SPECIAL FEATURE: PERSPECTIVE

Molecular rotary motors: Unidirectional motionaround double bondsDiederik Rokea, Sander J. Wezenberga,1, and Ben L. Feringaa,1

Edited by J. Fraser Stoddart, Northwestern University, Evanston, IL, and approved April 6, 2018 (received for review January 17, 2018)

The field of synthetic molecular machines has quickly evolved in recent years, growing from a fundamentalcuriosity to a highly active field of chemistry. Many different applications are being explored in areas suchas catalysis, self-assembled and nanostructured materials, and molecular electronics. Rotary molecularmotors hold great promise for achieving dynamic control of molecular functions as well as for poweringnanoscale devices. However, for these motors to reach their full potential, many challenges still need to beaddressed. In this paper we focus on the design principles of rotary motors featuring a double-bond axleand discuss the major challenges that are ahead of us. Although great progress has been made, furtherdesign improvements, for example in terms of efficiency, energy input, and environmental adaptability,will be crucial to fully exploit the opportunities that these rotary motors offer.

molecular motor | alkene |molecular machine

Control of motion at the molecular scale has intriguedchemists for a very long time. The quest for overcomingrandom thermal (Brownian) motion has culminated inthe emergence of synthetic molecular machines (1–7),including motors (8–12), muscles (13), shuttles (14),elevators (15), walkers (16), pumps (17–19), and assem-blers (20). By taking inspiration from the fascinatingdynamic and motor functions observed in biologicalsystems (e.g., ATPase and bacterial flagella) (21), thefield of synthetic molecular machines has evolved rap-idly in recent years. This is due to major advances insupramolecular chemistry and nanoscience, the emer-gence of the mechanical bond (7), and the develop-ment of dynamic molecular systems (22). A variety ofpotential applications is now being considered inareas ranging from catalysis (23) and self-assembly(24) to molecular electronics (25, 26) and responsivematerials (27, 28). Furthermore, translation of motionfrom the molecular scale to the macroscopic scale al-lows for dynamically changing material propertiesand the movement of larger objects. Cooperativityand amplification across several length scales can beachieved, for example, by incorporating molecularmachines in gels (29, 30), liquid crystals (27, 31), andpolymers (32) or by anchoring them to surfaces (33,34), allowing the control of a variety of properties in-cluding surface wettability (33, 34), contraction or

expansion of gels (29), and actuation of nanofibers inresponse to their environment (35). This leap fromstatic to dynamic materials clearly demonstrates thepotential of molecular machines. Although much ef-fort has already been devoted to the development ofmolecular motors and machines as well as the eluci-dation of their operational mode, a great deal of de-sign improvements are needed to fully exploit thepotential in practical applications. Ideally, molecularmotors can operate with high efficiency and durableenergy input (fuel), can be easily adapted to a specificenvironment or application, are compatible with spe-cific functions, and can be synchronized and act ina cooperative manner.

Where the pioneering work of Sauvage, Stoddart,and others successfully led to stimuli-controlled trans-lational and rotary motion in mechanically interlockedsystems (1, 36–39), the induction of unidirectional ro-tary motion posed a major challenge. Distinct ap-proaches, including those based on catenanes (39),surface confined systems (40), and aryl–aryl single-bond rotation (8, 41, 42), have been taken over thelast two decades to develop molecular motors capa-ble of such unidirectional rotation when supplied byenergy in the form of light, chemical stimuli, or elec-trons (11). In this paper, we focus on rotary motors thatcontain a double-bond axle (Fig. 1). Although it may

aStratingh Institute for Chemistry, University of Groningen, 9747 AG Groningen, The NetherlandsAuthor contributions: D.R., S.J.W., and B.L.F. wrote the paper.The authors declare no conflict of interest.This article is a PNAS Direct Submission.Published under the PNAS license.1To whom correspondence may be addressed. Email: s.j.wezenberg@rug.nl or b.l.feringa@rug.nl.Published online April 30, 2018.

www.pnas.org/cgi/doi/10.1073/pnas.1712784115 PNAS | September 18, 2018 | vol. 115 | no. 38 | 9423–9431

SPECIA

LFEATURE:

PERSPECTIV

E

Dow

nloa

ded

by g

uest

on

June

25,

202

0

seem unusual to use a double bond as rotary axle since the rota-tion is restricted, stimuli such as light can induce rotation (cf. isom-erization), as is most elegantly seen in the process of vision (43). Assuch, autonomous and repetitive unidirectional rotation has beensuccessfully achieved in multiple systems, all of them driven bylight. Here, we discuss the key design principles of these systemsand furthermore a perspective on key challenges and possiblefuture developments is provided.

Motors Based on Overcrowded AlkenesAt the very basis of overcrowded alkene-based molecular motorsis a photochemical cis–trans isomerization around their centralcarbon–carbon double bond. For stilbene, this process has beenstudied already for more than half a decade (44). Due to thesymmetry of stilbene, there is no directional preference in theisomerization process. It was shown in 1977 that the introductionof steric bulk around the double bond distorts the otherwiseplanar geometry, giving rise to helical chirality (45). This featurewas further exploited to develop a chiroptical switch, in which twopseudoenantiomeric forms with opposing helical chirality couldbe selectively addressed (46). This work formed the basis for thedesign and synthesis of the first molecule capable of undergoingunidirectional 360° rotation around a double bond, which ourgroup reported in 1999 (9). It is based on an overcrowded alkene,with two identical halves on each side of the double bond (therotary axle) [(P,P)-trans-1 in Fig. 2A]. Due to steric interactionsbetween the two halves, in what is referred to as the fjord region,the molecule is twisted out of plane, resulting in a helical shape.The first molecular motor featured two stereogenic methyl sub-stituents which are preferentially in a pseudoaxial orientation dueto steric crowding. These stereocenters dictate the helical chiralityin both halves of the molecule and hence the direction of rotation.A full rotary cycle consists of four distinct steps: two photo-chemical and energetically uphill steps and two thermally acti-vated and energetically downhill steps.

Starting from (P,P)-trans-1 (Fig. 2A), irradiation with UV light(>280 nm) induces a trans-to-cis isomerization around the doublebond leading to isomer (M,M)-cis-1 with opposite helical chirality.This photoisomerization is reversible and under continuous irra-diation a photostationary state (PSS) is observed in which for thisparticular case the ratio of cis to trans is 95:5. In (M,M )-cis-1 themethyl substituents end up in an energetically less favorablepseudoequatorial position. To release the built-up strain, a ther-mally activated process occurs, in which both halves slide along-side each other, inverting the helicity from left- (M,M ) to right-handed (P,P) and allowing the methyl substituents to readoptthe energetically favored pseudoaxial orientation. The photogenerated

states, which are prone to such a thermal helix inversion (THI)process, have often been referred to as “unstable” or “meta-stable” states. The THI is energetically downhill and effectivelywithdraws the higher-energy isomer such as (M,M )-cis-1 from thephotoequilibrium mixture and hence completes the unidirec-tional 180° rotary motion. The second part of the cycle proceedsin a similar fashion as a second photoisomerization step affords(M,M )-trans-1 (PSS trans to cis ratio of 90:10) with the methylsubstituents again in the pseudoequatorial position. A secondTHI then reforms (P,P)-trans-1 and a full 360° rotation cycle iscompleted.

The so-called second-generation light-driven molecular rotarymotors, consisting of distinct upper (rotor) and lower (stator)halves and bearing only a single stereogenic center (Fig. 2C), werepresented shortly after (47). Analogous to the first-generationmotors, 360° rotary motion can be achieved by a sequence ofphotochemical and thermal steps. The design, with nonidenticalhalves, allowed for a much broader scope of functionalization, inparticular for surface anchoring through the stator, and paved theway for many different applications (12), as shown in Fig. 3.

Fig. 1. Schematic representation of unidirectional rotary motionaround a double-bond axle.

Fig. 2. (A) Rotary cycle of a first-generation motor based onovercrowded alkene. (B) Top view of the rotary cycle. (C) Structuresof second- and third-generation molecular motors.

9424 | www.pnas.org/cgi/doi/10.1073/pnas.1712784115 Roke et al.

Dow

nloa

ded

by g

uest

on

June

25,

202

0

The development of second-generation motors revealed thatthe presence of only one stereogenic center is sufficient to induceunidirectional rotation. This raised the question of whether uni-directional rotation can be achieved in the absence of anystereocenters (48). To address this question, symmetrical motorswere synthesized bearing two rotor units. These third-generationmotors only have a pseudoasymmetric center and still unidirec-tional rotary motion around both axles was found to occur.

Since our first reports on molecular rotary motors based onovercrowded alkene, great effort has been dedicated to the un-derstanding of their functioning, especially the key parametersthat govern the isomerization processes, and the use of new in-sights to adapt the structural design (11). This has resulted in alarge collection of overcrowded alkene-based motors with dif-ferent properties, which have been applied successfully to inducemotion at the molecular scale as well as the nano scale and macroscale (27, 34, 49). The main aspects that have been investigatedand that will be discussed in the next sections are the speedof rotation, the excitation wavelength, and the efficiency of themotors.

Rotational Speed Adjustment. For every different application ofmolecular motors, for example in soft materials or biological sys-tems, often a distinct frequency of rotation is desired. The pho-tochemical steps proceed on a timescale of picoseconds, makingthe usually much slower THI the rate-limiting step (50). Consid-erable research effort in our group has been devoted to fullyunderstanding the THI and to altering the rotational speed bystructural modifications. Throughout the years, density functionaltheory (DFT) calculations have proven to be useful in predictingthermal barriers before the synthesis of new motors and in inter-preting experimental results (51, 52). Although many factors mayinfluence the rate of the THI, steric interactions within the mole-cule play a dominant role. Generally, two approaches have beentaken to speed up the THI (Fig. 4): (i) by lowering the thermalbarrier through a decrease in the steric hindrance in the fjord re-gion or (ii) by raising the energy of the unstable state relative tothe transition state. Additionally, electronic effects on the barrierof the THI were studied by introducing a strong push–pull systemover the central double bond in a second-generation motor (53).A decrease in the barrier of the THI was observed as well as for thethermal E–Z isomerization. Consequently, upon generation of theunstable state, both a “forward” THI and a competing “back-ward” thermal E–Z isomerization took place in this push–pullsystem, reducing the efficiency of the resulting motor.

For first-generation motors, the THI processes for unstable cisand trans isomers are different and therefore have different rates(9). Modifications to the design of the molecular motor can havecomplex and sometimes opposite effects on these rates. For ex-ample, the introduction of more steric bulk at the stereogeniccenter, by replacing the methyl substituent in motor 4 with anisopropyl group, accelerates the rate of thermal isomerizationfrom the unstable to the stable cis form but decreases the rategoing from the unstable to the stable trans isomer (Fig. 5) (54). Thelatter process was found to be so slow that an intermediate statecould be observed with mixed helicity, that is (P,M)-trans-5, sug-gesting that the THI is a stepwise process. This kind of behaviorwas already predicted for related overcrowded alkenes that donot have stereogenic methyl substituents, which according tocalculations racemize between (M,M ) and (P,P) helical structuresvia an intermediate structure with (P,M ) helicity (55). This example,however, remains so far the only case in which such a stepwisemechanism for the THI in first-generation motors is observed. Todecrease the amount of steric hindrance in the fjord region, andhence lower the energy barrier for THI, the central six-memberedring was changed to a five-membered ring (Motor 6 in Fig. 5) (56).It has to be noted that upon reducing the ring size from six- to five-membered, conformational flexibility is lost. As a consequence,the unstable states are most likely further destabilized as the steric

Fig. 3. Molecular motors based on overcrowded alkene in differenttypes of applications.

Fig. 4. Approaches for speeding up the THI either through (i) adecrease in steric hindrance in the fjord region or (ii) bydestabilization of the unstable state with respect to thetransition state.

Roke et al. PNAS | September 18, 2018 | vol. 115 | no. 38 | 9425

Dow

nloa

ded

by g

uest

on

June

25,

202

0

hindrance cannot be relieved by folding, which additionally con-tributes to increased rates for the THI.

In an attempt to further destabilize the unstable states, over-crowded alkene 7 was synthesized, which has two tert-butyl in-stead of methyl substituents at the stereogenic center (57).However, these substituents cause too much steric hindranceimpeding the unidirectional motion. Another approach is to re-place the naphthalene moieties with xylene moieties (motor 8)(58). In this design, the xylyl methyl substituents cause the nec-essary steric hindrance in the fjord region. X-ray analysis showsthat these methyl substituents are more sterically demanding inthe trans isomer, forcing the molecule in a more strained confor-mation, also leading to destabilization of the unstable trans iso-mer. The barrier of the THI from unstable trans to stable trans wasfound to be lower in motor 8 with respect to motor 6. However,the increased steric hindrance in the fjord region causes a higherbarrier for the unstable cis to stable cis isomerization, reflectingthe complex and opposite effect that (often subtle) changes in themolecular design may have on the rates of these two THI processes.

Similar systematic structural modifications have been made tosecond-generation motors to alter their speed of rotation. Initialstudies mainly involvedmotors of the general structure 9 (Fig. 6) inwhich the bridging atoms (X and Y) were varied (47, 51, 52). Half-lives of the unstable states ranging from 233 h [X = S, Y = C(CH3)]to 0.67 h (X = CH2, Y = S) were measured. The conformationallyflexible six-membered rings allow for the molecule to releasesome of the strain around the double bond that is built up in thephotoisomerization step. DFT calculations have shown that, incase of a six-membered ring, the THI is a stepwise process andmultiple transition states have been identified (51, 52).

Also here the change to a five-membered ring in the statormakes the molecule more rigid. Again, this results in an increasedbarrier for the THI in compound 10, up to the point where athermal E–Z isomerization becomes favored over the THI and abistable switch is obtained (59). When only in the upper-half rotora five-membered ring is introduced (motor 11), the rotationalspeed dramatically increases up to the megahertz regime (60).Motor 12, bearing two five-membered rings, however, has a lowerbarrier, in analogy to the first-generation motors, due to a de-crease in the steric hindrance in the fjord region (61). The sterichindrance, and as a consequence the THI barrier, was further re-duced by replacing the naphthalene moiety in the upper half withxylene or benzothiophene (motor 13) (58, 62). Furthermore, largersubstituents have been placed at the stereogenic center to in-crease the rate of the THI (63, 64) and DFT calculations showed

that this decrease is due to destabilization of the unstable state,effectively lowering the barrier for THI.

The speed of the rotary motors is also dependent on the sol-vent (65, 66). In a systematic study of motors with pending arms ofvarying flexibility and length it was established that solvent po-larity plays a minor role, but that in particular enhanced solventviscosity for motor systems with rigid arms decreases THI drasti-cally (67). The results were analyzed in terms of a free volumemodel and it is evident that matrix effects (solution, surface,polymer, and liquid crystal) comprise a challenging multiparam-eter aspect in applying molecular motors. In all these examples,changing the speed of rotation of a molecular motor requires aredesign of the molecule and multistep synthesis. Dynamic controlof the rotational speed would be the next logical step in the de-velopment of molecular motors. Locking the rotation by employingan acid–base-responsive self-complexing pseudorotaxane wasthe first example that allowed for such dynamic control over therotary motion (68). More recently, an allosteric approach wasreported in which metal complexation was used to alter the speedof rotation (69). Complexation of different metals to the statorpart of a second-generation molecular motor caused different de-grees of contraction of the lower half. As a consequence the sterichindrance in the fjord region decreased, which resulted in a lowerbarrier for the THI.

Precise control of the speed of rotation in a dynamic fashionremains challenging but the first examples have shown thatlengthy syntheses can be avoided and motor speed can be al-tered in situ. Controlling speed by external effectuators (metal/ionbinding, pH, and redox) or tuning in response to chemical con-versions (catalysis) or environmental (matrix and surface) constraintsoffer exciting opportunities for more advanced motor functioning.

Shifting the Excitation Wavelength. A major challenge in thefield of photochemical switches and motors is to move away fromthe use of damaging UV light because it limits the practicality insoft materials and biomedical applications (70, 71). For this rea-son, it is important to shift the irradiation wavelengths toward thevisible spectrum (72). The most straightforward method is to makechanges to the electronic properties of the motors in such a waythat the molecule is able to absorb visible light. However, suchchanges may be detrimental to the photoisomerization process.The first successful example of a visible light-driven molecularmotor made use of a push–pull system to red-shift the excitationwavelength (73). This second-generation motor comprised a nitro-acceptor and a dimethylamine-donor substituent in its lower half,which allowed for photoexcitation with 425-nm light. An alternativestrategy that is often used to red-shift the absorption of molecularphotoswitches relies on the extension of the π system. Indeed, uponextension of the aromatic system of the stator half of second-generation motors unidirectional rotation could be induced byirradiation at wavelengths up to 490 nm (74).

Apart from these methods that focus on altering the highestoccupied molecular orbital– lowest unoccupied molecular orbitalgap, alternative strategies based on metal complexes are highlypromising. For example, palladium tetraphenylporphyrin wasused as a triplet sensitizer to drive the excitation of a molecularmotor (75). Isomerization of the motor was shown to occur bytriplet–triplet energy transfer, upon irradiation of the porphyrinwith 530- to-550 nm light. In a related example, a molecular motorwas incorporated as a ligand in a ruthenium(II)–bipyridine com-plex (76). Irradiation with 450 nm into the metal-to-ligand chargetransfer band resulted in unidirectional rotation.

Fig. 5. Structural variations of first-generation molecular motors andeffects on t1/2 of THI process.

9426 | www.pnas.org/cgi/doi/10.1073/pnas.1712784115 Roke et al.

Dow

nloa

ded

by g

uest

on

June

25,

202

0

These examples illustrate that there are multiple viable strat-egies to red-shift the excitation wavelength of molecular motors.However, the change of the wavelength region at which thesemotors can be operated is still modest. Moving further away fromUV light toward red light or even near-IR remains a major chal-lenge. There are several strategies that have shown promisingresults for photoswitches, such as diarylethenes and azobenzenes,that have not been applied to molecular motors yet (72). For ex-ample, multiphoton absorption processes using upconvertingnanoparticles (77) or two-photon fluorophores (78) should beconsidered in future studies.

Improvement of the Photochemical Efficiency. Improving theefficiency of the photochemical isomerization process has provento be difficult as it not as well understood as the thermal isom-erization process. Typically, quantum yields below 2% are ob-served for E/Z photoisomerization of second-generation motors(52, 79). To improve the yield, a detailed understanding of theexcited state dynamics is required. Over the last decade, a com-bination of advanced spectroscopic studies and quantum chem-ical calculations has been used to gain insight in the mechanism ofthe photochemical isomerization. Using time-resolved fluores-cence and picosecond transient absorption spectroscopy, a two-step relaxation pathway was observed after the initial excitationto the Franck–Condon excited state (50, 79, 80). Within 100 fs, abright (i.e., fluorescent) state relaxes to an equilibrium with alower-lying dark (i.e., nonfluorescent) state. Based on femtosec-ond stimulated Raman spectroscopy, supported by quantumchemical calculations, it has been postulated that this process isaccompanied by elongation and weakening of the central double

bond (81–84). Solvent viscosity studies showed that this process isindependent of solvent friction, which is consistent with a volume-conserving structural change (79, 85). The dark excited state,formed after this first relaxation, has a lifetime of ∼1.6 ps and re-laxes back to the ground state to either the stable or unstable formvia conical intersections (CIs). Relaxation to the ground stateleaves excess vibrational energy which is dissipated to the solventat the tens of picoseconds timescale (81). The relatively longlifetime of the dark state is attributed to the fact that a high degreeof twisting and pyramidalization of one of the carbons of thecentral double bond is required to reach the CI (84). Recentstudies showed that this relaxation to the ground state, which isassociated with twisting and pyramidalization, is not dependent onthe size of the substituents (85), while it is dependent on viscosity.This observation suggests that the motion that accompanies therelaxation to the ground state is not necessarily a complete rotationof the halves but rather occurs only at the core of the molecule.

The ability to control CIs could lead to major improvements, asthey play an important role in the efficiency of the photochemicalstep. To improve the efficiency of molecular motors, Filatov andcoworkers investigated the factors influencing the CIs in a theo-retical study (86, 87). The calculations predict that by placingelectron withdrawing groups close to the central axle, such as animinium, the character of the CI changes from a twist-pyramidalizationto a twist-bond length alteration. This effectively changes themode of rotation from a precessional motion for current motors toan axial motion with higher efficiency. These computational de-signs have already inspired the development of new photoswitcheswith increased efficiency (88) but have not yet been applied to

Fig. 6. Structural variations of second-generation molecular motors.

Fig. 7. Proposed rotational cycle for a redox-driven molecular motor.

Roke et al. PNAS | September 18, 2018 | vol. 115 | no. 38 | 9427

Dow

nloa

ded

by g

uest

on

June

25,

202

0

motors and should be taken into account in attempts to increasethe efficiency.

Redox-Driven Motors. As an alternative to the use of (UV-visible)light to power molecular motors we considered designing anelectromotor taking advantage of redox processes. Preliminarystudies toward using overcrowded alkenes as redox-driven mo-lecular motors are promising (89). A rotary cycle is envisioned, inwhich consecutive oxidation/reduction cycles would electro-chemically form the unstable state, which then undergoes a THI toafford the stable state, completing 180° rotation (Fig. 7). Unfor-tunately, the stereogenic center was found to be susceptible todeprotonation, leading to an irreversible double-bond shift inwhich the central axis is converted to a single bond. As this type ofdegradation pathway impedes any successful directional motion,a different design has to be made. Quaternization of the stereo-genic center by replacing the hydrogen for a fluorine atom wouldprevent deprotonation. It was recently shown that such fluorine-substituted molecular motors with quaternary stereocenters arestill capable of unidirectional rotation when irradiated by light,making them excellent candidates to be studied as redox-drivenmolecular motors (90).

Alternative Motor DesignsIn 2006, Lehn (91) proposed a new type of light-driven molecularmotor derived from imines. The design is based on the two typesof E/Z-isomerization processes that imines can undergo, namelya photochemical isomerization and a thermal nitrogen inversion.A two-step rotational cycle was proposed, starting with a light-induced E/Z isomerization, which has an out-of-plane rotationalmechanism (Fig. 8). A thermally activated in-plane nitrogen in-version involves a planar transition state which would convertthe system back to the original state. These two combinedprocesses would lead to a net rotational motion as both followa different pathway. Placing a stereogenic center next to theimine leads to preferential rotation by favoring the direction ofthe photochemical isomerization. This is different from the otherexamples of double-bond motors, as directionality is induced

here in the photochemical step, rather than the thermal isomer-ization step.

Based on these design principles, Greb and Lehn (92) reportedin 2014 on the synthesis of the first rotary motor based on imines(i.e., N-alkyl imine 14) bearing a stereocenter next to the centralimine (Fig. 9). Because of the twisted shape of the lower half and E–Zisomerism of the imine four stereoisomers are formed. A helicityinversion does not occur under the experimental conditions due tothe relatively high barrier for this process compared with the nitro-gen inversion. Under thermodynamic equilibrium, there is a pre-ference for (S,M)-cis over (S,P)-trans, whereas there is not a majorpreference for either (S,P)-cis or (S,M)-trans. Photochemical isom-erization occurs upon irradiation with 254-nm light and at PSS theratio is shifted toward (S,P)-trans relative to (S,M )-cis, while theratio of the other two isomers remains unaffected by irradiation.Heating up the mixture of isomers to 60 °C for 15 h allows for thenitrogen inversion to occur, restoring the original distribution ofdiastereomers. Both processes are equilibrium reactions and there-fore both forward and backward reactions can occur. However, dueto preferred formation of one of the isomers in each step, overall a netrotation occurs.

During their investigations, it was found that annealing abenzene ring to the lower half is essential for this two-step motoras it effectively increases the thermodynamic barrier for the helicityinversion. Interestingly, when this inversion can occur, a molecularmotor with a four-step rotary cycle is obtained, reminiscent of thecycle for motors based on overcrowded alkenes. That is, two pho-tochemical isomerization steps and two thermally activated ringinversions give rise to 360° rotation. To show that imines can beused as two-step molecular motors in a more general sense and toprovide more experimental and theoretical proof for the rotationalbehavior, camphorquinone imines were synthesized in a follow-upstudy (93).

One of the major advantages of imine-based molecular motorsis that many types of imines with different properties can be easilysynthesized through simple condensation reactions starting fromcommercially available materials. Furthermore, fine-tuning of thespeed of these motors can be achieved through controlling thebarrier for N-inversion. The barrier for this process depends

Fig. 9. A two-step molecular rotary motor based on imines.

Fig. 8. Proposed mechanism for an imine-based molecular motor.

9428 | www.pnas.org/cgi/doi/10.1073/pnas.1712784115 Roke et al.

Dow

nloa

ded

by g

uest

on

June

25,

202

0

largely on the substituent at theN-atom, providing a good handleto alter the speed of rotation. The assumption that there is apreferred direction of rotation in the photochemical E/Z isomeri-zation in chiral imines due to the unsymmetrical excited statesurface is very plausible, but further experimental support to un-equivocally prove their preferred direction of rotation is warranted.These photoinduced isomerization processes often occur at thepicosecond timescale, making it very difficult to obtain directevidence. Potentially, quantum mechanical calculations can aid inexploring the excited state surface.

In 2015, Dube and coworkers introduced a light-driven mo-lecular motor based on a thioindigo unit fused with a stilbenefragment (Fig. 10) (94). The design and rotational cycle resemblethat of the molecular motors based on overcrowded alkeneswith the distinct difference that a sulfoxide stereogenic center ispresent. Due to the steric crowding around the central axle, thesubstituents of the central double bond are twisted out of plane,giving the molecule a helical shape. The helicity is dictated bythe chirality of the sulfoxide. The behavior of this motor wasexamined by UV-visible and 1H-NMR spectroscopy, showingthat, in analogy to the overcrowded alkene motors, the rota-tional cycle consists of four steps: two alternating sets of pho-tochemical and thermal isomerization steps. The photochemicalisomerization could be induced by light of wavelengths up to500 nm and a frequency of rotation of 1 kHz at room tempera-ture was determined. During the 1H-NMR studies, only the(E,S,M )-15 unstable state was observed, whereas the (Z,S,M )-15state could not be detected, presumably because the THI is toofast. This hypothesis was supported by detailed DFT calcula-tions showing a four-step unidirectional rotary cycle. More re-cently, a more sterically crowded and, therefore, slower motor

was synthesized, which allowed for the direct observation of thefourth state (95).

OutlookSince the first development of light-driven molecular rotarymotors two decades ago, great progress has been made incontrolling unidirectional rotation around double bonds. Inparticular, the overcrowded alkene-based molecular rotarymotors have been thoroughly investigated. Various designs arenow increasingly applied to control dynamic functions (12);however, for a wider range of applications of these motors fur-ther improvements are essential (e.g., the use of longer irradi-ation wavelengths, as usually the photochemical isomerizationsteps are induced by UV light, which is harmful and thus im-pedes application in chemical biology and materials science).The first visible-light-driven motors that can be powered withlight up to 500 nm have recently been introduced, but morepowerful strategies such as multiphoton absorption or photonupconversion need to be explored since they will afford a majorred shift in the irradiation wavelength, preferably even into thenear-IR region. Although the influence of structural changes onthe speed of rotation of these motors has been well established,supramolecular and metal-based approaches that allow forspeed adjustments with multiple stimuli are highly promising.Increasing the efficiency of molecular motors is a more complexproblem that offers another nice challenge also in view of theirpotential use in nanoscale energy conversion and storage aswell as performance of mechanical work by future rotary motor-based molecular machines. In this regard, theoretical studiescould aid at improving the efficiency and motor design. Anothermajor challenge for molecular motors that comprise a stereogeniccenter is to obtain sufficient quantities of enantiopure material,which is needed to explore new applications, in particular to-ward responsive materials. Enantioselective synthetic routes to-ward first- and second-generation motors have been recentlydeveloped (96, 97) and a chiral resolution method by crystalliza-tion of first-generation motors offers important perspectives (98).

All these fundamental challenges have to be considered in theperspective of molecular machines: control of functions and thedesign of responsive materials. Tuning molecular motors to oper-ate in complex dynamic systems will require, among others, syn-chronization of rotary and translational motion, precise organizationand cooperativity, and amplification of motion along length scales.A first approach toward coupled motion was recently reported byour group, in which the rotary motion of the molecular motor istransferred to the synchronized movement of a connected biarylrotor (99). The prospects for controlling motion at the nano scaleand beyond will continue to provide fascinating challenges for themolecular designer and many bright roads for the molecular mo-torist in the future.

AcknowledgmentsThis work was supported by NanoNed, The Netherlands Organization forScientific Research Top grant (to B.L.F.) and Veni Grant 722.014.006 (to S.J.W.),the Royal Netherlands Academy of Arts and Sciences, the Ministry of Education,Culture and Science (Gravitation programme 024.001.035), and EuropeanResearch Council Advanced Investigator Grant 694345 (to B.L.F.).

1 Balzani V, Credi A, Raymo FM, Stoddart JF (2000) Artificial molecular machines. Angew Chem Int Ed Engl 39:3348–3391.2 Kinbara K, Aida T (2005) Toward intelligent molecular machines: Directed motions of biological and artificial molecules and assemblies. Chem Rev105:1377–1400.

Fig. 10. Rotational cycle for hemithioindigo-based motors.

Roke et al. PNAS | September 18, 2018 | vol. 115 | no. 38 | 9429

Dow

nloa

ded

by g

uest

on

June

25,

202

0

3 Browne WR, Feringa BL (2006) Making molecular machines work. Nat Nanotechnol 1:25–35.4 Kay ER, Leigh DA (2015) Rise of the molecular machines. Angew Chem Int Ed Engl 54:10080–10088.5 Balzani V, Credi A, Venturi M (2008) Molecular Devices and Machines: Concepts and Perspectives for the Nanoworld (Wiley-VCH, Weinheim, Germany), 2nd Ed.6 Sauvage JP, Gaspard P, eds (2010) From Non-Covalent Assemblies to Molecular Machines (Wiley-VCH, Weinheim, Germany).7 Bruns CJ, Stoddart JF (2016) The Nature of the Mechanical Bond: From Molecules to Machines (Wiley, Inc., Hoboken, NJ).8 Kelly TR, De Silva H, Silva RA (1999) Unidirectional rotary motion in a molecular system. Nature 401:150–152.9 Koumura N, Zijlstra RW, van Delden RA, Harada N, Feringa BL (1999) Light-driven monodirectional molecular rotor. Nature 401:152–155.

10 Leigh DA, Wong JKY, Dehez F, Zerbetto F (2003) Unidirectional rotation in a mechanically interlocked molecular rotor. Nature 424:174–179.11 Kassem S, et al. (2017) Artificial molecular motors. Chem Soc Rev 46:2592–2621.12 van Leeuwen T, Lubbe AS, �Stacko P, Wezenberg SJ, Feringa BL (2017) Dynamic control of function by light-driven molecular motors. Nat Rev Chem 1:96.13 Jimenez MC, Dietrich-Buchecker C, Sauvage JP (2000) Towards synthetic molecular muscles: Contraction and stretching of a linear rotaxane dimer. Angew Chem

Int Ed Engl 39:3284–3287.14 Silvi S, Venturi M, Credi A (2009) Artificial molecular shuttles: From concepts to devices. J Mater Chem 19:2279–2294.15 Badjic JD, Balzani V, Credi A, Silvi S, Stoddart JF (2004) A molecular elevator. Science 303:1845–1849.16 Leigh DA, Lewandowska U, Lewandowski B, Wilson MR (2014) Synthetic molecular walkers. Top Curr Chem 354:111–138.17 Ragazzon G, Baroncini M, Silvi S, Venturi M, Credi A (2015) Light-powered autonomous and directional molecular motion of a dissipative self-assembling system.

Nat Nanotechnol 10:70–75.18 Cheng C, et al. (2015) An artificial molecular pump. Nat Nanotechnol 10:547–553.19 Erbas-Cakmak S, et al. (2017) Rotary and linear molecular motors driven by pulses of a chemical fuel. Science 358:340–343.20 Kassem S, Lee ATL, Leigh DA, Markevicius A, Solà J (2016) Pick-up, transport and release of a molecular cargo using a small-molecule robotic arm. Nat Chem 8:138–143.21 Schliwa M, ed (2003) Molecular Motors (Wiley-VCH, Weinheim, Germany).22 Lubbe AS, van Leeuwen T, Wezenberg SJ, Feringa BL (2017) Designing dynamic functional molecular systems. Tetrahedron 73:4837–4848.23 Wang J, Feringa BL (2011) Dynamic control of chiral space in a catalytic asymmetric reaction using a molecular motor. Science 331:1429–1432.24 van Dijken DJ, Chen J, Stuart MCA, Hou L, Feringa BL (2016) Amphiphilic molecular motors for responsive aggregation in water. J Am Chem Soc 138:660–669.25 Collier CP, et al. (1999) Electronically configurable molecular-based logic gates. Science 285:391–394.26 Green JE, et al. (2007) A 160-kilobit molecular electronic memory patterned at 10(11) bits per square centimetre. Nature 445:414–417.27 Eelkema R, et al. (2006) Molecular machines: Nanomotor rotates microscale objects. Nature 440:163.28 Foy JT, et al. (2017) Dual-light control of nanomachines that integrate motor and modulator subunits. Nat Nanotechnol 12:540–545.29 Li Q, et al. (2015) Macroscopic contraction of a gel induced by the integrated motion of light-driven molecular motors. Nat Nanotechnol 10:161–165.30 Goujon A, et al. (2017) Bistable [c2] daisy chain rotaxanes as reversible muscle-like actuators in mechanically active gels. J Am Chem Soc 139:14825–14828.31 Chen J, et al. (2014) Textures of cholesteric droplets controlled by photo-switching chirality at the molecular level. J Mater Chem C Mater Opt Electron Devices

2:8137–8141.32 Pijper D, Feringa BL (2007) Molecular transmission: Controlling the twist sense of a helical polymer with a single light-driven molecular motor. Angew Chem Int Ed

Engl 46:3693–3696.33 Berna J, et al. (2005) Macroscopic transport by synthetic molecular machines. Nat Mater 4:704–710.34 Chen K-Y, et al. (2014) Control of surface wettability using tripodal light-activated molecular motors. J Am Chem Soc 136:3219–3224.35 Coleman AC, et al. (2011) Light-induced disassembly of self-assembled vesicle-capped nanotubes observed in real time. Nat Nanotechnol 6:547–552.36 Dietrich-Buchecker CO, Sauvage JP, Kintzinger JP (1983) Une nouvelle famille de molecules : Les metallo-catenanes. Tetrahedron Lett 24:5095–5098.37 Anelli PL, Spencer N, Stoddart JF (1991) A molecular shuttle. J Am Chem Soc 113:5131–5133.38 Sauvage J-P (1998) Transition metal-containing rotaxanes and catenanes in motion: Toward molecular machines and motors. Acc Chem Res 31:611–619.39 Erbas-Cakmak S, Leigh DA, McTernan CT, Nussbaumer AL (2015) Artificial molecular machines. Chem Rev 115:10081–10206.40 Tierney HL, et al. (2011) Experimental demonstration of a single-molecule electric motor. Nat Nanotechnol 6:625–629.41 Fletcher SP, Dumur F, Pollard MM, Feringa BL (2005) A reversible, unidirectional molecular rotary motor driven by chemical energy. Science 310:80–82.42 Collins BSL, Kistemaker JCM, Otten E, Feringa BL (2016) A chemically powered unidirectional rotary molecular motor based on a palladium redox cycle. Nat

Chem 8:860–866.43 Mc Ilwain JT (1997) An Introduction to the Biology of Vision (Cambridge Univ Press, Cambridge, UK).44 Waldeck DH (1991) Photoisomerization dynamics of stilbenes. Chem Rev 91:415–436.45 Feringa B, Wynberg H (1977) Torsionally distorted olefins. Resolution of cis- and trans-4,4′-Bi-1,1′,2,2′,3,3′-hexahydrophenanthrylidene. J Am Chem Soc 99:602–603.46 Feringa BL, Jager WF, De Lange B, Meijer EW (1991) Chiroptical molecular switch. J Am Chem Soc 113:5468–5470.47 Koumura N, Geertsema EM, van Gelder MB, Meetsma A, Feringa BL (2002) Second generation light-driven molecular motors. Unidirectional rotation controlled by a

single stereogenic center with near-perfect photoequilibria and acceleration of the speed of rotation by structural modification. J Am Chem Soc 124:5037–5051.48 Kistemaker JCM, �Stacko P, Visser J, Feringa BL (2015) Unidirectional rotary motion in achiral molecular motors. Nat Chem 7:890–896.49 Kudernac T, et al. (2011) Electrically driven directional motion of a four-wheeled molecule on a metal surface. Nature 479:208–211.50 Conyard J, et al. (2012) Ultrafast dynamics in the power stroke of a molecular rotary motor. Nat Chem 4:547–551.51 Klok M, et al. (2008) New mechanistic insight in the thermal helix inversion of second-generation molecular motors. Chemistry 14:11183–11193.52 Cnossen A, Kistemaker JCM, Kojima T, Feringa BL (2014) Structural dynamics of overcrowded alkene-based molecular motors during thermal isomerization. J Org

Chem 79:927–935.53 Pijper D, van Delden RA, Meetsma A, Feringa BL (2005) Acceleration of a nanomotor: Electronic control of the rotary speed of a light-driven molecular rotor. J Am

Chem Soc 127:17612–17613.54 ter Wiel MKJ, van Delden RA, Meetsma A, Feringa BL (2005) Light-driven molecular motors: Stepwise thermal helix inversion during unidirectional rotation of

sterically overcrowded biphenanthrylidenes. J Am Chem Soc 127:14208–14222.55 Zijlstra RWJ, et al. (1999) Chemistry of unique chiral olefins. 4. Theoretical studies of the racemization mechanism of trans- and cis-1,1′,2,2′,3,3′,4,4′-octahydro-

4,4′-biphenanthrylidenes. J Org Chem 64:1667–1674.56 ter Wiel MKJ, van Delden RA, Meetsma A, Feringa BL (2003) Increased speed of rotation for the smallest light-driven molecular motor. J Am Chem Soc

125:15076–15086.57 Ter Wiel MKJ, Kwit MG, Meetsma A, Feringa BL (2007) Synthesis, stereochemistry, and photochemical and thermal behaviour of bis-tert-butyl substituted

overcrowded alkenes. Org Biomol Chem 5:87–96.58 Pollard MM, Meetsma A, Feringa BL (2008) A redesign of light-driven rotary molecular motors. Org Biomol Chem 6:507–512.59 Kistemaker JCM, Pizzolato SF, van Leeuwen T, Pijper TC, Feringa BL (2016) Spectroscopic and theoretical identification of two thermal isomerization pathways for

bistable chiral overcrowded alkenes. Chemistry 22:13478–13487.60 Klok M, et al. (2008) MHz unidirectional rotation of molecular rotary motors. J Am Chem Soc 130:10484–10485.61 Vicario J, Meetsma A, Feringa BL (2005) Controlling the speed of rotation in molecular motors. Dramatic acceleration of the rotary motion by structural

modification. Chem Commun (Camb) 105:5910–5912.62 Fernandez Landaluce T, London G, Pollard MM, Rudolf P, Feringa BL (2010) Rotary molecular motors: A large increase in speed through a small change in design.

J Org Chem 75:5323–5325.

9430 | www.pnas.org/cgi/doi/10.1073/pnas.1712784115 Roke et al.

Dow

nloa

ded

by g

uest

on

June

25,

202

0

63 Vicario J, Walko M, Meetsma A, Feringa BL (2006) Fine tuning of the rotary motion by structural modification in light-driven unidirectional molecular motors. J AmChem Soc 128:5127–5135.

64 Bauer J, Hou L, Kistemaker JCM, Feringa BL (2014) Tuning the rotation rate of light-driven molecular motors. J Org Chem 79:4446–4455.65 Lubbe AS, Kistemaker JCM, Smits EJ, Feringa BL (2016) Solvent effects on the thermal isomerization of a rotary molecular motor. Phys Chem Chem Phys

18:26725–26735.66 Kistemaker JCM, Lubbe AS, Bloemsma EA, Feringa BL (2016) On the role of viscosity in the Eyring equation. ChemPhysChem 17:1819–1822.67 Chen J, Kistemaker JCM, Robertus J, Feringa BL (2014) Molecular stirrers in action. J Am Chem Soc 136:14924–14932.68 Qu D-H, Feringa BL (2010) Controlling molecular rotary motion with a self-complexing lock. Angew Chem Int Ed Engl 49:1107–1110.69 Faulkner A, van Leeuwen T, Feringa BL, Wezenberg SJ (2016) Allosteric regulation of the rotational speed in a light-driven molecular motor. J Am Chem Soc

138:13597–13603.70 Lerch MM, Hansen MJ, van Dam GM, Szymanski W, Feringa BL (2016) Emerging targets in photopharmacology. Angew Chem Int Ed Engl 55:10978–10999.71 Russew M-M, Hecht S (2010) Photoswitches: From molecules to materials. Adv Mater 22:3348–3360.72 Bleger D, Hecht S (2015) Visible-light-activated molecular switches. Angew Chem Int Ed Engl 54:11338–11349.73 van Delden RA, Koumura N, Schoevaars A, Meetsma A, Feringa BL (2003) A donor-acceptor substituted molecular motor: Unidirectional rotation driven by visible

light. Org Biomol Chem 1:33–35.74 van Leeuwen T, Pol J, Roke D, Wezenberg SJ, Feringa BL (2017) Visible-light excitation of a molecular motor with an extended aromatic core. Org Lett

19:1402–1405.75 Cnossen A, et al. (2012) Driving unidirectional molecular rotary motors with visible light by intra- and intermolecular energy transfer from palladium porphyrin.

J Am Chem Soc 134:17613–17619.76 Wezenberg SJ, Chen KY, Feringa BL (2015) Visible-light-driven photoisomerization and increased rotation speed of a molecular motor acting as a ligand in a

Ruthenium(II) complex. Angew Chem Int Ed Engl 54:11457–11461.77 Carling CJ, Boyer JC, Branda NR (2009) Remote-control photoswitching using NIR light. J Am Chem Soc 131:10838–10839.78 Croissant J, et al. (2013) Two-photon-triggered drug delivery in cancer cells using nanoimpellers. Angew Chem Int Ed Engl 52:13813–13817.79 Conyard J, Cnossen A, Browne WR, Feringa BL, Meech SR (2014) Chemically optimizing operational efficiency of molecular rotary motors. J Am Chem Soc

136:9692–9700.80 Augulis R, Klok M, Feringa BL, Van Loosdrecht PHM (2009) Light-driven rotary molecular motors: An ultrafast optical study. Physica Status Solidi (C) Current Topics

in Solid State Physics (Wiley-VCH, Weinheim, Germany), pp 181–184.81 Hall CR, et al. (2017) Ultrafast dynamics in light-driven molecular rotary motors probed by femtosecond stimulated Raman spectroscopy. J Am Chem Soc

139:7408–7414.82 Kazaryan A, et al. (2010) Understanding the dynamics behind the photoisomerization of a light-driven fluorene molecular rotary motor. J Phys Chem A

114:5058–5067.83 Kazaryan A, Lan Z, Schäfer LV, Thiel W, Filatov M (2011) Surface hopping excited-state dynamics study of the photoisomerization of a light-driven fluorene

molecular rotary motor. J Chem Theory Comput 7:2189–2199.84 Pang X, et al. (2017) “Watching” the dark state in ultrafast nonadiabatic photoisomerization process of a light-driven molecular rotary motor. J Phys Chem A

121:1240–1249.85 Conyard J, et al. (2017) Ultrafast excited state dynamics in molecular motors: Coupling of motor length to medium viscosity. J Phys Chem A 121:2138–2150.86 Filatov M, Olivucci M (2014) Designing conical intersections for light-driven single molecule rotary motors: From precessional to axial motion. J Org Chem

79:3587–3600.87 Nikiforov A, Gamez JA, Thiel W, Filatov M (2016) Computational design of a family of light-driven rotary molecular motors with improved quantum efficiency.

J Phys Chem Lett 7:105–110.88 Briand J, et al. (2010) Coherent ultrafast torsional motion and isomerization of a biomimetic dipolar photoswitch. Phys Chem Chem Phys 12:3178–3187.89 Logtenberg H, et al. (2016) Towards redox-driven unidirectional molecular motion. ChemPhysChem 17:1895–1901.90 �Stacko P, Kistemaker JCM, Feringa BL (2017) Fluorine-substituted molecular motors with a quaternary stereogenic center. Chemistry 23:6643–6653.91 Lehn JM (2006) Conjecture: Imines as unidirectional photodriven molecular motors-motional and constitutional dynamic devices. Chemistry 12:5910–5915.92 Greb L, Lehn J-M (2014) Light-driven molecular motors: Imines as four-step or two-step unidirectional rotors. J Am Chem Soc 136:13114–13117.93 Greb L, Eichhöfer A, Lehn JM (2015) Synthetic molecular motors: Thermal N inversion and directional photoinduced C=N bond rotation of camphorquinone

imines. Angew Chem Int Ed Engl 54:14345–14348.94 Guentner M, et al. (2015) Sunlight-powered kHz rotation of a hemithioindigo-based molecular motor. Nat Commun 6:8406.95 Huber LA, et al. (2017) Direct observation of hemithioindigo-motor unidirectionality. Angew Chem Int Ed Engl 56:14536–14539.96 Neubauer TM, et al. (2014) Asymmetric synthesis of first generation molecular motors. Org Lett 16:4220–4223.97 van Leeuwen T, Danowski W, Otten E, Wezenberg SJ, Feringa BL (2017) Asymmetric synthesis of second-generation light-driven molecular motors. J Org Chem

82:5027–5033.98 van Leeuwen T, et al. (2016) Enantiopure functional molecular motors obtained by a switchable chiral-resolution process. Chemistry 22:7054–7058.99 �Stacko P, et al. (2017) Locked synchronous rotor motion in a molecular motor. Science 356:964–968.

Roke et al. PNAS | September 18, 2018 | vol. 115 | no. 38 | 9431

Dow

nloa

ded

by g

uest

on

June

25,

202

0