Electrostatic Potential Differences and Halogen-Bond Selectivity

Article pubs.acs.org/crystal

Electrostatic Potential Differences and Halogen-Bond Selectivity Christer B. Aakeröy,*,† Tharanga K. Wijethunga,† John Desper,† and Marijana Đaković‡ †

Department of Chemistry, Kansas State University, Manhattan, Kansas 66506, United States Department of Chemistry, Faculty of Science, University of Zagreb, Horvatovac 102a, HR-10000 Zagreb, Croatia

‡

S Supporting Information *

ABSTRACT: Molecular electrostatic potential based guidelines for selectivity of halogen-bond interactions were explored via systematic co-crystallizations of 9 perfluorinated halogen-bond donors and 12 ditopic acceptors presenting two binding sites with different electrostatic potentials. A total of 89 of the 108 reactions resulted in co-crystal formation (as indicated by IR spectroscopy), and 35 new crystal structures were obtained. Methanol was exclusively used as a solvent for crystal growth in order to avoid any potential solvent−solute bias throughout these experiments. The structures were organized into three different groups depending upon the specific nature of the observed halogen-bond connectivities in each case. The electrostatic potential difference between the two acceptor sites on each molecule was defined as the ΔE value. Group 1 comprised acceptor molecules with a ΔE value below 35 kJ/mol units, and in this category halogen bonding took place on both binding sites in all co-crystals (9/9). Ditopic acceptor molecules in Group 2 were characterized by a ΔE value in the 35−65 kJ/mol range, and in this group half the structures showed halogen bonding to the best acceptor (11/22) and half the structures showed halogen bonding to both binding sites (11/22). In Group 3 the ΔE value was >167 kJ/mol, and in all of the co-crystals found herein (7/7), the halogen-bond donor favored the best acceptor site. These results allow us to propose some tentative guidelines and rationales for halogen-bond preferences in competitive systems. If ΔE < 35 kJ/mol, the electrostatic potential difference is not large enough to allow the donor molecules to form halogen bonds of sufficiently different thermodynamic strength to result in any pronounced molecular recognition preference (typically both, or several acceptors are then engaged in halogen bonding). Upon the basis of data produced in this study, in combination with relevant structures from the Cambridge Structural Database, it seems reasonable to suggest that if the ΔE value between two geometrically accessible halogen-bond acceptor sites is greater than 75 kJ/mol, the thermodynamic advantage of forming halogen bonds to the best acceptor provides a strong enough driving force that the best donor consistently interacts with the best acceptor; intermolecular selectivity is the result. However, if the ΔE resides between these two proposed boundaries, the outcome is unpredictable, and other factors are then likely to be responsible for the path that a particular supramolecular reaction will follow.

■

INTRODUCTION

supramolecular chemistry are still at an early stage of realization.8 Fundamental work on halogen bonding spans a wide range including studies of electrostatics,9 geometric aspects,10 activation processes for halogen-bond donors,11 design of halogen-bond donors and acceptors,12 hierarchy of halogen bonding13 and the nature of halogen bonding in gas phase,14 solution phase15 and in the solid state.16 Furthermore, studies on controlling the halogen bonding in supramolecular architectures have been a focal point in some cases.17 Apart from these fundamental studies, halogen bonding has received attention from material chemists due to applications in supramolecular gels,18 nanoparticle self-assembly,19 liquid crystals,20 optical materials,21 separation processes,22 stabiliza-

The effective design of predictable and desired supramolecular architectures requires a clear understanding of the balance between intermolecular interactions.1 In this context, numerous studies on hydrogen bonding, the most well-known among noncovalent interactions, have been performed.2 Halogen bonding,3 a relatively recent addition to the tool box of supramolecular chemistry, has not yet received the same amount of attention. Halogen bonding is often viewed as a primarily (but not exclusively) electrostatic interaction between the electropositive σ hole4 of an “activated” halogen atom and an electronegative region of another atom or molecule.5 Even though the concept of halogen bonding dates back more than a century,6 the ability of halogen atoms to function as effective and reliable sites for structure directing molecular recognition processes remained largely underexplored until the 1990s.7 Thus, halogen bonding (XB) and its potentials in the field of © XXXX American Chemical Society

Received: December 14, 2015 Revised: March 10, 2016

A

DOI: 10.1021/acs.cgd.5b01770 Cryst. Growth Des. XXXX, XXX, XXX−XXX

Crystal Growth & Design

Article

Figure 1. Three possible outcome of co-crystallizations between asymmetric ditopic acceptors and halogen-bond donors.

Figure 2. Calculated MEP values on each acceptor site and the corresponding ΔE values for the 12 acceptors are given in kJ/mol.

tion processes,23 and electric and magnetic materials.24 In addition halogen bonding in biological systems25 is attracting much attention, and medicinal chemistry, drug discovery,26 and transmembrane anion transport also represent areas where this interaction may be of considerable practical benefit.27 Etter’s rules offer useful guidelines regarding selectivity28 and binding preferences for hydrogen-bond interactions,29 but a similar framework for preferred interactions among halogen bonds has yet to be fully realized. In order to determine if hydrogen- and halogen-bond preferences could both be described and rationalized in similar ways, we previously analyzed a possible molecular electrostatic potential (MEP) dependent selectivity of halogen bonding. Molecules decorated with a single XB donor were introduced to molecules carrying two different acceptor sites, and the results suggested that the halogen-bond donor favors the better acceptor (as ranked by MEP surface calculations), whereas the second best acceptor site did not participate in any notable structure directing interactions.30,31 Upon closer examination it became evident that no selectivity could be identified if the difference in MEP values for two competing sites was too low. Further support for this suggestion was obtained through a Cambridge Structural Database (CSD)32 survey conducted on halogen-bonded

systems comprising molecules with two different acceptor moieties. Although the acceptors sites could be readily ranked according to their calculated MEP values, the observed halogen-bond interactions did not always follow the expected or intended connectivities.33 In order to more clearly delineate when electrostatically based halogen-bond preferences can be relied upon, from a practical synthetic point of view, we need much more relevant and custom-designed structural data. Given a system that comprises a molecule with two different halogen-bond acceptor sites and a single-point halogen-bond donor, we first postulate that if a co-crystal is formed, then there are essentially three possible structural/geometric outcomes that may all be determined by the relative difference in MEP between the two binding sites. MEP between the two binding sites, Figure 1. The working hypothesis is that if the MEP difference is greater than some cutoff value, halogen bonding will favor the best acceptor, and if the difference is below a certain value, both acceptor sites will be engaged in halogen bonding. There may also be a more diffuse region (gray area) where no statistically significant trend can be identified. To test our hypothesis and, ideally, to establish a set of boundaries that allow us to define the synthetic space where coB



Article

produced a crystal suitable for single-crystal X-ray diffraction (13 of which we have reported previously).30 For the remaining 22 structures, X-ray experimental data and crystallographic data are given in the SI, and all halogen-bond geometries are provided in Table 2.

crystals with specific metrics and topologies can be reliably synthesized, we have carried out an extensive systematic study of halogen-bonded co-crystals using a combination of 12 ditopic acceptors (each presenting two binding sites with different MEP values) and 9 halogen-bond donors. We selected the ditopic acceptors with as broad a spread of MEP differences as possible, Figure 2. Probing and determining halogen-bond selectivity was achieved utilizing a systematic co-crystallization approach. Nine iodine-based perfluorinated halogen-bond donors were selected (including both aromatic and aliphatic donors), Figure 3. Co-crystallization experiments between the 12 acceptors and

■

RESULTS MEP values for the acceptor molecules are provided in Figure 2. For each acceptor molecule the MEP difference, ΔE, between the two sites was calculated as indicated below, and the relevant data are summarized in Figure 2. Table 1 summarizes the calculated MEPs values for the donors under study. ΔE value = |MEP value on the best acceptor| − |MEP value on the second best acceptor|

Table 1. Calculated MEP Values on Halogen-Bond Donors D1−D9

Figure 3. Halogen-bond donors in this study.

9 donors were conducted (108 reactions), and 35 of these reactions produced crystals suitable for single-crystal X-ray diffraction.

■

EXPERIMENTAL SECTION

molecule

DFT (kJ/mol)

1,2-diiodotetrafluoroethane, D1 1,4-diiodooctafluorobutane, D2 1,6-diiodoperfluorohexane, D3 1,8-diiodoperfluorooctane, D4 iodopentafluorobenzene, D5 1,2-diiodotetrafluorobenzene, D6 1,4-diiotetrafluorobenzene, D7 1,3,5-triiodotrifluorobenzene, D8 4,4′-diiodoperfluorobiphenyl, D9

+163 +168 +169 +169 +166 +162 +169 +158 +164

The IR analysis for identifying co-crystal formation focused on the C−F stretches of the halogen-bond donor. A shift of three wave numbers or more was considered to be significant and indicative of a positive result, i.e., co-crystal formation (the subsequent single-crystal X-ray diffraction analyses confirmed that the assignment based on IR data was correct). The results are summarized in the SI along with the success rate for each acceptor and donor; 89 of 108 reactions yielded co-crystals corresponding to a supramolecular yield of 82%. The IR analysis allows us to identify a successful co-crystal synthesis, but it does not provide unambiguous details about specific interaction sites which is crucial for determining the binding preferences for the different halogen-bond donors and, thus, single crystal data were required. In order to avoid any potential solvent−solute bias throughout these experiments we opted to consistently use methanol for all crystallizations (∼2 mL) and slow evaporation as the crystal growing technique. The 22 new structures are presented in order of increasing ΔE value of the ditopic acceptor with all halogen-bond geometries listed in Table 2. With A1 (ΔE = 22 kJ/mol) two crystal structures were obtained (A1:D7 and A1:D9), and in each case both binding sites imidazole nitrogen (N(im)) and pyridine nitrogen (N(py)) engage in XB formation, Figure 4. No interaction preference is observed. Two structures (A2:D2 and A2:D7) were obtained with A2 (ΔE = 26 kJ/mol), and again both binding sites form halogen bonds, Figure 5. A3 (ΔE = 31 kJ/mol) produced three structures (A3:D3, A3:D7, and A3:D9) with all acceptor moieties participating in halogen bonding, Figure 6.

All precursors, solvents, donors, and acceptor A11 were purchased from commercial sources and used without further purification. A9 and A10 were synthesized as reported in the literature.34 A1−A8 were synthesized following a modified procedure of a literature report.35 A1236 and D937 were synthesized following literature procedures. Synthetic procedures and characterization of all the acceptors are provided in the Supporting Information (SI). Infrared spectra were recorded with a Nicolet 380 FT-IR. 1H NMR spectra were recorded on Varian Unity plus 400 MHz spectrometer. Electrostatic Potential Calculations. Molecular electrostatic potentials were calculated with density functional B3LYP level of theory with a 6-311++G** basis set in a vacuum. Calculations were carried out using Spartan 8 software.38 Molecules were geometry optimized with the maxima and minima in the electrostatic potential surface (0.002 e/au isosurface) determined using a positive point charge in the vacuum as a probe. The numbers indicate the interaction energy (kJ/mol) between a positive point probe and surface of molecule at that particular point. A positive value for the interaction energy indicates a positive surface potential, while a negative value indicates a negative surface potential. Co-Crystal Screening and Crystallography. The initial cocrystal screening was carried out via solvent-assisted grinding, whereby each acceptor and donor was mixed together in stoichiometric ratios (1:2 acceptor/donor ratio with D5 and 1:1 stoichiometry with the other donors) and ground with a mortar and pestle with the assistance of a drop of methanol (a total of 108 experiments). In each experiment, about 10 mg of the acceptor was used with the appropriate amount of donor, and once the solvent had evaporated, the ground mixture was analyzed using IR spectroscopy to determine if a co-crystal had formed, see SI. The solids from the grinding experiments were then dissolved in a minimum amount (∼2 mL) of methanol and left in a vial for slow evaporation in order to obtain crystals suitable for single crystal X-ray diffraction; 35 experiments C



Article

Table 2. Halogen-Bond Geometriesa co-crystal

C−I···N

I···N (Å)

C−I ···N (°)

A1:D7

C(31)−I(1)···N(13) C(41)−I(2)···N(21) C(31)−I(1)···N(13) C(41)−I(2)···N(21)#1 C(41A)−I(1)···N(13) C(41B)−I(1)···N(13) C(44A)−I(2A)···N(31)#1 C(44B)−I(2B)···N(31)#1 C(41)−I(1)···N(13) C(51)−I(2)···N(31) C(10)−I(1)···N(3) C(15)−I(2)···N(1)#1 C(31)I(1)···N(13) C(34)−I(2)···N(21)#1 C(31)−I(1)···N(13) C(41)−I(2)···N(21)#1 C311−I11···N131 C341−I21···N211#1 C312−I32···N132 C(31)−I(1)···N(13) C(41)−I(4)···N(21) C(41)−I(1)···N(13) C(41)−I(1)···N(13) C(43)−I(2)···N(21)#1 C(41A)−I(1)···N(13) C(41B)−I(1)···N(13) C(51)−I(2)···N(31) C(41)−I(1)···N(13) C(51)−I(3)···N(31) C(41)−I(1)···N(13) C(51)−I(2)···N(31) C(41)−I(1)···N(13) C(51)−I(4)···N(31) C(41)−I(1)···N(13) C(26)−N(21)···I(2)#1 C(31)−I(1)···N(13) C(34)−I(2)···N(21)#1 C(31)−I(1)···N(13) C(34)−I(2)···N(21)#1 C(6)−I(6)···N(11) C(8)−I(8)···N(21) C211−I11···N111 C212−I22···N112 C(21)−I(1)···N(11) C(26)−I(2)···N(14)#1 C(31)−I(1)···N(11) C(32)−I(2)···N(21)

2.782(2) 2.762(3) 2.9658(16) 2.8379(16) 2.838(3) 2.838(3) 2.769(3) 2.868(3) 2.813(2) 2.764(2) 2.854(6) 2.866(5) 2.8336(16) 2.8149(15) 2.766(3) 2.768(3) 2.8096(19) 2.8502(19) 2.7891(18) 2.807(2) 2.820(2) 2.7601(9) 2.852(2) 2.889(2) 2.924(3) 2.924(3) 2.982(3) 2.793(2) 3.016(2) 2.8071(14) 2.8788(16) 2.965(3) 2.788(3) 2.814(2) 2.8063(19) 2.757(3) 2.815(2) 2.8357(17) 2.8688(18) 2.664(4) 2.689(4) 2.641(3) 2.668(3) 2.683(4) 3.440(5) 2.722(2) 2.776(2)

168.27(10) 177.29(10) 166.92(6) 177.30(6) 169.5(2) 174.6(3) 169.0(3) 175.6(2) 179.33(8) 176.80(8) 174.9(2) 174.0(2) 170.80(6) 174.00(6) 174.21(8) 177.05(9) 176.85(8) 170.39(7) 177.27(7) 175.70(11) 177.93(10) 176.67(4) 169.65(7) 172.31(8) 172.7(2) 171.5(5) 169.96(13) 177.63(9) 170.94(9) 177.59(5) 174.16(6) 165.81(10) 176.44(13) 171.73(7) 125.91(15) 178.49(10) 2.815(2) 171.67(7) 172.76(7) 176.48(15) 172.01(16) 178.10(11) 178.93(11) 176.4(2) 164.42(18) 178.30(10) 173.43(9)

A1:D9 A2:D2

A2:D7 A3:D3 A3:D7 A3:D9 A4:D7

A4:D8 A5:D7 A5:D8 A6:D4

A6:D6 A6:D7 A6:D8 A7:D8 A8:D2 A8:D7 A11:D1 A11:D2 A11:D3 A11:D6

Figure 4. Halogen bonds in the structures of (a) A1:D7 and (b) A1:D9; both acceptor sites are involved.

Figure 5. Halogen bonds in the crystal structures of (a) A2:D2 and (b) A2:D7.

Figure 6. Halogen bonds in the structures of (a) A3:D3, (b) A3:D7, and (c) A3:D9.

A1:D9: #1 x − 1, y − 1, z − 1; A2:D2: #1 x − 2, y, z + 1; A3:D3: #1 3/2 + x, 3/2 − y, z − 1/2; A3:D7: #1 x, y − 1, z; A3:D9: #1 x, y − 1, z; A4:D7: #1 −x + 1, −y, −z + 2; A5:D8: #1 x + 1, y + 1, z − 1; A7:D8: #1 −x, −y + 1, −z; A8:D2: #1 x + 1, y, z + 2; A8:D7: #1 x + 1, y − 1, z − 1; A11:D3: #1 x, −y + 1/2, z − 1/2.

Figure 7. Halogen bonds in the structures of (a) A4:D7 and (b) A4:D8.

a

C−H moiety of a nearby ring. A5 is inconsistent, since both acceptors in the structure of A5:D8 form halogen bonds, Figure 8. Four structures (A6:D4, A6:D6, A6:D7, and A6:D8) were obtained with A6 (ΔE = 37 kJ/mol), and again all acceptors were involved in halogen bonding, Figure 9. A7 (ΔE = 38 kJ/mol) gave one structure (A7:D8), Figure 10, and A8 (ΔE = 46 kJ/mol) gave two structures (A8:D2 and A8:D7), Figure 11, and in all of them every acceptor site was engaged in a halogen bond.

In the two structures (A4:D7 and A4:D8) of A4 (ΔE = 33 kJ/mol), all acceptors sites were again involved in halogen bonding, Figure 7. In contrast, the two structures of A5 (ΔE = 36 kJ/mol) (A5:D7 and A5:D8) behave differently. A5:D7 shows a halogen bond to the best acceptor, imidazole, while the pyridine nitrogen atom acts as a hydrogen-bond acceptor for a D



Article

Four structures (A11:D1, A11:D2, A11:D3, and A11:D6) were obtained with A11 (ΔE = 167 kJ/mol), and all of them show binding to only the best acceptor site, while in A11:D3, there is an additional short contact formed with the second best acceptor site, Figure 12. Even though the grinding results

Figure 8. Halogen bonds in the structures of (a) A5:D7 − only the best acceptor, N(im) participates; (b) A5:D8 − both acceptor sites participate.

Figure 12. Halogen bonds in the structures of (a) A11:D1, (b) A11:D2, (c) A11:D3, and (d) A11:D6.

showed a 67% success in forming co-crystals with A12 (ΔE = 175 kJ/mol), we were unable to grow crystals suitable for single-crystal X-ray diffraction of any co-crystal of this ligand.

■

DISCUSSION In order to complement the structural data provided by the 35 data points gathered in this manuscript, a CSD search was carried out to find any additional relevant structural information. The search yielded three crystal structures, A11:D5,39 A11:D7,40 and A11:D840 with all of them presenting a halogen bond to the best acceptor site only. These three structures will be included in the general discussion. By correlating the ΔE values with the specific molecular recognition events that took place in the 38 structures included in this study, it is possible to group them into three general categories, Figure 13. Group 1 contains nine structures with acceptors A1−A4. In each case, for all four ditopic molecules, both available acceptor sites are engaged in a halogen bond. The ΔE is below 35 kJ/ mol for A1−A4, which indicates that an electrostatic potential difference of this magnitude or below is not sufficient to induce selectivity as both acceptors are equally competitive. Group 2 contains 22 structures with acceptors A5−A10. In 11 of these structures, both acceptor sites form halogen bonds, and in the remaining 11 structures only one halogen bond is formed and it is exclusively to the best acceptor sites (as ranked by calculated MEP values). It is, at least at this point, unclear as to why half the structures contain one, and half the structures contain two, halogen bonds, and, consequently, we refer to this space where the ΔE values for the acceptor fall in the 35−65 kJ/mol range as a gray area where the outcome is unpredictable. However, it should be noted that the two molecules with highest ΔE values, A9 (ΔE = 55 kJ/mol), and A10 (ΔE = 64 kJ/mol), respectively, contribute 10 of the 11 structures in this group that contain only one halogen bond (and in all those cases it involves the best acceptor site). With A9, four of six structures, and with A10 six of seven structures show electrostatically controlled halogen-bond selectivity. Group 3 contains seven structures, all of them with ditopic acceptor A11 (ΔE = 167 kJ/mol). In all seven structures, the acceptor site with the higher electrostatic potential is preferred by the halogen-bond donor, whereas the second-best acceptor is not engaged in any short N···I contacts.

Figure 9. Halogen bonds in the structures of (a) A6:D4, (b) A6:D6, (c) A6:D7, and (d) A6:D8.

Figure 10. Primary halogen bond interactions in A7:D8.

Figure 11. Halogen bonds in the structures of (a) A8:D2 and (b) A8:D7.

Six structures (A9:D2,30 A9:D3,30 A9:D6,31 A9:D7,31 A9:D8,31 and A9:D931) with A9 (ΔE = 55 kJ/mol) and seven structures (A10:D1,30 A10:D2,30 A9:D3,30 A9:D4,30 A9:D6,31 A9:D7,31 and A9:D931) with A10 (ΔE = 64 kJ/mol) have previously been reported. In 10 of these structures only the best acceptor was engaged, while in the remaining 3 (A9:D8, A9:D9, and A10:D6) both sites formed halogen bonds. E



Article

Figure 13. Classification of structures based on ΔE values and halogen-bond preferences.

Figure 14. Proposed assignments.

We can now begin to summarize our results against the postulated outcomes that were outlined at the beginning of this manuscript in Figure 1. As there is a considerable gap between A9/A10 and A11 in terms of ΔE values, 55, 64, and 167 kJ/ mol, respectively, it is difficult to say exactly where the boundary between Groups 2 and 3 should be drawn. It is worth keeping in mind that any proposed boundary is not going to offer an unambiguous cutoff point since we are dealing with reversible and relatively weak interactions. However, we feel that there are substantial benefits to suggesting a ΔE that can serve as a practical guideline for co-crystal synthesis involving multitopic molecules; we expect that the proposed boundaries in this analysis can and will be modified as we gain access to more structural data, especially on molecules that carry ΔE in the range between 70 and 160 kJ/mol. Although it can be argued, based on existing data, that the border between Groups 2 and 3 can be anywhere between 65 and 166 kJ/mol, we believe that since 66% of structures with A9 (ΔE = 55 kJ/mol) and 86% of structures with A10 (ΔE = 64 kJ/mol) are selective toward the best acceptor, a tentative value of 75 kJ/mol as an estimate of a boundary between Groups 2 and 3 represents a reasonable starting point for further refinements. In the next

phase of this investigation we will focus our attention on the synthesis of appropriate co-crystals of ditopic acceptors that display ΔE values around the 65−100 kJ/mol range in order to get key structural for more refined boundaries, Figure 14. It is worth pointing out that the possible influence of the relative electrostatic potentials and molecular geometries of the halogen-bond donors D1−D9 on the structural outcome of these co-crystallizations is expected to be insignificant. All of them rely on an iodine atom that has been “activated” (its polarizability has been enhanced) through fluorine substitutions, and the electrostatic potentials for D1−D9 are very similar, Table 1, and none of them are sterically hindered from participating in short N···I contacts. As mentioned previously, different solvents will play a role in the way in which solute molecules form assemblies and aggregates in solution, but we deliberately employed the same solvent (methanol) in all crystallizations so the solvent−solute effect for all crystallizations would be same. Of course in a different solvent system the interaction energies will change, and, as a result, the ΔE values presented in Figure 14 may change. Finally, we wanted to test how well these proposed guidelines for halogen-bond preferences would work for F



Article

Table 3. Results Comparison for CSD Reported Structuresa electrostatic potentials (kJ/mol) donor

CSD code

best acceptor

second best acceptor

ΔE value (kJ/mol)

group assignment

D1

ULOKOV ULOLAI BEWXOS BEWXIM COKNOG COKNUM ECASAE COGKAM DIVCIV KABLAC JAQMEU LUKMIN PEFPAT QOLJIK TOJBUQ TOJCAX VABNUJ -

−160 −164 −175 −175 −171 −171 −182

−41 −152 −154 −154 −58 −58 −150

119 12 21 21 113 113 32

3 1 1 1 3 3 1

yes yes yes yes yes yes no (best acceptor)

−164 −152 −174 −191 −137 −212 −142 −178 −195 −175

−30 −108 −30 −145 −102 −57 −106 −174 −155 −149

134 44 144 46 53 155 36 4 40 26

3 2 3 2 2 3 2 1 2 1

yes no preference yes best acceptor best acceptor yes best acceptor no (best acceptor) no preference no (second best acceptor)

D2 D3 D4 D5 D6 D7

D8 D9 a

correct structural outcome according to assignment

No crystal structures were obtained with D5, D6, D8, or D9.

The remaining five structures contain acceptors with ΔE values in the range of 36−53 kJ/mol, which places them in Group 2, the “gray zone”, where we do not expect to find any consistent patterns of molecular recognition events. Two structures show multiple halogen bonds, three structures show a preference for the best acceptor, and this mixed outcome is consistent with our results summarized in Figure 13 (50% no selectivity; 50% best donor···best acceptor). This group clearly does merit a “gray area” descriptor because it is obviously not possible to a priori hypothesize what the likely intermolecular outcome is going to be in molecular solids involving multitopic acceptor molecules with this range of ΔE values. All in all, only two of the 17 test structures provided outliers with respect to the proposed scheme, assignment scheme, Figure 14, which means that 88% of the test structures (15/17) displayed halogen-bond interactions that were consistent with our proposed guidelines based upon an original set of 38 crystal structures.

predicting the structural outcome in halogen-bonded systems where different interaction outcomes are possible. We focused our search on co-crystals mined from the CSD that contained D1−D9 as donor molecules, and we were able to find 17 relevant crystal structures comprising acceptor molecules decorated with two or more acceptor sites with different electrostatic potentials. The group included ditopic as well as multitopic acceptors and the acceptor atom could be either a nitrogen, oxygen, or sulfur atom. Density functional theory (DFT) calculations were performed on all 17 acceptors to establish the relative ranking of the binding sites as well as for assigning the structures to one of the Groups 1−3, Table 3. Six of the co-crystals were found to belong to Group 3 since the ΔE values for these acceptors (113−155 kJ/mol) were significantly above the cutoff between Groups 2 and 3, and according to our proposed scheme for assigning/predicting the structural outcome, halogen-bond interactions would be expected to be selective for the best acceptor. We were pleased to note that in all six cases, only halogen bonds to the best acceptor (as ranked by calculated MEP values) were observed. Another six co-crystals belonged to Group 1 as they contained acceptors with ΔE values in the range of 4−32 kJ/ mol. Our proposed guidelines suggest that halogen-bond selectivity is unlikely and that the “best donor···best acceptor” outcome cannot be reliably expected. An examination of these six structures shows that three of them contain interactions with multiple acceptor sites, two favor the best acceptor, and one favors the second best acceptor. Consequently, four of the six structures do not display an electrostatically driven preference for the best acceptor which offers validation for our proposed boundary between Groups 1 and 2. If the ΔE between two different acceptor sites is below a certain value (currently standing at 35 kJ/mol) there is no longer a sufficient thermodynamic advantage for the formation of a co-crystal driven primarily by a halogen bond to the best acceptor.

■

CONCLUSIONS In order to explore if halogen-bond interactions display specific and predictable pattern preferences in competitive supramolecular systems, we have developed a few simple electrostatically based guidelines. The foundation for these guidelines was the result of a systematic co-crystallization study involving 9 halogen-bond donors and 12 asymmetric ditopic acceptors (108 experiments) which produced 35 crystal structures (and additional three relevant structures were mined from the CSD). The differences in calculated molecular electrostatic potential values between acceptor sites on the same molecule were correlated with the observed halogen-bond interactions which essentially led to a three-group classification scheme. If the ΔE values are less than 35 kJ/mol, the electrostatic potential difference is not large enough to allow the donor molecules to form halogen bonds of sufficiently different thermodynamic G



Article

(3) Metrangolo, P.; Meyer, F.; Pilati, T.; Resnati, G.; Terraneo, G. Angew. Chem., Int. Ed. 2008, 47, 6114−6127. Aakeröy, C. B.; Wijethunga, T. K.; Haj, M. A.; Desper, J.; Moore, C. CrystEngComm 2014, 16, 7218−7225. (4) Politzer, P.; Murray, J. S.; Clark, T. Phys. Chem. Chem. Phys. 2013, 15, 11178−11189. Clark, T.; Hennemann, M.; Murray, J. S.; Politzer, P. J. Mol. Model. 2007, 13, 291−296. (5) Desiraju, G. R.; Ho, S.; Kloo, L.; Legon, A. C.; Marquardt, R.; Metrangolo, P.; Politzer, P. A.; Resnati, G.; Rissanen, K. Pure Appl. Chem. 2013, 85, 1711−1713. (6) Colin, M. Ann. Chim. 1814, 91, 252−272. Guthrie, F. J. Chem. Soc. 1863, 16, 239−244. (7) Metrangolo, P.; Resnati, G. Cryst. Growth Des. 2012, 12, 5835− 5838. (8) Gilday, L. C.; Robinson, S. W.; Barendt, T. A.; Langton, M. J.; Mullaney, B. R.; Beer, P. D. Chem. Rev. 2015, 115, 7118−7195. (9) Auffinger, P.; Hays, F. A.; Westhof, E.; Ho, P. S. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 16789−16794. Brinck, T.; Murray, J. S.; Politzer, P. Int. J. Quantum Chem. 1992, 44, 57−64. Politzer, P.; Lane, P.; Concha, M.; Ma, Y.; Murray, J. J. Mol. Model. 2007, 13, 305−311. (10) Aakeröy, C. B.; Sinha, A. S.; Chopade, P. D.; Desper, J. Dalton Trans. 2011, 40, 12160−12168. Mukherjee, A.; Tothadi, S.; Desiraju, G. R. Acc. Chem. Res. 2014, 47, 2514−2524. Pavan, M. S.; Durga Prasad, K.; Guru Row, T. N. Chem. Commun. 2013, 49, 7558−7560. (11) Fourmigué, M. Curr. Opin. Solid State Mater. Sci. 2009, 13, 36− 45. Bauzá, A.; Quiñonero, D.; Frontera, A.; Deyà, P. M. Phys. Chem. Chem. Phys. 2011, 13, 20371−20379. (12) Alkorta, I.; Sanchez-Sanz, G.; Elguero, J. CrystEngComm 2013, 15, 3178−3186. Dumele, O.; Wu, D.; Trapp, N.; Goroff, N.; Diederich, F. Org. Lett. 2014, 16, 4722−4725. Aakeröy, C. B.; Wijethunga, T. K.; Desper, J.; Đakovic, M. Cryst. Growth Des. 2015, 15, 3853−3861. (13) Aakeröy, C. B.; Baldrighi, M.; Desper, J.; Metrangolo, P.; Resnati, G. Chem. - Eur. J. 2013, 19, 16240−16247. Aakeröy, C. B.; Chopade, P. D.; Desper, J. Cryst. Growth Des. 2013, 13, 4145−4150. (14) Legon, A. C. Phys. Chem. Chem. Phys. 2010, 12, 7736−7747. Legon, A. C. Phys. Chem. Chem. Phys. 2014, 16, 12415−12421. (15) Caballero, A.; Zapata, F.; White, N. G.; Costa, P. J.; Felix, V.; Beer, P. D. Angew. Chem., Int. Ed. 2012, 51, 1876−1880. (16) Rissanen, K. CrystEngComm 2008, 10, 1107−1113. Chandran, S. K.; Thakuria, R.; Nangia, A. CrystEngComm 2008, 10, 1891−1898. Metrangolo, P.; Pilati, T.; Terraneo, G.; Biella, S.; Resnati, G. CrystEngComm 2009, 11, 1187−1196. Itoh, T.; Nomura, S.; Nakasho, S.; Uno, T.; Kubo, M.; Tohnai, N.; Miyata, M. Macromolecules 2015, 48, 5450−5455. Aakeröy, C. B.; Spartz, C. L. Top. Curr. Chem. 2014, 358, 155−182. (17) Gonnade, R. G.; Bhadbhade, M. M.; Shashidhar, M. S.; Sanki, A. K. Chem. Commun. 2005, 5870−5872. Aakeröy, C. B.; Wijethunga, T. K.; Desper, J. CrystEngComm 2014, 16, 28−31. Aakeröy, C. B.; Rajbanshi, A.; Metrangolo, P.; Resnati, G.; Parisi, M. F.; Desper, J.; Pilati, T. CrystEngComm 2012, 14, 6366−6368. (18) Meazza, L.; Foster, J. A.; Fucke, K.; Metrangolo, P.; Resnati, G.; Steed, J. W. Nat. Chem. 2012, 5, 42−47. (19) Shirman, T.; Arad, T.; van der Boom, M. E. Angew. Chem., Int. Ed. 2010, 49, 926−929. Shirman, T.; Kaminker, R.; Freeman, D.; van der Boom, M. E. ACS Nano 2011, 5, 6553−6563. (20) Bertani, R.; Metrangolo, P.; Moiana, A.; Perez, E.; Pilati, T.; Resnati, G.; Rico-Lattes, I.; Sassi, A. Adv. Mater. 2002, 14, 1197−1201. Nguyen, H. L.; Horton, P. N.; Hursthouse, M. B.; Legon, A. C.; Bruce, D. W. J. Am. Chem. Soc. 2004, 126, 16−17. (21) Saha, B. K.; Nangia, A.; Nicoud, J. F. Cryst. Growth Des. 2006, 6, 1278−1281. Cariati, E.; Forni, A.; Biella, S.; Metrangolo, P.; Meyer, F.; Resnati, G.; Righetto, S.; Tordin, E.; Ugo, R. Chem. Commun. 2007, 2590−2592. (22) Farina, A.; Meille, S. V.; Messina, M. T.; Metrangolo, P.; Resnati, G.; Vecchio, G. Angew. Chem., Int. Ed. 1999, 38, 2433−2436. Takeuchi, T.; Minato, Y.; Takase, M.; Shinmori, H. Tetrahedron Lett. 2005, 46, 9025−9027.

strength to result in any pronounced molecular recognition preference (typically both, or several acceptors are then engaged in halogen bonding). Upon the basis of data produced in this study, in combination with relevant structures from the CSD, it seems reasonable to suggest that if the ΔE value between two geometrically accessible halogen-bond acceptor sites is greater than 75 kJ/mol, the thermodynamic advantage of forming halogen bonds to the best acceptor provides a strong enough driving force that the best donor consistently interacts with the best acceptor; intermolecular selectivity is the result. However, if the ΔE resides between these two proposed boundaries, the outcome is completely unpredictable and other factors are then likely to be responsible for the direction in which a particular supramolecular reaction and subsequent nucleation and crystallization will take. Our results have produced a set of very simple guidelines that may be readily utilized when developing practical supramolecular synthetic strategies or in small-molecule drug design. Exceptions to these guidelines are inevitable, and the specific boundaries may move slightly as more relevant data become available, but this simple approach can still offer important insight of relevance to practical co-crystal synthesis and to the understanding of the balance between competing intermolecular interactions.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.cgd.5b01770. Experimental data for synthesis, IR, PXRD, DSC, and crystallographic data (PDF) Accession Codes

CCDC 1442434−1442455 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing [email protected], or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This material is based upon work supported by the U.S. Army Research Laboratory and the U.S. Army Research Office under Contract/Grant Number W911NF-13-1-0387, and by the Croatian Science Foundation under the project UIP-11-20131809.

■

REFERENCES

(1) Lehn, J. M. Chem. Soc. Rev. 2007, 36, 151−160. Benedict, J. B. Recent Advances in Crystallography; InTech: Rijeka, Croatia, 2012. Meyer, F.; Dubois, P. CrystEngComm 2013, 15, 3058−3071. Cinčić, D.; Frišcǐ č, T.; Jones, W. Chem. - Eur. J. 2008, 14, 747−753. (2) Steiner, T. Angew. Chem., Int. Ed. 2002, 41, 48−76. Głowacki, E. D.; Irimia-Vladu, M.; Bauer, S.; Sariciftci, N. S. J. Mater. Chem. B 2013, 1, 3742−3753. Sweetman, A. M.; Jarvis, S. P.; Sang, H.; Lekkas, I.; Rahe, P.; Wang, Y.; Wang, J.; Champness, N. R.; Kantorovich, L.; Moriarty, P. Nat. Commun. 2014, 5, 3931. H



Article

(23) Aakeröy, C. B.; Wijethunga, T. K.; Benton, J.; Desper, J. Chem. Commun. 2015, 51, 2425−2428. (24) Metrangolo, P.; Resnati, G. Science 2008, 321, 918−919. (25) Lu, Y.; Wang, Y.; Zhu, W. Phys. Chem. Chem. Phys. 2010, 12, 4543−4551. Voth, A. R.; Hays, F. A.; Ho, P. S. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 6188−6193. (26) Lu, Y.; Shi, T.; Wang, Y.; Yang, H.; Yan, X.; Luo, X.; Jiang, H.; Zhu, W. J. Med. Chem. 2009, 52, 2854−2862. Ibrahim, M. A. A. J. Comput. Chem. 2011, 32, 2564−2574. (27) Vargas Jentzsch, A.; Emery, D.; Mareda, J.; Metrangolo, P.; Resnati, G.; Matile, S. Angew. Chem., Int. Ed. 2011, 50, 11675−11678. Jentzsch, A. V.; Emery, D.; Mareda, J.; Nayak, S. K.; Metrangolo, P.; Resnati, G.; Sakai, N.; Matile, S. Nat. Commun. 2012, 3, 905−913. Vargas Jentzsch, A.; Matile, S. J. Am. Chem. Soc. 2013, 135, 5302− 5303. (28) Etter, M. C. Acc. Chem. Res. 1990, 23, 120−126. Aakeröy, C. B.; Beatty, A. M.; Helfrich, B. A. J. Am. Chem. Soc. 2002, 124, 14425− 14432. Aakeröy, C. B.; Salmon, D. J. CrystEngComm 2005, 7, 439− 448. Bowers, J. R.; Hopkins, G. W.; Yap, G. P. A.; Wheeler, K. A. Cryst. Growth Des. 2005, 5, 727−736. (29) Aakeröy, C. B.; Desper, J.; Salmon, D. J.; Smith, M. M. CrystEngComm 2009, 11, 439−443. Lynch, D. E.; Sandhu, P.; Parsons, S. Aust. J. Chem. 2000, 53, 383−387. Hunter, C. A. Angew. Chem., Int. Ed. 2004, 43, 5310−5324. Hunter, C. A. Angew. Chem. 2004, 116, 5424−5439. Musumeci, D.; Hunter, C. A.; Prohens, R.; Scuderi, S.; McCabe, J. F. Chem. Sci. 2011, 2, 883−890. (30) Aakeröy, C. B.; Wijethunga, T. K.; Desper, J.; Moore, C. J. Chem. Crystallogr. 2015, 45, 267−276. (31) Aakeröy, C. B.; Wijethunga, T. K.; Desper, J. J. Mol. Struct. 2014, 1072, 20−27. (32) CSD ConQuest 1.17; Cambridge Crystallographic Data Centre: Cambridge, U.K., 2014. (33) Aakeröy, C. B.; Desper, J.; Fasulo, M.; Hussain, I.; Levin, B.; Schultheiss, N. CrystEngComm 2008, 10, 1816−1821. (34) Aakeröy, C. B.; Wijethunga, T. K.; Desper, J. New J. Chem. 2015, 39, 822−828. (35) Aakeröy, C. B.; Epa, K.; Forbes, S.; Schultheiss, N.; Desper, J. Chem. - Eur. J. 2013, 19, 14998−15003. (36) Felten, A. E.; Zhu, G.; Aron, C. D. Org. Lett. 2010, 12, 1916− 1919. (37) Espallargas, G. M.; Recuenco, A.; Romero, F. M.; Brammer, L.; Libri, S. CrystEngComm 2012, 14, 6381−6383. (38) Spartan’08; Wavefunction, Inc.: Irvine, CA. (39) Präsang, C.; Whitwood, A. C.; Bruce, D. W. Cryst. Growth Des. 2009, 9, 5319−5326. (40) Roper, L. C.; Präsang, C.; Kozhevnikov, V. N.; Whitwood, A. C.; Karadakov, P. B.; Bruce, D. W. Cryst. Growth Des. 2010, 10, 3710− 3720.

I


Electrostatic Potential Differences and Halogen-Bond Selectivity

Recommend Documents