pith. sign in

arxiv: 2605.22394 · v1 · pith:XS2W5ZXPnew · submitted 2026-05-21 · ⚛️ physics.chem-ph

Dynamic electron correlation energy for multireference wavefunction methods from one- and two-electron reduced density matrices

Micha{\l} Hapka , Aleksandra Tucholska , Katarzyna Pernal This is my paper

Pith reviewed 2026-05-22 02:10 UTC · model grok-4.3

classification ⚛️ physics.chem-ph
keywords dynamic correlationmultireference methodsreduced density matricesadiabatic connectionMC-PDFTspin-state energeticsbiradicals
0
0 comments X

The pith

Linearized AC0 recovers dynamic correlation from reduced density matrices as well as or better than second-order perturbation theory across multireference benchmarks.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This Perspective reviews methods that recover dynamic electron correlation for multireference wave functions using only the one- and two-electron reduced density matrices of a reference wave function. These fall into DFT-based approaches such as MC-srDFT and MC-srPDFT that rely on charge, spin, and on-top pair densities, and ab initio multireference adiabatic connection formulations. Direct benchmarks under identical active spaces and basis sets on singlet-triplet gaps in biradicals, excitation energies, and spin-state splittings in iron complexes show MC-srPDFT as the strongest DFT option thanks to the on-top pair density, yet all DFT variants prove unreliable for transition-metal spin energetics. Linearized AC0 stands out by rivaling or surpassing more expensive second-order perturbation theory methods on every tested set.

Core claim

Linearized AC0 within the multireference adiabatic connection framework extracts dynamic correlation energy directly from the reference reduced density matrices and achieves accuracy comparable to or exceeding that of second-order perturbation theories on singlet-triplet gaps, excitations, and transition-metal spin splittings, while DFT-based methods that incorporate the on-top pair density improve over simpler density-only functionals but still fail on metal complex energetics.

What carries the argument

The linearized AC0 approximation to the multireference adiabatic connection integrand, which evaluates dynamic correlation solely from the one- and two-electron reduced density matrices supplied by the reference wave function.

If this is right

  • MC-srPDFT emerges as the most accurate DFT-based method because it uses the on-top pair density in addition to charge and spin densities.
  • All DFT-based approaches remain unreliable for spin-state energetics in transition-metal complexes.
  • Linearized AC0 delivers competitive or superior results to second-order perturbation theory while depending only on reduced density matrices.
  • These outcomes highlight the value of on-top pair densities for functional-based methods and the scalability potential of adiabatic connection formulations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The density-matrix-only character of AC0 suggests it could integrate with low-scaling or quantum-computing frameworks for larger active spaces.
  • Improved functionals tuned for transition metals might close the gap between DFT-based and ab initio approaches on spin-state problems.
  • Testing the same methods on periodic systems or extended organic radicals would reveal whether the benchmark rankings hold beyond molecular cases.

Load-bearing premise

The selected benchmark problems with matched active spaces and basis sets supply an unbiased head-to-head comparison that extends to general chemical applications.

What would settle it

A new calculation on an additional iron complex or biradical system outside the current sets, using the same active spaces, in which linearized AC0 produces clearly larger errors than second-order perturbation theory would falsify the performance claim.

Figures

Figures reproduced from arXiv: 2605.22394 by Aleksandra Tucholska, Katarzyna Pernal, Micha{\l} Hapka.

Figure 1
Figure 1. Figure 1: Mean unsigned errors across all data sets (see also Tables 1–4) for the low-order density [PITH_FULL_IMAGE:figures/full_fig_p019_1.png] view at source ↗
read the original abstract

Efficiently recovering dynamic correlation in strongly correlated systems without incurring prohibitive computational costs remains a central challenge in quantum chemistry. In this Perspective, we review and benchmark methods capable of recovering dynamic correlation for multireference wave functions exclusively from low-order reduced density matrices and densities. These approaches require at most the two-electron reduced density matrix of the reference wave function and fall into two categories: density functional theory (DFT)-based methods and purely ab initio multireference adiabatic connection (AC) methods. The former include MC-srDFT, which recovers dynamic correlation through a short-range exchange-correlation functional depending on the charge and spin densities, as well as MC-PDFT and MC-srPDFT, which employ translated functionals that additionally depend on the on-top pair density. Within the post-CASSCF framework, we perform a direct, head-to-head benchmark of these approaches under identical computational settings (including active spaces and basis sets) against challenging multireference problems, including singlet-triplet gaps in organic biradicals, excitation energies, and spin-state splittings in iron complexes. Among the DFT-based methods, MC-srPDFT emerges as the most accurate, underscoring the benefit of incorporating the on-top pair density. However, all considered DFT-based methods fail to provide reliable spin-state energetics for transition-metal complexes. Conversely, linearized AC0 rivals or outperforms more computationally expensive second-order perturbation theory approaches across all benchmark sets. We discuss these findings in the context of alternative formulations and existing literature, highlighting critical limitations and identifying promising directions for the future development of scalable multireference methods.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. This Perspective reviews and benchmarks methods for recovering dynamic electron correlation in multireference systems using only one- and two-electron reduced density matrices. It categorizes approaches into DFT-based methods (MC-srDFT, MC-PDFT, MC-srPDFT) that employ short-range functionals or on-top pair densities and purely ab initio multireference adiabatic-connection (AC) methods. Under identical CASSCF settings, active spaces, and basis sets, the paper benchmarks these on singlet-triplet gaps in organic biradicals, excitation energies, and spin-state splittings in iron complexes. Key findings are that MC-srPDFT is the most accurate among DFT-based methods but all fail on transition-metal spin energetics, while linearized AC0 rivals or outperforms second-order perturbation theory across the sets.

Significance. If the benchmarks hold and the linear AC approximation is validated, the work provides a useful head-to-head comparison that positions AC-based methods as promising for scalable multireference treatments without the cost of full perturbation theory or higher-order expansions. The explicit use of RDMs and identical computational settings strengthens the comparative assessment, though the absence of direct error quantification for the linearization limits the strength of the outperformance claim.

major comments (1)
  1. [Abstract / AC methods discussion] Abstract and the section discussing AC methods: the headline claim that linearized AC0 rivals or outperforms second-order PT is load-bearing for the central conclusion, yet the manuscript provides no direct quantification of the linearization error (e.g., by comparing the straight-line or endpoint approximation to numerically integrated AC results on the same benchmark systems). Without this, it is unclear whether the reported gains over PT are larger than the approximation error itself.
minor comments (2)
  1. [Abstract] The abstract states that all DFT-based methods fail on spin-state energetics for transition-metal complexes; a brief table or figure summarizing the specific error magnitudes for MC-srPDFT versus experiment or reference values would improve clarity.
  2. [Methods / AC formulation] Notation for the adiabatic-connection integrand and the linearization procedure should be defined explicitly in the main text (rather than assuming familiarity with prior AC literature) to aid readers new to the approach.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their thorough review and valuable feedback on our manuscript. We address the major comment point by point below, providing clarifications and indicating where revisions will be made to strengthen the paper.

read point-by-point responses

  1. Referee: [Abstract / AC methods discussion] Abstract and the section discussing AC methods: the headline claim that linearized AC0 rivals or outperforms second-order PT is load-bearing for the central conclusion, yet the manuscript provides no direct quantification of the linearization error (e.g., by comparing the straight-line or endpoint approximation to numerically integrated AC results on the same benchmark systems). Without this, it is unclear whether the reported gains over PT are larger than the approximation error itself.

    Authors: We thank the referee for this observation. The manuscript does not include a direct comparison between the linearized AC0 and a numerically integrated AC for the benchmark systems presented. Such calculations are computationally demanding, particularly for the larger iron complexes, and were not performed in this work. However, the linear approximation in AC0 is well-established in the literature, and prior studies have quantified its accuracy on smaller multireference systems, showing that the error is typically small compared to the energy differences discussed here. We will update the manuscript to include a more explicit discussion of the linearization approximation and its expected error, referencing relevant prior work, to better contextualize the performance claims. This revision will be made in the section discussing AC methods. revision: partial

Circularity Check

0 steps flagged

No circularity: benchmarks rely on external comparisons under fixed settings

full rationale

The paper conducts a head-to-head benchmark of DFT-based and ab initio AC methods for dynamic correlation recovery using RDMs, comparing them directly to second-order perturbation theory on external test sets (singlet-triplet gaps, excitation energies, iron-complex spin splittings) with identical active spaces and basis sets. No derivation step reduces a claimed prediction or result to a fitted parameter or self-citation by construction; the linearized AC0 performance is reported as an empirical outcome from the numerical comparisons rather than an algebraic identity or internally normalized quantity. Self-citations to prior method formulations are present but not load-bearing for the central benchmark claims, which remain falsifiable against independent literature data.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

As a review and benchmark paper the central claims rest on standard quantum-chemistry assumptions about active-space selection and the transferability of benchmark sets; no new free parameters or invented entities are introduced in the abstract.

pith-pipeline@v0.9.0 · 5833 in / 1049 out tokens · 66044 ms · 2026-05-22T02:10:07.937279+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages

  1. [1]

    O.; Taylor, P

    (1) Roos, B. O.; Taylor, P. R.; Sigbahn, P. E. A complete active space SCF method (CASSCF) using a density matrix for- mulatedsuper-CIapproach.Chem. Phys. 1980,48, 157–173. (2) Roos, B. O. The complete active space self-consistent field method and its ap- plications in electronic structure calcula- tions.Adv. Chem. Phys.1987,69, 399–

    work page 1980

  2. [2]

    Striking the right bal- ance of encoding electron correlation in the Hamiltonian and the wavefunction ansatz.Faraday Discuss.2024,254, 359–

    (3) Szenes, K.; Mörchen, M.; Fischill, P.; Reiher, M. Striking the right bal- ance of encoding electron correlation in the Hamiltonian and the wavefunction ansatz.Faraday Discuss.2024,254, 359–

    work page 2024

  3. [3]

    E.; Mussard, B.; Holmes, A

    (4) Smith, J. E.; Mussard, B.; Holmes, A. A.; Sharma, S. Cheap and near exact CASSCF with large active spaces.J. Chem. Theory Comput.2017,13, 5468–

    work page 2017

  4. [4]

    iCISCF: An Iterative Configuration Interaction-Based Multiconfigurational Self-Consistent Field Theory for Large Active Spaces.J

    (5) Guo, Y.; Zhang, N.; Lei, Y.; Liu, W. iCISCF: An Iterative Configuration Interaction-Based Multiconfigurational Self-Consistent Field Theory for Large Active Spaces.J. Chem. Theory Comput. 2021,17, 7545–7561. (6) Garniron, Y.; Scemama, A.; Giner, E.; Caffarel, M.; Loos, P.-F. Selected config- uration interaction dressed by perturba- tion.J. Chem. Phys....

    work page 2021

  5. [5]

    (10) Sun, Q.; Yang, J.; Chan, G. K.-i. A gen- eral second order complete active space self-consistent-field solver for large-scale systems.Chem. Phys. Lett.2017,683, 291–299. 21 (11) Teale, A. M.; Helgaker, T.; Savin, A.; Adamo, C.; Aradi, B.; Arbuznikov, A. V.; Ayers, P. W.; Baerends, E. J.; Barone, V.; Calaminici, P. et al. DFT exchange: sharing perspec...

    work page 2017

  6. [6]

    Variational de- termination of the two-electron reduced density matrix: A tutorial review.Wi- ley Interdiscip

    (13) Eugene DePrince III, A. Variational de- termination of the two-electron reduced density matrix: A tutorial review.Wi- ley Interdiscip. Rev.: Comput. Mol. Sci. 2024,14, e1702. (14) Mazziotti, D. A. Quantum many-body theory from a solution of the N- representability problem.Phys. Rev. Lett.2023,130, 153001. (15) Maradzike, E.; Hapka, M.; Pernal, K.; De...

    work page 2024

  7. [7]

    (20) Piris, M.; Ugalde, J. M. Perspective on natural orbital functional theory.Int. J. Quant. Chem.2014,114, 1169–1175. (21) Piris, M. Exploring the potential of natu- ral orbital functionals.Chem. Sci.2024, 15, 17284–17291. (22) Piris, M. Global method for electron cor- relation.Phys. Rev. Lett.2017,119, 063002. (23) Piris, M. Global natural orbital func...

    work page 2014

  8. [8]

    M.; Sivalingam, K.; Neese, F

    (35) Kempfer, E. M.; Sivalingam, K.; Neese, F. Efficient Implementation of Approximate Fourth Order N-Electron Valence State Perturbation Theory. J. Chem. Theory Comput.2025,21, 3953–3967. (36) Forsberg, N.; Malmqvist, P.-Å. Multi- configuration perturbation theory with imaginary level shift.Chem. Phys. Lett. 1997,274, 196–204. (37) Ghigo, G.; Roos, B. O....

    work page 2025

  9. [9]

    M.; Wouters, S.; Pier- loot, K

    (40) Phung, Q. M.; Wouters, S.; Pier- loot, K. Cumulant approximated second- order perturbation theory based on the density matrix renormalization group for transition metal complexes: A bench- mark study.J. Chem. Theory Comput. 2016,12, 4352–4361. (41) Dyall, K. G. The choice of a zeroth-order Hamiltonian for second-order perturba- tion theory with a com...

    work page 2016

  10. [10]

    K.; Truhlar, D

    (66) Carlson, R. K.; Truhlar, D. G.; Gagliardi, L. Multiconfiguration pair- density functional theory: a fully translated gradient approximation and its performance for transition metal dimers and the spectroscopy of Re2Cl2− 8 . J. Chem. Theory Comput.2015,11, 4077–4085. (67) Rodrigues, G. L.; Scott, M.; Del- cey, M. G. Multiconfigurational pair- density ...

    work page 2015

  11. [11]

    J.; Gagliardi, L.; Truhlar, D

    (94) Ghosh, S.; Verma, P.; Cramer, C. J.; Gagliardi, L.; Truhlar, D. G. Combin- ing wave function methods with den- sity functional theory for excited states. Chem. Rev.2018,118, 7249–7292. (95) Pernal, K. Electron Correlation from the Adiabatic Connection for Multireference Wave Functions.Phys. Rev. Lett.2018, 120, 013001. (96) Langreth, D.; Perdew, J. E...

    work page 2018

  12. [12]

    M.; Coriani, S.; Helgaker, T

    (97) Teale, A. M.; Coriani, S.; Helgaker, T. Accurate calculation and modeling of the adiabatic connection in density func- tional theory.J. Chem. Phys.2010,132, 164115. (98) Ángyán, J. G.; Liu, R.-F.; Toulouse, J.; Jansen, G. Correlation Energy Expres- sions from the Adiabatic-Connection Fluctuation Dissipation Theorem Ap- proach.J. Chem. Theory Comput.2...

    work page arXiv 2010

  13. [13]

    (126) Davidson, E.Reduced density matrices in quantum chemistry; Elsevier, 2012; Vol

    work page 2012

  14. [14]

    Normal order and extended Wick theorem for a multiconfiguration reference wave func- tion.J

    (127) Kutzelnigg, W.; Mukherjee, D. Normal order and extended Wick theorem for a multiconfiguration reference wave func- tion.J. Chem. Phys.1997,107, 432–449. (128) Colmenero, F.; Del Valle, C. P.; Valde- moro, C. Approximating q-order reduced density matrices in terms of the lower- order ones. I. General relations.Phys. Rev. A1993,47,

    work page 1997

  15. [15]

    P.; Ernzerhof, M.; Burke, K.; Savin, A

    (130) Perdew, J. P.; Ernzerhof, M.; Burke, K.; Savin, A. On-top pair-density interpre- tation of spin density functional theory, with applications to magnetism.Int. J. Quant. Chem.1997,61, 197–205. (131) Stoll, H.; Savin, A. InDensity Func- tional Methods in Physics; Dreizler, R., da Providencia, J., Eds.; Plenum, New York, 1985; pp 177–207. (132) Pollet,...

    work page 1997

  16. [16]

    Combining long-range configu- ration interaction with short-range den- sity functionals.Chem

    (133) Leininger, T.; Stoll, H.; Werner, H.-J.; Savin, A. Combining long-range configu- ration interaction with short-range den- sity functionals.Chem. Phys. Lett.1997, 275, 151–160. (134) Gori-Giorgi, P.; Savin, A. Properties of short-range and long-range correlation energy density functionals from electron- electron coalescence.Phys. Rev. A2006, 73, 0325...

    work page 1997

  17. [17]

    K.; Luo, S.; Ma, D.; Olsen, J.; Truhlar, D

    (145) Li Manni, G.; Carlson, R. K.; Luo, S.; Ma, D.; Olsen, J.; Truhlar, D. G.; Gagliardi, L. Multiconfiguration pair- density functional theory.J. Chem. The- ory Comput.2014,10, 3669–3680. (146) Scott, M.; Rodrigues, G. L.; Li, X.; Del- cey, M. G. Variational pair-density func- tional theory: Dealing with strong corre- lation at the protein scale.J. Chem...

    work page 2014

  18. [18]

    D.; Ball, M

    (155) McLachlan, A. D.; Ball, M. A. Time- Dependent Hartree-Fock Theory for Molecules.Rev. Mod. Phys.1964,36,

    work page 1964

  19. [19]

    Exchange-correlation energy from pair- ing matrix fluctuation and the particle- particle random-phase approximation

    (156) Van Aggelen, H.; Yang, Y.; Yang, W. Exchange-correlation energy from pair- ing matrix fluctuation and the particle- particle random-phase approximation. Phys. Rev. A2013,88, 030501. (157) van Aggelen, H.; Yang, Y.; Yang, W. Exchange-correlation energy from pair- ing matrix fluctuation and the particle- particle random phase approximation.J. Chem. Ph...

    work page 2014

  20. [20]

    Spinless formula- tion of linearized adiabatic connection approximation and its comparison with the second order N-electron valence state perturbation theory.Faraday Discuss

    (161) Guo, Y.; Pernal, K. Spinless formula- tion of linearized adiabatic connection approximation and its comparison with the second order N-electron valence state perturbation theory.Faraday Discuss. 2024,254, 332–358. (162) Drwal, D.; Beran, P.; Hapka, M.; Mod- rzejewski, M.; Sokół, A.; Veis, L.; Per- nal, K. Efficient adiabatic connection ap- proach fo...

    work page 2024

  21. [21]

    J.; Knizia, G.; Manby, F

    (167) Werner, H.-J.; Knowles, P. J.; Knizia, G.; Manby, F. R.; Schütz, M. Molpro: a general-purpose quantum chemistry pro- gram package.Wiley Interdiscip. Rev.: Comput. Mol. Sci.2012,2, 242–253. (168) Sun, Q.; Zhang, X.; Banerjee, S.; Bao, P.; Barbry, M.; Blunt, N. S.; Bog- danov, N. A.; Booth, G. H.; Chen, J.; Cui, Z.-H. et al. Recent developments in the...

    work page 2012

  22. [22]

    L.; Bakken, V.; Bast, R.; Boman, L.; Chris- tiansen, O.; Cimiraglia, R.; Coriani, S.; Dahle, P

    (172) Aidas, K.; Angeli, C.; Bak, K. L.; Bakken, V.; Bast, R.; Boman, L.; Chris- tiansen, O.; Cimiraglia, R.; Coriani, S.; Dahle, P. et al. The Dalton quantum chemistryprogramsystem.WIREs Com- put. Mol. Sci.2014,4, 269–284. (173) Saito, T.; Nishihara, S.; Yamanaka, S.; Kitagawa, Y.; Kawakami, T.; Ya- mada, S.; Isobe, H.; Okumura, M.; Yam- aguchi,K.Symmetr...

    work page 2014

  23. [23]

    J.; Shen, J.; Ajala, A

    (175) Stoneburner, S. J.; Shen, J.; Ajala, A. O.; Piecuch, P.; Truhlar, D. G.; Gagliardi, L. Systematic design of active spaces for multi-reference calculations of singlet– triplet gaps of organic diradicals, with benchmarks against doubly electron- attached coupled-cluster data.J. Chem. Phys.2017,147, 164120. (176) Schreiber, M.; Silva-Junior, M. R.; Sau...

    work page 2017

  24. [24]

    M.; Guo, Y.; Riplinger, C.; Tuczek, F.; Neese, F

    (181) Floser, B. M.; Guo, Y.; Riplinger, C.; Tuczek, F.; Neese, F. Detailed pair nat- ural orbital-based coupled cluster stud- ies of spin crossover energetics.J. Chem. Theory Comput.2020,16, 2224–2235. (182) Pierloot, K. The CASPT2 method in inorganic electronic spectroscopy: From ionictransitionmetaltocovalentactinide complexes.Mol. Phys.2003,101, 2083–

    work page 2020

  25. [25]

    B.; Peterson, K

    (183) Balabanov, N. B.; Peterson, K. A. Ba- sis set limit electronic excitation en- ergies, ionization potentials, and elec- tron affinities for the 3d transition metal atoms: Coupled cluster and multiref- erence methods.J. Chem. Phys.2006,

    work page 2006

  26. [26]

    A.; Harrison, R

    (184) De Jong, W. A.; Harrison, R. J.; Dixon, D. A. Parallel Douglas–Kroll en- ergy and gradients in NWChem: Es- timating scalar relativistic effects using Douglas–Kroll contracted basis sets.J. Chem. Phys.2001,114, 48–53. (185) Reiher, M.; Wolf, A. Exact decoupling of the Dirac Hamiltonian. I. General theory. J. Chem. Phys.2004,121, 2037–2047. (186) Domi...

    work page 2001

  27. [27]

    Approximation to Second Order N-Electron Valence State Perturbation Theory: Limiting the Wave Function within Singles.J

    (198) Guo, Y.; Pernal, K. Approximation to Second Order N-Electron Valence State Perturbation Theory: Limiting the Wave Function within Singles.J. Chem. The- ory Comput.2025,21, 5545–5558. (199) Hapka, M.; Pastorczak, E.; Pernal, K. Self-Adapting Short-Range Correlation Functional for Complete Active Space- Based Approximations.J. Phys. Chem. A2024,128, 7...

    work page 2025