Phase Rule and Thermodynamics of Solutions and Fluids
CSIR NET Earth Sciences geochemistry: Gibbs and Goldschmidt phase rules, Clapeyron slopes, equilibrium constants, activity, regular solutions, fugacity.
- 18 min
On this page · 14 sections
Prerequisites
- The laws of thermodynamics and the Gibbs function at MSc level: Laws of Thermodynamics.
- The chemist’s version of the phase rule and the Clapeyron equation: Phase Transitions, Phase Equilibria and the Phase Rule. This page applies it to rocks.
- Solid solutions and coupled substitution from Atomic Structure, the Periodic Table and Ionic Substitution.
The Hook: Three Polymorphs at One Point
Thermodynamic Toolkit
For a closed system at equilibrium, the Gibbs energy \( G = H - TS \) is minimum at fixed \( T \) and \( P \), and
$$dG = V\,dP - S\,dT + \sum_i \mu_i\,dn_i$$where \( \mu_i = (\partial G/\partial n_i)_{T,P,n_{j\neq i}} \) is the chemical potential of component \( i \). Phases are in equilibrium when each component has the same \( \mu \) in every phase in which it is present.
For a reaction, with \( \Delta \) meaning products minus reactants:
$$\left(\frac{\partial \Delta G}{\partial P}\right)_T = \Delta V \qquad \left(\frac{\partial \Delta G}{\partial T}\right)_P = -\Delta S$$So raising pressure favours the side with smaller volume (denser phases: garnet, jadeite, kyanite, coesite), and raising temperature favours the side with larger entropy (fluids, melts, disordered phases, sillimanite).
The Phase Rule
Gibbs phase rule
$$\boxed{F = C - P + 2}$$- \( C \): the minimum number of chemical components needed to describe the composition of every phase (after subtracting independent reactions and fixed constraints).
- \( P \): number of phases (each mineral, the melt, the fluid).
- \( F \): degrees of freedom: intensive variables (\( T \), \( P \), compositions) that can be changed independently without changing \( P \).
- The “2” stands for \( T \) and \( P \). At fixed pressure (an isobaric diagram) use \( F = C - P + 1 \).
| Situation | C | P | F | Meaning |
|---|---|---|---|---|
| \( \text{Al}_2\text{SiO}_5 \) single polymorph | 1 | 1 | 2 | divariant field |
| Kyanite = sillimanite | 1 | 2 | 1 | univariant line |
| Ky + And + Sil | 1 | 3 | 0 | invariant (triple) point |
| Binary eutectic at fixed \( P \) (two solids + melt) | 2 | 3 | 0 (isobaric) | fixed \( T \) and melt composition |
| \( \text{CaCO}_3 + \text{SiO}_2 = \text{CaSiO}_3 + \text{CO}_2 \), all four present | 3 (CaO, SiO\(_2\), CO\(_2\)) | 4 | 1 | univariant curve in \( P \)–\( T \) (if fluid is pure CO\(_2\)) |
Goldschmidt’s mineralogical phase rule
A rock is normally found in a divariant field: it formed over a range of \( T \) and \( P \), not on a line or at a point. Setting \( F \ge 2 \) in the Gibbs rule:
$$\boxed{P \le C}$$“The number of minerals in a rock in equilibrium over a range of \( P \) and \( T \) does not exceed the number of components.” A rock with more phases than components records a reaction caught in progress, disequilibrium, or a component missed in the count.
Korzhinskii’s open-system version. If some components are perfectly mobile (their chemical potential is fixed by an external reservoir, for example \( \text{H}_2\text{O} \), \( \text{CO}_2 \), \( \text{K}_2\text{O} \) in metasomatism), only the inert components count:
$$P \le C_{\text{inert}}$$So metasomatic zones (skarns, alteration haloes) have few minerals, often monomineralic bands.
Clapeyron slopes of reactions
Along a univariant boundary, \( d(\Delta G) = 0 \), so
$$\boxed{\frac{dP}{dT} = \frac{\Delta S}{\Delta V} = \frac{\Delta H}{T\,\Delta V}}$$| Reaction type | \( \Delta S \) | \( \Delta V \) | Slope in \( P \)–\( T \) |
|---|---|---|---|
| Solid–solid (polymorphs, Ky = Sil) | small | small | moderate, straight lines; sign varies |
| Dehydration / decarbonation (fluid released) | large positive | positive at low \( P \), shrinking at high \( P \) | low slope at low \( P \), steepening upward; curve bends back at very high \( P \) |
| Melting (dry) | positive | positive | positive and steep |
| Melting with \( \text{H}_2\text{O} \) (wet solidus) | positive | strongly negative at low \( P \) as water dissolves in the melt | negative slope at low pressure (up to a few kbar), then positive and steep |
Units: 1 J = 1 Pa m\(^3\), so \( \Delta S \) in J mol\(^{-1}\) K\(^{-1}\) divided by \( \Delta V \) in m\(^3\) mol\(^{-1}\) gives Pa K\(^{-1}\). Remember 1 cm\(^3\) = \( 10^{-6} \) m\(^3\) and 1 J bar\(^{-1}\) = 10 cm\(^3\).
Equilibrium Constants and Activity
For each component, \( \mu_i = \mu_i^\circ + RT\ln a_i \), where \( a_i \) is the activity relative to a chosen standard state. For a reaction at equilibrium:
$$\boxed{\Delta G^\circ = -RT\ln K} \qquad K = \prod a_i^{\nu_i}\ (\text{products positive, reactants negative})$$and away from equilibrium \( \Delta G = \Delta G^\circ + RT\ln Q \). Pure solids and pure liquids in their standard state have \( a = 1 \).
The van ’t Hoff equation gives the temperature dependence:
$$\frac{d\ln K}{d(1/T)} = -\frac{\Delta H^\circ}{R} \qquad \ln\frac{K_2}{K_1} = -\frac{\Delta H^\circ}{R}\left(\frac{1}{T_2} - \frac{1}{T_1}\right)$$For geothermometers the equilibrium constant of an exchange reaction (Fe–Mg between garnet and biotite, for example) is calibrated against \( T \); pressure-sensitive net-transfer reactions (with large \( \Delta V \)) make geobarometers. That is the thermodynamic basis of thermobarometry.
Thermodynamics of Solutions
Ideal solutions
In an ideal solution the components mix randomly with no heat of mixing:
$$\Delta H_{\text{mix}} = 0 \qquad \Delta S_{\text{mix}} = -R\sum_i X_i\ln X_i \qquad \Delta G_{\text{mix}} = RT\sum_i X_i\ln X_i$$\( \Delta S_{\text{mix}} \) is always positive, so \( \Delta G_{\text{mix}} \) is always negative: mixing is spontaneous. For a 50:50 binary, \( \Delta S_{\text{mix}} = R\ln 2 = 5.76 \) J mol\(^{-1}\) K\(^{-1}\).
Raoult’s law (\( a_i = X_i \)) holds for the major component as \( X_i \to 1 \). Henry’s law (\( a_i = h_i X_i \), with a constant \( h_i \)) holds for a trace component at high dilution. Henry’s law is why trace-element partition coefficients are constant over a range of low concentrations (see the next topic).
Mixing on crystallographic sites
In minerals, ions mix on sites, not as whole formula units. For an end-member with \( n \) equivalent sites occupied by the ion in question, the ideal activity is
$$\boxed{a_i^{\text{ideal}} = (X_i^{\text{site}})^{n}}$$| Solid solution | Site | Ideal activity |
|---|---|---|
| Olivine \( (\text{Mg},\text{Fe})_2\text{SiO}_4 \) | 2 octahedral M sites | \( a_{\text{Fo}} = X_{\text{Mg}}^2 \) |
| Garnet \( (\text{Mg},\text{Fe},\text{Ca})_3\text{Al}_2\text{Si}_3\text{O}_{12} \) | 3 dodecahedral X sites | \( a_{\text{Prp}} = X_{\text{Mg}}^3 \) |
| Orthopyroxene \( (\text{Mg},\text{Fe})_2\text{Si}_2\text{O}_6 \) | M1 and M2 | \( a_{\text{En}} = X_{\text{Mg}}^{\text{M1}}X_{\text{Mg}}^{\text{M2}} \) |
For garnet with \( X_{\text{Mg}} = 0.6 \): \( a_{\text{Prp}} = 0.6^3 = 0.216 \), not 0.6. Forgetting the exponent is a standard error.
Non-ideal solutions: activity coefficients and the regular solution
Real solutions have \( a_i = \gamma_i X_i \). The simplest model, the symmetric regular solution, adds a heat of mixing with an interaction parameter \( W \) and keeps ideal entropy:
$$\Delta H_{\text{mix}} = W X_1 X_2 \qquad RT\ln\gamma_1 = W X_2^2 \qquad RT\ln\gamma_2 = W X_1^2$$- \( W > 0 \): unlike neighbours are disfavoured; positive deviation (\( \gamma > 1 \)); at low \( T \) the \( \Delta G_{\text{mix}} \) curve develops two minima and the solution unmixes.
- The two coexisting compositions define the solvus; its top is the critical (consolute) temperature:
- Below \( T_c \), a homogeneous crystal grown at high \( T \) exsolves lamellae on slow cooling: perthite (Na-rich lamellae in K-feldspar), antiperthite, exsolved pyroxenes (pigeonite or orthopyroxene lamellae in augite), and the plagioclase peristerite gap.
- Two-feldspar and two-pyroxene solvus thermometers use the compositions of coexisting phases on the limbs of the solvus.
Binary phase diagrams for melts and crystals
Isobaric \( T \)–\( X \) diagrams apply \( F = C - P + 1 \).
| Type | Example (at 1 atm) | Key feature |
|---|---|---|
| Eutectic | diopside–anorthite | two liquidus curves meet at the eutectic; melts of any bulk composition finish crystallising at the eutectic \( T \) |
| Complete solid solution | albite–anorthite (plagioclase); forsterite–fayalite (olivine) | liquidus and solidus loops; first crystals richer in the high-\( T \) end-member (An, Fo) |
| Peritectic (incongruent melting) | forsterite–silica | enstatite melts to forsterite + a more silica-rich liquid; reaction relation |
| Solvus + solid solution | alkali feldspar | homogeneous crystals at high \( T \), unmixing (perthite) at low \( T \) |
Lever rule. At a temperature inside a two-phase field, draw the tie line through the bulk composition \( X_0 \). With phase compositions \( X_s \) (solid) and \( X_l \) (liquid):
$$\boxed{\text{fraction of liquid} = \frac{X_0 - X_s}{X_l - X_s}} \qquad \text{fraction of solid} = \frac{X_l - X_0}{X_l - X_s}$$Equilibrium vs fractional crystallisation in a solid-solution loop. In equilibrium crystallisation, crystals continuously react with the melt and the last liquid disappears when the solid reaches the bulk composition. In fractional crystallisation, crystals are removed or zoned; the liquid moves further, and normal zoning (calcic cores, sodic rims in plagioclase) records it. This is the experimental basis of Bowen’s reaction series.
Thermodynamics of Fluids
Fugacity
For an ideal gas, \( \mu = \mu^\circ + RT\ln(P/P^\circ) \). For a real fluid, pressure is replaced by fugacity \( f \), the “effective pressure”:
$$\boxed{\mu = \mu^\circ + RT\ln\frac{f}{P^\circ}} \qquad f = \phi P$$where \( \phi \) is the fugacity coefficient (1 for an ideal gas). In a fluid mixture, \( f_i = \phi_i X_i P \) (ideal mixing of real gases: the Lewis–Randall rule). At crustal pressures \( \phi_{\text{H}_2\text{O}} \) and \( \phi_{\text{CO}_2} \) depart strongly from 1, which is why experimentally derived equations of state are used in thermobarometry.
Supercritical fluids
Above the critical point, liquid and vapour become one fluid phase. For pure water the critical point is about 374 °C and 22.1 MPa; for \( \text{CO}_2 \) about 31 °C and 7.4 MPa. Most metamorphic and hydrothermal fluids in the middle and deep crust are therefore supercritical: dense, with high solvent power for silica and metals. \( \text{H}_2\text{O} \)–\( \text{CO}_2 \) mixtures are miscible at high \( T \) but unmix at lower \( T \), especially when salts are present; this immiscibility is recorded by coexisting fluid inclusions.
Mixed-volatile reactions and \( T \)–\( X_{\text{CO}_2} \) diagrams
For a decarbonation such as
$$\text{CaCO}_3 + \text{SiO}_2 = \text{CaSiO}_3 + \text{CO}_2$$with pure solids, \( K = a_{\text{CO}_2} \) (or \( f_{\text{CO}_2} \) relative to a standard state). Diluting the fluid with \( \text{H}_2\text{O} \) lowers \( a_{\text{CO}_2} \) and lowers the reaction temperature at fixed \( P \). On a \( T \)–\( X_{\text{CO}_2} \) diagram:
- decarbonation curves rise from low \( T \) at \( X_{\text{CO}_2} \to 0 \) to their maximum at \( X_{\text{CO}_2} = 1 \);
- dehydration curves have their maximum at \( X_{\text{CO}_2} = 0 \);
- reactions releasing both \( \text{H}_2\text{O} \) and \( \text{CO}_2 \) have a maximum at an intermediate \( X_{\text{CO}_2} \) equal to the mole fraction of CO\(_2\) among the volatiles released (for example 0.5 if the reaction releases one CO\(_2\) and one H\(_2\)O).
An infiltrating water-rich fluid can therefore drive decarbonation at a lower temperature: the basis of skarn formation and wollastonite near intrusions.
graph LR
A[Count components C and phases P] --> B{Is a component perfectly mobile?}
B -- No --> C[Gibbs: F = C - P + 2]
B -- Yes --> D[Korzhinskii: count only inert components]
C --> E{Rock formed over a range of T and P?}
E -- Yes --> F[Goldschmidt: P not more than C]
E -- No --> G[Univariant line or invariant point]Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| Gibbs: \( F = C - P + 2 \) | Goldschmidt: \( P \le C \) (assumes \( F \ge 2 \)) | |
| Raoult (solvent, \( a = X \)) | Henry (trace solute, \( a = hX \)) | |
| Activity (effective concentration) | Fugacity (effective pressure of a gas or fluid) | |
| Solvus (unmixing of a solid solution on cooling) | Solidus (beginning of melting) | |
| Eutectic: two solids + melt, both solids stable to melting | Peritectic: one solid reacts with melt to give another solid | |
| Exchange reactions (\( \Delta V \approx 0 \)): thermometers | Net-transfer reactions (large \( \Delta V \)): barometers |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Take \( R = 8.314 \) J mol\(^{-1}\) K\(^{-1}\).
Level 1: Foundation
- How many degrees of freedom does a one-component system have at its triple point?
- Write the relation between \( \Delta G^\circ \) and the equilibrium constant.
- What does a fugacity coefficient of 1 mean?
Answers
- Zero (\( F = 1 - 3 + 2 = 0 \)): an invariant point.
- \( \Delta G^\circ = -RT\ln K \).
- The fluid behaves as an ideal gas (\( f = P \)).
Level 2: Part B Style
Practice Q1. A hornfels contains only quartz, calcite and wollastonite in a system described by CaO, SiO\(_2\) and CO\(_2\). The pore fluid was pure CO\(_2\) and is now gone. Treating CO\(_2\) as an inert component, the assemblage with the fluid present had:
(a) \( F = 0 \) (b) \( F = 1 \) (c) \( F = 2 \) (d) \( F = 3 \)
Answer: (b) F = 1
Phases: quartz, calcite, wollastonite and CO\(_2\) fluid = 4. Components = 3. \( F = 3 - 4 + 2 = 1 \): univariant. The rock lies on the reaction curve, which is why such assemblages are used as reaction “isograds”. By Goldschmidt’s rule (\( P \le C \)), four phases in three components cannot be a divariant assemblage.
Practice Q2. For a mineral reaction with \( \Delta S > 0 \) and \( \Delta V < 0 \), the univariant curve in \( P \)–\( T \) space:
(a) has a positive slope; products on the high-\( T \), low-\( P \) side (b) has a negative slope; products favoured by both higher \( T \) and higher \( P \) (c) is vertical (d) is horizontal
Answer: (b)
\( dP/dT = \Delta S/\Delta V < 0 \). The products have more entropy (favoured by \( T \)) and less volume (favoured by \( P \)), so they sit on the high-\( T \), high-\( P \) side of a negatively sloping line.
Practice Q3. In a garnet with \( X_{\text{Mg}} = 0.6 \) on the X site, the ideal activity of pyrope is:
(a) 0.60 (b) 0.36 (c) 0.216 (d) 0.072
Answer: (c) 0.216
Pyrope has three Mg on three equivalent X sites, so \( a = X_{\text{Mg}}^3 = 0.6^3 = 0.216 \). 0.36 would be correct for olivine (two sites).
Practice Q4. In a binary system, a bulk composition of 40 wt% B is at a temperature where liquid contains 60 wt% B and coexisting crystals 20 wt% B. The mass fraction of liquid is:
(a) 0.33 (b) 0.40 (c) 0.50 (d) 0.67
Answer: (c) 0.50
Liquid fraction \( = (X_0 - X_s)/(X_l - X_s) = (40 - 20)/(60 - 20) = 0.50 \). Option (d) takes the wrong lever arm.
Level 3: Part C Style
Practice Q5. A solid solution between end-members 1 and 2 follows the symmetric regular-solution model with \( W = 10 \) kJ mol\(^{-1}\).
Which of the following are correct?
- At 800 K and \( X_2 = 0.3 \), \( \gamma_1 \approx 1.14 \).
- The critical temperature of the solvus is about 601 K.
- At 1000 K the solution is continuous across all compositions.
- Because \( W > 0 \), \( \Delta H_{\text{mix}} \) is negative.
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2 and 4 only (d) 1, 3 and 4 only
Answer: (b) 1, 2 and 3 only
- 1 true: \( \ln\gamma_1 = W X_2^2/RT = 10000\times 0.09/(8.314\times 800) = 0.135 \); \( \gamma_1 = e^{0.135} = 1.145 \).
- 2 true: \( T_c = W/2R = 10000/16.628 = 601 \) K.
- 3 true: 1000 K is above \( T_c \), so there is no miscibility gap.
- 4 false: \( \Delta H_{\text{mix}} = W X_1X_2 \) is positive when \( W > 0 \).
Practice Q6. An exchange reaction used as a geothermometer has \( \Delta H^\circ = -50 \) kJ mol\(^{-1}\) and negligible \( \Delta V \). By what factor does its equilibrium constant change on cooling from 800 °C to 700 °C, and what does that imply?
(a) increases about 1.8 times; useful thermometer, insensitive to pressure (b) decreases about 1.8 times; useful barometer (c) increases about 18 times; useful thermometer (d) does not change; \( \Delta V \approx 0 \) makes it useless
Answer: (a)
\( \ln(K_{700}/K_{800}) = -\frac{\Delta H^\circ}{R}\left(\frac{1}{973.15} - \frac{1}{1073.15}\right) = \frac{50000}{8.314}\times 9.576\times 10^{-5} = 0.576 \), so \( K_{700}/K_{800} = e^{0.576} = 1.78 \). An exothermic reaction (\( \Delta H < 0 \)) has a larger \( K \) at lower \( T \). A large \( \Delta H \) with small \( \Delta V \) makes a good thermometer that depends little on pressure; (b) confuses this with a barometer, which needs large \( \Delta V \).
Quick Revision Table
| Concept | Key point |
|---|---|
| Gibbs phase rule | \( F = C - P + 2 \); isobaric \( F = C - P + 1 \) |
| Goldschmidt | \( P \le C \) for divariant rocks |
| Korzhinskii | \( P \le C_{\text{inert}} \) in open systems |
| Clapeyron | \( dP/dT = \Delta S/\Delta V \); 1 J = 1 Pa m\(^3\) |
| Equilibrium | \( \Delta G^\circ = -RT\ln K \); van ’t Hoff for \( T \) dependence |
| Ideal mixing | \( \Delta S_{\text{mix}} = -R\sum X\ln X \) (5.76 J mol\(^{-1}\) K\(^{-1}\) at 50:50) |
| Site activity | \( a = X^n \) (olivine \( n = 2 \), garnet \( n = 3 \)) |
| Regular solution | \( RT\ln\gamma_1 = WX_2^2 \); \( T_c = W/2R \) |
| Lever rule | liquid fraction \( = (X_0 - X_s)/(X_l - X_s) \) |
| Fugacity | \( f = \phi P \); \( \mu = \mu^\circ + RT\ln f \) |
| Critical points | H\(_2\)O about 374 °C, 22.1 MPa; CO\(_2\) about 31 °C, 7.4 MPa |
| \( T \)–\( X_{\text{CO}_2} \) | decarbonation maximum at \( X = 1 \); dehydration at \( X = 0 \) |
Key Takeaways
What’s Next
Equilibrium tells you where a system wants to go; kinetics tells you whether it gets there. Continue with Distribution Coefficients, Nucleation and Diffusion. For the electrochemical side of chemical potential (Eh and the Nernst equation), see Redox, Eh–pH, Waters and Aerosols.