Trace element thermobarometry now serves as a standard for characterizing P-T conditions spanning the upper mantle to upper crust. Some equilibria – especially Zr-in-rutile, Zr-in-titanite, Ti-in-quartz, Ti-in-zircon, Ni-in-garnet and Ni-in-olivine – have well-founded thermodynamic formulations, and applications have firm observational and theoretical bases. However, other trace element calibrations lack a sound thermodynamic foundation, and any P-T interpretations should be treated skeptically. Three key requirements for thermodynamically valid thermobarometers are that they must represent (1) a balanced chemical reaction, expressed (2) among mineral phase components that (3) satisfy crystal-chemical controls on trace element substitution. . These requirements are missing from many calibrations.

Problematic thermobarometers include REE exchange for garnet-clinopyroxene, orthopyroxene-clinopyroxene, and plagioclase-clinopyroxene, as well as Ti-in-biotite and Ti-in-hornblende. These thermobarometers lack chemically balanced reactions among phase components, so are thermodynamically faulty. For Ti-in-biotite and Ti-in-hornblende, thermodynamically valid calibrations for balanced chemical reactions do exist (Chambers and Kohn, 2012; Am Min), but Ti contents depend too strongly on mineral composition and too weakly on temperature to be useful for thermometry. Simplified chemical reactions proposed for REE thermobarometers violate crystal chemistry (they do not represent equilibria among mineral phase components) and do not reflect two-mineral systems, instead implicitly requiring the presence of melt. More generally, current REE calibrations rely on partition coefficients rather than equilibrium constants, which does not accurately account for the effects of chemical mixing in minerals or chemistry of all crystallographic sites. Application of the REE garnet-clinopyroxene thermobarometer to well-equilibrated experiments and natural rocks leads to thermometric errors of hundreds of degrees, and barometric errors of tens of kbar, compromising the usefulness of the current calibration.

Advances in trace element thermobarometry will require stricter adherence to thermodynamic principles and further investigation of substitution mechanisms of trace elements in minerals. Some substitutions are straightforward (e.g., Ni²⁺ substitutes simply for Mg²⁺ in olivine and garnet), others are not (olivine has two possible Al substitution mechanisms). Strategies for assessing substitution mechanisms and mixing models are straightforward, e.g., REE diffusion rates in typical garnet and pyroxene are too fast to permit charge-coupling to tetrahedral Al via tschermaks- or YAG-type components. Instead, chemical and partitioning data strongly support Na+REE coupled substitution to produce, for example, a (Na·REE·Ca)Al₂Si₃O₁₂ component in garnet. However, these methods must be implemented more consistently.