construction · materials · family: the wrong ruler
the wrong rulerfor concrete's lifetime
Durability rules judge low-clinker concrete by how fast it carbonates — but carbonation alone does not corrode rebar, 16 competing test standards cannot be converted to real-world rates, and a standard's conservative carbonation rates would write off reusable precast elements
Problem statement
Replacing Portland clinker with supplementary materials such as slag, fly ash or limestone is "one of the most feasible and rapidly implementable solutions" for cutting concrete's CO₂, but lowering the clinker content "can lead to an increasing risk of neutralisation of the concrete pore solution and potential reinforcement corrosion due to carbonation," and the durability rules that decide whether a concrete may be used in a given exposure class treat carbonation depth (the advance of the pH-neutralised front toward the reinforcement) as the measure of corrosion risk. After six years and 120+ members, RILEM Technical Committee 281-CCC concluded that this ruler is bent in three places. There are "at least 16 different standards or recommendations" for carbonation testing "with significant differences in sample curing, pre-conditioning, carbonation exposure conditions, and methods used for determination of carbonation depth"; an interlaboratory test across 22 laboratories found "the variation between laboratories was even larger than the potential effect of raising the CO₂ concentration from 1 to 4%," and "accelerated and natural carbonation rates cannot be converted into each other without applying correction factors"; "it remains unknown how best to translate carbonation rates obtained via accelerated tests to carbonation rates under natural conditions for different types of concrete." Worse, the premise is shaky: reviewing documented practical cases, the committee found that "a reduced pH alone is not sufficient to lead to significant steel corrosion in concrete," so "from practical experience there does not seem to be a direct correlation between reduction of pH associated with concrete carbonation, and corrosion of steel reinforcement." Meanwhile, at the other end of the life cycle, the ReCreate reuse project found that applying the standard carbonation rates in EN 16757:2022 to a 1967 Swedish precast building "led to the conclusion that most of the precast elements would not be reusable," whereas with contextual carbonation rates "all elements were deemed suitable for reuse." The unsolved problem is a qualification regime for concrete durability that measures a proxy, measures it inconsistently, and cannot yet be replaced because the real predictors of corrosion in carbonated low-clinker concrete are not established.
Why this matters
Cement is among the largest industrial sources of CO₂ and clinker substitution is its main near-term lever (a general-knowledge framing, not from the source); if low-clinker concretes are excluded from, or over-designed for, common exposure classes because of an accelerated test that over-predicts their carbonation, decarbonisation is slowed by the very rules meant to protect service life. The committee notes that "this mindset has long dictated the research priorities surrounding the developments towards new, low-emission binders" — an approach its members call "rather simplistic" — and that the finding "has profound consequences for the established approach of assessing the durability performance based on carbonation testing and modelling." Cover-depth requirements are in play too; the committee even suggests "the concrete cover prescribed in standards for design of concrete for different exposure classes, could be revised considering the loading in service condition," having found carbonation depths 9–16% lower under compressive load at 30% of failure load and up to 70% higher under tensile load at 60% of the failure load. Getting the ruler right also matters for CO₂ accounting (how much CO₂ concrete re-absorbs) and for the reuse economy that depends on estimating remaining service life of existing elements.
What’s been tried and why it hasn’t worked
The TC's five working groups produced a database of natural and accelerated carbonation data spanning years of exposure, a predictive equation in which the water-to-reactive-CaO ratio is "a decisive factor," a recommendation for testing carbonation under load, and a critical review of standards. The 22-lab round robin showed that 91-day accelerated tests rank cement types consistently and are "a good estimate of the potential resistance to carbonation," but "large variations within and between laboratories did not allow to draw clear conclusions regarding the effect of sample pre-conditioning and carbonation exposure conditions." For low-calcium alkali-activated materials, "even at the relatively modest concentration of 1% CO₂, accelerated testing may lead to inaccurate predictions of the carbonation resistance under natural exposure conditions." Natural exposure requires "at least one year" to yield a stable rate. And the corrosion review found the influencing factors that can "lead to considerable corrosion damage" are "the moisture state, the microstructure of the carbonated concrete, various species that may be present — even in minor amounts — in the concrete pore solution, and the cover depth" — none of which a carbonation-depth test measures. On the test side the letter's concluding remarks say that for accelerated carbonation testing "it is important to promote the harmonisation of the sample preconditioning, and testing exposure conditions, particularly CO₂ concentration, in existing standards." On the reuse side, ReCreate found EN 16757's carbonation rates "may be overly conservative and hinder the reuse of concrete elements," and argues for relying on contextual carbonation rates from condition assessment and new on-site measurements.
What would unlock progress
Progress needs (1) correction factors, or better a physics-based mapping, between accelerated and natural carbonation for each binder family, built from natural-exposure programmes that record the microclimate; (2) a qualification test that measures what actually drives corrosion in carbonated concrete — moisture state and pore-solution chemistry of the cover — rather than pH front position, particularly "in novel cementitious systems with a limited service track record"; and (3) standards revision (EN 12390-10/-12, EN 16757, CEN/TR 17310, exposure-class rules) that admits contextual measurement for existing elements. Adjacent precedents: fatigue and creep qualification in metals, where accelerated tests are tied to service conditions by validated equivalence models; and corrosion engineering in pipelines, which qualifies by moisture and electrochemical state rather than by a single chemical proxy.
Entry points for student teams
A materials team could set up a comparative programme on one or two low-clinker mixes — accelerated carbonation at 1% and 3–4% CO₂ under two of the standard protocols, plus indoor natural exposure — and quantify how protocol choice changes the durability verdict, adding a data point to the RILEM database. A team with electrochemical skills could instrument carbonated specimens of different binders at controlled moisture states to test the finding that a reduced pH alone is not sufficient to cause significant steel corrosion and prototype a moisture-based screening test. A structural-reuse team could re-run the ReCreate service-life assessment on a local existing precast building, comparing the EN 16757 rate with measured carbonation depths, and draft the assessment guidance a permitting authority would accept. Relevant skills: cement chemistry, corrosion electrochemistry, statistics, standards analysis.
Genes
Structural cousins: stuck the same way, in other fields
Sources
De Belie, N. and Bernal, S. A. (2025), "Closing Letter of RILEM TC 281-CCC: Carbonation of Concrete with Supplementary Cementitious Materials," RILEM Technical Letters 10: 22–32, doi 10.21809/rilemtechlett.2025.220, accessed 2026-08-18; ReCreate project (2025), "Reusing Precast Concrete for a Sustainable Future: Evaluating service life, carbonation and carbon footprint," accessed 2026-08-18 go to source 1 ↗ go to source 2 ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
All RILEM quotations and figures (2018–2024, 120+ members, five working groups, 16 standards, 22-lab interlaboratory test, 1→4% CO₂ comparison, 9–16% and up to 70% loading effects, corrosion-review conclusions, research needs) are from the open-access RILEM Technical Letters closing letter (11 pp.) read on 2026-08-18; the EN 16757 finding is from the ReCreate project's January 2025 research summary of a peer-reviewed study on a 1967 Swedish building, read the same day (the underlying paper was not read — flag). `failure:regulatory-mismatch` is applied under its "wrong measurement basis" sub-pattern (the qualification test measures a proxy that the evidence says does not predict the failure it is meant to prevent); `failure:wrong-problem` ⚠ was considered — the objective (durable reinforced concrete) is right, only the metric is wrong — and rejected per decision order A.1. `constraint:technical` (materials/chemistry sub-category) is applied because the committee states the true corrosion drivers in carbonated low-clinker concrete are not yet established, so a better test cannot simply be written; `constraint:regulatory` because the standards embed the proxy. `constraint:data` was considered for the missing natural-exposure records but the binding gap is mechanistic understanding plus standards. `temporal:window` was considered (second-generation Eurocode and EN 16757/CEN TR 17310 revisions are live) but not applied for lack of a dated deadline in the sources — a verifier may add it. Related collection briefs: `construction-concrete-service-life-multi-mechanism` (combined-mechanism prediction) and `materials-scm-batch-quality-variability` (batch quality); this brief is the mis-specified durability qualification proxy. Sibling intake brief `construction-reused-precast-concrete-product-approval-gap` covers the legal/product-approval side of precast reuse that the EN 16757 finding touches. Note the steer to avoid measurement-gap genes: the core here is a wrong measurement basis embedded in regulation, not an inability to measure.
Source type: Self-articulated (international technical committee summarising its own open questions)
Verified at intake 2026-08-18: gate (net) + adversarial source check + contested-tag second coding. Verifier re-read the RILEM Technical Letters closing letter PDF (all quotations confirmed: 16 standards, 22-lab ILT, 1→4% CO₂, 'cannot be converted', 'remains unknown', 'reduced pH alone is not sufficient', 'mindset', 9–16%/70%, 'at least one year') and the ReCreate January 2025 research summary; the EN 16757 sentence was re-quoted verbatim (the earlier 'incorrectly deemed' wording was a paraphrase in quotation marks).
Related briefs (distinct sub-problems, cross-referenced 2026-08-18): `construction-fly-ash-scm-supply-shortfall-alternative-acceptance`, `materials-scm-batch-quality-variability`.
Reconciliation 2026-10-01: Whole write-up read against both sources it names. Both were fetched on 2026-10-01 and read in full: the RILEM Technical Letters closing letter (https://letters.rilem.net/index.php/rilem/article/download/220/227/2228, the 11-page PDF, direct download) and the ReCreate project article of 15 January 2025 (https://recreate-project.eu/2025/01/15/reusing-precast-concrete-for-a-sustainable-future-evaluating-service-life-carbonation-and-carbon-footprint/, direct). The citation on the Source line was compared with the Crossref record for the DOI (https://api.crossref.org/works/10.21809/rilemtechlett.2025.220) and matches: De Belie and Bernal, RILEM Technical Letters 10: 22-32, 2025. (1) Long title. It ended "a Standard's Conservative Carbonation Rate Has Written Off Reusable Precast Slabs." The ReCreate article describes a study of one 1967 building in which the EN 16757 rates "led to the conclusion that most of the precast elements would not be reusable"; it speaks of precast elements, not slabs, and does not say any element was actually discarded. The title now ends "a Standard's Conservative Carbonation Rates Would Write Off Reusable Precast Elements." The Display title is unchanged. (2) The text said "every such blend carbonates faster than plain Portland cement" and listed calcined clay among the substitutes. The letter does not say this; it says lowering the clinker content "can lead to an increasing risk of neutralisation of the concrete pore solution and potential reinforcement corrosion due to carbonation," and it says the carbonation resistance of new materials such as clays has not been extensively studied. The sentence now quotes the letter and names slag, fly ash and limestone. (3) The text said the correction factors "do not yet exist." The letter says the rates cannot be converted without correction factors and that it remains unknown how best to translate accelerated rates to natural ones; the added words were taken out. (4) "The industry has been optimising against the wrong target" is not in the letter; it now gives the letter's own words, that committee members call the approach "rather simplistic." (5) "Cover-depth requirements, and therefore concrete volume and embodied carbon, follow directly" is not in either source; the sentence now says only that cover-depth requirements are in play, followed by the committee's suggestion, with the comma the source has restored inside the quotation. (6) "So standards default to accelerated tests" is not in the letter and was removed. (7) An altered quotation: the text quoted the letter as "leading to considerable corrosion damage"; the letter's words are "the influencing factors that can, however, lead to considerable corrosion damage." The quotation now matches. (8) "Harmonising 16 protocols is a multi-year committee process, and no agreed alternative metric exists to harmonise toward" is not in the letter. The sentence now only quotes, from the letter's concluding remarks, its call to promote harmonisation of preconditioning and exposure conditions in existing standards for accelerated carbonation testing. (9) "Yet the standard is what a permitting engineer must apply" is not in the ReCreate article (which lists EN 16757 as product category rules for environmental product declarations); the sentence now gives what the article argues for, contextual carbonation rates from condition assessment and new on-site measurements. (10) In the entry points, "carbonation alone does not corrode" stood in quotation marks as if it were the letter's wording; it is not, and it is now a plain paraphrase of the letter's sentence that "a reduced pH alone is not sufficient to lead to significant steel corrosion in concrete." (11) The loading figures were given as "under 30% compressive load" and "under 60% tensile load"; the letter says 30% failure load in compression and 60% of the tensile failure load, and the sentence now says so. Confirmed in the letter and left as written: active 2018 to 2024, over 120 members, five working groups; at least 16 standards or recommendations; the 22-laboratory test; the 1 to 4% CO₂ comparison; the 91-day and one-year findings; the same ranking of cement types across protocols; the database with several years of natural exposure; the equation and the water-to-reactive-CaO ratio; the recommendation for testing under load; the 9% to 16% and up to 70% loading figures; the corrosion-review passages; the low-calcium alkali-activated finding at 1% CO₂; EN 12390-10 and -12; and every other quotation. Confirmed in the ReCreate article and left as written: EN 16757:2022, the 1967 Swedish building, CEN/TR 17310, and the three quotations. Not checkable against these sources because they are the write-up's own framing rather than source claims: the opening sentence of "Why This Matters" (already marked as general knowledge) and the two adjacent precedents (metals fatigue and creep, pipeline corrosion). The underlying ReCreate journal paper was again not read. A second reader went over these repairs against both sources on the same day, fetching both again and testing every quotation in the body against the source text; all ten stand, two of them (8 and 10) in wording brought closer to the letter, and the second reader added item 11. Genome tags untouched.