Fig. 1. Microfluidic platform for synthesis of biopolymer membranes within standing pH gradients. (a) Schematic of a parallel membrane synthesis device with 4 X-channels showing co-flowing basic solutions (blue) and the acidic biopolymer solution (red). (b) Photo of the realized device with dye solutions injected to visualize the co-flow patterns described in (a). Inset image shows a single X-channel before membrane synthesis.

Fig. 2. Viscosity measurements for (a) collagen solution in acid acetic 0.02 N at 4 (▲), 2 (?) and 1 mg mL-1 (●) and (b) chitosan at 5 (▲), 2 (?), and 1 mg mL-1 (●) concentrations. The vertical axis is the same scale as in (a). Both collagen and chitosan solutions were maintained at pH 3.5.

Fig. 3. In situ formation of chitosan membranes studied by time-lapse imaging. (a) Sequential images from the time-lapse series showing propagation of a chitosan membrane (growth parallel to flow indicated by red arrow) during the initial contact of the chitosan solution (5 mL h-1, pH 3.5) with NaOH solution (pH 10). Flow rates were Qp = 0.4 mL h-1 and Qb = 7.8 mL h-1 (Qtotal = 8.2 mL h-1). Red arrows show the position of the propagation front after each 15 s interval. (b) Average propagation velocity of a chitosan membrane measured for different solution total flow rates (Qtotal) while maintaining a flow rate ratio of Qb/Qp = 20. (c) Widening chitosan membrane in time (growth perpendicular to flow indicated by red arrow). The direction of flow is from left to right in images (a) and (c). (d) Membrane width versus time for 4 mL h-1 collagen (black) at Qb = 1.83 mL h-1, 5 mL h-1 chitosan (blue) at Qb = 2.7 mL h-1, and 60:40 (v/v%) collagen/chitosan hybrid (red) at Qb = 2.1 mL h-1. All pH values were 3.5 and 10 for the polymer and NaOH solutions, respectively. Error bars in (b) and (d) were generated from the standard deviation of quadruplicate measurements under identical conditions.

Fig. 4. Growth of membrane width and changes to uniformity under different flow rates for chitosan (a, b), collagen (c, d) and collagen/chitosan hybrids (e, f) during the first 10 min of growth. Inset images showing typical membranes after 30 min. (scale bar = 500 μm). (a) Chitosan membrane formation was tested under five different flow rates: Qtotal = 2.7 (orange), 5.4 (blue), 8.1 (gray), 10.8 (yellow), and 13.5 mL h-1 (green). (b) Width measurements of chitosan membranes at the upstream (solid line) and downstream (dashed line) positions in the X-channel for fastest (red) and slowest (blue) flow rates used in (a). (c) Collagen membrane formation was tested under three different flow rates: Qtotal = 0.63 (orange), 1.23 (yellow), and 1.83 mL h-1 (green). (d) Width measurements of collagen at upstream and downstream positions, with line colours same as in (b). (e) Hybrid chitosan-collagen membrane formation for three different solution flow rates: Qtotal = 1.03 (orange), 2.1 (yellow), and 5 mL h-1 (green). (f) Width measurements of hybrid chitosan-collagen membrane at upstream and downstream positions, with line colours same as in (b). All flow rate ratios were Qb/Qp = 20.

Fig. 5. (a) Variation in membrane width (Δw) between up- and downstream positions as a function of total flow rate. (b) Volume swelling ratios for membranes formed at different flow rates for 0.4 mL mg-1 collagen (orange circles), 0.4 mL mg-1 chitosan (blue circles), and 60/40 collagen/chitosan hybrid (green circles). Dashed lines in (a) are included for eye guidance.

Fig. 6. (a) Live/dead (green/red) assays of 3T3 fibroblast cells one week after inoculation for (i) collagen, (ii) chitosan, and (iii) collagen/chitosan hybrid membranes. Yellow regions correspond to dead cells overlapping with a green background and are therefore, considered as red for the purposes of statistics. Membranes appear green due to retention of live stain and auto-fluorescence. Membrane synthesis was conducted with Qtotal = 0.63, 2.7, and 1.0 mL h-1 for collagen, chitosan, and collagen-chitosan hybrid extracellular membranes, respectively. Scale bar in (i) is common for (i-iii). (b) Cell viability calculated based on 4 repeated experiments for each membrane type. ANOVA analysis followed by post-hoc paired t-test for the same variance showed a significant difference between chitosan and collagen (*).