The adhesion mechanism of mucoadhesive tablets with dissimilar chain flexibility on viscoelastic hydrogels

Abstract

Mucosal membranes with strong variability in their viscoelastic properties line numerous organs and are often targeted by mucoadhesive formulations, e.g., highly swellable hydroxypropylmethylcellulose (HPMC) and slightly cross-linked poly(acrylic acid) (PAA) tablets. Although the factors determining the strength of mucoadhesion are hierarchical and affected by both reversible and irreversible processes, the currently available strategies generally view mucoadhesion as the individual performance of the mucoadhesive excipient. We propose an integrated concept that considers the viscoelasticity and tensile properties of both the adhesive interphase and the bulk phases. To reduce the complexity of the mucosal membrane and eliminate the effect of specific macromolecular interactions, we studied the adhesion on mucosa-mimetic freeze/thawed (FT) poly(vinyl alcohol) (PVA) hydrogels. Their viscoelastic properties were controlled by the number of FT cycles and the polymer concentration. The adhesive strength of HPMC tablets displayed a pronounced dependence on the viscoelasticity of PVA gels, explained by the limited chain flexibility and interpenetration of HPMC, resulting in the formation of a thin the adhesive interphase compared to PAA. We recognized scaling laws between toughness and strength for tensile and adhesive properties as well as general correlations between viscoelastic and adhesive properties, which can aid the more rational design of both mucoadhesive formulations and mucosa-mimetic materials.

Graphical abstract

Highlights

• •Freeze/thawed poly(vinyl alcohol) hydrogels were used as mucosa-mimetic surfaces.
• •Relationship between the viscoelasticity and adhesion of swellable tablets was mapped.
• •Scaling laws were found between toughness and strength for adhesive properties.
• •Weak interpenetration of rigid polymer chains limits the adhesion performance.

1. Introduction

Mucoadhesion is a specific type of bioadhesion, in which the polymer excipient of the drug formulation adheres to a mucosal membrane surface [1,2]. A prolonged retention of the formulation and therefore more efficient drug release is expected by this administration route, which is further facilitated by the avoidance of the metabolic effect of the liver and the stomach [3]. Despite the widespread application and development of mucoadhesive solid and semi-solid dosage forms [[4], [5], [6], [7], [8], [9]], there are no standardized and generally accepted in vitro protocols for screening of mucoadhesive performance.

The direct measurement of mucoadhesion is based on mechanical tests, e.g., tack measurement on ex vivo mucosal tissue samples, which carries a lot of measurement uncertainty originated from the natural biodiversity and differences arising during preparation and storage of animal samples. Furthermore, ex vivo tests have strict lab requirements and high costs. To replace animal tissues, mucosa-mimetic materials offer a solution with improved robustness and ease of use [10,11]. The first in vitro mucus models based on animal sources, e.g., lyophilised porcine dermis or tanned leather, displayed significant limitations and generally miss the viscoelastic and sticky nature of physiological mucus. Later on, the tissue mimetic properties of hydrogels were recognized, e.g., viscoelasticity, high water content, similar mechanical behaviour, and novel hydrogel-based mucosa mimetic materials as substrates were introduced for testing adhesion. Now, they can mimic the adhesive performance of certain dosage forms measured on excised mucosal membranes. Vitaliy et al. successfully mimicked the mucoadhesion of hydroxypropylmethylcellulose (HPMC) and cross-linked poly(acrylic acid) derivative (PAA) tablets on porcine buccal mucosa using a 2-hydroxyethylmethacrylate (HEMA) N-acryloyl glucosamine (AGA) copolymer hydrogel as a substrate [12]. In another work, they developed a method for testing thermogelling semi-solid dosage forms using the previously mentioned HEMA-AGA substrate [13]. Eshel-Green et al. designed poly(ethylene glycol) diacrylate hydrogels containing free thiols (PEGDA-QT) to study the effect of thiolation of polymers compressed to tablets for testing. The free thiol groups were included in the hydrogel substrate to mimic the – SH groups of the cysteine-rich subdomains of the mucin glycoproteins [14]. They reproduced the improved adhesion properties of thiolated formulation measured on porcine small intestine samples by using their synthetic hydrogel surfaces. Nevertheless, these works reported limited analysis of the possible mechanism of adhesion and the shape of adhesion curves is not discussed in detail. Therefore, the conclusion might not be general enough to understand the role of mechanical properties of substrates on adhesive failure and to develop mucosa-mimetic materials for other types of formulations resulting in the application of highly different models across the literature [1,10,15].

The rational design of more general mucosa-mimetic models requires the knowledge of driving forces of adhesion, which are not only the reversible macromolecular interactions [16], but also irreversible, time-dependent factors can significantly affect the adhesion [17,18]. A deeper analysis of adhesive failure generally lacks in the topic of mucoadhesion, but a more comprehensive approach can be taken from related fields. General adhesives (tapes), adhesive gels and surgical adhesives all share similar features, e.g., the determining role of viscoelastic properties in their adhesion and the complex interaction of adhesives, substrates and water [[19], [20], [21]]. The importance of chemical functionalities on mucoadhesion is possibly overemphasised, and physical interactions of the polymer chains at the interface may have greater importance in many cases [17,[22], [23], [24]]. Although interpenetration or polymer interdiffusion at the polymer-mucus interface is known [24,25], but this phenomenon has not been examined sufficiently, especially its consequences to the macroscopic adhesion. Accordingly, a controlled change of viscoelastic properties of the substrates and exploring the correlation between adhesive and viscoelastic properties could significantly improve our knowledge and understanding of mechanisms and key factors in mucoadhesion of dosage forms. This can contribute to the more rational design of mucosa-mimetic materials providing standardized measurement of mucoadhesion.

Our recent results demonstrated the importance of electrostatic interactions in the adhesion of chitosan tablets on mucin-containing hydrogels, in comparison with the adhesion of hydroxypropylmethylcellulose (HPMC, K15M) and poly(acrylic acid) (PAA, Carbopol® Ultrez 10 NF) tablets on the same hydrogel surface [26]. In the current work, we aimed to explore the role of chain flexibility in the adhesion of tablets on hydrogels, without the complexity of mucin-polymer secondary interactions. To this end, we prepared mucosa-mimetic hydrogel substrates of poly(vinyl alcohol) (PVA) to exclude specific secondary interactions. PVA was chosen because of its biocompatibility and the capability to form physical hydrogels with long-term stability under mild synthesis conditions. Its gelation is based on the so-called freeze-thaw (FT) process that enable us to preserve the functionality of mucin or other sensitive biomacromolecules in the resulting hydrogel, thus providing possibilities for future development of other mucosa-mimetic materials. Our central hypothesis is that the strength and toughness of adhesion for HPMC tablets significantly change with gel stiffness because of the limited interpenetration of rigid polysaccharide chains into the adhesive interphase. To test this hypothesis in a reliable way, we also studied the adhesion of slightly cross-linked PAA tablets with more flexible chains. To this end, we prepared mucosa-mimetic hydrogel substrates of poly(vinyl alcohol) (PVA) by freeze-thaw (FT) method. Their mechanical properties were modulated, on the one hand, by varying the number of the FT cycles with keeping the polymer concentration constant; on the other hand, by varying the polymer concentration with keeping the number of the FT cycles constant during the preparation of the gels (referred altogether further as synthesis conditions). The precursor PVA solutions and the PVA hydrogels were characterized comprehensively for their viscoelastic features, using both rotational and oscillatory rheology at small strains. The mechanical properties of the gels, limiting the adhesion properties, were also studied by applying large deformations in tensile tests. The adhesive performance was thoroughly analysed, general correlations between the mechanical properties and adhesion measured on hydrogels were found, and a distinct behaviour of the two mucoadhesive polymers was identified and explained with their dissimilar chain flexibility and the mechanical properties of the PVA hydrogels.

2. Materials and methods

2.1. Materials

Poly(vinyl alcohol) (PVA) (Mowiol® 10-98, Mw approx. 61 000 g/mol, degree of hydrolysis: 98.0–98.8 mol%) was purchased from Merck. Carbopol® Ultrez 10 NF, a slightly cross-linked poly(acrylic acid) derivative (PAA) from Lubrizol Advanced Materials Europe was kindly provided by Azelis Hungary Ltd. Hydroxypropylmethylcellulose (HPMC) Benecel™ K15M PH CR (M_w = 575 kDa) as a product of Ashland Inc. is used. It is a “K” type (2208) HPMC commonly used in controlled release applications [27]; it has 20−24 % methoxy („22″ from its identification number; degree of substitution, DS = 1.5) and 7−12 % hydroxypropoxy content („08″ from its identification number; molar substitution, MS = 0.25). The 15M stands for the range of the dynamic viscosity, which is 15 × 10³ mPa‧s (2 wt%, water, 20 °C) with a 75 %–140 % acceptance criteria, for the current product it is 18 × 10³ mPa‧s. All the information above is ascertained according to the USP standards and provided by the producer. Further specification for HPMC is shown in Table S1. Ultrapure water (ρ > 18.2 MΩ cm, Millipore) was used for the preparation of the solutions and the hydrogels. All experiments were performed at 25 °C unless otherwise stated.

2.2. Synthesis of hydrogels

PVA hydrogels were prepared by the freezing-thawing (FT) method. PVA was dissolved in water (20 wt%) at 80 °C under stirring for approximately 18 h. This stock solution was used to prepare diluted solutions, which were stirred for 8 h (80 °C) to reach complete homogeneity. The resultant PVA solutions were poured onto a glass plate bordered with a 4 mm thick silicone frame and covered with another glass plate. PVA solutions were converted into hydrogels by using 1 to 5 consecutive FT cycles, each of which consisted of a freezing step at −20 °C for 18 h and followed by a thawing step at room temperature (25 °C) for 6 h.

2.3. Tensile testing of the gels

Cylindrical FT PVA gel specimens were prepared for tensile tests according to Section 2.2. except that the polymer solutions were poured into glass tubes with an inner diameter and height of 10 mm, and the tubes were then covered with glass plates during gelation. The resultant hydrogels were fixed to the bottom and the upper plates of an Instron 68SC-1 instrument. To secure the gel samples on the plates, a layer of double-sided adhesive tape was used on both sides, and a small amount of Na₂CO₃ powder and cyanoacrylate glue were added between the gel and the tape, on both tapes. After curing of the glue (3 min), the gel was pulled apart with a deformation rate of 0.125 mm/s, and the force was measured as the function of distance using a 50 N measuring cell.

2.4. Tablet preparation

Tablets for testing adhesion were prepared by direct compression of polymer powders. The tablets were pressed using a Dott. Bonapace auto tablet press machine (CPR-6). The compression force was recorded via the inbuilt measurement cell of the tablet press. The crushing strength of the tablets was assessed using a tablet hardness tester (Dr. Schleuniger THP-4 M). The tablets had a 7.0 mm diameter with biconvex geometry characterized by its highest (H) and smallest (L) thickness. The most important properties of the tablets are summarized in Table S2, in which the successful preparation of tablets is shown by hardness values of 50 N and higher for all compositions.

2.5. Swelling of the tablets

The tablets were swollen in abundant water in Petri dishes and weighed after 1, 3, 5, 10, 15, 30, 45, and 60 min. The excess water was removed from the surface of the tablets with a lint-free tissue paper before measuring their weights on an analytical balance.

2.6. Rheological analysis of the gels and swollen tablets

Rheological measurements were performed with an Anton Paar Physica MCR301 rheometer in oscillatory mode using a Peltier device to keep the temperature at 25.0 ± 0.1 °C. The tablets after 1 h of swelling and the hydrogels were measured with a plate-plate geometry (PP-25; sample gap: 2.8 mm for the gels, 1.3 and 2.2 mm for the swollen PAA and HPMC tablets respectively). Dynamic moduli (storage, G’ and loss, G” modulus) were recorded at a strain (γ) of 1 % in the angular frequency (ω) range of 500–0.5 rad s⁻¹ with measuring 5 data points per decade.

2.7. Adhesion measurements

The adhesion of the tablets on PVA hydrogels was measured using a mechanical tester (Instron 68SC-1) with a circular holder for the PVA hydrogels. A 25 mm diameter disk was cut from the 4 mm thick gel sheet for each test, and clamped into the circular holder which has a 15 mm diameter cavity in the upper mount. The tablet was attached to a cylindrical upper probe (10 mm diameter). To ensure the sufficient adhesion of the tablet on the upper probe, a cylindrical (10 mm diameter) piece of laboratory filter paper was attached to one side of a double-sided adhesive tape. The tablet was then fixed to the filter paper using a few drops of a cyanoacrylate glue, and after complete curing (∼4 h), the whole system was attached to the upper probe of the mechanical tester using the other side of the double-sided tape. For testing adhesion of the tablets on the hydrogels, a 10 N measurement cell was used, with the following three-step program: compression at a deformation rate of 0.01 mm s⁻¹ until reaching a compressive force of 0.1 N, holding the force at 0.1 N for 60 s, and debonding at a deformation rate of 0.05 mm s⁻¹. The force (F) was recorded as a function of displacement and transformed into nominal stress (σ) – strain (ε) curves according to Eq. (1) and Eq. (2)

$$\sigma =\frac{F}{A_0}$$ (1)

$$\epsilon =\frac{h-h_0}{h_0}$$ (2)

where A₀ is the cross-sectional area of the accessible surface of the gel, h₀ is the initial height of the hydrogel and h is the height of the deformed gel (including the interphase) at the given point of the measurement.

To compare the curves, we determined the peak stress (σ_max), the maximum deformation until failure, i.e. the elongation-at-break (ε_max) and the toughness of adhesion (area under the curve, see Eq. (3)) determined [17].

$$Toughness={\int }_0^{{\epsilon }_{\mathit{max}}}\sigma \left(\epsilon \right)d\epsilon$$ (3)

3. Results and discussion

3.1. Synthesis of the gels and their mechanical characterization at small deformations

Physically cross-linked PVA hydrogels were prepared as robust model substrates based on the robust model substrates studied previously in our group [26]. Viscoelasticity is a substantial material property in adhesion as demonstrated for various soft materials and pressure-sensitive adhesives. The effect of the viscoelasticity as well as that of mechanical properties at large deformations were determined to identify general correlations between mechanical properties of the hydrogels and the adhesion of polymer tablets [18]. It is worth noting that the adhesion is not only determined by the polymer excipient, but the complex system of the hydrogel, the adhesive interphase and the tablet must be considered. Using this PVA-based hydrogel model enabled us to closely monitor the effect of viscoelasticity on adhesion, without the uncertainty of ex vivo mucosal samples. The mucosal membranes in different organs, species, and even individuals can differ hugely in their viscoelastic properties, which can significantly affect the adhesion, making the comparison of results impossible [18,28]. Nevertheless, the viscoelasticity of PVA hydrogels can be tuned by varying the preparation conditions, which may later enable us to mimic the mechanical properties of certain mucosal surfaces specifically [29].

Mucosa-mimetic materials are usually composed of 1) a polymer responsible for developing the gel structure in aqueous medium, characteristic for the native mucus covering the mucosal membrane, 2) large amount of water and 3) one or more biosimilar components to mimic the chemical functionality of the mucin as the main protein component of the mucus [15,26,30,31]. To preserve stability and functionality of the biosimilar components, chemical reaction during gel formation should be avoided if possible. Physical crosslinking may not result in the formation of hydrogels with sufficient thermal stability or requires the addition of ions in high concentrations [30], potentially affecting the physico-chemical properties of entrapped components. Previously, we used poly(vinyl alcohol) (PVA) as gel forming polymer to overcome such limitations and also because of the cost-efficient synthesis of PVA hydrogels, which can be important later if one thinks of a large-scale production of such hydrogels for standardized testing methods. Mucin as biomimetic component was included in the hydrogel to identify the possible electrostatic interactions during adhesion. In this project a simplified PVA hydrogel model is used to separate the effect of viscoelasticity, which is sporadically mentioned in the literature of mucoadhesion. The aqueous solution of PVA turned into a hydrogel in the FT process, and the physically crosslinked network was held together by microcrystallites [32] as cross-linking points in the polymer network resulting considerable thermal stability. These crystallites form upon freezing as a result of a spinodal liquid-liquid phase separation of the aqueous PVA solution and crystallites remain stable during thawing at room temperature [33]. The crystallinity, thus the mechanical properties of the hydrogels, can be further enhanced by consecutive freezing-thawing (FT) cycles [34].

First we characterized the concentration-dependent viscosity of the aqueous PVA solutions to identify any microstructural changes with concentration, potentially affecting the gel structure as well. The lowest concentration was chosen to be 7 % as this is the range of critical entanglement concentration (C_e) for the molar mass of the PVA used (see Eq. (4)) [35] and it is highly expected that chain entanglements are strongly required to achieve considerable adhesion.

$$C_e=\frac{\rho M_e}{M_w}=\frac{1.27\times 3750}{61000}=7.8wt\%\text{,}$$ (4)

where ρ is the density of PVA, M_e is the entanglement molecular weight and M_w is the weight-average molecular weight of the PVA.

PVA solutions showed close-to-Newtonian behaviour at concentrations less than or equal to 13 % (Fig. 1 a), whereas remarkable shear-thinning was observed at higher concentrations with a Newtonian plateau at low shear rates. Zero-shear viscosity values were used for comparison (Fig. 1 b) and two different regimes can be clearly identified on log-log representation of viscosity-concentration data points. The slope of the first linear line is 4.0, which is near 3.9, the exponent of the scaling law between viscosity and concentration for semi-dilute entangled polymer solutions [36]. Interestingly, the scaling exponent changed to around 6.1 at concentrations higher than 13 %, which might be a transition from semi-dilute to concentrated solution resulting in a strongly entangled structure. This concentration range is in good agreement with the finding of Norton et al. [37], who observed a significant change in mechanical properties at high concentrations, indicating a change in the microstructure of the gels.

Fig. 1.

Characterization of viscosity of PVA solutions. a) Viscosity curves of the precursor PVA solutions b) zero shear viscosity values of PVA solutions as a function of concentration.

Semicrystalline polymers often display light scattering observed as opaque or hazy appearance. Similarly, all PVA gels synthesized here were opalescent to various extent due to the formation of microcrystallines acting as cross-linking points [38]. The apparent absorbance of the gels at 500 nm (characteristic for the opacity) significantly increased after one FT cycle compared to the precursor solution (Fig. 2 a). An even larger change in absorbance was detected between the first and second FT cycle in correlation with the massive increase in G’ indicating a major change in the structure of the hydrogel including the serious increase of crystallinity. The further increase in the number of FT cycles resulted in a plateau of apparent absorbance indicating that the major part of crystallization occurs in the first two cycles. The increase of polymer concentration also caused a gradual increase of the apparent absorbance, particularly at lower concentrations (Fig. 2 b). Nevertheless, the effect of polymer concentration on absorbance at 3 FT cycles seemed to be small in this range in comparison with the effect of FT cycles, which implies that crystallinity of hydrogels with different polymer concentrations is similar, though its accurate and absolute measurement is troublesome by the conventional methods (X-ray diffraction, XRD or differential scanning calorimetry, DSC) due to its very low value [39].

Fig. 2.

The change in opacity of the FT PVA gels shown by the apparent absorbance values at 500 nm; as a function of a) the number of FT cycles and b) PVA concentration.

Viscoelastic materials with comparable elastic and viscous character, e.g., polymer networks in water near to their sol-gel transition can be accurately characterized by oscillatory rheology in the linear viscoelastic region (LVE), i.e., at small, reversible deformations. During the periodic deformations, the storage (G′) and the loss modulus (G″) are measured, characteristic of the elastic response and viscous response of the matter, respectively. A gel structure displays G′ value higher than G″ over a wide frequency range and a constant value of G′ at low deformation rate (frequency). In this case, the elastic response is characteristic for the permanent structure (infinite relaxation time) determined by cross-linking density and not affected by possible polymer chain entanglements or other time-dependent processes. As an example, PVA hydrogels with a polymer concentration of 13 % in their precursor solution showed G′ over the G″ even at low frequencies, indicating the presence of the gel structure (Fig. 3 a). In the absence of physical entanglements, in ideal networks [40], G’ and G” values are independent of frequency. The increased moduli of the gels at high frequencies suggests the presence of an entangled network, in addition to the cross-linked, microcrystalline structure, because polymer entanglements contribute to the elasticity of the network as the chains are unable to untangle within short time (or at high frequency).

Fig. 3.

Typical viscoelastic properties and thermal stability of FT PVA gels shown on a gel with 13 % of PVA concentration, after 3 FT cycles; a) frequency-dependent and b) temperature-dependent dynamic moduli of the gel and the precursor solution (at 1 rad s⁻¹).

Thermal stability of the PVA hydrogels was studied at constant strain and frequency and the gel to sol transition was monitored by the change of dynamic moduli (G′ and G″). The exemplar hydrogel with 13 % of PVA concentration after 3 FT cycles proved to be thermally stable and maintained a gel structure up to 50 °C indicated by the G’ value over the G” and the constant value of both over a wide temperature range (Fig. 3 b). The gel to sol transition was indicated by the fall of both moduli and the presence of an intersection at 60 °C, above which viscous character dominates over elastic one. Final values of moduli at temperatures higher than 70 °C are equal to those of the precursors solution suggesting the complete disruption of the gel structure and the formation of a viscoelastic liquid. These temperature values agree with the findings of Urushizaki et al. [41] from similar measurements and below the T_g (glass transition temperature) of solid PVA (70–80 °C) [42], attributed to the high water content of PVA gels [43,44].

PVA hydrogels with a wide range of viscoelastic properties were synthesized for further studies by varying polymer concentration in the precursor solutions and the number of FT cycles (altogether synthesis conditions). The polymer concentration was varied as follows: 7, 9, 11, 13, 15, 17, and 20 wt% with keeping the number of FT cycles at 3; the number of the FT cycles was varied between 1 and 5 in the second experimental series. The polymer concentration was kept at 15 % for which we observed excellent reproducibility for rheological and adhesive properties in our previous work, while significant change in those properties as a function of cycle number [26]. Accordingly, the strength of the gel network was sufficient to retain the original shape of the gel under its own weight for all compositions except the gels with only 1 FT cycle, where visible deformation was observed but the gel remained stable in this case as well. In agreement with the visual observation, all the gels showed G′ over the G” at a wide frequency range tested here, even at low frequencies, indicating the presence of a gel structure (Fig. 4a and b). The frequency-dependence of moduli was experienced for all samples, but it was more pronounced at lower number of FT cycles, especially for one FT cycle. The effect of the number of FT cycles on mechanical properties is demonstrated by comparing the storage modulus of the gels at a chosen, low frequency corresponding to the relatively low deformation rate during adhesion tests [17]. The storage modulus remarkably increased between the first and the second FT cycle, and then levelled in the subsequent FT cycles (Fig. 4 c). The increase of PVA concentration also resulted in the increase of G’ (Fig. 4 d) but in a steady and gradual manner without reaching a plateau value. As an additional characteristic of viscoelastic properties, degree of the phase shift (δ) of the mechanical response is also shown in (Fig. 4c and d) All hydrogels displayed δ values lower than 3° suggesting a highly elastic network. In addition, some viscous character was also observed, which is important for the energy-dissipation during adhesion. Increasing either PVA concentration or the number of FT cycles reduced the δ values to around 1° indicating the diminishing of viscous character (Fig. 4c and d). These changes can be attributed to both the change of crystallinity and the structure of the amorphous phase in the hydrogels.

Fig. 4.

Viscoelasticity of the FT PVA hydrogels with varying the synthesis conditions. Frequency-dependent moduli of hydrogels prepared with a) different number of FT cycles and b) different polymer concentrations (only selected data is shown, for the full version please see Fig. S1.) Change in characteristic viscoelastic properties at a specific frequency with a) different number of FT cycles and b) different polymer concentrations. Relaxation time spectra with a) different number of FT cycles and b) different polymer concentrations.

Dynamic moduli obtained from frequency sweep measurements were converted to relaxation time spectra in which relaxation modulus is plotted against the relaxation times. Characteristic relaxation times, i.e. peaks on the spectra give information on the time ranges of microstructural processes that account for stress relaxation. Relaxation processes related to entanglements and smaller units (Kuhn units) are above the time (frequency) range studied [35,36], but reptation movement of larger units, e.g., full chains can be detected in this time window. The peak around 0.5 s that spans over a time decade and is present for all the gels (Fig. 4e and f), which is in good agreement with the relaxation time observed for semidilute entangled PVA solutions. This peak is assigned to the movement of a mesh composed of several full chains which are connected by hydrogen bonds as the reptation of a single chain occurs on a much shorter time scale [35]. The increase of PVA concentration to 15 % or higher resulted in the appearance of shorter relaxation times around 0.05 s. The presence of multiple peaks indicates the formation of relaxation units of different sizes and the lower limit of this size range might even be related to the reptation of single chains. The characteristic change around 13–15 % PVA concentration in the gel coincides with the transition in zero-shear viscosity as a function of polymer concentration and might also have consequences on the mechanical properties of PVA hydrogels as studied in forthcoming chapters. In addition, a less pronounced but visible peak in the same time interval appears for gels prepared with at least three FT cycles, suggesting that the increase of cycle number also caused the formation of relaxation units of different sizes [37,45].

3.2. Tensile testing, applying large deformations on the gels

The testing of adhesion of solid polymer tablets is always accompanied by the deformation of the viscoelastic substrate. As a result of the soft and viscoelastic character of the latter, the mechanical stress leads to a large deformation and significant energy dissipation, measured as a part of the irreversible work of adhesion. The dissipated energy can make up a major part of the irreversible work of adhesion [46], therefore it is of high significance to study the mechanical performance of the viscoelastic substrate, which contribute to the accurate interpretation of the adhesion.

The tensile curves of PVA gels with various polymer concentrations or number of FT cycles are shown in Fig. 5a and b, respectively. A close to linear relationship between stress and deformation was observed in the initial part of the curves regardless of the composition, with an increasing modulus with polymer concentration and cycle number in accordance with the tendency of storage modulus values determined by rheology (Fig. 4c and d). Measurements were highly reproducible indicating the robust fixture of the gels, which enabled us to accurately characterize the mechanical failure of the unnotched bulk gel instead of any weak point of the fixture. A clear maximum in stress was observed for the different hydrogels, corresponding to the initial stress of the fracture, followed by a stepwise reduction in stress with the further increase of deformation, which agrees with the visual observation of the formation of fibrillary structure of the hydrogels at large deformation followed by their breakage. This kind of failure mechanism is characteristic of gels with the ability to dissipate a high amount of energy through breaking of dynamic bonds, e.g., secondary chemical bonds or chain entanglements. The continuous re-formation of these bonds increases the plastic zone around the tip of the crack, generally leading to a less rigid failure and a higher elongation at break than that of chemically crosslinked gels [47]. The modulus increased with the number of FT cycles and levelled off after the 4th cycle (Fig. 5 c), similarly to their storage modulus, which may imply that crystallinity did not increase further in the last cycle. A dissimilar tendency was found as a function of polymer concentration as the modulus did not reach a plateau value even at the highest polymer concentration (Fig. 5 d). The elongation-at-break remained almost constant as the function of the number of the FT cycles (Fig. 5 c), with a very slight increase between the 4th and the 5th FT cycles. In contrast, a substantial increase was detected in elongation-at-break as a function of the polymer concentration after a threshold concentration of around 15 % (Fig. 5 d). We hypothesise that the mechanical behaviour at large deformation can be explained by the amorphous phase of the hydrogels, not the crystalline one as the crystallinity might be similar after 3 cycles at different polymer concentrations or at the same concentration with further increasing the cycle number (see absorbance data in Fig. 2). Furthermore, the characteristic change in elongation-at-break corresponds to the PVA concentration of around 13 % which agrees well with the concentration related to the change of scaling law exponent for the viscosity of PVA solutions. In this concentration interval, entanglement of polymer chains may strongly increase resulting in the growth of apparent chain length in the amorphous phase of the gel, thus increasing elongation at break [48,49]. The increase of apparent chain length is also supported by the ordering of PVA chains in domains driven by hydrogen bonds at high polymer concentration [33]. Both the strength (maximus stress) and toughness (area under the curve) of the hydrogels increased with the increase of polymer concentration or cycle number. (Fig. 5e and f). The correlation between these two properties is surprisingly close and in both cases, the toughness as the function of the tensile strength follows a power law with an exponent around 1.1 (Fig. 6), shedding light on an identical failure mechanism among all the gels synthesized here by the FT process. The only outlier is the gel with 15 % PVA after 1 FT cycle, which can be explained by its imperfect network also suggested by its frequency-dependent storage modulus values.

Fig. 5.

Mechanical characterization of PVA hydrogels at large deformations. Tensile curves of gels with various a) FT cycles and b) concentrations. Elongation-at-break (ε_max) and the tensile modulus (E) of gels with various c) FT cycles and d) concentrations. The toughness and the strength (σ_max) of gels with various e) FT cycles and f) concentrations.

Fig. 6.

Correlation between the characteristic values of the tensile curves for various gel compositions, the only outlier (15 % 1C) is marked with a circle.

3.3. Swelling and rheology of the swollen polymer tablets

Slightly cross-linked poly(acrylic acid) (PAA, Carbopol® Ultrez 10 NF) and hydroxypropylmethylcellulose (HPMC, K15M) tablets are expected to display dissimilar adhesive strength on the PVA gels or even different adhesive mechanisms due to their distinct structure and chain flexibility [51] although both are hydrophilic, can be used in controlled release applications [27,50], and also water-swellable first generation mucoadhesives [21], [23], [52], [53]. PAA displayed a remarkable water uptake capacity without dissolution due the presence of carboxylate groups and slightly cross-linked structure. In comparison, HPMC, bearing only hydroxyl functional groups and practically linear, non-cross-linked polymer chains, showed quite similar swelling kinetics and integrity to PAA and reached up to 300 % degree of swelling (Fig. 7 a), to the same level as the PAA, which can be explained by the presence of physical entanglements due to its high molecular weight (M_w = 575 kDa). The swollen tablets, after 1 h of swelling time, were tested by frequency sweep measurements at small deformations to draw conclusions on their microstructure. Swollen PAA tablet showed viscoelastic behaviour characteristic for gels (Fig. 7 b), with a massive gap between the storage and the loss modulus and a moderate frequency dependence, which indicates some imperfections in the cross-linked network caused by dangling chain ends. HPMC displayed strongly frequency-dependent behaviour for both dynamic moduli characteristic for entangled polymer networks without chemical cross-links. The storage modulus decreased rapidly at low frequencies and approached loss modulus indicating significant energy dissipation. G’ reached a plateau at low frequencies indicating the presence of significant, cohesive forces even at very slow deformation rates, although with a stiffness much lower than that of PAA. Accordingly, we do not expect cohesive failure of either tablet during adhesion tests; therefore failure can occur only in the gel substrate or the adhesive interphase.

Fig. 7.

Swelling and adhesion of the tablets; a) swelling kinetics, b) frequency-dependent dynamic moduli of the tablets after 1 h of swelling.

3.4. Adhesion of PAA tablets on PVA gels

PAA tablets, generally considered as strong mucoadhesives, were tested first on PVA gels with various polymer concentration and number of FT cycles. Characteristic adhesion curves are shown in Fig. 8 a) and b). The strength of adhesion was considerably high which can be explained by the fast swelling of the tablets (see swelling kinetics in section 3.3) that enables them to uptake water from the hydrogel surface by capillary forces and enhance chain interpenetration. The cohesive failure of the hydrogel substrates was observed visually either at low cycle numbers (1 or 2 FT cycles in Fig. 8 a) or at low PVA concentration (7–11 % in Fig. 8 b) indicated by the presence of several local extremum in addition to the main peak. These curve shapes resemble the tensile curves of the same PVA gels suggesting identical failure mechanism with the breakage of fibrillar structures forming from the hydrogels during debonding. The toughness of adhesion increases with the stiffness of the gels in the cohesive region, while it shows a gradual decrease with increasing the stiffness further, either by the increasing cycle number or the increasing PVA concentration (Fig. 8c and d). The maximum adhesion stress is reached at the same cycle number and polymer concentration as in the case of the toughness but further increase of stiffness resulted in reaching a plateau instead of a maximum-type curve. These results are in agreement with the finding of Zosel [46], where the maximum toughness was reached at the boundary of cohesive and adhesive failure (in that case, right after the gel point of the substrate), and the toughness decreased significantly with further raising the density of crosslinks. At the maximum in stress and also in the toughness of adhesion, the gel is strong enough to make the interfacial interactions fail instead of its bulk. The elongation at the break is the highest in this intermediate region (Fig. 8e and f) possibly because both the adhesive interphase (Fig. 9) and the bulk of the gels are deformed significantly. With increasing the modulus of the gels further, the elongation at break decreases with a constant σ_max, thus smaller toughness of adhesion. The constant value of σ_max indicates the similar strength and/or density of the interactions in the interphase despite the development of structural changes in the gels causing their stiffness to increase. The transition from cohesive to adhesive failure is nicely demonstrated by plotting the σ_max values for adhesion against the σ_max values from the tensile tests (Fig. 8g and h). Data points representing adhesive failure fall clearly in the right-bottom region while those of cohesive failure scatters around the diagonal line, where adhesive strength or toughness equals to the corresponding tensile quantity. The toughness as the function of σ_max values follows a power law with an exponent around 1 (Fig. 8i and j) for cohesive failure, similarly to the tensile measurements. In contrast, data points describing adhesive failure on gels with different cycle numbers follow a different power law with an exponent of 2.5. The different characteristic values suggest that the failure mechanism in the adhesive region remains unchanged in the interphase for PVA gels with different stiffness.

Fig. 8.

Adhesion of PAA tablets on FT PVA hydrogels. Adhesion curves of gels with various a) FT cycles and b) PVA concentrations. Elongation-at-break (ε_max) and the modulus (E) calculated from the adhesion curves measured on gels with various c) FT cycles and d) PVA concentrations. The toughness and the strength of adhesion (σ_max) on gels with various e) FT cycles and f) PVA concentrations. Correlation between the σ_max values of adhesion and the tensile tests with various g) FT cycles and h) PVA concentrations. Correlation between the characteristic values of the adhesion curve with various i) FT cycles and j) PVA concentrations.

Fig. 9.

Hypothesised structures of the PAA – PVA and the HPMC – PVA aqueous interphase during adhesion of the tablets on PVA hydrogels. PAA is assumed to display more efficient interpenetration and thicker interphase than HPMC, which is a cellulose derivative with significantly less chain flexibility and more rigid main chain. The gray rectangles represent microcrystallites in the gel, and the green dots indicate the slightly crosslinked structure of the PAA. Meshes with various sizes are shown in the amorphous phase of PVA gels.

3.5. Adhesion of the HPMC tablet on PVA gels

HPMC, as a mucoadhesive cellulose derivative displayed dissimilar adhesive characteristics compared to PAA, even though the first steps of the process including swelling by capillary effects and chain interpenetration are assumed to be similar. Adhesive strength of HPMC on the hydrogels was significantly lower than measured for PAA, both for gels with different cycle number (Fig. 10 a) and polymer concentration (Fig. 10 b). Therefore, it was possible to analyse the interfacial failure on a wider scale of gel properties. Indeed, only the PVA gels with 1 or 2 FT cycles showed cohesive failure using HPMC tablets (Fig. 10 a) and no cohesive failure was experienced with 3 cycles independently of the polymer concentration (Fig. 10 b). The comparison with PAA is even more clear in Fig. 10c and d, which shows the reduction of adhesion toughness by around one, while that of the σ_max values by around half order in magnitude. The toughness value displayed a maximum at the transition of cohesive to adhesive failure when the number of FT cycles was varied (Fig. 10 c). More interestingly, we also observed a maximum in toughness as a function of polymer concentration despite the interfacial failure for all polymer concentration tested (Fig. 10 d). Moreover, σ_max does not reach a plateau as a function of each synthetic condition but a clear decrease was observed with the increase of either the cycle number or the PVA concentration, accompanied by a serious decrease in ε_max of adhesion (Fig. 10e and f). These facts led us to the conclusion that the strength of the interfacial interactions between HPMC tablets and PVA gels vary with the change in properties of the latter. HPMC has significantly lower chain flexibility than that of PAA [51], presumably reducing the possible chain interpenetration of HPMC into the dense amorphous phase of PVA gels, prepared with either higher cycle number or higher polymer concentration. The insufficient interpenetration results in a less entangled interphase ultimately resulting in lower toughness and deformability, considering also the possibly weak interaction between rigid HPMC chains (Fig. 9). The cohesive-adhesive transition is supported by comparing the σ_max values from adhesion and tensile tests (Fig. 10g and h). Plotting the toughness against σ_max of adhesion the exponent of the power law fitted after the maximum in σ_max and toughness is around 2.0 (Fig. 10i and j), which is lower by 0.5 than that of PAA. This exponent indicates a different behaviour for the adhesion of HPMC on PVA gels compared to PAA on the same gels and also supports our assumptions on the reduced strength of interfacial interactions for HPMC.

Fig. 10.

Adhesion of HPMC tablets on FT PVA hydrogels. Adhesion curves of gels with various a) FT cycles and b) PVA concentrations. Elongation-at-break (ε_max) and the modulus (E) calculated from the adhesion curves measured on gels with various c) FT cycles and d) PVA concentrations. The toughness and the strength of adhesion (σ_max) on gels with various e) FT cycles and f) PVA concentrations. Correlation between the σ_max values of adhesion and the tensile tests with various g) FT cycles and h) PVA concentrations. Correlation between the characteristic values of the adhesion curve with various i) FT cycles and j) PVA concentrations.

3.6. Correlation between adhesion and viscoelasticity

As a general comparison of adhesion of PAA and HMPC tablets on PVA hydrogels with different viscoelastic properties, the toughness of adhesion and viscoelastic phase angle of all the gels was plotted against the storage modulus of the gels in Fig. 11. The increase of gel stiffness could be achieved both by increasing cycle number and polymer concentration as shown previously, but the viscoelastic character differs for gels with low modulus prepared by using different synthesis conditions. The relatively high phase angle of the PVA gels after one FT cycle suggest an imperfect gel structure, e.g., the presence of significant sol fraction. One might expect the cohesive failure of these gels in adhesion tests; indeed, during the debonding of both tablets from these gels caused the breakage of the gels instead of an adhesive failure. Increasing either the cycle number or the polymer concentration in the hydrogels resulted in similar phase angle at the same modulus, i.e., a master curve exists between these two quantities regardless of the synthesis conditions (gray curves in Fig. 11). Nevertheless, microstructural differences can still exist between gels with the same viscoelastic properties prepared by different synthesis conditions. The increase of cycle number primarily increases the size of the crystallites but can also affect the structure of amorphous phase indicated by relaxation spectra (chapter 3.1). The increase of polymer concentration mainly enhances entanglement in the amorphous phase, but can also affect crystallinity (see optical properties in Fig. 2).

Fig. 11.

Correlation between adhesive and viscoelastic properties at a specific frequency value. A master curve with gray is shown for phase angle and storage modulus. Dissimilar adhesion tendencies are shown with green and red curves for PAA and HPMC tablets, respectively.

The adhesion of the two different tablets displays dissimilar tendencies. Toughness of adhesion for PAA tablets show a maximum at the boundary of cohesive and adhesive failure in strong agreement with the results of Zosel et al. [46] The adhesion of HPMC tablets, however, displays a more complex behaviour. In general, the toughness is much lower than that of the adhesion of PAA tablets indicating less efficient stress transfer in the interphase of HPMC and PVA. The failure mechanism remains cohesive in the first two FT cycles, in contrast to the adhesion of HPMC tablets on PVA gels with the same or even somewhat lower stiffness prepared with lower polymer concentration but three FT cycles. The PVA gels after two FT cycles displayed pronounced frequency-dependent dynamic moduli and damping factor (Fig. 4 a, Fig. S2 a), which indicates the presence of loosely connected chains or significant sol fraction that might result in fast crack propagation and ultimately the cohesive failure of the gels. This scenario is less probable for the PVA gels with similar modulus prepared with lower PVA concentration, but higher cycle number (Fig. 4 b, Fig. S2 b). It must be emphasized that phase angles in Fig. 11 plotted only for the slowest deformation rate that can be assumed in the interphase during the debonding, but the deformation rate can significantly increase with the deformation of the interphase, thus viscoelastic properties at small strains and frequencies cannot correlate completely with the large deformations during adhesion tests. However, in the region of adhesive failure for HPMC tablets show a similar tendency was observed between toughness and modulus as shown for PAA tablets, although with a stronger reduction in toughness with increasing stiffness. The dissimilarities in maximum toughness and tendencies indicate differences in the structure thus deformability of the adhesive interphase.

The maximum deformation in adhesive debonding was higher than that in the tensile tests for softer gels, especially for the debonding of PAA tablets (Fig. 12a and b). This trend suggests that a significant part of the deformation can be attributed to the interphase, but the bulk of the gels is also deformed. As a consequence, one part of the adhesion toughness can be deduced from the deformation of the bulk of the gel, and the other part is related to the deformation in the interphase. These contributions to the toughness cannot be separated completely, although the minimum elongation of the interphase can be estimated using the corresponding tensile curve of the PVA gels, with some assumptions. We applied an approximation only for the adhesive failure (both for PAA and HPMC tablets), in which case the maximum deformation is relatively small compared to that in the case of cohesive failure. Therefore, the rate of deformation can be assumed to be equal in the adhesion and in the tensile tests. We determined the maximum deformation of the gel from the tensile curve at which the area under the curve is the same as the toughness of adhesion measured on the same gel. This is the maximum elongation in the bulk of the gel (${\Delta h}_{\mathit{max},gel}$) at the adhesive failure. Subtracting that from the elongation corresponding to the complete debonding in adhesion ($\Delta h$), we obtain the minimum elongation of the interphase (${\Delta h}_{\mathit{min},interphase}$) in adhesive debonding (Eq. (5)):

$${\Delta h}_{\mathit{min},interphase}={\Delta h-\Delta h}_{\mathit{max},gel}\left[mm\right]$$ (5)

Fig. 12.

Correlation between the elongation at break values from the adhesion and tensile tests with various a) number of FT cycles and b) PVA concentrations. The minimum value for the elongation of the interphase in adhesive debonding of tablets with various c) FT cycles and d) PVA concentrations.

The results from the calculations above (Fig. 12c and d) support the theory that PAA chains have higher mobility in the PVA gels and reach a deeper interpenetration resulting in the formation of a more entangled interphase (Fig. 9). This eventually explains the generally higher elongations for the adhesion of PAA tablets compared to HPMC regardless of the viscoelastic properties of the gels. Nevertheless, large elongation in the interphase can also be reached for the adhesion of HPMC tablets if it is attached to less stiff, thus less dense networks in which a deep chain interpenetration of the rigid cellulose derivative is possible. Accordingly, the adhesion performance strongly depends on both the chain dynamics of polymer excipient and the mechanical properties (at small and large strains) of the substrate of adhesion. In a more general sense, careful mechanical characterization of mucosal membranes or mucus is inevitable for the correct interpretation of the results of adhesion tests and choose drug formulations with improved mucoadhesion.

4. Conclusion

We introduced a novel concept for studying the adhesion of polymer tablets on soft mucosa-mimetic hydrogels. The mechanical properties of the complex adhesive interphase was considered with the elimination of the effect of specific mucin-polymer interactions by using poly(vinyl alcohol) (PVA) hydrogels as substrate for adhesion. Physically cross-linked PVA hydrogels were prepared by varying two synthesis conditions: the number of freeze-thaw cycles (1–5) and the PVA concentration (7–20 %). The adhesion of the slightly crosslinked poly(acrylic acid) (PAA, Carbopol® Ultrez 10 NF) on the hydrogels showed a maximum in adhesion toughness at the boundary of cohesive and adhesive failure regardless the synthesis conditions of the hydrogels. The adhesive strength reached a plateau and did not change with increasing stiffness of the hydrogels. In contrast to PAA, hydroxypropylmethylcellulose (HPMC, K15M) with less flexible chains showed a more complex behaviour, and after reaching the adhesive region, the further increase in the stiffness of the hydrogels caused both the strength and toughness of adhesion to decrease rapidly, which indicates important changes in the adhesive interphase. Based on the elongation at break estimated in the interphase, the rigid HPMC chains are not capable of deep interpenetration into the dense amorphous phase of the PVA gels prepared with high polymer concentration or cycle number. Our explanation is supported by the generally higher minimum deformation of the interphase for PAA than HPMC. Finally, we recognized general scaling-law type correlations between adhesive toughness and strength. The results can help understand the role of chain flexibility and interpenetration in the adhesion of solid mucoadhesive polymer dosage forms and can contribute to a more knowledge-based design of novel mucoadhesive formulations and mucosa-mimetic materials.

CRediT authorship contribution statement

Gergely Stankovits: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Conceptualization. Kata Szayly: Methodology, Investigation, Formal analysis. Dorián László Galata: Resources, Methodology, Investigation. János Móczó: Methodology, Formal analysis. András Szilágyi: Resources, Funding acquisition, Conceptualization. Benjámin Gyarmati: Writing – original draft, Supervision, Methodology, Funding acquisition, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the supplementary data to this article:

Data availability

Data will be made available on request.