electronic supplementary material

LCA of Chemicals

A tiered approach to estimate inventory data and impacts of chemical products and mixtures

Gregor Wernet1*, Stefanie Hellweg2, Konrad Hungerbühler3

Received: 28 February 2011 / Accepted: 27 February 2012

© Springer-Verlag 2012

Responsible editor: Guido Sonnemann

G. Wernet ()

The ecoinvent centre, Empa, Ueberlandstrasse 129, 8600Dübendorf, Switzerland

e-mail:

S. Hellweg

ETH Zurich, Institute for Environmental Engineering, Wolfgang-Pauli-Strasse 15, 8093Zurich, Switzerland

K. Hungerbühler

ETH Zurich, Institute for Chemical and Bioengineering, Hönggerberg, HCI, 8093 Zurich, Switzerland

()Corresponding author:

e-mail:

1 Methodology for standard estimations in the process models

The process-based models use standard estimations for relevant flows in chemical productions if no better data are available. These estimations are based on published literature of energy uses in chemical productions and waste treatment. In addition, some models are applied for the treatment of waste. In some cases, these represent nothing more than best guesses. The estimations are described here (all values per functional unit of reference product):

  • Yield of chemical reactions is assumed to be 95% if no better data are available. The remaining 5% of all chemical inputs are considered to be wastes and are modeled for waste treatment.
  • For all volatile chemicals, emissions to air are estimated to be 0.2% of the total input.
  • All remaining chemical waste streams are treated in waste water treatment according to Koehler et al. 2007. This assumes that 10% are emitted to a river, while the remaining 90% are oxidized.CO2 emissions from this process are included in the inventory.
  • Use of heat (in the form of steam) is estimated to be 2 MJ per process step (based on Gendorf 2000).
  • Electricity use is estimated to be 0.333 kWh per process step (based on Gendorf 2000).
  • TOC, DOC, BOD5 and COD from aqueous emissions and the treatment of solid wastes are estimated using ecoinvent waste treatment models (available in the files section of the ecoinvent homepage).
  • Infrastructure input is estimated at 4∙10-10 units of a chemical plant.
  • Transport for all inputs is assumed to be 500km of freight train and 100 km of lorry >16t for a European dataset.

Using these estimations and models, it is possible to generate a functional inventory without any quantitative information about the production process. Of course, this will be a poor estimation and may not reflect reality well. Parts of this approach can be seamlessly integrated with actual process information.

For sources and further information see:

  • Hischier R, Hellweg S, Capello C, Primas A (2004) Establishing life cycle inventories of chemicals based on differing data availability. Int J Life Cycle Assess 10:59-67
  • Geisler G, Hofstetter TB, Hungerbühler K (2004) Production of fine and speciality chemicals: procedure for the estimation of LCIs. Int J Life Cycle Assess 9:101-113
  • Köhler A, Hellweg S, Recan E, Hungerbühler K (2007) Input-dependent life-cycle inventory model of industrial wastewater-treatment processes in the chemical sector. Environ Sci Technol 41:5515-5522
  • Gendorf (2000) Environmental declaration of the chemical plants at the Gendorf site in Burgkirchen, Germany. Available online at
  • Ecoinvent Report No. 8, Chemicals, available for free after registration at

page1of4

2 Detailed results for case study 2: tobacco flavor

The LCIA results for the process models and the Finechem tool that resulted in Figure 3 of the main article are reproduced in table SI1.

Table SI1 LCIA results of the process models and the Finechem tool for the tobacco flavor case study. In the contributions according to finechem, the chemicals for which the tiered approach recommends the application of process models are marked in yellow

page1of4