A structural and theoretical study of inter­molecular inter­actions in nicotinohydrazide dihydro­chloride

In the title compound [systematic name: 3-(aza­niumyl­car­bam­o­yl)pyridinium dichloride], C₆H₉N₃O²⁺·2Cl^-, the ions are connected by N-H...Cl hydrogen bonds to form layers and C-H...Cl inter­actions expand the layers into a three-dimen­sional net. The energies of the N-H...Cl inter­actions range from typical for very weak inter­actions (0.17 kcal mol^-1) to those observed for relatively strong inter­actions (29.1 kcal mol^-1). C-H...Cl inter­actions can be classified as weak and mildly strong (energies ranging from 2.2 to 8.2 kcal mol^-1). Despite the short contacts existing between the parallel aromatic rings of the cations, [pi]-[pi] inter­actions do not occur.

S1. Comment

Intermolecular interactions play a crucial role in chemical, catalytic and biochemical processes, chemical and crystal engineering, as well as in supramolecular chemistry (Jeffrey & Saenger, 1991; Jeffrey, 1997; Epstein & Shubina, 2002). Over the course of a [the last?] century new types of intermolecular interactions, such as π–π, anion···π, cation···π and other interactions, were found and they were extensively studied (Payer et al., 2007; Schnabel et al., 2007), especially their spectral (Tonge et al., 2007; Fecko et al., 2003), structural (Suresh, 2007; Fisher et al., 2007) and thermodynamic (Harmon & Nikolla, 2003; Villar et al., 2003) features. However, the energetic behaviour of intermolecular interactions has not received as much attention as other properties. Energetic characteristics are important because they are responsible for the stability of created assemblies (both supramolecular complexes and in-reaction intermediates) and, consequently, they govern the possibilities of applications of specific intermolecular interactions (Gavezzotti, 2008; Oliveira et al., 2006).

The group of aromatic hydrazides exhibits a wide variety of biological activities including the effective inhibition of oxidases (Artico et al., 1992) and peroxidases (Ouellet et al., 2004), antifungal and antimicrobial activity (Deep et al., 2010) and anticonvulsant activity (Ragavendran et al., 2007). In addition, these compounds have diverse technical applications, e.g. as anticorrosive materials (Al-Hazam, 2010) or stabilizers of polymers (Ahmed Mohamed, 1997). The isomers of nicotinic acid hydrazide are also biologically active: the nicotinic acid hydrazide (Ia) is an effective inhibitor of cyclooxygenase activity of the prostaglandin H₂ synthase-2, and the isonicotinic acid hydrazide (IIa) has very potent antitubercular activity (Zhang & Young 1993). The mechanism of the biological activity of (IIa) is well known (Bardou et al., 1998), but the route of action of (Ia) is still unresolved; thus studies of the intermolecular bonding properties of (Ia) can be crucial for both determining the inhibition mechanism and the design of drugs selectively inhibiting enzymatic oxidation reactions. The structures of ortho (IIIa; Wu et al., 2005; Zareef et al., 2006)-, meta (Ia; Priebe et al., 2008; Portalone & Colapietro, 2008)- and para (IIa; Jensen, 1954; Bhat, et al. 1974)-isomers of nicotinic acid hydrazide were previously determined, but the supramolecular adducts of (Ia) and (IIIa) are unknown and a very limited number of supramolecular complexes of such compounds have been reported for (IIa) (Meng et al., 2008; BenHamada & Jouini, 2006; Kupfer-Tsoucaris & Tsoucaris, 1964; Gel'mbol'dt et al., 2002; Xie, 2007). Additionally, in the literature the structures of four metal coordination compounds of (IIa) (Hanson, et al., 1981; Tsintsadze et al., 1980; Zinner et al., 1979; Jiang et al., 2009) and two of (Ia) (Van Hecke et al., 2007; Tsintsadze et al., 1979) have been found. The nicotinic acid hydrazide dihydrochloride, (I), the protonated form of (Ia), has been studied because physiological pH typically is lower than 7 and thus such protonation occurs in biological systems, and almost all drugs or bioactive molecules undergo protonation before they enter the reaction chain. The chloride was chosen and used as a counterion since it is the most typical inorganic anion existing in living organisms, and thus knowledge about the intermolecular interactions between the mentioned species can be crucial in both studies of reaction mechanisms and design-targeted drug molecules based on aromatic hydrazides.

The asymmetric unit of the title compound, (I), contains the inorganic anion acting as counterion, balancing the charge of the 2-(hydraziniumcarbonyl)pyridinium cation (Fig. 1). The cation is doubly protonated, similar to isonicotinic acid hydrazide dihydrochloride (Kupfer-Tsoucaris & Tsoucaris, 1964), 4-(hydraziniumcarbonyl)pyridinium hexafluorosilicate (Gel'mbol'dt et al., 2002) and in contrast to monoprotonated dichloro-(isonicotinic acid hydrazide)-copper(II) hydrochloride (Hanson et al., 1981), (pyridin-4-ylcarbonyl)diazanium 4-(hydrazinocarbonyl)pyridinium bis(dihydrogen phosphate) (BenHamada & Jouini, 2006), catena-(bis(µ₂-isonicotinehydrazido)-(isonicotinehydrazido)- -triaqua-chloro-di-manganese trichloride isonicotinehydrazide solvate) (Tsintsadze et al., 1980) and 4-(hydrazinocarbonyl)pyridinium 3-carboxy-4-hydroxybenzenesulfonate monohydrate (Xie, 2007). The cation of (I) is slightly distorted from planarity (Table 1), with a maximum deviation from the weighted least-squares plane calculated for its all non-hydrogen atoms of 0.1421 (11) Å existing for the N4 atom. Such a maximum deviation in (Ia) exists for O1 and it is distinctly larger (0.538 Å, Priebe et al., 2008; 0.541 Å, Portalone & Colapietro 2008). Considerations of (IIa) and its protonated dichloride, (II) (Kupfer-Tsoucaris & Tsoucaris, 1964), show maximum deviations of about 0.4 Å for both compounds, and (IIIa) can be considered planar. Because the intramolecular N—H···N/O hydrogen bonds in (I) and (II) are structurally forbidden [they can only exist in planar (IIIa)], thus it can be supposed that twist of the hydraziniumcarbonyl moiety in (I) is caused mainly by the intermolecular interactions and crystal packing and less by the electronic structure of the cation. The organic cation shows typical bond lengths and angles values comparable with the other, above cited, isomers of nicotinic acid hydrazide, their salts and their complex compounds.

The cations and anions of (I) are held together by N—H···Cl hydrogen bonds (Fig. 2, Table 2) and they create the DDDDDC(2) unitary graph set (Bernstein et al., 1995). At the second-level graph, N₂, the hydrogen bonds can be expressed as C₂¹(5)C₂¹(9) C(2)D[R₄²(10)C₂¹(9) C(2)D] basic graph sets. In this way, a layer parallel to the crystallographic (100) plane is created. The neighbouring layers are interlinked by short intermolecular C—H···Cl interactions (Table 2), which can be classified as weak hydrogen bonds (Desiraju & Steiner, 1999). In such a manner, a three- dimensional supramolecular structure is created. The hydrogen-bond scheme in (Ia) is distinctly different due to both the absence of a hydrogen-bond ionic acceptor (Cl^-) and the presence of two potential hydrogen-bond acceptors (—NH₂ and ≐N≐) instead of two additional hydrogen-bond donors (—NH₃⁺ and ≐NH⁺≐). The first-level graph set of (Ia) is expressed by the C(5) C(5) C(6) descriptor. A hydrogen-bonding scheme similar to that in (I) exists in (II): N₁ = DDDDD. The main dissimilarity originates from the different position of the protonated nitrogen atom in the aromatic ring and the creation of two-centre hydrogen bonds in (II) in contrast to two-centre and three-centre hydrogen bonds in (I). There are short contacts between the parallel aromatic rings of cations of (I) [the symmetry-generated rings are obtained by -x + 1, -y + 1, -z and -x + 1, -y + 1, -z + 1 symmetry transformations; the distances between ring centroids are 3.805 (2) and 4.054 (2) Å, respectively], but the possibility of π–π interactions was rejected on geometric grounds (neighbouring aromatic bonds do not overlap sufficiently).

The molecular electronic properties have been calculated at a single point for both the diffraction-derived coordinates and the optimized structure, and these are comparable within three standard deviations, although the geometrically optimized molecules show typical elongation of the N—H and the C—H bonds (from 0.07 to 0.16 Å). The total binding energies of intermolecular interactions were calculated for molecular sets containing from one to eight cation–anion pairs with usage of total self-consistent field energy. The cation[s] and anions within each molecular set were arranged as in hydrogen-bonded layers, along interlayer C—H···Cl intermolecular interactions and along piles created by cations. The structural parameters (with hydrogen-atom positions geometrically optimized) were a starting model in each calculation. To calculate the electrostatic interaction energy between anions and cations, additional computations were performed for subsets containing an odd number of cations and anions (such calculated interaction energy is composed from a purely electrostatic term associated with formal +1 and -1 charges and from total intermolecular interaction energy between uncharged molecules, thus the subtraction of the energy value calculated for uncharged subsets from analogous charged subsets gives the purely electrostatic interaction energy between ions, E_electr, Table 2). Basisset superposition error (BSSE) corrections were carried out using the counterpoise (CP) method (Boys & Bernardi,1970). The B3LYP functional (Becke, 1993; Lee et al., 1988) and Hartree–Fock calculation followed by a Møller–Plesset correlation energy correction (Møller & Plesset, 1934) truncated at second-order (Head-Gordon et al., 1988) in the triple-ζ 6–311++G(3df,2p) basis set was used, as implemented in GAUSSIAN03 (Frisch et al., 2004). In all cases, the differences in electronic properties and energies originating from the different number of cation–anion pairs used in the calculation and the differences between the above described methods are given in parentheses as standard deviations of the mean values. Where a deviation is not given, the values were the same within their range of reported precision. As expected, the cations and anions of (I) are attracted by the strong electrostatic interactions with binding energy ranging from 65 to 113 kcal mol-1, and the interaction energy strongly depends on the geometric arrangement of ions (Table 2). Noteworthy is the fact that electrostatic attraction between the hydrogen-bonded layers parallel to the crystallographic (100) plane is distinctly weaker than those existing within the layers. The electrostatic repulsion between the organic cations related by the twofold screw axis at 1/2, b, 1/4 (along C3—H3···O1^v interactions; symmetry code as in Table 2) is about -159 kcal mol-1. The values of N—H···Cl hydrogen-bond energies lie in ranges observed for similar systems such as 2,4-dinitrophenylhydrazine hydrochloride hydrate (Kruszynski, 2008), o-tolidinium dichloride dihydrate (Kruszynski, 2009a), 5-(2-halogenoethoxy)-2,3-dihydro-1,4-benzodioxines (Kruszynski, 2009b), 2,3-dihydro-1,3-benzothiazol-2-iminium monohydrogen sulfate, 2-iminio-2,3-dihydro-1,3-benzothiazole-6-sulfonate (Kruszynski & Trzesowska-Kruszynska, 2009), 2-amino-5-chloro-1,3-benzoxazol-3-ium inorganic salts (Kruszynski & Trzesowska-Kruszynska, 2010), 2,3-dihydro-1,3-benzothiazol-2-iminium hydrogen oxydiacetate (Trzesowska-Kruszynska & Kruszynski, 2009), but these energies are larger than typical ones observed for N—H···Cl bonds of non-ionic species. Because the energies of the hydrogen bonds containing donor atoms possessing a formal +1 charge (N1 and N3) are not distinctly stronger than those geometrically similar ones having a formal neutral charge (e.g. N2—H2N···Cl1ⁱ and N3—H3N···Cl1ⁱⁱ hydrogen bond; symmetry codes as in Table 2) thus it can be postulated that the energy increase is caused mainly by larger negative charge of ionic chlorine [for both anions -0.94 (6) A.U. derived from analysis of the atomic charges based on the Breneman radii (Breneman & Wiberg, 1990)] in comparison to covalently bonded one, and slightly by larger positive charge situated on hydrogen atoms of protonated amine group [from 0.50 (4) to 0.56 (3) A.U. calculated as above] in comparison to neutrally charged groups. Thus the hydrogen-bond acceptor–donor electron-density sharing in (I) is more privileged than in compounds containing non-ionic species (Kruszynski, 2009a) and, for creation of stronger hydrogen bonds, the additional electron density on a hydrogen-bond acceptor is more important than electron-density deficiency on a hydrogen-bond donor. The strength of the hydrogen bonds does not correlate directly with the enlargement of the D···A distance or D—H···A angle, but a general relationship of increasing hydrogen-bond energy with both decreasing D···A distance and increasing D—H···A angle is observed (Fig. 3). The flattering of the surface of the function E(D···A, D—H···A) at lower angles confirms the postulate that the orientation of the D—H bond relative to the line linking donor and acceptor is more important in the creation of stronger hydrogen bonds than the D···A distance (Kruszynski, 2008), e.g. at D—H···A angle = 100° all hydrogen-bond energies vanish regardless of D···A distances, and at D···A distance = 3.6 Å the increase of D—H···A angle leads to increase of hydrogen-bond energy. For both types of donors (N—H and C—O) the hydrogen-bond geometric parameters and energy undergo the same dependence; thus the relatively short, but exhibiting a narrow angle, N3—H3O···Cl2^iv' hydrogen bond (symmetry code as in Table 2) is distinctly weaker than the longer but more linear C—H···Cl hydrogen bond. The C—H···Cl hydrogen-bond energy is relatively large and for the C5—H5···Cl2^vi hydrogen bond (symmetry code as in Table 2) it is distinctly greater than the most typical values (about 4 kcal mol-1; Desiraju & Steiner, 1999) as a result of a large negative charge located on the Cl^- atom and in consequence the transfer of a considerable amount of electron density to the C—H donor of the hydrogen bond. It must be noted that the geometrically allowed C3—H3···O1^v hydrogen bond (symmetry code as in Table 2) is actually a non-bonding interaction. The non-bonding character of this interaction in comparison to the less linear and longer C—H···Cl bonding interactions again is proof that charge located on hydrogen-bond acceptor atoms has crucial importance for the creation of strong hydrogen bonds. The calculations of the intermolecular interaction energy of parallel aromatic rings related by -x + 1, -y + 1, -z + 1 and -x + 1, -y + 1, -z symmetry transformation show that these interactions are antibonding in character even if the dispersion energy was introduced to calculations. That confirms the absence of π–π stacking interactions in (I).

The non-covalent nature of the intermolecular interactions in (I) was analysed by the natural bond orbital (NBO) method (Foster & Weinhold, 1980; Reed & Weinhold, 1985; Reed et al., 1988). In this method the strength of the donor–acceptor charge transfer delocalization is characterized by the second-order stabilization energy, E_del. For hydrogen bonds the principal charge transfer interactions E_del(1) occur between the chloride lone pairs and antibonding orbitals of the N—H and C—H bonds. In the lateral charge transfer interactions the chloride lone pairs donate their electron density to the one-centre Rydberg orbitals of the hydrogen atom. Noteworthy is the fact that interactions between lone pair orbitals and Rydberg orbitals in C—H···Cl bonds contribute more to total hydrogen-bond energy than in N—H···Cl bonds. Thus changing the relatively strong N—H hydrogen-bond donor to the weak one (C—H) leads to redistribution of donor electron density and diminishes the energy of constituent interactions containing antibonding orbitals. Because the electron density of the hydrogen-bond acceptor is transferred from donor antibonding orbitals to donor Rydberg orbitals, the total energy of geometrically similar C—H···Cl and N—H···Cl hydrogen bonds is comparable.

Related literature

For related literature, see: Ahmed Mohamed (1997); Al-Hazam (2010); Artico et al. (1992); Bardou et al. (1998); Becke (1993); BenHamada & Jouini (2006); Bernstein et al. (1995); Bhat et al. (1974); Boys & Bernardi (1970); Breneman & Wiberg (1990); Deep et al. (2010); Desiraju & Steiner (1999); Epstein & Shubina (2002); Fecko et al. (2003); Fisher et al. (2007); Foster & Weinhold (1980); Frisch (2004); Gavezzotti (2008); Gel'mbol'dt, Davydov, Koroeva & Ganin (2002); Hanson et al. (1981); Harmon & Nikolla (2003); Head-Gordon, Pople & Frisch (1988); Jeffrey (1997); Jeffrey & Saenger (1991); Jensen (1954); Jiang et al. (2009); Kruszynski (2008, 2009a, 2009b); Kruszynski & Trzesowska-Kruszynska (2009, 2010); Lee et al. (1988); Møller & Plesset (1934); Meng et al. (2008); Oliveira et al. (2006); Ouellet et al. (2004); Payer et al. (2007); Portalone & Colapietro (2008); Priebe et al. (2008); Ragavendran et al. (2007); Reed & Weinhold (1985); Reed et al. (1988); Schnabel et al. (2007); Suresh (2007); Tonge et al. (2007); Trzesowska-Kruszynska & Kruszynski (2009); Tsintsadze et al. (1979, 1980); Van Hecke, Nockemann, Binnemans & Van Meervelt (2007); Villar et al. (2003); Wu et al. (2005); Xie (2007); Zareef et al. (2006); Zhang & Young (1993); Zinner et al. (1979).

S2. Experimental

Compound (I) (0.001 mol, 0.2101 g) was dissolved in a hot aqueous standard solution of HCl (20 ml, 0.2 mol dm^-3). The solution was stored in a gaseous HCl atmosphere and after 6 weeks crystals suitable for measurement had grown in a 94% yield.

S3. Refinement

The H atoms were found from difference Fourier syntheses after eight cycles of anisotropic refinement. All H-atom positions were refined freely and the U values were set at U_iso(H) = 1.2U_eq(C,N) in all (including the final) refinement cycles. The isotropic displacement parameters of the H atoms were then refined to check the correctness of their positions. After eight cycles, the refinement reached stable convergence with isotropic displacement parameters in the range 0.0147–0.0313. The values of the isotropic displacement parameters of the H atoms have reasonable values, proving the correctness of the H-atom positions.

Computing details

Data collection: CrysAlis CCD (UNIL IC & Kuma, 2000); cell refinement: CrysAlis RED (UNIL IC & Kuma, 2000); data reduction: CrysAlis RED (UNIL IC & Kuma, 2000); program(s) used to solve structure: SHELXS97 (Sheldrick, 2008); program(s) used to refine structure: SHELXL97 (Sheldrick, 2008); molecular graphics: XP in SHELXTL/PC (Sheldrick, 2008) and ORTEP-3 (Farrugia, 1997); software used to prepare material for publication: SHELXL97 (Sheldrick, 2008) and PLATON (Spek, 2009).

Figures

	Fig. 1. A view of the asymmetric unit of (I), showing the atom-labelling scheme. Displacement ellipsoids are drawn at the 50% probability level and H atoms are shown as spheres of arbitrary radii. Hydrogen bonds are indicated by dashed lines.
	Fig. 2. Part of the molecular packing of (I), showing the N—H···Cl hydrogen bonds (lines) creating the layer parallel to the crystallographic (100) plane. The symmetry codes are as given in Table 2.
	Fig. 3. A surface plot showing the relationship between hydrogen-bond energy (E, Table 2) and D···A distances/D—H···A angles. The values of E are indicated by dots.

3-(azaniumylcarbamoyl)pyridinium dichloride

Crystal data

C6H9N3O^2+·2Cl^−	F(000) = 432
Mr = 210.06	Dx = 1.604 Mg m^−3 Dm = 1.60 Mg m^−3 Dm measured by Berman density torsion balance
Monoclinic, P21/c	Mo Kα radiation, λ = 0.71073 Å
Hall symbol: -P 2ybc	Cell parameters from 893 reflections
a = 9.4516 (7) Å	θ = 2–50°
b = 12.5202 (12) Å	µ = 0.70 mm^−1
c = 7.5020 (6) Å	T = 90 K
β = 101.559 (4)°	Prism, orange
V = 869.75 (13) Å^3	0.21 × 0.20 × 0.18 mm
Z = 4

Data collection

Kuma KM-4-CCD diffractometer	1657 independent reflections
Radiation source: fine-focus sealed tube	1598 reflections with I > 2σ(I)
Graphite monochromator	Rint = 0.034
Detector resolution: 1048576 pixels mm^-1	θmax = 25.7°, θmin = 2.2°
ω scans	h = −11→11
Absorption correction: numerical (X-RED; Stoe & Cie, 1999)	k = −15→15
Tmin = 0.860, Tmax = 0.889	l = −9→9
12309 measured reflections

Refinement

Refinement on F^2	Primary atom site location: structure-invariant direct methods
Least-squares matrix: full	Secondary atom site location: structure-invariant direct methods
R[F^2 > 2σ(F^2)] = 0.024	Hydrogen site location: difference Fourier map
wR(F^2) = 0.065	Only H-atom coordinates refined
S = 1.12	w = 1/[σ^2(Fo^2) + (0.0289P)^2 + 0.5818P] where P = (Fo^2 + 2Fc^2)/3
1657 reflections	(Δ/σ)max = 0.001
136 parameters	Δρmax = 0.31 e Å^−3
0 restraints	Δρmin = −0.21 e Å^−3

Special details

Geometry. All e.s.d.'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.'s are taken into account individually in the estimation of e.s.d.'s in distances, angles and torsion angles; correlations between e.s.d.'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.'s is used for estimating e.s.d.'s involving l.s. planes.

Refinement. Refinement of F² against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F², conventional R-factors R are based on F, with F set to zero for negative F². The threshold expression of F² > σ(F²) is used only for calculating R-factors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F² are statistically about twice as large as those based on F, and R- factors based on ALL data will be even larger.

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

	x	y	z	Uiso*/Ueq
C1	0.65595 (16)	0.41947 (12)	0.2001 (2)	0.0121 (3)
C2	0.70810 (17)	0.50966 (13)	0.2973 (2)	0.0143 (3)
H2	0.806 (2)	0.5282 (15)	0.336 (2)	0.017*
N1	0.61479 (15)	0.57781 (11)	0.35014 (18)	0.0154 (3)
H1N	0.658 (2)	0.6373 (16)	0.425 (3)	0.019*
C3	0.47031 (17)	0.56288 (13)	0.3122 (2)	0.0153 (3)
H3	0.417 (2)	0.6164 (16)	0.353 (3)	0.018*
C4	0.41449 (16)	0.47281 (13)	0.2192 (2)	0.0148 (3)
H4	0.316 (2)	0.4622 (15)	0.196 (3)	0.018*
C5	0.50742 (17)	0.40067 (13)	0.1627 (2)	0.0145 (3)
H5	0.470 (2)	0.3406 (16)	0.092 (3)	0.017*
C6	0.75009 (16)	0.34300 (12)	0.1217 (2)	0.0124 (3)
O1	0.69854 (12)	0.27143 (9)	0.01911 (15)	0.0159 (2)
N2	0.89525 (14)	0.35789 (11)	0.17439 (18)	0.0132 (3)
H2N	0.935 (2)	0.4189 (17)	0.217 (2)	0.016*
N3	0.98193 (14)	0.29657 (11)	0.07688 (19)	0.0130 (3)
H3N	0.981 (2)	0.3265 (15)	−0.033 (3)	0.016*
H3P	0.948 (2)	0.2291 (17)	0.068 (3)	0.016*
H3O	1.073 (2)	0.2910 (15)	0.148 (3)	0.016*
Cl1	0.92096 (4)	0.06499 (3)	0.21749 (5)	0.01403 (12)
Cl2	0.75891 (4)	0.74545 (3)	0.61362 (5)	0.01442 (12)

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
C1	0.0133 (7)	0.0116 (7)	0.0120 (7)	0.0017 (6)	0.0041 (6)	0.0030 (6)
C2	0.0132 (8)	0.0155 (8)	0.0153 (7)	−0.0009 (6)	0.0054 (6)	0.0010 (6)
N1	0.0180 (7)	0.0132 (7)	0.0168 (7)	−0.0029 (5)	0.0076 (5)	−0.0017 (5)
C3	0.0172 (8)	0.0147 (8)	0.0160 (8)	0.0042 (6)	0.0086 (6)	0.0035 (6)
C4	0.0096 (7)	0.0190 (8)	0.0164 (8)	0.0015 (6)	0.0040 (6)	0.0023 (6)
C5	0.0133 (7)	0.0147 (8)	0.0155 (7)	−0.0010 (6)	0.0027 (6)	−0.0001 (6)
C6	0.0124 (7)	0.0117 (7)	0.0136 (7)	0.0001 (6)	0.0040 (6)	0.0034 (6)
O1	0.0135 (5)	0.0137 (5)	0.0204 (6)	−0.0003 (4)	0.0031 (4)	−0.0041 (5)
N2	0.0113 (6)	0.0119 (6)	0.0177 (7)	0.0005 (5)	0.0058 (5)	−0.0028 (5)
N3	0.0116 (6)	0.0135 (7)	0.0150 (7)	0.0017 (5)	0.0051 (5)	−0.0004 (5)
Cl1	0.01106 (19)	0.0127 (2)	0.0187 (2)	0.00066 (13)	0.00377 (14)	−0.00100 (13)
Cl2	0.01210 (19)	0.0165 (2)	0.0149 (2)	0.00023 (13)	0.00328 (14)	−0.00067 (13)

Geometric parameters (Å, º)

C1—C2	1.380 (2)	C4—H4	0.92 (2)
C1—C5	1.395 (2)	C5—H5	0.95 (2)
C1—C6	1.505 (2)	C6—O1	1.2174 (19)
C2—N1	1.343 (2)	C6—N2	1.362 (2)
C2—H2	0.94 (2)	N2—N3	1.4269 (18)
N1—C3	1.351 (2)	N2—H2N	0.88 (2)
N1—H1N	0.97 (2)	N3—H3N	0.90 (2)
C3—C4	1.375 (2)	N3—H3P	0.90 (2)
C3—H3	0.93 (2)	N3—H3O	0.92 (2)
C4—C5	1.384 (2)
C2—C1—C5	118.86 (14)	C4—C5—C1	120.19 (15)
C2—C1—C6	123.39 (13)	C4—C5—H5	120.0 (11)
C5—C1—C6	117.64 (14)	C1—C5—H5	119.7 (11)
N1—C2—C1	119.31 (14)	O1—C6—N2	122.44 (14)
N1—C2—H2	114.6 (12)	O1—C6—C1	121.49 (13)
C1—C2—H2	126.0 (12)	N2—C6—C1	116.06 (13)
C2—N1—C3	123.09 (14)	C6—N2—N3	115.20 (13)
C2—N1—H1N	115.7 (11)	C6—N2—H2N	123.4 (12)
C3—N1—H1N	121.1 (11)	N3—N2—H2N	114.1 (12)
N1—C3—C4	119.30 (14)	N2—N3—H3N	110.0 (12)
N1—C3—H3	115.3 (12)	N2—N3—H3P	108.3 (12)
C4—C3—H3	125.4 (12)	H3N—N3—H3P	112.4 (17)
C3—C4—C5	119.23 (14)	N2—N3—H3O	108.2 (12)
C3—C4—H4	119.0 (12)	H3N—N3—H3O	113.3 (17)
C5—C4—H4	121.8 (12)	H3P—N3—H3O	104.3 (17)
C1—C6—N2—N3	−170.21 (12)

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
N1—H1N···Cl2	0.97 (2)	2.05 (2)	3.0149 (14)	171.0 (16)
N2—H2N···Cl1^i	0.88 (2)	2.28 (2)	3.1359 (14)	165.1 (17)
N3—H3N···Cl1^ii	0.90 (2)	2.29 (2)	3.1596 (15)	161.3 (16)
N3—H3P···Cl1	0.90 (2)	2.38 (2)	3.1772 (15)	147.6 (16)
N3—H3O···Cl2^iii	0.92 (2)	2.19 (2)	3.0641 (14)	158.4 (17)
N3—H3O···Cl2^iv	0.92 (2)	2.817 (19)	3.1393 (13)	101.9 (13)
C2—H2···Cl1^i	0.94 (2)	2.73 (2)	3.5971 (16)	153.7 (15)
C3—H3···O1^v	0.93 (2)	2.51 (2)	3.4315 (19)	172.7 (16)
C4—H4···Cl1^v	0.92 (2)	2.77 (2)	3.4948 (16)	136.2 (15)
C5—H5···Cl2^vi	0.95 (2)	2.671 (19)	3.5098 (16)	147.8 (15)

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