The relationships between ferric iron content and the P-T-fH₂ conditions of formation were examined for two biotite compositions: annite (K₂Fe₅Al₄Si₅O₂₀(OH)₄) and siderophyllite (K₂Fe₆Al₂Si₆O₂₀(OH)₄). The synthesized phases were annealed at fixed hydrogen fugacities using both the solid oxygen buffering technique of Eugster (1957) and the H₂ buffering technique of Shaw (1967). Resulting hydrogen fugacities ranged from 0.004 bars (at T = 400°C, P_T = 2 kb) to 51 bars (at T = 750°C, P_T = 1 kb).

Ferrous iron contents of the annealed biotites were determined by wet chemical analyses. Total iron was determined by microprobe analyses to be equal to the stoichiometric values. The data confirm the predictions of Hazen and Wones (1972, 1978) that: 1) There is a structural limit imposed upon the Fe³⁺ content of annite due to the misfit between the octahedral and tetrahedral layers. This misfit requires a minimum of 11% Fe³⁺ in annite. 2) The steric misfit in annite can be corrected by a substitution of Al^vi + Al^iv for Fe^vi + Si^iv, so that there is no Fe³⁺ in siderophyllite at high hydrogen fugacities.

A model relating Fe³⁺/Fe²⁺ ratios, fH₂, and T is proposed. The model accounts for the amount of Fe³⁺ needed to correct the steric misfit in annite and allows for variation in Fe, Al, and Mg contents among biotites. A simple oxidation-reduction reaction is used to relate changes in the non-steric ferric iron to hydrogen fugacity and temperatures for the Fe-Mg-Al biotites. The equilibrium constant for the reaction can be expressed as:

(1) log K= 3607.2/T - 4.47

where, depending on composition, K is expressed as follows and R is a constant proportional to the U structurally required iron.

(2) Al^vi/0.3 + Mg/0.72 ≥ 1, K = Fe³⁺/Fe²⁺ fH₂^1/2

(Al^vi + Mg ≥ 1)

(3) Al^vi/0.3 + Mg/0.72 > 1, K = Fe³)/Fe²⁺ - (1-Al^vi-MG) fH₂^1/2

(Al^vi + Mg < 1)

(4) Al^vi/0.3 + Mg/0.72 < 1, K = [Fe³⁺/(Fe²⁺ - (1-Al^vi-MG)) -0.185] fH₂^1/2

(Al^vi + Mg < 1)