5-Iodo­uracil: structure of a reflection twin

The crystal structure of 5-iodo­uracil, C₄H₃IN₂O₂, has been determined in the noncentrosymmetric space group P2₁ on a nonmerohedrally twinned crystal. Both twin components are enanti­omorphically pure, but the twin element is a mirror plane perpendicular to c*. The mol­ecular structure is discussed and stacking faults in the two-dimensional packing are proposed as a reason for the twinning.

S1. Introduction

The crystal structure of the title compound, (I), is known from the literature (Sternglanz et al., 1975; Portalone, 2008). The achiral molecule crystallizes in the noncentrosymmetric space group P2₁. The correct absolute structure has been determined based on the refined R value of both enanti­omorphs (Sternglanz et al., 1975) or by determination of the Flack parameter (Flack, 1983), which resulted in a value close to zero (Portalone, 2008). A second polymorph has been reported by Valkonen et al. (2013). Here, the space group is Cmca and a crystallographic mirror plane is present perpendicular to the molecular plane. Consequently, the molecules in this structure are disordered.

In the original structure determination of Sternglanz et al. (1975), the displacement parameters of the I atom are larger than expected, about twice the magnitude of the other atoms. On the other hand, atom C5 which is connected to the iodine is extremely anisotropic. There is no explanation given in the paper, but these displacement parameters might be an indication for difficulties in the intensity determination at that time.

In an attempt to optimize the procedure for intensity integration of needle-shaped crystals, we therefore took the structure determination of (I) as an example. Overall, we have measured data sets of several crystals of (I). All of the structures show the monoclinic P2₁ structure described above. One of the crystals showed inter­esting nonmerohedral twinning properties, which we will describe in this communication.

S2. Experimental

S2.1. Synthesis and crystallization

5-Iodo­uracil was purchased from Aldrich. Crystals were obtained by slow cooling of a hot saturated solution in acetone.

S2.2. Refinement

Crystal data, data collection and structure refinement details are summarized in Table 1. Indexing of the diffraction pattern with the DIRAX program (Duisenberg, 1992) resulted in two orientation matrices that are related by a twofold rotation about uvw=[100]. These orientation matrices were used for the intensity determination with Eval15 (Schreurs et al., 2010) and they are included in the CIF file.

Initial structure refinements against the non-overlapping reflections of the two twin components resulted in two models with opposite absolute structure (Fig. 4). The corresponding Flack parameters are x = 0.098 (13) and 0.915 (17), respectively (Parsons et al., 2013). This situation was confirmed by a refinement against twinned reflection data in a HKLF-5 file format (Herbst-Irmer & Sheldrick, 1998). This HKLF-5 file contains the two twin domains and the corresponding inverted hkl indices, thus in total four components. Two components refined batch scale factors of approximately zero, BASF = -0.005 (16) and 0.00 (3). Only two components had significant BASF values of 0.61 (4) and 0.394 (16). This corresponds to the correct absolute structure of the first twin component and the inverted absolute structure of the second twin component.

As consequence of this absolute structure determination and large correlations present in the four-component case, an HKLF-5 reflection file was prepared for the final refinements, where the hkl indices of the second twin component have been inverted with respect to the measured indices. The TWINABS software was used for this purpose (Sheldrick, 2012). The BASF parameter for this two-component twin refined to 0.3891 (8).

S3. Comment

The asymmetric unit of (I) is shown in Fig. 1. Within standard uncertainties, the bond distances in (I) are equal to those reported by Portalone (2008). The older structure determination of Sternglanz et al. (1975) is less precise and accurate. In (I), the six-membered ring is essentially planar, with a maximum deviation of 0.024 (4) Å from the least-squares plane. The torsion angles in the ring vary between -4.7 (7) and 3.3 (6)° (Table 2). The molecule can consequently be described with an approximate noncrystallographic C_s symmetry in the molecular plane. This is in contrast to the Cmca structure (Valkonen et al., 2013), where the mirror plane is perpendicular to the molecule. The r.m.s. deviation from the exact C_s symmetry in (I) is 0.0505 Å (Pilati & Forni, 1998).

Despite the nonmerohedral twinning, the R values in (I) are improved compared to the previous single-crystal structure determinations (Sternglanz et al., 1975; Portalone, 2008). The displacement parameters of the I atom are now in the same range as the other atoms. None of the atoms fails to the Hirshfeld rigid-bond test with more than 5σ (Hirshfeld, 1976). It should be noted that the standard uncertainties of the refined parameters in (I) are larger than those by Portalone (2008). This is due to a new way of estimation which the program SHELXL2014 uses for noncentrosymmetric structures (Sheldrick, 2015) compared to the previous version (SHELXL97; Sheldrick, 2008) used by Portalone (2008).

In (I), there is still residual electron density in the neighbourhood of the I atom. This might indicate remaining problems in the data integration or the twin refinement, but it might also hint to anharmonic movements of the heavy atom which are not taken into account with the current displacement parameters. The values for the residual electron density in (I) are slightly larger than those reported by Portalone (2008). We were not able to reproduce the latter values from the deposited structure factors.

The high quality of the structure determination allowed us to perform a rigid-body analysis of the anisotropic displacement parameters using the THMA11 software (Schomaker & Trueblood, 1998). For this TLS analysis, we used unit weights for all atoms in order to avoid a too large influence of the exocyclic iodine, which has much smaller standard uncertainties in the displacement parameters. This resulted in a rigid-body R value of 0.11 (R = [Σ(U_obs - U_calc)²/ΣU_obs²]^1/2). This rather small R value confirms that the molecular movements can be approximated by a rigid-body motion. The L and T tensors are only slightly anisotropic with ratios of their maximum and minimum eigenvalues of L₁/L₃ = 2.6 and T₁/T₃ = 1.7. The eigenvectors of both the L and T tensors do not coincide with the molecular axes.

By inter­molecular N—H···O hydrogen bonding, (I) forms two-dimensional layers in the ab plane (Table 3 and Fig. 2), which can be characterized as a 4,4-grid (Barnett et al., 2008). As a consequence, the stacking in the c direction is an alternation between the hydrogen-bonded nets and iodine layers (Fig. 3). The shortest inter­molecular I···I distance is 4.1875 (4) Å. Assuming a van der Waals radius of 1.98 Å for iodine (Bondi, 1964), this distance is larger than the sum of van der Waals radii. On the other hand, it is well known that the shape of iodine is not spherical if inter­molecular contacts are considered. The radius varies between 1.76 Å in the direction of the I—C bond up to 2.13 Å perpendicular to the bond (Nyburg & Faerman, 1985). In (I), the corresponding angles are 114.49 (12)° for C—I···Iⁱ [symmetry code: (i) -x+2, y-1/2, -z+2] and 142.68 (12) ° for C—I···Iⁱⁱ [symmetry code: (ii) -x+2, y+1/2, -z+2]. A very weak halogen–halogen bond might thus be present.

The indexing of reflections resulted in a twofold rotation about uvw=[100] as twin operation. This relationship was used for the determination of the measured intensities (see Experimental). In the monoclinic crystal system, this twin operation can equivalently be expressed as twofold rotation about c*. Due to the presence of iodine as a strong anomalous scatterer, large Bijvoet differences are expected. A values of 306 is calculated for Friedif (Flack & Shmueli, 2007). This should allow the reliable determination of the absolute structure even in the twin case. Scatter plots of the Bijvoet differences for the two twin domains are shown in Fig. 4. It is evident that the two twin domains are different enanti­omorphs. The appropriate twin law is consequently not a rotation about c* but a mirror at hkl=(001). Considering the above discussion about the stacking of two-dimensional layers it is reasonable to explain the twinning in terms of stacking faults. Fig. 4 shows a model for such a stacking fault based on a reflection at the hkl=(001) plane.

The space group P2₁ of (I) does not contain symmetry elements of the second kind (inversion centres, mirrors or glide planes) and therefore belongs to the Sohnke space groups. Additionally, the space group is polar. Crystallization of achiral molecules in polar space groups is a topic of ongoing research (Hulliger et al., 2013). When the polar structure is obtained from a centrosymmetric high-temperature structure by a ferroelectric phase transition, the result is often a (merohedral) inversion twin. A similar but more seldom situation occurs in the so-called kryptoracemates. Here, the space group belongs to the Sohnke groups but the asymmetric unit contains two or more molecules of opposite chirality (Bernal & Watkins, 2015). In this communication, we describe a third case: we have obtained a nonmerohedrally twinned crystal of which both twin components are enanti­omorphically pure but are related to each other by a mirror operation.

Computing details

Data collection: APEX2 (Bruker, 2007); cell refinement: PEAKREF (Schreurs, 2013); data reduction: EVAL15 (Schreurs et al., 2010) and TWINABS (Sheldrick, 2012); program(s) used to solve structure: coordinates from the literature (Sternglanz et al., 1975); program(s) used to refine structure: SHELXL2014 (Sheldrick, 2015); molecular graphics: PLATON (Spek, 2009); software used to prepare material for publication: SHELXL2014 (Sheldrick, 2015).

Figures

	Fig. 1. The molecular structure of the title compound, (I), with displacement ellipsoids drawn at the 50% probability level. H atoms are drawn as small spheres of arbitrary radii.
	Fig. 2. Hydrogen bonds forming a two-dimensional network in (I), viewed along the c axis. The thickness of the layer corresponds to the length of the c axis [14.1348 (7) Å].
	Fig. 3. Alternation of N—H···O hydrogen-bonded nets and iodine layers in (I), stacked upon each other in the c direction. The view is along the b axis.
	Fig. 4. Scatter plots for the Bijvoet differences in the two twin components, prepared by the PLATON software (Spek, 2009). For preparing the plots, only nonoverlapping reflections unaffected by the twinning were used. Only reflections with Δobs > 0.25σ(Δobs) are drawn. Reflections confirming the correct absolute structure are drawn in black, relections with the wrong sign are shown in red. Left: first twin component; 291 reflections are black and 74 are red. Right: second twin component; 66 reflections are black and 260 are red.
	Fig. 5. Artistic impression of the stacking faults created by the twin operation in the crystal structure of (I). The structure is mirrored at the (001) plane.

5-Iodo-1H-pyrimidine-2,4-dione

Crystal data

C4H3IN2O2	F(000) = 220
Mr = 237.98	Dx = 2.589 Mg m^−3
Monoclinic, P21	Mo Kα radiation, λ = 0.71073 Å
a = 4.8929 (3) Å	Cell parameters from 4649 reflections
b = 4.4195 (2) Å	θ = 4.2–30.0°
c = 14.1348 (7) Å	µ = 5.16 mm^−1
β = 92.721 (3)°	T = 150 K
V = 305.31 (3) Å^3	Needle, colourless
Z = 2	0.24 × 0.05 × 0.04 mm

Data collection

Bruker Kappa APEXII diffractometer	2932 reflections with I > 2σ(I)
Radiation source: sealed tube	Rint = 0.023
φ and ω scans	θmax = 30.1°, θmin = 2.9°
Absorption correction: multi-scan (TWINABS; Sheldrick, 2012)	h = −6→6
Tmin = 0.313, Tmax = 0.433	k = −6→6
13501 measured reflections	l = −19→19
3122 independent reflections

Refinement

Refinement on F^2	Secondary atom site location: difference Fourier map
Least-squares matrix: full	Hydrogen site location: inferred from neighbouring sites
R[F^2 > 2σ(F^2)] = 0.019	H-atom parameters constrained
wR(F^2) = 0.046	w = 1/[σ^2(Fo^2) + (0.0254P)^2 + 0.0706P] where P = (Fo^2 + 2Fc^2)/3
S = 1.09	(Δ/σ)max < 0.001
3122 reflections	Δρmax = 0.93 e Å^−3
83 parameters	Δρmin = −0.54 e Å^−3
1 restraint	Absolute structure: Flack x determined using 728 quotients [(I+)-(I-)]/[(I+)+(I-)] (Parsons et al., 2013)
Primary atom site location: heavy-atom method	Absolute structure parameter: −0.021 (12)

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. Refined as a 2-component twin.

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

	x	y	z	Uiso*/Ueq
I1	0.74759 (6)	0.75499 (2)	0.90522 (2)	0.02837 (9)
O2	0.7068 (5)	0.2240 (12)	0.5053 (2)	0.0209 (7)
O4	1.0837 (7)	0.8996 (7)	0.7158 (2)	0.0201 (6)
N1	0.5260 (5)	0.2597 (17)	0.65075 (19)	0.0150 (5)
H1N	0.3972	0.1279	0.6348	0.018*
N3	0.8863 (6)	0.5619 (8)	0.6129 (3)	0.0142 (6)
H3N	0.9989	0.6252	0.5703	0.017*
C2	0.7056 (8)	0.3394 (9)	0.5844 (3)	0.0151 (7)
C4	0.9115 (7)	0.6981 (7)	0.7009 (3)	0.0123 (8)
C5	0.7221 (9)	0.5860 (9)	0.7685 (3)	0.0156 (7)
C6	0.5349 (8)	0.3729 (9)	0.7407 (3)	0.0161 (7)
H6	0.4083	0.3016	0.7845	0.019*

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
I1	0.04520 (15)	0.02549 (12)	0.01438 (11)	−0.0017 (3)	0.00102 (10)	−0.0005 (2)
O2	0.0146 (10)	0.025 (2)	0.0236 (12)	−0.0010 (17)	0.0041 (10)	−0.0091 (17)
O4	0.0161 (15)	0.0186 (14)	0.0256 (16)	−0.0060 (12)	0.0001 (13)	−0.0006 (11)
N1	0.0098 (10)	0.0140 (11)	0.0210 (12)	−0.006 (3)	0.0009 (9)	0.002 (3)
N3	0.0088 (15)	0.0167 (15)	0.0173 (16)	−0.0040 (12)	0.0024 (12)	0.0009 (12)
C2	0.0092 (15)	0.0149 (16)	0.0213 (18)	0.0020 (13)	0.0017 (16)	−0.0014 (11)
C4	0.0110 (14)	0.008 (2)	0.0182 (16)	−0.0008 (11)	−0.0001 (12)	0.0014 (12)
C5	0.0164 (17)	0.0153 (17)	0.0152 (17)	0.0014 (16)	0.0015 (16)	0.0019 (13)
C6	0.0137 (17)	0.0170 (15)	0.0179 (18)	0.0019 (14)	0.0031 (15)	0.0044 (14)

Geometric parameters (Å, º)

I1—C5	2.070 (4)	N3—C2	1.370 (5)
O2—C2	1.229 (5)	N3—C4	1.383 (5)
O4—C4	1.237 (4)	N3—H3N	0.8800
N1—C2	1.361 (5)	C4—C5	1.449 (5)
N1—C6	1.365 (6)	C5—C6	1.359 (6)
N1—H1N	0.8800	C6—H6	0.9500
C2—N1—C6	123.6 (4)	O4—C4—N3	119.8 (3)
C2—N1—H1N	118.2	O4—C4—C5	125.9 (4)
C6—N1—H1N	118.2	N3—C4—C5	114.3 (3)
C2—N3—C4	127.2 (3)	C6—C5—C4	119.3 (4)
C2—N3—H3N	116.4	C6—C5—I1	122.0 (3)
C4—N3—H3N	116.4	C4—C5—I1	118.7 (3)
O2—C2—N1	123.4 (4)	C5—C6—N1	121.0 (4)
O2—C2—N3	122.2 (4)	C5—C6—H6	119.5
N1—C2—N3	114.4 (4)	N1—C6—H6	119.5
C6—N1—C2—O2	175.6 (5)	N3—C4—C5—C6	−2.4 (5)
C6—N1—C2—N3	−4.7 (7)	O4—C4—C5—I1	−3.5 (5)
C4—N3—C2—O2	−177.0 (4)	N3—C4—C5—I1	177.2 (2)
C4—N3—C2—N1	3.3 (6)	C4—C5—C6—N1	1.1 (6)
C2—N3—C4—O4	−179.2 (4)	I1—C5—C6—N1	−178.5 (4)
C2—N3—C4—C5	0.1 (5)	C2—N1—C6—C5	2.7 (8)
O4—C4—C5—C6	176.9 (4)

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
N1—H1N···O4^i	0.88	2.20	2.872 (5)	133
N3—H3N···O2^ii	0.88	1.89	2.754 (4)	169

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