Wednesday, 1 April, 2026
Fractionation Database
The fractionation database is the scientific repository that feeds the default values of Nexus cards. This page describes its structure, its current content and the rules governing its quality.
Role and Positioning
The fractionation database is distinct from IsoFind's analytical database (isofind.db). It is stored in a dedicated SQLite file (fractionation_database.db) and contains no field data: only fractionation coefficients drawn from published experimental, theoretical or field literature.
Its role is to provide Nexus cards with reference δ values, documented ranges and associated experimental conditions, so that every parameter used in a calculation is traceable back to its original publication. When a process card is added to the canvas, its default values come from this database, filtered by process type, element and material.
Table Structure
The database contains one main table, fractionation_data, organised into six groups of fields.
Element and isotopic system identification
| Field | Type | Description |
|---|---|---|
| element | TEXT (required) | Element symbol (e.g. Sb, Fe, Zn). |
| isotope_system | TEXT (required) | Notation of the analysed isotopic system (e.g. 123Sb/121Sb, 56Fe/54Fe). |
Geochemical process
| Field | Type | Description |
|---|---|---|
| process_type | TEXT (required) | Process category: adsorption, redox, precipitation, dissolution, biological, evaporation, diffusion, complexation. |
| process_category | TEXT | Kinetic nature of the process: equilibrium, kinetic or mixed. Determines which model (equilibrium Rayleigh or kinetic) applies by default in the Nexus. |
| phase_A | TEXT | Source phase of the isotopic transfer (e.g. aqueous, Sb(III), stibnite). |
| phase_B | TEXT | Product phase (e.g. adsorbed, Sb(V), aqueous). |
| material_detail | TEXT | Specific mineral or adsorbant material (e.g. Ferrihydrite, γ-Al2O3, Sb2S3). |
Fractionation value
| Field | Type | Description |
|---|---|---|
| delta_notation | TEXT | Fractionation notation as expressed in the source publication. |
| delta_value | REAL (required) | Fractionation value in per mil. Required, unless delta_min and delta_max are provided (CHECK constraint). |
| delta_min / delta_max | REAL | Range of values observed in the study (used for processes with documented variability). |
| uncertainty | REAL | Analytical uncertainty associated with delta_value, in per mil. |
Experimental conditions
| Field | Description |
|---|---|
| temperature / temperature_unit | Experiment temperature (default: °C). |
| ph | pH of the experimental medium. |
| ionic_strength | Ionic strength in mol/L. |
| time_scale | Duration of the experiment or note that equilibrium was reached (e.g. <48 h, equilibrium). |
| equilibrium_reached | Boolean indicating whether thermodynamic equilibrium was reached in the experiment. Determines the choice of calculation model in the Nexus. |
| oxidation_state_initial / oxidation_state_final | Oxidation states of the element at the start and end of the process. |
| oxidant_reductant | Chemical agent used to induce the redox change. |
| concentration / concentration_unit | Initial concentration of the element in the experiment. |
Biological parameters
| Field | Description |
|---|---|
| biological_agent | Bacterial strain or organism responsible for the fractionation (filled only for process_type = biological). |
| metabolic_pathway | Metabolic pathway involved (e.g. enzymatic Sb(III) → Sb(V) oxidation). |
Bibliographic traceability and quality
| Field | Description |
|---|---|
| study_type | Nature of the source study: experimental, theoretical or field. |
| reference | Full citation of the publication (authors, year). |
| doi | DOI identifier of the publication. |
| year | Publication year. |
| notes | Additional information not covered by other fields (context, adsorbant category, etc.). |
| quality_flag | Entry reliability indicator: high, medium or low. Based on the precision of the reported uncertainty, reproducibility and study type. |
| date_added / added_by / last_modified | Audit metadata: date added, identifier of the user who entered the record, date of last modification. |
Sample Database Content
The version distributed with IsoFind contains 28 entries covering antimony (Sb) on the ¹²³Sb/¹²¹Sb isotopic system, from publications between 2003 and 2024. This is the first documented element, consistent with the domain of expertise of IsoFind's developer.
Breakdown by process type
| Process type | Entries | Mean δ (‰) | Observed range (‰) |
|---|---|---|---|
| Adsorption | 19 | +0.61 | [0.00 ; 2.35] |
| Redox | 5 | +0.62 | [0.00 ; 1.45] |
| Biological | 2 | −0.40 | [−0.60 ; −0.20] |
| Dissolution | 2 | +0.50 | [0.50 ; 0.50] |
Documented adsorbant materials
The 19 adsorption entries cover four families of mineral phases, all relevant to the environmental geochemistry of Sb:
| Family | Minerals | Entries |
|---|---|---|
| Iron oxyhydroxides | Ferrihydrite, Goethite, Haematite, Schwertmannite | 11 |
| Manganese oxides | β-MnO2, Birnessite | 3 |
| Aluminium oxides | γ-Al2O3 | 3 |
| Clays | Kaolinite, Montmorillonite | 2 |
Documented biological processes
Two biological entries are present, corresponding to bacterial Sb(III) → Sb(V) oxidation experiments. Both show negative fractionation, in contrast to abiotic processes, making them discriminating in a tracing workflow. The documented strains are Ensifer (Ferrari et al., 2023) and Pseudomonas sp. J1 (Jia et al., 2024).
Integrated bibliographic references
| Reference | Year | Processes covered |
|---|---|---|
| Rouxel et al. | 2003 | Redox |
| MacKinney | 2016 | Redox |
| Zhou et al. | 2022 | Kinetic adsorption (γ-Al2O3) |
| Ferrari et al. | 2022 | Theoretical redox (Senarmontite/Cervantite) |
| Ferrari et al. | 2023 | Redox, biological (Ensifer) |
| Zhou et al. | 2023 | Adsorption (Ferrihydrite, Goethite, Haematite) |
| Ferrari et al. | 2024 | Adsorption (Ferrihydrite, Schwertmannite), Sb(III) and Sb(V) |
| Luo et al. | 2024 | Adsorption (Ferrihydrite, Goethite) |
| Wu et al. | 2024 | Adsorption (Goethite, Haematite, γ-Al2O3, β-MnO2, Kaolinite) |
| Wen et al. | 2024 | Adsorption (β-MnO2) |
| Zhou et al. | 2024 | Adsorption (Birnessite) |
| Jia et al. | 2024 | Biological (Pseudomonas sp. J1) |
| Kaufmann et al. | 2024 | Stibnite dissolution (Sb2S3) |
Data Quality and Reliability Indicator
The quality_flag field allows the confidence accorded to each entry in Nexus calculations to be weighted. A high indicator is assigned to entries that combine an explicitly reported analytical uncertainty, demonstrated reproducibility and rigorous experimental study (well-controlled conditions, documented equilibrium). In the current version, the vast majority of entries do not yet have an assigned flag, reflecting the state of available literature on Sb: uncertainties are not always reported in a standardised way in publications.
The high-quality entry in the current version corresponds to stibnite dissolution (Kaufmann et al., 2024, δ = +0.50 ‰ ± 0.05), one of the few to provide an explicit uncertainty and complete experimental conditions.
In the Nexus, when multiple entries match the same process type and material, the engine calculates a representative fractionation value (mean or median depending on data scatter), weighted by the quality_flag when set. Entries without a flag contribute with a neutral weight. The user can always replace this default value with their own measurement in the card.
Extending the Database to Other Elements
The database is designed to be multi-element. Its schema is generic: the element and isotope_system fields have no constraints on accepted values, allowing any documented isotopic system (Fe, Zn, Cu, Mo, Cr, Hg, etc.) to be integrated without schema modification.
New entries can be added directly via the Process Library in the Nexus. Each entry added by the user is stored in the same table, with the added_by field set to the current user's identifier to distinguish data integrated by IsoFind from locally added data.
Data added locally by the user is not synchronised with updates to the database distributed by IsoFind. During a software update, new entries from the official database are added without overwriting existing entries. Any locally created entry is preserved as long as it is not manually deleted.
To contribute fractionation data to IsoFind's distributed database, the recommended channel is to submit entries in the table format (element, process, phases, δ value, conditions, DOI reference) via the IsoFind contact form. Data will be verified and integrated in the next update after validation.