EN

Calculate the Mineral Saturation Index (SI)

The saturation index (SI) compares what a solution actually contains (the ion activity product, IAP) with what it could hold at most before precipitation (the solubility product, Ksp, of the mineral in question). It shows whether a groundwater or brine is likely to dissolve or precipitate a given mineral.

SI < 0: undersaturated solution, the mineral may dissolve. SI ≈ 0: solution at equilibrium. SI > 0: supersaturated solution, the mineral tends to precipitate.

Method

Three minerals with simple stoichiometry (one cation per anion): IAP = [cation] × [anion].

Mineral and ion activities (mol/L)

Pre-filled for the selected mineral; editable if your reference differs or the temperature isn't 25 °C.

For calcite, CO₃²⁻ activity is typically far lower than Ca²⁺ in a natural water — most dissolved inorganic carbon sits as HCO₃⁻, not CO₃²⁻.

For any mineral, including more complex stoichiometries (dolomite, fluorite…) whose IAP has already been worked out from its ion activities.

IAP and Ksp

Terms marked * are defined in the glossary.

IAP versus Ksp, and why the comparison is a log

Two numbers, same algebraic shape, different jobs. Ksp* is the solubility product of the mineral at equilibrium — a thermodynamic constant for that solid, at that temperature and pressure. IAP* is the ion activity product of the water being examined, built from the same ions, the same exponents, but from the activities actually present. SI* is how far the water sits from the mineral.

AaBb (s) ⇌ a A + b B

Ksp = {A}a {B}b  ·  IAP = {A}watera {B}waterb

SI = log10(IAP / Ksp) = log10(IAP) − log10(Ksp)

Calcite is the 1:1 case this form is built for: IAP = {Ca2+}{CO32−}, log10 Ksp ≈ −8.48 at 25 °C, Ksp ≈ 3.3 × 10−9. Gypsum and halite, same pattern. Dolomite, fluorite, a clay: the IAP is worked out first, then the second tab.

Why the log? Because Ksp jumps from 101.6 (halite) to 10−8.5 (calcite) to 10−17 (something nastier). A raw ratio of 2 and a raw ratio of 0.5 are symmetric on a log scale (SI = +0.30 and −0.30) and unreadable as 6.6 × 10−9 versus 1.65 × 10−9. SI = 0 is equilibrium, SI < 0 is undersaturated (the mineral can dissolve), SI > 0 is supersaturated (it can precipitate). Direction, not a rate.

Activity is not concentration

Type 80 mg/L Ca into a calculator that wants activity, and the input is already a lie. Activity ai is the effective concentration, concentration times an activity coefficient γi. Only in a near-distilled water is γ ≈ 1 and the lie small.

ai = γi · (ci / c°)  ·  I = ½ Σ ci zi2

I is ionic strength*. Fresh limestone water, I ≈ 0.005–0.02, γ for Ca2+ is already ~0.7–0.8, for CO32− similar or a bit lower. Seawater, I ≈ 0.7, γ for a divalent ion can sit around 0.2–0.3 depending on the model. A brine: Davies and Debye–Hückel have left the building; Pitzer is the model. The catch is that IAP is a product. Two coefficients of 0.7 turn IAP into half of the concentration product. SI then drops by ~0.3. That is the difference between “slightly scaling” and “near equilibrium”. This page does not compute γ. If millimolar concentrations are pasted, that belongs on the printout.

Speciation is the other half. “Calcium” from the lab is total Ca. A good fraction may sit as CaSO40, CaHCO3+, CaCO30. Free Ca2+ is what Ksp is written with. Same for carbonate: most DIC at pH 7 is HCO3−, not CO32−. The field note under the anion box is not decoration. Carbonate speciation is a pH problem as much as a geology one.

Why SI = 0 is a band, not a point

The classifier on this page treats |SI| ≤ 0.05 as equilibrium. That band is a shrug, not a theorem. Stack the errors and 0.05 is already optimistic.

  • pH. Calcite SI moves almost 1:1 with pH once HCO3− is fixed, because {CO32−} follows aH. A 0.05 pH miss — dirty electrode, warm sample, degassing in the bottle — is ~0.05 on SI, and 0.1 pH is common in the field.
  • Ksp itself. Published log Ksp for calcite at 25 °C wanders from about −8.42 to −8.52 depending on the database. That is already ±0.05 before anything is measured.
  • γ and complexes. Davies versus extended Debye–Hückel versus Pitzer, 25 °C constants used on a 12 °C water: another few hundredths, more in saline water.
  • Alkalinity versus CO32−. Alkalinity is titrated. The machine wants {CO32−}. The conversion is pH, T, and the whole carbonate tree.

In short, |SI| < 0.1 on calcite in a real groundwater is “cannot tell from equilibrium”. Decisions about scale or lining wait for a clearer miss — SI > +0.2 to +0.3, or SI < −0.2 — and even then kinetics get a vote. After all, a number to two decimals is not two decimals of knowledge.

Calcite and dolomite in groundwater

Most drinking-water chemistry is the calcite system in disguise. Limestone, chalk, marble, a calcite cement in sandstone: the water sees CaCO3 and soil CO2.

CaCO3 + CO2 + H2O ⇌ Ca2+ + 2 HCO3−

CaMg(CO3)2 + 2 CO2 + 2 H2O ⇌ Ca2+ + Mg2+ + 4 HCO3−

Closed system (no extra CO2 from the soil, recharge already in the rock): pH climbs, SIcalcite heads toward 0, hardness is modest. Open system under a biologically active soil (PCO2 ten to a hundred times atmospheric): the water dissolves more calcite, Ca and alkalinity rise together, pH stays lower. That is why a chalk spring is hard, and why a granite-catchment stream is not.

Dolomite is slower. Groundwaters in dolostone often sit at SIcalcite ≈ 0 and SIdolomite still negative, because the mineral will not keep up. A high Mg/Ca with both indices positive is a different story — evaporation, or a long residence, or a pump-and-heat scheme. Pair the two SI values; do not name the aquifer from calcite alone. The thin section still wins if one is available.

Temperature and Ksp

Calcite is backwards compared with sugar in tea. Dissolution is exothermic: Ksp falls as the water warms. Same Ca, same carbonate, higher T: SI rises. Heat a groundwater and it wants to drop scale. Cool it and it wants to pick calcite off a lining.

d ln Ksp / dT = ΔH° / (R T2)

T (°C)log10 Ksp calcite (working figures)
0−8.38
15−8.43
25−8.48
50−8.64
80−8.90

The form pre-fills −8.48. That is 25 °C. A 60 °C heat exchanger is not 25 °C. Edit log10(Ksp), or scale is understated on the hot side and overstated on a cold spring. Gypsum goes the other way (more soluble when warm). Halite barely cares for ordinary T. One checkbox does not cover three minerals.

Pressure at borehole depth is a smaller term for calcite than T is, until a geothermal well. Then a proper database is wanted, not this page plus a hunch.

Scale in plant and wellbore

Scale in a kettle, a heat-exchanger bundle, a well screen, a dripper: most of the time it is calcite, sometimes aragonite when the water is hot and Mg is high, sometimes gypsum if sulphate is the story, rarely halite unless the water is evaporating. SI says whether precipitation is allowed. It does not say millimetres per year.

  • SI < −0.3 on calcite: aggressive water. It will pick at concrete, grout, old limestone mains, a marble slab. Lining chemistry, not a soap-scum problem.
  • |SI| < 0.1: leave it. A chemical programme cannot be defended on that.
  • SI > +0.3: scale is allowed. Heat, a pressure drop, a splash into air, and it will show up on the hot face first.

The usual trigger in a plant is not the groundwater as sampled. It is what is done to it: warmed 40 °C, CO2 stripped in a tower, two wells mixed, pH raised with soda ash. Recalculate SI at the process T and PCO2, not at the wellhead. A water that is fine in the aquifer can plaster a calorifier.

What PHREEQC adds to a pocket log-ratio

This calculator is a pocket log-ratio for a 1:1 mineral, or for an IAP already trusted. A natural water is dozens of elements, ion pairs, redox, and a mineral list longer than the menu. PHREEQC* (USGS) is what most hydrogeochemists run: speciation, activity models (Davies, extended Debye–Hückel, SIT, Pitzer), SI for every solid in the database, mixing, inverse models, a kinetic rate if one is fed in.

  • Speciation first. Total Ca from the bottle becomes free Ca2+ plus pairs. IAP uses the free ion. Hand calculation skips that unless someone is stubborn.
  • Many minerals at once. Calcite, dolomite, gypsum, anhydrite, fluorite, chalcedony, ferrihydrite, siderite. The surprise SI is often silica or a ferric oxide, not calcite.
  • The database is a choice. phreeqc.dat, wateq4f.dat, llnl.dat, minteq.v4.dat, pitzer.dat do not share one log Ksp. Quote the file. A 0.1 jump in calcite log K between databases is not a new aquifer.

Geochemist's Workbench, EQ3/6, PHREEQC’s cousins: same physics, different buttons. None of them repair a pH measured in the lab on a sample that fizzed in the car. Garbage activities in, beautiful SI out.

Worked example: scaling versus aggressive water

Two waters, 25 °C constants matching the pre-fill, activities estimated rather than hand-waved as concentrations. This is the sketch done before PHREEQC, not instead of it.

Water A — limestone borehole. pH 7.8, Ca 80 mg/L (0.00200 mol/L), alkalinity 250 mg/L as HCO3− (0.00410 mol/L), I ≈ 0.01. At pH 7.8, {CO32−}/{HCO3−} ≈ K2 / aH ≈ 0.003, hence free CO32− ≈ 1.2 × 10−5 mol/L. With γCa ≈ 0.75 and γCO3 ≈ 0.70:

IAP ≈ (0.75 × 0.00200) × (0.70 × 1.2 × 10−5) ≈ 1.3 × 10−8

SI = log10(1.3 × 10−8 / 3.3 × 10−9) ≈ +0.6

Supersaturated. Scale-forming if heated or allowed to degas. In the aquifer it may sit for years without dropping a crystal — Mg, phosphate, a lack of nuclei. On a 60 °C exchanger, Ksp is smaller and this SI is worse. Softening, acid, a threshold inhibitor: that is a plant decision, not a geology one.

Water B — granite-sandstone catchment. pH 6.6, Ca 25 mg/L (6.2 × 10−4 mol/L), HCO3− 60 mg/L (9.8 × 10−4 mol/L). CO32− is now ~2 × 10−7 mol/L. IAP lands near 8 × 10−11.

SI ≈ log10(8 × 10−11 / 3.3 × 10−9) ≈ −1.6

Aggressive. This water will etch calcite grout and eat a concrete culvert. It will not scale a kettle. Lining, not anti-scalant. If A and B are mixed in a network, do not average the SI values — mix the waters, then recompute. The element concentrations add; the log of a product does not.

PCO2, degassing, and the weir that grows travertine

Soil air under grass can sit at PCO2 = 10−2 to 10−1.5 atm. The atmosphere is ~10−3.4. A spring that equilibrated with soil CO2 in the vadose zone is loaded with DIC. Hit the weir, the cascade, the tap: CO2 leaves, pH jumps, {CO32−} jumps, SIcalcite slams positive, and the weir grows a crust. Travertine and calcareous tufa are that process with time and a lot of water.

The bottle in the van does the same on a smaller scale. Fill it, leave headspace, wait three hours, measure pH in the lab: Water A now looks more scaling than it was down the hole. Field pH, field alkalinity, a filtered sample for Ca, or a preserved bottle with no bubble. SI is not more precise than the pH that was actually in the formation.

Do not compute calcite SI from a laboratory pH on an unpreserved, half-full bottle of a sparkling limestone water. The sample has already degassed. The diagnosis will be scale the aquifer does not have, or missed aggression that it does.

Gypsum and halite: the other two minerals in the menu

Gypsum (CaSO4·2H2O), log10 Ksp ≈ −4.58 at 25 °C. IAP = {Ca2+}{SO42−} (water activity belongs in a careful Ksp; this page ignores that). Many waters on gypsum beds sit near SI = 0 for gypsum and still negative for calcite, or the reverse. Seawater is undersaturated in gypsum; evaporate it and gypsum drops before halite. A cooling tower that concentrates a sulphate water will hit gypsum scale even when calcite is inhibited.

Halite (NaCl) is the odd one in the list: log10 Ksp ≈ +1.58, Ksp ≈ 38. Ordinary groundwater has SIhalite around −5 to −2. A brine is needed — salt-pan, evaporite aquifer, some oilfield waters — before SI approaches 0. If this form returns SI ≈ 0 for halite on a tap-water analysis, activities were typed as if they were mol/L of a Dead Sea end-brine, or a milligram was swapped for a mole.

The 1:1 algebra is the same. The geology is not. Gypsum scale is sulphate chemistry and often a concentration step. Halite scale is desalination, evaporators, and winter salt, not a chalk spring.

North Sea produced water: calcite, barite, and a pressure drop

UKCS production chemists live on SI, even when they quote a proprietary scale package rather than this page. A Forties or Brent produced water is often near calcite equilibrium down-hole (high PCO2, 80–100 °C). At the wellhead the pressure falls, CO2 leaves, T drops, and SIcalcite slams positive — the same degassing story as the travertine weir, with a Christmas tree instead of a cascade. Barite (BaSO4) is the other North Sea special: mixing a barium-rich formation water with a sulphate-rich seawater flood gives SIbarite ≫ 0 even when each end-member was fine on its own. The IAP is not 1:1 in the menu here; the second tab takes a precomputed IAP from PHREEQC or ScaleChem.

BGS and operator formation-water catalogues (Ca, Ba, SO4, alkalinity) are the English-language input, not a French spa analysis. A positive SI is still only permission to precipitate. Nuclei on a choke, a seed of old scale, and a sudden splash drop crystals in an afternoon; a clean, Mg-bearing water can sulk at SI = +0.5 for a long time. Millimetres per year need a rate law, a surface area, and a flow. PHREEQC will integrate a rate if one is given. This form will not.

Langelier, Ryznar, and why the plant still quotes them

Operators older than PHREEQC grew up on the Langelier saturation index* (LSI = pH − pHs) and the Ryznar stability index (RSI = 2 pHs − pH). pHs is the pH at which that water would be in equilibrium with calcite, from a recipe that uses total dissolved solids, Ca, alkalinity and T. For a dilute, oxidizing water, LSI tracks thermodynamic SI fairly well. For a brine, a reducing water, or anything with ion pairs that cannot be ignored, it is a cousin, not a twin.

LSI > 0 is “scaling tendency”, LSI < 0 “aggressive”, in that language. RSI bins are even more empirical (RSI < 6 scaling, > 7 aggressive — recipes vary). They exist because a treatment works can compute them from the four numbers on the daily sheet without a speciation code. Use them as a plant index if the plant already does. Do not convert an LSI of +0.4 into a PHREEQC SI of +0.4 and call it validation.

This page’s SI is the thermodynamic one: log10(IAP/Ksp), activities in. If all that is in hand is mg/L and a pH, either estimate γ and the CO32− split by hand, or go to PHREEQC. The geochemistry is the same either way; the arithmetic is not. Redox sits outside the 1:1 menu and wrecks iron and manganese SI values if the water is pretended to be at equilibrium with the atmosphere. Fe(III) oxides look wildly supersaturated in a reduced borehole once the sample has rusted in the bottle. Sample iron for SI the way isotopes are sampled: on purpose, or not at all.

Glossary

Ksp
Solubility product of the mineral at equilibrium, at a stated temperature. For calcite at 25 °C, log10 Ksp ≈ −8.48. A thermodynamic constant of the solid, not of the water.
IAP
Ion activity product of the water, built from the same ions and exponents as Ksp but from the activities actually present. For a 1:1 salt, IAP = {cation}{anion}.
Saturation index (SI)
SI = log10(IAP/Ksp). Negative: undersaturated (dissolution allowed). Zero: equilibrium. Positive: supersaturated (precipitation allowed). Direction, not a rate.
Ionic strength (I)
I = ½ Σ ci zi². Controls activity coefficients. Fresh water I ≈ 0.01; seawater ≈ 0.7; a brine needs Pitzer, not Davies.
PHREEQC
USGS aqueous geochemistry code. Speciation, activity models, SI for every solid in the chosen database, mixing and inverse models. The pocket calculator on this page is the log-ratio step only.
Langelier saturation index (LSI)
Plant index LSI = pH − pHs, where pHs is an empirical calcite-equilibrium pH from TDS, Ca, alkalinity and T. Tracks thermodynamic SI in dilute water; not a substitute for IAP/Ksp in a brine.

Related calculators