ecoinvent Version 3.8

Focusing on user comfort, ecoinvent v3.8 includes enhanced documentation and expands the sectorial coverage of the ecoinvent database with around 360 new and 700 updated datasets.

Oct 28th, 2023

It further encompasses the first results of our collaboration with Rede ACV on the database’s Brazilian data. 

With version 3.8 of the ecoinvent database users of ecoinvent data can benefit from enhanced documentation and information on products and markets, data linked in the additional system model “Allocation, cut-off, EN15804ˮ, which supports specifically EPD practitioners, a calculation support to work with the GHG protocol as well as around 360 new and 700 updated datasets related to agriculture, bamboo, batteries, chemicals and plastics, electronics, metals, forestry and wood, transport, and electricity. This version also releases the first results of the collaboration between Rede ACV and ecoinvent aiming at enhancing the coverage and accuracy of Brazilian data in the database. 

The data used for the additions and updates were provided by various organisations including ETH, EMPA, Pforzheim University and ACV Brazil. 

Features and Documentation 

Product Information 

The update aims at offering comfort to the users while engaging with the database. All products and services included in the database now integrate a description that supports users in identifying them. The information includes a definition of the product or service as seen in the database. Regarding services, their documentation now incorporates supporting information on how to use them. Overall, ecoinvent v3.8 contains 3’300 unique products and services. 

Market Information 

With the release of ecoinvent v3.8, ecoinvent aims at improving the description and information in the market datasets included in the database. Specifically, all market datasets have been reviewed, and their existing meta information has been complemented, corrected or completed, in order to make the use and usability of these datasets more coherent and precise for the users of the ecoinvent database.

Database Overview 

To provide a more concise overview of the contents of the ecoinvent database, the information about the activities included has been centralized into one file; “Database Overview for ecoinvent 3.8”. The file contains the following information: 

  • The lists of datasets contained in each of the system models, as well as the list of unlinked and unallocated datasets. For version 3.8, each dataset has been assigned to one or more sectors, which helps filter for datasets that can belong to different sectors. 
  • The geographies used in the ecoinvent database. Each geography is assigned a classification, and the geographies that are contained in each geography are indicated. 
  • The list of the LCIA methods for which ecoinvent calculates impact scores. The version of the methods is indicated, and the original source used for defining the characterization factors is provided. 

Enhanced Support for GHG Protocol 

With the updates of ecoinvent v3.8, the ecoinvent database now offers the possibility to calculate emissions from electricity according to Scope 2 and Scope 3 Guidance of the GHG Protocol. For that, an additional file can be found in the ‘Files’ section ecoQuery. Separated scope 2 and 3 emission factors for electricity producing activities and electricity markets are now available in a spreadsheet. Emission factors are calculated based on the GWP100 (global warming potential) values from the IPCC 2013 AR5 report. They are representative for the end consumers of electricity and refer to the location-based reporting method, i.e., they represent grid average emission factors. On top of this, the new spreadsheet also provides disaggregated emission and emission factors for individual greenhouse gases or groups of greenhouse gases. 

New System Model for EPD Practitioners 

Acknowledging the increasing importance of Environmental Product Declarations (EPDs) in the construction and manufacturing sector and the need for standardized, accurate, non-misleading communication of the environmental information for products, the ecoinvent association introduced the ‘allocation, cut-off, EN15804’ system model. The aim of this system model is twofold; on the one hand, the ‘allocation, cut-off, EN15804’ facilitates the EPD practitioners to comply with the standard EN15804&A2:2019. On the other hand, the new system model contributes to a harmonization in the calculation of the indicators of the standard. The system model is fully compliant with ISO 14025, ISO 21930 and EN15804&A2:2019. It provides all Life Cycle Inventory (LCI) indicators required by the aforementioned standards and adheres to the end-of-waste criteria set by the European Commission. 

LCIA Methods 

With ecoinvent v3.8, the latest versions of CML (v4.8 2016) and EF (v3.0, and adapted for EN15804) were added to LCIA methods. More details about impact categories of all implemented methods can be found in the new database overview file. 

New and Updated Data

Brazilian Data 

The Cornerstone project was set up between Rede ACV (the Brazilian business network for LCA) and the ecoinvent association to improve the availability and access to background data representing key supply chains for Brazil. Rede ACV, through its contractor ACV Brasil, departed from existing content in the ecoinvent database to develop regionalized life cycle inventory (LCI) data representative of the situation in Brazil. The data in ecoinvent 3.8 cover the extraction, import and distribution of crude petroleum oil and natural gas, production of biodiesel (esterification of fatty acid methyl ester, FAME), distribution of diesel (for transport), and road freight transport by lorry. Older ecoinvent datasets have been updated for v3.8 to better reflect the country-specific supply of crude petroleum oil to petroleum refinery operation in Brazil. 


The sector features a small update with the representation of new products, such as lentils and peas, and the recontextualization of agricultural services for Canada. Data for these updates were provided by the University of British Columbia and the Food Systems PRISM Laboratory. 


ecoinvent v3.8 includes new datasets for three types of Li-Ion batteries (NCA, NMC111, NMC811) and their production. These datasets were created by the Swiss Federal Laboratories for Materials Science and Technology (EMPA) and incorporate various chemicals, such as cobalt sulphate, disodium disulphite, vinyl carbonate and oxides, and hydroxides for the three battery types. The individual battery components (cathode, anode, high and low voltage systems, cells and battery management systems) are split, and each is represented in an individual dataset. Complementary, a new dataset on lead-acid batteries created by the University of Pforzheim is also included in the v3.8 update. 

Chemicals and Plastics 

In ecoinvent 3.8, we continued our efforts to update data and strengthen the collaboration with the industry. Data for styrene production was updated with the support of PlasticsEurope and new and updated data for polyurethane foam was provided by EUROPUR. 


ecoinvent 3.8 contains updated, attributional electricity markets for most countries with a reference year of 2018. In the case of US and Canada, the reference year is 2019. Additionally, the Swiss electricity markets as well as transformation losses (high to medium and medium to low voltage) for all updated markets were adjusted. Data creation and review was done in collaboration with ETH Zurich and the Paul Scherrer Institute (PSI). 


Version 3.8 includes new datasets on consumer electronic devices. Based on the latest literature data, datasets for smartphone and tablet were added and the existing dataset for laptop was updated. For the mobile devices datasets for a new printed wiring board and liquid crystal display were created. Components in the corresponding supply chain have been adjusted as needed. 

Forestry and Wood 

The Wood sector has been updated with new data for bamboo forestry and the production of (intermediate) bamboo construction materials. i.e., bamboo pole, flattened bamboo, and woven bamboo mat. These new datasets were developed by Dr. Edwin Zea Escamilla with the support of the International Organization for Bamboo and Rattan INBAR. 


New and updated datasets for the metals sector have been provided by EMPA (sponsored by the Swiss Federal Office for the Environment, BAFU) for ecoinvent version 3.8. The new data cover the production beryllium, ferroniobium and scandium oxide produced from rare earth tailings. Updates were provided by EMPA for the production of ferrochromium, lithium carbonate, strontium sulfate, tantalum, titanium, as well as for the mining of heavy mineral sands. 


In collaboration with INFRAS, the datasets for air freight (belly and dedicated) transport are updated. The consumption of kerosene and the corresponding emissions were too low (version 3.6 to 3.7.1). 

Resource Correction for Biogenic Carbon 

In the “allocation, cut-off by classification” system model, the data for the forestry, wood, paper and cardboard sectors, biogenic carbon balances are restored after allocation using a new elementary exchange that allows to follow the process transparently. 

Relevant Documents and Files

Report of Changes

This report covers all changes made in the ecoinvent database between version 3.7.1 (2020) and version 3.8 (2021). It reflects both the database-wide changes, as well as the sector-specific changes. 

Report of Changes

This automatically generated excel file provides an overview of changes made on the undefined unit process datasets between versions 3.7.1 and 3.8 of the ecoinvent database. The Report of Changes XLS can be found in the «Files» section of ecoQuery, the online version of the ecoinvent database version 3. It is available to licensees of the ecoinvent database. 

Correspondence File 

The correspondence file is a spreadsheet that lists all datasets in versions 3.7.1 and 3.8 and matches corresponding datasets between versions. 

Database Overview File 

The Database Overview file describes the contents of the database. The following information is contained in the file: 

Activity Overview(s) 

The lists of datasets contained in each of the three system models, as well as the list of unlinked and unallocated datasets. For version 3.8, each dataset has been assigned to one or more sectors, which helps filter for datasets that can belong to different sectors. 

ecoinvent Geographies 

The geographies used in the ecoinvent database. Each geography is assigned a classification, and the geographies that are contained in each geography are indicated. 

LCIA Methods 

The list of the LCIA methods for which ecoinvent calculates impact scores. The version of the methods is indicated, and the original source used for defining the characterization factors is provided. 

Known Issues

This section lists all known data errors in version 3.8 of the ecoinvent database. 

Data Issues 

  • The activities “electricity voltage transformation from high to medium voltage” and “electricity voltage transformation from medium to low voltage” for the RoW geography appear in the Allocation, cut-off and APOS system model. As the RoW markets for electricity do not exist for these system models, the datasets have LCIA scores of zero and should not be used. 
  • The exchange amount of the input from the environment “Tellurium, in ground” (compartment: “natural resource”; subcompartment: “in ground”) in the activity “copper mine operation and beneficiation, sulfide ore” for the geographies AU, CA, CL, CN, ID, KZ, RU, US, and ZM was found to be inconsistent with the amount in the GLO geography. The exchange amounts in the local geographies have been scaled down based on the assumption that these geographies extract 70% of the global Te that is extracted in copper ores (modelled in the GLO dataset).The new amounts for the undefined datasets are listed in the table below. They will be available with the release of v3.9. 



Previous Value 

New Value 

copper mine operation and beneficiation, sulfide ore, AU 




copper mine operation and beneficiation, sulfide ore, CA 




copper mine operation and beneficiation, sulfide ore, CL 




copper mine operation and beneficiation, sulfide ore, CN 




copper mine operation and beneficiation, sulfide ore, ID 




copper mine operation and beneficiation, sulfide ore, KZ 




copper mine operation and beneficiation, sulfide ore, RU 




copper mine operation and beneficiation, sulfide ore, US 




copper mine operation and beneficiation, sulfide ore, ZM 




  • The dataset styrene production was updated with support from PlasticsEurope for version 3.8. A company survey of three producers and four sites was conducted, as a consequence the benzene emissions to water were removed from the dataset and the benzene emissions to air reduced by four orders of magnitude. This update had a significant impact on selected human toxicity indicators (e.g. ecological scarcity 2013, carcinogenic substances into air-total). The other key contribution to human toxicity impact categories, e.g. in ReCiPe Midpoint (H) – human toxicity, is the material input of benzene. The market for benzene in the ecoinvent database represents a mix of benzene produced through catalytic reforming and pyrolysis gasoline as well as a share produced as a by-product of coking. The latter technology has significant human toxicity impact scores. Market shares will be revised within upcoming releases. 
  • The dataset “adipic acid production” is under review regarding the dinitrogen monoxide (N2O) emissions generated during its production. New emission abatement methods are examined, which would improve the removal efficiency and therefore reduce N2O emissions in both the RER and GLO datasets.” 
  • “The datasets “cathode production, NCA, for Li-ion battery” and “cathode production, NMC811 for Li-ion battery” have an incorrect input of the intermediate exchange “NMC111 oxide”. For the NCA cathode production the exchange should be “NCA oxide” and for NMC811 cathode production “NMC811” respectively. This leads to an overestimation of impact scores for the cathode production datasets as well as the subsequent battery cell and battery production datasets for NCA and NMC811. The correct scores for the affected datasets were recalculated and are accessible in the “Files” section of ecoQuery in the file: Corrected LCIA scores – Oxide input in cathode production, li-ion batteries v3.8.xlsx” 
  • The dataset “methylene diphenyl diisocyanate production” (GLO and RER) treats the formed hydrochloric acid as a waste, this does not reflect reality. Hydrochloric acid is a valuable by-product which is used or sold by the MDI producer.The ecoinvent association advices users to either use the LCIA scores from the current ecoProfile available on the PlasticsEurope website or to correct the unit process in ecoinvent by removing the waste neutralisation and treating HCl as a by-product. 
    To correct the ecoinvent dataset, we would advise user to use the undefined dataset available through ecoQuery and remove the exchanges of the input of sodium hydroxide and the outputs of sodium and chloride. 
  • The file “electricity emission factors – scope 2 – 3 in ecoinvent v3.8.xlsx” published on the 21st of September 2021 has been replaced by an updated version in December 2022. The new file “electricity emission factors – scope 2 – 3 in ecoinvent v3.8 (updated).xlsx” contains several corrections which are described in detail in the “read me” tab of the updated file. The updated file can be found under the v3.8 supporting document in the Files section of ecoQuery. 

Life Cycle Impact Assessment Issues 

  • In the EF v3.0 methods (EF v3.0, EF v3.0 no LT, and EF v3.0 EN15804), the unit for the impact category “eutrophication: freshwater” should be “kg P-Eq” instead of “kg PO4-Eq”. 
    For some LCIA methods which are no longer updated but still published, names of elementary exchanges that were renamed were not updated in the files containing the characterization factors. This caused some of the characterization factors to be excluded from the implementation used for version 3.8. If wished by the user, this could be updated manually by changing from the old to the new name in the “LCIA implementation 3.8.xlsx”-file for the following exchanges: 

Old name 

New name 












Information about errors in the documentation reported in the PDF files for the ecoinvent three attributional system models 

  • We have identified an issue in some pdfs of datasets in the cut-off, EN15804 or APOS system models that are multioutput datasets in their undefined form. Specifically, allocation factors were wrongly reported in these pdfs. The error only concerns the allocation factors, scores and all exchanges were reported correctly. 
    Please be informed that this issue only concerns a few pdfs, and it does not affect the ecoinvent database.