Crystal X-ray diffraction and molecular modelling considerations

Subscriber access provided by University of Glasgow Library

Article

Crystal X-ray diffraction and molecular modelling considerations elucidate the factors responsible for the opposing host behaviour of two isostructural xanthenyl- and thioxanthenyl- derived host compounds Benita Barton, Cedric W. McCleland, Mino R. Caira, Lize de Jager, and Eric C Hosten Cryst. Growth Des., Just Accepted Manuscript • DOI: 10.1021/acs.cgd.9b00078 • Publication Date (Web): 19 Feb 2019 Downloaded from http://pubs.acs.org on February 19, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

1

Crystal X-ray diffraction and molecular modelling considerations elucidate the factors responsible for the opposing host behaviour of two isostructural xanthenyl- and thioxanthenyl- derived host compounds

Benita Barton,*a Cedric W. McCleland,*a Mino R. Caira,b Lize de Jager,a Eric C. Hostena

a

Department of Chemistry, PO Box 77000, Nelson Mandela University, Port Elizabeth, 6031, South

Africa. E-mail: [email protected] b

Department of Chemistry, University of Cape Town, Rondebosch, 7701, South Africa

Abstract

In this work, we compare the host behaviour of two structurally related compounds, N,N'-bis(9phenyl-9-xanthenyl)ethylenediamine and N,N'-bis(9-phenyl-9-thioxanthenyl)ethylenediamine when each one was recrystallized from mixtures of guest solvents pyridine, morpholine, piperidine and dioxane.

(These solvents were clathrated individually in single solvent

experiments.) In these conditions, the two hosts displayed distinctly opposing selectivities: when employing an equimolar quaternary guest mixture, the resulting mixed complex of the oxygen host analogue contained a significant amount of dioxane (68%) and only 8% pyridine, while the sulfur host derivative discriminated inordinately against dioxane (12%) in favour of pyridine (57%). These results were more remarkable when considering the crystal packing of the two apohost compounds, which revealed them to be isostructural. The reasons for the opposing host behaviours were explored through detailed crystal X-ray diffraction and computational studies at the molecular mechanics and DFT levels. Furthermore, thermal analyses indicated that complexes containing preferred guest solvents possessed increased thermal stabilities. Finally, the two host compounds were allowed to compete by dissolving equimolar mixtures of these in each of the heterocyclic guest solvents, and the results of these experiments were related back to their relative selectivity preferences.

Keywords

Host-guest Chemistry; Selectivity; Molecular Modelling; X-ray Crystallography; Isostructural.

ACS Paragon Plus Environment

Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 59

2

1. Introduction

Supramolecular chemistry may be defined as chemistry beyond the molecule, and relies on molecular recognition phenomena between host and guest compounds to facilitate the complexation of the guest compounds within the host framework.1,2 Guest molecules are usually, but not always, held in the host crystal by means of non-covalent interactions which may include, but are certainly not limited to, hydrogen bonding, π−π stacking and CH−π interactions. These are then often responsible for the inclusion phenomena.

Host compounds may enclathrate selectively when recrystallized from mixtures of guest compounds that are individually clathrated. For example, Nassimbeni et al3 employed hosts 9,9′bianthryl, 9,9′-spirobifluorene and trans-2,3-dibenzoylspiro[cyclopropane-1,9-fluorene], while Lusi and Barbour4 used a Werner complex Ni(NCS)2(para-phenylpyridine)4, to investigate the feasibility of separating xylene isomers. Furthermore, Toda and his research team5 showed that 1,1,2,2-tetraphenylethane-1,2-diol selectively includes para-xylene from a para-xylene/metaxylene mixture, and our research team has also recently focused on xylene inclusion selectivities using a host compound derived from naturally-occurring tartaric acid.6,7 The separation of the xylene isomers is a tangible challenge since fractional distillations are usually inefficient owing to the very similar boiling points of these isomers. Although this discussion has focused primarily on the xylenes, analogous concerns arise when considering many other isomeric compounds that have important industrial applications.

The selectivity displayed by host compounds is not limited to experiments involving achiral isomeric guest compounds alone, but also has application in the resolution of racemates. α,α,α',α'-Tetraphenyl-2,2-dimethyl-1,3-dioxolane-4,5-dimethanol (TADDOL) 1 (Scheme 1), also tartaric acid-derived, is highly proficient in the resolution of a wide range of racemic compounds, and separations were achieved with very high enantiomeric excesses.8

Compounds N,N'-bis(9-phenyl-9-xanthenyl)ethylenediamine 2 and the sulfur analogue, N,N'bis(9-phenyl-9-thioxanthenyl)ethylenediamine 3 (Scheme 1), have demonstrated an extremely rich host-guest chemistry. These compounds were first synthesized in our laboratories and demonstrated inclusion ability, forming complexes with a few guests in these initial


Page 3 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


3

investigations.9,10 Additionally, the thioxanthenyl derivative 3 was functionalized further by altering the diamino alkyl linker or substituting hydrogen for chlorine on the A ring; both these adaptations negatively affected the host ability.10 More recently, 3 was reported to form complexes with the dihaloalkanes CH2Cl2, CH2Br2 and CH2I2.11 The host packing in the three complexes was determined to be isostructural (monoclinic, P21/n), while the apohost material experienced an altogether different packing mode (monoclinic, P1). These results confirmed that the host structure collapses in the absence of guest and, therefore, does not possess zeolitelike character. Both 2 and 3 were also subjected to these dihaloalkanes but in their vapour phase: interestingly, only 3 demonstrated the ability to absorb the gaseous guests and form complexes with each one, while 2 was not able to do so.12 Owing to its highly efficacious host ability, the behaviour of 3, when presented with mixed pyridines or xylenes, was also examined.7,13 This host displayed near-complete selectivity, more especially when it was recrystallized from an equimolar o-/m-/p-xylene mixture, where 95% of the p-isomer was found in the resultant complex.

It is therefore clear from these previous reports that compounds 2 and 3 possess contrasting host abilities (despite the only difference being in their two B rings, where position 10 corresponds to an oxygen atom in the former, and a sulfur atom in the latter). This behaviour difference was affirmed in the present investigation when considering recrystallization experiments of 2 and 3 from mixtures of guests pyridine, morpholine, piperidine and dioxane (which were individually included by both host analogues in single solvent experiments) (Scheme 1); observations made from these experiments were remarkable in the extent of these differences. In order to understand this phenomenon, we carried out an in-depth comparative analysis of the two hosts in both the absence and presence of guest by considering both their host crystal structures, obtained from crystal diffraction experiments, and computational conformational and electronic analyses.

The outcomes of these investigations are reported

here, as well as the crystal structures of each of the host-guest complexes. Additionally, thermal analyses were employed to ascertain the relative thermal stabilities of the complexes and, finally, interesting observations were made from experiments where the two host compounds were allowed to compete for guest inclusion (as opposed to competition experiments involving crystallization of a single host compound in the presence of two or more potential guest compounds).



Page 4 of 59

4

Scheme 1. Structures of TADDOL (1), N,N'-bis(9-phenyl-9-xanthenyl)ethylenediamine (2), N,N'bis(9-phenyl-9-thioxanthenyl)ethylenediamine (3) and the four six-membered guest heterocyclics

2. Results and discussion

The syntheses of host compounds 2 and 3 are simple and high yielding, and have been published previously.9,10 (In the Supplementary Information, Figures S1−S6 depict the 1H NMR, 13C NMR and IR spectra for both host compounds.)

2.1 Assessment of the host ability of compounds 2 and 3 for potential guest solvents pyridine, morpholine, piperidine and dioxane

Xanthenyl compounds 2 and 3 were recrystallized from the four heterocyclic organic solvents in order to determine whether complexes would form: these host compounds were consequently dissolved in each of the guests in open vials under ambient conditions. Once crystallization occurred, the crystals were collected by suction filtration and thoroughly rinsed with petroleum


Page 5 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


5

ether (40−60 °C) in order to remove any superficial guest solvent. 1H NMR analysis revealed that each of the guest compounds was enclathrated in the host:guest ratios shown in Table 1.

Table 1: Host:guest ratios for the complexes of host 2 and 3 formed with the four heterocyclic guests compounds Host:guest (H:G) ratioa

Complex Host 2∙2Pyridine

1:2

Host 2∙Morpholine

1:1

Host 2∙Piperidine

1:1

Host 2∙Dioxane

1:1

Host 3∙Pyridine

1:1

Host 3∙Morpholine

1:1

Host 3∙Piperidine

1:1

Host 3∙Dioxane a

1:1 1

Determined using H NMR spectroscopy.

With the exception of host 2∙pyridine (H:G 1:2), all of the complexes crystallized with a 1:1 H:G ratio.

2.2 Hosts’ 2 and 3 selectivities in the presence of equimolar binary, ternary and quaternary mixtures of the four guest solvents

Since each of the four solvents was included by hosts 2 and 3, we subsequently investigated their inclusion selectivities when each was recrystallized from equimolar binary, ternary and quaternary mixtures of these guest compounds. In order to maintain the equimolar condition, these experiments were carried out in closed vials at approximately 0 °C. After crystallization occurred, the solid material was treated in an identical fashion to the products from the single solvent recrystallization experiments. Table 2 summarizes the data obtained, with the H:G ratios being determined through 1H NMR spectroscopy as previously described; the preferred guest compound is shown in bold italic red font face.



Page 6 of 59

6 Table 2: Results from competition experiments for hosts 2 and 3 using equimolar binary, ternary and quaternary mixtures of the four guest solventsa,b Entry

Host Guest Compound Pyridine

Guest Morpholine

1

2

x

2

2

x

3

2

x

4

2

5

2

x

6

2

x

7

2

x

x

8

2

x

9

2

x

10

2

11

2

x

12

3

x

13

3

x

14

3

x

15

3

16

3

x

17

3

x

18

3

x

x

19

3

x

Guest Piperidine

Guest Dioxane

Guest ratios (%e.s.d.s)c

Overall H:G ratio

x

0.10:0.90 (0.68)

1:1

d

x x

0.15:0.85 (1.00)

1:1

x

0.05:0.95 (0.42)

1:1

x

0.30:0.70 (0.95)

1:1

x

0.94:0.06 (1.99)

1:1

x

0.08:0.84:0.08 (0.41:0.28:0.13)

1:1

x

0.08:0.04:0.88 (0.36:0.74:1.10)

1:1

x

0.11:0.23:0.66 (0.39:0.60:0.22)

1:1

x

x x x

x

x

0.23:0.05:0.72 (0.91:0.37:0.53)

1:1

x

x

x

0.08:0.20:0.04:0.68 (0.11:0.17:0.60:0.88)

1:1

x

0.82:0.18 (0.72)

1:1

0.81:0.19 (0.47)

1:1

0.76:0.24 (1.10)

1:1

x

0.54:0.46 (0.90)

1:1

x

0.50:0.50 (0.66)

1:1

x

0.52:0.48 (1.29)

1:1

x

0.66:0.18:0.16 (0.21:0.61:0.39)

1:1

0.69:0.17:0.14 (0.33:0.58:0.25)

1:1

x x x

x

x


Page 7 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


7 20

3

21

3

22

3

x

x

x

x

0.61:0.17:0.22 (0.39:0.23:0.16)

1:1

x

x

x

0.33:0.34:0.32 (0.53:0.37: 0.91)

1:1

x

x

x

0.57:0.16:0.14:0.12 (0.08:0.13:0.23:0.09)

1:1

a

H:G ratios determined using 1H NMR spectroscopy using CDCl3 as the solvent; bexperiments were always conducted in duplicate for confirmation purposes; cestimated standard deviations (e.s.d.s) are provided in parentheses; dthese mixtures did not crystallize, and a gum remained in the recrystallization vessel.

It is clear from Table 2 that hosts 2 and 3 behave entirely differently in the presence of the various combinations of equimolar guest solvents. Host 2, when recrystallized from binary guest mixtures comprising dioxane, displayed high selectivity for this guest compound (90, 95 or 70% when the other guest solvent was pyridine, piperidine or morpholine, respectively; entries 1, 4 and 5). Furthermore, an analysis of the experiments where pyridine was present revealed that inclusion of this guest was disfavoured [pyridine/dioxane and pyridine/morpholine furnished crystals with only 10 and 15% pyridine, respectively (entries 1 and 3); notably, crystallization was not successful in the pyridine/piperidine experiment (entry 2)]. These results are in stark contrast to similar experiments involving host 3. The binary experiments with pyridine showed that this guest compound was now highly preferred (pyridine/dioxane, pyridine/piperidine and pyridine/morpholine afforded mixed complexes containing 82%, 81% and 76% pyridine; entries 12, 13 and 14); therefore only 18% of the complex contained dioxane when 3 was recrystallized from pyridine/dioxane (entry 12). Also evident is that host 3 is significantly less selective than 2. Comparison of entries 5 and 10 with entries 16 and 21 shows that host 3 is unselective when recrystallized from morpholine/dioxane (entry 16, 50:50) and morpholine/piperidine/dioxane (entry 21, 33:34:32), while 2 is highly selective for dioxane in both instances (entry 5, morpholine/dioxane 30:70, and entry 10, morpholine/piperidine/dioxane 23:5:72).

From the data in Table 2, and particularly where the host compounds were recrystallized from equimolar quaternary mixtures of the solvents, it is apparent that the relative guest selectivities of hosts 2 and 3 may be ranked as:

Host 2: dioxane (68%) > morpholine (20%) > pyridine (8%) > piperidine (4%) (entry 11); Host 3: pyridine (57%) > morpholine (16%) > piperidine (14%) > dioxane (12%) (entry 22).



Page 8 of 59

8

We subsequently conducted binary competition experiments in which the guest solvents were present in unequal amounts to observe whether the host selectivities vary in the presence of changing guest molar ratios.

2.3 Selectivity investigations of hosts 2 and 3 using binary solvent mixtures with varying molar ratios

Each host compound was dissolved in a mixture of known amounts of any two of the four relevant solvents, the vessels closed and retained at approximately 0 °C. Upon crystallization, both the mother liquor from which crystallization had occurred and the crystalline material itself (after washing) were analysed using 1H NMR spectroscopy. We were therefore able to construct selectivity profiles, which display the selectivity behaviour of the host over a guest concentration range, by plotting the mole fraction of guest 1 present in the crystal after recrystallization of the host compound from the two-guest mixture (Z) against the mole fraction of guest 1 in the twoguest mixture (X) [Figures 1a−e (host 2) and 2a−f (host 3)].# For two guests A and B, the selectivity coefficient is defined as K(A:B) = ZA/ZB*XB/XA.14 An average of these values was calculated for each

selectivity profile. The straight lines shown in all of these figures would represent the case for unselective host compounds, where K = 1.

Surprisingly, remarkable differences in the selectivity profiles for the two host compounds are immediately evident when comparing Figures 1 with 2, despite their comparable behaviour in the

single

solvent

experiments.

From

Figures

1a

(pyridine/dioxane)

and

1b

(pyridine/morpholine), it is clear that host 2 is selective for dioxane and morpholine, respectively, over the entire concentration range assessed [average K values were determined to be 11.2 (in favour of dioxane) and 8.3 (for morpholine), respectively]. In sharp contrast are Figures 2a and 2c constructed for host 3 (also pyridine/dioxane and pyridine/morpholine experiments): in these cases, pyridine is consistently preferred [K = 5.5 and 3.5 (both for pyridine)]. (The average K value for the selectivity profile of Figure 2b for host 3 (pyridine/piperidine) was 4.9 in favour of pyridine.) Also evident from these figures and the average computed K values is the higher selectivity of host 2 compared to 3: selectivity profiles from the morpholine/dioxane (Figure 2d)

#

Note that the pyridine/piperidine experiment with host 2 yielded no crystals and a selectivity profile could thus not be constructed.


Page 9 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


9

(K = 1.0), piperidine/morpholine (Figure 2e) (K = 1.3) and piperidine/dioxane (Figure 2f) (K = 1.5) experiments have data points always very close to the hypothetical line that represents no selectivity (K = 1), in contrast to Figures 1c−e, where high selectivities remain for guest solvents dioxane, morpholine and dioxane, respectively (K = 3.2, 7.1 and 28.5).


Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

Page 10 of 59

10

(a)

(b)

(d)

(c)

(e)

Figure 1: Selectivity profiles for host 2 using binary mixtures of (a) pyridine/dioxane, (b) pyridine/morpholine, (c) morpholine/dioxane, (d) piperidine/morpholine and (e) piperidine/dioxane; no crystals formed during the pyridine/piperidine experiment and hence no selectivity profile could be constructed in this case


Page 11 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46


11

(a)

(d)

(b)

(e)

(c)

(f)

Figure 2: Selectivity profiles for host 3 using binary mixtures of (a) pyridine/dioxane, (b) pyridine/piperidine, (c) pyridine/morpholine, (d) morpholine/dioxane, (e) piperidine/morpholine and (f) piperidine/dioxane



Page 12 of 59

12

The divergent behaviour of the host compounds when recrystallized from guest mixtures was unexpected, prompting us to compare the crystal structures of the two apohost compounds.

2.4 Crystal structures of apohosts 2 and 3

Using a Bruker Kappa Apex II diffractometer with graphite-monochromated Mo Kα radiation (λ = 0.71073 Å), these experiments were conducted at 200 K. APEXII and SAINT were used for data collection, and cell refinement and data reduction, respectively.15 SHELXT-201416 was used to solve the structures which were refined by least-squares procedures using SHELXL2017/117 together with SHELXLE18 as a graphical interface. All non-hydrogen atoms were refined anisotropically. Carbon-bound hydrogen atoms were added in idealized geometrical positions in a riding model. Nitrogen-bound hydrogen atoms were located on the difference Electron Density map and refined freely. Data were corrected for absorption effects using the numerical method implemented in SADABS.15

Table 3 contains the relevant crystallographic data for the two apohost compounds (these data have been reported before for 311 but are repeated here for ease of comparison with those for 2). Both crystallize in the triclinic crystal system and space group P1 and, upon closer inspection, appear isostructural owing to their very similar unit cell dimensions. These data were deposited at the Cambridge Crystallographic Data Centre [CCDC reference numbers 1587301 (apohost 2) and 1540116 (apohost 3)11].

In order to confirm the isostructural nature of the crystal packing in these two crystals, we employed the Mercury software program19 to draw best-fit superimposed host molecules, obtained by least-squares, for extended structures of the two apohost compounds (Figure 3a). Also provided here is a stereoview for better clarity (Figure 3b). Furthermore, we computed the powder X-ray diffraction (PXRD) patterns for the two solids (Figure 4).


Page 13 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


13 Table 3: Single crystal X-ray crystallographic data for apohosts 2 and 311

Chemical formula Formula weight Crystal system Space group µ (Mo-Kα)/mm-1 a/Å b/Å c/Å alpha/° beta/° gamma/° V/Å 3 Z F(000) Temp./K Restraints Nref Npar R wR2 S θ min−max/° Tot. data Unique data Observed data [I > 2.0 (I)] Rint Completeness (%) Min. resd. dens. (e/ Å3) Max. resd. dens. (e/ Å3)

(a)

Apohost 2

Apohost 311

C40H32N2O2 572.67 Triclinic P 1 0.077 9.0854(4) 12.2325(5) 14.9806(6) 76.534(2) 79.787((2) 69.738(2) 1510.39(11) 2 604 200 0 7504 405 0.0442 0.1064 1.03 1.8, 28.4 39884 7504 5175

C40H32N2S2 604.79 Triclinic P 1 0.205 9.0912(5) 12.3688(7) 14.9416(8) 77.362(2) 82.375(2) 70.793(2) 1544.81(15) 2 636 200 0 7386 405 0.0391 0.1043 1.03 1.8, 27.9 39535 7386 6497

0.032 99.9 −0.19 0.24

0.017 99.9 −0.78 0.50

(b)

Figure 3: Overlaid (a) extended crystal structures (O – red, S – yellow) and (b) stereoviews for hosts 2 and 3 (O and S – yellow, hydrogen atoms removed for clarity)



Page 14 of 59

14

Figure 4: Computed PXRD patterns for hosts 2 (top) and 3 (bottom)

Figures 3a and 3b show that the host molecules occupy very similar positions in their respective crystal structures, while the computed traces (Figure 4) appear, at first glance, completely dissimilar. However, the specific values of the six unit cell dimensions are what determine the angular positions (2ϴ values) of the peaks, and every atom in the crystal contributes towards the PXRD profile at every angular position, depending on its positional coordinates x, y, z as well as its X-ray scattering power. Furthermore, the intensity contribution from each atom is very sensitive to a small change in its location in space. In the case of the two host compounds 2 and 3, the unit cell parameters are superficially practically the same (in fact, there are differences spanning a wide range, from < 0.1% to about 3.2%). Despite these variations, one would still expect to see PXRD peaks appearing at more or less the same 2ϴ values (with differences of a couple of tenths of a degree for corresponding peaks). From Figure 4, the correspondence, in general, is reasonably good between the positions of the peaks, but the relative intensities are in very poor correlation. However, if the two crystals were isostructural, one would, in fact, expect significant differences in peak intensities, since the O atoms in host 2 are replaced by S atoms in 3, and sulfur has a scattering power approximately twice that of oxygen at low angles. The coordinates of corresponding atoms in the two crystals are in reasonable agreement as one would expect for isostructural crystals, but the replacement of O ACS Paragon Plus Environment

Page 15 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


15

atoms by S atoms is a big perturbation on the molecule itself. For example, S–C bond lengths are significantly greater than O–C lengths, so that the O and the S atoms themselves do not have very similar x,y,z coordinates. Furthermore, this difference induces a distortion of the tricyclic system so that the C atoms are splayed out more in the S analogue. Hence with such significant changes in the atomic coordinates of numerous corresponding atoms, the intensity matches become worse. In addition, there are differences of a few degrees in the orientations of the aromatic moieties, and this means also that the coordinates of corresponding atoms do not have identical values.

To further investigate this phenomenon, the O and S atoms were removed from the calculations, and the resultant computed powder patterns for these modified structures are provided in Figure 5.

Figure 5: Computed PXRD patterns for artificially ‘modified’ hosts 2 (top) and 3 (bottom) after omitting atoms O and S atoms from the computation

From this figure, the powder patterns now appear more alike overall, but there are still significant intensity differences and, as stated above, these are due to differences in the coordinates of the C and N atoms resulting ultimately from the different covalent radii of the O and S atoms. Therefore, the two crystals are indeed isostructural as we use the term loosely ACS Paragon Plus Environment


Page 16 of 59

16

/qualitatively, the deviations from perfect isostructurality being due to the geometrical differences induced when O is replaced by S.

We were determined to investigate the reasons for the contradictory host behaviours despite their isostructural crystal packing owing to very similar molecular structures, and therefore carried out single crystal X-ray diffraction (SCXRD) studies on the eight complexes using suitable quality crystals from the single solvent recrystallization experiments.

2.5 SCXRD analyses for complexes of hosts 2 and 3 with pyridine, morpholine, piperidine and dioxane

Using identical SCXRD experimental parameters as for apohosts 2 and 3, we obtained the diffraction data for the eight complexes and, subsequently, deposited the CIF files at the CCDC [CCDC

reference

numbers 1587302 (2∙2pyridine),

1587304

(2∙morpholine),

1587305

(2∙piperidine), 1587303 (2∙dioxane), 1549682 (3∙pyridine), 1551195 (3∙morpholine), 1551196 (3∙piperidine) and 1551197 (3∙dioxane).] Note that nitrogen-bound hydrogen atoms were, once more, located on the difference Electron Density map and refined freely except for 2∙morpholine, where all the N−H bond distances were restrained to be identical, and 3∙morpholine, which is disordered and the nitrogen hydrogen atom could not be located.

Tables 4 and 5 summarize the relevant crystallographic data and refinement parameters for these SCXRD experiments, for complexes with hosts 2 and 3, respectively.


Page 17 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46


17 Table 4: Single crystal X-ray crystallographic data for the 2∙2pyridine, 2∙morpholine, 2∙piperidine and 2∙dioxane complexes Chemical formula Formula weight Crystal system Space group µ (Mo-Kα)/mm-1 a/Å b/Å c/Å alpha/° beta/° gamma/° V/Å 3 Z F(000) Temp./K Restraints Nref Npar R wR2 S θ min-max/° Tot. data Unique data Observed data [I > 2.0 (I)] Rint Completeness (%) Min. resd. dens. (e/ Å3) Max. resd. dens. (e/ Å3)

2∙2Pyridine C40H32N2O2∙2(C5H5N) 730.88 Triclinic P-1 0.078 9.3906(4) 13.9667(7) 15.8401(8) 112.772(2) 92.849(2) 93.679(2) 1905.05(16) 2 772 200 0 8482 513 0.0397 0.1052 1.02 1.7, 28.4 68864 9492 7634

2∙Morpholine C40H32N2O2∙C4H9NO 659.80 Triclinic P-1 0.080 9.5795(8) 16.4528(12) 22.9013(16) 75.579(2) 81.758(2) 88.166(2) 3459(5) 4 1400 200 15 17227 925 0.0577 0.1631 1.02 0.9, 28.4 108771 17227 11185

2∙Piperidine C40H32N2O2∙C5H11N 657.82 Triclinic P-1 0.076 8.5909(3) 9.6596(3) 22.6958(7) 88.176(2) 88.027(2) 68.416(2) 1749.93(10) 2 700 200 6 8704 462 0.0536 0.1531 1.03 0.9, 28.3 47758 8704 6495

2∙Dioxane C40H32N2O2∙C4H8O2 660.78 Triclinic P-1 0.082 8.2083(4) 12.7907(6) 17.7256(8) 69.522(2) 87.087(2) 79.972(2) 1716.71(14) 2 700 200 0 8510 459 0.0452 0.1216 1.03 1.2, 28.3 51503 8510 6464

0.021 99.8 −0.25 0.33

0.034 100.0 −0.66 0.81

0.027 100.0 −0.39 0.44

0.023 99.8 −0.27 0.47


Crystal Growth & Design 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

Page 18 of 59

18

Table 5: Single crystal X-ray crystallographic data for the 3∙pyridine, 3∙morpholine, 3∙piperidine and 3∙dioxane complexes 3∙Pyridine 3∙Morpholine 3∙Piperidine 3∙Dioxane Chemical formula C40H32N2S2∙C5H5N C40H32N2S2∙C4H8NO C40H32N2S2∙C5H11N C40H32N2S2∙C4H8O2 Formula weight 683.90 690.91 689.94 692.90 Crystal system Monoclinic Monoclinic Monoclinic Monoclinic Space group P21/n P21/n P21/n P21/n µ (Mo-Kα)/mm-1 0.195 0.194 0.187 0.195 a/Å 10.1347(3) 10.2905(7) 13.9591(11) 10.3115(7) b/Å 13.3006(3) 13.3919(9) 13.7868(10) 13.3115(7) c/Å 25.3821(7) 25.2637(17) 19.7327(15) 25.2886(17) alpha/° 90 90 90 90 beta/° 91.964(2) 92.547(3) 109.750(3) 91.892(3) gamma/° 90 90 90 90 V/Å 3 3419.44(2) 3478.1(4) 3574.2(5) 3481.0(4) Z 4 4 4 4 F(000) 1440 1460 1464 1464 Temp./K 200 200 200 200 Restraints 0 12 0 6 Nref 8508 8662 8852 8660 Npar 459 454 463 429 R 0.0343 0.0440 0.0348 0.0782 wR2 0.0970 0.1221 0.0918 0.2361 S 1.06 1.03 1.03 1.03 θ min-max/° 1.6, 28.3 1.6, 28.3 1.8, 28.3 1.6, 28.4 Tot. data 74142 130407 82206 127974 Unique data 8508 8662 8852 8660 Observed data 6832 7392 7194 7645 [I > 2.0 (I)] Rint 0.035 0.021 0.021 0.020 Completeness (%) 100.0 100.0 100.0 100.0 Min. resd. Dens. (e/ Å3) −0.25 −0.40 −0.30 −1.02 Max. resd. Dens. (e/ Å3) 0.31 0.81 0.34 2.29


Page 19 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


19

Host compound 2 consistently crystallizes in the triclinic crystal system and space group P1, whether guest compound is present or not, but none of these solids displays isostructural host packing with any other in the series of complexes. In the 2∙morpholine single crystal experiment, residual electron density indicated that the guest molecule is positionally-disordered but only to a minor extent (this disorder could not be modelled), while during resolution of the 2∙piperidine crystal structure, it was observed that one of the host nitrogen hydrogens is disordered over two positions. On the other hand, the host packing in three of the complexes of host 3, namely 3∙pyridine, 3∙morpholine and 3∙dioxane, was determined to be isostructural. The same crystal system and space group (monoclinic, P21/n) were also favoured for the 3∙piperidine complex, but its unit cell dimensions were significantly different compared with the others. Both inclusions with morpholine and dioxane display disorder: in the former, it was not possible to determine the position of the guest nitrogen-bound hydrogen atom, while the disorder in the latter was around its centre of gravity with significant overlap of the disordered components such that this could not be modelled. As a result, the value of wR2 is abnormally high.

Surprisingly, given the multiple ring structures present in these host compounds, neither 2 nor 3 experienced any significant inter- or intramolecular host−host π−π interactions. However, comparable host−host CH−π contacts involving both host aromatic and methylene hydrogen atoms were observed, while the only hydrogen bonding present was exclusively intramolecular in nature [N−H∙∙∙N, 2 – 2.8680(15) Å, 105.5(10)ᵒ; 3 – 2.8953(15) Å, 104.1(12)ᵒ, in addition to a small number of non-classical host−host ArC−H∙∙∙N hydrogen bonding]. Any minor differences in the packing interactions in these host crystals are possibly as a result of the slight geometrical differences between the two analogues, as stated earlier.

Table S1 (Supplementary Information) contains a summary of the significant host−guest interactions present in the complexes of hosts 2 and 3. Immediately evident is the difference in the number of these interactions present for the two host compounds, with 2 experiencing a significantly larger number with each guest solvent compared with 3. For example, hydrogen bonding is prevalent in each of the four complexes with 2, while 3 participates in only one such interaction, that with piperidine [(host)N−H∙∙∙N−C(guest) 3.264(2) Å, 156.5(1)°], and is absent in the remaining three complexes.



Page 20 of 59

20

Additionally, the mode of guest accommodation was investigated by removing the guest molecules from the packing calculations (using the Mercury software19) and displaying the resultant voids (dark yellow, Figures 6a−f). All the guest compounds in the four complexes of host 2 experience discrete cavity occupation with the exception of pyridine, which is accommodated in multi-directional, infinite and open channels along the directions [100], [111] and [−111]. It has been reported by Tanaka et al20 that the relative thermal stabilities of the complexes of a particular host compound are dependent on whether the guest molecules are accommodated in channels or discrete cavities, with lower stabilities being associated with the former type. The discrimination displayed by 2 for pyridine may therefore be as a result of the channel occupation experienced by this guest compound. Host 3 accommodates each of its guest molecules within well-defined cavities: while two guest molecules occupy each cavity in the 3∙pyridine, 3∙morpholine and 3∙dioxane crystals (all three being isostructural), only one guest molecule is found in each void in the 3∙piperidine complex.


Page 21 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46


21 (a)

(b)

(d)

(e)

(c)

(f)

Figure 6: Voids (dark yellow) present in the (a) 2∙2pyridine, (b) 2∙morpholine, (c) 2∙piperidine, (d) 2∙dioxane, (e) 3∙morpholine (as a representative example of the three isostructural complexes) and (f) 3∙piperidine complexes after the guest molecules were removed from the packing calculation.



Page 22 of 59

22

In the hope of elucidating the contrasting behaviours of the host compounds, we engaged in a computational study in which conformational analyses of the host molecules were performed, along with the calculation of electrostatic potential surfaces.

2.6 Computational analyses

Pyramidal nitrogen atoms are usually configurationally labile owing to low inversion barriers, but inversion may be restricted in certain circumstances,21 potentially creating amino stereocentres. With one exception, the configurations of the N-atoms in the host compounds 2 and 3, both in the absence and presence of guest compounds, were found to be well defined in this study, indicative of arrested inversion of the amino centres in the crystal structures. We are confident about the configurations as the amino hydrogen atoms were located explicitly in difference electron density maps and allowed to refine freely. In only one case was a disordered host amino hydrogen atom evident, but the quality of the data was nevertheless sufficient to allow refining the site-occupancies of its two components of disorder, the resulting two positions being associated with inversion of the N-atom. Consequently, the relative configurations of the pairs of amino stereocentres in 2 and 3 are of interest since syn or anti diastereomers# can arise, with the implication of differing energies.

While the host compounds 2 and 3 in the absence of guest compounds have common features in their crystal structures, there are nevertheless key differences. Variations are also observed in the conformations adopted by 2 and 3 in their host-guest complexes and they furthermore display different selectivities towards potential guest compounds.

These observations

prompted us to perform conformational analyses of the two host structures. The study has entailed a detailed comparison of the compounds in their various crystal structures, as well as a computational investigation at the molecular mechanics and DFT levels in which their conformational distributions and associated energies were determined. The DFT methodology included application of the semi-empirical range-separated hybrid ωB97X-D and ωB97X-V density functionals that capture both short and long-range exchange and correlation

#

The configuration of each amino stereocentre was designated as R or S, with the syn diastereomer corresponding to the R,R or S,S configuration and anti to R,S. The configurations were established in terms of the Cahn–Ingold– Prelog sequence rules with the tricyclic unit being assigned the highest priority, followed by the ethylenediamine chain, the hydrogen atom and lastly, the lone pair. ACS Paragon Plus Environment

Page 23 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


23

interactions. A rigorous assessment22 against fifteen existing density functionals with respect to main group thermochemistry and non-covalent interactions revealed that ωB97X-V was the best functional tested for non-bonded interactions by a significant margin, as well as offering very good performance for thermochemistry. We considered that the application of these functionals in our study was appropriate since dispersion effects are undoubtedly pertinent in host-guest chemistry.

2.6.1 Apohosts 2 and 3

Single crystal X-ray diffraction experiments have revealed that the host compounds 2 and 3 crystallize isostructurally in the absence of guest compounds. The overlaid structures are displayed in Fig. 7 and selected structural features are listed in Table 6 by way of confirmation.

Figure 7. Views of overlaid crystal structures of apohost 2 and apohost 3

(a)

(b)*

(c)*

*Images (b) and (c) display alternative views of the correspondence between the respective tricyclic termini in 2 and 3 as apparent in (a).



Page 24 of 59

24

Table 6. Selected structural parameters and relative energies* for hosts 2 and 3 in their various crystal structures

*𝜔B97X-V/6-311+G(2DF,2P) single point energies were calculated on 𝜔B97X-D/6-31G* optimised geometries.

Apohost 2 and apohost 3 both crystallize as anti diastereomers. The Ph–C–N(R)–C–C–N(S)–C– Ph chains adopt similar conformations, with the torsion angles II – VI varying as -ap, -ap, sc, ap and ap, respectively,23 when considered from the end nearer the R amino stereocentre. Further similarities are evident in their crystal packing where their unit cells show pairs of aligned molecules in each case, but with one molecule rotated with respect to the other along its alignment axis. As a result, an aryl ring of one xanthenyl unit nestles in the fold of the xanthenyl system of its neighbour, while this interaction is reciprocated at the other termini (Fig. 8).


Page 25 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


25

Figure 8. Unit cells of apohost 2 and apohost 3 viewed along the a* axes

apohost 2

apohost 3

However, subtle structural differences can be discerned between xanthene 2 and thioxanthene 3, chiefly in the geometries of the tricyclic rings. The S–C bonds are about 0.4 Å longer compared to O–C, and the C–S–C bond angles 15−16° more acute than C–O–C, distorting the reasonably regular hexagonal geometry of the central ring of the xanthene unit, and resulting in folding about the S---CH2 axis in the thioxanthene system and a boat-like structure for its central ring.

The tricyclic units attached to the respective R amino stereocentres are flatter (folded by 3° in 2 and 10° in 3) while greater folding is found in the units adjacent to the S stereocentres (19° and 33°, respectively#). In the latter case the phenyl groups adopt pseudo axial orientations where they are tilted perpendicularly with respect to the plane of the tricyclic system, while the amino groups are pseudo equatorial and nearly eclipsing the peri C–H bonds. In apohost 3 the ring plane of the pseudo axial phenyl group is further twisted from alignment with the C(9)– N bond than is the case at the other terminus or in 2. Notably, in the less folded second thioxanthenyl unit of apohost 3, the amino group is inclined pseudo axially and phenyl, pseudo equatorially.

#

The folding angle is defined as the angular deviation from planarity for the two planes hinging about the X---CH2 axis. ACS Paragon Plus Environment


Page 26 of 59

26

2.6.2 Structures of compounds 2 and 3 in their host-guest complexes

Variations are observed in the conformations adopted by compounds 2 and 3 in their hostguest complexes (Table 6).

The dixanthenyl compound 2 maintains the anti configuration in its complexes and, with two exceptions, adopts all antiperiplanar arrangements for the torsion angles of its ethylene diamine linkage, with flat or only weakly folded xanthenyl units.

The exceptions occur in the pyridine complex where two conformations of the host molecule are observed in the unit cell. The key difference in the first conformer, 2∙(pyridine)-B, is significant folding in its xanthenyl units, where the C(9) amino groups are pseudo equatorially orientated. In the second, 2∙(pyridine)-A, torsion angles II and VI are synclinal and both xanthenyl systems strongly folded with their amino groups now in pseudo axial situations. The latter conformer ranks higher in energy than the first.

Both the syn and anti configurations were considered for the host molecule in its piperidine complex since the hydrogen atom of one of the amino groups was found to be disordered over two positions with 69:31 syn:anti weightings. The anti configuration 2∙(piperidine)-RS was computed to be of significantly higher energy than its syn counterpart owing to a closer contact between the disordered amino hydrogen atom and an ortho hydrogen on the adjoining C(9) phenyl group in the former (1.98 Å) compared to a larger separation for the latter (2.49 Å).

In contrast to apohost 3 which is anti, host 3 adopts the syn configuration in the four complexes studied, three of whose amino configurations are R,R and the fourth, S,S. The conformations adopted by 3 in the complexes are otherwise very similar. The only significant differences compared to apohost 3 are that both thioxanthenyl units are strongly folded in the complexes and the planes of both C(9) phenyl groups twisted out of alignment with the C(9)–N bonds.

There is reordering of the relative energies of 2 and 3 as manifested in their various host-guest complexes when compared at the MMFF94, B3LYP/6-31G* and ωB97X-V/6-311+G(2df,2p)


Page 27 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


27

levels.# The ranges of conformational energies obtained at the DFT levels were found to be significantly larger for host 2 than for 3.

2.6.3 Conformational features of compounds 2 and 3

The variations observed in the conformations adopted by 2 and 3 in their host-guest complexes prompted us to perform detailed conformational analyses of the two host structures. Our approach was to start with the simple parent xanthene and thioxanthene structures and systematically determine the effects of introducing phenyl and amino substituents on C(9), thereby extending to the hosts 2 and 3.

Conformer distributions for the compounds were determined through molecular mechanics calculations, followed by geometry refinement at the DFT level. Selected geometrical features for the optimized structures of the xanthene and thioxanthene series are highlighted in Table 7 (see also Fig. 9).

#

The latter being single point energies based on 𝜔B97D-V/6-31G* geometries. ACS Paragon Plus Environment


Page 28 of 59

28

Table 7. Computed structural parameters for xanthene, thioxanthene and derivatives determined at the 𝜔B97XV/6-311+G(2df,2p) level Erel/kJ.mol−1

Bond lengths/Å

C–X

Angles/°

C–CH2

C–X–C

Tricyclic ring folding

X…C(9)–Ph

X…C(9)–N

Xanthene

–

1.37

1.51

119

0

127 (H)

9-Phenylxanthene

–

1.37

1.52

118

18

115

138 (H)

9-Aminoxanthene

9-(N-Methylamino)xanthene

9-Amino-9-phenylxanthene

9-(N-Methylamino)-9-phenylxanthene

–

0.00

1.37 1.51/1.52

119

14

137 (H)

117

3.91

1.37

118

23

146 (H)

102

0.00

1.37 1.51/1.52

119

13

135 (H)

119

3.10

1.37/1.38 1.51/1.52

118

24

146 (H)

107

1.51

0.00

1.37

1.53

119

6

132

119

8.22

1.37

1.53

118

22

147

98

0.00

1.37

1.52/1.53

119

6

130

121

5.26

1.37

1.53

119

26

147

103

Thioxanthene

–

1.77

1.51

99

48

82 (Ha) 171 (Hb)

9-Phenylthioxanthene

–

1.77

1.52

102

30

107

147 (H)

0.00

1.77

1.52/1.53

99

49

75 (H)

177

0.03

1.77

1.51/1.52

100

41

161 (H)

92

0.00

1.77

1.51/1.52

100

42

162 (H)

91

4.39

1.77

1.52/1.53

98

51

73 (H)

178

0.00

1.77

1.53/1.54

99

47

85

169

4.95

1.76

1.54

101

40

161

89

0.00

1.76

1.54

101

42

162

87

3.00

1.77

1.54

98

48

83

169

3.00

1.76

1.54

103

13

135

115

119

4/5

130/132

122/119

104/100

1/39

128/162

123/83

9-Aminothioxanthene

9-(N-Methylamino)thioxanthene

9-Amino-9-phenylthioxanthene

9-(N-Methylamino)-9-phenylthioxanthene

Host 2*

–

Host 3*

–

1.36/1.37 1.52/1.53 1.77

1.53/1.55

*Equilibrium conformer, 𝜔B97X-D/6-31G* optimised geometry


−

Page 29 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


29

Figure 9. Selected conformations of xanthene and thioxanthene and C(9)-substituted derivatives Xanthene

Thioxanthene

9-Phenylxanthene

9-Phenylthioxanthene

9-Aminoxanthene

9-Aminothioxanthene

9-Amino-9-phenylxanthene

9-Amino-9-phenylthioxanthene

Conformer 1, Erel = 0.00 kJ∙mol−1




Page 30 of 59

30

Conformer 2, Erel = 8.22 kJ∙mol−1 9-Methylamino-9-phenylxanthene

Conformer 3, Erel = 4.95 kJ∙mol−1 9-Methylamino-9-phenylthioxanthene





Xanthene was calculated to be planar with its central ring describing an irregular hexagon, narrowed towards the oxygen atom apex as a result of shorter C–O bonds compared with C– CH2. By contrast, in thioxanthene the C–S bonds are longer than C–CH2, giving a hexagon widened at the sulfur atom apex. Furthermore, the C–S–C bond angle is 20° more acute than C–O–C, imparting further strain and resulting in folding of the molecule about the S- - -CH2 axis and the central ring assuming a boat shape. Consequently one of the methylene hydrogens is tilted towards the centre of the ring in a pseudo axial orientation while the other is pseudo equatorial and nearly eclipsing the peri C–H bonds. A further effect is greater congestion around the outer face of the C(9) atom in thioxanthene compare to xanthene, with the


Page 31 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


31

separation between the peri hydrogen atoms in thioxanthene being almost 0.7 Å smaller than in xanthene.

Introduction of a 9-phenyl group in xanthene results in some folding of the xanthenyl framework and displacement of the phenyl bond towards the pseudo axial location, giving a O…C(9)–Ph angle of 115°. Similar effects occur with 9-amino and 9-methylamino groups although in each case there is a nearby more folded conformer where the amino group is even closer to a true pseudo axial position. In contrast, when both 9-phenyl and 9-amino or 9methylamino groups are present, comparatively little distortion of the planar parent xanthenyl system is observed in the lowest energy conformers. However, in both instances second higher energy conformers were detected where the xanthenyl systems are more folded and with the amino substituents significantly more inclined towards the pseudo axial sites.

In thioxanthene, a 9-phenyl substituent appreciably reduces the ring folding angle, resulting in a more distinctly pseudo axial orientation of the substituent as evidenced by the more acute S…C(9)–Ph angle (107°), compared with its oxygen analogue. A 9-amino group marginally prefers the pseudo equatorial orientation rather than pseudo axial, but the energy difference relative to the alternative conformer is only 0.03 kJ∙mol−1. The energy difference between the lowest adjacent conformers of 9-methylaminothioxanthene is 4.39 kJ∙mol−1, but in this case the pseudo axial orientation is again preferred while the substituent is pseudo equatorial in the higher energy conformer.

In contrast to the xanthenyl system, there are striking differences between 9-amino-9phenylthioxanthene and its 9-methylamino-9-phenyl analogue. In the first case, the two lowest energy conformers both contain strongly folded thioxanthenyl systems with the amino group orientated pseudo equatorially in the lower energy conformer and pseudo axially in the other. In the 9-methylamino-9-phenyl derivative the lowest energy conformer contains a pseudo axial amino group. Then follow two conformers of equal energy, the first of which is similar to the lowest energy conformer except that the amino group is now pseudo equatorial. In the second, the thioxanthenyl system is substantially flattened and the amino group is only marginally pseudo axial.



Page 32 of 59

32

The structural parameters for the equilibrium conformers identified through conformational searches on host compounds 2 and 3 are shown in Table 7 for comparison and discussed further below.

2.6.4 Conformations of compound 2

A molecular mechanics conformational search carried out for N,N’-bis(9-phenyl-9xanthenyl)ethylenediamine 2 up to a relative energy limit of 100 kJ∙mol−1 gave rise to a large set of conformers including enantiomers and other symmetrically related isomers. The data set was simplified by deleting such isoenergetic conformers. Relative energies, configurations of the N-atoms, torsion angles in the ethylenediamine linkage, the degree of folding in the xanthenyl units, and the orientations of the 9-amino and 9-phenyl groups with respect to xanthenyl system are listed in Table 8.


Page 33 of 59


33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60



Page 34 of 59

34

Table 8. MMFF94 structural parameters for host compound 2 conformers

Table 9. DFT structural parameters for host compound 2 conformers

The set of lowest energy conformers (2-I to 2-IV, Erel 0−7.75 kJ∙mol−1) are characterized by ethylenediamine chains where the torsion angles are antiperiplanar, except for the central bond IV which is synclinal, while one of the adjoining bonds (III or V) is anticlinal in the lowest energy conformer 2-I as well as in 2-III. In the case of 2-II and 2-IV, all the bonds except for IV are antiperiplanar. Conformers displaying an all-antiperiplanar array of torsion angles occur in the range 9.11−11.48 kJ∙mol−1. The xanthenyl units are all nearly planar.

Higher energy conformers are characterized by the increasing occurrence of synclinal bonds and folding in the xanthenyl moieties. The highest energy conformers tend to have synclinal bonds II and VI and both xanthenyl systems significantly folded. In the folded xanthenyl systems, the 9-amino groups are always found to be pseudo axial and the 9-phenyl groups pseudo equatorial. There are only a few instances of synperiplanar bonds, with the first occurring at Erel 25.06 kJ∙mol−1 (2-XVIII) and in all cases limited to bond IV.

For comparison, the structural features of the crystalline structure of host compound 2, as well as where it is involved in various host-guest complexes, are included in Table 8 and illustrated in Fig. 10. All have relatively high energies.


Page 35 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


35

Figure 10. Host compound 2 in crystal structures and after DFT geometry optimisation

Apohost 2

2∙(pyridine)-A

2∙(pyridine)-B

2∙(dioxane)

2∙(piperidine)-RS

2∙(morpholine)-A

2∙(morpholine)-B

2∙(morpholine)-C

2∙(piperidine)-RSa

2∙(pyridine)-B

2∙(piperidine)-RS

Crystal structures

Apohost 2

DFT optimised structures a Isostructural with 2∙(dioxane), 2∙(morpholine)-A, 2∙(morpholine)-B and 2∙(morpholine)-C.

When the geometries of the host crystal structures were optimized without constraints at the molecular mechanics level, Apohost 2-all_M relaxed to the lowest energy conformer 2-I located in the conformational search, six others settled into the all periplanar conformer 2-V, and another reorganized similarly except with a slightly higher energy (2-VII); the xanthenyl systems are generally flattened. In the final case, structure 2∙(pyridine)-A-all_M settled into a significantly higher energy conformer (Erel 42.97 kJ∙mol−1) with bonds II and VI remaining synclinal and the xanthenyl rings folded. The anti configuration of the piperidine complex, 2∙(piperidine)-RS-all_M, was now found to be about 2.4 kJ∙mol−1 lower in energy than its syn analogue.

The molecular mechanics structures were then refined at the DFT level (Table 9), resulting in a more varied array of torsion angles in the ethylenediamine linkage and a greater tendency for ACS Paragon Plus Environment


Page 36 of 59

36

folding in the xanthenyl units among the lower energy conformers. The lowest energy conformer contains a synclinal central bond IV while the remainder of its bonds are antiperiplanar, and both xanthenyl ring systems are only marginally folded. These are followed by conformers containing up to three synclinal or anticlinal bonds as well as combinations of these orientations. In most cases one or both of the xanthenyl units are significantly folded. The orientations of the C(9) amino and phenyl groups as a function of the extent of folding in the xanthenyl ring systems are displayed graphically in Fig. 11. It is apparent that where the xanthenyl systems are folded, the preferred orientations of the C(9) amino and phenyl substituents are respectively pseudo axial and pseudo equatorial, although there are several exceptions. Interestingly, pseudo equatorial amino groups never occur simultaneously on both xanthenyl groups.


Page 37 of 59

37 Ph

N

Linear (Ph)

Linear (N)

160 150

R² = 0.4475

O–C9–X angle/°

140 130 120 110 R² = 0.4836

100 90 80 0

5

10

Ph

N

15 20 Xanthene A fold angle/°

25

Linear (Ph)

30

Linear (N)

160

150

R² = 0.455

140

O–C9–X angle/°

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


130 120 110 100 R² = 0.4909 90 80 0

5

10

15 20 Xanthene B fold angle/°

25

30

Figure 11. Host 2 angular orientations of the C(9) substituents with respect to the O…C(9) axis


35


38

The first conformer displaying an all antiperiplanar array of torsion angles is 2-CXXXII, located at Erel 19.99 kJ∙mol−1; its xanthenyl rings are flat.

The relative energies of the various conformers were compared against the relative configurations of their amino centres (Fig. 12). While the average energy of the anti conformers displayed is about 4 kJ∙mol−1 lower than the syn diastereomers, evidently the energy distributions of the conformers are not strongly influenced by the relative configurations of the amino centres. 60

50 Relative energy (kJ∙mol−1)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 59

40

30

20

10

0 syn configurations Host 2 crystal structures

anti configurations Optimised crystal structures

Figure 12. Host 2 conformer energy distributions as a function of N-atom relative configurations

Fig. 12 also shows that in the DFT calculations the host 2 crystal structures rank as relatively high energy conformers, relaxing to lower energy arrangements when allowed to optimize without constraint. In the latter case, all antiperiplanar torsion angles are again preferred, except for the lowest energy conformer, Apohost 2-all_D (Erel 5.80 kJ∙mol−1), where the torsion angles of bonds IV and V are synclinal and anticlinal respectively, as well as in 2∙(pyridine)-Aall_D, where bonds II and VI are synclinal. The xanthenyl systems are essentially flat in the optimized crystal structures, except for 2∙(pyridine)-A-all_D where there is marked folding of both xanthenyl units. ACS Paragon Plus Environment

Page 39 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


39

The anti configuration of the piperidine complex, 2∙(piperidine)-RS-all_D, was calculated to be only marginally lower in energy than its syn analogue. It is interesting that exclusively the anti diastereomer of host 2 was consistently selected during crystallization with the other guest compounds.

2.6.5 Conformations of compound 3

Conformational analysis of the dithioxanthenyl host 3 resulted in broadly similar trends to those found for 2 at both the molecular mechanics and DFT levels (Tables 10 and 11). The various crystal structures of 3 are included for comparison and are also illustrated in Fig. 13.



Page 40 of 59

40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Page 41 of 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


41

Table 10. MMFF94 structural parameters for host compound 3 conformers

Table 11. DFT structural parameters for host compound 3 conformers

Figure 13. Host compound 3 in crystal structures and after DFT geometry optimisation

Apohost 3

3∙(dioxane)a

3∙(dioxane) and 3∙(piperidine) enantiomer

Crystal structures a Isostructural with 3•(pyridine) and 3•(morpholine)

Apohost 3

Apohost 3 and 3∙(dioxane)

3∙(dioxane), 3∙(pyridine), 3∙(morpholine) and 3∙(piperidine) enantiomer

DFT optimised structures

The lowest energy molecular mechanics conformers (3-I to 3-V, Erel 0−8.86 kJ∙mol−1) are distinguished by antiperiplanar bonds in the ethylenediamine chains, except for the central bond IV which is synclinal, while in most cases one of the adjoining bonds (III or V) is anticlinal. The first conformer with all antiperiplanar torsion angles occurs at Erel 9.58 kJ∙mol−1 (3-VI). Throughout this range the thioxanthenyl ring systems are only slightly folded (

Crystal X-ray diffraction and molecular modelling considerations

Recommend Documents