Experimental melting curve of iron revisited

With new experimental data presented in the last 2 years, it becomes possible to resolve conflicts in the data sets used in constructing the melting curve of iron, T-m(P). On the basis of these new data, several data sets have been dropped: the Williams et al. [1987] melting curve up to 100 GPa and the Bass et al. [1987] and Yoo et al. [1993] shock-wave-derived T-m(P) in the 200-300 GPa range based on light emissivity measurements. The Boehler [1993] T-m(P) curve to 200 GPa ana the Brown and McQueen [1986] shock-wave-determined T-m(240) remain, leaving a gap between 240 and 330 GPa. We fill this gap using the Lindemann law of melting. The Lindemann law and the temperature values along the Brown and McQueen [1986] Hugoniot require the value of the Gruneisen ratio, gamma; thus gamma connects T-m at 330 GPa with T-m found for values of the Hugoniot. It is further shown that the heat of crystallization, Delta H-m, is dependent on gamma. Thus, through gamma, a connection is made between the melting curve and the power generated within the inner core. The effect of all these connections of physical properties through gamma leads us to recommend 5600-6500 K as the T-m of iron at inner-outer core boundary conditions.