liquidus temperature hi mg basalt

[roman@ipgp.fr]

Dear all,
I am trying to simulate a simple equilibrium crystallization of the following hi mg basalts at 10kbar without any constrained buffer. The input file is attached below.
However, I get very different liquidus temperatures if I use pmelts, rhyolite melts, and petrolog.

Tl = 1357 °C with pmelts
Tl = 1446 °C with rhyolite melts
Tl=1260°C with petrolog

Moreover, even if I reduce pressure up to 5kbar the liquidus temperature with rhyolite melts is always pretty high (>1330 °C) and the phase in the liquid is always orthopyroxene, while I'd expect olivine. Does anybody know from where these differences come?
Thank you!
alberto


Title: dummy
Initial Composition: SiO2 51.57
Initial Composition: TiO2 0.78
Initial Composition: Al2O3 13.02
Initial Composition: Fe2O3 1.02
Initial Composition: FeO 7.62
Initial Composition: MnO 0.17
Initial Composition: MgO 11.96
Initial Composition: CaO 10.08
Initial Composition: Na2O 2.33
Initial Composition: K2O 0.55
Initial Composition: P2O5 0.13
Initial Temperature: 1200.00
Final Temperature: 800.00
Increment Temperature: 1.00
Initial Pressure: 10000
Final Pressure: 10000
Increment Pressure: 0.00
dp/dt: 0.00
log fo2 Path: None

[Paula]

Hi Alberto,

Sorry for the delay in replying. A few thoughts:

• There is a known systematic offset of ~100 ^oC in T versus F for (Rhyolite-)MELTS. Near the peridtotie solidus, pMELTS overestimates T(F) by about 60 ^oC so that give you a rough estimate of the offset for the calculated basalt liquidus. See Ghiorso et al. (2002) for more details.
• See Hirschmann et al. (1999) for a discussion of how MELTS overstabilises opx at the expense of olivine. Note that the Rhyolite-MELTS and MELTS models behave identically when quartz and Ksp are absent, as here.
• Although, you don't have Cr2O3 in this bulk composition, some of Paul's comments on your other Find Liquidus post do still apply here.
• Also note that pMELTS is not calibrated for MnO or P2O5 so those should not be included in this comparison.
• You might consider putting K and P as trace species too (this shouldn't affect the liquidus temperature, it may affect whether you can completely crystallize the liquid).

Cheers,
Paula