day carbon intensity. Here the background changes too.

The production volume behind the numbers

One feature of the pathway shapes everything that follows: under Carbon Looping, total polymer production at the port falls from about 1.5 Mt per year in 2025 to around 0.8 Mt in 2050 — close to half. The reason is economic: bio-based and synthetic feedstock routes are more energy-intensive, which makes production in Rotterdam less competitive than in regions with cheaper renewable electricity. A substantial share of polymer production relocates to other European regions and beyond.

The product mix changes too. PET remains the largest single output throughout, holding steady at roughly 410 kt per year. Conventional polyol production is phased out and replaced by retrofit-based routes, which dominate polyol supply after 2040. Bio-based PLA grows.

This means that a substantial part of the environmental improvement reported below reflects producing less in Rotterdam, not producing more cleanly. The burden associated with the relocated production does not disappear — it moves along the supply chain to other world regions, where this assessment does not follow it. Every figure that follows should be read with that in mind.

Climate and other environmental impacts fall

The pathway delivers on its central purpose. Climate change impacts across the whole life cycle fall from approximately 3.8 Mt CO₂-equivalent per year in 2025 to between 1.4 and 2.2 Mt in 2050, with the steepest reduction between 2025 and 2045.

The higher estimate in the range for 2050 excludes biogenic carbon storage, while the lower one counts the carbon temporarily held inside bio-based polymers. Since most polymers are short-lived and release that carbon on incineration or degradation, the more conservative figure is the safer guide to long-term climate benefit.

The reductions come from several directions at once: process changes at the cluster, the shifting polymer mix, and the progressive decarbonisation of the electricity, hydrogen and materials supplying it. Some of the benefit is earned upstream, in a cleaner European grid, rather than at the port itself.

Several other indicators improve alongside. Human toxicity holds roughly steady to 2045 then drops sharply. Particulate matter formation declines. Acidification stays within a narrow band before falling substantially by 2050. Freshwater ecotoxicity peaks around 2045 at roughly 194 CTUe per year, then falls to about 75.

But climate impacts are not eliminated. PET production remains a major contributor throughout, and retrofit polyol production adds significantly between 2035 and 2045. Even an ambitious transformation leaves residual life cycle impacts that would need further mitigation or removal.

Critical material demand increases

The trade-off appears in the metals.

Lithium extraction rises by more than an order of magnitude between 2025 and 2045, driven mainly by PET production and retrofit polyol routes — a consequence of battery storage and electrification spreading through upstream supply chains. Cobalt follows the same shape less steeply. Copper and nickel rise moderately to 2045.

Iridium shows one of the strongest relative increases of any material assessed, concentrated between 2035 and 2045, with retrofit polyol production the dominant contributor. Iridium is essential to the membrane electrolysers that make renewable hydrogen, so this points to a potential bottleneck in hydrogen-based industrial decarbonisation — a scarce metal sitting on the critical path of a strategy the EU is betting on.

Dysprosium climbs steadily to 2045, reflecting demand for the permanent magnets used in wind turbines and efficient motors. Titanium peaks around 2040–2045 and stays elevated, and is the one material where polypropylene rather than polyols dominates by 2050. Palladium stays comparatively flat.

The increases are concentrated in the materials that make electrification and hydrogen possible.

Aggregate mineral and metal resource depletion follows the same trajectory: rising from 2025 to 2045, then declining. Two things contribute to that decline — the fall in production volume, and the higher share of recycled metal supply assumed in the background database — and this assessment does not separate them.

Two products account for most of the burden

Across nearly every indicator, the same two contributors appear.

Retrofit polyol production is a recurring hotspot. It dominates acidification from around 2040, becomes the largest contributor to terrestrial eutrophication and to resource depletion from 2035, and drives the iridium and dysprosium increases. Since it is the route replacing conventional polyol production, this matters: the substitute carries environmental costs the original did not, concentrated in categories nobody was tracking.

PET is a persistent contributor rather than a spike. It is the largest single product by volume throughout and shows up consistently across climate change, particulate matter, terrestrial eutrophication, copper, nickel, lithium and palladium. It does not dominate any single category the way polyols dominate toxicity, but it never leaves.

There is a caveat worth stating: the sharp improvements in 2050 coincide with retrofit polyol production being phased out and overall volumes falling. Some of what looks like a technology becoming clean is a technology being switched off.

What this means

The main finding is methodological: assessing industrial decarbonisation on greenhouse gases alone will systematically miss where the burden goes.

Trade-offs are structural, not incidental. The metals appear precisely because the strategy works — electrification, hydrogen and circular feedstocks are what drive both the emissions reduction and the material demand. They cannot be separated by better implementation.

Burden shifting crosses borders. Nearly half of Rotterdam's polymer production relocates under this pathway, and the associated impacts go with it. A cluster-level or national inventory would record that as an improvement. From a global perspective it is a transfer, and only a life cycle approach makes it visible.

The method transfers. The environmental assessment module is modular by design and can be applied to other industrial clusters — the Rhine-Ruhr area and the Basque Country are candidates already identified. That is arguably the more valuable output of this case study than any individual number in it.

Limitations

These results are preliminary and illustrate what the framework can do rather than what policy should do. Four constraints matter.

One scenario, one background. A single industrial pathway (Carbon Looping) against a single background scenario (SSP2-NDC). With no alternative pathway to compare against, the results show a trajectory, not a choice between trajectories.

The model coupling is still being validated. Mapping between industrial model variables and life cycle inventories has been established through a first systematic matching procedure, but needs further expert validation — particularly for novel production routes and intermediate chemicals not represented in existing databases. The scenarios across the modelling framework are not yet fully harmonised.

Future recycling rates are assumptions, not results. Primary-versus-secondary metal supply shares were compiled from literature and expert judgement, and this assessment used the most optimistic of three variants. Linking these directly to material flow modelling would be more consistent, and would matter: the metal results depend on them.

Use and end-of-life are outside the boundary. Because much of the output is exported and its final fate unknown, the assessment stops at the factory gate. For products whose climate benefit depends on how long they store biogenic carbon, that is a real gap.

How this was modelled

The assessment uses an attributional life cycle assessment with ecoinvent v3.12 as the background database (allocation, cut-off by classification), implemented in the Brightway framework. Prospective modifications to background processes are made with premise, which updates technology market shares, process efficiencies, emissions and supply mixes across sectors according to external scenario information. The pathways framework is used to assess entire transformation pathways rather than individual products.

The functional unit is one year of chemical production at the Port of Rotterdam — the annual output of all final-demand products, including MDI, PBR, PET, polypropylene, nylon 6.6, three polyol routes (conventional, retrofit, lignin-based), PLA, PBS and PMMA. Intermediates and by-products are not treated as separate functional units but are used to build upstream supply chains, avoiding double counting. Imports are included in the market mix of intermediates; no credits are taken for exported intermediates. Use phase and end-of-life are excluded.

Five scenario packages were combined into the prospective database: the REMIND SSP2-NDC global pathway as the overarching background; OPEN-PROM for worldwide electricity and hydrogen mixes; ITOM for the European steel sector; FORECAST for country-specific ammonia and methanol markets in Europe; and the ITOM Port of Rotterdam module for the chemical cluster itself, including precursor market mixes and final energy demand.

Impact categories follow Environmental Footprint v3.1, with climate change also reported using IPCC 2021 factors including biogenic CO₂ and short-lived climate forcers. Categories assessed: climate change, human toxicity (carcinogenic), particulate matter formation, acidification, freshwater ecotoxicity, terrestrial eutrophication, and mineral and metal resource depletion. Extraction of eight metals — copper, lithium, nickel, cobalt, iridium, dysprosium, palladium and titanium — is tracked separately using RELICS.

Future metal circularity was represented by compiling primary and secondary supply shares for material families from IEA, JRC, OECD, International Resource Panel and material-flow literature, under conservative, baseline and optimistic 2050 variants. The optimistic variant was used here. Because published sources mostly report end-of-life recycling rates rather than the secondary market shares that life cycle modelling requires, these projections are best read as scenario-based representations rather than forecasts.

Data and citation

• Scenario results are available on IAM PARIS.
• Models: ITOM · OPEN-PROM · FORECAST · ecoinvent v3.12 · premise · Brightway
• Region: Port of Rotterdam, Netherlands · Years: 2025–2050 · Pathway: Carbon Looping + SSP2-NDC
• Full methodology: TRANSIENCE Deliverable D8.1 (will be soon available on the TRANSIENCE website)
• Licence: CC BY 4.0 — charts and text may be reused with attribution.

This explainer accompanies work carried out under the TRANSIENCE project, funded by the European Union. The underlying deliverable has not yet been approved by the European Commission and its findings may be revised.

The AI assistant Claude (Opus 5, Anthropic) was used to support the writing of this explainer. All findings, figures and interpretations derive from Deliverable D8.1 and have been reviewed and approved by the authors, who take full responsibility for the content.