Tunable Light-Responsive Polyurethane-urea Elastomer Driven by Photochemical and Photothermal Coupling Mechanism

Light-driven soft actuators based on photoresponsive materials can be used to mimic biological motion, such as hand movements, without involving rigid or bulky electromechanical actuations. However, to our knowledge, no robust photoresponsive material with desireable mechanical and biological properties and relatively simple manufacture exists for robotics and biomedical applications. Herein, we report a new visible-light-responsive thermoplastic elastomer synthesized by introducing photoswitchable moieties (i.e., azobenzene derivatives) into the main chain of poly(ε-caprolactone) based polyurethane urea (PAzo). A PAzo elastomer exhibits controllable light-driven stiffness softening due to its unique nanophase structure in response to light, while possessing excellent hyperelasticity (stretchability of 575.2%, elastic modulus of 17.6 MPa, and strength of 44.0 MPa). A bilayer actuator consisting of PAzo and polyimide films is developed, demonstrating tunable bending modes by varying incident light intensities. Actuation mechanism via photothermal and photochemical coupling effects of a soft–hard nanophase is demonstrated through both experimental and theoretical analyses. We demonstrate an exemplar application of visible-light-controlled soft “fingers” playing a piano on a smartphone. The robustness of the PAzo elastomer and its scalability, in addition to its excellent biocompatibility, opens the door to the development of reproducible light-driven wearable/implantable actuators and lightweight soft robots for clinical applications.


Synthesis of PCL-PUU (PAzo without Azo content)
The syntheses of PAzo and PCL-PUU were adapted from previous work 1 .In brief, Polycaprolactone diol (14.48 g, 7.24 mmol) and 4,4′-Methylenebis(cyclohexyl isocyanate) (6.80 g, 25.92 mmol) were dissolved in 30 mL of anhydrous dimethylacetamide in a 250 mL three-neck flask under nitrogen.After the solution was degassed with nitrogen for 0.5 h, bismuth neodecanoate (0.26 g, 0.36 mmol) was added.The reactant was then heated to 80 ℃ for 6 h to form a solution of the prepolymer.After the mixture was cooled down to 40 ℃, ethylenediamine (1.12 g, 13.49 mmol) in 100 mL of dimethylacetamide was added dropwise under vigorous stirring for about 1 h.Upon completion of the polymerization, 1-butanol (1.00 g, 13.49 mmol) was added for reaction termination.After the polymer solution was then stirred for 1 more hour and cooled to room temperature before being transferred to a 500 ml plastic screw top bottle.The resulting polymer solution was poured into a mould and cured in a 60 °C vacuum oven for 24 h before further processing.

In vitro experiments
Cell proliferation and viability Mouse embryonic dermal fibroblasts (NIH/3T3 cells, ATCC) were cultured in growth medium, which consisted of high glucose (4.5 g/L) Dulbecco's modified Eagles medium (DMEM) supplemented with 10% foetal bovine serum (FBS) and 1% penicillin/streptomycin, and incubated at 37 ℃ in a humidified atmosphere containing 5% CO 2 .
PAzo film was cut into discs (10 mm diameter and 0.2 mm thickness, n = 8 per group) and sterilised with 5% v/v hydrogen peroxide for 2 hours, then washed three times with sterile Phosphate Buffered Saline (PBS) for 5 minutes each time, and air-dried in a class II laminar flow hood.Coverslips were used as the positive control and underwent the same sterilisation procedure.Under sterile conditions, PAzo discs and coverslips were placed in non-adherent sterile Falcon TM 48-well plates (Corning, 734-0956), and incubated with 500 µL of the growth medium for 24 h prior to cell seeding.
Scaffolds were seeded with six-passage (P6) cells at a density of 1 10 4 cells/scaffold in × 500 µL of cell growth medium in 48 well plates.Fresh medium was replenished every other day, and the metabolic activity was evaluated by the PrestoBlue TM assay (Thermo Fisher Scientific, USA) on day 1, 3, 7, and 10, following the manufacturer's instructions, and standardised by total DNA content.Total-DNA content was quantified at the same time points as above using fluorescent Hoechst 33258 stain (Sigma-Aldrich, UK) according to the manufacturer's instructions.To investigate the biocompatibility of PAzo, the embryonic mouse 3T3-J2 fibroblasts were seeded on the 2D PAzo discs and glass coverslips (as positive control), then evaluated cell viability and proliferation on day 1, 3 7, and 10.The cell morphology was observed to investigate the cell-substrate interaction via immunofluorescence confocal microscopy.The metabolic activities and proliferations of 3T3-J2 cells were quantified by PrestoBlue assay and Total DNA assay, respectively.In Figure S4a, the overall metabolic activity for the PAzo group dropped first due to the adaption of the new microenvironment in the initial days, and then increased significantly.As shown in Figure S4b, total DNA contents increased significantly in the first 7 days in both PAzo and coverslips, followed by a plateau due to cell confluence.Figure S4c demonstrates the metabolic activity per ng DNA on PAzo discs and control groups, where cells on both substrates showed the highest values on Day 1, then witnessed a moderate decrease at early time points and rebounded at the later time point (Day 10).The initially high metabolic activity could be attributed to the adaptation of the new microenvironment right after seeding, while the rebound at later time point could be relevant to the stress of over confluence.

Cell morphology imaging via confocal microscopy
In

Coefficient of Thermal Expansion (CTE) of PAzo determined by thermal contraction/expansion experiment
The thermal contraction/expansion test (Figure S7) was conducted using a DMA machine (DMA Q800, TA Instrument) following a reported method 3 .Briefly, a small stretching force of was applied to ensure the sample was straight.The temperature was then 1 × 10 -4  decreased from 80 °C to 10 °C at a rate of 2 °C min -1 and heated back to 80 °C at the same rate.
From Figure S7, the CTE of PAzo was determined to be within the 204.4 × 10 -6 ℃ -1 temperature range of approximately 12 °C to 32 °C, via linearly fitting the experimental data.This CET result was used for theoretical calculations and FEA simulations.The dynamic light-driven bending performance of the actuator were systematically characterised using the light intensity up to 4.38 W/cm 2 and cycle period of 4s (2s on, 2s off).
To evaluate the repeatability of the light-driven actuator, we performed cyclic bending tests on the actuator, as shown in Figure S11a and b.These tests indicate that the actuator bending behaviour and related temperature changes are stable and repeatable during 100-cycle light stimuli, except for the 1 st and 2 nd cycles.The initial variation of the bending angle and temperature of the actuator in the 1 st and 2 nd cycles of warm-up may be due to the relaxation of internal stress induced by thermal history during fabrication of the actuator.Figure S11c and d display the cyclic bending angle and related temperature changes plotted with cycle time at different light intensity irradiation.As expected, the actuator deflected higher bending angles as the light intensity increased.With the same setup at a constant light intensity of 4.38 W/cm 2 , we tested the deflection and temperature of the actuator at different cycle period (i.e., frequency) as shown in Figure S11e and f.It is interesting to see that the time-dependence of photo-driven dynamic behaviour, i.e., the longer cycle of light stimulates, the larger bending angle is deflected, which is reminiscent of the characteristic time-temperature dependence of polymer viscoelasticity.Overall, these four figures show the fully recoverable actuation with excellent controllability and stability properties of the light-driven actuator.

Theoretical analysis of bending angle of bilayer actuator upon light irradiation with low intensity
Although we have explained the preliminary mechanisms of these two bending behaviours, theoretical analysis can help us better understand the bending behaviour under low-intensity light irradiation.Under low light irradiation with a low light intensity of 0.73 W/cm 2 , the maximum temperature of the PAzo/Kap is around 30.4 ℃, which is lower than the melting temperature of 38 ℃ (Figure 2e).We make the first assumption that the photothermal effect of the bilayer dominates the bending behaviour under low light intensity irradiation.
Based on Timoshenko's theory of bending of a bimetal strip subjected to uniform heating 2 , the general equation for the curvature of the PAzo/Kap bilayer upon light irradiation (Figure S14a) is where ρ is the radius of curvature; and are the coefficients of expansion of the Kapton  1  2 and PAzo, respectively; and are the original and given temperatures, respectively; is the  0  ℎ thickness of the bilayer strip; is the thickness ratio of Kapton layer ( ) and PAzo layer ( ); is the Young's modulus ratio. There are two more assumptions: 1) bending happens in the whole strip after light irradiation; 2) the strip is subjected to uniform heating when reaching the highest temperature.
where L is the arc length of the (Here, L = 10 mm). Combining equations ( 1)-( 5), we obtained: Hence, if , m and n remain constant, the bending angle can only be regulated by  2 ,  1 (the change of temperature before and after light illumination).(Figure S15 shows the ∆  storage modulus of PAzo at various temperatures.As the temperature increases from 20 ℃ to 40 ℃, the storage modulus of PAzo exhibits a slight decrease, from 53 MPa to 37 MPa.Therefore, we assume that the modulus of PAzo remains relatively constant for theoretical calculations under low-intensity light illumination.)Substituting all parameters (in Table S4) into equation (6), when under light irradiation with low intensity ( ), .The ∆  = 9 ℃  ≈ 6.14°t heoretical formula of the bending angle at different increased temperature is plotted in Figure S14b.
It is of note that the theoretically calculated angle is less than that by experiments (~ 20°), which indicates that photothermal mechanism is not the predominant contributor to the bending behaviour upon low intensity light irradiation.In other words, it is speculated that photochemical mechanism or the azobenzenes paly an essential role in the resulted bending angle.

Energy conversion of PAzo/Kap bilayer actuators
The photothermal conversion efficiency (defined as the generated elastic energy accumulated during deformation divided by the incident light energy) of PAzo/Kap bilayer actuator was investigated, since energy loss must be considered when it turns into mechanical deflection.Follow the reported the calculation method 4 based on the bimetallic strip thermostat model. 2 The elastic energy stored in the actuator is defined by equation (7), where are the strain and stress during bending performance, which are depicted as      follow (according to the reported model 5 ): ) [E H (3H The subscripts and represented the PAzo layer and Kapton layer, respectively.refers to    Young's modulus, is the thickness and is bending curvature: The elastic energy of the bilayer actuator: where is the surface area of the bilayer, is the thickness of the bilayer.

𝑆 𝑠𝑎𝑚𝑝𝑙𝑒 𝐻
Light input energy was calculated by: where is the light intensity and is the light illumination area, t is the actuation   ℎ   ℎ time.
Finally, the energy conversion efficiency ( ) is the ratio between elastic energy and light  input energy: We calculated the energy conversion efficiency of the PAzo/Kap actuator under low intensity (0.73 W/cm 2 ) 470 nm light irradiationto be around 0.02%, which is comparable to the reported efficiency of other light-driven bilayer actuators 4,6 .When the light is on, the actuator will bend, and its tip will touch the salt drop, removing ions and producing the note sound from the phone.Our "photo-finger" is not conductive and only serves the bending function to help the conductive wire touch the iPhone screen.

Storage Modulus of
The morphology of 3T3 fibroblast was imaged utilising Leica TSC SP8 upright confocal microscope (Leica, Germany).Briefly, cellladen scaffolds (n = 2 per group) were harvested at each day point, fixed with 4% paraformaldehyde (PFA; Sigma-Aldrich, UK) in PBS for 30 min at room temperature, and rinsed with PBS twice.Then, scaffolds were permeabilised with 0.1% Triton-X 100 (Sigma-Aldrich, UK) for 5 min and rinsed with PBS twice, then blocked with 1% bovine serum albumin (BSA) in PBS solution for 30 min.Following further rinsing, cells were stained with Alexa Fluor TM 488 Phalloidin (Thermo Fisher Scientific, USA) overnight for F-actin and DAPI solution (Thermo Fisher Scientific, USA) 15 min for nuclei and then kept in PBS before imaging.When imaging, samples were taken out from PBS and placed on a slide, covered with a glass coverslip to ensure flatness.Images were taken using the Leica TCS SP8 upright confocal microscope, with LASX software, 10 and 20 objective dry lens.To avoid × × sample dehydration, imaging process was limited to 5 min for each sample.

Figure S1 .
Figure S1.Optical image of PCL-PUU shown in the blue dashed box.

Figure S2 .
Figure S2.TGA curves of PAzo (a) Weight plotted with temperature; (b) Derivative weight plotted with temperature.
Figure S5.UV-Vis absorption spectra (deduct the background of DMAc solvent) (a) Recovery in the absorption of PAzo solution after the irradiation of 470 nm (low intensity) for 60 s; (b) Changes in the absorption of PAzo solution upon the irradiation with 470 nm light (0.73 )   2 for 20 s then 365 nm UV light (0.73 ) for different time; (c) Changes in the absorption   2 of PAzo solution upon the irradiation with 365 nm light (0.73 ) for different time period;   2 (d) absorption of PCL-PUU (PCLU) under three conditions i.e., without light irradiation, 365 nm light irradiation for 20 s and 470 nm light irradiation for 20 s.

Figure
Figure S6.X-ray diffraction spectra of (a,b) Extra Small Angle X-ray Scattering (ESAXS) and (c) Middle Angle X-ray Scattering (MAXS) of PAzo films before and after light irradiation treatment.
Figure S7.The experimental results of thermal contraction/expansion test for PAzo.(For determining CTE of PAzo)

Figure S12 .
Figure S12.Actuation stress analysis of the PAzo/Kap bilayer (length of 10 mm, width of 2 mm, no specific statement, thickness: 0.16 mm, cycle period: 40s, light intensity: 4.38 W/cm 2 ) under the 470 nm light irradiation at different states.(a) Actuation stress of the bilayer as a function of cycle time and light intensity stimulated by periodic light for on (20s) and off (20s) (1% pre-strain); (b) Actuation stress as a function of irradiation cycle number and cycle period (frequency) stimulated by periodic light (4.38 W/cm 2 ); (c) Actuation stress as a function of cycle time and the thickness of PAzo/Kap bilayer actuator stimulated by period light (4.38 W/cm 2 ) for on (20 s) and off (20 s) (thickness of Kapton is kept at constant of 0.05 mm); (d) Actuation stress decreases with increasing the thickness of PAzo/Kap bilayer.

Figure
Figure S14.(a) Schematic diagram of double layer actuator in theoretical model; (b) Theoretical formula of bending angle plotted with the change of temperature for PAzo/Kap under 470nm light irradiation with low light intensity.

Figure S15 .
Figure S15.Storage Modulus of PAzo plotted with temperature.

Table S1 .
Comparisons of light responsive materials reported in literatures.

Table S4 .
Material properties of PAzo and Kapton for theoretical calculations