Towards first-principles prediction of valence instabilities in mixed stack charge-transfer crystals

Strongly correlated electrons delocalized on one-dimensional (1D) soft stacks govern the complex physics of mixed stack charge-transfer crystals, a well-known family of materials composed of electron-donor (D) and acceptor (A) molecules alternating along the 1D chain. The complex physics of these systems is well captured by a modified Hubbard model that also accounts for the coupling of electrons to molecular and lattice vibrational modes and for three-dimensional electrostatic interactions. Here we study several experimental systems to estimate relevant model parameters via density-functional theory calculations on DA units and isolated molecules and ions. Electrostatic intermolecular interactions, an important quantity not just to define the degree of charge transfer of the ground state but also to predict the propensity of the system towards multistability and hence towards discontinuous phase transitions, are also addressed. Results compare favorably with experimental data.

Figure 1

Chemical structures of the systems of interest in this work. Chemical names associated with the abbreviations are as follows: DMTTF = 4,$4^{\prime }$-dimethyltetrathiafulvalene; TTF = tetrathiafulvalene; CA = chloranil; $2,5{\mathrm{Cl}}_2\mathrm{BQ}=2$,5-dichloro-$p$-benzoquinone; BA = bromanil; Per = perylene; 10BTBT = 2,7-didecyl[1]benzothieno[3,2-$b]\left[1\right]\mathrm{benzothiophene};\mathrm{TCNQ}=\mathrm{tetracyanoquinodimethane};{\mathrm{TCNQF}}_4$ = tetrafluoro-tetracyanoqui-nodimethane; DBTTF = dibenzotetrathiafulvalene; TMPD = $N,N,N^{\prime },N^{\prime }$-tetramethyl-$p$-phenylenediamine; NNTMB = $N,N,N^{\prime },N^{\prime }$-tetramethyl-benzidine; ClMePD = 2-chloro-5-methyl-$p$-phenylenediamine; DMeDCNQI = 2,5-dimethyl-dicyanoquinone-diimine.

Figure 2

Results of MHM calculations for ten CT crystals defined in Fig. 1. The different panels group the crystals according to the ${\epsilon }_T/t$ values: (a) ${\epsilon }_T<1.8t$, all values of $\rho$ are accessible; (b) ${\epsilon }_T\sim 1.8t$ marginally bistable systems; (c) ${\epsilon }_T>1.8t$ bistable systems; and (d) ${\epsilon }_T\gg 1.8t$ extreme bistability. In all panels the dashed black line is the universal $\rho (z/t)$ curve; colored lines refer to the specific systems numbered as in Fig. 1 and obtained with the parameters in Table 2. Diamonds show the ionicity computed for the $z/t$ value obtained from the atomistic parametrization.

Figure 3

Experimental $\rho$ values vs calculated results for ten CT crystals defined in Fig. 1. A generally good agreement is found between experiment and theory, with most points falling on the ideal 1:1 (dashed) line within the uncertainties of calculations and experiments (error bars). Experimental ionicities are taken from Ref. [61] for TTF-CA, [29] for DMTTF-CA, [51] for $\text{TTF-}2,5{\mathrm{Cl}}_2\mathrm{BQ}$, [62] for TTF-BA, [63] for Per-TCNQ, [53] for 10BTBT-TCNQ and $10{\text{BTBT-F}}_4\mathrm{TCNQ}$, [64] for DBTTF-TCNQ, [65] for NNTMB-CA, and [66] for ClMePD-DMeDCNQI.