Introduction
Shape-memory materials are intelligent materials that can transition from a temporary state to a primitive state under external stimuli. Compared with shape-memory materials such as hydrogels, liquid crystal elastomers, shape-memory alloys, and shape-memory ceramics, shape-memory polymers and their composites (SMPs/SMPCs) have the advantages of mechanical robustness, large deformability, low cost, easy preparation, and multiple stimulus actuation (Xiao Kuang et al., Reference Xiao Kuang, Wu, Hamel, Ding, Wang, Dunn and Qi2018). Stimulations such as light (Lendlein et al., Reference Lendlein, Jiang, Jünger and Langer2005; Cortés et al., Reference Cortés, Cosola, Sangermano, Campo, González Prolongo, Pirri, Jiménez-Suárez and Chiappone2021; Liang et al., Reference Liang, Yu, Wang, Wang and Amin2021), heat (Ge et al. 2013, 2014, Reference Ge, Sakhaei, Lee, Dunn, Fang and Dunn2016), electricity (Zarek et al., Reference Zarek, Layani, Cooperstein, Sachyani, Cohn and Magdassi2016; Zhang et al., Reference Zhang, Li, Cheng, Ye, Sakhaei, Yuan, Rao, Zhang, Chen, Wang, He, Liu, Xiao, Qu and Ge2021a, Reference Zhang, Li, Li, Li, Xiong, Zhu, Lan and Ge2021b; Wang et al., Reference Wang, Zhao, Li, Dong, Zhang, Zou and Gu2023; Zhang et al., Reference Zhang, Li, Long, Xiong and Ge2023), magnetism (Kim et al., Reference Kim, Yuk, Zhao, Chester and Zhao2018; Ma et al., Reference Ma, Wu, Ze, Kuang, Zhang, Qi and Zhao2020; Ze et al., Reference Ze, Kuang, Wu, Wong, Montgomery, Zhang, Kovitz, Yang, Qi and Zhao2020; Hu et al., Reference Hu, Yasa, Ren, Goudu, Ceylan, Hu and Sitti2021), pH (Han et al., Reference Han, Dong, Fan, Liu, li, Wang, Yuan, Li and Zhang2012), and humidity (Sessini et al., Reference Sessini, Arrieta, Fernandez-Torres and Peponi2018) can all be used to drive SMPs/SMPCs. The controllability of external stimulus such as heat/magnetic field/PH and humidity are poor, making it difficult to achieve quantitative driving deformation at specific locations. Dielectric elastomers cannot achieve specific position-driven deformation (Wang et al., Reference Wang, Zhao, Li, Dong, Zhang, Zou and Gu2023). Embedded conductors in the SMP structure can achieve electric heating at specific positions, but the driving position cannot be changed (Zarek et al., Reference Zarek, Layani, Cooperstein, Sachyani, Cohn and Magdassi2016; Zhang et al., Reference Zhang, Li, Cheng, Ye, Sakhaei, Yuan, Rao, Zhang, Chen, Wang, He, Liu, Xiao, Qu and Ge2021a; Zhang et al., Reference Zhang, Li, Long, Xiong and Ge2023). Light stimulation can realise the remote control of the deformation of light-responsive SMPC structure at a specific position (Liang et al., Reference Liang, Yu, Wang, Wang and Amin2021).
According to the deformation mechanism, light-responsive SMPC can be divided into photochemical type and photothermal type. Photochemical SMPC usually uses UV light or visible light to activate photochemical groups to form new chemical crosslinks, which leads to macro deformation of materials (Jin et al., Reference Jin, Song, Jiang, Song, Zhao and Xie2018; Peng et al., Reference Peng, Zhang, Zhao and Xie2021). This process is very slow (usually takes more than 1 hour), and the temporary deformation is unstable. Photothermal SMPC usually converts the energy of infrared light or sunlight into heat to drive the shape change (Leng et al., Reference Leng, Lan, Liu and Du2011; Wang et al., Reference Wang, Liu and Leng2016; Wang et al., Reference Wang, Wang, Wei and Zhang2022). Photothermal SMPC usually has excellent shape fixation rate, shape recovery rate, and shape recovery speed, so it has been widely used in aerospace (Li et al., Reference Li, Liu, Lan, Pan, Liu, Leng and Xie2019b), medical devices (Xie et al., Reference Xie, Shao, Ma, Wang, Huang, Yang, Wang, Ruan, Luo, Wang, Chu and Yu2018; Chu et al., Reference Chu, Xiang, Wang, Xie, Xiang and Zhou2020; Yang et al., Reference Yang, Zheng, Younis, Shao, Fu, Zhang, Lin, Li and Huang2021), soft robots (Ji et al., Reference Ji, Jiang, Chang and Sun2014; Zhang et al., Reference Zhang, Yu, Wang, Zarrouk, Seo, Cheng, Buchan, Takei, Zhao, Ager, Zhang, Hettick, Hersam, Pisano, Fearing and Javey2014; Ding et al., Reference Ding, Yuan, Peng, Wang, Qi and Dunn2017; Toncheva et al., Reference Toncheva, Khelifa, Paint, Voue, Lambert, Dubois and Raquez2018; Xu et al., Reference Xu, Ding, Wei, Bao, Ke, Liu, Yang and Yang2019; Wang et al., Reference Wang, Zhao, Weng, Wang, Liu, Han and Zhang2021; Shan et al., Reference Shan, Chen, Tan, Cao, Fang, Lu and Xu2022), and self-healing SM structures (Li et al., Reference Li, Li, Wang, Li and Sun2019a; Du et al., Reference Du, Jin, Shi, Shen, Lai and Zhou2020; Yan et al., Reference Yan, Li, Wang, Chen, Ma and Yang2020).
Common photothermal conversion materials include metal nanoparticles (Toncheva et al., Reference Toncheva, Khelifa, Paint, Voue, Lambert, Dubois and Raquez2018; Liang et al., Reference Liang, Yu, Wang, Wang and Amin2021; Yang et al., Reference Yang, Zheng, Younis, Shao, Fu, Zhang, Lin, Li and Huang2021), carbon-based materials (Ji et al., Reference Ji, Jiang, Chang and Sun2014; Zhang et al., Reference Zhang, Yu, Wang, Zarrouk, Seo, Cheng, Buchan, Takei, Zhao, Ager, Zhang, Hettick, Hersam, Pisano, Fearing and Javey2014; Ding et al., Reference Ding, Yuan, Peng, Wang, Qi and Dunn2017; Li et al., Reference Li, Liu, Lan, Pan, Liu, Leng and Xie2019b; Wang et al., Reference Wang, Zhao, Weng, Wang, Liu, Han and Zhang2021), dyes (Fang et al., Reference Fang, Fang, Liu, Chen, Lu and Xu2016), rare earths (Shan et al., Reference Shan, Chen, Tan, Cao, Fang, Lu and Xu2022), and so forth. Carbon-based nanomaterials, that is, graphene (Ji et al., Reference Ji, Jiang, Chang and Sun2014)/carbon nanotubes (CNTs; Zhang et al., Reference Zhang, Yu, Wang, Zarrouk, Seo, Cheng, Buchan, Takei, Zhao, Ager, Zhang, Hettick, Hersam, Pisano, Fearing and Javey2014; Ding et al., Reference Ding, Yuan, Peng, Wang, Qi and Dunn2017)/carbon black (Wang et al., Reference Wang, Zhao, Weng, Wang, Liu, Han and Zhang2021, are the first choice of photothermal conversion materials due to their excellent photothermal conversion efficiency and good dispersion. CNTs are attractive for light-responsive SMPCs due to their excellent infrared light absorption and thermal conductivity (1,000–6,000 W/mK) (Kim et al., Reference Kim, Shi, Majumdar and McEuen2001; Liao et al., Reference Liao, Liu, Liu, Deng and Yang2015; Han et al., Reference Han, Zhang, Chen and Sun2018). Numerous studies have mixed CNT with shape-memory polymer matrix, such as thermoplastics (Zhang et al., Reference Zhang, Yu, Wang, Zarrouk, Seo, Cheng, Buchan, Takei, Zhao, Ager, Zhang, Hettick, Hersam, Pisano, Fearing and Javey2014; Xu et al., Reference Xu, Ding, Wei, Bao, Ke, Liu, Yang and Yang2019), PDMS elastomer (Li et al., Reference Li, Hou, Yin and Jiang2018), and epoxy-based SMP, to prepare near-infrared activated SMPCs. However, the manufacturing methods of most matrix materials are injection moulding (Li et al., Reference Li, Hou, Yin and Jiang2018), hot press moulding (Xu et al., Reference Xu, Ding, Wei, Bao, Ke, Liu, Yang and Yang2019; Li et al., Reference Li, Liu, Lan, Pan, Liu, Leng and Xie2019a; Du et al., Reference Du, Jin, Shi, Shen, Lai and Zhou2020), and vacuum filtration (Ding et al., Reference Ding, Yuan, Peng, Wang, Qi and Dunn2017; Zhang et al., Reference Zhang, Yu, Wang, Zarrouk, Seo, Cheng, Buchan, Takei, Zhao, Ager, Zhang, Hettick, Hersam, Pisano, Fearing and Javey2014), resulting in light-responsive SMPC structures being thin sheets. The combination of shape-memory polymer precursor solutions containing acrylic functional groups and 3D printing technology can produce structurally complex and functionally diverse light-responsive intelligent structures.
Here, we prepared a UV curable infrared responsive SMP–CNT composite material. Two monomers (IBoA, BA) and one crosslinker (aliphatic urethane diacrylate [AUD]) with acrylic functional groups form the UV curable SMP matrix (IBBA). This novel light-responsive SMPC (IBBA–CNT) was prepared by fully mixing IBBA with CNT. The thermal mechanical performance experiment, shape-memory effect experiment, and photothermal conversion experiment show that IBBA–CNT has good mechanical properties, shape-memory performance, and photothermal response performance. Selective near-infrared laser activation of specific regions of IBBA–CNT sheets can result in different folding structures. We explored the photothermal driving behaviour of near-infrared laser in the depth direction of light-responsive SMPC using 3D-printed IBBA–CNT cube. Based on the photothermal driving effect in the depth direction, we demonstrated the advantages of selective near-infrared light-driven three-dimensional SMP structures through a 3D-printed multi-material movable type.
Results and discussion
Figure 1a presents the chemical structures used to prepare the IBBA precursor solution which consists of 50 wt.% isobornyl acrylate (IBoA) and 30 wt.% benzyl acrylate (BA) as linear chain builder, 20 wt.% AUD as crosslinker. After the mixture of monomers and crosslinker is stirred and mixed evenly, add 2 wt.% diphenyl(2,4,6-trimethylbenzoly) phosphine oxide (TPO) as a photoinitiator. Light-responsive SMPC was prepared by adding 0.05 wt.% CNTs into IBBA solution and mixing them through ultrasound. After 30 min of ultrasonic mixing, the CNTs in IBBA–CNTs were uniformly dispersed and remained stationary for 12 h without significant precipitation. The addition of 0.05 wt.% CNTs has no significant effect on the rheological properties of IBBA precursor solution, and the viscosities are lower than 0.4 Pa·s before and after the addition of CNTs (Figure 1b, Supplementary Figure S1). The reduction of CNTs reduces the UV curing efficiency of IBBA precursor solution. The curing time of a 100-μm-thick IBBA–CNT layer is 11.6 s (UV light with an energy density of 8 mW/cm2 and a wavelength of 385 nm, curing time: tc = tgel − ts ), which is much longer than the time required to cure an IBBA layer of the same thickness (Figure 1c).
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240824023312487-0868:S2752807324000046:S2752807324000046_fig1.png?pub-status=live)
Figure 1. The thermal mechanical properties and precursor rheological properties of IBBA–CNT. (a) Chemical composition and UV curing products of shape-memory polymer precursors. (b) The rheological properties of the precursors. (c) Photorheological properties of the precursors. (d,e) Dynamic thermodynamic performance test results. (f) Quasi-static tensile test results at room temperature. (g) Quasi-static tensile test results at programmed temperature.
We printed IBBA dog-bone samples and IBBA–CNT dog-bone samples to explore the effect of CNTs on the mechanical properties of IBBA (Supplementary Figure S1). DMA Q850 was used to measure their thermal mechanical properties (Figure 1d,e). The addition of 0.05 wt.% CNT reduced the glass modulus (storage modulus at room temperature) of IBBA and increased the rubbery modulus (storage modulus above the glass transition temperature) of IBBA. The addition of CNT also reduced the glass transition temperature (Tg ) of the material (62°C to 58°C). During the DMA testing process, the temperature rises uniformly, and IBBA–CNT with higher thermal conductivity absorbs heat faster, resulting in a lower measured glass transition temperature. The quasi-static tensile tests shown that the addition of CNT reduces the modulus of the material at room temperature (388 to 293 MPa; Figure 1f) and increases its modulus at programming temperature (Tg + 25°C) (0.20 to 0.26 MPa; Figure 1g), but has no significant effect on the elongation at break of the materials (450.1% to 413.6%; Figure 1g). The test results of thermal mechanical properties show that the addition of CNT has a significant effect on the strength of the material, but the material still has a high modulus at room temperature and maintains a large deformation capacity at programming temperature.
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240824023312487-0868:S2752807324000046:S2752807324000046_fig2.png?pub-status=live)
Figure 2. Shape-memory programming of IBBA–CNT through near-infrared photothermal effects. (a) Shape-memory cycle experiment of IBBA–CNT. (b) Near-infrared photothermal effect testing equipment. (c) Infrared camera window. (d) Experimental results of near-infrared photothermal effect. (e) Selective near-infrared laser programming for IBBA–CNT sheet.
The shape-memory experiment and near-infrared laser thermal effect experiment demonstrate that IBBA–CNT is an excellent light-responsive SMPC. Figure 2a shows two shape-memory cycles of IBBA–CNT. The shape fixation rates in two shape-memory cycles are all 99.9%. The shape recovery rate of the first cycle is 89.8%, and after training in the first cycle, the shape recovery rate of the second cycle is 99.8%. The testing device for the surface photothermal effect of IBBA–CNT is shown in Figure 2b. The infrared camera records the temperature changes of IBBA–CNT sheet irradiated by near-infrared laser. Figure 2c shows the screen of an infrared camera, which can only record the temperature changes of the IBBA–CNT sheet without recording the surrounding environment by adjusting the temperature measurement range (white rectangular box in the centre of the screen). The surface temperature variation curve of IBBA–CNT sheet (Figure 2d) was plotted based on the infrared camera video during the experimental process. Due to the excellent photothermal effect of CNT, near-infrared laser can rapidly heat IBBA–CNT, reaching the programming temperature (88°C) in just 5 s and 150°C in just 29 s. After reaching 150°C, turn off the laser, and the IBBA–CNT sheet rapidly cools to the programming temperature (6 s), and then slowly cools to room temperature. To achieve programming of IBBA–CNT sheets, it is necessary to irradiate at a specific position for more than 5 s and quickly complete the programming operation within 6 s after turning off the laser.
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240824023312487-0868:S2752807324000046:S2752807324000046_fig3.png?pub-status=live)
Figure 3. Near-infrared laser programming for IBBA–CNT three-dimensional pillar. (a) Experimental schematic diagram of the effective thermal response depth of near-infrared laser on IBBA–CNT. (b) Design drawing of a three-dimensional multi-material pillar. (c) Compression experiments of three-dimensional multi-material pillars with different heights at programming temperature. (d) Snapshots of 50% compressive strain of three-dimensional multi-material pillars. (e) Near-infrared laser programming for 3D multi-material pillar.
We demonstrated the selective region activation and programming of the IBBA–CNT sheet by near-infrared laser, as well as the selective region activation and recovery (Figure 2e). After selectively irradiating the centre of the sheet, both bending and twisting deformations can be completed (Figure 2e(i,ii)). After selectively irradiating multiple areas, bending deformation can be achieved at multiple locations in the sheet to form an ‘M’ shape (Figure 2e(iii)), and controllable sequential recovery can be achieved through selective near-infrared laser irradiation (in Supplementary Video S1, the left bend of the ‘M’-shaped sheet is first unfolded, then the right bend is unfolded, and finally the middle bend is unfolded).
Further, 0.05 wt.% CNT not only endows IBBA shape-memory polymers with the ability to respond to near-infrared laser, but also enables near-infrared laser transmission to a certain depth in the material. This will endow the material with near-infrared laser response capability in the depth direction, which has not been addressed in previous work. We tested the near-infrared photothermal depth influence range of IBBA–CNT (0.05 wt.% CNT) using the method shown in Figure 3a. As shown in Figure 3a, a near-infrared laser is irradiated on the edge of the right side of the 3D-printed IBBA–CNT cube, and the infrared camera records the temperature changes on the front face of the cube. When the maximum temperature reaches 150°C, turn off the near-infrared laser and record the infrared camera image at this moment. Based on the temperature scale and the size of the cube, we measured the size of the area where the temperature on the front face of the cube exceeded 88°C. This area is called an effectively programmable heat affected zone, with a length of approximately 8 mm and a width exceeding 8 mm. This provides design parameters for the design of three-dimensional IBBA–CNT structures.
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240824023312487-0868:S2752807324000046:S2752807324000046_fig4.png?pub-status=live)
Figure 4. Selective near-infrared laser programming for IBBA–CNT shape-memory movable type plate 4D printing. (a) Schematic diagram of selective infrared laser programming movable type plate. (b) Digital model of multi-material movable type plate. (c) DLP 3D-printed model of multi-material movable type plate. (d) Movable type plate in compressed state. (e) Movable type ‘E’ obtained by selective near-infrared laser programming.
As shown in Figure 3b, we design three-dimensional multi-material pillars with 3D-printed transparent photocurable resin material at both ends (cubes with a side length of 4 mm) and 3D-printed IBBA–CNT cylinder in the middle (diameter: d = 3 mm, heights: h = 4, 5, 6, 7, or 8 mm). Figure 3c depicts the force displacement curve of compressing three-dimensional multi-material pillars with different heights to 50% strain at a strain rate of 0.01 s−1 at programming temperature. Under 50% compressive strain, the 4- and 5-mm-high IBBA–CNT cylinders showed significant barrelling deformation and creasing, and the 7- and 8-mm-high cylinders showed significant buckling and creasing. However, the 6-mm column has barrelling deformation and small creases under 50% compressive strain (Figure 3d). Therefore, a diameter of 3 mm and a height of 6 mm are ideal designs for IBBA–CNT cylinders. Figure 3e illustrates the shape-memory experiment of IBBA–CNT cylinder with h = 6 mm. Under the photothermal effect of near-infrared laser, the IBBA–CNT cylinder is compressed and programmed. After turning off the laser, the material quickly cools and the compressed shape is fixed. The IBBA–CNT cylinder freely recovers to its initial height by turning on the near-infrared laser again.
Multiple multi-material 3D pillars are combined to form a movable type plate, and different printing movable types can be obtained through selective infrared laser programming. The schematic diagram of the selective infrared laser programming movable type board is shown in Figure 4a. The side view of the multi-material 3D-printed movable type plate is shown in Figure 4a(i). After heating to the programming temperature, the height of the IBBA–CNT cylinders decreases after external force compression (Figure 4a(ii)). Maintain the compressed state and cool to room temperature, and the movable type plate is locked in a flat shape (Figure 4a(iii)). Near-infrared laser can penetrate the transparent resin at both ends of the pillars and directly irradiate the IBBA–CNT cylinders. The IBBA–CNT cylinders activated by photothermal activation can be freely restored to their original height (Figure 4a(iv)). Figure 4b,c shows the digital model and 3D printing model of movable type plate composed of five rows and five columns of multi-material pillars, respectively. Figure 4d shows a movable type plate programmed to a compressed flat state. Selective near-infrared laser irradiation of the pillars can decode the flat-shaped movable type plate to form the required upward protruding movable type (Figure 4e and Supplementary Video S2).
Conclusion
We have developed a near-infrared responsive SMPC material, IBBA–CNT, with high mechanical strength, strong deformation ability, and UV curability. It can be used to produce near-infrared-driven 4D printing structures through digital light processing. IBBA–CNT containing 0.05 wt.% CNT has good shape-memory performance and photothermal effect: shape fixation rate is 99.9%, shape recovery rate is 99.8%, and programming temperature can be reached after 5 s of irradiation with a 250-mW 808-nm near-infrared laser. We demonstrated the selective near-infrared laser programming of IBBA–CNT sheets into various origami and their recovery process in a controllable sequence. More importantly, for the first time, we explored the photothermal effect of near-infrared laser in the depth direction of light-responsive SMP. We printed IBBA–CNT multi-material pillars and conducted experiments to demonstrate the feasibility of near-infrared laser programming in the depth direction. Finally, we designed a movable type board to demonstrate the application of selective near-infrared laser programming in deep direction activation of light-responsive SMPC. It can be predicted that there will be more research on the design and application of three-dimensional light-responsive SMPC structure in the future.
Materials and methods
Materials
Isobornyl arcylate (IBoA), Benzyl acrylate (BA), diphenyl(2,4,6-trimethylbenzoly) phosphine oxide (TPO) were purchased from Sigma-Aldrich (Shanghai, China). CNTs were purchased from XFNANO (Nanjing, China). Ebecryl 8413 (AUD) was kindly provided by Allnex (Frankfurt am Main, Germany).
Rheological test
The viscosity (η) of IBBA and IBBA–CNT precursors were measured by using a controlled-stress rheometer (DHR2, TA Instruments, Inc., Elstree, UK) with an aluminium plate geometry (diameter 25 mm, gap 100 μm).
Photorheological test
The storage modulus and loss modulus of materials were measured on a DHR2 machine with an aluminium plate geometry (diameter 20 mm, gap 100 μm). First, 20 s were detected without light, then 20 s were exposed in 385-nm UV light with 8-mW/cm2 light intensity, and more 20 s were detected after the end of exposure. Aluminium plate rotated at a speed of 5 rad s−1 throughout the 60-s detection process. The intersection of the loss modulus and storage modulus curves is the gel point, and the corresponding time minus 20 s is the curing time.
Dynamic mechanical analysis experiments
Samples with dimensions of 10 mm × 5 mm × 1 mm were tested at a frequency of 1 Hz and an amplitude of 10 μm using a DMA analyser (Q850 DMA, TA Instruments). The temperature was first equilibrated at −20°C for 3 min, and then gradually increased to 100°C at a heating rate of 3°C/min. The glass transition temperatures (Tg ) were assigned as the temperature at which tanδ value was maximum.
Uniaxial tensile experiments
Tension experiments on dog-bone samples with a gauge length of 20 mm and a cross section of 5 mm × 2 mm were conducted using MTS machine at a strain rate of 0.01 s−1.
Shape-memory behaviour tests
Figure 2a presents the result from typical shape-memory cyclic tests for calculating shape fixation ratio (Rf = εu /εp ) and shape recovery ratio (Rr = (εu − εr )/εu ). First, an IBBA–CNT sample is stretched to 100% at a constant strain rate (0.001 s−1) at Tg + 25°C (88°C). Second, the sample is cooled to 25°C (−2.5°C/min) and held 2 min while it is kept stretched. Third, the external load is suddenly released at 25°C, and the temporary fixed strain εu can be measured. Last, the sample is heated to Tg + 25°C (2.5°C/min) and held at Tg + 25°C for 1 h where the recovery strain εr is measured.
3D printing
A self-assembled multi-materials DLP printer (Cheng et al., Reference Cheng, Wang, Sun, Liu, He, Li, Ye, Yang, Wei, Li, Jian, Deng and Ge2022) was used to print multi-materials structures. The slice thickness of IBBA–CNT layers is 100 μm, and the exposure time of each layer is 12 s (exposure intensity 8 mW/cm2).
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/pma.2024.4.
Data availability statement
The data that support the findings of this study are available from the first author and the corresponding author upon reasonable request.
Funding statement
This work was supported by the National Natural Science Foundation of China (H.L.; Grant No. 52305312), the Natural Science Foundation of Guangdong Province (H.L.; Grant No. 2022A1515010047), the Science, Technology Innovation Program of the Ministry of Trade, Industry & Energy (MOTIE, Korea) (Q.G.; Grant No. 20009618), and the Technology and Innovation Commission of Shenzhen Municipality (Q.G.; Grant No. ZDSYS20210623092005017).
Competing interest
The authors declare none.
Author contributions
Q.G. and H.L. conceived and designed the study. H.L. and Z.C. conducted experiments with assistance from S.Y., B.J. and H.Y. H.L. and Z.C. conducted data analyses. H.L., Z.C. and Q.G. wrote the article.