An Old Crystallization Technique as a Fast, Facile, and Adaptable Method for Obtaining Single Crystals of Unstable “Li2TCNQF4” and New Compounds of TCNQ or TCNQF4: Syntheses, Crystal Structures, and Magnetic Properties

Detailed structural information is essential for understanding the properties of TCNQ and TCNQF4 compounds (TCNQ = 7,7,8,8-tetracyanoquinodimethane; TCNQF4 = 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane). The ineludible requirement of obtaining crystals of a size and quality sufficient to yield a successful X-ray diffraction analysis has been challenging to satisfy because of the instability of many of these compounds in solution. Crystals of two new complexes of TCNQ, [trans-M(2ampy)2(TCNQ)2] [M = Ni (1), Zn (2); 2ampy = 2-aminomethylpyridine], as well as unstable [Li2(TCNQF4)(CH3CN)4]·CH3CN (3), can be prepared in minutes by a horizontal diffusion technique and can be harvested easily for X-ray structural studies. Compound 3, previously described as “Li2TCNQF4,” forms a one-dimensional (1D) ribbon. Compounds 1 and 2 can also be obtained as microcrystalline solids from methanolic solutions of MCl2/LiTCNQ/2ampy. Their variable-temperature magnetic studies confirmed a contribution of strongly antiferromagnetically coupled pairs of TCNQ•— anion radicals at higher temperatures with exchange coupling J/kB = −1206 K and J/kB = −1369 K for 1 and 2, respectively, estimated using a spin dimer model. The presence of magnetically active anisotropic Ni(II) atoms with S = 1 in 1 was confirmed, and the magnetic behavior of 1, representing an infinite chain of alternating S = 1 sites and S = 1/2 dimers, was described by a spin-ring model suggesting ferromagnetic exchange coupling between Ni(II) sites and anion radicals.


■ INTRODUCTION
Compounds based on TCNQ [TCNQ = 7,7,8,8-tetracyanoquinodimethane; IUPAC 2,2′-(cyclohexa-2,5-diene-1,4-diylidene)dipropanedinitrile] and its derivatives manifest an ample variety of physical properties, with their electronic and magnetic behavior having been the subject of innumerable studies. Key to understanding the properties of TCNQ derivatives, particularly in solids, is the determination of their structures, with accurate information on bonding and intermolecular interactions involving TCNQ n− (n = 0, 1, 2, and intermediate between 0 and 2), which in turn requires the preparation of crystals adequate for structure determination, usually by in-house X-ray diffraction.
Some of the widely used procedures for the preparation of crystals such as solvent evaporation, solvent exchange, and vertical diffusion of different solutions require the compound to be in solution for hours to days, or even months because a slow evaporation or diffusion process favors the formation of fewer nuclei with more crystal growth. 1 TCNQ (Scheme 1) as a diamagnetic solid can be reduced easily, yielding the TCNQ •− anion radical (AR), and by further reduction, a diamagnetic dianion TCNQ 2− can be formed. 2 TCNQ in the form of the AR is interesting from a magnetic point of view since the presence of an unpaired electron makes it magnetically active (S = 1/2). The magnetic properties of numerous compounds based on the TCNQ AR have been studied in detail. 3−12 Besides magnetic properties, TCNQcontaining compounds were studied as promising semiconducting materials, 13,14 materials for quantum computing, 15 electric capacitors, 16 and gas sensors, 17 and as porous materials for gas storage and separation. 18 The TCNQ AR can act as a simple counter anion, or due to the presence of four cyanide groups, it can act as a ligand in coordination complexes; in the latter case, the TCNQ AR can be bound as a terminal ligand via an N atom from one of the cyanide groups, or it can act as a μ 2 -, μ 3 -, or μ 4 -bridging ligand. 2 From a magnetic point of view, complexes containing magnetically active 3d or 4f central atoms and TCNQ AR are heterospin systems in which the overall magnetic properties are determined by the magnetic properties of both constituents and by their interaction. In addition, the magnetic properties will also be affected by the arrangement of the TCNQ ARs; in the case of the frequently occurring π-dimerization of TCNQ units, their strong antiferromagnetic (AFM) coupling leads to a singlet ground state of the π-dimerized unit. 19 Heterospin complexes with Ni(II) possess the combination of spins S = 1 and S = 1/2, which can be considered as analogues to Ni(II)/ Cu(II) heterobimetallic complexes (S = 1, S = 1/2). 20 The Cambridge Structural Database 21 records more than 30 complexes of Ni(II) with TCNQ. Among these complexes, monodentate coordination of TCNQ AR to the Ni(II) central atom is infrequent, having been observed to date only in the complexes [NiL R (TCNQ) 2 ] (L R = 1,8-(C 2 H 4 OH) 2 -1,3,6,8,10,13-hexaazacyclotetradecane), 22 [Ni(cyclam)-(TCNQ) 2 ] (cyclam = 1,4,8,11-tetraazacyclotetradecane), 23 and [Ni(L2)(TCNQ) 2 ](TCNQ)·CH 3 COCH 3 (L2 = pentaazamacrocyclic ligand containing the 1-hexadecyl alkyl chain). 24 In all of these complexes, the TCNQ ARs occupy relative trans-positions in the octahedral coordination sphere of the central Ni(II) atom, while a cis-arrangement of TCNQ ARs was observed only in [Ni(trien)(TCNQ) 2 ] (trien = 1,4,7,10tetraazadecane), 23 as forced by the character of the blocking ligand trien.
It was observed some time ago 25 that recrystallization of these compounds frequently produces their degradation, and thus for structure analysis, it is important to obtain the compounds directly as single crystals to avoid the necessity of recrystallization. For instance, Li + salts of TCNQ or TCNQF 4 (TCNQF 4 = 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane) can be used as reactants for the preparation of TCNQ/ TCNQF 4 compounds; however, there have been few reports of structures of Li + with TCNQ or TCNQF 4 . Besides some organic salts in which Li + is bonded to crown ethers, 26−28 there are just two structures; in one of them, TCNQ is bonded to Zn, 29 and in the other one, 30 each Li + cation is coordinated by four TCNQF 4 2− groups forming a two-dimensional (2D) metal−organic framework.
Within our broader interest in the crystal structures and magnetic properties of heterospin Ni(II) complexes with the TCNQ AR, 12,31 we have prepared a novel heterospin complex [Ni(2ampy) 2 (TCNQ) 2 ] (1; 2ampy = 2-aminomethyl-pyridine) with two TCNQ ARs coordinated to the Ni(II) central atom. This compound has been prepared in two different ways, as microcrystals from a methanol solution of NiCl 2 ·6H 2 O and 2ampy with LiTCNQ and as single crystals of sufficient quality for X-ray structure determination from the reaction in methanol of crystals of an insoluble 2ampy coordination compound, [Ni(2ampy) 2 (NO 3 ) 2 ], with soluble LiTCNQ, in a fast, facile procedure designed to avoid decomposition. The phase identity of the compounds obtained through the two different procedures was established using X-ray powder diffraction. We have further tested this procedure with the preparation and structure analysis of single crystals of [Li 2 (TCNQF 4 )(CH 3 CN) 4 ]·CH 3 CN (3). The preparation of this compound in acetonitrile and its instability were described some ten years ago, 32 at which time it was reported that attempts to grow crystals suitable for diffraction analysis were unsuccessful.
Single crystals for the X-ray study were prepared by a horizontal diffusion method as follows: A few light purple crystals of [Ni(2ampy) 2 (NO 3 ) 2 ] of approximate dimensions 0.7 mm × 0.7 mm × 0.7 mm were added to 3 mL of MeOH in a Petri dish of 30 mm diameter. Then, 3.5 mg of solid LiTCNQ was added about 15 mm away from the purple crystals. A green/blue color developed instantaneously in the solution, and within 10 min, shiny dark crystals of compound 1 appeared on the surface of the purple crystals ( Figure 1).

Synthesis and Crystallization of [trans-Zn(2ampy) 2 (TCNQ) 2 ]
(2). The synthesis of 2 was performed by the same procedure as for 1, using ZnCl 2 (0.3 mmol; 0.045 g) instead of nickel chloride hexahydrate. Yield: 68%. CHN analysis (exp./calc., in %) for C 36  Single crystals of 2 for the X-ray study were prepared by a similar procedure to that used for the single crystals of 1, by horizontal diffusion using the complex [Zn(2ampy) 2 (NO 3 ) 2 ] prepared according to Tandon. 34 Within an hour, dark violet prisms of 2 had formed in the space near the insoluble Zn-containing starting material. These were harvested from the solution and analyzed.
Synthesis and Crystallization of [Li 2 (TCNQF 4 )(CH 3 CN) 4 ]· CH 3 CN (3). Single crystals of appropriate size and quality for the X-ray study were prepared by the horizontal diffusion method. In a crystallization dish with a radius of 30 mm were placed 2 mL of acetonitrile, 45 mg (0.36 mmol) of solid LiI, and 2 mg (7.24 × 10 −3 mmol) of solid TCNQF 4 , with LiI and TCNQF 4 separated by approximately 15 mm. Within 10 min, transparent needle-like crystals of 3 were formed near the original location of TCNQF 4 . The acetonitrile solution was withdrawn, and FomblinY oil was added immediately in order to cover the crystals and prevent decomposition. The crystals are stable under the oil for a time span of 10−30 min (Figures 2 and S1).
Instrumental Methods. C, H, and N elemental analyses for 1 and 2 were performed on a PerkinElmer 2400 Series II CHNS/O Analyzer. The infrared spectra for 1 and 2 in the range of 4000−300 cm −1 were recorded on a Jasco FT-IR 4600 spectrophotometer using the attenuated total reflectance (ATR) method. The X-ray powder diffraction patterns of bulk samples of 1 and 2 were measured on a RIGAKU D-Max/2500 diffractometer with a rotating anode and an RINT2000 vertical goniometer in the 2θ range 3.0−40°using Cu Kα radiation (λ = 1.5406 Å) and a 2θ step size of 0.02°; the model powder diffraction pattern was calculated using the program MERCURY. 35 X-ray Structural Analysis of [M(2ampy) 2 (TCNQ) 2 ] (M = Ni, Zn). Single-crystal X-ray data for [Ni(2ampy) 2 (TCNQ) 2 ] (1) were collected using the same crystal at 100, 173, and 296 K, and data for [Zn(2ampy) 2 (TCNQ) 2 ] (2) were collected at 100 K on an Oxford Diffraction Xcalibur diffractometer equipped with a Sapphire3 CCD detector and a graphite monochromator utilizing MoKα radiation (λ = 0.71073 Å). The CrysAlis software package 36 was used for data collection and reduction. Absorption correction was done using ABSPACK (multiscan method). 37 The structures were solved using SHELXT 38 and refined against F 2 using the full-matrix least-squares method with the program SHELXL-2018/3 39 incorporated in the WinGX program package. 40 Anisotropic displacement parameters were refined for all non-hydrogen atoms. The hydrogen atoms bonded to carbon atoms were included at idealized positions and refined as riders with isotropic displacement parameters assigned as 1.2 times the U eq values of their corresponding bonding partners.
The crystal that was used for all three analyses of the Ni(II) complex was found to be an aggregate of two components related by a rotation calculated to be 2.88°about the real-space vector (0.8498, − 0.4396, 0.2909) at T = 100 K, with a similar relationship holding at the other two temperatures. For the data at 100 K, the refined volume fraction for the second component was 0.369 (2). In addition, because of the nature of the twinning, with the accompanying difficulty of separating the intensities of partially overlapped low-angle reflections, and in view of the presence of outliers, after examination of the data and results, it was decided to exclude reflections outside the resolution range of 0.82−5.00 Å (8.15−51.36°2θ, ShelxL instruction "SHEL 5 0.82"). The low-angle limit excluded 12 reflections in addition to the 11 reflections that were rejected on the grounds of having been, in our judgment, affected by the beamstop. The high-angle cutoff was chosen on the basis of the inspection of quality indicators, including the CC(1/2) statistic. 41 It was hoped that the refinement of 1 under these conditions would give a result less affected by bias caused by the twinning. Other refinement conditions were also tested, including refinement with no exclusion based on resolution alone; the latter gives R 1 = 0.0697 and wR 2 = 0.1779, so any benefit from fine-tuning the data set in this manner is not reflected in the residuals. For the refinement of 1 at T = 296 K, the same resolution limits were used, while for T = 173 K, the high-angle limit was not used.
Structure graphics were drawn using the program Diamond. 42 The crystal data and final parameters of the structure refinements are summarized in Table 1, while selected geometric parameters are given in Table 2. Possible hydrogen bonds are gathered in Table 3. (3), covered with a thin layer of perfluorinated oil (FomblinY) and mounted on a MiTeGen micromount, was placed in the nitrogen cold stream of the diffractometer with as little exposure to the ambient atmosphere as was practically possible during the transfer. The cold-stream temperature was maintained at 103 K for the entirety of the diffraction measurements, and the sample showed no signs of decomposition.
We were unable to derive a recognizable structure solution using orthorhombic symmetry [and not limiting the space group choice to C222 1 ]; the analysis thereupon proceeded under the hypothesis that the sample was a twin by pseudo-merohedry. 43,44 In this stage of the analysis, the problem was treated as a case of monoclinic symmetry with pseudo-orthorhombic metrics in the twin. The structure solution and refinement proceeded routinely from that point. For the structure solution, the cell was reduced from the initial C-centered orthorhombic (oC) setting to primitive monoclinic (mP), with the cell parameters given in Table 1. The transformation from oC to the mP setting, by rows and based on unit-cell axes, is (−1 0 0/0 0 1/0.5 0.5 0), and the twin law for the monoclinic setting, also by rows, is (−1 0 0 / 0 −1 0 / 1 0 1). Both of these matrices can be derived using the program Platon. 45 On this basis, the original pseudoorthorhombic c-axis of 8.4555(3) Å (final refined value) corresponds to the monoclinic symmetry direction. The first space group choice for the monoclinic setting was P2 1 /n, based upon "systematically weak" groups of reflections. The twin law for this system obscures systematic absences for putative n-or c-glide planes since it produces an overlap of absent reflections from one twin individual with nonabsences from the other.
The structure was solved using ShelxD 46 with space group P2 1 /n and with the twin information incorporated into the peak list optimization stage of the calculations. The program revealed all 37 non-H atomic sites of the asymmetric unit, and where needed, those sites were assigned their correct element types manually. For the sake of curiosity, we also used ShelxT, which assigned the correct space group, P2 1 /n, and also revealed all 37 unique atomic sites with their correct elements assigned.
The solution from ShelxD was used to commence the refinement. The first test consisted of isotropic refinement of the 37 non-H atoms, with no twin law applied. This gave R 1 = 0.2345 and wR 2 = 0.6071, with a significant number of systematic absence violations (356 out of 1132) and with all of the 50 worst-fit reflections having Fo 2 significantly larger than Fc 2 . This preliminary refinement was repeated with the twin law included, giving R 1 = 0.1166 and wR 2   parameter for the second individual refined to a value near one-half. These and other considerations indicate a strong possibility of twinning, as has been expounded by Herbst-Irmer and Sheldrick. 47 In the final refinement, the non-hydrogen atoms were refined anisotropically, and no restraints or constraints were used for them. H atoms, all of which are in the methyl groups of five acetonitrile fragments, were placed with an idealized geometry about their carrier carbon atoms but with the initial torsion angle about the neighboring C−C bond established by a local difference Fourier calculation. During refinement, the H atoms were treated as riders and also allowed to rotate about the C−C bond but not to tilt.
The refinement converged with the residuals given in Table 1. The final refined value of the population parameter for the second individual was 0.480(2). Selected geometric parameters and possible hydrogen bonds are given in Tables 4 and 5.
Magnetic Studies. Magnetic properties were investigated using a Quantum Design MPMS-XL5. Measurements of the temperature dependence of the magnetic moment were performed in a magnetic field of 1 kOe in zero-field-cooled (ZFC) and field-cooled (FC) regimes at temperatures from 1.8 to 300 K. The field dependence of magnetization was measured in magnetic fields up to 5 kOe at temperatures of 1.8 and 4.5 K. The diamagnetic contribution of the (1) in a microcrystalline form. With the aim of better understanding the magnetic contribution from the TCNQ ARs, we also synthesized the analogous Zn(II) complex (2) (S = 0). None of our attempts to recrystallize the microcrystalline solids in order to obtain samples of acceptable size and quality for X-ray diffraction was successful, including using different solvents or diffusion of different combinations of solvents. Moreover, not even the compounds themselves were recovered from these attempts, an     indication that once the compounds are in solution, they decompose. This apparently is not unusual in this kind of chemistry, as similar behavior was observed long ago. 25 A different strategy was designed with the aim of producing the compounds directly in the form of crystals large enough and of sufficient quality for single-crystal X-ray diffraction analysis, thus avoiding recrystallization. However, one of the possible strategies, using a fast reaction, can produce microcrystals unsuitable for single-crystal X-ray diffraction; that is the case for the preparation using MCl 2 , 2ampy, and LiTCNQ in MeOH, as described above.
We found that a sufficiently fast reaction that produces single crystals of 1 and 2 that yield successful X-ray diffraction analysis occurs in methanol at room temperature when LiTCNQ is added in close proximity to insoluble crystals of [M(2ampy) 2 (NO 3 ) 2 ] (M = Ni, Zn). 33,34 The process can be followed under a microscope. Figure 1 shows crystals of 1 grown in about 10 min on the surface of insoluble purple crystals of [Ni(2ampy) 2 (NO 3 ) 2 ]. This is an easy, simple, and, importantly, fast procedure for obtaining crystals, which is often the most challenging step in preparing and characterizing a new product. As it involves a time frame of 10 min as opposed to several hours or even days, we were able to forestall decomposition. This process is not adequate, however, for producing a bulk sample because the relative insolubility of the Ni-containing starting material slows the production of larger quantities of product, once again permitting decomposition or the formation of a mixture of products. Thus, the two procedures, one for synthesizing the bulk product and the other for obtaining single crystals for structure analysis, must both be used. The former is a fast solution-based reaction with soluble reagents, and the latter has soluble LiTCNQ placed in close proximity to crystals of an insoluble precursor. The insolubility of one of the precursors seems to impede the formation of a large number of crystallization nuclei and therefore favors the growth of fewer crystals, as can be seen in Figure 1. To summarize, the method consists of using the reactants in close proximity, with one of them quite insoluble. Obviously, the new compound also has to be insoluble in the solvent used.
The identity of the bulk solids was confirmed by comparing their X-ray powder diffractograms to the calculated patterns based on the single-crystal X-ray structure analyses (see the Supporting Information, Figures S2 and S3). The purity and chemical identity of 1 and 2 were also corroborated by C, N, and H elemental analyses. We note that Ballester et al. 23 reported the synthesis and characterization of analogous Ni(II) complexes with TCNQ and tetra N-donor ligands (cyclam = 1,4,8,11-tetraazacyclotetradecane, trien = 1,4,7,10-tetraazadecane).
The infrared spectra of compounds 1 and 2 ( Figures S4 and  S5) are dominated by absorption bands located around 2200 cm −1 assigned to ν(C�N) stretching vibrations from TCNQ units. In the case of 1, two absorption bands at 2178 and 2159 cm −1 were observed, while in the case of compound 2, one additional weaker band was located (2173, 2158, and 2123 cm −1 ). Similar values (2186, 2182, and 2161 cm −1 ) were observed in the IR spectrum of [Ni(cyclam)(TCNQ) 2 ] with monodentate N-coordinated TCNQ ARs. 23 Also, in the spectra of both 1 and 2, other strong absorption bands originating from vibrations of ν(C�C), δ(CH), ω(C(CN) 2 ), and ν(C−C) were observed at around 1570, 825, 480, and 1330 cm −1 ; their tentative assignment was done with reference to the literature. 19,48 Synthesis

and Crystal Formation of [Li 2 TCNQF 4 -(CH 3 CN) 4 ]·CH 3 CN (3).
In order to test the applicability of the simple horizontal diffusion procedure for obtaining crystals of some unstable compounds for which speed can be key, we tested the method with the Li/TCNQF 4 system. The synthesis of pure bulk "Li 2 TCNQF 4 " in acetonitrile at 50−60°C has been reported; 32 however, attempts to obtain crystals were unsuccessful.
Crystals of [Li 2 TCNQF 4 (CH 3 CN) 4 ]·CH 3 CN (3) can be prepared by the addition of LiI and TCNQF 4 to acetonitrile under normal laboratory conditions. Figure 2 shows selected photos of the process as viewed under a microscope. The full process is recorded in the Supporting Information ( Figure S1). Initially (Figure 2a), the yellow solid TCNQF 4 is added to acetonitrile in a Petri dish (30 mm diameter); (Figure 2b), after LiI (white cream color, left superior part of the photo of the Petri dish) is added, very quickly (less than 30 s), a dark color appears in the TCNQF 4 zone. This color is typical of the TCNQF 4 •− radical, which is the first reduction step. In addition, a part of the methanol solution is now green (yellow + dark blue). In less than 1 min, some needles of compound 3 can be observed in the lower right part of the photo in Figure  2c,g. These needles continue appearing, now over the whole Petri dish, for some minutes (Figure 2d,h) and then disappear in approximately 10 min (Figure 2e). Finally, most of the dark blue solid disappears (Figure 2f).
A time span of around 10 min transpires from when the first crystals appear to the disappearance of the last one. We estimate that the "lifespan" of a given crystal, from when it visually appears under the microscope until it disappears, can be from 3 to 6 min. This agrees with the already described instability of "Li 2 TCNQF 4 " in acetonitrile. However, the elimination of the acetonitrile solution (removal by a pipette is sufficient) and the immediate addition of FomblinY oil serves to preserve the crystals long enough to select a single crystal and transfer it to the cold stream of a diffractometer for structure analysis. This simple procedure enables the identification even of species that disappear in a matter of minutes, as in the present case.
Horizontal diffusion techniques are not new, 49,50 although they do not seem subject to as much use as other methods, except for high-temperature applications. 51 For the present study, the application of a simple variant of this family of techniques has proved capable of producing direct crystallization with the limited nucleation needed to permit sufficient crystal growth. The method is also easily adaptable in terms of controlling diffusion speed, and importantly, in its simplest form, enabling facile crystal harvesting and mounting.

Crystal Structures of [M(2ampy) 2 (TCNQ) 2 ] [M = Ni(1), Zn(2)]. As the two crystal structures, [Ni(2ampy) 2 (TCNQ) 2 ]
(1) (Figure 3 for T = 100 K and Supporting Figures S6 and S7 for T = 173 and 296 K, respectively) and [Zn(2ampy) 2 -(TCNQ) 2 ] (2) (Figure S8), are isostructural, we will describe the crystal structure of 1 and the relevant information for 2 will be given in parentheses. The structure of 1 is formed by molecules of the complex [Ni(2ampy) 2 (TCNQ) 2 ], which reside on crystallographic centers of symmetry. The central Ni(II) [Zn(II)] atom is coordinated by a distorted octahedron consisting of two chelating 2ampy ligands in the equatorial plane and monodentate TCNQ •− ARs in the axial positions. Each axial ligand is coordinated via one N atom yielding an {N 6 } donor set. As of this writing, as has been mentioned the CSD 21 24 In contrast, we do not find any entries in the CSD for an analogous Zn(II) complex with an {N 6 } donor set and two monodentate TCNQ ARs. The nearest similarity to 2 is displayed by the dinuclear complex [Zn 2 (bpy) 4 (TCNQ) 2 (TCNQ-TCNQ)] (CUM-MOO), in which each Zn(II) atom is coordinated by two chelating 2,2′-bipyridine (bpy) ligands and one terminal TCNQ AR with a σ-dimerized (TCNQ) 2 2− dianion bridging the two metals. 52 We note that in this compound, the nitrogen donor atoms from TCNQ are in cis It was reported that KTCNQ exists in two polymorphic phases with a first-order phase transition at 122°C; the lowand high-temperature phases are both monoclinic, but they differ in cell parameters. 54−56 We explored the possibility of a temperature-dependent phase transition in 1, which, if present, could be expected to have interesting consequences for the magnetic properties. To this end, we analyzed the structure of 1 at T = 100, 173, and 296 K; no qualitative structural change was observed. Selected geometric parameters for 1 at T = 173 and 296 K are given in Table S1. We also note that the twinned nature of the crystal did not change with temperature. Selected crystallographic data for these measurements are listed in Table 1, where we observe that the changes are those that are expected with varying temperatures, an example being the unit-cell volume.
The extended structures of 1 and 2 are consolidated by N− H···N hydrogen bonds (Figures 4 and S9 and Table 3); these mediate the formation of 2D supramolecular aggregates dominated by hydrogen-bonded ring systems with a graphset symbol R 4 2 (26) (Figure 4, R1). The π-dimerization and consequent spin pairing of the quinoid rings of TCNQ ARs have implications for the magnetic properties of 1 and 2. The neighboring quinoide rings are coplanar and eclipsed, and the interplanar distances are 3.1236(2) Å for 1 and 3.1266(2) Å for 2. These short distances reflect strong π-dimerization. In the analogous complex [Ni(cyclam)(TCNQ) 2 ], 23 a similar distance between planes was observed [3.18(1) Å]. Even shorter distances were  4 2− are the internal part in such a way that the Li cations act as bridges between two TCNQF 4 2− and each TCNQF 4 2− fragment is bonded to four Li cations. The ribbon is thus formed by Li + and TCNQF 4 2− in a 2:1 ratio. The Li cations are also contacted by two terminal CH 3 CN groups, which complete a tetrahedral environment about Li + ( Figure 5). There is also one interstitial acetonitrile molecule per TCNQF 4 2− . The main distances and angles are collected in Table 4, and the possible hydrogen bonds are in Table 5.
Successive C 6 F 4 fragments along the ribbon are not coplanar, forming a dihedral angle of 35.02(11)°, as can be seen in Figure 6. The chains run parallel to the crystallographic a-axis with the interstitial acetonitrile molecules between the chains (Figure 7).
As was mentioned before, there are very few structures with Li and TCNQ or TCNQF 4 , and of those, only in the following cases is TCNQ or TCNQF 4 bonded to the Li cation. TCNQ •− is bonded to Li [15] 30 in which Li + is bonded to 4 TCNQF 4 2− groups. Compound 3, to our knowledge, is the first structurally characterized 1D ribbon formed by Li + and TCNQ -/2or TCNQF 4 -/2-. A related chain can be found in some Cu(I) 2 TCNQ or Cu(I) 2 TCNQF 4 compounds. 59,60 In fact, "Cu(I) 2 TCNQF 4 " with acetonitrile has been described; 60 however, its X-ray single-crystal structure could not be determined, and its structure was related to that of Cu(I) 2 TCNQF 4 (EtCN) 2 , which has CH 3 CH 2 CN instead CH 3 CN. The "Cu(I) 2 TCNQF 4 " chain in the propionitrile compound is similar to the one in compound 3, although with two interesting differences: (a) the Cu(I) centers have a coordination index of just 3, and (b) the TCNQF 4 groups in the ribbons are coplanar.
Magnetic Properties. The magnetic properties of 1 and 2 were studied in the temperature range of 1.8−300 K. No difference between ZFC and FC responses was observed, suggesting no magnetic ordering in both complexes. The value of χT = 0.04 emuK/mol of 2 at 300 K (Figure 8, inset), corresponding to the effective magnetic moment μ eff = 0.57μ B (μ B is the Bohr magneton), is evidence of the greatly reduced magnetic moment originating from two TCNQ ARs in a molecule of 2, each carrying spin 1/2. This reduction is expected due to the presence of strong π-dimerization between TCNQ ARs from neighboring molecules, as shown in Figure  S9. The appropriate magnetic model to describe magnetic properties represents an AFM spin dimer with a strong exchange coupling J, defined by a Hamiltonian

Crystal Growth & Design
which allows only a small population of magnetic spin-triplet states to be accessible at ambient temperatures. A small, almost constant value of χT below 200 K can be related to the presence of a minor concentration of nondimerized paramagnetic TCNQ ARs. As a result, the temperature dependence of the susceptibility was described by a model consisting of a paramagnetic contribution described by the Curie law and the Bleaney−Bowers formula 61 for AFM spin dimers.
where parameter c represents the reduced concentration of dimers. The best fit to the experimental data was obtained for exchange coupling J/k B = −1369 K and c = 0.995 (only 0.5% of TCNQ AR pairs are noninteracting, contributing to the paramagnetic susceptibility), with the g-factor fixed to g TCNQ = 2 in all presented calculations. The exchange coupling strength falls between the ones suggested in similar complexes with central Ni(II) ions by Ballester et al., 23,57 following the trend in distances between TCNQ ARs from neighboring molecules forming π-dimers. The low-temperature magnetization of 2 shown in Figure 9 can be described by the Brillouin function for the paramagnetic contribution of 0.7% of noninteracting (broken) TCNQ AR pairs (c = 0.993).
Regarding complex 1, one has to include the contribution of the Ni(II) ion (3d 8 , spin S = 1) present in the molecular unit, yielding χT = 1.262 emuK/mol at 300 K ( Figure 10, inset), corresponding to the effective magnetic moment μ eff = 3.178 μ B . A small increase in χT above 200 K is expected due to the strong π-dimerization between TCNQ ARs from neighboring molecules forming AFM spin dimers. At low temperatures, a drop in χT can be the consequence of zero-field splitting of Ni(II) ions and/or the effective AFM exchange interaction between them. Since at low temperatures, the magnetic moment of Ni(II) ions dominates, a negligible contribution of paramagnetic TCNQ ARs (expected to be similar to 2) will be omitted in the further analysis. In the first step, we neglected a possible interaction between Ni(II) ions and TCNQ ARs within a molecule, similar to refs 24, 57 but an effective coupling could be transmitted between Ni(II) in neighboring molecules through coordinated TCNQ ARs (πdimerized TCNQ ARs) and total susceptibility was defined by a simplified model.    A possible correction to the estimation of the diamagnetic contribution using Pascal's constants and temperatureindependent paramagnetism is included using χ 0 . The contribution of Ni(II) ions, χ Ni , is represented by the formula  (4) where N A is the Avogadro constant, g is the average g-factor, J′ is an exchange interaction with neighboring z ions introduced in the frame of mean-field theory 62 (in our case, z = 2 is assumed), and χ ZFS is the susceptibility of isolated Ni(II) ions described by the effective spin Hamiltonian, taking into account the zero-field splitting (ZFS) with parameters D and E.
Since the correct determination of the sign of a small D parameter from the analysis of the magnetic response of a polycrystalline sample may be ambiguous, we have used a simple approach to predict its sign and magnitude from the deformation of the coordination octahedron as proposed in ref 63. Using the structural data obtained at 296, 173, and 100 K, one can find that the sign of the D parameter changes between 296 and 173 K from positive to negative. Moderate values of D/k B = 4.9 K at 296 K, D/k B = −5.4 K at 173 K, and D/k B = −4.3 K at 100 K were obtained, and a strong rhombicity E/D is expected. Since the susceptibility is affected by such values of ZFS parameters only at low temperatures, we assumed a single negative D parameter as the starting value for our analysis. The calculation of temperature dependence of the susceptibility χ ZFS and corresponding field dependence of magnetization M ZFS based on eq 5 was performed using the EasySpin 64 toolbox in the MATLAB environment. The average magnetic response was calculated from the distribution of the magnetic-field vector over the orientational grid to account for the polycrystalline characteristics of the sample. The χT was then calculated using eq 3. The contribution of AFM spin dimers in low-temperature magnetization is negligible and was omitted. The calculation of M ZFS was performed, also neglecting effective interactions between Ni(II) ions since the inclusion of both the rhombic E parameter and effective interaction in the framework of the mean-field theory for the magnetization of a polycrystalline sample is not trivial. The simultaneous fit of χT and magnetization shown in ··· ··· ··· ··· and applying periodic boundary conditions (exchange coupling J between the endchain spins). The Hamiltonian of such a spin ring can be defined in the EasySpin toolbox, including the ZFS parameters for S = 1 sites, energy eigenvalues found by the exact diagonalization method, and the susceptibility and magnetization, can be calculated. The best agreement with the experimental data, as shown in Figures 10 and 11 by dashed lines, was then obtained for the set of parameters D/k B = −6.5 K, E/D = 0.2, average g-factor g Ni = 2.12, J/k B = −1105 K, J 1−1/2 /k B = 22 K, and χ 0 = 2 × 10 −4 emu/mol. A strong ferromagnetic (FM) exchange interaction J 1−1/2 between Ni(II) and ARs was also observed in Ni(TCNQ) 2 65 or a Ni(II)-verdazyl radical complex. 66 It is interesting that the ground state of such a model results in an effective antiparallel orientation of S = 1 spins for both possible FM and AFM coupling J 1−1/2 , which is in accordance with negative J′ obtained from a simplified model of weakly interacting Ni(II) ions and AFM spin dimers.

■ CONCLUSIONS
A fast and facile variation of well-traveled diffusion methods, consisting of adding the reactants in close proximity in a Petri dish with a solvent in which one of the reactants and the final product are insoluble or only slightly soluble, can produce suitable crystals for X-ray diffraction studies of TCNQ or TCNQF 4 compounds. This method can be extended to other compounds and can be useful when any of the chemicals is unstable in solution. In addition, intermediate compounds can be identified and isolated when following the reaction under a microscope.
Using this method, the following compounds have been prepared and identified crystallographically: (a) unstable [Li 2 TCNQF 4 (CH 3 CN) 4 ]·CH 3 CN (3) was prepared and analyzed structurally for the first time showing that Li + and TCNQF 4 2− form a 1D ribbon with two acetonitriles coordinated to each Li + and (b) two new isostructural complexes with TCNQ •− ARs, namely, [Ni(2ampy) 2 -(TCNQ) 2 ] (1) and [Zn(2ampy) 2 (TCNQ) 2 ] (2). Both compounds are neutral and formed by a central atom of Ni(II) or Zn(II) coordinated equatorially by two chelating 2ampy ligands and in the axial positions by monodentate TCNQ •− ARs. Measurements of the magnetic susceptibility and magnetization of compounds 1 and 2 were performed. A strong AFM exchange coupling within the pairs of TCNQ •� ARs from neighboring molecules was obtained, J/k B = −1206 K and J/k B = −1369 K for 1 and 2, respectively, from the spin dimer model. The presence of magnetically active anisotropic Ni(II) atoms with S = 1 in 1 was included in the model of an infinite chain of alternating S = 1 sites and S = 1/2 dimers; its simple version, a spin-ring model, was used to estimate the exchange coupling between Ni(II) and TCNQ •− ARs. ■ ASSOCIATED CONTENT
Full process of preparation of single crystals; powder patterns; IR spectra; molecular structures; hydrogen bonding system; and geometric parameters (PDF).