Synthesis 2023; 55(02): 297-306
DOI: 10.1055/a-1921-8664
paper
Special Issue dedicated to Prof. Alain Krief

Crystal Structures of Organoselenium Compounds: Structural Descriptors for Chalcogen Bonds

Laurie Bodart
,
Johan Wouters

Authors thank the Fonds National de la Recherche Scientifique (FNRS), Belgium, for financial support.
 


Abstract

Less conventional non-covalent interactions such as chalcogen bonds attract the attention of researchers in various fields (organocatalysis, material sciences, biological chemistry, …). We present here useful descriptors to easily discriminate the structures in which chalcogen bonds involving selenium are observed. Our study focused on organoselenium compounds as chalcogen bond donors and on molecular entities, as chalcogen bond acceptors, containing N, O, S, Se, and Te atoms or aromatic rings. For conventional chalcogen bonds (C–Se⋯X, with X = N, O, S, Se, or Te), the combination of the C–Se⋯X angle and the distance between X and the C–Se-C plane proved to be most relevant for identification of chalcogen bonds. For chalcogen⋯π bonds, the most relevant parameters are a combination of the C–Se⋯X angle and the angle between the C–Se bond and the normal to the aromatic ring plane.


#

Introduction

For a long time, crystal engineering focused on intermolecular hydrogen bonds, which are now well studied.[1] However, over the past two decades, researchers have become increasingly interested in other, less conventional types of non-covalent interactions, firstly halogen bonds and more recently chalcogen (but also pnictogen and tetrel) interactions.[2] [3] [4] [5] These interactions result from an anisotropic distribution of the electron density around the donor, which can generate a region of lower electronic density, commonly called σ-hole, as well as another region with higher electronic density.[6,7] A σ-hole often corresponds to a region of positive electrostatic potential except if the element has a high electronegativity and low polarizability.[6] Because of their anisotropic electron density, halogens and chalcogens are therefore amphiphilic species, which can act either as an electrophile or as a nucleophile.

To avoid any confusion when using the terms halogen/chalcogen bond, it is necessary to recall the IUPAC recommendation according to which halogen and chalcogen interactions refer to ‘interactions in which an element of families VII/17 for halogens or VI/16 for chalcogen behaves as an electrophile (just like the hydrogen atom in an H bond) and interacts with an electron rich site (i.e., a Lewis base)’.[8] Typically, halogen and chalcogen bonds are, respectively, depicted as R–Hal⋯X or R–Ch⋯X, where the electron-deficient halogen or chalcogen atom is the donor and the Lewis base (X) the acceptor.[8] The main difference between halogen (HalB) and chalcogen bonds (ChB) is that chalcogens can have two substituents, which results in the existence of two σ-holes (located in the extension of the R–Ch and R′–Ch bonds), where halogens have only one (located in the extension of the R–Hal bond).[9] [10]

The σ-hole interactions, including the chalcogen interactions, are often explained by different contributions (i.e., electrostatic and orbital mixing).[11] Thus, the presence of a region with a positive electrostatic potential (σ-hole, due to the polarization of the electronic charge to the covalent bond) on the chalcogen atom allows an electrostatic interaction with an electron rich region (a free pair of a Lewis base, for example).[12] According to Murray et al. this σ-hole refers to the electron-deficient region of the outer lobe of a p-orbital involved in a covalent bond.[7] , [12] [13] [14] [15]

The orbital mixing contribution (n → σ*) corresponds to an interaction between an occupied orbital (n) of a Lewis base and the antibonding orbital (σ*R–Ch) of the chalcogen atom. This interaction results in the directionality of the chalcogen interaction.[11] , [16] [17] [18]

Therefore, several factors influence the formation of chalcogen interactions:[7]

- the electronegativity of the chalcogen atom (the less electronegative the chalcogen is, the stronger the positive zone).[19] A σ-hole may have a positive or negative electrostatic potential depending on the Ch atom and its substituent. For example, the strongly electronegative and weakly polarizable oxygen atom, which is often not considered as chalcogen bond donor, often has a σ-hole corresponding to a negative electrostatic potential;[7] [14] [15]

- the polarizability of the chalcogen atom and of its substituent (a more polarizable atom will have a stronger positive zone);

- the electron-attracting character of the chalcogen substituents (the positive zone is stronger when the substituent is more electron-attracting, and considering the orbital mixing contribution to ChB, electron-withdrawing substituent lower the LUMO energy level, making the σ* a better acceptor of electron). Recent studies have shown that oxygen can indeed act as a chalcogen bond donor if it is bound to electron-withdrawing groups (which can then generate a positive electrostatic potential at its surface);[20] [21]

- the hybridization of the chalcogen atom (for example, in the case of sp-hybridization, the electron-deficient lobe of the bonding p-orbital is partially filled by the electronic charge of the s-orbital, which can prevent the formation of a positive electrostatic potential zone).[7]

Chalcogen interactions, and in general σ-hole bonds, are directional (close to linearity).[14] Politzer reports angles between 170 and 180°, more particularly, for chalcogen interactions Lundemba reports R–Ch⋯X bond angles above 160° (with Ch being a chalcogen bond donor and X a chalcogen bond acceptor).[13] [22] This directionality has an important potential for applications in various fields such as, among others, chemistry, biochemistry, or materials sciences. In the field of biochemistry, it has been shown that ebselen is able to interact by chalcogen bonding with the electronic lone pairs of nitrogen and oxygen atoms. It is thus suspected that this interaction may play a crucial role in its binding to biological targets.[23] Chalcogen bonding has also been shown to stabilize biological motifs in proteins and chalcogen interactions are also being studied in the field of drug design.[11,24,25] In chemistry and material sciences, chalcogen bonds are reported, among others, in the fields of organocatalysis,[26] [27] [28] [29] anion recognition, optoelectronics,[30] [31] [32] or supramolecular chemistry.[9] , [33] [34] [35] [36] [37]

Intermolecular interactions between (heavy) chalcogens, have long been known as ‘secondary bonding interactions’, a term originally coined by Alcock.[38] This concept has been more recently replaced by halogen, chalcogen, pnicogen, or tetrel bondings. Several reviews about chalcogen bonds in small organic molecules and in proteins have already been published.[19] [22] , [39] [40] [41] [42] [43] However, the literature mentioning and studying chalcogen-π interactions is quite scarce.[41] [43] In this study, we report the results of a geometrical analysis of crystal structures retrieved from a search of the Cambridge Structural Database (CSD) to highlight the different parameters to be considered for crystal engineering based on chalcogen bonds. In particular, we focus on the study of organoselenated compounds (a similar approach has been applied recently by Tiekink on Te⋯N secondary bonding interactions).[44] In the first part of the study, we will focus on ‘classical’ chalcogen interactions to review the geometrical parameters commonly used to identify such contacts in crystallographic databases. The second part of this manuscript deals with the less commonly studied chalcogen-π interactions, with the objective to define relevant geometrical parameters to easily identify such contacts.


#

Results and Discussion

Search for C–Se⋯X ChB (X =N, O, S, Se, and Te)

The occurrence and spatial distribution of interactions between chemical functional groups can be visualized through 3D scatterplots generated by the knowledge-based library IsoStar.[45] It can be accessed through the web (https://isostar.ccdc.cam.ac.uk/html). In the frame of the present study, frequencies and directionalities of intermolecular contacts involving a selenium atom are available from pre-calculated scatterplots in IsoStar as illustrated in Figure [1] in the case of the methylseleno functional group.

Zoom Image
Figure 1 Two perpendicular views of IsoStar 3D scatterplots characterizing the methylseleno functional group (in stick) in interaction with C (cross, grey), N (blue spheres), O (red spheres), S (yellow spheres), or Se (green spheres) extracted from the CSD (A) or the PDB (B). Selection of the contacts is made through definition of a van der Waals distance range (fixing the longer contact at maximum distance equals to the sum of van der Waals radii).

Structures of both organic (CSD database, Figure [1]A) or protein molecules (PDB database, Figure [1]B) can be analyzed and show that close contacts involving a selenium atom are relatively frequent. Also, the relative position of atoms around the selenium is not random. A more systematic and quantitative analysis of these interactions has been possible through systematic searches among the CSD using ConQuest.[46] [47]

We have searched the CSD for structures of organoselenated compounds with short Se⋯X contacts (Se⋯X distance less than the sum of the van der Waals radii, with X = N, O, S, Se, Te). We limited the search to structures in which Se is bonded to two carbon atoms, as described further. We did not search for molecules containing halogen atoms as acceptor to avoid including structures where the chalcogen moiety would act as acceptor of halogen bonds.

In this part, we will discuss the frequency and geometry of these short distance contacts as well as parameters that can help identify the actual presence of chalcogen bonds in the structures. However, without an in-depth structural analysis, it remains difficult to distinguish the effective contributions of these chalcogen bonds to the overall stabilization of the structure. It remains indeed a challenge to unambiguously identify the presence of chalcogen bonding versus indirect close contacts resulting from crystal packing or from the presence of other stronger bonds with other atoms in the system.

In the literature, different geometrical parameters are used to describe a chalcogen bond and in particular, we note the use of spherical coordinates or the definition of a centroid between the substituents of the chalcogen atom, which is then used to define a centroid–chalcogen–X angle.[42] [48] In this work, we have opted for parameters that are easy to implement in ConQuest and ready to visualize.

The main geometrical parameters that describe the chalcogen bonds in the present study [Figure [2], and Figure S1 in the Supporting Information (SI)] are:

∙ distances (Å): d (Se⋯X, with maximum value being the sum of van der Waals radii of Se and X, or Se⋯Y, with Y the centroid of an aromatic ring), dZ (Se⋯Z, with Z the closest atom of the aromatic ring) dXP (distance between X and the plane passing through C1–Se–C2); dHZ (distance between the orthogonal projection of the Se atom on the aromatic plane and the centroid of the aromatic ring)

∙ angles (°): A (C1–Se–C2), A1 (C1–Se⋯X or C1–Se⋯Y, with Y the centroid of an aromatic ring), A2 (C2–Se⋯X or C2–Se⋯Y), A3 (angle between the plane passing through an aromatic ring and the one passing through C1–Se–C2) and A4 (angle between the C–Se bond vector and the normal to the aromatic plane);

∙ torsion angle (°): T (Se–C1–C2⋯X or Se–C1–C2⋯Y), which also proved to be an interesting parameter to position the Se atom and the chalcogen bond acceptor.

Zoom Image
Figure 2 Selected parameters for (A) analysis of Se⋯X and (B) Se⋯aromatic short contacts. Y: Centroid of aromatic ring; A: Angle C1–Se–C2; A1: angle C1–Se–X or C1–Se–Y; A2: Angle C2–Se–X or C2–Se–Y; d: distance Se-X or Se-Y; dZ: distance Se to the closest atom of the aromatic ring; dXP: distance X to the plane passing through C1–Se–C2; A3: angle between the plane passing through the aromatic ring and the one passing through C1–Se–C2; and A4: angle between the C–Se bond vector and the normal to the aromatic plane.

First, we searched the CSD for short Se⋯N/O/S/Se contacts. This search was performed in two ways: without specifying the number of atoms attached to the ChB donor Se atom (Nb n.s.) and then by setting its coordination to 2 (Nb = 2). The number of retrieved entries for each search is shown in Table [1].

Table 1 Number of Structures Retrieved from the CSD Database Searcha

X = N

X = O

X = S

X = Se

X = Te

Nb n.s.

Nb = 2

Nb n.s.

Nb = 2

Nb n.s.

Nb = 2

Nb n.s.

Nb = 2

Nb n.s

Nb = 2

115

108

197

135

88

85

325

309

3

3

a Search restricted to non-disordered single crystal of organics. Nb is the coordination number of the Se atom. n.s. means that Nb was not specified for the CSD search.

This initial research indicates that most of organoselenated compounds (in which Se is bonded to at least two carbon atoms) presenting short Se⋯X distance have a coordination number of 2. For the sake of clarity and conciseness, therefore, the remainder of this study will focus only on structures in which the Se atom has a coordination number of 2. It is interesting to note that several short Se⋯X contacts can be observed in the same structure, as a result, some CSD refcodes may appear several times in the analyses that follow.

Zoom Image
Figure 3 Values of dXP (Å) (blue) and C–Se···X angle (°) (grey) in CSD retrieved entries for A, X = N; B, X = O; C, X = S; D, X=Se and E, X = Te. Black lines are indicative of a C–Se⋯X angle value of 150° (dotted line) and a dXP value of 1 Å (straight line). To facilitate trend analysis CSD entries have been sorted in descending order of C–Se⋯X value.

It is necessary to discriminate the structures, within the retrieved CSD entries, in which these short contacts actually correspond to chalcogen bonds. This should be possible thanks to the parameters such as the distances d, dXP, and the angles A1 and A2 defined above. We also defined the Nc parameter, which is the quotient obtained by dividing the chalcogen bond distance, d, by the sum of the van der Waals radii of the Se and X atoms.[49] This parameter is equivalent to the reduced or normalized distance parameter (RXB) commonly used in the studies of halogen bonds.[50] [51] [52] Chalcogen bonds are directional, so in a strong chalcogen bond, the C–Se⋯X angle (A1 or A2) is close to 180° (most often above 160°).[13,14,22] On the other hand, a strong ChB requires the atoms C, Se, and X to be almost in the same plane, which implies a distance dXP close to 0 Å. Finally, the shorter the distance Se⋯X (i.e., the smaller the parameter Nc) the stronger the chalcogen bond should be. The analysis of the CSD structures in relation to the values of these different parameters (d, dxp, A1, A2) should confirm the presence of a chalcogen bond. It also allows the evaluation of the distribution of structures involving such intermolecular interactions as a function of the nature of the ChB acceptor (i.e., containing N, O, S, Se, or Te atom).

Zoom Image
Figure 4 A: Short Se⋯Se contacts in structure of 2,3,6,7-tetramethoxyselenanthrene radical cation triiodide (CSD refcode: RIKZIU) do not correspond to ChB but are rather the consequence of π-stacking. B: Observed interactions in structure 1,3-selenazolidine-2,4-dione (CSD refcode: BOJCOS), carbon atoms in light pink highlight a chain stabilized through ChB and H-bonds.

In general, in the different structures recovered from the CSD a decrease in the angle A1 or A2 (C–Se⋯X) is accompanied by an increase in the distance dXP, this is observed for all ChB acceptor studied (X = N, O, S, Se or Te; Figure [3] A, B, C, D, and E, respectively). The analysis reveals, as expected that the Nc parameter alone is not sufficient to discriminate chalcogen bonds from other short contacts generated by other interactions. A closer look at the structures with short Se⋯X contacts (small Nc values) indicate that small Nc values do not all correspond to strong chalcogen bonds (Figure S2, in SI). A short Nc parameter can correspond to A1 angle values well away from 180° as well as to high dXP distances (e.g., RIKZIU, short Se⋯Se distance due to π-stacking (Figure [4]A), Nc = 0.85, dXP = 3.23 Å and C–Se⋯Se angle = 89.5°; WANKAX, short Se⋯S contact, Nc = 0.93, dXP = 3.40 Å and C–Se⋯S angle = 96.4°). This was expected as crystal packing is the result of a variety of non-covalent interaction (H-bonds, π-stacking, …); ChB being only one, weak, component of the full stabilization effects. Furthermore, Nc parameter does not include the directional aspect of chalcogen bonds. Consequently, it is more relevant to combine the Nc value with the dXP distance and/or with the C–Se⋯X angle (Figure S3 in SI), to easily discriminate structures associated with strong chalcogen bonds (small Nc value, C–Se⋯X angle close to 180°, and small dXP distance (Figure S3 in SI and Table [2]).

Table 2 Strongest ChB Identified from the Plot of dXP and A1 in Function of the Normalized Contact Parameter (Nc)

X = N

CSD refcode

Nc

dXP (Å)

A1 (°)

UQAXUH

0.815

0.03

166.3

FEDJAC

0.820

0.07

177.2

UQAXUH

0.822

0.02

170.4

WERYAT

0.824

0.03

174.0

X = O

CSD refcode

Nc

dXP (Å)

A1 (°)

IDOFUE

0.793

0.33

162.7

FUJTOU

0.806

0.35

172.6

QUHYAV

0.840

0.01

167.3

APIQEX

0.844

0.21

168.0

KUDSIN

0.846

0.26

174.8

APIQEX

0.847

0.12

168.2

X = S

CSD refcode

Nc

dXP (Å)

A1 (°)

TITLOX

0.914

0.96

163.4

RUDPEL01

0.930

0.18

168.9

RUDNOT01

0.931

0.76

164.7

X = Se

CSD refcode

Nc

dXP (Å)

A1 (°)

BOJCOS

0.771

0.03

158.7

NOQWEY

0.872

0.80

160.4

NOQWAU

0.872

0.27

172.0

X = Te

CSD refcode

Nc

dXP (Å)

A1 (°)

MIVYIB

0.980

0.49

172.7

Among the structures with strong ChB, the 1,3-selenazolidine-2,4-dione structure (CSD reference code: BOJCOS) attracted our attention because it came out in two searches, those of the short Se⋯O and Se⋯Se contacts. Moreover, its Nc parameter (0.88 for Se⋯O and 0.77 for Se⋯Se) is, in the case of the Se⋯Se contact, significantly smaller than that of other structures with strong C–Se⋯Se ChB (Nc around 0.87). This structure is stabilized by various chalcogen interactions. Just as the amide-amide dimer is formed by R2 2(8) hydrogen bonds, the proximity of a ChB acceptor (O) and donor (Se) allows the formation of another dimer stabilized by cyclic C–Se⋯O chalcogen interactions [R2 2(6) interaction pattern, if one uses a notation analogous to the well-known graph sets for H-bonds].[53] Combination of this amide-amide H-bond interaction with the cyclic C–Se⋯O ChB stabilizes chains of 1,3-selenazolidine-2,4-dione in the structure. These chains are further interacting through C–Se⋯Se ChB (Figure [4]B).

The torsion angle T is an alternative descriptor to dXP that also positions the Lewis base/ChB acceptor with respect to the Se donor atom. For short d (Se⋯X) distances, most frequent values for T are around 0°, corresponding to the X in the same plane that the C1, C2 and Se atoms. The polar scatter plot (Figure [5] and Figure S4 ion SI) combines information of both the values of T and the strength of the ChB (via d). The shortest Se⋯X contacts are found at the center of the plot and most frequent ones correspond to T values close to 0°.

Zoom Image
Figure 5 Polar scatter plot of the torsion [T(°)] values and of the Se⋯X distance [d (Å)]

This scatter plot (Figure [5] and Figure S4 in SI) as well as the combined analysis of parameters Nc, dXP and A (Figure [3] and Figure S2 in SI) indicate significant differences in the proportion of structures stabilized by chalcogen bond depending on the nature of the X atom. For X = N (Figure [3]A and Figure S2 A in SI), most retrieved structures do indeed appear to involve chalcogen bonds. We observe that structures in which the C–Se⋯X angle (A1 or A2 depending on whether the chalcogen bond is along the σ* orbital of the C1–Se or C2–Se bond) is between 150 and 180° also have a dXP parameter smaller than or very close to 1 Å, indicating a small deviation from the planarity of the C, Se, and X atoms. These same structures also tend to have a smaller Nc parameter than those with a C–Se⋯X angle of less than 150° (indicating a priori a weaker or no chalcogen bond). This proportion of short contacts with geometries optimal for ChB is lower for the other X atoms studied (proportion of short contacts with geometries indicative of ChB: N (91%) >O (78%) >S (44%) ~Se (41%) (Figure [3] and Figure S2 in SI).

For X=Te, there are not enough structures to draw relevant conclusions. However, of the three structures found, only one (CCDC refcode: MIVYIB) appears to have a Se⋯Te ChB with Se as the ChB donor. Since Te is normally a better ChB donor than Se, it is at first glance surprising to find this intermolecular interaction. However, this can be explained by the existence of a strong intramolecular C–Te⋯Se chalcogen bond (with Te as the ChB donor, Figure [6] and Table [3]). In this structure, Te and Se are therefore both ChB donors and acceptors. In the structure MIVZAU, the important dXP value indicates that Se is not a ChB donor, and indeed, it acts as a ChB acceptor (Figure [6] and Table [3]). Similarly, in the third retrieved structure (CSD refcode ECITEP), the high dXP and the low C–Se⋯Te angle values indicate that there is no ChB with Se as ChB donor. However, the structure is rather stabilized by C–Se⋯Se (as found when searching for close Se–Se short contacts) and C–Te⋯Te ChB (Figure [6] and Table [3]).

Zoom Image
Figure 6 Chalcogen bonds in structures of A, MIVYIB; B, MIVZAU; and C, ECITEP. Inter- and intra-molecular ChB are shown as blue and magenta dased lines respectively.

Table 3 Parameters of ChB Observed in Structures with CSD Refcode MIVYIB, MIVZAU, and ECITEP

CSD refcode

C–Ch⋯X

Ch⋯X (Å)

C–Ch⋯X (°)

Distance X to C–Ch–C plane (dXP) (Å)

MIVYIB

C13–Te1⋯Se1a

3.218(1)

172.01(8)

0.049

C19–Se1-Te1

3.899(1)

172.75(8)

0.489

MIVZAU

C13–Te1⋯Se1a

3.213(2)

171.3(3)

0.274

C1–Te1⋯Se1

3.928(2)

168.6(2)

0.191

ECITEP

C4–Se2⋯Se3

3.720(3)

172.2(2)

0.398

C4–Te2⋯Te3

3.710(2)

172.6(2)

0.314

C1–Te1⋯Te3

3.747(2)

170.4(2)

0.463

a Intramolecular ChB.


#

Search for C–Se⋯π Chalcogen Bonds

The second part of the study is dedicated to the search of the less explored Se⋯π ChB. The search was limited to 5- or 6-atom aromatic rings, defining the ring atoms as C or N (6-atom rings) or as C, N, O, or S (5-atom rings). The effect of the number of atoms attached to the Se atom was also studied. This was done by comparing the number of CSD entries extracted by setting Nb to 2 with that without specifying the number of atoms attached to Se. The number of retrieved entries for each search is shown in Table [4].

Table 4 Number of Structures Retrieved from the CSD Database Searcha

Nb

Aromatic ring size

(number of atoms)

Atoms in ring

Retrieved CSD entries

2

6

C

1175

n.s.

6

C

1267

2

6

C, N

1259

n.s.

6

C, N

1353

2

5

C, N, O, S

  57

n.s.

5

C, N, O, S

  59

a Search restricted to non-disordered single crystal of organics. Nb is the coordination number of the Se atom, n.s. means that Nb was not specified for the CSD search.

The results of this preliminary research indicate that most of the recovered structures contain a 6-atom aromatic ring. In addition, most organoselenated compounds (in which Se is attached to at least two carbon atoms) with a dZ distance shorter than the sum of the van der Waals radii + 0.5 Å have a coordination number of 2. The study of C–Se⋯π ChB will focus only on structures in which the Se atom is attached to 2 carbon atoms. As in the previous research (about C–Se⋯X short contacts), several Se⋯π short contacts can be observed in the same structure, and thus some CSD refcodes can appear several times in the following analyses.

From the results of this preliminary research, it is necessary to distinguish the structures in which the short Se⋯π distance corresponds to ChB. In this case, the Se atom and the aromatic ring are expected to act as the electron-deficient and electron-rich species, respectively. From a geometrical parameter point of view, this implies the torsion angle (T) to have a value close to 0°, angle A1 (C–Se⋯π) to have a value close to 180° (applied cutoff: 150–180°), but also the C–Se bond to be almost perpendicular to the plane of the aromatic ring (A4 close to 0°, applied cutoff: 0–40°). Actually, A1 is more relevant than T because a value of A1 close to 180° implies T to be close to 0° (Figure S5 in SI).

Plotting of A4 against A1 (Figure [7]) shows that most of the retrieved structures deviate from these ideal parameters. Only 344 structures on 1175 (6-atom aromatic ring) and 13 on 57 (5-atom aromatic ring) fulfil these geometric criteria. Combination of these two criteria seems to be quite good to discriminate C–Se⋯π ChB as illustrated by the analyses of the retrieved structures. For example, in structure NUBHUQ, two C–Se⋯π ChB are observed (Figure [8] A, Table [5]), with one having more optimal parameters (A4 = 11.09°, dHZ = 0.12 Å) than the other (A4 = 39.812°, dHZ = 2.57 Å). Other examples are shown in Figure [8] and Table [5].

Zoom Image
Figure 7 Plots of A4 against A1 angles (°) for A: 6-atom aromatic ring (atoms in ring = C only or C and N) and B: 5-atoms aromatic ring (atoms in ring = C and N, O or S). Green frames highlight CSD entries fulfilling the geometric criteria for chalcogen bond.
Zoom Image
Figure 8 C–Se⋯π ChB (blue and magenta dashed lines) in structures of CSD refcode. A: NUBHUQ; B: NAHROH; C: VADVON, and D: ROYXEK.

Table 5 Parameters of Se⋯π ChB Illustrated in Figure [8a]

CSD refcode

A1 (°)

A4 (°)

A3 (°)

dHZ (Å)

NUBHUQ

167.1

11.1

89.2

0.12

NUBHUQb

177.2

39.8

51.9

2.57

NAHROH

172.0

13.9

84.4

0.50

VADVON

172.1

18.8

72.4

1.06

ROYXEK

165.1

17.7

74.9

1.12

a The dHZ parameter corresponds to the horizontal displacement of the Se donor and the π-system and is calculated as the distance between the aromatic plane centroid and the orthogonal projection of Se on the aromatic plane.

b Parameters of Ch⋯π highlighted in magenta in Figure [8].

Furthermore, applying the cutoffs 150°<A1<180° and 0°<A4<40°, the average d-distances (i.e., Se⋯centroid) are 4.12 ± 0.35 Å and 3.73 ± 0.19 Å for 6- and 5-atom aromatic rings, respectively. These d-values can be paralleled with the d distances reported for cation⋯π interactions (i.e., 3.4 Å in small molecule structures or 4.1–4.3 in protein structures).[54] These cutoffs further correspond to dHZ values between 0.12 and 4.66 Å (97% dHZ values between 0.12 and 3.6 Å). Such dHZ values can also be put in parallel with the offsets observed for cation⋯π interactions (i.e., 0–3.6 Å).[55]

Parameter A4 appears to be more relevant than A3 for discriminating structures with and without C-Se⋯π ChB. Although there seems to be a correlation between these two parameters (Figure [9]A, B), A3 is less straightforward to use in the analysis because it lacks information about the direction of the C–Se bond (unlike the A4 angle). Indeed, for A3 close to 90°, a large variety of A4 angles is possible and therefore the criterion A3 = ~90° is not sufficient. For example, if only A3 = ~90° is used, or even A3 = ~90° and 150°<A1<180°, the short Se⋯π contacts shown in Figure [9] C, D would have been considered as Se⋯π ChB.

Zoom Image
Figure 9 Top: Plot of A3 vs A4 angle (A and B: Se⋯π short contacts with 6-atom and 5-atom aromatic rings, respectively). Bottom: Se⋯π short contacts not relevant for ChB despite A3 being close to 90°. C: Structure with CSD refcode AHUNIB (A4 = 60.7°, A1 = 155.8°, A3 = 88.6°) and D: structure with CSD refcode IZUNIA (A4 = 62.5°, A1 = 153.0°, A3 = 87.4°).

When analyzing the plots of A3 in function of A4, we note that the entries with an A3 angle significantly lower than 90° all correspond to high A4 angles (close to 90°). Thus, all the CSD entries plotted in the lower right corner of the plot of A3 versus A4 do not correspond to chalcogen bonds (since the plane passing through the aromatic ring and that passing through C1–Se–C2 are almost parallel). Furthermore, when analyzing this graph, it must be noted that the angle A4 can be defined in 2 ways (angle between the C1–Se bond and the normal to the plane of the aromatic ring or angle between the C2–Se bond and this same normal). In fact, these two definitions are the corollary of the definitions of angles A1 and A2 (C1–Se⋯Y and C2–Se⋯Y). To be sure not to miss structures where the short contacts correspond to a chalcogen bond (depending on the definition of the C1 and C2 atoms during the search), all the CSD searches performed in this work have been done with these two possible definitions for the A4 angle. In consequence, for each A3 angle, there are two A4 values and structures in the upper right corner of the graph of A3 versus A4 (i.e., structures with A3 and A4 close to 90°) do not necessarily correspond to outliers. Indeed, given that the angles A (C1–Se–C2) have values between 110 and 67°, if the angle A4 (C1–Se-normal to the aromatic ring plane) is close to 90°, then it is that the C2–Se bond forms, with the normal to the aromatic plane, an angle close to 0°. And in this case, the chalcogen bond involves C2–Se⋯π instead of C1–Se⋯π.


#
#

Methods

Search of the Cambridge Structural Database

Crystal structures were searched systematically in the Cambridge Structural Database (CSD) restricting the search to entries in the CSD whose intermolecular distance d (Se⋯X) is less than the sum of the van der Waals radii. For C–Se⋯π ChB, we limited the search to structures for which the intermolecular distance dZ (i.e., Se⋯closest atom of the aromatic ring) is shorter than the sum of the van der Waals radii + 0.5 Å.[56] Search was also restricted to non-disordered single crystal structures of organic compounds. The 3D parameters defined in ConQuest[46] are described in the main manuscript and illustrated in Figure [2] and Figure S1 in SI.

For classical ChB (i.e., C–Se⋯X, with X = N, O, S, Se, Te), a normalized contact parameter (Nc) was also included and is calculated as the quotient obtained by dividing the ChB distance, d, by the sum of the van der Waals radii of the donor, Se, and acceptor, X, atoms.[49] Standard van der Waals radii were taken from the article of Bondi.[57]

For C–Se⋯π interactions, we limited the search to 5- or 6-atom aromatic rings, defining the ring atoms as C or N (6-atom rings) or as C, N, O, or S (5-atom rings). The horizontal displacement parameter (dHZ, distance between the aromatic plane centroid and the orthogonal projection of Se on the aromatic plane) has been calculated.

Influence of the number of atoms covalently attached to the selenium chalcogen bond donor (noted Nb) has been analyzed by defining the ‘number of bonded atoms’ to either ‘undefined’ or fixing it to ‘2’.


#
#

Conclusion

We cite here from the preface of Alain Krief’s book on organoselenium chemistry: ‘organoselenium chemistry has undergone a spectacular mutation, from an exotic area of science practiced by a few scientists it became a relatively well mastered and widely used methodology of synthetic organic chemistry’.[58] In addition to its versatile reactivity, the selenium atom is also a key player in the field of non-covalent interactions and acts as donor atom in chalcogen bonds with a variety of electron rich acceptors. Chalcogen bonds involving selenium have been observed in proteins as well as in organoselenium compounds. A better characterization of this non-covalent interaction is crucial in different fields of chemistry and biology and, in particular, for further crystal engineering of future supramolecular materials.

In the present contribution, simple descriptors (distances and angles) were identified and statistical analysis of crystal structures of organoselenium compounds allowed to define criteria to distinguish chalcogen bonds from other non-specific interactions. Combination of the Se⋯X distance and C–Se⋯X angle proved efficient in retrieving strong ChB between selenium and N, O, S, Se, or Te containing acceptors. Position of the X atom in the plane of the C–Se–C moiety is also essential to discriminate ChB from other close contacts and can equally be assessed by determining the T torsion angle that should be close to 0° or the dXP distance expected to be close to 0 Å. Further studies to include other acceptors, including halogen atoms, are welcome to verify how general the descriptors retained here are informative. Our work also points to less studied chalcogen bonds between selenium and aromatic π-systems. In order to discriminate between parallel π-π, non ChB contacts, and T-shaped Seπ chalcogen bonds, the A4 descriptor, combined with the C–Se⋯π angle, proved interesting.


#
#

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgment

J.W. thanks A. Krief who inspired him and motivated the study of Se containing molecules. Access to the PTCI (UNamur) scientific platform and the CECI is acknowledged.

Supporting Information


Corresponding Author

Johan Wouters
Department of Chemistry, University of Namur (UNamur)
61 Rue de Bruxelles, 5000 Namur
Belgium   

Publication History

Received: 05 July 2022

Accepted: 09 August 2022

Accepted Manuscript online:
09 August 2022

Article published online:
13 September 2022

© 2022. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


Zoom Image
Figure 1 Two perpendicular views of IsoStar 3D scatterplots characterizing the methylseleno functional group (in stick) in interaction with C (cross, grey), N (blue spheres), O (red spheres), S (yellow spheres), or Se (green spheres) extracted from the CSD (A) or the PDB (B). Selection of the contacts is made through definition of a van der Waals distance range (fixing the longer contact at maximum distance equals to the sum of van der Waals radii).
Zoom Image
Figure 2 Selected parameters for (A) analysis of Se⋯X and (B) Se⋯aromatic short contacts. Y: Centroid of aromatic ring; A: Angle C1–Se–C2; A1: angle C1–Se–X or C1–Se–Y; A2: Angle C2–Se–X or C2–Se–Y; d: distance Se-X or Se-Y; dZ: distance Se to the closest atom of the aromatic ring; dXP: distance X to the plane passing through C1–Se–C2; A3: angle between the plane passing through the aromatic ring and the one passing through C1–Se–C2; and A4: angle between the C–Se bond vector and the normal to the aromatic plane.
Zoom Image
Figure 3 Values of dXP (Å) (blue) and C–Se···X angle (°) (grey) in CSD retrieved entries for A, X = N; B, X = O; C, X = S; D, X=Se and E, X = Te. Black lines are indicative of a C–Se⋯X angle value of 150° (dotted line) and a dXP value of 1 Å (straight line). To facilitate trend analysis CSD entries have been sorted in descending order of C–Se⋯X value.
Zoom Image
Figure 4 A: Short Se⋯Se contacts in structure of 2,3,6,7-tetramethoxyselenanthrene radical cation triiodide (CSD refcode: RIKZIU) do not correspond to ChB but are rather the consequence of π-stacking. B: Observed interactions in structure 1,3-selenazolidine-2,4-dione (CSD refcode: BOJCOS), carbon atoms in light pink highlight a chain stabilized through ChB and H-bonds.
Zoom Image
Figure 5 Polar scatter plot of the torsion [T(°)] values and of the Se⋯X distance [d (Å)]
Zoom Image
Figure 6 Chalcogen bonds in structures of A, MIVYIB; B, MIVZAU; and C, ECITEP. Inter- and intra-molecular ChB are shown as blue and magenta dased lines respectively.
Zoom Image
Figure 7 Plots of A4 against A1 angles (°) for A: 6-atom aromatic ring (atoms in ring = C only or C and N) and B: 5-atoms aromatic ring (atoms in ring = C and N, O or S). Green frames highlight CSD entries fulfilling the geometric criteria for chalcogen bond.
Zoom Image
Figure 8 C–Se⋯π ChB (blue and magenta dashed lines) in structures of CSD refcode. A: NUBHUQ; B: NAHROH; C: VADVON, and D: ROYXEK.
Zoom Image
Figure 9 Top: Plot of A3 vs A4 angle (A and B: Se⋯π short contacts with 6-atom and 5-atom aromatic rings, respectively). Bottom: Se⋯π short contacts not relevant for ChB despite A3 being close to 90°. C: Structure with CSD refcode AHUNIB (A4 = 60.7°, A1 = 155.8°, A3 = 88.6°) and D: structure with CSD refcode IZUNIA (A4 = 62.5°, A1 = 153.0°, A3 = 87.4°).