**1. Introduction**

In recent decades, the goal of creating artificial molecule-scale mechanical systems (e.g., molecular machines, shuttles, gears, rotors, turnstiles) has attracted enormous attention within the synthetic community [1,2]. Herein, we focus on one specifically selected approach-transition metal catalysed coupling [3], towards the synthesis of the "skeletal backbones" of molecular mechanical systems, their basic construction elements. Incorporation of these rigid fragments allows us to investigate the functional behaviour of a wide selection of individual and mechanically interlocked systems, such as those illustrated schematically in Figure 1. Even a brief perusal of the vast literature on the syntheses of molecular machines leads one to conclude that transition metal-catalysed (especially palladium-catalysed) coupling reactions play a prominent role in this area of modern synthetic chemistry, as exemplified below.

Following the pioneering work on corannulene by Siegel et al. [4], the study of an umbrella-type inversion process in the nitrogen-embedded buckybowl, 2, from its precursor, 1, was made possible via three sequential Pd-catalysed coupling reactions (shown in red, Scheme 1). The bowl-to-bowl inversion barrier was reported to be 98 kJ mol−<sup>1</sup> [5].

**Figure 1.** Examples of fluxional molecular systems: (**i**) bowl-to-bowl inversion, (**ii**) shuttling, (**iii**) pirouetting, and (**iv**) rotating.

**Scheme 1.** Efficient synthesis of the buckybowl, **2,** via Pd-catalysed C-C bond formation.

Palladium-catalysed aryl-acetylene coupling reactions have been used by Stoddart's group to construct long rigid molecular spacers of the [2]rotaxane shuttle, 3, (Figure 2) whereby low shuttling energy barriers were reported unless a "speed bump" moiety was attached to the naphthalene ring [6].

**Figure 2.** Fast degenerate shuttling of a cyclobis(paraquat-*p*-phenylene) ring along a rigid dialkyne framework.

In a similar vein, we preceded these results with our non-degenerate tripodal two-station rotaxane shuttle, 4 (Figure 3), which was designed to be adsorbed in an oriented way on a titanium dioxide surface [7]. Unexpectedly, the synthesis of the rigid spacer, B, via Suzuki-type cross-coupling proved problematic, as no terphenyl product could be isolated for R = H, X = OH; however, when R = Me, X = OH, the terphenyl linker was isolable in 75% yield, and was fully characterised by X-ray crystallography [8]. Similarly, catalysed coupling reactions were recently used by Loeb and co-workers in the synthesis of oligo-*p*-phenylene components of their solid-state shuttle 5 [9].

In the present report, we collect together some of our own work and focus chiefly on molecular machines and rotors, whereby a defined pair of molecular fragments is connected by a rotatable single C-C bond (Figure 1, part iv) to attain restricted intramolecular rotation. Our goal was to prepare molecular gearing systems, incorporating a paddlewheel-shaped triptycene fragment attached to a molecular shuttle. The dynamic behaviour of the shuttle could, in principle, allow free rotation or function as a molecular brake, as in Figure 4.

**Figure 3.** Palladium-catalysed formation of C–C bonds in the tripodal rigid non-degenerate shuttle, **4** [7,8], and the rigid ring-in-ring shuttle, **5** [9].

**Figure 4.** Schematic of a molecular gearing system in which free rotation of the paddlewheel P is controlled by sliding the "latch" Y of the shuttle S.

The planned approach was to prepare 9-(3-indenyl)anthracene, 6, and 9-(2-indenyl)anthracene, 7. The synthesis of such molecules requires the attachment of the bicyclic unit to anthracene, with subsequent Diels-Alder addition of benzyne to form the corresponding indenyl triptycenes, 8 and 9, respectively. One could then envisage incorporation of a bulky organometallic fragment that could undergo a reversible η<sup>6</sup> Ɂ η<sup>5</sup> haptotropic shift across the indenyl framework, whereby the pentahapto-complexed system would hinder paddlewheel rotation. An appropriate cross-coupling procedure appeared to offer the most propitious route to 6 and 7.

#### **2. Indenyl Anthracenes and Triptycenes**

#### *2.1. Synthetic Aspects*

Our initial target, 9-(3-indenyl)anthracene, 6, was successfully obtained in 42% yield via the Stille coupling of 9-bromoanthracene and 1-(tributylstannyl)indene catalysed by Pd(PPh3)4 in DMF [10]. Interestingly, benzyne addition to 6 occurred not only at the C(9) and C(10) positions of the anthracene to give the desired triptycene, 8, but also to the indenyl substituent, thus furnishing the [4 + 2] cyclo-adduct, 10 (Scheme 2). These three structures were validated by X-ray crystallography [10,11], and are shown in Figures 5 and 6. These data revealed the extent of steric crowding between the indenyl and the peri-hydrogens at C(1), C(8) and C(13) of the triptycyl unit in 8 that engenders a 50 kJ mol−<sup>1</sup> rotational barrier, as measured by variable-temperature NMR, even before incorporation of the organometallic moiety.

**Scheme 2.** Palladium-catalysed route to **6**, and its benzyne adducts, **8** and **10**.

**Figure 5.** Molecular structures of 9-(3-indenyl)anthracene, **6**, and 10-(anthracen-9 -yl)-[4a,9]-methano-4a,9-dihydrophenanthrene, **10**.

**Figure 6.** Molecular structure of 9-(3-indenyl)triptycene, **8**, and a space-fill view, from the reverse angle, to emphasise its molecular crowding.

In light of these observations, it was decided to focus on the preparation of 9-(2-indenyl)anthracene, 7, the precursor to 9-(2-indenyl)triptycene. A previous report indicated that Heck-type palladium cross-couplings of indene with iodoarenes led primarily to formation of 2-arylindenes and, to a lesser extent, 3-arylindenes [12]. However, since their characterisations were based only on NMR data, we chose to perform the reaction of indene with 1-bromo-4-iodobenzene, using palladium acetate as the catalyst and triethylamine as the base [11]. Gratifyingly, the major product was identified unequivocally by X-ray crystallography as 2-(4-bromophenyl)indene, 11, whose structure appears as Figure 7.

**Figure 7.** Molecular structure of 2-(4-bromophenyl)indene, **11**.

As anticipated, the iodo substituent, rather than bromo, was replaced (Scheme 3, upper). Nevertheless, we found that bromobenzene and indene in the presence of dichloro-bis(tri-*O*tolylphosphine)palladium(II) in DMF at 100 ◦C delivered 2-phenylindene in 90% yield. Therefore, with some confidence (unjustified, as it turned out), we carried out the reaction of indene with 9-bromoanthracene under these same conditions in the expectation of forming the desired 9-(2-indenyl)anthracene, 7, as the major isomer. However, as depicted in Scheme 3 (lower), the already known 9-(3-indenyl)anthracene, 6, together with the indeno-dihydroaceanthrylene, 12, (Figure 8) were the only cross-coupled products [11].

**Scheme 3.** Pd-catalysed formation of arylindenes. Reagents and conditions: (**i**) indene, Pd(OAc)2, Et3N, 100 ◦C; (**ii**) as for (**i**) but using (*O*-tolyl3P)2PdCl2.

**Figure 8.** Molecular structure of indeno [1,2-α]-10,16-dihydroaceanthrylene, **12**.

These results may be rationalised in terms of the two different modes of insertion of indene into the palladium-aryl linkage of (9-anthracenyl)Pd(*O*-tolyl3P)2Br, thereby generating intermediates 13 and 14 (Scheme 4). In the former case, *syn*-elimination of HPd(*O*-tolyl3P)2Br occurs readily to yield 9-(1-indenyl)anthracene that subsequently rearranges to its 9-(3-indenyl)anthracene counterpart, 6. However, intermediate 14 lacks a suitably positioned hydrogen *syn* to palladium and so instead undergoes intramolecular palladation of the adjacent anthracene ring to form the aceanthrylene, 12. Such a mechanism should leave the two bridgehead hydrogens in a *syn* arrangement, as is indeed evident in the X-ray crystal structure shown in Figure 8, which emphasizes the folded nature of the molecule about the common bond linking the two five-membered rings.

We note that such a cyclopentenylation to form an aceanthrylene was first observed by Dang and Garcia-Garibay [13]. They found, surprisingly, that their attempt to prepare 9- (trimethylsilylethynyl)anthracene, 15, via the Sonogashira reaction between 9-bromoanthracene and ethynyltrimethylsilane in the presence of (Ph3P)2PdCl2, CuSO4/Al2O3 and Et3N in refluxing benzene gave instead 2-(trimethylsilyl)aceanthrylene, 16, as the major product. Once again, it is now evident that the lack of a *syn*-disposed hydrogen led to intramolecular palladation, as illustrated in Scheme 5. Since that time, Plunkett has elegantly exploited this cyclopentenylation process to produce polycyclic systems that represent fragments of fullerenes [14,15]. We note in passing that, depending on the particular Sonogashira conditions selected, the yield of 9-(trimethylsilylethynyl) anthracene, 15, has not only been improved to 98% [16], but also that one can produce 4-(9-anthracenyl)- 1,3-bis(trimethylsilyl)-but-3-en-1-yne, 17, in moderate yields; a detailed mechanism has been advanced [17,18].

**Scheme 4.** Proposed mechanisms for the indenylation of 9-bromoanthracene (L = PAr3).

**Scheme 5.** Formation of multiple products from the palladium-catalysed reaction of 9-bromoanthracene and ethynyltrimethylsilane.

Having now established that the Heck-type reaction of indene with 9-bromoanthracene provides an improved route to 9-(3-indenyl)anthracene, 6, rather than its desired 2-indenyl counterpart, 7, we chose instead to attempt a Suzuki-type cross-coupling, as illustrated in Scheme 6. The reaction of 9-bromoanthracene with 2-indenylboronic acid catalysed by dichloro-bis(diphenylphosphinoferrocene)palladium(II), (dppf)PdCl2, in ethanol-toluene at 75 ◦C, using Na2CO3 as the base, delivered 9-(2-indenyl)anthracene, 7, in 52% yield, together with the homo-coupled product, 2,2 -biindenyl (9%) [11].

**Scheme 6.** Pd-catalysed Suzuki-type coupling of aryl bromides. Reagents and conditions: (**i**) 2 indenylboronic acid, (Ph3P)2PdCl2, (1 mol%), ethanol-toluene, Na2CO3, 30 h, 75 ◦C; (**ii**) as for (**i**) but using (dppf)PdCl2.

With 9-(2-indenyl)anthracene in hand, Diels-Alder addition of benzyne to form 9-(2 indenyl)triptycene, 9, proceeded in 81% yield. This much improved yield of 9 compared to that of 9-(3-indenyl)triptycene, 8 (42%) may be attributed to the trajectory of approach for a potential [4 + 2] cycloaddition of benzyne to the five-membered ring of either 6 or 7. As depicted in Schemes 2 and 7,

in the former case, access to both the anthracene and the indene is available, whereas, in the 2-indenyl system, the latter process would be blocked by the proximity of the planar anthracenyl, thus leading to a single product in enhanced yield [11].

**Scheme 7.** Reaction of benzyne and 9-2(indenyl)anthracene yields only a single adduct, **9**.
