Contrib. Mineral. Petrol. 72, 415-424 (1980) Contributions to Mineralogy and Petrology 9 by Springer-Verlag 1980 Stability of Titanian Clinohumite: Experiments and Thermodynamic Analysis Martin Engi* and Donald H. Lindsley Department of Earth and Space Sciences, State University of New York, Stony Brook, New York 11794, USA Abstract. Reversed hydrothermal experiments on a natural titanoclinohumite [77-Cl; approximate for- mula Mg7.sFeTio.5016(OH)] show that it breaks down at 475 ~ + 11 ~ C (3.5 kbar), 620 ~ + 11 ~ C (14 kbar) and 675~ ~ (2l kbar) to the assemblage olivine +ilmenite+vapor. An internal-consistency analysis of the data yields ArG o (298 K, 1 bar) = 36,760 • 3,326 cal (mole 77- CI) t. A~ S~ (298 K, 1 bar) = 34.14 • 5.91 cal deg- 1 (mole 77- Cl) ~. Linear correlation coefficient rG_s~ 1.0. A solution model that accounts for TiO 2-M(OH)2 and F-OH substitution shows that the results for our nearly F-free 77-Cl are in reasonable agreement with the unreversed breakdown experiments of Mer- rill et al. (1972) on a F-bearing 77-CI. Because fluorine is necessary to stabilize 77-Cl under mantle conditions, we suggest that Ti-Cl is much more likely to be a "storage device" for fluo- rine than for water in the mantle. I. Introduction Previous work regarding the stability of clinohumite minerals has concentrated on natural occurrences, constraints imposed by crystal structure refinements, and some direct experimental investigations of phase equilibria. Some of these studies have yielded ap- parently conflicting results, particularly regarding the thermal stability of titanian clinohumite (abbreviated to Ti--CI) at high pressures and the possible geo- chemical/geophysical implications of this phase. Trommsdorff and Evans (1980) summarize and discuss the contrasting conclusions reached by dif- * Present address: Department of Geological Sciences, The Uni- versity of British Columbia, Vancouver, B.C. V6T lW5, Canada ferent investigators of natural phase assemblages: While the typical occurrences of 77-Cl in kimberlite xenoliths suggest a stability field for that phase at upper mantle conditions, a much lower thermal stabil- ity is inferred from the observed breakdown of Ti- C1 to olivine+ilmenite in antigorite schist. Tromms- dorff and Evans suggest that partial substitution of F for OH may be responsible for the stabilization to higher temperatures. Recent crystal structure refinements lend inde- pendent support to this idea. Fujino and Takeuchi (1978) indicate that substitution of F for OH in the brucite-layers of the clinohumite structure reduces the hydrogen-hydrogen repulsion (due to the short H -H bonds: 0.87 ,~), thus stabilizing the structure. It is noted, however, that the substitution of Ti for H2M (M =Mg, Fe 2+) has essentially the same effect. Most recently, Ribbe (1979) explored further and more quantitatively the structural aspects of these sub- stitutions of F and Ti in the homologous humite series; his conclusions generally support the con- tentions cited above. Furthermore, Robinson et al. (1973) observc a slight Fe-enrichment of the M3-site in 77-C/, which contrasts with the Fe-depletion of M3 in Ti-free clinohumitcs. The energetic effects of this change in ordering characteristics with com- position are not a priori predictable. Based on available crystal-chemical data, all three cxchange componcnts operating in Ti-C1 (i.e., OHF_I, H2MgTi_I, FeMg_l) are thus potentially important in determining its stability. By analogy to other silicates, e.g., biotite and amphibole, OHF 1 may bc expected to be gcologically most important, however. The marked effect of that exchange com- ponent on the thermodynamic properties of several Ti-free humite group minerals has recently been de- termined accurately by analysis of phase-equilibrium data (Duffy and Greenwood 1979). Clinohumite, chondrodite, and norbergite turn out to have identi- 0010-7999/80/0072/0415/$02.00