This article explains how to find the right dataset for a chemical compound.
Sometimes chemical compounds cannot be found in databases, especially when they are more complex or less commonly used chemicals. How can you choose a fitting dataset in this case?
Step 1. Check whether the use of the chemical contributes significantly to your composition or potential impacts (top 5 or top 10 chemicals).
If the chemical at hand contributes a tiny amount to the total impact of the product, a less accurate approximation is practical. It will take less time to model a less accurate approximation, without having a significant effect on the end result. By contrast, if the chemical is one of the key components of the product, then a more precise approach is necessary.
Step 2. Check alternative chemical names.
Sometimes chemicals are present in ecoinvent with a different name, for example:
- 2-Aminoethanol = monoethanolamine
- Phenylethane = ethylbenzene
- Esterquats = Quaternary ammonium compounds
- Etc...
A useful website we often use at Ecochain for finding chemical synonyms is PubChem.
Step 3. Check whether the chemical belongs to a more generic category
Check whether the chemical belongs to a more generic category and can be modelled as such:
- Non-ionic surfactant
- Ethoxylated alcohol
- Organic solvent
- Foaming agent
- Etc…
Step 4. Use a generic dataset (chemical organic / chemical inorganic)
Use generic datasets for organic (carbon containing) or inorganic chemicals such as “market for chemical, organic | chemical, organic | Global” or “market for chemicals, inorganic | chemical, inorganic | Global” dataset from ecoinvent.
Alternate option: Stoichiometric modelling
Model the production process of the chemical yourself. This means taking the components required to produce that chemical in the necessary ratio, and adding a production process (incl. % yield) or energy use to it.
See an example of producing pyrogenic silica (SiO2) from Silicium tetrachloride (SiCl4)
Step 1: Atom weights of common elements
Chemical element |
atom weight |
Unit |
C |
12,011 |
g·mol−1 |
H |
1,008 |
g·mol−1 |
O |
15,999 |
g·mol−1 |
Cl |
35,453 |
g·mol−1 |
Si |
28,086 |
g·mol−1 |
Step 2: Determine atom weights (molar masses) of chemical reaction
Chemical element |
atom weight |
Unit |
Input |
||
SiCl4 | 169,898 | g·mol−1 |
2H2 | 4,032 | g·mol−1 |
O2 | 31,999 | g·mol−1 |
Sub-total | 205,928 | g·mol−1 |
Output | ||
SiO2 | 60,084 | g·mol−1 |
4HCl | 145,844 | g·mol−1 |
Sub-total | 205,928 | g·mol−1 |
Step 3: Relate atom weights (molar masses) of chemical reaction to the reference chemical
Chemical element |
atom weight |
Unit |
Calculation | Kg chemical /kg pyrogenic silica (SiO2) |
Input | ||||
SiCl4 | 169,898 | g·mol−1 | 169,898/60,084 | 2,828 |
2H2 | 4,032 | g·mol−1 | 4,032/60,084 | 0,067 |
O2 | 31,999 | g·mol−1 | 31,999/60,084 | 0,533 |
Output | ||||
SiO2 | 60,084 | g·mol−1 | 60,084/60,084 | 1,000 |
4HCl | 145,844 | g·mol−1 | 145,844/60,084 | 2,427 |
Step 4: Correction for yield
Assumed yield = 90%
Chemical element |
atom weight |
Unit |
Calculation | Kg chemical /kg pyrogenic silica (SiO2) |
Input | ||||
SiCl4 | 169,898 | g·mol−1 | 2,828/0,9 | 3,142 |
2H2 | 4,032 | g·mol−1 | 0,067/0,9 | 0,075 |
O2 | 31,999 | g·mol−1 | 0,533/0,9 | 0,592 |
Output | ||||
SiO2 | 60,084 | g·mol−1 | - | 1,000 |
4HCl | 145,844 | g·mol−1 | - | 2,427 |
Step 5: find correct ecoinvent datasets
- SiCl4 = silicon tetrachloride production | silicon tetrachloride | Global
- 2H2 = chichibabin amination | hydrogen, liquid | Europe
- O2 = -
- 4HCl = market for hydrochloric acid, without water, in 30% solution state | hydrochloric acid, without water, in 30% solution state | Europe
Step 6: Add chemical reaction to Mobius and put ' minus (-) in front of property for 4HCl
This should be done because the system is multifunctional. No allocation should be applied. Instead, the impact of 4HCl should be corrected. This can easily be done in Mobius by simply putting a minus (-) in front of the property. This gives -1 kg mass.