Hydrogen-bonding and carbon­yl–carbonyl inter­actions in violuric acid methanol solvate

Violuric acid [systematic name: pyrimidine-2,4,5,6(1H,3H)-tetrone 5-oxime] crystallized from a methanol solution stored at approximately 278 K as a monosolvate, C₄H₃N₃O₄·CH₃OH, in the form of very small and fragile needles. Synchrotron radiation was needed to collect an adequate data set. Analysis of the crystal structure reveals that the isonitroso group of violuric acid is disordered over two positions with refined occupancies of approximately 3:1 for the major and minor disorder components. This fact has some important consequences for the hydrogen-bonding motifs found in the crystal packing, which are different for each component, although the overall packing pattern does not change. The crystal packing consists of closely stacked hydrogen-bonded sheets. Between the sheets are found carbonyl–carbonyl dipolar inter­actions, which are the principal inter­molecular forces holding the sheets together.

Comment

We have been inter­ested in s-block metal complexes of cyanuric acid, barbituric acid and other related ligands for some time. Violuric acid is a derivative of barbituric acid, having a C=N—OH substituent at the 5-position on the barbiturate ring. This substituent allows extra coordination and hydrogen-bonding possibilities compared with barbituric acid itself. Many transition metal complexes of violuric acid are known (Abraham et al., 1980; Hamelin, 1972; Tamaki & Okabe, 1996; Faus et al., 1996), along with some s-block metal complexes (Gillier, 1965; Hamelin, 1976). We have sought to plug the gaps in the literature, and also to look again at those reported structures (some of which are many years old) in order to evaluate trends in the coordination and hydrogen bonding of violuric acid with s-block metals.

For a reliable comparison of the mol­ecular structure of coordinated versus free violuric acid, we sought to determine the crystal structure of violuric acid in the absence of any other mol­ecule. Although the crystal structure of violuric acid monohydrate has been known for many years (Craven & Mascarenhas, 1964; Craven & Takei, 1964), we recently redetermined it in order to resolve several issues in the original reports (Nichol & Clegg, 2005). However, we still wished to obtain the structure of unsolvated violuric acid. We thought that crystallization from a methanol solution might achieve our goal. As this report shows, this was not the case. What the result does allow us to do, however, is observe the effect of changing the solvent on the resulting hydrogen-bonding motifs found within the structure.

Unlike the straightforward crystallization from aqueous solutions, violuric acid does not readily crystallize from methanol. Storage of a methanol solution at 278 K for around one month resulted in only a few very delicate needle crystals. These were taken to Station 9.8, SRS, Daresbury Laboratory, in an ice-box and data were collected via the EPSRC-funded UK National Crystallography Service.

The mol­ecular structure of violuric acid methanol solvate, (I), is shown in Fig. 1. As is clear from the figure, the isonitroso group is disordered over two positions, with relative occupancies of approximately 3:1 for the major and minor components. This disorder is not uncommon and is something that we have observed in violurate–metal complexes, although it is not observed at all in violuric acid monohydrate. Although this disorder has little bearing on the mol­ecular structure, it is of great importance to the crystal packing. Other than this disorder, the mol­ecular dimensions (Table 1) are largely unexceptional and the rest of the violuric acid mol­ecule appears to be well ordered; the exocyclic angles at atom C4 deviate somewhat from the ideal trigonal value of 120° because of steric inter­action between the isonitroso group and adjacent ring substituents.

The most elegant aspect of the analysis is to be found not in the mol­ecular structure but in the crystal packing. In essence, this is rather simple and consists of three basic hydrogen-bonding (Table 2) graph-set motifs (Bernstein et al., 1995). Firstly, there is the very common R₂²(8) motif – found in almost all barbiturate crystal structures – involving two N—H⋯O hydrogen bonds; this is located about a crystallographic inversion centre and connects the violuric acid mol­ecules into tapes, which run parallel to the ac diagonal. Secondly – and ignoring the minor disorder component for the moment – the methanol mol­ecule is connected to the violuric acid mol­ecule by a more unusual but not uncommon R₁²(6) motif, involving the methanol O-bound H atom acting as a bifurcated donor to one carbonyl O atom and the O atom of the isonitroso group. Finally, an R₄⁴(8) motif links the aforementioned methanol and isonitroso atoms together with their symmetry equivalents about a crystallographic inversion centre. It is this final link that joins the tapes together to form a two-dimensional network sheet structure, as shown in Fig. 2.

Given that the minor disorder component is present in around one-quarter of the crystal structures, we cannot simply ignore how it affects the crystal packing. Whilst the R₂²(8) tapes are unaffected by this disorder, the way they are linked together is very much at the heart of the disorder analysis. Fig. 3 gives two views of the methanol–violuric acid inter­actions, with the major and minor disorder components shown separately. At the top is the major component, with the two hydrogen-bonding motifs already discussed clearly visible. At the bottom is the hydrogen bonding involving the minor component of the disorder only. Although the only atoms displaced are those of the isonitroso substituent (the methanol mol­ecule is not significantly disordered, despite the large ellipsoids which suggest that this might be a possibility), the difference in the hydrogen-bonding motifs is quite striking. The methanol O-bound H atom can now no longer act as a bifurcated donor; were it to form an O—H⋯N inter­action, the H⋯A distance would be an unacceptably long 2.86 Å. We therefore have a single large R₄⁴(16) connection between the violuric acid tapes, which themselves are unaffected by the disorder. The effect on the overall form of the crystal packing is only slight; Fig. 4 shows the crystal packing involving the minor component only and the patterns formed are still the same.

Whilst there is extensive hydrogen bonding within each sheet, there are no hydrogen-bonding inter­actions linking different sheets together. The distance between adjacent sheets is around 3 Å, which is very close and is less than the sum of the van der Waals radii of the respective atoms. As a result, what we see are many dipolar carbonyl–carbonyl inter­actions (Allen et al., 1998) between the tightly packed sheets (Fig. 5). Multipolar inter­actions have often been overlooked when considering those inter­actions that play a significant role in crystal packing. However, with the development of information-rich structural databases and powerful search algorithms, the subject has received increasing attention in recent years, in both small-mol­ecule and macromolecular crystallography, so much so that a review has been published recently (Paulini et al., 2005). The present structure offers a further example of one in which dipolar inter­actions are vital to the integrity of the crystal packing. Despite this, the difficulty with which the crystals grew and their relative instability once formed suggests that the attractive strength of these purely electrostatic inter­actions cannot be relied upon as the basis of controllable crystal growth in the same manner as hydrogen bonds are now widely considered.

Figure 1 The asymmetric unit of (I), with 50% probability displacement ellipsoids. Open bonds show the minor disorder component.

Figure 2 The hydrogen-bonded sheet structure in (I), viewed along the ab diagonal; the minor disorder component has been ignored. Dashed lines represent hydrogen bonds and the continuation of hydrogen bonding.

Figure 3 A representation of how the disorder affects the hydrogen-bonding motifs. (a) The hydrogen bonding involving the major component and (b) the hydrogen bonding involving the minor component.

Figure 4 The hydrogen-bonded sheet structure in (I), viewed along the ab diagonal; the major disorder component has been ignored. Dashed lines are as in Fig. 2.

Figure 5 Some highlighted carbonyl–carbonyl inter­actions within the crystal packing of (I).

Experimental

Violuric acid (0.17 g, 1 mmol) was dissolved in methanol (10 ml) with gentle heating. Storage for around one month at 278 K resulted in the growth of very small and fragile feather-like crystal agglomerations.

Crystal data

• C₄H₃N₃O₄·CH₄O

• M_r = 189.14

• Triclinic,

• a = 4.9650 (9) Å

• b = 8.6208 (17) Å

• c = 10.2613 (17) Å

• α = 100.474 (2)°

• β = 102.939 (2)°

• γ = 99.823 (2)°

• V = 410.56 (13) ų

• Z = 2

• D_x = 1.530 Mg m⁻³

• Synchrotron radiation

• λ = 0.6751 Å

• Cell parameters from 1809 reflections

• θ = 4.0–28.8°

• μ = 0.14 mm⁻¹

• T = 120 (2) K

• Needle, colourless

• 0.17 × 0.06 × 0.04 mm

Data collection

• Bruker APEX-2 CCD diffractometer

• ω scans

• Absorption correction: multi-scan(SADABS; Sheldrick, 2004)T_min = 0.950, T_max = 0.995

• 3107 measured reflections

• 1462 independent reflections

• 1290 reflections with I > 2σ(I)

• R_int = 0.013

• θ_max = 24.0°

• h = −5 → 5

• k = −10 → 10

• l = −12 → 12

Refinement

• Refinement on F²

• R[F² > 2σ(F²)] = 0.041

• wR(F²) = 0.113

• S = 1.08

• 1462 reflections

• 152 parameters

• H atoms treated by a mixture of independent and constrained refinement

• w = 1/[σ²(F_o²) + (0.0612P)² + 0.1629P] where P = (F_o² + 2F_c²)/3

• (Δ/σ)_max < 0.001

• Δρ_max = 0.29 e Å⁻³

• Δρ_min = −0.28 e Å⁻³

Table 1
Selected geometric parameters (Å, °)

O1—C1	1.218 (2)
O2—C2	1.2202 (19)
O3—C3	1.2218 (19)
O4—N3	1.361 (3)
N3—C4	1.293 (4)
O4′—N3′	1.299 (13)
N3′—C4	1.349 (16)
N1—C1	1.376 (2)
N1—C2	1.374 (2)
N2—C2	1.372 (2)
N2—C3	1.369 (2)
C1—C4	1.490 (2)
C3—C4	1.490 (2)
O5—C5	1.395 (3)

O4—N3—C4	114.6 (2)
O4′—N3′—C4	111.7 (8)
C1—N1—C2	126.63 (15)
C2—N2—C3	126.80 (14)
N1—C1—C4	115.10 (14)
N1—C2—N2	116.56 (14)
N2—C3—C4	115.32 (14)
N3—C4—C1	129.00 (17)
N3—C4—C3	111.45 (17)
N3′—C4—C1	109.4 (5)
N3′—C4—C3	131.1 (5)
C1—C4—C3	119.53 (14)

Table 2
Hydrogen-bond geometry (Å, °)

D—H⋯A	D—H	H⋯A	D⋯A	D—H⋯A
O4—H4⋯O5	0.98 (4)	1.59 (4)	2.567 (2)	174 (3)
O4′—H4′⋯O5	1.01 (12)	1.64 (12)	2.574 (5)	152 (9)
O5—H5⋯O1^i	0.84 (3)	1.98 (3)	2.7361 (18)	150 (2)
O5—H5⋯O4^i	0.84 (3)	2.21 (3)	2.761 (2)	124 (2)
N1—H1N⋯O2^ii	0.86 (2)	1.97 (2)	2.8322 (19)	178 (2)
N2—H2N⋯O3^iii	0.88 (2)	1.98 (2)	2.8519 (18)	172 (2)
Symmetry codes: (i) -x, -y+1, -z+1; (ii) -x+1, -y, -z; (iii) -x+2, -y, -z+1.

The NOH group was refined as disordered, with a refined occupancy of 0.749 (4) for the major component. The anisotropic displacement parameters of atoms N3 and N3′ were constrained to be equal. All H atoms were located in a difference electron-density map and, with the exception of the methyl H atoms, were freely refined. Methyl H atoms were refined using a riding model, with U_iso(H) values of 1.5U_eq(C) and a C—H distance of 0.98 Å. There is a short H⋯H contact between this methyl group and the H atom of the minor component of the disordered isonitroso group; this indicates that the methyl group is also disordered by rotation about the C—O bond, but the minor component was not included in the refinement.

Data collection: APEX2 (Bruker, 2004); cell refinement: APEX2; data reduction: SAINT (Bruker, 2001); program(s) used to solve structure: SIR2002 (Burla et al., 2003); program(s) used to refine structure: SHELXTL (Sheldrick, 2001); molecular graphics: SHELXTL and MERCURY (Bruno et al., 2002); software used to prepare material for publication: SHELXTL and local programs.

Supporting information

S1. Comment

We have been interested in s-block metal complexes of cyanuric acid, barbituric acid and other related ligands for some time. Violuric acid is a derivative of barbituric acid, having a C═N—OH substituent at the 5-position on the barbiturate ring. This substituent allows extra coordination and hydrogen bonding possibilities compared with barbituric acid itself. Many transition metal complexes of violuric acid are known (Abraham et al., 1980; Hamelin, 1972; Tamaki & Okabe, 1996; Faus et al., 1996), along with some s-block metal complexes (Gillier, 1965; Hamelin, 1976). We have sought to plug the gaps in the literature and also to look again at those reported structures (some of which are many years old) in order to evaluate trends in the coordination and hydrogen bonding of violuric acid with s-block metals.

For a reliable comparison of the molecular structure of coordinated versus free violuric acid, we sought to determine the crystal structure of violuric acid in the absence of any other molecule. Although the crystal structure of violuric acid monohydrate has been known for many years (Craven & Mascarenhas, 1964; Craven & Takei, 1964), we recently redetermined it in order to resolve several issues in the original reports (Nichol & Clegg, 2005). However, we still wished to obtain the structure of unsolvated violuric acid. We thought that crystallization from a methanol solution might achieve our goal. As this report shows, this was not the case. What the result does allow us to do, however, is observe the effect of changing the solvent on the resulting hydrogen-bonding motifs found within the structure.

Unlike the straightforward crystallization from aqueous solutions, violuric acid does not readily crystallize from methanol. Storage of a methanol solution at 278 K for around one month resulted in a few very delicate needle crystals only. These were taken to Station 9.8, SRS, Daresbury Laboratory, in an ice-box and data were collected via the EPSRC-funded UK National Crystallography Service.

The molecular structure of violuric acid methanol solvate, (I), is shown in Fig. 1. As is clear from the figure, the isonitroso group is disordered over two positions with relative occupancies of approximately 3:1 for the major and minor components. This disorder is not uncommon and is something that we have observed in violurate metal complexes, although it is not observed at all in violuric acid monohydrate. Although this disorder has little bearing on the molecular structure, it is of great importance to the crystal packing. Other than this disorder, the molecular dimensions are largely unexceptional and the rest of the violuric acid molecule appears to be well ordered; the exocyclic angles at atom C4 deviate somewhat from the ideal trigonal value of 120° because of steric interaction between the isonitroso group and adjacent ring substituents.

The most elegant aspect of the analysis is to be found not in the molecular structure but in the crystal packing. In essence, this is rather simple and consists of three basic hydrogen-bonding graph-set motifs (Bernstein et al., 1995). Firstly, there is the very common R²₂(8) motif – found in almost all barbiturate crystal structures – involving two N—H···O hydrogen bonds, which is located about a crystallographic inversion centre and connects the violuric acid molecules into tapes, which run parallel to the ac diagonal. Secondly – and ignoring the minor disorder component for the moment – the methanol molecule is connected to the violuric acid molecule by a more unusual but not uncommon R²₁(6) motif, involving the methanol O-bound H atom acting as a bifurcated donor to one carbonyl O atom and the O atom of the isonitroso group. Finally, an R⁴₄(8) motif links the aforementioned methanol and isonitroso atoms together with their symmetry equivalents about a crystallographic inversion centre. It is this final link which joins the tapes together to form a two-dimensional network sheet structure, as shown in Fig. 2.

Given that the minor disorder component is present in around one-quarter of the crystal structure, we cannot simply ignore how it affects the crystal packing. Whilst the R²₂(8) tapes are unaffected by this disorder, the way they are linked together is very much at the heart of the disorder analysis. Fig. 3 gives two views of the methanol–violuric interactions, with the major and minor disorder components shown separately. At the top is the major component, with the two hydrogen-bonding motifs already discussed clearly visible. At the bottom is the hydrogen bonding involving the minor component of the disorder only. Although the only atoms displaced are those of the isonitroso substituent (the methanol molecule is not significantly disordered, despite the large ellipsoids which suggest that this might be a possibility), the difference in the hydrogen bonding motifs is quite striking. The methanol O-bound H atom can now no longer act as a bifurcated donor; were it to form an O—H···N interaction, then the H···A distance would be an unacceptably long 2.86 Å. We therefore have a single large R⁴₄(16) connection between the violuric acid tapes, which themselves are unaffected by the disorder. The effect on the overall form of the crystal packing is only slight; Fig. 4 shows the crystal packing involving the minor component only and the patterns formed are still the same.

Whilst there is extensive hydrogen bonding within each sheet, there are no hydrogen bonding interactions linking different sheets together. The distance between adjacent sheets is around 3 Å, which is very close and is less than the sum of the van der Waals radii of the respective atoms. As a result, what we see are many dipolar carbonyl–carbonyl interactions (Allen et al., 1998) between the tightly packed sheets (Fig. 5). Multipolar interactions have often been overlooked when considering those interactions that play a significant role in crystal packing. However, with the development of information-rich structural databases and powerful search algorithms, the subject has received increasing attention in recent years, in both small-molecule and macromolecular crystallography, so much so that a review has been published recently (Paulini et al., 2005). This structure offers a further example of one in which dipolar interactions are vital to the integrity of the crystal packing. Despite this, the difficulty with which the crystals grew, and their relative instability once formed, suggests that the attractive strength of these purely electrostatic interactions cannot be relied upon as the basis of controllable crystal growth in the same manner as hydrogen bonds are now widely considered.

S2. Experimental

Violuric acid (0.17 g, 1 mmol) was dissolved in methanol (10 ml) with gentle heating. Storage for around one month at 278 K resulted in the growth of very small and fragile feather-like crystal agglomerations.

S3. Refinement

The N═OH group was refined as disordered, with a refined occupancy of 0.749 (4) for the major component. The anisotropic displacement parameters of atoms N3 and N3' were constrained to be equal. All H atoms were located in a difference electron density map and, with the exception of the methyl H atoms, freely refined. Methyl H atoms were refined using a riding model, with U_iso(H) values of 1.5U_eq(C) and a C—H distance of 0.98 Å. There is a short H···H contact between this methyl group and the H atom of the minor component of the disordered isonitroso group; this indicates that the methyl group is also disordered, but the minor component was not included in the refinement.

Computing details

Data collection: APEX2 (Bruker, 2004); cell refinement: APEX2; data reduction: SAINT (Bruker, 2001); program(s) used to solve structure: SIR2002 (Burla et al., 2003); program(s) used to refine structure: SHELXTL (Sheldrick, 2001); molecular graphics: SHELXTL and Mercury (Bruno et al., 2002); software used to prepare material for publication: SHELXTL and local programs.

Figures

	Fig. 1. The asymmetric unit of (I), with 50% probability displacement ellipsoids. Open bonds show the minor disorder component.
	Fig. 2. The hydrogen-bonded sheet structure in (I), with the minor disorder component ignored, viewed along the ab diagonal. Dashed lines (blue in the on-line version) represent hydrogen bonds and (red) the continuation of hydrogen bonding.
	Fig. 3. A representation of how the disorder affects the hydrogen-bonding motifs. (a) The hydrogen bonding involving the major component, and (b) the hydrogen bonding involving the minor component.
	Fig. 4. The hydrogen-bonded sheet structure in (I), with the major disorder component ignored, viewed along the ab diagonal. Dashed lines are as in Fig. 2.
	Fig. 5. Some highlighted carbonyl–carbonyl interactions within the crystal packing of (I).

pyrimidine-2,4,5,6(1H,3H)-tetrone 5-oxime methanol solvate

Crystal data

C4H3N3O4·CH4O	Z = 2
Mr = 189.14	F(000) = 196
Triclinic, P1	Dx = 1.530 Mg m^−3
Hall symbol: -P 1	Synchrotron radiation, λ = 0.6751 Å
a = 4.9650 (9) Å	Cell parameters from 1809 reflections
b = 8.6210 (17) Å	θ = 4.0–28.8°
c = 10.2610 (17) Å	µ = 0.14 mm^−1
α = 100.474 (2)°	T = 120 K
β = 102.939 (2)°	Needle, colourless
γ = 99.823 (2)°	0.17 × 0.06 × 0.04 mm
V = 410.56 (13) Å^3

Data collection

Bruker APEX2 CCD diffractometer	1462 independent reflections
Radiation source: Daresbury SRS station 9.8	1290 reflections with I > 2σ(I)
Silicon 111 monochromator	Rint = 0.013
thin–slice ω scans	θmax = 24.0°, θmin = 4.0°
Absorption correction: multi-scan (SADABS; Sheldrick, 2004)	h = −5→5
Tmin = 0.950, Tmax = 0.995	k = −10→10
3107 measured reflections	l = −12→12

Refinement

Refinement on F^2	Primary atom site location: structure-invariant direct methods
Least-squares matrix: full	Secondary atom site location: difference Fourier map
R[F^2 > 2σ(F^2)] = 0.041	Hydrogen site location: difference Fourier map
wR(F^2) = 0.113	H atoms treated by a mixture of independent and constrained refinement
S = 1.08	w = 1/[σ^2(Fo^2) + (0.0612P)^2 + 0.1629P] where P = (Fo^2 + 2Fc^2)/3
1462 reflections	(Δ/σ)max < 0.001
152 parameters	Δρmax = 0.29 e Å^−3
0 restraints	Δρmin = −0.28 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.

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

	x	y	z	Uiso*/Ueq	Occ. (<1)
O1	0.1767 (3)	0.28166 (15)	0.21124 (12)	0.0303 (3)
O2	0.7259 (3)	−0.06990 (15)	0.11945 (11)	0.0301 (3)
O3	0.8000 (2)	0.12759 (14)	0.56769 (11)	0.0253 (3)
O4	0.2014 (3)	0.37904 (18)	0.47908 (17)	0.0258 (5)	0.749 (4)
H4	0.185 (7)	0.426 (4)	0.570 (4)	0.060 (10)*	0.749 (4)
O4'	0.4430 (10)	0.3260 (6)	0.6113 (5)	0.0294 (16)	0.251 (4)
H4'	0.30 (2)	0.370 (13)	0.654 (11)	0.06 (3)*	0.251 (4)
N1	0.4500 (3)	0.10589 (17)	0.16874 (15)	0.0250 (4)
H1N	0.393 (4)	0.096 (2)	0.081 (2)	0.028 (5)*
N2	0.7611 (3)	0.03382 (17)	0.34269 (14)	0.0231 (3)
H2N	0.884 (5)	−0.024 (3)	0.367 (2)	0.034 (5)*
N3	0.4059 (6)	0.2948 (4)	0.5106 (4)	0.0247 (7)	0.749 (4)
N3'	0.345 (2)	0.3137 (16)	0.4805 (15)	0.0247 (7)	0.251 (4)
C1	0.3519 (3)	0.20861 (19)	0.25692 (17)	0.0242 (4)
C2	0.6492 (3)	0.01775 (19)	0.20473 (17)	0.0237 (4)
C3	0.6900 (3)	0.12542 (19)	0.44808 (16)	0.0222 (4)
C4	0.4706 (3)	0.21783 (19)	0.40561 (17)	0.0231 (4)
C5	0.2034 (7)	0.4555 (4)	0.8319 (3)	0.0753 (9)
H5A	0.3635	0.4022	0.8378	0.113*
H5B	0.0310	0.3764	0.8253	0.113*
H5C	0.2416	0.5425	0.9141	0.113*
O5	0.1666 (3)	0.52024 (17)	0.71566 (14)	0.0389 (4)
H5	0.024 (6)	0.561 (3)	0.710 (3)	0.059 (8)*

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
O1	0.0303 (7)	0.0368 (7)	0.0293 (6)	0.0213 (5)	0.0083 (5)	0.0076 (5)
O2	0.0355 (7)	0.0356 (7)	0.0236 (6)	0.0224 (6)	0.0083 (5)	0.0031 (5)
O3	0.0235 (6)	0.0318 (6)	0.0223 (6)	0.0115 (5)	0.0064 (5)	0.0047 (5)
O4	0.0248 (9)	0.0269 (9)	0.0303 (10)	0.0160 (7)	0.0107 (6)	0.0042 (7)
O4'	0.034 (3)	0.039 (3)	0.016 (3)	0.014 (2)	0.008 (2)	0.001 (2)
N1	0.0268 (8)	0.0305 (8)	0.0197 (8)	0.0157 (6)	0.0048 (6)	0.0037 (6)
N2	0.0217 (7)	0.0274 (7)	0.0234 (7)	0.0140 (6)	0.0062 (5)	0.0055 (6)
N3	0.0176 (16)	0.0217 (12)	0.038 (2)	0.0104 (10)	0.0096 (14)	0.0053 (10)
N3'	0.0176 (16)	0.0217 (12)	0.038 (2)	0.0104 (10)	0.0096 (14)	0.0053 (10)
C1	0.0212 (8)	0.0255 (8)	0.0286 (9)	0.0101 (7)	0.0093 (7)	0.0047 (7)
C2	0.0233 (8)	0.0253 (8)	0.0241 (8)	0.0104 (7)	0.0068 (6)	0.0043 (6)
C3	0.0191 (8)	0.0231 (8)	0.0252 (9)	0.0056 (6)	0.0080 (6)	0.0037 (6)
C4	0.0210 (8)	0.0228 (8)	0.0259 (9)	0.0068 (6)	0.0075 (6)	0.0032 (6)
C5	0.102 (2)	0.0693 (18)	0.0590 (16)	0.0507 (17)	0.0029 (15)	0.0176 (13)
O5	0.0338 (7)	0.0422 (8)	0.0407 (8)	0.0194 (6)	0.0130 (6)	−0.0044 (6)

Geometric parameters (Å, º)

O1—C1	1.218 (2)	N1—C1	1.376 (2)
O2—C2	1.2202 (19)	N1—C2	1.374 (2)
O3—C3	1.2218 (19)	N2—H2N	0.88 (2)
O4—H4	0.98 (4)	N2—C2	1.372 (2)
O4—N3	1.361 (3)	N2—C3	1.369 (2)
N3—C4	1.293 (4)	C1—C4	1.490 (2)
O4'—H4'	1.01 (12)	C3—C4	1.490 (2)
O4'—N3'	1.299 (13)	C5—H5A	0.980
N3'—C4	1.349 (16)	C5—H5B	0.980
O5—H5	0.84 (3)	C5—H5C	0.980
N1—H1N	0.86 (2)	O5—C5	1.395 (3)
H4—O4—N3	101 (2)	N1—C2—N2	116.56 (14)
O4—N3—C4	114.6 (2)	O3—C3—N2	121.02 (14)
H4'—O4'—N3'	105 (6)	O3—C3—C4	123.65 (14)
O4'—N3'—C4	111.7 (8)	N2—C3—C4	115.32 (14)
H5—O5—C5	108.1 (19)	N3—C4—C1	129.00 (17)
H1N—N1—C1	120.4 (13)	N3—C4—C3	111.45 (17)
H1N—N1—C2	112.9 (13)	N3'—C4—C1	109.4 (5)
C1—N1—C2	126.63 (15)	N3'—C4—C3	131.1 (5)
H2N—N2—C2	117.3 (13)	C1—C4—C3	119.53 (14)
H2N—N2—C3	115.8 (13)	O5—C5—H5A	109.5
C2—N2—C3	126.80 (14)	O5—C5—H5B	109.5
O1—C1—N1	119.99 (15)	O5—C5—H5C	109.5
O1—C1—C4	124.90 (14)	H5A—C5—H5B	109.5
N1—C1—C4	115.10 (14)	H5A—C5—H5C	109.5
O2—C2—N1	122.35 (15)	H5B—C5—H5C	109.5
O2—C2—N2	121.09 (14)
C2—N1—C1—O1	178.94 (15)	O1—C1—C4—N3	3.5 (3)
C2—N1—C1—C4	−2.3 (2)	O1—C1—C4—N3'	2.6 (6)
C3—N2—C2—O2	−178.78 (15)	O1—C1—C4—C3	−178.37 (15)
C3—N2—C2—N1	1.7 (2)	N1—C1—C4—N3	−175.2 (2)
C1—N1—C2—O2	−179.43 (15)	N1—C1—C4—N3'	−176.1 (6)
C1—N1—C2—N2	0.1 (3)	N1—C1—C4—C3	2.9 (2)
C2—N2—C3—O3	178.33 (15)	O3—C3—C4—N3	−2.3 (3)
C2—N2—C3—C4	−0.9 (2)	O3—C3—C4—N3'	−1.9 (8)
O4—N3—C4—C1	−0.3 (4)	O3—C3—C4—C1	179.28 (14)
O4—N3—C4—C3	−178.5 (2)	N2—C3—C4—N3	176.93 (19)
O4'—N3'—C4—C1	−179.8 (7)	N2—C3—C4—N3'	177.3 (7)
O4'—N3'—C4—C3	1.4 (13)	N2—C3—C4—C1	−1.5 (2)

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
O4—H4···O5	0.98 (4)	1.59 (4)	2.567 (2)	174 (3)
O4′—H4′···O5	1.01 (12)	1.64 (12)	2.574 (5)	152 (9)
O5—H5···O1^i	0.84 (3)	1.98 (3)	2.7361 (18)	150 (2)
O5—H5···O4^i	0.84 (3)	2.21 (3)	2.761 (2)	124 (2)
N1—H1N···O2^ii	0.86 (2)	1.97 (2)	2.8322 (19)	178.2 (19)
N2—H2N···O3^iii	0.88 (2)	1.98 (2)	2.8519 (18)	172.2 (19)

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