The first "Science" in polymer direction in 2021: Controlling the nano-scale inhomogeneity in seawater desalination membranes can be enhanced

Nearly 80% of the world's fresh water is used for agriculture, animal husbandry and energy applications, which puts tremendous pressure on existing water sources in developed and developing countries
.
Technologies such as membrane filtration, distillation, and ion exchange are widely used to purify water sources; however, the energy required to remove dissolved solutes, especially salts, from water is still high
.
? Reverse osmosis technology (RO) produces nearly 21 billion gallons of water every day, accounting for 66% of the global desalination capacity.
RO is also playing an increasing role in the recovery of fresh water from wastewater or other waste liquids for human and industrial purposes.
The more important the role
.
In recent years, some progress in the synthesis of reverse osmosis membranes has provided an effective way to obtain highly permeable seawater desalination membranes.
Studies have found that by controlling the internal morphology, thickness, and feed surface area of ​​the fully aromatic polyamide (PA) active layer, Greatly improve the water permeability
.
However, it is not clear how the current nanoscale morphology of PA relates to the high performance observed in such films
.
To this end, Manish Kumar, Department of Civil, Architecture and Environmental Engineering, University of Texas, and Enrique D.
Gomez, Department of Chemical Engineering, Pennsylvania State University, published a titled "Nanoscale control of internal inhomogeneity enhances water transport in desalination membranes" in "Science".
The work proposed that the control of nano-scale polyamide inhomogeneity is the key to maximizing water permeability without sacrificing salt selectivity
.
?? This study describes a method to quantify changes in polymer quality at the three-dimensional (3D) nanoscale and explores its impact on water transport in the active layer of four reverse osmosis membranes
.
In order to individually explore the influence of the nano-morphology of the membrane on the water transport characteristics, the difference in chemical composition between the membranes used was minimized
.
Fourier transform infrared spectroscopy confirmed that the four reverse osmosis membranes have an almost constant ratio of carboxylic acid to amide
.
Through electron tomography, energy-filtered transmission electron microscopy, and solution diffusion simulations, this article found that nano-scale density changes are detrimental to the water transport in these membranes, and controlling these density fluctuations is essential to improve the performance of reverse osmosis membranes.

.
? [Quantify the three-dimensional nano-scale inhomogeneity of PA RO film] Transmission electron microscopy (TEM) cannot quantitatively associate the microstructure of PA with the desalination performance
.
When imaging in scanning TEM (STEM) mode using a high-angle annular dark field (HAADF) detector, for a single-component system, the pixel intensity is directly related to the sampling quality
.
Furthermore, for a single PA film, the pixel intensity of the HAADF-STEM image is a function of sample thickness, density and pixel size
.
In order to separate the thickness and density of PA in an electron microscope, three-dimensional reconstruction of the topography of nano-level PA is required
.
Through HAADF-STEM tomography, we have achieved the creation of a 3D model describing the nanoscale surface and internal morphology (Figure 1A, 1B)
.
The quantitative analysis of the three-dimensional model shows that the porosity and surface area of ​​PA are consistent with the analysis results of similar industrial reverse osmosis membranes
.
? High-resolution HAADF-STEM tomography can separate the density and thickness of PA, so that the nano-level 3D distribution of each parameter can be determined independently
.
Although the average value of the relevant membrane characteristics is usually used to estimate the transmission rate of the membrane, the nano-scale mass distribution is likely to control the transmission of water through the reverse osmosis membrane
.
Therefore, the changes in membrane resistance and water flux will be combined by nano-scale changes in PA thickness and density
.
In this paper, energy-filtered TEM and HAADF-STEM are used to map the changes in the nano-scale inhomogeneity in PA film density and its relationship with the changes in film thickness
.
In this paper, the 3D nano-scale intensity distribution (Figure 1C, 1D) is converted into a nano-scale density distribution ρ(r), and the water diffusivity (Dw) in the PA membrane is extracted from it.

.
Figure 1.
Quantify the three-dimensional nano-scale inhomogeneity of PA RO film through the combination of energy filtering TEM and electron tomography? [Water diffusion] The premise of measuring the nano-level three-dimensional PA inhomogeneity is to obtain the average PA density (ρavg) , Average sFFV (sFFVavg) and average diffusion coefficient of water (Dw,avg)
.
In short, by calculating the elastic and inelastic scattering components of the TEM image (Figure 2A-2C), the mean free path of electrons can be obtained.
The mean free path can be used to determine the ρavg in the PA film, which can be extended to sFFVavg and Dw,avg
.
In a series of membranes measured, the water permeability rose from 6.
39 ± 0.
22 to 8.
36 ± 0.
15 liters m? 2 hour? 1 bar? 1, and correlated with ρavg from 1.
15 ± 0.
14 to 0.
86 ± 0.
09 g cm? 3 (Figure 2F)
.
These average density values ​​are consistent with the density values ​​of bulk PA reported in the literature
.
In addition, under the same increase in permeability, sFFVavg increased from 0.
35 ± 0.
04 to 0.
52 ± 0.
05 (Figure 2G), indicating that the increase in angstrom free volume is positively correlated with water flux
.
The large sFFV value is consistent with the FFV prediction of the glass polymer, indicating that the PA free volume elements may be connected to each other
.
Dw, avg are compiled from Dw and 1/sFFV data from free volume theory and molecular dynamics simulation
.
? The Dw distribution from PA1 to PA4 gradually narrows with the increase of water permeability, indicating that the local distribution of mass affects water transport
.
However, since the diffusivity distributions of water at the nanoscale in Figures 1I and 1J cannot explain the changes in membrane resistance, it is not possible to predict the transport characteristics of water from these distributions alone
.
The spatial distribution of changes in local membrane resistance plays a vital role in determining the diffusion path of water
.
The diffusion of water molecules in the lower thickness and lower density PA area is easier than the diffusion in the thicker and dense PA area, that is, the water transport will take the path of least resistance.

.
In addition, these resistance changes can cause distribution in the flow and cause flux hotspots, which cannot be explained solely based on the simple average Dw
.
Figure 2.
Measurement of the average density, free volume, and diffusion coefficient of water in PA membranes with an energy-filtered transmission electron microscope.
[Three-dimensional model of water transport] In order to predict the transport properties, this paper uses a three-dimensional model to calculate the diffusion of water.
The model shows how the thickness and Dw vary locally
.
Through a series of calculations, the diffusion path of water through the PA membrane can be determined (Figure 3), and the influence of the nano-level PA morphology on the three-dimensional water transport can be seen
.
The light gray area corresponds to the ultra-low water diffusivity area in the membrane, that is, the high PA density and low sFFV area
.
The water diffusion coefficient corresponding to the dark gray area is between 1.
2 and 1.
5 × 10? 5 cm2 s?1
.
The area with the greatest resistance is near the top surface of the PA, which emphasizes the importance of the surface area of ​​the PA for water transport
.
Figure 3.
A three-dimensional model for calculating water transport obtained by energy filtering TEM and electron tomography? [Water permeability prediction] In order to avoid these areas with high PA density and low sFFV, the diffusion path of water is in the x direction There are changes in both the y direction and the y direction
.
Using the flow chart, the water permeability can be reliably predicted with zero adjustable parameters
.
The inset in Figure 4 is based on the unevenness of the density and thickness of the PA1 and PA4 membranes to predict the water flux distribution (Jw, p) of the reconstructed PA volume element on the xy plane
.
Although all membranes show some local inhomogeneities in water flux, we found that the low flux area of ​​the high flux membrane (PA4) is the smallest
.
Comparing several xy planes in the membrane and the flow distribution on each surface, evidence of lateral water transport (x and y directions) can be found, which indicates that the nano-level PA morphology affects water transport in all three dimensions
.
Using the calculated flow chart, we can reliably evaluate the predicted water permeability Pw,p, which is qualitatively consistent with the experimentally measured water permeability Pw,m (Figure 4)
.
Figure 4.
Nano-scale water transport calculation predicts water permeability under zero adjustable parameters, and compares the membrane performance with the most advanced membrane materials.
[Summary] The highest permeable membrane (PA4) has the lowest average density and the narrowest density distribution , So as to maximize the overall permeability while maintaining selectivity
.
This shows that by minimizing the quality fluctuation that is unfavorable to water permeation, and limiting the average density value to a level that is as low as possible but still able to ensure ion selectivity, the water permeability of the permeable membrane can be increased as much as possible
.
This article combines energy filtering TEM and electron tomography to create a key tool for predicting the correlation between high-performance reverse osmosis membrane morphology and water transport
.
These associations can be extended to other aspects such as molecular separation and polymerization systems to improve design strategies for various applications, including gas and hydrocarbon separation, carbon capture, blue energy production, and seawater desalination
.