Hot Fingers: Individually Addressable Graphene-Heater Actuated Liquid Crystal Grippers

Liquid crystal-based actuators are receiving increased attention for their applications in wearables and biomedical or surgical devices, with selective actuation of individual parts/fingers still being in its infancy. This work presents the design and realization of two gripper devices with four individually addressable liquid-crystal network (LCN) actuators thermally driven via printed graphene-based heating elements. The resistive heat causes the all-organic actuator to bend due to anisotropic volume expansions of the splay-aligned sample. A heat transfer model that includes all relevant interfaces is presented and verified via thermal imaging, which provides good estimates of dimensions, power production, and resistance required to reach the desired temperature for actuation while maintaining safe electrical potentials. The LCN films displace up to 11 mm with a bending force of 1.10 mN upon application of 0–15 V potentials. The robustness of the LCN finger is confirmed by repetitive on/off switching for 500 cycles. Actuators are assembled into two prototypes able to grip and lift objects of small weights (70–100 mg) and perform complex actions by individually controlling one of the device’s fingers to grip an additional object. Selective actuation of parts in soft robotic devices will enable more complex motions and actions to be performed.


S2. Differential Scanning Calorimetry
Figure S2: DSC heating and cooling ramps of the LCN mixture.The heating rate is 10 °C min -1 and cooling rate 5 °C min -1 .Three heat ramps and two cooling ramps were performed with a 30 min break in between cooling and heating.The first cycle of heating and cooling has been excluded from the plot.The polarized optical microscopy (POM) results confirm that at temperatures beyond 85 °C (Figure S3a), the LC mix is in the isotropic phase.In Figure S3b, the change in phase from isotropic to nematic is presented by the appearance of colorful domains.These domains indicate an increase in order in the LC mixture at 82.5 °C, which is in line with the DSC results in Figure S2.To emulate the method used to prepare the LCNs, specifically the step in which the glass cell filled with LC mix is cooled down to 55 °C, the sample was also imaged at 55 °C (Figure S3c).The image shows that the sample remained in the nematic phase after cooling, as no crystallization was observed.Figure S3d shows the sample region after ten minutes of holding at 55 °C.This confirms that the LC mix remained in the nematic phase for the full duration of the UV treatment/polymerization step (180 s).

S5. Heat transfer model
The model input parameters are provided in Table S1

S8. Bend angle and bending strain estimation
The bending strain ε was approximated as  = ℎ/2, 2,3 with h the combined height of the substrate and the printed track and r the bending radius, which was measured with Tracker software as is shown in Figure S8.The bend angle could take on several values depending on the chosen bend radius due to the sample curvature varying along the substrate, most likely as a consequence of the graphene track not covering the full substrate length.This is illustrated in Figure S8d-e, where two different choices of bend radius (r = 2.0 vs r =1.3) for bending observed at 15 V yield bend angles of 142° versus 127°, giving strain levels of ε S ≈ 0.032 or ε S ≈ 0.048, respectively.It should be noted that the used strain equation might not be fully applicable to our situation as the printed layer was thicker than the substrate itself, nor did we account for the bilayer structure, which would require measuring the Young's moduli of both the tracks and the substrates. 2,3Therefore, the reported strain levels should only be used as a rough approximation.The annotated dataset is provided via the 4TU.ResearchData repository. 1

Figure S3 :
Figure S3: Optical microscopy images of the liquid crystal mixture in this work under crossed polarizers.a) After heating to 85 °C; b) after subsequent cooling to 82.5 °C; c) after further cooling to 55 °C; d) after 10 minutes holding at 55 °C.The scale bar in all panels is 250 µm.

Figure S4 :
Figure S4: Dynamic mechanical thermal analysis (DMTA) of a planar LCN actuator along the alignment direction,with the storage modulus on the left axis (black) and Tan δ (loss tangent) on the right axis (green).

Figure S5 :
Figure S5: Initial trial prints on PET substrate used to get an estimate of the sample resistance for the heat model (R ≈ 2000 Ω over a total path length of 26 mm).

Figure S6 :
Figure S6: Box plots of base widths, FWHM (left axis) and heights (right axis) of 35 printed graphene tracks on top of LCN substrates characterized with profilometry.

Figure S8 .
Figure S8.Displacement and bend angle measurements measured with Tracker software.a) Overlaid photographs of an LCN-G sample while flat and while bent with annotations of the substrate edge displacement with respect to the substrate origin before bending (0 V). b,d,e) Bend angle α and bend radius r (red squares) determination at 14 V (b) and 15 V (d, e).(d) and (e) illustrate two different choices of bend radius yielding different bend angles and hence different bending strain levels.c,f) Bend angle α and displacement versus voltage (c) and temperature (f), where at 15 V, the bend angle as determined in (b) and (d) is reported.

Table S1 .
, while the full model calculations are provided via the 4TU.ResearchData repository. 1 Model input parameters.The Excel sheet containing the full calculations is accessible via the 4TU.ResearchData repository. 1 Resistance values were obtained with a multimeter (N = 27), and track dimensions with profilometry (N = 35).