Multi-stimuli-responsive programmable biomimetic actuator

Untethered small actuators have various applications in multiple fields. However, existing small-scale actuators are very limited in their intractability with their surroundings, respond to only a single type of stimulus and are unable to achieve programmable structural changes under different stimuli. Here, we present a multiresponsive patternable actuator that can respond to humidity, temperature and light, via programmable structural changes. This capability is uniquely achieved by a fast and facile method that was used to fabricate a smart actuator with precise patterning on a graphene oxide film by hydrogel microstamping. The programmable actuator can mimic the claw of a hawk to grab a block, crawl like an inchworm, and twine around and grab the rachis of a flower based on their geometry. Similar to the large- and small-scale robots that are used to study locomotion mechanics, these small-scale actuators can be employed to study movement and biological and living organisms.


Supplementary Figure 2.
The relationship between the width of PDMS mold and PPy stripe. The average width differences between PDMS mold (100, 200 and 500 μm in width) and corresponding PPy stripes (n=10) are 6.7 μm, 6.5 μm and 8.1 μm respectively. The width of PPy stripes are presented as mean (SD).  Figure 4A showed the schematic diagram of precise modification of PANI onto the GO film. Similar to the modification of PPy, FeCl3 was transferred onto the GO film by a FeCl3-loaded agarose hydrogel stamp, and FeCl3 was present in only specific patterns replicated from the agarose hydrogel stamp. After that, the mixture of aniline and ethanol (V (aniline): V (ethanol) = 3:2) was dropped onto the GO film and reacted with FeCl3 to generate PANI with specific patterns on the GO film. Raman, FT-IR and XPS test were conducted to demonstrate the successful modification of PANI on the GO film. Supplementary Figure 4B exhibited the optical images of PANI patterns with different sizes and shapes on the GO film. We can see clearly that light black PANI patterns are introduced onto the GO film with high precision. Supplementary Figure 4C displayed the Raman spectra of GO and GO/PANI. The spectrum of GO has two dominant peaks at 1350 and 1593 cm -1 , corresponding to its D and G bands. Compared to the Raman spectrum of GO, some characteristic Raman bands of PANI appeared in the Raman spectrum of PANI. The Raman band at 443 cm −1 is associated with the phenazine-like segment.
The band related to out-of-plane C-H deformation of quinonoid ring appears at around 512 cm −1 . The 799 and 1160 cm −1 bands are assigned to the C-H bending in quinonoid ring and C-H bending deformation in the benzenoid ring. The 1509 cm −1 band is associated with C=N stretching vibrations of the quinonoid units. The appearance of the above bands suggests the formation of PANI on the surface of GO film 2-3 . In order to further confirm the formation of PANI on the surface of GO film, the FT-IR analysis was also employed to characterize the GO/PANI. Supplementary Figure 4D, a group of typical bands of PANI appeared 2, 4 , C=N stretching of the quinonoid ring and C=C stretching of the benzenoid ring at 1601 and 1494 cm -1 respectively, C-N stretching of secondary aromatic amines at 1286 cm -1 , C-H bendings of the benzenoid ring and the quinonoid ring at 1245 and 1124 cm -1 respectively, and C-C stretching of the quinonoid at 794 cm -1 . The XPS spectra revealed the surface element compositions of GO and GO/PANI, exhibiting bands at 281.2, 396.4 and 529.8 eV, corresponding to C1s, N1s, and O1s, respectively (Supplementary Figure 4E). The existence of nitrogen element for GO/PANI also demonstrate the formation of PANI on the GO film.

Discussion for Supplementary Figure 5:
Supplementary Figure 5A showed the schematic diagram of precise modification of PEDOT onto the GO film. Similar to the modification of PPy and PANI, FeCl3 was transferred onto the GO film by a FeCl3-loaded agarose hydrogel stamp, and FeCl3 was present in only specific patterns replicated from the agarose hydrogel stamp. After that, the mixture of EDOT and ethanol (V (EDOT): V (ethanol) = 4:1) was dropped onto the GO film and reacted with FeCl3 to generate PEDOT with specific patterns on the GO film. Raman, FT-IR and XPS test were conducted to demonstrate the successful modification of PEDOT on the GO film. Supplementary Figure 5B exhibited the optical images of PEDOT patterns with different sizes and shapes on the GO film. We can see clearly that black PEDOT patterns are introduced onto the GO film with high precision. Supplementary Figure 5C displayed the Raman spectra of GO and GO/PEDOT. The spectrum of GO has two dominant peaks at 1350 and 1593 cm -1 , corresponding to its D and G bands. Compared to the Raman spectrum of GO, some characteristic Raman bands of PEDOT appeared in the Raman spectrum of PEDOT. The Raman bands at 1430 and 1505 cm −1 are attributed to the symmetric and asymmetric stretching vibration of the C=C bond in PEDOT, respectively. The band at 1263cm −1 is assigned to the stretching modes of single C-C inter-ring bonds in PEDOT. Other weaker bands at around 990, 856 and 703 cm −1 are assigned to C-C asymmetric bond, C-H bending of 2, 3, 5-trisubstituted thiophene and C-S-C bond of PEDOT. The appearance of the above bands suggests the formation of PEDOT on the surface of GO film [5][6][7] . In order to further confirm the formation of PEDOT on the surface of GO film, the FT-IR analysis was also employed to characterize the GO/PEDOT. As shown in Supplementary Figure 5D Supplementary Figure 6A showed the schematic diagram of precise modification of calcium alginate onto the GO film. Similar to the modification of PPy, PANI and PEDOT, CaCl2 was transferred onto the GO film by a CaCl2-loaded agarose hydrogel stamp, and CaCl2 was present in only specific patterns replicated from the agarose hydrogel stamp. After that, 2% sodium alginate was dropped onto the GO film and reacted with CaCl2 to generate calcium alginate with specific patterns on the GO film.
Supplementary Figure 6B exhibited the optical images of calcium alginate patterns with different sizes and shapes on the GO film. We can see clearly that grey white calcium alginate patterns are introduced onto the GO film with high precision. Furthermore, the GO/PPy (e) possesses maximum response speed, which is slightly larger than GO/PPy (b). This is because the PPy thickness reduces with the decreasing microcontact time, so the stiffness of actuator reduces synchronously. However, GO/PPy (d) is smaller in response speed than GO/PPy (b), which is owing to the minor difference of water-adsorption ability between GO and PPy layer at the microcontact time of 10s. Therefore, the proper choice of GO thickness and microcontact time is important for the actuator.

Supplementary Note 5
Calculation of energy conversion efficiency: The energy conversion efficiency (η) of a bilayer actuator can be defined as the total elastic energy generated by the GO/PPy (b) actuator divided by the input laser energy (Q Laser). The elastic energy of our actuators can be calculated as follow 10 respectively. The Young's modulus of the GO layer is measured by an All-Electrodynamic Dynamic Test Instrument (Instron Model E1000, England). The Young's modulus of the PPy layer has been reported in many paper with the same value [12][13] .
The calculated total elastic energy of GO/PPy (b) actuator is 0.00137 J.
The actuating time (ԏ) is 3s. Aactuator represent the area exposed on the IR light. The input laser energy applied to the actuator can be expressed as: The calculated input laser energy of GO/PPy (b) actuator is 0.07479 J.
Hence, the energy conversion efficiency η is given by: The energy conversion efficiency η is then calculated to be 1.832 %.