Comparative environmental assessment of limestone calcined clay cements and typical blended cements

Decarbonization of the cement and concrete industries is one of the top priorities on the path to a carbon-neutral economy. This article presents a novel model for evaluating the emissions from the production of metakaolin (MK) as a supplementary cementitious material used in ternary blended cements (e.g., 35% metakaolin, 15% limestone, and 50% portland cement) and an accompanying decisions-support tool (MKC-Tool). Applications with a case study in California showed 36%–39% reductions in greenhouse gas (GHG) emissions from ternary blends with MK compared to portland cement. Compared to commercially available blended cements, the ternary blend showed the lowest global warming potential. All the cements containing fly ash showed higher GHG intensities than the ternary blend (16%–42% higher GHG emissions). The development of cements made with portland cement, metakaolin, and limestone at an industrial scale will have the potential to contribute 5%–50% to the global reduction of GHG emissions from the cement industry.


Introduction
The rapid industrial growth and urbanization of the past 70 years have significantly increased the consumption of concrete. Concrete production impacts the environment in many ways, significantly by releasing greenhouse gas (GHG) emissions into the atmosphere. The increasing demand for this construction material is related to concrete's numerous attributes such as relatively affordable cost, ease of manufacturing, well-established building codes, and excellent track record. Such high demand translates into the annual consumption of 4 Gt of cement, 17.5 Gt of aggregate, and over 2 Gt of water worldwide [1,2]. Raw material acquisition and concrete manufacturing are responsible for about 8% of the annual global carbon footprint [2] and 18% of the global annual industrial water consumption (in which most of its use is associated with the production of aggregates and the electricity needed throughout concrete's value chain) [1].
Cement is a powder that reacts with water, producing hydration products capable of binding the isolated aggregate particles, forming a complex structural material. Current statistics estimate that the annual production of cement surpassed 4 Gt in 2017, and it will continue to grow in the next decades [3]. Cement production increased more than ten times since 1950, in contrast to smaller, three-fold growth in steel production and constant production of wood, as reported by Monteiro et al [4]. During cement production, approximately 0.9 ton of CO 2 is released into the atmosphere for each ton of cement clinker produced anywhere in the world. (1.5 Gt CO 2 -eq were reportedly released in 2018 from the global cement industry [5]). These emissions are mainly from the decarbonation of limestone (one of the cement's raw materials, composed of calcium carbonate, CaCO 3 ), which transforms into calcium oxide (CaO) and carbon dioxide (CO 2 ) under thermal treatment at around 900°C. CaO is further heated up to 1450°C along with aluminosilicates from clays to produce clinker, the principal component of cement, whereas the CO 2 gases are released into the atmosphere and are considered to be the largest non-combustion source of CO 2 emissions from the industrial sectors [6,7].
Multiple strategies have been put in place and continue to be studied for the decarbonization of concrete's life cycle: (a) reducing CO 2 emissions through electrifying the calcination process [8] or through the electrochemical synthesis of lime during the calcination process [9]; (b) incorporating material properties and environmental considerations into structural design [10] and [11]; (c) performing mix design optimization using artificial intelligence techniques [12] and [13]; among others (e.g., [6] and [14]). However, those which aim to reduce the fraction of clinker in cement are the most reliable when considering their ready application in emerging economies where the demand for construction materials-especially cement and concrete-is expected to increase in the next decades.
Clinker substitution can be done by partially replacing clinker with supplementary cementitious materials (SCMs). Also known as mineral admixtures, SCMs are finely ground siliceous or alumino-silicate materials that could have cementitious or pozzolanic characteristics or both. Pozzolans (a subset of SCMs) consist of reactive aluminosilicates that react with the calcium hydroxide from cement hydration to produce additional cementitious phases (e.g., calcium (alumino)silicate hydrates). Reducing CO 2 emissions comes from reducing clinker content in the final cement product. SCMs can offset CO 2 releases into the atmosphere from clinker production in two ways, depending on whether clinker substitution is performed in cement manufacturing after clinker formation or at the concrete production stage. Both practices are limited to a maximum standard percentage of replacement to avoid drops in the early strength, workability of the mix, and durability of the concrete. In addition, an energy reduction is also achieved since these SCMs require significantly lower or no calcination processing to achieve the material's reactivity. The most common SCMs are ground granulated blast furnace slag (GGBFS), fly ash, silica fume, natural pozzolans, rice husk ash, and calcined clays. Adding pure limestone is also commonly used in ternary and quaternary portland cement (PC) blends with other SCMs. A study [15] used US Geological Survey data to suggest that natural pozzolans such as calcined clays (with high content of Al and Si) are plentiful alternative candidates for cement production. Furthermore, due to its widespread local availability, calcined clays overcome the issue of ready availability for use in the future. Among all clays, those rich in kaolinite mineral have proven to be more reactive when mixed with cement and limestone [16]. Kaolinitic clays calcined to gain reactivity are also known as metakaolin (MK) clays.
The substitution of clinker with mineral additions that are either cementitious or pozzolanic is a common practice in Europe and, to a lesser extent, in North America. Blended cements in the United States are specified according to the standards ASTM C150 and ASTM C595. However, the most common mineral materials specified in those standards correspond to ground granulated blast furnace slag (GGBFS) and fly ash (e.g., slag modified portland cement I (SM), portland pozzolan cement IP/P). The availability of the aforementioned materials is limited; thus, they alone do not contribute to a long-term solution in reducing clinker content in cement-based materials to improve cement's environmental impact and meet carbon reduction targets. Portland cement (Types I to V) sold in the United States dominated the market with more than 90% of the total 2017 sales in the country. At the same time, blended cements constituted only 2.1% of the total sales during the same year. However, SCMs in the United States are predominately used with cement to mix concrete at the concrete ready-mix plants.
Because the reduction of the clinker-to-cement ratio is one of the most effective strategies to mitigate cement's environmental burden worldwide [17], this article aims to compare the GHG emissions and energy demand of the most common blended cements in the United States with those from ternary blends of cement made with clinker, metakaolin, and limestone. Table S1.1 in the Supporting Information (SI) shows the most recent data on the market share of different types of cements sold in the United States in 2017 [3].
The case study presented herein focuses on California, the fifth largest economy in the world [18, 19] and one of the top two states in the United States for portland cement consumption in the past 30 years, with 9,400,000 metric tons in 2015 (11% of the national cement consumption) [20]. The environmental value in substituting clinker with MK and limestone to reduce clinker content in cement and its associated emissions could be translated to other regions based on the Californian case results.

Environmental performance of calcined clay as a supplementary cementitious material
At industrial scale, in terms of environmental, social, and economic impact, questions such as: 'Will different processing of metakaolin vary its environmental impact and therefore the overall environmental load of the blended cement?' have been raised among researchers. To answer this and other questions, a careful analysis of the GHG footprint of the calcination of clays should be performed. The importance of calculating the environmental impact of concrete binder made of cement, limestone, and metakaolin would allow us to understand their performance in reference to conventional concrete made of portland cement and commercially available Pozzolan-Portland-Cement (PPC) concretes. Life-cycle assessment (LCA) of cement with metakaolin must be completed to evaluate its environmental performance.
First, it is essential to identify the environmental value of implementing metakaolin as a partial replacement of cement in concrete production. Current emission factors of MK are estimates based on approximations from cement's manufacturing processes. Mikhailenko [21] reported 175 kg CO 2 /tonne-MK emissions from raw material extraction and fuel-derived emissions. This value was previously stated by Cassagnabere et al in [22,23] based on values reported by Gartner [15]. Unfortunately, details of calculations of this value are not mentioned by Gartner. Moreover, Cassagnabère reports this emission factor later on in [24] and states that this value is supported by confidential data based on an environmental impact assessment. Heath et al [25] reported a value of 423 g CO 2 -eq/kg of metakaolin based on the assumption that 1.16 kg of kaolin are required to obtain 1 kg of metakaolin after calcination. In addition, calculations of this figure were based on energy consumption of 2.5 MJ/kg of metakaolin using natural gas as fuel for clay calcination [26]. Similar values have been published by other authors, such as Jones et al [27], whose cradle-to-gate analysis reported a value of 330 g CO 2 -eq/kg of metakaolin based on 400 kWh/t of embodied energy based on private communication with materials suppliers. Habert and Ouellet-Plamondon [28] emphasizes the high variability of MK's embodied energy due to the type of fuel used during its calcination process. Reduced values of global warming potential (GWP) of MK could be achieved when alternative heat sources (e.g., biomass) are used. These authors reported a value of 92.4 g CO 2 -eq/kg of MK when using biogas (agricultural waste) as fuel for the calcination process, whose emissions are approximately five times lower than the values reported above. Gettu et al [29] proposed a value of 2.6 MJ/kg of clay required for its calcination based on estimates made by experts from the cement industry and calculations of specific heat and calcination energy from thermogravimetric analysis of several samples of clay. For their study, an LCA was performed to compare the environmental profile of over 30 mix proportions of various concretes, considering typical values for India.
These researchers have pointed out that using metakaolin as an SCM has a more significant impact on cement's environmental profile than using known (by)products such as fly ash and slag. To activate these clays, energy from fuel combustion during calcination is required to produce a reactive material (metakaolin). However, even though metakaolin's production requires energy input for its calcination, its widely and ready availability makes this material a great potential substitute for other commercially available SCMs (e.g., fly ash, slag). Additionally, most LCAs of blended cements using fly ash or slag do not include impacts associated with the processing of these materials in their analyses due to the consideration that these are waste products of the energy and steel industry. However, a recent study by Habert and Ouellet-Plamondon [28] points out that slag and fly ash are considered byproducts of these industrial plants, so an environmental burden should be allocated to them. When these impacts are considered, the total environmental influence of concretes made with these SCMs could increase and even supersede those made with cements blended with metakaolin. Furthermore, LCA impact categories other than GWP during MK production have only been reported by Heath et al [25] based on the CML 2 (2002) baseline method [30] with natural gas as the main heating source. The use of natural gas as fuel has also been studied by Habert and Ouellet-Plamondon [28]. A summary of emission factors and embodied energy per unit of mass of metakaolin found in the recent peer-reviewed literature is shown in table 1.
It is important to note that there is still a lack of rigorous data in the literature on the environmental impacts of metakaolin and its use as SCM in cement and concrete production. Thus, specific data on life-cycle inventory emissions for calcined clay needs to be fully assessed. This article describes a life-cycle inventory framework to calculate the life-cycle environmental impacts of metakaolin production which in turn will allow for a complete assessment of the environmental burden of concretes made with blends of PC, MK, and limestone. A decision-  [32] support tool called the MKC-LCA Tool was designed to evaluate such impacts using a comprehensive LCA model. The tool was created with the aim of estimating the environmental impacts of metakaolin production from cradle to gate, based on real and locally specific parameters as much as possible.

Scope and system boundary
Metakaolin is the final product of the calcination of clay rich in kaolin mineral. The production of metakaolin is similar to that of cement manufacturing. The raw clay with kaolinite content is mined and brought to the plant, where processing and calcination occur. Figure 1 compares each stage of the processing of cement and metakaolin.
The GHG emissions and energy demand of the most common blended cements in the United States (i.e., slag-modified portland cement I (SM), portland blast furnace slag cement IS, slag cement S, pozzolan modified portland cement I (PM), and portland pozzolan cement IP/P) are calculated and compared with those from ternary blends of cement made with clinker, metakaolin, and limestone. Therefore, the functional unit in this study is one tonne of metakaolin manufacturing and its use in various cement blends. The assessment ends at an intermediate stage of metakaolin's life cycle (i.e., cradle to gate), as opposed to a cradle-to-grave assessment, due to the lack of published literature on the long-term durability performance, service-life modeling, and end of life of concretes made with cement blends containing metakaolin under several conditions. Because such information is needed to accurately account for expected maintenance activities, materials demolition and disposal, and potentially sequestered CO 2 during concrete's service life, a cradle-to-grave assessment might be dependent on assumptions that may not represent actual conditions. Therefore, a cradle-to-gate assessment is a widely accepted framework for the assessment of concrete and other building materials. This type of analysis is very useful to develop complete life-cycle assessments of concrete in buildings and other applications. Figure  S4.1 in the SI shows the system boundary of MKC-LCA Tool. For comparison, the GWP of portland cement (95% clinker and 5% gypsum) is also included in the model. A breakdown of each cement's components by mass per unit weight of cement is shown in table S2.1 in the SI based on average values specified in ASTM C150 and ASTM C595 [33,34]. Data were collected to reflect production conditions in California for the latest available year.

Modeling and parameters
To understand the differences in each cement's environmental impact, the modeling parameters corresponding to the technology used for clinker production, calcination of SCMs-when applicable-and fuel mix used during those processes is kept constant in all the modeled scenarios (i.e., preheater-precalciner kiln and California's average fuel mix) as shown in table S3.1. in the SI. The average kiln fuel mix used in the model corresponds to the average fuel mix used in cement kilns in California as shown in table S3.2 in the SI. California's cement plants and their location are detailed in table S3.3 in the SI. In view of this, extraction of raw materials and production of all the cements were modeled in California. For cements containing metakaolin and limestone, two calcination scenarios are modeled with the Californian average fuel mix (i.e., dry rotary kiln and flash calciner). Portland cement was modeled as ASTM Type II as this is the most common type of cement commercially available in California. Life-cycle inventory data were used from the Green Concrete LCA Web Tool developed by Gursel and Horvath [35] to model the energy demand and GWP of portland cement on mass basis (i.e., MJ/tonne and kg CO 2 eq/tonne, respectively). The manufacturing technology for the calculation of the energy demand and GWP of portland cement is summarized in table S3.4 in the SI. In addition, metakaolin and limestone are modeled to be processed at the cement plant and blended with cement at the last stage of production (i.e., separate grinding and interblending with portland cement). Table S3.5 in the SI shows the transportation modes and distances associated with all cement-based materials' production. Note that at industrial scale, metakaolin can either be processed at a different processing plant or at the cement plant as an independent line running in parallel with clinker production.
Aside from the raw materials to produce metakaolin, fuel and electricity are the main inputs during metakaolin manufacturing. The MKC-LCA Tool does not only consider direct emissions from metakaolin processing from direct fuel combustion and electricity use, but it also considers the supply-chain effects of fuel procurement and processing (pre-combustion) and electricity generation (construction and operation of power plants) in its assessment.
For this purpose, data were obtained from different studies developed both in the United States and internationally to generate a complete life-cycle inventory (LCI) of the most common fuels used in metakaolin extraction and production processes as well as an LCI of electricity generation. The main databases, international statistics, studies, and literature used as main sources for material quantities, inventories, and methodology are summarized in table S3.6 in the SI.
Due to the word limit of this article, detailed descriptions of the environmental assessment model and parameters are found in section S3 of the SI.

Decision-support tool
To assess the environmental impact of metakaolin production as an SCM, a comprehensive LCA model has been created and captured in a decision-support tool, the MKC-LCA Tool, designed to estimate the environmental impacts of metakaolin production from cradle to gate based on real, local parameters, as specific to California as possible. With this tool, cement and concrete manufacturers, decision-makers in the construction sector, and researchers can obtain validation of the benefits of using metakaolin in their concrete mix designs. A detailed description of the MKC-LCA tool is found in section S4 of the SI.

Results and discussion
3.1. Energy demand and global warming potential Figure 2 shows the GWP per tonne of each blended cement modeled (energy demand is shown in figure S5.1 of the SI). Compared to PC, ternary cements made with MK and LS in a 35:15 replacement ratio showed a significant reduction of GHG emissions and life-cycle energy demand (i.e., 36% lower emissions and 19% lower energy demand when MK was calcined using a dry rotary kiln and 39% lower GHG emissions and 24% lower energy demand when using a flash calciner). This is particularly interesting given that a significant reduction in GHG emissions and energy demand could still be achieved by repurposing decommissioned rotary kilns without investing in the best available technology (i.e., flash calciners). When compared to commercially available blended cements, these ternary blends of cement with MK and LS showed the lowest GWP, and energy demand was only surpassed by slag cement (S) since the latter comprises an average replacement ratio of 85% of clinker by slag.
The GWP of portland slag cement with 47.5% slag and 52.5% PC was only 9% lower than that of 35MK:15LS (i.e., 503 kg CO 2-eq /tonne versus 549 kg CO 2-eq /tonne). These two cement blends consisted of a similar amount of total replacement of PC with an SCM. However, at an industrial scale, the lack of readily available slag worldwide renders this SCM incapable of meeting the high demand for cement. Furthermore, when the metakaolin in 35MK:15LS cement is calcined using a flash calciner, its GHG emissions are only 4% higher than those of portland slag cements. Finally, all of the blended cements containing fly ash (i.e., PC I (PM) and IP/P) showed higher GHG intensities compared to 35MK:15LS cement due to their higher cement content and lower replacement ratios with SCMs (i.e., 27.5% and 72.5%, respectively). Similar energy demands are obtained when comparing 35MK:15LS with portland pozzolan cement IP/P (only an 8% difference). However, the total replacement of clinker by SCMs is almost twice the amount by weight of binder for the 35MK:15LS than for the IP/P cement, which indicates a better use of raw materials and lower GHG emissions. Ultimately, the use of commercially available blended cements in many regions faces a lack of availability of the SCMs commonly used in their formulations (i.e., slag, fly ash), and those alone may not meet the reduction targets outlined in the IPCC and UNEP reports for the global cement industry by 2050 [17].
According to the case study results, cement blends that contain metakaolin and limestone as a partial replacement for cement could potentially reduce the GHG emissions from cement manufacturing by 5% to 40%, depending on the overall replacement ratio and the availability of other SCMs that do not need to be processed (i.e., calcined at high temperatures) to exploit their pozzolanicity and react with clinker during hydration of cement. However, SCMs such as fly ash and slag are limited to coal-fired electricity generation and steel manufacturing, making their distribution to all regions limited, as is the case in California where neither of these technologies are in place. At a global scale, when different formulations of blends with portland cement, metakaolin, and limestone are used, up to a 50% reduction in global GHG emissions could potentially be achieved compared to other cements. Note that these conclusions could not be fully drawn when considering these cements in concrete applications since other parameters should also be taken into consideration, such as mix designs, strength targets, curing time, unreinforced or reinforced applications, environmental exposure, and durability of the concrete members. A worldwide analysis could be done in the future, but it has been excluded from this assessment due to the ample variability in the data and the uncertainty that exists during each production process at an industrial scale.

Influence of fuel mix input on the GWP and energy demand of metakaolin production
The input variability available to the user could influence the model's total GHG emissions and energy demand calculations. However, as shown previously, the GWP and energy demand are influenced mainly by the fuel-mix input used for the calcination of the raw clay and in the pyroprocessing of clinker during cement manufacturing. Figure 3 shows the life-cycle GHG emissions of metakaolin's production in California while varying the type of fuel used for calcination. Even though the calcination process is the most relevant and energy intensive in the metakaolin production chain, the results compared to PC do not vary greatly, and the overall trend is unchanged. The total GWP of metakaolin production in California ranges from 295 kg CO 2-eq /tonne to 350 kg CO 2-eq /tonne. Significant reductions are observed when the calcination process is performed using the best available technology (i.e., flash calciner) and waste biomass for calcination (46 kg CO 2-eq /tonne). For completeness, the average fuel mix used in the United States is also modeled and included in figure 3 and figure  S6.1 of the SI. California's average fuel mix for kiln firing comprises mainly coal (58%), petroleum coke (i.e., petcoke-24%), and waste tires (6%) that could be acquired whole or shredded. Additional energy should be considered for processing whole waste tires if they arrive at the manufacturing plant in such condition. However, all case studies modeled herein assumed that the tires are shredded. On the other hand, the United States average fuel mix is composed mainly of bituminous coal (64%) and petcoke (21%), which explains the slight difference in the GHG emissions produced during metakaolin calcination when comparing both fuel mix inputs. Even though a transition to increased use of waste tires for fuel during calcination could lead to small savings in energy (i.e., 0.3%), the reduction in GHG emissions is still noticeable (i.e., 11%) from this production stage, as shown in both figures.
Furthermore, the use of conventional fossil fuels for kiln firing represents 25%-30% of the cement production costs [36][37][38][39]. The potential replacement of conventional fossil fuels with alternative waste fuels (e.g., waste tires) for kiln firing has been estimated to reach an average of 12% worldwide, with potentially higher replacement ratios in developing countries, therefore, adding an economic incentive due to the reduction in fuel purchase cost [36]. More energy savings are expected when using waste biomass, such as waste wood, for fuel calcination, as shown in figure S6.2.
Energy demand and GHG emissions from transportation are the same for metakaolin manufacturing since the variations in the production technology and transportation distances of raw materials for PC production within the system boundary are considered similar to those for MK production. Thus, the sensitivity analysis did not consider variation in transportation distances as those do not significantly influence the overall results. However, when biomass is used in great quantities as the principal fuel for calcination, impacts from transportation could start playing a more significant role in the overall energy demand and climate change impacts.

Influence of fuel mix input on the GWP and energy demand of blended cements
Changes in the fuel mix for calcination and pyroprocessing could lead to savings in energy demand and lower GHG emissions from binder production. Figure 4 shows the influence of changing the fuel mix for calcination and clinker production on the GWP of these blended cements (energy demand is shown in figure S6.3 of the SI). Utilizing waste tires as the main fuel for metakaolin calcination and clinker pyroprocessing causes a 5% to 10% reduction in the GWP and less than a 1% reduction in the energy demand. For binders with little process-based CO 2 emissions (e.g., slag cement), switching to a waste fuel only reduces GHG emissions by 5%. Such binders contain higher replacement ratios of clinker with SCMs that do not need to be subjected to thermal treatment, nor do they release significant amounts of CO 2 during the process.
Conversely, binders with higher process-based CO 2 emissions and whose SCM content requires thermal activation benefit from using waste fuels to reduce the overall fossil energy demand and GHG intensity. It is worth noting that even when including these variations of fuel mix used during calcination of metakaolin and the pyroprocessing of clinker, the slag cements and the cements with metakaolin and limestone remain the blended cements with the lowest GWP and energy demand. For blended cements with metakaolin and limestone, further reductions in GHG emissions and energy demand could be achieved with changes in the efficiency of the calcination equipment (i.e., dry rotary kiln) for the best available technology (i.e., flash calciner). Considering the use of waste tires as fuel for calcination and pyroprocessing, reduction of 4% in the GHG emissions and 6% in the energy demand of these blended cements (i.e., MK:LS cements) are achieved when using a flash calciner for metakaolin's thermal treatment. It is important to highlight that factories that produce clinker through wet processing tend to use more waste fuel to control their fuel cost and increase their competitiveness against other cement plants with better and more efficient technologies. This could lead to different results in the overall assessment of GHG emissions and energy demand for regions in the world that still maintain active wet kilns in their cement plants. Because all cement plants in California utilize dry rotary kilns with preheaters and precalciner, the analysis is not included herein.
Transportation distances have less influence on these binders because their GHG emissions and energy demand are highly related to the combustion of fuels and the calcination and clinkerization processes. Not included in this work is the consideration of other emissions that could be relevant when assessing the impact on human health and the evaluation and control of criteria and toxic air pollutants (e.g., NO x , CO, acetaldehyde, benzene, formaldehyde, etc.) that are generated during the burning process of tires.

Influence of the total replacement ratio of metakaolin and limestone on the GWP and energy demand of blended cements in California
The formulation of ternary cement blends composed of PC, metakaolin, and limestone is another variable that should be explored to assess how different formulations compete with portland cement and other commercially available blended cements in California. For this scenario, all formulations of PC and blended cements were modeled with a fuel mix composed of 100% waste tires, and their GWPs are detailed in figure 5 (energy demand is shown in figure S6.4 of the SI).
Results indicate that increasing the limestone content to reduce the clinker-to-cement ratio further reduces the GWP and energy demand of cement blends made of clinker, metakaolin, and limestone. For the cements containing metakaolin and limestone, an overall increase in SCMs by 25% (i.e., 40% replacement versus 50% replacement or 35MK:5LS versus 35MK:15LS) resulted in a 13% reduction in GHG emissions and 10% reduction in energy demand. A maximum increase of SCMs by 37.5% (i.e., 35MK:5LS versus 35MK:20LS) resulted in a 15% reduction in energy demand and a 20% reduction in GHG emissions. In addition, smaller yet meaningful reductions are achieved when increasing the percentage by mass of limestone from 5% to 10% (i.e., 35MK:5LS versus 35MK:10LS): energy demand and GHG emissions are reduced by 5% and 7%, respectively. Compared to other cement products (e.g., blended cements containing slag or fly ash), cements with metakaolin and limestone offer the lowest GWP and energy demand on a mass basis. The latter are only outperformed by portland slag cements (IS) and slag cement (S), whose clinker substitution ratios correspond to 47.5% and 85%, respectively.
As previously mentioned, the assessed energy demand and GHG emissions correspond to those by unit mass of each blended cement and PC investigated (i.e., kg CO 2-eq /tonne and MJ/tonne) and did not consider the structural capacity (i.e., technical performance) of these cements in plain and reinforced concrete applications. Determination of GWP and energy demand is based on specific model parameters, location of raw materials and processing, and energy used, hence results could vary depending on the variation of such inputs.

Review of inventory data quality and data gaps
Data quality assessment is essential in any environmental assessment to ensure that the study uses appropriate types of data and that the data match the needs of the study [40]. The quality of the data used in the development of the MKC-Tool is approached qualitatively following the pedigree matrix developed by Weidema et al [41]. The expected effects of various types of uncertainty and data variability are described in the pedigree matrix shown in table S7.1 of the SI.
Higher uncertainty is intrinsically connected to existing data gaps in LCAs of emerging technologies due to the lack of knowledge at the initial phases of product development (i.e., the R&D phase). This is the case of metakaolin production for use as an SCM in blended cements made with clinker at industrial scale. In many cases, the primary data for key processes were not available. Hence, the results are considered best obtainable estimates. On that account, when secondary data are used, accurate referencing has been included for all the data collected to build the life-cycle inventory for the MKC-Tool. Consequently, data are unrounded but considered accurate to no more than one significant digit.
Emissions to air mainly come from the releases through the combustion of fuels during the pyroprocessing of metakaolin. Because the calcination of metakaolin consumes approximately 79% of the energy demand throughout the entire manufacturing process, data on the life-cycle inventory heavily rely on the quality of the data on fuel consumption. Particulate matter (PM) emissions are attained due to PM control devices used in manufacturing. Most uncontrolled PM releases occur during raw material quarrying and transportation to the manufacturing facility. Emissions from diesel-fueled vehicles used for transportation of materials within the system boundary are calculated from peer-reviewed studies that are compatible with the geospatial and temporal scope of this study [42][43][44].
The effects on the methodology chosen to calculate the global warming potential of metakaolin production ( i.e., TRACI 2.1 [45]) are valid for application in North America. Changing the impact assessment methodology (e.g., GWP 20 years) would not drastically change the results, given that these changes would be marked primarily on how methane is evaluated. Methane is not an important gas released during cement production as it is in other industries and economic sectors such as agriculture. Direct CO 2 emissions from cement production account for 98%-99% of the total GHG emissions in cement plants in California [46] and direct methane emissions for 2% of total GWP, according to the case scenarios evaluated herein.

Conclusions
The use of SCMs as a partial replacement for clinker in cement should be the focus of attention in the short term to help reduce GHG emissions and energy demand. SCMs such as calcined clay and limestone have been identified to be available in the quantities we globally need to provide a sustainable decrease of the clinker-tocement ratio without compromising cement's and concrete's performance. However, questions such as What is our reference point for baseline performance of cement-based materials? What is the range of embodied carbon we could expect in this category? How does this particular supplementary cementitious material stack up against this reference point and the rest of the potential options? remain the focus of attention in the present research.
The MKC-Tool has been designed to perform LCAs involving metakaolin used as SCM in concrete or as a pozzolanic addition to blended cements. The main goals were to (1) identify the hotspots or activities with the highest emissions throughout the manufacturing process and (2) support comparing multiple cement-based materials and concretes made with 100% portland cement and/or other cementitious or pozzolanic materials. The MKC-Tool was created to incorporate LCA into decision making as it helps the user to decide quantitatively about the types of cements that could be used in different projects and their overall environmental impacts.
The MKC-LCA Tool was used to model the GWP and metakaolin's energy demand for different calcination technology scenarios and fuel mix input. Emphasis on GWP and energy demand is given since the emissions derived from the manufacturing of metakaolin are considered to contribute in a major capacity to the potential cause of climate change. In addition, cement blends with a fixed ratio of replacement with metakaolin and limestone are also presented and evaluated against portland cement and commercially available blended cements in California. With eight cement plants distributed along the state and an annual production of 9.4 million tons of clinker in 2017 [3], California is the second largest cement-producing state in the United States after Texas. The consequential life-cycle assessment described in this study answers the question of how likely it is that cement-based materials comprising metakaolin, limestone, and portland cement have a lower environmental impact than conventional blended cements and/or portland cement.
LCAs of emerging technologies are subject to uncertainties as the product and its manufacturing process mature and become more widespread. Accordingly, future development of the MKC-LCA Tool should consider: • The performance of each alternative under study. In this context, the development of mechanical properties of mortars and concretes made with ternary blends of portland cement, metakaolin, and limestone may be different from those made with traditional portland cement. A broader overview of these cement-based materials' environmental impacts would be possible when accounting for parameters such as age of the composite, compressive strength, quality, and grade of the calcined clay, etc. Moreover, while proper mechanical strength is necessary, other properties may be more appropriate for specific exposure of the structure. Therefore, durability parameters should be introduced in the assessment. These are not currently included in the tool due to lack of a standard protocol.
• It would be valuable to make the tool relevant to widespread local decision making by incorporating local energy supplies and electricity mixes and variations in manufacturing technologies and supply chains [47][48][49][50].
• In the present work, only portland cement clinker was used in the analysis of ternary cement blends with metakaolin and limestone. However, a new generation of cements with reduced carbon footprint (e.g., belite cements, sulfoaluminate cements) could further reduce the overall environmental impact of cement when used in conjunction with metakaolin and limestone. Hence the interactions with these alternative binders and coupled substitutions of metakaolin and limestone should be further investigated. A related point to consider is the optimization of the ternary blend cements with varying metakaolin content.