Quantitative Description of Strongly Correlated Materials by Combining Downfolding Techniques and Tensor Networks
Daan Vrancken, Simon Ganne, Daan Verraes, Tom Braeckevelt, Lukas Devos, Laurens Vanderstraeten, Jutho Haegeman, Veronique Van Speybroeck · 2025
Read Quantitative Description of Strongly Correlated Materials by Combining Downfolding Techniques and Tensor Networks by Daan Vrancken, Simon Ganne, Daan Verraes, Tom Braeckevelt, Lukas Devos, Laurens Vanderstraeten, Jutho Haegeman, Veronique Van Speybroeck (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-2502-19588.txt.
Introduction
Daan Vrancken and co-authors combine two advanced methods to model materials where electrons interact strongly, in this 2025 paper. Standard methods struggle with such materials, whose behavior drives magnets and superconductors. The team first shrinks the problem using density functional theory plus a screening step, building a simpler Hubbard-style model.
Tensor networks then solve that model with fine control over accuracy. The team applied the pipeline to two carbon chain polymers and one copper oxide chain material. Predicted energy gaps matched both high-end calculations and lab measurements closely. The networks also yield spin patterns and excitation energies. Costs grow gently with system size, which suits long chain materials.
Catalog record
| CREATOR | Daan Vrancken, Simon Ganne, Daan Verraes, Tom Braeckevelt, Lukas Devos, Laurens Vanderstraeten, Jutho Haegeman, Veronique Van Speybroeck |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 96 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-19588
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Quantitative Description of Strongly Correlated Materials by Combining Downfolding Techniques and Tensor Networks Daan Vrancken, Simon Ganne, Daan Verraes, Tom Braeckevelt, Lukas Devos, Laurens Vanderstraeten, Jutho Haegeman, Veronique Van Speybroeck Abstract We present a high-accuracy procedure for electronic structure calculations of strongly correlated materials. To address limitations in current electronic structure methods, we employ density functional theory in combination with the constrained random phase approximation to construct an effective multi-band Hubbard model, which is subsequently solved using tensor networks. Our work focuses on one-dimensional and quasi-one-dimensional materials, for which we employ the machinery of matrix product states. We apply this framework to the conjugated polymers trans-polyacetylene and polythiophene, as well as the quasi-one-dimensional charge-transfer insulator Sr 2 CuO 3 . The predicted band gaps show quantitative agreement with state-of-the-art computational techniques and experimental measurements. Beyond band gaps, tensor networks provide access to a wide range of physically relevant properties, including spin magnetization and various excitation energies. Their flexibility supports the implementation of complex Hamiltonians with longer-range interactions, while the bond dimension enables systematic control over accuracy. Furthermore, the computational cost scales efficiently with system size, demonstrating the framework’s scalability. I Introduction Simulating molecules and materials from first principles is essential not only for predicting properties of interest but also for understanding the mechanisms behind emergent phenomena. Techniques like density functional theory (DFT) [ 1 , 2 ] have enabled the quantitative description of the electronic structure of many materials [ 3 , 4 ] . The local density approximation (LDA) and generalized gradient approximation (GGA) to the DFT exchange-correlation functional are commonly used to investigate weakly correlated systems. However, the accuracy of these estimations is often insufficient [ 5 , 6 ] , particularly for strongly correlated materials, which exhibit some of the most intriguing and least understood features, such as unconventional superconductivity [ 7 , 8 , 9 , 10 , 11 ] , colossal magnetoresistance [ 12 , 13 , 14 ] , and metal-insulator transitions [ 15 , 16 ] . Addressing the many-body problem in strongly correlated materials requires more advanced methods [ 17 ] . DFT +U , for example, attempts to compensate for the neglected correlation by introducing an interaction parameter U to the energy functional [ 18 , 19 ] . Determining the optimal value of U is challenging, often necessitating experimental data, which limits DFT +U from being a fully first-principles approach. The same issue arises with hybrid functionals [ 20 , 21 ] , where the appropriate mixing of exact exchange and LDA/GGA exchange-correlation can be difficult to predict for strongly correlated materials. Another method is the GW approximation [ 22 , 23 ] , which utilizes many-body perturbation theory and often provides significant improvements over DFT calculations, particularly for conventional semiconductors and insulators [ 24 ] . Yet, even this approach frequently falls short in adequately describing materials with strong electron correlation [ 25 , 26 , 27 ] . In wave function-based quantum chemistry, coupled-cluster (CC) techniques [ 28 , 29 ] and full configuration interaction [ 30 ] provide a systematic and precise framework for handling strong correlation, but their high computational cost makes them infeasible for solids. Alternatively, downfolding methods aim to construct an effective model that captures the essential physics by retaining only the low-energy degrees of freedom, typically the electronic states close to the Fermi level [ 24 ] . A suitable target space includes these states, which dominate the material’s behavior at the low-energy scale. The resulting effective model, often represented as a Hubbard model [ 31 ] , can then be solved with greater accuracy to extract key physical properties. Although the concept is straightforward, several technical challenges complicate the procedure. One key issue, common to most downfolding routines, is double counting [ 32 ] . The screening effects by electrons outside the target space should be incorporated into the two-body interaction parameters, while screening within the low-energy space must be excluded, as it is already accounted for in the effective model. Various constrained methods have been developed to deal with this issue [ 33 , 34 , 35 , 36 , 37 , 38 , 39 ] . Among them, the constrained random phase approximation (cRPA) [ 40 , 41 ] is perhaps the most widely used technique for calculating the screened interaction parameters in solids, in part due to its ability to retrieve the frequency-dependence [ 42 ] . A second source of double counting arises from the fact that one-body hopping terms are determined from an initial computation that already includes some electron interactions through the Hartree term and the exchange-correlation functional. To address this, correction terms can be subtracted from the parameters. Exclusion of the Hartree contribution within the target space is relatively straightforward, but correcting for the exchange-correlation is more difficult. While multiple correction schemes exist [ 19 , 43 , 44 , 45 ] , their impact on the accuracy of simulations remains questionable [ 32 , 46 , 47 ] . On the other hand, a constrained GW (c GW ) calculation effectively removes double counting up to the perturbation order within the GW scheme [ 38 ] . Once the model is constructed, the next challenge is solving it. Substantial progress has been made with dynamical mean field theory (DMFT) [ 46 , 48 , 49 ] , which maps the Hubbard model to an Anderson impurity model [ 50 ] . Impurity models can be solved with even greater precision using methods like quantum Monte Carlo (QMC) [ 51 , 52 ] or coupled-cluster techniques. Conventional DFT+DMFT [ 46 , 19 , 53 , 54 ] has been modified to extended DMFT (EDMFT), GW +DMFT, and even GW +EDMFT, offering solutions to some of the initial limitations [ 55 , 56 , 57 , 58 ] . However, challenges remain, such as the absence of non-local spin fluctuations and non-local vertex corrections beyond RPA [ 24 ] . Ongoing research is exploring further extensions to overcome these issues [ 59 , 60 , 61 , 62 , 63 , 64 , 65 , 66 , 67 , 68 , 69 ] . Besides DMFT, other methods like density matrix embedding theory (DMET) [ 70 , 71 ] and self-energy embedding theory (SEET) [ 72 , 73 ] , which are closely related to DMFT, have been developed, alongside the direct application of solvers such as QMC to the constructed Hubbard model. Directly solving the Hubbard model allows for a more intuitive interpretation as it closely resembles the physical system. A key limitation of DMFT and DMET is that improving their approximations primarily relies on increasing cluster size, leading to exponential growth of the computational cost. Recent work has focused on systematic improvability in quantum embedding that extends beyond merely enlarging fragment sizes [ 74 ] . While the applicability of QMC is limited by the sign problem [ 75 ] , coupled cluster methods are generally impractical for realistic multi-band Hubbard models beyond the singles and doubles (CCSD) approximation. In the past two decades, tensor networks (TNs) [ 76 ] have emerged as powerful tools for describing interacting quantum lattice and molecular systems [