No abstract

Background, aim, and scope Pharmaceuticals have been recently discussed in the press and literature regarding their occurrence in rivers and lakes, mostly due to emissions after use. The production of active pharmaceutical ingredients (APIs) has been less analyzed for environmental impacts. In this work, a life cycle assessment (LCA) of the production of an API from cradle to factory gate was carried out. The main sources of environmental impacts were identified. The resulting environmental profile was compared to a second pharmaceutical production and to the production of basic chemicals. Materials and methods Detailed production data of a pharmaceutical production in Basel, Switzerland were used as the basis of this work. Information about the production of precursor chemicals was available as well. Using models and the ecoinvent database to cover remaining data gaps, a full life cycle inventory of the whole production was created.Using several life cycle impact assessment methods, including Cumulative Energy Demand (CED), Global Warming Potential (GWP), Eco-Indicator 99 (EI99), Ecological Scarcity 2006, and TRACI, these results were analyzed and the main sources of environmental burdens identified. Results Pharmaceutical production was found to have significantly more environmental impacts than basic chemical production in a kilogram-per-kilogram basis. Compared to average basic chemical production, the API analyzed had a CED 20 times higher, a GWP 25 times higher and an EI99 (H/A) 17 times higher. This was expected to a degree, as basic chemicals are much less complex molecules and require significantly fewer chemical transformations and purifications than pharmaceutical compounds. Between 65% and 85% of impacts were found to be caused by energy production and use. The fraction of energy-related impacts increased throughout the production process. Feedstock use was another major contributor, while process emissions not caused by energy production were only minor contributors to the environmental impacts. Discussion The results showed that production of APIs has much higher impacts than basic chemical production. This was to be expected given the increased complexity of pharmaceutical compounds as compared with basic chemicals, the smaller production volumes, and the fact that API production lines are often newer and less optimized than the production of more established basic chemicals. The large contributions of energy-related processes highlight the need for a detailed assessment of energy use in pharmaceutical production. The analysis of the energy-related contributions to the overall impacts on a process step level allows a comprehensive understanding of each process' contribution to overall impacts and their energy intensities. Responsible editor: Roland HischierElectronic supplementary material The online version of this article (

The chemical industry is increasing its efforts to reduce the environmental burdens of chemical production. One focus is to implement energy-efficient processes and green technologies early in the process design to maximize environmental efficiency and to reduce costs. However, as data on many chemical products are scarce, many sustainability studies are hampered by the lack of information on production processes, and chemicals are often neglected or only crudely estimated. Models that estimate production data and environmental burdens can be vital tools to aid sustainability efforts. In addition, they are useful for the environmental assessment of chemicals without access to production data, i.e. in supply-chain management or for the assessment of products using chemicals as materials. Using mass and energy flow data on the petrochemical production of 338 chemicals, we developed models that can estimate key production parameters directly from the molecular structure. The data sources were mostly production data provided by industrial partners, extended by data from the ecoinvent database. The predicted parameters were the Cumulative Energy Demand (CED), the Global Warming Potential (GWP), the Eco-indicator 99 score, a Life Cycle Assessment (LCA) method, and the electricity and heat use over the production cycle. Model outputs include a measure of the prediction uncertainty. The median relative errors of the models were between 10% and 30%, within acceptable ranges for estimations. The modelled parameters offer a thorough insight into the environmental performance of a production process and the model estimates can be of great service in process design, supply-chain management and environmental assessments of chemical products in the early planning and design stages where production data are not available.