Example datasets

Parent Previous Next

Here are brief descriptions of  the example datasets that are distributed with the programs, along with an explanation of the format used.



Dataset name: PV_quartz.dat



Title: Quartz


Comment: 

Data are from Angel et al (1997) J. Appl. Crystallography.


This is a simple P-V dataset of  the compression of a single phase, without phase transitions. The data are of high quality and include a determination of V at room pressure. Therefore there are no problems at all in fitting this data. So, if you have 'problems' with the program, come back and test it with this dataset.


System: It is not specified in the file, because only volume data are given. If you put in the crystal system, that is ok, but it would not be used because there are no cell parameters.


Tscale: Not specified because this is an isothermal dataset.

Format: The data include just P, V and their estimated uncertainties.




Dataset name: Pcell_hexagonal.dat



Title: Kalsilite High P Data from Gatta et al. (2011) Am Min 96:1363


Comment: This is unit-cell data, no volume data, from a single-crystal high-pressure study of kalsilite. The material undergoes a first-order phase transition just above 3.6 GPa. 


System: Hexagonal. The program can then calculate the unit-cell volume from the given unit-cell parameters for each data point by assuming you have used the conventional cell setting for the hexagonal setting (i.e. c-unique)


Tscale: Not specified because this is an isothermal dataset.

Format: The data include just P, a, c and their estimated uncertainties.




Dataset name: Pcell_triclinic.dat

Title: Albite


Comment: 

Data are from Benusa et al (2005) American Mineralogist.


This illustrates how to load the complete unit-cell parameters in one file. 


The data are of high quality and include a determination of V at room pressure. The data are interesting in that they exhibit volume softening at high pressures, so you will need a 4th-order EoS to obtain good fits. In addition, the compression is very anisotropic. The softest direction is close to the (100) plane normal; compare a fit to this direction to, for example, the cell axes b or c.


System: It is not specified in the file, because it is triclinic.


Tscale: Not specified because this is an isothermal dataset.

Format: The data include all cell parameters and their estimated uncertainties.




Dataset name: rutile.dat


Title: Data are taken from Zaffiro et al (2019)  Mineralogical Magazine, 


Comment: These are the data selected from the literature by Zaffiro et al. (2019) on the basis of mutual consistency. Full details of the data selection are in the paper.


System: tetragonal has been specified.


Tscale: specifies that the current temperature reported in the dataset is in  K. 


Vscale: because the various sources of the data have different values of the cell volume at 298K, each set of cell parameters has been scaled to its own volume at 298K, and listed as V/V(298K). This allows the data to be fitted together. So Vscale is given as V/V0


Lscale: This is for the two cell parameters, a and c, and is given as L/L0 because the same rescaling as for volume was applied.


Format: The high temperature dataset contains unit cell parameters and volume as a function of temperature, therefore in the format line is specified that the data columns will contain in the order temperature, esd(T), P, esd(P), V, esd(V), a lattice, esd(a)  c lattice and esd(c).




Dataset name: T_triclinic.dat


Title: An27 plagioclase


Comment: 

Data are from Tribaudino et al (2011) American Mineralogist.


These are just V-T data, that extend from 90 to about 900 K. The low-temperature regime exhibits a clear reduction in the thermal expansion, so you need to use either a Salje equation or the thermal pressure model to fit the data.


System: It is not specified in the file, because it is triclinic.


Tscale: Not specified, because the temperatures are given in K

Format: The data includes just T and V, without estimated uncertainties. This example also shows that you can separate the data with spaces; commas are not required.




Dataset names: spessartine_Vmolar.dat and spessartine_Vcell.dat

Title: Data are taken from Greaux & Yamada (2014)


Comment: Two versions of this data are included as example datafiles. One has the volumes given as unit cell volumes, and the other has the volumes converted to molar volumes (cm3/mol). Only the file with molar volumes can be used to fit the MGD EoS, but both can be used to fit the Holland-Powell thermal pressure EoS and isothermal EoS. Fits of identical EoS to the two datasets should give you identical parameters, apart from the value of Vo. 




System: cubic has been specified.



Format: The dataset shown is for molar volumes. It was generated from the original data with the edit utility of EosFit7c which then writes the data to a fixed number of decimal places in columns separated by spaces.





Dataset name: Akermanite.dat

Title: Akermanite McConnell et al (2000)


Comment: This is unit-cell data, from a single-crystal high-pressure study of the mineral akermanite. The material undergoes a continuous phase transition at about 1.72 GPa. The low-pressure phase is the low-symmetry phase.


System: Tetragonal. The program will use the input unit-cell parameters and volume.


Tscale: Not specified because this is an isothermal dataset.

Format: The data include P, a, c, and V and their estimated uncertainties.


Dataset name: quartz_moduli.dat


This file contains the bulk and linear adiabatic moduli recalculated from the cij values reported by Wang et al. (2015). This illustrates all of the features of an input data file with moduli for EosFit.


Rules:


1) You cannot mix different types of data in one file. If you want to do simultaneous fits (e.g. to V and K) you have to load these from two different .dat files.


2) If you have a data file with moduli, you must use the datatype to specify the data are moduli


3) On the datatype line, you must also write isothermal or adiabatic. See here for more information.


4) The format line is the same as for volume and cell data. Label a column which contains K, with V. A column of moduli for the a axis is labeled a.


In this example we have used comment lines (with #) to label the columns with more details.