Spin-adapted open-shell time-dependent density functional theory: towards a simple and accurate method for spin-flip-down excitations
Hewang Zhao and Zhendong Li · 2025
Read Spin-adapted open-shell time-dependent density functional theory: towards a simple and accurate method for spin-flip-down excitations by Hewang Zhao and Zhendong Li (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-16906.txt.
Introduction
Hewang Zhao and Zhendong Li introduce XSF-TDA, a method for computing spin-flip excitations of open-shell molecules without the spin contamination that plagues spin-flip time-dependent density functional theory. Flipping a spin during excitation is the natural route to bond breaking states and inverted singlet-triplet gaps, but contaminated answers are hard to trust.
The method grows out of the authors' earlier spin-adapted open-shell theory for spin-conserving excitations. The paper compares strategies for suppressing contamination across unrestricted and mixed-reference variants of the spin-flip approach. Pilot calculations on prototype systems show XSF-TDA describing bond breaking and inverted gap systems more faithfully than its predecessors. It is a step toward a spin-flip method that is both simple enough for routine use and accurate enough for chemistry's hard cases.
Catalog record
| CREATOR | Hewang Zhao and Zhendong Li |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-16906
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Spin-adapted open-shell time-dependent density functional theory: towards a simple and accurate method for spin-flip-down excitations Hewang Zhao and Zhendong Li Abstract A major challenge in using spin-flip time-dependent density functional theory (SF-TD-DFT) for spin-flip-down excitations is the presence of spin contamination. While several improved methods have been developed in the past, a simple and accurate method remains elusive. Here, based on our previous development on spin-adapted open-shell TD-DFT for spin-conserving excitations (X-TD-DFT) [Z. Li and W. Liu, J. Chem. Phys. 135, 194106 (2011)], we introduce a method termed as XSF-TDA for modeling spin-flip-down excitations, and provide an in-depth comparison of different methodologies for mitigating spin contamination in SF-TDA. Pilot applications to prototype systems demonstrate the promise of XSF-TDA over existing SF-TDA methods, including unrestricted SF-TDA (USF-TDA) and mixed-reference SF-TDA (MRSF-TDA), in describing bond breakings and inverted singlet-triplet gap systems. keywords spin-flip; time-dependent density functional theory; spin contamination; spin adaptation † † affiliation: Key Laboratory of Theoretical and Computational Photochemistry, Ministry of Education, College of Chemistry, Beijing Normal University, Beijing 100875, China 1 Introduction Spin-flip time-dependent density functional theory (SF-TD-DFT) [ 1 ] , which extends the earlier spin-flip framework [ 2 ] to TD-DFT, has garnered significant attention for its ability to describe electronic transitions beyond the reach of conventional linear response spin-conserving TD-DFT. For instance, it can be used to study spin-flip transitions in spin-flip emitters [ 3 , 4 ] , which represent a distinct class of luminescent materials that exhibit unique photophysical properties due to spin‐forbidden yet radiative transitions. Apart from spin-flip transitions, SF-TD-DFT is also capable of capturing certain double excitations, single-bond dissociation processes, and the correct topology of conical intersections [ 5 , 6 , 7 ] , which are notoriously challenging for DFT and spin-conserving TD-DFT within the adiabatic approximation [ 8 ] . The initial implementation of SF-TD-DFT by Shao et al. [ 1 ] employed a collinear hybrid exchange–correlation (XC) kernel within the Tamm–Dancoff approximation (TDA). This formulation necessitates the use of hybrid XC functionals; pure density functionals would otherwise yield excitation energies that are simply orbital energy differences. Subsequently, Wang and Ziegler introduced a noncollinear XC kernel [ 9 , 10 ] , which was originally proposed in the context of relativistic TD-DFT for spinor excitations [ 11 , 12 , 13 ] , within the local density approximation (LDA) to describe spin-flip excitations in a nonrelativistic framework. Later, full noncollinear XC kernels were implemented for generalized gradient approximation (GGA) and hybrid functionals, both within TDA [ 14 , 15 ] and in full SF-TD-DFT [ 16 ] . However, numerical instabilities can arise with noncollinear GGA kernels. To address this, an adiabatic local density approximation (ALDA0) [ 16 , 17 ] was introduced, neglecting density gradients in the XC kernel evaluation to enhance numerical stability. More recently, alternative strategies based on a multicollinear formulation have been proposed [ 18 , 19 , 20 ] . A major drawback of SF-TD-DFT is the presence of excited-state spin contamination [ 21 , 22 ] , which arises from the incompleteness of the excitation manifold. We define a subspace as spin complete if it contains all the determinants required to form spin eigenfunctions, and a method as spin-adapted if all its configurations are spin eigenfunctions. Note a subtle distinction: spin-adaptation requires that only spin eigenfunctions are included, but does not necessitate inclusion of all the configurations within a spin-complete subspace. To illustrate the spin contamination issue in SF-TD-DFT, consider a triplet reference system as depicted in Fig. 1 . The orbitals of the high-spin open-shell state can be categorized into closed-shell (C), open-shell (O), and virtual-shell (V). We denote a single excitation from an \alpha core orbital to a \beta virtual orbital as CV( \alpha\beta ). It is easy to find that only OO( \alpha\beta )-type spin-flip-down excitations generate a spin-complete subspace. In contrast, the other three types - CV( \alpha\beta ), OV( \alpha\beta ), and CO( \alpha\beta ) - result in spin contamination due to the spin incompleteness of their generated configurations. The determinants shown in red in Fig 1 , which are higher excitations than singles with respect to the reference, are necessary to achieve spin completeness/adaptation. Within wavefunction theory, methods such as spin-complete spin-flip configuration interaction singles (SC-SF-CIS) [ 23 ] have been developed to address this issue by directly incorporating the missing higher excitations. However, adapting such approaches to TD-DFT is challenging due to the absence of an explicit wavefunction. Although propagator corrections [ 21 ] and generalized excitation operators [ 24 ] have been proposed to include higher-order excitations in TD-DFT, these methods have not seen practical application. Figure 1: Illustration of the excited-state spin contamination problem in SF-TD-DFT/SF-TDA due to the incompleteness of the excited state manifold generated by spin-flip-down single excitations. Configurations in red color represent the missing higher excitations with respect to the reference to achieve spin completeness/adaptation. Building on the insight that the missing configurations for spin completeness can be accessed from other components of the open-shell reference, one of the present authors developed [ 22 ] a spin-adapted random phase approximation (S-RPA) and time-dependent density functional theory (S-TD-DFT) using the tensor equation-of-motion (TEOM) formalism originally introduced in nuclear physics [ 25 ] . A key feature of S-RPA is that only single excitations explicitly enter the final matrix elements, while the contributions from higher excitations required for spin-adaptation are implicitly incorporated via the Wigner-Eckart theorem [ 22 ] . This crucial property allows for a direct translation of the matrix elements into the TD-DFT framework, leading to S-TD-DFT [ 26 ] without the need to go beyond the adiabatic approximation. Subsequently, a more accurate and simpler version, termed X-TD-DFT [ 27 ] , was introduced to better account for spin degeneracy conditions and to eliminate double counting of correlation for those already spin-adapted states. X-TD-DFT has been shown to achieve accuracy comparable to conventional closed-shell TD-DFT for spin-conserving excitations [ 28 ] . For spin-flip-down excitations, a similar strategy can in principle be applied to achieve spin adaptation, as proposed in the original work [ 22 ] and later implemented by Zhang and Herbert [ 5 ] and Chima and Visscher [ 29 ] . However, addressing double counting of correlation in this context is more challenging, as spin contamination affects three types of excitations, and adding the effects of higher excitations for spin-adaptation may deteriorate the excited states dominated by the OO( \alpha\beta ) type of excitations, which are already spin-complete. Another relevant approach is the mixed-reference SF-TD-DFT within TDA (MRSF-TDA) proposed by Lee et al. [ 30 , 31 ] , which uses only the M_{S}=\pm 1 components of a triplet reference to mitigate spin contamination, but is not applicable to other spin states. In summary, a general, simple, and accurate spin-adapted method for spin-flip-down excitations is still lacking. In this work, we present a method termed as XSF-TDA, which extends our previous developments in spin-adapted TD-DFT [ 22 , 26 , 27 ] to spin-flip-down excitations. We introduce a practical approach to alleviate the double counting of correlation resulting from the implicit inclusion of high-order excitations for spin adaptation. Pilot applications on small systems, where high-level theoretical or experimental reference