Sometimes the Same, Sometimes Different: Understanding Self-Assembly Algorithms in Coordination Networks

Abstract

The syntheses and characterizations of three new ligands containing two 4,2′:6′,4″-tpy or two 3,2′:6′,3″-tpy metal-binding domains are reported. The ligands possess different alkyloxy functionalities attached to the central phenylene spacer: n-pentyloxy in 3, 4-phenyl-n-butoxy in 4, benzyloxy in 5. Crystal growth under ambient conditions has led to the formation of {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n, with structures confirmed by single crystal X-ray diffraction. Both the cobalt(II) center and ligand 3 or 4 act as 4-connecting nodes and both {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n possess a 3D cds net despite the fact that 3 and 4 contain two 4,2′:6′,4″-tpy and two 3,2′:6′,3″-tpy units, respectively. Taken in conjunction with previously reported data, the results indicate that the role of the alkyloxy substituent is more significant than the choice of 4,2′:6′,4″- or 3,2′:6′,3″-tpy isomer in determining the assembly of a particular 3D net. The combination of Co(NCS)₂ with 5 resulted in the formation of the discrete molecular complex [Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH in which 5 acts as a monodentate ligand. The pendant phenyls and both coordinated and non-coordinated 4,2′:6′,4″-tpy units are involved in efficient π-stacking interactions.

1. Introduction

Polymeric materials are essential to mankind, both as the crucial biomolecules enabling life processes and as the materials at the core of our socioeconomic structures. Metallopolymers are an important subset of polymers, in which the incorporation of metal centers allows the blending of new properties which are not readily accessed through polymeric structures based only on C, H, N, O, S and P atoms. An important class of metallopolymers are the coordination polymers and networks, in which the metal centers are incorporated through coordination rather than covalent bonds. Specifically, coordination polymers and networks are infinite assemblies in which the metal centers are connected by coordinated organic molecules. The coordination number and coordination geometry of the metal centers provide an exquisite control over the dimensionality as well as the topographical and topological features of the coordination assemblies. Ligands with pyridine nitrogen donors comprise some of the most commonly utilized linkers, for example [1,2,3,4,5,6,7], with 4,4′-bipyridine (4,4′-bpy, Scheme 1) used extensively as a rigid organic linker in coordination assemblies. An [M(pytpy)₂]ⁿ⁺ (pytpy = 4′-(4-pyridyl)-2,2′:6′,2″-terpyridine) is an expanded 4,4′-bpy since both can bind two metal ions and possess similar vectorial properties (Scheme 1) [8]. Newkome and coworkers have demonstrated the assembly of metallocages based upon [M(tpy)₂]ⁿ⁺ building blocks with other vectorial properties [9]. By changing from the 2,2′:6′,2″-isomer of terpyridine to either 3,2′:6′,3″- or 4,2′:6′,4″-terpyridine (3,2′:6′,3″- or 4,2′:6′,4″-tpy, Scheme 2) it should be possible to modify the consequences of the coordination interactions on the network assembly in a controlled and predictable manner. The ligands 3,2′:6′,3″- or 4,2′:6′,4″-tpy both bind metal ions only through the four outer N-donors and there is no literature precedent for metal-coordination through the central pyridine ring. We and others have made widespread use of 4,2′:6′,4″-tpy ligands, typically functionalized in the 4′-position for the assembly of 1D-coordination polymers and 2D-networks [10,11,12]. Examples of 3D-architectures incorporating 4,2′:6′,4″-tpy ligands are rare [13] and only a very few examples of 2D- and 3D-assemblies directed by 3,2′:6′,3″-tpy ligands have been reported [14,15,16,17,18,19,20]. The most significant difference between a 4,2′:6′,4″-tpy and 3,2′:6′,3″-tpy is the effect that inter-ring C–C bond rotation has on the mutual directionality of the outer nitrogen donors. For 4,2′:6′,4″-tpy, there is no change in the vectorial orientation of the nitrogen donors, while for 3,2′:6′,3″-tpy there is significant variation with the interannular bond rotation [11]. This in turn increases the structural diversity to be expected with 3,2′:6′,3″-tpy, at the expense of the predictability of the self-assembly algorithm. In Scheme 2, the 3,2′:6′,3″-tpy unit is drawn in an orientation preorganized for the formation of metallomacrocycles, as observed with 1-(3,2′:6′,3′′-terpyridin-4′-yl)ferrocene [21] and 5-(3,2′:6′,3′′-terpyridin-4′-yl)-N,N-diphenylthiophen- 2-amine [22].

Covalent linkage of two 4,2′:6′,4″- or 3,2′:6′,3″-tpy domains combines two ditopic metal-binding domains into a tetratopic ligand as exemplified with the general ligands 1 and 2 in Scheme 2. We have developed a suite of such ligands, typically with 1,4-phenylene spacers functionalized with alkyloxy groups to increase solubility in organic solvents [23]. The nature of the OR group is critical in determining the outcome of the self-assembly. For example, the reaction of 1a or 1b (Scheme 2) with zinc(II) halides leads to (4,4) nets in which 1a or 1b is the 4-connecting node. However, in the case of 1a, simple 2D-sheets are observed, whereas the presence of the longer n-octyl groups in 1b leads to 2D→2D parallel interpenetration with the alkyloxy chains in extended conformations and directed within each corrugated sheet [23,24]. We have also demonstrated that introducing a pendant aryl group can facilitate face-to-face π-interactions within the solid-state assembly. In the case of [Zn₂Br₄(1c)^.H₂O]_n (see Scheme 2 for 1c), 2-fold interpenetrating nbo nets are observed with inter-net π-stacking resulting in tightly associated interpenetrated nets and, as a consequence, large void spaces in the lattice [25].

In this paper, we describe reactions of ligands 3–5 (Scheme 3) with Co(NCS)₂. Upon interacting with pyridine donors, Co(NCS)₂ typically generates octahedral cobalt(II) centers with trans-thiocyanato ligands and four pyridine nitrogen donors defining the equatorial plane. It is therefore expected to act as a 4-connecting node giving access to a range of different networks [26,27]. We have shown that the reaction of Co(NCS)₂ with 1d (Scheme 2) leads to a 3D cds net in which both metal and ligand are 4-connecting nodes [28]. Increasing the length of the alkyloxy chain to R = n-octyl and replacing the 4,2′:6′,4″- by a 3,2′:6′,3″-tpy (2a, Scheme 2) switches the 3D assembly from a cds to an lvt net [20]. In the present work, 3 was selected in order to investigate the effect of lengthening the alkyloxy chain from three to five carbon atoms—would the cds net be retained? Ligand 4 is comparable with 2a in terms of the 3,2′:6′,3″-tpy donor set but introduces terminal phenyl domains in the alkyloxy substituents. Both 3 and 5 possess 4,2′:6′,4″-tpy units, but the inclusion of the pendant phenyls in 5 was expected to affect the assembly by enabling arene–arene π-interactions.

2. Materials and Methods

2.1. General

¹H and ¹³C NMR spectra were recorded on a Bruker DRX-500 NMR spectrometer (Bruker BioSpin AG, Fällanden, Switzerland) with chemical shifts referenced to residual solvent peaks (TMS = δ 0 ppm). Electrospray ionization (ESI) mass spectra were measured on a Shimazu LCMS 2020 instrument (Shimadzu Schweiz GmbH, Roemerstr., Switzerland) and high resolution ESI (HR-ESI) mass spectra on a Bruker maXis 4G QTOF instrument (Bruker BioSpin AG, Fällanden, Switzerland); samples were introduced as MeCN solutions containing 0.1% formic acid or 0.1% trifluoroacetic acid (TFA) for electrospray mass spectrometry (ESI MS) and high-resolution ESI MS, respectively. Commercially available precursors were purchased from Fluorochem (Hadfield, UK), Sigma-Aldrich (Buchs SG, Switzerland), TCI (Eschborn, Germany) or Acros (Chemie Brunschwig AG, Basel, Switzerland) and used without further purification. Compound 4b was prepared following the literature and NMR spectroscopic data were in accord with those reported [29].

2.2. Compound 3a

Compound 3a was prepared by adapting a literature method [29]. 2,5-Dibromohydroquinone (2.0 g, 7.47 mmol), 1-bromopentane (2.36 mL, 2.88 g, 18.7 mmol) and anhydrous K₂CO₃ (3.1 g, 22.4 mmol) were dissolved in dry DMF (100 mL) and the mixture was heated at 100 °C for ca. 15 h. The reaction mixture was cooled to room temperature, added to a beaker containing 100 mL of ice-water and stirred for 30 min. The precipitate that formed was separated by filtration and washed with water (3 × 30 mL) and dried in vacuo. Compound 3a was isolated as a white powder (2.85 g, 6.98 mmol, 93.5%). ¹H NMR (500 MHz, CDCl₃) δ/ppm 7.09 (s, 2H, H^C3), 3.95 (t, J = 6.5 Hz, 4H, H^a), 1.81 (m, 4H, H^b), 1.46 (m, 4H, H^c), 1.39 (m, 4H, H^d), 0.94 (t, J = 7.2 Hz, 6H). ¹³C{¹H} NMR (126 MHz, CDCl₃) δ/ppm 150.2 (C^C2), 118.6 (C^C3), 111.3 (C^C1), 70.5 (C^a), 29.0 (C^b), 28.3 (C^c), 22.5 (C^d), 14.2 (C^e). In reference [29], ¹H NMR signals were unassigned, and ¹³C NMR spectroscopic data were not reported.

2.3. Compound 4a

Compound 4a has been reported but we find the following method more convenient. The method was as for 3a starting with 2,5-dibromohydroquinone (1.5 g, 5.6 mmol), 1-bromo-4-phenylbutane (0.96 mL, 3.04 g, 14 mmol) and anhydrous K₂CO₃ (2.32 g, 16.8 mmol) in dry DMF (100 mL). Then, 4a was isolated as a white powder (2.67 g, 5.02 mmol, 89.6%). The NMR spectroscopic data agreed with those published [29].

2.4. Compound 5a

Compound 5a has previously been reported [30], but we find the following synthesis more convenient. The method was as for 3a starting with 2,5-dibromohydroquinone (1.5 g, 5.6 mmol), benzyl chloride (1.61 mL, 1.77 g, 14.0 mmol) and anhydrous K₂CO₃ (2.32 g, 16.8 mmol) in dry DMF (100 mL). Then, 5a was obtained as a white powder (2.19 g, 4.89 mmol, 87.3%). ¹H NMR spectroscopic data were in accord with those reported [30]. ¹³C{¹H} NMR (126 MHz, CDCl₃) δ/ppm 150.2 (C^C2c), 136.3 (C^D1), 128.8 (C^D3), 128.3 (C^D4), 127.4 (C^D2), 119.4 (C^C3), 111.7 (C^C1), 72.1 (C^a).

2.5. Compound 3b

Compound 3a (1.8 g, 4.41 mmol) and dry Et₂O (150 mL) were added to a dried flask and cooled to 0 °C in an ice bath. ⁿBuLi solution (1.6 M in hexanes, 8.27 mL, 13.2 mmol) was slowly added to the solution of 3a over 20 min and the temperature maintained at 0 °C for 6 h. Dry DMF (1.02 mL, 13.2 mmol) was then added and the solution stirred for 15 h, while warming up to room temperature. The reaction was neutralized with saturated aqueous NH₄Cl solution and extracted with CH₂Cl₂ (200 mL). The organic phase was dried over MgSO₄ and concentrated in vacuo. Compound 3b was obtained as a yellow solid (1.14 g, 3.72 mmol, 84.4%) and used without further purification. ¹H NMR (500 MHz, CDCl₃) δ/ppm 10.51 (s, 2H, H^CHO), 7.42 (s, 2H, H^C3), 4.08 (t, J = 6.5 Hz, 4H, H^a), 1.87–1.76 (m, 4H, H^b), 1.52–1.31 (m, 8H, H^c+d), 0.93 (t, J = 7.2 Hz, 6H, H^e). ¹³C{¹H} NMR (126 MHz, CDCl₃) δ/ppm 189.6 (C^CHO), 155.3 (C^C2), 129.4 (C^C1), 111.7 (C^C3), 69.3 (C^a), 28.8 (C^b), 28.3 (C^c), 22.5 (C^d), 14.1 (C^e).

2.6. Compound 5b

The method was the same as for 3b but starting with 5a (2.0 g, 4.46 mmol) and ⁿBuLi (1.6 M in hexanes, 8.36 mL, 13.4 mmol). Compound 2b was obtained as a yellow solid (1.23 g, 3.55 mmol, 79.6%) and was used without further purification. ¹H NMR spectroscopic data matched those reported [29]. ¹³C{¹H} NMR (126 MHz, CDCl₃) δ/ppm 189.2 (C^CHO), 155.2 (C^C2), 135.8 (C^D1), 129.7 (C^C1), 128.9 (C^D2/D3), 128.5 (C^D4), 127.7 (C^D2/D3), 112.5 (C^C3), 71.3 (C^a).

2.7. Compound 3

Compound 3b (0.3 g, 0.98 mmol) was dissolved in EtOH (100 mL), then 4-acetylpyridine (0.45 mL, 0.49 g, 3.92 mmol) and crushed solid KOH (0.22 g, 3.92 mmol) were added in one portion. Aqueous NH₃ (32%, 7.8 mL) was added dropwise and the reaction mixture was stirred at room temperature for ca. 15 h. The precipitate that had formed was collected by filtration and washed with water, EtOH and Et₂O (3 × 10 mL, each). Compound 3 was obtained as a white solid (0.13 g, 0.18 mmol, 18.8%). M.p. = 297 °C. ¹H NMR (500 MHz, CDCl₃) δ/ppm 8.81 (m, 8H, H^A2), 8.15–8.07 (m, 12H, H^B3+A3), 7.16 (s, 2H, H^C3), 4.07 (t, J = 6.3 Hz, 4H, H^a), 1.87 (m, 4H, H^b), 1.37 (m, 4H, H^c), 1.32–1.19 (m, 4H, H^d), 0.77 (t, J = 7.3 Hz, 6H, H^e). ¹³C{¹H} NMR (126 MHz, CDCl₃) δ/ppm 154.8 (C^B2), 150.7 (C^C2), 150.6 (C^A2), 148.4 (C^B4), 146.4 (C^A4), 129.2 (C^C1), 121.7 (C^B3), 121.3 (C^A3), 115.3 (C^C3), 69.9 (C^a), 29.27 (C^b), 28.61 (C^c), 22.53 (C^d), 14.02 (C^e). ESI-MS m/z 754.40 [M + H + MeCN]⁺ (base peak, calc. 754.39), 713.35 [M + H]⁺ (calc. 713.36). HR ESI-MS m/z 713.3600 [M + H]⁺ (calc. 713.3599).

2.8. Compound 4

The method was as for compound 3 but starting with 4b (0.69 g, 1.62 mmol) and 3-acetylpyridine (0.82 mL, 0.89 g, 7.25 mmol), crushed KOH (0.41 g, 7.25 mmol) and aqueous NH₃ (32%, 12.4 mL). Compound 4 was isolated as a white solid (0.47 g, 0.56 mmol, 34.7%). M.p. = 253 °C. ¹H NMR (500 MHz, CDCl₃) δ/ppm 9.38 (d, J = 2.2 Hz, 4H, H^A2), 8.73 (m, 4H, H^A6), 8.50 (dd, J = 7.9, 2.0 Hz, 4H, H^A4), 8.01 (s, 4H, H^B3), 7.45 (dd, J = 8.0, 7.9 Hz, 4H, H^A5), 7.17–7.12 (m, 6H, H^C3+D3), 7.09 (m, 2H, H^D4), 6.98 (m, 4H, H^D2), 4.07 (t, J = 6.1 Hz, 4H, H^a), 2.55 (t, J = 7.5 Hz, 4H, H^d), 1.87–1.78 (m, 4H, H^b), 1.75–1.67 (m, 4H, H^c). ¹³C{¹H} NMR (126 MHz, CDCl₃) δ/ppm 154.9 (C^B2), 150.7 (C^C2), 150.4 (C^A6), 148.6 (C^A2), 148.1 (C^B4), 141.9 (C^D1), 134.8 (C^A3), 134.6 (C^A4), 129.5 (C^C1), 128.45 (C^D3), 128.4 (C^D2), 125.9 (C^D4), 123.8 (C^A5), 120.3 (C^B3), 115.5 (C^C3), 69.7 (C^a), 35.5 (C^d), 29.1 (C^b), 28.0 (C^c). ESI-MS m/z 837.40 [M + H]⁺ (base peak, calc. 837.39). HR ESI-MS m/z 837.3901 [M + H]⁺ (calc. 837.3912).

2.9. Compound 5

The method was as for 3, starting with 5b (1.06 g, 3.06 mmol), 4-acetylpyridine (1.56 mL, 1.7 g, 13.8 mmol), crushed KOH (0.77 g, 13.8 mmol), and aqueous NH₃ (32%, 24 mL). Compound 5 was isolated as a white solid (0.87 g, 1.16 mmol, 37.8%). Decomp. > 280 °C. ¹H NMR (500 MHz, CDCl₃) δ/ppm 8.77 (m, 8H, H^A2), 8.07 (s, 4H, H^B3), 7.96 (m, 8H, H^A3), 7.40–7.33 (m, 10H, H^D2+D3+D4), 7.31 (s, 2H, H^C3), 5.19 (s, 4H, H^a). ¹³C{¹H} NMR (126 MHz, CDCl₃) δ/ppm 154.9 (C^B2), 150.7 (C^C2), 150.5 (C^A2), 147.7 (C^B4), 146.3 (C^A4), 136.2 (C^D1), 129.5 (C^C1), 129.0 (C^D3), 128.7 (C^D4), 127.9 (C^D2), 121.5 (C^B3), 121.4 (C^A3), 116.1 (C^C3), 72.1 (C^a). ESI-MS m/z 794.30 [M + H + MeCN]⁺ (calc. 794.32), 753.35 [M + H]⁺ (calc. 753.30). HR ESI-MS m/z 753.2967 [M + H]⁺ (calc. 753.2973).

2.10. {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n

A solution of Co(NCS)₂ (1.75 mg, 0.01 mmol) in MeOH (8 mL) was layered over a solution of 3 (7.13 mg, 0.01 mmol) in 1,2-dichlorobenzene (5 mL). A few pink crystals of [Co(NCS)₂(3)·0.8C₆H₄Cl₂]_n were obtained after 2–4 weeks. Insufficient material was obtained for powder X-ray diffraction.

2.11. {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n

A solution of Co(NCS)₂ (3.5 mg, 0.02 mmol) in MeOH (8 mL) was layered over a solution of 4 (8.37 mg, 0.01 mmol) in 1,2-dichlorobenzene (5 mL). Pink crystals of [Co(NCS)₂(4)·1.6H₂O·1.2C₆H₄Cl₂]_n (2.2 mg, 0.0018 mmol, 18% based on 5) were obtained after 2–4 weeks. The bulk sample was characterized by powder X-ray diffraction (Figure S1).

2.12. [Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH

A solution of Co(NCS)₂ (3.5 mg, 0.02 mmol) in MeOH (8 mL) was layered over a solution of 5 (15.1 mg, 0.02 mmol) in CHCl₃ (5 mL). A small crop of pink crystals of [Co(NCS)₂(MeOH)₂(5)₂·2CHCl₃·2MeOH] was obtained after 2–4 weeks. Insufficient material was obtained for powder X-ray diffraction

2.13. Crystallography

Single crystal data were collected on a Bruker APEX-II diffractometer (Bruker Biospin AG, Fällanden, Switzerland); data reduction, solution and refinement used APEX2, SuperFlip and CRYSTALS respectively [31,32,33]. Structure analysis used Mercury v. 3.10 [34,35]. In {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n, the dichlorobenzene molecule was refined as rigid body (occupancy = 0.8 per formula unit). In addition, the aliphatic chain showed high thermal motion, and was refined isotropically using chemically reasonable restraints to bond distances and angles. In {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n, the side chain of the organic ligand is disordered with high thermal motion; as a result, the phenyl ring was refined with rigid body restraints and the whole chain was refined isotropically. Powder diffraction data were collected on a Stoe Stadi P powder diffractometer (Stoe & Cie GmbH, Darmstadt, Germany).

{[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n: C_52.80H_47.20Cl_1.60CoN₈O₂S₂, M = 1005.60, pink block, monoclinic, space group P2₁/c, a = 10.3756(6), b = 19.1855(11), c = 16.2699(9) Å, β = 106.881(3)^o, U = 3099.2(3) ų, Z = 2, D_c = 1.078 Mg m^–3, μ(Cu-Kα) = 3.749 mm⁻¹, T = 123 K. Total 31,013 reflections, 5731 unique, R_int = 0.033. Refinement of 4672 reflections (282 parameters) with I > 2σ (I) converged at final R1 = 0.1176 (R1 all data = 0.1247), wR2 = 0.1340 (wR2 all data = 0.1341), gof = 0.9888. CCDC 1877183.

{[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n: C_65.20H_56.00Cl_2.40CoN₈O_3.60S₂, M = 1217.36, pink block, monoclinic, space group P2₁/c, a = 13.7465(6), b = 15.7832(7), c = 16.2872(8) Å, β = 112.147(2)^o, U = 3273.0(3) ų, Z = 2, D_c = 1.235 Mg m^–3, μ (Cu-Kα) = 3.953 mm⁻¹, T = 123 K. Total 23,855 reflections, 5937 unique, R_int = 0.035. Refinement of 5459 reflections (313 parameters) with I >2σ (I) converged at final R1 = 0.1460 (R1 all data = 0.1460), wR2 = 0.1415, gof = 0.9888. CCDC 1877182.

[Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH: C₁₀₈H₉₀Cl₆CoN₁₄O₈S₂, M = 2047.77, pink block, triclinic, space group P–1, a = 14.1006(10), b = 14.5250(10), c = 14.5512(11) Å, α =62.394(3), β = 89.108(3), γ = 64.293(3)^o, U = 2314.5(3) ų, Z = 1, D_c = 1.469 Mg m^–3, μ(Cu-Kα) = 4.035 mm⁻¹, T = 123 K. Total 26,118 reflections, 8388 unique, R_int = 0.026. Refinement of 7965 reflections (618 parameters) with I > 2σ (I) converged at final R1 = 0.1090 (R1 all data = 0.2656), wR2 = 0.1117 (wR2 all data = 0.2664), gof = 0.9524. CCDC 1877184.

3. Results

3.1. Synthesis and Characterization of Ligands

Compounds 3–5 were prepared following the one-pot strategy of Wang and Hanan [36] (right-side of Scheme 4), an approach that requires the appropriate dicarbaldehyde. Precursors 3a–5a (Scheme 4) were prepared by treatment of 2,5-dibromohydroquinone with 1-bromopentane, 1-bromo-4-phenylbutane or benzyl chloride under basic conditions. Subsequent lithiation and treatment with DMF followed by addition of aqueous NH₄Cl yielded the dicarbaldehydes 3b–5b. After work-up, compounds 3–5 were isolated as white solids in yields ranging from 18.8 to 37.8%. In the HR-ESI mass spectrum of each compound, the highest-mass peak corresponded to the [M + H]⁺ ion (m/z = 713.3600, 837.3901 and 753.2967, respectively). For 3 and 5, this was the base peak, and lower intensity peaks at 357.1841 and 377.1522, respectively, arose from the [M + 2H]²⁺ ion. In the HR-ESI mass spectrum of 4, the [M + 2H]²⁺ ion gave the base peak (m/z = 419.1994), and a peak assigned to the [M + 3H]³⁺ ion (m/z = 279.8021) was also observed. The mass spectra are shown in Figures S2–S4. The ¹H and ¹³C NMR spectra of the ligands were assigned by COSY, NOESY, HMQC and HMBC methods, and the atom labelling scheme is defined in Scheme 4. NOESY cross peaks between the signals for protons H^A3 and H^B3 in 3 and 5 distinguished H^A3 from H^A2. Figure 1 displays the aromatic regions of the ¹H NMR spectra and Figures S5–S10 show full ¹H and ¹³C NMR spectra.

3.2. Crystal Growth

All crystal growth experiments were carried out under ambient conditions by layering a methanol solution of Co(NCS)₂ over a solution of the ligand in either 1,2-dichlorobenzene or chloroform (see Materials and methods). Parallel crystal-setups using 1,2-dichlorobenzene or chloroform for each ligand/Co(NCS)₂ combination were made, and the ratio of Co(NCS)₂: ligand was either 1:1 or 2:1. For ligands 3 and 4, pink crystals were harvested after 2–4 weeks, although in the case of 3, few crystals formed. X-ray quality crystals were only obtained when 1,2-dichlorobenzene was used during crystal growth, and single crystal X-ray diffraction established the assembly of coordination networks (see below). For ligand 5, crystals were only obtained when using a solution of 5 in chloroform, and X-ray diffraction revealed the unexpected discrete coordination compound [Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH.

3.3. {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n

Single crystal X-ray diffraction confirmed that reactions of Co(NCS)₂ with 3 and 4 yielded the coordination assemblies {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n. For the latter, verification that the single crystal was representative of the bulk sample was obtained from powder X-ray diffraction (Figure S1). Both {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n crystallize in the monoclinic space group P2₁/c. The cobalt(II) center in each compound is in a similar six-coordinate environment with trans-thiocyanato ligands and nitrogen donors from four different ligands defining the equatorial sites. The asymmetric unit in each structure contains one cobalt atom and half of a ligand 3 or 4. In {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n, the asymmetric unit contains 0.4 C₆H₄Cl₂, while in {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n, the asymmetric unit has 0.6 C₆H₄Cl₂ disordered over two orientations modelled with equal occupancies. For the compound with ligand 4, the pyridine ring containing N1 is disordered and was modelled over two sites of occupancies 0.6 and 0.4. The O atom of the 4-phenyl-n-butoxy chain is also disordered, and was again modelled over two sites with occupancies 0.6 and 0.4. In the discussion below, we focus only on the major occupancy sites in {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n. Figure 2 and Figure 3 illustrate that each of ligands 3 and 4 binds four cobalt (II) centers, thereby acting as a 4-connecting node.

Table 1. Important bond parameters in {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n.

Bond Parameter	{[Co(NCS)2(3)]·0.8C6H4Cl2}n	{[Co(NCS)2(4)]·1.6H2O·1.2C6H4Cl2}n
Distance/Å
Co1–N1	2.208 (2)	2.238 (4)
Co1–N4	2.062 (3)	2.065 (4)
Co1–N3v	2.179 (2)	2.169 (4)
Angle/deg
N1–Co1–N4	90.18 (10)	88.67 (15)
N3v–Co1–N1	84.73 (9)	93.81 (17)
N3v–Co1–N4	90.20 (11)	90.32 (16)
N3vi–Co1–N1	95.27 (9)	86.19 (17)
N3vi–Co1–N4	89.80 (11)	89.68 (16)
N1iv–Co1–N4	89.82 (10)	91.33 (15)

gives bond parameters within each cobalt (II) coordination sphere. These parameters and all other bond distances and angles in the structures are unremarkable, and do not warrant further comment.

The dihedral angles between the planes through adjacent aromatic rings in coordinated 3 and 4 are given in the first two entries in

Table 2. Dihedral angles between pairs of bonded aromatic rings in {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n, and in {[Co(NCS)₂(1d)]·2C₆H₄Cl₂}_n [28] and {[Co(NCS)₂(2a)]·4CHCl₃}_n [20].

Compound	Dihedral Angle between Planes/Deg
	Ring with N1/Ring with N2	Ring with N2/Ring with N3	Ring with N2/Phenylene Ring
{[Co(NCS)2(3)]·0.8C6H4Cl2}n	25.3	30.7	40.2
{[Co(NCS)2(4)]·1.6H2O·1.2C6H4Cl2}n	34.6	6.4	44.3
{[Co(NCS)2(1d)]·2C6H4Cl2}n	19.5	31.3	40.5
{[Co(NCS)2(2a)]·4CHCl3}n	29.9; 22.7	12.7; 12.2	58.5; 55.9

. The twist of the ring with N2 with respect to the phenylene ring in each compound is typical of covalently bonded arene rings and minimizes H…H repulsions. There are small differences in the conformations of the 4,2′:6′,4″- and 3,2′:6′,3″-tpy units in coordinated 3 and 4, as displayed in Figure 4a in the overlay of the {Co₄(3)} and {Co₄(4)} units. However, both 3 and 4 function as planar 4-connecting nodes and, despite the isomeric donor sets, each propagates into a {6⁵.8} cds net (Figure 5 and Figure S12) in which half of the adjacent nodes are mutually perpendicular and half are coplanar [26]. This resembles the 3D net observed in {[Co(NCS)₂(1d)]·2C₆H₄Cl₂}_n [28] and, for comparison, the corresponding dihedral angles for this structure are given in Table 1. In fact, the unit cell dimensions of {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(1d)]·2C₆H₄Cl₂}_n (both in the space group P2₁/c) are very similar (a = 10.3756(6), b = 19.1855(11), c = 16.2699(9) Å, β = 106.881(3)^o, U = 3099.2(3) ų versus a = 10.2136(9), b = 19.3452(17), c = 16.2214(15) Å, β = 107.027(3)^o, U = 3064.6(5) ų), revealing that the network can accommodate either n-pentyloxy or n-propoxy chains without significant perturbation (Figure S11).

The phenyl groups in the 4-phenylbutoxy chains in {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n do not play a critical role in the assembly process and are not involved in π-stacking interactions. This is in contrast to the 3-phenylpropoxy groups in [Zn₂Br₄(1c)^.H₂O]_n (see introduction) [25]. Instead, the 1,2-dichlorobenzene solvate in {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n engages in a face-to-face π-interaction with the pyridine ring containing atom N3 (Figure S13). Pertinent parameters are centroid…centroid and centroid…pyridine-plane separations of 3.74 and 3.49 Å, respectively, and an angle between the ring planes of 9.7°. However, since the solvent site is only partially occupied, one must be cautious in over-discussing both these interactions and the Cl…H–C contacts shown in Figure S13.

It is instructive to compare the structures of {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n (this work) and {[Co(NCS)₂(2a)]·4CHCl₃}_n [20]. These compounds both have 3,2′:6′,3″-tpy donor sets, but differ in the nature of the alkyloxy chain (4-phenylbutoxy versus n-octyloxy). A second difference is the solvent system for crystal growth (1,2-dichlorobenzene versus chloroform). Figure 4b illustrates an overlay of the {Co₄(4)} and {Co₄(2a)} units in {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n and {[Co(NCS)₂(2a)]·4CHCl₃}_n, and reveals a significant difference in ligand conformation. This arises from the larger dihedral angles between the 3,2′:6′,3″-tpy ring with N2 and the phenylene ring (Table 1). As previously described, {[Co(NCS)₂(2a)]·4CHCl₃}_n possesses a {4².8⁴} lvt net. It is now clear that the assembly of the lvt net is not simply a consequence of introducing the 3,2′:6′,3″-tpy in place of the 4,2′:6′,4″-tpy domain. Unfortunately, we have not been able to obtain good quality X-ray diffraction data from crystals grown from a combination of Co(NCS)₂ with 1b (Scheme 2) and are unable, therefore, to verify whether the switch from the cds to lvt net is a consequence of the steric demands of the long n-octyl chain.

3.4. [Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH

The discussion above demonstrates that consistent networks can be obtained with a combination of Co(NCS)₂ and tetratopic ligands incorporating either 3,2′:6′,3″- or 4,2′:6′,4″-tpy metal-binding domains. The observations suggest that the nature of the alkyloxy substituents, rather than the choice of tpy isomer, may be the critical factor in determining the outcome of the coordination assembly. On going from ligand 3 to 5, we gain the potential for π-stacking without significantly lengthening the alkyloxy chain. Layering an MeOH solution of Co(NCS)₂ over a CHCl₃ solution of 5 produced only a few crystals, and there was insufficient material to obtain powder XRD data for the bulk sample. Single crystal X-ray diffraction revealed the unexpected formation of the discrete molecular coordination compound [Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH. The compound crystallizes in the triclinic space group P–1, and the centrosymmetric structure is shown in Figure 6 with selected bond parameters given in the figure caption. Each ligand 5 acts as a monodentate N-donor ligand and MeOH molecules complete the octahedral coordination sphere of Co1. While {Co(NCS)₂(MeOH)₂(py)₂} (py = pyridine) motifs are well established [37,38,39,40,41,42], there is only one example in the CSD (v. 5.39 with updates [43]) featuring a monotopic 4,2′:6′,4″-tpy ligand 6 (Scheme 5) [44]. In this work, crystal growth by layering a MeOH/Co(NCS)₂ solution over a solution of 6 in MeOH/CH₂Cl₂ with an intermediate MeOH/CH₂Cl₂ layer resulted in the formation of [Co(NCS)₂(MeOH)₂(6)₂]. The preference for this discrete molecular assembly was attributed to the introduction of the sterically demanding pyrenyl group which is involved in efficient intermolecular pyrene…4,2′:6′,4″-tpy π-stacking interactions [44]. In the case of [Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH, packing interactions (Figure 7) involve centrosymmeric pairs of phenyl rings containing C20 and C20ⁱⁱ (symmetry code ii = 2 − x, 2 − y, 1 − z), pairs of 4,2′:6′,4″-tpy units containing N2/N4 and N2ⁱⁱⁱ/N4ⁱⁱⁱ (symmetry code iii = 1 − x, 2 − y, −z), and pairs of 4,2′:6′,4″-tpy units containing N6/N7 and N6^iv/N7^iv (symmetry code iv = 2 − x, −y, 2 − z). The first two sets of interactions lock together four molecules (Figure 7a). For the phenyl…phenyl π-stacking interaction, the separations of the ring planes and centroids are 3.35 and 3.65 Å, respectively. The coordinated 4,2′:6′,4″-tpy unit with N2/N3/N4 is virtually planar (dihedral angles between the ring planes = 7.4 and 6.0°), and 4,2′:6′,4″-tpys with N2/N4 and N2ⁱⁱⁱ/N4ⁱⁱⁱ engage in a π-stacking interaction (Figure 7a) with an angle between the ring planes of 6.0° and centroid…centroid distance of 3.60 Å. For the non-coordinated 4,2′:6′,4″-tpy unit, the dihedral angles between pairs of rings with N5/N6 and N6/N7 are 24.1 and 3.2°, respectively. Figure 7b depicts the face-to-face π-stacking interaction pairs of 4,2′:6′,4″-tpys containing N6/N7 and N6^iv/N7^iv (centroid…centroid = 3.83 Å). The overall packing in the lattice is, therefore, dominated by efficient π-stacking [45], but, because of the low yield of crystals, we are unable to say whether this leads to a preference for a discrete molecular structure in which three of the four potential metal binding sites of ligand 5 remain non-coordinated, or whether the isolation of crystals of [Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH is serendipitous.

4. Conclusions

We have prepared three new ligands containing two 4,2′:6′,4″-tpy or two 3,2′:6′,3″-tpy metal-binding domains with the expectation that they would function as 4-connecting nodes in coordination networks. The ligands possess different alkyloxy functionalities attached to the central phenylene spacer: n-pentyloxy in 3, 4-phenylbutoxy in 4, benzyloxy in 5. Reactions with Co(NCS)₂ under ambient conditions with MeOH and 1,2-C₆H₄Cl₂ as solvents resulted in the formation of {[Co(NCS)₂(3)]·0.8C₆H₄Cl₂}_n and {[Co(NCS)₂(4)]·1.6H₂O·1.2C₆H₄Cl₂}_n, both of which exhibit 3D cds nets. These data taken along with previously reported related structures suggests that the assembly of a particular net (cds or lvt) with 4-connecting Co and bis(tpy) ligands is independent of the choice of 4,2′:6′,4″- or 3,2′:6′,3″-tpy. The role of the alkyloxy substituent appears to be more significant. The combination of Co(NCS)₂ with ligand 5 resulted in the formation of the discrete molecular complex [Co(NCS)₂(MeOH)₂(5)₂]·2CHCl₃·2MeOH in which 5 behaves as a monodentate ligand. The pendant phenyls and both coordinated and non-coordinated 4,2′:6′,4″-tpy units engage in efficient π-stacking interactions. However, we are unable to identify whether the isolation of the mononuclear complex is preferred over network assembly, or is a serendipitous result. Further systematic investigations are needed in order to delineate the assembly algorithms in these systems.