On this page
Changes to ecoinvent 3.9.1
Part of What was changed in each source. What is still open about ecoinvent is on the issue tracker under the label ecoinvent.
ecoinvent 3.9.1 is extracted and curated — 4,718 rows, corrected by
ecoinvent-3.9.1-manual-fixes.json and the ecoinvent-match-overrides.json
shared by every ecoinvent release — and it can be merged, but it is not part
of the standard build: the 2026-08-29 build of 54b0f9d merged
ecoinvent 3.12 and ecoinvent 3.8
only. An ecoinvent uuid is the same in every release, so what those two pages
say about a row is what the merge would do with 3.9.1's copy of it.
3.9.1 sits between the two merged releases in one way that matters here: it
is the release in which ecoinvent renamed most of its dissolved metals from
the bare element (Aluminium, Copper) to a named ion (Aluminium III,
Copper ion), and a handful of elements were still waiting for that rename.
One name, two substances
Which rows of a name land on which substance is measured on a build, and this release is not in one. For the rows it shares with the merged releases, read the same section on the 3.8 and 3.12 pages.
An ionic charge was added or corrected
Whether a row's published name carries a charge its source name did not is
measured on a build, and this release is not in one. What can be said is how
the release names dissolved metals, and where 3.12 has since changed its
mind about a row — because the rule for a release with no correspondence
table is that ecoinvent's newest name for a uuid is the one that counts, and
that rule is written down in ecoinvent-canonical-identities.json.
This release names 24 metal ions with a Roman numeral — Aluminium III,
Chromium VI, Zinc II — and six with the word ion and no charge —
Antimony ion, Arsenic ion, Copper ion, Iron ion, Tin ion,
Titanium ion. Publishing Chromium VI as Chromium(6+) is notation only
and is not a change. Every bare element name in the release carries the
element's own registry number; none carries an ion's.
Four elements were renamed by ecoinvent after 3.9.1, and 3.12's name is the one the merge would use for these rows:
| 3.9.1 name and number | Rows | 3.12 name and number for the same uuids | What the merge would publish |
|---|---|---|---|
Sodium, 7440-23-5 (the metal) |
8 emission rows to air and soil (the stratosphere row and the two water rows keep the metal's name in 3.12) | Sodium I, 17341-25-2 |
Sodium(1+) |
Calcium, 7440-70-2 (the metal) |
8 emission rows to air and soil (the stratosphere row keeps the metal's name in 3.12) | Calcium II, 14127-61-8 |
Calcium(2+) |
Caesium, 7440-46-2 |
5 emission rows to water | Caesium I, same number |
Caesium, ion — the CAS number is the element's, so the charge is left unstated |
Strontium, 7440-24-6 |
13 emission rows to air, soil and water | Strontium II, same number |
Strontium, ion, for the same reason |
The resource rows (natural resource / in ground) of these four keep the
element's name and number in every release and are untouched. Two further
rows go the other way: five Manganese-55 rows to water (7439-96-5, in
kBq) are Manganese in 3.12 — manganese-55 is the one stable isotope, so
the nuclide name was saying nothing — and thirteen of the fourteen Molybdenum VI
rows carry 16065-87-5 here and 7439-98-7, the element's number, from 3.10.1
on (the fourteenth, to the stratosphere, is dropped in 3.10.1).
Rows merged into one flow
Which rows of this release land together on one consensus flow is measured
on a build, and this release is not in one. The
3.8 and
3.12 pages list the merges
for the rows those releases share with this one — the ore-grade resource
rows (Copper, 0.52% in sulfide, …), the [Deleted] tombstones and the
alternative spellings are the same uuids here.
Registry numbers corrected
Each of these is written down in ecoinvent-3.9.1-manual-fixes.json with
the evidence in full; the reasons below are condensed.
| ecoinvent name | Rows in 3.9.1 | Number shipped | What it designates | Corrected to | What that designates | Why |
|---|---|---|---|---|---|---|
Uranium-238 |
10 (543421e8-f0b3-45a2-b0ed-03489e878138 and nine others, in kBq) |
7440-61-1 | Uranium, the element, by mass | 24678-82-8 | The nuclide U-238 | The rows measure the activity of one nuclide; every list that ships the row (3.8 through 3.12, EF 3.1) carries the element's number and is corrected the same way, so the lists agree about what the nuclide's number is (#200). |
Water, salt, sole |
1 (79238018-8ec1-4615-9469-2b0df95a43c3, natural resource / in water, m³) |
7732-18-5 | The water molecule | removed | — | Sole is German for brine, a saturated salt solution; a mixture does not carry the registry number of one of its constituents. EF 3.1's copy of the flow is corrected the same way. The synonym oxidane, the IUPAC name of water, is removed for the same reason. |
Nitrogen, organic bound |
5 (water: ground-, ground- long-term, ocean, surface water, unspecified) | 7727-37-9 | Dinitrogen, N₂, the gas | removed | — | Organically bound nitrogen is a measurement over many compounds rather than one compound, and with the gas's number left in place the rows fuse with the gas (#355). The synonyms dinitrogen and molecular nitrogen, which 3.9.1 copies onto these rows from its Nitrogen rows, are removed for the same reason. |
Gas, mine, off-gas, process, coal mining |
1 (3ed5f377-344f-423a-b5ec-9a9a1162b944, natural resource / in ground, sm³) |
8006-14-2 | Natural gas | removed | — | Coal-seam methane released during mining is a different resource from natural gas, which ecoinvent ships beside it as Gas, natural (7c337428-fb1b-45c7-bbb2-2ee4d29e17ba) with the same number; ecoinvent itself took the CAS number off this uuid in 3.10.1. With it in place the merge publishes the two as one substance (#412). |
Zirconia, as baddeleyite |
1 (e07b4402-abe3-4346-8c42-051c5983bd1e, natural resource / in ground, kg) |
1314-23-4 | Zirconium dioxide, the compound | 12036-23-6 | Baddeleyite, the mineral form of zirconium dioxide | The row is the mined mineral, its name says so, and ecoinvent registers the same uuid 12036-23-6 from 3.10.1 on; both CAS numbers are ZrO₂, so this is one substance under two registrations (#596). |
Stibnite |
1 (3e0034cd-21d6-4582-9fbf-09c26edd05df, natural resource / in ground, kg) |
1345-04-6 | Antimony trisulfide, the manufactured chemical | 1317-86-8 | Stibnite, the mineral | The row is ore taken out of the ground; IARC separates the two registrations as compound and mineral, and BAFU independently registers its stibnite 1317-86-8. EF 3.1's thirteen antimony trisulfide rows are emissions and keep the compound's number (#604). |
Granite |
1 (a4375a18-172c-4f82-90b7-bca972f75548, natural resource / in ground, kg) |
219714-96-2 | Penoxsulam, a rice herbicide sold under the trade name Granite | removed | — | The row is the rock; the herbicide's number reached it through the trade name, and ecoinvent removed it in 3.11. Left in place, the merged row was relabelled Penoxsulam (#670). |
Fenpropimorph |
1 (c8896e68-6c78-43be-a31b-bc5d2307f40a, soil / agricultural, kg) |
67306-03-0 | The morpholine skeleton with stereochemistry unstated — EF 3.1's uncharacterised duplicate of the fungicide | 67564-91-4 | The racemic fungicide, the entry the JRC characterises | 3.10.1 onward register the same uuid 67564-91-4; without the correction one substance is split by release (#671). |
Borax |
1 (eead2933-c2be-4a53-a0bd-bd33b67e4145, natural resource / in ground, kg) |
1330-43-4 | Anhydrous sodium tetraborate, the manufactured salt | 1303-96-4 | The decahydrate — mineral borax | A resource row extracts the mineral; 3.10.1 onward register the same uuid 1303-96-4. |
Diatomite |
1 (9877ce00-65f8-4c0c-9fcf-92aa53a2c9c0, natural resource / in ground, kg) |
7631-86-9 | Silicon dioxide, the pure substance | 61790-53-2 | Diatomite, the sedimentary rock | 3.10.1 onward register the same uuid 61790-53-2; without the correction the row lands on Silicon Dioxide while 3.12's identical row stays on Diatomite. |
Fluorspar |
1 (0fa4f51e-b0dc-5d11-84d3-b32f0f3c88d5, natural resource / in ground, kg) |
7789-75-5 | Calcium difluoride, the pure substance | 14542-23-5 | Fluorite, the mined mineral | 3.12 registers the same uuid 14542-23-5, the CAS number 3.8's own Fluorspar, 92%, in ground sibling already carries. |
Beta-cyfluthrin |
5 (soil, air and water rows, kg) | 68359-37-5 | Cyfluthrin, the commercial eight-isomer mixture | 1820573-27-0 | Beta-cyfluthrin, the product enriched in the two most insecticidally active diastereoisomeric pairs | PPDB registers the two products separately (records 74 and 192). ecoinvent's own transition table retired the name Cyfluthrin for these uuids in favour of Beta-cyfluthrin, while EF 3.1 ships plain cyfluthrin under the shared number, so the beta rows fused with the plain substance; the shared EC number 269-855-7 stays. Every release ships the rows and takes the same correction (#854). |
In each of the last four the newer release's registration of the same uuid is taken as evidence of what the older one meant, the reading every cross-release correction in these files takes.
Other corrections
These change a name, a synonym or a unit rather than a CAS number.
| ecoinvent name | Rows in 3.9.1 | Change | Why |
|---|---|---|---|
Silver-110 |
10, in kBq | Renamed Silver-110m; Silver-110 kept as a searchable alternative name |
Silver-110 is a 24.56-second ground state; the species a reactor-effluent inventory reports is the isomer Ag-110m, at 249.863 days, which is what EF 3.1's characterisation factors describe (#219). The number 14391-76-5 stays: it indexes the mass number and there is no nuclide-specific number for the isomer. |
Ioxynil methyl ester, 2436-73-9 |
1 (soil / agricultural) |
Renamed MCPA-methyl, and the old name is not kept |
2436-73-9 contains no iodine: it is methyl 2-(4-chloro-2-methylphenoxy)acetate, the methyl ester of MCPA. ecoinvent's three other ioxynil flows carry correct numbers, so the name slipped and the CAS number is kept (#326). |
Zirconia, as baddeleyite |
1 | Renamed Baddeleyite |
Zirconia is zirconium dioxide, and the row is the mineral being mined; the mineral's own name is shorter and more accurate (#596). |
Peat |
1 (c5035ce2-5ee5-431f-a287-4b25da42be74, natural resource / biotic, kg) |
Renamed Peat, horticulture; Peat kept as an alternative name |
EF 3.1 has two peats: fuel peat in megajoules, and this one, dug for growing media and counted by mass. Under the bare name the kilogram row reached the megajoule flow and its own factor stopped being published (#450). |
Gypsum, 13397-24-5 |
1 (11a2a7b1-ab2f-47b8-9e29-6f33d5207fa6) |
Synonyms sulfate and dihydrate removed |
They are the words of the real synonym Calcium sulfate dihydrate split apart, and sulfate names a substance the list publishes, which gypsum reached by it (#597). |
Occupation, arable, conservation tillage (obsolete), … conventional tillage (obsolete), … reduced tillage (obsolete) |
3 (fdb1b2d0-f537-401e-b845-1d93da512174, e489cce4-a80f-417d-9ae6-9fc14cc7dd49, 81e07a67-28e0-4392-a553-d86e54a9b8a9) |
Unit m² → m²·a, factor 1.0 | Every other land-occupation row in ecoinvent and in EF 3.1 is an area held for a time; ecoinvent's own correspondence tables map these three at 1.0 with the comment assumed conversion based on land use through an entire year (#565). |