Double Rotors with Fluxional Axles: Domino Rotation and Azide–Alkyne Huisgen Cycloaddition Catalysis

Abstract The simple preparation of the multicomponent devices [Cu4(A)2]4+ and [Cu2(A)(B)]2+, both rotors with fluxional axles undergoing domino rotation, highlights the potential of self‐sorting. The concept of domino rotation requires the interconversion of axle and rotator, allowing the spatiotemporal decoupling of two degenerate exchange processes in [Cu4(A)2]4+ occurring at 142 kHz. Addition of two equiv of B to rotor [Cu4(A)2]4+ afforded the heteromeric two‐axle rotor [Cu2(A)(B)]2+ with two distinct exchange processes (64.0 kHz and 0.55 Hz). The motion requiring a pyridine→zinc porphyrin bond cleavage is 1.2×105 times faster than that operating via a terpyridine→[Cu(phenAr2)]+ rupture. Finally, both rotors are catalysts due to their copper(I) content. The fast domino rotor (142 kHz) was shown to suppress product inhibition in the catalysis of the azide–alkyne Huisgen cycloaddition.

Dynamic functional devices composed of multiple molecular components are attracting ever-increasing interest for two major reasons. [1][2][3] Firstly, the impressive properties of multicomponent machines have been amply demonstrated by nature, [4] and secondly, the facile exchange of components in such systems opens the door for self-repair during operation or even evolution toward novel emerging properties.
While fluxional molecules involving bond cleavage/formation have a long-standing history, [12] related dynamics within multiple degenerate structures in supramolecular multicomponent devices is scarce. [13,14] Herein, for the first time, multicomponent rotors [15][16][17][18][19] are equipped with fluxional axles allowing them to undergo domino [20] rotation. Such domino rotors are characterized by a unique feature: the rotator arm of the first rotor subunit intra(supra)molecularly interconverts into the axle of the second rotor subunit. The homodimeric double rotor [Cu 4 (A) 2 ] 4+ is held together by two pyridine (py)!copper(I) phenanthroline interactions (= HETPYP binding: [21] Heteroleptic Pyridine and Phenanthroline complexation). Two further copper(I) phenanthroline sites without additional ligands serve as recipient stations for the rotator arm (Scheme 1). By design, only one of the pyridine arms of A serves as a rotational axis at a given time.
In the homodimeric rotor the domino rotations occur via isoenergetic transition states.
In order to generate two diverse domino rotational exchange processes with distinct activation barriers, we constructed the heteromeric double rotor [Cu 2 (A)(B)] 2+ using two different binding motifs, a HETTAP (= Heteroleptic Terpyridine and Phenanthroline complexation) [22] and a pyridine!zinc porphyrin (N py !ZnPor) [10] coordination. Finally, in both rotors the effect of rotational motion on suppression of product inhibition was probed in a copper(I)catalyzed [23,24] azide-alkyne Huisgen [25] cycloaddition. Although both rotors have the same kind of copper(I) centers, their catalytic activity is different due to dissimilar rotational rates.
The design of ligands A and B was guided by the geometric fit at the coordination sites in both aggregates [Cu 4 (A) 2 ] 4+ and [Cu 2 (A)(B)] 2+ . With ligand B being known from former work, [26] we had only to synthesize ligand A, in which two shielded phenanthrolines and one pyridyl unit are Scheme 1. Two two-axle double rotors. LogK denotes the binding constant of the single-step pyridine association and logb describes the overall stability.
connected to the 1,3,5-positions of a benzene core. The geometry of A should lead to a dimeric parallelogram-type structure in [Cu 4 (A) 2 ] 4+ with two antiparallel pyridyl units operating as axles. After mixing two equiv of copper(I) and one equiv of A, the two-component dimer [Cu 4 (A) 2 ] 4+ was furnished quantitatively, as evidenced by spectroscopic data ( 1 H NMR, 1 H-1 H COSY, UV/Vis, ESI-MS). The single peak in the ESI mass spectrum at m/z = 684.8 with correct isotopic distribution (SI, Figure S37), the single set of signals in the 1 H-DOSY {D = 4.1 10 À10 m 2 s À1 , r % 12.9 , (Supporting Information, Figure S21 Kinetic analysis provided the exchange frequency (k) at different temperatures with k 298 = 142 kHz (at 298 K). The activation data were determined as DH°= 51.1 AE 0.7 kJ mol À1 , DS°= 25.5 AE 3.0 J mol À1 K À1 , and DG°2 98 = 43.5 kJ mol À1 (Supporting Information, Figure S24). Since the kinetic data for exchange in [Cu 4 (A) 2 ] 4+ are very similar to those of previously reported nanorotors that operate via single N py ![Cu(phenAr 2 )] + dissociation (DG°2 98 = 46.6 kJ mol À1 ), [1b,16, 27,32] there is convincing evidence that only one HETPYP interaction is cleaved at any given time in [Cu 4 (A) 2 ] 4+ . Hence, the mechanism of exchange seems to follow an intramolecular nondirectional rotation involving two axles with dissociation of one N py ![Cu(phenAr 2 )] + interaction being the rate-limiting step. Due to symmetry, both rotational axles in the homodimeric rotor [Cu 4 (A) 2 ] 4+ have equal probability to dissociate.
For the heteromeric rotor, we have chosen ligands A and B held together by two orthogonal dynamic interactions, [26] that is, the weak N py !ZnPor interaction (Scheme 2; [(2)·(4)], logK 2!4 = 4.3) and the robust HETTAP linkage (for [Cu(1)(3)] + , log b = 9.3). [19] Ligand B was designed in a way that the pyridine terminus of ligand A should bind to the ZnPor of B, while the tridentate terpyridine (tpy) ligand of B is connected with the copper(I) phenanthroline unit of A. When the HETTAP complex is intact it will serve as rotational axle, leading to an exchange of the pyridine terminus of A between both ZnPor sites of B. Similarly, when the N py !ZnPor interaction serves as rotational axle, both copper(I)-loaded phenanthrolines will exchange their position by HETTAP dissociation/association. The N py !ZnPor bond cleavage should have a rather low energy of activation, [26] whereas the HETTAP binding is known for its slow dynamics as demonstrated by Sauvage using interlocked molecular machines. [7a, 28] The  tion motif was further validated by a redshift of the ZnPors Q-band from 537 to 544 nm in the UV/Vis spectrum (Supporting Information, Figure S43). Finally, DOSY and elemental analysis confirmed quantitative formation of [Cu 2 -(A)(B)] 2+ (Supporting Information, Figure S22).
Rapid exchange of the pyridyl head of A between both ZnPor sites of B in [Cu 2 (A)(B)] 2+ was indicted by a single set of ZnPor signals in the 1 H NMR spectrum. In contrast, two individual signal sets for both phenanthrolines suggested that exchange at the HETTAP sites was slow on the NMR timescale (at 25 8C).
In order to measure the fast rotational dynamics in rotor [Cu 2 (A)(B)] 2+ Figure S23). Using the kinetic data over the whole temperature range furnished the activation parameters as DH°= 46.9 AE 0.4 kJ mol À1 and DS°= 4.8 AE 1.2 J mol À1 K À1 and the activation free energy for spinning at 298 K as DG°2 98 = 45.5 kJ mol À1 (Supporting Information, Figure S23).
The kinetics of the slow exchange requiring cleavage at the HETTAP site was evaluated by a 1 H-1 H ROESY (Figure 3 B) experiment in CD 2 Cl 2 /CD 3 CN (5:1) because a cross correlation was observed between proton 9u-H of the copper(I)-loaded phenanthroline (at d = 6.96 ppm) and proton 9c-H of the HETTAP complexed phenanthroline signal (at d = 5.98 ppm). [29] The activation parameters for the corresponding exchange were determined at 298 K (k 298 = 0.55 s À1 and DG°2 98 = 74.8 kJ mol À1 ) (SI, Figure S25).
From the activation data of rotor [Cu 2 (A)(B)] 2+ (DG°2 98 = 45.5 kJ mol À1 ) and those of previously reported rotors (46.6 kJ mol À1 ), [26] we may safely conclude that the fast exchange is the result of the cleavage of a single axial N py ! ZnPor interaction. In order to achieve a better understanding of the rate-limiting step in the slow exchange process involving the HETTAP site, we self-assembled the mono-  Figure S26). This finding suggested that the copper(I) ion does not travel along with the terpyridine (tpy) unit. Thus, in [Cu 2 (A)(B)] 2+ the tpy! [Cu(phenAr 2 )] + dissociation seems to be relevant in the ratelimiting step of exchange at the HETTAP site and not the phenAr 2 ![Cu(tpy)] + dissociation. This interpretation is in line with the relative binding constants in mixed terpyridinephenanthroline copper(I) complexes, since phenanthroline is more strongly bound (logK ½Cuð1Þ þ = 5.1) than terpyridine (logK 3!½Cuð1Þ þ = 4.2). [19,30,28] After a successful demonstration of the dynamic behavior of homodimeric and heteromeric domino rotors, their catalytic activity as catalysts was tested. Both two-axle rotors display at any given time coordinatively free copper(I) phenanthroline units that have the potential to catalyze a click reaction. Thus, a mixture of [Cu 4 (A) 2 ] 4+ (0.90 mm), 5, and 6 in 1:20:20 ratio was dissolved in CD 2 Cl 2 /CD 3 CN (5:1) and heated at 50 8C ( Figure 5). After 2 h of heating, 63 % of the click product 7 was detected by 1 H NMR analysis (Supporting Information, Chapter 6). Analogously, [Cu(1)] + (3.6 mm ; the fourfold concentration accounts for a proper comparison with [Cu 4 (A) 2 ] 4+ containing four copper centers) as reference catalyst, 5, and 6 in a 1:5:5 ratio were dissolved in   Figure S20). In addition, it was shown that deliberate addition of product 7 to [Cu 4 (A) 2 ] 4+ reduced the yield of the catalytic reaction (Supporting Information, Figure S31). Actually, a linear correlation was seen between loss in the catalytic yield (%) and the amount (mol %) of externally added product (Figure 5 b). The effect of higher turnover number due to increased product liberation in the fast rotor [Cu 4 (A) 2 ] 4+ comes as no surprise. Similar multicomponent rotors [31] have emerged as an attractive class of catalysts because they are able to mimic sophisticated machinery from nature, such as ATP synthase, [4] in their ability to suppress product inhibition through a nanomechanical motion. Thus rotating catalysts, even when rotation is stochastic, are more efficient than their static prototypes, [32] a result that we see confirmed by comparing [Cu 4 (A) 2 ] 4+ with [Cu(1)] + using accurately the same amount of copper(I) ( Table 1).
In comparison, a mixture of the heteromeric two-axle rotor [Cu 2 (A)(B)] 2+ (1.8 mm), 5, and 6 in a 1:10:10 ratio of furnished rather low yield (28 %) under analogous conditions. This effect on the yield and the low rotational exchange (k 298 = 0.55 s À1 ) at the copper(I) sites are consonant to those of the reference reaction; apparently a slow motion does not provide a significant reduction of product inhibition. The faster the rotational exchange at the active copper(I)-loaded phenanthroline stations, the higher the catalytic activity, [32] actually also with other acetylenes ( Figure S33).
In conclusion, we have fabricated a new class of domino nanorotors with two fluxional axles. The rotational dynamics in the homodimeric two-axle double rotor is governed by the dissociation of the N py ![Cu(phenAr 2 )] + interaction. The full exchange at all four copper phenanthroline sites requires that both pyridine arms act as fluxional axles. In the case of the heteromeric two-axle rotor, rotational exchange at the HETTAP site is 1.2 10 5 times slower than the swapping of the pyridine arm between the two ZnPor units. Here we were able to individually control both fluxional axles and their dynamics.
Finally, we have utilized the speed change on going from the homodimeric to heteromeric two-axle rotor to modulate the product inhibition of a click reaction. Attractive future goals are to modify the distinct rotational modes in heteromeric domino rotors by changing the components (A or B or metal ions, for example, Ag + , Zn 2+ ) and to implement more than two fluxional axles in domino rotation.  Table 1: Exchange frequency of catalysts, their catalytic reaction yields, and the rates v 0 at time t = 0.
[a] Yields determined from three independent runs. For conditions see text and the Supporting Information, Chapter 6.