Keywords

1 Introduction

The interaction between light and matter lies at the heart of the most important processes of life [1]. Light consists of photons which are exploited by natural systems as both quanta of energy and elements of information. All the natural phenomena related to the interaction between light and matter and the great number of applications of photochemistry in science and technology can ultimately be traced back to these two aspects of light. Living examples of this double-faced nature of light are provided by the two most important photochemical processes taking place in the biological world: photosynthesis and vision.

A variety of functions can also be obtained from the interaction between light and matter in artificial systems [2]. The type and utility of such functions depend on the degree of complexity and organization of the chemical systems that receive and process the photons. Indeed, understanding the interaction between light and molecules, together with the progress in chemical synthesis, has led to the point where one can conceive and assemble artificial multicomponent systems capable of using light as an energy supply and/or as an input signal. The construction of such ultraminiaturized devices and the comprehension of their working mechanisms are based on the concepts of supramolecular chemistry [3] and photochemistry [4], and have become topics of great interest for nanoscience.

In general terms, a molecular machine can be defined [5, 6] as an assembly of a discrete number of molecular components designed to perform specific mechanical movements in response to appropriate external stimuli. The nanomachines of the biological world [7, 8], such as ATPase and myosin, are the premier, proven examples of the feasibility and utility of nanotechnology [9], and constitute a sound rationale for attempting the realization of artificial molecular machines. As the bottom-up construction of devices of such a complexity is currently a prohibitive task, chemists have been trying to make much simpler systems, to understand the principles and processes at the basis of their operation, and to investigate the problems posed by interfacing them with the macroscopic world.

The photochemical tweezers described [10] by Shinkai and coworkers in the early 1980s can be considered as primitive examples of molecular machines; in fact, systems of this type have evolved to a considerable degree of complexity [11]. The concept of artificial photochemically driven molecular machine, however, was illustrated and realized for the first time in 1993 in a landmark paper by Balzani, Stoddart and coworkers [12]. Since then, the development of powerful synthetic methodologies, combined with a device-driven ingenuity evolved from the attention to functions and reactivity, have led to remarkable achievements. Several reviews [6, 13–18] and a monograph [5] dealing with light-operated molecular machines are available.

In this chapter we discuss the possibilities offered by the use of light to operate molecular machines, namely, to supply energy to the system and/or gain information on its state. After some general considerations, a few recent examples of photochemically controlled molecular machines based on rotaxanes and related compounds, most of them studied in our laboratory, are described. They present some small, although interesting, advancements in exploiting the peculiar properties of light and its interaction with matter to obtain useful functions.

2 The Role of Light

2.1 Energy Supply and Monitoring Signals

As happens in the macroscopic world, molecular-level devices and machines need energy to operate and signals to communicate with the operator [19]. The most obvious way to supply energy to a chemical system is through an exergonic chemical reaction. Not surprisingly, the majority of the molecular motors of the biological world are powered by chemical reactions (e.g., ATP hydrolysis) [7–9]. Richard Feynman observed [20] that “an internal combustion engine of molecular size is impossible. Other chemical reactions, liberating energy when cold, can be used instead.” This is exactly what happens in our body, where the chemical energy supplied by food is used in long series of slightly exergonic reactions to power the biological machinery that sustains life.

If an artificial molecular machine has to work by inputs of chemical energy, it will need addition of fresh reactants (“fuel”) at any step of its working cycle, with the concomitant formation of waste products. Accumulation of waste products, however, will compromise the operation of the device unless they are removed from the system, as happens in our body as well as in macroscopic internal combustion engines. The need to remove waste products introduces noticeable limitations in the design and construction of artificial molecular machines based on chemical fuel inputs.

Chemists have long known that photochemical and electrochemical energy inputs can cause the occurrence of endergonic and reversible reactions. In the last few years the outstanding progress made by supramolecular methodologies [21, 22], in particular photochemistry [23] and electrochemistry [24], has led to the design and construction of molecular machines powered by light or electrical energy which work without formation of waste products.

In the context of artificial nanomachines, light energy stimulation possesses a number of further advantages, particularly in comparison with chemical stimulation. First of all, the amount of energy conferred to a chemical system by using photons can be carefully controlled by the wavelength and intensity of the exciting light, in relation to the absorption spectrum of the targeted species. Such energy can be transmitted to molecules without physically connecting them to the source (no “wiring” is necessary), the only requirement being the transparency of the matrix at the excitation wavelength. Other properties of light, such as polarization, can also be utilized. Lasers provide the opportunity of working in very small spaces and extremely short time domains, and near-field techniques allow excitation with nanometer resolution. On the other hand, the irradiation of large areas and volumes can be conveniently carried out, thereby allowing the parallel (or even synchronous) addressing of a very high number of individual nanodevices.

Because molecules are extremely small, the observation of motion at the molecular level, which is crucial for monitoring the operation of a molecular machine, is not trivial. In general, the motion of the component parts should cause readable changes in some chemical or physical properties of the system. Photochemical methods are also useful in this regard. As a matter of fact, photons can play with respect to chemical systems the dual role of writing (i.e., causing a change in the system) and reading (i.e., reporting the state of the system) [23]. As discussed in the introduction, this is primarily true in nature, where sunlight photons are employed both as energy quanta in photosynthetic processes, and as information elements in vision and other light-triggered processes. For example, luminescence spectroscopy is a valuable method because it is easily accessible and offers good sensitivity and selectivity, along with the possibility of time- and space-resolved studies [25]. In particular, flash spectroscopic techniques with laser excitation enable the study of extremely fast processes.

The use of light to power nanoscale devices is relevant for another important reason. If and when a nanotechnology-based industry will be developed, its products will have to be powered by renewable energy sources, because it has become clear that the problem of energy supply is a crucial one for human civilization for the years ahead [26]. In this frame, the construction of nanodevices, including natural-artificial hybrids [27] that harness solar energy in the form of visible or near-UV light, is indeed an important possibility.

2.2 Autonomous Operation

An important feature of a molecular machine is its capability to exhibit an autonomous behavior, i.e., to operate without external intervention (in other words, in a constant environment) as long as the energy source is available. Hence, autonomous molecular machines are free-running devices that do not require controlled and repeated addition of chemicals or other environmental changes. Natural motors are autonomous, in most cases acting as catalyzers for the fueling reaction. Except for some work on DNA nanodevices [28, 29], the chemically powered artificial molecular machines reported so far are not autonomous because after the mechanical movement induced by a given input they need another, opposite, input to be reset.

The design of autonomous nanomachines can take advantage from reversible photochemical processes. For instance, the operation of the machine could be based on a photoinduced sequence of processes that lead the system through transient electronic and nuclear (mechanical) states; the final deactivation of the system to the ground state provides an automatic reset and closes the cycle of operation. Alternatively, the mechanical motion could be related to the light-triggered switching between two stable states as happens in photochromic systems. These approaches will be more convenientely discussed in the examples illustrated in Sects. 4 and 5.

3 Threaded and Interlocked Multicomponent Species

In principle, molecular machines can be designed starting from several kinds of molecular and supramolecular systems [5, 6, 13, 30–42], including DNA [43, 44]. However, for the reasons mentioned below, most of the systems constructed so far are based on interlocked molecular species such as rotaxanes and related species. The names of these compounds are derived from the Latin words rota and axis for wheel and axle, respectively. Rotaxanes [45] are minimally composed (Fig. 1a, b) of a dumbbell-shaped molecule surrounded by a macrocyclic compound (the “ring”) and terminated by bulky groups (“stoppers”) that prevent disassembly. If the stoppers are not present, the assembled species is denoted as a pseudorotaxane and in solution it equilibrates with the separated axle-type and ring components (Fig. 1c). Important features of these systems arise from noncovalent interactions between the components that contain complementary recognition sites. Such interactions are responsible for the self-assembly of pseudorotaxanes and efficient template-directed syntheses of rotaxanes, and include electron donor–acceptor ability, hydrogen bonding, hydrophobic–hydrophylic character, π–π stacking, electrostatic forces and, on the side of the strong interaction limit, metal-ligand bonding.

Fig. 1
figure 1

Schematic representation of (a) ring shuttling and (b) ring rotation in rotaxanes, and (c) the threading/dethreading equilibrium involving the axle-type and ring components of a pseudorotaxane

Full size image

Rotaxanes are appealing systems for the construction of molecular machines because (1) the mechanical bond enables a large variety of mutual arrangements of the molecular components while conferring stability to the system, (2) the interlocked architecture limits the amplitude of the intercomponent motion in the three directions, (3) the stability of a specific arrangement (co-conformation) is determined by the strength of the intercomponent interactions, and (4) such interactions can be modulated by external stimulation. Two interesting molecular motions can be envisaged in rotaxanes, namely translation, i.e., shuttling of the ring along the axle (Fig. 1a), and rotation of the ring around the axle (Fig. 1b). Hence, rotaxanes are good prototypes for the construction of both linear and rotary molecular machines. Systems of the first type, termed molecular shuttles, constitute the most common implementation of the molecular machine concept with rotaxanes.

The assembly–disassembly of the axle-type and ring components of a pseudorotaxane (Fig. 1c) is reminiscent of the threading/dethreading of a needle and can be controlled by external stimulation [5, 46]. Studies on switchable pseudorotaxanes are important for the development of less trivial unimolecular machines based on rotaxanes and related interlocked compounds.

It should be noted that, unless one of the molecular components of the assembly is attached to a fixed support (e.g., a surface) in a rigid manner, “moving” and “static” parts cannot be meaningfully identified. Hence, for systems working in solution, it is correct to refer to the processes shown in Fig. 1 as relative movements of the components with respect to one another.

Moreover, the drawings shown in Fig. 1, while providing a simple structural and topological representation, are somewhat misleading because they give the impression that rotaxanes are made of rigid molecular components, which is not the case for the vast majority of the systems reported so far. However, in order to obtain clear-cut mechanical movements the molecular components should exhibit at least some stiffness. As will be evidenced by the examples in the following sections, this feature for molecular machines is most commonly fulfilled by utilizing molecular components that possess rigid subunits in their structures.

Interestingly, the dumbbell component of a molecular shuttle exerts on the ring motion the same type of directional restriction imposed by the protein track for linear biomolecular motors (an actin filament for myosin and a microtubule for kinesin and dynein) [7]. It should also be noted that interlocked molecular architectures are largely present in natural systems – for instance, DNA catenanes and rotaxanes are known [45]. Many processive enzymes, that is, enzymes that remain attached to their biopolymer substrates (DNA, RNA, or proteins) and perform multiple rounds of catalysis before dissociating, are thought to exhibit a rotaxane structure, as confirmed for example by the observation of the crystal structure of DNA λ-exonuclease ([47] and references therein). Clearly, the unique aspect of the rotaxane architecture, that is, the mechanical binding of the catalyst with the substrate which leaves the former free to displace itself along the latter without losing the system’s integrity, is utilized by Nature to enhance the activity of processive enzymes.

4 Molecular Shuttles and Related Systems

Molecular shuttles are rotaxanes in which the ring component can move along the axle portion of the dummbell-shaped component [48]. These types of systems constitute a common implementation of the molecular machine concept with artificial chemical systems [5, 6, 13]. The minimal design for a controllable molecular shuttle [49] involves the incorporation of two different recognition sites (stations) on the axle. The ring originally encircles the most efficient station until a chemical reaction, properly activated, changes the relative affinity between the ring and the stations, thereby causing the translation of the ring to the other station.

Light stimulation of molecular shuttles can have two main effects: (1) modify the interactions of the ring with the stations, i.e., the thermodynamic stability of the mechanical states involved in the operation of the machine, and/or (2) control the transition rates between such states, i.e., the kinetics of the mechanical motions. Photoinduced electron-transfer processes, both intramolecular [50, 51] and intermolecular [52, 53], and photoisomerization reactions [54–58] have been largely employed for purpose (1) above. Photoisomerization processes can also be profitably used for controlling the kinetics of molecular motions (point (2)) [59], as shown in some of the examples reported in the following.

4.1 Molecular Shuttling Operated by Photoswitching of Radical–Radical Interactions

An example of a molecular shuttle based on photoinduced electron transfer and exploiting radical–radical interactions in water was recently described [60]. The design of the system relies on previously investigated rotaxanes utilizing ruthenium(II) polypyridine-type complexes as electron-transfer photosensitizers [50, 61, 62]. The investigated rotaxane 1 8+ (Fig. 2) consists of the electron acceptor macrocycle cyclobis(paraquat-p-phenylene), CBPQT4+, and a dumbbell-shaped component that contains four units: (1) a Ru(bpy)3 2+-type complex (bpy = 2,2'-bipyridine) which plays the dual role of a photosensitizer and a stopper (Ru2+), (2) a 1,5-dioxynaphthalene electron donor unit (DNP), (3) a 4,4'-bipyridinium electron acceptor site (BPY2+), and (4) a diisopropylbenzene moiety as the second stopper. The behavior of rotaxane 1 8+ and of its molecular components was investigated by means of NMR and UV-visible spectroscopy, and by electrochemical techniques.

Fig. 2
figure 2

Structural formula and schematic representation of the light-induced and chemically assisted shuttling movements of bistable rotaxane 1 8+

Full size image

In the starting co-conformation of 1 8+, the CBPQT4+ macrocycle encircles the DNP site on account of π-donor–acceptor interactions (Fig. 2). Irradiation of a deoxygenated water solution of the rotaxane with visible light (absorbed exclusively by the Ru2+ unit) causes an electron transfer from the excited state of Ru2+ to the bipyridinium dications present both in the CBPQT4+ ring and in the axle (BPY2+). Triethanolamine (TEOA) added at a sufficiently high concentration acts as a sacrificial reductant, thereby preventing the back electron transfer from the bipyridinium radical cations to the oxidized Ru3+ center. As a consequence, upon exhaustive irradiation, all three of the bipyridinium units (two in the ring and one in the axle) are converted to the monoreduced forms and are accumulated in the solution. The resulting diradical dicationic CBPQT2( •+) ring undergoes translation in order to encircle the BPY•+ unit in the axle, on account of stabilizing radical–radical interactions (signaled by the characteristic absorption bands in the visible region [63, 64]) and the weakening of the π-donor–acceptor interaction between the reduced ring and the DNP donor unit. As soon as the bipyridinium radical cations are oxidized by oxygen, the restoration of the donor–acceptor interactions of CBPQT4+ and DNP, together with the increased electrostatic repulsion between CBPQT4+ and BPY2+, induces the ring to shuttle back and encircle the DNP unit (Fig. 2).

In principle, the shuttling mechanism of rotaxane 1 8+ should enable light-powered autonomous operation of the molecular machine, as observed earlier with similar systems [51, 52]. The experiments performed so far, however, do not allow one to gather information in this regard. The investigation of such an important issue would require time-resolved spectroscopic experiments, possibly replacing TEOA with a non-sacrificial reductant as an electron relay capable of electronically resetting the system when ring shuttling has occurred [51, 52].

4.2 Extension-Contraction of a Rotaxane Dimer

An exciting development in the field of artificial molecular machines has been the construction of molecules that can stretch and contract upon external stimulation [65]. Systems of this kind remind the operation of the sarcomere, that is, the basic functional element of skeletal muscles, in which the simultaneous sliding of the stacked filaments of myosin and actin powered by ATP hydrolysis causes a change in the sarcomere length [7, 8]. In a muscle fiber the extension and contraction of a large number of sarcomeres is summed up, causing a macroscopic mechanical effect. Similarly, it can be imagined that polymers obtained from monomers that can perform a reversible change in length (or volume) upon stimulation could enable the amplification of nanoscale motions up to the macroscopic scale, thus obtaining artificial molecule-based muscles [16].

The first attempts to develop artificial molecular muscles was based on the combination of the concept of controllable molecular shuttle [49, 66] with the topology of a rotaxane dimer (Fig. 3) [67]. Rotaxane-based muscles operated by chemical stimulation have been reported [68] and have recently been utilized to make covalent and coordination polymers in which collective extension and contraction effects are obtained [69, 70].

Fig. 3
figure 3

Schematic representation of the stimuli-induced extension-contraction motion of a bistable rotaxane dimer

Full size image

An interesting example of a molecular muscle operated by light is represented by compound 2 shown in Fig. 4 [71]. This rotaxane dimer comprises α-cyclodextrin (α-CD) as the macrocyclic components and photoisomerizable stilbene units in the axle components. It is well known that stilbene in the E-configuration is effectively included within α-CD. Indeed, 2D NMR spectroscopy showed that in the E,E isomer of 2 the stilbene units are surrounded by the α-CD rings. Irradiation of E,E-2 at 350 nm in water afforded a photostationary state composed of the E,E-, E,Z-, and Z,Z-isomers of 2 in a ratio of ca. 2:2:1. Subsequent irradiation at 254 nm to target the Z-stilbene units yielded a photostationary mixture containing E,E- and E,Z-2 in a ratio of ca. 6:1. The reversibility of the process was monitored by alternating irradiation at 350 nm and at 254 nm. The photoproducts were isolated by HPLC and subjected to NMR analysis, which showed that in Z,Z-2 the α-CD rings are moved off the Z-stilbene moieties towards the propyl and blocking groups. The movement of the CD rings away from the stilbene units shortens the length of the rotaxane 2 and the distance between the blocking groups. Therefore, the isomers E,E-, E,Z-, and Z,Z-2 represent, respectively, the extended, intermediate, and contracted states of a reversible, light driven molecular mimic of a muscle sarcomere (Fig. 4).

Fig. 4
figure 4

Photochemical switching between extended (E,E-), intermediate (E,Z-), and contracted (Z,Z-) structures in the rotaxane dimer 2

Full size image

4.3 A Molecular Information Ratchet

As anticipated above, by appropriate design of the axle component it is possible to control both the thermodynamics and the kinetics underlying the ring shuttling motion in a rotaxane, thus mimicking energy ratcheting mechanisms which are the basis of the operation of biomolecular machines [7–9, 72]. An interesting example is represented by compound 3 in Fig. 5a [73]. The system starts at chemical equilibrium, that is, with 85% of macrocycles on the fumaramide station and 15% on the succinamide station, as a consequence of the different affinities of the ring for these two sites (Fig. 5b). Because of the presence of the bulky silyl group R on the axle, ring shuttling in this rotaxane is associated with a large kinetic barrier. When statistical balance is broken by light excitation that causes E→Z isomerization of the fumaramide station and the kinetic barrier is removed, the system moves towards the new equilibrium. Restoring the barrier produces a system that is balanced but cannot equilibrate because shuttling is prevented. Hence, the thermal Z→E back isomerization makes the system statistically unbalanced and out of equilibrium. In this rotaxane the axle can perform the task of directionally changing the net position of the macrocycle. Eventually, the removal of the bulky substituent allows the system to reach equilibrium and restore the original ring distribution.

Fig. 5
figure 5

(a) Structural formula of rotaxane 3, in which the ring movement caused by E→Z photoisomerization is made irreversible by insertion of a barrier after the E→Z transformation and shuttling. The result is a change in the net position of the macrocycle (b)

Full size image

The same principle has been exploited to obtain the first example of a molecular information ratchet [74]. The described system is rotaxane 4 2+ which consists of a dibenzo[24]crown-8-based macrocycle mechanically locked onto a linear molecular axle (Fig. 6). Along the axle there are two stations for the ring − a dibenzylammonium (DBA+) and a monobenzylammonium (MBA+) unit − which bind the ring with comparable affinities, but are distinguishable for the purpose of monitoring the system. An α-methylstilbene unit divides the axle asymmetrically into two compartments, each containing a station. When the stilbene unit is in the E form, the macrocycle can move randomly along the full length of the axle, whereas when the stilbene unit is in the Z form, the ring is trapped in one of the two compartments. Therefore, the stilbene unit plays the role of a photoswitchable gate for the ring movement between the two stations. With the stilbene unit is in the E form (i.e., gate open) an equilibrium distribution of the ring between the two stations of 65(DBA+):35(MBA+) is established (Fig. 6a).

Fig. 6
figure 6

Scheme for operation of a molecular information ratchet based on the photoisomerizable bistable rotaxane 4 2+. Dashed arrows indicate processes that are unlikely to occur

Full size image

To drive the ring distribution away from equilibrium, the gate should be closed for most of the time, and opened preferentially when the macrocycle occupies a specific position (in this case the DBA+ station). The first requirement is obtained by adding to the solution a suitable photosensitizer (benzil) which leads to a photostationary state rich in the cis form of α-methylstilbene (82:18 Z:E under the conditions employed) by intermolecular triplet sensitization. The second requirement is accomplished by appending another photosensitizer (benzophenone) to the macrocycle, capable of causing the Z–E photoisomerization of the stilbene gate by intramolecular triplet sensitization. Benzophenone was chosen because it leads to a photostationary state more rich in the E form of α-methylstilbene (55:45 Z:E) compared to benzil. A key feature of the system is that the DBA+ station is very close to the stilbene gate, whereas the MBA+ station is relatively distant from the gate. Therefore, it can be expected that intramolecular (benzophenone) sensitization (i.e., gate opening) is more efficient when the macrocycle is in the DBA+ compartment, whereas the efficiency of intermolecular (benzil) sensitization should be independent on the location of the macrocycle. Conditions are chosen so that the benzophenone-sensitized isomerization dominates (gate open) when the ring is in the DBA+ compartment – that is, held near to the gate (Fig. 6b) – whereas the benzil-sensitized reaction dominates (gate closed) when the ring is in the MBA+ compartment – that is, held far from the gate (Fig. 6c).

The system starts with the stilbene gate open (E) and an equilibrium ring distribution of 65(DBA+):35(MBA+). Irradiation in the presence of a suitable amount of benzil leads the system to an 80:20 Z:E photostationary state. Under this condition, the ring distribution becomes 45(DBA+):55(MBA+); that is, about one-third of the macrocycles which occupied the more energetically favorable DBA+ compartment at equilibrium have been moved to the less favorable MBA+ compartment. Ultimately, the different photoreactivity of the various interconverting isomers of 4 2+ (Fig. 6) leads to a ring distribution between the two compartments under light irradiation which is different from that observed at the equilibrium in the dark.

It should be pointed out that in this system photons are not used to modify the binding energy between the ring and either station, but to power an information transfer process, as schematically indicated in Fig. 7 [6]. In other words, driving the ring distribution away from its equilibrium value is ensured only by the fact that the macrocycle “signals” its position to the gate, which opens (or closes) accordingly. The similarity of these processes with the hypothetical task performed without an energy input by a “demon” in Maxwell’s famous thought experiment has been extensively discussed [6].

Fig. 7
figure 7

Schematic representation of an information ratchet mechanism for the directional transport of a Brownian particle along a potential energy surface. If the particle signals its position in a distance-dependent manner, then only the barrier closer to the particle may be lowered

Full size image

4.4 Light-Induced Memory Effects in a Bistable Molecular Shuttle

Recently, the bistable rotaxane 5 4+ in which orthogonal chemical and optical stimuli are used to gain full control on the thermodynamics (i.e., the distribution of the rings between the two sites located along the axle) and the kinetics (i.e., the translation rate of the rings between the sites) of molecular shuttling was described (Fig. 8) [75]. A system of this kind is interesting not only from the view point of molecular machines but also from that of signal processing and storage. As a matter of fact, controllable molecular shuttles can be considered as bistable mechanical switches at the nanoscale [76]. While the operation of bistable molecular switches is based on classical switching processes between thermodynamically stable states, the development of molecular memories – which rely on a sequential logic behavior [77] – also requires a control of the rates of the mechanical movement between such states.

Fig. 8
figure 8

Structural formula and schematic representation of rotaxane 5 4+

Full size image

The functional units incorporated in the multicomponent rotaxane 5 4+ (Fig. 8) are: (1) the π-electron-deficient CBPQT4+ ring; (2) the π-electron donor recognition sites of the dumbbell component, constituted by a tetrathiafulvalene (TTF) unit and a 1,5-dioxynaphthalene (DNP) unit; and (3) a photoactive 3,5,3′,5′-tetramethylazobenzene (TMeAB) moiety, located in between the TTF and DNP units, which can be reversibly and efficiently switched between its E and Z configurations by photochemical stimuli. Since the TTF unit is more π-electron rich than the DNP one, the CBPQT4+ macrocycle prefers to encircle the TTF unit rather than the DNP one in the starting co-conformation of 5 4+ (Fig. 9). This preference is evidenced by the presence of a charge-transfer (CT) band peaking at 850 nm in its absorption spectrum (Fig. 10). Upon chemical or electrochemical oxidation of the TTF unit to its radical cation (TTF•+) form, signaled by the appearance of the TTF•+ absorption features in the 400–650 nm region (Fig. 10), CBPQT4+ shuttles to the DNP recognition site on account of the electrostatic repulsion caused by the TTF•+ radical cation and the loss of π-donor–acceptor interactions with CBPQT4+. Such a process can be monitored by the disappearance of both the CT band at 850 nm and the sharp absorption features at around 320 nm, typical of the DNP site not surrounded by CBPQT4+ (Fig. 10).

Fig. 9
figure 9

Chemically and photochemically triggered memory switching cycle of rotaxane 5 4+

Full size image
Fig. 10
figure 10

Absorption spectrum of a 19 μM solution of E-5 4+ (CH3CN, 295 K) before (solid curve) and after (dashed curve) oxidation with 1 equiv. of Fe(ClO4)3 to obtain E-5 5+

Full size image

Steady-state and time-resolved UV-visible spectroscopic experiments showed that, upon quick chemical reduction of the TTF•+ unit to its neutral state, the CBPQT4+ ring immediately shuttles back to encircle the TTF site if the TMeAB unit is in the E configuration, whereas it remains trapped on the DNP site if the TMeAB unit has been photoisomerized to the Z isomer prior to the TTF•+→TTF back reduction (Fig. 9). Indeed, the first-order rate constant for replacement of the CBPQT4+ ring onto the regenerated TTF site in the photoisomerized rotaxane, obtained by monitoring the recovery of the CT absorption band at around 850 nm, is in very good agreement with the first-order rate constant corresponding to the thermal Z→E isomerization of the TMeAB unit, measured by observing the recovery of the absorption band of the E-TMeAB unit at 344 nm (Fig. 11). This behavior can be explained considering that the E–Z isomerization of TMeAB brings about a large geometrical change capable of affecting substantially the free-energy barrier for the shuttling of CBPQT4+ along the axle component [78]. The azobenzene unit in its Z configuration indeed poses a much greater steric hindrance to the shuttling of the ring than does an E azobenzene unit.

Fig. 11
figure 11

Time-dependent absorption changes (CH3CN, 295 K), monitored at (a) 344 nm (E-TMeAB absorption) and (b) 842 nm (TTF-CBPQT4+ CT absorption), showing the regeneration of E-5 4+ from the metastable state Z-5 4+. The lines represent the data fitting according to a first-order kinetic equation. Adapted with permission from [75]

Full size image

In summary, the switching cycle of rotaxane 5 4+ (Fig. 9) consists of the following steps: (1) oxidation of TTF, causing the shuttling of CBPQT4+ from the TTF•+ to the DNP site; (2) UV light irradiation, converting the TMeAB unit from the E to the Z configuration (gate closed); (3) back reduction, regenerating the neutral TTF unit with the CBPQT4+ still residing on the DNP unit, and (4) successive photochemical or thermal Z→E back isomerization, opening the gate and enabling the replacement of CBPQT4+ onto the TTF primary recognition site.

In other words, in a “write-lock-erase” experiment based on the cycle shown in Fig. 9, the data is written on the rotaxane by an oxidation stimulus, and locked by UV light irradiation; after the writing session, the oxidized species can be reduced back to the original form without losing the written data for a remarkably longer time compared to thermodynamically controlled molecular switches. Indeed, the data remain stored for a few hours in the dark at room temperature until the thermal opening of the azobenzene gate occurs. Therefore, 5 4+ operates as a bistable memory element under light-triggered kinetic control. It is also important to note that light irradiation not only locks the data previously recorded by oxidation but also protects the nonoxidized rotaxanes from accidental writing. These properties have positive implications for the use of such molecules in engineered test devices [79, 80].

Another elegant strategy to affect the shuttling kinetics by using light makes use of an intramolecular photocycloaddition reaction and its thermal reversion to shrink the macrocyclic component of a rotaxane [81].

5 Molecular Threading/Dethreading with Directional Control

The examples discussed in the previous sections demonstrate that the principles and methods of supramolecular chemistry applied to the construction of working devices and molecular machines represent a powerful strategy for the development of nanoscience and nanotechnology as well as for the comprehension of the several biological processes in which natural motors and machines operate [8, 9, 82].

Pseudorotaxanes, supramolecular complexes minimally composed of a ring-type host surrounding an axle-type guest [45], are the simplest prototypes of artificial molecular level machines [5]. As already mentioned, their working mechanism is based on the assembly/disassembly of the axle and ring components and it resembles the threading/dethreading of a needle. Studies on switchable pseudorotaxanes, in which these movements can be controlled by external stimulation [5, 46], are of great interest not only for the development of functional materials [83] but also for the realization of molecular machines based on rotaxanes, catenanes, and related interlocked compounds. Specifically, the development of a pseudorotaxane motif capable of performing unidirectional threading and dethreading processes [84–86] under control of external stimuli (Fig. 12a) would be important for the construction of processive linear motors based on rotaxanes (Fig. 12b) and, at least as a perspective, rotary motors based on catenanes (Fig. 12c) [6, 13].

Fig. 12
figure 12

Representation of unidirectional threading/dethreading of a [2]pseudorotaxane with nonsymmetric components (a), a processive linear motor based on a [2]rotaxane (b), and a rotary motor based on a [2]catenane (c). Reprinted with permission from [102]

Full size image

The essential feature of molecular motors is indeed the directional control of the motion, which is achieved by modulating not only the thermodynamics but also the kinetics of the transition between the mechanical states of the device. This result can be achieved by applying ratcheting concepts to the design of the systems [6, 13, 87]. A few examples of artificial molecular rotary motors and DNA-based linear motors [44, 88, 89] have been described, and fully synthetic linear motor molecules are available (see chapters by Leigh and Willner in this volume) [90]. Such systems, however, are based on sophisticated chemical species and/or their operation relies on a complex sequence of chemical reactions. Therefore, the development of concepts and structures for the construction of linear supramolecular motors characterized by simple, efficient and reversible operation is still an open problem and an important challenge, because linear movements are essential both in Nature and technology.

5.1 Photocontrolled Unidirectional Transit of a Molecular Axle Through a Macrocycle

As discussed above, the control of the relative direction of threading and dethreading in pseudorotaxanes cannot be achieved simply by means of classical switching processes between thermodynamically stable states (disassembled components/assembled complex) [13]. The rates of the transition processes between such states also have to be controlled in order to make the desired processes faster than the undesired ones.

A first step forward towards this goal is the ability to adjust the threading/dethreading kinetics by modulating the corresponding energy barriers through external stimulation. A system that can be reversibly photoswitched between thermodynamically stable (pseudorotaxane) and kinetically inert (rotaxane) states is based on the molecular components 6 + and 7 shown in Fig. 13 [91]. The molecular axle 6 + comprises a secondary ammonium center as a hydrogen-bonding donor and two photoswitchable azobenzene end groups. In organic solution, 6 + and the dibenzo[24]crown-8 ring 7 self-assemble to yield a pseudorotaxane complex (Fig. 14), on account of hydrogen bonding between the secondary ammonium center of the axle and the oxygen atoms of the crown ether, with possible contributions from π-stacking interactions between the aromatic moieties of 7 and the azobenzene units. Remarkably, the E→Z photoisomerization of the azobenzenes at both ends of axle 6 + slows down the threading/dethreading of ring 7 by at least four orders of magnitude. The threading time constants under the conditions adopted (room temperature, 5 mM acetonitrile solutions) are >2.5 s and 20 h for E,E-6 + and Z,Z-6 +, respectively. Another interesting feature of this system is that the thermodynamic stability of the complex is also affected by light: the association constant of 6 + with 7 drops by a factor of two on going from the E,E to the Z,Z isomer of the axle. Moreover, the host–guest recognition can be switched off by deprotonating the ammonium center of 6 + with a base, i.e., using a stimulus orthogonal to that employed for switching the azobenzene end units [91].

Fig. 13
figure 13

Structure formulas and schematic representation of the examined axle 6 + and ring 7

Full size image
Fig. 14
figure 14

Operation scheme of a self-assembling system that can be reversibly photoswitched between thermodynamically stable (pseudorotaxane) and kinetically inert (rotaxane) forms

Full size image

Other rotaxane-type systems in which the threading/dethreading kinetics can be photocontrolled according to strategies similar to that described above have been reported in the recent literature [92, 93].

Building upon the results described above, the operation of a simple supramolecular assembly in which a molecular ring translates unidirectionally along a molecular axle in response to photochemical and chemical stimulation has recently been reported [94]. As discussed in Sect. 3, it should be remembered that in solution only the movements of the ring and axle components relative to one another can be considered; therefore, an equally valid view of the device operation would be to consider that the molecular axle passes through the cavity of the macrocyclic component. In fact, such a convention will be used to describe the systems presented in this and the next sections. Besides the implications for the realization of molecular linear motors based on rotaxanes and rotary motors based on catenanes (Fig. 12), a system of this kind constitutes a first step towards the construction of an artificial molecular pump.

The device (Fig. 15) is based on a non-symmetric axle molecule comprising three different functional units: (1) a passive pseudo-stopper (D), (2) a central ammonium recognition site (S) for the ring, and (3) a photoswitchable azobenzene unit (P) at the other end. As for the system described in Fig. 14, the ring-axle recognition relies primarily on hydrogen-bonding interactions involving the ammonium center of the axle and the crown ether oxygen atoms.

Fig. 15
figure 15

Strategy for the photoinduced unidirectional transit of a molecular ring along a non-symmetric molecular axle. Simplified potential energy curves (free energy vs ring-axle distance) for the states shown, describing the operation of the system in terms of a flashing ratchet mechanism, are also reported

Full size image

The strategy at the basis of the operation of this ensemble is shown schematically in Fig. 15. In acetonitrile solution, for kinetic reasons, the axle threads the ring exclusively from the side of the photoactive gate in its starting E configuration (Fig. 15a), affording a pseudorotaxane in which the molecular ring encircles the recognition site S. Light irradiation converts the E–P end group into the angular Z form, a process which is also expected to destabilize the supramolecular complex (Fig. 15b) [91]. Therefore, the dethreading of a fraction of the axle molecules in the pseudorotaxane population is expected, which occurs by exit of the axles from the side carrying the D moiety (Fig. 15c). The system is brought back to its initial state by photochemical or thermal conversion of the Z–P gate back to the E configuration (Fig. 15d). Overall, the photoinduced directionally controlled transit of the axle through the ring would be obtained according to a flashing energy ratchet mechanism [6, 72].

There are two basic requirements for this strategy to work: (1) the kinetic barriers for the slippage of the ring through the axle end groups should follow the ΔG ‡(E–P) < ΔG ‡(D) < ΔG ‡(Z–P) order, and (2) the ring should form a more stable pseudorotaxane when the axle has the photoswitchable end group in its E configuration compared to the Z one. It is also important that the differences in the kinetic and stability constants are sufficiently large, and that the photochemical interconversion of the P gate between its E and Z forms is fast, efficient, and reversible.

It is clear that the choice of the pseudo-stopper unit D is crucial for the successful operation of the mechanism shown in Fig. 15. It was previously reported [95] that the bis(cyclopentylmethyl)ammonium ion 8 + (Fig. 13) is complexed by 7 to form a pseudorotaxane, with threading and dethreading rate constants that fall nicely in between those observed for E,E-6 + and Z,Z-6 + with the same ring [91]. Hence, the strategy shown in Fig. 14 can be implemented with the non-symmetric axle E-9 +, derived from the symmetric guests E,E-6 + and 8 + (Fig. 13).

1H NMR spectroscopic titration experiments showed that in acetonitrile E-9 + pierces ring 7 exclusively from the E-azobenzene terminus. Irradiation of E-9 + with UV light affords Z-9 + in an almost quantitative way. The increased hindrance of the azobenzene end group upon photoisomerization forces Z-9 + to thread 7 through its cyclopentyl terminus. It should be noted that the E→Z photoisomerization of the azobenzene end group of 9 + also takes place efficiently when it is surrounded by 7. Therefore kinetic control of the threading/dethreading side of 9 + can be achieved by photoadjusting the steric hindrance of its azobenzene end group.

In contrast with the results found for the [E,E-6⊂7]+ and [Z,Z-6⊂7]+ pseudorotaxanes [91], the stability constants of [E-9⊂7]+ and [Z-9⊂7]+ are identical within errors. Therefore, in acetonitrile the dethreading of Z-9 + from the ring cannot be caused by the same photochemical stimulus that triggers the azobenzene E→Z isomerization. Because deprotonation of the ammonium recognition site of Z-9 + with a base causes the fast dethreading from 7, thereby neutralizing the stoppering ability of the Z-azobenzene unit, K+ ions were used as competitive guests for 7 [96] in order to promote the disassembly of the complexes. The addition of 2 equiv. of KPF6 causes the complete dethreading of both [E-9⊂7]+ and [Z-9⊂7]+; however, while at room temperature the K+-induced disassembly of the former complex is immediate, the latter one exhibits a dethreading half-life of 51 min (Fig. 16). Considering that the K+-induced dethreading of [Z,Z-6⊂7]+ occurs with a half-life of ca. 40 h (Fig. 16) [91], these observations indicate that the chemically induced disassembly of Z-9 + from 7 takes place exclusively by slippage of the ring through the cyclopentyl end of the axle.

Fig. 16
figure 16

Time-dependent concentration changes, obtained from 1H NMR data in CD3CN at 298 K, showing the K+-induced dethreading of (a) [Z-9⊂7]+ and (b) [Z,Z-6⊂7]+. Conditions: (a) 5.1 mM Z-9H+, 5.4 mM 7 (about 40% complexation of the axle molecules), 15.2 mM KPF6; (b) 4.8 mM Z,Z-6H+, 7.6 mM 7 (about 65% complexation of the axle molecules), 16.7 mM KPF6. Reprinted with permission from [94]

Full size image

The results of an experiment that illustrates the directional transit of the ring along the axle are summarized in Fig. 17: (1) E-9 + pierces 7 with its E-azobenzene side to form the pseudorotaxane [E-9⊂7]+; (2) irradiation in the near UV region converts quantitatively [E-9⊂7]+ into [Z-9⊂7]+, characterized by slow assembly-disassembly kinetics; (3) the successive addition of K+ ions promotes the dethreading of Z-9 + from 7 by the passage of the cyclopentyl moiety through the cavity of the ring. It should be noted that equilibration of the [Z-9⊂7]+ complex with its separated components, that would cause the loss of the information on the threading direction of E-9 +, is much slower than the time required for the activation of the dethreading stimulus (addition of K+). Therefore, after the threading event the system is “locked” by photoisomerization, and the successive addition of potassium ions causes dethreading in the same direction along which threading of E-9 + has initially occurred. The starting species E-9 + can be fully regenerated by thermal Z→E back isomerization, and sequestration of K+ by an excess of [18]crown-6 affords the re-assembly of [E-9⊂7]+ and the full reset of the system.

Fig. 17
figure 17

Representation of the photochemically and chemically controlled transit of 7 along 9 +

Full size image

This supramolecular system, however, if it were to be incorporated in a compartmentalized structure (e.g., embedded in the membrane of a vesicle), could not be used to “pump” the molecular axle and generate a transmembrane chemical potential because the ring component has two identical faces. Despite this deficiency the described system is characterized by a minimalist design, facile synthesis, convenient switching, and reversibility: all these features constitute essential requirements for real world applications.

In the next section a strategy similar to that just described is applied to supramolecular assemblies based on three-dimensional non-symmetric macrocycles whose lengths can approach the thickness of a bilayer membrane, [97] and in which face-selective threading can be realized.

5.2 Solvent- and Light-Controlled Unidirectional Transit of a Non-symmetric Molecular Axle through a Non-symmetric Three-Dimensional Macrocycle

An increase in the structural complexity of the self-assembled pseudorotaxane/rotaxane systems can be obtained by using as a ring component the macrocycle tris(phenylureido)calix[6]arene [98] derivative 10 (Fig. 18). Because of the non-symmetric nature of this “wheel” it is possible to thread selectively suitable axles from the “upper” or “lower” rim of the macrocycle leading to “up” and “down” oriented isomers (Fig. 19). It is known [99] that, in apolar media, macrocycle 10 is able to be threaded exclusively from the upper rim by axles derived from 4,4′-bipyridinium salts [100]. This behavior can be explained by the peculiar chemical and structural features of compound 10 as a host, which are (1) a π-donor cavity that, because of its width, can include the positively charged bipyridinium unit of the axle, but not together with its counter anions, (2) three efficient hydrogen-bonding donor ureidic groups at the upper rim that, by complexing the counter anions of the axle, can assist the insertion of the cationic portion of the latter, and (3) three methoxy groups at the lower rim that, in apolar media, are oriented towards the interior of the cavity [101], thereby hindering the access of the guest from this direction. The use of more polar solvents has a profound effect on these interactions. In fact, the solvent polarity, by changing the extent of ion pairing of the axle and decreasing the pivoting role of the three ureidic groups of the host, affects both the concentration of the active guest available in solution and the binding ability of the macrocycle.

Fig. 18
figure 18

Structural formulas of the examined ring and axle components

Full size image
Fig. 19
figure 19

Formation of “up” and “down” oriented pseudorotaxane isomers by self-assembly of the nonsymmetric ring and axle components

Full size image

To obtain pseudorotaxane systems based on ring 10, capable of undergoing relative unidirectional threading/dethreading motion, the axles 11 2+ and 12 2+ (Fig. 18) were employed [102]. They are composed of a central electron-acceptor 4,4′-bipyridinium unit functionalized with a hexanol chain on one side and either a stilbene (11 2+) or a tBu-substituted stilbene (12 2+) on the other. The terminal OH group has been selected because it can be involved in stoppering reactions, while the stilbene and the tBu-substituted stilbene head groups have been employed because they are not too bulky to prevent their slippage [103, 104] through 10, but large enough to enable a kinetic control of the threading/dethreading motions of the ring. NMR spectra of the equilibrium mixture obtained after mixing 10 with an excess of axle 11 2+ or 12 2+ in C6D6 at room temperature prove that only one pseudorotaxane type complex forms predominantly, confirming that both axles enter the calixarene cavity through the calixarene upper rim with their OH terminus to yield the oriented species [11⊂10]2+ up and [12⊂10]2+ up (Fig. 20). This orientational control during the formation of the pseudorotaxane indicates that the macrocycle promotes the threading of the axles from the upper rim and that the latter components access the macrocycle through the less bulkier OH terminus in a process that is kinetically controlled by the different size of the end groups of the axles. When the OH end group of pseudorotaxanes [11⊂10]2+ up and [12⊂10]2+ up is replaced by bulky diphenylacetyl moieties, the pseudorotaxanes are converted to semirotaxanes [13⊂10]2+ up and [14⊂10]2+ up (Fig. 20). These species exhibit a more pronounced rotaxane-like behavior than [11⊂10]2+ up and [12⊂10]2+ up because one end of their axle is stoppered by the presence of the bulky diphenylacetyl moiety, while dethreading from the side carrying the stilbene-type unit is greatly slowed down by the steric hindrance of the latter.

Fig. 20
figure 20

Self-assembly of pseudorotaxanes [11⊂10]2+ up and [12⊂10]2+ up, and synthesis of semirotaxanes [13⊂10]2+ up and [14⊂10]2+ up

Full size image

In polar solvents the multiple interactions that stabilize complexes [13⊂10]2+ up and [14⊂10]2+ up are weakened and, therefore, dissolution of these species in such solvents induces the dethreading of the axle. Because of the presence of the diphenylacetyl stopper at one end of the axles, the dethreading occurs through the slippage of the sufficiently slim stilbene-type unit from the calixarene lower rim. The dethreading rate in DMSO for semirotaxane [13⊂10]2+ up (Fig. 21a) is two orders of magnitude lower than that of [14⊂10]2+ up, owing to the higher hampering effect of the tBu-substituted stilbene present in axle 14 2+ compared with the unsubstituted axle 13 2+.

Fig. 21
figure 21

Time-dependent absorption changes at 460 nm observed upon dissolution of [13⊂10]2+ up in DMSO at room temperature, before (a, full circles) and after (b, open circles) exhaustive irradiation at 334 nm, causing the E→Z isomerization of the stilbene end group. The first-order fitting curves are also shown

Full size image

In these semirotaxanes the dethreading rate can also be photocontrolled upon UV light-induced E–Z isomerization of the stilbene end group of the axles. For both [13⊂10]2+ up and [14⊂10]2+ up at the photostationary state about 70% of the stilbene units are converted from the E to the Z isomer. In the case of [13⊂10]2+ up the higher hampering effect of the Z isomer compared with the E one results in a much more difficult slippage of this unit through the lower rim of the macrocycle once it is dissolved in a polar solvent (Fig. 21b). It is interesting to notice that the E–Z photoisomerization affects the dethreading rate constant to a greater extent than does the incorporation of the tBu group on the stilbene unit. The rate constant observed upon isomerization is indeed about one order of magnitude slower than that observed for the E-isomer of the substituted stilbene.

Remarkably, in the case of [14⊂10]2+ up the photoisomerized compound does not undergo dethreading in polar solvents at all. The Z isomer of the tBu-stilbene is too bulky to pass through the calixarene lower rim, and [14⊂10]2+ up in its Z configuration exhibits a genuine rotaxane behavior.

According to the transformation sequence described above, and as schematized in Fig. 22, the unidirectional transit of the non-symmetric molecular axles 11 2+ and 12 2+ through the non-symmetric macrocycle 10 (or the translation of the ring along the axles) is achieved. The strategy is based on the use of appropriately designed molecular components, an essential feature of which is their nonsymmetric structure, and exploits the following steps: (1) in apolar solvents axles 11 2+ or 12 2+ thread calixarene 10 from its upper rim, leading to an oriented pseudorotaxane structure in which the OH group is positioned at the lower rim of the ring; this threading mode is favored because of the small hampering effect of the OH group being substantially lower than that of the stilbene moiety (Fig. 22a); (2) by a stoppering reaction that introduces a bulky diphenylacetyl moiety, the pseudorotaxane is converted into a semirotaxane-like species (Fig. 22b); and (3) replacement of the apolar solvent with a polar one weakens the interactions that stabilize the assembled structure and induces the axle dethreading from the calixarene lower rim (Fig. 22c), that is, in the same direction of the axle threading.

Fig. 22
figure 22

Simplified potential energy curves representing the relative unidirectional transit of the axle through the ring. The horizontal coordinate of the diagrams represents the axle–ring distance when they approach one another along the direction and with the orientation shown in the drawings. (a) Threading of the axle through the macrocycle upper rim in apolar solvents. (b) The stoppering reaction that converts the pseudorotaxanes into rotaxane-like species. (c) Dethreading of the axle from the macrocycle lower rim in polar solvents

Full size image

It is also important to stress the essential role played by the stilbene unit incorporated at one end of the axle. It enables one: (1) to achieve the unidirectional transit of the axle through the macrocycle, because its dimensions are not too large to prevent slippage, but big enough to induce a kinetic control of the axle threading/dethreading processes; and (2) to tune the dethreading rate because of the possibility of modifying its hindering effect upon the use of stilbene unit substituted with relatively bulky groups, or, more interesting, upon photoisomerization.

6 Conclusions

One of the most interesting aspects of supramolecular (multicomponent) systems is their interaction with light. The systems described in this chapter show that, in the frame of research on supramolecular photochemistry, the design and construction of nanoscale machines capable of performing useful light-induced functions can indeed be pursued. Their potential applications are various – from energy conversion to sensoring and catalysis – and, to a large extent, still unpredictable. As research in the area is progressing, non-conventional functionalities that could be enabled by these systems are emerging. For example, (1) their behavior can be exploited for processing information at the molecular level, and (2) their mechanical features can be utilized for transportation of nanoobjects, mechanical gating of molecular-level channels, and nanorobotics.

It should be noted, however, that the species described here, as most multicomponent systems developed so far, operate in solution, that is, in an incoherent fashion and without control of spatial positioning. Although the solution studies are of fundamental importance to understanding their operation mechanisms and for several uses (e.g., intelligent sensoring, catalysis, drug delivery), it seems reasonable that before such systems can find applications in some fields of technology, they have to be interfaced with the macroscopic world by ordering them in some way. The next generation of multicomponent molecular species will need to be organized so that they can behave coherently and can be addressed in space. Viable possibilities include deposition on surfaces, incorporation into polymers, organization at interfaces, or immobilization into membranes or porous materials. Recent achievements in this direction [18, 69, 70, 105–108] let one be optimistic that useful devices based on functional supramolecular systems could be obtained in a not too distant future.

Apart from foreseeable applications related to the development of nanotechnology, investigations on photochemical molecular machines are important to increase the basic understanding of photoinduced reactions and other important processes such as self-assembly, as well as to develop reliable theoretical models, for example to interpret energy- and electron-transfer processes in complex systems.

It is also important to point out that machines capable of performing useful light-induced functions are of the highest importance in current times. It has become clear that products and services in the years ahead, including those of a nanotechnology-based industry, will have to exploit solar energy [109].

Finally, this research also has the merit of stimulating the ingenuity of chemists, thereby instilling new life into chemistry as a scientific discipline.