Partial melting of basaltic rocks

[Hufangyang]

Hello Paula,

I have been trying to make equilibrium melting calculations from basaltic rocks to see whether I can produce andesitic-granitic melts that I expect. However, I have trouble in deciding which software to use and how to use it. Now I have download and install rhyolite-Melts, but I don't know which option to select in Option Menu. I have read the instruction and all of them are about crystallization and partial melting of mantle rocks. So is there a kind of Melts can solve my problem? And how can I use it?

You can also contact me via my email: hufangyang@gmail.com

Thanks a lot
Fangyang

[Paula]

Hi Fangyang,

The rhyolite-MELTS GUI should work fine for your calculations. The GUI has a superliquidus initial guess (alphaMELTS has the option to do a subsolidus initial guess) so it is simplest to do crystallization runs. However, as you want to do equilibrium/batch melting, this is just the reverse of batch crystallization. I suggest you try the tutorial for rhyolite-MELTS on this forum:

magmasource.caltech.edu/forum/index.php/topic,133.0.html

If that works then you can try modifying the Morb.melts file to your basalt composition. Change the Initial Pressure to something suitable and remove the "Fractionate: solids" line. You can double check that no phases will be fractionated by clicking the Options menu, after reading in the .melts file. Note that if your basalt composition includes H2O then you should not try the Find Liquidus part of the procedure - just start at a temperature high enough that the basalt is (partially) molten.

Hope that helps but please get back to me if there are more queries.

Best,
Paula

[Hufangyang]

Thank you very much for your prompt reply, Paula.

I will try this method very soon and to see the results. I will contact you if there are more queries.

Thank you again for your help!

Fangyang

[Hufangyang]

Hello Paula,

I have run the program under various pressures and get useful results. However, I have three questions about the results that I obtained.

First, this is actually a crystallization process, so the temperature here has no actual meaning to the partial melting, right?

Second, I added 0.5-2.0% H2O to the initial composition of the basalt, and got biotite out of the system, but no amphibole or hornblende out of the system which I expected. Is this related to the CaO content or some other factors?

Third, the TiO2 content are very high for a felsic melt, and the TiO2 content increased at first and then decreased resulting in usually >1 wt.% for melts with SiO2 >65 wt.%. In addition, I got lower MgO, FeO, Fe2O3 compared to experiment melts. Moreover, the H2O content increased from the start and reached a relatively high level (2-6 wt%). The Al2O3, CaO, Na2O and K2O contents are normal. Is these results related to my initial composition or other factors?

I would be very grateful if you can provide any clues to these questions.

Thank you for your help!

Fangyang

[Paula]

Hi Fangyang,

Sorry for the delay in replying. It has been busy in the run up to the AGU meeting.

First, this is actually a crystallization process, so the temperature here has no actual meaning to the partial melting, right?

As long as there is no fractionation of liquid or solid phases, it does not matter whether you run up temperature or down temperature. So the temperature will have exactly the same meaning for crystallization as for partial melting. However, it is worth bearing in mind that there is a systematic offset between the MELTS-calculated solidus and liquidus temperatures and the experiments (for example, this AGU abstract talks about the offset for peridotite melting at high pressure - I know that is not exactly what you are interested in, but it gives the idea...). It is often better to compare calculations and experiments for a similar melt fraction rather than the same temperature.

Second, I added 0.5-2.0% H2O to the initial composition of the basalt, and got biotite out of the system, but no amphibole or hornblende out of the system which I expected. Is this related to the CaO content or some other factors?

The amphibole and hornblende models in MELTS are not that great and it can be difficult to stabilise them even when they should be present. Rhyolite-MELTS is not calibrated for intermediate compositions where amphibole is important (for more info see the FAQ on the MELTS website). There are also problems with the biotite model, as detailed in the alphaMELTS software review.

Without actually seeing your starting composition or knowing which experiments you are comparing the results with, it is a bit difficult to speculate what is going on. Some details of the calculated compositions could be artefacts of the Rhyolite-MELTS model, and in particular the range of mineral phases available in the model. Some minor components (e.g. TiO₂ or H₂O) only go into a few solid phases and this can mean that the modelled liquid ends up too rich in these components or it can artificially stabilise a particular solid phase (e.g. Cr in spinel - again not relevant to your composition, but just given as an example). If you want to send me the composition I can have a go at running it.

Paula

[Hufangyang]

Hi Paula, thanks for your reply! I have a better understanding of this now.

It will be great if you have time running it for a try.

Here is the initial composition:
SiO2: 49.8
TiO2: 2.39
Al2O3: 16.10
Fe2O3: 1.57
FeO: 8.91
MnO: 0.23
MgO: 6.11
CaO:  7.55
Na2O: 2.53
K2O: 1.26
P2O5: 0.32
H2O: 0.50

It should be note that the FeO and Fe2O3 contents are estimated based on the formula (TFe2O3=Fe2O3+1.1111*FeO) for the reason that I have only obtained TFe2O3 content.

The pressure of 10-20 kbar would be OK and the FMQ could be set as -1.

Thank you for your precious time and help!

Fangyang