Influence of the Size and Shape of Halopyridines Guest Molecules G on the Crystal Structure and Conducting Properties of Molecular (Super)Conductors of (BEDT ‐ TTF) 4 A + [M 3+ (C 2 O 4 ) 3 ] ∙ G Family

: New organic (super)conductors of the β″‐ (BEDT ‐ TTF) 4 A + [M 3+ (C 2 O 4 ) 3 ]G family, where BEDT ‐ TTF is bis(ethylenedithio)tetrathiafulvalene; M is Fe; A is the monovalent cation NH 4+ ; G is 2 ‐ fluoropyridine (2 ‐ FPy) (1); 2,3 ‐ difluoropyridine (2,3 ‐ DFPy) (2); 2 ‐ chloro ‐ 3 ‐ fluoropyridine (2 ‐ Cl ‐ 3 ‐ FPy) (3); 2,6 ‐ dichloropyridine (2,6 ‐ DClPy) (4); 2,6 ‐ difluoropyridine (2,6 ‐ DFPy) (5), have been prepared and their crystal structure and transport properties were studied. All crystals have a layered structure in which the conducting layers of BEDT ‐ TTF radical cations alternate with paramagnetic supramolecular anionic layers {A + [Fe 3+ (C 2 O 4 ) 3 ] 3 − G 0 } 2 − . Crystals 1 undergo a structural phase transition from the monoclinic ( C 2/ c ) to the triclinic ( P 1 ) symmetry in the range 100–150 K, whereas crystals 2 – 5 have a monoclinic symmetry in the entire range of the X ‐ ray experiment (100– 300 K). The alternating current (ac) conductivity of salts 1 − 4 exhibits metallic behavior down to 1.4 K, whereas the salt 5 demonstrates the onset of a superconducting transition at 3.1 K. The structures and conducting properties of 1 – 5 are compared with those of the known monoclinic phases of the family containing different monohalopyridines as “guest” solvent molecules G. the β″‐ ‐ TTF) 4 O)[Fe(C


Introduction
The large family of layered molecular (super)conductors (BEDT-TTF)4A + [M 3+ (C2O4)3]G, where BEDT-TTF is bis(ethylenedithio)tetrathiafulvalene; M 3+ is a magnetic or non-magnetic metal cation (M = Fe, Cr, Mn, Ru, Rh, Ga, Co, Al); G is a neutral "guest" molecule; A + is a small monovalent cation (A + = NH4 + , K + , H3O + , Rb + ), continues to be actively investigated . This is due to the fact that structural and conducting properties of this four-component system can be widely modified by changing the components A + , M 3+ and G. All crystals have a layered structure in which the conducting radical cation layers of BEDT-TTF alternate with insulating supramolecular anionic layers {A + [M 3+ (C2O4)3] 3− G 0 } 2− . Cationic and anionic layers interact with each other through the formation of a large number of hydrogen bonds between the components of the anionic layer and terminal ethylene groups of BEDT-TTF. The anionic layers have a familiar honeycomb-like structure in which M 3+ and А + cations linked by oxalate bridges form the hexagonal cavities in which neutral "guest" molecules G are incorporated. Cationic layers have different BEDT-TTF packing types depending on the chemical composition of the anionic layers.
The main feature of (BEDT-TTF)4A + [M 3+ (C2O4)3]G crystals is that the crystal symmetry, packing type of conducting BEDT-TTF layers, and therefore, the conducting properties of crystals depend mainly on the size and shape of the included "guest" molecule G [1,3,[17][18][19][20]. The variation of M (keeping the G the same) leads to a noticeable change of conducting properties, but does not affect the crystal symmetry and the type of packing of cation layers. The nature of the singly charged cation A + affects the properties of the crystals weakly. Currently, four groups of (BEDT-TTF)4A + [M 3+ (C2O4)3]G crystals are known, which have different symmetry and BEDT-TTF packing type.
Monoclinic β″-(BEDT-TTF)4A + [M 3+ (C2O4)3]G crystals were obtained with a large number of guest molecules G (G = PhCN, PhNO2, PhBr, PhCl, PhF, PhI, DMF, py, CH2Cl2, 1,2-dichloro-and 1,2-dibromorobenzene, different isomers of halopyridines). The conducting properties of these crystals vary from semiconducting to metallic and superconducting depending on the size, shape, and orientation of G in the hexagonal cavity, while the dimensions of the hexagonal cavity weakly depend on the size of G and are determined by the size of the M 3+ cation and the distance from the outer oxalate oxygen atoms to the cation A + . As a result, the large molecules G occupy almost the entire volume of the cavity and their position is strictly fixed. The smaller molecules G can occupy different positions. This leads to an increase in the structural disorder in the anionic layers and terminal ethylene groups of BEDT-TTF and to the suppression of the conductivity of molecular crystals. It was shown later that the conductivity of these crystals depends not so much on the molecular volume of G, as on its length along the direction c of the unit cell, [17]. Not so long ago, a correlation between the value of the c parameter of the unit cell and conducting properties of BEDT-TTF crystals with tris(oxalato)metallate anions was found [29].
The majority of these crystals keep the monoclinic symmetry in the entire range of the X-ray experiment (90-300 K). An exception is several β″-(BEDT-TTF)4(NH4 + )[Fe(C2O4)3]G crystals containing PhHal (Hal = F, Cl, Br) or a mixture of PhHal with PhCN as G. In these crystals, the superconducting transition is preceded by a structural phase transition from the monoclinic to triclinic state with decreasing temperature [21,22,25]. This structural transition arises from noticeable positional shifts of all components of the complex anion, giving rise to two nonequivalent organic β″-layers and the partial ordering of the ethylene groups of BEDT-TTF molecules. The consequence of these changes is the phase transitions of these crystals from metallic to mixed metallic/insulating states [32]. According to the theoretical concepts of Merino and McKenzie, the appearance of this state precedes the appearance of superconductivity in layered molecular conductors with a quarter-filled conduction band [33]. All of the aforementioned crystals containing PhHal as G support this conclusion.
The group of orthorhombic "pseudo-k"-crystals includes several semiconducting crystals (M = Fe, Cr, Co, Al; A + = K + , NH4 + , H3O + ; G = PhCN and its mixtures with PhNO2 or C6H4Cl2) [1,4,19]. These crystals grow in "dry" solvents G and also together with monoclinic crystals in the presence of traces of water in the reaction medium. The organic layer of these crystals is formed by charged [(BEDT-TTF)2] 2+ dimers surrounded by neutral [BEDT-TTF] molecules that are perpendicular to the dimers. Due to the strong localization of a charge on dimers, all orthorhombic crystals are semiconductors.

Synthesis
The reaction medium for the growth of (BEDT-TTF)4A + [M 3+ (C2O4)3]G crystal consists of the guest solvent either neat or mixed with other solvents (H2O, CH3OH, C2H5OH, CH3CN, etc.) which are usually not incorporated in the structure but have a considerable effect on the electrocrystallization process. In particular, they increase the solubility of the inorganic electrolyte in the guest organic solvent and, hence, enable variation of the current and the donor/electrolyte molar ratio, i.e., parameters that largely determine the crystal structure, growth rate and quality of crystals formed.
For example, in the medium of anhydrous C6H5CN, only semiconducting orthorhombic crystals grow, while both orthorhombic and monoclinic superconducting crystals (Tc = 7-8.5 K) in the form of needles were obtained by P. Day et al. using C6H5CN with a small addition of water [1]. These were the first superconducting crystals in this large family. Until now, the Tc of these crystals remains the highest among the crystals of this family. However, usually, only orthorhombic crystals form in this medium. In addition, needle crystals are of poor quality.
Interesting results were obtained by us when we used the mixture of C6H5CN with 1,2,4-trichlorobenzene (or 1,3-dibromobenzene) and C2H5OH as a reaction medium [19]. The molecules of 1,2,4-trichlorobenzene and 1,3-dibromobenzene do not enter the structure of crystals due to their size (1,2,4-trichlorobenzene) or geometry (1,3dibromobenzene), but their presence facilitates the formation of high-quality monoclinic crystals. In this case, monoclinic crystals grow in the form of thick plates or parallelepipeds. Unlike most crystals of BEDT-TTF salts, in which the conducting layers are arranged in the plate plane, in these crystals, the conducting layers are perpendicular to this plane.
In this work, the new monoclinic β″-(BEDT-TTF)4(NH4 + )[Fe(C2O4)3](HalPy) crystals (1-5) were also obtained by using the mixtures of 1,2,4-trichlorobenzene with different mono-and dihalopyridines as reaction medium (Table 2). Note that unlike all previously used guest components G, which are liquids, 2,6dichloropyridine (crystals 4) is a solid substance. It is possible that some other solids with interesting physical properties can also be included in the structure of the anionic layers for obtaining new multifunctional compounds combining two or more physical properties in the same crystal lattice.

Structure
Salts 1-5 of common formula β″-(BEDT-TTF)4(NH4 + )[Fe(C2O4)3](HalPy) crystallize in the monoclinic space group C2/c, with two crystallographically independent BEDT-TTF (A and B) molecules, half a [Fe(C2O4)3] 3− anion, half a NH4 + cation, and half a halopyridine guest molecule. The full cell contains two formula units. Salts 1-5 are isostructural with other monoclinic C2/c BEDT-TTF tris(oxalato)metallates packed by β″ type [27,29]. The structure consists of β″-type packed BEDT-TTF conducting radical cation layers separated by complex insulating anion layers {(NH4 + )[Fe(C2O4)3] 3-(HalPy)} 2-interleaved along the c axis ( Figure 1). Within the layers, BEDT-TTF radical cations form dimerized stacks with nearly coplanar molecular mean planes, incorporating radical cations expanding in a AABBAABBAABB manner with a shift of around a half of molecular short axis. Radical cations from adjacent stacks are arranged into chains expanding in the same dimerized manner. In the chains, BEDT-TTF radical cations are placed side-by-side and are nearly coplanar ( Figure 2).  An average charge on the BEDT-TTF is +1/2 obtained from the charge balance requirements. In the salt 5, the radical cation A is completely ordered, both terminal ethylene groups of the radical cation B are disordered over two positions with the occupancies of main positions of 0.61 (atoms C17, C18) and 0.74 (atoms C19, C20) for each terminal correspondingly, Figure S1. Minor positions are complementary to the main positions with the occupancies of 0.39 (atoms C17′, C18′) and 0.26 (atoms C19′, C20′). Mutual arrangements of ethylene groups in a BEDT-TTF respect eclipsed conformation in an ordered molecule A and staggered conformation in a molecule B for both main and minor disordered configurations ( Figure 3). The data of the other salts are summarized in Table 3.
It is known that the order and configuration of BEDT-TTF radical cation terminal ethylene groups play an important role in the charge transport of BEDT-TTF-based molecular conductors. From Table 3, it becomes immediately clear that, in all salts 1-5, the molecule A is fully ordered and adopts eclipsed conformation. Table 3. BEDT-TTF radical cation terminal ethylene group occupancies and mutual conformation for the salts 1-5. In the case of disordered groups, the occupancies of the main positions are presented. In the case of completely ordered groups, occupancies are equal to unity. Terminal A1 is defined as the ethylene group closest to the Fe 3+ ion, A2-an opposite side of the molecule. Terminals B1/B2 are defined by the same manner in the relation to the NH 4+ ion. Letter e denotes the eclipsed conformation while s denotes the staggered one. An example of e/s conformation is presented in Figure 3. Experiment temperature is 100 K unless otherwise noted.  Another clear observation is that the configuration of salt 1 at 240 K, and, especially at 150 K, resembles that of 5 at 100 K. Upon cooling in the range 100-150 K, salt 1 experiences a crystal lattice transformation from higher (monoclinic) to lower (triclinic) symmetry. A similar phase transition was observed earlier in the family of halobenzene solvents (PhF, PhCl, PhBr) [21] or their mixtures with PhCN [22] as well as with the guest molecule types closest to the current research-monosubstituted halopyridines, namely, 2-Cl-Py and 2-Br-Py (Table 1). In these cases, the structural transition was associated with a superconductive one. However, no superconductive transition is observed in salt 1. Let us speculate about possible reasons from the perspective of the crystal structure.
The structural transition results in halving of the unit cell volume and appearance of four independent BEDT-TTF radical cations (A, B, C, D). Molecule C is generated from A by the loss of symmetry, and molecule D is generated from B. It is seen from Table 3 that both molecules A and C are still ordered and adopt the same configuration while in the pair B/D the former is almost fully ordered but the latter, in contrast, demonstrates almost randomly distributed occupancies of the terminal group carbon atoms. Such a redistribution of occupancies is most likely accompanied by a charge redistribution that, in turn, results in weak charge localization and loss of superconductive transition. Indeed, analysis of the charge distribution over independent BEDT-TTF radical cations based on bond lengths enables us to roughly estimate the charges on A/B/C/D as +0.5/+0.33/+0.5/+0.67, respectively, indicating a charge redistribution in the pair B/D [37]. This formula, when applied to the structures with high R-values, often provides wrong absolute values and should not be taken too seriously. However, that gives us a good qualitative insight into the intermolecular charge distribution.
Within each anion layer, the Fe 3+ and NH4 + cations are linked by oxalate bridges forming a hexagonal packing arrangement where Fe 3+ is octahedrally coordinated by oxalate ligands. Outer oxygens of oxalates form hydrogen bonds with the NH4 + cations ( Figure 4). This anionic network forms large hexagonal cavities where halopyridine guest molecules reside. Fe 3+ cations and nitrogen of NH4 + cations reside in special positions. There are twofold symmetry axes connecting them.
There are two chemically and structurally reasonable alternatives for a cation incorporating into an anion layer-ammonium (NH4 + ) or oxonium (H3O + ). It should be noted that the unambiguous identification of the cation is a challenging task that has a long history. That is especially a source of doubt for compounds demonstrating superconductive properties, as there is an empirical rule that H3O + favors the superconducting transition [1,27,29]. In the particular case of superconducting salt 5, we consider the cation to be NH4 + , guided by the following arguments: 1. Both symmetrically unequivalent hydrogen atoms were found on the electronic differential map; 2. Adding each of the hydrogens results in a subtle but sensible decrease of the Rvalue; 3. Replacing N by O does not improve the R-value; 4. NH4 + cation is residing on a twofold axis and adopts distorted tetrahedral geometry that has been restored to a symmetrical tetrahedron by geometrical constraints without loss of the R-value.
It should be emphasized that, while molecules of 2,6-DFPy and 2,6-DClPy have their own molecular twofold symmetry axes, they still reside in the cavities with the substantial displacement of approximately 0.8 Å with no atoms occupying special positions, resulting in a positional disorder over two positions around the lattice twofold axis ( Table 4, Figures  5 and 6). It is most likely a consequence of the requirement to completely fill up the cavity void space by a smaller molecule. However, the larger chlorine-containing 2,6-DClPy is almost 0.1 Å less displaced than the fluorine-containing 2,6-DFPy but requires a larger tilt to be accommodated.   Figure 5.  is the angle between the pyridine ring plane and the plane of the hexagonal cavity. e is a mean guest molecule displacement defined as (e1 + e2)/2, where e1 is a displacement of the outer halopyridine halogen atom nearest to the cusp Fe 3+ ion around the twofold axis, e2 is a displacement of the inner halopyridine ring carbon/nitrogen atom closest to the nadir nitrogen atom of NH4 + ion. e' is an alternative mean displacement value (e1 + e2+ ... + en)/n where averaging is taken for all displaced atoms. S is a square unit of the hexagonal cavity taken as w(b + h)/2.  It appears that the smallest 2-FPy guest molecule that lacks any molecular symmetry presents the most interesting case. At higher temperatures, the arrangement of 2-FPy in the cavity resembles 2,6-DFPy/2,6-DClPy cases with a much smaller displacement. However, upon cooling, the fluorine atom and the two pyridine carbon atoms (C41 and C44) are situated on the lattice twofold axis (Figure 7). This effect is already clear at 150 K-far away from the structural transition. Thus, 2-FPy localization precedes or even induces a complete structural transformation. Nitrogen and carbon positions of 2-FPy appear to be inseparable at temperatures 150 and 100 K; that is probably due to the formation of a domain superstructure. The elevated electric field strength along the domain grains may serve as a factor causing failure of the superconductive transition. However, such speculations require further deeper investigations.

Salts
Unlike all other salts presented in the current work, salt 3, with the most asymmetrical halopyridine molecule-2-Cl-3-FPy, contains an NH4 + cation displaced from the lattice twofold axis and shows an oxalic ligand disorder in the tris(oxalato)ferrate anion ( Figure 8). Guest solvent molecules could not be identified by means of crystal structure tools and were refined as a set of separated carbon atoms plus fluorine with released occupancies. It appears that a highly asymmetric guest solvent molecule with substituents of different sizes does not favor growth of high-quality crystals.
Though the structure-property correlation is a very complex and ambiguous topic, some preliminary conclusions can be provided: 1. Use of highly nonsymmetrical halopyridine 2-Cl-3-FPy leads to low quality single crystals; 2. Guest molecules of higher symmetry and larger volume may produce superconductive single crystals; 3. Ordering of non-symmetrical guest molecules at lower temperatures may impede the SC transition by casting a constant dipole moment at the place of the localization.

Сonducting Properties
The temperature dependences of the out-of-plane resistivity for the samples 1-5 are presented in Figure 9. All these samples are metals despite the existence of the negative slopes on the low-temperature parts on the (T) curves for the samples 5 and 1. Moreover, the sample 5 demonstrates a sharp resistance drop at low temperature which one can surely attribute to the onset of a superconducting transition at Tc = 3.1 K. Unfortunately, we could not reach the real zero resistance state, even at our minimal temperature 1.4 K, because the superconducting transition is not very narrow. This is typical for organic metals, for which the transition width is usually more or about 2 K. The behavior of the sample 1 at T < 20 K is similar to that of the sample 5, so we do not exclude the possibility that this sample could also be a superconductor at temperatures below 1 K, but we could not check this statement because our temperature region was restricted by 1.4 K.

Synthesis
Crystals of five new radical cation salts of the β″-(BEDT-TTF)4A + [M 3+ (C2O4)3]G family were obtained by electrochemical oxidation of BEDT-TTF at a platinum electrode in organic solvents, at constant current and temperature (25 °C), in the presence of supporting electrolytes, namely, ammonium tris(oxalato)metallates combined with a 18crown ether. The electrocrystallization process was performed in conventional twocompartment U-shaped cells separated by a porous glass filter. BEDT-TTF, (NH4)3[Fe(ox)3]•3H2O, 18-crown ether and solvents were placed in the cathode compartment and, then, the solution obtained was distributed between the two compartments of the cell.
The exact conditions for the synthesis of each salt are described below. A total of 18 mg of BEDT-TTF, 150 mg of (NH4)3Fe(ox)3, 450 mg of 18-crown-6 were dissolved in the mixture of 15 mL of 2,6-DFPy, 5 mL 1,2,4-trichlorobenzene and 2 mL of 96% ethanol. J = 0.9 μA. Several thick crystals in the form of plates were collected from the anode after 6 days.

Structure
X-ray diffraction analyses of the salts 1-5 were carried out on a CCD Agilent XCalibur diffractometer with an EOS detector (Agilent Technologies UK Ltd., Yarnton, Oxfordshire, England). Data collection, determination and refinement of unit cell parameters were carried out using the CrysAlis PRO program suite [38]. X-ray diffraction data at 100(1) K for the salts 1-5 were collected using MoKα (λ = 0.71073 A) radiation. The same single crystal for the salt 1 was used for the data collection at different temperatures. Data at 100(1) K for the experiment 1a, data at 150(1) K for the experiment 1b, and data at 240(1) K for the experiment 1c were seamlessly collected without removing the crystal from the diffractometer.
The structures 1-5 were solved by the direct methods. The positions and thermal parameters of non-hydrogen atoms were refined isotropically and then anisotropically by the full-matrix least-squares method. The positions of the hydrogen atoms were calculated geometrically. The guest molecule was found to be disordered over two positions. The geometry of guest molecule was recovered for salts 1-5 except salt 3. Guest solvent molecules in salt 3 were refined as a set of separated carbon atoms plus fluorine with released occupancies.

Conducting Properties
The temperature dependences of the electrical resistance of single crystals were measured using a four-probe technique by a lock-in detector at 20 Hz alternating current J = 1 mkA in the temperature range (1.4-300 K). Two contacts were attached to each of the two opposite sample surfaces with conducting graphite paste. In the experiment, we have measured the out-of-plane resistance R with the current running perpendicular to the conducting layers. Because of the high anisotropy of the samples, we did not succeed in the measurements of the in-plane sample resistance. The out-of-plane resistivity  of the samples was calculated from R taking into account that the out-of-plane current is uniform due to the high sample anisotropy.
Crystals 4 (G = 2,6-DClPy) and 5 (2,6-DFPy) contain the highly symmetric G molecules. Each of these molecules contains two identical substituents. In this case, there were no problems with the quality of the crystals formed. Crystals 4 demonstrate metallic properties down to 1.4 K, while crystals 5 show the onset of a superconducting transition at 3.1 K. This is the first superconductor among crystals of the β″-(BEDT-TTF)4(A + )[M 3+ (C2O4)3]G family containing dihalopyridines or dihalobenzenes as G.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.