Figure 3. Fabrication process of the gripper. a) Electroadhesive pad fabrication. A three-layer structure, which consists of a bottom silicone layer (i), a conductive silicone layer (ii), and a top silicone layer (iii) is blade-casted and cured in an oven one on top of the other. The electrode pattern (ii) is engraved using CO2 ablation before the top silicone layer (iii) is cast. b) i) The fabricated pad is placed in an empty mold. Then, liquid silicone is poured into the mold to form the silicone bag and simultaneously connect the bag and the pad in a monolithic structure. ii) The silicone is solidified by thermal curing and the gripper is removed from the mold. iii) The silicone bag is filled with coffee granules (in brown).

Grasping force characterization

The grasping performance of the gripper in GJ mode results from three different mechanisms: physical interlocking between the gripper and object, static friction caused by normal stresses in the contact regions, and suction effect due to a reduced air pressure over part of the contact surface (the latter can only occur when the gripper generates a vacuum seal against the object).[11]Static friction is the main factor because it contributes to both interlocking and suction modes.[5] Thus, the characterization of the grasping force of a GJ gripper is usually made for the static friction mechanism, which depends on the size and shape of the manipulated object, the contact angle, surface conditions, and applied force at the moment of grasp.[5,11]
Figure 4. Characterization of grasping performance for an object with hemispheric shape. Each data point corresponds to the average measurement of three grippers tested five times each (15 measurements). a) Grasping force for granular jamming and combined (granular jamming with electroadhesion) modes as a function of object size. Diameters of hemispheres used in these tests range from 33% to 89% of the gripper’s active area. The gripper was pushed down and contacted the hemisphere until it reaches 5N of applied force in all experiments. In the combined mode, the EA pad was activated by applying 2.5 kV and GJ was triggered by the air evacuation with the pressure drop of 84 KPa. Grasping forces increase with increasing object diameter, but the combined mode displays 15.2% superiority for largest object diameter compared to granular jamming alone. b) The measured total electroadhesion force of the gripper as a function of applied voltages on a flat surface. The normal force scales from 0 to 0.2 N with the voltage applied from 0 to 5 kV with a step size of 500V. c) The applied force as a function of a grasped object size. In experiments b and c, the gripper was pulled down to a certain absolute coordinate above all hemispheres. The applied force increases with the growth of the object size from 1.2 to 2.2 N. d) The contact angle as a function of a grasped object size. The contact angle changes from 73° to 30°.
In this study, the static friction mechanism in the combined mode is enhanced by EA, which increases the friction forces and normal forces of the gripper.[1,28] Thus, to quantify the grasping performance, we characterize the grasping force in GJ mode and in combined mode for different manipulated object sizes (Figure 4a ). The gripper is connected to the tensile testing machine equipped with a load cell to perform vertical displacement and applied force measurements on the grasped object. The objects to be grasped are fixed on a plate beneath the gripper and in line with its central axis. The objects consist of a set of hemispheres printed in polylactic acid (PLA) with the diameters from 12 to 32 mm that corresponds to 33 to 89% of the gripper diameter. The shape of the hemispheres does not allow interlocking grasp. The gripper is moved down to a preprogrammed height by the tensile machine and presses against the object (see Experimental Section). After that, the GJ or combined mode is activated, and the tensile machine lifts the gripper up until it releases the object. In the combined mode, the EA pad is activated by applying 2.5 kV and GJ is triggered by the air evacuation with a pressure drop of 84 KPa. Although the EA pad can operate up to 5 kV (Figure 4b), the applied voltage here is set to 2.5 kV to prevent possible breakdowns that may happen when squeezing the EA pad during the grasping operation. The grasping force data for three experiments is collected using the load cell of the tensile machine.
In the GJ mode, the grasping force varies from 5.8 (standard deviation (SD) = 0.38 N) to 10.6 N (SD = 0.62 N) for object sizes from 33 to 89% of the gripper diameter. Interestingly, the maximum grasping force is achieved for the hemisphere with the size of 78% of the gripper diameter and then starts decreasing. In the case of combined mode, the grasping force varies from 6.1 (SD = 0.18 N) to 12.1 N (SD = 0.83 N) for object sizes from 33 to 89% of the gripper diameter. In this case, there is a continuous increase in grasping force with the size of the hemisphere. For the hemisphere size of 89% of the gripper diameter, the grasping force of the combined mode is 15.2% larger than the GJ mode’s force.
In the GJ mode, the grasping force depends not only on the size of the object but also on the applied force with which a gripper touches the object. This applied force determines a contact angle between the gripper surface and the object. The contact angle and size of the object define the contact surface i.e., amount of granules in the contact zone between the gripper and the object. To characterize the relationship between the applied force and the contact angle, we use the same setup with the tensile machine, hemispheres, and an additional camera placed in front of the gripper to capture the contact angle. The gripper is lowered to a preprogrammed height by the tensile machine. After reaching the hemisphere, the applied force data is collected from the load cell.
The applied force increases from 1.2 (SD = 0.03 N) to 2.2 (SD = 0.61 N) N for the object size from 33 to 89% of the gripper diameter (Figure 4c). The applied force grows with an increase of the object size because larger object size results in the larger contact surface area across which more granules can cover the target. We noticed that an increase in the applied force correlates with a decrease in the contact angle. The lower contact angle shows that a larger area of the hemisphere is in the contact with the gripper causing more granules to flow towards the surface of the hemisphere, press onto it resulting in a higher applied force. The contact angle decreases from 73° (SD = 9.6°) to 30° (SD = 4.9°) for the object size from 33 to 89% of the gripper diameter (Figure 4d).
The EA mode allows grasping of flat objects, which is not possible in GJ mode, by means of the normal force generated between the gripper and the object when voltage is applied. The amount of normal force generated by the interdigitated electrode structure depends on the design of electrodes, dielectric constants of contact materials, thickness of the dielectric layer, and effective area of contact. It also scales with the square of the applied voltage.[1,29] Other physical parameters, such as surface conditions and roughness can also influence the normal force.[18] Thus, we performed experimental characterization of the EA generated force (Figure 4b). The tensile machine lowers the gripper onto the flat surface and a voltage is applied from 0 to 5 kV with a step increment of 0.5 kV. The measured normal force scales with the voltage as reported in the literature.[1,28,30] The inherent stiction between the outer dielectric layer and the flat surface causes high standard deviation of normal force measurements at zero voltage. Normal force reaches 0.2 N (SD = 0.02 N) at 5kV. Normal force can also be enhanced either by decreasing the outer dielectric layer or increasing the density of generated fringe fields by decreasing the width of the electrodes and the gaps between electrodes.
The grasping force of conventional GJ grippers varies up to an order of magnitude depending on the shape of object.[5]Thus, we characterized the grasping force generated by the proposed gripper in GJ mode for different object geometries, including flat and deformable objects (Figure 5a ). For these experiments, we used seven different 3D shapes with the cross-sectional diameter of 80-90% of the gripper’s diameter, which corresponds to maximum grasping force (Figure 4a) consisting of a sphere, cube, pyramid, disks with 20 and 50 mm in a diameter, cylinder, and parallelepiped. All shapes were 3D printed with the highest layer thickness of 0.4 mm (instead of a usual 0.06 mm). The objects were fixed at the bottom of the tensile machine, while the gripper grasped and pulled the object up and measures the holding force. We tested three grippers three times for each shape. We observed the highest grasping forces of 6.1, 10.4, and 8.9 N in both modes for a sphere, cube, and disk of 20 mm. The combined mode increases the adhesion for these objects by 5.6%, 3.7%, and 0.5%. In comparison to spheres, these shapes allow granules to flow more easily around the target object and envelope it completely. Yet, when the manipulated object design is less suitable for GJ mode, the influence of EA on the resulted grasping force is higher. We observed the highest differences (35%, 16%, and 15%) in grasping forces between granular jamming mode (0.83, 1.33, 1.35 N) and combined mode (1.12, 1.54, 1.55 N) for a cylinder, pyramid, and parallelepiped respectively. In the case of deformable water balloon, we observed grasping forces of 1.8 and 2 N for the GJ and combined mode indicating the grasping enhancement of 11%.
Figure 5. Characterization of grasping forces for different geometries and surface conditions. a) The grasping force for granular jamming (GJ) and combined mode for different object geometries that have the same volume: sphere, cube, pyramid, disks with diameters of 20 and 50 mm, cylinder, parallelepiped, and deformable balloon with liquid. The flat disk with a diameter of 50 mm has a diameter larger than the diameter of a gripper and cannot be grasped using GJ mode. The highest grasping forces of 6.1, 10.4, and 8.9 N in both modes were observed for a sphere, cube, and disk of 20 mm. The combined mode increases the adhesion for these objects by 5.6%, 3.7%, and 0.5%. These shapes allow granules to flow more easily around the target object and grasp it from all sides. The highest differences of 35%, 16%, and 15% in grasping forces between GJ mode 0.83, 1.33, 1.35 N and combined mode 1.12, 1.54, 1.55 N are observed for a cylinder, pyramid, and parallelepiped respectively. The combined mode shows worse performance only in the case of a disk of 50 mm (flat object). The activation of GJ in a combined mode makes the bottom surface involved in grasping rigid, thus, less compliant. b) Grasping forces of a flat object as a function of four grasping modes. Electroadhesion mode (EA) shows the highest grasping force of 0.8 N compared to combined, granular jamming, and nature stickiness forces of 0.15, 0.27, and 0.45 N resulting in the grasping enhancement of 430%, 194%, and 78% respectively. c) The grasping force for a combined mode for different surface conditions: dry (grey), oily (yellow), porous (pink), moistened (red), and powdered (brown). The hemisphere with the dimeter of 28 mm, which corresponds to 78% of the gripper diameter was used. The same setup and testing protocol as for Figure 4a was used. The grasping of objects with dry, moistened, and oily surfaces resulted in four to five times higher grasping force compared to porous and powdered surfaces due to suction effect formation.
We also characterized the grasping force for each of the objects when GJ, EA, and the combined mode are turned off, and the object is lifted only due to the inherent stickiness of silicone (Figure S1, Supporting Information). Interestingly, for a flat object larger than the gripper size, we observed that two highest forces of 0.42 and 0.80 N are achieved in the soft state of the gripper using only the stickiness of silicone and EA mode, respectively (Figure 5b). In these modes, the gripper is in the soft state and is more compliant than GJ mode, resulting in a better contact between the gripper and the object that leads to higher grasping performance. The modes that includes GJ, thus, stiff state of the gripper, result in the maximal grasping force of 0.27 N, which is three times lower than EA. The activation of GJ makes the gripper rigid and the external contact surface of the gripper becomes rough (Figure S2, Supporting Information). The higher surface roughness in a rigid state makes the contact area between the device and the manipulated object lower than in a soft state.[18]
Contrary to EA-based grippers, GJ-based grippers can grasp objects that are powdered, moistened, and porous.[11] In this study, we chose different gripper design and material composition from already existing devices.[31] Thus, we performed characterization of the grasping force at different surface conditions in the combined mode (Figure 5c) as this mode shows the highest gasping force among all modes (Figure 4a). We used the hemisphere with a dimeter of 28 mm, which corresponds to 78% of the gripper diameter, following the same setup and testing protocol as for Figure 4a. We measure the grasping force for the gripper in combined mode under different surface conditions of the object: dry, oily, porous, moistened, and powdered. Porous and powdered surfaces resulted in the lowest grasping forces (1.1 N and 1.9 N, respectively). Oily, moistened, and dry objects instead yielded substantially higher forces (14.8 N, 12.9 N, 12.6 N, respectively). The porous and powdered surfaces prevent the formation of sealed interface between the gripper and object surfaces, eliminating the suction effect. In the case of oily, moistened, and dry objects, the silicone surface of the gripper fastens tightly around the object causing a suction effect, which magnifies the grasping force. These results are in line with the performance of a conventional GJ gripper measured with similar surface conditions.[11]These results indicate that the outer surface of the gripper, which is three times thicker than conventional GJ grippers due to the presence of the EA electrodes, does not limit grasping capabilities for different surface conditions.
Grasping and manipulation of fragile, deformable, and flat objects.
GJ is an effective method for grasping solid objects with different shapes.[5,11,12] However, the grasping of deformable, and flat objects can be challenging and is never discussed in the literature for GJ-based grippers. Here we show that the combined use of GJ and EA enables successful pick and place of such objects (Figure 6 and Video S1, Supporting Information). In a first example, we pick up two fragile quail eggs with diameters of 27 mm and a weight of 17 grams each. The gripper in the soft state approaches the eggs until the granules in the gripper flow around both eggs