Desalination Membranes: Characterization Techniques

INTRODUCTION

Membrane characterization is vital to understanding the key fundamental physical, chemical, morphological, and mechanical properties that ensure the best membrane performance. An understanding of such membrane structural and morphological properties as porosity, pore size and distribution, surface topography and morphology, surface chemistry, as well as the basic mechanical strength, resistance to deformity, and hydraulic resistance, is essential to control the separation properties and usefulness of the resultant membrane.

The difficulty is in gaining such detailed information about the membrane characteristics, especially when the required dimensions and characteristics could be at atomic scale, which requires the latest analytical and spectroscopic technologies. Thus, this chapter provides a comprehensive overview of available membrane characterization technologies . The first section focuses on physical, chemical, thermal and mechanical characterization techniques, while the second section examines the membrane performance in terms of flux and salt rejection rates.

MEMBRANE CHARACTERIZATION

Surface Morphology and Physical Characteristics

The surface morphology and the physical characteristics are usually the first characteristics to be examined in the prepared membranes; these include studying the porous structure of the membrane, which covers the membrane porosity, the pore shape, the average pore size, and the pore size distribution. These characteristics are considered key information, since they are affected by the preparation approach and affect the membrane performance, Such principal information as the microstructure of the membrane and the surface topography can also be obtained. The microstructure inspection provides information regarding the binding between the different layers, along with the binding between the main matrix and any loaded fillers. Besides, it is important to examine the surface roughness as it affects the membrane performance, either by affecting the water flux through the membrane, or by affecting fouling. In addition, foulant contents and types are required to be characterized.

Porous Structure Characterization

Characterizations of the porous structure of the membrane include determination of the overall porosity, the average pore size, and the pore size distribution. These characteristics are considered important information since they are influenced by the preparation techniques and any modification that is made on the membrane; moreover, they affect membrane performance (Capannelli et al. 1983). Many techniques can be used to characterize the porous structure. Usually the gravimetric (Archimedes) method is used to determine the porosity (Hou et al. 2009), using the following equation:

where Wj is the wet membrane weight; w₂ is the dry membrane weight; p_w is the density of wetting agent; and p_p is the density of polymer. Water is usually used as the wetting agent for hydrophobic membranes, while isopropyl alcohol is used as the wetting agent for hydrophilic membranes.

Another method of porosity evaluation is the mercury intrusion method (Lorente-Ayza et al. 2017). In this method, the intrusion volume of mercury in the membrane pores is recorded. However, this technique uses high pressure, which can cause sample deformation. In addition, in this technique, the pores that determine the flux in the selective layer cannot be distinguished from those in the support (the large pores).

Porosity can also be determined by measuring the polymer material density using isopropyl alcohol that penetrates inside the membrane pores and the density of the membrane using water that does not penetrates through the pores. For density determination a balance and a pycnometer are used. Then, the following equation is used:

where p_m is the membrane density, while p_r is the polymer density (Khayet and Matsuura 2001).

Regarding pore size and pore size distribution, several methods are used. SEM is one approach, in which a software is used on the scanned images of the surface, in order to measure the pores sizes (Abdullah et al. 2014. Wyart et al. 2008). Gas permeation method is another technique (Hou et al. 2009). In this technique, the gas is tested using nitrogen, and a flow meter is used to measure the permeation flux. The permeate flux is plotted versus the pressure, from which the intercept (K„) and the slope (P„) are determined. The average pore size is calculated using the following equation:

where R, T, M and p are the gas constant, the absolute temperature, the molecular weight and the gas viscosity, respectively.

The bubble point method is used for determination of the largest pore size in the membrane (Dickenson 1997). The membrane is put in a container with gas from the bottom side and liquid on the topside of the membrane. The gas increases steadily with time. The bubble point is reached when a constant flow of bubbles is seen on the membrane’s topside. The pore size is then determined as the pore diameter (d) using the following equation:

where у is the surface tension; в is the contact angle; and P is the pressure at which the bubbles are formed.

Another approach for determination the pore size is by applying permeability data in mathematical models, such as the Capillary Tube (CT) model (Eygeris et al. 2018). In the CT model, the pore radius is calculated as follows:

where К is the permeability that can be determined using Darcy’s law (Eygeris et al. 2018); r is the tortuosity; and f is the porosity of the membrane.

Regarding pores at nanoscale, nanopermporometry is used for the determination of pore sizes in the range of 0.5-50 nm. This technique depends on the capillary condensation of a vapor such as water, hexane, or isopropyl alcohol, and on its ability to block the permeation of noncondensable gases such as He and N₂ permeation (Albo et al. 2014, Tsuru et al. 2003). For small pore size, the vapor condenses at a vapor pressure (P_v) that is lower than the saturation vapor pressure (/*,). The pore radius is then calculated using the Kelvin equation, as follows:

where v is the molar volume; a is the surface tension; 0 is the contact angle; and i'k is the radius of the pore.

Electron Microscopy

Electron microscopy is used to obtain very high-resolution images using a beam of accelerated electrons, which have very short wavelength under vacuum, as the source of illuminating radiation. Two main types of electron microscopy are scanning electron microscope (SEM) and transmission electron microscope (ТЕМ).

SEM

Scanning electron microscope (SEM) produces images when a focused electron beam is scanned across a specific surface area of the specimen. When the electron beam interacts with the surface, it loses energy, which is converted into other forms that provide signals carrying information about the properties of the surface, such as its composition and topography.

For sample preparation, membranes are usually dried overnight at room temperature (Abdullah et al. 2014). After that, samples are sectioned into small strips and mounted on an SEM stage for skin layer (top view), while they are frozen in liquid nitrogen, and fractured to obtain flakes for cross-section imaging, which give information about the membrane morphology and structure. This is followed by sputter coating by gold, palladium, platinum, osmium, or other metals to enhance the surface electrical conductivity, thereby avoiding the charging effect for SEM observation (Hou et al. 2009, Huang et al. 2019, U.S. Department of the Interior Bureau of Reclamation 2009).

SEM is usually used for characterization of membrane topography and morphology. It can be used to quantitatively characterize the membrane surface roughness and the porosity, the average pore diameter, the pore size distribution, and the pore aspect ratio, which can be determined with the help of such software products as MATLAB and Image J (Abdullah et al. 2014, Wyart et al. 2008). In one study, (Wei et al. 2013) a thin-film composite nanofiltration hollow fiber membrane was examined using SEM to study the surface morphology and the cross sections of the composite’s membranes. It was found that there is a difference in surface morphology between the outer surface and the inner surface of the hollow fiber membrane. The outer surface was found to be smooth, while the inner one was found to be rough surface that was packed with tiny nodules. In addition, the middle and outer edge of the hollow fiber membrane were porous and full of finger-like configurations, as seen in Figure 7.1.

In addition, SEM can be used to examine the foulants’ nature at the microscopic level (US Department of the Interior Bureau of Reclamation 2009) and

FIGURE 7.1 FE-SEM images of the fabricated composite NF hollow fiber membrane (a), outer surface (b), inner surface (c), cross section x 200 (d), and cross section x 5 K. (Wei et al. 2013).'

1 Reprinted from Chemical Engineering Journal 223. Wei, Xiuzhcn, Xin Kong, Chengtian Sun, and Jinyuan Chen, “Characterization and application of a thin-film composite nanofiltration hollow fiber membrane for dye desalination and concentration.” 172-182, Copyright (2013), with permission from Elsevier to make a qualitative analysis of the foulant types that are deposited on the surface of the membranes (Butt et al. 1997). SEM images are capable of showing the different thickness of fouled materials on different membranes, for a variety of membrane materials and water conditions (Sachit and Veenstra 2017).

It can also be used to examine the surface binding in case of composite membranes, or detect any nanoparticles included in the membrane matrix, in addition of determination the nanoparticle size in the polymer matrix (Kadhom and Deng 2019, Khulbe and Matsuura 2017). Khadom et al (Kadhom and Deng 2019) used SEM to measure the size of bentonite in polymeric membranes. SEM images showed that bentonite nanoparticles were around 50nm.

Transmission Electron Microscopy (ТЕМ)

In ТЕМ, the image is formed as a result of interaction between the electrons and the specimen, which occurs when the focused ion beam is transmitted through the specimen.

Sample preparation depends on the type of the specimen, which is deposited on a support grid in case it is a powdered substance or a nanotube. Other specimens such as biological tissues are embedded in a resin to withstand the high vacuum, before being cut into very thin sections (less than 100 nm thickness), using ultra-microtome. In cases where the specimens can withstand the high vacuum, such as some polymers; it is ultra-microtomed as it is, without being embedded. Sometimes, specimens need to be stained.

ТЕМ is used in membrane characterization to study the morphology and microstructure of the samples (Huang et al. 2019). Xu et al (Xu et al. 2019) used ТЕМ to examine graphene quantum dots incorporated in thin film composite (TFC) membranes. ТЕМ images showed that the membranes had thin skin layers and fingerlike pores, as seen in Figure 7.2. They were able to determine the thickness of the substrate layer, which was found to be 60 nm on average.

In addition. ТЕМ can be used to examine the nanoparticles’ shape and size inside the polymer matrix of the membrane. In one study, ТЕМ images showed the bentonite nanoparticles as grey blocks inside the membrane. The effect of the nanoparticles on the structure of the membrane when compared to unfilled thin film membrane was revealed, in which the thickness of the membrane was increased in the presence of nanoparticles from (100-300 nm) to 500 nm (Kadhom and Deng 2019).

Atomic Force Microscopy (AFM)

AFM is another common tool for studying surface topography. In AFM, atomic force is used to plot the probe - sample interaction. This interaction between the sample and the probe tip is used to form a three-dimensional image of the sample surface at a very high resolution.

AFM is usually used to examine the roughness of the membrane, and the pore size and distribution, in addition to the phases on the membrane surface (Khulbe and Matsuura 2017).

FIGURE 7.2 The cross-sectional ТЕМ images of the H-PAN, the pristine TFC-Omembrane, and the TFC-0.05 membrane (Xu et al. 2019).²

2 Reprinted from Desalination 451. Xu, Shengjie, Feng Li. Baowci Su, Michael Z Hu, Xucli Gao, and Congjie Gao. “Novel graphene quantum dots (GQDs)-incorporatcd thin film composite (TFC) membranes for forward osmosis (FO) desalination.” 219-230, Copyright (2019), with permission from Elsevier.

It is important to check the roughness because it is expected to positively affect the water flux owing to higher specific area as found by Xu et al (Xu et al. 2019) who examined thin film membranes that are incorporated with graphene quantum dots for forward osmosis desalination. They examined the roughness of the membranes using AFM. Those filled with graphene quantum dots showed higher surface roughness and higher flux.

The roughness of the membrane surface is also an important characteristic that affects fouling/antifouling properties. Fouling occurrence increases at surfaces with high roughness (Mahmoud et al. 2015). Zhao et al (Zhao et al.

2013) examined the surface of PVDF and GO/PVDF membranes using AFM. They found that pure PVDF membranes had high roughness, while the GO/PVDF membranes had smoother surfaces, which indicated that addition of GO to PVDF decreases the fouling potential because higher membrane surface roughness leads to contaminants accumulation. Similar results were found by Safarpour et al (Safarpour et al. 2015) who incorporated rG0/Ti0₂ nanocomposites in a polyamide layer. The roughness of the membranes was determined using AFM. It was found that increasing the content of the nanocomposite in the polymer layer reduced the roughness of the membranes.

In another study, composite polysulfone membranes were examined for the effect of grafting on the roughness of the membranes using AFM. It was found that the grafted polysulfone membranes had rougher surfaces than the untreated polysulfone membranes, which indicated that the graft polymerization was successful, as shown in Figure 7.3. (Akbari et al. 2010).

AFM can be used to calculate the root mean square roughness of reverse osmosis (RO) membranes at different directions over different samples (Ferrero et al. 2011). Moreover, it is used to calculate the pore size and pore size distribution, and surface roughness of the membranes (Shirazi et al. 2013).

Another use of AFM is the understanding the mechanism of membrane formation, as found by Huang et al (Huang et al. 2019) who formed hydro- thermal reduction with time for reduced graphene oxide (rGO) membranes. They found that regularly intact rGO membranes that have black appearance were formed under treatment times of 0.5, 1, and 2 h. However, when treated for more than 4h, rGO sheets agglomerated and formed particles instead of films.

Fourier-Transform Infrared (FTIR) Microscopy

FT1R microscopy techniques are used to characterize the contaminants that accumulate on the membrane surface. They provide information about fouling distribution on the membrane surface (Ferrero et al. 2011), and are used together with FTIR spectroscopy to allow both visualization of the surface and analysis of the contents accumulating on the surface. Ferrero et al (Ferrero et al. 2011) used FTIR microscopy in the characterization of fouling distribution on the reverse osmosis membrane surface. They were able to conclude that the membrane surface is covered by proteins, polysaccharides, and inorganic species as shown in Figure 7.4.

FIGURE 7.3 AFM images of pristine polysulfone (PSF) ultrafiltration (UF) membrane and PSF-grafted-PAA after (a) 1 hour, (b) 2 hours, (c) 3 hours, (d) UV irradiation time, (Akbari et al. 2010).³

Reproduced from A. Akbari et al. (2010) Water Science & Technology, 62.11 (2655-2663) with permission from the copyright holders, IWA Publishing.

FIGURE 7.4 Infrared images for the main components found in the membrane surface: proteins (1711-1480 cm'¹), inorganics (1291-936 cm'¹), polysaccharides (1165-1076 cm'¹). Combination of the three FTIR images in a RGB graphic: proteins (green), inorganics (red), polysaccharides (blue). (Source: Ferrero et al. 2011).