Air Emissions Impacts of Fuel Substitution in the United States Cement Industry Authors: Anaís Ostroski, Vi H. Rapp, Corinne D. Scown Published: 2026 ABSTRACT High-heat industrial processes cannot be easily electrified, often rely on comparatively high-emitting solid fuels, and are essential for economic growth. Cement production in particular is critical to supporting construction and the expansion and maintenance of infrastructure systems in countries around the world. Fuels for cement production include petcoke, coal, and waste solvents, and their combustion contributes approximately 40% of the CO2 emissions from global cement production. Broadening the fuel resources to include agricultural residues and other wastes can reduce emissions and divert waste from landfills. We assess the potential to use 16 different alternative fuels across the 88 integrated cement plants in the United States. The share of alternative fuels in the cement production sector could increase from the current level of 17% to as high as 73%. Fuel substitution can reduce the current fuel-related life-cycle emissions (28.6 ± 5.7 Mton CO2/year) by 19–45%. The greatest national-scale human health cobenefits occur when alternative fuels are prioritized at facilities currently using coal and petcoke. NO x emissions outcomes from this strategy are mixed, but SO x and particulate matter emissions reductions drive net human health benefits. These results suggest that targeted fuel substitution in heavy industries can yield outsized emissions benefits. Introduction Domestic cement production is critical to supporting growth in the construction industry, repairing aging infrastructure systems, and addressing the chronic undersupply of housing. As a sector, it is also emissions-intensive. Approximately 60% of carbon dioxide (CO2) emissions stem from the calcination process, an unavoidable reaction in which calcium carbonate (limestone) is heated to form calcium oxide (lime) and CO2. The remaining 40% of CO2 emissions are attributed to the combustion of fuels needed to achieve the high temperatures required for clinker production. Process innovations, such as the replacement of limestone with alternative raw materials, the use of supplementary cementitious materials (SCMs) to replace clinker, and carbon capture and storage (CCS), offer promising pathways for long-term emissions mitigation. In contrast, the use of alternative fuels represents the most practical near-term strategy for reducing combustion-related emissions. This approach involves substituting conventional fuels (coal, natural gas, and petroleum coke) with biomass, municipal solid waste, fats, oils, and greases (FOGs), hazardous wastes, and other materials. About 73% of the cement plants in the United States use some alternative fuels, mostly tires and waste solvents, but the total share of alternative fuels is only 16%. In contrast, the share of alternative fuels in the European Union (EU) cement industry is 53%. − However, the total potential for fuel substitution within the U.S. cement industry has not yet been comprehensively quantified. The adoption of alternative fuels is associated with concerns beyond climate effects due to emissions of criteria air pollutants (PM2.5, NO x , and SO x ), as such exposures are known to have negative impacts on human health. Air pollutant emissions are a function of the interactions between chemical composition, kiln design and operation, and emission control technologies. Several studies have sought to characterize air pollutant emissions associated with the use of alternative fuels. Fyffe et al. compared coal combustion with partial replacement with an alternative fuel consisting of postconsumer plastics and fiber. They found that life-cycle CO2 and SO2 emissions reduced, while NO x emissions increased, likely due to increased temperature near the fuel injection. Prisciandaro et al. observed divergences when analyzing two cement plants with petcoke as a baseline: emissions of NO x and SO2 decreased in one plant and increased in another plant. Richards and Agranovski reached broadly similar conclusions via monitoring cement plants using alternative fuels, finding that alternative fuels either sustained or reduced emissions of CO2, NO x , and SO x under normal operating conditions. Despite this growing body of evidence, few studies have estimated the population-level health impacts of criteria pollutant emissions associated with widespread alternative fuel use in cement production in the United States, a gap that the present study aims to address. Burning alternative fuels can be an operational challenge. The physical and chemical characteristics of the fuel are critically important because fuel comes into direct contact with raw materials that are converted to clinker. Physical characteristics of the fuels, such as particle size, can delay ignition or cause incomplete combustion, leading to uneven temperature profiles and flame instability. The chemical composition of the fuels has direct effects on process energy efficiency and cement quality. Alternative fuels may vary in moisture content, level of contamination, and heating value, impacting fuel handling, feeding, and emissions compliance. Preprocessing steps such as drying, shredding, grinding, and densifying can partially address these challenges. For example, refuse-derived fuels (RDF) and, more recently, solid recovered fuel (SRF) use preprocessing to transform municipal solid waste into densified pellets of more predictable quality. − There are 88 clinker-producing cement plants across the United States, and there are open questions as to how much additional alternative fuels these facilities could access and how the use of such fuels would impact air emissions. Here, we address these questions by (i) estimating the availability of suitable alternative fuels within practical shipping distances of cement plants in the United States; (ii) developing emissions estimates for conventional and alternative fuels as a function of fuel composition, kiln systems, and emissions control technologies; and (iii) estimating current facility-level fuel mixes and modeling scenarios to evaluate the consequences of fuel switching. Methodology Figure a shows a schematic representation of the methodology. To quantify emissions associated with fuel combustion in cement kilns, we first simulated business-as-usual (BAU) fuel usage at the facility level. Data on industry-wide fuel consumption and types of fuels used by each facility were obtained from the American Cement Association (ACA), formerly Portland Cement Association (PCA), , which also breaks down total fuel consumption by categories of kiln, year modernized, capacity, and region. In addition, the Resource Conservation and Recovery Act provides receipts of hazardous waste for use as fuels at individual cement plants. The amount of energy needed to produce one ton of clinker was obtained from facility-specific environmental product declarations (EPD). When EPDs were not available, we estimated the energy demand based on a hierarchical imputation by leveraging the ACA categories. With this information, we built an iterative optimization model to simulate potential fuel mixes under the uncertainty. We rely on the BioSiting Tool and the Environmental Protection Agency (EPA) Recycling Infrastructure and Market Opportunities database to supply information about the availability of alternative fuels. Subsequently, we developed air emission factors for each fuel and kiln combination, which were coupled with our estimate of fuel use under different scenarios to obtain facility-level emissions. We considered emissions of carbon dioxide (CO2), nitrogen oxides (NO x ), sulfur oxides (SO x ), and particulate matter (PM2.5). We used the intervention model for air pollution (InMAP) to translate emissions into estimated changes in human health and monetary impacts. (a) Assessment framework for the impact of fuel switching in cement kilns on greenhouse gas and air pollutant emissions; (b) map with estimated fuel mix for each facility and total fuel utilized (in MJ); (c) share of alternative fuel availability and total feedstock availability (MJ). Alternative Fuel Availability We considered 16 different alternative fuels within 4 broader categories: (1) agricultural residues (corn stover, sorghum stover, wheat straw, oat straw), (2) forest residues (softwood, hardwood, mixedwood), (3) municipal solid waste (tires, grease trap waste, plastics #3–7, paper, textiles, urban wood, yard trimmings), and (4) Resource Conservation and Recovery Act (RCRA) waste (waste solvents, hazardous waste, waste oil). While many other types of waste materials are potentially available, such as carpets, wind turbine blades, and railway tracks, we have selected the fuels based on data availability and the supply predictability. We leveraged the BioSiting Tool to gather availability of most feedstocks around a radius of 80 km around cement facilities. For plastic bales and tires, we utilized the Recycling Infrastructure and Market Opportunities Map from the EPA to obtain estimates of the amount of unrecycled plastics and tires by zip code. We also assumed that 2% of plastics bales #3–7 are PVC, which is not allowed to be combusted in cement kilns due to chlorine content. After adjusting for PVC, we refer to this resource as plastics #4–7 or simply plastics. Waste covered by the Resource Conservation and Recycling Act was obtained via the biennial report of 2023 and the Toxic Release Inventory Program, whose wastes were categorized as waste solvents (W202: Halogenated Solvent, W203: Nonhalogenated Solvent, W204: Solvent Mixture), waste oil (W206: Waste Oil), and hazardous waste (W210: Reactive organic liquid, W209: Paint, inks, etc., W211: Paint Thinner, W406: Dried paint, W604: Paint, ink sludge). Business-as-Usual Fuel Mix Estimate The initial list of cement plants was obtained from the Global Database of Cement Production Assets and Upstream Suppliers and reports from the American Cement Association. ,, The cement facilities were categorized according to their type of kiln (preheater, preheater/precalciner, long wet, long dry), capacity (smaller or greater than 1 Mt/year), and region. Kiln-type information was obtained from Title V operating permits. Out of the 115 cement plants in the United States, 88 of them are integrated plants that produce clinker, while the remaining facilities are grinding only. For the integrated plants, we obtained 56 Environmental Product Declarations to obtain the average energy demand of each plant. We estimated the amount of energy consumed per unit of cement by summing up energy declared as (i) Renewable primary resources used as an energy carrier (fuel), (ii) Nonrenewable primary resources used as an energy carrier (fuel), (iii) Renewable secondary fuels, and (iv) Nonrenewable secondary fuels. For the facilities that did not have accessible EPDs, energy requirements were estimated with a hierarchical imputation approach, leveraging the aforementioned four categories. Averages were taken for each combination of categories. If no data were available for a given group, we assumed that the kiln type is a robust predictor of energy use. The current facility-level fuel mix is built upon a set of best guesses herein referred to as the prior (fi,j^) . First, the most foundational piece of information is the amount of each fuel used by each category, as it was used to derive fuel mixes by kiln type, capacity, and region. For each facility, we estimated a fuel mix by normalizing the average for their respective categories. Subsequently, the prior was updated based on more specific information. The PCA plant information summary was used to obtain which types of fuels are utilized by each facility. Most importantly, this provided which fuels were not used and thus the respective percentage for that fuel should be kept at zero. For waste fuels classified as hazardous, the Resource Conservation and Recovery Act Biennial Report was used to obtain receipts of hazardous waste, waste solvents, and waste oil. A permit is necessary to handle hazardous wastes, so only facilities listed in the biennial data set were assumed to combust RCRA waste. The amount of hazardous waste and waste solvents received was converted to energy based on average lower heating values, and the remaining fuels utilized by the plants were adjusted proportionally. For other waste types, such as SRF or waste tires, public information about adapted infrastructure for this type of fuel was considered. For example, the presence of the Pelican System by Entsorga or the HOTDISC by FLSmidth was taken as evidence for the combustion of SRF and tires, respectively. The posterior facility-level fuel mix was obtained with an iterative optimization model. This approach was chosen to make sure certain constraints would be respected even under uncertainty. The objective function (eq 1a) was set up to minimize the difference between the posterior and the prior with the addition of a regularization factor. Constraint in eq 1b ensures that the total amount of fuel for all facilities respects the maximum fuel usage reported by the Portland Cement Association, constraint in eq 1c guarantees that the sum of fractions is one, constraint in eq 1d makes sure that fuel j for each facility i is kept at zero if that fuel is not used by facility i, constraint in eq 1e ensures non-negativity. minimize∑i,j(fi,j−fi,j^)2+λ∑i,jfi,j(1a)subjectto∑iEifi,j=Ftotal,j⁣∀j(1b)∑jfi,j=1⁣∀j(1c)fi,j=0⁣∀(i,j)∈gi,j(1d)fi,j≥0⁣∀i,j(1e) 1a For each of the 1000 iterations, the nonzero values of the prior fi,j^ were perturbed with values from a uniform distribution 2 fi,jnois^y=fi,j^+U(a,b)∑jfi,j^+U(a,b) Fuel Substitution Scenarios We built substitution scenarios to meet energetic demands under different constraints based on a fuel availability network. We analyzed six national scenarios and compared them to a business-as-usual case (Table ). The scenarios were designed along two sets: the first set reflects near-term regulatory and industry constraints that cement plants are likely to face, while the second set explores theoretical widespread alternative fuel adoption and availability. 1Scenario Names and Descriptions #scenariodescription1BAUbusiness-as-usual fuel mix estimate220% fossil reductionsubstitution of up to 20% of fossil fuel (coal, petcoke, natural gas) with alternative fuels3moderate deploymentcaps biomass at 5%, tires at 20%, plastics at 10%4aggressive deploymentcaps biomass at 15%, tires at 20%, plastics at 10%5natural gas lock-inkeep current natural gas levels, only coal and petcoke are substituted6unrealized potentialmaximize alternative fuel use with feedstocks available today7SRFmaximize alternative fuel use with feedstocks available today and increased SRF productionaAll constraints on alternative fuel blending are implemented at the individual facility level. Scenario Names and Descriptions All constraints on alternative fuel blending are implemented at the individual facility level. The first three scenarios (20% fossil reduction, moderate deployment, and aggressive deployment) are heavily constrained. The 20% fossil reduction scenario caps alternative fuel blending at 20% of the current total fossil fuel use in each facility. Facilities that already blend more than this amount are capped at their current blending level. A 20% substitution represents a conservative target that can be achieved with minimal infrastructure changes and acknowledges the practical limits and near-term goals of fuel replacement. The Moderate and Aggressive Deployment are based on goals set by the American Cement Association roadmap to carbon neutrality and compiled by the U.S. Department of Energy as scenarios for 2030. Moderate deployment has a goal to satisfy 35% of fuel needs with waste-based fuels, with facility-level limits of biomass to 5%, tires to 20%, and plastics to 10% of total fuel needs. The Aggressive Deployment scenario builds upon the Moderate Deployment scenario by increasing the biomass limit to 15%. The fuel-specific caps reflect current supply chain constraints and operational limits, including the presence of zinc and sulfur in tires, possible chlorine in plastics, and the cost and handling of biomass. The latter three scenarios (Natural Gas Lock-In, Unrealized Potential, and SRF) are more flexible and generally aim to maximize the use of alternative fuels while meeting each facility’s existing energy requirements. In the Natural Gas Lock-In scenario, alternative fuels are only used to displace petcoke and coal, while facility-level natural gas consumption is kept at its current levels. Natural gas has become increasingly common in U.S. cement plants, and many facilities invested in gas-firing infrastructure. This scenario reflects the realistic possibility that facilities will continue using natural gas and will only displace solid fuels. The Unrealized Potential scenario maximizes fossil fuel substitution with alternative fuels available to each cement plant and applies no other constraints. It serves as a theoretical upper bound based on today’s availability of alternative fuels, stripped of operational and supply limitations. The SRF scenario is similar to Unrealized Potential but assumes that additional SRF production capacity can be built to increase SRF supply beyond current levels, constrained only by availability of its waste inputs. It is motivated by the fact that SRF is not constrained by municipal solid waste generation rates but by processing infrastructure. It identifies the benefits if policy or market incentives stimulate SRF production capacity. Below, we provide the mathematical details of scenario modeling. Sets S sources of alternative fuels. C set of conventional fuels. Ca set of current alternative fuels of unknown origin as defined by ACA that will remain unchanged (biomass, renewable wood and agricultural, renewable seeds and shells, tire fluff, other solids). A set of alternative fuels. B set of biomass fuels. Decision Variables n s,i,j network of the flow of fuel j between fuel source s and facility i. e i,j energy of fuel j used by facility i Mathematical Formulation 3a minimize∑i,j∈Cei,j subjectto∑jei,j=Di⁣∀i(3b)∑iei,j=∑sMs,j⁣∀j(3c)∑ins,i,j≤Ms,j⁣∀s,j(3d)ns,i,j≤Ns,i,j⁣∀(s,i,j)(3e) 3b 3f ∑sns,i,j≤fi,j⁣∀j∈C 3g ∑sns,i,j≥fi,j⁣∀j∈A 3h ∑sns,i,j=fi,j⁣∀j∈Ca where D i is the energetic demand of facility i, M s,j is the maximum availability of fuel j at source s, N s,i,j is the original fuel availability network, and f i,j is the current usage of fuel j by facility i. All variables are expressed in MJ. Equality constraints were modeled to give the model flexibility of ±2%. The original fuel availability network (N s,i,j ) is a compilation of source–destination information on the alternative fuel as detailed in the Alternative Fuel Availability section. For most feedstocks, it is a local network of the sources and maximum amounts of feedstock within 80 km of the cement plant. For fuels such as waste solvents that are already transported for longer distances, it refers to the amount of waste solvents transferred to cement plants according to the RCRA biennial report. In short, it dictates the maximum amount of feedstock that can be transferred from a given source to the cement plant. For the Fossil 20% scenario, constraint in eq was added to indicate that the alternative fuel usage must be equal to or lower than the current alternative fuel usage plus a maximum of 20% of fossil fuel usage. 4 ∑j∈Aei,j≤0.2∑j∈Cfi,j+∑j∈Afi,j⁣∀i For the Natural Gas Lock-In scenario, natural gas was treated as part of the set Ca , not C, meaning it was not allowed to change. For the SRF scenario, the model remained the same as Unrealized Potential, but the input data changed to reflect higher production of SRF. This was modeled with an assumption that each metro area with more than one million inhabitants would have the capacity to produce SRF from paper, textiles, plastics, and urban wood that currently go to the landfill. The Moderate Deployment scenario caps biomass use at 5%, plastics at 10%, and tires at 20% of fuel mix to depict a short-term scenario with consideration of possible chlorine contamination in plastics and zinc and sulfur content of tires, as well as logistical hurdles to increase biomass use as fuel. The Aggressive Deployment increases biomass use allowance to 15%. The goals described by the PCA roadmap and DOE refer to industry-wide fuel mix; however, since the reasoning for capping plastics and tires are due to composition, we set a maximum at the facility-level. 5 ei,j≤0.2Di⁣forj=‘Tires’,⁣∀i 6 ei,j≤0.1Di⁣forj=‘Plastics’,⁣∀i 7 ei,j≤0.05Di⁣for ⁣j∈B,⁣∀i Development of Emission Factors Emissions generated by fuel combustion in cement kilns are dependent on many different factors, such as the composition of fuels and operational conditions, including injection velocity, temperature profiles, and emission controls. This indicates that there is a level of uncertainty in emission estimates associated with the combustion of a given fuel that is difficult to capture with emission factors. The objective here is to estimate potential emissions resulting from fuel combustion as they relate to fuel composition. Emission factors are also a function of the type of emission control available and enforced by regulatory agencies. We utilized Title V operation permits to gather information about the use of emission controls for criteria air pollutants. Emission Factors for the Combustion Process We conducted a literature search to gather information on the composition of feedstocks. We searched for studies that conducted proximate and ultimate analysis on our target feedstocks and selected 105 studies. The compiled information and their respective references can be found in the Supporting Information. Carbon Dioxide For the purposes of CO2 accounting, we assumed 100% of the carbon in each fuel is converted to CO2 during combustion in the kilns, as the fraction of unconverted carbon is typically negligible. Biogenic carbon was not included in the CO2 emissions, and biogenic carbon content varied. Agricultural and forest residues, as well as animal fats and greases have all of their carbon as biogenic. Textiles were assumed to be 55% biogenic. For the engineered fuels (SRF and RDF), we assumed a composition of 30% paper, 20% wood, 35% plastic, and 15% textilestherefore, a biogenic carbon content of approximately 58%. Nitrogen Oxides Nitrogen oxides (NO x ) are primarily generated via the oxidation of nitrogen in the fuel (fuel NO x ) and oxidation of atmospheric nitrogen under high temperatures (thermal NO x ). On the basis of fuel chemistry, coal has the potential to produce more NO x than natural gas per unit of heat since it contains nitrogen. However, the equipment used to combust each fuel can significantly impact the level of NO x production. In the case of natural gas, which has no inherent nitrogen, combustion strategies may be focused on reducing the flame temperature to control thermal NO x formation. Because coal contains a substantial amount of nitrogen, strategies may focus on staged combustion or reducing the amount of oxygen available to prevent fuel and thermal NO x formation. In this study, we first estimate total uncontrolled NO x emissions associated with the combustion of each fuel regardless of combustion strategy, and then we apply reductions based on the efficiency of each control technology. The conversion of fuel-bound nitrogen to NO x during combustion is a complex process influenced by local oxygen availability (air-fuel ratio), temperature, residence time, and fuel-nitrogen speciation. It is not reasonable to assume that all fuel-bound nitrogen is converted to NO x because a fraction may react with other fuel components, bind to ash particles forming particulate matter, or be emitted as N2. We obtained ratios of fuel nitrogen to the total nitrogen emitted (prior to any emissions controls) as NO x from a literature search on experimental studies that have measured NO x emissions from fuel combustion in different temperatures and oxygen levels (Table S4). This information was used to estimate the amount of uncontrolled NO x that is expected to form during combustion of a given fuel. We attempted to capture differences in NO x formation by kiln type based on EPA’s estimation of emission factors for each type of kiln and the difference relative to preheater/precalciner kilns. Where ratios of fuel nitrogen to total NO x formation were not available, we adopted NO x emission factors from the Argonne National Lab’s GREET Model. Therefore, we estimate the NO x (as NO2) emissions as 8 NOx=cr×Nf×46.005514.0067 where cr is the fraction of fuel-bound nitrogen emitted as NO x , and Nf is the mass of nitrogen in the fuel per unit energy. Control technologies to mitigate release of nitrogen oxides include flame cooling, low-NO x burners, midkiln firing, staged combustion, selective noncatalytic reduction (SNCR), and selective catalytic reduction (SCR). Low-NO x burners are designed to reduce both thermal and fuel-bound NO x by controlling local stoichiometry (air-fuel ratio) and temperature. Midkiln firing refers to a secondary fuel injection mechanism to burn part of the fuel at lower temperatures. Staged combustion refers to the process of staging the introduction of fuel, air, and raw material to minimize NO x formation. SNCR utilizes an ammonia solution, which, in the presence of oxygen, reacts with NO x to form N2 and water. Similarly, SCR also relies on ammonia, but with the addition of a catalyst such as titanium oxide. We adopted removal efficiencies from studies by the EPA, − as detailed in Table S5. Finally, the NO x emission factors (EFN) for each fuel and control technology were calculated as 9 EFN=NOx×(1−Rc) where Rc is the removal efficiency of control technology c. Sulfur Dioxide Sulfur is introduced in the kiln system in the form of organic compounds, sulfides, sulfates, or elemental sulfur from both raw materials and fuel. Once in the kiln, sulfur is rapidly converted to SO2, reaches a peak, and gradually decreases with time. The form of this conversion is dependent on temperature, oxygen concentration, moisture, and particle size. Typically, approximately 40 to 70% of sulfur dioxide is released due to sulfur in the raw materials and the remaining from fuel sulfur. Cement rotary kilns are characterized by stoichiometry excess, with Ca to S molar ratio often greater than 10:1, which results in the formation of free CaO. Under high temperatures and an oxidizing environment, sulfur dioxide reacts with CaO and forms CaSO4, which is stable under oxidizing conditions. This inherent SO2 removal ranges in efficiency due to chemical balance, chemical form of sulfur, and fuel characteristics. For example, when the sulfur is introduced mostly in the form of sulfides, removal may be lower. Because the conditions that affect sulfur absorption vary with the type of kiln, we adopted inherent removal rates accordingly, as shown in Table S1. This inherent removal ranges from 38% to as high as 98% depending on kiln type. In addition to inherent removal, there are control technologies for additional removal of sulfur dioxide. Here, we consider four types of control: dry sorbent injection, hot meal injection, spray dryer absorber, and wet SO2 scrubber. Dry sorbent injection systems introduce reagents such as hydrated lime or sodium bicarbonate. Hot meal injection utilizes calcined raw material from the kiln as a source of calcium oxide. The spray dryer absorber involves spraying a lime slurry into the flue gas. Wet scrubbers utilize a liquid, such as a water-based lime slurry solution, to capture SO2 as the flue gas passes a scrubbing chamber. We adopt efficiency ranges from Miller and Young (Table S2). Control technologies have efficiencies of reaching 95%. The removal rate for combinations of kiln type and control technologies is calculated as 10 Rt=Ri+(1−Ri)×Rc where Rt represents the total SO2 removal rate, R i the inherent SO2 removal according to the type of kiln, and Rc the removal from control technologiesapplied to the portion of sulfur dioxide that is not inherently removed. The SO2 emission factors (EF s ) was calculated per kiln per control technology, on the amount of sulfur dioxide that is not removed: 11 EFs=Sf×64.06632.065×(1−As)×(1−Rt) where Sf is the sulfur content in the fuel, As is the amount of sulfur that is absorbed by ash in the kiln assumed to be 25% according to Liu et al., and the fraction 64.066/32.065 is the molecular ratio of sulfur dioxide to elemental sulfur. Particulate Matter We relied on the Argonne National Lab’s GREET Model for PM2.5 emission factors. While the model offers specific emission factors for natural gas and coal combusted in kilns, we adopted emission factors for combustion in industrial boilers, assuming that they provide the closest estimate of PM2.5 emissions for other fuels. For municipal solid wastes, the model is limited to PM2.5 from incinerators. These assumptions add substantial uncertainty as the combustion technologies are fundamentally different, with different temperature profiles, oxygen levels, and pressure, for example. This is a limitation to our particulate matter emissions estimates and must be interpreted as approximations. In addition, these emission factors were assumed to represent uncontrolled combustion. Most cement kilns have a series of fabric filters, or baghouses, according to the emission controls information gathered from the Title V permits. These dust controls capture at least 99% of PM2.5, and we used this 99% capture value to estimate the controlled PM2.5 emissions. Emission Factors for Upstream Processes Alternative fuels need to be preprocessed before being ready to be combusted in the kilns. This preparatory stage ensures optimal energy release and minimizes the risk of operational issues. For instance, agricultural residues and forest residues were specifically identified as materials benefiting from densification. This densification process transforms loose, low-density biomass into a more compact and uniform fuel. This not only improves their handling, storage, and transportation efficiency but also enhances their combustion characteristics. Municipal solid wastes were assumed to be transformed into RDF or SRF with processing steps as detailed by Ferronato et al. Grease and livestock fat were assumed to go through oil extraction and waste treatment according to Hums et al. We relied on a series of studies to obtain estimates of carbon dioxide emissions of preprocessing steps, ,− with their respective values laid out in Table S6. Counterfactuals Counterfactuals are hypothetical scenarios to represent what is currently happening to the alternative fuels. Here, they are utilized to better understand the consequences of adopting the materials as alternative fuels in cement kilns. This approach allows for a comprehensive understanding of potential consequences, both positive and negative, that might arise from such an adoption. For instance, agricultural and forest residues are frequently left on the field to decompose. While this process contributes to soil health and nutrient cycling, it also represents a missed opportunity for energy recovery, and trade-offs must be better understood. Yard trimmings are typically managed through composting, but this highly depends on local infrastructure. Municipal solid wastes are predominantly disposed of in landfills. RCRA wastes were assumed to be incinerated without energy recovery. By systematically comparing the hypothetical adoption of these materials as alternative fuels against their current management practices, the analysis aims to provide a robust framework for decision-making. This involves quantifying potential emissions reductions while also understanding shifts in waste management infrastructure and regulatory considerations that would accompany such a transition. Again, we relied on published studies to obtain estimates of emissions associated with each counterfactual, ,− with values detailed in Table S7. Results We estimated the sector-wide energetic requirements to be approximately 418 million GJ (lower heating value, LHV). Most facilities rely on coal and natural gas to supply energy for operations, but a few facilities rely on hazardous waste and waste solvents to meet a majority of their energy needs (Figure b). We estimate that approximately 735 million GJ (LHV) of alternative fuel is available within practical shipping distances (80 km radius for solid fuels) of cement plantsequivalent to 1.75 times the sector-wide energy demand (Figure c). The most abundant and widely available alternative fuel is corn stover. Emission Factors from Fuel Combustion Carbon dioxide emissions vary widely by the type of fuel. Average CO2 emissions from fuel combustion in cement kilns range from zero, if biogenic (all carbon is treated as part of the contemporary cycle), to 107 g CO2/MJ of fuel according to the carbon content of the material. Figure shows the net change in emissions per MJ of fuel if each alternative fuel is used in place of coal (left-hand bar for each fuel) and the net emissions associated with its counterfactual (business-as-usual management practices for the waste, on the right-hand bar for each fuel). Most alternative fuels emit less CO2 than coal (Figure ). The majority of alternative fuels reduce CO2 emissions if used in place of coal, with two exceptions: nos. 4–7 plastics and tires. Because tires and plastics do not break down in the landfill on a time frame that would cause any significant methane emissions, the counterfactual emissions for those materials are comparatively low. There is additional uncertainty associated with the counterfactual for tires, which are sometimes recycled to produce materials such as rubberized mulch and playground surfaces, but it is not well documented what materials are being displaced in that case. If, however, these materials were incinerated in the counterfactual (as is the case for waste solvents), their use as alternative fuels would reduce net emissions. Waste solvents cannot be landfilled; we assume that their counterfactual disposal would be incineration without energy recovery. This process requires the addition of fuel oil (ancillaries), which further increases the emissions footprint of the counterfactual and thus the net benefit of using solvents in cement kilns. Net change in life-cycle CO2 emissions per MJ of fuel if each alternative fuel is used in place of coal (lefthand bars) and the net counterfactual emissions from business-as-usual management practices for the waste (right-hand bars). Substitution is net beneficial for emissions if each fuel’s lefthand net value is below the value on the right-hand bar. This comparison assumes the same combustion technology throughout and that 100% of the fuel carbon is converted to CO2. SRF: solid-recovered fuel, RDF: refuse-derived fuel, C&D: construction and demolition. Unlike CO2, the emissions for criteria air pollutants and their relation to alternative fuel chemical composition are less straightforward to estimate. Figure shows the emission factors for SO x and NO x for combinations of fuel, kiln type, and emission controls. Most of the SO x formed from fuel sulfur during combustion is scrubbed in the kiln due to ideal conditions for reaction with calcium oxide, forming CaSO4, which is stable under kiln conditions. Emission factors for SO x can reach 0.6 kg of SO2/GJ (Figure a), with the highest-emitting fuel being petcoke, due to its higher sulfur content ranging from 1 to 6%, when paired with a long wet kiln with no sulfur emissions control technology. In contrast, petcoke combustion in preheater/precalciner kilns without additional control emits 0.04 kg of SO2/GJ, on average. Combustion of coal, tires, and waste solvents in kilns without a preheater, precalciner, or sulfur control technology may emit between 0.1 and 0.3 kg of SO2/GJ of fuel combusted. Emission factors for (a) SO x and (b) NO x for each combination of fuel, kiln system, and emission control technology. SCR: selective catalytic reduction; SNCR: selective noncatalytic reduction; SRF: solid-recovered fuel; RDF: refuse-derived fuel. Figure b shows the emission factors for each fuel, type of kiln, and NO x control technology. Here, the worst case is a long wet kiln without any controls, while the best case is preheater/precalciner kilns with both low-NO x burner and selective catalytic reduction system (SCR). SRF and RDF had the highest NO x emission factors. This is mainly due to the nitrogen content of the components, which can potentially be manipulated in manufacturing to meet specific needs. More specifically, the presence of PE plastics may cause high NO x emissions due to potential elevated combustion temperatures and the conversion of atmospheric nitrogen to NO x (thermal NO x ), and textiles such as leather may be composed of about 10% nitrogen, resulting in high fuel NO x emissions. Effects of Fuel Switching Net life-cycle CO2 emissions associated with fuel use in cement kilns in the BAU scenario were estimated to be 28.6 ± 5.7 Mton CO2e annually. Figure a shows net life-cycle CO2e emissions, including upstream, combustion, and credit for counterfactuals for each scenario. Figure b shows the results of the optimization models to estimate fuel mix, averaged for the sector for each scenario. Figure c shows the relative difference in CO2 emissions associated with fuel combustion for each facility. As the industry-wide fossil fuel share of total fuel combustion decreases from 83% (BAU) to 60% (Moderate Deployment), net CO2 emissions decline by 18%, to 23.2 Mton CO2. Further relaxing the constraint on biomass to allow up to 15% blending of agricultural residues results in net life-cycle emissions of 21.9 Mton CO2 in the Aggressive Deployment scenario. The flexible scenarios may result in relative emissions reduction between 41 and 52% because of the larger shares of biogenic sources and credits for counterfactuals. In general, facilities located in the Midwest have higher potentials to reduce CO2 emissions due to the abundance of agricultural residues. The fuel combustion emissions, as in what would be practically measured at the kiln stack, could be potentially reduced by 44% between BAU and Unrealized Potential. Carbon dioxide equivalent emissions by scenario. (a) Industry-wide net life-cycle emissions, with credit for counterfactuals and upstream processes, dots represent net value; (b) industry-wide average fuel mix by scenario with the percentage of conventional fuel shares, where conventional fuels are defined as coal, petcoke, and natural gas; (c) facility-level relative difference of CO2 emissions from fuel combustion from BAU. BAU: business-as-usual, SRF: solid-recovered fuel. In addition to CO2 emissions, a key consideration in fuel substitution is the impact on air quality. NO x , SO x , and primary PM2.5 are particularly important for estimating total changes in atmospheric PM2.5 concentration and the resulting impacts on human health. The changes in concentration and resulting population impacts are calculated using InMAP. An increase of PM2.5 by 10 μg/m3 is associated with a 14% increase in risk of premature mortality. Monetary damages are based on the value of statistical life ($13.7 million), a measure of the cost-benefit of regulations, or the willingness of society to pay for reductions in mortality risks, whose value is determined by the U.S. Department of Health and Human Services. Figure compares the mean of human health costs for each scenario (a), maps the human health costs for BAU (b), and shows the relative change in combustion emissions at the facility-level as conventional fuels are replaced. The more constrained scenarios (Fossil 20%, Moderate Deployment, and Aggressive Deployment) do not result in a statistically significant difference in human health costs from the BAU scenario (p = 0. 69,0. 91,0. 23, respectively). The more unconstrained scenarios (Natural Gas Lock-in, Unrealized Potential, SRF) do not produce human health impact differences that are statistically significant among themselves; however, the SRF case has a slightly higher mean than the Natural Gas Lock-In and Unrealized Potential scenarios because it can lead to higher emissions of NO x (Figure c). Nevertheless, there is a clear distinction between the two sets of scenarios (p < 0.0001), with average monetized human health damages dropping from $36,082.54 ± 370.55 (BAU) to $25,581.38 ± 370.17 (Natural Gas Lock-In), suggesting that eliminating the alternative fuel blending constraints produces meaningful improvements in human health impacts from air pollution. Comparison of human health cost and combustion emissions by scenario. (a) Mean human health cost, where each value was obtained by averaging the costs of all cells in the grid for a given scenario. Cells vary in size to capture spatial variation in population, smaller cells in urban areas and larger cells in rural areas. (b) BAU human health cost map, where black dots show the location of the cement facilities. (c) Relative change from BAU of the combustion emissions of SO x , NO x , and PM2.5each dot represents emissions change in an individual facility, colored by scenario. Lower numbers mean higher fossil fuel displacement, e.g., 100% means all fossil fuel was replaced by alternative fuels and 0 means fossil fuels in the fuel mix remain unchanged. Although each facility’s impacts differ depending on the kiln design, control technologies, and locally available fuels, some general trends do emerge (Figure c). When the share of fossil fuels in the fuel mix decreases, PM2.52 emissions generally decrease (R 2 = 0.83). SO x and NO x emissions may either increase or decrease depending on which alternative fuels displace fossil fuels. In the SRF scenario, NO x emissions increase most of the time due to higher NO x emission factor for SRF. The Fossil 20% scenario shows that, for the same level of fossil fuel substitution, emissions of criteria air pollutants have the potential to increase or decrease. Prioritizing replacement of coal and petcoke without changing natural gas levels may lead to the highest returns for SO x reduction as emissions can decrease at a higher rate than fossil fuels are substituted (dark teal area). This is of particular importance because SO x is the primary driver of human health costs, with a total of $676 million, making up 54% of total costs, while NO x is linked to $571 million (45%) and primary PM2.5 to $4 million (<1%). For every kilogram of SO x emitted, we found an associated human health cost of $187.36. This is lower than the rate for PM2.5 ($373.05/kg PM2.5) and higher than NO x ($44.95/kg NO x ). Figure c highlights that substituting fossil fuels with alternative fuels will almost always lead to a decrease in PM2.5 emissions; however, the choice of alternative fuels must be done systematically according to the kiln design, control technologies, and fuel characteristics to ensure NO x and SO x can be mitigated via operational decisions and will not surpass emission limits. Discussion In this study, we estimate life-cycle GHG and combustion criteria air pollutant emissions associated with fuel use in the cement industry, taking into consideration that each of the 88 facilities has different energetic requirements, kiln designs, and local feedstock availability. Our results show that decreasing fossil fuel shares leads to a reduction in life-cycle CO2e emissions, especially in the Midwest where agricultural residues are abundant. The substitution of up to 20% of fossil fuels in each facility with alternative fuels resulted in a 10% decrease in facility-level combustion CO2e emissions and a 19% reduction in net life-cycle fuel-related (combustion and upstream) CO2e emissions for the entire U.S. sector. In comparison, Hasanbeigi and Bhadbhade came to different conclusions. They found that switching to alternative fuels had a small impact on total GHG emissions. In one scenario, they replaced 20% of coal and petcoke with tires, resulting in a 1% reduction in total sector-wide CO2 emissions (including calcination emissions). They modeled biomass as an alternative fuel but treated biogenic carbon as indistinguishable from fossil carbon, which is not aligned with common practices in life-cycle emissions accounting. Thus, they found that substituting biomass for 20% of coal and petcoke use would reduce sector-wide GHG emissions by only 0.4%. Fyffe conducted experimental tests to substitute 10 to 20% of the precalciner energy requirements with SRF, but this study did not define a counterfactual for the waste material or model biogenic carbon differently from fossil carbon; thus, they also found very modest life-cycle GHG reductions between 1.4 and 1.6%. Here, we solely compared emissions of fuel use, excluding calcination, and included counterfactuals to measure the change caused by the diversion of waste and residue materials to use as fuel. Emissions of criteria air pollutants can either decrease or increase as alternative fuels are introduced into the kiln, but there was a statistically significant reduction in air pollution-related health costs when the industry-wide fossil fuel share is below 50%. Emissions of PM2.5 decreased with rising alternative fuel shares and, consequently, diminished impacts to human health costs. However, chemical composition and size distributions of PM2.5 vary across fuel types, and their respective toxicity levels have been shown to differ depending on the source. , InMAP does not yet incorporate differences in PM2,5 toxicity. Instead, it focuses on primary and secondary PM2.5 because such concentrations are considered to be the primary driver of premature mortality from air pollution and are comparable across studies. Future work may update health assessments as more data on the PM2.5 composition become available for each fuel. The combustion of fuel derived from waste may increase NO x and SO x emissions; however, the emission factors developed here are based on stoichiometry and ratio of fuel-bound nitrogen and NO x formation and do not capture changes in operational conditionssuch as temperature and air-fuel ratio, which influence NO x and SO x formation. For example, Courtemanche et al. showed that the fraction of nitrogen and sulfur that converts to NO x and SO x , respectively, varies significantly with the air-fuel ratio and temperature profile. For temperatures similar to a cement kiln (1600 K), they showed that the conversion of fuel N to NO x from coal and tire-derived fuel can range from 50 to 10% with increasing equivalence ratios, because fuel-rich environments reduce NO x formation. As such, the temperature and air-fuel ratios can be manipulated during operations to control emissions. Similarly, operators may choose to burn specific fuels in the calciner or in the rotary kiln accordingly. NO x emissions tend to increase with higher fuel substitution rates, especially when energy is supplied by fuels derived from municipal solid waste (SRF and RDF). This is in agreement with Fyffe et al., who estimated a fractional increase of measured NO x emissions ranging from 25 to 93% with SRF injection rates of 1 to 2 ton/hour in comparison to a coal reference case (10–20% SRF as a fraction of total fuel input). In conversation with cement plant operators, Fyffe et al. pointed out that the spike in NO x emissions can be explained by a feeding system that is not designed for alternative fuels and lack of adjustments in burn conditions to control air-fuel ratios and residence time to limit NO x formation. Emission factors are a best estimate, and emissions are very specific to kiln design and other facility-specific conditions. In addition, the combustion of fuels in the rotary kiln or in the calciner leads to different outcomes. Here, we attempted to capture variation across kiln systems based on previously developed emission factors from AP-42. In terms of SO x emissions, oxidized sulfur tends not to contribute substantially to sulfur dioxide emissions, as most SO2 emissions are a result of pyritic sulfur in the raw materials. Nevertheless, the fuel also influences the release of sulfur from the raw material. Mut showed that fuel particle size resulted in an increase of SO2 emissions and validated a model that predicts SO2 stack emissions as a function of fuel particle size and other operating conditions. Fyffe et al. showed that life-cycle SO2 emissions decreased between 19 and 44% depending on the scenario and SRF substitution rate. In addition, kiln systems with an in-line preheater/precalciner and no alkali bypass do not release SO2 because the flue gas from the kiln is directed to the precalciner. Due to the lack of information on the specific design of in-line and separate line calciners, we have not incorporated this factor into emissions calculations. Therefore, SO2 emissions estimated in this study must be interpreted as an upper bound, which can be reduced through changes in the operating conditions. While the cement sector is not a primary source of persistent organic pollutants (POPs), it generates unintentional POPs such as dioxins and furans (PCDD/Fs). The mechanisms that cause the formation of POPs in cement kilns are complex; studies indicate that emission rates depend more on raw material organic content, residence time, and air pollution control-device temperature zones than on fuel type. , Ames et al. performed extensive testing in two cement facilities, testing coal, petcoke, and petcoke with hazardous waste and found that PCDD/F emission profiles did not significantly differ with hazardous waste but with facility characteristics. Conesa et al. observed no systematic differences in emissions with SRF, though sulfur in petcoke did correlate with PCDD/Fs. Chen et al. noted that coprocessing of waste chemicals can increase PDCC/Fs emissions in the humidifier tower and require operational control. Richards et al. found that firing tires generally lowered PCDD/F emissions, especially with decreasing inlet temperatures. Ye et al. indicated that emissions of PCDD/Fs can increase with hazardous waste but depend on type, copper content, and sulfur content. Given that the formation of POPs is complex and InMAP does not support its assessment, they were not considered in this study. Future research is needed to establish causal effects of fuel characteristics and quantify the respective emission factors. Beyond the need to adjust operating conditions to minimize criteria air pollutant emissions, the biggest challenge to alternative fuel adoption is logistical. While emissions reductions are achievable, fuel supply needs to be constant and reliable as cement kilns operate 7 days a week, most weeks, with fuel injections ranging from 10 to 22 tons of fuel per hour. , The availability of agricultural residues is inherently seasonal and their massive amounts require infrastructure to avoid losses or degradation. Implementing alternative fuels relies on the interest and technical expertise of cement producers, but also on the waste management sector. Current specialized alternative fuel production is limited, forcing cement plants to develop parallel expertise in sourcing, fuel processing, and chemical analysissuch as Geocycle, a subsidiary of Holcim, or Cadence Environmental Energy, the fuel supplier for Ash Grove. Incentives and trading platforms can play a crucial role in accelerating the adoption of alternative fuels. For example, the European Union’s Emissions Trading System (EU ETS) is based on a “cap and trade,” incentivizing a reduction in emissions via carbon market approach. In addition, regulations on waste management may be helpful. Some processed fuels may be regulated as nonhazardous secondary material, but oftentimes, they are regulated as waste–requiring companies to comply with waste-to-energy or incinerator regulations and permits. Financial support to overcome initial investment hurdles, tax incentives, and research on fuel sources and combustion technologies can also accelerate adoption. Ultimately, developing supportive and streamlined standards and comprehensive strategies may encourage and enable use of a broader set of resources for industrial fuels in the cement production sector and beyond. Supplementary Material