I am doing non-spin polarized doping calculation in La1-xEuxVO4. Eu having less ionic radii than La so after doping volume should shrink and lattice constants should be decrease. Experimentally after doping it is showing decrease lattice constants than pure LaVO4 but from DFT I am getting increased lattice constants and increased relaxed optimized volume after doping.
Thank you Dr. Mohamed for reply !
I got the answer. I was doing non-spin polarized calculation since Eu having magnetic moment when I did spin -polarized calculation with proper Eu magnetic moment then I got the right trend similar to experiment.
Thank you all for nice suggestions.
I will try to implement these .
Mr. Sharma,
Are you sure, generally, that you have performed DFT calculations in predicting the volumes of your systems; which theoretical methods you have utilized; have you preliminary geometry optimization to whole systems or you have used any initial crystallographic coordinates; and can you provide information about the approximation for computing of the volume, particularly?
Because of such as tendency is obtained, using numerical integration techniques, mainly employed in QSAR, where the volume is calculated by cubic grids taking into consideration van der Waals radii of a hard sphere. The value of 3D grid is explicitly fixed by mentioned radii as boundary condition, computing V. You could pay attention to shown equation in the attachment and [ref. 2].
However there are not taken into consideration the complex nature of the bonds of f-elements, particularly those of 5f-elements, and any charge transfer effects. For this reason you have a simple linear regression relationship where with increasing in the number of Eu atoms an increasing in the volume has occurred.
Please pay attention to attached data where there are computed volumes of LaVO4, using the crystallographic coordinates given in [ref. 1] and the corresponding Eu1-Eu4 derivative systems. The relationship number of Eu-atoms versus volume is linear at a level of confidence 99.982%. As you could expect replacing Eu with Tb, U, Am, Ac, Bk, or Lr in linear relationships with opposite slopes have been obtained with a confidence levels > 99.9 %, where particularly for Ac and Lr the slope is the same, while for Am and Bk it is slightly different, meaning that from a chemical point of view are you shall interact Am or Bk, would be the same according this model - but it is wrong according experimental ones about the chemical reactivity, including known data about their chemical compositions!
If you pay attention to our papers [refs. 3 and 4] on complexes of d- and f-elements, there are obtained data that, from about 150 computed cases only in 3 complexes of Zn2+ the electronic configuration of the metal ion corresponds to the classical concepts for d10 ion according (ref. 3, Table 2) to the crystal field theory, where the atoms/ions are described as spheres as well as accounting only electrostatic interactions between the ions/atoms in the complexes. Given that only to those 3 cases the discussed numerical integration approach would be applicable, however, only approximating the Zn2+ ion.
Remains the questions posted above, thereby, about your approximation computing the volume; and additional questions, have you any NBO data about the state of Eus atoms in the unit cell as well as are all Eus and Las are equivalent as coordination environment? Because of in the shown LaVO4 crystal [ref. 1], the different Las are not equivalent from a coordination chemical point of view. They they have different geometry of the inner coordination sphere. Is your case is the same or all f-elements/ions are equivalent as coordination environment?
It would be also interesting if you have replaced Eu with U or Am are you shall obtain the attached relationships. Most probably yes!
[Ref. 3] Gas-phase CT-stabilized Ag(I) and Zn(II) metal-organic complexes - Experimental versus theoretical study, Polyhedron, 30, 2011, 2564-2573, M. Lamshoeft, J. Storp, B. Ivanova, M. Spiteller
[Ref. 4] Adsorption of uranium composites onto saltrock oxides - experimental and theoretical study, B. Ivanova, M. Spiteller, Journal of Environmental Radioactivity, 135, 2014, 75-83
I, and another 10000 scientists, dissent with Henk Smid.
DFT is based on solid mathematical foundations as the original work of W. Kohn has shown. However, as with many mathematical results to make it useful you need to introduce an approximation (and Physics is built on approximations) for the exchange and correlation energy. Also, if you go beyond the ground state properties, then you move away from what is the realm of the theory and results become more qualitative than quantitative. Stating that DFT plainly "does not work" is neglecting the last 20 years of research (and successes) in Material Science.
Thank you Dr. Bojidarka B. Ivanova for very knowledgeable discussion and suggestions.
I will go through all suggestions and references.
Ref 1 and 2 are not visible from attachment. Please tell me ref.1 and 2. again.
Mr. Smid,
What represent the so-called "true" for you, reflecting in mind a quantum state?! - The mixed quantum state is a probabilistic function of pure states or a math function expressing the maximum probability to find e in a 3D space; while the pure state is in fact a vector, identified with a quantum number within the frame of an electron in hydrogen atom for example, which is expressed by the well known "atomistic" model. Therefore you are talking for math approximation within the frame of a model! Given that your 'trueness' is model-approximation, in fact.
In this respect, the object of all those discussions is among various approximations/models which set expresses maximum realistic the real state of the matter.
Particularly the discussed approximation of volume cannot express realistic the existing chemical data, data from chemical analysis, therefore this model/approximation suffers of drawbacks.
Among the advantages of DFTs are the usage of experimental parametrization. On this base we can predict other systems of interest at a high level of accuracy. Therefore DFTs have involved themselves experimental set of data.
DFTs have also balanced between accuracy, efficiency, computational times and cost. It is well known that those methods use scaling factors, and in this topic largest part of information about accuracy of different DFTs is well know and taken into consideration to, in fact, all publishing literature data using DFTs.
If you have any new theoretical approximation/model balancing above stated factors making each theory applicable to the real chemical experimental observations, meaning to the real world, nature and s.o. than we could discuss it/them.