WorksPhysics and Astronomy
Mechanism of Superconductivity in the Hubbard Model at Intermediate Interaction Strength
Abstract
We study the fluctuations responsible for pairing in the -wave superconducting state of the two-dimensional Hubbard model at intermediate coupling within a cluster dynamical mean-field theory with a numerically exact quantum impurity solver. By analyzing how momentum and frequency dependent fluctuations generate the wave superconducting state in different representations, we identify antiferromagnetic fluctuations as the pairing glue of superconductivity both in the underdoped and the overdoped regime. Nevertheless, in the intermediate coupling regime, the predominant magnetic fluctuations may differ significantly from those described by conventional spin-fluctuation theory.
Introduction
The microscopic mechanism of unconventional high-temperature superconductivity has been one of the most controversially debated topics in condensed matter physics since the discovery of superconductivity in layered copper-oxides in 1986. While several aspects of the observed physics, such as the -wave symmetry of the order parameter and the proximity to an antiferromagnetic Mott phase, clearly suggest that superconductivity must emerge from strongly correlated electronic processes, the intrinsic quantum many-body nature of the problem has hitherto prevented a rigorous identification of the pairing glue. To explicitly address this point, we present a focused study of the origin of superconductivity in the two dimensional single band Hubbard model. The Hubbard Hamiltonian, which includes a kinetic term describing the hopping between neighbouring sites on a lattice and a potential-energy term encoding a local electrostatic repulsion, is a minimal theoretical model believed to capture the salient aspects of cuprate superconductivity. To provide a conclusive answer, we perform an analysis of the anomalous self energy in the -wave superconducting state within the method of fluctuation diagnostics [14]. We note that, unlike other diagrammatic approaches, which postulate a specific physical mechanism, analyze its consequences, and then compare to experiments, the fluctuation diagnostics procedure treats fluctuations of all kinds, including those possibly driving superconductivity, on equal footing, and is applicable in all parameter regimes, independent of the degree of correlation. However, the fluctuation diagnostics procedure as derived in [14] was only applicable to the highly symmetric normal state, and thus cannot be used to analyze superconductivity. Hence, we will first generalize this approach to the case of phases with spontaneously broken symmetries and then apply it to identify the dominant fluctuations driving the anomalous self energy in the superconducting state.
Among the theoretical explanations proposed for the origin of the high-temperature superconductivity in this context, spin fluctuations have been a prominent scenario since the beginning [1, 2, 3, 4]. In particular, within in the weak-coupling regime of the Hubbard model, renormalization group techniques [5, 6, 7, 8] find -wave superconductivity in qualitative agreement with spin-fluctuation exchange studies [1, 8], consistent with diagrammatic Monte Carlo calculations [9]. At the same time, other qualitatively different microscopic pictures of superconductivity exist besides the spin fluctuations, including the RVB theory [10], nematic fluctuations [11], loop current order [12], or the “intertwining” of orders of different types [13]. In fact, to what extent the weak-coupling spin fluctuation results apply to the much stronger interaction values, which are typical of cuprate materials, and whether there are other competing or intertwining fluctuations driving the superconductivity remains unresolved.
Method
The Hamiltonian of the two-dimensional single band Hubbard model is
with a lattice site, momentum, annihilation (creation) operators, and the density. is the dispersion with hopping , the interaction strength, and the chemical potential. We use the dynamical cluster approximation (DCA) [15] on a cluster with size with a numerically exact continuous time auxiliary field (CTAUX) [16, 17] impurity solver to enter the superconducting state non-perturbatively [18], and obtain Green’s functions, self-energies, and vertex functions.
In the DCA, the momentum structure of the Hubbard model self-energy is approximated by basis functions which retain the full frequency dependence [15]. To enter the superconducting phase, we allow for order and provide a superconducting bias field at the first iteration, removing it in subsequent iterations to converge to the equilibrium state. Within the 8-site DCA approximation, the model exhibits a large and stable d-wave superconducting region [18, 19]. The model is known to also exhibit a stripe phase [20, 21, 22, 23, 24], to which our calculation is not sensitive, since its periodicity is larger than the cluster.
To identify the superconducting glue, we apply the fluctuation diagnostics scheme [14] to the anomalous self-energy in the superconducting state. This approach, which so far has been derived [14] and applied [14, 25] only in the paramagnetic normal state, allows for a rigorous identification of the dominant scattering mechanisms responsible for the observed self-energy. Fluctuation diagnostics exploits symmetries in the Hamiltonian that lead to different expressions for the Schwinger-Dyson (SD) equation of the self energy
where we use the Einstein summation notation and introduce the short-hand notation for momentum, spin and time indices. The two indices on the self energy and the Green’s functions represent their normal () and anomalous () components, with . is the static Hartree contribution for the normal self energy, and is the full two-electron scattering amplitude. denotes the antisymmetrized interaction which, in the Hubbard model, is proportional to the local interaction . This expression is exact and relates two-particle fluctuations to single-particle quantities.
Panel (a) of Figure 1 shows a diagrammatic representation of the anomalous self-energy of Eq. (2) in frequency space, with representing fermionic and representing bosonic indices. A choice of , , , , , , , , and , satisfies momentum, energy, and spin conservations. Panel (b) shows two low-order terms. , in the case of superconducting order, can have eight possible combinations of incoming or outgoing legs, as illustrated in panel (c). It then contains all allowed scattering processes, some of which are illustrated in panel (d).
Figure 1. Anomalous self-energy diagrams. Panel a: Schwinger Dyson equation (Eq. (2)). Shaded box denotes vertex with one fixed outgoing leg. Dotted lines: normal or anomalous Green’s function. Panel b: Low-order anomalous self energy diagrams. Panel c: Explicit representation of index combinations of . Panel d: Some of the low-order diagrams contributing to .
The different expressions of the SD are equivalent when all internal summations are performed. Important information about the role played by the different scattering channels can be gained by comparing the expressions of the SD equation after partial summations over the internal variables .
From a physical point of view, each expression can be associated to one of the possible collective modes (e.g., density, magnetic, singlet/triplet pairing) of the electronic system. A large contribution to the final sum over at low transfer frequency and at a definite momentum signifies a dominant collective mode, in contrast to contributions more evenly distributed over a wide range of frequencies and momenta.
We now derive Eq. (2) in more detail. To have access to all the single- and two-particle quantities in the SD equation, we introduce the Bogoliubov-de Gennes spinors in momentum space [26]
The single-particle Green’s function in the singlet superconducting state is
where is the (imaginary) time-ordering operator. SU(2) symmetry reduces the number of independent terms in Eq. (4) to four, such that Eq. (4) can be written in a compact form as
The single particle Green’s function matrix contains both normal (N) and anomalous (A) entries, with the normal terms defined as , , and the anomalous terms defined as , . Fourier transforming to frequency space and introducing short hand notations , we can define . For d-wave superconductivity on a lattice with inversion symmetry, the anomalous Green’s function can be chosen to be real [27], such that . The self energy can then be computed with
where
The two-particle Green’s function takes the form
where we use as a short-hand notation for momentum, spin and imaginary time indices . is either a creation operator or an annihilation operator .
In the paramagnetic state, the number of creation and annihilation operators in the two-particle Green’s function must be equal to preserve charge conservation. In the superconducting state, the broken U(1) symmetry gives in total combinations of different creation or annihilation operators, which can be written in matrix form as
with , .
Each term in the two-particle Green’s function matrix can be decomposed into connected () and disconnected parts as
This relation can be written in matrix form
where includes terms of the form , and includes terms of the form and . The full vertex can be computed from the connected part of the two-particle Green’s function [28, 29]
Due to momentum and energy conservation, and following the particle-hole convention [29] of the Fourier transform, the momentum and frequency indices in the matrix above can be assigned as , , , and , with short hand notations for fermionic and for bosonic indices. Three independent spin combinations are possible in a SU(2) symmetric system, ,
and , with or [29]. We can then define
where quantities in Eq. (14b) can be obtained from those in Eq. (14a) via SU(2) and crossing symmetries [29]. Introducing the full vertex matrix , Eq. (13) can be written in matrix form as
In order to extend fluctuation diagnostics to the superconducting state, we identify all independent scattering channels of the symmetry broken state and derive the corresponding equivalent expressions of the SD equation. In the basis of BdG spinors, the creation of pairs of particles and holes in the spin singlet state is then described by the matrices , [ being the Pauli matrices, ], while the corresponding terms for the triplet state are given by , . The two matrices and [with the projector in the particle (hole) subspace] define the density operator of particles and holes. Analogously, and yield the spin operator, from which magnetic fluctuations originate. With these definitions, the two-particle Green’s function in different physical channels can be defined as
where the two time-dependent operators on the r.h.s. are defined as , with corresponding to one of the eight matrices , , , , , , , . All terms defined by Eq. (17) can be computed from the linear combination of terms in and . The physical channels of the full vertex function can be defined with the same linear combinations as the two-particle Green’s function.
The underlying symmetries of the system block-diagonalize the matrix , such that each block identifies one independent scattering channel [30]. In the paramagnetic state, where both the (global) symmetry and the symmetry hold, only terms that conserve spin and particle number will be non-zero [29] giving rise to four independent scattering channels: define the density, the magnetic, the singlet-pairing, and the triplet-pairing channel.
In the superconducting state, the spontaneously broken symmetry allows for processes violating particle number conservation, and mix different scattering channels that are independent in the paramagnetic state. However, the symmetry still holds and allows to identify two independent scattering channels: the density/spin-singlet channel and the magnetic/spin-triplet channel . Hence, in the superconducting phase Eq. 2 can be rewritten in two equivalent ways as
with
Eqs. 18 and 19 are the decompositions that enable application of the fluctuation diagnostics scheme to the superconducting state.
Results
Figure 2 gives the phase diagram of the 2D Hubbard model on the hole-doped side within the DCA approximation at intermediate interaction strength [18] showing the pseudogap (PG), superconducting (SC), and metallic (M) regime. The pseudogap regime is characterized by a suppression of the single particle spectral function, and the superconducting phase corresponds to the region where the anomalous Green’s function is non-zero. We present the results for two representative parameter sets without next-nearest neighbor hopping on an eight-site cluster with , , i.e. (, corresponding to UD for this value of ) and (, corresponding to OD for this value of ), see Ref. 18 for a phase diagram. In DCA, both cases considered lie deep in the superconducting phase where the anomalous Green’s function is non-zero for and , with relation . Cluster momentum points are shown in the inset of Figure 3. The inset of Figure 2 shows the momentum () and frequency () distribution of , which is computed by summing over fermionic indices but not over in Eqs. 18, 19. The pie chart insets show that for both UD and OD there is a dominant contribution from and in the magnetic/triplet channel . In the density/singlet channel , contributions from different momenta and frequencies are evidently distributed much more evenly.

Figure 2. Phase diagram sketch, with pseudo-gap (PG, orange), metal (M, red), and superconducting (SC, blue) regimes. Black diamonds denote UD and OD data points analyzed in detail. Inset: Pie chart of in the density () and magnetic () channels. Counter-clockwise from the top, pieces represent contributions for momentum , and summation over the remaining momenta in an eight-site cluster. In each slice, separation indicates bosonic frequency with .

Figure 3. for several transfer momenta in the density () and magnetic () channels. OD and UD correspond to the diamonds in Figure 2. Black line with crosses: total anomalous self-energy after summation over all . Inset: non-interacting Fermi surface and location of momentum points corresponding to colors in main panels.
We first focus on the momentum distribution of within the two physical channels in Figure 3, computed by summing over all indices in Eq. 18, 19 except for the transferred momentum . The inset in the upper left panel shows the momentum points in an eight-site cluster, and the Fermi surface in the non-interacting system for dopings of 0.2 (corresponding to hole doping), 0, -0.2, and -0.4 (corresponding to electron doping). The left two panels show the contribution of different in the density channel. The weak dependence indicates the absence of a dominant mode in this channel. Results for the magnetic/triplet channel are shown in the right two panels. The transfer momentum associated with AFM fluctuations is clearly the dominant mode in both the UD and the OD regime. We note that a sub-leading, though still sizable, negative contribution to the anomalous self energy is originated by a ferromagnetic mode with . The black lines with crosses are computed by summing over all different momenta in the cluster, resulting in and of Eq. 17.
Important insight can be gained by a complementary analysis in frequency space: Figure 4 shows the frequency dependence of , corresponding to the result at the lowest fermionic Matsubara frequency
Figure 4. Re in density () and magnetic () channels. OD and UD correspond to the two data points of Figure 2. Inset: Re in density channel with rescaled axis.
i after summation over all indices except for bosonic frequency in Eq. 18, 19. The low-frequency peak in the magnetic channel identifies the corresponding fluctuation as a well-defined and long lived “mode”. In the density channel, the contributions are more evenly distributed across a wide frequency range, corresponding to short lived fluctuations.
Discussion and Conclusion
By extending the fluctuation diagnostics approach to the superconducting phase, we have been able to unambiguously identify spin-fluctuations [1, 2, 3, 4] as the dominant contribution to the –wave pairing in the Hubbard model at interaction strengths believed to be relevant for the cuprates, i.e., beyond the weak-coupling regime [1, 6, 8, 9]. At the same time, consistent with the existing work in the normal state [14, 31, 32], we do not find any indication supporting the alternative scenarios mentioned in the introduction, such as nematic fluctuations [11], loop current order [12], or “intertwining” of different orders [13]. In the latter case, multiple fluctuations such as density or magnetic ones would contribute synergistically to the pairing, rather than compete, in contradiction with our results. We emphasize that the fluctuation diagnostic is capable of detecting the occurrence of this situation, when it is realized, e.g. in the attractive Hubbard model [14].
Our identification of the superconducting glue agrees with the findings of several experiments. Ref. 33 finds good quantitative agreement between the spectral function computed from conventional spin fluctuation theory with magnetic susceptibility measured by inelastic neutron scattering, and the spectral function measured from angle-resolved photoemission spectroscopy in the superconducting phase of YBCO. Inelastic photon scattering experiments [34] on Hg1201 and Hg1212 infer that the superconducting temperature can be determined by the strength of the magnetic interactions (“paramagnon signals”), supporting the theory of magnetically mediated high-temperature superconductivity. Other experiments suggest a relation between superconductivity and charge density wave [35], or that the pseudogap and superconductivity may have different origins [36]. Thus, the numerical findings of our study suggest the possibility that the latter class of experiments may be probing aspects of cuprates physics beyond those encoded in the single-orbital Hubbard model on an eight-site DCA cluster.
Independently of the agreement with this multifaceted experimental evidence, our identification of the superconducting glue in terms of spin-fluctuations touches a delicate and important aspect of the theoretical description of high- superconductivity. In particular, conventional spin fluctuation theory as described in Refs. 37–40 appears only able to capture a fraction of the pairing contribution [41]. The origin of this discrepancy can be ascribed to the RPA-like one-loop spin fluctuations expressions used in conventional approaches which, outside of the weak-coupling regime, do not capture all spin-fluctuation mediated processes [42].
The microscopic picture of superconductivity emerging from our analysis agrees well with recent studies of the description of the non-superconducting pseudogap regime: While spin-fluctuations were identified as the predominant mechanism of the pseudogap [14, 31, 32], differences with respect to the predictions of conventional spin-fluctuation theory were found and traced [43] to the imaginary part of the dynamical scattering amplitude between electrons and spin fluctuations, which is absent in conventional approaches [43].
In conclusion, our fluctuation diagnostics of the superconducting order in the Hubbard model precisely identifies antiferromagnetic spin fluctuations as the glue of the –wave pairing. This conclusion applies to the intermediate-to-large values of the electronic interaction relevant to cuprate physics. For this reason, the spin-fluctuations-driven pairing found in our calculations are expected to differ from conventional spin-fluctuation theories.
Acknowledgments
XD and EG are supported by NSF DMR 2001465. LD acknowledges DE-SC0019469. AT acknowledges financial support from the Austrian Science Fund (FWF) through the Project I 5868-N – FOR 5249 (QUAST). We thank A.J. Millis, Kai Sun and F. Krien for detailed and insightful discussions.
References
- [1]D. J. Scalapino, E. Loh, and J. E. Hirsch, d-wave pairing near a spin-density-wave instability, Phys. Rev. B 34, 8190 (1986).pubmed.ncbi.nlm.nih.gov/9939525
- [2]K. Miyake, S. Schmitt-Rink, and C. M. Varma, Spin-fluctuation-mediated even-parity pairing in heavy-fermion superconductors, Phys. Rev. B 34, 6554 (1986).pubmed.ncbi.nlm.nih.gov/9940551
- [3]D. J. Scalapino, Superconductivity and spin fluctuations, Journal of Low Temperature Physics 117, 179 (1999).DOI
- [4]D. J. Scalapino, A common thread: The pairing interaction for unconventional superconductors, Rev. Mod. Phys. 84, 1383 (2012).DOI
- [5]C. J. Halboth and W. Metzner, d-wave superconductivity and pomeranchuk instability in the two-dimensional hubbard model, Phys. Rev. Lett. 85, 5162 (2000).DOI
- [6]D. Zanchi and H. J. Schulz, Weakly correlated electrons on a square lattice: Renormalization-group theory, Phys. Rev. B 61, 13609 (2000).
- [7]C. Honerkamp, M. Salmhofer, N. Furukawa, and T. M. Rice, Breakdown of the landau-fermi liquid in two dimensions due to umklapp scattering, Phys. Rev. B 63, 035109 (2001).
- [8]S. Raghu, S. A. Kivelson, and D. J. Scalapino, Superconductivity in the repulsive hubbard model: An asymptotically exact weak-coupling solution, Phys. Rev. B 81, 224505 (2010).DOI
- [9]Y. Deng, E. Kozik, N. V. Prokof'ev, and B. V. Svistunov, Emergent BCS regime of the two-dimensional fermionic hubbard model: Ground-state phase diagram, EPL (Europhysics Letters) 110, 57001 (2015).
- [10]P. W. Anderson, The resonating valence bond state in La₂CuO₄ and superconductivity, Science 235, 1196 (1987).pubmed.ncbi.nlm.nih.gov/17818979
- [11]E. Fradkin, S. A. Kivelson, M. J. Lawler, J. P. Eisenstein, and A. P. Mackenzie, Nematic fermi fluids in condensed matter physics, Annual Review of Condensed Matter Physics 1, 153 (2010).
- [12]A. Allais and T. Senthil, Loop current order and d-wave superconductivity: Some observable consequences, Phys. Rev. B 86, 045118 (2012).DOI
- [13]E. Fradkin, S. A. Kivelson, and J. M. Tranquada, Colloquium: Theory of intertwined orders in high temperature superconductors, Rev. Mod. Phys. 87, 457 (2015).DOI
- [14]O. Gunnarsson, T. Schäfer, J. P. F. LeBlanc, E. Gull, J. Merino, G. Sangiovanni, G. Rohringer, and A. Toschi, Fluctuation diagnostics of the electron self-energy: Origin of the pseudogap physics, Phys. Rev. Lett. 114, 236402 (2015).pubmed.ncbi.nlm.nih.gov/26196809
- [15]T. Maier, M. Jarrell, T. Pruschke, and M. H. Het tler, Quantum cluster theories, Rev. Mod. Phys. 77, 1027 (2005).DOI
- [16]E. Gull, P. Werner, O. Parcollet, and M. Troyer, Continuous-time auxiliary-field monte carlo for quantum impurity models, Euro. phys. Lett. 82, 57003 (2008).
- [17]E. Gull, A. J. Millis, A. I. Lichtenstein, A. N. Rubtsov, M. Troyer, and P. Werner, Continuous-time monte carlo methods for quantum impurity models, Rev. Mod. Phys. 83, 349 (2011).DOI
- [18]E. Gull, O. Parcollet, and A. J. Millis, Superconductivity and the pseudogap in the two-dimensional hubbard model, Phys. Rev. Lett. 110, 216405 (2013).pubmed.ncbi.nlm.nih.gov/23745902
- [19]X. Chen, J. P. F. LeBlanc, and E. Gull, Superconducting fluctuations in the normal state of the two-dimensional hubbard model, Phys. Rev. Lett. 115, 116402 (2015).DOI
- [20]B.-X. Zheng, C.-M. Chung, P. Corboz, G. Ehlers, M.-P. Qin, R. M. Noack, H. Shi, S. R. White, S. Zhang, and G. K.-L. Chan, Stripe order in the underdoped region of the two-dimensional hubbard model, Science 358, 1155 (2017).DOI
- [21]F. W. Huang, C. B. Mendl, H.-C. Jiang, B. Moritz, and T. P. Devereaux, Stripe order from the perspective of the hubbard model, npj Quantum Materials 3, 22 (2018).DOI
- [22]M. Qin, C.-M. Chung, H. Shi, E. Vitali, C. Hubig, U. Schollwöck, S. R. White, and S. Zhang (Simons Collaboration on the Many-Electron Problem), Absence of superconductivity in the pure two-dimensional hubbard model, Phys. Rev. X 10, 031016 (2020).DOI
- [23]P. Mai, S. Karakuzu, G. Balduzzi, S. Johnston, and T. A. Maier, Intertwined spin, charge, and pair correlations in the two-dimensional hubbard model in the thermodynamic limit, Proceedings of the National Academy of Sciences 119, e2112806119 (2022).DOI
- [24]A. Wietek, Y.-Y. He, S. R. White, A. Georges, and E. M. Stoudenmire, Stripes, antiferromagnetism, and the pseudogap in the doped hubbard model at finite temperature,
- [25]T. Schaefer and A. Toschi, How to read between the lines of electronic spectra: the diagnostics of fluctuations in strongly correlated electron systems, Journal of Physics: Condensed Matter 33, 214001 (2021).
- [26]J. Zhu, Bogoliubov-de Gennes Method and Its Applications (2016).DOI
- [27]A. I. Lichtenstein and M. I. Katsnelson, Antiferromagnetism and d-wave superconductivity in cuprates: A cluster dynamical mean-field theory, Phys. Rev. B 62, R9283 (2000).DOI
- [28]J. Negele and H. Orland, Quantum Many-Particle Systems (1998).
- [29]G. Rohringer, A. Valli, and A. Toschi, Local electronic correlation at the two-particle level, Phys. Rev. B 86, 125114 (2012).DOI
- [30]L. Del Re and A. Toschi, Dynamical vertex approximation for many-electron systems with spontaneously broken su(2) symmetry, Phys. Rev. B 104, 085120 (2021).
- [31]W. Wu, M. Ferrero, A. Georges, and E. Kozik, Controlling fermion diagrammatic expansions: Physical nature of the pseudogap in the two-dimensional hubbard model, Phys. Rev. B 96, 041105 (2017).
- [32]X. Dong, X. Chen, and E. Gull, Dynamical charge susceptibility in the hubbard model, Phys. Rev. B 100, 235107 (2019).DOI
- [34]L. Wang, G. He, Z. Yang, M. Garcia-Fernandez, A. Nag, K.-J. Zhou, M. Minola, M. L. Tacon, B. Keimer, Y. Peng, and Y. Li, Paramagnons and high-temperature superconductivity in mercury-based cuprates (2020), arXiv:2011.05029 [cond-mat.supr-con].DOI
- [35]H. Chu, S. Kovalev, Z. X. Wang, L. Schwarz, T. Dong, L. Feng, R. Haenel, M.-J. Kim, P. Shabestari, H. L. Phuong, K. Honasoge, R. D. Dawson, D. Putzky, G. Kim, M. Puviani, M. Chen, N. Awari, A. N. Ponomaryov, I. Ilyakov, M. Bluschke, F. Boschini, M. Z. Zhonno, S. Zhdanovich, M. Na, G. Christiani, G. Logvenov, D. J. Jones, A. Damascelli, M. Minola, B. Keimer, D. Manske, N. Wang, J.-C. Deinert, and S. Kaiser, Fano interference of the higgs mode in cuprate high-tc superconductors (2021), arXiv:2109.09971 [cond-mat.supr-con].DOI
- [36]Z.-B. Wu, D. Putzky, A. K. Kundu, H. Li, S. Yang, Z. Du, S. H. Joo, J. Lee, Y. Zhu, G. Logvenov, B. Keimer, K. Fujita, T. Valla, I. Božović, and I. K. Drozdov, Homogeneous superconducting gap in DyBa₂Cu₃O₇−δ synthesized by oxide molecular beam epitaxy, Phys. Rev. Materials 4, 124801 (2020).DOI
- [41]X. Dong, E. Gull, and A. J. Millis, Quantifying the role of antiferromagnetic fluctuations in the superconductivity of the doped hubbard model (2022), arXiv:2202.10577 [cond-mat.supr-con].DOI
- [42]M. Kitanani, T. Schäfer, H. Aoki, and K. Held, Why the critical temperature of high-T_c cuprate superconductors is so low: The importance of the dynamical vertex structure, Phys. Rev. B 99, 041115 (2019).
- [43]F. Krien, P. Worm, P. Chalupa, A. Toschi, and K. Held, Spin scattering turns complex at strong coupling: the key to pseudogap and fermi arcs in the hubbard model (2021), arXiv:2107.06529 [cond-mat.str-el].

