The swelling behavior can be attributed to competing interactions between the ions present in the gastric simulant and water molecules with the polymeric backbone of the absorbent material. The material's absorption property was evaluated by measuring the mass swelling capacity, that translates to the amount of fluid contained per gram of absorbent material (Figure S3). For instance, the absorbent material captured 45.97 ± 1.71 g of DI water/g absorbent, 44.62 g/g ± 0.85 of buffer solution and, 34.90 g/g ± 1.99 of gastric simulant. The absorbent material is hydrophilic, hence it quickly interacts with the water molecules in aqueous solutions. This interaction results in the expansion of the coiled network of the polymer, and consequent swelling (Figure 3a). The polymer-solvent interactions can then be tuned based on the solution properties, such as ionic strength or pH, that will determine the solvent quality, water/polymer interactions, and consequent swelling/de-swelling behavior. This phenomenon can be employed to provide reversible de-swelling of the absorbent matrix by placing the fully swollen hydrogel matrix in a high concentration of specific ions (e.g., Na+ or Ca2+), as illustrated in Figure S4. The osmotic pressure caused by ions in solution leads to a shift in water molecules toward the more hypertonic environment, leaving the polymeric chains free to coil back to their original position, resulting in hydrogel de-swelling, and heightened biomarker release.\cite{Liu2019}
To demonstrate that liquid-based biomarkers are specifically contained within the absorption matrix, the specific contribution of each component of the robotic pill to biomarker uptake was evaluated. Fully-assembled robotic pills were incubated for one hour in a solution containing 2 µm beads under constant shaking. Following the incubation, the pills were disassembled into their components, and the capture contribution of each component was analyzed by measuring the number of beads recovered from each separate robotic pill component, including (i) a negative control pill incubated without beads, (ii) the plastic backbone, (iii) porous membrane, and (iv) absorbent material, as shown in Figure 3b. These measurements indicate that the absorbent material provides a larger number of beads trapped (113 ± 29) when compared to the plastic backbone (6 ± 3), and porous membranes (20 ± 4) (Figure 3c). The sum of all beads recovered by the different pill components represents a capture yield of 29.8 percent from stock solution. In addition, the release of the microbeads was assisted using a 2M CaCl2 solution that helped to deswell the gel and release the trapped particles. We should note that this release method would help accelerate processing time but could affect biological targets viability after release. Figure S5 illustrates the colorimetric quantification of the recovered bovine serum albumin (BSA) protein captured overnight by a pill collector under different recovery protocols, including i) placing the absorbent material in a 2M CaCl2 solution and shaking for 30 min, and i) placing the absorbent material in a PBS solution and shaking for 30 min. Although both batches were incubated following the same protocol, the presence of a salt-rich environment promoted precipitation of the recovered protein affecting the analytical measurement compared to the use of buffer solution. Thus, the recovery method can be tailored for a different type of biomarkers.
The robotic pill was also evaluated for the capture of motile bacteria. The expansion of the absorbent material can serve to trap the bacteria within the three-dimensional porous polymer matrix,\cite{DiGiacomo2017,Bhattacharjee2019} retaining them in the structure under a moist environment, as shown in the scheme in Figure 3d and Video S2.