Absence of superconductivity and valence bond order in the Hubbard–Heisenberg model for organic charge-transfer solids

Abstract

A frustrated, effective 1 2 -filled band Hubbard-Heisenberg model has been proposed to describe the strongly dimerized charge-transfer solid families κ-(ET)2X and Z[Pd(dmit)2]2. In addition to unconventional superconductivity these materials also exhibit antiferromagnetism, candidate spinliquid phases, and in the case of Z=EtMe3P, a so-called valence-bond solid phase. We show that neither superconductivity nor the valence-bond solid phase occurs within the Hubbard-Heisenberg model, indicating that the effective 1 2 -filled band model is unsuitable for these materials.PACS numbers: 71.10. Fd, 71.10.Hf, 74.20.Mn, 74.70.Kn Low-dimensional organic charge transfer solids (CTS) are being intensively studied because of their many unusual competing and coexisting electronic phases. The most studied among them are probably the κ-(ET) 2 X and Z[Pd(dmit) 2 ] 2 families, which, depending on the anion X − or cation Z + exhibit unconventional superconductivity (SC), Néel antiferromagnetic (AFM) order, charge ordering (CO), candidate quantum spin liquid (QSL) behavior, and valence-bond solid (VBS) order [1]. The apparent similarity between these with the cuprate superconductors have been noted by many investigators. The semiconductor-SC transition in the CTS occurs under the application of pressure at constant carrier density, which suggests that the transition is driven by a small modification of a particular parameter of an appropriate Hamiltonian. The key questions then are, what is the minimal model, and which is the parameter whose changes give the competing phases.Experimental observations appear to give a simple answer to these questions. The 2:1 (1:2) stoichiometry of semiconductor-paramagnetic metal (PM) transition occurs with increasing frustration. D-wave SC mediated by fluctuations of the AFM ordering at the AFM-PM boundary has been also proposed based on mean-field and dynamic mean-field theories (DMFT) [3][4][5][6][7][8][9][10].Numerical calculations have, however, failed to find SC within the triangular lattice 1 2 -filled band Hubbard model [11][12][13]. Numerical studies have also failed to find a VBS phase in the model [14]. Although the 1 2 -filled Hubbard model on the anisotropic triangular lattice does not appear to support SC, closely related models continue to be suggested as the appropriate theoretical model for describing the SC transition in the CTS. It has been claimed that the simple Hubbard model does not include all the spin-spin interactions that play an important role in the CTS, and that additional spin exchange unrelated to the Hubbard U must be incorporated to correctly capture AFM fluctuation effects [8][9][10]. This has led to theoretical works on the so-called Hubbard-Heisenberg model given below. The goal of this Letter is to critically exam-

Theory of metal-intercalated phenacenes: Why molecular valence 3 is special

“…Capone et al have performed DMFT calculations within the 1 2 -filled band Jahn-Teller-Hubbard model, and have proposed that pressure-induced increased bandwidth leads to intramolecular pairing of electrons 41 . This conclusion is in apparent disagreement with theoretical works [27][28][29][30][31][32] showing the absence of SC within the 1 2 -filled band (the latter however do not include Jahn-Teller distortion). An alternate explanation of the antiferromagnetic-to-SC transition can be given within our approach, wherein the antiferromagnet at ambient pressure is indeed effective 1 2 -filled, but the pressure-induced increase in bandwidth leads to equalization of the electron population among the two lowest occupied MOs (as would occur in trianions of coronene within our theory, see Appendix).…”

“…These theories do not provide the correct starting point for any theory of SC in the trianions. Although mean field and DMFT calculations had found antiferromagnetic-to-SC transition within the frustrated 1 2 -filled Hubbard band model in the past [23][24][25][26] , more recent numerically precise calculations, using a variety of techniques, have universally found absence of SC within this model [27][28][29][30][31][32] . From an experimental perspective, the absence of SC in the Mott-Hubbard semiconductor K 1 pentacene 19 confirms the theories that predict absence of SC within the 1 2 -filled band Hubbard model.…”

Coulomb enhancement of superconducting pair-pair correlations in a34-filled model forκ−(BEDT-TTF)2X

“…As mentioned above, previous works have shown (i) suppression of pair-pair correlations within the effective ρ = 1 model [17][18][19][20] , and (ii) the possibility of a fluctuating CO within models focusing on the monomer molecules 33,34,[36][37][38] . We therefore construct pair creation operators that allow unequal charge densities on the monomer molecules that constitute a dimer in Fig.…”

“…Enhanced pair-pair correlations is a necessary though not sufficient condition for SC. We previously used this criterion to evaluate the possibility of SC within the ρ = 1 Hubbard model on triangular lattices 17,19,20 . Suppression of pairpair correlations by the Hubbard U was found for all the lattices we investigated.…”

“…A necessary condition for SC within any interacting electrons model however is that interactions must enhance superconducting pair-pair correlations relative to the noninteracting model. Precise numerical calculations within the 1 2 -filled band Hubbard model on triangular lattices have shown that pair-pair correlations decrease with Hubbard U for all anisotropy [17][18][19][20] , thus indicating the need for going beyond the effective 1/2-filled band model in our search for the the mechanism of SC in κ-(ET) 2 X.…”