Petrochemical products are essential parts of modern society, found in everyday items (such as clothing and packaging), fuels and fertilizers. The petrochemical industry consumes a large portion of global oil and gas and emits considerable greenhouse gas (GHG) emissions. The global chemical sector emits approximately 1.5 gigatons (Gt) of CO2 emissions annually, 60% of which are attributed to primary petrochemicals such as ammonia, methanol and olefins1. Thus, it is urgent to decarbonize the global petrochemical industry.

Accurate GHG emissions assessment is crucial for mitigating climate change. Current methods rely on data on energy and material flows and emission intensity factors (that is, GHG emissions per unit mass of product) obtained from life-cycle assessment (LCA) databases, literature, process simulations or proxy data. LCA is a method to evaluate the environmental impact across a product’s life cycle, for example, a cradle-to-grave LCA includes raw material acquisition, manufacturing, use phase and end of life. Previous country-specific LCAs have reported uncertainties in petrochemical carbon footprints2,3. However, a comprehensive understanding of uncertainty sources and reduction strategies across the global petrochemical industry is lacking. This gap hinders accurate LCAs of common petrochemical products such as plastics and challenges climate policy and decarbonization efforts.

Now, writing in Nature Chemical Engineering, Fanran Meng, Jonathan Cullen and co-workers address this critical knowledge gap by examining 81 chemicals across 37,000 facilities globally4. Their cradle-to-gate analysis examines GHG emissions along petrochemical supply chains from raw material acquisition to chemical production, including upstream and downstream chemicals (Fig. 1). They employed an automated algorithm to match each facility to possible production technologies for estimating facility-specific GHG emissions factors. The study identified six uncertain sources (Fig. 1) and quantified their impacts. The global annual cradle-to-gate GHG emissions from petrochemicals were estimated to be 1.9 ± 0.6 Gt of CO2 equivalent (CO2e), with life-cycle GHG estimates of various chemicals showing 15–40% uncertainties. The inability to pinpoint specific production processes is the largest uncertainty source. Feedstock and indirect energy use are another two main contributors, each accounting for about half of the remaining uncertainty. This study also explored uncertainty mitigation strategies and evaluated their potential benefits.

Fig. 1: Uncertainty sources and petrochemical groups.
figure 1

There are six uncertainty sources, including two model uncertainties and four data uncertainties. Two model uncertainties are caused by the lack of process-specific information and the selection of allocation methods. Data uncertainty is classified based on four types of GHG emission sources along petrochemical supply chains, including emissions associated with upstream feedstock, on-site fuel consumption (direct energy use), off-site energy generation (indirect energy use) and chemical reactions (direct processes). Petrochemicals are categorized into upstream and downstream product groups. Upstream chemicals are input materials of downstream chemicals and some upstream chemicals (for example, some intermediate chemicals).

Full size image

The study first explored the effects of different allocation methods, a crucial step in LCAs of multi-output systems (for example, petrochemical plants) to partition burdens among products. Mass, energy and economic allocations are common methods that split environmental burdens among products based on their mass, energy content or economic values, respectively. The authors applied three allocation methods and found little differences in the GHG estimates of most chemicals, except for some chemicals (for example, hydrogen and methanol) with large differences between mass and economic allocation methods. For these chemicals and other systems where the economic values of co-products differ substantially, uncertainties caused by allocation method selection need to be considered.

The inability to determine process specificity was assessed by comparing cradle-to-gate GHG emissions from various processes of producing the same chemical. Limited knowledge of the exact process used in a chemical plant often leads to the use of generic emission factors, which can result in a large deviation from actual emissions given the wide range of possible emission factors for chemicals with many alternative processing technologies. For instance, due to greater diversity in processing technologies, primary chemicals show broader emission factor ranges than downstream chemicals such as thermoplastics. The study specifically highlighted ethylene and methanol, two commodity chemicals with large production volumes and critical roles as building blocks for the chemical industry1. Their GHG emissions intensities can vary by more than tenfold, influenced by feedstock (for example, coal, naphtha and natural gas) and processing technologies, with feedstock having a more pronounced effect.

The impact of process-specificity uncertainty is larger than the overall impact of the four data uncertainty sources shown in Fig. 1. However, this uncertainty is eliminated when specific processes are known. Data uncertainties were explicitly evaluated across four main sources: feedstock supply chains, indirect and direct energy use, and direct processes. Each uncertainty source and its contribution to the overall uncertainty of the cradle-to-gate GHG emissions of a process was quantified. On average, data uncertainties associated with feedstock and indirect energy use have much larger impacts than those related to direct energy use and processes.

In assessing the relevance of these uncertainties for global decarbonization, the authors propagated uncertainties from the facility level to the global level, showing a substantial uncertainty of 0.6 GtCO2e out of 1.9 ± 0.6 GtCO2e for the total sector. Primary chemicals are major uncertainty sources (459 MtCO2e), affecting downstream chemicals such as polyethylene — one of the most widely used plastics. Among downstream chemical groups, thermoplastics show the largest uncertainty with 238 MtCO2e. Thermoplastics are widely used for everyday products. Thus, GHG estimates of thermoplastics have determinantal impacts on the accuracy of LCAs for consumer goods.

The authors further explored the benefits and priorities of reducing uncertainty. Eighty per cent of uncertainty can be reduced by knowing the specific process technologies of only 20% of chemical facilities. Such knowledge may be challenging to acquire due to confidentiality. Without detailed process information, up to 61% uncertainty could still be reduced by collecting more precise data on feedstock and indirect energy input for 25% of chemical plants globally.

Robust GHG accounting is the foundation of developing effective mitigation strategies. This is the first, comprehensive analysis of global petrochemical plants and uncertainties in their life-cycle GHG emissions. This study identifies two priority areas to improve data transparency and precision; one is to tackle major uncertainty sources (process specificity, feedstock and indirect energy use), the other is to focus on primary chemicals and facilities with large uncertainty (for example, olefins, methanol and ammonia). Addressing these uncertainties needs precise, facility-specific data that are often restricted by confidentiality. Novel approaches (for example, blockchain) and efforts in chemical industry digitalization may address this issue by providing secure data sharing5,6,7. However, real-world exploration and applications of these approaches in the chemical sector will be needed. More research is also needed to investigate the uncertainty of the use phase and end of life of petrochemicals, which are not included in this study but are essential parts of the full life cycles of chemicals8. This study demonstrates the powerful use of uncertainty analysis for the global petrochemical industry. Future research may consider similar approaches for other sectors to drive more accurate GHG emissions tracking and reporting across industries.