– for DFT formalism, the type of exchange–correlation functional
LDA functionals (LDA, LSDA, LDA + U, LSDA + U, …)
GGA functionals (PW91, PBE, PBESol, BLYP, WC 2006, …),
Meta-GGA functionals (M06-L, TPSS, SSB-D, MVS, …),
Hybrid functionals (B3LYP, PBE0, HSE, …),
Non-local functionals (ADA, WDA, …).
• Function basis set used to represent molecular or crystal orbitals: (PW, (L)APW,
ASW, (L)ASO, (L)MTO, localized atomic basis functions). In case of the latter,
also the appropriate basis set (minimal, split valence, Pople style, polarized, diffused, correlation consistent, polarization consistent, etc.) has to be additionally
chosen—this choice strongly influences the time of calculations, which scales as
N
X , (x > 3, N—number of basis functions for various QM theory levels);
The choice of level theory is a crucial step in every calculations, and it has to be
done thoughtfully after careful consideration supported by literature reports on
similar systems, before the “production” calculations start. In many cases, due to the
lack of sufficient information on similar system studied from given properties
viewpoint, it is necessary to run preliminary calculations to properly determine
required theory level for given formalism to ensure the ability to simulate the
properties of the system under study in acceptable time, using reasonable computer
resources and simultaneously to obtain useful, physically meaningful results. It
should be, however, mentioned here that in many cases the available options to
choose from are strongly limited by the prior choice of the computer software, which
will be used for calculations, since in various programs only limited number of listed
above options is implemented (currently a huge number of quantum chemistry and
solid-state physics programs, both commercial and free is available [144]).
As final remark, it is worth to mention that whatever theory level is chosen for
practical calculations, when a series of calculations is carried out for a group of
model structures, in order to analyze the influence of the chemical composition,
defects, local order, etc, on the physic-chemical properties of interest, it is a must to
use for all calculations the same theory level and convergence criteria, since
otherwise, one will get the results which simply cannot be sensibly compared and
the whole endeavor of in silico experiment will simply end up in wasted time and
computer resources.
1.4 Conclusions
The use of theoretical methods as tools supporting experimental research is associated with many crucial limitations resulting, on the one hand, from the limitations
of the methods themselves, and on the other from the complexity of the materials
studied. In the case of spectroscopy investigating the interaction of electromagnetic
radiation with matter, a number of practically available theoretical methods that
1 Computational Methods in Spectroscopy
41
LDA functionals (LDA, LSDA, LDA + U, LSDA + U, …)
GGA functionals (PW91, PBE, PBESol, BLYP, WC 2006, …),
Meta-GGA functionals (M06-L, TPSS, SSB-D, MVS, …),
Hybrid functionals (B3LYP, PBE0, HSE, …),
Non-local functionals (ADA, WDA, …).
• Function basis set used to represent molecular or crystal orbitals: (PW, (L)APW,
ASW, (L)ASO, (L)MTO, localized atomic basis functions). In case of the latter,
also the appropriate basis set (minimal, split valence, Pople style, polarized, diffused, correlation consistent, polarization consistent, etc.) has to be additionally
chosen—this choice strongly influences the time of calculations, which scales as
N
X , (x > 3, N—number of basis functions for various QM theory levels);
The choice of level theory is a crucial step in every calculations, and it has to be
done thoughtfully after careful consideration supported by literature reports on
similar systems, before the “production” calculations start. In many cases, due to the
lack of sufficient information on similar system studied from given properties
viewpoint, it is necessary to run preliminary calculations to properly determine
required theory level for given formalism to ensure the ability to simulate the
properties of the system under study in acceptable time, using reasonable computer
resources and simultaneously to obtain useful, physically meaningful results. It
should be, however, mentioned here that in many cases the available options to
choose from are strongly limited by the prior choice of the computer software, which
will be used for calculations, since in various programs only limited number of listed
above options is implemented (currently a huge number of quantum chemistry and
solid-state physics programs, both commercial and free is available [144]).
As final remark, it is worth to mention that whatever theory level is chosen for
practical calculations, when a series of calculations is carried out for a group of
model structures, in order to analyze the influence of the chemical composition,
defects, local order, etc, on the physic-chemical properties of interest, it is a must to
use for all calculations the same theory level and convergence criteria, since
otherwise, one will get the results which simply cannot be sensibly compared and
the whole endeavor of in silico experiment will simply end up in wasted time and
computer resources.
1.4 Conclusions
The use of theoretical methods as tools supporting experimental research is associated with many crucial limitations resulting, on the one hand, from the limitations
of the methods themselves, and on the other from the complexity of the materials
studied. In the case of spectroscopy investigating the interaction of electromagnetic
radiation with matter, a number of practically available theoretical methods that
1 Computational Methods in Spectroscopy
41