Supramolecular Chirogenesis in Bis-Porphyrin: Crystallographic Structure and CD Spectra for a Complex with a Chiral Guanidine Derivative

The complexation of (3aR,7aR)-N-(3,5-bis(trifluoromethyl)phenyl)octahydro-2Hbenzo[d]imidazol-2-imine (BTI), as a guest, to ethane-bridged bis(zinc octaethylporphyrin), bis(ZnOEP), as a host, has been studied by means of ultraviolet-visible (UV-Vis) and circular dichroism (CD) absorption spectroscopies, single crystal X-ray diffraction, and computational simulation. The formation of 1:2 host-guest complex was established by X-ray diffraction and UV-Vis titration studies. Two guest BTI molecules are located at the opposite sides of two porphyrin subunits of bis(ZnOEP) host, which is resting in the anti-conformation. The complexation of BTI molecules proceed via coordination of the imine nitrogens to the zinc ions of each porphyrin subunit of the host. Such supramolecular organization of the complex results in a screw arrangement of the two porphyrin subunits, inducing a strong CD signal in the Soret (B) band region. The corresponding DFT computational studies are in a good agreement with the experimental results and prove the presence of 1:2 host-guest complex as the major component in the solution (97.7%), but its optimized geometry differs from that observed in the solid-state. The UV-Vis and CD spectra simulated by using the solution-state geometry and the TD-DFT/ωB97X-D/cc-pVDZ + SMD (CH2Cl2) level of theory reproduced the experimentally obtained UV-Vis and CD spectra and confirmed the difference between the solid-state and solution structures. Moreover, it was shown that CD spectrum is very sensitive to the spatial arrangement of

As a part of our ongoing studies towards application of bis(ZnOEP) for sensing polyfunctionalized chiral organic molecules, here we report the complexation of a chiral guanidine compound, (3aR,7aR)-N-(3,5-bis(trifluoromethyl)phenyl)octahydro-2H-benzo [d]imidazol-2-imine (BTI, Figure 2), with bis(ZnOEP) [36]. Complexation and supramolecular chirogenesis phenomena in the selected host-guest system were studied by means of UV-Vis, CD spectroscopies, and single crystal X-ray analysis. The experimental results were fully rationalized with the aid of computational simulation. Special attention was paid to the reasons why experimental CD spectra in many cases are different from the simulated, especially in the case of chiral porphyrin-based supramolecular systems with a certain degree of conformational flexibility. Schematic representation of complexes of bis(ZnOEP) with both enantiomers of a chiral amine and the corresponding CD signal induction. Ethyl substituents in bis(ZnOEP) are omitted for clarity. "S" and "L" denotes small and large substituents in a chiral amine guest, respectively. See [30][31][32] for the experimental CD spectra.
As a part of our ongoing studies towards application of bis(ZnOEP) for sensing polyfunctionalized chiral organic molecules, here we report the complexation of a chiral guanidine compound, (3aR,7aR)-N- (3,5- Figure 2), with bis(ZnOEP) [36]. Complexation and supramolecular chirogenesis phenomena in the selected host-guest system were studied by means of UV-Vis, CD spectroscopies, and single crystal X-ray analysis. The experimental results were fully rationalized with the aid of computational simulation. Special attention was paid to the reasons why experimental CD spectra in many cases are different from the simulated, especially in the case of chiral porphyrin-based supramolecular systems with a certain degree of conformational flexibility.

Materials and Methods
General methods. UV-Vis absorption spectra were recorded on a Jasco V-730 double-beam spectrophotometer in a 1-cm thermally stabilized screw-cap quartz cuvette with a septum cap. CD spectrum were recorded on a Jasco J-1500 spectrophotometer in a 1-cm screw cap quartz cuvette in analytical-grade CH2Cl2 at 20 °C. Data acquisition was performed in the 375-475-nm range with a scanning rate of 50 nm/min, bandwidth of 2.6 nm, response time of 4 s, and accumulations in 3 scans. 1 H NMR spectra of BTI were recorded on a Bruker Avance III 400 MHz spectrometer. The chemical shifts (δ) are reported in ppm and referenced to a CHCl3 residual peak at 7.26 ppm for 1 H NMR, and a CDCl3 peak at 77.16 ppm for 13 C NMR. HRMS measurement for BTI was performed on an Agilent 6540 UHD Accurate-Mass Q-TOF LC/MS system (Agilent Technologies, Santa Clara, CA, USA) equipped with an AJS-ESI source.
(3aR,7aR)-N- (3,5- Spectroscopic Titrations. All the solutions were prepared and mixed by using properly calibrated analytic glassware (Hamilton ® Gastight syringes, volumetric flasks). All weights were balanced with a Radwag MYA 11.4 microbalance (accuracy ± 6 μg). The concentration of zinc porphyrin was held constant throughout the titration sequence. The titration data were fitted globally by using online software Bindfit [39][40][41]. UV-Vis spectrophotometric titration experiments were performed in analytical-grade CH2Cl2. To a solution of zinc porphyrin, a solution of guest (dissolved in a stock solution of the host to keep the concentration of the host constant) was added portion-wise using a gastight syringe at 20 °C. The changes in the bathochromic shift of the Soret band were monitored at different concentrations of the guest.
Single crystal X-ray analysis. The data was measured using a dual-source Rigaku SuperNova diffractometer equipped with an Atlas detector and an Oxford Cryostream cooling system using mirror-monochromated Mo-Kα radiation (λ = 0.71073 Å). Data collection and reduction for all complexes were performed using the program CrysAlisPro [42] and the Gaussian face-index absorption correction method was applied [42]. All structures were solved with Direct Methods (SHELXS) [43][44][45] and refined by full-matrix least squares based on F 2 using SHELXL-2013 [43][44][45]. Non-hydrogen atoms were as-

Materials and Methods
General methods. UV-Vis absorption spectra were recorded on a Jasco V-730 doublebeam spectrophotometer in a 1-cm thermally stabilized screw-cap quartz cuvette with a septum cap. CD spectrum were recorded on a Jasco J-1500 spectrophotometer in a 1-cm screw cap quartz cuvette in analytical-grade CH 2 Cl 2 at 20 • C. Data acquisition was performed in the 375-475-nm range with a scanning rate of 50 nm/min, bandwidth of 2.6 nm, response time of 4 s, and accumulations in 3 scans. 1 H NMR spectra of BTI were recorded on a Bruker Avance III 400 MHz spectrometer. The chemical shifts (δ) are reported in ppm and referenced to a CHCl 3 residual peak at 7.26 ppm for 1 H NMR, and a CDCl 3 peak at 77.16 ppm for 13 C NMR. HRMS measurement for BTI was performed on an Agilent 6540 UHD Accurate-Mass Q-TOF LC/MS system (Agilent Technologies, Santa Clara, CA, USA) equipped with an AJS-ESI source.
Materials. Bis(ZnOEP) was prepared as described in [36]. BTI was prepared by intramolecular cyclization of 1-((1R,2R)-2-aminocyclohexyl)-3-(3,5-bis(trifluoromethyl)phenyl)thiourea [37] following the experimental procedure described in [38]. Spectroscopic Titrations. All the solutions were prepared and mixed by using properly calibrated analytic glassware (Hamilton ® Gastight syringes, volumetric flasks). All weights were balanced with a Radwag MYA 11.4 microbalance (accuracy ± 6 µg). The concentration of zinc porphyrin was held constant throughout the titration sequence. The titration data were fitted globally by using online software Bindfit [39][40][41]. UV-Vis spectrophotometric titration experiments were performed in analytical-grade CH 2 Cl 2 . To a solution of zinc porphyrin, a solution of guest (dissolved in a stock solution of the host to keep the concentration of the host constant) was added portion-wise using a gastight syringe at 20 • C. The changes in the bathochromic shift of the Soret band were monitored at different concentrations of the guest.
Single crystal X-ray analysis. The data was measured using a dual-source Rigaku SuperNova diffractometer equipped with an Atlas detector and an Oxford Cryostream cooling system using mirror-monochromated Mo-K α radiation (λ = 0.71073 Å). Data collection and reduction for all complexes were performed using the program CrysAlisPro [42] and the Gaussian face-index absorption correction method was applied [42]. All structures were solved with Direct Methods (SHELXS) [43][44][45] and refined by full-matrix least squares based on F 2 using SHELXL-2013 [43][44][45]. Non-hydrogen atoms were assigned anisotropic displacement parameters unless stated otherwise. Hydrogen atoms were placed in idealized positions and included as riding. Isotropic displacement parameters for all H atoms were constrained to multiples of the equivalent displacement parameters of their parent atoms with U iso (H) = 1.2 U eq (parent atom). Enhanced rigid bond restraints (RIGU) [46,47] with standard uncertainties of 0.001 Å 2 were applied for several atom pairs as well as distance restraints (DFIX). Positional disorder of the trifluoromethyl (CF 3 ) groups is observed. Split positions were assigned with isotropic displacement parameters: site occupancy refinement converged to 54.2(6)% to 45.8(6)% with the sum of the site occupancies of both alternative positions constrained to unity. The X-ray single crystal data, experimental details, and CCDC number (2051302) are given below.
Crystal data for the 1:2 complex of bis(ZnOEP) and BTI: CCDC-2051302, C 104 H 120 F 12 N 14 Zn 2 , M = 1924.87 gmol −1 , purple plate, 0.10 × 0.08 × 0.03 mm 3 , triclinic, space group P1 Computational details. Structural optimization was performed using resolution of identity (RI) approximation [48][49][50], PB86 functional [51,52] with D3 dispersion correction [53] and def2-SV(P) basis set [54] implemented in Turbomole 7.0 [55], which showed good agreement with experimental data [56][57][58], as in a gas phase and in solvent. To include solvent effect the COSMO solvent model [59] was used. To confirm that the optimized geometry corresponds to a local minimum, the respective vibrational frequencies were calculated using the same program and level of theory. To get a more accurate energy value for Gibbs free energies calculations, a single point calculation was done using the RI-BP86/def2-TZVP [60] level of theory and COSMO solvent model (ε = 8.93). The optimized geometries of supramolecular host-guest complexes are given in Geometries.xyz (provided in the Supplementary Materials).
The UV-Vis and CD spectra were simulated using the Gaussian16 [61] software and TD-DFT method [62][63][64]. For spectra simulations, the ωB97X-D functional [65] and cc-pVDZ basis set [66][67][68] with the SMD solvent model [69] were used, since this level of theory showed good agreement with experimental data [56,57,70]. The first 10 excited states were calculated in order to ensure that the B band region of the spectrum was covered. The corresponding data are given in the Supplementary Materials.
The UV-Vis and CD spectra and host-guest geometries were visualized using GaussView 6.0.16 [71]. For plotting the simulated spectra, a bandwidth of 0.04 eV was used due to its best agreement with the experimental spectra. The rotatory strengths were calculated on the basis of dipole velocity formalism.

Absorption UV-Vis and CD Spectroscopy
As reported before, bis(ZnOEP) adopts a syn-conformation ( Figure 3) in non-coordinating solvents (e.g., CH 2 Cl 2 ) because of strong intramolecular π-π interactions between two porphyrin subunits [72]. Complexation with an external ligand results in the conformational switch to the corresponding anti form, which also causes dramatic changes in the UV-Vis spectra [31]. Similar to previously studied amine ligands [30][31][32], a portion-wise addition of BTI (Figure 3) resulted in a noticeable bathochromic shift of the Soret band and its split into two distinct absorption peaks at 426 and 437 nm that clearly indicates the formation of 1:2 host-guest complex with bis(ZnOEP) resting in the anti-conformation ( Figure 2). As additional evidence, a Q(1,0) band at 559 nm was noticeably enhanced in comparison to the ligand free bis(ZnOEP) [31]. The resultant UV-Vis spectrum obtained at the final point of spectroscopic titration is shown on Figure 4A and consists of the following wellresolved absorption peaks, λ max , nm (log ε): 426 (5.46), 437 (5.47), 561 (4.49), and 597 (4.00). Curve fitting of the absorbance change observed during the spectroscopic titration with the 1:2 binding isotherm [39] yielded two association constants K 1 = (0.51 ± 0.01)·10 3 M −1 , K 2 = (3.1 ± 0.1)·10 3 M −1 at 293 K for the first and the second ligation events, respectively (Figure 3, Figures S1-S3 and Table S1, see the Supplementary Materials). More than a 6-fold increase of the second association constant is indicative of a highly positive cooperativity of the complexation process as it was found previously for other amine guests [73]. consists of the following well-resolved absorption peaks, λmax, nm (log ε): 426 (5.46), 437 (5.47), 561 (4.49), and 597 (4.00). Curve fitting of the absorbance change observed during the spectroscopic titration with the 1:2 binding isotherm [39] yielded two association constants K1 = (0.51 ± 0.01)·10 3 M −1 , K2 = (3.1 ± 0.1)·10 3 M −1 at 293 K for the first and the second ligation events, respectively (Figure 3, Figures S1-S3 and Table S1, see the Supplementary Materials). More than a 6-fold increase of the second association constant is indicative of a highly positive cooperativity of the complexation process as it was found previously for other amine guests [73]. The CD spectrum of the resulting 1:2 host-guest complex ( Figure 4B) was measured in CH2Cl2 at the end point of UV-Vis titration, after addition of the 1300-fold excess of BTI. This corresponds to 97% conversion of free bis(ZnOEP) host into the corresponding complex, and was calculated based on the values of the corresponding association constants and initial concentrations of the host and guest. While the parent bis-porphyrin is achiral and thus being CD silent, addition of BTI ligand induced a strong optical activity in the Soret transition region ( Figure 4B). The observed CD profile consists of four Cotton effects (CEs): two positive peaks at 412 and 438 nm (with the intensities of 45 and 52 M −1 ·cm −1 , correspondingly), and two negative peaks at 426 and 451 nm (with the intensities of −106.5 and −13.5 M −1 ·cm −1 , correspondingly). Surprisingly, such a complicated CD profile is contrastingly distinguishable from more simple CD spectra of bis(ZnOEP) induced by conventional chiral amines [30][31][32]. Apparently, this is a result of the more complex structure of BTI in comparison to monodentate guests. To understand the origin of the observed chirogenic phenomenon, the corresponding computational studies were undertaken (see Section 3.3).

Single Crystal X-ray Structure
Crystallographic data provided an additional proof of the structure of 1:2 complex between bis(ZnOEP) and BTI, resting in the anti-conformation in the solid-state ( Figure  5). Complexation proceeds via coordination of the imine nitrogen of BTI ligands to the zinc ions of bis(ZnOEP), with an average Zn-N bond distance of 2.18 Å. Commonly to zinc porphyrin complexes, zinc is penta-coordinated and slightly shifted out of the mean porphyrin plane towards the imine nitrogen of BTI ligand. Coordination of two guest molecules occurs from the opposite sides of bis(ZnOEP). In the solid-state, two porphyrin units are arranged nearly parallel to each other, with the Zn-Cmeso-Cmeso-Zn dihedral angle of 179.0°, hence indicating only a slight anticlockwise turn of the porphyrin moieties. The CD spectrum of the resulting 1:2 host-guest complex ( Figure 4B) was measured in CH 2 Cl 2 at the end point of UV-Vis titration, after addition of the 1300-fold excess of BTI. This corresponds to 97% conversion of free bis(ZnOEP) host into the corresponding complex, and was calculated based on the values of the corresponding association constants and initial concentrations of the host and guest. While the parent bis-porphyrin is achiral and thus being CD silent, addition of BTI ligand induced a strong optical activity in the Soret transition region ( Figure 4B −1 , correspondingly). Surprisingly, such a complicated CD profile is contrastingly distinguishable from more simple CD spectra of bis(ZnOEP) induced by conventional chiral amines [30][31][32]. Apparently, this is a result of the more complex structure of BTI in comparison to monodentate guests. To understand the origin of the observed chirogenic phenomenon, the corresponding computational studies were undertaken (see Section 3.3).

Single Crystal X-ray Structure
Crystallographic data provided an additional proof of the structure of 1:2 complex between bis(ZnOEP) and BTI, resting in the anti-conformation in the solid-state ( Figure 5). Complexation proceeds via coordination of the imine nitrogen of BTI ligands to the zinc ions of bis(ZnOEP), with an average Zn-N bond distance of 2.18 Å. Commonly to zinc porphyrin complexes, zinc is penta-coordinated and slightly shifted out of the mean porphyrin plane towards the imine nitrogen of BTI ligand. Coordination of two guest molecules occurs from the opposite sides of bis(ZnOEP). In the solid-state, two porphyrin units are arranged nearly parallel to each other, with the Zn-C meso -C meso -Zn dihedral angle of 179.0 • , hence indicating only a slight anticlockwise turn of the porphyrin moieties.

Single Crystal X-ray Structure
Crystallographic data provided an additional proof of the structure of 1:2 complex between bis(ZnOEP) and BTI, resting in the anti-conformation in the solid-state ( Figure  5). Complexation proceeds via coordination of the imine nitrogen of BTI ligands to the zinc ions of bis(ZnOEP), with an average Zn-N bond distance of 2.18 Å. Commonly to zinc porphyrin complexes, zinc is penta-coordinated and slightly shifted out of the mean porphyrin plane towards the imine nitrogen of BTI ligand. Coordination of two guest molecules occurs from the opposite sides of bis(ZnOEP). In the solid-state, two porphyrin units are arranged nearly parallel to each other, with the Zn-Cmeso-Cmeso-Zn dihedral angle of 179.0°, hence indicating only a slight anticlockwise turn of the porphyrin moieties.

DFT Modelling of the Complex
The obtained crystal structure was used as a starting point for further geometry optimization. Three standard protocols were used as follows: (1) refining only the positions of hydrogens and fluorine atoms [74,75], (2) full optimization in the gas phase [74,[76][77][78], and (3) full optimization in dichloromethane (COSMO solvent model) [76,77,79,80]. Optimization was performed using the RI-PB86-3D/def2-SV(P) level of theory followed by further simulation of the UV-Vis and CD spectra by using the ωB97X-D/cc-pVDZ level of theory with the SMD solvent model (see Figure 6 and Table S2 in the Supplementary Materials).

DFT Modelling of the Complex
The obtained crystal structure was used as a starting point for further geometry optimization. Three standard protocols were used as follows: (1) refining only the positions of hydrogens and fluorine atoms [74,75], (2) full optimization in the gas phase [74,[76][77][78], and (3) full optimization in dichloromethane (COSMO solvent model) [76,77,79,80]. Optimization was performed using the RI-PB86-3D/def2-SV(P) level of theory followed by further simulation of the UV-Vis and CD spectra by using the ωB97X-D/cc-pVDZ level of theory with the SMD solvent model (see Figure 6 and Table S2 in the Supplementary Materials). The UV-Vis spectra calculated on the basis of the crystal structure (λmax at 391 and 399 nm) and optimized geometries (λmax at 401 and 415 nm) are blue shifted in comparison to the experimentally obtained spectrum (λmax at 426 and 437 nm) ( Figure 6A). Additionally, it is of note that the absorption profile of the crystal structure-based spectrum shows a non-resolved split of the Soret band with only a shoulder at 391 nm due to the close proximity of the simulated electronic transitions. However, the optimized geome-  The UV-Vis spectra calculated on the basis of the crystal structure (λ max at 391 and 399 nm) and optimized geometries (λ max at 401 and 415 nm) are blue shifted in comparison to the experimentally obtained spectrum (λ max at 426 and 437 nm) ( Figure 6A). Additionally, it is of note that the absorption profile of the crystal structure-based spectrum shows a nonresolved split of the Soret band with only a shoulder at 391 nm due to the close proximity of the simulated electronic transitions. However, the optimized geometries in both the gas phase and the solution give nearly the same absorption profile with a well-resolved split Soret band, which is similar to the experimental spectrum.
In CD spectra, which are more sensitive to any geometry changes, the spectrum simulated using the crystal structure has three CEs: a positive peak of +670.0 cgs at λ max at 380 nm, a negative peak of -2062.2 cgs at λ max at 396 nm, and a positive peak of +1752.3 cgs at λ max at 403 nm ( Figure 6B). This spectrum is somewhat similar to the experimental one ( Figure 4B), except the difference in the intensities of the positive bands and an absence of the low-energy negative band at 451 nm. As in the case of UV-Vis spectra, the CD spectral profiles simulated by using the geometries optimized in the gas phase and in dichloromethane solution are essentially the same ( Figure 6B). Both spectra have four CEs: two positive peaks at 393 (393) and 490 (410) nm and two negative peaks at 409 (402) and 416 (417) nm in a gas phase and in dichloromethane (in brackets). However, the calculated intensities and shape of the bands are different from the experimental one.
To explain the differences in the calculated and experimental spectra, we assumed that the experimental spectrum represents the combined contribution of various conformers or even differently organized host-guest complexes. We attempted to define these species, since X-ray analysis usually defines only the most energetically favorable conformer. In addition to the lowest energy conformer A (which also dominates in the solid-state, according to X-ray analysis), three 1:2 complexes (B-D) differing by the spatial orientation of the guest molecules and two 1:1 complexes differing by the coordination mode to the second porphyrin moiety were built and their geometries were fully optimized. The calculated Gibbs free energies of these complexes are presented in Figure 7. In the complex B, two BTI molecules are placed asymmetrically at the opposite sides of the bis(ZnOEP) host. In the complexes C and D, the guests are located at the same side of bis(ZnOEP); however, both BTI molecules are facing each other either by the octahydrobenzo[d]imidazole fragments (complex C) or by the bis(trifluoromethyl)phenyl fragments (complex D). In addition to 1:2 complexes, two possible 1:1 complexes with clockwise and anticlockwise positions of two porphyrin subunits were also modelled. In the 1:1 complexes E and F, the host-guest interaction occurs via a two-point coordination mode in a tweezer fashion by placing the BTI molecule between two porphyrin moieties. As the most plausible binding modes, the corresponding tweezer conformation is fixed either by simultaneous coordination of two nitrogens of BTI with two zinc ions of bis(ZnOEP) (complex E) or by interaction of imino nitrogen and one fluorine atom of the CF 3 group with zinc ions of bis(ZnOEP) (complex F, Figure 7). All of the corresponding energies and geometries are given in Table S3 and Geometries.xyz (see the Supplementary Materials).
According to the Boltzmann distribution, the lowest energy complex A is a dominant species in solution (97.7%), whilst another complex B makes up only 1%, with the Gibbs free energy being by 2.8 kcal mol −1 higher than that of the complex A. The Gibbs free energies of the complexes C and D are even higher by 4.5 and 22.7 kcal mol −1 , which is quite reasonable considering the fact that the approach of two BTI molecules from the same side of bis(ZnOEP) is sterically hindered by the porphyrin's ethyl peripheral substituents. Formation of the 1:1 complexes E and F is also unfavorable, since their Gibbs free energies are 2.5 and 11.0 kcal mol −1 higher as compared to the major complex A. Based on these data, it is obvious that the formation of other complexes could not be the reason why the simulated spectra differ from the experimental one.
As the most plausible binding modes, the corresponding tweezer conformation is fixed either by simultaneous coordination of two nitrogens of BTI with two zinc ions of bis(ZnOEP) (complex E) or by interaction of imino nitrogen and one fluorine atom of the CF3 group with zinc ions of bis(ZnOEP) (complex F, Figure 7). All of the corresponding energies and geometries are given in Table S3 and Geometries.xyz (see the Supplementary Materials). According to the Boltzmann distribution, the lowest energy complex A is a dominant species in solution (97.7%), whilst another complex B makes up only 1%, with the Gibbs free energy being by 2.8 kcal mol -1 higher than that of the complex A. The Gibbs free energies of the complexes C and D are even higher by 4.5 and 22.7 kcal mol −1 , which is quite reasonable considering the fact that the approach of two BTI molecules from the same side of bis(ZnOEP) is sterically hindered by the porphyrin's ethyl peripheral sub- In order to understand the differences with the experimental data, porphyrin geometries obtained from the crystal structure and optimized in solvent were compared and two major distinctions were found. In particular, one of the porphyrin planes is more deformed, with the C β -C β -C βopp -C βopp dihedral angles being 18.1 • and −15.2 • (optimization in solution) and 9.4 • and −9.3 • (crystal structure), and porphyrin planes are shifted in respect to each other by 17.0 • (based upon the Zn-C meso -C meso -Zn dihedral angle), as compared to the crystal structure geometry ( Figure 8A and Table 1). Additionally, it turned out that solvation plays an insignificant role, resulting in the minor conformational changes found for the complexes optimized as in the gas phase and in CH 2 Cl 2 . For example, in the complex optimized in a gas phase, the Zn atoms became a bit closer to the average porphyrin planes (N-Zn-N opp angle), but the C β -C β -C βopp -C βopp and Zn-C meso -C meso -Zn dihedral angles remained almost unchanged. Therefore, it was reasonable to conclude that the differences observed between CD spectra, calculated using the geometries optimized in solution, and the crystal structure are attributed to altering the spatial position of the two porphyrin units. Indeed, it is known that the CD spectra of bis-porphyrins are highly sensitive to the orientation of porphyrin rings relative to each other [30][31][32]. Furthermore, it was previously established that the CD amplitude has a parabolic-like dependence on the dihedral angle between the coupling electronic transitions, with zero values at 0 • and 180 • and a maximum value at around 70 • [81]. In the case of the porphyrin chromophores, there are two B electronic transitions orientated along the corresponding meso (5-15 and 10-20) positions ( Figure 8B) and any directional deviation from the coplanar conformation makes these transitions optically active. In the crystal structure, both porphyrin planes are situated almost on the same line, with the dihedral angle Zn-C meso -C meso -Zn being 179.0 • , which can be attributed to a slight anticlockwise orientation. In the structure optimized in dichloromethane, two porphyrin planes are orientated clockwise, with the dihedral angle Zn-C meso -C meso -Zn equaling −163.5 • .
tional changes found for the complexes optimized as in the gas phase and in CH2Cl2. For example, in the complex optimized in a gas phase, the Zn atoms became a bit closer to the average porphyrin planes (N-Zn-Nopp angle), but the Cβ-Cβ-Cβopp-Cβopp and Zn-Cmeso-Cmeso-Zn dihedral angles remained almost unchanged. Therefore, it was reasonable to conclude that the differences observed between CD spectra, calculated using the geometries optimized in solution, and the crystal structure are attributed to altering the spatial position of the two porphyrin units. Indeed, it is known that the CD spectra of bis-porphyrins are highly sensitive to the orientation of porphyrin rings relative to each other [30][31][32]. Furthermore, it was previously established that the CD amplitude has a parabolic-like dependence on the dihedral angle between the coupling electronic transitions, with zero values at 0° and 180° and a maximum value at around 70° [81]. In the case of the porphyrin chromophores, there are two B electronic transitions orientated along the corresponding meso (5-15 and 10-20) positions ( Figure 8B) and any directional deviation from the coplanar conformation makes these transitions optically active. In the crystal structure, both porphyrin planes are situated almost on the same line, with the dihedral angle Zn-Cmeso-Cmeso-Zn being 179.0°, which can be attributed to a slight anticlockwise orientation. In the structure optimized in dichloromethane, two porphyrin planes are orientated clockwise, with the dihedral angle Zn-Cmeso-Cmeso-Zn equaling −163.5°.    To confirm this hypothesis unambiguously, a relaxed coordinate scan along the Zn-C meso -C meso -Zn dihedral angle with the step of 2 • using the RI-PB86-3D/def2-SV(P) level of theory and COSMO solvent model was carried out (corresponding energies are given in Table S4, see the Supplementary Materials). Further, for all these structures, the corresponding CD spectra using the ωB97X-D/cc-pVDZ level of theory with the SMD solvent model were simulated (corresponding data are given in Table S5, see the Supplementary Materials). The spectrum calculated was average weighted and based on the conformers' electronic energies (entropies and vibrational energies are not taken into account), as shown in Figure 9A. In this spectrum, the intensity of the 4th CE decreased by 63 cgs, and the intensities of the 3rd, 2nd, and 1st CEs increased by 158, 23, and 210 cgs, respectively, as compared to the CD spectrum of complex A in solution. These values better match the experimental data; however, the intensities of the 2nd and 1st CEs are still relatively small (+265 and −335 cgs). solvent model were simulated (corresponding data are given in Table S5, see the Supplementary Materials). The spectrum calculated was average weighted and based on the conformers' electronic energies (entropies and vibrational energies are not taken into account), as shown in Figure 9A. In this spectrum, the intensity of the 4th CE decreased by 63 cgs, and the intensities of the 3rd, 2nd, and 1st CEs increased by 158, 23, and 210 cgs, respectively, as compared to the CD spectrum of complex A in solution. These values better match the experimental data; however, the intensities of the 2nd and 1st CEs are still relatively small (+265 and −335 cgs). In turn, the energy scan showed that the potential energy surface is flat and continues for the range of −167.5° to −151.5°, with the energies varying within just 0.4 kcal mol −1 ( Figure 9B and Table S4, see the Supplementary Materials), which makes difficult to determine the exact minimum. Nevertheless, the minimum found by the geometry optimization process (and confirmed by frequency calculations) corresponds to the dihedral angle of -163.5° and it is 0.2 kcal mol −1 higher than in the case of the complex with the dihedral angle of −155.5°. Such a small energy difference is within an accuracy error of the DFT method and could be a result of the numerical noise. Therefore, it is plausible to assume that the Zn-Cmeso-Cmeso-Zn dihedral angle of bis(ZnOEP) complex in solution is about -155.5°, since the UV-Vis and CD spectra of this structure ( Figure 10) has the best match with the experimentally measured spectrum ( Figure 4). All other simulated spectra are presented in Figure S4, see the Supplementary Materials. In turn, the energy scan showed that the potential energy surface is flat and continues for the range of −167.5 • to −151.5 • , with the energies varying within just 0.4 kcal mol −1 ( Figure 9B and Table S4, see the Supplementary Materials), which makes difficult to determine the exact minimum. Nevertheless, the minimum found by the geometry optimization process (and confirmed by frequency calculations) corresponds to the dihedral angle of −163.5 • and it is 0.2 kcal mol −1 higher than in the case of the complex with the dihedral angle of −155.5 • . Such a small energy difference is within an accuracy error of the DFT method and could be a result of the numerical noise. Therefore, it is plausible to assume that the Zn-C meso -C meso -Zn dihedral angle of bis(ZnOEP) complex in solution is about −155.5 • , since the UV-Vis and CD spectra of this structure ( Figure 10) has the best match with the experimentally measured spectrum ( Figure 4). All other simulated spectra are presented in Figure S4, see the Supplementary Materials.

Conclusions
In the present work, the complexation of bis(trifluoromethyl)phenyl)octahydro-2Hbenzo[d]imidazol-2-imine (BTI), as a guest, with bis(ZnOEP), as a host, was studied from the experimental and theoretical points of view. It was found that the host-guest complexation ratio is 1:2 and the BTI binding occurs via coordination of the imine group to

Conclusions
In the present work, the complexation of bis(trifluoromethyl)phenyl)octahydro-2Hbenzo[d]imidazol-2-imine (BTI), as a guest, with bis(ZnOEP), as a host, was studied from the experimental and theoretical points of view. It was found that the host-guest complexation ratio is 1:2 and the BTI binding occurs via coordination of the imine group to the zinc ions of bis(ZnOEP). In agreement with the crystallographic results and DFT study, in solution the main host-guest complex is a supramolecule, in which the two guest molecules are located symmetrically at the opposite sides of bis(ZnOEP) host.
Complexation of the guests to bis(ZnOEP) causes the formation of a screw conformation in the bis-porphyrin host due to the shift of the porphyrin planes relative to each other, which results in low-energy shifts of the porphyrin electronic transitions and the appearance of a strong CD signal in the Soret band region. Although the simulated UV-Vis spectra based on solid-state geometry and structure optimized in the CH 2 Cl 2 solution are similar, the CD spectra are more sensitive to geometry changes and differ drastically. The relative orientation of porphyrin planes changes upon solvation with the resultant complex, adopting a clockwise screw, with the Zn-C meso -C meso -Zn dihedral angle being ca −155.5 • . The CD spectrum simulated for this spatial orientation is a good match with the experimental data and showed four clearly observed Cotton effects in the Soret band region induced by the chirogenic process of asymmetry transfer from a chiral guest to an achiral host.
This study is one of the rare examples of comprehensive CD analysis of chirality induction in bis-porphyrins caused by external chiral ligands, which can be a benchmark approach for the rationalization of supramolecular chirogenesis in bis-porphyrins. Furthermore, the obtained results demonstrate the necessity of careful consideration of all external and internal factors that influence the supramolecular organization of complex to attain the best match between experimental and simulated CD spectra.