Poly[μ-2-amino­pyrazine-κ²N¹:N⁴-μ-cyanido-copper(I)]: a three-dimensional network from laboratory powder diffraction data

In the title compound, [Cu(CN)(C₄H₅N₃)]_n or [Cu(μ-CN)(μ-PyzNH₂)]_n (PyzNH₂ is 2-amino­pyrazine), the Cu^I center is tetra­hedrally coordinated by two cyanide and two PyzNH₂ ligands. The Cu^I–cyano links give rise to [Cu–CN]_∞ chains running along the c axis, which are bridged by bidentate PyzNH₂ ligands. The three-dimensional framework can be described as being formed by two inter­penetrated three-dimensional honeycomb-like networks, both made of 26-membered rings of composition [Cu₆(μ-CN)₂(μ-PyzNH₂)₄].

S1. Comment

Metal-organic networks of Cu^I–cyano and bridging diimine ligands have potential applications in gas storage and catalysis. Metal-organic materials containing luminescent metals, such as copper(I), and having suitable network porosity can potentially be used in gas molecule sensing systems, since the inclusion of small molecules into the network can alter the luminescent behavior of the material. The wide variety of bonding modes of Cu^I–cyano units allows the preparation of new materials of a large number of structural types, where the Cu^I coordination often varies from 2 to 5. Furthermore, the assembly of Cu^I and asymmetrically substituted diimine ligands can lead to chiral two- and three-dimensional networks, which could find applications in asymmetric catalysis and nonlinear optics (Teichert & Sheldrick, 1999, and references therein).

We have recently prepared a variety of new Cu^I–cyano diimine compounds by open reflux reactions and hydrothermal syntheses, and have investigated their luminescence properties (Tronic et al., 2007). The ligands studied included pyrazine (Pyz), 2-aminopyrazine (PyzNH₂), quinoxaline (Qox), phenazine (Phz), 4,4'-bipyridyl (Bpy), pyrimidine (Pym), 2-aminopyrimidine (PymNH₂), 2,4-diaminopyrimidine [Pym(NH₂)₂], 2,4,6-triaminopyrimidine [Pym(NH₂)₃], quinazoline (Qnz), pyridazine (Pdz), and phthalazine (Ptz). As part of this study, [CuCN(µ-PyzNH₂)], (I), was prepared by an open reflux reaction. This reaction did not produce single crystals suitable for structure determination, but the structure of (I) (Fig. 1) has been solved from the X-ray powder diffraction pattern collected at room temperature.

Only a few 1:1 [CuCN(µ-ligand)] complexes have been previously reported, all containing Cu^I with distorted tetrahedral coordination. In the crystal structure of [CuCN(µ-pdvb)] [pdvb is bis(4-pyridyl)-trans-1,4-divinylbenzene], the long bidentate pdvb ligand favors the formation of corrugated sheets, each sheet being composed of parallel-running [Cu–pdvb]∞ zigzag chains (Hanika-Heidl et al., 2003). The sheets are linked together through cyano bridges that give rise to [Cu–CN]∞ chains running along the direction perpendicular to the plane of the sheets. This structure exhibits remarkably different Cu^I···Cu^I separations of 4.804 (via CN^-) and 20.366 Å (via pdvb). The pdvb Cu—N interatomic distances are 2.249 and 2.173 Å, whereas the cyano Cu—C and Cu—N distances are 1.880 and 1.941 Å respectively.

[CuCN(µ-2-MePyz)] (2-MePyz is 2-methylpyrazine) and [CuCN(µ-4-MePym)] (4-MePym is 4-methylpyrimidine) form three-dimensional frameworks in which one-dimensional [Cu–CN]∞ chains are bridged by the linear bidentate aromatic ligands (Teichert & Sheldrick, 1999). In [CuCN(µ-2-MePyz)], 22-membered rings of [Cu₆(CN)₄(µ-2-MePyz)₂] composition form a chiral honeycomb-like two-dimensional network; these two-dimensional units are connected by additional 2-MePyz ligands, giving rise to a porous and chiral three-dimensional framework in P2₁2₁2₁. On the other hand, the linear ligand 4-MePym bonds to Cu^I atoms at shorter distances owing to its intramolecular geometrical disposition of N atoms, and it forms tetramers in a centrosymmetric structure containing disordered cyano C and N atoms. Furthermore, regardless of the asymmetric nature of the 4-MePym ligand, [CuCN(µ-4-MePym)] crystallizes in the non-chiral space group P4₂/n.

[CuCN(µ-Pyz)] forms a three-dimensional network in P2₁/c [a = 6.208 (2) Å, b = 9.158 (2) Å, c = 11.198 (2) Å and β = 90.89 (3)°], wherein each Cu^I center is tetrahedrally coordinated to two cyano and two Pyz units. [Cu–CN]∞ chains run along the b axis and they are connected by two Pyz ligands, giving rise to 22-membered rings of composition [Cu₆(CN)₄(µ-Pyz)₂]. Alternatively, we can describe the three-dimensional network as formed by two interpenetrated, but otherwise symmetrically equivalent, three-dimensional honeycomb-like arrays made of 26-membered rings of composition [Cu₆(CN)₂(µ-Pyz)₄] [see inset (a) in Fig. 2]. Offset face-to-face (π–π) interactions within Pyz ligands are present, with a centroid-to-centroid distance of 4.58 (6) Å. The Pyz Cu^I—N distances are 2.159 (12) and 2.134 (12) Å, whereas the cyano Cu^I—C and Cu^I—N distances are 1.920 (18) and 1.944 (19) Å, respectively, and the Cu^I tetrahedral angles are 99.3 (5), 133.1 (8), 99.8 (6), 111.1 (8), 101.4 (6) and 107.2 (8)°. The existence of some C/N disorder (not refined) in the cyano ligands is reported for this structure (Kuhlman et al., 1999).

A similar packing was found for (I). The refined unit-cell parameters (after axes permutations) are also similar, even though [CuCN(µ-Pyz)] and [CuCN(µ-PyzNH₂)] belong to the monoclinic and orthorhombic crystal systems, respectively. Fig. 2 shows the distorted tetrahedral coordination adopted by atom Cu1. This figure also shows one of the two interpenetrated three-dimensional honeycomb-like arrays formed by 26-membered rings of [Cu₆(CN)₂(µ-PyzNH₂)₄] composition, which in turn generate the three-dimensional network of [CuCN(µ-PyzNH₂)]. The compound crystallizes in the noncentrosymmetric and nonchiral space group Pca2₁ (29), and the cyano C and N positions for (I) refine as ordered. Additionally, offset face-to-face (π–π) interactions between the PyzNH₂ ligands with a centroid-to-centroid distance of 4.16 (5) Å are found.

In view of the chemical similarity of Pyz and PyzNH₂, as well as of the crystal packing, it is reasonable that the ligand Cu—N distances found are very close for the two complexes. The differences in the tetrahedral coordination angles around Cu^I are slightly larger.

Related literature

For related literature, see: Altomare et al. (2004); Boultif & Louër (2004); Giacovazzo et al. (2002); Hanika-Heidl, El-din, Etaiw, Ibrahim, El-din & Dieter Fisher (2003); Larson & Von Dreele (2000); Le Bail, Duroy & Fourquet (1988); Scott (1983); Teichert & Sheldrick (1999); Tronic et al. (2007).

S2. Experimental

Compound (I) was prepared by open reflux synthesis as described by Tronic et al. (2007). Equimolar quantities of Cu^I cyanide and KCN were suspended in water and warmed. PyzNH₂ was added in half of the previous molar quantity and the suspension was refluxed overnight under N₂. The reaction mixture was then filtered, and the solid was washed with water, ethanol and diethyl ether and then dried under vacuum. A yellow powder was isolated. The C, H and N elemental compositions were measured by standard techniques, and the Cu content was determined by atomic absorption spectroscopy.

S3. Refinement

The powder diffraction pattern was indexed without impurity peaks with the program DICVOL04 (Boultif & Louër, 2004) to the orthorhombic unit-cell parameters a = 11.5353 Å, b = 8.3093 Å, c = 6.7841 Å, α = β = γ = 90.0° and M₁₉ = 51, with unit-cell volume of 650.3 ų. Three additional monoclinic unit cells were found with similar a, b and c values (differently permuted), β = 90.194, 90.338 and 90.159°, and M₁₉ = 12.1, 14.2 and 14.2.

Le Bail fits (Le Bail et al. 1988) performed with the program GSAS (Larson & Von Dreele, 2000) in P222 confirmed the validity of the orthorhombic unit-cell parameters. Additional fits were carried out in P2 with monoclinic lattice parameters. Starting from the fits with the largest β angle, the refined monoclinic parameters were a = 11.531 (1) Å, b = 6.7832 (6) Å, c = 8.3018 (8) Å, α = 90°, β = 90.001 (2)° and γ = 90°, showing essentially the same agreement factors as the orthorhombic unit cell. The estimated density value of 1.9 Mg cm^-3 suggested Z = 4.

The choice of space group symmetry was not immediate from the powder pattern. The observation of the systematic absences first suggested P2₁2₁2 (18), and a plausible structural model was found with the direct methods program EXPO2004 (Altomare et al., 2004) using a data set collected in the 2θ range 14–100°, with a step size of 0.025° and time of 20 s per step. In this model, the Cu^I coordination was distorted tetrahedral and all PyzNH₂ atoms as well as the cyano C and N atoms were found. However, a 1.48 Å N—N contact within the amine groups of the PyzNH₂ ligands was present, and a Rietveld refinement of this model was not satisfactory. Since it was likely that less obvious systematic absences had been overlooked, an exhaustive search of other possible space groups yielded only Pb2₁a (29) or Pbma (57). Pbma was discarded after an EXPO2004 run which did not give rise to an alternative model. However, it was possible to find all atoms of the asymmetric unit of (I) in a default run of EXPO2004 in Pb2₁a. The unit cell was transformed to the standard space group setting Pca2₁ (29). Once approximate Cu1 atomic positions had been determined using EXPO2004, the program PSSP (P. W. Stephens & S. Pagola, Powder Structure Solution Program, https://powder.physics.sunysb.edu/programPSSP/pssp.html), using direct-space methods and a simulated annealing algorithm, was applied to determine the location of the cyano and PyzNH₂ ligands. The crystal structure was found in almost all runs with an agreement factor S of 0.012, using 366 reflections and 50 000 cycles, and with initial temperature, final temperature and decrement factor values of 50, 0.001 and 0.8, respectively. The molecular geometry of the PyzNH₂ ligand was taken from the Cambridge Structural Database (Allen, 2002) entry AMPYRZ (provide full reference), including H atoms for structure solution.

The Le Bail fit obtained using a second data set collected with higher counting statistics in the 2θ region 80–135° had R_wp = 1.89% and χ² = 1.6. The Rietveld refinement was carried out using the program GSAS (Larson & Von Dreele, 2000), confirming the structural model obtained in Pca2₁. The atomic positions of the PyzNH₂ ligand were refined as a rigid body, and H-atom positions were replaced by the ones obtained with the program WinGX (Farrugia, 1999). A positive-definite set of anisotropic displacement parameters could be determined for atom Cu1, whereas the remaining non-H-atom isotropic displacement parameters were constrained to the same value. U_iso(H) values were constrained to a value of 1.2 times the equivalent isotropic displacement parameter of the non-H atoms.

As a general rule, the coexistence of light and heavy atoms in a crystal structure is deemed to decrease the accuracy of the crystallographic parameters determined if small errors in the heavy-atom parameters are present (Giacovazzo et al., 2002). Additionally, the crystallographic parameters obtained from powder diffraction experiments are less accurate than those determined from single-crystal diffraction owing to a lower observation/parameter ratio. For this structure in particular, only a soft bond length restraint of 1.160 (1) Å had to be included in order to obtain a reasonable cyano bond length, although no restraints were necessary for PyzNH₂ Cu—N distances or angles. The possibility of disorder in the cyano C9 and N10 atoms was investigated at intermediate refinement stages as well as at the end of the Rietveld refinements, and in more than one set of atomic positions obtained from PSSP runs. This was achieved by switching C– and N-atom identities and refining the respective C and N occupancy factors for both positions. As a result, the cyano C9 and N10 atoms refine as ordered, as sometimes is the case in this type of compound.

In the Rietveld refinements, only the scale factors and background coefficients of all histograms were refined without being subjected to constraints. The following parameters were constrained to equivalent values for the three histograms: lattice parameters, 2θ zero error, transparency, sample displacement error, profile parameters, atomic positions, isotropic and Cu1 anisotropic displacement parameters, preferred orientation and absorption coefficients. The standard deviations of crystallographic parameters have been corrected following the procedure reported by Scott (1983). A plot of the observed and calculated powder diffraction intensities and their difference (at the bottom) is shown in Fig. 3.

Computing details

Data collection: Philips X'Pert Data Collector Software; cell refinement: GSAS (Larson & Von Dreele, 2000); data reduction: GSAS (Larson & Von Dreele, 2000); program(s) used to solve structure: EXPO2004 (Altomare et al., 2004) and PSSP (Stephens & Pagola, unpublished results); program(s) used to refine structure: GSAS (Larson & Von Dreele, 2000); molecular graphics: Mercury (Version 1.4.2; Macrae et al., 2006); software used to prepare material for publication: publCIF (Version 1.5.1_c_beta; Westrip, 2008).

Figures

Figure 1. The molecular structure of the title compound, showing the atom-labeling scheme and displacement ellipsoids at the ???% probability level. Figure 2. View of the distorted tetrahedral coordination around Cu^I, with atom labelling for one asymmetric unit (H-atom labels have been omitted for clarity). (a) [Cu6(CN)2(µ-Pyz)4] 26-membered rings in the three-dimensional honeycomb-like arrays of [CuCN(µ-Pyz)]. H atoms are not shown for clarity. Figure 2. The final Rietveld refinement. Observed intensity (points), calculated profile (solid line) and peak positions (| symbols). The difference plot (Iobserved - Icalculated) is shown at the bottom.

Poly[µ-2-aminopyrazine-κ²N¹:N⁴-µ-cyanido-copper(I)]

Crystal data

[Cu(CN)(C4H5N3)]	F(000) = 368.0
Mr = 184.68	Dx = 1.886 Mg m^−3
Orthorhombic, Pca21	Cu Kα radiation, λ = 1.5418 Å
Hall symbol: P 2c -2ac	T = 298 K
a = 11.536 (1) Å	Particle morphology: fine powder
b = 6.7868 (7) Å	yellow
c = 8.3073 (9) Å	cylinder, 27 × 27 mm
V = 650.4 (1) Å^3	Specimen preparation: Prepared at 298 K and 101 kPa, cooled at 0 K min^−1
Z = 4

Data collection

Philips Analytical X'Pert Pro MRD diffractometer	Data collection mode: reflection
Radiation source: x-ray	Scan method: step
Curved graphite monochromator	2θmin = 14.00°, 2θmax = 135.00°, 2θstep = 0.025°
Specimen mounting: packed powder in flat plate holder

Refinement

Least-squares matrix: full	1 restraint
Rp = 0.017	41 constraints
Rwp = 0.026	Primary atom site location: structure-invariant direct methods
Rexp = 0.015	Secondary atom site location: structure-invariant direct methods
RBragg = 0.083	Hydrogen site location: inferred from neighbouring sites
χ^2 = 2.958	H-atom parameters constrained
? data points	Weighting scheme based on measured s.u.'s
Excluded region(s): none	(Δ/σ)max = 0.04
Profile function: Pseudovoigt (Thompson et al., 1987), with asymmetry correction from Finger et al. (1994) and microstrain broadening as reported by Stephens (1999).	Preferred orientation correction: March–Dollase as implemented in GSAS along the (111) axis
149 parameters

Special details

Refinement. The z-coordinate of C9 does not have standard uncertainty. By the use of the bond length restrain for the cyano atoms in the polar space group Pca21(29) (polar direction [001]), the parameter has been fixed as unrefinable and no standard uncertainty is calculated.

The powder diffraction pattern was indexed without impurity peaks with the program DICVOL04 (Boultif & Louër, 2004) to the orthorhombic unit-cell parameters a = 11.5353 Å, b = 8.3093 Å, c = 6.7841 Å, α = β = γ = 90.0° and M₁₉ = 51, with unit-cell volume of 650.3 ų. Three additional monoclinic unit cells were found with similar a, b and c values (differently permuted), β = 90.194, 90.338 and 90.159°, and M₁₉ = 12.1, 14.2 and 14.2.

Le Bail fits (Le Bail et al. 1988) performed with the program GSAS (Larson & Von Dreele, 2000) in P222 confirmed the validity of the orthorhombic unit-cell parameters. Additional fits were carried out in P2 with monoclinic lattice parameters. Starting from the unit cells with the largest β angle, the refined monoclinic parameters were a = 11.531 (1) Å, b = 6.7832 (6) Å, c = 8.3018 (8) Å, α = 90°, β = 90.001 (2)° and γ = 90°, showing essentially the same agreement factors as the orthorhombic unit cell. The estimated density value of 1.9 Mg cm^-3 suggested Z = 4.

The choice of space group symmetry was not immediate from the powder pattern. The observation of the systematic absences first suggested P2₁2₁2 (18), and a plausible structural model was found with the direct methods program EXPO2004 (Altomare et al., 2004) using a data set collected in the 2θ range 14–100°, with a step size of 0.025° and data collection time of 20 s per step. In this model, the Cu^I coordination was distorted tetrahedral and all PyzNH₂ atoms as well as the cyano C and N atoms were found. However, a 1.48 Å N—N contact within the amine groups of the PyzNH₂ ligands was present, and a Rietveld refinement of this model was not satisfactory. Since it was likely that less obvious systematic absences had been overlooked, an exhaustive search of other possible space groups yielded only Pb2₁a (29) or Pbma (57). Pbma was discarded after an EXPO2004 run which did not give rise to an alternative model. However, it was possible to find all atoms of the asymmetric unit of (I) in a default run of EXPO2004 in Pb2₁a. The unit cell was transformed to the standard space group setting Pca2₁ (29). Once approximate Cu1 atomic positions had been determined using EXPO2004, the program PSSP (P. W. Stephens & S. Pagola, Powder Structure Solution Program, https://powder.physics.sunysb.edu/programPSSP/pssp.html), using direct-space methods and a simulated annealing algorithm, was applied to determine the location of the cyano and PyzNH₂ ligands. The crystal structure was found in almost all runs with an agreement factor S of 0.012, using 366 reflections and 50 000 cycles, and with initial temperature, final temperature and decrement factor values of 50, 0.001 and 0.8, respectively. The molecular geometry of the PyzNH₂ ligand was taken from the Cambridge Structural Database (Allen, 2002) entry AMPYRZ (provide full reference), including H atoms for structure solution.

The Le Bail fit obtained using a second data set collected with higher counting statistics in the 2θ region 80–135° had R_wp = 1.89% and χ² = 1.6. The Rietveld refinement was carried out using the program GSAS (Larson & Von Dreele, 2000), confirming the structural model obtained in Pca2₁. The atomic positions of the PyzNH₂ ligand were refined as a rigid body, and H-atom positions were replaced by the ones obtained with the program WinGX (Farrugia, 1999). A positive-definite set of anisotropic displacement parameters could be determined for atom Cu1, whereas the remaining non-H-atom isotropic displacement parameters were constrained to the same value. U_iso(H) values were constrained to a value of 1.2 times the equivalent isotropic displacement parameter of the non-H atoms.

As a general rule, the coexistence of light and heavy atoms in a crystal structure is deemed to decrease the accuracy of the crystallographic parameters determined if small errors in the heavy-atom parameters are present (Giacovazzo et al., 2002). Additionally, the crystallographic parameters obtained from powder diffraction experiments are less accurate than those determined from single-crystal diffraction owing to a lower observation/parameter ratio. For this structure in particular, only a soft bond length restraint of 1.160 (1) Å had to be included in order to obtain a reasonable cyano bond length, although no restraints were necessary for PyzNH₂ Cu—N distances or angles. The possibility of disorder in the cyano C9 and N10 atoms was investigated at intermediate refinement stages as well as at the end of the Rietveld refinements, and in more than one set of atomic positions obtained from PSSP runs. This was achieved by switching C– and N-atom identities and refining the respective C and N occupancy factors for both positions. As a result, the cyano C9 and N10 atoms refine as ordered, as sometimes is the case in this type of compound.

In the Rietveld refinements, only the scale factors and background coefficients of all histograms were refined without being subjected to constraints. The following parameters were constrained to equivalent values for the three histograms: lattice parameters, 2θ zero error, transparency, sample displacement error, profile parameters, atomic positions, isotropic and Cu1 anisotropic displacement parameters, preferred orientation and absorption coefficients. The standard deviations of crystallographic parameters have been corrected following the procedure reported by Scott (1983). A plot of the observed and calculated powder diffraction intensities and their difference (at the bottom) is shown in Fig. 3.

Giacovazzo, C., Monaco, H. L., Artioli, G., Viterbo, D., Ferraris, G., Gilli, G., Zanotti, G. & Catti, M. (2002). Fundamentals of Crystallography, 2nd ed. Oxford University Press.

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
Cu1	0.5088 (9)	0.2041 (6)	0.939 (6)	0.06738
N2	0.160 (2)	0.619 (3)	0.956 (7)	0.066 (7)*
C3	0.260 (1)	0.688 (2)	0.897 (7)	0.066 (7)*
C4	0.362 (2)	0.573 (3)	0.907 (7)	0.066 (7)*
N5	0.364 (3)	0.397 (3)	0.973 (7)	0.066 (7)*
C6	0.263 (3)	0.330 (2)	1.031 (7)	0.066 (7)*
C7	0.164 (3)	0.440 (3)	1.023 (7)	0.066 (7)*
N8	0.262 (1)	0.869 (3)	0.831 (7)	0.066 (7)*
H4	0.4309	0.6243	0.8657	0.079*
H6	0.2603	0.2057	1.0786	0.079*
H7	0.0961	0.3884	1.0655	0.079*
H8A	0.1981	0.9374	0.8269	0.079*
H8B	0.3242	0.9157	0.7921	0.079*
C9	0.498 (9)	0.01 (1)	1.10621	0.066 (7)*
N10	0.496 (5)	−0.075 (6)	1.227 (5)	0.066 (7)*

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Cu1	0.045 (6)	0.032 (5)	0.126 (9)	−0.007 (7)	0.015 (9)	−0.01 (1)

Geometric parameters (Å, º)

N8—C3	1.35 (4)	C4—C3	1.41 (3)
N5—C4	1.32 (4)	Cu1—C9	1.91 (8)
N5—C6	1.34 (5)	Cu1—N10^ii	1.97 (7)
N5—Cu1	2.14 (3)	C9—N10	1.16 (5)
C7—N2	1.34 (4)	N8—H8A	0.860
C7—C6	1.37 (4)	N8—H8B	0.860
N2—C3	1.34 (4)	C7—H7	0.930
N2—Cu1^i	2.12 (2)	C4—H4	0.930
C4—N5	1.32 (4)	C6—H6	0.930
C4—N5—C6	116 (3)	C9—Cu1—N10^ii	110 (2)
C4—N5—Cu1	121 (3)	N5—Cu1—N2^iii	106.7 (7)
C6—N5—Cu1	121 (1)	Cu1—C9—N10	167 (6)
N2—C7—C6	123 (3)	Cu1^iv—N10—C9	176 (5)
C7—N2—C3	116 (2)	N5—Cu1—C9	106 (3)
C7—N2—Cu1^i	124 (1)	C3—N8—H8A	120.0
C3—N2—Cu1^i	119.1 (9)	C3—N8—H8B	120.0
N5—C4—C3	123 (3)	H8A—N8—H8B	120.0
N8—C3—N2	118.9 (16)	N5—C4—H4	118.8
N8—C3—C4	121 (2)	C3—C4—H4	118.8
N2—C3—C4	120 (3)	N2—C7—H7	118.4
N5—C6—C7	122 (3)	C6—C7—H7	118.4
N5—Cu1—N10^ii	112 (2)	N5—C6—H6	119.4
N2^iii—Cu1—C9	113 (3)	C7—C6—H6	119.4
N2^iii—Cu1—N10^ii	109 (2)

Symmetry codes: (i) x−1/2, −y+1, z; (ii) −x+1, −y, z−1/2; (iii) x+1/2, −y+1, z; (iv) −x+1, −y, z+1/2.