Design of Polyether sulfone flat-sheet membranes for multi-layered hemodiafiltration devices

Abstract

Multilayered hemodialysis devices enable blood flow without the need for an external pump, whereas hollow fiber-type devices require a pump due to significant pressure loss across the fibers. This highlights a key advantage of multilayered devices for implantable applications, where a simpler and lighter system reduces the burden on patients. This study investigates the mechanical strength and dialysis efficiency of polyether sulfone (PES) flat-sheet membranes for multilayered devices. PES membranes, with thicknesses ranging from 40 to 160 µm, were prepared using the liquid inversion method, with thickness controlled via spin coating. The mechanical strength of the membranes was tested following the ISO 8637-1:2017 protocol, and membranes thicker than 80 µm were experimentally verified to withstand pressures of up to 500 mmHg, making them suitable for dialysis applications. Furthermore, the study demonstrates the successful use of 80 µm membranes in both in vitro and ex vivo experiments with rats, identifying this thickness as optimal for multilayered dialysis devices.

Fig 1. Illustration of the multi-layered hemodiafiltration device.

Fig 2. (a) Fabrication process of PES membrane. (b) Photograph of the fabricated device.

Fig 3. The 3D models of (a) hollow fiber membrane and (b) flat-sheet-sheet membrane.

Fig 4. Illustration of the experimental setup for pressure resistance test and dialysis performance test.

Pressures at the inlet and outlet of the device are ($P_{Bin}$ and $P_{Bout}$ for the blood channel and $P_{Din}$ and $P_{Dout}$ for the dialysate channel) and measured with pressure gages and a polygraph system.

Fig 5. Cross sectional ESEM images of the PES membranes: (a) 50 μm, (b)100 μm (c)125 μm, (d)150 μm (e)200 μm, (f)250 μm.

The scale bar represents 100μm.

Fig 6. Thickness of skin layer and support layer at each film thickness.

Fig 7. Photographs of PES flat-sheet membrane after the membrane rupture for the thickness of (a) 40, (b) 80, (c) 120, and (d) 160 μm.

Fig 8. (a) The creatinine clearance and (b) the filtration coefficient as a function of the time. Error bar: means± S.D. n = 3. The commercially available dialyzers were reported to have

$L_p$ of 20–60 mL/h mmHg m2 [25].

Fig 9. (a) Photograph of an ex vivo experiment with SD rats. The animal is under anesthesia.

The device is connected to femoral artery and vein. $P_{Bin}$, $P_{Bout}$, $P_{Din}$ and $P_{Dout}$ are continuously measured with polygraph system. Blood is collected to measure the creatinine concentration. (b) Comparison of creatinine removal rate in in vitro and ex vivo experiments.

Fig 10. (a) AFM image of an 80-µm-thick membrane. (b)

$R_a$, (c) $R_v$, and (d) $R_q$ values for membranes with thicknesses of 40, 80, 120, 160, 200, and 240 µm. Black dots indicate the measured values at three different locations, while the averaged values are represented by diamonds. No significant differences in surface roughness were observed with respect to membrane thickness. Furthermore, the measured roughness values were an order of magnitude lower than those reported in previous studies [–29].