S.
Khorasani
,
D. S.
Botes
,
M. A.
Fernandes
* and
D. C.
Levendis
Molecular Sciences Institute, School of Chemistry, University of the Witwatersrand, PO Wits 2050, Johannesburg, South Africa. E-mail: Manuel.Fernandes@wits.ac.za; Fax: +27 11 7176749; Tel: +27 11 7176723
First published on 10th September 2015
Electron donor/acceptor (EDA) interactions have been found to be very useful in engineering reactive heteromolecular crystals, but few examples have been reported in the literature. By utilising EDA interactions, crystals of charge-transfer (CT) complexes were formed with bis(N-allylimino)-1,4-dithiin as the electron acceptor and 9-bromoanthracene as the electron donor. The CT complex crystallised in the monoclinic P21/n space group with the crystal structure consisting of stacks of alternating electron donor and acceptor molecules in a 1:1 ratio. These crystals are able to undergo a solid-state Diels–Alder reaction with bis(N-allylimino)-1,4-dithiin as the dienophile and 9-bromoanthracene as the diene. Examination of close contacts indicates that the diene can theoretically react with the dienophile above or below it within a stack as the reaction distances are less than 3.5 Å in both directions. A single crystal was selected and allowed to react at 30 °C, was analysed at various states of conversion by single-crystal X-ray diffraction, and was found to react by approximately 10% every 6 days, with the reaction occurring in a single direction along the CT stack axis. The solid-state reaction creates a void space which leads to a molecular conformational change within the crystal. Consequently, the single crystal started to show significant signs of deterioration after approximately 28% conversion but remained intact upon further reaction and was found to anneal as 100% conversion was approached, leading to the formation of new intermolecular interactions not present in the starting crystal. The solid-state reaction occurs topochemically when fewer than 28% or more than 80% of the molecules have reacted, with minimal motion during the reaction. In the conversion range of 28–80%, the reaction occurs in an almost topotactic manner with significant molecular motion and associated crystal deterioration.
In the original formulation of the topochemical principle, Schmidt and his co-workers5 suggested that solid-state reactions occur with minimal motion and that in solid-state photochemical reactions the atoms involved in the reaction have to be within 4.2 Å of each other; the distance and orientation criteria have since been referred to as Schmidt's criterion. Over the past 20 years, there has been some debate on what exactly minimal motion actually means. Kaupp6,7 has suggested that solid-state reactions occur with maximal motion that sometimes occurs when the reaction is heterogeneous rather than homogeneous.8 For a heterogeneous reaction, it has been suggested that the surface behaviour of all such reactions should be examined via AFM to determine if the solid-state reaction indeed involves minimal motion.6 Since the tail-end absorbed radiation (the sample irradiated with the lowest energy wavelength able to drive the reaction) technique pioneered by Enkelmann and co-workers9 in 1994, single-crystal-to-single-crystal (SCSC) photochemical reactions have become more common. The use of tail-end radiation results in a crystal being reacted more evenly and in many cases allows photochemical reactions that would usually not occur as SCSC reactions to be studied as SCSC reactions.10,11 The study of SCSC reactions (both thermal and photochemical) has led to some interesting reaction mechanism discoveries for some of these reactions and allowed the mechanism of reaction cooperativity to be identified in a few solid-state reactions.12–14 Even so, there is still some debate as to whether reactions occur with minimal motion and whether this view is perhaps limiting progress in the field of solid-state chemistry. Reactions that are influenced by the starting coordinates of the parent crystal are currently classified as either topochemical or topotatic.15 Reactions that occur as SCSC reactions are typically regarded as topochemical reactions and viewed as occurring with minimal motion, while reactions that lead to significant loss of crystal quality (and usually crystal disintegration) but where the structure of the product is influenced by the starting coordinates of the reagents are classified as topotactic reactions and occur with significant molecular motion resulting in crystal disintegration. A topotactic pathway is taken by most solid-state reactions where the starting coordinates of the reagents determine the structure of the resulting product. The solid-state dimerisation of nitrosobenzene derivatives recently reported by Varga et al. has been carried out in different topochemical environments,16 which allowed them to propose different possible theoretical explanations of the thermal organic reaction mechanisms in the solid state.
The study of SCSC solid-state reactions by conventional ‘slow’ diffraction techniques only allows us to see the average picture of what is happening in the crystal, compared to the vast array of ever increasing new technologies and spectroscopies becoming available in structural dynamics and photocrystallography.17 However, considerable mechanistic and other information can be gleaned from the careful analysis of these data.8,18–22
Into a 20 ml round bottom flask, 0.200 g of N-allyldichloromaleimide (0.972 mmol) was added, dissolved in ethanol, and heated for five minutes at 80 °C. 0.074 g of thiourea (0.972 mmol) was dissolved in ethanol and added to the flask. The solution was then heated under reflux for 2 hours during which green crystals began precipitating out of the orange yellow solution. The ethanol was then removed in vacuo, and the product was dissolved in dichloromethane. The organic layer was washed with dichloromethane, water and brine in order to thoroughly extract the product from the aqueous layer after which the dichloromethane solution was dried with anhydrous magnesium sulfate and filtered. The solvent was removed in vacuo with bis(N-allylimino)-1,4-dithiin crystallising out as green crystals with a yield of 0.162 g (76%).
Crystals of the recrystallised product were prepared by reacting 0.020 g of the CT at 130 °C, followed by recrystallisation from dichloromethane by slow evaporation, which resulted in a solvated crystal form and a non-solvated crystal form.
For the two recrystallised product structures, it was found that the bromine atom was disordered over two positions due to full molecule disorder, with the remaining atoms approximately in the same position as the dominant structure. While the bromine atom was refined over two positions, it was not possible to place an alternative position for the second orientation for the molecule. Consequently, the R-factors for the recrystallised products are high. Analysis using PLATON24 did not indicate the presence of twinning in either case. Particular details for the final refinements for all structures in the form of SHELXL-2014 RES files can be found in the CIF file associated with the structure in the ESI.† The software programs used in this work were as follows: data collection: APEX2;25 cell refinement and data reduction: SAINT;26 program suite used to solve and refine structures: SHELX-2014;23 molecular graphics: ORTEP-3 for Windows,27SCHAKAL-9928 and CrystalExplorer-3.1;29 software used to prepare material for publication: WinGX-2014.130 and PLATON.24 Crystallographic information for the CT before and at various stages of conversion can be found in Table 1, while the crystallographic information for the recrystallised product can be found in Table 2. The atom numbering scheme for the CT and product structures can be found in Fig. 2. The same numbering scheme has been used for the recrystallised product crystals.
Conversion/% | 0 (starting CT) | 9.4 (6 days) | 20.1 (12 days) | 29.4 (19 days) | 41.9 (25 days) | 58.1 (31 days) | 80.4 (37 days) | 88.2 (43 days) | 91.0 (49 days) | 93.4 (54 days) | 94.0 (61 days) | 96.7 (70 days) | 97 (84 days) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Crystal system | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic | Monoclinic |
a/Å | 12.8529(5) | 12.8561(4) | 12.8650(6) | 12.8617(8) | 12.8605(14) | 12.905(2) | 12.8578(11) | 12.8049(10) | 12.7386(12) | 12.6987(10) | 12.668(2) | 12.6830(16) | 12.6357(10) |
b/Å | 13.5669(4) | 13.6180(4) | 13.6783(6) | 13.7077(8) | 13.7943(15) | 14.015(2) | 14.1427(12) | 14.1886(9) | 14.1989(11) | 14.2111(11) | 14.193(3) | 14.2527(17) | 14.2062(11) |
c/Å | 14.6676(5) | 14.6351(5) | 14.6129(7) | 14.5846(9) | 14.5372(16) | 14.524(3) | 14.4105(13) | 14.4999(11) | 14.6254(12) | 14.6928(12) | 14.740(3) | 14.782(2) | 14.7531(12) |
β/° | 107.7130(10) | 107.9060(10) | 108.152(2) | 108.297(2) | 108.642(4) | 109.180(4) | 110.053(4) | 110.453(3) | 110.594(4) | 110.648(3) | 110.608(5) | 110.713(4) | 110.689(4) |
Unit cell volume/Å3 | 2436.40(15) | 2438.12(13) | 2443.5(2) | 2441.3(3) | 2443.6(5) | 2481.1(7) | 2461.6(4) | 2468.3(3) | 2476.3(4) | 2481.2(3) | 2480.6(8) | 2499.4(6) | 2477.5(3) |
Temperature/K | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) | 173(2) |
Space group | P21/n | P21/n | P21/n | P21 | P21 | P21 | P21 | P21/n | P21/n | P21/n | P21/n | P21/n | P21/n |
Density (calc.)/g cm−3 | 1.613 | 1.612 | 1.608 | 1.609 | 1.608 | 1.583 | 1.596 | 1.592 | 1.587 | 1.583 | 1.584 | 1.572 | 1.586 |
Z | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 |
Radiation type | MoKα | MoKα | MoKα | MoKα | MoKα | MoKα | MoKα | MoKα | MoKα | MoKα | MoKα | MoKα | MoKα |
Absorption coefficient, μ/mm−1 | 1.899 | 1.898 | 1.894 | 1.895 | 1.893 | 1.865 | 1.880 | 1.874 | 1.868 | 1.865 | 1.865 | 1.851 | 1.868 |
Absorption correction | Integration | Integration | Integration | Integration | Integration | Integration | Integration | Integration | Integration | Integration | Integration | Integration | Integration |
No. of reflections measured | 85876 | 86184 | 62425 | 61508 | 56880 | 96915 | 42789 | 65093 | 34735 | 62291 | 23758 | 32098 | 53951 |
No. of independent reflections | 5873 | 5879 | 5888 | 8587 | 8583 | 8691 | 8630 | 4346 | 4364 | 4369 | 4336 | 4396 | 3888 |
R int | 0.0479 | 0.0613 | 0.0512 | 0.0829 | 0.1335 | 0.2648 | 0.2129 | 0.3217 | 0.3077 | 0.2943 | 0.2266 | 0.2398 | 0.3034 |
Final R1 (I > 2σ(I)) | 0.0263 | 0.0333 | 0.0364 | 0.0696 | 0.1032 | 0.1463 | 0.1038 | 0.0781 | 0.1037 | 0.0830 | 0.0799 | 0.0725 | 0.0907 |
Final wR(F2) (I > 2σ(I)) | 0.0667 | 0.0756 | 0.0906 | 0.1632 | 0.2494 | 0.3412 | 0.2351 | 0.1613 | 0.1538 | 0.1447 | 0.1465 | 0.1200 | 0.1432 |
Final R1 (all data) | 0.0349 | 0.0526 | 0.0554 | 0.1057 | 0.1724 | 0.2358 | 0.2224 | 0.1656 | 0.2066 | 0.1621 | 0.1864 | 0.1588 | 0.1502 |
Final wR(F2) (all data) | 0.0715 | 0.0836 | 0.1018 | 0.1880 | 0.2982 | 0.4108 | 0.2969 | 0.1988 | 0.1942 | 0.1747 | 0.1909 | 0.1482 | 0.1663 |
Goodness of fit on F2 | 1.032 | 1.020 | 1.042 | 1.062 | 1.043 | 1.039 | 1.045 | 1.016 | 1.040 | 1.030 | 1.019 | 1.027 | 1.128 |
CCDC number | 1409360 | 1409361 | 1409362 | 1409363 | 1409364 | 1409365 | 1409366 | 1409367 | 1409368 | 1409369 | 1409370 | 1409371 | 1417564 |
Non-solvated | Dichloromethane solvate | |
---|---|---|
Molecular formula | C28H19BrN2O4S2 | C28H20BrN2O4S2, CH2Cl2 |
Formula mass | 591.48 | 677.41 |
Crystal system | Monoclinic | Triclinic |
a/Å | 12.7053(8) | 8.2573(9) |
b/Å | 23.1346(14) | 12.5878(13) |
c/Å | 8.5162(5) | 14.5411(16) |
α/° | 90 | 110.120(6) |
β/° | 103.246(4) | 90.583(7) |
γ/° | 90 | 99.971(6) |
Unit cell volume/Å3 | 2436.6(3) | 1393.8(3) |
Temperature/K | −100 | −50 |
Space group | P21/c | P |
Density (calc.)/g cm−3 | 1.612 | 1.614 |
Z | 4 | 2 |
Radiation type | MoKα | MoKα |
Absorption coefficient, μ/mm−1 | 1.899 | 1.856 |
Absorption correction | Integration | Integration |
No. of reflections measured | 22108 | 21563 |
No. of independent reflections | 5878 | 5449 |
R int | 0.0861 | 0.1251 |
Final R1 (I > 2σ(I)) | 0.0664 | 0.1111 |
Final wR(F2) (I > 2σ(I)) | 0.1549 | 0.3219 |
Final R1 (all data) | 0.1324 | 0.1864 |
Final wR(F2) (all data) | 0.1709 | 0.3543 |
Goodness of fit on F2 | 1.029 | 1.075 |
CCDC number | 1409373 | 1409372 |
Hirshfeld surfaces were generated using CrystalExplorer-3.1.29
Molecules in the CT structure arrange in stacks where donor and acceptor molecules alternate. The allyl groups on the acceptor molecules are arranged trans to each other. Each allyl group on an acceptor points towards the allyl group on the next acceptor within a stack, with the donor molecule located between them. This results in the donor molecule within a stack effectively being surrounded on three sides by a pocket formed by two acceptor molecules as shown by the blue d-norm Hirshfeld surfaces in Fig. 3b which surround the green de surface of the bromoanthracene. The solid-state reaction discussed in this paper occurs at the open end of this pocket. As also shown in Fig. 3b, the bromine atom of the bromoanthracene molecule is surrounded by the allyl groups from the acceptor molecules in two neighbouring CT stacks. An unusual feature of this stack – compared to such stacks in previously reported dithiin charge-transfer crystals12,34 – is that the dienophiles are pseudo related through a mirror plane defined by the anthracene molecule, which means that the diene interacts with the opposite sides of the same double bond in the dienophile above and below it. The usual situation is that the dienophiles are related (or pseudo related) by an inversion centre located on the central ring of the anthracene. Distances between the reacting atoms are between 3.41 and 3.49 Å at -100 °C and therefore well within Schmidt's criterion (<4.2 Å).5 The reaction can therefore theoretically occur in either direction, but at 30 and 50 °C, it occurs along the longer direction indicated in Fig. 3b. This is also unusual, as all other solid-state reactions involving dithiin derivatives with anthracene have been found to occur in both directions along the CT stack, although there is evidence suggesting that there is considerable cooperativity in the process and that the reaction once initiated occurs to some extent in a single direction before changing direction again.12,34 In the case of the AD:9BrA crystal, the reaction seems to be completely cooperative as the reaction only occurs in one direction despite the two distances being approximately equal (differ less than 0.08 Å) to each other. It is likely that this may be a consequence of asymmetry in the local packing environment, this asymmetry being indicated by the previously mentioned pseudo mirror symmetry around the anthracene molecules.
The CT stacks after about 20% and 93% conversion (monomers deleted for clarity in this case) to the Diels–Alder product are shown in Fig. 4. A distinctive feature of the reaction is the rotation of one of the allyl groups from a trans position (with respect to the other allyl group) to a cis position. From a crystal quality point of view, the consequence of this motion is that it leads to crystal degradation after about 20% conversion has occurred. In our first attempt to carry out a SCSC reaction at 30 °C, the crystal disintegrated, while the second attempt resulted in the work reported here. The reaction at 75 °C for 24 hours resulted in a crystal that had maintained its crystal habit (though mostly opaque) but no longer diffracted, suggesting that at this temperature the reaction is more random, and the molecular motions (the rotation of the allyl group being one of these) are more significant.
Fig. 4 Molecules in the charge-transfer stack of the crystal (a) at 20% conversion (product in orange) and (b) after 93% conversion (CT monomers deleted for clarity) to the Diels–Alder cycloadduct. Note the position of the allyl groups in the final product which are orientated trans with respect to each other initially (a; see also Fig. 3b) but orientated cis with respect to each other in the final product crystal (b). Note also that the orientation of the allyl group in the 20% reacted crystal is similar to that of the starting material. It is only after 28% conversion that the conformation of this allyl group starts resembling that of the final product. In addition, the product molecules tilt slightly off the original CT stacking axis in the SCSC product crystal as highlighted by the arrows in (b). |
While the crystal may experience a significant amount of stress due to the rotation of the allyl group, the process is required as the formation of the Diels–Alder product itself leads to some space (a void) being created between the anthracene and dithiin rings, which bend away from each other during the reaction as shown in Fig. 5a. This space is filled by the allyl group rotating towards the void as the reaction proceeds beyond 20% conversion.
Fig. 5 Overlay by least squares fit of N and S atoms of the unreacted CT structure (blue), the CT molecules in the 28% reacted crystal (red), the product conformation after about 80% conversion [major conformation in grey (65%), minor conformation in orange (15%)], and the product conformation after about 91% conversion [major conformation in green (83%), minor conformation in tan (8%)]. The presence of the minor components indicates the probable path taken by the allyl group during the rotation assuming it is the intermediate allyl conformation. The sole conformation of the product at 20% conversion (orange in Fig. 4a) is also shown. In this conformation, the allyl group lies in the same plane as the imide group it is bonded to. In addition, only the major (final) conformation is present in the crystal after 70 days (97% conversion) or more of reaction. |
Analysis of the allyl conformation in the starting CT structure, the product conformation at 20% conversion, and two almost completely reacted crystals also indicates the path taken by the allyl group during the reaction (Fig. 5b). This concerted motion probably also leads to the slight tilt of the product molecules relative to the CT axis visible in Fig. 4b. The tilt alternates from +5 to −5° relative to the b axis for each successive product molecule, with the rotation occurring parallel to the c axis (see Fig. 3a for axes). In order to determine the most stable allyl conformation likely to be adopted by the product, some reacted product was recrystallised by slow evaporation from dichloromethane leading to two crystal forms: non-solvated product crystals and dichloromethane solvate product crystals. The resulting product conformation from the recrystallised product structures and the product from the 91% and 93% reacted crystal structures are shown in Fig. 6. Ironically, the allyl group that does not undergo rotation during the SCSC reaction is the one that shows the most variation in the molecules from the various structures. However, the allyl group that does undergo rotation during the solid-state reaction does adopt approximately the same conformation in all these structures. This suggests that one of the driving forces for the molecular rotation is minimisation of the intramolecular energy, with the creation of a void space during the reaction allowing the process to occur. In the case of the allyl group that does not move, the conformation in both the CT and the reacted product is almost the same but different from the recrystallised structures. C–H⋯O interactions involving this allyl group increase in strength as the crystal anneals near the end of the reaction and may hold it in place during the reaction (see H-bonds C14–H14B⋯O3 and C14B–H14D⋯O3B in Table 3).
D–H⋯A | D–H | H⋯A | D⋯A | <D–H⋯A |
---|---|---|---|---|
*Converted from aromatic to methine hydrogen in the reaction. | ||||
CT | ||||
C11–H11⋯O3i | Not possible in this orientation | |||
C14–H14B⋯O3ii | 0.95 | 2.76 | 3.602(2) | 148 |
C22–H22*⋯O1iii | 0.95 | 2.54 | 3.350(2) | 144 |
C27–H27⋯O4iv | 0.95 | 2.48 | 3.239(2) | 136 |
C11–H11B⋯Br1v | Not possible in this orientation | |||
C8–O4⋯S1vi | 3.487(1) | 145.8(1) | ||
54 day (93%) product structure | ||||
C11C–H11E⋯O3Bi | 0.95 | 2.33 | 3.270(13) | 171 |
C14B–H14D⋯O3Bii | 0.95 | 2.48 | 3.330(16) | 149 |
C22B–H22B*⋯O1Biii | 1.00 | 2.97 | 3.765(10) | 137 |
C27B–H27B⋯O4Biv | 0.95 | 2.62 | 3.204(10) | 120 |
C11C–H11F⋯Br1Bv | 0.95 | 3.02 | 3.849(12) | 147 |
C8B–O4B⋯S1Bvi | 3.297(6) | 145.3(6) | ||
70 day (97%) product structure | ||||
C11C–H11E⋯O3Bi | 0.95 | 2.36 | 3.308(14) | 178 |
C14B–H14D⋯O3Bii | 0.95 | 2.48 | 3.331(16) | 149 |
C22B–H22B*⋯O1Biii | 1.00 | 2.98 | 3.762(9) | 136 |
C27B–H27B⋯O4Biv | 0.95 | 2.63 | 3.205(9) | 120 |
C11C–H11F⋯Br1Bv | 0.95 | 3.03 | 3.846(13) | 145 |
C8B–O4B⋯S1Bvi | 3.302(5) | 145.2(5) |
The mobile allyl group also forms new weak interactions as it changes position as can be seen by examining close contacts involving the affected allyl group before and after the reaction. While the strongest interactions in the CT and reaction product structures are dominated by charge-transfer and π⋯π interactions, the structure is also stabilised by C–H⋯O and C–H⋯Br interactions. The weakening of some of these present in the CT structure followed by the strengthening of other hydrogen bonds and the appearance of a new one in the reacted crystal can be seen in the Hirshfeld surfaces in Fig. 7 and the H-bond list in Table 3. The Hirshfeld surfaces for the CT before the reaction (Fig. 7a and c) indicate that the carbonyl groups are involved in strong C–H⋯O interactions with the aromatic hydrogen atoms of the anthracene molecules from neighbouring stacks. One of these – C22–H22⋯O1 – lengthens dramatically as C22 is directly involved in the reaction, changing from an aromatic sp2 carbon to an alkyl sp3 carbon in the process; the C22⋯O1 distance which is 3.350 Å before the reaction lengthens to 3.762 Å at 97% and is therefore absent in the Hirshfeld surface of the product.
At the beginning of the reaction, there are no significant C–H⋯O or C–H⋯Br interactions involving either of the allyl groups as can be seen in Table 3, where hydrogen bonds involving C11 or C14 with O or Br are either long or non-existent. However, after 80% conversion, very significant C–H⋯O interactions appear, involving both allyl groups now interacting with the carbonyl groups on molecules in neighbouring stacks. In the case of the conformationally stable allyl group, strong C–H⋯O interactions involving C14 are formed in the direction of a nearby carbonyl group, presumably due to the relaxation (annealing) of the structure near the end of the reaction. In this case, the C14⋯O3 distance shortens from 3.602 Å in the CT to 3.331 Å in the reaction product (Table 3). No conformational change is required by this allyl group to bring about the interaction, and the C–H⋯O interaction is centrosymmetric resulting in the formation of a molecular dimer that can be seen in Fig. 7d. On the other hand, the allyl group that does undergo conformational change forms two new interactions which are not present in the original CT structure: a weak C–H⋯Br (C11C–H11F⋯Br1B) interaction indicated in Fig. 7b and a C–H⋯O interaction (C11C–H11E⋯O3B) visible in Fig. 7d. These can only form if the allyl group rotates into the correct position as can be seen by comparing the allyl groups in Fig. 6c and d. The overall effect of the conformational change is the optimisation of the intramolecular energy of the molecule as well as the intermolecular energy of the crystal. The geometrical parameters of these hydrogen bonds are listed in Table 3. In addition to changes in the hydrogen bonding, there is also a short contact involving S1 and O4 which shortens from 3.487 Å (C8–O4⋯S1) in the CT to 3.302 Å (C8B–O4B⋯S1B) at 97% conversion.
Fig. 8 (a) Plot of R1 (R-factor) values for structures at different states of conversion solved in P21/n (all structures) and in P21 for conversions between 28% and 80%. The structures solved in P21/n with R-factor values less than 5% up to 20% conversion, but after this, solved with much higher R-factors reaching a maximum of 19.7% at 60% conversion. Solving the structures at 28–80% conversion in P21 instead leads to significantly lower R-factor values. (b) Plot of percent conversion against reaction time. Points 4–6 (28–80% conversion) are from the P21 structure solutions [green points in (a)], with the remainder from the P21/n structure solutions. The graphs in Fig. 10 are plotted with respect to the conversion values shown in (b). |
The deterioration of crystal quality due the solid-state reaction can also be seen in the diffraction patterns of the sample. Reconstructions of the h0l layer are shown in Fig. 9. As can be seen, the diffraction range of the crystal decreases from diffracting beyond 0.75 Å resolution at 10 seconds of exposure in the starting CT (Fig. 9a) to barely diffracting to 1.0 Å resolution after more than 90% conversion at 60 seconds of exposure on the same instrument (Fig. 9c). A broad powder ring around the beam stop which indicates that the crystal exterior is also undergoing a phase change (or becoming amorphous) while still maintaining the crystal habit as the reaction continues is also visible in Fig. 9c.
Despite the large change in the R-factor during intermediate stages of the reaction (about 28–80% of conversion), the unit cell parameters do not change very much. The changes in unit cell parameters, as well as the percent change in unit cell parameters with reaction conversion, are shown in Fig. 10. The a unit cell parameter decreases slightly [about ~0.22 Å (~1.7%)] over the course of the reaction. The c unit cell parameter decreases by about ~0.26 Å (~1.7%) until 80% conversion is reached. At this point, c starts to lengthen, and at 97% conversion, it is ~0.11 Å (0.78%) longer than that in the CT. These axes are orientated perpendicular to the reacting stack axis (the CT stack is parallel to b) and would therefore not be expected to change much. The b axis increases by ~0.64 Å (~4.7%). This large change is probably due to the Diels–Alder cycloadduct being formed in this direction within the stack axis. The molecules expand along the b axis, and the movement of the allyl groups also occurs along the stack axes. The β angle increases by ~3.0° (~2.8%) as the CT stacks relax (anneal) into new positions with respect to each other near the end of the reaction. While the volume data in Fig. 10c are erratic due to the weak data, the overall trend is an increase in cell volume of ~41 Å3 (~1.7%).
Unlike the solid-state reacted crystal where molecules stack on each other and where electron poor (ex-dithiin) regions of the product molecules interface with the electron rich (ex-anthracene) regions of neighbouring molecules within a stack (Fig. 4b), molecules in the non-solvated recrystallised product first overlap on each other on the ex-anthracene regions to form π⋯π dimers which then pack in an edge-to-face manner to form a zig-zag or herring bone type structure (Fig. 12a). In the solvated product crystal, the molecules arrange in stacks similar to those in the solid-state reaction product. However, while the molecules are vertically arranged on top of each other in the SCSC product – this being a consequence of the starting CT structure – the product molecules in the solvated crystal structure are displaced relative to each other, with an allyl group sitting over the aromatic rings of the product below it in the stack and with dichloromethane molecules filling up the space between the product molecules (Fig. 12b).
Lattice energy calculations carried out using PIXEL31 involving only the dominant conformation in the non-solvated recrystallised product crystal yield an energy of −213.4 kJ mol−1. Carrying out the same calculation for only the product (dominant conformation) in the SCSC reacted crystals which have reacted 91% or more yields energies of about -195(1) kJ mol−1. Ignoring the difference in the intramolecular energies (which may be significant), the solution grown product is clearly more stable from an intermolecular point of view. The solid-state reaction has, however, resulted in a very different crystal from those obtained from solution and provides a method for producing metastable crystals.
For comparison, molecule⋯molecule interactions stronger than −35 kJ mol−1 present in each of the three product crystals (SCSC derived and both recrystallised forms) are shown in Fig. 13. For the SCSC product, the strongest interactions are with molecules in neighbouring stacks related by an inversion axis [data from the 54 day reacted (93% conversion) crystal, lattice energy of 195.7 kJ mol−1]; all contribute −55 kJ mol−1 or more, with the strongest interaction contributing −60.2 kJ mol−1 (Fig. 13a). Molecules along the stacking (ex-CT axis) contribute to the fourth strongest interaction at −50.6 kJ mol−1. The molecule⋯molecule interaction bringing the centrosymmetric C–H⋯O interaction shown in Fig. 7d into alignment contributes −26.7 kJ mol−1 in this structure, while the molecule⋯molecule interaction associated with the new C–H⋯O hydrogen bond contributes −18.7 kJ mol−1. The molecule⋯molecule interaction associated with the S1⋯O4 interaction listed in Table 3 contributes −15.1 kJ mol−1, with the energy contributions due to Coulombic, polarization, dispersion and repulsive forces being −4.1, −3.5, −18.8 and 11.4 kJ mol−1, respectively, indicating that the S⋯O interaction is attractive in this case, although these numbers are for the full molecule⋯molecule interaction.
In the non-solvated recrystallised product, the strongest interaction is between the π⋯π interacting dimers discussed previously at −71.0 kJ mol−1, while the edge-to-face interaction contributes about −42.9 kJ mol−1 (Fig. 13b). For the solvated recrystallised product, the strongest interaction is with a molecule related by an inversion centre in a neighbouring stack at −69.6 kJ mol−1, while molecules along the stack discussed previously have a molecule⋯molecule interaction energy of −50.7 kJ mol−1 (Fig. 12c). Interactions involving the dichloromethane contribute less than −25 kJ mol−1 each to the lattice energy.
Footnote |
† Electronic supplementary information (ESI) available. CCDC 1409360–1409373 and 1417564. For crystallographic data in CIF or other electronic format see DOI: 10.1039/c5ce01301a |
This journal is © The Royal Society of Chemistry 2015 |