Magneto‐Responsive Polymeric Soft‐Shell‐Based Capsule Endoscopy for High‐Performance Gastrointestinal Exploration via Morphological Shape Control

Capsule endoscopy is a valuable tool for diagnosis of the gastrointestinal tract. As a result, active‐control capsule endoscopy, which can be controlled via external stimulation, has emerged as a prominent diagnosis tool, unlike traditional passive capsule endoscopy. However, owing to its rigid cover, the high rotating radii limit its observation range at anchor points, such as polyps. There are also persistent challenges with pass‐through in narrow spaces in the intestines. This study proposes a soft‐shell capsule endoscope (SSCE) to improve this limitation. Experiments ranging from virtual phantoms to ex vivo investigations in the stomach, small intestine, and large intestine were conducted to ensure comprehensive coverage of the entire digestive system by the SSCE. Additionally, in the stomach and intestine experiments, we compared the performances of the SSCE and hard‐shell capsule endoscope (HSCE) to highlight the advantages of the SSCE. The results demonstrate that the SSCE achieves a 37.3% increase in the observable angles through bending motions at the anchor point compared to the HSCE, in addition to increased efficiency of pass‐through in the intestines. This revolutionary innovation is expected to significantly impact and promote high efficiency via pass‐through and the observation area in the clinical application of capsule endoscopy.


Introduction
Since the approval of the first wireless capsule endoscope (CE) of the small intestine by the U.S. Food and Drug Administration (FDA) in August 2001, [1,2] it has been widely used in clinical applications for early monitoring and diagnosis of the gastrointestinal (GI) system. [3,4]Owing to its small size, the device has allowed patient convenience, low pain, wide detection area, and ease of operation. [5,6]The CE was initially designed to solve the limitations of traditional endoscopes, with which it was difficult to observe some lesions in the small intestine and colon. [7,8]raditional gastroscopies and colonoscopies posed challenges in terms of watching the full coverage of the GI tract.There is a linear endoscopy, known as enteroscopy, specifically designed for examining the small intestine.Doctors use this to access deep into the small intestine, including double balloon enteroscopy, [9] single balloon enteroscopy, [10] and spiral enteroscopy. [11]Balloon and spiral enteroscopies are collectively known as deep enteroscopy.The solution for complete detection of the GI tract using traditional wired endoscope is to combine deep enteroscopy and colonoscopy.Deep enteroscopy in the GI tract allows examination and marking a point; then, a colonoscopy can be performed at the marked point for full-coverage GI examination.However, the potential for secondary injury to patients during insertion is a concern with artificial long-distance manipulation of the digestive tract.To mitigate this, James et al. developed the invention of a wire-controlled endoscope, controlled by an external magnetic field. [12]This innovation allows for a more refined of the camera within the endoscope, facilitated by a non-linear smooth control system, thereby diminishing patient discomfort during the procedure.However, the limitation that the entire digestive system can be observed with just one wired endoscope still exists.The emergence of CE thus allows convenient and comprehensive inspection from the esophagus to rectum in clinical applications.
Passive capsule endoscopes relies on the motility movement of the digestive organs for its downward progression.It can only be applied to observe and detect the digestive system passively. [13,14]By adding a permanent magnet inside the capsule, [15] the endoscope can be actively controlled using a magnetic field generated by an external electromagnetic system to improve movements that are otherwise difficult to perform under passive locomotion, [16] which is limited to only being able to move downward with the peristalsis of the digestive system. [17]While maintaining the essential functions of endoscopes, CEs endowed with more functions are successively becoming new research achievements in this field.Hoang et al. developed a tattooing capsule endoscope (TCE) for application to organs with sharp curvatures and geometrical limitations. [18]Lee et al. developed a mucoadhesive patch built in the ingestible shell that offers drug delivery and release, showing potential for use in CE for GI tract treatment. [19]Le et al. use a gripper tool to add a miniature biopsy module on an active locomotive intestinal capsule endoscope (ALICE). [20]As a promising alternative to endoscopy, several companies, such as Jinshan (China), [21] Given Imaging (now Medtronic, Israel), [22] and Intromedic (South Korea), [23] have also developed and commercialized devices.
[26][27] However, previous research has focused mainly on demonstrating motion through biomimicry, [28] so progress has to be further made for actual applications.Although a few soft biomedical applications exist, [29,30] including catheters [31][32][33] and embolisms, [34] most presentations revolve around movements through motion. [35]Capsule-related research is even more scarce; Son et al. introduced a soft CE with a biopsy function, wherein the softness is only in the hinge component to facilitate deformation during collapsing motions rather than the entire device. [36]Given these limitations, we developed a soft-shell capsule endoscope (SSCE) to overcome the challenges (Figure 1a) and combined it with practical applications.The SSCE uses a complete soft-shell design to address the limitations associated with rigid CEs.Such a departure from the cylindrical hardcover structure provides several advantages, including improved patient safety and reduced risk of injury: The CE's soft design reduces the chance of mucosal damage compared to hard-shelled versions.Its flexibility lets it move safely through the GI tract, adjusting to its shape without creating harmful pressure, which lowers the risk of injuries like tears.A soft capsule can also pass more easily through tight spots, decreasing the risk of complications like blockages.
In the case of hard-shell CEs (HSCEs), the shortest straightline distance hinders the ability to observe in a 360°view around a fixed point.To enable rotational movements about the vertical axis, the observed inner wall or inner diameter length must be greater than or equal to the total length of the CE.In the domain of GI imaging, CE has emerged as a revolutionary non-invasive technology, enabling detailed visualization of the gastric interior by navigating within the stomach.Despite the advancement, a pivotal challenge that persists is the capsule's tendency to lose track of its relative position during transit.This is exacerbated by the variable orientations it undergoes, which are influenced by external control mechanisms.The limitation of maintaining accurate positional tracking hinders the capsule's ability to offer a complete and contiguous mapping of the confined gastric spaces, which is critical for thorough examination and diagnosis.In contrast to the CE, the SSCE presents distinct advantages for achieving 360°comprehensive coverage and observation.By increasing the scanning area of the capsule at a single location, when dealing with areas having diameters larger than 1.5 cm, the SSCE can achieve comprehensive coverage by rotating 360°after tilting the head under the control of an external magnetic field.This mechanism enables a complete visual assessment of the surrounding areas.
Instead of mimicking the structured cylindrical hardcover, the proposed SSCE has a modified shape based on the molding technique, [37,38] which can inevitably produce active locomotion under an external electromagnetic field as a premise (Figure 1b).At the head and tail of the SSCE, we reserve spaces for the camera module, battery, and data transmission module.The diameter of the hollow body part in the middle is reduced to 2 mm, which provides space for connecting wires between the head and tail.We magnetized the integrated body after deformation and fixation in advance, as required for more experiments.
In the present study, we provide the motivation for the SSCE that can be deformed under an external electromagnetic field. [39]e also fully use this feature to conduct further experiments and demonstrations.First, we tested the deformation abilities of the SSCE, such as bending, tumbling, and rotating, for an applied electromagnetic field to show the advantages of the soft properties.Second, overall spatial scanning is performed as a relatively intuitive demonstration of the advantages; here, we scanned the inner area of a hemispherical structure.Finally, we demonstrate the aforementioned motions in the stomach, small bowel, and large intestine.The experiments start with a phantom module and progress to a porcine ex vivo test.The goal here is to prove that the functions of the SSCE can cover the entire GI system (Figure 1c).

Design and Fabrication of Soft Capsule Shell
The SSCE fabrication is illustrated in Figure 2a as a concept schematic.The soft shell consists of three distinct parts: head, body, and tail.The head section accommodates a camera module, while the body is hollowed out to facilitate connection between the head and tail, allowing for the inclusion of wiring and other components.The bottom portion of the shell primarily houses modules such as batteries and facilitated data transmission.In addition, a small permanent magnet is embedded if enhanced control under an external electromagnetic field is required.Aside from the essential transparent hemispherical glass utilized at the head and the seal at the tail, the remaining parts were fabricated using molding technology in an integrated production process (Figure 2a(i)).We utilized a mixture of silicate polymer materials and permanent magnetic powder [40] to achieve controlled movements under an external electromagnetic field.
To align the magnetic moment of the embedded permanent magnetic powder, we employed a vibrating sample magnetometer (VSM) for magnetization of the soft shell. [41,42]To optimize the magnetization direction for each part of the soft robot, we maintained a bent configuration during the magnetization process, as depicted in Figure 2a(ii).Specifically, the main body of the soft shell was bent at approximately 180°to facilitate generation under the electromagnetic actuation (EMA) system.We applied a magnetic field with a magnitude of 1.3 T, significantly exceeding the remanence value of the magnetic powder, which was measured as 891 mT.Following the magnetization process, we assembled the camera module and battery, as previously mentioned (Figure 2a(iii)).The resulting dimensions of the SSCE were as follows: height of 30 mm, head and tail radii of 12 mm each, and outer diameter of the body averaging 3 mm (Figure 2a(vi)).

Characterization and Optimization
The magnetic properties of the soft shell were assessed using a VSM.The results of the VSM measurements was presented in Figure 2b, showing the relationship between the magnetic flux density and magnetization magnitude.Notably, the red point on the graph indicates the point of intersection between the curve and y-axis, corresponding to a remanence value of 0.962 emus.This value represents the magnetic magnitude of the soft-shell material after magnetization.Notably, the magnetization was determined to be 32.497emu g À1 at this remanence point.This magnitude is significant as it indicates that the soft-shell material can generate sufficient magnetization when exposed to an applied magnetic field.
The SSCE composition is illustrated in Figure 2c.A camera holder part was designed to accommodate a small camera instead of the hemisphere glass traditionally assembled at the head of the capsule.The assembled SSCE is depicted in Figure 2d.To enable the desired motions, we employed an EMA system, as depicted in Figure 2e.The SSCE was positioned within the workspace of the EMA system for subsequent experimentation.
The remote actuation scheme for the magnetically guided active-control SSCE is presented in Figure 2f, which utilizes a global coordinate system for the traditional head and go (HAG) motion.The directions of the vectors in the scheme are determined by two angles, yaw angle (θ) and pitch angle (α), which are employed for alignment purposes.Specifically, the alignment of the capsule is achieved by adjusting the pitch angle α_B within the θ_B-tilted plane.The angle θ defines the relationship between the xz-plane and desired control plane of the robot.In contrast, the angle α represents the discrepancy between the controlled vector and its projection on the xy-plane.
The structure optimization process of the SSCE was carried out next.Initially, our focus was on modifying the thickness of the middle part of the hollow cylinder as it influences the stiffness during bending motions, as verified through testing.We fabricated three prototypes with varying thicknesses, namely 1, 0.75, and 0.5 mm, as displayed in Figure 2g.Subsequently, we conducted bending motion tests under the EMA system to establish the relationship between the input values and output angles.The deformation of the soft capsule shell was evaluated under a 5 mT magnetic field.The results demonstrate that the 1 mm sample could achieve a maximum bending angle of 76.87°when the α_B value was set to 45°.Further increasing the input value resulted in a decrease in the bending angle.A similar behavior was observed for the 0.75 mm sample, but with a maximum bending angle of 95.19°when the input value reached 45°.In the thinnest model, the 0.5 mm prototype, the head ultimately achieved contact with the tail portion, representing the theoretical maximum bending achievable.This motion allows the SSCE to observe the entire area in a single-sided view.The angle between the head and tail sections was measured, yielding a minimum value of 7.98°at an input of À120°and a maximum angle of 134.29°when the input value was 45°.The strategic placement of the battery at the SSCE's rear not only biases its center of gravity to the back end, but also stabilizes the capsule against rotations that could be induced by minimal angles in the external electromagnetic field.This prevents overall overturning and facilitates the intended one-point anchoring motion, which is vital for achieving a wide range of scanning without the necessity of additional anchoring mechanisms.The images on the right side are the maximum bent shapes for different thicknesses of the SSCE.It is important to note that the measurement of the included angle is based on the centerlines of the head and tail in the side view.Given that the camera used in the SSCE has a visible area of 60°, the model enables complete surround scanning.Therefore, we selected this particular model for further experiments.Achieving a thinner structure, such as 0.25 mm, presents challenges with the molding technique, but the 0.5 mm prototype already demonstrates the theoretically maximum bending capability.

Magnetic-Field-Guided Experimental Performance Demonstration
The concept schematic of the SSCE's bending behavior was presented in Figure 3a, accompanied by experimentally captured screenshots; the red arrow is the magnetic field direction of input and the magnetic strength is also 5 mT.The prototype, with a body thickness of 0.5 mm, exhibits varying degrees of bending at different input angles, covering an entire single-side area.The bending motion of the 0.5 mm SSCE is depicted in Figure 3b.The soft capsule demonstrates an additional falling motion when subjected to a vertical field under a uniform magnetic field.At a magnetic field direction α_B of 90°, the CE displays n-shaped bending, while at À90°, it exhibits u-shaped bending.The arc radii of the bending curvatures were measured at different magnetic field strengths.In Figure 3c, n-and u-shaped bending were tested under a uniform magnetic field ranging from 0 to 25 mT with 5 mT increments, revealing a decrease in bending curvature as the magnetic field increased.The maximum spans of the nand u-shaped bending arc radii were approximately 23.46 mm and 34.51 mm, respectively.
The tumbling motions of the SSCE under a rotating magnetic field is shown in Figure 3d and Movie S1, Supporting Information.The capsule was placed on a flat plate with a fixed field strength of 5 mT, and the value of α_B was smoothly adjusted to induce front or rear tumbling.The value of θ_B was changed to determine the tumble direction.The video demonstrates the tumbling motions of the SSCE along the x-axis and y-axis under an external magnetic field.Additionally, the SSCE was placed inside a hollow cylinder tube, enabling observation of the rotating motions using a similar principle, as shown in Figure 3e and Movie S2, Supporting Information.By smoothly adjusting the value of α_B under the control of the external rotating magnetic field, the SSCE achieved rotation in a simulated environment.The inner side of the tube was marked with letters A to G, allowing clockwise or counterclockwise viewing by changing the rotation direction.This demonstration highlights the advantages of the SSCE for efficient observation and accurate positioning in small and closed areas, such as the small intestine.The bending of the body and head enable wide-angle observations, covering the side blind areas entirely in a small-radius cylinder.
In clinical detection by CE, the endoscope may get blocked by dirt, such as debris, digest, or luminal content hindering inspection of the digestive system.Therefore, effective cleaning of the blocked dirt is crucial.Friction with the surrounding tissue has been identified as a preferred solution for this owing to the simplicity and high efficiency.The application of the SSCE's collapsible motion for cleaning blocked dirt is presented in Figure 3f and Movie S3, Supporting Information.The endoscope undergoes repeated bending by the switched external magnetic field until the dirt obstructing the camera is cleared from the viewing range.A polluted polymer film was used for the test, initially confirming the presence of dirt on the camera surface through a top view.A controlled external magnetic field was then utilized to make the endoscope lie down.When a vertical upward magnetic field with a magnetic field strength of 30 mT and a switching frequency of 3 Hz was applied, the SSCE repeatedly executed an n-shaped bending.Real-time observation of the camera surface allowed confirmation of dirt cleaning through friction.The endoscope was then made to stand using an external magnetic field, and the strength of the magnetic field was adjusted to verify thorough cleaning of the surface dirt from the top view.This demonstration effectively utilizes the deformable soft shell's advantages and addresses the issue of camera dirt in practical applications through a large bending angle.
We also evaluated the crawling efficiency of the SSCE in a fixed hollow tube.We placed the SSCE in a hollow tube of diameter 15 mm, poured 1 mL of corn oil to simulate the environment of the small intestine, and applied a sawtooth wave for 180 amplitudes.We conducted experiments under external magnetic field strengths of 10 and 30 mT.It can be seen from the results in Figure S1 and Movie S4, Supporting Information, that the velocity of the SSCE also changed under different frequencies of the external magnetic fields.At 10 mT, the best speed of 11.23 mm s À1 was realized at an input frequency of 2 Hz; at 30 mT, the best speed of 44.94 mm s À1 was realized at an input frequency of 5 Hz.The speeds at the two frequencies were 0.37 and 1.50 body length/s correspondingly.After achieving the best speeds under different magnetic field strengths, the moving speeds dropped sharply, contrary to increasing the magnetic field because the response speeds of the SSCE were not as high as the change in the external magnetic field.Hence, we also call 2 Hz and 5 Hz as the step-put frequencies.

Surround-Scanning-Based Stomach Ex Vivo Test
Observing multiple fields within one position can improve the inspection efficiency in CE.One of the challenges in gastric monitoring with an actual CE is the difficulty in determining the capsule's position using the camera on the endoscope, which often leads to loss of coordinates.To address this problem, we conducted phased advanced experiments (Figure 4a).As a preliminary experiment for full-area scanning, we selected a hemisphere with a 100 mm inner diameter and placed five colored ribbons inside.The hemisphere, with the colorful ribbons arranged from top to bottom as purple, dark green, light green, yellow, and orange were illustrated in Figure 4b.A blue ribbon spans all these colors and serves as the starting and ending points for the scan at a magnetic field strength of 5 mT.The experiment begins at the top with the purple ribbon as the initial point, then descends along the blue ribbon to locate the second layer, the dark green ribbon.The scanning process involves rotation and observation until the entire bottom orange section is scanned.By sequentially scanning all the ribbons, we have effectively scanned the interior of the hemisphere in all directions.Figure 4c and Movie S5, Supporting Information, represent a portion of the observed field captured by the SSCE camera and the complete scanning process.
We further analyzed the planar scanning efficiency of the CE in Figure 4d,e, and Movie S6, Supporting Information.The HSCE is restricted in its ability to bend downward, resulting in a maximum bending angle of 90°relative to the vertical position.In contrast, the SSCE can reach a bending angle of 123.6°.This property allows the CE to observe the proximal end of the anchor point, expanding the observation field and minimizing  blind spots.In clinical applications, observing the bottom region involves anchoring and observing from the upper side.However, it is often challenging to maintain the position of the first anchor point and determine the relative position accurately.To support the aforementioned theoretical analysis, we compared the SSCE and HSCE at the centers of multiple concentric circles of various colors at a magnetic field strength 5 mT.From the result images, it is evident that the SSCE can observe five areas, whereas the HSCE can only observe three.
Capitalizing on this advantage, we conducted further experiments utilizing a stomach model to simulate practical applications.We prepared an isometric model of the stomach via 3D printing and differentiated various regions by applying distinct colors, as shown in Figure 4f.In this experiment, we still compared the HSCE and SSCE.The CE was initially anchored on the lesser curvature (left double arrow) by non-uniform mass, and a magnetic field controlled its bending through the EMA system at 8 mT that followed a greater curvature (right double arrow).In Figure 4g and Movie S7, Supporting Information, the HSCE's observation from an anchor point only covers a part of the fundus and part of the pyloric antrum owing to its limited swing range.However, Figure 4h and Movie S7, Supporting Information, show that the SSCE covers most of the stomach's interior by leveraging its comprehensive observation range.It successfully scans the inside of the stomach, esophagus, starting position of the stomach, and part of the duodenal bulb.Furthermore, the SSCE achieves full-range scanning from a single anchor position.Through these comparative experiments, we conclude that the bending capability of the SSCE not only facilitates full-coverage scanning of the surrounding environment, but also enables observation near the anchor point.Consequently, this improved scanning efficiency can significantly enhance clinical detection efficiency.
To demonstrate the mobility of the SSCE within an actual stomach, we conducted an ex vivo experiment using a porcine stomach.To better observe the movements of the SSCE inside the stomach, we cut the stomach along the lesser and greater curvatures, dividing it into two parts; one part was laid flat for observation.As shown in Figure 4i and Movie S8, Supporting Information, the SSCE starts at the cardia and tumbles its way through the body to the pyloric area of the stomach at an external magnetic field 30 mT.The red dotted line represents the actual trajectory of the SSCE within the stomach.

Experiments on Passage in the Small Intestine
In this section, we focus on movements within the small intestine.First, we performed locomotion experiments in silicone molds of the small intestine.The mold and SSCE view's inner images are shown in Figure 5a.From the SSCE view, it is easy to observe that the photos at the initial position, end position, and two curved corners all face the front.Movie S9, Supporting Information, illustrates images when the SSCE moves through the phantom; the time-lapse images match the trajectory SSCE images in the phantom movement.The green dotted line is the trajectory of the capsule.In clinical applications, this property is of great help to observe the efficiency of the small intestine system.It is difficult to confirm the position of the CE in the body without the help of an external real-time positioning device, so we attempt to maintain the viewing direction in the front as much as possible.Before performing CE on patients, we usually perform a preliminary unobstructed capsule examination on groups with high incidence rates or suspected subjects to rule out problems with failure of passage due to blockages or very narrow digestive tract.By dissolving and discharging the capsule from the body within a predetermined time, the capsule helps doctors judge whether the patient is suitable for CE.However, this method has the disadvantage that it increases the inspection cost.
We also conducted a movement comparison experiment between the SSCE and HSCE in a small intestine phantom with different inner diameters and curve radii.Figure 5b(i) shows that when the inner diameter is 20 mm and curve radius is 17.5 mm, both the SSCE and HSCE can pass through the S-shaped curve smoothly; Figure 5b(ii) shows that when the inner diameter is 15 mm and curve radius is also 15 mm, the SSCE can still pass through the curve smoothly but the HSCE cannot.The black dotted line is the trajectory of the capsule (Movie S10, Supporting Information).The SSCE is superior because of its soft shell and can crawl forward in a curved tube; however, if the radius of the curve is less than or equal to the length of the HSCE, it cannot pass through the curved tube.These results show that the SSCE can pass through a phantom with a small curve radius by crawling forward.In clinical trials, the CE can follow the digestive movements of the small intestine to move forward.However, a small intestine with very small curve angles may pose hidden dangers that may cause difficulties in expelling the CE.
We also conducted stomach experiments with the porcine small intestine to verify the feasibility.The experiment was performed while the porcine small intestine was in the collapsed state (Figure 5c and Movie S11, Supporting Information).In these experiments, the SSCE passed through the straight and curved regions of the small intestine via a mixture of crawling, rotating, and n-and u-shaped bending motions with repeatability.We successfully achieved passage efficiency through the small intestine that was comparable to that of the HSCE.

Experiments on Passage in the Large Intestine
In the locomotion of the CE inside the large intestine, the challenge is moving forward by the large intestine's voluntary movements.Herein, we briefly explain the large intestine's internal structure.The lining of the large intestine is pleated, with valves and circular muscles that prevent the stool from moving down.It is precisely because of this structure that the movement of the CE in the large intestine is affected even if it can be actively driven.We prepared a 3D-printed phantom mimicking the folds of the large intestine and carried out the SSCE experiment on it (Figure 6a and Movie S12, Supporting Information).Under the control of an external magnetic field; we used the tumbling motions to pass over three folds with different heights and widths from right to left.A rotating magnetic field centered on the x-axis was applied at a strength of 30 mT.When the magnetic field was applied, the SSCE's raised head will be stuck on the high point of the barrier.Under a continuously applied rotating magnetic field, the tail on the right side of the barrier will also turn over.Based on the same principle, the SSCE also successfully passes the last two barriers.In this experiment, the direction of travel of the SSCE is not a straight line but deviates towards the heavier tail.This problem will be solved in the large intestine because of a border.This point of view is proved in the following experiment with the large intestine mold.
The large intestine phantom used in this experiment (Figure 6b), where we compare the pass-through abilities of the HSCE and SSCE (Figure 6c and Movie S13, Supporting Information).For the HSCE, it is not easy to cross over a fold.Figure S3, Supporting Information, shows that the device head is first lifted higher than the wall before being pulled forward using external stimuli, such as a magnetic field, and the head is then stuck in a groove behind a fold.Finally, it is necessary to raise the head again and repeat the previous motion to overcome the folds and pass through the interior of the large intestine.Using the SSCE to pass through the cavity of the large intestine also involves falling into the next groove after turning over a fold.However, this does not prevent observing the surrounding environment simultaneously while advancing through the soft outer under the control of an external magnetic field.
To achieve the experimental purpose of full coverage of the SSCE on the digestive system, we added an ex vivo experiment after that with the large intestine.This section compares the SSCE and HSCE passage efficiencies in the same porcine large intestine segment.As shown in Figure 6d and Movie S14, Supporting Information, the passage of the HSCE in the porcine large intestine is blocked by the influence of the circular muscles and folds.Although under the control of an external magnetic field, it is challenging to continue moving forward by constantly changing the direction and method of advancement.However, the SSCE is superior owing to its soft characteristics, so that mixed motions and movements within the large intestine can be realized successfully.Similar to the movement in the small intestine, straight and curved segments were successfully traversed.In this experiment, passage after collapse of the large intestine demonstrates that the SSCE benefits greatly from the soft shell to enhance passage.

Conclusions
CEs are crucial for visualizing the GI tract for diagnostic purposes.However, the limitations of traditional CEs in achieving comprehensive observations within confined spaces have prompted the development of alternative solutions.In this study, we propose an SSCE controlled by an external electromagnetic field.First, optimal prototypes were determined based on different body thicknesses and bending angles.The prototypes were then integrated with camera modules and batteries to demonstrate their deformation and motion capabilities under an external applied magnetic field.Subsequent experiments were conducted in the stomach as well as small and large intestines.All experiments were initially performed using models and later validated through ex vivo experiments.Collectively, these findings provide strong evidence toward the feasibility of achieving comprehensive coverage of the digestive system using the SSCE.However, certain limitations still exist, particularly regarding specific motions.The current challenge with SSCE technology is the need for consistent and unobstructed forward visualization as the capsule navigates the complex environment of the large intestine-a feature that is well-managed within the simpler structure of the small intestine.The variable and often intricate colonic folds present obstacles that impede the capsule's ability to maintain this necessary visual orientation.To surmount this challenge, we propose the integration of a modular component that can variably adjust the volume of the capsule's soft body within the intestinal environment.Such a balloon-like expandable mechanism would support the colonic wall from inside during the capsule's inflation, facilitating a smoother traversal and stable imaging as the capsule progresses through the colon.In conclusion, utilizing a soft magnetic shell as the outer wall of a CE is a hitherto unprecedented development.In future experiments, we will strive to address the abovementioned limitations and make significant strides in improving the detection efficiency for practical clinical applications.

Experimental Section
Materials: Neodymium-iron-boron particles (NdFeB, Neo Magnequench, Canada) with a mean size of 5 μm were used as the magnetic material.The unmagnetized particles with a magnetic remanence of 891 mT were utilized and magnetized after the fabricating the shell.We used silicone rubber (Ecoflex 0020, Smooth-On, Inc., USA) to form the soft part of the shell.The utilized material has a tensile strength of 160 psi and elongation at break of 845%, which provide good strength and ductility to the capsule cover.The capsule camera built into the soft shell (DST1616, DASHING SOLAR TECHNOLOGY CO., LTD., Taiwan) has 1.6 mm diameter, 4 mm length, 120°viewing angle, 160 K pixel resolution, and 9 M USB for realtime image confirmation.
Preparation of SSCE: We fabricated a soft capsule cover by the molding technique.The mold includes three parts, namely an inner cylinder, an inner pot, and a peripheral hollow column.The inner cylinder constitutes the body part of the SSCE, and the pot is divided into three segments whose purpose is to better separate the cured soft shell from the mold.Therefore, the order of combination is to first restore the three parts of the inner tank and then limit them to the outer hollow column.Finally, the inner column is inserted after pouring the mixture.The designed mold was fabricated using a 3D printer (Objet 30 pro, Stratasys, USA) and coated with a thin layer of release agents (Ease release 200, Smooth-On, Inc., USA) to prevent adhesion between the mold and created soft robot.The silicone rubber and magnetic particles were then mixed in a mass ratio of 1:1.The mixture was first stirred and placed in a vacuum pump for 10 min to eliminate bubbles, poured into the prepared mold, and cured in an oven at 70°for 4 h.The fabricated soft robot has a height of 30 mm, an outer diameter of 12 mm, and a thickness of 1 mm at the head and tail regions.
Characterization: The magnetization curves of the SSCE were measured using a VSM (Lake Shore Cryotronics 7404, Westerville, OH, USA).
Composition of HSCE: HSCEs generally have three parts: head, body, and tail (Figure S2, Supporting Information).Compared with the SSCE, except for the shell, the difference is a permanent-magnet cylinder of length 10 mm embedded in the body.The CE's magnetically guided aspect allows precise control of its movement within the GI tract.External magnets are used to generate magnetic fields, thereby manipulating the capsule's position and orientation.
EMA System: The EMA system comprising one pair each of Maxwell, Helmholtz, and rectangular coils can accommodate magnetic fields up to 35 mT (Figure S4, Supporting Information).The system can perform full 5-degrees of freedom (DOF) motion, including 3-DOF translation and 2-DOF rotation; its magnetic workspace is 60 mm Â 60 mm Â 60 mm at the center of the coil configuration.The EMA system is controlled by eight power supplies consisting of four NX15 units and four 3001iX units (AMETEK, USA) operated using LabVIEW program (Figure S5, Supporting Information).

Figure 1 .
Figure 1.Schematic of the concept of a SSCE.a) Changing the HSCE to SSCE to increase flexibility.b) Forward movement by tumbling motion under a rotating magnetic field; forward movement by crawling motion under a rectangular magnetic field input.c) Detection by SSCE can cover the GI tract, including the stomach, small intestine, and large intestine.

Figure 2 .
Figure 2. Fabrication, characterization, and optimization of the SSCE.a) Fabrication of the soft shell via molding technique, followed by magnetization to align the magnetic moment in the soft shell.Final scale when built into the camera module and battery to achieve an actual SSCE.b) Magnetic property of soft shell measured by VSM.c) Final image showing assembly of the parts.d) Actual composition of the SSCE in this study (scale bar: 12 mm).e) Eightcoil EMA system in the study and the experimental setup.f ) Angle θ is used to specify the relationship between the xz-plane and desired control plane of the robot.The angle α describes the difference between the controlled vector and its projection on the xy-plane.g) Optimization process for three kinds of prototypes based on body thickness difference, and relationship between the input α and measured absolute bending angle.The images on the right side are the maximum bent shapes (scale bar: 12 mm).

Figure 3 .
Figure 3. Magnetic-field-guided experimental performance demonstration.a) Illustration of the SSCE bending motion under a magnetic field and captured images of real shapes.The red arrow is the actual input magnetic field direction.b) Illustrations of the n-, u-shaped bending (red curve) and their captured images under different magnetic field strengths.c) Relationship between magnetic field strength input and measured arc radius.d) Movement via rotating magnetic field on a plate by tumbling motion through the x-axis and y-axis.e) Rotational scanning of labeling alphabet via a rotating magnetic field in a cylindrical tube.f ) Schematic of the experimental dirt cleaning, where the SSCE's bending followed the magnetic field on and off states; the figure also shows the time-lapse images (scale bar: 12 mm).

Figure 4 .
Figure 4. Surround-scanning-based stomach ex vivo test.a) Schematic of the experimental setup for scanning motion in a hemispherical space.b) Real hemisphere structure with colorful ribbons used in the experiment.c) SSCE time-laps images in the surround-scanning experiment.d) Theoretical observation values of the HSCE when bending, and actual capsule images showing the observed areas and top view image.e) Theoretical observation values of the SSCE when bending, and actual capsule images showing the observed areas and top view image.f ) Stomach phantom fabricated by 3D-printer, painted with different colors to distinguish different areas.g) When anchoring, the results of the inner stomach scanning with HSCE and the shape schematics.h) Results of inner stomach scanning with SSCE and the shape schematics.i) Ex vivo test on a cut porcine stomach, where the movement was by tumbling motion under a rotating magnetic field, and the red dotted line is the trajectory of SSCE movement (scale bar: 12 mm).

Figure 5 .
Figure 5. Experiments on passage in the small intestine.a) Small intestine phantom and illustration images where SSCE moves through the phantom.The time-lapse images match the trajectory SSCE images in the phantom movements (the green dotted line is the trajectory of the capsule).b) Comparison between SSCE and HSCE for different curvature scales in cylindrical tubes (the black dotted line is the trajectory of the capsule).i)The inner diameter of the tube is 20 mm, and the curvature radius is 17.5 mm; the capsules move from the upper right to lower left corners and can smoothly pass through the S-shaped phantom including two curved corners.ii) The inner diameter of the tube is 15 mm, and the curvature radius is also 15 mm; the capsules still move from the upper right to lower left corners; the HSCE cannot pass through the first curved corner and is stuck, but the SSCE can pass through the entire phantom.c) Ex vivo test in the porcine small intestine for movement efficiency (scale bar: 30 mm).

Figure 6 .
Figure 6.Experiments on passage in the large intestine.a) Climbing over the printed wall to mimic circular muscles in the real large intestine via tumbling motions.I) 10 mm high; 5 mm thickness; II) 5 mm high; 3 mm thickness; III) 7.5 mm high; 4 mm thickness (scale bar: 12 mm).b) Large intestine phantom used in this work.c) Comparison between HSCE and SSCE in the large intestine phantom and corresponding endoscope images.d) Ex vivo tests between HSCE and SSCE in the porcine large intestine (scale bar: 30 mm).