MTDATA – Phase Diagram Software from the National Physical Laboratory







Training and Support 




Other Services 





Databases, Definitions and Models

Three main types of database/datafile are used in MTDATA. The foundation is provided by the database source files which are ascii text files identifiable by the preferred extensions .loa. or .dbl. They are used as the source for generating the databases themselves each of which comprises a binary database and an index designed for fast retrieval of data. These have the extensions .dbs and .inx respectively. Finally, when data are retrieved for a system they are transferred to another ascii file which by default is given the extension .mpi. With limited exceptions only the source files are of interest to the user, since it is to these files that all calculations are ultimately traceable. Full details are given elsewhere. It is possible to edit the .mpi files but this is not recommended. Users can amend and add to their own databases (.inx and .dbs files) for pure substances using the THERMOTAB module. For other purposes the UTILITY module must be used.

MTDATA generates a substantial number of output and working files. Information on these relevant to their use and management is given elsewhere.

Elements, charge and vacancies

Elements are represented by their normal chemical symbols using upper and lower case.


The electron is represented by /- and positive charge by /+. The use of two different symbols for charge avoids the need for negative stoichiometry numbers.


Vacancies are represented by Va. They do not need to be included in lists of components when systems are defined.

Substances of fixed composition

For the purpose of developing new data substances must conform to the following rules. The following explanations should be read in detail particularly by those concerned with the data storage or amendment.

  • Up to eight elements are allowed in one substance.

  • Normal chemical conventions are observed for the representations of symbols and formulae.

  • The phase of each substance is described by appending at least one symbol to the formula inside pointed brackets <>. The allowed state symbols are:
    <c>    condensed, ie liquid or solid, the default option and normally omitted,
    <aq>  aqueous species,
    <g>    gaseous species (<g> is expanded to <gas> by ACCESS, MULTIPHASE etc but not by THERMOTAB).

All constituents of solution phases and some pure substances need to be further defined by a more definite phase label, for example:

Carbon   C<graphite>, C<diamond>
Iron       Fe<FCC_A1>,  Fe<BCC_A2>

If a pure substance other than a gaseous or aqueous species has no phase label one is generated by capitalisation of the formula name on data retrieval, eg MnO2 becomes MNO2. This does not apply to THERMOTAB.

Isomers are distinguished by a different convention in which an isomer identifier is tagged to the formula by an underscore, eg: C2H2Cl2_1,1<gas>, C2H2Cl2_trans<gas>, C2H2Cl2_cis<gas>. In THERMOTAB or in files for loading data in to databases a different convention is used in which the isomer name is contained in the phase label, eg<g,cis>.

Charged species may be represented by using the symbols for negative and positive charge exactly as if they were element symbols, ie symbol before stoichiometry number.

Examples of substance names are:


In general, use of generic names for phases is recommended. Thus in the last of these examples ’melilite’  is preferred to  ’gehlenite’. On the other hand, if no solution data are available for mixtures with other melilites, ’gehlenite’ is the better name to use.


Unary data are required for end members of solution phases. For this reason it is convenient to use the word unary as a noun to describe a constituent of a solution phase. A unary may be a normal stoichiometric substance that is stable in the pure state: Fe<bcc> and SiO2<alpha_quartz> are examples. In other cases it may be essentially equivalent but unstable in the pure state, for example Cr<fcc>. In addition to these self-evidently necessary unariesit is sometimes necessary to add hypothetical associated species that are invoked in order to model the properties of solution phases in which there is strong interaction between the components. Phases for which there is explicit recognition of sublattice structure also need to be considered and these require an extension of the formalism.


Colons are used to separate the sublattices. The formula for normal spinel MgAl2O4<spinel> is written:

Mg/+2:Al/+3:O/-2<spinel:1:2:4>   normal spinel

but because real spinel is partly inverse data will also be required for the notional substances:

Mg/+2:Mg/+2:O/-2<spinel:1:2:4>    charge=-2
Al/+3:Al/+3:O/-2<spinel:1:2:4>       charge=+1
Al/+3:Mg/+2:O/-2<spinel:1:2:4>     charge=-1

A combination of the last two gives inverse spinel. The four substances are the stoichiometric formulae for the end member constituents, unaries, of a real spinel of overal lstoichiometric composition. In order to model the properties of spinels containing more or less MgO or Al2O3 than the 1:2 ratio a third sublattice for cations and the possibility of vacancies on cation sublattices may be introduced. These are also required for oxygen deficient or oxygen excessspinels.

In the above examples the species Mg/+2 on the second sublattice is identified in MULTIPHASE as Mg/+2:2<spinel>.This symbolism is used in the data files generated by ACCESS. There is no facility for identifying the individual sublattices by their structural position in the crystal lattice so care needs to be taken to maintain consistency in the order of assigning data to the various combinations.

As seen from the examples above the unaries necessary to model certain phases may carry a net charge and may have vacancies on one or more sublattices. The user should refer to ACCESS and UTILITY pages for further information.

Interactions in solution

The user does not normally have to write down "formulae" identifying particular interactions in solution. The solution data are retrieved automatically by ACCESS. Unlike the unaries they are not identified in the list of substances and of course they are not assigned any mass in the results of calculation.

Thus the user needs to be aware of the existence of interactions only in the development of new binary or higher order interaction data. Reference should be made to the UTILITY and ACCESS pages for further information. Following normal convention, mixing in a phase as a whole or on an individual sublattice is indicated by commas separating the species that are mixing. Sublattices are separated by colons in the same way as for unaries. Mixing generally occurson only one sublattice at a time. Examples are:


The first example is a simple solution between Fe and Ni in the liquid phase. The second relates to mixing between the last two of the unaries of spinel given above. The third example refers to interactions between ferrous ions and vacancies in wüstite, which is given the generic phase name ’halite’. It would also be necessary to consider interactions between ferrous and ferric ions and between ferric ions and vacancies.

The generation of data for loading in to a database needs more detailed information than is given here. Reference should be made to the UTILITY pages. More general information on terminology is given in the Glossary.

Thermodynamic models used in MTDATA

The temperature and pressure dependence of the Gibbs energy and interactions between the constituents of solution phases are described by mathematical models. Those used in MTDATA are well established in the scientific literature and are supported to varying degrees by a body of data. They are incorporated in a modular way into the software anddata structures. The models include:

  • A system for describing temperature dependence of either Gibbs energy or heat capacity, that has the flexibility to deal with all the equations normally used forrepresenting the data for pure substances and their interaction in solution.

  • The Murnaghan model for the thermodynamic data as a function of pressure.

  • The Inden model formagnetic contributions.

  • The compound energy model allowing solution on the individual sublattices of crystalline compounds.

  • The associated solution model for liquids in which there is strong bonding between the components, the bonding having a significant covalent character.

  • The two-sublattice ionic liquid model used for molten salts and other mainly ionic systems.

  • The extended Redlich-Kister model for predicting the multicomponent thermodynamic data of non-ideal solutions including those on individual sublattices.

  • An extended Kapoor-Frohberg model for slags.

  • The Quasi-chemical model

  • The Pitzer model for aqueous solutions

Updated 26 May 2010