True mol­ecular conformation and structure determination by three-dimensional electron diffraction of PAH by-products potentially useful for electronic applications

The true mol­ecular conformation and the crystal structure of benzo[e]di­naphtho­[2,3-a;1′,2′,3′,4′-ghi]fluoranthene, 7,14-di­phenyl­naphtho­[1,2,3,4-cde]bis­anthene and 7,16-di­phenyl­naphtho­[1,2,3,4-cde]heli­anthrene were determined ab initio by 3D electron diffraction. All three mol­ecules are remarkable polycyclic aromatic hydro­carbons. The mol­ecular conformation of two of these com­pounds could not be determined via classical spectroscopic methods due to the large size of the mol­ecule and the occurrence of multiple and reciprocally connected aromatic rings. The mol­ecular structure of the third mol­ecule was previously considered provisional. These com­pounds were isolated as by-products in the synthesis of similar products and were at the same time nanocrystalline and available only in very limited amounts. 3D electron diffraction data, taken from submicrometric single crystals, allowed for direct ab initio structure solution and the unbiased determination of the inter­nal mol­ecular conformation. Detailed synthetic routes and spectroscopic analyses are also discussed. Based on many-body perturbation theory simulations, benzo[e]di­naphtho­[2,3-a;1′,2′,3′,4′-ghi]fluoranthene may be a promising candidate for triplet–triplet annihilation and 7,14-di­phenyl­naphtho­[1,2,3,4-cde]bis­anthene may be a promising candidate for inter­molecular singlet fission in the solid state.

1. Introduction

Polycyclic aromatic hydro­carbons (PAHs) are organic mol­ecules containing a series of conjugated carbon rings. The term `PAH' is generally restricted to com­pounds consisting of only C and H atoms, and com­prising two or more aromatic rings bonded in various arrangements (Lawal, 2017). Early syntheses of PAHs (Anschütz, 1886; Scholl et al., 1910; Scholl & Meyer, 1932) attracted considerable inter­est due to their opto­electronic properties, which led to the creation of a number of functionalized com­pounds based on PAH backbones (Buchlovič et al., 2013; Ko et al., 2018; Castro et al., 2019). In particular, in the mid-1900s, Clar and co-workers synthesized and identified a number of PAHs and related com­pounds with unique physical and chemical properties (Clar et al., 1964, 1981; Clar & Schmidt, 1975).

In most cases, the identity of the newly synthesized organic com­pounds can be determined from combustion analysis and by a combination of NMR, IR, UV and mass spectrometry. In particular, a com­parison of the UV and photoelectron (PE) spectral data provides information about the purity and helps to distinguish `isotopic PAHs', i.e. com­pounds with different shapes but identical UV band positions (Clar & Schmidt, 1976, 1979; Clar et al., 1981).

It is often impossible to characterize PAHs by single-crystal X-ray diffraction (XRD) because of the difficulty in obtaining suitable sufficiently large crystals. Therefore, the final confirmation is generally attained by powder XRD analysis. How­ever, when there are many possible mol­ecular structures that can match a particular stoichiometry, it may be difficult to determine the exact conformation based on spectroscopic data alone. Furthermore, PXRD is intrinsically limited to one-dimensional data and hampered by the overlap of independent reflections, which may become rather severe for unit-cell parameters longer than 20 Å. As a consequence of this unsatisfactory state of affairs, the conformation of several large PAHs mentioned in the literature is uncertain or unknown, and many more are considered suspect (Clar et al., 1981; Fetzer, 2000, 2007; Wilkes, 2010).

In recent years, electron diffraction (ED) has become a routine option for crystal structure determination when crystals sufficiently large for single-crystal XRD cannot be grown. The key development was the establishment of procedures for collecting and recombining 3D electron diffraction (3D ED) data (Kolb et al., 2007; Wan et al., 2013; Gemmi et al., 2019; Nannenga & Gonen, 2019). The 3D ED method has been used for the structure determination of hundreds of previously unknown structures, including inorganic materials (Kaiukov et al., 2020; Krysiak et al., 2021; Kapaca et al., 2021), small-mol­ecule organics (Andrusenko et al., 2019; Jones et al., 2019; Brázda et al., 2019; Cui et al., 2020; Levine et al., 2020; Bruhn et al., 2021; Hall et al., 2021; Andrusenko et al., 2021; Papi et al., 2021), peptides and macromolecules (Sawaya et al., 2016; Krotee et al., 2018; Lanza et al., 2019; Xu et al., 2019; Warmack et al., 2019). It is noteworthy that ab initio crystal structure determination does not require information about the mol­ecular conformation, but only a rough estimation of the atomic content of the unit cell. In this article, we report the true mol­ecular conformation and the crystal structure determined by the 3D ED method for three PAHs belonging to two different systems.

Fluoranthenes are one of the most studied classes of PAHs (Haritash & Kaushik, 2009). They are also known as `non-alternant PAHs' because they contain six-membered benzene rings fused with an additional five-membered ring (Gupte et al., 2016). Fluoranthene-based PAHs serve as a basic unit in preparing chromophores for organic light-emitting diodes (OLEDS) and field-effect transistors (Saranya et al., 2011). Benzo[e]di­naphtho­[2,3-a;1′,2′,3′,4′-ghi]fluoranthene (I) is form­ed by pyrolysis together with two other hydro­carbons, namely, benzo[a]phenanthro[9′,10′-c]tetra­cene (II) and di­benzo[a,k]naphtho[1,2,3,4-ghi]perylene (III). Fig. 1 illustrates a stepwise model of the corresponding synthetic route des­cribed by Clar and co-workers (Clar et al., 1964). The mol­ecular conformation of II is supported by several spec­tro­scopic data, its reactive behaviour and, most conclusively, by an alternative synthesis with Zn dust that produces II exclusively (Fig. S1). Conversely, the mol­ecular conformations of I and III were only reported as tentative (Clar et al., 1964). Spectroscopic data for II and III are presented and discussed in the supporting information.

Figure 1 A stepwise model of the synthesis of benzo[e]di­naphtho­[2,3-a;1′,2′,3′,4′-ghi]fluoranthene (I) and related products (II and III). The mol­ecule of I was considered provisional before its structure determination by 3D ED, whereas the mol­ecular structure of III is still speculative.

The second system studied in this article involves both bis­anthene and heli­anthrene com­pounds. Initially, 4,11-di­phenyl­bis­anthene (IV) was synthesized by Sauvage, who only determined the maximum positions of the longwave absorption bands (Sauvage, 1947a,b). Renewed inter­est in this com­pound was sparked in the 1970s, as it was shown to be one of the few mol­ecules to fluoresce at IR wavelengths (Maulding, 1970; Rauhut et al., 1975). Following the procedure of Sauvage's synthesis (Sauvage, 1947a,b), we attempted to synthesize new IR emitters with a high fluorescent quantum yield starting from helianthrone. Fig. 2 illustrates a stepwise model of the corresponding synthetic route. During this reaction, in addition to 4,11-diphenylbis­anthene (IV), a number of by-products occur, among which only 7,16-di­phenylheli­anthrene (V) was identified previously (Arabei & Pavich, 2004). This com­pound appears in a greater yield when N₂ is bubbled through the system during the reaction (Fort, 2010). Two additional by-products of the reaction, VI and VII, were not identified via spectroscopic data, but appear to form via a Diels–Alder condensation reaction. The mol­ecular structures of by-products VI and VII have hitherto remained unknown, due to the large size of the mol­ecules and the occurrence of multiple and reciprocally connected aromatic rings.

Figure 2 A stepwise model of the synthesis of 4,11-di­phenyl­bis­anthene (IV). The by-product V was identified and determined in Arabei & Pavich (2004). The mol­ecular conformation and crystal structure of by-products VI and VII were unknown prior to their determination by the 3D ED method. For each product of the reaction, the formula, mol­ecular weight (M) and yield (in bold) are reported.

Here, we present the ab initio structure determination of com­pounds I, VI and VII, obtained through 3D ED data. This method also allowed for the confirmation of the predicted mol­ecular conformation for com­pound I and for the true mol­ecular identification of com­pounds VI and VII. Moreover, detailed synthetic routes and spectroscopic analyses of all the synthetic products are reported. Having solved the structures of these com­pounds, we eventually use many-body perturbation theory calculations within the GW approximation and the Bethe–Salpeter equation (BSE) (Golze et al., 2019; Blase et al., 2020) to assess their potential usefulness in photovoltaics.

2. Methods

2.1. Synthesis

Samples I, II and III were obtained starting from a Friedel–Crafts reaction (Friedel & Crafts, 1877) between dibenzo[g,p]chrysene (10 g) and o-toluyl chloride, with AlCl₃ as catalyst. The two reactants were connected by a ketone group, either 2- or 3-(o-tolu­yl)dibenzo[g,p]chrysene (12 g). The 1- and 4-positions are excluded on steric grounds. After Elbs pyrolysis (15 g, 420–430 °C, 10 min under CO₂), fractional crystallization and chromatography showed that three products had formed (Fig. 1): I (0.46 g, violet–red needles, m.p. 348–350 °C), II (1.5 g, thick orange–yellow needles, m.p. 258–259 °C) and III (0.1 g, fibrous orange–yellow needles, m.p. 282°C).

Multiple syntheses of samples IV, V, VI and VII (Fig. 2 and Fig. S2) started from adding phenyl–Li to helianthrone, which can be prepared on a 100 g basis (Scholl & Mansfeld, 1910). The resulting diol was pyrolyzed with ten times the qu­antity of Cu powder for 45 min at 360 °C under N₂. Gradient sublimation gave an excellent yield of compound V and a smaller yield of IV, both not free of each other. There were also two major by-products: VI (m.p. >427 °C) and VII (m.p. 426–427 °C). Repeated gradient sublimation and chromatography gave four pure (>99%) products. Besides VI and VII, we eventually identified at least four other by-products, but these were not obtained free of each other.

2.2. Spectroscopy

UV spectroscopic analyses were conducted using an Agilent Cary 300 spectrophotometer. UV spectra were collected at room tem­per­ature using benzene (above 275 nm) or cyclo­hexane (above 200 nm) solvents (Fig. 3, Figs. S4–S7 and Tables S1–S3). Over 150 UV spectra had to be recorded to ensure the identity and homogeneity of the various fractions. Fractional sublimation was obtained using a Heraeus tube oven ROK 3/30 with a Kelvitron REK 19-20 controller and a melting-point correction. Fluorescence (FL) spectroscopic analyses were carried out with an Aminco–Bowman SPF-500 spectro­fluoro­meter (Table S4). Compound I was insufficiently soluble in cyclo­hexane solvent, so for the determination of reliable extinction coefficients, only benzene was used as solvent. Compound VI is well soluble in benzene and less soluble in cyclo­hexane. Compound VII is well soluble in both solvents. The gas-phase PE spectrum for com­pound I was obtained using a PerkinElmer PS-18 spectrometer with a Helectros Developments photon source (Fig. 3(b)). Samples were also analysed by positive ion mode Matrix-assisted laser desorption/ionization on a Bruker Daltonics ultrafleXtreme II mass spectrometer using Colloidal Graphite as the matrix (Figs. S8–S10).

Figure 3 (a)/(c)/(e) UV and (b)/(d)/(f) simulated GW@PBE0 spectra of (a)/(b) com­pound I, (c)/(d) com­pound VI and (e)/(f) com­pound VII. On the UV spectra, numbers indicate zoomed areas. Comparison of the simulated GW@PBE0 spectrum (in blue) with the experimental PE spectrum (in purple) of com­pound I are highlighted (in (b)). Frontier orbitals of all three species are also visualized. Gaussian broadening of 0.2 eV was applied to the com­puted ionization energies to simulate the resolution of the experiment.

2.3. TEM microscopy and 3D electron diffraction

High-angle annular dark-field scanning transmission electron microscopy (HAADF–STEM) imaging and ED data were recorded with a Zeiss Libra 120 TEM operating at 120 kV and equipped with a LaB₆ thermionic source. 3D ED was performed in STEM mode after defocusing the beam. Therefore, the beam for the 3D ED experiments was almost parallel when crossing the sample (Köhler illumination). The TEM was set in STEM mode, but the beam, which is normally very convergent in this configuration, was purposely defocused (Benner & Probst, 1994). The actual beam convergence was 60 µrad. ED patterns were collected with a beam size of about 150 nm in diameter, obtained using a 5 µm C2 condenser aperture. Data were recorded by a single-electron ASI MEDIPIX detector (Nederlof et al., 2013). An extremely low dose illumination, corresponding to 0.01 e⁻ Å⁻² s⁻¹, was used in order to avoid beam damage. The total dose during data collection depends on many experimental parameters, such as the number of frames, the exposure time and the image tracking mode, and is therefore different for each data collection. A rough estimation of the total dose during a step­wise data collection is in a range of about 1–5 e⁻ Å⁻². When working with such low doses, the availability of a hybrid-pixel single-electron detector is crucial for acquiring a diffraction pattern with a satisfactory signal-to-noise ratio.

3D ED acquisitions were performed rotating the sample around the TEM goniometer axis in steps of 1°, in total tilt ranges up to 120° (Figs. 4–6). The exposure time per frame was 1 s. The camera length was 180 mm, allowing a resolution in real space of up to 0.7 Å. During the experiment, the beam was precessed around the optical axis by an angle of 1°. Precession was obtained using a Nanomegas Digistar P1000 device. After each tilt, the crystal position was tracked by STEM imaging and a diffraction pattern was acquired. All data acquisitions were performed at room tem­per­ature.

Figure 4 (a) HAADF–STEM image of a typical crystal of com­pound I, used for 3D ED data acquisition. (b)/(c)/(d) Reconstructed 3D ED data viewed along the main crystallographic directions. Note that these are projections of a 3D diffraction volume and not 2D oriented ED patterns, and therefore axial extinctions are not visible.

Figure 5 (a) HAADF–STEM image of a typical crystal of com­pound VI, used for 3D ED data acquisition. (b)/(c)/(d) Reconstructed 3D ED data viewed along the main crystallographic directions.

Figure 6 (a) HAADF–STEM image of a typical crystal of com­pound VII, used for 3D ED data acquisition. (b)/(c)/(d) Reconstructed 3D ED data viewed along the main crystallographic directions. Note, in the case of com­pound VII, the HAADF image shows more than one single crystal, while the 3D ED data were collected from a single-crystal area (upper right) that appear as thin as the other two com­pounds.

3D ED data were analysed using the software PETS2.0 (Palatinus et al., 2019) and ADT3D (Kolb et al., 2011). Structure determination was carried out by standard direct methods (SDM) as implemented in the software SIR2014 (Burla et al., 2015), using a fully kinematical approximation (I_hkl proportional to |F_hkl|²). Kinematical least-squares structure refinement was performed with the software SHELXL (Sheldrick, 2015) using electron atomic scattering factors. Structures were finally refined considering dynamical effects (Palatinus et al., 2013, 2015) using the software JANA2006 (Petrícek et al., 2014) and assuming a simple platelet shape.

2.4. Computational details

Geometry relaxations of the crystal structures of com­pounds I, VI and VII, and of the single-mol­ecule structure of com­pound I, were conducted using density functional theory (DFT) with the FHI-aims code (Blum et al., 2009). The Perdew–Burke–Ernzerhof (PBE) (Perdew et al., 1996, 1997) generalized gradient approximation was combined with the Tkatchenko–Scheffler (TS) (Tkatchenko & Scheffler, 2009) pairwise dispersion method. Tight numerical settings and tier 2 basis sets were used. Full unit-cell relaxation was performed until no force com­ponent on any atom exceeded 0.01 eV Å⁻¹.

GW+BSE calculations for the crystal structures of com­pounds I, VI and VII were performed using the BerkeleyGW code (Deslippe et al., 2012). Quantum ESPRESSO (Giannozzi, 2009) was used to com­pute the mean-field eigenvectors and eigenvalues, and to generate the mean-field coarse-grid and fine-grid wave functions with the PBE exchange-correlation functional. For com­pound I, we used a coarse k-grid of 8 × 2 × 1 and a fine k-grid of 8 × 4 × 2. For com­pound VI, we used a coarse k-grid of 4 × 1 × 1 and a fine k-grid of 8 × 2 × 2. For com­pound VII, we used a coarse k-grid of 2 × 2 × 2 and a fine k-grid of 4 × 4 × 4. Troullier–Martins norm-conserving pseudopotentials (Troullier & Martins, 1991) were used and the kinetic energy cut-off was set at 50 Ry. About 550 unoccupied bands were included in the GW calculation. The static remainder correction was applied to accelerate the convergence with respect to the number of unoccupied states (Deslippe et al., 2013). The polarizability, inverse dielectric matrix and GW self-energy operator were constructed based on the mean-field eigenvalues and eigenfunctions using the coarse k-point settings. Optical properties, including excitation energies, exciton wave functions and absorption spectra were calculated by solving the BSE within the Tamm–Dancoff approximation (TDA). 24 valence bands and 24 conduction bands were included in the BSE calculation (Figs. S11–S16). Taking the full dielectric matrix as input to screen the attraction between the electron (e) and hole (h), the e–h inter­action kernel was constructed on the coarse k-point grid. To construct the Bethe–Salpeter Hamiltonian, the GW quasi­particle energies and e–h inter­action kernel calculated with coarse k-point settings were inter­polated onto the fine k-point grid. The subsequent diagonalization yielded the excitation energies and wave functions. The exciton wave functions of com­pounds I and VI were converged using supercells of 8 × 4 × 2 and 8 × 2 × 2, respectively, based on the criterion proposed in Liu et al. (2020c). The degree of singlet exciton charge-transfer character (%CT) was calculated by double-Bader analysis (DBA) (Wang et al., 2018; Liu et al., 2020c). The results for rubrene and penta­cene are from Wang et al. (2016), the results for quaterrylene, perylene and tetra­cene are from Wang et al. (2018), the results for anthracene are from Liu et al. (2020c), and the results for terrylene are from Hall et al. (2021).

GW+BSE calculations for the single mol­ecule of com­pound I were conducted using the FHI-aims code, using augmented tier-2 basis sets (Liu et al., 2020a). PBE was used as the DFT starting point. A detailed account of the GW implementation in FHI-aims is provided in Ren et al. (2012) and Caruso et al. (2013). Briefly, the self-energy is first calculated on the imaginary frequency axis and then analytically continued to the real frequency axis. A 16-parameter Pade approximation was used in the analytical continuation (Ren et al., 2012; Van Setten et al., 2015). Using the GW quasiparticle energies, BSE calculations were performed to obtain the singlet and triplet excitation energies (Liu et al., 2020c). Results for the known TTA chromophores, which we com­pare to com­pound I here, are from Wang et al. (2020), in which the same methodology was used. For GW calculations of the gas-phase PE spectra of I, VI and VII, we used the PBE-based hybrid functional (PBE0) (Adamo & Barone, 1999) as the mean-field starting point, with tight numerical settings and tier 4 basis sets. This method was benchmarked in Marom et al. (2012) and Knight et al. (2016), and shown to yield spectra in good agreement with experiments for a variety of organic com­pounds. For com­pounds VI and VII, only calculated GW@PBE0 spectra are presented, in the absence of experimental data.

3. Results and discussion

3.1. Benzo[e]di­naphtho­[2,3-a;1′,2′,3′,4′-ghi]fluoranthene

The mol­ecular conformation of com­pound I was reported as provisional (Clar et al., 1964). New absorption measurements (Fig. S3) match well with those reported by Clar et al. (1964), apart from the spurious long-wavelength band at 584 nm. The alleged com­position from mass spectrometry was C₃₄H₁₈ (Fig. S8), also consistent with the related UV spectrum (Fig. 3(a) and Table S1). The occurrence of cyclo­de­hy­drog­en­ation, with the formation of a five-membered ring, is accom­panied by a red shift of the first UV band by 74.5 nm. In particular, the sharp band at about 440 nm is often found in fluoranthene-type PAHs, and never in alternant cata- or peri-condensed PAHs. The strongest argument in favour of Clar's assignment comes from the PE spectrum, which fits the calculated one (Clar et al., 1981). Also, our GW@PBE0 calculation, based on the structure determination by 3D ED data, is in excellent agreement with experimental PE data (Fig. 3(b)). We note that the com­puted ionization energies are vertical values. We do not consider the relaxation of the atomic positions in response to the electronic excitation, vibrational effects, scattering cross-section effects, detector resolution and other sources of noise in the experiment, which may affect the peak breadth and intensity. The com­parison to experiment is focused on peak positions, with the dashed lines at the experimental peak maxima serving as a guide to the eye. Additionally, this spectrum shows an impurity band at 6.95 eV. This peak is not related to the presence of III, which should instead give a band at about 6.67 eV (Clar et al., 1981).

3D ED data were recorded from three flat crystal fragments of I, with sizes less than 1 µm (Fig. 4(a)). All data sets were consistent with a primitive ortho­rhom­bic unit cell, with parameters a = 5.1 (1), b = 17.7 (4) and c = 23.2 (5) Å (Figs. 4(b)–(d)). Such a cell would likely contain four mol­ecules. Upon inspection of the reciprocal space reconstruction, the extinction rules 0k0: k = 2n and 00l: l = 2n were observed. All 3D ED acquisitions miss information about h00, because this direction is always parallel to the main surface of the platelet and, standing vertical, cannot be sampled inside the TEM goniometer range. The possible space groups for com­pound I were therefore P22₁2₁ (No. 18) or P2₁2₁2₁ (No. 19), both with multiplicity 4.

Structure solution was performed by SDM using the 3D ED data set with the highest angular range and the greatest resolution. The solution eventually converged in the space group P2₁2₁2₁. All but one of the 34 non-H atoms of the mol­ecule were located ab initio. The missing C and all H atoms were generated during least-squares refinement against 3D ED data imposing constraints on the aromatic rings. More details about structure determination and refinement are reported in Table S5.

The structure determination of com­pound I allowed for unequivocal confirmation of the mol­ecular model inferred from spectroscopic data, which was previously considered provisional (Clar et al., 1964). Mol­ecules of I are flat and arranged in four columns per unit cell, which extend along a. Inside each column, mol­ecules form a herringbone stack, parallel to one of the four planes of the {3,10,6} family (Figs. 7(a) and 7(b)). The inter­molecular distance is about 3.5 Å, similar to that observed in terrylene, another PAH recently solved by the 3D ED method and potentially promising for electronic applications (Hall et al., 2021).

Figure 7 The mol­ecular configuration and crystalline structure of (a)/(b) com­pound I, (c)/(d) com­pound VI and (e)/(f) com­pound VII. C atoms are shown in brown and H atoms are shown in light grey.

3.2. 7,14-Di­phenyl­naphtho­[1,2,3,4-cde]bis­anthene and 7,16-di­phenyl­naphtho­[1,2,3,4-cde]heli­anthrene

The two side products VI and VII obtained during the synthesis of IV were separated by repeated gradient sublimation. Compound VI appears as dark-green crystals that do not melt up to 427 °C and shows a deep-red colour in concentrated benzene solution or a carmine colour in dilute solution. While exposed to UV light, it appears violet–blue and has an intense orange–red fluorescence in daylight. Based on mass spectrometry (Fig. S9) and combustion data, VI has the com­position C₄₆H₂₄. The mol­ecular conformation could not be deduced by spectroscopic data alone and no crystal produced was sufficiently large for single-crystal XRD analysis. In the absence of experimental PE data, we present here the com­puted GW@PBE0 spectra of com­pounds VI and VII (Figs. 3(d) and 3(f)).

3D ED data collected on three VI micrometric crystals (Fig. 5(a)) consistently indicated a primitive ortho­rhom­bic cell with parameters a = 9.9 (2), b = 26.5 (5) and c = 20.7 (4) Å (Figs. 5(b)–(d)). Such a cell would likely contain eight VI mol­ecules. Upon inspection of the reciprocal space reconstruction, the extinction rules uniquely indicated the space group Pbca (No. 61), with multiplicity 8. Despite the remarkable com­plexity of the mol­ecule, its structure solution was obtained ab initio by SDM on the basis of 3D ED data. All 46 non-H atoms of the mol­ecule were located automatically in the first Fourier map, eventually revealing the mol­ecular conformation that could not be obtained by spectroscopic data (Fig. 7(c)). More details about the structure determination and refinement are reported in Table S5.

Mol­ecules of VI are flat and pack as a sandwich herringbone structure parallel to the (102) and (10) planes (Fig. 7(d)). The phenyl groups act as nonplanar satellites of the mol­ecule, forming with it an angle of about 66°. The inter­molecular distance is about 3.5 Å.

Apparently, during the synthesis, phenyl–Li acts as a masked benzyne and connects to the reactive diene positions. A literature search supports the 3D ED result. In fact, the dimesityl analogue of VI was synthesized in an unequivocal manner by benzyne addition to 7,14-bis­(mesit­yl)bis­anthene (Konishi et al., 2013), when benzyne was liberated in situ from o-anthranilic acid and isoamyl nitrite. Due to severe twisting about the formal single bond, the phenyl and mesityl groups show little difference; hence, the UV spectra reported by Konishi et al. (2013) agrees with that obtained for com­pound VI (Fig. 3(c)). Compared to IV, lateral annelation shifts the first UV band by 79 nm to the blue (from 684.5 to 613 nm), due to the gain of one Clar sextet. There is a good mirror relationship between the absorption and fluorescence emission, and the Stokes shift is 6 nm.

Compound VII results in ochre crystals with a melting point of 426–427 °C and shows a permanganate colour in solution with benzene or cyclo­hexane. Under UV light it emits a green fluorescence and is moderately stable in daylight. This feature contrasts with the extreme sensitivity of heli­anthrene, whose fluorescence fades in daylight within seconds (Seip & Brauer, 1992). The combustion and mass spectrometry (Fig. S10) analysis pointed to a com­position of C₄₆H₂₆. As for com­pound VI, the mol­ecular conformation could not be deduced by spectroscopic data and no crystal was sufficiently large for single-crystal XRD analysis.

3D ED data were recorded from six crystal fragments with a size of less than 1 µm (Fig. 6(a)). All of them delivered a triclinic unit cell with parameters a = 10.5 (2), b = 11.6 (3), c = 12.8 (3) Å, α = 85.3 (5), β = 76.1 (5) and γ = 84.9 (5)° (Figs. 6(b)–(d)). No possibility of cell centring was envisaged, and no extinction rule pointing to higher symmetry was recognised. Such a cell would conveniently host only two VII mol­ecules. Structure solution of this com­pound was indeed obtained ab initio by SDM in the space group P (No. 2). All 46 non-H atoms of the mol­ecule were automatically spotted in the first Fourier map, allowing also, in this case, the mol­ecular conformation to be established, which could not be obtained by spectroscopic data (Fig. 7(e)). More details about the structure determination and refinement are reported in Table S5.

The two terminal phenyl rings in VII act again as nonplanar satellites, forming an angle of about 65° with the main part of the mol­ecule. Their presence partially breaks the aromaticity of the heli­anthrene group. The nonplanarity of VII allows for both C—H⋯π and π–π inter­actions, leading to the reduction of symmetry down to triclinic. No herringbone motif is present (Fig. 7(f)).

In the corresponding UV spectrum, a first sharp band was observed at 501 nm (Fig. 3(e)). Consistent with the 3D ED results, in going from V to VII, there is a blue shift by 79 nm due to the gain of one Clar sextet in the mol­ecule. The Stokes shift is 16.5 nm (Fig. S7, Tables S2 and S3).

3.3. Electronic properties

Having definitively resolved the mol­ecular and crystal structures of com­pounds I, VI and VII by 3D ED, we are now able to use com­puter simulations to assess their potential usefulness for photovoltaic applications. In particular, we are inter­ested in discovering new materials capable of undergoing singlet fission (SF), the down-conversion of one singlet exciton into two triplet excitons (Smith & Michl, 2010, 2013; Monahan & Zhu, 2015; Michl, 2019), and its reverse process, triplet–triplet annihilation (TTA), the up-conversion of two triplet excitons into one singlet exciton (Baldo et al., 2000; Schmidt & Castellano, 2014; Schulze & Schmidt, 2015). Both of these processes can be harnessed to reduce losses and thus increase the conversion efficiency of solar cells. In a typical solar cell, the absorption of one photon produces one charge carrier. The absorption of a high-energy photon generates a highly excited singlet exciton, which subsequently thermalizes to the lowest excited state, such that the excess photon energy is lost to heat. SF can be potentially utilized to convert that excess photon energy into an additional charge carrier by generating two triplet excitons from one high-energy singlet exciton. Another source of losses in solar cells is that photons with energies below the gap of the absorber cannot be absorbed and their energy is lost. TTA can be utilized to convert two low-energy photons into an additional charge carrier. Hence, supplementing the traditional absorber with SF and TTA materials can significantly enhance the efficiency of solar cells by reducing losses at both the high end and the low end of the solar photon energy spectrum. SF and TTA materials are rare because few chromophores meet the stringent requirements for the excited-state energetics, as detailed below. Most of the known and predicted SF and TTA chromophores are PAHs, mainly acene and perylene derivatives. Therefore, we evaluate the PAHs studied here for these purposes.

To assess the potential of mol­ecular crystals to undergo SF in the solid state, we consider a two-dimensional descriptor (Wang et al., 2018; Liu et al., 2020b,c), as shown in Fig. 8. The primary descriptor, displayed on the x axis, is the thermodynamic driving force for SF, which is the difference between the singlet exciton energy and twice the triplet exciton energy (E_s − 2E_t). A high driving force indicates that a material is likely to undergo SF with a high rate. However, an overly high driving force would lead to losses in solar energy conversion. Therefore, it has been suggested that materials with E_s ≃ 2E_t may be preferable (Wu et al., 2014). Owing to the approximations used in GW+BSE calculations, the values of E_s − 2E_t are systematically underestimated. Hence, we restrict the discussion to qualitative com­parisons between materials.

Figure 8 A two-dimensional descriptor for assessing SF candidates. The thermodynamic driving force for SF (Es − 2Et) is displayed on the x axis and the singlet exciton charge-transfer character (%CT) is displayed on the y axis. The error bars correspond to the range of %CT values obtained by using different hole positions when performing the double-Bader analysis. Compounds I and VI (the latter highlighted in red) are com­pared to some acene and rylene phases.

The secondary descriptor, displayed on the y axis, is the degree of charge-transfer character (%CT) of the singlet exciton wave function. Exciton wave functions have two spatial variables corresponding to the electron and hole probability distributions. If the hole is located on one mol­ecule, the degree of charge-transfer character corresponds to the probability of finding the electron on other mol­ecules. This descriptor is motivated by the growing body of experimental evidence for the involvement of an inter­mediate charge-transfer state in the SF process (Monahan & Zhu, 2015; Chan et al., 2013; Kim et al., 2019; Miyata et al., 2019; Margulies et al., 2017). A singlet exciton with a high degree of charge-transfer character is thought to be favourable for SF (Monahan & Zhu, 2015; Sharifzadeh et al., 2013, 2015; Broch et al., 2018; Hart et al., 2018). To evaluate %CT, we use double-Bader analysis (DBA), an extension of the Bader charge-partitioning scheme, to exciton wave functions with two spatial variables. The %CT is calculated by performing nested sums over the electron distributions obtained for different hole positions within a mol­ecule. The error bars correspond to the range of %CT values obtained for different hole positions within the double-Bader analysis (Wang et al., 2018; Liu et al., 2020c).

In Fig. 8, com­pounds I and VI are com­pared to perylenes and acenes with respect to these two descriptors. Penta­cene has the highest SF driving force of the materials shown here and is known to undergo fast SF with a high triplet yield (Chan et al., 2011; Wilson et al., 2011). SF in tetra­cene is experimentally known to be slightly endoergic (Tomkiewicz et al., 1971; Grumstrup et al., 2010; Burdett et al., 2010; Chan et al., 2012; Burdett & Bardeen, 2012, 2013; Arias et al., 2016). Compound VI has a somewhat higher SF driving force, as well as a higher singlet exciton CT character than tetra­cene. Based on this, it may be a promising candidate for inter­molecular SF in the solid state. Compound I has a lower SF driving force than anthracene and perylene, whose derivatives are known to undergo TTA (Renaud et al., 2013; Jiang et al., 2013; Eaton et al., 2013; Mirjani et al., 2014; Renaud & Grozema, 2015; Le et al., 2016; Würthner et al., 2016). Therefore, it may be able to undergo TTA. The com­puted singlet and triplet excitation energies of com­pound VII are 2.90 and 2.00 eV, respectively. It is thus not a likely candidate for either SF or TTA; therefore, it is not shown in Figs. 8 and 9.

Figure 9 Assessment of com­pound I (highlighted in red) as a TTA candidate, com­pared to several known TTA chromophores. The energy release in the singlet pathway, = 2T1 − S1, is plotted on the x axis and the energy release in the com­peting triplet pathway, = 2T1 − T2, is plotted on the y axis. The dashed line on the left correspond to the of rubrene and the dashed line on the right corresponds to the of pyrene. The region delineated by the two diagonal lines is where most experimentally known TTA chromophores are found in Wang et al. (2020).

To further assess the potential of com­pound I to undergo TTA, we com­pare its excited-state energetics as an isolated mol­ecule to several chromophores experimentally known to undergo TTA (Fig. 9). The com­parison is based on the criteria proposed in Wang et al. (2020). The primary criterion for a chromophore to undergo TTA is that the energy release in the singlet pathway, = 2T₁ − S₁, plotted on the x axis, should be positive to provide thermodynamic driving force, but not overly large to avoid significant energy losses. As discussed in detail in Wang et al. (2020), GW+BSE@PBE systematically underestimates . Therefore, we assess new chromophores based on a com­parison with known TTA chromophores, similar to our procedure for assessing SF candidates. In rubrene, TTA is experimentally known to be approximately isoergic (Cheng et al., 2010). Therefore, we consider the of rubrene as the lower limit for the singlet pathway to be open. Mol­ecules with a smaller than rubrene may be more likely to undergo SF than TTA. This is indicated by the left dashed vertical line in Fig. 9. Of the experimentally known TTA chromophores calculated in Wang et al. (2020), pyrene has the highest . We then consider pyrene as the upper limit, above which the energy loss is too high for efficient TTA. This is indicated by the right vertical dashed line in Fig. 9.

The main process com­peting with TTA is the combination of two triplet excitons into a higher-energy triplet state, T₂, which decays nonradiatively. Hence, a secondary criterion for the quantum yield of TTA to be as high as possible is for the energy release in the triplet pathway, = 2T₁ − T₂, to be as small as possible com­pared to . In Wang et al. (2020), we have found that most of the experimentally known TTA chromophores fall into the range between = and = + 0.32 eV. This is indicated by the two diagonal lines in Fig. 9. Compound I (highlighted in red) is well within this range, between perylene and an anthracene derivative. Therefore, based on energetic considerations, com­pound I may be a promising candidate for TTA. We note, however, that favourable energetics are a necessary but not sufficient condition for good TTA chromophores (Wang et al., 2020).

4. Conclusions

Organic syntheses typically yield several side products, which are generally discarded even if their qu­antity amounts to 20–50% of the total. Such side products are often unwanted. Moreover, the assessment of their potential use is com­plicated by the difficulty of their characterization. Spectroscopic data may be insufficient to unambiguously determine the mol­ecular conformation, especially when mol­ecules are com­plex or the presence of more than one species is suspected. Powder X-ray diffraction requires a sufficient amount of material and may suffer from peak overlap, in particular when the phase of inter­est cannot be properly purified, its unit-cell parameters are long and/or small crystal size causes severe peak broadening. In these cases, 3D electron diffraction emerges as a promising technique for conclusive determination of the mol­ecular conformation and crystal structure, as we have demonstrated here for benzo[e]di­naphtho­[2,3-a;1′,2′,3′,4′-ghi]fluoranthene (com­pound I), 7,14-diphenyl-naphtho­[1,2,3,4-cde]bis­anthene (com­pound VI) and 7,16-di­phenyl­naphtho­[1,2,3,4-cde]heli­anthrene (com­pound VII).

Once the structure is determined, com­puter simulations may be used to predict the electronic and optical properties of side products, in particular when the amount of material initially produced is insufficient for detailed experimental characterization. This can help to assess the potential usefulness of side products for various applications. In some cases, unintended side products may turn out to be as useful as the main reaction products. For example, the second system analysed here com­prises four related mol­ecules. Of these, IV exhibits exceptional optical properties (Gorelenko et al., 1977) and V found practical application as a pigment for visible-light-sensitive actinometers (Brauer et al., 1983). Based on the many-body perturbation theory simulations presented here, com­pounds I and VI may be useful for photovoltaic applications. Compound I may exhibit triplet–triplet annihilation, which may enable harvesting photons with energies below the gap of the absorber in a solar cell. Compound VI may exhibit inter­molecular singlet fission in the solid state, which may enable the harvesting of two charge carriers from one high-energy photon in a solar cell. Compound VII is a wide-gap insulator.

We hope that the results reported in this article will inspire others to pursue further characterization of organic side products by 3D ED and com­puter simulations. This may lead to the discovery of potentially useful com­pounds for various applications. Thus, 3D ED is a promising new avenue for enriching our knowledge of organic synthetic routes and exploiting side products that would otherwise be discarded.

Supporting information

Computing details

For all structures, program(s) used to solve structure: SIR2014 (Burla et al., 2015); program(s) used to refine structure: JANA2006 (Petricek et al., 2014) (Palatinus et al., 2015); molecular graphics: VESTA Momma & Izumi, 2011); software used to prepare material for publication: JANA2006 (Petricek et al., 2014).

(I)

Crystal data

C34H18	Z = 4
Mr = 426.5	F(000) = 888
Orthorhombic, P212121	Dx = 1.344 Mg m^−3
Hall symbol: P 2xab;2ybc;2zac	Electron radiation, λ = 0.0335 Å
a = 5.14 (10) Å	µ = 0 mm^−1
b = 17.7 (4) Å	T = 293 K
c = 23.2 (5) Å	Nanocrystal, dark orange
V = 2110 (80) Å^3	0.001 × 0.0003 × 0.0001 mm

Data collection

Zeiss Libra 120 diffractometer	θmax = 1.2°, θmin = 0.1°
data have been collected by precession–assisted 3D electron diffration scans	h = −6→6
71669 measured reflections	k = −19→20
8927 independent reflections	l = −28→28
1138 reflections with I > 3σ(I)

Refinement

Refinement on F	105 constraints
R[F > 3σ(F)] = 0.172	H-atom parameters constrained
wR(F) = 0.201	Weighting scheme based on measured s.u.'s w = 1/(σ^2(F) + 0.0001F^2)
S = 1.82	(Δ/σ)max = 0.045
8927 reflections	Δρmax = 0.77 e Å^−3
181 parameters	Δρmin = −1.24 e Å^−3
78 restraints	Absolute structure: 2000 of Friedel pairs used in the refinement

Special details

Experimental. 3D electron diffration recorded by TEM and dynamically refined

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	1.0194 (11)	0.2845 (5)	0.5772 (4)	0.0290 (7)*
H1	0.881851	0.245779	0.569831	0.0349*
C2	1.0787 (13)	0.3383 (5)	0.5346 (3)	0.0290 (7)*
H2	0.983346	0.337812	0.496972	0.0349*
C3	1.2717 (15)	0.3927 (5)	0.5449 (3)	0.0290 (7)*
H3	1.314004	0.430955	0.514626	0.0349*
C4	1.4055 (12)	0.3934 (5)	0.5978 (4)	0.0290 (7)*
H4	1.54292	0.432133	0.605152	0.0349*
C5	1.3463 (14)	0.3396 (6)	0.6403 (3)	0.0290 (7)*
C6	1.1533 (15)	0.2852 (5)	0.6300 (3)	0.0290 (7)*
C7	1.0625 (9)	0.2219 (3)	0.67166 (19)	0.0290 (7)*
C8	1.2304 (8)	0.2393 (3)	0.71765 (19)	0.0290 (7)*
C9	1.4301 (11)	0.2924 (5)	0.7321 (4)	0.0290 (7)*
C10	1.5055 (15)	0.3487 (6)	0.6929 (3)	0.0290 (7)*
C11	1.7016 (16)	0.4002 (5)	0.7080 (3)	0.0290 (7)*
H11	1.755371	0.44037	0.680049	0.0349*
C12	1.8223 (12)	0.3955 (5)	0.7622 (4)	0.0290 (7)*
H12	1.96199	0.43215	0.772971	0.0349*
C13	1.7468 (15)	0.3391 (6)	0.8014 (2)	0.0290 (7)*
H13	1.832696	0.335733	0.840113	0.0349*
C14	1.5507 (16)	0.2876 (5)	0.7864 (3)	0.0290 (7)*
C15	0.8811 (8)	0.1179 (3)	0.72719 (18)	0.0290 (7)*
C16	0.7066 (8)	0.0573 (4)	0.7319 (2)	0.0290 (7)*
H16	0.701965	0.026436	0.768081	0.0349*
C17	0.5387 (10)	0.0400 (4)	0.6860 (2)	0.0290 (7)*
H17	0.414347	−0.00316	0.689373	0.0349*
C176	0.5453 (11)	0.0834 (4)	0.6352 (2)	0.0290 (7)*
H176	0.425659	0.071042	0.602512	0.0349*
C18	0.7199 (10)	0.1440 (4)	0.6305 (2)	0.0290 (7)*
H18	0.724619	0.174829	0.594341	0.0349*
C19	0.8879 (9)	0.1613 (4)	0.67644 (19)	0.0290 (7)*
C20	1.2238 (7)	0.1960 (3)	0.76839 (17)	0.0290 (7)*
C21	1.3920 (7)	0.2133 (3)	0.81444 (17)	0.0290 (7)*
C22	1.3854 (6)	0.1699 (3)	0.86530 (16)	0.0290 (7)*
C23	1.2105 (6)	0.1092 (3)	0.87009 (15)	0.0290 (7)*
C24	1.0424 (6)	0.0918 (3)	0.82404 (16)	0.0290 (7)*
H24	0.918038	0.048659	0.827447	0.0349*
C25	1.0491 (7)	0.1353 (3)	0.77317 (16)	0.0290 (7)*
C26	1.2039 (7)	0.0658 (3)	0.92096 (17)	0.0290 (7)*
H26	1.079533	0.022623	0.924374	0.0349*
C27	1.5537 (7)	0.1873 (3)	0.91136 (18)	0.0290 (7)*
H27	1.678101	0.230442	0.907945	0.0349*
C28	1.5470 (8)	0.1439 (4)	0.96223 (19)	0.0290 (7)*
C29	1.7153 (9)	0.1613 (4)	1.0083 (2)	0.0290 (7)*
H29	1.839666	0.204425	1.004844	0.0349*
C30	1.7086 (10)	0.1179 (5)	1.0591 (2)	0.0290 (7)*
H30	1.828189	0.130232	1.091868	0.0349*
C31	1.5337 (10)	0.0571 (5)	1.0639 (2)	0.0290 (7)*
H31	1.529099	0.026269	1.100066	0.0349*
C32	1.3654 (9)	0.0398 (4)	1.0179 (2)	0.0290 (7)*
H32	1.240975	−0.003354	1.021328	0.0349*
C33	1.3722 (8)	0.0832 (4)	0.96702 (18)	0.0290 (7)*

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
?	?	?	?	?	?	?

Geometric parameters (Å, º)

C1—H1	1	C17—H17	1
C1—C2	1.40 (4)	C17—C176	1.40 (5)
C1—C6	1.40 (5)	C176—H176	1
C2—H2	1	C176—C18	1.40 (4)
C2—C3	1.40 (4)	C18—H18	1
C3—H3	1	C18—C19	1.40 (4)
C3—C4	1.40 (5)	C20—C21	1.41 (4)
C4—H4	1	C20—C25	1.41 (4)
C4—C5	1.40 (4)	C21—C22	1.41 (5)
C5—C6	1.40 (4)	C22—C23	1.41 (4)
C5—C10	1.48 (5)	C22—C27	1.41 (4)
C6—C7	1.55 (5)	C23—C24	1.41 (4)
C7—C8	1.40 (4)	C23—C26	1.41 (5)
C7—C19	1.40 (4)	C24—H24	1
C8—C9	1.43 (4)	C24—C25	1.41 (5)
C8—C20	1.40 (5)	C26—H26	1
C9—C10	1.40 (4)	C26—C33	1.41 (4)
C9—C14	1.40 (5)	C27—H27	1
C10—C11	1.40 (4)	C27—C28	1.41 (5)
C11—H11	1	C28—C29	1.41 (4)
C11—C12	1.40 (5)	C28—C33	1.41 (4)
C12—H12	1	C29—H29	1
C12—C13	1.40 (4)	C29—C30	1.41 (5)
C13—H13	1	C30—H30	1
C13—C14	1.40 (4)	C30—C31	1.41 (4)
C15—C16	1.40 (4)	C31—H31	1
C15—C19	1.40 (5)	C31—C32	1.41 (4)
C15—C25	1.40 (4)	C32—H32	1
C16—H16	1	C32—C33	1.41 (5)
C16—C17	1.40 (4)
H1—C1—C2	120	C17—C176—H176	120
H1—C1—C6	120	C17—C176—C18	120.0 (9)
C2—C1—C6	120.0 (12)	H176—C176—C18	120
C1—C2—H2	120	C176—C18—H18	120
C1—C2—C3	120.0 (12)	C176—C18—C19	120.0 (9)
H2—C2—C3	120	H18—C18—C19	120
C2—C3—H3	120	C7—C19—C15	120.0 (8)
C2—C3—C4	120.0 (9)	C7—C19—C18	120.0 (9)
H3—C3—C4	120	C15—C19—C18	120.0 (12)
C3—C4—H4	120	C8—C20—C21	120.0 (12)
C3—C4—C5	120.0 (13)	C8—C20—C25	120.0 (8)
H4—C4—C5	120	C21—C20—C25	120.0 (9)
C4—C5—C6	120.0 (12)	C20—C21—C22	120.0 (12)
C4—C5—C10	112.6 (14)	C21—C22—C23	120.0 (8)
C6—C5—C10	127.4 (9)	C21—C22—C27	120.0 (12)
C1—C6—C5	120.0 (9)	C23—C22—C27	120.0 (9)
C1—C6—C7	112.9 (12)	C22—C23—C24	120.0 (9)
C5—C6—C7	127.1 (11)	C22—C23—C26	120.0 (8)
C6—C7—C8	97.4 (13)	C24—C23—C26	120.0 (12)
C6—C7—C19	142.6 (8)	C23—C24—H24	120
C8—C7—C19	120.0 (9)	C23—C24—C25	120.0 (12)
C7—C8—C9	139.5 (7)	H24—C24—C25	120
C7—C8—C20	120.0 (12)	C15—C25—C20	120.0 (9)
C9—C8—C20	100.5 (11)	C15—C25—C24	120.0 (12)
C8—C9—C10	121.0 (13)	C20—C25—C24	120.0 (8)
C8—C9—C14	119.0 (9)	C23—C26—H26	120
C10—C9—C14	120.0 (12)	C23—C26—C33	120.0 (12)
C5—C10—C9	107.6 (14)	H26—C26—C33	120
C5—C10—C11	132.4 (9)	C22—C27—H27	120
C9—C10—C11	120.0 (12)	C22—C27—C28	120.0 (12)
C10—C11—H11	120	H27—C27—C28	120
C10—C11—C12	120.0 (9)	C27—C28—C29	120.0 (12)
H11—C11—C12	120	C27—C28—C33	120.0 (8)
C11—C12—H12	120	C29—C28—C33	120.0 (9)
C11—C12—C13	120.0 (12)	C28—C29—H29	120
H12—C12—C13	120	C28—C29—C30	120.0 (13)
C12—C13—H13	120	H29—C29—C30	120
C12—C13—C14	120.0 (12)	C29—C30—H30	120
H13—C13—C14	120	C29—C30—C31	120.0 (9)
C9—C14—C13	120.0 (9)	H30—C30—C31	120
C16—C15—C19	120.0 (8)	C30—C31—H31	120
C16—C15—C25	120.0 (9)	C30—C31—C32	120.0 (10)
C19—C15—C25	120.0 (12)	H31—C31—C32	120
C15—C16—H16	120	C31—C32—H32	120
C15—C16—C17	120.0 (9)	C31—C32—C33	120.0 (13)
H16—C16—C17	120	H32—C32—C33	120
C16—C17—H17	120	C26—C33—C28	120.0 (9)
C16—C17—C176	120.0 (13)	C26—C33—C32	120.0 (12)
H17—C17—C176	120	C28—C33—C32	120.0 (8)

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
C1—H1···C18	**	2.43	3.18 (10)	130.76
C16—H16···C24	1.00	2.47	2.81 (8)	99.51
C17—H17···C12^i	1.00	2.44	3.38 (10)	156.54
C18—H18···C1	1.00	2.50	3.18 (10)	125.03
C24—H24···C16	1.00	2.47	2.81 (8)	99.37

Symmetry code: (i) −x+2, y−1/2, −z+3/2.

(VI)

Crystal data

C46H24	Z = 8
Mr = 576.7	F(000) = 2400
Orthorhombic, Pbca	Dx = 1.402 Mg m^−3
Hall symbol: -P -2xab;-2ybc;-2zac	Electron radiation, λ = 0.0335 Å
a = 9.9 (2) Å	µ = 0 mm^−1
b = 26.5 (5) Å	T = 293 K
c = 20.7 (4) Å	Nanocrystal, blue
V = 5460 (180) Å^3	0.001 × 0.001 × 0.0001 mm

Data collection

Zeiss Libra 120 diffractometer	θmax = 1.2°, θmin = 0.1°
data have been collected by precession–assisted 3D electron diffration scans	h = −11→11
132615 measured reflections	k = −33→33
9870 independent reflections	l = −25→25
1463 reflections with I > 3σ(I)

Refinement

Refinement on F	141 constraints
R[F > 3σ(F)] = 0.249	H-atom parameters constrained
wR(F) = 0.284	Weighting scheme based on measured s.u.'s w = 1/(σ^2(F) + 0.0001F^2)
S = 2.11	(Δ/σ)max = 1.056
9870 reflections	Δρmax = 0.27 e Å^−3
208 parameters	Δρmin = −0.32 e Å^−3
120 restraints

Special details

Experimental. 3D electron diffration recorded by TEM and dynamically refined

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	−0.431 (3)	0.2203 (3)	−0.1858 (4)	0.0016 (8)*
H1	−0.413136	0.197668	−0.223426	0.0019*
C2	−0.437 (3)	0.2730 (3)	−0.1946 (4)	0.0016 (8)*
H2	−0.430941	0.28765	−0.238973	0.0019*
C3	−0.451 (3)	0.3051 (3)	−0.1407 (4)	0.0016 (8)*
H3	−0.447945	0.342505	−0.14655	0.0019*
C4	−0.468 (2)	0.2843 (3)	−0.0787 (4)	0.0016 (8)*
H4	−0.482469	0.307048	−0.040826	0.0019*
C5	−0.466 (2)	0.2316 (3)	−0.0702 (4)	0.0016 (8)*
H5	−0.475327	0.216951	−0.025955	0.0019*
C6	−0.451 (3)	0.1996 (2)	−0.1239 (4)	0.0016 (8)*
C7	−0.0556 (8)	0.06042 (18)	0.0768 (2)	0.0016 (8)*
H7	−0.052168	0.098081	0.074594	0.0019*
C8	0.0321 (8)	0.0341 (2)	0.1184 (3)	0.0016 (8)*
H8	0.097686	0.053117	0.145695	0.0019*
C9	0.0272 (8)	−0.0190 (2)	0.1215 (3)	0.0016 (8)*
H9	0.089408	−0.037666	0.150939	0.0019*
C10	−0.0653 (7)	−0.04586 (18)	0.0829 (2)	0.0016 (8)*
H10	−0.068808	−0.083521	0.085114	0.0019*
C11	−0.1530 (7)	−0.01955 (14)	0.04131 (19)	0.0016 (8)*
C12	−0.1482 (7)	0.03359 (14)	0.03824 (19)	0.0016 (8)*
C13	−0.2359 (6)	0.05990 (13)	−0.00337 (18)	0.0016 (8)*
C14	−0.2310 (7)	0.11302 (15)	−0.0065 (2)	0.0016 (8)*
H14	−0.165347	0.132033	0.020832	0.0019*
C15	−0.3187 (7)	0.13932 (14)	−0.0480 (2)	0.0016 (8)*
H15	−0.315223	0.176985	−0.050193	0.0019*
C16	−0.4112 (7)	0.11251 (12)	−0.0865 (2)	0.0016 (8)*
C17	−0.4160 (6)	0.05938 (10)	−0.08346 (16)	0.0016 (8)*
C18	−0.3284 (6)	0.03308 (10)	−0.04188 (15)	0.0016 (8)*
C19	−0.3332 (6)	−0.02008 (10)	−0.03881 (15)	0.0016 (8)*
C20	−0.2456 (6)	−0.04638 (12)	0.00277 (18)	0.0016 (8)*
C21	−0.2504 (7)	−0.09951 (14)	0.0058 (2)	0.0016 (8)*
H21	−0.188296	−0.118156	0.035322	0.0019*
C22	−0.3430 (7)	−0.12633 (14)	−0.0327 (2)	0.0016 (8)*
H22	−0.346383	−0.163992	−0.030521	0.0019*
C23	−0.4306 (6)	−0.10003 (12)	−0.0743 (2)	0.0016 (8)*
C24	−0.4257 (6)	−0.04689 (10)	−0.07732 (16)	0.0016 (8)*
C25	−0.5231 (7)	−0.12684 (14)	−0.1128 (2)	0.0016 (8)*
C26	−0.5134 (6)	−0.02059 (9)	−0.11891 (15)	0.0016 (8)*
C27	−0.5085 (6)	0.03256 (9)	−0.12197 (15)	0.0016 (8)*
C28	−0.4988 (7)	0.13881 (13)	−0.1281 (2)	0.0016 (8)*
C29	−0.5913 (7)	0.11199 (13)	−0.1666 (2)	0.0016 (8)*
C30	−0.6789 (8)	0.13829 (16)	−0.2082 (3)	0.0016 (8)*
H30	−0.675503	0.175956	−0.210364	0.0019*
C31	−0.7715 (9)	0.11148 (17)	−0.2467 (3)	0.0016 (8)*
H31	−0.833618	0.130131	−0.27618	0.0019*
C32	−0.7763 (8)	0.05835 (15)	−0.2436 (2)	0.0016 (8)*
H32	−0.841903	0.03934	−0.270952	0.0019*
C33	−0.6886 (7)	0.03205 (12)	−0.20207 (19)	0.0016 (8)*
C34	−0.5962 (6)	0.05886 (11)	−0.16355 (18)	0.0016 (8)*
C35	−0.6059 (6)	−0.04740 (11)	−0.15742 (18)	0.0016 (8)*
C36	−0.6935 (7)	−0.02110 (13)	−0.19900 (19)	0.0016 (8)*
C37	−0.7860 (8)	−0.04791 (16)	−0.2375 (2)	0.0016 (8)*
H37	−0.848196	−0.02926	−0.266989	0.0019*
C38	−0.7909 (8)	−0.10103 (18)	−0.2344 (3)	0.0016 (8)*
H38	−0.856508	−0.120047	−0.26175	0.0019*
C39	−0.7032 (8)	−0.12734 (17)	−0.1929 (3)	0.0016 (8)*
H39	−0.706663	−0.165009	−0.190707	0.0019*
C40	−0.6107 (7)	−0.10053 (13)	−0.1544 (2)	0.0016 (8)*
C41	−0.6202 (15)	−0.2719 (4)	−0.1014 (6)	0.0016 (8)*
H41	−0.700471	−0.293382	−0.09199	0.0019*
C42	−0.6257 (15)	−0.2195 (4)	−0.0903 (6)	0.0016 (8)*
H42	−0.70776	−0.204093	−0.070543	0.0019*
C43	−0.5156 (17)	−0.1887 (3)	−0.1072 (7)	0.0016 (8)*
C44	−0.3974 (15)	−0.2107 (4)	−0.1321 (5)	0.0016 (8)*
H44	−0.318653	−0.188954	−0.143416	0.0019*
C45	−0.3903 (15)	−0.2633 (4)	−0.1412 (5)	0.0016 (8)*
H45	−0.306309	−0.278973	−0.158618	0.0019*
C46	−0.5017 (17)	−0.2939 (3)	−0.1256 (7)	0.0016 (8)*
H46	−0.496516	−0.331292	−0.13185	0.0019*

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
?	?	?	?	?	?	?

Geometric parameters (Å, º)

C1—H1	1	C22—C23	1.41 (3)
C1—C2	1.41 (5)	C23—C24	1.41 (5)
C1—C6	1.41 (5)	C23—C25	1.41 (3)
C2—H2	1	C24—C26	1.41 (3)
C2—C3	1.41 (4)	C25—C40	1.41 (3)
C3—H3	1	C26—C27	1.41 (5)
C3—C4	1.41 (5)	C26—C35	1.41 (3)
C4—H4	1	C27—C34	1.41 (3)
C4—C5	1.41 (5)	C28—C29	1.41 (3)
C5—H5	1	C29—C30	1.41 (3)
C5—C6	1.41 (4)	C29—C34	1.41 (5)
C7—H7	1	C30—H30	1
C7—C8	1.41 (3)	C30—C31	1.41 (3)
C7—C12	1.41 (3)	C31—H31	1
C8—H8	1	C31—C32	1.41 (5)
C8—C9	1.41 (5)	C32—H32	1
C9—H9	1	C32—C33	1.41 (3)
C9—C10	1.41 (3)	C33—C34	1.41 (3)
C10—H10	1	C33—C36	1.41 (5)
C10—C11	1.41 (3)	C35—C36	1.41 (3)
C11—C12	1.41 (5)	C35—C40	1.41 (5)
C11—C20	1.41 (3)	C36—C37	1.41 (3)
C12—C13	1.41 (3)	C37—H37	1
C13—C14	1.41 (5)	C37—C38	1.41 (5)
C13—C18	1.41 (3)	C38—H38	1
C14—H14	1	C38—C39	1.41 (3)
C14—C15	1.41 (3)	C39—H39	1
C15—H15	1	C39—C40	1.41 (3)
C15—C16	1.41 (3)	C41—H41	1
C16—C17	1.41 (5)	C41—C42	1.41 (5)
C16—C28	1.41 (3)	C41—C46	1.41 (5)
C17—C18	1.41 (3)	C42—H42	1
C17—C27	1.41 (3)	C42—C43	1.41 (4)
C18—C19	1.41 (5)	C43—C44	1.41 (5)
C19—C20	1.41 (3)	C44—H44	1
C19—C24	1.41 (3)	C44—C45	1.41 (5)
C20—C21	1.41 (5)	C45—H45	1
C21—H21	1	C45—C46	1.41 (4)
C21—C22	1.41 (3)	C46—H46	1
C22—H22	1
H1—C1—C2	120.06	C22—C23—C25	120.0 (12)
H1—C1—C6	120.06	C24—C23—C25	120.0 (7)
C2—C1—C6	119.9 (9)	C19—C24—C23	120.0 (7)
C1—C2—H2	120.05	C19—C24—C26	120.0 (12)
C1—C2—C3	119.9 (12)	C23—C24—C26	120.0 (7)
H2—C2—C3	120.04	C23—C25—C40	120.0 (12)
C2—C3—H3	120	C24—C26—C27	120.0 (7)
C2—C3—C4	120.0 (14)	C24—C26—C35	120.0 (12)
H3—C3—C4	120	C27—C26—C35	120.0 (7)
C3—C4—H4	120	C17—C27—C26	120.0 (7)
C3—C4—C5	120.0 (8)	C17—C27—C34	120.0 (12)
H4—C4—C5	120	C26—C27—C34	120.0 (7)
C4—C5—H5	120	C16—C28—C29	120.0 (12)
C4—C5—C6	120.0 (12)	C28—C29—C30	120.0 (12)
H5—C5—C6	120	C28—C29—C34	120.0 (8)
C1—C6—C5	120.0 (14)	C30—C29—C34	120.0 (8)
H7—C7—C8	120	C29—C30—H30	120
H7—C7—C12	120	C29—C30—C31	120.0 (12)
C8—C7—C12	120.0 (12)	H30—C30—C31	120
C7—C8—H8	120	C30—C31—H31	120
C7—C8—C9	120.0 (8)	C30—C31—C32	120.0 (8)
H8—C8—C9	120	H31—C31—C32	120
C8—C9—H9	120	C31—C32—H32	120
C8—C9—C10	120.0 (8)	C31—C32—C33	120.0 (8)
H9—C9—C10	120	H32—C32—C33	120
C9—C10—H10	120	C32—C33—C34	120.0 (12)
C9—C10—C11	120.0 (12)	C32—C33—C36	120.0 (7)
H10—C10—C11	120	C34—C33—C36	120.0 (7)
C10—C11—C12	120.0 (7)	C27—C34—C29	120.0 (7)
C10—C11—C20	120.0 (12)	C27—C34—C33	120.0 (12)
C12—C11—C20	120.0 (7)	C29—C34—C33	120.0 (7)
C7—C12—C11	120.0 (8)	C26—C35—C36	120.0 (12)
C7—C12—C13	120.0 (12)	C26—C35—C40	120.0 (7)
C11—C12—C13	120.0 (7)	C36—C35—C40	120.0 (7)
C12—C13—C14	120.0 (7)	C33—C36—C35	120.0 (7)
C12—C13—C18	120.0 (12)	C33—C36—C37	120.0 (8)
C14—C13—C18	120.0 (7)	C35—C36—C37	120.0 (12)
C13—C14—H14	120	C36—C37—H37	120
C13—C14—C15	120.0 (8)	C36—C37—C38	120.0 (8)
H14—C14—C15	120	H37—C37—C38	120
C14—C15—H15	120	C37—C38—H38	120
C14—C15—C16	120.0 (12)	C37—C38—C39	120.0 (8)
H15—C15—C16	120	H38—C38—C39	120
C15—C16—C17	120.0 (7)	C38—C39—H39	120
C15—C16—C28	120.0 (12)	C38—C39—C40	120.0 (12)
C17—C16—C28	120.0 (7)	H39—C39—C40	120
C16—C17—C18	120.0 (7)	C25—C40—C35	120.0 (8)
C16—C17—C27	120.0 (7)	C25—C40—C39	120.0 (12)
C18—C17—C27	120.0 (12)	C35—C40—C39	120.0 (8)
C13—C18—C17	120.0 (12)	H41—C41—C42	120
C13—C18—C19	120.0 (7)	H41—C41—C46	120
C17—C18—C19	120.0 (7)	C42—C41—C46	120.0 (13)
C18—C19—C20	120.0 (7)	C41—C42—H42	120
C18—C19—C24	120.0 (7)	C41—C42—C43	120.0 (14)
C20—C19—C24	120.0 (12)	H42—C42—C43	120.01
C11—C20—C19	120.0 (12)	C42—C43—C44	120.0 (15)
C11—C20—C21	120.0 (7)	C43—C44—H44	120.01
C19—C20—C21	120.0 (7)	C43—C44—C45	120.0 (12)
C20—C21—H21	120	H44—C44—C45	120.01
C20—C21—C22	120.0 (8)	C44—C45—H45	120.01
H21—C21—C22	120	C44—C45—C46	120.0 (14)
C21—C22—H22	120	H45—C45—C46	120.01
C21—C22—C23	120.0 (12)	C41—C46—C45	120.0 (15)
H22—C22—C23	120	C41—C46—H46	120.01
C22—C23—C24	120.0 (7)	C45—C46—H46	120.01

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
C7—H7···C14	1.00	2.48	2.82 (7)	99.53
C10—H10···C21	1.00	2.48	2.82 (7)	99.53
C14—H14···C7	1.00	2.48	2.82 (7)	99.56
C15—H15···C5	1.00	2.13	2.89 (9)	131.67
C15—H15···C6	1.00	2.12	2.60 (6)	106.98
C21—H21···C10	1.00	2.48	2.82 (7)	99.55
C22—H22···C43	1.00	2.41	2.84 (7)	105.43
C22—H22···C44	1.00	2.50	3.09 (8)	117.66
C32—H32···C37	1.00	2.48	2.82 (11)	99.54
C37—H37···C32	1.00	2.48	2.82 (11)	99.54
C45—H45···C1^i	1.00	2.43	3.36 (13)	154.10

Symmetry code: (i) −x−1/2, y−1/2, z.

(VII)

Crystal data

C46H30	V = 1500 (60) Å^3
Mr = 582.7	Z = 2
Triclinic, P1	F(000) = 604
Hall symbol: -P 1	Dx = 1.287 Mg m^−3
a = 10.5 (2) Å	Electron radiation, λ = 0.0335 Å
b = 11.6 (3) Å	µ = 0 mm^−1
c = 12.8 (3) Å	T = 293 K
α = 85.3 (5)°	Nanocrystal, yellow
β = 76.1 (5)°	0.001 × 0.0004 × 0.0001 mm
γ = 84.9 (5)°

Data collection

Zeiss Libra 120 diffractometer	θmax = 1.2°, θmin = 0.1°
data have been collected by precession–assisted 3D electron diffration scans	h = −12→12
51475 measured reflections	k = −14→14
8265 independent reflections	l = −15→15
1315 reflections with I > 3σ(I)

Refinement

Refinement on F	149 constraints
R[F > 3σ(F)] = 0.162	H-atom parameters constrained
wR(F) = 0.196	Weighting scheme based on measured s.u.'s w = 1/(σ^2(F) + 0.0001F^2)
S = 1.84	(Δ/σ)max = 0.023
8265 reflections	Δρmax = 0.27 e Å^−3
241 parameters	Δρmin = −0.44 e Å^−3
93 restraints

Special details

Experimental. 3D electron diffration recorded by TEM and dynamically refined

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	0.4415 (9)	0.9191 (7)	0.3132 (6)	0.0174 (5)*
H1	0.534328	0.918917	0.269097	0.0209*
C2	0.3691 (11)	1.0244 (6)	0.3401 (6)	0.0174 (5)*
H2	0.410318	1.099122	0.315146	0.0209*
C3	0.2388 (11)	1.0247 (6)	0.4019 (6)	0.0174 (5)*
H3	0.187211	1.099662	0.421042	0.0209*
C4	0.1809 (10)	0.9198 (7)	0.4368 (6)	0.0174 (5)*
H4	0.088076	0.920048	0.480871	0.0209*
C5	0.2533 (11)	0.8146 (6)	0.4100 (6)	0.0174 (5)*
H5	0.212087	0.73988	0.434909	0.0209*
C6	0.3837 (11)	0.8143 (6)	0.3482 (5)	0.0174 (5)*
C7	0.2089 (6)	0.4264 (4)	0.0166 (4)	0.0174 (5)*
H7	0.161845	0.50474	0.027119	0.0209*
C8	0.1656 (7)	0.3471 (4)	−0.0435 (5)	0.0174 (5)*
H8	0.088009	0.369441	−0.075477	0.0209*
C9	0.2319 (7)	0.2365 (4)	−0.0584 (5)	0.0174 (5)*
H9	0.201335	0.180573	−0.100848	0.0209*
C10	0.3417 (6)	0.2048 (4)	−0.0130 (4)	0.0174 (5)*
H10	0.38862	0.126356	−0.023431	0.0209*
C11	0.3850 (5)	0.2839 (3)	0.0468 (3)	0.0174 (5)*
C12	0.3185 (5)	0.3948 (3)	0.0617 (3)	0.0174 (5)*
C13	0.3618 (5)	0.4741 (3)	0.1217 (3)	0.0174 (5)*
C14	0.2951 (6)	0.5854 (3)	0.1366 (4)	0.0174 (5)*
H14	0.217584	0.607792	0.104612	0.0209*
C15	0.3385 (7)	0.6649 (4)	0.1969 (5)	0.0174 (5)*
H15	0.291463	0.743256	0.207409	0.0209*
C16	0.4485 (7)	0.6331 (4)	0.2424 (4)	0.0174 (5)*
C17	0.5153 (6)	0.5219 (3)	0.2275 (4)	0.0174 (5)*
C18	0.4719 (5)	0.4423 (3)	0.1672 (3)	0.0174 (5)*
C19	0.5383 (5)	0.3315 (3)	0.1524 (3)	0.0174 (5)*
C20	0.4949 (5)	0.2521 (3)	0.0922 (3)	0.0174 (5)*
C21	0.5616 (6)	0.1408 (4)	0.0773 (4)	0.0174 (5)*
H21	0.531022	0.08485	0.034884	0.0209*
C22	0.6717 (7)	0.1091 (4)	0.1225 (5)	0.0174 (5)*
H22	0.71875	0.030812	0.111916	0.0209*
C23	0.7150 (6)	0.1885 (4)	0.1827 (4)	0.0174 (5)*
C24	0.6493 (9)	0.3003 (5)	0.1964 (8)	0.0174 (5)*
C25	0.8280 (7)	0.1545 (6)	0.2284 (6)	0.0174 (5)*
C26	0.6938 (7)	0.3802 (6)	0.2594 (4)	0.0174 (5)*
C27	0.6275 (6)	0.4932 (7)	0.2754 (4)	0.0174 (5)*
C28	0.4915 (9)	0.7156 (6)	0.3040 (5)	0.0174 (5)*
C29	0.5472 (13)	0.6595 (7)	0.3904 (8)	0.0174 (5)*
C30	0.5541 (14)	0.7225 (6)	0.4791 (8)	0.0174 (5)*
H30	0.51413	0.803775	0.477049	0.0209*
C31	0.6299 (14)	0.6753 (6)	0.5527 (7)	0.0174 (5)*
H31	0.634806	0.719693	0.615191	0.0209*
C32	0.6989 (13)	0.5649 (7)	0.5375 (8)	0.0174 (5)*
H32	0.752389	0.531657	0.589262	0.0209*
C33	0.6921 (13)	0.5019 (6)	0.4487 (8)	0.0174 (5)*
H33	0.740698	0.42419	0.437942	0.0209*
C34	0.6162 (13)	0.5491 (7)	0.3752 (7)	0.0174 (5)*
C35	0.8323 (8)	0.3632 (7)	0.2613 (7)	0.0174 (5)*
C36	0.9015 (10)	0.4550 (5)	0.2816 (6)	0.0174 (5)*
H36	0.858471	0.53528	0.285997	0.0209*
C37	1.0317 (10)	0.4331 (6)	0.2956 (7)	0.0174 (5)*
H37	1.080603	0.497889	0.309823	0.0209*
C38	1.0928 (8)	0.3192 (7)	0.2893 (7)	0.0174 (5)*
H38	1.184711	0.303761	0.299242	0.0209*
C39	1.0236 (10)	0.2274 (5)	0.2691 (7)	0.0174 (5)*
H39	1.066642	0.147158	0.264759	0.0209*
C40	0.8934 (10)	0.2494 (6)	0.2550 (7)	0.0174 (5)*
C41	1.0589 (9)	−0.1197 (7)	0.1467 (6)	0.0174 (5)*
H41	1.13812	−0.147224	0.091323	0.0209*
C42	1.0090 (9)	−0.0039 (6)	0.1411 (6)	0.0174 (5)*
H42	1.052548	0.051013	0.081674	0.0209*
C43	0.8978 (9)	0.0347 (5)	0.2189 (7)	0.0174 (5)*
C44	0.8367 (9)	−0.0424 (7)	0.3023 (6)	0.0174 (5)*
H44	0.757472	−0.014851	0.357728	0.0209*
C45	0.8866 (10)	−0.1581 (6)	0.3080 (6)	0.0174 (5)*
H45	0.843102	−0.213073	0.367427	0.0209*
C46	0.9977 (10)	−0.1968 (6)	0.2302 (6)	0.0174 (5)*
H46	1.033299	−0.279308	0.234149	0.0209*

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
?	?	?	?	?	?	?

Geometric parameters (Å, º)

C1—H1	1	C23—C24	1.42 (6)
C1—C2	1.40 (6)	C23—C25	1.45 (4)
C1—C6	1.40 (6)	C24—C26	1.45 (4)
C2—H2	1	C25—C40	1.45 (5)
C2—C3	1.40 (6)	C25—C43	1.52 (7)
C3—H3	1	C26—C27	1.44 (6)
C3—C4	1.40 (6)	C26—C35	1.45 (6)
C4—H4	1	C27—C34	1.45 (6)
C4—C5	1.40 (6)	C28—C29	1.45 (5)
C5—H5	1	C29—C30	1.42 (5)
C5—C6	1.40 (6)	C29—C34	1.42 (6)
C7—H7	1	C30—H30	1
C7—C8	1.42 (4)	C30—C31	1.42 (4)
C7—C12	1.41 (4)	C31—H31	1
C8—H8	1	C31—C32	1.42 (6)
C8—C9	1.41 (6)	C32—H32	1
C9—H9	1	C32—C33	1.42 (5)
C9—C10	1.42 (4)	C33—H33	1
C10—H10	1	C33—C34	1.42 (4)
C10—C11	1.41 (4)	C35—C36	1.42 (5)
C11—C12	1.42 (6)	C35—C40	1.42 (7)
C11—C20	1.42 (4)	C36—H36	1
C12—C13	1.42 (4)	C36—C37	1.42 (5)
C13—C14	1.42 (6)	C37—H37	1
C13—C18	1.42 (4)	C37—C38	1.42 (7)
C14—H14	1	C38—H38	1
C14—C15	1.42 (4)	C38—C39	1.42 (5)
C15—H15	1	C39—H39	1
C15—C16	1.42 (4)	C39—C40	1.42 (5)
C16—C17	1.42 (6)	C41—H41	1
C16—C28	1.45 (4)	C41—C42	1.40 (7)
C17—C18	1.42 (4)	C41—C46	1.40 (5)
C17—C27	1.45 (4)	C42—H42	1
C18—C19	1.41 (6)	C42—C43	1.40 (5)
C19—C20	1.42 (4)	C43—C44	1.40 (5)
C19—C24	1.42 (4)	C44—H44	1
C20—C21	1.42 (6)	C44—C45	1.40 (7)
C21—H21	1	C45—H45	1
C21—C22	1.42 (4)	C45—C46	1.40 (5)
C22—H22	1	C46—H46	1
C22—C23	1.42 (4)
H1—C1—C2	120	C19—C24—C23	120.0 (10)
H1—C1—C6	120	C19—C24—C26	120.6 (12)
C2—C1—C6	120.0 (19)	C23—C24—C26	119.4 (14)
C1—C2—H2	120	C23—C25—C40	115.3 (14)
C1—C2—C3	120.0 (12)	C23—C25—C43	122.4 (9)
H2—C2—C3	120	C40—C25—C43	120.0 (17)
C2—C3—H3	120	C24—C26—C27	120.2 (13)
C2—C3—C4	120.0 (12)	C24—C26—C35	115.2 (10)
H3—C3—C4	120	C27—C26—C35	119.1 (10)
C3—C4—H4	120	C17—C27—C26	117.3 (9)
C3—C4—C5	120.0 (19)	C17—C27—C34	113.4 (12)
H4—C4—C5	120	C26—C27—C34	120.9 (10)
C4—C5—H5	120	C16—C28—C29	112.4 (17)
C4—C5—C6	120.0 (12)	C28—C29—C30	120.1 (18)
H5—C5—C6	120	C28—C29—C34	118.8 (14)
C1—C6—C5	120.0 (12)	C30—C29—C34	120.0 (14)
H7—C7—C8	120	C29—C30—H30	114.25
H7—C7—C12	120	C29—C30—C31	120.0 (16)
C8—C7—C12	120.0 (12)	H30—C30—C31	125.08
C7—C8—H8	120	C30—C31—H31	120
C7—C8—C9	120.0 (14)	C30—C31—C32	120.0 (13)
H8—C8—C9	120	H31—C31—C32	120
C8—C9—H9	120	C31—C32—H32	120
C8—C9—C10	120.0 (9)	C31—C32—C33	120.0 (14)
H9—C9—C10	120	H32—C32—C33	120
C9—C10—H10	120	C32—C33—H33	120
C9—C10—C11	120.0 (12)	C32—C33—C34	120.0 (16)
H10—C10—C11	120	H33—C33—C34	120
C10—C11—C12	120.0 (13)	C27—C34—C29	119.9 (12)
C10—C11—C20	120.0 (12)	C27—C34—C33	119.3 (16)
C12—C11—C20	120.0 (8)	C29—C34—C33	120.0 (13)
C7—C12—C11	120.0 (8)	C26—C35—C36	121.2 (13)
C7—C12—C13	120.0 (12)	C26—C35—C40	118.4 (11)
C11—C12—C13	120.0 (13)	C36—C35—C40	120.0 (17)
C12—C13—C14	120.0 (13)	C35—C36—H36	120
C12—C13—C18	120.0 (12)	C35—C36—C37	120.0 (13)
C14—C13—C18	120.0 (8)	H36—C36—C37	120
C13—C14—H14	120	C36—C37—H37	120
C13—C14—C15	120.0 (13)	C36—C37—C38	120.0 (10)
H14—C14—C15	120	H37—C37—C38	120
C14—C15—H15	120	C37—C38—H38	120
C14—C15—C16	120.0 (12)	C37—C38—C39	120.0 (17)
H15—C15—C16	120	H38—C38—C39	120
C15—C16—C17	120.0 (9)	C38—C39—H39	120
C15—C16—C28	119.2 (13)	C38—C39—C40	120.0 (13)
C17—C16—C28	120.8 (14)	H39—C39—C40	120
C16—C17—C18	120.0 (13)	C25—C40—C35	121.4 (17)
C16—C17—C27	118.1 (9)	C25—C40—C39	118.6 (13)
C18—C17—C27	121.9 (12)	C35—C40—C39	120.0 (10)
C13—C18—C17	120.0 (12)	H41—C41—C42	120
C13—C18—C19	120.0 (8)	H41—C41—C46	120
C17—C18—C19	120.0 (13)	C42—C41—C46	120.0 (15)
C18—C19—C20	120.0 (13)	C41—C42—H42	120
C18—C19—C24	120.0 (8)	C41—C42—C43	120.0 (11)
C20—C19—C24	120.0 (12)	H42—C42—C43	120
C11—C20—C19	120.0 (12)	C25—C43—C42	128.9 (10)
C11—C20—C21	120.0 (8)	C25—C43—C44	111.1 (17)
C19—C20—C21	120.0 (13)	C42—C43—C44	120.0 (17)
C20—C21—H21	120	C43—C44—H44	120
C20—C21—C22	120.0 (9)	C43—C44—C45	120.0 (15)
H21—C21—C22	120	H44—C44—C45	120
C21—C22—H22	120	C44—C45—H45	120
C21—C22—C23	120.0 (12)	C44—C45—C46	120.0 (11)
H22—C22—C23	120	H45—C45—C46	120
C22—C23—C24	120.0 (14)	C41—C46—C45	120.0 (17)
C22—C23—C25	119.3 (13)	C41—C46—H46	120
C24—C23—C25	120.7 (10)	C45—C46—H46	120

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
C33—H33···C35	1.00	2.37	3.00 (11)	120.33