The NewMag crystal-field code for f-element systems: Implementation for extended active spaces, second-order correlated energies, and generalisation to fn configurations
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.
Kai: We're looking at the paper: "The NewMag crystal-field code for f-element systems".
Mira: Here's the gist of the paper. The NewMag crystal-field code for f-element systems enables the extraction of Stevens crystal-field parameters after relativistic and multiconfigurational calculations, which is important because it provides a tool to understand the magnetic and optical properties of f-element systems.
Kai: Code Capabilities The code can post-treat ORCA and OpenMolcas outputs that contain state-average complete active space self-consistend field (SA-CASSCF) calculations with minimal or extended active spaces, second-order NEVPT2 or CASPT2 calculations, and spin-orbit configuration interaction (SOCI) calculations.
Mira: Stevens parameters are extracted following the original Stevens convention with the so-called “extended” parameters.
Kai: Rotation invariant indicators are also computed, after applying a correction of the parameter values to ensure normalization, following Rudowicz.
Title and authors: Kai: We started by looking at this paper titled "The NewMag crystal-field code for f-element systems: Implementation for extended active spaces, second-order correlated energies, and generalisation to fn configurations". The title itself tells us immediately that they’re expanding the NewMag program.
Mira: It signals an update on the existing code that focuses on extending its capabilities to handle more complex theoretical inputs from programs like ORCA and OpenMolcas.
Lev: So, what does this extension actually mean in terms of computation? Are we talking about a massive jump in required memory or just a change in how the calculation is structured?
Kai: It’s not necessarily a massive jump, but they are post-treating outputs that already have state average complete active space self consistent field calculations with minimal or extended active spaces.
Mira: Minimal or extended active spaces means they can take results from calculations that might be less demanding than the full calculation, which is smart for accessibility.
Lev: If you’re running this on a real quantum computer, having a way to ingest these already calculated intermediates saves you from doing the most expensive part of the job from scratch.
Kai: The paper emphasizes that they are including second order correlated energies like NEVPT2 or CASPT2 in their post-treatment capabilities, which is important for getting more accurate results.
Mira: Including those second order methods means they are aiming for higher precision than just using the single point estimates from the first level of calculation.
Lev: Higher precision usually means more time spent on correlation effects, and we have to be careful about how that scales with the size of these active spaces.
Kai: The paper also mentions spin-orbit configuration interaction SOC calculations as part of what this code can post-treat, which is a key ingredient for magnetic studies.
Mira: That SOC calculation is where things get interesting, because spin-orbit coupling is often the dominant factor in determining the magnetic anisotropy in these systems.
Lev: And from an engineering perspective, if you need to model SOC effects accurately on hardware, you really need to be careful about how those terms are scaled up when you move from simpler approximations to the full interaction.
Kai: So, they’re essentially creating a pipeline that takes existing high-level calculations and refines them for magnetic property extraction.
Mira: It sounds like they are trying to make the analysis part of this process more rigorous, rather than just relying on empirical fitting models which I see in some older literature.
Lev: That’s where the value is; if you can post-treat those outputs, you’re not just getting a number; you’re validating the entire theoretical structure used to get that number.
Kai: The paper sets up this new framework for extracting Stevens crystal field parameters after relativistic and multiconfigurational calculations are performed Gwenhaël Duplaix-Rata,one Dumitru-Claudiu Sergentu,two three Boris Le Guennic,one and Rémi Maurice*one one)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR six thousand two hundred twenty-six F-thirty-five thousand Rennes France <ref:2609.37541#pg1,Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic>.
Mira: And the authors are clearly building on previous work by NewMag but adding significant new features to make it more powerful for modern computational chemistry.
Lev: It’s interesting that they're focusing on f n configurations where n is not zero, not seven or fourteen, which suggests they are targeting a wider range of systems than just the most common ones.
The paper's summary: Kai: Now moving into the summary of this paper, it’s about how NewMag works and what it actually achieves with all these new features. They want to show that this code is a viable tool for extracting crystal field parameters from these complex ab initio calculations.
Mira: The summary explains that they can extract Stevens parameters following the original convention with the so called extended parameters, plus rotation invariant indicators after applying a correction by Rudowicz to ensure normalization.
Lev: Normalization corrections always sound like necessary evils, but you have to ask what those corrections actually are and how much they change the final derived values.
Kai: They also detail how they compute rotation invariant indicators after applying a correction of the parameter values to ensure normalization following Rudowicz, which is an important step for comparison.
Mira: And at the SOCI level, they extract the spin-orbit coupling constant within a spherical approximation of that SOC operator, which is another key magnetic property.
Lev: Extracting that SOC constant through a specific approximation means you have to be aware of the assumptions made about the SOC operator structure; it’s not always exact.
Kai: They describe their methodology for building the effective Hamiltonian at the SR level in the basis of ML components of the ground two S plus one L term of the reference free ion, and this involves five specific steps to process ab initio wave functions.
Mira: Those five steps are where they handle everything from root selection to transforming wave functions into complex spherical harmonics and many electron orbital bases.
Lev: That sounds like a lot of transformation matrices; from an engineering standpoint, you have to ensure those transformations are numerically stable so you don't introduce huge numerical errors during the process.
Kai: For parameter extraction at the SR level, they follow the irreducible tensor operator procedure of Chibotaru and Ungur for the crystal field B q k parameters.
Mira: And at that same SOCI level, they apply the ITO procedure to extract lambda of SOCI.
Lev: So you have two different procedures—ITO for one level and a coupled basis approach for the other—which means we’re comparing results between two slightly different theoretical frameworks.
Kai: They also mention that their resulting SOC constants at the SOCI level are similar to those obtained by AILFT since NewMag uses the spherical approximation.
Mira: That similarity is a good sign, suggesting that their approach is consistent with other established methods when looking at the spin-orbit coupling effects.
Lev: Consistency across different extraction methods usually means you’re in a safer spot computationally, but we still need to check if that consistency holds true for all f n configurations they are testing.
Kai: They show that this methodology is always successful with lanthanide complexes, and they further discuss its applicability to transition metal and actinide complexes.
Mira: So the paper is demonstrating a broad applicability by showing success across those different element series, which is a significant claim if true for the transition metals too.
Lev: Showing success across f n configurations where n isn't zero, not seven or fourteen means they’re covering a lot of ground in terms of system diversity.
The paper's improvements: Kai: The paper really lays out several improvements they are making to NewMag, and these are all about expanding what the code can handle.
Mira: First big thing is the extension to f n configurations where n is not zero, not seven or fourteen, including f two through f six and f eight through f thirteen configurations with extended active spaces.
Lev: That’s covering a lot of ground in terms of element series diversity, which is exactly what we need when we are trying to model a wide variety of real-world systems.
Kai: They also improved the methodology for how they build the effective Hamiltonian at the SR level by letting users choose between retaining pseudo-L roots or providing an output from an independent self consistent field calculation.
Mira: That choice gives users flexibility in deciding how much computational effort they want to invest based on their specific needs, which is very practical.
Lev: Flexibility is good, but it means the user has to be careful about what’s actually inside those pseudo-L roots if they’re not running the full CASSCF calculation.
Kai: They also introduced a workflow that processes ORCA or OpenMolcas outputs by first localizing active space orbitals, then feeding them to subsequent CASCI calculations to identify SR wave functions through "localized" and "purified" active orbitals.
Mira: That localization step sounds like it’s an extra filtering mechanism to ensure you are working with the most relevant parts of the wavefunction before proceeding.
Lev: Filtering out irrelevant components can reduce noise, but we need to make sure that this filtering doesn't accidentally remove a crucial piece of physics.
Kai: They also have these improvements for parameter extraction, such as re-scaling extracted CFPs based on normalized versions, like B prime q k parameters defined using rank two four and six operators to ensure consistency with Rudowicz normalization.
Mira: That’s an attempt to enforce a more standardized way of presenting the results, which helps when comparing results from different theoretical sources.
Lev: Standardizing the output format is helpful for benchmarking, but we have to be sure that the re-scaling doesn't introduce systematic errors that aren't accounted for in the original theory.
Kai: They also improved how they handle J mixing due to second order CF couplings by constructing the CF Hamiltonian in the ML S basis prior to transforming it to the J basis, which naturally introduces J mixing and allows its quantification at both SR and SOCI levels.
Mira: So they’re not just extracting parameters; they are actively looking for those subtle effects like J mixing that happen due to second order coupling, which is where the physics gets really interesting.
Lev: Quantifying that J mixing is crucial because it tells you how much the simple model Hamiltonian deviates from reality, and we need to see if this quantification is meaningful for experimental comparison.
Conclusion: Kai: So to wrap up this paper on "The NewMag crystal-field code for f-element systems: Implementation for extended active spaces, second-order correlated energies, and generalisation to fn configurations". They’ve shown a lot of capabilities in terms of post-treatment.
Mira: They proved that the code can successfully extract Stevens parameters following the original convention with the so called extended parameters and rotation invariant indicators after applying a correction by Rudowicz for normalization.
Lev: So, we have to keep an eye on those normalization corrections because they are a necessary part of getting those final, comparable numbers.
Kai: And at the SOCI level, they successfully extract the spin-orbit coupling constant within a spherical approximation of that SOC operator as well.
Mira: This gives us another way to see how spin-orbit effects are influencing the magnetic behavior directly from their calculations.
Lev: It’s interesting that they managed to keep that SOC constant consistent with AILFT, which means the spherical approximation is a relatively good assumption for those specific cases.
Kai: They demonstrate broad applicability across lanthanides and discuss its applicability to transition metal and actinide complexes, even with the noted limitations in those latter cases.
Mira: It’s a solid tool for detailed analysis of magnetic properties in these f-element systems when you have the right input calculations ready to go.
Lev: If you’re looking for a tool to help map out the landscape of crystal field parameters, this code is definitely worth checking out for your next project.
Kai: We've seen how they extend NewMag to handle more f configurations and incorporate second order correlated energies into the extraction process.
Lev: That’s fair; for hardware implementation, you need to know exactly where those coupling errors land so you can design better error mitigation strategies.
Kai: Overall, this paper on "The NewMag crystal-field code for f-element systems: Implementation for extended active spaces, second-order correlated energies, and generalisation to fn configurations" provides a detailed method for extracting Stevens parameters from complex ab initio calculations Gwenhaël Duplaix-Rata,one Dumitru-Claudiu Sergentu,two Dumitru Le Guennic,one and Rémi Maurice*one one)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR six thousand two hundred twenty-six F-thirty-five thousand Rennes France <ref:2609.37541#pg1,Le Guennic,1 and Rémi Maurice*1 1)Univ Rennes, CNRS, ISCR>.
Mira: It’s a helpful tool for pushing the analysis of magnetic properties in f-element systems further than before.
Lev: We need to keep an eye on those normalization corrections because they are a necessary part of getting those final, comparable numbers.
University of Rennes CNRS ISCR Institute des Sciences Chimiques de Rennes · Faculty of Chemistry Alexandru Ioan Cuza University of Ias,i Institute for Computational Science and Research in Romania
physics.chem-ph
Submitted: 2026-09-29
Updated: 2026-11-06
Code: https://github.com/clausserg/newmag
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
Importance score: 80/100
The gist: The NewMag crystal-field code for f-element systems enables the extraction of Stevens crystal-field parameters after relativistic and multiconfigurational calculations…
Key concepts
- Stevens Crystal Field Parameters
- These are specific parameters that describe how the electric field of surrounding ions splits the energy levels of electrons in an f-element atom. NewMag extracts them using advanced calculations to accurately model magnetic properties.
- Extended Active Spaces
- This refers to using a larger set of electronic orbitals in quantum chemistry calculations than standard methods require. This allows for a more complete description of the electron correlation and better modeling of complex f-element systems.
- Spin-Orbit Coupling (SOC)
- SOC is an interaction between an electron's spin and its orbital motion around the nucleus. NewMag extracts this constant at the SOCI level, which is crucial for understanding magnetic properties in many f-element complexes.
Terminology
Summary
The NewMag crystal-field code for f-element systems enables the extraction of Stevens crystal-field parameters after relativistic and multiconfigurational calculations, which is important because it provides a tool to understand the magnetic and optical properties of f-element systems. p. 0 fr)
(Dated: 30 September 2026)
We report a massive update of the NewMag program, which enables to extract Stevens crystal-field parameters (CFPs)
after relativistic and multiconfigurational calculations are performed. The code can now post-treat ORCA and OpenMolcas outputs that contain state-average complete active space self-consistend field (SA-CASSCF) calculations with
minimal or extended active spaces, second-order NEVPT2 or CASPT2 calculations, and spin-orbit configuration interaction (SOCI) calculations.Where the paper says it
Code Capabilities
The code can post-treat ORCA and OpenMolcas outputs that contain state-average complete active space self-consistend field (SA-CASSCF) calculations with minimal or extended active spaces, second-order NEVPT2 or CASPT2 calculations, and spin-orbit configuration interaction (SOCI) calculations. Stevens parameters are extracted following the original Stevens convention with the so-called “extended” parameters. p. 0 The code can now post-treat ORCA and OpenMolcas outputs that contain state-average complete active space self-consistend field (SA-CASSCF) calculations with
minimal or extended active spaces, second-order NEVPT2 or CASPT2 calculations, and spin-orbit configuration interaction (SOCI) calculations. Stevens parameters are extracted following the original Stevens convention with the
so-called “extended” parameters. Stevens parameters are extracted following the original Stevens convention with the
so-called “extended” parameters. Rotation invariant indicators are also computed, after applying a correction of the
parameter values to ensure normalization, following Rudowicz. Rotation invariant indicators are also computed, after applying a correction of the
parameter values to ensure normalization, following Rudowicz. At the SOCI level, the spin-orbit coupling (SOC)
constant is also extracted within a spherical approximation of the SOC operator. For all the reference calculation levels, the model spectrum is reconstructed, allowing a direct assessment of its quality. The reported implementation is
successfully applicable to f-element systems with a non-void, non-half-filled/empty or non-full f shell, that is for f
n
configurations with n ̸= 0, ̸= 7 and ̸= 14.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Methodology and Model Spaces
The model spaces considered include the pseudo-L approximation where the number of roots depends on the configuration. With NewMag, since it works on the basis of all the possible J’s that can be formed from the ground 2S+1L term of the reference free ion, the model space includes all of its spin components. p. 2 f
n
f
1
f
2
f
3
f
4
f
5
f
6
f
8
f
9
f
10 f
11 f
12 f
13
2L+1 7 11 13 13 11 7 7 11 13 13 11 7
2J+1 6 9 10 9 6 1 13 16 17 16 13 8
∑
J
2J+1 14 33 52 65 66 49 49 66 65 52 33 14
Once the SOC is considered, the model space size naturally increases. With NewMag, since it works on the basis of
all the possible J’s that can be formed from the ground 2S+1L
term of the reference free ion, the model space includes all the
spin components of the 2L+1 orbital roots, that is ∑
J
2J+1 =
(2L+1)(2S+1) = (2L+1)( ˜n+1) where ˜n is the (maximum) number of unpaired electrons ( ˜n=n if n<7 or ˜n=14−n if n>7). With SINGLE_ANISO, the pseudo-J
approximation implies that the ground 2S+1LJ term of the reference free ion is used to define the model space. For a less
than half-filled shell (f
1– f
6
), J
min is the ground state, while
for a more than half-filled shell, J
max is the ground state. Moreover, as already mentioned,
the pseudo-J approximation for the f
6
configuration leads to
only one energy level in the model space, and thus to the practical impossibility of extracting CFPs. To conclude, prior to any extraction of the CFPs with
NewMag, the previous SR ab initio calculations must deal with
at least 2L+1 roots while the SOCI ones must deal with at least
(2L+1)( ˜n+1) roots, since by definition the ab initio space cannot be smaller than the model one.Where the paper says it
Where the paper says it
Where the paper says it
Effective Hamiltonian Construction
The effective Hamiltonian at the SR level is built in the basis of ML components of the ground 2S+1L term of the reference free ion, and it undergoes five steps to process ab initio wave functions. Step 1 allows treatments beyond the pseudo-L approximation by letting users choose between retaining pseudo-L roots or providing an output from an independent SA-CASSCF calculation. p. 3 Step 1, in
contrast, is a key feature of the present NewMag development,
as it allows treatments beyond the pseudo-L approximation
(which is not possible with SINGLE_ANISO). NewMag allows
the user to choose among the following options:
a. To generalize, let us denote by U1 the 7×7 matrix that transforms the one-electron
representation from the RSHs to the CSHs. In f
n
cases, the
transformation matrix U is constructed as the tensor product
of U1 with itself, repeated ˜n−1 times:
U = U1
n˜−1
⊗ U1 = U1 ⊗.Where the paper says it
Where the paper says it
Parameter Extraction and Analysis
The extraction of the CF Bq k parameters at the SR level follows the irreducible tensor operator (ITO) procedure of Chibotaru and Ungur. At the SOCI level, two operations are performed to extract parameters: extracting CF Bq k parameters from the SOCI effective Hamiltonian and applying the ITO procedure to extract λ(SOCI). p. 5 f
n
f
1
f
2
f
3
f
4
f
5
f
6
f
8
f
9
f
10 f
11 f
12 f
13
Highest k 6 10 12 12 10 6 6 10 12 12 10 6
Number of CFPs 27 65 90 90 65 27 27 65 90 90 65 27
At the SOCI level, two operations are performed to extract
parameters:
a. The CF B
q
k
parameters are extracted from the SOCI effective Hamiltonian similar to the SR level:
B
q
k
(SOCI) = Tr(Heff(SOCI)·O
q
k
)
Tr(O
q
k
·O
q
k
)
(6)
b. f 8– f
13). The resulting SOC constants are similar to those obtained
by AILFT since NewMag also uses the spherical approximation.Where the paper says it
Where the paper says it
Case Studies and Limitations
The paper demonstrates capabilities through six case studies, including a model dysprosium(III) complex (f9), which serves as an ideal configuration for testing the NewMag development. For transition metal complexes like a model nickel(II) complex (d8), NewMag shows limitations because it neglects couplings between the SR states of the model space with those outside it, and it also neglects second-order SOCs for instance with states of different spin multiplicities. p. 7 COMPUTATIONAL DETAILS
In this article, we report six case studies aiming at illustrating the capabilities of the current NewMag development for
lanthanide complexes (four cases), as well as exploring potential limitations when applying the code for other purposes, for
instance to study transition metal complexes (two cases). Because the implementation works with both OpenMolcas and
ORCA, results generated with both these programs are reported, but a detailed comparison between data generated with
these two programs is given only in Section IV B 1.Where the paper says it
Conclusion and Comparison
NewMag is a complementary approach to SINGLE ANISO and AILFT, retaining explicit information on the nature of the many-electron states throughout the procedure. It performs very well for any trivalent lanthanide complex of the f1–f6 and f8–f13 configurations. p. 7 COMPUTATIONAL DETAILS
In this article, we report six case studies aiming at illustrating the capabilities of the current NewMag development for
lanthanide complexes (four cases), as well as exploring potential limitations when applying the code for other purposes, for
instance to study transition metal complexes (two cases). Because the implementation works with both OpenMolcas and
ORCA, results generated with both these programs are reported, but a detailed comparison between data generated with
these two programs is given only in Section IV B 1. COMPUTATIONAL DETAILS
In this article, we report six case studies aiming at illustrating the capabilities of the current NewMag development for
lanthanide complexes (four cases), as well as exploring potential limitations when applying the code for other purposes, for
instance to study transition metal complexes (two cases). Because the implementation works with both OpenMolcas and
ORCA, results generated with both these programs are reported, but a detailed comparison between data generated with
these two programs is given only in Section IV B 1.Where the paper says it
Where the paper says it
Conflict of Interest
The authors have no conflicts to disclose.
Acknowledgments
The authors thank Rémi Marchal for the administration of the HPC cluster of ISCR as well as for precious technical assistance.
Data Availability Statement
The data that support the findings of this study are available within the article and at the NewMag code repository, which is publicly available on GitHub at https://github.com/clausserg/newmag. p. 7 The output files analyzed in this
article are available at https://github. com/clausserg/
newmag, where further details on the individual calculations
can be found.Where the paper says it
References
1N. F. Chilton, “Molecular magnetism,” Annu. p. 1 Note that the other mentioned programs typically
execute an empirical extraction using a point-charge model to
represent the ligands. The SINGLE_ANISO program, developed by Ungur and
Chibotaru9,12, was the first available for extracting CF parameters from relativistic and multiconfigurational ab initio wave
functions, and played a key role in the development of the
field of computational magnetism with transition-metal and
lanthanide complexes; the code allowed many researchers to
rationalize magnetic properties across a plethora of SIMs, and
contributed decisively to the design and characterization of
current state-of-the-art lanthanide-based SIMs.Where the paper says it
3J. D. Rinehart and J. p. 0 A potentially “good” SIM may be expected if the |J,±Mmax
J
⟩
energy levels are the lowest in energy, which translates into
the oblate vs. prolate rule of Rinehart and Long3
. Actually, optical and magnetic properties of f-element systems can successfully be understood within the framework of
CF theory1,2. For instance, a “good” single ion magnet (SIM)
may be obtained by following specific design rules that are
directly related to the CF model3
.Where the paper says it
Where the paper says it
4W. Martin and W. Wiese, “Atomic spectroscopy,” in Springer Handbook of Atomic, Molecular, and Optical Physics, edited by G. p. 2 Assuming SA-CASSCF calculations, the SR wave functions are typically expressed in the following way:
ψ
SR
i = ∑
l
al
|φl⟩ (2)
where each root ψ
SR
i
is expressed as a linear combination of
Slater determinants φl with real coefficients al
. Within each
Slater determinant, the molecular (or atomic, for free atoms or
ions) orbitals may have occupations of 2, 1 or 0, with the total
number of electrons in a given Slater determinant equal to n. ion, although the intention is to implicitly account for most of
these effects. Alternatively, we have recently reported NewMag: a new
code to compute parameters for understanding optical and
magnetic properties in the context of ab initio multiconfigurational calculations11,14.Where the paper says it
Where the paper says it
5K. W. H. Stevens, “Matrix elements and operator equivalents connected with the magnetic properties of rare earth ions,” Proc. p. 0 Note that various frameworks coexist in the literature and that we are working within the framework of Stevens equivalent operators5
. The general expression of the CF Hamiltonian, HˆCF, is:
HˆCF = ∑
k
αk(J,L,l)
k
∑
q=−k
B
q
kO
q
k
(J,L,l) (1)
where the k’s and q’s are the operator ranks and orders,
respectively, the O
q
k
’s are the “extended” Stevens operators,
the B
q
k
’s are the CF parameters, and the αk’s are the appropriate reduced-matrix elements (these are also called α, β and γ
for rank 2, 4 and 6).Where the paper says it
6J. J. Baldoví, S. Cardona-Serra, J. M. Clemente-Juan, E. Coronado, A. Gaita-Ariño, and A. Palii, “SIMPRE: A software package to calculate crystal field parameters, energy levels, and magnetic properties on mononuclealanthanoid complexes based on charge distributions,” J. p. 1 Note that the other mentioned programs typically
execute an empirical extraction using a point-charge model to
represent the ligands. The SINGLE_ANISO program, developed by Ungur and
Chibotaru9,12, was the first available for extracting CF parameters from relativistic and multiconfigurational ab initio wave
functions, and played a key role in the development of the
field of computational magnetism with transition-metal and
lanthanide complexes; the code allowed many researchers to
rationalize magnetic properties across a plethora of SIMs, and
contributed decisively to the design and characterization of
current state-of-the-art lanthanide-based SIMs.Where the paper says it
7A. Scheie, “PyCrystalField: software for calculation, analysis and fitting of crystal electric field Hamiltonians,” J. p. 4 Although this step-by-step procedure may seem tedious, involving the localization of the orbitals following an initial
SA-CASSCF calculation, a subsequent CASCI calculation,
and several successive basis transformations, it is necessary
to track the character of the many-electron states throughout
the entire procedure, without introducing any a priori assumptions regarding their nature. By doing so, one can extract CF
Hamiltonians fully ab initio; the approach may allow one to
go beyond the phenomenological approach if required, since
the ab initio calculations can then support the choice of the
model Hamiltonian or help revising it if necessary28.Where the paper says it
8L. Fiorucci and E. Ravera, “Not just another crystal field software,” J. p. 0 The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes.Where the paper says it
9L. F. Chibotaru and L. p. 5 6
procedure as described earlier11, based on the irreducible tensor operator (ITO) procedure of Chibotaru and Ungur9,12. At
the SR level, the B
q
k
’s are directly extracted from the effective
Hamiltonian:
B
q
k
(SR) = Tr(Heff(SR)·O
q
k
)
Tr(O
q
k
·O
q
k
)
(5)
The highest possible value for a k rank relates to the degrees
of freedom that arise from the size of the model space, comprising 2L+1 elements: it is simply equal to 2L (2L operations
are necessary to couple the Mmin
L
and Mmax
L
configurations). Note that the other mentioned programs typically
execute an empirical extraction using a point-charge model to
represent the ligands. The SINGLE_ANISO program, developed by Ungur and
Chibotaru9,12, was the first available for extracting CF parameters from relativistic and multiconfigurational ab initio wave
functions, and played a key role in the development of the
field of computational magnetism with transition-metal and
lanthanide complexes; the code allowed many researchers to
rationalize magnetic properties across a plethora of SIMs, and
contributed decisively to the design and characterization of
current state-of-the-art lanthanide-based SIMs. Moreover, as already mentioned,
the pseudo-J approximation for the f
6
configuration leads to
only one energy level in the model space, and thus to the practical impossibility of extracting CFPs. To conclude, prior to any extraction of the CFPs with
NewMag, the previous SR ab initio calculations must deal with
at least 2L+1 roots while the SOCI ones must deal with at least
(2L+1)( ˜n+1) roots, since by definition the ab initio space cannot be smaller than the model one.Where the paper says it
Where the paper says it
Where the paper says it
10M. Atanasov, D. p. 1 With the J basis, only
the MJ components of the ground 2S+1LJ term are included
(2J+1 components). The AILFT (for Ab Initio Ligand Field
Theory) program, developed by Atanasov et al. Note that the other mentioned programs typically
execute an empirical extraction using a point-charge model to
represent the ligands. The SINGLE_ANISO program, developed by Ungur and
Chibotaru9,12, was the first available for extracting CF parameters from relativistic and multiconfigurational ab initio wave
functions, and played a key role in the development of the
field of computational magnetism with transition-metal and
lanthanide complexes; the code allowed many researchers to
rationalize magnetic properties across a plethora of SIMs, and
contributed decisively to the design and characterization of
current state-of-the-art lanthanide-based SIMs.Where the paper says it
Where the paper says it
11D.-C. Sergentu, G. p. 0 The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. maurice@univ-rennes. fr)
(Dated: 30 September 2026)
We report a massive update of the NewMag program, which enables to extract Stevens crystal-field parameters (CFPs)
after relativistic and multiconfigurational calculations are performed. maurice@univ-rennes. fr)
(Dated: 30 September 2026)
We report a massive update of the NewMag program, which enables to extract Stevens crystal-field parameters (CFPs)
after relativistic and multiconfigurational calculations are performed.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
12L. Ungur and L. p. 1 Note that the other mentioned programs typically
execute an empirical extraction using a point-charge model to
represent the ligands. The SINGLE_ANISO program, developed by Ungur and
Chibotaru9,12, was the first available for extracting CF parameters from relativistic and multiconfigurational ab initio wave
functions, and played a key role in the development of the
field of computational magnetism with transition-metal and
lanthanide complexes; the code allowed many researchers to
rationalize magnetic properties across a plethora of SIMs, and
contributed decisively to the design and characterization of
current state-of-the-art lanthanide-based SIMs. Keywords: crystal-/ligand-field; Stevens operators and parameters; ab initio calculations; ORCA; OpenMolcas. INTRODUCTION
The crystal field (CF) and ligand field (LF) models are key
to understand various properties of f-element complexes, in
particular concering the lanthanides (4 f).Where the paper says it
Where the paper says it
13D.-C. Sergentu, F. p. 0 The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. The active orbitals are essentially of 4 f
character. The second one, denoted CAS(5,9), augments the
minimal CAS with the bonding orbitals between the ligand
πδ orbitals and the lanthanide 4 fδ ones, denoted (πδ –4 fδ
)+
where the “+” symbol indicates in-phase bonding character13
. 37541v1 [physics. p. 0 The latter two are clearly the closest ones to our own work
arXiv:2609. 37541v1 [physics.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
14D.-C. Sergentu, D. p. 0 The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. ion, although the intention is to implicitly account for most of
these effects. Alternatively, we have recently reported NewMag: a new
code to compute parameters for understanding optical and
magnetic properties in the context of ab initio multiconfigurational calculations11,14. 37541v1 [physics. chem-ph] 29 Sep 2026Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
15F. Neese, “Software update: The ORCA program system—version 6. p. 0 fr)
(Dated: 30 September 2026)
We report a massive update of the NewMag program, which enables to extract Stevens crystal-field parameters (CFPs)
after relativistic and multiconfigurational calculations are performed. The code can now post-treat ORCA and OpenMolcas outputs that contain state-average complete active space self-consistend field (SA-CASSCF) calculations with
minimal or extended active spaces, second-order NEVPT2 or CASPT2 calculations, and spin-orbit configuration interaction (SOCI) calculations.Where the paper says it
16G. Li Manni et al., “The OpenMolcas web: A community-driven approach to advancing computational chemistry,” J. p. 0 prolate rule of Rinehart and Long3
. The CF theory has undergone various conceptual developments over the past century and continues to evolve today, particularly within the computational chemistry community. For instance, a “good” single ion magnet (SIM)
may be obtained by following specific design rules that are
directly related to the CF model3
. Regarding optical properties, the CF theory may allow to understand why formallyforbidden electronic transitions4 become allowed, often quite
intense, and may be observed in the laboratory.Where the paper says it
Where the paper says it
17B. O. Roos and P. R. Taylor, “A complete active space SCF method (CASSCF) using a density matrix formulated super-CI approach,” Chem. p. 3 These are first used to perform SA-CASSCF
calculations and localize the active space orbitals, which are
fed to subsequent CASCI calculations. With ORCA, localized
orbitals can be generated with the “actorbs” keyword, which
in fact generates the orbital set that is also used by AILFT,
or manually, using the “rotate” directive in the scf block.Where the paper says it
18B. O. Roos, “The complete active space SCF method in a Fock-matrixbased super-CI formulation,” Int. p. 3 These are first used to perform SA-CASSCF
calculations and localize the active space orbitals, which are
fed to subsequent CASCI calculations. With ORCA, localized
orbitals can be generated with the “actorbs” keyword, which
in fact generates the orbital set that is also used by AILFT,
or manually, using the “rotate” directive in the scf block. 37541v1 [physics. chem-ph] 29 Sep 2026Where the paper says it
Where the paper says it
20K. Andersson et al., “Second-order perturbation theory with a complete active space self-consistent field reference function,” J. p. 0 fr)
(Dated: 30 September 2026)
We report a massive update of the NewMag program, which enables to extract Stevens crystal-field parameters (CFPs)
after relativistic and multiconfigurational calculations are performed. The code can now post-treat ORCA and OpenMolcas outputs that contain state-average complete active space self-consistend field (SA-CASSCF) calculations with
minimal or extended active spaces, second-order NEVPT2 or CASPT2 calculations, and spin-orbit configuration interaction (SOCI) calculations.Where the paper says it
30A. Abragam and B. Bleaney, Electron paramagnetic resonance of transition ions (Oxford University Press, 1970). p. 0 The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes.Where the paper says it
35N. C. Chang et al., “Optical spectra, energy levels, and crystal-field analysis of tripositive rare earth ions in Y2O3. I. Kramers ions in C2 sites,” J. p. 2 Assuming SA-CASSCF calculations, the SR wave functions are typically expressed in the following way:
ψ
SR
i = ∑
l
al
|φl⟩ (2)
where each root ψ
SR
i
is expressed as a linear combination of
Slater determinants φl with real coefficients al
. Within each
Slater determinant, the molecular (or atomic, for free atoms or
ions) orbitals may have occupations of 2, 1 or 0, with the total
number of electrons in a given Slater determinant equal to n.Where the paper says it
41G. Jansen and B. A. Hess, “Revision of the Douglas-Kroll transformation,” Phys. p. 7 OpenMolcas calculations
All calculations performed with OpenMolcas16 used
the release version 25. The Douglas-Kroll-Hess
Hamiltonian39–41 was employed, together with the “compatible with it” ANO-RCC basis sets42–44.Where the paper says it
42B. O. Roos et al., “New relativistic ANO basis sets for lanthanide atoms with applications to the Ce diatom and LuF3,” J. p. 8 9
ity basis sets on all atoms (ANO-RCC-VTZP). Two active spaces were considered.Where the paper says it
43B. O. Roos et al., “New relativistic ANO basis sets for transition metal atoms,” J. p. 7 ORCA calculations
All calculations performed with ORCA15 used version
6. The Douglas-Kroll-Hess Hamiltonian39–41 was also
employed, together with appropriate basis sets: DKH-def2
basis sets48,49 for light (H, C, N, O) as well as for transition
metal atoms (Zn) and SARC2-DKH basis sets50 for the lanthanide atoms (Eu, Dy).Where the paper says it
44B. O. Roos et al., “New relativistic atomic natural orbital basis sets for lanthanide atoms with applications to the Ce diatom and LuF3,” J. p. 7 ORCA calculations
All calculations performed with ORCA15 used version
6. The Douglas-Kroll-Hess Hamiltonian39–41 was also
employed, together with appropriate basis sets: DKH-def2
basis sets48,49 for light (H, C, N, O) as well as for transition
metal atoms (Zn) and SARC2-DKH basis sets50 for the lanthanide atoms (Eu, Dy).Where the paper says it
45N. Forsberg and P.-Å. Malmqvist, “Multiconfigurational perturbation theory with imaginary level shift,” Chem. p. 5 f
n
f
1
f
2
f
3
f
4
f
5
f
6
f
8
f
9
f
10 f
11 f
12 f
13
Highest k 4 8 8 8 4 0 12 14 16 14 12 6
Number of CFPs 14 44 44 44 14 0 90 119 152 119 90 27
This is completely bypassed by the NewMag approach, since
the direct ITO extraction of the CFPs in the |J,MJ ⟩ basis,
based on the full manifold that correlates with the 2S+1L
ground state of the reference free ion, allows the extraction
of the same number of CFPs as at the SR level (see Table
II). This aspect is perfectly consistent with the CF theory,
since the CF should only act on the orbital degrees of freedom
and since the SOC is treated as a perturbation of the SR picture within the SOCI framework.Where the paper says it
46G. Ghigo et al., “A modified definition of the zeroth order Hamiltonian in multiconfigurational perturbation theory (CASPT2),” Chem. p. 7 However, we considered only
minimal active space calculations and only SA-CASSCF and
SO-CASSCF calculations11 because of limitations of the first
released version. Here, we aim at considering a larger active
space and/or more correlated energies based on second-order
perturbation theory, with OpenMolcas, and with triple-ζ qual-Where the paper says it
51F. Neese, “Efficient and accurate approximations to the molecular spinorbit coupling operator and their use in molecular g-tensor calculations,” J. p. 2 Assuming SA-CASSCF calculations, the SR wave functions are typically expressed in the following way:
ψ
SR
i = ∑
l
al
|φl⟩ (2)
where each root ψ
SR
i
is expressed as a linear combination of
Slater determinants φl with real coefficients al
. Within each
Slater determinant, the molecular (or atomic, for free atoms or
ions) orbitals may have occupations of 2, 1 or 0, with the total
number of electrons in a given Slater determinant equal to n.Where the paper says it
52M. D. Walter et al., “Cerocene revisited: The electronic structure of and interconversion between Ce2(C8H8)3 and Ce(C8H8)2,” Organometallics, 28, 698–707 (2009). p. 7 To this end, we revisit the cerocene
anion, i. the bis(cyclooctatetraenyl)cerium(III) anion,
[Ce(C8H8)2]
−52.Where the paper says it
53J. J. Le Roy et al., “Structural and magnetic conformation of a cerocene [Ce(COT)2]− exhibiting a uniconfigurational f1 ground state and slow-magnetic relaxation,” Dalton Trans. p. 0 J, L, l are the possible bases: J is the polyelectronic orbital+spin basis (J=L+S), L is the polyelectronic
orbital basis and l is the monoelectronic orbital basis. For
describing the magnetic properties of lanthanide SIMs, it is
common practice to work with the ground-state |J,MJ ⟩ basis.Where the paper says it
54F. Gendron et al., “Single-ion 4f element magnetism: an ab-initio look at Ln(COT)−2,” Dalton Trans. p. 1 Note that the other mentioned programs typically
execute an empirical extraction using a point-charge model to
represent the ligands. The SINGLE_ANISO program, developed by Ungur and
Chibotaru9,12, was the first available for extracting CF parameters from relativistic and multiconfigurational ab initio wave
functions, and played a key role in the development of the
field of computational magnetism with transition-metal and
lanthanide complexes; the code allowed many researchers to
rationalize magnetic properties across a plethora of SIMs, and
contributed decisively to the design and characterization of
current state-of-the-art lanthanide-based SIMs.Where the paper says it
55F. Gendron et al., “Magnetic coupling in the Ce(III) dimer C2(COT)3,” Inorg. p. 0 Rotation invariant indicators are also computed, after applying a correction of the
parameter values to ensure normalization, following Rudowicz. At the SOCI level, the spin-orbit coupling (SOC)
constant is also extracted within a spherical approximation of the SOC operator.Where the paper says it
61R. Alessandri et al., “Crystal field in rare-earth complexes: From electrostatics to bonding,” Chem. p. 0 Keywords: crystal-/ligand-field; Stevens operators and parameters; ab initio calculations; ORCA; OpenMolcas. INTRODUCTION
The crystal field (CF) and ligand field (LF) models are key
to understand various properties of f-element complexes, in
particular concering the lanthanides (4 f).Where the paper says it
62C. Görller-Walrand and K. Binnemans, “Chapter 155 Rationalization of crystal-field parametrization,” (Elsevier, 1996) pp. p. 0 The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes.Where the paper says it
63R. Micová et al., “Magnetic properties of a europium(III) complex – possible multiplet crossover,” Dalton Trans. p. 0 J, L, l are the possible bases: J is the polyelectronic orbital+spin basis (J=L+S), L is the polyelectronic
orbital basis and l is the monoelectronic orbital basis. For
describing the magnetic properties of lanthanide SIMs, it is
common practice to work with the ground-state |J,MJ ⟩ basis.Where the paper says it
64J. P. Fackler Jr. and A. Avdeef, “Crystal and molecular structure of tris(2,4-pentanedionato)manganese(III), Mn(O2C5H7)3, a distorted complex as predicted by Jahn-Teller arguments,” Inorg. p. 9 First, we have determined a genuine structure for the
[Ce∗
(H2O)8 ·(H2O)]3+
complex, based on TD-DFT, obtained
by optimizing the 8th SR root (the first 7th corresponding
to the 4 f
1 manifold). The structure is displayed in Figure III (coordinates are given in the GibHub repository of
NewMag).Where the paper says it
65R. Maurice et al. p. 0 The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. The NewMag crystal-field code for f-element systems: Implementation
for extended active spaces, second-order correlated energies, and
generalisation to f
n configurations
Gwenhaël Duplaix-Rata,1 Dumitru-Claudiu Sergentu,2, 3 Boris Le Guennic,1
and Rémi Maurice*1
1)Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) – UMR 6226, F-35000 Rennes,
France
2)Faculty of Chemistry, ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
3)ICI-RECENT AIR (RA03), ”Alexandru Ioan Cuza” University of Ias,i, Ias,i, Romania
(*Electronic mail: remi. maurice@univ-rennes. Keywords: crystal-/ligand-field; Stevens operators and parameters; ab initio calculations; ORCA; OpenMolcas. INTRODUCTION
The crystal field (CF) and ligand field (LF) models are key
to understand various properties of f-element complexes, in
particular concering the lanthanides (4 f).Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
66Y. Zhang and D.-S. Yang, “Spin purity and spin mixing in lanthanide–cyclooctatetraene complexes: SO-MCQDPT2 and axial-field analysis,” J. p. 2 They merely provide the freedom to rotate the orbitals
into a basis that facilitates the analysis and interpretation of
the wave functions, which is the approach adopted here. If minimal-active-space SA-CASSCF calculations are performed, and if the CASCI space for the high-spin states spans
only the orbital configurations that correlate with the 2S+1L maurice@univ-rennes. fr)
(Dated: 30 September 2026)
We report a massive update of the NewMag program, which enables to extract Stevens crystal-field parameters (CFPs)
after relativistic and multiconfigurational calculations are performed.Where the paper says it
Where the paper says it
67R. Maurice et al., “Zero-field splitting in transition metal complexes: Ab initio calculations, effective hamiltonians, model hamiltonians, and crystal-field models,” in Handbook of Relativistic Quantum Chemistry, edited by W. p. 1 Note that the other mentioned programs typically
execute an empirical extraction using a point-charge model to
represent the ligands. The SINGLE_ANISO program, developed by Ungur and
Chibotaru9,12, was the first available for extracting CF parameters from relativistic and multiconfigurational ab initio wave
functions, and played a key role in the development of the
field of computational magnetism with transition-metal and
lanthanide complexes; the code allowed many researchers to
rationalize magnetic properties across a plethora of SIMs, and
contributed decisively to the design and characterization of
current state-of-the-art lanthanide-based SIMs.Where the paper says it
68R. Maurice et al., “Analytical derivations for the description of magnetic anisotropy in transition metal complexes,” in Computational Modelling of Molecular Nanomagnets, edited by G. p. 1 Note that the other mentioned programs typically
execute an empirical extraction using a point-charge model to
represent the ligands. The SINGLE_ANISO program, developed by Ungur and
Chibotaru9,12, was the first available for extracting CF parameters from relativistic and multiconfigurational ab initio wave
functions, and played a key role in the development of the
field of computational magnetism with transition-metal and
lanthanide complexes; the code allowed many researchers to
rationalize magnetic properties across a plethora of SIMs, and
contributed decisively to the design and characterization of
current state-of-the-art lanthanide-based SIMs.Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved extraction of crystal field parameters (CFPs) from ab initio calculations by enabling post-treatment of ORCA and OpenMolcas outputs containing state-average complete active space self consistent field SA-CASSCF calculations with minimal or extended active spaces, second order correlated energies, and spin orbit configuration interaction SOCI calculations.
-
Ability to extract Stevens parameters following the original Stevens convention with the so called extended parameters and compute rotation invariant indicators after applying a correction of the parameter values to ensure normalization following Rudowicz.
-
Capability to extract the spin orbit coupling SOC constant within a spherical approximation of the SOC operator at the SOCI level, as well as reconstruct model spectra for all reference calculation levels for direct assessment of quality.
-
Extension of NewMag to process f n configurations with n not equal to 0, not equal to 7 and not equal to 14, including f 2–f 6 and f 8–f 13 configurations with extended active spaces and root selection beyond the pseudo-L or pseudo J approximation.
-
Implementation of a workflow that processes ORCA or OpenMolcas outputs by first localizing active space orbitals, then feeding them to subsequent CASCI calculations, allowing for the identification of SR wave functions through
localized
andpurified
active orbitals. -
Ability to build the effective Hamiltonian at the SR level in the basis of ML components of the ground 2S+1L term of the reference free ion by performing five steps including selection of roots, retention of components on targeted f orbitals, and application of transformations to express wave functions in complex spherical harmonics and many-electron orbital bases.
-
Functionality to build the effective Hamiltonian at the SOCI level in a coupled basis using Clebsch-Gordan coefficients based on the ground 2S+1L term, allowing for extraction of CF B q k parameters and SOC parameters like lambda(SOCI).
-
Ability to analyze the quality of model spectra by calculating mean absolute error MAE and root mean square deviation RMSD to ab initio energies, as well as calculating the impact of various k ranks on the model spectrum by reconstructing the model Hamiltonian with different combinations of CFPs.
-
Capability to compare results with existing codes like SINGLE ANISO and AILFT by demonstrating that NewMag is complementary, combining their strengths while enabling additional analyses such as explicit evaluation of pseudo-J approximation relevance or assessment of the impact of rank 8 and higher CFPs.
-
Support for analyzing J mixing due to second order CF couplings by constructing the CF Hamiltonian in the ML S basis prior to transforming it to the J basis, which naturally introduces J mixing and allows for its quantification at both SR and SOCI levels.
-
Implementation of rotationally invariant Stevens parameters by re-scaling extracted CFPs based on normalized versions, such as B prime q k parameters, defined using rank 2, 4, and 6 operators to ensure consistency with Rudowicz normalization.
Abstract
We report a massive update of the NewMag program, which enables to extract Stevens crystal-field parameters (CFPs) after relativistic and multiconfigurational calculations are performed. The code can now post-treat ORCA and OpenMolcas outputs that contain state-average complete active space self-consistend field (SA-CASSCF) calculations with minimal or extended active spaces, second-order NEVPT2 or CASPT2 calculations, and spin-orbit configuration interaction (SOCI) calculations. Stevens parameters are extracted following the original Stevens convention with the so-called "extended" parameters. Rotation invariant indicators are also computed, after applying a correction of the parameter values to ensure normalization, following Rudowicz. At the SOCI level, the spin-orbit coupling (SOC) constant is also extracted within a spherical approximation of the SOC operator. For all the reference calculation levels, the model spectrum is reconstructed, allowing a direct assessment of its quality. The reported implementation is successfully applicable to f-element systems with a non-void, non-half-filled/empty or non-full f shell, that is for f n configurations with n != 0, != 7 and != 14. Similarities and differences with the SINGLE ANISO and AILFT codes are discussed. With selected examples, we showcase the interest of determining CFPs in f-element systems to understand their magnetic and optical properties in general, and more specifically the added value of NewMag. Finally, application of this approach to d-element systems is discussed, to reveal when it readily and successfully applies and when it may fail in reproducing satisfactorily the ab initio energies.
Related papers
- A System-Independent Metadynamics Strategy for Reactive Training Data: Application to Gas-Phase Organic Reactions
- A Machine Learning Model for the Chemistry of a Solvated Electron
- Compression of virtual spaces in transcorrelated methods via singular value decomposition: application to the G2 set
- Charge Transfer with a Spin. II: A Framework for Diabatization which Localizes Charge and Spin
- Diabatic Seam Space Sampling for Hydrogen Tunneling Systems with Nuclear-Electronic Orbital Theory
- Augmented Roothaan-Hall Hessian Applied to Spin-Restricted Open-Shell Density-Functional Theory