A bio-inspired expandable soft suction gripper for minimal invasive surgery—an explorative design study

Gripping slippery and flexible tissues during minimal invasive surgery (MIS) is often challenging using a conventional tissue gripper. A force grip has to compensate for the low friction coefficient between the gripper’s jaws and the tissue surface. This study focuses on the development of a suction gripper. This device applies a pressure difference to grip the target tissue without the need to enclose it. Inspiration is taken from biological suction discs, as these are able to attach to a wide variety of substrates, varying from soft and slimy surfaces to rigid and rough rocks. Our bio-inspired suction gripper is divided into two main parts: (1) the suction chamber inside the handle where vacuum pressure is generated, and (2) the suction tip that attaches to the target tissue. The suction gripper fits through a ∅ 10 mm trocar and unfolds in a larger suction surface when being extracted. The suction tip is structured in a layered manner. The tip integrates five functions in separate layers to allow for safe and effective tissue handling: (1) foldability, (2) air-tightness, (3) slideability, (4) friction magnification and (5) seal generation. The contact surface of the tip creates an air-tight seal with the tissue and enhances frictional support. The suction tip’s shape grip allows for the gripping of small tissue pieces and enhances its resistance against shear forces. The experiments illustrated that our suction gripper outperforms man-made suction discs, as well as currently described suction grippers in literature in terms of attachment force ( 5.95±0.52  N on muscle tissue) and substrate versatility. Our bio-inspired suction gripper offers the opportunity for a safer alternative to the conventional tissue gripper in MIS.


Surgical grippers versus suction grippers
Gently but securely gripping and handling of slippery, compliant and irregular tissues remains one of the most challenging procedures during minimal invasive surgery (MIS). In MIS, slender instruments are guided into the abdomen via trocars. These trocars act as portals through small incisions ( figure 1(a)). They typically have a diameter between 5 and 10 mm. Grippers used in current practice commonly use a force grip to manipulate tissues. This force grip is applied by two jaws that grip the tissues sideways and is subsequently translated into a pull force, to relocate or stretch the gripped tissue ( figure 1(b)) [1]. To prevent slip and to compensate for the low friction coefficients between the jaws and tissue, high pinch forces are often applied [2], which in turn can lead to unwanted tissue trauma [1,3,4]. Furthermore, force feedback is limited in the currently available MIS grippers, thus creating the risk of excessive force application during surgery and unintended tissue injury.
Tissue trauma due to excessive clamping forces of the gripper can lead to severe implications [1][2][3]5]. Bowel perforation is an example of an implication with a high mortality rate due to improper gripper use [6]. Furthermore, Joice et al [7] showed that the largest number of surgical errors during gall bladder removal is associated with gripper jaw induced tearing of the tissue. Next to the bowel and gall bladder, the bile duct, ureters, fallopian tubes and spleen are examples of delicate organs prone to damage by gripping [8]. In order to overcome these challenges, a new type of gripper can be introduced that does not need to enclose the tissue in order to apply high pinch forces for sufficient grip. An interesting direction for this challenge is suction attachment, in which a pressure differential is used to grip different substrates ( figure 1(b)). Suction grippers can exert pull forces directly in line with the instrument shaft, which limits damage caused by compressive pinch forces from the sides. This makes suction technology not only potentially relevant for gripping delicate and slippery tissue, but also for gripping large organs that can not be clamped between the jaws of a grasping forceps.

State of the art
Suction grippers are widely used in industry to grasp a variety of objects, such as cardboard packages, sheet metal, bottles or electronic components. In the medical field, however, commercially available suction stabilizers only exist for (open) cardiac surgery. Examples are the Octopus Heart Stabilizer (Medtronic), the Starfish (Medtronic) and the Xpose 3 (Getinge). These devices have been designed to locally stabilize heart tissue during off-pump coronary artery bypass surgery [9][10][11][12]. Next to these commercially available devices, suction grippers are described in literature for tissue stabilization, drug delivery and tissue dissection [13][14][15]. In addition, suction discs have been successfully tested in vivo as a locomotion strategy on heart tissue to perform injections in the heart wall [16,17].
The devices mentioned above clearly show the potential of suction technology to interact with tissue in MIS. However, the forces that these devices can exert are relatively low, in the range of 1 N, and thus do not allow for tissue lifting or stretching. Only two suction grippers specifically designed for tissue manipulation during MIS have been identified in literature: (1) a laparoscopic suction gripper for bowel manipulation (developed at Delft University of Technology) [18] and (2) a lung positioner for thoracoscopic surgery (developed at Tokyo institute of technology) [19]. The laparoscopic suction gripper consists of a rigid suction nozzle and was able to generate an attachment force of 5.6 N on bowel tissue using a vacuum pressure of 80 kPa [18]. The lung positioner consists of an array of rigid suction chambers that form a triangular shape inside the body after insertion through a small incision [19]. Even though both devices showed that suction technology is promising for tissue manipulation in a minimally invasive setting, they were challenging to manufacture and vulnerable to air leaks. This vulnerability was substantiated by the lung positioner requiring a continuous vacuum inflow in order to maintain grip on the tissue [19]. Therefore, further development in terms of grip, substrate diversity and manufacturing is needed to bring the concept of vacuum gripping during MIS to the operating room.

Inspiration from biological suction discs
Rubber suction discs are commonplace in our daily environment. Unfortunately, the possibilities of suction attachment of current man-made suction discs are still limited since these discs do not adhere to rough substrates with a roughness greater than 35 µm (where roughness is defined as the average grain size) [21], or remain stationary on slippery substrates when subjected to external forces. This is mainly due to their inability to create a sufficient seal or sufficient friction between the disc and the substrate, making them less suited for attachment to rough surfaces and slippery and soft tissues. Also, the suction attachment of these discs is dependent on deformation due to an initial pressing force, which is undesired for applications on tissues during surgery.
In a number of scientific research groups in the world, research is being carried out unraveling the working mechanisms of suction discs found in nature, as these outperform man-made suction discs in terms of versatility [22]. Many biological suction discs are able to attach to a wide range of substrates, varying from soft and slimy surfaces to rigid and rough rocks [21,23]. Examples of animals with such powerful suction discs are clingfish, octopods, aquatic insect larvae and starfish [23][24][25][26][27][28] (figure 2). The process of attachment and detachment can be quite dynamic, with some gobiidae fish able to climb the vertical cliffs of waterfalls using a pelvic-fin-derived suction disc [29]. Hill-stream loaches even use their entire body as sucker for anchoring in high-speed flows in stream environments [30,31].
Researchers reported varying ways to design bioinspired suction discs. The main focus of current research efforts is on mimicking the material characteristics and morphology of the suction disc rim to enable improved attachment performance, for example used for fruit handling in industry. It must be noted that none of these bio-inspired discs has been developed specifically for MIS. Tramacere et al [32] developed octopod inspired suction discs made of soft elastomeric materials and with a network of grooves on their surface. Ditsche and Summers [21] and Sandoval et al [33] developed suction discs inspired by the northern clingfish that featured highly elastic contact surfaces, various types of inclusions, slits and bio-inspired body geometry. Sadeghi et al [34] developed a sucker pad inspired by the tube feet of sea urchins, where multiple individually sealed suction discs are actuated by one vacuum chamber. Designed for the manipulation of fragile, rough or wet objects, these studies evaluated the attachment performance of their bio-inspired suction discs on varying substrates compared to man-made suction discs. The results showed that the bio-inspired suction discs outperformed conventional man-made suction discs in terms of substrate diversity, as they were able to adhere to rough and wet substrates where man-made discs failed [21,33].

Aim of this study
Even though major strides have been made in the development of bio-inspired suction discs, a number of challenges need to be researched and overcome for tissue manipulation in MIS, such as the generation of sufficient forces for tissue lifting and stretching while incorporating outer dimensions applicable for MIS and maintaining stable sealing. This study uses biological suction discs as inspiration to propose a novel design for a minimally invasive tissue gripper actuated by suction. We will investigate how nature is able to create a strong seal on soft surfaces and how to translate this into a small diameter soft suction disc design for MIS. The goal of our study is, therefore, to develop and test a bio-inspired suction gripper for manipulating flexible and slippery surfaces during MIS.

Decoupling functions
In order to understand how man-made and biological discs differ from one another, the working principle of a regular man-made suction disc and an octopod suction disc is analyzed. Figure 3 visualizes this analysis in a stepwise manner. The first row shows how both suction disc types generally generate suction. An octopod suction disc has a clear distinction between a dedicated suction chamber that generates the vacuum force and an optimized contact surface that maintains the seal and increases grip generation, whereas manmade suction discs generally combine these functions in one single flexible component [20,22,26,[35][36][37].

Suction generation
In the design of regular man-made suction discs, the suction chamber is usually integrated with the contact surface. The deformation of this contact surface generates the suction force. The concave contact surface is pressed towards the substrate (F i ) to expel the air between the two surfaces [38]. The stiffness force of the body of the suction disc (F 0 ), generated by its internal strain, tends to restore the original shape of the disc (figure 3(a)), which then generates the vacuum force [38,39]. This means that a trade-off is made between the required stiffness of the disc to restore its original shape and the required flexibility of the disc to create a seal.
The suction chamber and the contact surface of an octopod suction disc are physically separated. Therefore, they are specialized for their own function, suction generation and seal-and grip generation, respectively. The shape, flexibility and musculature of the suction chamber are optimized for generating powerful volume change. The highly specialized musculature is subdivided into radial, circular and meridional muscles (figure 3(b)) [35,37]. As the chamber acts as a muscular hydrostat, muscle contraction in one direction results in the elongation of a muscle in another direction [35,37]. The suction chamber's wall thins by contraction of the radial muscles, which increases the internal volume of the suction chamber and results in a vacuum force. Cross connective fibers prevent the collapse of the suction chamber when being subjected to vacuum pressure [20]. This form of dimensional stability is also identified in the northern clingfish's suction disc, where collagen bundles and flexible rays in the suction chamber's wall add resistance against vacuum pressure [40,41].

Seal generation
The contact surface of a regular man-made suction disc is usually unable to adapt sufficiently to curved or rough substrates, as the disc as a whole, including the part in contact with the substrate, requires a certain amount of stiffness for sufficient vacuum generation (figure 3(c)). Due to the physical separation of the suction chamber from the contact surface, the contact surface of an octopod suction disc can be optimized for its sealing properties (figure 3(d)) [42][43][44]. The contact surface of an octopod disc contains ridges and grooves that usually mold with the contours of the substrate, thereby closing leakage channels [20,26]. The surface consists of a highly flexible material, allowing it to follow the curves and irregularities of the substrate. Tramacere et al [36] found that the contact surface of an octopod suction disc has a Young's modulus of 7.7 kPa, which is very low as compared to rubber with a Young's modulus between 1-10 MPa [45]. For this reason, an octopod suction disc is able to easily attach to curved and rough substrates. Furthermore, the radial muscles in the disc's contact surface are oriented in such a way that contraction results in flattening and enlargement of the contact surface [20]. This lowers the disc's susceptibility to leakage as it enables the increasing of the sealing region [42]. Tramacere et al tested the influence of suction cup flexibility by mimicking the structured suction disc's contact surface of octopi using Ecoflex 00-30 silicone (Smooth-On Inc. USA). This resulted in approximately 65% higher pull-off forces as compared to a regular suction disc [32].

Grip generation
Whereas the regular man-made suction disc generally contains a smooth contact surface (figure 3(e)), the contact surface of an octopod suction disc reveals microstructures (figure 3(f)) [22,32]. These microstructures are also identified on suction discs of a number of other animal classes like decapods [20,46], clingfish [21,23,24,40], remoras [47,48], lumpsuckers [41], river loaches [31], catfish [49][50][51][52], blepharicerid larvae [25,53] and parasites [54]. They can appear as rigid hooks that extend from the skeleton, or flexible structures that extend from the epidermis [23,24,40,47,54]. The rigid hooks interlock with the asperities of the substrate, thereby acting as a ratchet system [48]. The flexible microstructures usually mold over the asperities of the substrate [44]. This increases the contact area between the suction disc and substrate and generates an interlocking effect, which both enhance the disc's grip [24,31]. Figure 4 shows that the suction disc's contact surface of the Octopus bimaculoides is covered with grooves and protrusions. The radius of the protrusions is approximately 3-4 µm [46]. The microstructures on the contact surface of the Clingfish is even more hierarchically layered, where the smallest protrusions can adapt to surface asperities of just a few nanometers [21,24]. These microstructures form a layer on top of the contact surface that magnifies its grip. The magnification of the grip enhances the resistance against shear forces and inward slip. This plays an essential role in the disc's attachment performance, as the man-made suction disc tends to slip inwards when it is pulled off, causing the seal to break rather easily [23,31]. Also, the interlocking effect results in better grip on slippery surfaces.

Schematic disc design
A schematic redesign of a regular man-made suction disc had been made inspired by the decoupled functions in an octopod suction disc, which are suction generation, seal generation and grip generation (figure 3(g)). Here, suction generation and seal generation are similarly decoupled in a separate suction chamber that conducts volume change and in a dedicated flexible contact surface that creates a seal. By decoupling these functions, no trade-off has to be made between the required stiffness for the vacuum generation and the required flexibility for the seal generation. The contact surface is optimized to maximize the sealing region, which lowers the disc's susceptibility to leakage. The contact surface is reinforced by a dedicated grip surface optimized on the magnification of grip counteracting shear stresses.

Design constraints
The schematic redesign of the regular man-made suction disc in the previous section acts as base for our suction gripper's design. Some additional design constraints are related to the gripper's use in MIS to ensure safe and reliable handling of slippery tissue. The vacuum force of the disc has a theoretical limit defined by its maximal suction surface and vacuum pressure. Our suction gripper will be actuated by a vacuum pressure of 50 kPa, as the use of this amount of vacuum pressure on human tissue has been extensively tested by the commercially available Octopus Heart Stabilizer (Medtronic) [10,11,55]. Higher vacuum pressures could risk tissue damage. Next to a fixed vacuum pressure, the space available as vacuum surface limits the maximum vacuum force further. The vacuum surface is constrained by the allowed dimensions for surgical instruments during MIS. The vacuum gripper should be guided into the body by a trocar with a maximal diameter of 10 mm, which defines the maximally allowed outer dimensions of the vacuum gripper. The maximum space for the suction surface, a circle 10 mm in diameter, with a maximum vacuum pressure of 50 kPa, results in a theoretical maximum vacuum force of only 3.9 N, assuming that the entire allowed space is used as suction surface, meaning that the wall thickness of the disc equals zero. To place this in perspective, a short literature survey was conducted to list the mass of the different organs inside the human body. Table 1 shows an overview of the weight of some of the largest organs, showing the need for roughly 50% higher vacuum forces for tissue lifting, depending on the organ. A strategy for higher vacuum force generation is integrating a folding mechanism into the suction disc, such that it unfolds into a larger suction surface when guided into the abdomen through the trocar.

Design suction tip 3.2.1. Layered disc design
The octopod suction disc analysis in the previous section reveals that a regular man-made suction disc can be improved by separating its suction chamber and contact surface and optimizing these for their own function. We will first focus on the design of the suction tip of the suction gripper, which enhances seal generation and grip generation on slippery tissue. The design of the suction chamber is less critical, as there is much design freedom to integrate this chamber in the suction gripper, e.g. in the handle or the shaft of the device.
The suction tip's design integrates five functions that it needs to fulfil in order to allow for safe and effective tissue handling: • Allow the suction tip to expand into a larger suction surface (foldability). • Create an air-tight suction tip (air-tightness).
• Allow for a smooth transition of the suction tip through the trocar (slideability). • Magnify the friction coefficient between the suction tip and tissue (friction magnification). • Allow secure sealing region formation (seal generation).
These functions were translated into separate layers of the suction tip, see figure 5.
In order to expand the suction surface of the suction tip, a number of folding mechanisms were considered, varying in their way to transform from an unfolded into a folded state. The challenge was to integrate a mechanism with easy actuation, small required space and the ability to resist high vacuum pressures. A suction tip containing pre-bent rays proved most optimal. These flexible rays return to their natural bent shape after guidance through the trocar. This means that no external actuation is needed to transform the tip into its unfolded state, minimizing usage steps for the surgeon. Additionally, these rays act as a support structure that strengthens the tip's dimensional stability when its subjected to high vacuum pressures. Moreover, the tip maintains its overall flexibility which is needed for folding.
Leakage of the foldable suction tip was prevented by an air-tight layer that covered the folding mechanism. The challenge of this layer's design was to minimize its thickness while maintaining its air-tightness, easy manufacturing and providing a smooth integration with the folding mechanism. A thin layer of silicone turned out to act best considering silicone's high elongation rate, its ability for casting as a thin film and its bio-compatibility. The silicone was reinforced by a finely woven mesh, which acted as a base to cast the silicone film in a three-dimensional shape. The silicon layer interlocks between the mesh fibers after curing to create a secure bond. Also, the mesh provides additional dimensional stability, similar to the collagen fibers of the octopod's suction disc. The silicone enclosure was covered by a thin foil with a low friction coefficient with the trocar.
The design of the footprint of the suction tip was focused on the generation of a secure seal and the  magnification of the friction coefficient with the surface of tissue. For the generation of a secure sealing region, the challenge was to lower the risk of the formation of leakage channels between the suction tip's footprint and tissue surface. This is achieved by attaching highly flexible material to the sturdy edge of the suction tip to make the tip better adaptable to the tissue surface. Figure 6(a) shows three possible footprint configurations of this flexible layer. The first configuration shows the suction tip with only its edge covered by flexible material, which creates short leakage channels that can result in failure. In the second configuration, a flexible membrane is attached to the outside of the tip's edge. Because of the membrane, atmospheric air has to intrude a longer distance before causing leakage. The downside of this second footprint configuration is the occupied space of the suction tip. This problem is avoided by placing the flexible membrane at the inside of the tip's edge, as shown in the third configuration. This configuration has the additional benefit that the tissue is pressed onto the footprint of the suction tip when the vacuum pressure pulls the tissue inside the tip ( figure 6(b)). The flexible membrane incorporates a high friction coefficient with tissue to enhance the suction tip's frictional support against shear forces.

Intended use
The intended use of the suction gripper is visualized in figure 7. The gripper's tip unfolds after being guided into the abdomen through the trocar (figures 7(a) and (b)). A vacuum pressure is generated to fixate the suction gripper on the tissue surface ( figure 7(c)). The tissue is pulled into the suction tip due to the applied vacuum pressure. The firm grip secures stable tissue lifting and stretching during tissue manipulations ( figure 7(d)). At the end of the procedure, the vacuum pressure is released and the suction tip returns into its folded state when retracted into the trocar (figure 7(e)).

Fabrication suction tip
A prototype of the suction tip was built to validate its attachment performance. Dimensional stability and foldability of the suction tip were provided by flat nylon rays. The nylon rays were pre-bent into a semi circular shape by heating them while being pressed around a rod. A finely woven mesh was glued between the rays (universal contact glue, Bison). The air-tight film was fabricated using silicone (Smooth Sil, Smooth-on Inc. USA) with a shore hardness of 40A. This type of silicone has the perfect balance between flexibility, allowing the tip to fold and stiffness to maintain dimensional stability. For the deposition of a thin silicone layer on the mesh, a fabrication method inspired by gravity-based dip-coating was used ( figure 8). First, the silicone mixture was poured into the suction tip, creating a silicone basin. Secondly, the suction tip was hung above a drainage platform. Gravitational forces drained the mixture from the silicone basin. During curing, the silicone mechanically interlocked with the mesh. The outer layer of the suction tip was fabricated by revolving a thin polyethylene foil around the suction tip.
Ecoflex 00-30 (Smooth-on Inc. USA) was used to fabricate the footprint. This silicone was chosen due to its high elongation rate of 900%, making it suitable to adapt to the shape of the tissue. In order to create the footprint, a thin film of uncured Ecoflex was smeared on a flat plate. The suction tip was placed on this uncured layer to attach the Ecoflex silicone air-tight to the Smooth Sil silicone while curing ( figure 9).
The suction surface of the suction tip was made as large as possible, while still being able to slide in and out a ∅ 10 mm trocar ( figure 10). This resulted in a suction surface with a 15 mm diameter in the unfolded state.

Experimental goal
The main goal of the experiments was to investigate the performance of the flexible suction tip, by (1) assessing the attachment performance of the suction tip on diverse (phantom) tissues and (2) measuring the effect of the vacuum pressure on the attachment performance.

Experimental setup
The experimental setup is illustrated in figures 11(a)-(d). In this experimental setup, the suction tip was connected to a 30 ml syringe by tubing with high buckling resistance (FESTO PAN-6x1). The suction tip was connected to an s-beam junior load cell (FUTEK, model FLLSB200) via a wire. This wire assured that the suction tip could move freely and, therefore, prevented premature detachment due to peel-off forces. Figure 11(a) shows a closeup of this part of the setup. The load cell, in turn, was connected to a linear stage via a 3D-printed base plate. The linear stage (ALMOTION LT50-TR-G8-200) was connected to a 3D-printed support that functioned as a containment unit for the gelatin tissue phantoms and animal tissues. The linear stage allowed for controlled motion of the load cell with the attached suction tip and thus for measuring the detachment force. Next to pull-off forces, this setup was able to measure slide-off forces exerted on the suction tip by connecting the load cell to the tip using a pulley system ( figure 11(d)). The friction between the pulley and the wire was negligible. The slide-off forces were exerted in line with the contact surface of the suction tip, preventing the generation of a moment that would tilt the suction tip. The vacuum pressure was measured by a vacuum sensor (NXP, model MPX4115AP) that was connected to the tubing of the suction unit using a T-piece connector. Both the load cell and the vacuum sensor were connected to an analogue signal conditioner (CPJ RAIL, SCAIME) and a data acquisition unit set at a sampling rate of 20 Hz (NI USB-6008, National Instruments Corporation).

Experimental design 4.3.1. Experimental variables
The detachment force of the suction tip was the main performance measure of the suction tip. The detachment force was defined as the force required to detach the suction tip from its substrate and could either be orientated perpendicular (pull-off force) or parallel (slide-off force) to the contact surface. The detachment force was tested on a variety of substrates and vacuum pressures. Phantom tissues were manufactured to mimic tissues with varying stiffness, see table 2. A plexiglass plate was used as reference to compare the measurements of the phantom tissues with a fully stiff substrate. All phantoms and the plexiglass substrate were tested in dry and lubricated conditions, as it was expected that the mucus layer of tissues might influence the performance of the suction tip. Vacuum pressures of 30 kPa, 40 kPa, 50 kPa, 60 kPa and 70 kPa were used. Additionally, experiments were conducted at the Skillslab in Erasmus  Medical Centre (Rotterdam, the Netherlands) using ex vivo chicken liver tissue and ex vivo cow muscle tissue.

Experimental protocol
At the start of each experiment, the load cell was calibrated by subtracting the weight of the suction tip from the measured pull-off force. Next, the suction tip was placed on the substrate and the vacuum pressure was generated by pulling the syringe manually. The syringe was pulled until the desired vacuum pressure was reached, which was monitored in LabVIEW.
3D-printed clamps fixated the syringe at the desired position. The linear stage moved upward with a velocity of 1 mm s −1 . In order to achieve constant starting conditions, the following two steps were taken. First, the suction tip was placed on the substrate with an untensioned wire. This ensures that no initial traction was applied to the load cell at the moment of vacuum initiation. Secondly, the tubing connected to the suction base was supported, as its weight could otherwise tilt the suction tip and break the seal. The gelatin was poured into containers with horizontal grooves on the inside to prevent the gelatin from getting lifted during the experiment. As gelatin degrades over time, only gelatin made the same day was used for experiments. The animal tissue samples were fixated between two surfaces of a 3D-printed clamp.
The main experiment was subdivided into two sub-experiments: (1) Experiment 1: Effect of the substrate and (2) Experiment 2: Effect of vacuum pressure. In Experiment 1, the suction tip was tested on different gelatin tissue phantoms, the plexiglass reference substrate and the ex vivo tissue samples. The vacuum pressure was set to 50 kPa for all measurements of Experiment 1, based on the vacuum pressure of the commercially used Octopus heart stabilizer (Medtronic). The phantom tissues were lubricated with a thin layer of water-based lubricant. The slide-off measurements were performed on lubricated phantom tissues as these generally reassemble the surface properties of tissues closest, being surrounded by a slippery mucus layer. Measurements of Experiment 2 were all performed on lubricated phantom tissue made of 10 wt% gelatin while being subjected to pull-off forces. Vacuum pressures were generated in a range of 30 kPa-70 kPa. Five trials per condition were executed for all experiments.
• Experiment 2. Pull-off measurements using vacuum pressures in a range of 30 kPa-70 kPa.

Data analysis
The data of the pressure sensor and load cell was obtained using Labview 2013. From this data, the average detachment force ± standard deviation was calculated. The data from Experiment 1.1 was analyzed by a one-way ANOVA test to examine if the attachment performance when being subjected to pull-off forces is significantly different for the different substrates. Also, the average detachment force onto lubricated substrates obtained in Experiment 1.1 was compared with the average detachment force onto dry substrates obtained in Experiment 1.2 by an independent two-sided t-test. The data from Experiment 1.3 was analyzed using an independent twosided t-test to examine if the attachment performance when being subjected to pull-off forces is significantly different for the muscle and liver tissue samples. Furthermore, the data from Experiment 2 was analyzed by a one-way ANOVA test in order to identify if the attachment performance is significantly different for the different vacuum pressures. The statistical significance was set at an α value of less than 0.05. The linearity of the relationship between vacuum pressure and detachment force was analyzed by the correlation coefficient.

Pull-off measurements
5.1.1. Experiment 1.1: lubricated phantom tissue Experiment 1.1 showed that the suction tip induced higher detachment forces to the substrates with a lower stiffness, as illustrated in figure 12. The average detachment force against pull-off forces and its standard deviation on the lubricated plexiglass plate, lubricated 15 wt% gelatin, lubricated 10 wt% gelatin and lubricated 5 wt% gelatin were 2.02 ± 0.49 N, 2.7 ± 0.13 N, 3.27 ± 0.15 N and 3.26 ± 0.16 N, respectively. A one-way ANOVA test shows that the difference between performance outcomes of the substrates is significant (F(3, 16) = 23.35, p = 4.4 × 10 −6 ). It was noticeable that the suction tip failed a few times in its initial attachment on the plexiglass plate, while this did not happen with the gelatin phantom tissues. Furthermore, the observation was made that detachment from the rigid substrates (the plexiglass plate and 15 wt% gelatin) was often due to a V-shaped indentation (see figure 13), while the perimeter of the suction tip kept its circular shape during detachment from the flexible substrates.

Experiment 1.2: dry phantom tissue
The attachment performance of the suction tip against pull-off forces on dry substrates is illustrated in figure 14. The average detachment force and its standard deviation on the dry plexiglass plate, dry 15 wt% gelatin, dry 10 wt% gelatin and dry 5 wt% gelatin were 2.04 ± 0.12 N, 3.79 ± 0.41 N, 4.03 ± 0.28 N and 3.68 ± 0.33 N, respectively. Oneway ANOVA testing shows that the difference in performance outcomes between the substrates is significant (F(3, 16) = 17.51, p = 2.6 × 10 −5 ). An independent t-test shows an insignificant effect on

Experiment 1.3: ex vivo tissue samples
The attachment performance of the suction tip against pull-off forces on ex vivo animal tissue samples is illustrated in figure 15. The average detachment force and its standard deviation on the liver tissue sample and the muscle tissue sample are 1.98 ± 0.90 N and 5.95 ± 0.52 N, respectively. An independent t-test shows that the difference in performance outcomes between the tissues is significant (t(8) = 8.5783, p = 2.6333 × 10 −05 ). The suction tip performed less well on the liver tissue sample compared to the liver's phantom counterpart, whereas it performed better on the muscle tissue sample compared to the muscle's phantom counterpart.

Experiment 1.4: lubricated phantom tissues
The average detachment force against slide-off forces and its standard deviation on the lubricated plexiglass plate was 4.61 ± 0.96 N. The suction tip was observed to slide over the plexiglass plate's surface before it detached at once. No detachment force could be measured for the lubricated 15 wt% gelatin and the lubricated 10 wt% gelatin, as the suction tip kept sliding over the gelatin phantoms without detaching. The average attachment force before the suction tip started sliding for the 15 wt% and 10 wt% gelatin phantoms were 2.06 ± 0.35 N and 1.38 ± 0.20 N, respectively. The suction tip did stop sliding on the lubricated 5 wt% gelatin. However, the gelatin broke before the detachment force could be reached. The reached average attachment force before the 5 wt% gelatin broke was 1.99 ± 0.35 N.

Experiment 1.5: ex vivo tissue samples
The slide-off measurements on the ex vivo animal tissue samples showed that the suction tip did not find grip on both surfaces. The suction tip kept sliding over the liver tissue until the end of the sample was reached, or until the tissue slipped out of the suction tip during sliding. The average attachment force during sliding was 1.74 ± 0.8 N. The suction tip kept sliding over the muscle tissue sample until it touched a bump in the tissue surface. The average detachment force of the suction tip on muscle tissue was 6.37 ± 0.41 N. One-way ANOVA testing shows that the difference in performance outcomes between the pressures is significant at p < 0.05 (F(4, 20) = 47.62, p = 6 × 10 −10 ).
An independent t-test shows no significant difference in the detachment force with a vacuum pressure of 60 kPa and 70 kPa (t(8) = 1.0047, p = 0.3445).

Main findings
In this study, we have developed the first bioinspired suction disc design for a suction gripper in MIS. A proof-of-principle experiment showed that the developed suction tip was able to successfully and safely grip slippery tissues without enclosing them.
Noticeably, the attachment performance of the suction tip was higher on flexible substrates than on rigid substrates. The muscle tissue sample showed clear imprints of the suction tip after a measurement, which reveals that the suction tip is performing a shape grip onto the flexible tissue surface ( figure 17(a)). The tissue is pulled into the suction tip, where the tip's perimeter grasps the flexible tissue, somewhat similar to how a conventional gripper grips tissues ( figure 17(b)). This results in improved sealing and additional grip when attached to flexible substrates. Also, pulling the tissue into the suction tip results in an expansion of the suction surface, as more tissue surface is directly affected by the vacuum pressure. Figure 18 shows the force-time graphs and vacuum pressure-time graphs of a trial with the lubricated plexiglass plate and a typical trial with lubricated 10 wt% gelatin, subjected to pull-off forces. It shows that more leakage occurs during the attachment onto the rigid plexiglass plate, whereas less vacuum pressure is lost during attachment on the flexible gelatin. This substantiates that the suction tip seals better on flexible substrates.
When subjected to shear forces, the suction tip's shape grip again enhanced its attachment performance. The highly flexible 5 wt% gelatin was deformed sufficiently into the suction tip to prevent the tip from sliding over the slippery surface. The shape grip was insufficient on the 10 wt% gelatin, the 15 wt% gelatin and the plexiglass plate to prevent the suction tip from sliding. In future research, we propose to look into strategies that increase the grip of the footprint layer to lower the risk of sliding. Nature shows that this could be done by adding a microstructure on the footprint's surface [24,31]. Optionally, the vacuum pressure could be increased when slip is detected, either by eye or an integrated sensing system. Furthermore, it was observed that the suction tip had more grip on the dry substrates than on the lubricated substrates. This could be caused by the higher friction coefficient between the footprint and the dry substrate, which counteracts inward slip of the rim and thereby results in delayed detachment. This would explain why detachment forces of the suction tip are higher on the muscle tissue sample than the liver tissue sample for both the pull-off measurements and the shear measurements. Although the liver tissue is more flexible than the muscle tissue, its surface is covered by a highly slippery mucus layer. Overall, these results show that the suction tip performs best on flexible and dry tissues.
Increasing the vacuum pressure from 30 kPa to 70 kPa showed a relationship with the attachment force that acted linearly for low vacuum pressures, and flattens out for high vacuum pressures. It could Figure 18. Occurrence of leakage during measurements when subjected to pull-off forces. Left: force-time and pressure-time graphs of a measurement with the lubricated plexiglass plate as substrate. Right: force-time and vacuum pressure-time graphs of a measurement with lubricated 10 wt% gelatin as substrate. More leakage before detachment occurs with the rigid substrate compared to the flexible substrate, which means the suction tip seals better with flexible substrates. Both measurements were performed with an initial pressure difference of 50 kPa. Figure 19. Occurrence of leakage during measurements when being subjected to pull-off forces. Left: force-time and vacuum pressure-time graphs of a measurement with an initial pressure difference of 30 kPa. Right: force-time and vacuum pressure-time graphs of a measurement with an initial pressure of 70 kPa. More leakage before detachment occurs using a low initial pressure difference compared to a high initial pressure difference. Both measurements were performed with a lubricated 10 wt% substrate.
be hypothesized that the substrate is pulled more into the suction tip using high vacuum pressures, increasing the foot print surface and thus the resistance to leakage. At some point, the substrate is maximally deformed and no longer affected by higher vacuum pressures. This would explain why the curve in figure 16 flattens out for higher vacuum pressures. The force-time graphs of a typical measurement ( figure 19) indeed show more leakage during attachment at lower actuation pressures. The developed suction tip was able to exert a maximal detachment force of 3.27 ± 0.15 N on slippery, flexible phantom tissue, 1.98 ± 0.90 N on liver tissue and 5.95 ± 0.524 N on muscle tissue. Theoretically, all listed organs in table 1 can be lifted by these attachment forces. It could be argued that heavy organs would be lifted by multiple suction grippers anyway, to prevent local stress accumulation. Figure 20 shows the suction tip's ability to lift some arbitrary objects with varying shapes and surface characteristics. The tip is able to grip objects smaller than the suction tip's perimeter by adapting its shape to the object. Figure 21 shows how the suction tip is able to lift a slippery chicken heart. During surgery, the vacuum pressure should be monitored and optimized in order to gain sufficient grip on different tissue types, or multiple smaller suction gripper could be used to ensure proper attachment.

Application fields
In some specific applications, it could be favorable to use a suction gripper over a conventional gripper. For example, literature mentions the merit of using suction gripping in bowel and colon manipulations, as these thin membranes are most delicate and easily damaged by squeezing forces of the gripper's jaws [13,18]. Another application for the suction gripper would be the fixation of voluminous slippery tissue, e.g. liver tissue, during the execution of a biopsy.

Limitations
This study is subjected to some limitations. The exploratory nature of this study provides only initial insights into the performance behavior of the suction gripper. Future research, including in vivo studies, should show in which procedures the use of our suction gripper is favorable over a conventional gripper.

Design recommendations
For the device, some development steps are recommended to increase its potential. First, more research should be focused on optimizing the vacuum actuation, either manually or by a vacuum pump. Figure 22 shows a first prototype in which a vacuum chamber is integrated into the suction gripper's handle to facilitate manual actuation. Furthermore, future research should optimize the properties of the footprint layer in terms of grip and flexibility. Future studies should focus on in vivo experiments to evaluate the performance of the suction gripper in a realistic clinical setting. Biocompatible alternatives should be found for the materials that are currently incorporated in the prototype, before testing on porcine models. Miniaturization of the suction gripper, striving for an outer diameter of 5 mm, is required to function as an equal alternative to the conventional gripper in minimally invasive procedures. Optimization of the folding mechanism could result  in a higher gain in the vacuum surface when converting from a folded into an unfolded state. Here, the nylon rays of the current folding mechanism could be replaced by superelastic nitinol wires. These nitinol wires have as an additional advantage that they can be accurately shaped using heat, which increases the shape consistency and durability of the suction grippers. These design recommendations will contribute to increasing the potential of our suction-based tissue gripper.

Conclusion
This study presented the development and successful first validation of a bio-inspired suction gripper for MIS. Similar to biological suction discs, our developed gripper is divided into two main parts: (1) the suction chamber inside the handle in which vacuum pressure is generated and (2) the suction tip that attaches to the target tissue and allows for tissue manipulation during MIS. The suction tip consists of five layers to allow the suction tip to expand into a bigger surface area (foldability), to create an air-tight tip (air-tightness), to allow for a smooth transition of the suction tip through the trocar (slideability) and to create an air-tight seal with the tissue (seal generation) while magnifying frictional support (friction magnification). The experiments illustrated that our suction gripper outperforms manmade suction discs, as well as currently described suction grippers in literature in terms of attachment force (5.61 versus 5.95 N [18]) and substrate versatility (figures 20 and 21). The suction tip's shape grip enhances its resistance against slipping during the shear experiments and adapts the tip's shape to grip objects smaller than the suction tip's outer diameter. Our bio-inspired suction gripper offers the opportunity for a safer alternative to the conventional tissue gripper in MIS.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.