For an even number of rings ([4], [6], [8], [10], [12]…), regardless of the size of the DTH, the bond topology is similar to that shown in Figure 2. The three types of thiophene ring fusions are shown at the top of the diagram and correspond to exo–exo, endo–endo, and exo–endo topologies. If we assume bonding through the a-carbon (represented by the red arrows), the three types of conjugated systems may be further simplified to the circled structures. For n even, exo–exo-[n]DTH reduces to a cross-conjugated π system exemplified as 1,8-divinylnaphthalene, and so does endo–endo-[n]DTH to 2,7-divinylnaphthalene, in which disrupted conjugation has been represented by a green circuit ending in a red one. Conversely and still for n even, exo–endo-[n]DTH reduces to a linear conjugated π system exemplified as 1,7-divinylnaphthalene, in which efficient conjugation has been represented by a green circuit connecting both ends.

Next, we consider the odd number of rings. A complete symmetric pattern emerges here. In this case, again regardless of the DTH size ([3], [5], [7], [9], [11]…), the bond topology is represented in Figure 3. The three types of thiophene ring fusions displaying exo–exo, endo–endo, and exo–endo topologies are shown at the top of the diagram. These types of conjugated systems may be further simplified to the circled structures. For n odd, exo–exo-[n]DTH reduces to a linear-conjugated π system exemplified as 1,2-divinylbenzene, and so does endo–endo-[n]DTH to 1,4-divinylnaphthalene, in which efficient conjugation has been represented by a green circuit from end to end. Conversely, exo–endo-[n]DTH reduces to a cross-conjugated π system exemplified as 1,3-divinylbenzene, in which disrupted conjugation has been again represented by a green circuit ending in a red one.

In short, we can draw expected circuits of conductance originated by the bond topology. Linear conjugation facilitates efficient electron delocalization, enabling π-electrons to flow freely along the entire length of the π system, from end to end, as depicted by the green circuit between the red arrows. On the contrary, cross-conjugation disrupts this process, disabling π-electrons from flowing freely between both ends.

The synthetic strategies used for preparing these series of dithia[n]helicenes and a theoretical study of the binding properties of these molecules to a gold electrode are presented.

In previous works, our group proposed a modular synthetic route using different fragments to synthesize exo- and endo-dithia[7] and [9]helicenes. (9) Furthermore, by STM, the self-assembly on the gold surface of these molecules was analyzed to observe one single molecule in the case of the exo-compound, exposing its electrical conductance and even allowing visual chiral assignment (Baciu et al., 2020). The odd number of ring series was later expanded using a similar modular assembly to dithia[9]helicenes, also with the exo- and endo-topology on both ends (Baciu et al., 2022). With the development of these new enlarged helicenes, interesting photophysical and chiroptical properties arise, such as significant mirror-image CPL activity, ranking among the largest reported for a helicene derivative. This established our modular synthetic protocols as reliable and robust, enough to face new and more demanding challenges. This modular synthetic pathway relies on a final step of photochemical reaction to create armchair-type ring fusions between specific variable-sized fragments. These fragments can occupy either a central or a terminal position. (9) The fragments used as central building blocks, 3,6-dibromophenanthrene 10, naphthalene-2,7-diylbis(trifluoromethanesulfonate) 11, and 2,11-dibromobenzo[c]phenanthrene 12, were prepared using different synthesis strategies (see Schemes 1, 2). The synthesis method of 10 was reported by our group in previous work (Baciu et al., 2021). Fragment 11 was easily obtained from 2,7-dihydroxynaphthalene by double trifluoromethanesulfonation using triflic anhydride (see SI). Central fragment 12 was obtained through oxidative photocyclization, as described in Scheme 1. Hydroboration of 4-bromoethynylbenzene provided the vinylboronic derivative 13 needed for the subsequent Suzuki coupling with 14. The stilbenic precursor 15 obtained this way was irradiated with UV light using the LED technology to afford 12.

For the synthesis of the terminal fragments, a route consisting of a Sonogashira reaction and a hydroboration was used to obtain fragments 7 and 8. These contain the specific naphtho[2,1-b]thiophene and naphtho[1,2-b]thiophene moieties, respectively, ready to be assembled into central fragments by Suzuki coupling (Baciu et al., 2020). Furthermore, the smaller benzo[b]thiophene derivative 9 was obtained through analogous synthetic pathways (Baciu et al., 2022).

With these fragments in hand, our synthetic pathway adapted state-of-the-art methods for the palladium-catalyzed coupling reaction with the pioneering work on the photochemical synthesis of thiahelicenes, which was developed by Wynberg and Groen (1968), which inspired the key steps of this synthetic method. All (E)-stilbenic precursors were obtained by using the Suzuki coupling reaction using vinylboronic esters 7–9 and the different central fragments described above (Scheme 2).

Then, oxidative photocyclization (Jørgensen, 2010) of the bis-stilbenic precursors 16, 17, 18, and 19 under Mallory–Katz conditions was the last synthetic step required. It was done by use of 365 nm LED irradiation in benzene to improve the solubility of the substrate, under an Ar atmosphere, with iodine as the final oxidizer and 1,2-butylene oxide as a HI scavenger. This afforded exo-dithia[10]helicene 1, endo-dithia[10]helicene 2, exo-dithia[11]helicene 3, and endo-dithia[11]helicene 4 in fair yields (Scheme 3).

By adapting the synthetic strategy explained before, our group decided to synthesize dithiahelicenes with both an exo and an endo sulfur atom in the same molecule or mixed topology. Exo-endo-dithia[7]helicene 5 and exo-endo-dithia[10]helicene 6 were the selected candidates. In both syntheses, fragment 8 was used to introduce the endo thiophene terminal moiety. In the first case, an (E)-stilbenic precursor was obtained by coupling 8 and 20 (synthesized in previous work) (9) under typical Suzuki conditions (Scheme 4). In the second case, using the same principles, our group prepared fragment 23. Subsequent coupling of 23 with 7 under Suzuki reaction conditions provided the (E)-stilbenic precursor ready for the final photocyclization. In both cases, the final helicenes were obtained by UV irradiation under Mallory–Katz conditions by using an LED source (Scheme 5).

Calculations were performed with Kohn–Sham density functional theory (DFT) and time-dependent DFT (TDDFT) linear response methods. Ground-state equilibrium structures were optimized with spin-restricted DFT. All structures were characterized as minima via harmonic vibrational frequency calculations. For the binding energies (ΔE _f (kcal/mol), DFT calculations were done at the B3LYP/6-311G(d,p) level for the organic moiety (C, H, and S) and LANL2DZ for heavy atoms (Au₁₀ cluster). For accuracy, the basis set superposition error (BSSE) was taken into account. Two interacting fragments, dithiahelicene (fragment 1) and Au₁₀ (fragment 2), were defined. First, both fragments were allowed to interact until minimum energy was located by optimization of the complex and corroborated by vibrational analysis. Second, we optimized the isolated fragments, placed them in the complex orientation, and separated them along the direction of the bond from the equilibrium binding distance until a negligibly small interaction was found, and further away (typically up to 20 Å of separation), evaluating at this point the energy by a single-point calculation. A counterpoise calculation of these unbound complexes was carried out to corroborate the accuracy of the calculation (confirming a BSSE = 0). The binding energy was then obtained by subtracting both calculated electronic energies. For the dithiahelicene fragment, full optimization afforded C ₂ symmetries for the exo and endo cases and C ₁ for the mixed exo–endo configuration. The gold tip (Au₁₀) was taken from the bulk fcc gold crystal (Fm−3m) and relaxed at the theory level used for gold (B3LYP/LANL2DZ) while imposing T _d symmetry as a constraint. Then, it was used as a rigid fragment. For the optical gaps (TDDFT) and HOMO-LUMO (DFT) calculations, geometry optimizations and vibrational analysis were performed with the CAM-B3LYP functional and the def2SVPP basis set. This work was done using Gaussian 16 version C.01 (G16) (Frisch et al., 2022).

Our group analyzed the binding energy of two of the largest dithiahelicenes available both in exo- and endo-configurations, namely, dithia[10]helicenes and dithia[11]helicenes ([10] or [11]DTH) to a gold tip, by DFT calculation. The tip consisted of 10 gold atoms forming a tetrahedron (T _d symmetry), with its vertex pointing in the (1, 1, 1) direction of a standard fcc gold crystal. Figure 4 shows the most significant results obtained for these molecules. Focusing on the exo-isomer, several general features are revealed. First, as anticipated, there is a preference for the terminal thiophene ring when binding to the Au₁₀ cluster. Binding at the inner carbo-rings (rings 2–5 or 6) is less favorable for approximately 1–1.5 kcal/mol, with these having binding energies comparable to those of homologous carbo[n]helicenes. The favored position within the terminal ring is the α-carbon position, slightly favored over the sulfur atom by approximately 0.5 kcal/mol. A significant degree of pyramidalization is evident at the α-carbon during binding at this position. These binding energy patterns are reproducible and comparable regardless of the parity (odd or even number of rings) of the helicene. There is little difference in the binding of the tip at the inner rigs (rings 2–5), which always occurs in an η ² manner, i.e., involving a pair of CH carbons of the outer helix, while for α-C or S binding in the thiophene, the first ring is always η ¹. For the exo-[11]DTH, it is difficult for the tip to fit snugly into the innermost ring (ring 6) because of crowding, resulting in reduced binding energy.

For the endo-isomers, the binding scenario is less meaningful and its analysis more intricate, resulting in a less straightforward picture compared to that of the exo-isomer. Distortion of the initial thiophene ring, caused by a bigger endo-sulfur atom, leads to a variety of configurations where a portion of the gold cluster’s edge rests on the ring’s surface. This results in unusually large binding energies as more than one gold atom of the cluster cooperatively binds to the helicene surface. However, in instances where the tip is attached via a single gold atom (η ¹), the binding energies are similar to those of the exo-isomer, which is a good model of the interaction. The complete array of configurations and their respective binding energies are detailed in SI.

These data suggest several things. First, thiahelicenes are preferably grafted to the gold electrode through the α-carbon, rather than to the sulfur atoms as initially anticipated. This is in line with our study findings, shown in Figures 2, 3 with a red arrow. Still, the small difference in binding energies (0.5–1.5 kcal/mol) restrains the possibility of displaying selective binding (through the α-carbon) only to experiments conducted at very low temperatures. Second, in absolute terms, binding energies are relatively small, and it is likely the possibility of a complex scenario of multiple binding sites acting cooperatively on a gold surface, so dissecting a single-atom contact in a break-junction experiment may be extremely difficult. That makes the actual experiment quite intricate and challenging to perform. Attempts to measure conductance in such molecules at room temperature have afforded unexpectedly poor results (de Ara et al., 2024), presumably in our opinion due to, in addition to all of the above, thermal molecular jiggling between the electrodes. We hope that future, more accurate low-temperature experiments may allow us to differentiate between the different binding topologies reported in this article.

Now, a final note must be made to assess the effect of DTH size within the same parity series. The interface coupling between the molecule and the electrode greatly affects the conducting behavior in conjugated molecular systems. This interaction is governed by many factors, such as the HOMO-LUMO gap and the alignment of this gap to the Fermi energy level of the electrode (Liu et al., 2009; Wang et al., 2009). For maximizing the electron transmission, a certain degree of alignment between the molecular orbital levels and the Fermi level is required (Xin et al., 2019). For completeness, we report here a DFT calculation of the molecular orbitals usually dominating the conductance, which are the HOMO and the LUMO, as well as the optical gap of all the DTHs covered in this work (Figure 5).

The optical gap of a molecule corresponds to the energy of the lowest electronic transition accessible via absorption of a single photon and accounts for the stabilization energy of the electron–hole pair created during the transition (Bredas, 2014). This leads to smaller and usually more reliable gaps compared to the gross estimate provided by DFT HOMO-LUMO gaps. In our systems, similar to other conjugated molecules, the optical gap decreases with increasing molecular size. However, this effect is not very pronounced here, i.e., less than a tenth of an eV (ca. 0.08 eV) per added ring, displaying a similar value for all the series (exo, exo–endo, or endo). For a given size, differences in the optical gap are equally very small, following the order exo > exo–endo > endo. This gap is a crucial parameter in determining the conductance of molecular wires since it is directly correlated to the barrier that must be crossed by the electrons in the tunneling process (Liu et al., 2008).