Fingerprints of preformed pairs in two-electron angle-resolved photoemission spectroscopy

Janez Bonča, Andrea Damascelli, Mona Berciu · 2026

Read Fingerprints of preformed pairs in two-electron angle-resolved photoemission spectroscopy by Janez Bonča, Andrea Damascelli, Mona Berciu (2026) 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-2604-18560.txt.

Introduction

Bonča, Damascelli, and Berciu compute the two-electron removal weight for the Hubbard-Holstein model via variational diagonalization, from a paired ground state on a chain. They argue this weight proxies ARPES intensity. Against the larger separate-pair signal, the weaker same-pair signal sits at lower binding energy with a momentum pattern of other shape. They verify these fingerprints for pairs of every kind, proving they arise from conservation laws and hold for electron-boson models. Seeing them in the lab confirms the reality of pairs.

Catalog record

CREATORJanez Bonča, Andrea Damascelli, Mona Berciu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28288
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.18560

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Fingerprints of preformed pairs in two-electron angle-resolved photoemission spectroscopy Janez Bonča, Andrea Damascelli, Mona Berciu Abstract We use variational exact diagonalization (VED) to calculate the two-electron removal spectral weight for the Hubbard-Holstein model, starting from the ground-state with two electrons on a one-dimensional chain. We argue that this spectral weight provides a valuable proxy for the intensity of 2eARPES processes. Our results show that when contrasted to the presumably larger signal due to two electrons ejected from two different pairs, the presumably weaker signal due to two electrons ejected from the same pair (i) is segregated in energy, appearing at a lower binding energy, and (ii) has a very characteristic momentum dependence, with a different symmetry than that of the signal corresponding to two electrons emitted from two different pairs. We verify that these fingerprints appear for pairs with different symmetries, and prove that they arise as a direct consequence of momentum and energy conservation, therefore they are generic for any model with electron-boson coupling that can lead to formation of electron pairs. Experimental observation of these fingerprints will confirm the existence of pairs. Moreover, the momentum dependence map allows one to distinguish whether the pairs are coherent (superconducting) or not. Finally, we argue that these considerations generalize to finite but low electron concentrations, finite temperatures and higher dimensions. Introduction: Angle-resolved photoemission spectroscopy (ARPES) is a well-established and extremely successful experimental technique that provides direct access to the quasiparticle band structure and Fermi surfaces of a variety of interesting materials Damascelli et al. 2003 . Particularly useful for strongly-correlated electron systems is the fact that ARPES also enables access to the quasiparticle self-energy. This quantity can be compared against self-energies corresponding to various correlated model Hamiltonians, allowing theorists to gain insights into the most relevant physics of the system of interest. While access to these single-particle properties is extremely valuable, it is not sufficient to fully characterize the state of a correlated system. For example, direct signatures of pair formation, e.g. in the context of superconductivity, is seen with ARPES if the superconducting gap opened in the single-particle density of states is within the energy resolution of the ARPES system. However, generically ARPES cannot identify whether the gap is due to superconductivity or some other order; additional measurements are needed to settle that Damascelli 2004 . This becomes problematic when trying to identify new correlated states that are not well understood theoretically. Of primary interest are liquids of pre-formed pairs, hypothesized to appear in systems where boson exchange favours pairing of the electrons, however these pairs are not coherent (the system is not superconducting) because the electron density is too low or the temperature is too high. We recently argued that signatures of such a state could be seen in ARPES Kovač et al. 2025 , however it is recognized that one of the most direct diagnostics for pairing correlations is provided by two electron coincidence ARPES (2eARPES); other possible correlation-focussed spectroscopies are reviewed in Ref. Su et al. 2025 . Indeed, analysis of the probability of coincidence detection of a pair of ejected electrons upon absorption of one photon has allowed the observation of the exchange-correlation hole surrounding an electron in a correlated (but unpaired) metal Schumann et al. 2007 , while recent theoretical work argued that 2eARPES can be used to identify the pairing state in BCS superconductors whose Cooper pairs have a nonzero center-of-mass momentum Mahmood et al. 2022 , and to extract dynamic electron interactions in other ordered phases, such as a charge density wave Kemper et al. 2025 . Exact calculation of two-photons in, two-electrons out 2eARPES intensities has revealed clear signatures of pre-formed pair formation for an 8-site cluster attractive Hubbard model at half-filling Devereaux et al. 2023 . In this Letter, we use the formalism developed in Refs. Stahl and Eckstein 2019 ; Su and Zhang 2020 ; Devereaux et al. 2023 to calculate 2eARPES intensities. First, we argue that when compared to the presumably larger signal due to two electrons ejected from two different pairs, the signal due to two electrons ejected from the same pair is always (i) segregated in energy, and (ii) has a characteristic momentum-dependence. These fingerprints are a direct consequence of momentum and energy conservation, therefore they must appear in any model hosting pre-formed pairs. Their observation in 2eARPES will directly confirm the presence of pairs, allow for an estimate of their real-space radius, and also indicate whether the pairs are coherent or not. We then verify this by calculating numerically the 2eARPES intensity of a pre-formed pair (a bipolaron) in the one-dimensional (1D) Hubbard-Holstein model. Varying the electron-phonon coupling strength and the on-site Coulomb repulsion, we can contrast the results for ground-states (GS) consisting of two unbound polarons (strong repulsion, weak el-ph coupling) from those for GS with a bipolaron (weak repulsion, strong el-ph coupling). We confirm that the two fingerprints are observed both for s - and p -wave pairs. Model and Method: We study an N -site 1D chain with electrons that experience Hubbard repulsion and are also Holstein-coupled to dispersive optical phonons, as described by the Hamiltonian: \displaystyle\vskip 0.0ptH \displaystyle= \displaystyle-t_{\mathrm{el}}\sum_{j,\sigma}(c^{\dagger}_{j,\sigma}c_{j+1,\sigma}+\mathrm{H.c.})+{g}\sum_{j}\hat{n}_{j}(a_{j}^{\dagger}+a_{j}) (1) \displaystyle+ \displaystyle t_{\mathrm{ph}}\sum_{j}(a^{\dagger}_{j}a_{j+1}+\mathrm{H.c.})+\omega_{0}\sum_{j}a_{j}^{\dagger}a_{j} \displaystyle+ \displaystyle U\sum_{j}n_{j\uparrow}n_{j\downarrow}, where c^{\dagger}_{j\sigma} and a^{\dagger}_{j} are electron and phonon creation operators at site j , respectively, \hat{n}_{j}=\sum_{\sigma}c^{\dagger}_{j\sigma}c_{j\sigma} is the density operator and U is the Hubbard repulsion. The free-electron dispersion \epsilon_{k}=-2t_{el}\cos(k) is controlled by the nearest-neighbor hopping amplitude taken as the energy unit t_{\mathrm{el}}\equiv 1 , while the optical phonon energy \Omega_{q}=\omega_{0}+2t_{\mathrm{ph}}\cos(q) is such that \omega_{0}>|2t_{ph}| . We set the lattice constant a=1 , also \hbar=1 . The electron-phonon coupling strength is hereafter characterized by the dimensionless \lambda=g^{2}/[2t_{\mathrm{el}}\sqrt{\omega_{0}^{2}-4t_{\mathrm{ph}}^{2}}] Marchand and Berciu 2013 . We study this Hamiltonian for N_{e}=2 electrons on a finite chain with periodic boundary conditions employing variational exact diagonalization (VED) Bonča et al. 1999 ; Bonča et al. 2000 ; Ku et al. 2002 ; Bonča and Trugman 2021 ; Kovač and Bonča 2024 ; Kovač and Bonča 2025 . We typically use N_{h}\sim 18 iterations to generate the variational space by repeated application of H ; we verified that the results are converged and correspond to the thermodynamic limit N\rightarrow\infty . 2eARPES Intensity: We calculate the spectral weight A_{2}(\omega,k_{1},k_{2})={1\over\pi}\mbox{Im}{\cal G}_{k_{1}k_{2}}(\omega) of the two-particle propagator: {\cal G}_{k_{1}k_{2}}(\omega)=\sum_{n}\frac{|\langle\psi^{(n,N_{e}-2)}_{-k_{1}-k_{2}}|c_{k_{1}\sigma_{1}}c_{k_{2}\sigma_{2}}|\psi^{(GS,N_{e})}_{0}\rangle|^{2}}{\omega-i\eta+E_{-k_{1}-k_{2}}^{(n,N_{e}-2)}-E_{0}^{(GS,N_{e})}} (2) Here, {\cal H}=|\psi_{k}^{(n,N_{e})}\rangle=E_{k}^{(n,N_{e})}|\psi_{k}^{(n,N_{e})}\rangle are the eigenstates and eigenenergies with N_{e} particles and total momentum k ; n labels other quantum numbers. The ground-state GS corresponds to n=0 and has total momentum 0. For results in the singlet sector we set \sigma_{1}=-\sigma_{2} and verify that double occupancy has a non-vanishing probability. For the triplet sector, we set \sigma_{1}=\sigma_{2} . In this case double occupancy is forbidden by the Pauli principle, and the value of U becomes irrelevant. This two-particle propagator can be linked directly to the quantity defined in Eq. (23) of Ref. Devereaux et al. 2023 : D^{(0)}_{{\bf k_{1}k_{2}}}(\omega)\!=\!\int_{-\infty}^{\infty}\!\!\!dte^{-i\omega t}\!\!\int_{-\infty}^{\infty}\!\!\!\!d\tau\langle c^{\dagger}_{k_{1}\sigma_{1}}(t)c^{\dagger}_{k_{2}\sigma_{2}}(t+\tau)c_{k_{2}\sigma_{2}}(\tau)c_{k_{1}\sigma_{1}}(0)\rangle (3) Up to a factor containing photoexcitation matrix elements, which is ignored for simplicity, D^{(0)}_{{\bf k_{1}k_{2}}}(\omega) was shown by Devereaux et al. Devereaux et al. 2023 to represent the 2eARPES coincidence detection rate integrated over all possible differences \Delta\omega=\omega_{1}-\omega_{2} at a fixed total \omega=\omega_{1}+\omega_{2} , where \omega_{i}=\omega_{ph}-{\hbar^{2}{\bf k}_{i}^{2}\over 2m}-W is the energy imparted to the

Read the complete text