622 | J. Mater. Chem. C, 2019, 7, 622--629 This journal is©The Royal Society of Chemistry 2019

Cite this: J.Mater. Chem. C, 2019,

7, 622

Photo-activated bimorph composites of Kaptonand liquid-crystalline polymer towards biomimeticcircadian rhythms of Albizia julibrissin leaves†

Xiao Li,a Shudeng Ma,a Jing Hu,a Yue Ni,a Zhiqun Lin b and Haifeng Yu *a

Circadian rhythm is a built-in bioclock widely existing in living organisms, not only in animals but also in

plants. Particularly, circadian rhythm is of great importance for the growth of plants. To mimic the

circadian rhythm behavior of Albizia julibrissin leaves, we designed photo-activated bimorph composites

with several kinds of photoresponsive liquid-crystalline polymers and commercially-available polyimide

(Kapton). Compared with conventional photo-actuators, the fabricated bimorph composite possesses

good mechanical properties, a large displacement angle and a fast photoresponsive rate at room

temperature. Upon irradiation with actinic light, unique photomechanical behaviors were observed, in

which the bimorph composites always bent towards the Kapton layer side independent of the incident

direction of UV light, as a result of the photoinduced volume expansion of the liquid-crystalline polymer

layer. To further explore the photomechanical properties, the F (photoinduced driving force)–I (light intensity)

and y (displacement angle)–I (light intensity) relationships of the photo-activated bimorph composites

were theoretically proposed based on a classical double beam model. Taking advantage of their sensitivity

to light intensities, artificial Albizia julibrissin leaves exhibiting circadian rhythms upon UV irradiation with

time varying light intensities (simulating the sunlight change from sunrise to sunset) were successfully

fabricated, which may extend the versatility of biomimetic research studies.

Introduction

Over millions of years of evolution, plants have adapted to thenatural world in a perfect manner. Several movements of plantshave been simulated by soft actuators, including the stimuli-responsive motion of Mimosa pudica,1 the gripping of flytraps,2

the burst of seedpods,3,4 and the winding of cucumber tendrilcoils.5,6 One typical example of plant movement is the ‘‘sleeping’’behavior (circadian rhythm) of Albizia julibrissin (known as silktree in China) leaves.7 This ‘‘sleeping’’ movement is of pivotalimportance to the growth of plants, which helps them to preventwater evaporation and keep warm at night. Although the

molecular mechanism of controlling the circadian rhythmhas been discovered by the 2017 Nobel Prize laureates,8 biomimeticresearch on plant circadian rhythm has not been reported yet. Theopening and closing movements of symmetric leaflets are thoughtto be strongly related to the inherent structure of the leaves andenvironmental stimuli (in this case, sunlight).9 As shown inFig. 1a, a wide variety of photoreceptors, such as phytochromes(for IR in sunlight) and cryptochromes (for UV and blue light),are located in the palisade tissue to ‘‘sense’’ sunlight. In addi-tion, the structure of the pinnate leaflets is crucial for such‘‘sleeping’’ behavior. The cells at the stipule of each leaflet actas the movement executors, propelling the leaflets to open orclose by turgor pressure changes based on the illuminationinformation offered by the photoreceptors. To mimic thesesunlight-stimulated movements, it is necessary to introducestimuli-responsive functional materials into the elegantlydesigned structures. Several works have reported bilayer-structured actuators composed of liquid crystalline elastomerlayers and stimuli-responsive layers,10–14 e.g. thermal responsive10

or humidity-responsive materials.12

Due to the excellent photoresponsive characteristics, azobenzene-containing liquid-crystalline polymer (LCP) is one of the mostfascinating choices as the photoreceptor.15–31 Upon irradiation withactinic light, the molecular shape change of rod-like azobenzenes

a Department of Material Science and Engineering, College of Engineering and

Key Laboratory of Polymer Chemistry and Physics of Ministry of Education,

Peking University, Beijing, 100871, China. E-mail: yuhaifeng@pku.edu.cnb School of Materials Science and Engineering, Georgia Institute of Technology,

Atlanta, GA, 30332, USA

† Electronic supplementary information (ESI) available: Detailed information ofthermal properties, photoresponsive properties, miscellaneous data of LC copoly-mers and supplementary note for derivation of eqn (3) (PDF) mechanicalbehaviour of the bilayer film. (avi) Opening and closing movements of artificialleaves with UV irradiation simulating sunlight change. (avi) Opening and closingmovement of artificial pinnate compound leaves. (avi). See DOI: 10.1039/c8tc05186k

Received 14th October 2018,Accepted 7th December 2018

DOI: 10.1039/c8tc05186k

rsc.li/materials-c

Journal ofMaterials Chemistry C

PAPER

Publ

ishe

d on

07

Dec

embe

r 20

18. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 2

/13/

2019

10:

51:0

7 PM

.

View Article OnlineView Journal | View Issue

This journal is©The Royal Society of Chemistry 2019 J. Mater. Chem. C, 2019, 7, 622--629 | 623

upon trans–cis photoisomerization can be autonomously amplifiedinto visible macroscopic motion in LCP films by the synergisticeffect, ensuring that the polymer system is suitable for light-drivenactuators.15–18 Thus, a series of macroscopic motions have beenobserved such as oscillation,18,19 wave propagation,20–22 self-origami,23 controlled microfluidics,24 helical movement,25–28

motor, rotational,29,30 and so on.21,31 Moreover, LCPs could alsoprovide robust mechanical properties and resistance to hostileenvironments after a chemical crosslinking treatment. Theseunique features enable light-responsive LCPs to be desirable forsimulation of the circadian rhythms of silk tree leaves. To the bestof our knowledge, there are no reports about such biomimeticwork so far.

In this paper, we designed biomimetic leaves with photo-activated bimorph composites (Fig. 1b), which could accuratelyreproduce the innate structure and the circadian rhythm functionof silk tree leaves. Specifically, the artificial leaves were crafted witha bilayer composite film composed of commercial polyimide(Kapton) substrate and one layer of photoresponsive LCP by afacile drop casting method. The LCP layer acts as both thephotoreceptor and the movement executor. The bimorph compo-site film displayed a dramatic photomechanical bending towardsthe Kapton layer due to the photoinduced volume expansion of thephotoresponsive LCP layer. In addition, photomechanical behaviorof the bimorph composite films with varied thickness was system-atically studied, and the relationship between the displacementangle (y) and the light intensity (I) of the photo-activated bimorphwas quantitatively analyzed for the first time. As y is sensitive to UVintensity change, smart artificial leaves displayed gradual openingand closing as the UV intensity switched. This work may expandthe diversity of artificial autonomous movement systems andprovide theoretical insights into photomechanical properties ofphoto-activated bimorph composites.

Results and discussionDesign of the biomimetic system

As shown in Fig. 1a, the circadian rhythm behaviour of naturalsilk tree leaves is usually accomplished via the synergy effectbetween the photosensitization process and the movementcontrolling process.9 The former process is achieved throughthe photoreceptors distributed in the palisade tissue under theblade surface of the leaflet (Fig. 1a, right panel). While the latterprocess relies on the turgor pressure change of cells inducedby sunlight. Accordingly, we designed a bilayer structurewith a Kapton bottom surface and LCP top surface of theartificial leaves, as shown in Fig. 1b. The photoresponsiveLCP layer was designed as the outer surface of the artificialleaves, working as the photoreceptor and the movementexecutor, while the Kapton layer was a photo-inert layer.Furthermore, the outer surface was chemically crosslinkedto enable the artificial leaves to accommodate harsh weatherand environments.

Design of molecular structure and bimorph preparation

Fig. 2a shows the molecular structure of random LC copolymersP(M6ABOC2-r-M6AzPy) obtained by copolymerization of twomonomers, M6ABOC2 and M6AzPy. The homopolymer PM6ABOC2

exhibiting a typical nematic phase has been widely applied in thedesign of photoresponsive LCP,32,33 endowing the copolymer withLC ordering. The introduction of the M6AzPy monomer contain-ing one active pyridine group provides the possible crosslinkingsite. Two random copolymers P82 (80 mol% of M6ABOC2 and20 mol% of M6AzPy) and P55 (50 mol% of M6ABOC2 and 50 mol%of M6AzPy) were synthesized, respectively. Their molecular weightsand thermal properties are listed in Table 1.

Fig. 1 (a) A scheme of silk tree leaves with symmetrical leaflets along therachis. Left panel: Unfolded leaflets during the daytime; right penal: cross-sectional structure of a typical leaflet. (b) A scheme of light-triggeredartificial leaves designed in this work.

Fig. 2 (a) Molecular design of the LCP. (b) A scheme of the drop castingmethod to prepare the LCP–Kapton bimorph composites.

Paper Journal of Materials Chemistry C

Publ

ishe

d on

07

Dec

embe

r 20

18. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 2

/13/

2019

10:

51:0

7 PM

. View Article Online

624 | J. Mater. Chem. C, 2019, 7, 622--629 This journal is©The Royal Society of Chemistry 2019

Both P55 and P82 displayed a typical Schlieren texture ofnematic phase under observation of a polarizing optical micro-scope (POM, Fig. S1, ESI†). The homopolymer PM6ABOC2 hadthe broadest LC temperature range of about 54 1C. As the molarratio of M6AzPy increased to 50%, the LC range of the copolymerreduced to only 4 1C, accordingly. To keep the liquid crystallineproperties, the ratio of M6AzPy should be kept no morethan 50%.

As shown in Fig. 2b, the bilayer film was fabricated by afacile drop-casting method following our previously-reportedstrategy.34,35 The LCP solution was first drop casted onto theKapton film. After solvent evaporation, the composite film wasannealed at the LC temperature (P55 at 105 1C and P82 at130 1C) and slowly cooled to room temperature to allow themesogens to distribute in a nematic manner. Finally, theannealed film was crosslinked by vapor–solid intermolecularquaternization in a sealed container at 60 1C. It is worth notingthat thickness of the LCP layer could be tuned by altering theconcentration of the polymer solution. The thickness of theKapton layer was fixed at 12.5 mm, as indicated in Fig. S2 (ESI†).The corresponding LCP layer thickness, the whole bilayer filmthickness and the ratio of the LCP layer prepared from the LCPsolutions with concentrations ranging from 4 mg mL�1 to12 mg mL�1 are summarized in Fig. S3 (ESI†).

Photoresponsive behaviour of the crosslinking system

Successful crosslinking of random copolymers can be provedby the appearance of colour changes, and FTIR and UV-visabsorption spectroscopy. Digital images of a series of compo-site films during the preparation process are shown in Fig. S4(ESI†). The colour change of yellow-to-brownish occurred asthe azopyridine group gradually quaternized with DIB.36 Thiscolour change became more prominent by increasing theazopyridine ratio in the random copolymers. In the UV-visabsorption spectra of the films recorded before and aftercrosslinking, the obvious shift of the maximum absorptionpeak was easily obtained. To get rid of the influence of theazobenzene chromophores in PM6ABOC2, the homopolymerPM6AzPy in thin film was first studied. As the quaternizationreaction time was elongated, a progressive peak shift from355 nm to 397 nm was observed in the absorption spectra ofthe PM6AzPy thin film (Fig. 3a). This 42 nm red-shift afterquaternization should be relevant for a strong charge transferfrom the alkyl group to the positively charged nitrogenatom.37–39 In addition, as shown in Fig. 3b, the new peak at1634 cm�1 in the FTIR spectra is characteristic of the positivelycharged pyridine group, indicating the occurrence of thequaternization reaction.40,41

Herein, the chemical crosslinking has two functions:(1) enhancement of the solvent resistance. The random copolymerbefore treatment can be easily dissolved in DMF, THF or chloro-form, forming clear transparent solutions. While the crosslinkedLCP film is intact in organic solvents without dissolution orpeeling off from the Kapton layer (Fig. S5, ESI†). (2) Accelerationof the photoresponse rate by changing the azopyridine group intothe famous type of push–pull azobenzene. Then the photo-responsive behaviours of the crosslinked and un-crosslinkedpolymer films were studied upon irradiation with 365 nm UVlight (100 mW cm�2).

As shown in Fig. 4b, the absorption of the quaternized P55film can be regarded as the superimposition of two types ofazobenzene chromophores: ethyoxy azobenzene (ABOC2) andquaternized azopyridine (AzPy). As the polymer reaches aphotostationary state, the peak at 311 nm is consistent withthe responsive feature of the ABOC2 part;42 while the peak at403 nm is attributed to quaternized AzPy groups. The formerdidn’t participate in the quaternization reaction, retaining theoriginal responsive features of PM6ABOC2 shown in Fig. S6a(ESI†) (the p–p* transition absorption peak decreased andshifted to lower wavelength). The quaternization reaction con-verted azopyridine into a pseudo-stilbene type of electron-push–pull azobenzene, which thus had a great impact on thephotoresponsive behaviour. Extremely fast cis-to-trans thermalisomerization (in several ms) occurs in quaternized PM6AzPyaccording to former reports.39,43 As light was turned off,cis-molecules instantly recovered to the initial trans state,before the spectrometer could detect it, thus leading to stableUV-vis absorption spectra in Fig. S6c (ESI†).

In terms of response rate, pure PM6ABOC2 film reachedits photostationary state within 5 min of photoirradiation

Table 1 Molecular weight and thermal properties of the polymers

Polymer Tg (1C) Tm (1C) TNI (1C) LC range Mn PDI

PM6ABOC2 94 — 148 54 24k 1.6P82 75 105 132 27 17.9k 1.63P55 60 102 106 4 19.2k 1.62PM6AzPy 47 — — 0 19k 1.53

Fig. 3 (a) UV-vis absorption spectra of PM6AzPy homopolymer as thecrosslinking time increases. (b) FTIR spectra of P55 before and afterchemical crosslinking.

Fig. 4 UV-vis absorption spectra of the P55 film before (a) and after(b) the chemical crosslinking. The films were exposed to UV light at 365 nm(100 mW cm�2).

Journal of Materials Chemistry C Paper

Publ

ishe

d on

07

Dec

embe

r 20

18. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 2

/13/

2019

10:

51:0

7 PM

. View Article Online

This journal is©The Royal Society of Chemistry 2019 J. Mater. Chem. C, 2019, 7, 622--629 | 625

(Fig. S6a, ESI†). While PM6AzPy film afforded a faster isomer-ization process, 5 s of light exposure was enough to obtain thecis-rich equilibrium state (Fig. S6b, ESI†). Thus, the photo-responsive rate of P55 which became stable with UV irradia-tion for 20 s (Fig. 4a) was enhanced compared with that of thehomopolymer PM6ABOC2. The response rate of crosslinkedP55 was further accelerated to 10 s owing to the existence ofthe electron-push–pull azobenzene induced by the quater-nized azopyridine. In terms of the UV-response rate, thesepolymers follow a sequence of: crosslinked PM6AzPy 4PM6AzPy 4 crosslinked P55 4 P554 PM6ABOC2. The photo-responsive behaviour of P82 was almost the same as that ofP55, which was employed in the following work for its betteroptical transparency.

Photomechanical behaviours of the bimorph composites

As the photoresponsive features have been clarified above, wenow turn to the photomechanical behaviour of the bimorphcomposite. Upon irradiation with UV light, they exhibited rapidand substantial photoinduced bending motion towards Kaptonside. Fig. 5a and b show representative snapshots of thebimorph composite film under photoirradiation and recoveryprocesses after removal of light (Movie S1, ESI†). With a rapidresponse rate, the film reached its maximum displacementangle (defined as the angle between normal line and the lineconnecting two ends of the film, inserted in Fig. 5a) within 2 sand recovered back within 1.5 s under room light. After the filmreached the maximum bending angle, further elongation ofexposure time did not increase the displacement angle (Fig. S7,ESI†). Based on the response rate difference between PM6ABOC2,PM6AzPy, P55, crosslinked P55 and crosslinked PM6AzPy we havediscussed above, it is expected that the addition of azopyridinegroup and crosslinking would enhance the photoresponsive rateof the bimorph composite. However, we found that there wasalmost no obvious difference in the response rate once the

bimorph composites with different kinds of LCPs were prepared.As indicated in Fig. S8 (ESI†), all the bimorph composite filmsdemonstrate a response time between 1.3 s to 2.0 s and a reversetime of 1.5–2.8 s. It could be concluded from the results that theUV-response rate of the bimorph composite film was not mainlydetermined by the response rate of the photoactive layer of LCPs.

Now we turn to the bending direction of the bimorphcomposite, as UV light shined on the bimorph, the film alwaysbent towards the Kapton side independent of the incidentdirection, as shown in Fig. 5a and b (the inserted schemes).As UV light irradiated from either side of the bilayer film, bothdisplacement angle and driving force were almost equal withinthe margin of error (Table S1, ESI†). This unique photomechanicalbehavior of the bimorph composite film is different from thatof the mono-domain oriented LC elastomers or homogeneousLC/polymer composite films,2,15,44–48 which often bend towardsthe light source due to the anisotropic volume contractionalong the long axis of mesogens. In our strategy, the bimorphcomposite film containing crosslinked LCP demonstrates a largedisplacement angle even though the mesogens were randomlydistributed. Barrett et al. proved the light-induced expansioneffect of azobenzene-containing polymers via ellipsometry andneutron reflectometry.49–51 As our group has recently reported,one crosslinked LCP system with azobenzene groups located inthe side chain instead of the crosslinker expanded upon UVirradiation as free volume is needed for the photochemicalprocess of trans–cis isomerization.51,52 Taking account of thephotoinduced expansion theories and similar response rates ofthe bimorph composites with different types of LCPs, we proposethat the photo-actuation behavior should be rationalized asfollows: UV irradiation triggered volume expansion of the photo-active LCP layer and induced stress within a short period betweenthe two layers, thus the stress pushes the bimorph compositefilm to bend towards the photo-inert Kapton side (Fig. 5c). Thatis, the photomechanical behavior of the bimorph compositeshould originate from the bilayer structure instead of the typeof photoresponsive LCPs.

Theoretical model for photomechanics of the bimorphcomposites

Since the bilayer structure of the composite film has a majoreffect on its photomechanical properties, we then systematicallyinvestigated the influence of the LCP layer thickness and lightintensity of UV light. Fig. 6a and b illustrate the photoinduceddriving force with varied LCP layer thickness at different lightintensities. UV illumination (100 mW cm�2) triggered animmediate increase of the driving force to �72 mN (the negativevalue means the strip was bent rather than straightened). All thesamples displayed very good reversibility without detectableattenuation as the cycle time increased.

As shown in Fig. 6a, a thicker LCP layer and a strongerlight intensity could produce a larger driving force. Ideally, thephotoinduced driving force should be proportional to thethickness of the LCP layer when the Kapton layer thickness wasfixed. Consequently, we established the F–I relationship accordingto the classical double-layer beam theory. An assumption

Fig. 5 Photomechanical behaviours of the LCP–Kapton bimorph com-posite films. (a) Snapshots of a composite film during the UV exposureprocess, UV light was shined on the LCP side from the left. From left toright: before UV, UV on for 2 s, recovery for 1.5 s. (b) UV light shined on theKapton side from the left. Inset is a scheme for the bending behaviour ofbilayer film (green: LCP layer; yellow: Kapton layer). (c) Scheme illustratingphotoisomerization induced expansion of the LCP layer.

Paper Journal of Materials Chemistry C

Publ

ishe

d on

07

Dec

embe

r 20

18. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 2

/13/

2019

10:

51:0

7 PM

. View Article Online

626 | J. Mater. Chem. C, 2019, 7, 622--629 This journal is©The Royal Society of Chemistry 2019

should be made that the cross-sections of the strip originallyplane and perpendicular to the axis remain plane during thephotomechanical bending and become perpendicular to thecurved axis of strip.53 The volume expansion induced strain ofthe LCP layer could be calculated as follows:

edrive = F/E1h1b = aI (1)

F = aE1h1bI (2)

E1 Young’s modulus of the LCP layer; h1 thickness of theLCP layer; b width of the bilayer strip; F driving force; andI UV light intensity.

E1 was measured by AFM in contact mode as 533 MPa(the force curve is given in Fig. S8, ESI†). a is a constant thatcould be determined from the measured data. To ensure theaccuracy of the constant, a thin LCP film with a thickness of 2.5 mmunder UV irradiation was chosen to obtain a = 0.0094. Accordingly,the scale factors for samples with thicknesses of 4.7 mm, 7.2 mm and9.5 mm were k4.7 = 0.706, k7.2 = 1.082, and k9.5 = 1.353, respectively.The calculated relationship for LCP samples with varied thicknessesare shown as solid lines. For all samples, the driving force is inproportion to the UV light intensity. As the LCP film thickens,the scale factor k also increases. That is, a larger driving forcecan be produced in a thicker film upon irradiation using thesame UV light. For comparison, the experimental data weredepicted as a scattergraph in Fig. 6c. Since a was obtained froma thin sample (a thickness of 2.5 mm), the scattergraph and thecalculated solid line were exactly consistent with each other(the green line). While for thicker films, the measured drivingforce deviated a little from the calculated data. The thickerthe film, the larger the deviation. This phenomenon could be

rationalized as follows: as the LCP layer becomes thicker, thetransparency of the film decreases owing to the scattering effectof randomly-distributed mesogens. Thus, the incident lightdissipates within the LCP layer, leading to a driving force thatwas slightly smaller than the theoretical values.

Fig. 6d shows variation of the displacement angle y with thelight intensity. Obviously, the thicker film displayed a largerdisplacement angle at the fixed incident light, which correspondswell to the change of driving force in Fig. 6c. The displacementangle y can be fitted according to eqn (3):

y ¼ �3E1E2h1h2 h1 þ h2ð ÞL3 E1h12 � E2h22ð Þ2�4 E1h13 þ E2h23ð Þ E1h1 þ E2h2ð Þ

aI (3)

Here, E2 and h2 represent the Young’s modulus and thicknessof the Kapton film, respectively. L is the length of the bilayerstrip. On account of eqn (3), the displacement angle (y) shouldbe proportional to light intensity (I). In addition, the slope isrelated to the thickness of the LCP layer h1, as L, a, E1, E2 and h2

are all constants. Based on the calculation, as h1 increases, theslope will also elevate accordingly. This theoretical analysis fitswell with the results shown in Fig. 6d.

Biomimetic ‘‘sleeping’’ behaviours of silk tree leaves

The unique photomechanical bending behaviour of the bimorphmakes them a perfect choice for mimicking the ‘‘sleeping’’behaviour of silk tree leaves. Thus, we designed artificial leaveswith two pieces of the bimorph composite strips fixed face-to-face (LCP layer-to-LCP layer), and one end of the twin strips wasfixed on a foldback clip. As light was incident, the photoreceptorsof the LCP layer absorbed the light energy, and photoinducedexpansion occurred, propelling the twin ‘‘leaflets’’ to open.In nature, silk tree leaves start to unfold at dawn, when thesunlight is still dim. Then, the leaves gradually open to itsmaximum angle as the sun rises above the horizon. Finally, asthe sun goes down at dusk, the opened leaves gradually returnto the closed state for ‘‘sleeping’’. To simulate the cycle ofsunlight change, we applied UV light with a gradient lightintensity (from 20 mW cm�2 to 100 mW cm�2, 5 s for each stage),as shown in Fig. 6a, b and Movie S2 (ESI†).

Just as expected, the structure of symmetrically-distributedpinnately compound leaves is critical for their ‘‘sleeping’’ beha-viours. Although the effective area exposed to sunlight is smallerin pinnately compound leaves compared with that of a singlewhole leaf, most of the ‘‘sleeping’’ plants, such as peanuts,Oxalidaceae and Mimosa pudica, have pinnately compoundleaves instead of a whole single leaf. To this end, we’ve alsodesigned artificial ‘‘leaflets’’ along the ‘‘rachis’’. This design wasachieved by cutting a bilayer film as illustrated in Fig. 7c. Toobtain sufficient light, displacements of leaflets on the twosides along the rachis were allowed. We demonstrated theopening and closing movements of artificial compound leaveswith four pairs of leaflets, as marked on the top view in Fig. 7dand Movie S3 (ESI†). Side views of the movements are also givenin Fig. 7e.

Fig. 6 (a) Driving force of the bilayer film when exposed to UV light withgradient intensities and the reversibility of the driving force change as UVlight goes through on-and-off cycles. (b) The light-produced driving forceand reversibility of the bilayer films with varied thicknesses (UV intensity:100 mW cm�2). (c) The relationship between the driving force and the UVlight intensity for films with varied thickness. Scattergraph: measured dataat gradient light intensities; solid line: theoretically calculated data witheqn (2). (d) Displacement change for each LCP thickness in the bilayer filmas the light intensity increased from 20 to 100 mW cm�2 with an interval of10 mW cm�2.

Journal of Materials Chemistry C Paper

Publ

ishe

d on

07

Dec

embe

r 20

18. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 2

/13/

2019

10:

51:0

7 PM

. View Article Online

This journal is©The Royal Society of Chemistry 2019 J. Mater. Chem. C, 2019, 7, 622--629 | 627

Conclusions

In summary, we have designed artificial leaves which could mimicthe circadian rhythm behaviour of Albizia julibrissin leaves, byemploying a robust photo-activated bimorph composited with acrosslinked LCP layer and Kapton substrate. The crafted bilayer filmdemonstrated fast and reversible photomechanical bending (within2 s) towards the Kapton layer with a tremendous displacement angle(601 at 100 mW cm�2) at room temperature. The bending behaviouris not related to the type of LCPs, but the bilayer structure. To furtherexplore the photomechanical properties of bimorphs, the relation-ship between the photoinduced displacement angle (y) and lightintensity (I) at varied thickness of the LCP layer was quantitativelydiscussed by using a bilayer thermoset model, providing a deepinsight into the photomechanical process of the photo-activatedbimorph. Upon irradiation with gradient UV light, the artificialleaves displayed circadian rhythm movements according to the lightchange from ‘‘dawn’’ to ‘‘dusk’’. This smart photo-sensing andmovement execution system paves the way to mimic the delicatemovement control system within natural plants.

ExperimentalMaterials

1,4-Diidonetetra-butane (DIB) and all reactants and solvents forsynthesis were purchased from TCI and used directly without

further purification. The monomer 6-(4-((4-ethoxyphenyl)diazenyl)-phenoxy)hexyl methacrylate (M6ABOC2) and 6-(4-(pyridylazo)-phenoxy)hexyl methacrylate (M6AzPy) were synthesized accordingto the literature report.32,33,54–56 M6ABOC2: 1H NMR (d, CDCl3):7.81 (m, 4H), 6.90 (m, 4H), 6.03 (s, 1H), 5.43 (s, 1H), 4.10 (t, 2H),4.04 (m, 2H), 3.95 (t, 2H), 1.87 (t, 6H), 1.79–1.74 (m, 2H), 1.69–1.63(m, 2H), 1.37–1.49 (m, 4H). M6AzPy 1H NMR (d, CDCl3): 8.79(d, 2H), 7.96 (d, 2H), 7.69 (d, 2H), 7.03 (d, 2H), 6.12 (s, 1H), 5.56(s, 1H), 4.19 (t, 2H), 4.08 (t, 2H), 1.96 (s, 3H), 1.80–1.90 (m, 2H),1.65–1.80 (m, 2H), 1.43–1.65 (m, 4H).

Synthesis of the random copolymers

The random copolymers were synthesized by free radical poly-merization of the two monomers with different feed ratios. Toobtain the random copolymer (P55 as an example), M6ABOC2

(410 mg, 1.0 mmol), M6AzPy (368 mg, 1.0 mmol), azodiiso-butyronitrile (AIBN, 4.2 mg, 0.025 mmol) and anhydrousdimethylformamide (DMF, 2.0 mL) were added into a flask.After freezing and drawing 3 times, the reaction was proceededat 70 1C with stirring for 48 h. After reaction, the product wasprecipitated in methanol 3 times, and dried in a vacuum ovenfor 24 h. Yield: 51.4%. P55: 1H NMR (d, CDCl3): 8.63 (s, 2H), 7.77–7.71 (m, 6H), 7.54 (s, 2H), 6.81 (m, 6H), 3.92–3.83 (m, 10H), 1.66(m, 4H), 1.54 (m, 4H), 1.43–1.24 (m, 12H) 0.83–0.98 (m, 9H).

Preparation of the bimorph composite film

Commercially-available Kapton film (thickness: 12.5 mm) wasrinsed in ethanol and deionized water successively. 12.0 mg ofthe random copolymer P55 was dissolved in a mixed solvent(DMF/THF = 1 : 9) and passed through a tetra-fluoroethylenefilter to remove insoluble impurities. After that, the filteredsolution was drop casted onto Kapton film with the size of2.5 cm � 7.5 cm. After evaporation of the solvent at roomtemperature (25 1C) for 24 hours, the composite film wasannealed at the LC temperature (105 1C for P55; 130 1C forP82) in a vacuum oven for 12 hours so that the mesogens homo-distributed randomly on the Kapton substrate. At the sametime, the residual solvent can be removed during the annealingprocess. The film thickness of the LCP layer can be tuned byaltering the concentration of the polymer solution.

Crosslinking of LCP by quaternization reaction

Quaternization reaction was carried out within sealed petridishes. A sponge with DIB crosslinker was placed at the bottomof the container, and the bimorph composite film was made tostick on top of the container, the reaction was conducted at70 1C for 6 hours.31 Finally, the crosslinked film was cut intostrips of 3 mm � 20 mm for further study.

Characterizations

FTIR measurement was conducted on a FTIR spectrometer(spectrum II, Perkin Elmer) in transmission mode. UV-visabsorption spectrum was recorded on a Lambda 750 UV/vis/NIRspectrophotometer (Perkin Elmer). AFM force curve wasobtained by MultiMode 8 (Bruker) with contact mode in air.The Young’s modulus was obtained by the average of 5 points

Fig. 7 The opening and closing movements of the artificial leaves.(a) Artificial leaves gradually opened at sun rise from the horizon. Fromleft to right, the UV light intensities were 0, 20 mW cm�2, 40 mW cm�2,60 mW cm�2, 80 mW cm�2, and 100 mW cm�2, respectively. Thecorresponding displacement angle were, 0, 71, 251, 361, 501, and 611. (b)Artificial leaves gradually closed as the sun goes down. From left to right,the UV light intensities were 100 mW cm�2, 80 mW cm�2, 60 mW cm�2,40 mW cm�2, 20 mW cm�2, and 0, respectively. (c) Scheme illustrating thedesign of artificial pinnately compound leaves. The leaflets were cut alongthe dotted line. (d and e) Opening and closing movement of artificialpinnate compound leaves upon UV on-and-off cycles. (d) Top view, thefour pairs of leaflets were marked in the middle panel. (e) Side view.

Paper Journal of Materials Chemistry C

Publ

ishe

d on

07

Dec

embe

r 20

18. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 2

/13/

2019

10:

51:0

7 PM

. View Article Online

628 | J. Mater. Chem. C, 2019, 7, 622--629 This journal is©The Royal Society of Chemistry 2019

on the same sample. Thermal property of the random copolymerswas measured via DSC 8000 (Perkin Elmer). Molecular weightof the copolymer was obtained via GPC (Waters 1515). Drivingforce of the bilayer film was tested using a tensile test machine(Jinan LianGong). Polarizing optical microscopic (POM) imageswere observed via Zeiss AXIO scope optical microscope equippedwith a hot stage. The thickness of the bilayer film was measuredusing a scanning electron microscope (SEM, S-4800, HITACHI) atan accelerating voltage of 2.0 kV. The thickness of the bilayerfilm was measured using a micrometer.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

H. Y. acknowledges financial support from the National NaturalScience Foundation of China (Grant No. 51573005, 51773002).

Notes and references

1 B. Su, S. Gong, Z. Ma, L. W. Yap and W. Cheng, Small, 2015,11, 1886–1891.

2 O. M. Wani, H. Zeng and A. Priimagi, Nat. Commun., 2017,8, 15546.

3 S. J. Aßhoff, F. Lancia, S. Iamsaard, B. Matt, T. Kudernac,S. P. Fletcher and N. Katsonis, Angew. Chem., Int. Ed., 2017,56, 3261–3265.

4 Y. Forterre and J. Dumais, Science, 2011, 333, 1715–1716.5 S. Iamsaard, S. J. Aßhoff, B. Matt, T. Kudernac, J. J. Cornelissen,

S. P. Fletcher and N. Katsonis, Nat. Chem., 2014, 6, 229–235.6 S. J. Gerbode, J. R. Puzey, A. G. McCormick and L. Mahadevan,

Science, 2012, 337, 1087–1091.7 C. R. McClung, Plant Cell, 2006, 18, 792–803.8 T. A. Bargiello, F. R. Jackson and M. W. Young, Nature, 1984,

312, 752.9 M. Ueda and Y. Nakamura, Plant Cell Physiol., 2007, 48,

900–907.10 R. R. Kohl and J. Chen, Angew. Chem., 2013, 125, 9404–9407.11 H. Xing, J. Li, Y. Shi, J. Guo and J. Wei, ACS Appl. Mater.

Interfaces, 2016, 8, 9440–9445.12 J. M. Boothby and T. H. Ware, Soft Matter, 2017, 13, 4349–4356.13 B. Zuo, M. Wang, B. Lin and H. Yang, Chem. Mater., 2018,

30, 8079–8088.14 L. Chen, M. Wang, L. Guo, B. Lin and H. Yang, J. Mater.

Chem. C, 2018, 6, 8251–8257.15 T. Ikeda, M. Nakano, Y. Yu, O. Tsutsumi and A. Kanazawa,

Adv. Mater., 2003, 15, 201–205.16 H. Yu and T. Ikeda, Adv. Mater., 2011, 23, 2149–2180.17 M. Wang, C. Zou, J. Sun, L. Zhang, L. Wang, J. Xiao, F. Li,

P. Song and H. Yang, Adv. Funct. Mater., 2017, 27, 1702261.18 T. J. White, N. V. Tabiryan, S. V. Serak, U. A. Hrozhyk,

V. P. Tondiglia, H. Koerner, R. A. Vaia and T. J. Bunning,Soft Matter, 2008, 4, 1796–1798.

19 K. M. Lee, M. L. Smith, H. Koerner, N. Tabiryan, R. A. Vaia,T. J. Bunning and T. J. White, Adv. Funct. Mater., 2011, 21,2913–2918.

20 A. H. Gelebart, D. Jan Mulder, M. Varga, A. Konya,G. Vantomme, E. W. Meijer, R. L. B. Selinger and D. J.Broer, Nature, 2017, 546, 632–636.

21 F. Ge, R. Yang, X. Tong, F. Camerel and Y. Zhao, Angew.Chem., Int. Ed., 2018, 57, 11758–11763.

22 W. Feng, D. J. Broer and D. Liu, Adv. Mater., 2018, 30, 1704970.23 J. Mu, C. Hou, H. Wang, Y. Li, Q. Zhang and M. Zhu, Sci.

Adv., 2015, 1, e1500533.24 X. Liu, S. Kim and X. Wang, J. Mater. Chem. B, 2016, 4,

7293–7302.25 J. Sun, L. Yu, L. Wang, C. Li, Z. Yang, W. He, C. Zhang,

J. Xiao, X. Yuan, F. Li and H. Yang, J. Mater. Chem. C, 2017,5, 3678–3683.

26 M. Wang, B. P. Lin and H. Yang, Nat. Commun., 2016, 7, 13981.27 J. Sun, R. Lan, Y. Gao, M. Wang, W. Zhang, L. Wang,

L. Zhang, Z. Yang and H. Yang, Adv. Sci., 2018, 5(5), 1700613.28 J. Hu, X. Li, Y. Ni, S. Ma and H. Yu, J. Mater. Chem. C, 2018,

6, 10815–10821.29 M. Yamada, M. Kondo, J. I. Mamiya, Y. Yu, M. Kinoshita,

C. J. Barrett and T. Ikeda, Angew. Chem., Int. Ed., 2008, 47,4986–4988.

30 Y. Yue, Y. Norikane, R. Azumi and E. Koyama, Nat. Commun.,2018, 9, 3234.

31 C. Ahn, K. Li and S. Cai, ACS Appl. Mater. Interfaces, 2018, 10,25689–25696.

32 H. Yu, C. Dong, W. Zhou, T. Kobayashi and H. Yang, Small,2011, 7, 3039–3045.

33 H. Yu, T. Kobayashi and G. Hu, Polymer, 2011, 52, 1554–1561.34 Z. Liu, R. Tang, D. Xu, J. Liu and H. Yu, Macromol. Rapid

Commun., 2015, 36, 1171–1176.35 R. Tang, Z. Liu, D. Xu, J. Liu, L. Yu and H. Yu, ACS Appl.

Mater. Interfaces, 2015, 7, 8393–8397.36 J. A. H. Gelebart, D. J. Mulder, G. Vantomme, A. P. H. J.

Schenning and D. J. Broer, Angew. Chem., Int. Ed., 2017, 56,13436–13439.

37 Q. Bo and Y. Zhao, Langmuir, 2007, 23, 5746–5751.38 W. Zhou, T. Kobayashi, H. Zhu and H. Yu, Chem. Commun.,

2011, 47, 12768–12770.39 W. Zhou and H. Yu, ACS Appl. Mater. Interfaces, 2012, 4, 2154–2159.40 X. Zhou, Y. Zhou, J. Nie, Z. Ji, J. Xu, X. Zhang and B. Du,

ACS Appl. Mater. Interfaces, 2014, 6, 4498–4513.41 J. Garcia-Amoros, W. A. Massad, S. Nonell and D. Velasco,

Org. Lett., 2010, 12, 3514–3517.42 X. Li, B. Li, M. He, W. Wang, T. Wang, A. Wang, J. Yu,

Z. Wang, S. Hong, M. Byun, S. Lin, H. Yu and Z. Lin,ACS Appl. Mater. Interfaces, 2018, 10, 4961–4970.

43 X. Zhou, Y. Zhou, J. Nie, Z. Ji, J. Xu, X. Zhang and B. Du,ACS Appl. Mater. Interfaces, 2014, 6, 4498–4513.

44 J. Mamiya, A. Yoshitake, M. Kondo, Y. Yu and T. Ikeda,J. Mater. Chem., 2008, 18, 63–65.

45 M. Yamada, M. Kondo, R. Miyasato, Y. Naka, J. I. Mamiya,M. Kinoshita, A. Shishido, Y. Yu, C. J. Barrett and T. Ikeda,J. Mater. Chem., 2008, 19, 60–62.

Journal of Materials Chemistry C Paper

Publ

ishe

d on

07

Dec

embe

r 20

18. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 2

/13/

2019

10:

51:0

7 PM

. View Article Online

This journal is©The Royal Society of Chemistry 2019 J. Mater. Chem. C, 2019, 7, 622--629 | 629

46 T. Ikeda, J. Mamiya and Y. Yu, Angew. Chem., Int. Ed., 2007,46, 506–528.

47 Z. Cheng, S. Ma, Y. Zhang, S. Huang, Y. Chen and H. Yu,Macromolecules, 2017, 50, 8317–8324.

48 O. M. Tanchak and C. J. Barrett, Macromolecules, 2005, 38,10566–10570.

49 K. G. Yager, O. M. Tanchak, C. Godbout, H. Fritzsche andC. J. Barrett, Macromolecules, 2006, 39, 9311–9319.

50 O. S. Bushuyev, M. Aizawa, A. Shishido and C. J. Barrett,Macromol. Rapid Commun., 2018, 39, 1700253.

51 Z. Cheng, T. Wang, X. Li, Y. Zhang and H. Yu, ACS Appl.Mater. Interfaces, 2015, 7, 27494–27501.

52 F. L. L. Visschers, M. Hendrikx, Y. Zhan and D. Liu, Soft Matter,2018, 14, 4898–4912.

53 S. Timoshenko, J. Opt. Soc. Am. Rev. Sci. Instrum., 1925, 11, 233–255.54 H. Yu, H. Liu and T. Kobayashi, ACS Appl. Mater. Interfaces,

2011, 3, 1333–1340.55 H. Zhang, R. Hao, J. K. Jackson, M. Chiao and H. Yu, Chem.

Commun., 2014, 50, 14843–14846.56 X. Liu and X. Wang, Acta Polym. Sin., 2017, 10, 1549–1556.

Paper Journal of Materials Chemistry C

Publ

ishe

d on

07

Dec

embe

r 20

18. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 2

/13/

2019

10:

51:0

7 PM

. View Article Online