Handsheet Coated by Polyvinyl Acetate as a Drug Release System

The present work aims to prepare a controlled release matrix, using biodegradable materials containing different active ingredients (AI) in order to deliver only the necessary dose for a specific time and a well-defined target. The matrix is a manual sheet, composed of lignocellulosic fibers. The membrane, which regulates the release, is obtained by immersing the sheet in an emulsion of polyvinyl acetate. Different thicknesses were deposited by varying the concentration of the emulsion. Five AIs, which differ in volume and solubility, were used as probes: salicylic acid (SA), acetylsalicylic acid (ASA), metoclopramide (MCP), ketoprofen (KP), and thiocolchicoside (TCC). Kinetic studies have shown that the release of these substances is strongly related to the pH, the solubility of the AI, the thickness, and the permeability of the polymer membrane. The kinetics of release of AI are zero order in an acidic medium and complex in a basic medium. The images taken by scanning electron microscopy (SEM) show the variation in thickness with the number of layers of emulsion deposited. Our system, paper coated with a polyvinyl-acetate (PVAc) emulsion film without a coalescing agent has proven to be a promising matrix for achieving the controlled release of different IAs. It suffices only to adjust the thickness of the membrane. The process is simple and easy compared with other recently developed extended release matrices.


Introduction
Several strategies were used in the past to fight against famine and epidemics of all kinds. Environmental problems and soil toxicity were not a primary concern, given the enormous needs of a growing world population and the necessity to provide food and health for everyone. The research was more concerned to find the substance that can eliminate and/or prevent the proliferation of pests (viral, bacteria, fungi, rats…), neither were the vectors used to reach the target, and organic solvents were in common use to solubilize the active substances. These solvents were often aromatic (carcinogenic), chlorinated solvents that end up causing irritation and harmful consequences to the environment. For good dispersion and good wetting of the substrates, the use of surface agents, such as phenolic ethers, has also been used, resulting in high toxicity [1].
Nowadays, the international laws on the use of toxic products are very severe. To solve such problems, scientists are redirecting to more environmentally friendly and healthier strategies in order to avoid drug toxicity [2]. One of the strategies is the use of bioproducts and the use of water as a solvent or dispersion medium [3,4]. The creation of efficient processes is pivotal, in order to avoid unnecessary losses and to design a system delivering only the necessary dose for a specific time and a well-defined target, to treat a specific condition. The release must be effective, in a way that it lasts the necessary amount of time, without being accumulated to avoid unnecessary toxicity.
The structure of these designed controlled release systems is generally composed of a reservoir matrix in the form of a tablet, particle, or vesicle [5,6]. It can be mineral, ceramic, or organic. Often, in order to increase the retention time and decrease the release rate, a coating is applied with a well-determined thickness and, if necessary, pore-forming agents are introduced into this coating. The five AIs studied in this work have been the subject of several controlled release studies with matrices in the form of tablets, gels, or patches. Obtaining a certain release kinetics is achieved by coating them, often with a polymeric layer having a certain porosity. We cite, without being exhaustive, a few matrices used to control the release of these 5 active ingredients, and this, in order: SA, ASA, MCP, KP, TCC. Salicylic acid (SA) was dispersed in Karya gum [7], grafted onto polyvinyl alcohol [8], used as a crosslinking agent for poly (xylitol adipate) [9], and encapsulated with poly (2-hydroxyethyl methacrylate) in functionalized chitosan microgels [10]. The second active ingredient (ASA) was dispersed in the form of particles in polymeric matrices [11,12], in sodium alginate alone or combined with phosphate of dibasic calcium [13], grafted on starch and coated with polyvinyl alcohol [14], in a semi-conductive gel, polythiophene/carrageenan, the release of the ASA occurs under the action of an electrical stimulus [15]. Metoclopramide hydrochloride (MCP) has been studied as hydrochloride in matrix tablets based on mixtures of hydroxypropyl methylcellulose/sodium carboxymethylcellulose/sodium starch glycolate [16], in xanthan/ethyl cellulose mixtures [17] and hydroxylpropylmethylcellulose/ xanthan/polyethylene glycol [18]. Regarding ketoprofen, it was formulated as microcapsules coated with an acetatecellulose butyrate-polystyrene composite membrane with a pore-forming agent [19], tablets based on mucilage/hydroxypropyl methylcellulose [20], microparticles composed of ketoprofen/Eudragit calcium coated with ethylcellulose and carboxymethylethylcellulose release [21]. The last probe, thiocolchicoside (TCC), one of the examples is a two-layer tablet based on hydroxypropyl methylcellulose as a gelling agent, mucoadhesive microspheres based on sodium alginate/chitosan/Carbopol [19].
The preparation of controlled release systems (tablets, granules, tablets, gels, particles) requires several stages for their preparations, namely, (i) the choice and dosage of the components, (ii) the homogeneity of the mixture, (iii) the technology behind the controlled release form (granulation, compression,…), and (iv) the addition of agents or the use of the laser to create pores.
Unlike these conventional systems, we focused our study in a very simple, inexpensive, and automatable system. Our system offers several advantages: (i) a core consisting solely of lignocellulosic fibers (handsheet), (ii) insoluble and pHinsensitive encapsulation in the release medium, and (iii) extremely simple production technology. It is a reservoir containing an active ingredient whose release is carried out by diffusion through the porous and microcracked membrane.

Kraft Pulp
The fibrous matrix is made from a commercial Kraft pulp from bleached softwood provided by Kruger Inc.-Wayagamack-Trois-Rivières mill (Canada).

Characteristics and Production of Handsheets
Handsheets were prepared according to the standard method Tappi T205 sp-954. The commercial Kraft pulp is placed into a refiner containing distilled water. After disintegration and defibrillation of the pulp, the mix is introduced into a British handsheet former to obtain a wet sheet. Handsheets are then dried between blotters, using a hydraulic press, and then kept overnight in a room with controlled temperature and humidity (23 °C, 50% RH). Handsheets were tested according to the following PAPTAC standard methods: dryness (A. The tensile test provides several information such as load to the maximum point (kg), energy at break (g/cm), and tensile energy absorption (g/cm). Scanning electron microscopy (SEM) images of the sheet structure were obtained using a JSM-5500 system (JEOL Ltd).

The Polyvinyl Acetate Emulsion
The emulsion polymerization was carried out in a glass reactor composed of a 500-mL container and a lid with 4 openings for mechanical agitation, nitrogen flush, feed, and refrigeration. The reactor's container is immersed in a thermostatically controlled bath at 85 °C and continuous stirring at 720 rpm. After the introduction of 90 mL of deionized water followed by the surfactant sodium dodecyl sulfate Flucka Germany (0.001 mol), we introduced 20 mL of vinyl acetate, i.e., 0.325 mol (Merck Germany) (the monomer). After stabilization of the mixture, 8.4 × 10 −4 mol of the initiator Na 2 S 2 O 8 (Prolabo Chemicals) is added.
Once the polymerization is complete, a part is precipitated in water saturated with sodium chloride, filtered, and washed thoroughly. After the Mark-Houwink constants a and K of vinyl acetate-THF pair are known, the viscosimetric measurements enable the determination of the viscosimetric mass of the obtained polyvinyl acetate.
The average viscosimetric mass of the synthesized polyvinyl acetate is determined by the Mark-Houwink relation (Eq. 1).
where [η] is the intrinsic viscosity and M is the molecular weight. The constants of the Mark-Houwink parameters, a and K, depend on the particular polymer-solvent system. In this study, K = 16 × 10 −5 and a = 0.7, for THF solvent [22].

Production of the Controlled Release Form
From a dried sheet, a parallelepiped structure is cut with the following dimensions: 2 mm length, 2 mm width, and 0.32 mm thickness. The parallelepiped is soaked in a solution of a known concentration of active ingredient (AI) without stirring, to preserve the geometric shape and to avoid dispersion of fibers in the sample. After water evaporation and drying at room temperature, the AI-loaded paper is protected by coating with polyvinyl acetate emulsion without a coalescing agent. In fact, polyvinyl acetate in solution or in the presence of a coalescing agent (e.g., glycol ethers) gives an impermeable film devoid of pores. The release rate, in this case, is zero.
In contrast, the polyvinyl acetate particles of an aqueous emulsion, without a coalescing agent, will juxtapose without merging, obtaining a honeycomb-like structure. The particles are juxtaposed to each other, leaving pores between them and consequently allowing the release of the AI.

Release Studies
The release kinetics of each probe is evaluated as reported in a previous work [23]. The study system consists of a 200-mL vessel, a magnetic stirrer set at 50 rpm, a temperature regulator (37 ± 1 °C), a pH meter, and a sample holder in the form of a porous basket. It allows the substance released from the sample to diffuse into the release bath. The AI is quantified using the Specord 210 Plus UV-visible double-beam spectrophotometer from Analytic Jena (Germany), covering the wavelength range from 190 to 1100 nm, with quartz cells from 'a thickness of 0.2 cm. The concentrations of each AI released are determined at the maximum absorption length (Table 1).

Controlled Release Kinetics Models
The designed delivery systems present planar geometric shapes coated by a polymeric membrane, with watersoluble probes. The release of these AI can be correctly described by at least one of the kinetic models: zero order, Higuchi model, or Korsmeyer-Peppas model. These models can fit the experimental data describing several release profiles.

Zero Order
There is only a diffusion of the probe without decay of the matrix form. This happens with certain transdermal systems, with poorly soluble AIs in coated matrices and osmotic systems, and the equation that governs this release is where Q t is the quantity of AI dissolved at time t, Q 0 is the initial quantity of AI in the solution (often Q 0 = 0), and K 0 is the zero order rate constant in concentration/time units. The results are presented by the cumulative quantity of the released probe as a function of time.

Higuchi Model
It is applied to all porous systems. The expression of the model is given by the simplified equation: where KH is the Higuchi dissolution constant. The data obtained is represented as the percentage of drug release relative to the square root of time. This model can be applied to transdermal systems, matrices, and tablets containing water soluble drugs.

Korsmeyer-Peppas Model
This model has proposed an equation to describe the release of AI from polymeric system and is only valid for the released 60% of AI using the following equation: where M t /M ∞ is the fraction of drug released at time t, k is the release rate constant, and n is the release exponent. To study the release kinetics, log is represented as a cumulative percentage as a function of log time. In this model, the value of n characterizes the AI release mechanism (Table 2).

Matrix Microstructures
This work is devoted to the study of forms intended for the controlled release of different probes. The matrix is composed of delignified kraft cellulosic fibers. To have a large volume and reservoir, we have developed a paper weighting 226.5 g/m 2 , which is higher than the conventional 150 g/m 2 .
The results of the various physicochemical tests carried out on this paper are grouped together in Table 3.
The fibers, with a weighted average length of 2.16 mm, small indices of curls and folds, and a significant hydrophilic character, will create tangles in the sheet in addition to the strong hydrogen bonds. This explains the relatively high resistance of the sample to tensile force.
Handsheets were tested according to the following PAPTAC standard methods: dryness (A.2), thickness (D.3), basis weight (D.4), porosity (D.31), roughness (D.31), tensile test (D.34), and tear test (D.9). The tensile test provides several information such as load to the maximum point (kg), energy at break (g/cm), and tensile energy absorption (g/cm); the results are shown in Table 4.  The high moisture content, in the order of 10%, reflects their strong hydrophilic character, which is the capacity of adsorption of water with hydroxyl groups. The SEM image (Fig. 2) shows a state of entanglement and the presence of empty space between the fibers, confirming the high porosity and roughness of the samples. The presence of loops and folds along the fibers confirms the rugged nature of the material's surface. We can also notice on the SEM image that the micro-fibrils form bridges between the macro-fibers.
Applying a uniform coating, from a thickness point of view, is not easy. A section through the polyvinyl acetatecoated sheet shows a non-uniform deposited thickness. On the other hand, not adding coalescing agents, generally glycol ethers, which allow the polymer particles to coalesce, will result in a film with microcracks. These cracks will allow the passage of small molecules (solvent, probe, etc.). The addition of suitable coalescing agents makes it possible to overcome the so-called "honeycomb" stage and the disappearance of the boundaries between the particles (Fig. 3). In other words, the formation of an impermeable film will not release the AIs.
The interstices between the particles will allow the diffusion of the solvent towards the interior of the matrix, the solvation, and the release of AI towards the outside (Fig. 3).

Analysis of Release Results
The released amount of each AI, by the controlled release form, is followed by UV spectrophotometry at the maximum wavelength absorbance of the AI (Table 1). At a predetermined time, 4 mL of solution is withdrawn from the release medium, and immediately replaced by the same dissolving liquid.
The probe molecules used in this study are polar with ionizable groups. In the cellulosic matrix, they will interact with the hydroxyl, carboxylate, and sulfonate groups present on the surface of the fibers. The interactions will be essentially in the form of hydrogen bonds. In addition to these strong interaction forces, these molecules must cross the polymeric membrane to diffuse outside the matrix form. The kinetics of diffusion through the protective layer in these systems is a function of the thickness of the polymeric layer, the presence or absence of pores, the pH, the volume of the probe, and its diffusion coefficient. Studies have shown that these coated reservoir systems (i) often allow zero order release as long as AI exists, (ii) have difficulty releasing high molecular weight AIs, (iii) can occur a risk of toxic overdose due to a production error or accidental breakdown of the polymer layer, exposing the contents of the reservoir to the external environment of the system, and (iv) are considered relatively complex and expensive systems [24,25].
The data obtained was studied using the most common models (zero order, Higuchi, and Korsmeyer-Peppas). The release evolution of each active ingredient, as a function of time at different pH, is illustrated in Fig. 4. It emerges from the examination of these two figures, percentage of release as a function of time (Fig. 4a, b) that each AI has its own release curve. At the beginning, except for the Ketoprofen, there is a kind of burst effect, in each curve. It is not an instant release; it is only the AI present in the protective membrane. Indeed, during the immersion of the paper containing the AI, there is migration of the AI when the paper is immersed in the emulsion. In both media, acidic and basic, it can be seen that the release continues at decreasing rates through time, except for ASA, where the release rate is constant. Considering both pH studies, it is possible to observe two stages of release. A first stage with a sharp slope for a period of 1 h followed by a second period with a milder slope, characterizing a slower release speed. As for the burst effect, the first release is certainly related to active molecules located at the membrane level. The AI trapped in the paper requires more time to leave the matrix. The water must first wet them and dissolve them. Once hydrated and under the action of external osmotic pressure, these molecules will be able to exit. Obviously, they will not come out at the same speed, as it is shown by the different release curves. Many parameters will take part in the release rate: the ionization state, the hydrodynamic volume/degree of solvation, the porosity, the interactions with both the fibrous network and the membrane, the thickness of the membrane, and its degree of hydration. Thus, the order of release, in both acidic and basic media, is SA > TCC > MCP > ASA > KP. The highest release rate observed in the case of AS can be attributed to the smaller size of the molecule and its solubility, especially in basic medium. At a pH of 1.3, the protonated form of AS will predominate and allow the possibility of establishing hydrogen bonds with the fibers. This explains its slower release compared with alkaline medium. For other AIs, their solubility in the medium is certainly a determining factor. For both pH media, it shows that there are two stages of release. A first stage with a sharp slope for a period of 1 h, followed by a second period with a milder slope characterizing a slower speed. As for the burst effect, the first release is certainly that of the active molecules located at the membrane level. The active ingredients trapped in the paper require more time to leave the matrix. The water must first wet them and dissolve them. Then, these molecules, once hydrated under the action of external osmotic pressure, will be able to exit. Obviously, they will not come out at the same speed as the different curves show. There is the ionization state, the hydrodynamic volume in other words the degree of solvation, the porosity, the interactions with both the fibrous network and the membrane, the thickness of the membrane, and its degree of hydration.
Thus, the order of release, in both acidic and basic media, is SA > TCC > MCP > ASA > KP ( Table 5). The highest release rate observed in the case of AS can be attributed to the smaller size of the molecule and its solubility, especially in basic medium. At a pH of 1.3, the protonated form of AS will predominate and allow the possibility of establishing hydrogen bonds with the fibers. This explains its slower release compared with basic medium. For other AIs, their solubility in the medium is certainly a determining factor. Indeed, the two probes which have lower solubilities are the least released in the medium. AI volume, apparently, is not involved since the larger colchicine is released more easily. The molecular weight, the porosity of the membrane, and especially the degree of solvation, linked to the presence of polar groups which dress the molecule, determine the rate of release. Indeed, the two probes which have low solubilities are the least released in the medium.
The ionization state depends on the pKa of the probe and the pH of the medium in which it is in solution. pKa is defined as the pH at which a compound is present at 50% in ionized form. The [ionized]/[nonionized] ratio is defined by the Henderson-Hasselbalch equations: The acidic AIs (ASA, SA, and KP), in solution in a medium where the pH is equal to 1.3, will be in the protonated form while alkaline medium will be in the form of an anion. The protonated form can establish hydrogen bonds with the cellulosic fibers of the paper whereas the basic medium will be hydrated but less bound to the fibers. Another factor which must be taken into account is that these molecules in their protonated form have solubility parameters closer to polyvinyl acetate (Table 5). Also, they can impregnate the membrane and be retained as plasticizers. This explains the higher release rate in basic medium. As regards the other two probes, MCP and TCC, in acidic and alkaline medium, will be protonated and therefore it is their solubility which will condition their release. The  (Table 5; Fig. 4).

Release Mechanism
The use of the Higuchi (Fig. 4a, b) and Korsmeyer-Peppas (Fig. 5a, b) models will allow us to have more information concerning the release mechanism of the different AIs. The planar shape of the designed matrices, enveloped by a polymeric film, must describe a release by these two models.
Also, they were used to represent the obtained experimental release data. As shown in Figs. 4 and 5, the variation is almost linear with a correlation coefficient almost equal to unity (R 2 ≈ 1) ( Table 6). This is valid for the two quantities considered, % release or Ln (% release), versus t 1/2 or Ln (t). We have grouped the values of the rate constants, the correlation coefficients, as well as the Korsmeyer-Peppas n value in Table 6. These delivery systems are in accordance with Higuchi's conditions: (i) the core is composed of lignocellulosic fiber coated with a dispersion of polyvinyl acetate particles, both inert in the delivery medium, (ii) the AIs are soluble and small in size, (iii) diffusion will occur through pores and cracks in the polymeric wall, (iv) the matrix contains an initial drug concentration much higher than the drug solubility, and (v) seen in the dimensions of the shape used, it can be said that the diffusion is unidirectional. Also, the fact that the release of active ingredients is linear as a function of time, using the Higuchi model, shows that the active ingredient is released by diffusion at a constant rate.
Lignocellulosic fibers alone cannot regulate the release of an active principle. It is a very porous and polar fibrous network. It can only retain the probes by Van der Waals interactions and by hydrogen bonds. It is the polymeric membrane that is responsible for controlling the release. With the Korsmeyer-Peppas model, we can determine the nature of the phenomenon involved in the release. It allows the mechanism to be determined when more than one type of active principle release phenomenon is involved [28].
Depending on the value of n, we can say that the release is performed according to Fick's law or not. If the plot of Ln (% release) as a function of Ln (t) gives a value of n close to 0.5, it can be stated that the release is governed by diffusion. In the case where n is close to 1, we are dealing with a non-Fickian release and the release of the probe follows a kinetic of zero order. When 0.5 < n < 1, there is an abnormal transport (non-Fickian); the drug release mechanism is governed by diffusion and swelling. Examination of the figures (Ln (% release) as a function of Ln (t)) shows lines with different slopes. We have reported the values of the n obtained, for the two pH values, in the table (Fig. 6).  The values of n obtained by the Korsmeyer-Peppas kinetic model lie in the range [0.  and this is for both pH values. These values indicate that the release mechanisms of the different probes are not the same. There is mainly the influence of the nature of the probe and the pH. This, of course, assuming we are working with the same membrane thickness in all forms of release. There are two groups of values of n reported in Table 6. Those which are less than 0.5, both in acidic and alkaline medium, with a correlation coefficient close to unity, indicate that the mechanism which controls the release of the two active ingredients MCP and TCC is a diffusion mechanism controlled by the membrane. These two AIs do not carry an acid function and have very high pKa compared with the other probes. At release pH values, they will not be protonated, and it is their solubilities that will compromise their releases. Also, only the porosity of the membrane will have an influence on the regulation.
The other probes are organic acids, of similar pKa but differ in their solubilities in water and their ionization states in different pH media. Their n values are between 0.5 and 1, and according to Korsmeyer-Peppas, the mechanism involved in their release is diffusional. It is managed both by the porosity of the membrane and the interactions between the probes and the cellulosic fibers.
The thickness of the membrane plays an important role in regulation. It is a juxtaposition of polymeric particles. Their arrangement in space will influence the porosity of the membrane. Also, to clarify this evidence, we developed forms of release by increasing the thickness of the membrane by immersing the paper containing AI in a more concentrated emulsion. The results obtained are shown in Fig. 7a, b. It is clear that the release rate is lower in the case of the double layer. This release rate is proportional to time and depends on the pH. The lines obtained have correlation coefficients R 2 , respectively, for pH 1.3 of 0.9912 and 0.9840 for pH 8.3. There is also the fact that the probes (SA, ASA, and KP), having solubility parameters (δ) closer to polyvinyl acetate, can have interactions or even impregnate the membrane and this will slow their release, which is of zero order.

Conclusions
Our system, paper coated with a polyvinyl-acetate (PVAc) emulsion film without a coalescing agent, has proven to be a promising matrix to achieve controlled release of different IAs. It needs only to adjust the thickness of the membrane. The process is simple and easy compared with other recently developed extended release matrices.
In this study, different parameters were studied for each active principle such as the pH of the dissolution medium, the thickness of the protective membrane (PVAc), the ionization state of the probe, and its solubility parameter. Various kinetic models tested showed that the release mechanism is essentially diffusional and the release rate of the various AIs is close to zero order.
The simplicity of our system makes it possible to envisage numerous applications: (i) in the agricultural field as a system delivering in a controlled manner fertilizers, pesticides, fungicides, and vitamins; (ii) in the medical field, they can be used as patches for drug delivery by the cutaneous or transdermal route; and (iii) the production of these matrices can be automated.