Mixing and flow-induced nanoprecipitation for morphology control of silk fibroin self-assembly

Abstract

Tuning silk fibroin nanoparticle morphology using nanoprecipitation for bottom-up manufacture is an unexplored field that has the potential to improve particle performance characteristics. The aim of this work was to use both semi-batch bulk mixing and micro-mixing to modulate silk nanoparticle morphology by controlling the supersaturation and shear rate during nanoprecipitation. At flow rates where the shear rate was below the critical shear rate for silk, increasing the concentration of silk in both bulk and micro-mixing processes resulted in particle populations of increased sphericity, lower size, and lower polydispersity index. At high flow rates, where the critical shear rate was exceeded, the increased supersaturation with increasing concentration was counteracted by increased rates of shear-induced assembly. The morphology could be tuned from rod-like to spherical assemblies by increasing supersaturation of the high-shear micro-mixing process, thereby supporting a role for fast mixing in the production of narrow-polydispersity silk nanoparticles. This work provides new insight into the effects of shear during nanoprecipitation and provides a framework for scalable manufacture of spherical and rod-like silk nanoparticles.

Fig. 1. The nanoprecipitation workflow for the preparation and purification of silk nanoparticles via desolvation in isopropanol. (a) In semi-batch format, the four formulation processing steps are: (1) loading of a bubble-free aqueous silk solution into a syringe equipped with blunt needle. (2) The silk feed position. (3) The flow rate control of silk solution. (4) Control of mixing time via the stirring rate during addition. In continuous format, the three formulation processing steps are: (1) loading of bubble-free aqueous silk and isopropanol into syringes and the NanoAssemblr™ microfluidic chip. (2) The flow rate control of silk solution and the flow rate ratio control of isopropanol: silk. (3) Control of mixing time via the micromixer chip design.

Fig. 2. Exemplary characterisation of flow and mixing properties in the closed semi-batch system. Circulatory flow in silk droplets produced at all flow rates and feed heights was observed using silk doped with iron oxide nanoparticles. The mixing time in the reactor decreased as stirring rate and feed height increased and was measured using the colour change of a methyl red and thymol blue mixture from acidic pH (red) to neutral pH (yellow). (a) Processed binary images of the circulatory flow field of droplets extruded from the closed semi-batch system. Insets show the two-dimensional direction of flow. (b) The variation of standard deviation of the normalized green channel (σ_G) across scale in the open semi-batch system and across feed height and stirring rate in the closed semi-batch system. (c) The raw images and processed masks showing the colour evolution within the flask at the optimal stirring rate of 400 rpm and feed height of 1.75 cm.

Fig. 3. Impact of feed height and stirring rate on nanoprecipitation of 3% w/v aqueous silk in the semi-batch closed-system at 1 mL min⁻¹ flow rate. (a) Hydrodynamic diameter, polydispersity index (PDI), zeta potential and yield of silk nanoparticles. Two-way ANOVA was used to compare multiple groups across feed height and stirring rate, followed by Tukey's pairwise multiple comparison post hoc test for size, polydispersity index and yield and Tukey's simple effect multiple comparison post hoc test for zeta potential. (b) Scanning electron microscopy showed that lower curvature morphologies were obtained as stirring rate and addition height decreased, due to an increasing degree of secondary self-assembly. Error bars are hidden in the bars and plot symbols when not visible, ±SD, n = 3. Asterisks denote statistical significance determined using post hoc tests as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Scale bars 200 nm.

Fig. 4. The impact of increasing the flow rate and silk feed concentration in the staggered herringbone micromixer and in semi-batch systems of high and low mixing time. The hydrodynamic diameter, polydispersity index, zeta potential and yield for (a) semi-batch format and (b) microfluidic format. For the unstirred semi-batch processes, two-way ANOVA was used to compare multiple groups across concentration and flow rate, followed by Tukey's pairwise multiple comparison post hoc test for yield and Tukey's simple effect multiple comparison post hoc test for size, polydispersity index and zeta potential. In stirred semi-batch format and microfluidic format, multiple groups were evaluated by two-way ANOVA, followed by Tukey's pairwise multiple comparison post-hoc test for all properties error bars are hidden in the bars when not visible, ±SD, n = 3. Asterisks denote statistical significance determined using post hoc tests as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. For ease, all statistically significant interactions have been omitted and are shown in Fig. S3.

Fig. 5. Scanning electron microscopy supported the DLS results and confirmed that for low shear processes in the (a) semi-batch format and (b) microfluidic format, the extent of self-assembly varied inversely with silk concentration while for high shear processes, the extent of assembly was maximised at the 2% silk concentration. Due to an extended growth phase associated with 0.5% silk feeds, rod-like silk nanoparticle morphologies were produced in the micromixer between 0.059–1.0 mL min⁻¹, and the particle polydispersity decreased with increasing flow rate.

Fig. 6. Schematic of protein–protein association and β-sheet assembly of silk fibroin via anti-solvent and shear-induced desolvation. The mechanism under shear flow is adapted from Dunderdale et al. and Zhang et al. Silk molecules, nanoparticles and stoichiometry of association are not drawn to scale.