An organometallic route to long helicenes

Abstract

Along with the recent progress in the development of advanced synthetic methods, the chemical community has witnessed an increasing interest in promising carbon-rich materials. Among them, helicenes are unique 3D aromatic systems that are inherently chiral and attractive for asymmetric catalysis, chiral recognition and material science. However, there have been only limited attempts at synthesizing long helicenes, which represent challenging targets. Here, we report on an organometallic approach to the derivatives of undecacyclic helicene, which is based on intramolecular [2 + 2 + 2] cycloisomerization of aromatic hexaynes under metal catalysis closing 6 new cycles of a helicene backbone in a single operation. The preparation of nonracemic compounds relied on racemate resolution or diastereoselective synthesis supported by quantum chemical (density functional theory) calculations. The fully aromatic [11]helicene was studied in detail including the measurement and theoretical calculation of its racemization barrier and its organization on the InSb (001) surface by STM. This research provides a strategy for the synthesis of long helical aromatics that inherently comprise 2 possible channels for charge transport: Along a π-conjugated pathway and across an intramolecularly π-π stacked aromatic scaffold.

Carbon-rich materials of an ordered structure such as fullerenes (1), carbon nanotubes (2), and graphene (3) have recently attracted considerable attention and stimulated broad multidisciplinary research toward a plethora of envisaged applications. Due to the unique properties of the carbon element, manifold carbon allotropes or, generally, carbon-rich materials, networks and nanostructures can be imagined (4, 5). Moreover, modern synthetic tools have already allowed for the preparation of many carbon-rich structures or at least their advanced precursors.

In the quest for other members of the family of carbon-rich materials, we proposed focusing on unique helicenes (6–9) whose longest representatives definitely deserve more attention. Helicenes are 3-dimensional polycyclic aromatic systems, which consist of all-ortho-fused aromatic rings and are inherently chiral. Considering a hypothetical model, a sufficiently long helicene molecule would represent an ordered all-carbon (or, alternatively, carbon-heteroatom) spring “wrapped” by hydrogen atoms. Accordingly, such systems are expected to exhibit remarkable chiroptical, mechanical, (semi)conductive and perhaps magnetic properties either in bulk or as single molecule devices. Despite the fact that helicenes have been known since 1956 (10), there have been only a few attempts at synthesizing long aromatic helices. In this regard, the world record among carbohelicenes belongs to Martin and Baes (11), who published the synthesis of [14]helicene 1 consisting of 14 benzene rings in 1975 (Fig. 1). Later, in a heterohelicene series, Yamada et al. (12) went even beyond such limits preparing the pentadecacyclic thia[15]helicene 2 (Fig. 1) in 1981. Although photoinduced double cyclodehydrogenation of a bis(arylvinyl)arene substrate, which was used to create the helical backbone, proceeded well in the former case (providing 1 in up to a 45% yield), the latter case indicated the capriciousness of this methodology (providing 2 in only a 5% yield).

Fig. 1.

The longest helicenes and their analogs 1–4 prepared so far.

Alternative nonphotochemical routes to the highest helical (hetero)aromatics and their congeners have recently been revived. Based on versatile triple Co^I-catalyzed alkyne [2 + 2 + 2] cycloisomerization, Vollhardt et al. (13) reported on the preparation of a unique heptadecacyclic angular [9]phenylene 3 in 2002 (Fig. 1), which is the longest “heliphene” representative known to date. Notwithstanding that they pushed the limits of forming long helical aromatics onward, the final operation forming the helical scaffold resulted in only a 2% yield of 3 and a 3.5% yield of its 4-methoxymethyl derivative. Such outcomes of the key multiple cyclization certainly confine any further effort to deal with even longer angular phenylenes. The last attempt by Rajca et al. (14) in 2005 at making an oligomeric (C₂S)_n helix shows a different strategy as well as repertoire of nonphotochemical synthetic tools used. They succeeded in preparing the shorter but remarkable carbon-sulfur [11]helicene 4 (Fig. 1), which consists of 11 fused thiophenes having all sulfur atoms at the molecular periphery, in a good 47–59% yield (for mono-annelation of a biaryl precursor by a sulfur reagent) or a low 1.3–3% yield (for tri-annelation of a quateraryl precursor). They could control helicity to a low extent using asymmetric lithiation (ee 11–19%).

As an alternative to long helicenes, the synthesis of helicene-based oligomers and polymers has been attempted. Giving up material monodispersity, Katz et al. explored the ways of polymerizing bifunctional helicene units. They succeeded in preparing both optically active [7]helicene-derived cobaltocenium oligomers (15) and the nickel salen [6]helicene-derived conjugated ladder polymer (16). An effort by Katz et al. to elongate the helicene backbone stepwise employing a helicene aryne intermediate failed since the unwanted intramolecular Diels-Alder addition prevailed over the intermolecular one (17). In addition, there are also other ways of forming helical aromatic structures or super structures, such as helicene aggregates (18), foldamers (19), helicates (20) or twisted aromatics (21). They might be promising surrogates for individual helicenes, although practical ways of tackling problems such as polydispersity, conformational instability, interrupted conjugation and an inefficient synthesis have yet to be mastered.

These efforts and results indicate serious difficulties connected with the synthesis of long helicenes. Solving this perennial problem can both meet the criteria of a milestone in the total synthesis of complex nonnatural molecules and supply unique materials for technology-oriented research.

Results and Discussion

Recently, we have proven that constructing helicene scaffolds by using triyne [2 + 2 + 2] cycloisomerization constitutes a robust and flexible methodology for the synthesis of helically chiral aromatics (22–26). Conceptually, the advantage of this methodology consists in the formation of 3 rings of a helicene backbone in a single step. Provided that cycloisomerization takes place twice or, ultimately, in a multifold way, it might simplify the synthesis of long helicenes. As mentioned above, Vollhardt et al. (13) manifested the strengths of multiple cyclizations while confronting the problem of low yields when attempting benzene-cyclobutane annulation in strained heliphene systems.

Being aware of these facts, we designed the structure of the hexayne precursors of the proposed [11]helicene molecules in such a way as to perform intramolecular double [2 + 2 + 2] cycloisomerization closing energetically more favorable 6- or 7-membered rings rather than the 4-membered one. To verify the hypothesis for the effective formation of 6 rings of helicene backbones in a single step, we attempted the synthesis of the [11]helicene-like molecule 13 (Scheme 1).

Scheme 1.

A reaction scheme illustrating the synthesis of the racemic [11]helicene-like molecule 13. (a) DIBAL-H (5.2 equiv.), CH₂Cl₂, −78°C, 30 min, then room temperature, 30 min, 91%; (b) CBr₄ (3.4 equiv.), PPh₃ (3.0 equiv.), acetonitrile, room temperature, 1 h, 61%; (c) But-2-yn-1-ol 8 (10.0 equiv.), NaH (10.0 equiv.), THF, 50°C, 30 min, 38%; (d) Aryl iodide 10 (2.5 equiv.), Pd(PPh₃)₄ (10 mol %), CuI (30 mol %), diisopropylamine, 85°C, 45 min, 47%; (e) ⁿBu4NF (3.0 equiv.), THF, room temperature, 30 min, 90%; (f) CpCo(CO)₂ (1.2 equiv.), PPh₃ (2.0 equiv.), decane, halogen lamp, 140°C, 1 h, 60%; (g) CpCo(C₂H₄)₂ (2.1 equiv.), THF, room temperature, 30 min, 45%; (h) Ni(cod)₂ (2.6 equiv.), PPh₃ (4.2 equiv.), THF, room temperature, 18 h, 24%.

Since 1,2,4,5-tetrasubstituted benzene building blocks are synthetically better accessible than 1,2,3,4 ones, we decided to embed the former structure into helicene backbones instead of the latter one to simplify the synthesis. Despite the fact that such a structural modification led to a disruption of the all-ortho annulation of skeletal rings, the molecular shape of the synthetic targets did not substantially deviate from the shape of the parent helicenes.

Starting from the known diyne 5 (27), we prepared tetrayne 9 within 3 routine synthetic steps. Since the reaction of 7 with 8 under basic conditions resulted in the attachment of alkyne sidearms as well as the removal of the TMS protecting groups, we could right away couple tetrayne 9 with the known iodide 10 (24) under Pd⁰/Cu^I catalysis to obtain hexayne 11. After desilylation of the terminal alkyne units, hexayne 12 was subjected to the final cyclization, which was mediated by various metal complexes. The use of CpCo(CO)₂ with PPh₃ was superior to CpCo(C₂H₄)₂ alone or Ni(cod)₂ with PPh₃ in providing the desired [11]helicene-like compound 13 in a good 60% yield. X-ray analysis confirmed the molecular structure of 13 (Fig. 2). Thus, double [2 + 2 + 2] cycloisomerization enabled the construction of 6 cycles of the helicene-like backbone in a single operation.

Fig. 2.

A view of the helicene molecule of 13 [an enantiomer with (M) helicity is shown]. The displacement ellipsoids are drawn on the 50% probability level (PLATON). The dihedral angle of the ring planes defined by atoms C1–C5, C48, and C33–C39 is 6.27 (11) ° (the first and the last 6-membered circle). The distance of their corresponding centroids is 7.890 Å.

We have recently developed an asymmetric synthesis of penta-to-heptacyclic helicene-like compounds, which is based on a diastereoselective [2 + 2 + 2] cycloisomerization of chiral triynes (25, 28). Therefore, we decided to follow this paradigm to prepare [11]helicene-like molecule 16 in a nonracemic form (Scheme 2). Indeed, optically pure hexayne (−)-(S,S)-14, which was prepared analogously to the achiral hexayne 11 (Scheme S1), underwent metal-mediated cyclizations in a stereoselective fashion providing the cyclized products in moderate yields. Employing CpCo(CO)₂ with PPh₃ at 140°C, we received diastereomers (M,S,S)-16 and (+)-(P,S,S)-16 in a 10 : 90 ratio and a 26% yield. Applying microwave irradiation at 150 to 200°C instead of irradiation with a halogen lamp, we recorded the same stereochemical outcome but a better yield (33%) after a shorter reaction period. The use of more reactive CpCo(C₂H₄)₂ allowed the performance of cyclization at room temperature, but the diastereoselectivity of the reaction and the yield (17%) were lower. Diastereomer with (P) helicity prevailed again. Since suitable crystals for a single-crystal analysis were not obtained, the helicity of the major diastereomer (+)-(P,S,S)-16 (Fig. S1) was assigned by interpreting its ROESY ¹H NMR spectrum (for details, see SI Text) in accord with our earlier observations (25, 28).

Scheme 2.

A reaction sequence illustrating the diastereoselective formation of the nonracemic [11]helicene-like molecule (+)-(p,s,s)-16. (a) CpCo(CO)₂ (4.2 equiv. in 2 successive portions), PPh₃ (4.1 equiv.), decane, halogen lamp, 140°C, 7.5 h, 26%, (m,s,s)-16 : (p,s,s)-16 = 10 : 90; (b) CpCo(CO)₂ (3.1 equiv.), PPh₃ (4.0 equiv.), THF, microwave oven irradiation, 200°C, 10 min, then 150°C, 1 h, 33%, (m,s,s)-16 : (p,s,s)-16 = 10 : 90; (c) CpCo(C₂H₄)₂ (2.7 equiv.), THF, room temperature, 15 min, 17%, (m,s,s)-16 : (p,s,s)-16 = 25 : 75.

The formation of the [11]helicene-like molecules from corresponding hexaynes is obviously a 2-step process comprising tandem [2 + 2 + 2] cycloisomerizations. In the case of nonracemic (+)-(P,S,S)-16, however, it raises a question of at what stage the helicity control takes place. Although we could not prove the occurrence of the intermediate (M,S,S)-15 or (P,S,S)-15, the explanation of the diastereoselectivity relies on the assumption of their occurrence as well as the supporting calculations using the DFT(B3LYP)/TZV+P method and conductor-like screening model (COSMO) to account for the solvation. We have recently found that the diastereoselectivity of [2 + 2 + 2] cycloisomerization of related triynes at elevated temperature is governed by the thermodynamic factors (25). Indeed, the reaction produced the more stable diastereomer (+)-(P,S,S)-16, which was expected to be formed from the more stable intermediate (P,S,S)-15 (Table 1). We assume that the helicity of the half-cyclized intermediate determines the helicity of the final product. The alternative pathway through the energetically rich quasi-meso conformer of (S,S)-16 (Fig. S2) and its thermal equilibration to the more stable (+)-(P,S,S)-16 seems to be unlikely as they differ substantially in Gibbs free energy (≈8.6 kcal/mol).

Table 1.

The calculated differences in Gibbs free energies between the (M) and (P) pairs of diastereomers (keeping the stereogenic centers unchanged)

Entry	Diastereomers	ΔGcalc*†, kcal/mol	(M):(P)calc†	(M):(P)exp‡
1	(M,S,S)-15 vs. (P,S,S)-15	1.65	12:88	(intermediate)
2	(M,S,S)-16 vs. (P,S,S)-16	1.79	10:90	10:90

*The positive value indicates the higher stability of the diastereomer with (P) helicity.

^†Calculated at the DFT(B3LYP)/TZV + P level.

^‡The stereochemical outcome of CpCo(CO)₂-mediated [2 + 2 + 2] cycloisomerization at 140 °C or 200 °C.

We have demonstrated that the synthesis of racemic as well as nonracemic [11]helicene-like compounds can be straightforward and efficient. However, parent helicenes are fully aromatic systems while molecules such as 13 or (+)-(P,S,S)-16 are not. Therefore, we concentrated our effort on the synthesis of [11]helicene 28, which consists of fused benzene rings only (Scheme 3). Starting from the known diester 17 (27), we prepared within 3 steps dialdehyde 21, which was treated with lithiated propyne 22. The double addition led to a 1:1 mixture of the d,l pair (R*,R*)-23 and the meso form (R,S)-23 in a good yield (with the structure of individual stereomers being assigned afterward). Both diastereomers could be separated by column chromatography and transformed to the key hexaynes (R*,R*)-25 and (R,S)-25. The d,l pair (R*,R*)-25 was subjected to [2 + 2 + 2] cycloisomerization mediated by various metal complexes affording mixtures of diastereomers (M*,R*,R*)-26 and (P*,R*,R*)-26 generally in moderate yields (up to 30%). The use of CpCo(CO)₂ with PPh₃ at 140°C resulted in the formation of diastereomers (M*,R*,R*)-26 and (P*,R*,R*)-26 in a 17:83 ratio. Using more reactive CpCo(C₂H₄)₂, which allowed the reaction to be run at room temperature, the diastereoselectivity of the cyclization was low, obtaining (M*,R*,R*)-26 and (P*,R*,R*)-26 in a 57 : 43 ratio. Finally, Ni(cod)₂ with PPh₃ mediated the reaction at room temperature to receive (M*,R*,R*)-26 and (P*,R*,R*)-26 in a 72:28 ratio. Such results showed that stereogenic centers in the starting racemic hexayne (R*,R*)-25 can effectively control the diastereoselectivity of tandem [2 + 2 + 2] cycloisomerization depending on the prevailing kinetic or thermodynamic factors. The diastereomers (M*,R*,R*)-26 and (P*,R*,R*)-26 were not possible to separate by column chromatography and, therefore, their helicity was assigned by interpreting the ¹H and ROESY ¹H NMR spectra of their mixtures (for details, see Fig. S3).

Scheme 3.

A reaction scheme illustrating the synthesis of racemic [11]helicene 28.

The meso hexayne (R,S)-25 could be cyclized to the racemic [11]helicene derivative (M*,R*,S*)-26 in a moderate yield (up to 40%) employing CpCo(CO)₂ with PPh₃ at 140°C. The use of CpCo(C₂H₄)₂ or Ni(cod)₂ with PPh₃ at room temperature also resulted in the formation of the desired product but in lower yields. The aromatization of (M*,R*,S*)-26 was then straightforward. Acid-assisted elimination of acetoxy groups followed by dehydrogenation with trityl tetrafluoroborate completed the synthesis of racemic [11]helicene 28 (Fig. 3). It is worth noting that, unlike many planar polyaromatic systems, this helical polyaromatic compound was stable and well soluble in various organic solvents. Higher solubility can be attributed to limited face-to-face π-π stacking interactions between individual molecules of 28 because only terminal naphthalene segments are not sterically shielded and/or significantly twisted.

Fig. 3.

A molecular model of (+)-(P)-28 (the structure optimized at the B3LYP/TZV+P level).

For the sake of receiving 28 in a nonracemic form, we performed the separation of enantiomers by HPLC on a semipreparative Eurocel O1 column (250 mm × 20 mm) using heptane-isopropanol (3:1) as a mobile phase (Fig. S4). The individual enantiomers exhibited remarkable optical activity: For (+)-28, we measured [α]_D = +7 142 deg cm³ g⁻¹ dm⁻¹ in CH₂Cl₂ (enantiomerically pure) and for (−)-28 [α]_D = −6 661 deg cm³ g⁻¹ dm⁻¹ in CH₂Cl₂ (containing 3.5% of the opposite enantiomer). Having the enantiomers separated, we could determine the barrier to racemization (ΔG) of 28 to be 37.5 ± 0.1 kcal mol⁻¹ (230°C). Such a value is much closer to the barrier of [6]helicene (36.2 kcal mol⁻¹, 27°C) (29) than to the barrier of [9]helicene (43.5 kcal mol⁻¹, 27°C) (29), which is the highest helicene studied in this way. In fact, the central ring in [11]helicene 28 is annulated in a meta fashion and 28 can thus be viewed as a 2,3-disubstituted [6]helicene, hence exhibiting a comparable flexibility. The computational treatment of the barrier revealed that this is a complex conformational process (Fig. 4). A comprehensive study on computing barriers to racemization of helicenes including 28 will be published separately.

Fig. 4.

A calculated pathway of racemization of 28 (the structures optimized at the BP86/def2-SVP level with the empirical correction for the dispersion energy, relative energies given).

To assign helicity to the individual enantiomers of 28, we correlated their CD spectra with those of the closest (−)-(M)-[9]helicene (Fig. S5), whose CD spectrum has been published (30). The latter levorotatory helicene uniformly exhibits the longest wavelength band, which is intensive and associated with a negative dichroism, and an adjacent band at shorter wavelengths, which is also intensive but associated with a positive dichroism. Since the (−)-enantiomer of [11]helicene 28 displays the same features, we can tentatively assign (M) helicity to (−)-28 and (P) helicity to (+)-28. Other bands at the shortest wavelengths are rather difficult to correlate.

With the ultimate goal being to develop short molecular wires that are transversal to the (semi)conducting layers, we attempted the deposition and characterization of [11]helicene 28 on a solid substrate. We succeeded in obtaining an STM image of the racemic [11]helicene 28 assembled on c(8 × 2) reconstructed InSb (001). The sample was prepared at room temperature and subsequently measured at 77 K (Fig. 5A). The molecules do not appear to be present at the flat terrace areas but to be preferentially arranged along the substrate atomic step edges forming molecular chains. This finding indicates that the helicene diffusion length at room temperature is larger than the flat terrace lateral dimensions and that the InSb step edges act as the adsorption traps.

Fig. 5.

(A) A STM image of racemic [11] helicene 28 assembled on c(8 × 2) reconstructed InSb (001) at room temperature and subsequently cooled down for imaging to 77 K; the sample bias was −1.25 V. (B) A topographic image of nonracemic [11]helicene (−)-28 adsorbed in the vicinity of the InSb (001) c(8 × 2) surface step edge; the sample bias was −3 V and the sample imaging temperature was 77 K.

There is an important difference in the organization of the molecules at the step edge when comparing arrangements of racemic 28 and nonracemic (−)-28 (Fig. 5 A and B). The STM technique is able to image a local density of states, thus making the imaging of molecular orbitals of adsorbed molecules feasible. Accordingly, we obtained a high resolution image of the occupied states of the 3 [11]helicene molecules (−)-28 (sample bias: −3 V) adsorbed in the vicinity of the substrate step edge (Fig. 5B).

Conclusions

In summary, we synthesized and fully characterized the [11]helicene derivatives 13, 16 and 26-28 belonging to the unique family of long helicenes. Such results demonstrated that intramolecular double [2 + 2 + 2] cycloisomerization of the hexayne precursors 12, 14 and 25 allowed for the effective construction of 6 cycles of the helicene backbone in a single operation. An asymmetric version of this cobalt- or nickel-catalyzed cyclization was developed to prepare the nonracemic [11]helicene-like derivative (+)-(P,S,S)-16 in good diastereomeric purity that would be in accord with the DFT calculations in terms of the thermodynamic control of the reaction. The enantiomers of the fully aromatic [11]helicene 28 were separated by liquid chromatography on a chiral column and, accordingly, the barrier to racemization could be measured for the longest helicene. The theoretical calculation of the barrier was in a good agreement with the experimental value and racemization could be described as a conformational process. The LT UHV STM study revealed that racemic [11]helicene 28 and enantiopure (−)-28 are organized differently at the step edges of InSb (001). High resolution images of the occupied states of the individual molecules of [11]helicene (−)-28 were obtained. These results extend our knowledge on the organization of helical aromatics on solid surfaces (31, 32), which is a topic of significant importance, related, for instance, to the amplification of chirality in 2D systems (33). This research provides a strategy for the synthesis of long helical aromatics that inherently comprise 2 possible channels for charge transport both along a π-conjugated pathway and across an intramolecularly π-π stacked aromatic scaffold. Directions for the research to take in the future could include attempts at the preparation of even longer and functionalized helicenes, their characterization on solid surfaces by SPM techniques and the measurement of the conductance of these unique molecular wires.

Materials and Methods

Synthesis of Helicene Derivatives.

The multistep synthesis of the key hexaynes 12, 14 and 25, their conversion to [11]helicene derivatives 13, 16 and 26–28, the single-crystal analysis of 13, separation of enantiomers of 28 as well as the determination of its racemization barrier are described in SI Text. For further experimental details, see SI Text.

Quantum Chemical Calculations.

Geometry optimizations were carried out using the RI-PBE method (34) and 6–31G(d) basis set, whereas the single-point energies were recomputed in the larger basis set TZV+P (triple-ζ valence with 1 polarization function on each atom) using the B3LYP method (35–38). To account for the solvation effects, the conductor-like screening model (COSMO) method (39, 40) was used. The Gibbs free energy of the system was then calculated as the sum of the in vacuo energy, solvation free energy, zero-point energy and entropic term. The Turbomole program package (41) was used throughout this study. For further computational details, see SI Text. The transition state (TS) search was performed by the QST3 method (42), i.e., using optimized geometries of (M)- and (P)-[11]helicene 28 and a guess structure of the transition state. This was done using the B3LYP functional with the cc-pVDZ basis set at first. The resulting TS and minima were subsequently reoptimized at the BP86/def2-SVP level with the empirical correction for the dispersion energy (43, 44), which improved the results significantly.

Scanning Tunneling Microscopy.

The self-assembling and characterization of the molecules were performed in a multichamber UHV system. The system was equipped with a low-temperature scanning tunneling microscope (Omicron NanoTechnology GmbH). The base pressure of the system was in the 10⁻¹¹ mbar range. Epi-ready InSb (001) wafers (Kelpin Crystals, undoped) were used as substrates. A c(8 × 2) reconstructed surface was obtained by subsequent cycles of annealing and Ar ion bombardment (with the sample temperature being 700 K, ion beam energy 700 eV, angle of incidence 60° off-normal and ion current density ≈0.5 μA/cm²). Typically, the cleaning cycles were repeated until a sharp LEED diffraction pattern reflecting c(8 × 2) symmetry could be seen. The InSb sample quality was verified by the STM technique. It was found that at temperatures below 180 K, a new phase is developed with a surface symmetry reduced to p2 in comparison to the room-temperature case, showing the c(8 × 2) reconstruction and c2-mm symmetry (45). In addition, the low-temperature phase of the InSb (001) surface reveals characteristic wavy bands and bright protrusions superimposed on the high resolution STM image, presumably of electronic origin. The racemic as well as nonracemic [11]helicene 28 were evaporated on the InSb surface kept at room temperature from a standard effusion cell. At the cell temperature of 500 K, the flux of the sublimating molecules was 1 ML/10 min, as determined by a quartz crystal microbalance. For the evaporation of individual molecules (ultra-low coverage), the cell temperature was lowered to 490 K. In this last case, the microbalance reading was approximately 1 order of magnitude lower.