Desalination of Seawater Using Cationic Poly(acrylamide) Hydrogels and Mechanical Forces for Separation

and SO 42 − ) drastically decrease the swelling capacity of previously employed anionic poly(acrylic acid- co -sodium acrylate) hydrogels, whereas no influence on the swelling behavior of the synthesized cationic hydrogels is found. The swelling behavior and mechanical properties are studied by varying the degree of crosslinking and degree of ionization systematically in the range of 1–5 and 25–75 mol%, respectively. Finally, artificial seawater ( c sea = 0.171 mol L − 1 ) is desalinated in a custom-built press setup with an estimated efficiency of E m³ = 17.6 kWh m − ³ by applying an external pressure on the swollen hydrogels. of salt depleted water. Moreover, the mechanical moduli have to be minimized to reduce the energy demand for deswelling the particles. This decoupling of the mechanical strength from the charge density is the aim of our next studies. Furthermore, we want to investigate blocking phenomena related to biofouling and clogging that drastically reduce the membrane efficiency of the reverse osmosis processes. Since the particles in the here proposed technique exhibit alternating swelling and deswelling cycles, a backwashing and self-cleaning effect could achieve a reduction of blocking phenomena.


Introduction
Nowadays, one of the major global risks on human health and economic growth is the lack of potable water supply. It is estimated that 80% of the population is exposed to major threats to water security, which is especially pronounced for areas with rapid population growth and low access to freshwater reservoirs. [1] Since only 0.5% of the world's water is available as freshwater source, a potential method to overcome water scarcity is to implement cost-efficient desalination techniques. [2] Currently, the total desalination capacity stands at 95.37 million m³ per day with ≈16 000 operating desalination plants. [3] Historically, thermal desalination processes, such as multistage flash distillation (MSF), have been predominantly used in the www.advancedsciencenews.com www.mme-journal.de finally, salt depleted water is recovered by applying an external force on the swollen hydrogel particles using a custom-built press setup, as depicted in Figure 1. Applying this desalination approach, potable water from a 35 g L −1 (0.171 mol L −1 ) NaCl solution having similar ionic strength as seawater could be obtained with an estimated energy demand of 8.9 kWh m −3 , utilizing highly crosslinked (degree of crosslinking (DC) = 5 mol%) poly(sodium acrylate) hydrogels. [22] In addition, the impact of various network architectures, such as surface crosslinked and interpenetrating hydrogels, on the separation behavior has been investigated. [23] Other studies demonstrated the utilization of temperature (50 °C) instead of pressure as an alternative, renewable stimulus with thermally responsive N-isopropylacrylamide hydrogels. [24] Furthermore, analytical thermodynamic models based on the early work of Katchalsky were developed and used to calculate the total energy cost that is required to reduce the salt concentration from 0.6 down to 0.001 mol L −1 over multiple swelling-deswelling cycles. The energy demand was estimated to around 1.5-3 kWh m − ³. [25] Previous studies were predominantly aimed on the technical realization and improvement of the desalination efficiency toward salt water by optimizing the hydrogel properties as well as the experimental parameters. Pure sodium chloride solutions were used as a rather simple model system for mimicking seawater which however is a much more complex system, containing multivalent ions, such as Mg 2+ , Ca 2+ , and SO 4 2− . These multivalent ions can interact with the negatively charged polyelectrolyte backbone and influence the swelling behavior by reducing the swelling capacities of previously employed anionic poly(acrylic acid-co-sodium acrylate) (PSA) hydrogels. [26,27] Thus, we propose the use of cationic hydrogels as an alternative separation medium to overcome the aforementioned limit.
The here presented study is organized as follows: first, a series of cationic poly(acrylamide-co-(3-acrylamidopropyl) trimethylammonium chloride) (PAPTAC) hydrogels are synthesized by varying the DC and degree of ionization (DI) systematically in the range of 1-5 and 25-75 mol%, respectively. Second, the swelling behavior of the synthesized cationic hydrogel is investigated in different salt solutions (MgCl 2 , CaCl 2 , and Na 2 SO 4 ) as well as artificial seawater with concentrations varying from 0.017 to 0.17 mol L −1 and compared to an anionic PSA hydrogel (DC = 1 mol%; DI = 75 mol%) that is used as a reference system. In addition, rheological measurements are performed to study the impact of the hydrogel composition on the network formation and mechanical properties. Finally, salt rejection measurements are conducted to investigate the influence of charge density on salt partitioning. As a proof of concept, the most promising sample is used for the desalination of artificial seawater in the press setup which is displayed in Figure 1.

Synthesis of Cationic Poly(acrylamide-co-(3-acrylamidopropyl)trimethylammonium chloride) Hydrogels
PAPTAC hydrogels were synthesized by free radical polymerization (FRP) of acrylamide (AM) and (3-acrylamidopropyl) trimethylammonium chloride (APTAC) in water with N,N′methylenebis(acrylamide) (MBA) as difunctional crosslinking agent, as displayed in Figure 1. The synthesis procedure was adapted from literature and slightly optimized as follows. [28] First, a 20 wt% solution of monomers (AM, APTAC and MBA) in water was prepared and the polymerization started after the Left: Schematic drawing of the press setup for the desalination of seawater utilizing cationic polyelectrolyte hydrogel particles as a separation medium. Salt depleted water is recovered by deswelling the particles with an external force. The cationic poly(acrylamide-co-(3-acrylamidopropyl) trimethylammonium chloride) (PAPTAC) hydrogel particles are synthesized by free radical copolymerization using the respective monomers in water. Right: Technical realization of the press setup. [22] The height of the piston x, elution volume V, and the pressure p are determined with an uncertainty of Δx = 1 μm, ΔV = 0.5 mL, and Δp = 0.01 bar. The ion content is measured offline via inductively coupled plasma optical emission spectroscopy (ICP-OES) and ion chromatography.
www.advancedsciencenews.com www.mme-journal.de addition of APS and TEMED as radical-redox initiating system. To take into account the amount of 4-methoxyphenol (MEHQ) inhibitor that is present in the APTAC monomer solution, the amount of APS und TEMED was adjusted and calculated according to Equation (1). Samples with a varying degree of crosslinking (DC = 1, 3, and 5 mol%) and degree of ionization (DI = 25, 50, and 75 mol%) were synthesized, which are given by Equations (2) The synthesis of the sample with DC = 1 mol% and DI = 25 mol% is here given as an example. First, MBA (0.103 g, 0.667 mmol) was dissolved in deionized water (25.3 g). Subsequently, AM (3.55 g, 0.05 mol) and APTAC (3.45 g, 0.0167 mol) as 75 wt% solution were added. TEMED (0.0516 g, 0.444 mmol) was added and the mixture was cooled with an ice bath and purged with nitrogen for 30 min. Then, APS (0.101 g, 0.444 mmol) was dissolved in deionized water (2 g) and added to the mixture to initiate the polymerization. The gelation typically started within 1 h and was allowed to complete overnight. The hydrogels were cut into small pieces and placed in a large excess of deionized water for 4 days to remove unreacted chemicals and sol content. The purified samples were dried in a vacuum oven at 60 °C for 3 days. The dry particles were grinded and sieved to obtain particles with sizes between 300 and 600 μm. The anionic PSA hydrogel reference system with DI = 75 mol% and DC = 1 mol% was synthesized in a similar fashion, using sodium hydroxide to deprotonate carboxylic acid and generate negative charges in the network. The procedure has been described in previous work in more detail. [17] An overview of the synthesized samples is displayed in Table 1.

Water Absorbency Measurements
The degree of swelling at equilibrium (Q eq ) was measured gravimetrically. ≈10 mg of the dried hydrogel sample (m p ) with a particle diameter between 350 and 650 μm was placed on a metal sieve (m sieve ) with a mesh size of 120 μm. The sieve was put on a metal rack, which was then placed in a Petri dish filled with the solution to let the particles swell overnight. Then, the sieve with the swollen particles was pressed on a paper towel to remove excess solution and weighed (m swollen ). The degree of swelling was calculated according to Equation Hydrogel samples were measured in aqueous sodium chloride solutions with concentrations ranging from 0.017 to 0.71 mol L −1 . Artificial seawater containing the most prominent salts (NaCl, MgCl 2 , Na 2 SO 4 , and CaCl 2 ) was prepared according to the composition of ASTM international (ASTM D1141) with concentrations similar to the sodium chloride solutions between 0.017 and 0.71 mol L −1 . [29] The ratio between the salts was kept constant for various concentrations. Absorbency measurements in pure multivalent salt solutions (MgCl 2 , Na 2 SO 4 , and CaCl 2 ) were performed with concentrations ranging from 0.34 to 128 mmol L −1 . The reported results are the average of three independent measurements with a standard deviation of σ Q < 5%.

Rheology and Sample Preparation
Mechanical properties of the hydrogels were analyzed via oscillatory shear experiments on the strain-controlled rheometer ARES G2 (TA Instruments, Eschborn, Germany). Hydrogel samples were prepared in a cylindrical mold with a diameter of 30 mm to obtain uniform disc-shaped specimens with a height of 5 mm. The mold was sealed and the crosslinking reaction was allowed to proceed overnight. The disc-shaped specimens were measured directly without further purification. The test geometry was a 30 mm diameter plate made of aluminum. The geometry was lowered until a constant axial force of 0.5 N (177 Pa) was applied to the sample and the temperature was controlled to 25 ± 0.1 °C by a Peltier element (Advanced Peltier System, TA Instruments). First, an oscillatory strain sweep from γ 0 = 0.01 to 1000% was performed for every sample at a constant frequency of ω = 6.3 rad s −1 ( f = 1 Hz) by varying the strain to determine the linear viscoelastic (LVE) regime. The values at a strain of γ 0 = 0.1% were chosen to be representative for the LVE regime. Frequency sweeps for three different specimens of the hydrogel sample were conducted at a fixed strain of γ 0 = 0.1% and the frequency was varied from ω = 0.2 to 630 rad s −1 . From these experiments the mean value of the elastic (G′) and viscous The varied synthetic parameters degree of crosslinking (DC) and degree of ionization (DI) are given in columns 2 and 3 while the sample names are derived from these quantities. The abbreviations PAPTAC and PSA refer to the cationic poly(acrylamide-co-(3-acrylamidopropyl)trimethylammonium chloride) and anionic PSA hydrogel, respectively.
www.advancedsciencenews.com www.mme-journal.de (G″) moduli were taken at a frequency of ω = 6.3 rad s −1 (f = 1 Hz) to compare the rheological behavior of different hydrogel samples. The error bars in the respective graphs show the standard deviation around the mean value.

Salt Rejection Measurements
The salt partitioning between the gel phase and the external solution was analyzed via salt rejection measurements in a 1 wt% NaCl (0.171 mol L −1 ) solution. The dried hydrogel samples were mixed overnight with the sodium chloride solution (total mass of the mixture m total = 40 g) in a specific ratio Q rel = m s /m absorb = 2, where m s refers to the mass of the brine solution and m absorb to the maximum water uptake according to Q eq . The fixing of the parameter Q rel = 2 ensures that one half of the brine volume is located in the swollen hydrogel and the other half in the supernatant phase, as depicted in the second step of Scheme 1. The salt concentration of the supernatant phase was analyzed by conductivity measurements (SevenMulti, Mettler Toledo, Gießen, Germany) upon an appropriate calibration. [9] The ratio of the salt concentration in the supernatant phase, c out , to the initial concentration, c 0 , defines the salt rejection (SR)

Desalination Experiment
The desalination of seawater was conducted on a custombuilt hydraulic press setup, as illustrated in Figure 1. The apparatus consists of a press chamber with a total volume of 400 mL and a piston that transfers the pressure to the sample, while a sieve unit (3-5 μm pore size) holds back the hydrogel particles. A detailed description of the construction is found in previous work. [9] During the compression of the particles the pressure inside the chamber and the volume flux is measured by a pressure sensor (SD-40, Suchy Messtechnik, Lichtenau, Germany) and a distance gauge (MarCator 1086, Mahr, Göttingen, Germany), respectively. In the first step of the experiment, dry hydrogel particles were swollen overnight in artificial seawater with a concentration of c sea = 0.171 mol L −1 (11.6 g L −1 ) that resembles the concentration of 1 wt% NaCl solution c NaCl = 0.171 mol L −1 (10 g L −1 ). The mass ratio of seawater to swollen hydrogel was kept constant (Q rel = 2), as described in the salt rejection measurement and in previous work. [23] The mixture is transferred into the press chamber and the supernatant phase is removed by applying a small pressure (1 bar). The sample is then compressed with an increasing pressure of 1 bar min −1 to a maximum pressure of 80 bar to ensure a constant volume flux. The emerging water was continuously collected in fractions and the ion content measured offline by inductively coupled plasma optical emission spectroscopy (ICP-OES) and ion chromatography (IC) for cations and anions, respectively (see the Supporting Information for more details). The expended energy E of the desalination process was calculated by the numerical integration of the pressure-volume work according to Equation (6) while the removed salt mass Δm (salt) is calculated by Equation (7) salt salt, initial salt, fractions where V i is the volume of fraction i. Subsequently, the ratio of the expended energy E to removed salt mass Δm (salt) was defined as the parameter κ, which is used to quantify the energy demand of the desalination and calculate the specific energy E m³ to remove 35 kg of salt in one cubic meter seawater by Equations (8) and (9) (NaCl) A detailed description of the calculation can be found in previous studies. [22,23] In total, the basic principle of the desalination follows a three-step process, as displayed in Scheme 1.

Swelling Behavior
During the desalination process salt depleted water is recovered by deswelling the particles (see Scheme 1), hence the hydrogel must exhibit a certain minimum degree of swelling at equilibrium (Q eq > 10 g g −1 ) in seawater to obtain a reasonable outflow. Water absorbency measurements are conducted for anionic www.advancedsciencenews.com www.mme-journal.de PSA and cationic PAPTAC hydrogels to compare the swelling behavior as the function of the artificial seawater concentration c sea , as shown in Figure 2.
The Q eq values decrease monotonically with higher salt concentrations due to the reduction of the difference in the osmotic pressure between hydrogel and the external solution. [20] The obtained values for Q eq are fitted by a power law following the general equation, eq salt = × Q a c n , which refers to the screening of charges along the polymer backbone according to the reduction of the Debye length, D salt 1/2 λ ∼ − c . [30,31] As a result, the electrostatic repulsion between monomer units is reduced and consequently Q eq decreases. However, lower values for the exponent are also found. The power law is valid for weakly crosslinked polyelectrolyte chains and values for the exponent n < −0.5 are expected to arise with increasing DC due to the restricted motion of the chains. [23] In contrast to the cationic hydrogel with n = −0.32 in seawater, a distinct deviation for the anionic hydrogel with a sharp decrease of Q eq in seawater with n = −0.83 is observed. This decrease in seawater is more than two times higher than in sodium chloride solutions. Seawater contains multivalent ions, such as Mg 2+ , Ca 2+ and SO 4 2− . During the swelling process the sodium counterions of the anionic carboxylate groups are exchanged with Mg 2+ and Ca 2+ ions which induce a reduction of the swelling capacity through ionic interactions with the polymer backbone. [26,32] The cationic hydrogel, however, shows a strong salt resistance toward seawater and exhibits similar values for Q eq as in sodium chloride solution. Moreover, this result suggests that divalent sulfate anions are not affecting the swelling capabilities of the cationic PAPTAC hydrogel. To further verify the salt resistant behavior of cationic PAPTAC toward divalent sulfate ions, the water absorbency is measured in pure Na 2 SO 4 salt solution. As a comparison the water absorbency of anionic PSA hydrogels is measured in MgCl 2 and CaCl 2 solutions, as depicted in Figure 3.
In pure Mg 2+ and Ca 2+ solutions, the anionic PSA hydrogel exhibits a strong salt sensitivity in the degree of swelling. Consequently, at low salt concentrations (c salt < 0.01 mol L −1 ) a sharp volume transition is found where Q eq is decreased approximately by a factor of 25. At this concentration a three times higher value is found in seawater, as shown in Figure 2. This deviation from seawater shows that the influence of multivalent ions depends on the ratio of salts in the external solution, as monovalent ions and divalent ions compete for the negatively charged carboxylate groups. [27,30,31,33] In contrast, an interaction of cationic hydrogels with divalent SO 4 2− anions is not observed. Hence, utilizing APTAC as a charged monomer with a quaternary amine functional group (-N + Me 3 ) for the introduction of positive charges in the polymer backbone facilitates a resistance of Q eq toward multivalent ions that are present in seawater.
To study the influence of synthetic parameters on the hydrogel swelling behavior, water absorbency measurements are performed for different PAPTAC compositions with varying DC and DI, as displayed in Figure 4.
The swelling behavior of the cationic hydrogels in NaCl solution and seawater shows similar values, which further indicates that the degree of swelling is mainly influenced by the synthesis parameters and is not altered by additional salts that are present in seawater. The Q eq decreases with higher DC which influences the average molecular weight, c M , between two adjacent crosslinking points. Therefore, at lower DC a looser network with longer elastic chains and higher absorption capacity is formed. [34,35] For large swelling ratios (Q eq > 10 g g −1 ) this  where β is a constant related to the specific volume of the polymer ν , the molar volume of water V 1 , and the Flory polymer-water interaction parameter χ 12 . [36,37] The average molecular weight between crosslinking points is proportional to the degree of crosslinking, c M ∼ DC −1 . Accordingly, the PAPTAC hydrogels follow this scaling law with exponents ranging from −0.46 to −0.62 decreasing with higher DI values. To explain the dependency on DI, two effects have to be considered. Primarily, DI determines the amount of positive charges in the polymer backbone. At higher DI the osmotic potential of the hydrogel is increased, and thus higher Q eq are observed. In addition, DI refers to the molar fraction of the positive charged monomer APTAC in the hydrogel composition and with increasing DI lesser amounts of AM are present. It is known that the use of AM as a copolymer increases the mechanical properties and stiffness of polyelectrolyte hydrogels. [38,39] Thus, at lower DI the highest strengthening effects and lower Q eq values are expected which is in agreement with the water absorbency measurements.

Rheological Investigation
To further study the effect of the initial hydrogel composition on the mechanical properties, oscillatory shear measurements are performed. [40,41] Frequency sweeps for all hydrogel samples show that the elastic modulus (G′) is constant over the whole frequency range (0.1-100 Hz) while exceeding the loss modulus (G″) by a factor of 10 which confirms network formation throughout the whole sample range (see Figure S1 in the Supporting Information). Values for G′ range from 1.5 to 24 kPa and increase with higher DC. However, a distinct deviation of G′ for different DI is observed, as shown in Figure 5.
As predicted by the rubber elasticity theory, G′ increases as a function of DC. [36,42,43] With increasing concentration of the crosslinking agent, more but shorter elastic chains are introduced into the network, exerting higher elastic forces during deformation. Additionally, G′ increases with decreasing DI, indicating that the hydrogels become more rigid as the concentration of AM rises in the initial composition. The highest values for G′ are found for hydrogel compositions with high amounts of AM (DI = 25 mol%), which exceed samples with low amounts (DI = 75 mol%) by a factor of at least 5 at equivalent DC. This deviation indicates that the initial monomer composition distinctly influences the mechanical properties and is further demonstrated in the values of tan δ that is used as a measure to quantify the extent of viscous contributions in the material (tan δ = G″/G′). With increasing DC, tan δ decreases in a linear fashion for high DI (75 mol%) and is constant for lower DI values. Note that the hydrogel samples are analyzed in the as-prepared state without purification. Hence, unreacted moieties and non-crosslinked chains (sol) are present during the measurement. Increased tan δ values and lower elastic moduli (G′) further indicate that the crosslinking efficiency is reduced as the concentration of APTAC increases in the reaction mixture. The presence of enhanced electrostatic repulsion effects between APTAC monomer units and the propagating polymer chains as well as the repulsion of neighboring polymer chains during network formation may explain a reduction in the crosslinking efficiency. [38,39]   www.advancedsciencenews.com www.mme-journal.de

Salt Rejection Measurements
Utilizing polyelectrolyte hydrogels as a separation medium for desalination involves two mechanisms or stages. First, the dry hydrogel particles absorb water due to the hydrophilicity and the difference in the osmotic pressure. Then, according to the Donnan theory the fixed charges and counterions in the polymer backbone hinder mobile salt ions from further entering the hydrogel network and establishing an equal concentration inside the gel phase and the surrounding solution. [44] Therefore, the concentration of mobile salts inside the gel is lower compared to the external solution as there are additional charge contributions from the network. This distribution of mobile salts between the gel phase and supernatant phase is referred to as salt partitioning. [45] Consequently, the combination of both effects, water absorption and the salt partitioning, leads to a salt enrichment in the supernatant phase. The ratio of the concentration in the supernatant phase (c supern ) to the initial salt concentration (c o ) can be further defined as salt rejection (SR), see Equation (5). As discussed in previous work, [9,22] the efficiency criterion for the desalination considers the energy expended during compression of the hydrogel particles as well as the amount of removed salt according to Equations (8) and (9). Hence, the extent of the salt partitioning is a crucial quantity for the desalination efficiency, and SR measurements can be utilized as a rather easily accessible tool for the estimation of the desalination efficiency in the press setup. The SR is a function of the fixed charge concentration, and thus highly depends on Q eq . At higher Q eq values, more water is absorbed and SR decreases. By plotting SR as the function of the fixed charge concentrations, the obtained values can be compared to the Donnan theory as shown in Equation (11) [9,23] where c 0 is the initial concentration of the external solution and c + is the concentration of APTAC monomer units (fixed positive charges) inside the hydrogel according to DI and the parameter f eff resembles the effective charge fraction with values ranging from f eff = 0 to 1. The experimental data obtained for SR is plotted as the function of Q eq and c + , as depicted in Figure 6.
Considering the simplicity of the Donnan model, the experimentally obtained SR values can be described as the function of the charge density using approximately an effective charge fraction of f eff = 0.4. Correspondingly, high SR values up to 17 mol% are achieved at low Q eq values as less water is absorbed. Multiple sources attribute to the deviation from the ideal Donnan theory. Previous studies showed that the Donnan theory fails to describe quantitatively the salt partitioning of highly charged hydrogels as the salt rejection is overpredicted. [9,45] Additionally, the effective charge density along the polymer backbone may be reduced by the so called Manning condensation, which occurs when the distance between two neighboring charged monomer units is lower than the Bjerrum length (l Bjerrum = 0.7 nm in water). [46] At this critical threshold counterions condense on the polyelectrolyte, limiting the effective charge density. In the herein presented cationic hydrogel system, the distance between positive charges is expected to be larger than the Bjerrum length, since six chemical bonds in the APTAC side group separate the positive charges of the quaternary amine moieties from the polymer backbone. However, the theory of Manning describes an idealized linear charged chain of zero radius and does not take into account any molecular properties of polyelectrolyte chains, for instance the stiffness and restricted rotation of the amide bonds. Hence, deviations from the theory are expected especially at highly crosslinked hydrogels, which additionally contribute to the reduction of the charge density. Furthermore, the APTAC concentration, and thus the concentration of fixed positive charges is calculated from the mass fraction of polymer. We estimate the content of APTAC from the initial hydrogel composition obtained by the DI. The highest value for SR is found for the sample with Figure 6. Salt rejection (SR) which is measured in 0.171 mol L −1 NaCl solution for PAPTAC hydrogel with varying degree of crosslinking (DC) and degree of ionization (DI) is plotted as the function of the degree of swelling at equilibrium (Q eq ) (left) and the concentration of APTAC monomer (positive charges, c + ) inside the swollen hydrogel (right). The experimental data is compared to the Donnan model according to Equation (11) with effective charge fractions of f eff = 1, 0.4, and 0.25. The latter was obtained in a previous study for poly(sodium acrylate) hydrogels and is used for comparison. [23] www.advancedsciencenews.com www.mme-journal.de DC = 5 mol% and DI = 25 mol%, which exhibits the highest APTAC concentration and the lowest Q eq . Maximizing SR is crucial to optimize the desalination efficiency. Therefore, it is required to synthesize highly charged hydrogels while maintaining low degrees of swelling to increase the fixed charge concentration.

Desalination Experiments
As a proof of concept, the sample with the highest SR (DC = 5 mol%; DI = 25 mol%) is chosen for the desalination of artificial seawater with a concentration similar to 1 wt% NaCl solution, c sea = 0.171 mol L −1 (11.26 g L −1 ) in the demonstrated press setup (see Figure 1). The results of the desalination process are displayed in Figure 7.
The concentration of the outflow (c out ) decreases linearly with a rate of dc out /dV = −0.13 g ml L −1 as the pressure is increased to deswell the particles, as displayed in Figure 7 (left). A salt reduction of 40% compared to the initial concentration is achieved at the last fraction. The energy demand for the desalination experiment is calculated from the pressure and elution volume by means of the volume work where the numerical integration of the p-V curve gives the expended energy E. [22,23] The energy demand for the desalination of seawater with a concentration of 11.6 g L −1 is estimated according to Equation (9) to E m³ = 17.6 kWh m − ³.
Additionally, a distinct difference of the salt partitioning for different ions is found, as shown in Figure 7 (right). While a strong salt rejection is observed for chloride ions (c rel (Cl − ) = c out /c 0 = 1.1), the salt rejection for divalent sulfate ions is slightly lower (c rel (SO 4 2− ) = 1.05). The counterion for the cationic APTAC monomer during synthesis is chloride with a concentration of c Cl,syn = 0.277 mol L −1 . The chloride concentration inside the gel is therefore already high, prior to the desalination experiment, and thus the concentration gradient from the external solution toward the inside of the hydrogel is weaker compared to sulfate anions. The tendency to diffuse into the hydrogel is stronger for sulfate ions. As a result of the diffusion, two chloride ions are exchanged with sulfate to maintain electroneutrality which further enhances the salt rejection of chloride. A distinct salt rejection for the cations is not observed and the cations distribute evenly between the gel and the supernatant phase. However, as shown in the water absorbency measurements the multivalent ions do not influence the swelling behavior of the synthesized cationic hydrogels which makes them suitable as separation medium in multivalent ion rich solutions. In contrast, PSA hydrogels show a network collapse in the presence of multivalent cations, fully preventing the desalination of seawater. Hence, our emphasis in future studies is to extend the application of cationic hydrogels as separation medium while optimizing the desalination efficiency, for instance, by decoupling the charge density from mechanical properties.

Conclusion
In this study, a series of cationic poly(acrylamide-co-(3-acrylamidopropyl)trimethylammonium chloride) hydrogels with varying DC and DI are synthesized for the application as separation medium in the desalination of seawater. The swelling behavior is studied in sodium chloride solution and artificial seawater while comparing it to an anionic PSA reference system. By studying the swelling behavior, we found that the cationic hydrogels are resistant toward the seawater composition that contains multivalent ions, such as Mg 2+ , Ca 2+ , and SO 4 2− , whereas a strong decrease of the swelling capacity is observed for the anionic hydrogel. Artificial seawater with a concentration of 11.6 g L −1 was desalinated with a custom-built press setup in a threestep process utilizing cationic hydrogel particles as separation medium. The energy demand of the desalination approach was estimated to E m³ = 17.6 kWh m − ³ with a total reduction of the salt concentration by up to 40%. To further increase the efficiency of the process, the charge density of the hydrogel particles has to be maximized while simultaneously maintaining reasonable degrees of swelling (Q eq > 10 g g −1 ) to allow a sufficient outflow Figure 7. Left: Desalination of seawater (c 0,total = 0.171 mol L −1 , 11.6 g L −1 ) with the custom-built press setup in a three-step process using cationic PAPTAC hydrogel particles with DC = 5 mol% and DI = 25 mol% as separation medium. An external pressure (p < 1 bar) by lowering the piston is applied to remove the salt enriched supernatant phase where c sea > c 0,total . Thereafter, the pressure is linearly increased (1 bar min −1 ) to deswell the particles and recover salt depleted water. The outflow is collected continuously in 3-10 mL fractions and subsequently their ion content measured. Right: Relative concentration (c rel = c out /c 0,salt ) of mobile ions shows the salt partitioning between the gel phase and the external solution during the desalination experiment.
www.advancedsciencenews.com www.mme-journal.de of salt depleted water. Moreover, the mechanical moduli have to be minimized to reduce the energy demand for deswelling the particles. This decoupling of the mechanical strength from the charge density is the aim of our next studies. Furthermore, we want to investigate blocking phenomena related to biofouling and clogging that drastically reduce the membrane efficiency of the reverse osmosis processes. Since the particles in the here proposed technique exhibit alternating swelling and deswelling cycles, a backwashing and self-cleaning effect could achieve a reduction of blocking phenomena.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.