Tailoring and visualising pore openings in gelatin-based hydrogel foams

Abstract

Hypothesis

While tailoring the pore diameters in hydrogel foams has been demonstrated in numerous studies, fine control over the diameters of the pore openings is still a challenge. We hypothesise that this can be achieved by controlling the size of the thin films which separate the bubbles in the liquid foam template. If this is the case, systematic changes of the template’s gas fraction $\varphi$ (the higher $\varphi$, the larger are the thin films) will lead to corresponding changes of the pore opening diameter.

Experiments

Since the size of the thin films depends on both bubble size $〈D_{\mathrm{b}}〉$ and gas fraction $\varphi$, we need to decouple both parameters to control the film size. Thus, we generated foams with constant bubble sizes via microfluidics and adjusted the gas fractions via two different techniques. The foams were solidified using UV light. Subsequently, they were analysed with confocal fluorescence microscopy.

Findings

We were able to change the pore opening diameter $〈d_{\mathrm{p}}〉$ at a constant pore diameter $〈D_{\mathrm{p}}〉$ by adjusting the gas fraction of the foam template. The obtained $〈d_{\mathrm{p}}〉$/$〈D_{\mathrm{p}}〉$ ratios are between those obtained theoretically for disordered foams and FCC ordered foams, respectively.

Introduction

Hydrogel foams based on biopolymers such as alginate, hyaluronic acid or gelatin have gained increasing attention in the field of biomaterials science because they combine the porous foam structure with the unique material properties of a hydrogel. Hydrogels are three-dimensionally cross-linked hydrophilic polymer networks that swell to equilibrium in aqueous media [1], [2]. Due to their high water content [3], as well as their biocompatibility and biodegradability [4], [5], [6], hydrogels based on biopolymers are studied extensively as functional materials in various fields, e.g. tissue engineering or drug delivery [7], [8], [9], [10]. What is of particular interest in this context is the biopolymer gelatin, which contains amino acid sequences relevant for cell adhesion as well as cleavage sites for biodegradation [11], [12]. However, the diffusion of small molecules such as drugs or nutrients through hydrogels is often hindered by the small mesh sizes (up to tens of nanometres) of the material [13], [14], [15]. Thus, the generation of hydrogel foams with pore diameters ranging from tens to hundreds of micrometres has been investigated thoroughly.

One suitable method for the generation of hydrogel foams based on biopolymers is liquid foam templating [16], [17], [18], [19]. With the help of this technique, a liquid foam is generated from an aqueous polymer solution and is subsequently solidified by cross-linking the polymer. As the morphology of the hydrogel foam is determined by the structure of the liquid template, the key parameters of the material, i.e. the pore diameter and the pore opening diameter, can be fine-tuned by tailoring the liquid template. Both pore diameter and pore opening diameter play a key role for the material properties of polymeric foams [20], [21]. In the field of tissue engineering, relations between the differentiation and colonization behaviour of cells on the one hand and the pore diameter as well as the pore opening diameter on the other hand have been demonstrated in numerous studies [22], [23], [24], [25]. To tailor the structure of the liquid template and thus the morphology of the hydrogel foam, microfluidic foaming has been examined extensively [16], [18], [21], [26]. With the help of this technique, a gas and a liquid phase are pushed through a chip with a micrometric constriction, which results in the pinch-off of bubbles. As this pinch-off occurs periodically, the generated bubbles have the same diameter, i.e. they are monodisperse. The bubble diameter of the liquid template and thus the pore diameter of the hydrogel foam can be conveniently tailored by varying the gas pressure and/or the liquid flow rate inside the microfluidic chip. This has been demonstrated for hydrogel foams based on chitosan, alginate and gelatin [16], [18], [21], [26]. The bubble diameter is further determined by the size of the constriction. While previous studies focussed on the variation of the pore diameter, the variation of the pore opening diameter has received little attention up until now. The mechanism of the pore opening process is still not very well understood. However, it can be assumed that the contact of two neighbouring bubbles leads to the formation of a contact zone between them. As these contact zones are subject to capillary forces generated by the surrounding struts, the thickness of the contact zone is reduced to a surfactant double layer. During cross-linking and freeze-drying, these films break and pore openings are generated (see Fig. 1).

Generally speaking, in liquid foam templating, the pore opening diameter correlates, to a first approximation, directly with the diameter of the thin liquid film which separates neighbouring bubbles. For foams with a high gas fraction ϕ (>90%) one can show by means of a simple scaling argument that the diameter $〈d_{\mathrm{p}}〉$ of the pore opening is related to the pore diameter $〈D_{\mathrm{p}}〉$. It holds

$$〈d_{\mathrm{p}}〉~\left(1-\sqrt{1-\mathrm{\Phi }}\right)〈D_{\mathrm{p}}〉.$$

For lower gas fractions, this relationship becomes a bit more complicated, as will be discussed in Section 3.3. Nevertheless, it shows the importance of decoupling the gas fraction and the pore size for explicit pore opening control. This was shown qualitatively for polyurethane foams by Testouri et al. [27]. A more systematic approach was developed for geopolymer- or gelatin-based foams that were generated from a foamed precursor solution which was then mixed with the geopolymer- or gelatin-containing dispersion/solution [21], [28], [29]. However, this approach is time-consuming, requires multiple formulations, and has high demands on foam stability. These requirements cannot always be met when using photo cross-linkable hydrogel formulations. The decoupling of pore size and gas fraction can also be achieved by varying the surfactant concentration, which was examined for systems based on alginate and methacrylated dextran [18], [30]. Despite the fact that this is a very elegant concept, high surfactant concentrations are required, which limits potential applications.

In the present study, we used microfluidics for the generation of monodisperse liquid foams with a specific bubble diameter. The continuous phase of the foams consisted of an aqueous solution of gelatin methacryloyl. In order to tailor the pore opening diameter, we systematically adjusted the gas fraction of the foams independently of the bubble size using two different methods. In the first method, the polymer solution was added to the liquid foam template made at high gas fraction from the very same polymer solution. In the second method, the bubbling frequency during foam generation was tailored by varying the liquid flow rate and gas pressure while keeping the ratio of these parameters constant. This effect was examined by Garstecki et al. [31] for an aqueous surfactant solution, but has never been used to tailor the pore opening diameters in polymer foams. Another question we want to address is how hydrogel foams can be characterised. Typically, hydrogel foams are characterized by scanning electron microscopy, µ-CT or bright field microscopy. The downside of these methods is that they are expensive and time-consuming or do not adequately visualise the equilibrium state, i.e. the swollen state of the material. Here, we use confocal fluorescence microscopy to characterise the material in the swollen state. The paper is organised as follows: Firstly, we briefly report and discuss the use of confocal fluorescence microscopy for the characterisation of gelatin-based hydrogel foams. Secondly, we apply the two aforementioned approaches to decouple gas fraction and pore size. Finally, we compare the ratios of our experimentally determined pore opening diameters and pore diameters with predictions by Princen [32] and Arditty [33] for disordered foams, as well as with predictions by Pitois et al. [28], [34] for foams with an FCC order.