Performance-Normalized Cradle-to-Gate Carbon Screening of Polymer-Modified Cementitious Repair Mortars Using Standardized Functional Units

Abstract

Polymer-modified cementitious repair mortars improve adhesion and durability but usually increase cradle-to-gate (A1–A3) emissions. This study develops a performance-normalized embodied-carbon screening framework that compares repair materials on standardized functional units (1 m³, 1 kg dry mix, 1 m² repair layer) and on declared repair functions. A database of 55 studies, 141 mixture records, and 243 endpoints was compiled, and 19 studies met predefined criteria for quantitative extraction. Two within-study case studies were analyzed, and emission factor and test-result uncertainties were propagated by Monte Carlo simulation. In epoxy-modified mortars (reconstructed inventory) exposed to sulfuric acid, modification reduced mass loss by 57.9–96.0% but increased emissions from 264 to 422–857 kg CO₂e/m³; the 5% resin mixture had the lowest incremental carbon cost and the 20% mixture the lowest deterioration-rate-normalized GWP in >99% of simulations. In styrene–butadiene rubber mortars, the cationic latex reduced the chloride migration coefficient by 31.5% and, with a regional EPD, lowered the chloride-normalized GWP by 26%; this advantage is lost above a latex emission factor of 2.66 kg CO₂e/kg. In a geopolymer benchmark, a declared desirability index ranked one of the same two mixtures first in all predefined scenarios. The framework supports screening rather than definitive ranking.

1. Introduction

Repair mortars are fundamental to restoring deteriorated concrete surfaces, protecting reinforcement, improving local load transfer, and extending the useful service life of existing infrastructure. Concrete is the principal material for infrastructure worldwide; global cement and concrete use continues to grow, particularly in rapidly developing regions [1]. However, concrete’s inherent brittleness and heterogeneous porous microstructure confer high permeability and crack sensitivity, making it vulnerable to environmental deterioration [2]. The resulting maintenance and replacement costs are substantial: in the United States, roughly 7.5% of 617,000 bridges have structural deficiencies, with an estimated cumulative repair demand of USD 125 billion [3].

In practice, the sustainability benefit of a repair intervention is not determined only by the embodied carbon of the repair material itself. A material with higher cradle-to-gate emissions may still reduce embodied carbon if it improves bond, lowers permeability, delays chloride or sulfate ingress, reduces shrinkage cracking, or extends the maintenance interval. This issue is especially relevant for polymer-modified systems because the polymer phase often contributes a relatively high material-level carbon intensity but also provides performance functions that ordinary cement-sand mortars may not achieve [4,5].

The construction sector is a major source of global energy use and greenhouse-gas emissions, and cement production is a dominant contributor to the embodied carbon of cementitious materials. Cement manufacturing is one of the most energy-intensive sectors, consuming approximately 12 EJ globally—about 3% of global energy consumption—and accounting for about 6% of global greenhouse-gas emissions [6]. The calcination process for clinker production accounts for a large portion of total emissions, while thermal energy and electricity use also contribute substantially [7]. The production of ordinary Portland cement, which contains more than 90% clinker, generates on average about 842 kg CO₂ per tonne of clinker [8]. With annual global concrete production expected to continue rising, decarbonization of the cement and concrete sector remains an urgent sustainability priority [9].

At the same time, aging infrastructure creates a growing need for repair, overlay, and rehabilitation materials. Traditional cement-based repair materials may suffer from shrinkage, long curing times, unstable strength development, and poor bonding at the repair–substrate interface [10]. The environmental question is therefore not only how to reduce the carbon footprint of each cubic meter of repair mortar, but how to minimize the carbon required to deliver a specified repair function for a specified design period.

Polymer-modified cementitious materials can improve adhesion, flexibility, cohesion, crack bridging, water resistance, chemical resistance, and durability [11,12]. Polymer modification is regarded as an effective route to enhance durability and interfacial performance of cement-based materials because the adhesion and film-forming ability of polymers can increase cohesion among phases, improve interfacial bonding, and enhance tensile strength, toughness, and long-term durability [13]. Widely used synthetic polymers include styrene–butadiene rubber, waterborne polyurethane, polyacrylate, vinyl acetate–ethylene copolymer powder, waterborne epoxy, and polyvinyl alcohol [14].

The extracted literature dataset shows a broad formulation space that includes water-dilutable epoxy resins, styrene–acrylic and styrene–butadiene latex systems, vinyl acetate–ethylene copolymer powders, polyvinyl alcohol, hydrophobic silane/silicone modifiers, self-healing epoxy microcapsules, hybrid fiber/polymer gel mortars, and alternative binders such as one-part alkali-activated mortars and metakaolin–slag geopolymers [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29].

This broad formulation space creates a methodological problem. Published studies report mixture proportions using inconsistent bases such as kg/m³, cement-to-sand ratios, polymer-to-cement ratios, water-to-binder ratios, percentage replacement, and orthogonal factor levels. Performance endpoints also differ across studies and include compressive strength, flexural strength, bond strength, setting time, flow value, capillary water absorption, chloride migration, carbonation resistance, sulfate resistance, shrinkage, mass loss, healing efficiency, and high-temperature residual strength. Without standardization, a mixture with high polymer content may appear environmentally unfavorable on a mass basis even if it provides substantial durability gains per repair layer or per unit service period.

Recent studies demonstrate that polymer-modified repair mortars can achieve significant performance improvements. For example, ion-exchange functionalized waterborne acrylate copolymer-modified mortar has been reported to improve interfacial shear strength, flexural bond strength, compressive strength, flexural strength, chloride migration resistance, and carbonation resistance [30]. A chloride-exposure study reported an increase in estimated service life from 14.9 years for a reference mortar to approximately 21–35 years for selected polymer-modified mortars [31]. These service-life values should not be generalized across all polymer-modified mortars, but they illustrate why durability benefits must be considered when evaluating repair-material sustainability. At the microstructural level, polymer-modified mortars disperse a polymer phase throughout the matrix, sealing pores and microcracks formed during hardening; beyond improved chemical resistance, polymers can also enhance workability at lower water-binder ratios, which may improve strength and durability [32].

Life-cycle assessment is standardized through ISO 14040 and ISO 14044 and is widely used to quantify environmental impacts in the construction sector [33,34]. However, repair-material life-cycle assessment has several methodological difficulties. First, system boundaries vary from cradle-to-gate production-only analysis to cradle-to-grave or consequential assessments that include maintenance and avoided replacement. Second, functional-unit selection strongly influences conclusions; one cubic meter of material does not necessarily represent equivalent repair service [35]. Third, durability data are often reported as laboratory mass loss, absorption, migration coefficient, or cycles to deterioration rather than as validated service life. Fourth, polymer products and admixtures frequently lack product-specific environmental declarations in experimental papers. Prior research has emphasized that the choice of functional unit can substantially alter the environmental ranking of mortars [36].

The broader performance-normalized carbon efficiency concept emphasizes service life, regional carbon factors, and performance-based functional units. Durability is a fundamental aspect of sustainability because reduced maintenance and restoration can reduce material demand, energy demand, downtime, and demolition waste [37]. Product-specific environmental product declarations for cementitious repair mortars are increasingly available, and this improves the feasibility of more robust performance-normalized carbon assessments [38,39].

This study addresses the above gap by combining database-driven extraction with a performance-normalized embodied-carbon screening framework. Unlike conventional comparisons that rank repair mortars only by kg CO₂e/m³ or kg CO₂e/kg, this study evaluates embodied carbon relative to declared repair functions, including strength, bond, repair-layer area, and laboratory durability improvement. The contribution is methodological and quantitative: it provides a transparent workflow for converting heterogeneous repair mortar data into standardized functional units and for evaluating embodied carbon together with governing mechanical and durability functions. Three elements distinguish the framework from existing functional-unit approaches (Section 2.4): (i) the durability-normalized indicators are calculated within each source study against its own control mixture, so that specimen size, curing, exposure, and test method are held constant; (ii) emission factor and test-result uncertainties are propagated jointly to estimate the probability that a ranking is retained; and (iii) the emission factor at which a polymer-modified mixture breaks even with its control is reported, which turns a missing product-specific environmental product declaration into an explicit decision threshold. The study is intentionally positioned as a screening embodied-carbon assessment, not as a definitive commercial product ranking or a full cradle-to-grave life-cycle assessment, because many polymer products in the published literature do not report product-specific environmental declarations and laboratory durability endpoints cannot be converted directly into field service life without validation.

2. Literature Background and Research Gaps

2.1. Repair Mortar Technologies

Modern repair mortars can be broadly classified into polymer-modified cementitious mortars, epoxy-based systems, alkali-activated/geopolymer systems, and unmodified cementitious systems. Polymer-modified cementitious mortars are attractive because they combine cementitious compatibility with improved adhesion, water retention, toughness, and permeability resistance. Epoxy-based systems can provide high adhesion and chemical resistance, but cost, thermal compatibility, vapor impermeability, and carbon intensity must be considered. Alkali-activated systems may reduce Portland cement content, but require careful control of activator content, shrinkage, efflorescence, curing, and early-age strength [14,22,26,27,28,29].

Previous research on repair materials has emphasized mechanical compatibility, substrate preparation, bond mechanisms, polymer–cement film formation, waterproofing, and long-term durability [40,41,42,43,44,45,46,47,48,49,50,51]. However, many studies focus on one or two performance indicators and do not connect the measured property improvement to environmental impact. This limits the ability of practitioners to select repair materials on a functionally equivalent basis. Table 1 summarizes the repair mortar technology classes and the functional-unit issues governing their environmental ranking.

Table 1. Repair mortar technology classes and the functional-unit issue that controls their environmental ranking.

System	Typical Advantages	Typical Limitations	Functional-Unit Emphasis
Unmodified cementitious mortar	Low material carbon and cost; high compatibility with concrete	Lower bond, permeability resistance, and chemical durability in aggressive exposure	kg CO2e/m3; kg CO2e/MPa; repair frequency
Polymer-modified cementitious mortar	Improved bond, water retention, toughness, crack bridging, impermeability, and durability	Higher initial carbon; possible strength reduction or delayed hydration at high dosage	kg CO2e per bond/durability gain; repair-layer GWP
Epoxy-rich repair systems	High adhesion, waterproofing, chemical resistance, and rapid repair potential	High polymer carbon and cost; compatibility and vapor movement concerns	Durability-normalized carbon; avoided replacement
Alkali-activated/geopolymer mortar	Low Portland cement content; potential carbon reduction; high early strength possible	Activator carbon, shrinkage, curing sensitivity, and field robustness	Binder-normalized GWP; strength–workability–shrinkage desirability

2.2. Polymer Types and Performance Mechanisms

Polymer modification affects cementitious mortars through polymer film formation, pore blocking, interfacial adhesion, improved cohesion, and crack-bridging mechanisms. Styrene–butadiene rubber, vinyl acetate–ethylene copolymer, acrylic emulsions, waterborne epoxy systems, polyurethane, and polyvinyl alcohol have all been reported to influence mechanical and transport behavior [23,24,25,27,52,53,54,55,56,57,58,59,60,61,62,63,64,65]. The direction and magnitude of property change depend on polymer type, polymer-to-cement ratio, water-to-binder ratio, curing condition, substrate moisture state, and compatibility with cement hydration.

Recent research has clarified mechanisms underlying polymer-induced performance enhancement. Ion-exchange reactions between waterborne acrylate copolymers and cement hydration products can form chemically bonded interpenetrating networks that densify the microstructure and strengthen interfacial adhesion [30].

A recurring result in the polymer-modified mortar literature is that higher polymer dosage does not automatically produce a better repair material. Small additions can substantially improve bond or durability, while further additions may introduce air, delay hydration, reduce compressive strength, increase carbon intensity, or cause diminishing returns. For this reason, the design target should be the minimum polymer dosage that satisfies the governing repair function rather than the maximum polymer content.

Research has shown that low polymer dosages may not form continuous polymer films, leaving the interfacial transition zone more connected and crack-sensitive under freeze–thaw cycling [31]. In contrast, selected mid-to-high polymer dosages can improve freeze–thaw resistance and transport blocking. The optimum dosage is therefore exposure-specific and should be linked to performance-normalized environmental indicators rather than to polymer content alone. Table 2 summarizes the polymer classes and suitable performance-normalized indicators.

Table 2. Polymer classes and performance-normalized indicators suitable for repair mortar assessment.

Modifier Class	Common Performance Contribution	Potential Risk at High Dosage	Best Normalized Metric
Styrene–butadiene rubber latex	Bond improvement, water resistance, flexibility, toughness [43,50,51,52]	Compressive-strength reduction, entrained air, delayed hydration [52,53,56]	kg CO2e per bond or permeability improvement
Vinyl acetate–ethylene/redispersible powder	Flexibility, adhesion, workability, compatibility in repair mortars [5,43,55]	Lower early strength or increased setting time depending on dosage [12,58]	kg CO2e per bond/workability-balanced score
Acrylic/styrene–acrylic	Waterproofing, adhesion, weathering resistance, permeability reduction [30,46,57]	Carbon increase and possible air entrainment [57,58]	kg CO2e per absorption reduction
Waterborne epoxy	Chemical resistance, high adhesion, dense polymer–cement network [32,59,61,62]	High embodied carbon, cost, and possible compatibility issues [47,60,66]	Incremental kg CO2e per durability gain; deterioration-rate-normalized GWP
Polyvinyl alcohol and fiber/polymer hybrids	Toughness, crack control, post-cracking behavior, self-healing support [24,27,65]	Workability loss, cost, and complex mixture design [24,27]	kg CO2e per toughness or healing efficiency
Silane/silicone hydrophobic modifiers	Water repellency and corrosion protection through reduced ingress [23,26]	May affect vapor movement or bonding to subsequent layers [23,25]	kg CO2e per absorption/chloride reduction

2.3. Life-Cycle Assessment Limitations in Repair Materials

As outlined in the Introduction, four methodological difficulties limit life-cycle assessment of repair materials: variable system boundaries, the strong influence of the functional unit, the use of laboratory durability proxies instead of validated service life, and the lack of product-specific environmental declarations for polymer products [33,34,35].

The broader performance-normalized carbon efficiency framework emphasizes service life, regional carbon factors, and performance-based functional units. Those concepts are retained in this manuscript as background and methodology, while numerical conclusions are kept conservative and tied to traceable extracted mixture and performance data. This avoids overstating generic polymer rankings that would require product-specific environmental product declarations, validated durability models, and consistent field-service assumptions.

2.4. Research Gaps and Objectives

Most repair mortar life-cycle assessment studies compare materials on a mass or volume basis rather than on equivalent repair function. Polymer-modified systems are often penalized for higher initial carbon without quantifying durability, adhesion, or service-life benefits. Many published mixtures cannot be compared directly because their mix designs are reported using different bases and incomplete inventories. There is also no widely adopted functional-unit framework for integrating embodied carbon, repair thickness, mechanical performance, bond, and durability proxies in polymer-modified repair mortars.

Several established approaches already relate the carbon footprint of cementitious materials to their function. Binder- and CO₂-intensity indices divide binder content or emissions by compressive strength [67]; mix-design studies compare carbon-mitigation strategies at equal strength class [68]; functional units that combine strength and service life have been shown to change the environmental ranking of concretes [35,37]; and environmental product declarations of commercial repair mortars report GWP per declared kilogram of product [38,39]. These approaches suit structural concrete, where compressive strength and a specified service life define equivalence. They are less suited to repair mortars, whose governing function is often bond, impermeability, or resistance to a specific aggressive agent, and whose durability evidence usually consists of short-term laboratory endpoints obtained with study-specific protocols. Table 3 compares these approaches with the framework proposed here; the novelty lies not in the functional-unit concept itself but in its within-study application to repair-specific durability endpoints, combined with probabilistic ranking and break-even emission factors.

Table 3. Comparison of existing function-based carbon assessment approaches with the framework of this study.

Approach	Functional Basis	Treatment of Durability	Treatment of Uncertainty	Limitation for Repair Mortars
CO2/binder intensity index [67]	kg CO2 or kg binder per MPa	Not included	Not included	Strength rarely governs repair performance
Equal-strength mix-design comparison [68]	1 m3 at a given strength class	Not included	Scenario comparison	Requires alternatives of equal strength
Service-life-based functional unit [35,37]	1 m3 × service life	Modeled service life	Scenario or sensitivity analysis	Requires validated deterioration models
Product EPD [38,39]	1 kg declared product	Not included (A1–A3)	Not reported for rankings	Not comparable across repair functions or layer thicknesses
This study	1 m3, 1 kg dry mix, 1 m2 layer, and per declared repair function	Within-study laboratory proxies (ICE, GWPdur); Fick-based initiation-time ratio for chloride	Monte Carlo on emission factors and test results; rank probabilities; break-even emission factor	Screening only; laboratory proxies are not validated for service life

The objectives of this study are to:

• Structure a literature-based database of polymer-modified and advanced cementitious repair mortars;
• Define standardized functional units for volume, dry-mix, repair-layer, and performance-normalized comparisons;
• Calculate cradle-to-gate embodied-carbon indicators using product-specific environmental product declarations or regional life-cycle inventory values where available;
• Demonstrate performance-normalized carbon analysis through two within-study case studies: epoxy-modified mortars under sulfuric acid exposure and SBR-modified repair mortars under chloride and freeze–thaw exposure;
• Quantify the effect of emission factor and test-result uncertainty on the indicators and rankings, together with repair-layer thickness effects; and
• Show how transparent multi-criteria desirability ranking, including weight and normalization sensitivity, can be applied to alternative-binder mixtures.

No mathematical optimization model (an objective function with constraints solved over a continuous design space) is formulated in this study. The terms “best-ranked” and “preferred” refer to the highest-ranked alternative among the tested mixtures under an explicitly declared decision rule.

3. Scope, System Boundary, and Database Construction

3.1. System Boundary and Assessment Position

The numerical examples use a screening cradle-to-gate material boundary. The analysis includes raw-material quantities multiplied by constituent emission factors but excludes transport to site, substrate preparation, mixing equipment, curing energy, installation losses, future maintenance, demolition, end-of-life treatment, and avoided replacement of the parent concrete element. In the modular notation of EN 15804 [69], the calculated values correspond to modules A1–A3 only. These stages are important for final comparative or consequential life-cycle assessment; however, they cannot be extracted consistently from the available experimental literature. Therefore, the results are interpreted as embodied-carbon screening indicators for material selection rather than complete life-cycle rankings. Statements about repair recurrence, maintenance, or avoided replacement would require modules B3–B5 and C1–C4; in this study they appear only as conditional break-even checks (Section 5.3 and Section 5.4 and Appendix B), not as life-cycle results.

3.2. Systematic Database-Construction Protocol

The database was constructed using a reproducible screening protocol. Searches were performed in publisher platforms and scholarly databases including ScienceDirect, SpringerLink, MDPI, Taylor & Francis, Google Scholar, Crossref/DOI look-up pages, and EPD program/operator databases. Search strings combined repair-material terms with polymer and sustainability terms, including: “polymer-modified repair mortar”, “cementitious repair mortar”, “waterborne epoxy cement mortar”, “SBR latex mortar”, “acrylic modified mortar”, “VAE redispersible powder mortar”, “PVA modified mortar”, “alkali-activated repair mortar”, “geopolymer repair mortar”, “sulfate resistance”, “acid resistance”, “bond strength”, “shrinkage”, “water absorption”, “life-cycle assessment”, “environmental product declaration”, and “global warming potential”.

Records were retained in the index when they reported at least one usable mixture proportion, modifier dosage, repair mortar functional endpoint, durability endpoint, or environmental factor. Records were excluded when they lacked mixture proportions, reported only qualitative discussion, focused on non-repair bulk concrete without transferable mortar information, duplicated another source, or used performance endpoints that could not be connected to a functional unit. Screening was performed in four stages: identification of candidate sources, title/abstract screening, full-text eligibility checking, and data extraction. The resulting index contains 55 studies. A study was selected for curated quantitative extraction only when it met all of the following criteria: (C1) a DOI-traceable, peer-reviewed source; (C2) numerical values, not only graphical results, for the quantities to be extracted; and (C3) at least one of (a) a complete or reconstructable mixture inventory, (b) a mechanical, bond, or durability endpoint reported for modified mixtures together with a control mixture tested under the same protocol, or (c) a material-level emission factor or binder information needed for carbon calculation. Eighteen studies satisfied C1–C3 in the original curation; the SBR mortar study used for the second case study [70], which belongs to the index, was added to the curated subset during the revision because it reports chloride migration and freeze–thaw endpoints together with a control mixture (Table 4). The other 36 indexed studies were used only for the qualitative synthesis in Section 2 and Section 5.6.

Table 4. Curated studies with authors and DOI links used for the database-driven analysis.

ID	Year	Authors	Contribution to Database	DOI Link
S002	2026	W. Kostrzewski; I. Laks; M. Sybis	Water-soluble epoxy resins as an innovative method of protecting concrete against sulfate corrosion [15]	https://doi.org/10.3390/ma19020364
S003	2024	J. Szymanowski; L. Sadowski	Mechanical performance of high-strength nano-modified cement mortars for overlays [16]	https://doi.org/10.3390/buildings14010248
S004	2025	N. Z. Muhammad et al.	Waterproofing performance of cement mortar incorporating nano silicon [17]	https://doi.org/10.3390/buildings15132227
S009	2026	G. D. M. Pillai et al.	Rice husk ash for flow and mechanical properties of ultra-high-strength mortar [18]	https://doi.org/10.3390/engproc2026130001
S010	2026	Y. Li; Y. Deng	Low-temperature self-healing cement mortar enabled by composite microcapsules [19]	https://doi.org/10.3390/ma19050933
S012	2022	A. Szewczak; G. Łagód	Adhesion of modified epoxy resin to a concrete surface [20]	https://doi.org/10.3390/ma15248961
S016	2026	S. Çalbıyık; N. Kabay; T. Omur; H. Ozkan	Modifiers in one-part alkali-activated rapid-hardening repair mortar [21]	https://doi.org/10.3390/su18147047
S022	2025	Y. Yin; Y. Mei	Influence of polymers on cement-based coating materials [14]	https://doi.org/10.3390/ma18143321
S030	2026	Y. Guo et al.	Early-strength admixtures in repair mortar containing recycled concrete powder and GGBFS [22]	https://doi.org/10.3390/buildings16122421
S035	2026	L. Wang et al.	Polymer-modified high-belite sulfoaluminate cement repair mortar [12]	https://doi.org/10.3390/buildings16122352
S036	2025	Z. Li; C. Guo; L. Xu; R. Wang	Waterproofing, impermeability, and vapor transmission of mortar with modifiers [23]	https://doi.org/10.3390/ma18102363
S040	2026	S. Jing et al.	Review on polymer-modified cementitious materials for underwater repair [13]	https://doi.org/10.3390/buildings16142751
S044	2026	H. Song et al.	PVA-modified NHL-based mortars for restoration of historical buildings [24]	https://doi.org/10.3390/ma19081567
S047	2024	A. Karozou; F. Kesikidou; E. Pavlidou; M. Stefanidou	Alkali activator and hydrophobic agents in clay-based mortars [25]	https://doi.org/10.3390/min14101035
S048	2026	Y. Luo et al.	Polyether-modified silicone oil micelles for waterproofing and corrosion resistance [26]	https://doi.org/10.3390/polym18101153
S049	2026	P. Zhang et al.	Hybrid PVA-steel fiber in nano-SiO2-modified epoxy-resin gel mortar [27]	https://doi.org/10.3390/gels12050424
S051	2026	Z. Zeng; P. Yu; Z. Chen; J. Zhou; H. Xin; L. Yu; A. Lin	Magnesium ammonium phosphate cement modified with gold tailings [28]	https://doi.org/10.3390/ma19122684
S052	2026	P. Liang; L. Zhu; R. Yin; R. Gao	Optimization of metakaolin–slag geopolymer mortar based on orthogonal experiment [29]	https://doi.org/10.3390/ma19102004
Added	2026	A. Xu et al.	Emulsifier type in SBR-modified cement-based materials; chloride migration and freeze–thaw [70]	https://doi.org/10.3390/polym18091128

For each retained source, extraction rules were applied consistently. Mixture quantities were recorded in their original form, units were preserved, and conversion eligibility was classified as direct 1 m³ inventory, reconstructable 1 m³ inventory, 1 kg dry-mix normalization, repair-layer scaling, performance-only extraction, or unsuitable for quantitative normalization. Extracted fields included study ID, authors, year, DOI, material system, modifier type and dosage, cement/binder/sand/aggregate/water quantities, functional unit, performance endpoint, exposure condition, data-source table or figure, emission factor status, and notes on assumptions. When product-specific environmental product declarations were unavailable, the factor was labeled as a proxy and retained only for sensitivity analysis.

The curated subset is purposive rather than exhaustive: it was assembled to demonstrate and test the normalization workflow, not to estimate an average carbon or performance effect of polymer modification. All 19 curated studies are open-access MDPI articles published between 2022 and 2026. This reflects criterion C2, because recent open-access articles more often provide complete numerical mixture tables, but it also introduces publisher and recency bias, so the subset should not be regarded as representative of the whole field. Two features limit the effect of this bias on the conclusions. First, every performance-normalized indicator is calculated within a study relative to that study’s own control mixture; no absolute performance values are pooled across studies. Second, the quantitative conclusions concern the behavior of the indicators and decision rules in two fully documented case studies, not the prevalence of any polymer family. The database-level counts (Figure 1 and Table 5) describe the extracted literature only.

Figure 1. Distribution of extracted mixture rows by modifier category.

Table 5. Most frequent performance endpoints in the extracted database.

Most Frequent Extracted Endpoint	Rows
Compressive strength	41
Flexural strength	27
Fluidity	16
28 d flexural strength	16
28 d compressive strength	16
28 d compressive-to-flexural ratio	16
Shrinkage	16
Capillary water absorption	13
Initial cracking load	9
Peak load	9
Ultimate deflection	9
Flow value	7

3.3. Extracted Database Structure

The structured dataset contains 55 indexed studies, 141 mixture/inventory records, and 243 mechanical, bonding, durability, and mixture-design endpoints. Each curated row includes the study identifier, authors, publication year, DOI URL, material system, original functional unit, key mixture quantities, modifier type and dosage, data-source table/figure, normalization usefulness, and notes. The Supplementary Materials provide the input data, emission factors, and Python (version 3.13.2) scripts that reproduce all case-study calculations, the Monte Carlo analysis, the geopolymer ranking, and the figures; the extracted study index is available from the author on request. Figure 1 shows the distribution of extracted mixture rows by modifier category, and Table 4 lists the curated studies used in the analysis.

4. Standardized Functional-Unit and Carbon Methodology

4.1. Functional-Unit Conversion

Four functional-unit levels were used. The first is a volume basis of 1 m³ of fresh or hardened repair material, preferred when a study directly reports kg/m³ or when density can be reconstructed. The second is 1 kg of dry mix, useful for ratio-based mortars where density is absent. The third is a repair-area basis, defined as 1 m² of repair layer with thickness t. The fourth is a performance-normalized functional unit, in which embodied carbon is divided by a strength, bond, healing, or durability function.

Standardizing the functional unit makes inventories and carbon values comparable, but it does not make performance data obtained with different specimen sizes, curing regimes, exposure durations, and test standards directly comparable. For this reason, performance values are never compared in absolute terms across studies. Each mechanical or durability indicator (Equations (4)–(9)) is computed relative to the control mixture of the same study, which was cast, cured, exposed, and tested with the same protocol, so that protocol-specific effects largely cancel in the ratio P_i/P₀. Rankings are reported only among mixtures of the same study, and the test conditions of each case study are stated in Section 5.2 and Section 5.3. Cross-study statements in Section 5.6 are qualitative.

$${\mathrm{m}}_{\mathrm{i},\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}}={\mathrm{m}}_{\mathrm{i},{\mathrm{m}}^3}\times \mathrm{t}$$

(1)

$$\mathrm{G}\mathrm{W}{\mathrm{P}}_{\mathrm{m}\mathrm{i}\mathrm{x}}=\sum ({\mathrm{m}}_{\mathrm{i}}\times \mathrm{E}{\mathrm{F}}_{\mathrm{i}})$$

(2)

$$\mathrm{G}\mathrm{W}{\mathrm{P}}_{\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}}=\mathrm{G}\mathrm{W}{\mathrm{P}}_{\mathrm{m}\mathrm{i}\mathrm{x}}\times \mathrm{t}$$

(3)

where m_i,area is the mass of constituent i required for a repair layer (kg/m²), m_i,m³ is the mass of constituent i per cubic meter (kg/m³), t is repair thickness in meters, GWP_mix is the cradle-to-gate global warming potential per cubic meter (kg CO₂e/m³), GWP_area is the corresponding value per square meter of repair layer (kg CO₂e/m²), and EF_i is the emission factor of constituent i in kg CO₂e/kg. Table 6 presents the emission factors used in the calculations, and Table A1 summarizes the functional-unit equations for spreadsheet implementation.

4.2. Performance-Normalized Indicators

For strength-governed repair functions, the simplest normalized indicators are carbon intensity per unit compressive strength or bond strength:

$$\mathrm{C}{\mathrm{I}}_{\mathrm{f}\mathrm{c}}=\mathrm{G}\mathrm{W}{\mathrm{P}}_{\mathrm{m}\mathrm{i}\mathrm{x}}/{\mathrm{f}}_{\mathrm{c}}^{\prime }$$

(4)

$$\mathrm{C}{\mathrm{I}}_{\mathrm{b}\mathrm{o}\mathrm{n}\mathrm{d}}=\mathrm{G}\mathrm{W}{\mathrm{P}}_{\mathrm{m}\mathrm{i}\mathrm{x}}/{\mathsf{\tau }}_{\mathrm{b}\mathrm{o}\mathrm{n}\mathrm{d}}$$

(5)

For deleterious durability metrics such as mass loss, penetration depth, water absorption, shrinkage, or chloride migration, lower values are better. The relative improvement of mixture i against the control mixture 0 is calculated as

$${\mathrm{D}}_{\mathrm{i}}={\displaystyle \frac{{\mathrm{P}}_0-{\mathrm{P}}_{\mathrm{i}}}{{\mathrm{P}}_0}}\times 100$$

(6)

where P is the deleterious performance metric. The incremental carbon required to obtain one percentage point of durability improvement is

$$\mathrm{I}\mathrm{C}{\mathrm{E}}_{\mathrm{i}}={\displaystyle \frac{\mathrm{G}\mathrm{W}{\mathrm{P}}_{\mathrm{i}}-\mathrm{G}\mathrm{W}{\mathrm{P}}_0}{{\mathrm{D}}_{\mathrm{i}}}}$$

(7)

A second durability-normalized screening indicator was used when the measured deleterious response can be interpreted as a proxy for deterioration rate:

$$\mathrm{G}\mathrm{W}{\mathrm{P}}_{\mathrm{d}\mathrm{u}\mathrm{r},\mathrm{i}}=\mathrm{G}\mathrm{W}{\mathrm{P}}_{\mathrm{i}}\times ({\mathrm{P}}_{\mathrm{i}}/{\mathrm{P}}_0)$$

(8)

The GWP_dur indicator is not intended to represent a direct service-life prediction, because the relationship between accelerated laboratory mass loss and field deterioration requires long-term validation. Rather, it is used as a screening metric to estimate the embodied carbon associated with an equivalent laboratory deterioration burden, under the simplifying assumption that measured mass loss is proportional to deterioration rate.

For diffusion-controlled chloride ingress, the error-function solution of Fick’s second law gives a time to reach a critical chloride concentration at a given cover depth that is inversely proportional to the apparent diffusion coefficient, provided that the surface concentration, critical concentration, and cover depth are equal [71]. Under these assumptions, the ratio of the corrosion-initiation times of mixture i and control 0 is

$${\mathrm{t}}_{\mathrm{i}}/{\mathrm{t}}_0={\mathrm{D}}_0/{\mathrm{D}}_{\mathrm{i}}$$

(9)

where t is the initiation time, and D is the chloride migration (or diffusion) coefficient. Equation (9) neglects aging of the diffusion coefficient, differences in chloride binding, and cracking; it is therefore used as a conditional indicator only. When the migration coefficient is used as P in Equation (8), GWP_dur is the material-stage GWP per unit of initiation time relative to the control, and modification is carbon-favorable when t_i/t₀ exceeds the carbon multiplier GWP_i/GWP₀.

4.3. Conservative Use of the PNCE Concept

A performance-normalized carbon efficiency (PNCE) index, in which embodied carbon is divided by a composite performance index combining strength, bond, service life, and durability, is useful as a design logic, but the present manuscript uses it conservatively. Multiplying several performance ratios can over-amplify differences when variables are correlated or estimated from different studies. Therefore, the main quantitative conclusions rely on single-function indicators and transparent multi-criteria desirability rather than universal polymer-category rankings.

4.4. Uncertainty Propagation

Input uncertainty was propagated by Monte Carlo simulation with 100,000 draws (random seed 2026). The cement factor followed a triangular distribution with mode and lower limit equal to the net A1–A3 value of the regional EPD (0.523 kg CO₂e/kg [72]; the gross value including emissions from waste-derived fuels, 0.700 kg CO₂e/kg, lies within the sampled range) and upper limit equal to the average clinker emission of 842 kg CO₂/t [8] multiplied by a clinker factor of 0.95 for CEM I (0.800 kg CO₂e/kg). The waterborne epoxy factor followed a triangular distribution with mode and lower limit equal to the product EPD proxy (2.75 kg CO₂e/kg [73]) and upper limit equal to the conservative model EPD (4.12 kg CO₂e/kg [66]); the SBR latex factor was treated in the same way, with mode and lower limit equal to a regional EPD of an SBR latex admixture (0.578 kg CO₂e/kg [74]) and upper limit equal to the model EPD for dispersion-based products (4.97 kg CO₂e/kg [75]). Database proxies for fly ash, aggregates, and water were sampled from lognormal distributions with a geometric standard deviation of 1.5, consistent with pedigree-based uncertainty factors for non-specific inventory data [76]. Measured responses were sampled from truncated normal distributions: acid mass loss with the specimen-level standard deviation reported for each mixture (n = 2), chloride migration coefficients with a coefficient of variation of 13%, equal to the inter-laboratory reproducibility of the non-steady-state migration method [77], and compressive strength with a coefficient of variation of 5%. For each draw, all indicators and the resulting rankings were recalculated. Results are reported as medians, 95% intervals (2.5th–97.5th percentiles), and the probability that each mixture is ranked first. These probabilities are conditional on the assumed input distributions; they measure the robustness of a ranking under these assumptions and are not general guarantees. A ranking retained in at least 95% of draws is described as robust, one retained in 50–95% of draws as conditional, and one retained in fewer than 50% as unresolved with the available data.

Table 6. Hybrid EPD/regional-LCI emission factors used in the calculations and the distributions assigned in the Monte Carlo analysis.

Material	Emission Factor	Unit	Status	Monte Carlo Distribution
Ordinary Portland cement, CEM I 42.5R	0.523	kg CO2e/kg	ITB EPD No. 703/2024, CEM I 42,5R/N-NA (Górażdże Cement); declared unit 1 t; A1–A3 GWP net total 523 kg CO2e/t (gross total 700 kg CO2e/t, including waste-derived fuels); issued 15.11.2024, valid to 15.11.2029 [72].	Triangular (0.523, 0.523, 0.800)
Fly ash	0.019	kg CO2e/kg	Database-based LCI; no matching product-specific EPD used in the screening calculation [78].	Lognormal, GSD 1.5
Fine aggregate/sand-gravel proxy	0.00103	kg CO2e/kg	ITB EPD No. 674/2024; program operator ITB; declared unit 1 t aggregate; A1–A3 GWP-total 1.03 kg CO2e/t for natural gravel/sand proxy; valid to 20.09.2029 [79].	Lognormal, GSD 1.5
Coarse aggregate/crushed basalt proxy	0.00328	kg CO2e/kg	ITB EPD No. 674/2024; program operator ITB; declared unit 1 t aggregate; A1–A3 GWP-total 3.28 kg CO2e/t for natural crushed aggregate proxy; valid to 20.09.2029 [79].	Lognormal, GSD 1.5
Water	0.0003	kg CO2e/kg	Negligible screening value; not ranking-sensitive	Lognormal, GSD 1.5
Waterborne epoxy product/dispersion	2.75	kg CO2e/kg	IBU EPD-SIK-20240380-CBA1-EN; program operator IBU; declared unit 1 kg Sikafloor 2510 W with packaging; A1–A3 GWP-total 2.75 kg CO2e/kg; valid to 03.03.2030 [73].	Triangular (2.75, 2.75, 4.12)
Epoxy product, conservative model	4.12	kg CO2e/kg	Deutsche Bauchemie/IBU model EPD for epoxy-resin products, group 4; declared unit 1 kg product; A1–A3 GWP-total 4.12 kg CO2e/kg; sensitivity case [66].	Upper limit of the epoxy distribution
SBR latex admixture	0.578	kg CO2e/kg	EPD-IES-0017048, Weberad 220 SBR (Saudi Arabia); program operator EPD International; declared unit 1 kg (1 L); A1–A3 GWP-total 0.578 kg CO2e; valid to 19.06.2030 [74]	Triangular (0.578, 0.578, 4.97)
Polymer dispersion, conservative model	4.97	kg CO2e/kg	Model EPD for dispersion-based products, group 4 (worst-case product of the group); sensitivity upper limit [75]	Upper limit of the SBR distribution

Table 6. Hybrid EPD/regional-LCI emission factors used in the calculations and the distributions assigned in the Monte Carlo analysis.

Material	Emission Factor	Unit	Status	Monte Carlo Distribution
Ordinary Portland cement, CEM I 42.5R	0.523	kg CO2e/kg	ITB EPD No. 703/2024, CEM I 42,5R/N-NA (Górażdże Cement); declared unit 1 t; A1–A3 GWP net total 523 kg CO2e/t (gross total 700 kg CO2e/t, including waste-derived fuels); issued 15.11.2024, valid to 15.11.2029 [72].	Triangular (0.523, 0.523, 0.800)
Fly ash	0.019	kg CO2e/kg	Database-based LCI; no matching product-specific EPD used in the screening calculation [78].	Lognormal, GSD 1.5
Fine aggregate/sand-gravel proxy	0.00103	kg CO2e/kg	ITB EPD No. 674/2024; program operator ITB; declared unit 1 t aggregate; A1–A3 GWP-total 1.03 kg CO2e/t for natural gravel/sand proxy; valid to 20.09.2029 [79].	Lognormal, GSD 1.5
Coarse aggregate/crushed basalt proxy	0.00328	kg CO2e/kg	ITB EPD No. 674/2024; program operator ITB; declared unit 1 t aggregate; A1–A3 GWP-total 3.28 kg CO2e/t for natural crushed aggregate proxy; valid to 20.09.2029 [79].	Lognormal, GSD 1.5
Water	0.0003	kg CO2e/kg	Negligible screening value; not ranking-sensitive	Lognormal, GSD 1.5
Waterborne epoxy product/dispersion	2.75	kg CO2e/kg	IBU EPD-SIK-20240380-CBA1-EN; program operator IBU; declared unit 1 kg Sikafloor 2510 W with packaging; A1–A3 GWP-total 2.75 kg CO2e/kg; valid to 03.03.2030 [73].	Triangular (2.75, 2.75, 4.12)
Epoxy product, conservative model	4.12	kg CO2e/kg	Deutsche Bauchemie/IBU model EPD for epoxy-resin products, group 4; declared unit 1 kg product; A1–A3 GWP-total 4.12 kg CO2e/kg; sensitivity case [66].	Upper limit of the epoxy distribution
SBR latex admixture	0.578	kg CO2e/kg	EPD-IES-0017048, Weberad 220 SBR (Saudi Arabia); program operator EPD International; declared unit 1 kg (1 L); A1–A3 GWP-total 0.578 kg CO2e; valid to 19.06.2030 [74]	Triangular (0.578, 0.578, 4.97)
Polymer dispersion, conservative model	4.97	kg CO2e/kg	Model EPD for dispersion-based products, group 4 (worst-case product of the group); sensitivity upper limit [75]	Upper limit of the SBR distribution

5. Results

5.1. Database-Level Observations

The extracted dataset is dominated by mechanical strength, flexural strength, flow/workability, shrinkage, water absorption, and specialized durability endpoints. Epoxy/waterborne epoxy systems, styrene–butadiene or acrylic latex systems, VAE-modified mortars, PVA-modified mortars, hydrophobic silane/silicone systems, nano-SiO₂/inorganic modifiers, and alkali-activated/geopolymer systems are all represented. This diversity confirms that a single performance index is not sufficient. A repair mortar designed for sulfate resistance should not be ranked only by compressive strength, and a self-healing mortar should not be ranked only by kg CO₂e/m³. Table 5 summarizes the most frequent performance endpoints in the extracted database.

5.2. Case Study 1: Epoxy-Modified Mortar Under Sulfuric Acid Exposure

The water-soluble epoxy-resin study was selected as the first numerical case because it reports mixture proportions and acid-exposure durability data [15]. The resin (a water-dilutable dispersion with approximately 41% solids) was dosed at 5%, 10%, and 20% resin solids by mass of cement, and the water carried by the dispersion was deducted from the mixing water. The acid mass-loss data used here (W-0 to W-20) are from the short-term test series of [15], which was carried out on standard cement mortar prisms rather than on the concrete cubes of the long-term series: two specimens per mixture were exposed for 6 months to sulfuric acid at pH 2, dried, brushed free of corroded material, and weighed. The source reports the per-1 m³ inventory only for its concrete mixtures (300 kg/m³ cement, 51.3 kg/m³ fly ash, 492.9 kg/m³ sand, and 1369.1 kg/m³ gravel up to 16 mm). The inventory of the tested mortar was therefore reconstructed for an assumed standard mortar according to EN 196-1 [80] (cement:sand:water = 1:3:0.5 by mass) by absolute volume with 2% air, giving 501 kg/m³ cement, 1504 kg/m³ sand, and 251 kg/m³ water for W-0 and 59.8, 116.9, and 223.8 kg/m³ of dispersion for W-5, W-10, and W-20 (Table 7). Because the emission factor of the EPD proxy is declared per kilogram of water-based product [73], it is applied to the dispersion mass. The concrete inventory of [15] is retained as a sensitivity case. Figure 2 shows the cradle-to-gate GWP and the measured mass loss, and Table 7 presents the performance-normalized carbon results.

Table 7. Inventory (kg/m³) and performance-normalized carbon results for the epoxy-modified mortars (reconstructed standard-mortar inventory, central emission factors).

Mix	Resin Solids/Cement (%)	Cement	Sand	Water	Dispersion	GWP (kg CO2e/m3)	GWP, 25 mm Layer (kg CO2e/m2)	Mass Loss (%)	D (%)	ICE (kg CO2e/m3 per % Point)	GWPdur (kg CO2e/m3)
W-0	0	501	1504	251	0.0	263.8	6.60	8.591 ± 0.381	0.0	—	263.8
W-5	5	490	1470	210	59.8	422.2	10.55	3.619 ± 0.247	57.9	2.74	177.8
W-10	10	479	1437	171	116.9	573.4	14.34	1.685 ± 0.312	80.4	3.85	112.4
W-20	20	459	1376	97	223.8	856.7	21.42	0.343 ± 0.133	96.0	6.17	34.2

Figure 2. Epoxy-modified mortars: (a) cradle-to-gate GWP of the reconstructed standard-mortar inventory (bars: central EPD/LCI factors; points and whiskers: Monte Carlo median and 95% interval); (b) mass loss after 6 months of sulfuric acid exposure (mean ± standard deviation, n = 2) [15].

The earlier version of this analysis combined the concrete inventory of [15] with the mortar durability data. Because the mass-loss data were obtained on mortar, the mortar inventory is now used as the primary case, so that the carbon and durability values refer to the same material. The EN 196-1 proportions are an assumption, because the source describes the specimens as standard mortar without listing their proportions; the resulting inventory is therefore an assumed scenario, not the experimentally reported composition. With the concrete inventory, the absolute GWP values are about 38% lower (163–565 kg CO₂e/m³), but all rankings are identical (Table A2). This is expected, because in both inventories the non-resin contributions change in proportion to the resin dosage and therefore affect all modified mixtures by the same fraction (Section 5.4).

The carbon-only ranking favors W-0 because it has no epoxy-resin contribution. However, W-0 also has the highest acid-exposure mass loss. Under the central EPD/LCI factor set, W-5 increases GWP by approximately 158 kg CO₂e/m³ while reducing mass loss by 57.9%, giving the lowest incremental carbon cost of approximately 2.74 kg CO₂e/m³ per percentage point of mass-loss reduction. W-10 reduces mass loss by 80.4% at an incremental cost of approximately 3.85 kg CO₂e/m³ per percentage point. W-20 delivers the greatest measured durability benefit, reducing mass loss by 96.0%, but has the highest incremental carbon cost per percentage point (6.17) because of the large resin-product dosage.

When the deterioration-rate-normalized carbon indicator is used, the interpretation changes. Because W-20 has very low measured mass loss, its GWP_dur value is the lowest among the four mixtures. This result demonstrates why repair-material assessment depends on the selected functional unit. If the decision rule is minimum initial embodied carbon, W-0 is preferred. If the decision rule is maximum durability gain per kg of additional carbon, W-5 is preferred. If the decision rule is to minimize a laboratory deterioration-rate proxy, W-20 is preferred. For balanced decisions, W-10 is a practical compromise because it captures most of the durability improvement of W-20 while avoiding roughly half of the incremental resin-related carbon burden. Mass loss under accelerated acid exposure is not proportional to field deterioration rate in general, so these results should not be interpreted as service-life evidence without exposure-specific validation. Figure 3 shows the uncertainty of both indicators (Section 5.4).

Figure 3. Monte Carlo distributions (100,000 draws; boxes: interquartile range; whiskers: 95% interval) of (a) the incremental carbon efficiency ICE and (b) the deterioration-rate-normalized indicator GWP_dur for the epoxy-modified mortars; diamonds show the deterministic values of Table 7.

5.3. Case Study 2: SBR-Modified Repair Mortar Under Chloride and Freeze–Thaw Exposure

The second case uses the SBR-modified repair mortars of Xu et al. [70], who compared latexes stabilized with anionic (MA), cationic (MB), and nonionic (MC) emulsifiers at SBR emulsion dosages of 10% (suffix -2) and 15% (suffix -3) by mass of cement. According to the mix table of the source, the mortars contain cement and sand in a mass ratio of 1:3 and have a total water-to-cement ratio of 0.45; the emulsion contains about 60% solids, and its water was deducted from the mixing water (added water: 45, 41, and 39 parts per 100 parts of cement). Strength was measured on 160 × 40 × 40 mm prisms, the chloride migration coefficient D_RCM on Φ100 × 50 mm disks by a non-steady-state migration test, and freeze–thaw resistance on 70.7 mm cubes after 100 cycles between −20 °C and +20 °C. All specimens were cured at 20 ± 2 °C and a relative humidity of at least 95%, and at least three specimens were tested per group [70]. Chloride data are available for the 10% mixtures and freeze–thaw data for the 15% mixtures, so the two exposures are evaluated separately, each against the same control M0.

Because the source reports mixture ratios rather than masses per cubic meter, inventories were reconstructed by absolute volume with particle densities of 3100 (cement), 2650 (sand), and 1000 kg/m³ (water and polymer solids) and an assumed air content of 2% (Table 8). The text of the source states a cement-to-sand ratio of 1:2, whereas its mix table gives 1:3; the table was used, and the 1:2 ratio changes the break-even latex factor of the cationic mixture only from 2.66 to 2.71 kg CO₂e/kg. The latex was assigned the regional EPD of an SBR latex admixture produced in Saudi Arabia (0.578 kg CO₂e/kg [74]); the solids content of that product is not declared, so the conservative dispersion model EPD is used as the upper limit.

Table 8. Reconstructed inventories (kg/m³), cradle-to-gate GWP, and strength-normalized carbon intensity of the SBR-modified mortars [70] (central emission factors).

Mix	Emulsifier	P/C (%)	Cement	Sand	Water	Latex	GWP (kg CO2e/m3)	fc (MPa)	CIfc (kg CO2e/m3 per MPa)
M0	—	0	514	1544	232	0	271	38.5	7.03
MA-2	anionic	10	499	1496	204	50	291	40.1	7.26
MB-2	cationic	10	499	1496	204	50	291	41.2	7.07
MC-2	nonionic	10	499	1496	204	50	291	40.0	7.28
MA-3	anionic	15	491	1474	192	74	301	37.0	8.14
MB-3	cationic	15	491	1474	192	74	301	38.0	7.92
MC-3	nonionic	15	491	1474	192	74	301	37.5	8.03

At 10% emulsion dosage, latex addition increases the material-stage GWP from 271 to 291 kg CO₂e/m³ (carbon multiplier 1.08), while compressive strength increases only slightly, from 38.5 to 40.0–41.2 MPa; the strength-normalized intensity CI_fc therefore remains close to that of the control. The chloride migration coefficient decreases by 19.4% (anionic), 31.5% (cationic), and 29.7% (nonionic). The cationic mixture MB-2 gives the lowest incremental carbon cost (0.65 kg CO₂e/m³ per percentage point, against 0.69 for MC-2 and 1.06 for MA-2) and, according to Equation (9), extends the corrosion-initiation period by a factor of 1.46, which clearly exceeds its carbon multiplier. Its chloride-normalized GWP is therefore 26% lower than that of the control (200 vs. 271 kg CO₂e/m³; Table 9).

Table 9. Durability-normalized indicators of the SBR-modified mortars: chloride migration (10% mixtures) and freeze–thaw strength loss after 100 cycles (15% mixtures), each relative to the control M0.

Mix	Exposure	P (Control → Mix)	Reduction D (%)	ICE (kg CO2e/m3 per % Point)	GWPdur (kg CO2e/m3)	ti/t0 (Equation (9))	GWPi/GWP0	Break-Even Latex EF (kg CO2e/kg)
MA-2	Chloride, DRCM	7.47 → 6.02	19.4	1.06	235	1.24	1.08	1.47
MB-2	Chloride, DRCM	7.47 → 5.12	31.5	0.65	200	1.46	1.08	2.66
MC-2	Chloride, DRCM	7.47 → 5.25	29.7	0.69	205	1.42	1.08	2.46
MA-3	Freeze–thaw, strength loss	16.0 → 9.2%	42.5	0.71	173	—	1.11	2.88
MB-3	Freeze–thaw, strength loss	16.0 → 7.2%	55.0	0.55	136	—	1.11	4.66
MC-3	Freeze–thaw, strength loss	16.0 → 7.8%	51.2	0.59	147	—	1.11	4.03
M0	Control	—	0	—	271	1.00	1.00	—

This advantage depends on the latex emission factor. Figure 4b shows that MB-2 remains below the control while the latex factor is below 2.66 kg CO₂e/kg; the corresponding break-even factors are 2.46 kg CO₂e/kg for the nonionic and 1.47 kg CO₂e/kg for the anionic mixture. Under freeze–thaw exposure at 15% dosage, the compressive-strength loss after 100 cycles decreases from 16.0% to 7.2–9.2%, and the break-even latex factors are higher (2.88–4.66 kg CO₂e/kg), again with the cationic mixture ranked first. The regional EPD value lies well below all break-even factors, whereas the conservative dispersion model EPD lies above them. The SBR case therefore shows a pattern different from that of the epoxy case: at a low polymer dosage the carbon penalty is small, and the durability gain is large, but whether modification is carbon-favorable depends on the product-specific latex factor, which the break-even factor makes explicit.

Figure 4. SBR-modified repair mortars [70]: (a) chloride migration coefficient at 28 days; (b) chloride-normalized GWP (Equation (8) with D_RCM as P) as a function of the latex emission factor. Dashed line: control M0; dots: break-even factors; shaded band: range between the latex EPD and the conservative dispersion model EPD.

5.4. Uncertainty Propagation, Emission Factor Sensitivity, and Repair-Thickness Effects

Mixing emission factors from different EPDs, databases, and regions changes the absolute GWP values but has a limited effect on the within-study rankings, and this can be shown analytically. In the reconstructed epoxy mortars, fly ash is absent, and adding resin displaces cement, sand, and water in fixed proportion to the dispersion mass (about 0.19, 0.57, and 0.69 kg per kg of dispersion). The non-resin terms therefore do not cancel exactly, but GWP_i − GWP₀ = m_ep,i·(EF_ep − 0.19·EF_cem − 0.57·EF_sand − 0.69·EF_w), so that they offset the same fraction of the resin contribution for every mixture (3.7% with the central factors). Equation (7) thus becomes ICE_i = m_ep,i·(EF_ep − 0.19·EF_cem − …)/D_i, and the ICE ranking is determined by the ratio m_ep,i/D_i (1.032, 1.454, 2.331 kg/m³ per percentage point for W-5, W-10, and W-20) for any combination of emission factors for which the bracket is positive, which holds throughout the sampled ranges. For GWP_dur, solving GWP_dur,i = GWP_dur,j for the epoxy factor gives a negative value for both adjacent pairs (W-20/W-10 and W-10/W-5), so no admissible factor reverses the order. In the SBR case, the ranking against the control depends on the latex factor, which is why the break-even factor is reported.

Table 10 and Figure 3 summarize the Monte Carlo results. Joint uncertainty widens the intervals of the epoxy indicators (for example, GWP of W-20: 880–1165 kg CO₂e/m³; ICE of W-5: 2.60–4.06 kg CO₂e/m³ per percentage point), but W-5 had the lowest ICE in 99.998% of draws, W-20 had the lowest GWP_dur in 99.9%, and the complete GWP_dur order W-20 < W-10 < W-5 < W-0 was retained in 99.6%. Separate runs show that the small residual overlap is caused by mass-loss variability (order retained in 99.6% of draws with emission factors fixed) rather than by emission factor uncertainty (100.0% with mass loss fixed), and the concrete inventory gives the same probabilities within 0.1 percentage point. Both epoxy rankings are therefore robust. In the SBR case, MB-2 was ranked first by ICE in 49.5% of draws and MC-2 in 39.8%, because the difference between their migration coefficients (5.12 vs. 5.25 × 10⁻¹² m²/s) is small compared with the 13% test reproducibility; this choice is unresolved with the available data. MB-2 satisfied the break-even condition t_i/t₀ ≥ GWP_i/GWP₀ in 71.9% of draws, and the control had the lowest chloride-normalized GWP in 17.7%, mainly in draws from the upper tail of the latex-factor distribution; the carbon benefit of SBR modification is therefore conditional on a low product-specific latex factor.

Table 10. Monte Carlo results (100,000 draws): medians, 95% intervals, and probability of being ranked first under each decision rule.

Mix	GWP (kg CO2e/m3)	Reduction D (%)	ICE (kg CO2e/m3 per % Point)	GWPdur (kg CO2e/m3)	P (First by ICE) (%)	P (First by GWPdur) (%)
W-0	304 (266–381)	—	—	304.4 (265.7–380.9)	—	0.0
W-5	491 (434–574)	57.9 (50.8–64.3)	3.15 (2.60–4.06)	207.2 (167.8–258.3)	100.0	0.0
W-10	667 (591–774)	80.4 (72.9–87.6)	4.43 (3.71–5.62)	130.8 (81.5–187.1)	0.0	0.1
W-20	991 (880–1165)	96.0 (92.9–98.9)	7.08 (6.13–8.83)	39.8 (10.8–72.2)	0.0	99.9
M0	313 (273–391)	—	—	313 (273–391)	—	17.7
MA-2	405 (312–540)	—	4.10 (0.78–50.24)	327 (205–524)	10.7	8.0
MB-2	405 (312–540)	—	2.77 (0.60–16.42)	278 (175–445)	49.5	41.2
MC-2	405 (312–540)	—	2.92 (0.62–19.09)	285 (179–455)	39.8	33.0

In a deterministic check, the central waterborne epoxy EPD proxy (2.75 kg CO₂e/kg product) was replaced with the conservative epoxy-resin model EPD value (4.12 kg CO₂e/kg product). Under the conservative epoxy factor, W-5, W-10, and W-20 increase to 504, 734, and 1163 kg CO₂e/m³, respectively, and their incremental carbon efficiencies become 4.15, 5.84, and 9.37 kg CO₂e/m³ per percentage point. Therefore, W-5 remains the best option under the incremental-carbon decision rule, while W-20 remains the lowest GWP_dur option when mass loss is treated as the deterioration-rate proxy, consistent with the analytical result above.

The repair-thickness check confirms that repair-layer GWP scales linearly with thickness but does not change the mixture ranking. For W-0, W-5, W-10, and W-20, the GWP per m² is 2.64, 4.22, 5.73, and 8.57 kg CO₂e/m² at 10 mm; 6.60, 10.55, 14.34, and 21.42 kg CO₂e/m² at 25 mm; and 13.19, 21.11, 28.67, and 42.83 kg CO₂e/m² at 50 mm. The incremental carbon efficiency per percentage point of mass-loss reduction also scales linearly with thickness: for W-5 it is 0.027, 0.068, and 0.137 kg CO₂e/m² per percentage point at 10, 25, and 50 mm, respectively.

5.5. Metakaolin–Slag Geopolymer Benchmark

A third, purely multi-criteria benchmark was taken from the metakaolin–slag geopolymer mortar study [29], because it reports a complete orthogonal data set. The L16(4⁵) design varied slag content (20–80%), sodium silicate modulus (1.0–1.6), alkali equivalent (6–12%), water-to-binder ratio (0.43–0.49), and sand-to-binder ratio (1.5–3.0), and reports fluidity, 28-day flexural and compressive strength, and 28-day drying shrinkage for all 16 mixtures (Table 4 of [29]; three specimens per mixture). The source reports a carbon intensity of 339 kg CO₂e/t for its recommended geopolymer mortar against 820 kg CO₂e/t for an ordinary cement mortar [29], but it does not report per-mixture inventories; the benchmark is therefore used here only to demonstrate a transparent and verifiable multi-criteria ranking, not as a carbon comparison. Each criterion k was transformed into an individual desirability d_k between 0 and 1 following Derringer and Suich [81]. The greater the compressive and flexural strength, the better:

$${\mathrm{d}}_{\mathrm{k}}={\displaystyle \frac{{\mathrm{y}}_{\mathrm{k}}-{\mathrm{L}}_{\mathrm{k}}}{{\mathrm{U}}_{\mathrm{k}}-{\mathrm{L}}_{\mathrm{k}}}}$$

(10)

and the smaller the drying shrinkage, the better:

$${\mathrm{d}}_{\mathrm{k}}={\displaystyle \frac{{\mathrm{U}}_{\mathrm{k}}-{\mathrm{y}}_{\mathrm{k}}}{{\mathrm{U}}_{\mathrm{k}}-{\mathrm{L}}_{\mathrm{k}}}}$$

(11)

and fluidity y is target-the-best, with target T = 190 mm, which lies in the middle of the observed range and corresponds to a workable, non-segregating repair mortar:

$${\mathrm{d}}_{\mathrm{f}\mathrm{l}}={\displaystyle \frac{\mathrm{y}-\mathrm{L}}{\mathrm{T}-\mathrm{L}}}(\mathrm{y}\le \mathrm{T});{\mathrm{d}}_{\mathrm{f}\mathrm{l}}={\displaystyle \frac{\mathrm{U}-\mathrm{y}}{\mathrm{U}-\mathrm{T}}}(\mathrm{y}>\mathrm{T})$$

(12)

where y_k is the measured value and L_k and U_k are the lower and upper normalization bounds, taken in the base case as the minimum and maximum observed among the 16 mixtures (132 and 270 mm for fluidity, 3.6 and 11.0 MPa for flexural strength, 15.8 and 86.9 MPa for compressive strength, and 0.060% and 0.143% for shrinkage). The overall desirability is the weighted geometric mean

$$\mathrm{D}=\prod {\mathrm{d}}_{\mathrm{k}}^{{\mathrm{w}}_{\mathrm{k}}},\sum {\mathrm{w}}_{\mathrm{k}}=1$$

(13)

with equal weights w_k = 0.25 in the base case. The geometric mean ensures that a mixture that fails one criterion completely (d_k = 0) cannot be compensated by the others. Table 11 lists the eight highest-ranked mixtures with all individual desirabilities, so that the ranking can be verified directly from the source data.

Table 11. Highest-ranked geopolymer mixtures [29] under the base-case desirability index (equal weights, geometric mean, min–max bounds, fluidity target 190 mm), with individual desirabilities d_k.

Rank	Mix	Slag (%)	Modulus	Alkali (%)	W/B	S/B	Flow (mm)	ff (MPa)	fc (MPa)	Shrinkage (%)	dfc	dff	dfl	dsh	D
1	X13	80	1.0	12	0.45	2.5	183	9.6	84.1	0.088	0.96	0.81	0.88	0.66	0.821
2	X7	40	1.4	12	0.43	2.0	170	10.3	74.2	0.075	0.82	0.91	0.66	0.82	0.795
3	X4	20	1.6	12	0.49	3.0	155	11.0	67.5	0.075	0.73	1.00	0.40	0.82	0.697
4	X12	60	1.6	8	0.43	2.5	176	10.7	72.3	0.111	0.80	0.96	0.76	0.39	0.687
5	X10	60	1.2	12	0.47	1.5	225	7.5	75.2	0.075	0.83	0.53	0.56	0.82	0.671
6	X3	20	1.4	10	0.47	2.5	165	8.9	56.6	0.071	0.57	0.72	0.57	0.87	0.671
7	X9	60	1.0	10	0.49	2.0	221	7.8	65.1	0.090	0.69	0.57	0.61	0.64	0.626
8	X8	40	1.6	10	0.45	1.5	220	8.2	62.8	0.095	0.66	0.62	0.62	0.58	0.621

In the base case, X13 (D = 0.821) and X7 (D = 0.795) are ranked first and second, followed by X4, X12, and X10 (Figure 5a). To test the subjective choices of the index, the ranking was recalculated with arithmetic instead of geometric aggregation, with fixed engineering normalization bounds (20–90 MPa, 3–12 MPa, 0.05–0.15%, and T ± 70 mm), with fluidity targets of 170 and 210 mm, and with the weight of each criterion doubled in turn (Table 12). In addition, 10,000 random weight vectors were drawn from a uniform Dirichlet distribution, so that all combinations of non-negative weights summing to one were equally likely (Figure 5b).

Figure 5. Geopolymer benchmark [29]: (a) base-case desirability of the 16 mixtures; (b) median rank and 5th–95th percentile range under 10,000 random weight sets, with the probability P₁ of being ranked first.

Table 12. Sensitivity of the geopolymer ranking to the formulation of the desirability index.

Scenario	Ranked First	Top Three
Base case (equal weights, geometric mean, min–max bounds, T = 190 mm)	X13	X13, X7, X4
Arithmetic instead of geometric mean	X13	X13, X7, X4
Fixed engineering bounds	X13	X13, X7, X4
Fluidity target 170 mm	X7	X7, X13, X4
Fluidity target 210 mm	X13	X13, X7, X10
Double weight on compressive strength	X13	X13, X7, X12
Double weight on flexural strength	X13	X13, X7, X4
Double weight on fluidity	X13	X13, X7, X12
Double weight on shrinkage	X7	X7, X13, X4
10,000 random weight sets (Dirichlet)	X13 (58.9%), X7 (29.2%)	X7 (99.3%), X13 (87.2%), X12 (41.6%)

X13 or X7 was ranked first in every scenario: X7 moves to first place only when shrinkage receives double weight or when the fluidity target is lowered to 170 mm. Under random weights, X13 was ranked first in 58.9% and X7 in 29.2% of the weight sets, and X7 remained in the top three in 99.3% (X13: 87.2%). Positions three to five (X4, X12, X10, and X3) change with the weights and are therefore conditional. The leading pair (X13 and X7) is thus stable with respect to the weights, normalization ranges, and aggregation rule in the scenarios tested, but which of the two is first depends on the weights (X13 first in 58.9% of random weight sets), and the order of the intermediate mixtures should be reported only together with the declared weights. Figure 6 shows the strength–shrinkage space of all 16 mixtures.

Figure 6. Compressive versus flexural strength of the 16 geopolymer mixtures [29], colored by 28-day drying shrinkage; the three base-case top-ranked mixtures are labeled in bold.

5.6. Cross-Study Synthesis

Across the curated database, polymer modification generally improves functions that are not captured fully by compressive strength alone. The extracted review and experimental studies report improved bonding, water retention, crack bridging, permeability resistance, sulfate or acid resistance, waterproofing, self-healing response, and toughness for selected polymer systems. At the same time, excessive polymer addition can reduce or delay cement hydration, entrain air, increase setting time, lower compressive strength, or raise material carbon intensity. The design problem is therefore not to maximize polymer dosage, but to identify the minimum polymer content that satisfies the governing repair function.

Four recurring dosage–response patterns were observed. First, dosage–response relationships are often nonlinear: small polymer additions can improve bond or durability substantially, but further addition may yield diminishing returns [31]. Second, the best mixture depends on the governing performance endpoint. A mixture with the highest compressive strength may not be the best for bond or permeability. Third, alternative binders can reduce carbon intensity, but they require careful control of workability, shrinkage, activator content, and early-age strength. Fourth, performance-normalized carbon indicators can reverse the ranking produced by kg CO₂e/m³ alone, especially when durability benefits are substantial.

6. Discussion

6.1. Implications for Sustainable Repair Design

The findings support a shift from material minimization to function minimization. Repair mortar should not be selected solely because it has the lowest kg CO₂e/m³. The relevant screening question is how much embodied carbon is required to deliver a specified repair function. A conventional mortar may have lower initial embodied carbon, whereas a polymer-modified mortar may be justified in aggressive exposure when the measured improvement is significant and repair recurrence would otherwise be likely. This second statement lies outside the A1–A3 boundary of the present assessment. The break-even checks of Section 5.3 and Section 5.4 and Appendix B do not quantify avoided repairs; they state the minimum durability benefit, or the maximum polymer emission factor, required for a modified mixture to offset its higher material-stage emissions. They become life-cycle results only when combined with a validated service-life model and modules B and C of EN 15804 [69].

Durability is a fundamental aspect of sustainability because reduced maintenance and restoration can reduce material demand, energy demand, time demand, and demolition waste [37]. Selected polymer-modified mortars have been reported to improve freeze–thaw resistance, transport blocking, chloride resistance, and estimated service life [31]. These durability benefits could offset higher initial embodied carbon only when exposure severity, application quality, and validated deterioration models support a longer repair interval.

In the epoxy case study, the choice among W-5, W-10, and W-20 depends on the decision rule. For a low-exposure environment or short design life, the incremental benefit of W-20 may not justify its additional polymer carbon. For severe acid exposure where repair access is difficult, the deterioration-rate-normalized result suggests that a high resin dosage could become environmentally rational if a validated service-life model confirms that it prevents repeated repair; this is a conditional interpretation and not a result of the cradle-to-gate assessment. This is precisely why the functional unit and the decision rule must be declared explicitly.

In the SBR mortar case, a low latex dosage gave a large reduction in chloride migration for a small carbon penalty, and the emulsifier type affected the chloride indicators more than strength. Whether the modification is carbon-favorable, however, hinges on the latex emission factor. For practitioners, the break-even factor is more useful than a single ranking, because it can be compared directly with the GWP declared in a supplier EPD. The broad PNCE concept remains useful for future service-life-based assessment, but universal polymer rankings should be avoided until product-specific environmental product declarations, field-validated durability models, and uncertainty analyses are available.

6.2. Practical Screening Workflow

A practical performance-normalized carbon workflow for polymer-modified repair mortar selection can be summarized as follows. First, mixture proportions, polymer dosage, curing conditions, exposure regime, and performance endpoints should be extracted from each study or supplier dataset. Second, mixtures should be converted to a common functional unit, preferably 1 m³ when absolute quantities are reported, or otherwise 1 kg dry mix with a clearly stated density assumption. Third, emission factors should be assigned by constituent and classified as extracted, EPD-based, database-based, or assumed. Fourth, cradle-to-gate GWP and repair-layer GWP should be calculated for the selected thickness. Fifth, performance-normalized indicators should be calculated for the governing repair function, always relative to a control tested under the same protocol. Sixth, emission factor and test-result uncertainties should be propagated, and rank probabilities and break-even emission factors reported. Seventh, non-dominated Pareto solutions should be identified across carbon, strength, bond, shrinkage, workability, and durability metrics. Finally, the preferred mixture should be selected according to exposure class, repair accessibility, target service life, and carbon budget.

6.3. Data Limitations and Uncertainty

The extracted database includes heterogeneous specimen sizes, curing regimes, exposure durations, and testing standards. Many studies report ratios rather than complete kg/m³ inventories; some cannot be converted to absolute volume without density assumptions. Polymer products rarely have product-specific environmental declarations in the source experimental papers. In the calculation, publicly available environmental product declarations were used where a sufficiently close product class was available. No public product-specific environmental product declaration was identified for the water-soluble epoxy product used in the primary case study; therefore, a verified environmental product declaration for a comparable water-based two-part epoxy coating was used as the central proxy [73], and a European model environmental product declaration for epoxy-resin products was retained as a conservative sensitivity case [66].

The present numerical analysis is cradle-to-gate (A1–A3) and does not include transport, surface preparation, installation, curing, maintenance, or end-of-life. Mass loss, migration coefficients, and freeze–thaw strength loss are treated as durability proxies; they should not be converted directly into service life without validation, and Equation (9) is a conditional link for diffusion-controlled chloride ingress only. The mortar inventories of both case studies were reconstructed from reported ratios (the epoxy mortar from the EN 196-1 standard proportions); the epoxy mass-loss data rest on two specimens per mixture, and the solids content of the latex in the regional EPD is not declared. The curated studies are recent open-access articles from one publisher, so they illustrate the framework rather than represent the field. The performance-normalized carbon efficiency concept should also be used carefully because multiplicative performance ratios can exaggerate differences when the performance variables are correlated or partly estimated.

6.4. Uncertainty Controls and Validation Considerations

The key uncertainty controls are addressed explicitly. Screening factors were replaced where public environmental product declarations or regional life-cycle inventory data were available: cement and aggregate values were updated using regional/product declarations [72,79], and a waterborne epoxy environmental product declaration proxy replaced the earlier resin-solids screening factor [73]. Remaining non-product-specific inputs are labeled, assigned probability distributions, and propagated by Monte Carlo simulation together with the measured test variability (Section 4.4). Repair-layer scaling is included, and durability-to-service-life conversion is avoided unless validated by field data. The EPD references used in the carbon calculation report the program operator, declaration number when available, declared unit, A1–A3 GWP module, and validity period where stated by the declaration.

The Supplementary Materials include the input data, the emission factor table with the assigned distributions, and the Python scripts and results for the case-study calculations, thickness and emission factor sensitivity, break-even checks, Monte Carlo analysis, geopolymer ranking, and figures, together with a script that checks the regenerated values against those reported here. This traceability is necessary because many mixture records require normalization assumptions and several environmental factors remain product-class proxies rather than exact manufacturer-specific declarations.

7. Conclusions

A structured database was developed from the extracted literature, containing 55 indexed studies, 141 mixture/inventory records, and 243 performance endpoints relevant to polymer-modified and advanced cementitious repair mortars; 19 studies met the criteria for quantitative extraction.

A standardized functional-unit framework is essential because published repair mortar studies report mixtures using inconsistent bases such as kg/m³, cement-to-sand ratio, water-to-binder ratio, polymer-to-cement ratio, and percentage replacement. Performance-normalized indicators should be computed within each study against its own control, because test protocols differ between studies.

Carbon-only ranking can be misleading for repair materials. Performance-normalized metrics are needed to account for strength, bonding, shrinkage, self-healing, permeability, acid resistance, and other repair functions; however, laboratory durability improvements should be treated as proxies unless translated through validated service-life models.

In the epoxy-modified mortars exposed to sulfuric acid, W-5 provided the best incremental carbon efficiency per percentage point of mass-loss reduction, W-10 a balanced compromise, and W-20 the best deterioration-rate-normalized indicator. Both rankings were retained in more than 99% of Monte Carlo draws, were identical for the mortar and concrete inventories, and the ICE ranking was shown analytically to be practically independent of the emission factors.

In the SBR-modified repair mortars, the cationic latex gave the largest reduction in chloride migration coefficient (31.5%) and the largest freeze–thaw improvement. With a regional latex EPD, its chloride-normalized GWP was 26% lower than that of the control, but the benefit is conditional on the latex emission factor, with break-even values of 2.66 kg CO₂e/kg for chloride and 4.66 kg CO₂e/kg for freeze–thaw exposure.

In the metakaolin–slag geopolymer benchmark, a desirability index with declared normalization, weights, and aggregation ranked mixtures X13 and X7 first. One of the two was ranked first in all predefined scenarios and in 88.0% of 10,000 random weight sets (X13: 58.85%, X7: 29.18%); the leading pair is therefore stable, whereas which of the two is first, and the intermediate positions, depend on the declared weights.

Overall, the framework supports screening decisions rather than a definitive ranking of polymer-modified repair mortars. Final project-level selection should replace proxy factors with supplier-specific environmental product declarations, apply exposure-specific durability models, and consider transport, installation, maintenance, and end-of-life stages when those data are available.