Potassium permanganate, KMnO₄, (ACS reagent grade, 99%) was purchased as dark purple odorless crystals from Sigma Aldrich. The hydrophobic biodegradable polymer, Polycaprolactone, was also purchased from Sigma Aldrich. All solutions used in this study were prepared using ultrapure water from a bench ultrapure water system. As a biocompatible and biodegradable polymer, PCL was used to encapsulate KMnO₄ to formulate a CRBP system to observe controlled-release kinetics and characteristics of encapsulated KMnO₄ when introduced to aqueous media.

In this study, the first phase consisted of formulating CRBP pellets using the baking method for encapsulation described by the literature [32]. To produce the CRBP delivery system, KMnO₄, as the chemical oxidant, and PCL, as the hydrophobic polymer, were utilized. The encapsulation of this KMnO₄ oxidant allowed for stabilizing control of chemical release. The study modeled basic reproducible components of Teasley et al. studies [6] [21] [24] for establishment of standard measurements including a KMnO₄ to PCL at a ratio of 1:5 or 20 wt%. The same standard measurements were used for all CRBP pellet formulations. Initially, using a laboratory muffle furnace (Thermo Scientific, US), PCL was heated at 90˚C. Once PCL was molten, KMnO₄ particles were added to the soften polymer and stirred thoroughly to achieve a homogeneous mixture using a stainless-steel spatula. After blending for uniform particle dispersion, and while malleable, the molten mixture was then placed in a Parr stainless steel pellet press and pelletized. Solid circular pellet structures were produced with 20 wt% KMnO₄-PCL and size of 1.3 cm in length capable of the controlled-release (Figure 2(a)). Pellets were then preserved and stored in a dry cool place until use for batch release experiments. Pellets were designed based on a combined mechanism [32] consisting of controlled dissolution and diffusion of KMnO₄ for an extended period of time.

The study was conducted in the form of a series of batch release experiments, similar to what is seen in Figure 2(b). Experiments were conducted with 1:5 ratio of CRPB KMnO₄/PCL (also referred to as 20 wt%), approximately 0.60 g in mass, pellets submerged in 1 L of ultra pure water and synchronously released over the 96 hour period. Each of 12 Pyrex glass beakers were used as batch reactors for individual pellets. These reactors were tightly capped at room temperature under constant mixing with a mechanical stirrer to ensure a homogenous mixed solution. Aluminum foil was wrapped around the glass reactors for protection from ambient light. Preliminary experiments confirmed the findings of prior studies noted that the water volume did not affect the amount of relative mass lost over a 24 h period [7]. The release of KMnO₄ from the PCL was observed periodically by measuring the oxidant concentration in the aqueous solution. Samples of aqueous solution were withdrawn using a sterile serological pipette at various times within the 1-week period. At each time interval after

release, pellets were collected from reactors, desiccated for 24 hours and weighed for calculating changes and averages in mass.

Water samples collected from the batch experiments were analyzed for KMnO₄ concentrations using a UV-Vis spectrophotometric by measuring the absorption at λ = 525 nm wavelength at each sampling time interval, which is within the maximum absorbance range for KMnO₄. UV-Vis spectrophotometric analysis indicated the amount of additive KMnO₄ integrated in the aqueous solution. UV-Vis spectrophotometric used the absorbance measurements to determine the levels of KMnO₄ concentration in the aqueous solution. The analyzed absorbance and concentration for the samples were taken for each interval of time. The surface morphology of the encapsulated KMnO₄ CRBP before and after release experiments were characterized using FEI Helious G4 UC field emission scanning electron microscope (FESEM).

The release of the KMnO₄ particles from the PCL matrix can readily be seen in the SEM images (Figure 3). Increasing oxidant removal from the PCL surface was evidenced by the increase in size of cavities in the waxy surface of the polymer as time from release increased. These crevices during oxidant release also mirror what was seen from other CR systems utilizing KMnO₄ release [1] [27]. KMnO₄ progresses initially from the preliminary contact between KMnO₄ particles on the surface of the matrix and water (Figure 3(a)) to 24 hours (Figure 3(b)). The dissolution of the polymer bearing KMnO₄ particles occurred as the surfaces were freely accessible to water in the batch reactors. The PCL polymer degraded slowly in the presence of aqueous solution, similar to what has been stated in the literature in that the embedded KMnO₄ was released by dissolution-diffusion which resulted in the formation of KMnO₄ solution that would be available to oxidize contaminants in the aqueous phase [1] [27] [29].

At the 24 hour mark, the SEM image indicates a maximum imbibition of the shell matrix (Figure 3(b)). It is at this point that the amount of hydrolysis in the PCL shell peaks, KMnO₄ release decreased as the system transitioned from the dissolution phase to the diffusion phase followed by a controlled and slower release of KMnO₄. Though beyond the sampling intervals for this study, a 1-week SEM picture was also taken of the system to show long-term effects of release on the polymer surface (Figure 3(c)). After a week of uninterrupted contact with water, KMnO₄ dissolution left empty pores and larger crevices allowing the embedded KMnO₄ to diffuse in aqueous solution, clearly seen in the SEM images and in agreement to what has been noted in the literature [29]. These results support the conjecture that the combined control CRBP delivery mechanism for PCL is more effective for long term release of the oxidant.

The experimental release data for the KMnO₄ from the PCL matrix (Figure 4) was found to best fit the following equation (R² = 0.93):

Q t = k t n (1)

whereQ represents the mass fraction of KMnO₄ released over time t, k is a constant

incorporating polymer and oxidant characteristics, and n is the diffusional exponent. The fit of the data has been used to describe the release process of other oxidants from wax matrices as non-Fickian diffusion with constant pseudo-convection due to stress within the wax or polymer matrix itself [1] [27] [34]. These values of n and k can be gathered once Equation (1) has been linearized in logarithmic form (Equation (2)) and by plotting the experimental data to determine the slope and the intercept on the log Q_t axis, respectively.

log Q t = log k + n log t (2)

Even with the linearized model, it has been shown that the estimated values might not necessarily simulate the nonlinear release kinetics [1] [34]. Therefore, estimation of parameters can be done with an optimization method that minimizes the sum of the square of errors (SSE) between the experimental and calculated data (Equation (3)).

S S E = ∑ t ( Q exp − Q c a l c ) 2 (3)

The present study measured release of KMnO₄ from PCL at a ratio of 1:5, otherwise known as 20 wt%. Results from this present study were compared to parallel release studies with the same oxidant, KMnO₄, encapsulated in different wax matrices at similar ratios. Kang et al. [1] examined the use of paraffin wax (Figure 5) to encapsulate the KMnO₄ at a ratio of 1:5, similar to that of the present work.

In another study, Ighere and Chawla [27] used of Poly-methyl-methacrylate (PMMA) as polymer at a ratio of 1:8 KMnO₄ to PMMA. PMMA is a hydrophobic synthetic resin and an ester of methacrylic acid generally produced from propylene (Figure 5). These comparable polymers had different chemical structures (Figure 5, Figure 6) than PCL which makes the information from their release compared to its release of KMnO₄ valuable.

Figure 7 shows the data in this present study with 1:5 ratio of KMnO₄ to PCL agrees favorably with data from the compared studies which included Kang et al. [1] with 1:5 ratio KMnO₄ to paraffin wax and that of Ighere and Chawla [27] with 1:8 ratio of KMnO₄ to PMMA. The data from our present study was normalized by a factor of 20, since its initial mass of oxidant in the PCL was 1/20^th of that utilized by the other comparable studies.

Furthermore, the PCL release data modeled also compare favorably with the model parameters from the comparative studies (Table 1). The values of n and k for all the studies were calculated from Equation 2 based on established research on this model by Sinclair and Peppas [34]. Evaluating these model parameters further indicated that the PCL release of KMnO₄ of the present study aligned with this oxidant’s release from the comparable studied polymers. The slight differences can be attributed to both the different chemical structures and physical characteristics of the different polymers used, as well as the higher mass amounts utilized in the data from the previous literature.

Though the present study paralleled the release from the compared studies, it most agreed with the Kang et al. [1] study. This was expected since that study and the present research utilized the same ratio of 1:5 or 20wt%. Further agreement

between the present study and Kang et al. [1] study. This was noted in the calculation of SSE from Equation 3. The data of the present study was calculated to have a SSE value of 1.9 × 10⁻⁵, while Kang et al. was 1.7 × 10⁻⁴. This indicated that not only was the data relationships between these studies agreed, but also, that the data from the present data better fit the model as it was lower than that of the comparable study. Overall, the similarity of release data between the diversity of polymers shows that the CRBP with PCL provides effective release of the KMnO₄ with the added benefit biodegradable nature of PCL, as mentioned previously [6] [21] [32].