Tangent-space Krylov computation of real-frequency spectral functions: Influence of density-assisted hopping on 2D Mott physics
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.
Kai: We're looking at the paper: "Tangent space Krylov computation of real-frequency spectral functions".
Mira: Here's the gist of the paper. Tangent space Krylov computation of real-frequency spectral functions: Influence of density-assisted hopping on 2D Mott physics is a novel method that efficiently computes zero-temperature real-frequency spectral functions on top of ground state matrix product states by projecting resolvents into the tangent space of the ground state MPS, allowing for a direct computation of spectral weights and positions The paper introduces a novel tangent-space Krylov (TaSK) method to compute zero-temperature real-frequency spectral functions on top of ground state matrix product states obtained from DMRG It relies on projecting resolvents to the tangent space of the GS-MPS, where they can be efficiently represented using Krylov space techniques This allows for a direct computation of spectral weights and their corresponding positions on the real-frequency axis The method restricts computations to the tangent space of the MPS manifold at the ground state, enabling a Lanczos scheme devoid of orthogonality issues because the tangent space has a vector-space structure The core computation involves projecting all involved states to V1⊥, which is the space of MPS tangent vectors orthogonal to Ψg⟩ Lanczos Implementation and Error Control The Lanczos scheme iteratively generates an orthonormal basis of the Nkr + 1-dimensional Krylov space K (Ψ0) = span Ψ0⟩, H1⊥Ψ0⟩,..., H1⊥) The resulting tridiagonal representation of H1⊥ in K (Ψ0) is diagonalized to obtain the approximate eigenpairs ωα, Ψ1⊥α ⟩ The main computational cost of TaSK scales as O(L D3dw), which is comparable to one single-site DMRG or time-dependent variational principle (TDVP) sweep The main source of error in TaSK is the projection of the Hamiltonian to V1⊥, H → H1⊥, and this can be quantified through the 2-site variance ∆2⊥α = P2⊥HΨ1⊥α ⟩2 Benchmarking on Spin Chains TaSK was benchmarked on the Haldane–Shastry and Heisenberg chains, finding excellent agreement with analytical results for the dynamical structure factor For the L = 40 HaldaneShastry model, TaSK results agreed remarkably well with analytical expressions for the dynamical structure factor The convergence of low-lying excitations error estimates δα decreased monotonically with increasing D∗, suggesting that the projection error is reduced as dim V1⊥ ∝ D2 gro For the spin-1/2 Heisenberg chain, TaSK results closely followed the Bethe-ansatz curve corresponding to the full 2 + 4-spinon contribution in the thermodynamic limit Influence of Density-Assisted Hopping (DAH) The study applied TaSK to the Hubbard model on a cylinder augmented by a density-assisted hopping (DAH) term This DAH term leads to particle-hole asymmetric single-particle mobilities and lifetimes in the resulting Mott insulator, and identifies responsible scattering processes Specifically, the DAH influences the dispersion of Green’s function zeros beyond its range, which has a frustrating effect on the Mott insulator studied here The next-nearest neighbor DAH generates a sizeable next-nearest neighbor term τ′—a remarkable result arising from the fact that the DAH tn vertex, unlike the U vertex, delocalizes electrons beyond single-particle hopping and suggesting it may have a similar frustration effect as t′ for the t-t′ Hubbard model Spectral Broadening and Physical Insight To obtain continuous spectral functions, a Gaussian broadening scheme is used to broaden discrete spectra, preserving low-order moments like µb0 and µb1 The analysis of the t-tn-U Hubbard model showed that mixed tn-U scattering leads to enhanced hole mobilities and a possibly frustrating nextnearest neighbor contribution to the dispersion of the selfenergy pole This emergent next-nearest neighbor contribution suggests that tn may stabilize superconductivity in the t-tn Hubbard model The TaSK method is deemed an efficient and accurate way for computing real-frequency spectral functions, especially useful for cases where ground-state calculations are challenging Comparison with TDVP A comparison between TaSK results and those obtained via time-dependent variational principle (TDVP) in the 64-site spin-1/2 Heisenberg chain showed that the data matches perfectly, and the deviation is within the expected Trotter error of TDVP This handshake between TaSK and TDVP shows that TaSK not only captures low-energy features accurately but also describes well high-energy features governing short-time dynamics Future Outlook The method is particularly useful for strongly correlated electron systems on cylinders, geared towards cuprates, heavy fermions, and Moir´e systems The findings regarding mixed tn-U scattering leading to enhanced hole mobilities and a possibly frustrating nextnearest neighbor contribution may be relevant for high-Tc superconductivity Further work is needed to include doping and calculate other dynamical properties such as spin or pair spectra The method's main source of error comes from the projection to the tangent space, which can be estimated via the introduced two-site variance Supplemental Details on DAH Physics The analysis of second-order self-energy for the t-tn-U model revealed that particle-hole asymmetry arises from a "cooperative" tn-U scattering event, where a particle-hole pair is emitted via Hn, followed by reabsorption via HU This leads to an emergent nextnearest neighbor term τ′ = 0.17 in the t-tn-U model, which suggests that tn may have a similar frustration effect as t′ for the t-t' Hubbard model The pole dispersion of a Mott insulator in the presence of DAH contains terms that exceed the nearest-neighbor range of the DAH term itself, revealing that the pole dispersion can be thought of as the effective dispersion of a dressed fermion Self-Energy Pole Dispersion The self-energy pole dispersion ξk is determined by high-frequency expansion, where Gk(z) = 1/z + Ek/z squared + E2k + ∆k/z cubed + E3k + 2Ek∆k + ∆kξk/z 4 For the t-U model, this expansion yields τ = −2.03 and τ′ = 0, confirming that the pole dispersion should contain only nearest-neighbor terms In contrast, for the t-tn-U model, we find τ = −1.86 and τ′ = 0.17, indicating that next-nearest neighbor DAH generates a sizeable nextnearest neighbor term Error Estimation and Convergence The convergence of the discrete frequencies ωα with the number of Krylov steps Nkr for both benchmark spin-chain models was shown to be rapid, saturating around Nkr = 30 for the HSM The convergence for the Heisenberg chain was slower but systematic, and some low-frequency spectral peaks were identified as genuine excitations of H rather than artifacts of the tangent space projection The error measure δα = (∆2⊥α)1/2 is used to estimate the error in ωα Broadening Strategy The broadening scheme uses a Gaussian kernel, gσ(ω, ω′) = 1/√2π exp − (ω − ω′)2/2σ2 This scheme conserves the zeroth and first moments of the discrete spectral function Ad(ω), which is desirable for controlling short-time dynamics The scheme preserves the first 6 moments of Ak(ω) when broadening parameters are set to σ = 0.5 and NCFE = 2 Increasing NCFE sharpens main spectral peaks, while increasing σ too small leads to a uniformly underbroadened spectrum with wiggles appearing across the entire frequency range Comparison to TDVP The time-dependent correlation function Sπ(t) obtained via TaSK matches perfectly with the TDVP result, and the deviation is within the expected Trotter error of TDVP This demonstrates that TaSK describes well high-energy features governing short-time dynamics Second-Order Self-Energy Origin The second-order self-energy is decomposed into Σ(2)k(ω) = ΣU2k(ω) + ΣUtnk(ω) + Σt2nk(ω), where the term ΣUtnk(ω) is not particle-hole symmetric This asymmetry arises via a "cooperative" tn-U scattering event, which is a leading-order correction to the self-energy in the presence of DAH The resulting correction term δΣk(ω) = −4tnγk/U(Σteff-U k(ω) − ΣHF k) is a simplified approximation that gives insight into the origin of particle-hole asymmetries Next-Nearest Neighbor Term Emergence The non-locality of qe†iσ in the presence of DAH produces contributions to
ξ∆: ij (and therefore also ξij) that extend beyond the range of either tij or t n, leading to next-nearest neighbor terms in ξk For the t-tn-U model, this results in τ = −1.86 and τ′ = 0.17, demonstrating that the next-nearest neighbor term is sizeable This emergence of nonzero τ′ suggests that tn may have a similar frustration effect as t' for the t-t' Hubbard model TaSK is an efficient and accurate method for computing real-frequency spectral functions, with its main source of error stemming from the projection to the tangent space It is especially useful in cases where ground-state calculations are challenging, such as strongly correlated electron systems on cylinders The findings regarding mixed tn-U scattering leading to enhanced hole mobilities and a possibly frustrating nextnearest neighbor contribution may be relevant for high-Tc superconductivity Future work should include doping and the calculation of other dynamical properties such as spin or pair spectra The method's success in matching TDVP results confirms its utility for describing both low-energy features and short-time dynamics Acknowledgments The authors thank Seung-Sup Lee, Sebastian P¨ackel, Felipe Picoli, Dai-Wei Qu, and Steven White for helpful discussions This work was supported in part by the Deutsche Forschungsgemeinschaft under grants INST 86/1885-1 FUGG, LE 3883/2-2 and Germany’s Excellence Strategy EXC-2111 (Project No.
Kai: 390814868) It is part of the Munich Quantum Valley, supported by the Bavarian state government with funds from the Hightech Agenda Bayern Plus The National Science Foundation supported JvD in part under PHY-1748958 AG acknowledges support from the Abrahams Postdoctoral Fellowship of the Center for Materials Theory at Rutgers University The work is supported by various grants and fellowships, including those from the Deutsche Forschungsgemeinschaft and the National Science Foundation References
1: S.
Mira: R.
Kai: White, Density matrix formulation for quantum renormalization groups, Phys.
Paper summary: Kai: So we're looking at this paper titled "Tangent space Krylov computation of real-frequency spectral functions: Influence of density-assisted hopping on 2D Mott physics" <ref:2510.07279#pg1,Tangent space Krylov computation of real-frequency spectral functions: Influence of density>. Essentially, they've developed a new way to compute these real-frequency spectral functions zero temperature on top of ground state matrix product states from density matrix renormalization group.
Mira: Right, so the core idea is using this tangent space Krylov method, or TaSK. They project the resolvents into the tangent space of that ground state matrix product state, which lets them compute the spectral weights and their positions directly on the real frequency axis.
Lev: The big claim here is that they can do this efficiently by using Krylov space techniques to represent those projected vectors, which avoids some orthogonality issues you usually see with standard Lanczos schemes.
Kai: And they show off their accuracy by benchmarking it on 1D Haldane-Shastry and Heisenberg chains, where the TaSK results agree really well with analytical results for the dynamical structure factor <ref:2510.07279#pg1,1D Haldane-Shastry and Heisenberg>.
Mira: They also look at the influence of density-assisted hopping on a Hubbard model on a cylinder, finding that this term creates particle hole asymmetric single particle mobilities in the resulting Mott insulator.
Lev: What's interesting is how this density assisted hopping leads to a next nearest neighbor term, tau prime, in the t sub t Hubbard model, and it suggests that this might cause frustration similar to the standard t prime term.
Kai: This method is claimed to be efficient because the main computational cost scales at big O of L D cubed dw, which is comparable to one single site density matrix renormalization group sweep or a time dependent variational principle sweep.
Mira: They use a Gaussian broadening scheme to get continuous spectral functions, and they mention that the mixed t sub t n u scattering leads to enhanced hole mobilities and this possibly frustrating next nearest neighbor contribution to the dispersion of the self energy pole.
Lev: They also compare their results with time dependent variational principle methods on a sixty-four site spin-one/two Heisenberg chain, where the data matches perfectly within the expected error of that TDVP method <ref:2510.07279#pg3,spin-1/2 Heisenberg chain>.
Kai: So, it seems this TaSK method is being positioned as an efficient and accurate way to compute these spectral functions, especially for cases where calculating the ground state itself is pretty hard.
Paper summary: Mira: The paper discusses how this approach can be particularly useful for strongly correlated electron systems on cylinders, like those found in cuprates or heavy fermions.
Lev: They flag that the main source of error in this TaSK method comes from projecting the Hamiltonian onto that tangent space, and they quantify this error using a two site variance, which is called delta two perpendicular alpha.
Kai: The convergence of these discrete frequencies with respect to the number of Krylov steps is shown to be rapid for some models, saturating around Nkr equal to thirty for the Haldane-Shastry model <ref:2510.07279#pg3>.
Mira: They also show how increasing that D star value reduces the error estimate delta alpha because the dimension of that tangent space V one perpendicular scales as big O of L D two d.
Lev: That scaling suggests that as you increase those variational parameters, you get better control over the projection error in this method.
Kai: So, when we look at how they've applied this to the t sub t n u model, they found that the second order self energy has a term not particle hole symmetric arising from a cooperative interaction between t sub t and u terms.
Mira: That asymmetry is caused by a specific scattering event where a particle hole pair is emitted via h n followed by reabsorption via h u, which leads to an emergent next nearest neighbor term tau prime equal to zero point one seven in the t sub t n u model.
Lev: That finding about tau prime being size zero for the standard t u model and non-zero in the t sub tn u model is significant because it suggests that tn might have a similar frustration effect as standard t prime for the full two dimensional Hubbard model.
Kai: The paper also discusses how the pole dispersion of a Mott insulator with density assisted hopping contains terms that go beyond the nearest neighbor range of those hopping terms, meaning you can think of the pole dispersion as an effective dispersion of a dressed fermion.
Mira: They use an expansion for the self energy pole dispersion where it shows that for the t u model, tau is minus two point zero three and tau prime is zero, which confirms that only nearest neighbor terms should be there.
Paper summary: Lev: But in the t sub tn u model, they find tau is minus one point eight six and tau prime equals zero point one seven, indicating a next nearest neighbor hopping term does appear in this case.
Kai: The broadening strategy they used is a Gaussian kernel with sigma equal to zero point five, and it preserves the zeroth and first moments of the discrete spectral function, which is good for controlling short time dynamics.
Mira: They say that setting sigma too small leads to a uniformly underbroadened spectrum with wiggles appearing across the entire frequency range, which gives you some insight into what happens when you don't broaden enough.
Lev: The paper also notes that while convergence for the Heisenberg chain was slower, they found that some low frequency spectral peaks were genuine excitations of the Hamiltonian rather than just artifacts from the tangent space projection.
Kai: So, looking at the whole picture of this "Tangent space Krylov computation of real-frequency spectral functions: Influence of density-assisted hopping on 2D Mott physics", it seems like they've found a method that is both efficient and accurate for getting those spectral functions <ref:2510.07279#pg1,Tangent space Krylov computation of real-frequency spectral functions: Influence of density>.
Mira: It’s an efficient way to handle the complexity when you have challenging ground state calculations, which is definitely where this kind of tool becomes useful for strongly correlated systems on cylinders.
Lev: The main thing to watch is that the error estimate delta alpha relies entirely on that projection to the tangent space, so if you want higher accuracy, you're really looking at improving how well you project the Hamiltonian onto that space.
Kai: The results connecting this method to time dependent variational principle methods on spin chains suggests it captures both low energy features accurately and also describes well high energy features governing short time dynamics.
Mira: The findings regarding mixed t sub t n u scattering leading to enhanced hole mobilities and possibly a frustrating next nearest neighbor contribution are relevant if you're looking at high temperature superconductivity.
Lev: For future work, they mention that they still need to include doping and calculate other dynamical properties, like spin or pair spectra, because that would give a more complete picture of the system.
Kai: So, yeah, this paper gives us a solid tool for calculating real frequency spectra in these complex areas.
Conclusion: Kai: So we're wrapping up this look at the paper "Tangent space Krylov computation of real-frequency spectral functions: Influence of density-assisted hopping on 2D Mott physics <ref:2510.07279#pg1,Tangent space Krylov computation of real-frequency spectral functions: Influence of density>."
Mira: Yeah, basically, they’ve built a new tool to figure out the actual energy spectrum of these complex materials without needing a full ground state calculation.
Lev: They use this tangent space Krylov method to project those resolvents into a smaller, manageable space around the ground state matrix product state.
Kai: It's about making spectral function calculations efficient by sticking close to the ground state structure instead of trying to solve the whole system at once.
Mira: The authors are showing how this works on models like the Hubbard model with density-assisted hopping, and they find interesting things about particle movement in those Mott insulators.
Lev: They found that this density-assisted hopping actually generates a next nearest neighbor term, a tau prime, which can be quite significant for the physics of the system.
Kai: It seems they’ve established a solid way to get real frequency data that’s comparable to more intensive methods like time dependent variational principle sweeps.
Mira: The big picture here is that this technique gives us a new lens to see how these hopping terms influence superconductivity and other correlated phenomena in 2D systems <ref:2510.07279#pg1>.
Lev: But they are careful, they admit the main limitation is that the accuracy really hinges on how well you project the Hamiltonian onto that specific tangent space, which we call the two site variance.
Kai: So it’s an efficient way to get spectral information, but you still have to manage that projection error if you want super high precision.
Mira: Exactly, and this tool is clearly useful for studying things like cuprates or heavy fermions where those ground state calculations are just too tough otherwise.
Lev: That's why the next step in their work will be looking at how doping affects these results and calculating other properties like spin or pair spectra to see if these insights hold up there.
Arnold Sommerfeld Center for Theoretical Physics · Center for NanoScience · Munich Center for Quantum Science and Technology · Department of Physics and Astronomy, Rutgers University
cond-mat.str-el
Submitted: 2025-10-08
Updated: 2026-11-09
Comments: 22 pages
DOI: 10.1103/6f9c-gb69
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 82/100
The gist: Tangent space Krylov computation of real-frequency spectral functions: Influence of density-assisted hopping on 2D Mott physics is a novel method that efficiently computes zero-temperature…
Key concepts
- Tangent Space Krylov (TaSK) method
- A novel technique that projects quantum states onto the tangent space of the ground state matrix product state. This allows for efficient computation of spectral functions by using Krylov space techniques, avoiding orthogonality issues inherent in standard Lanczos schemes on curved manifolds.
- Ground State Matrix Product States (GS-MPS)
- A representation of a quantum many-body system's ground state that is highly effective for studying strongly correlated systems. The TaSK method operates by projecting calculations onto the tangent space defined by this specific ground state, making it suitable for complex geometries like cylinders.
- Density-Assisted Hopping (DAH)
- An added term in the Hubbard model that influences particle-hole asymmetry and single-particle mobilities in a Mott insulator. DAH introduces next-nearest neighbor interactions ($ au'$) that can frustrate standard hopping terms, potentially stabilizing superconductivity.
- Real-Frequency Spectral Functions
- These functions describe the energy distribution of excitations in a quantum system. The TaSK method computes these functions directly on the real frequency axis, providing insight into how particles move and interact within the material's electronic structure.
Terminology
Summary
Tangent space Krylov computation of real-frequency spectral functions: Influence of density-assisted hopping on 2D Mott physics is a novel method that efficiently computes zero-temperature real-frequency spectral functions on top of ground state matrix product states by projecting resolvents into the tangent space of the ground state MPS, allowing for a direct computation of spectral weights and positions
Methodology and Core Innovation
The paper introduces a novel tangent-space Krylov (TaSK) method to compute zero-temperature real-frequency spectral functions on top of ground state matrix product states obtained from DMRG It relies on projecting resolvents to the tangent space of the GS-MPS, where they can be efficiently represented using Krylov space techniques This allows for a direct computation of spectral weights and their corresponding positions on the real-frequency axis The method restricts computations to the tangent space of the MPS manifold at the ground state, enabling a Lanczos scheme devoid of orthogonality issues because the tangent space has a vector-space structure The core computation involves projecting all involved states to V1⊥, which is the space of MPS tangent vectors orthogonal to Ψg⟩
Lanczos Implementation and Error Control
The Lanczos scheme iteratively generates an orthonormal basis of the Nkr + 1-dimensional Krylov space K (Ψ0) = span Ψ0⟩, H1⊥Ψ0⟩,..., H1⊥) The resulting tridiagonal representation of H1⊥ in K (Ψ0) is diagonalized to obtain the approximate eigenpairs ωα, Ψ1⊥α ⟩ The main computational cost of TaSK scales as O(L D3dw), which is comparable to one single-site DMRG or time-dependent variational principle (TDVP) sweep The main source of error in TaSK is the projection of the Hamiltonian to V1⊥, H → H1⊥, and this can be quantified through the 2-site variance ∆2⊥α = P2⊥HΨ1⊥α ⟩2
Benchmarking on Spin Chains
TaSK was benchmarked on the Haldane–Shastry and Heisenberg chains, finding excellent agreement with analytical results for the dynamical structure factor For the L = 40 HaldaneShastry model, TaSK results agreed remarkably well with analytical expressions for the dynamical structure factor The convergence of low-lying excitations error estimates δα decreased monotonically with increasing D∗, suggesting that the projection error is reduced as dim V1⊥ ∝ D2 gro For the spin-1/2 Heisenberg chain, TaSK results closely followed the Bethe-ansatz curve corresponding to the full 2 + 4-spinon contribution in the thermodynamic limit
Influence of Density-Assisted Hopping (DAH)
The study applied TaSK to the Hubbard model on a cylinder augmented by a density-assisted hopping (DAH) term This DAH term leads to particle-hole asymmetric single-particle mobilities and lifetimes in the resulting Mott insulator, and identifies responsible scattering processes Specifically, the DAH influences the dispersion of Green’s function zeros beyond its range, which has a frustrating effect on the Mott insulator studied here The next-nearest neighbor DAH generates a sizeable next-nearest neighbor term τ′—a remarkable result arising from the fact that the DAH tn vertex, unlike the U vertex, delocalizes electrons beyond single-particle hopping and suggesting it may have a similar frustration effect as t′ for the t-t′ Hubbard model
Spectral Broadening and Physical Insight
To obtain continuous spectral functions, a Gaussian broadening scheme is used to broaden discrete spectra, preserving low-order moments like µb0 and µb1 The analysis of the t-tn-U Hubbard model showed that mixed tn-U scattering leads to enhanced hole mobilities and a possibly frustrating nextnearest neighbor contribution to the dispersion of the selfenergy pole This emergent next-nearest neighbor contribution suggests that tn may stabilize superconductivity in the t-tn Hubbard model The TaSK method is deemed an efficient and accurate way for computing real-frequency spectral functions, especially useful for cases where ground-state calculations are challenging
Comparison with TDVP
A comparison between TaSK results and those obtained via time-dependent variational principle (TDVP) in the 64-site spin-1/2 Heisenberg chain showed that the data matches perfectly, and the deviation is within the expected Trotter error of TDVP This handshake between TaSK and TDVP shows that TaSK not only captures low-energy features accurately but also describes well high-energy features governing short-time dynamics
Future Outlook
The method is particularly useful for strongly correlated electron systems on cylinders, geared towards cuprates, heavy fermions, and Moir´e systems The findings regarding mixed tn-U scattering leading to enhanced hole mobilities and a possibly frustrating nextnearest neighbor contribution may be relevant for high-Tc superconductivity Further work is needed to include doping and calculate other dynamical properties such as spin or pair spectra The method's main source of error comes from the projection to the tangent space, which can be estimated via the introduced two-site variance
Supplemental Details on DAH Physics
The analysis of second-order self-energy for the t-tn-U model revealed that particle-hole asymmetry arises from a cooperative
tn-U scattering event, where a particle-hole pair is emitted via Hn, followed by reabsorption via HU This leads to an emergent nextnearest neighbor term τ′ = 0.17 in the t-tn-U model, which suggests that tn may have a similar frustration effect as t′ for the t-t' Hubbard model The pole dispersion of a Mott insulator in the presence of DAH contains terms that exceed the nearest-neighbor range of the DAH term itself, revealing that the pole dispersion can be thought of as the effective dispersion of a dressed fermion
Self-Energy Pole Dispersion
The self-energy pole dispersion ξk is determined by high-frequency expansion, where Gk(z) = 1/z + Ek/z squared + E2k + ∆k/z cubed + E3k + 2Ek∆k + ∆kξk/z 4 For the t-U model, this expansion yields τ = −2.03 and τ′ = 0, confirming that the pole dispersion should contain only nearest-neighbor terms In contrast, for the t-tn-U model, we find τ = −1. p. 386 and τ′ = 0.17, indicating that next-nearest neighbor DAH generates a sizeable nextnearest neighbor term
Error Estimation and Convergence
The convergence of the discrete frequencies ωα with the number of Krylov steps Nkr for both benchmark spin-chain models was shown to be rapid, saturating around Nkr = 30 for the HSM The convergence for the Heisenberg chain was slower but systematic, and some low-frequency spectral peaks were identified as genuine excitations of H rather than artifacts of the tangent space projection The error measure δα = (∆2⊥α)1/2 is used to estimate the error in ωα
Broadening Strategy
The broadening scheme uses a Gaussian kernel, gσ(ω, ω′) = 1/√2π exp − (ω − ω′)2/2σ2 This scheme conserves the zeroth and first moments of the discrete spectral function Ad(ω), which is desirable for controlling short-time dynamics The scheme preserves the first 6 moments of Ak(ω) when broadening parameters are set to σ = 0.5 and NCFE = 2 Increasing NCFE sharpens main spectral peaks, while increasing σ too small leads to a uniformly underbroadened spectrum with wiggles appearing across the entire frequency range
Comparison to TDVP
The time-dependent correlation function Sπ(t) obtained via TaSK matches perfectly with the TDVP result, and the deviation is within the expected Trotter error of TDVP This demonstrates that TaSK describes well high-energy features governing short-time dynamics
Second-Order Self-Energy Origin
The second-order self-energy is decomposed into Σ(2)k(ω) = ΣU2k(ω) + ΣUtnk(ω) + Σt2nk(ω), where the term ΣUtnk(ω) is not particle-hole symmetric This asymmetry arises via a cooperative
tn-U scattering event, which is a leading-order correction to the self-energy in the presence of DAH The resulting correction term δΣk(ω) = −4tnγk/U(Σteff-U k(ω) − ΣHF k) is a simplified approximation that gives insight into the origin of particle-hole asymmetries
Next-Nearest Neighbor Term Emergence
The non-locality of qe†iσ in the presence of DAH produces contributions to [ξ∆]ij (and therefore also ξij) that extend beyond the range of either tij or t n, leading to next-nearest neighbor terms in ξk For the t-tn-U model, this results in τ = −1.86 and τ′ = 0.17, demonstrating that the next-nearest neighbor term is sizeable This emergence of nonzero τ′ suggests that tn may have a similar frustration effect as t' for the t-t' Hubbard model
Conclusion and Outlook
TaSK is an efficient and accurate method for computing real-frequency spectral functions, with its main source of error stemming from the projection to the tangent space It is especially useful in cases where ground-state calculations are challenging, such as strongly correlated electron systems on cylinders The findings regarding mixed tn-U scattering leading to enhanced hole mobilities and a possibly frustrating nextnearest neighbor contribution may be relevant for high-Tc superconductivity Future work should include doping and the calculation of other dynamical properties such as spin or pair spectra The method's success in matching TDVP results confirms its utility for describing both low-energy features and short-time dynamics
Acknowledgments
The authors thank Seung-Sup Lee, Sebastian P¨ackel, Felipe Picoli, Dai-Wei Qu, and Steven White for helpful discussions This work was supported in part by the Deutsche Forschungsgemeinschaft under grants INST 86/1885-1 FUGG, LE 3883/2-2 and Germany’s Excellence Strategy EXC-2111 (Project No. 390814868) It is part of the Munich Quantum Valley, supported by the Bavarian state government with funds from the Hightech Agenda Bayern Plus The National Science Foundation supported JvD in part under PHY-1748958 AG acknowledges support from the Abrahams Postdoctoral Fellowship of the Center for Materials Theory at Rutgers University The work is supported by various grants and fellowships, including those from the Deutsche Forschungsgemeinschaft and the National Science Foundation
References
[1] S. R. White, Density matrix formulation for quantum renormalization groups, Phys. p. 4 Rev. Lett. 69, 2863 (1992) [2] U. p. 4 Schollw¨ock, The density-matrix renormalization group in the age of matrix product states, Annals of Physics 326, 96 (2011) [3] G. p. 4 Vidal, Efficient simulation of one-dimensional quantum many-body systems, Phys. p. 4 Rev. Lett. 93, 040502 (2011) [4] S. p. 4 R. White and A. p. 4 E. Feiguin, Real-time evolution using the density matrix renormalization group, Phys. p. 4 Rev. Lett.> 93, 076401 (2510) [5] A J Daley, C Kollath, U Schollw¨ock, and G Vidal, Time-dependent density-matrix renormalization-group using adaptive effective Hilbert spaces Journal of Statistical Mechanics: Theory and Experiment 2004 P04005 [6] S Paeckel, T K¨ohler, A Swoboda, S R Manmana, U Schollw¨ock, and C Hubig Time-evolution methods for matrix-product states Annals of Physics 411 167998 (2019) [7] M Grundner, P Westhoff, F B Kugler, O Parcollet, and U Schollw¨ock Complex time evolution in tensor networks and time-dependent Green’s functions Phys. p. 4 Rev. B 109 155124 (2024) [8] T D K¨uhner and S R White Dynamical correlation functions using the density matrix renormalization group Phys. p. 4 Rev. B 60 335 (1999) [9] E Jeckelmann, Dynamical density-matrix renormalization-group method Phys. p. 4 Rev. B 66 045114 (2002) [10] C Raas and G S Uhrig Spectral densities from dynamic density-matrix renormalization Eur Phys J B 45 293 (2005) [1] A Nocera and G Alvarez Spectral functions with the density matrix renormalization group Krylov-space approach for correction vectors Phys. p. 4 Rev. E 94 053308 (2016) [16] S Paeckel, Spectral decomposition and high-accuracy Greens functions Overcoming the Nyquist-Shannon limit via complex-time Krylov expansion arXiv:2411. p. 409680 [cond-mat. p. 4str-el] [17] K A Hallberg Density-matrix algorithm for the calculation of dynamical properties of low-dimensional systems Phys. p. 4 Rev. B 52 R9827 (1995) [18] E Koch, The Lanczos method in The LDA+DMFT Approach to Strongly Correlated Materials Modeling and Simulation Vol. p. 4 1 edited by E Pavarini E Koch D Vollhardt and A Lichtenstein Forschungszentrum J¨ulich J¨ulich 2011 pp 8. p. 41–8.30 [19] P E Dargel, A Honecker, R Peters, R M Noack, and T Pruschke Adaptive Lanczos-vector method for dynamic properties within the density matrix renormalization group Phys. p. 4 Rev. B 83 161104 (2011) [20] P E Dargel, A W¨ollert, A Honecker, I P McCulloch, U Schollw¨ock, and T Pruschke Lanczos algorithm with matrix product states for dynamical correlation functions Phys. p. 4 Rev. B 85 205119 (2012) [21] Y Wang, Z Yang, and C B Mendl Improved Lanczos algorithm using matrix product states arXiv:2504. p. 421786 [cond-mat. p. 4str-el] [30] L Vanderstraeten J Haegeman and F Verstraete Tangent-space methods for uniform matrix product states SciPost Phys Lect Notes 7 2019 [31] W. p. 5-L Tu, H.-K. Wu, N Schuch, N Kawashima, and J. p. 5-Y Chen Generating function for tensor network diagrammatic summation Phys. p. 5 Rev. B 103 205155 (2021) [44] Eigenenergies of the Hamiltonian are extensive whereas the excitation energy differences probed by Ok are not our Ok is a superposition of short-ranged operators [45] See Supplemental Material at url for additional information on technical details of TaSK; computation of the error estimates; comparison between spectral data obtained with TaSK and TDVP; convergence and broadening analysis; analytical analysis of density-assisted hopping effects on Mott-insulating physics of the 2D Hubbard model. p. 5
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
-
Better spectral function computation for Mott insulators using Tangent space Krylov (TaSK) to accurately capture particle-hole asymmetry resulting from density-assisted hopping (DAH). This allows AI systems to model strongly correlated materials like those relevant to cuprates with physical insights beyond Hartree-Fock approximations, specifically identifying
emergent next-nearest neighbor contribution
to the Mott pole dispersion. -
Enhanced accuracy in characterizing quasiparticle dynamics by using TaSK convergence analysis based on the two-site variance, which is estimated as delta alpha = (Delta 2⊥ alpha) 1/2. This allows AI to quantify how projection errors from tangent space restriction affect spectral features and estimate necessary bond dimension or Krylov step requirements for reliable results.
-
Improved understanding of scattering processes by computing the second-order self-energy correction δSigma k (Eq. 12). This enables AI systems to identify the
cooperative
tn-U scattering event that generates particle-hole asymmetry, which is crucial for modeling complex electronic instabilities in doped Mott insulators. -
Determination of effective hopping parameters in DAH systems by computing the next-nearest neighbor term tau prime (tau') in the self-energy pole dispersion ξk. This allows AI to predict if DAH introduces frustrating effects similar to t prime in other models, potentially stabilizing superconductivity or heavy fermion behavior.
-
Development of a robust method for obtaining continuous spectral functions from discrete TaSK data by using Gaussian broadening of the residual function R d m(z) via a continued fraction expansion (CFE). This enables AI to generate smooth, physically relevant real-frequency spectra without introducing spurious sharp peaks in the continuum, balancing resolution and smoothness.
Sources
- Spectral decomposition and high-accuracy Greens functions: Overcoming the Nyquist-Shannon limit via complex-time Krylov expansion
- An Accurate Lanczos Method for the Matrix Product State Representation
- Inexact subspace projection methods for low-rank tensor eigenvalue problems
- Reply to comment on "Controlled bond expansion for Density Matrix Renormalization Group ground state search at single-site costs"
Related papers
- Coupled Magnetic and Lattice Excitations in Multiferroic Ba3HoRu2O9: Insights from Neutron Scattering, Raman Spectroscopy, and Machine-Learned Phonon Calculations
- Gossamer Superconductivity in Moir'e WSe 2 Bilayer
- Spin-density wave of ferrimagnetic building blocks masking the ferromagnetic quantum-critical point in NbFe2
- Raman Circular Dichroism and spinon Quantum Geometry of Chiral Quantum Spin Liquids
- Twin Algebras: Condensable Algebras beyond Anyons
- Spin supersolid in triangular lattice easy-axis spin-1/2 antiferromagnet: a fermionic parton perspective