Bulk Calculator for Fractional Crystallization
Estimate trace-element evolution during fractional crystallization using Rayleigh fractionation. You can enter a known bulk distribution coefficient (Dbulk) directly, or compute it from mineral modes and mineral-melt partition coefficients.
Primary Inputs
Optional Mineral-Mode Builder
If method is set to “Calculate D_bulk from mineral modes,” D_bulk = Σ(mode × Kd). Enter modes in percent.
How to Calculate Bulk for Fractional Crystallization: A Practical Expert Guide
Fractional crystallization is one of the most important differentiation mechanisms in igneous petrology. If you are trying to model melt evolution in basaltic, andesitic, or rhyolitic systems, an accurate estimate of the bulk distribution coefficient (Dbulk) is central to any defensible calculation. In simple terms, Dbulk tells you how strongly an element partitions into crystals relative to melt during crystallization. Once Dbulk and the remaining melt fraction F are known, the Rayleigh equation allows you to estimate the concentration of that element in the residual liquid and associated solids.
This page’s calculator is designed for rapid scenario testing with geologically meaningful outputs: residual melt concentration, instantaneous crystal concentration, and cumulative removed solid concentration. Below is a full explanation of the equations, assumptions, limitations, and interpretation strategies used by professionals in geochemistry and petrology.
Why “bulk” matters
Individual minerals each have their own partition coefficient Kd for a given element. In nature, however, crystallizing assemblages are rarely single-mineral systems. You usually have mixtures such as olivine + clinopyroxene + plagioclase, and each phase scavenges elements differently. The bulk coefficient combines all active phases into one effective term:
Dbulk = Σ (Xi × Kdi), where Xi is mineral modal fraction (as a decimal).
That single number drives first-order enrichment or depletion trends during fractional crystallization. If Dbulk is less than 1, the element is incompatible and tends to concentrate in melt as crystals are removed. If Dbulk is greater than 1, it is compatible and tends to be stripped from melt into crystals.
Core Rayleigh equations used in the calculator
- Residual melt concentration: Cl = C₀ × F(D-1)
- Instantaneous crystal concentration: Cs,inst = D × Cl
- Cumulative removed solids concentration: Cs,cum = C₀ × (1 – FD) / (1 – F), when F < 1
Here, C₀ is initial concentration and F is remaining melt mass fraction (for example, F = 0.35 means 35% melt left and 65% removed as crystals).
Step-by-step process for reliable calculations
- Define your starting composition. Use a representative C₀ from whole-rock, glass, or melt-inclusion data.
- Choose the right Dbulk approach. If you have phase proportions and mineral Kd values, compute Dbulk. If not, use literature-based approximations as a sensitivity bracket.
- Estimate F realistically. F can come from petrographic constraints, major-element modeling, or phase-equilibrium data.
- Run forward calculations. Compare predicted liquid trends to observed geochemical arrays.
- Stress test assumptions. Vary D and F to evaluate model uncertainty.
Interpreting D values in real systems
D values depend on pressure, temperature, oxygen fugacity, composition, and mineral chemistry. For example, Sr can behave moderately compatible in plagioclase-rich crystallization but more incompatible if plagioclase is suppressed. Likewise, Ni can be strongly compatible when olivine dominates but much less so once olivine saturation drops out. This is why “single fixed D” models are best seen as first-order approximations and not complete process reconstructions.
| Element | Typical petrologic behavior | Representative D range (bulk or mineral-dominated context) | Modeling implication |
|---|---|---|---|
| Rb | Highly incompatible | ~0.01 to 0.05 | Strong enrichment in residual melt as F decreases |
| Ba | Incompatible to mildly compatible | ~0.03 to 0.3 | Usually increases in evolved melts unless feldspar control is strong |
| La | Incompatible REE | ~0.05 to 0.3 | Good tracer for melt enrichment trajectories |
| Sr | Sensitive to plagioclase | ~0.5 to 2.5 | Can flip trend depending on mineral assemblage |
| Ni | Compatible in olivine-rich systems | ~1.5 to 10+ | Drops rapidly in melt during early mafic fractionation |
| Cr | Highly compatible in spinel/pyroxene control | ~2 to 20+ | Very strong melt depletion when compatible phases crystallize |
Ranges are compiled from commonly reported experimental and petrologic modeling intervals used in igneous geochemistry; exact values vary by P-T-X conditions and mineral chemistry.
Numerical comparison: how F changes melt enrichment and depletion
The table below uses the Rayleigh equation to illustrate two contrasting behaviors: an incompatible element (D = 0.1) and a compatible element (D = 2.0). Both start at C₀ = 50 ppm.
| Remaining melt fraction F | Incompatible case D=0.1, Cl (ppm) | Compatible case D=2.0, Cl (ppm) | Interpretation |
|---|---|---|---|
| 0.80 | ~61.1 | 40.0 | Early differentiation mildly enriches incompatible elements |
| 0.60 | ~79.2 | 30.0 | Divergence accelerates as crystallization proceeds |
| 0.40 | ~114.0 | 20.0 | Incompatible inventory concentrates strongly in melt |
| 0.20 | ~212.8 | 10.0 | Late-stage melts can become strongly enriched or depleted |
| 0.10 | ~397.2 | 5.0 | Extreme fractionation generates steep compositional gradients |
Worked example using the calculator
Assume a basaltic parent melt contains 50 ppm of an incompatible element, and petrologic constraints suggest that only 35% melt remains (F = 0.35). If your assemblage-weighted Dbulk is 0.2, then:
- Residual melt concentration Cl = 50 × 0.35-0.8 ≈ 116 ppm
- Instantaneous crystal concentration Cs,inst = 0.2 × 116 ≈ 23 ppm
- Cumulative removed solids concentration Cs,cum is much lower than melt concentration, consistent with incompatible behavior
This pattern is exactly what field and laboratory observations show for many incompatible trace elements during prolonged differentiation.
Common mistakes to avoid
- Mixing equilibrium and fractional frameworks: Rayleigh equations assume continuous removal of crystals from melt.
- Using fixed D across all stages: D often changes when the mineral assemblage changes.
- Ignoring recharge, mixing, or assimilation: Natural systems are open more often than we wish.
- Comparing unlike datasets: Whole-rock trends, glass trends, and inclusions can capture different process scales.
Best-practice workflow for professional modeling
- Build a first-pass Rayleigh model using 2 to 3 plausible D scenarios.
- Use modal mineralogy to compute stage-specific Dbulk.
- Compare predicted trajectories against multi-element suites, not single elements.
- Integrate with petrographic and thermodynamic constraints (phase appearance/disappearance).
- Report uncertainty ranges rather than single deterministic numbers.
Authoritative learning resources
For further technical reading and classroom-grade references, consult:
- USGS: Magma, Crystallization, and Igneous Processes
- MIT OpenCourseWare: Trace Element Geochemistry
- Carleton College (SERC): Equilibria and Geochemical Modeling Resources
Final takeaway
To calculate bulk for fractional crystallization correctly, focus on three things: a realistic initial concentration, defensible Dbulk values tied to mineral assemblage, and geologically constrained F. The calculator above gives you rapid, transparent outputs using established Rayleigh relationships and visual trend plotting. For advanced studies, use it as a first-pass tool before moving to stage-wise or open-system mass balance models. Done carefully, this approach is a powerful bridge between field data, petrology, and quantitative geochemical interpretation.