Novel draw solutes of iron complexes easier recovery in forward osmosis process

Along with the rise in sea levels, the status of salinity is occurring in many countries in the world including Vietnam, which seriously impacts on agricultural production and life in these areas. As a result, many people will lack clean water to use in their daily life. To cope with these issues, seawater desalination technologies have been found by scientists worldwide. Desalination can be defined as any process that removes salt from water. At present, two types of technologies that are widely used around the world for desalination can be broadly classified as either thermal technology or membrane technology. These technologies need energy to operate and produce fresh water. According to Zheng et al. (2015), membrane technology is commonly applied for industrial wastewater treatment in China with around 6.7 million m³ of wastewater per day (2.4 billion m³ per year) treatment capacity applying membrane technology. Han (2013) used a reverse osmosis (RO) membrane filtering system for seawater desalination in the first phase of the Hongyanhe Nuclear Power Plant in China. The system has a maximum capacity of 16,000 m³/day and purifies some 5.8 million m³ of seawater per annum for power generation and domestic water. Forward osmosis (FO) membrane is a valuable technology which drives clean water permeates from feed solution to draw solution by the osmotic pressure difference across a semi-permeable membrane. With a low operation cost, the FO membrane technology has been widely applied in wastewater treatment (Achilli et al. 2009), brackish water and seawater desalination (McCutcheon et al. 2006), food processing (Petrotos & Lazarides 2001; Garcia-Castello et al. 2009), and power generation (Seppala & Lampinen 1999; Achilli et al. 2009). According to Phuntsho et al. (2013), the major factors affecting the performance of FO include membrane properties, draw solution properties, feed solution properties and the operating conditions. Determining the suitable draw solutes becomes particularly crucial for the FO process. Ge et al. (2013) and Klaysom et al. (2013) indicated that the ideal draw solute should include the principal traits such as: high solubility, minimal reverse draw solute leakage, easy regeneration, non-toxicity, low cost, and compatibility with membrane.

Inorganic salts were investigated as draw solutions in the FO process in previous studies due to their low cost and high osmotic pressure potential which creates a high flux (Achilli et al. 2010; Alnaizy et al. 2013). According to Kiriukhin & Collins (2002), the low charge and small hydrated radius of monovalent and divalent ions in the draw solution can result in a high reverse flux of salts, such as 0.6 M NaCl (7.2 GMH), 0.6 M NH₄HCO₃ (18.2 GMH) when deionized (DI) water was used as the feed solution. On the other hand, a high amount of energy as a driving force was required by the RO membrane to produce freshwater and recover these draw solutions (Zhao et al. 2012). In the past few decades, many studies have reported on the development of suitable draw solutions for an FO process; and a few reviews on draw solutions have been published to date (Chekli et al. 2012; Ge et al. 2013). Draw solutions can be generally classified as volatile compounds, organic compounds, inorganic compounds and novel synthetic compounds, including MNPs and polymer hydrogel. Some draw solutions showed justifiable water flux such as: 2-methylimidazole-based organic compounds (Yen et al. 2010), switchable polarity solvents (Stone et al. 2013), ferric and cobaltous hydro acid complexes (Ge et al. 2014) and poly (sodium 4-styrenesulfonate) (Tian et al. 2015). However, high reverse salt flux and relatively energy-intensive regeneration make them impractical in FO desalination. Therefore, identifying novel draw solutes with characteristics of high water flux, low reverse salt flux, and easy recovery is necessary in the FO process.

Thus, in this study, a series of ammonium complexes were investigated as draw solutes in the FO process on parameters such as water flux, reserve flux, solute rejection, and the results of the FO process after a 1-hour experiment. Exploration of suitable draw solutes plays an important role in developing the further advancement of FO processing and even commercialization of an FO membrane (Ren & McCutcheon 2014; Shaffer et al. 2015).

The pH value, total dissolved solids (TDS) and electrical conductivity of the iron complexes solutions were determined by pH, TDS and conductivity sensors (LAQUA pH/Ion/Cond Meter F-74BW, HORIBA, Japan) at different concentrations: as 0.05, 0.10, 0.15 and 0.20 M (A2S), as 0.05, 0.10, 0.15, 0.20, 0.40, 0.50 and 1.00 M (A3S) and 0.05, 0.10, 0.15, 0.20 and 0.40 M (A3C), respectively.

The experimental setup is shown in Figure 1. The active layer of FO membrane was oriented to face the feed solution in order to reduce internal concentration polarization and thus obtain high water flux, with the support layer facing the draw solution. The total membrane surface area was 32 cm². During experiments, the feed and draw solutions on both sides of the FO membranes were re-circulated by using two separate pumps with the same flow rate of 1.2 L/min, respectively. The system was operated at laboratory conditions. Each treatment was repeated three times over 1 hour at room temperature (around 25 ± 5 °C). pH, TDS and conductivity sensors (LAQUA pH/Ion/Cond Meter F-74BW, HORIBA, Japan) were used to monitor any changes in the containers of feed and draw solutions periodically every 15 min during experiments. The initial volume of draw solution was 0.2 L, and a 0.2 L feed solution glass was placed on a digital scale balance (Sartorius, Goettingen, Germany) to monitor the weight and volume changes at specified time intervals (Figure 1(a)). After the FO test, the diluted draw solution was recovered for reuse through an NF-90 membrane by using a cross flow module at room temperature (see Figure 1(b)). The NF-90 (DOWN) membrane was made from polyamide materials which can be used in pH ranging from 2 to 11 with 6.68 LMH/bar of pure water permeability. The filtration experiments were repeated three times using fresh membranes. The selection of the most appropriate experimental concentration for each complex is made before comparing test results between complexes.

Water flux, reserve flux and solute rejection were determined according to Hau et al. (2014) and Tian et al. (2015) as shown below.

Two different cations such as Fe⁺² and NH₄⁺ were found in these iron complexes. A2S is classified as a double salt of ferrous sulfate and ammonium sulfate. It forms monoclinic crystals, and is soluble in water at approximately 269 g/L. About 1,240 g/L of solubility in water and the weakly violet and octahedrical crystals were noted in A3S complex whereas A3C appeared in yellow crystals and was very soluble in water.

The solution properties of various iron complexes with different concentrations were characterized and are shown in Figure 2, and this played an important role to further check whether iron complexes could act as draw solutes. There is little effect on concentrations on the pH of various iron complexes. pH of different concentrations of A3S ranged from 1.96 to 2.53, then was higher in A2S (3.79–3.88) and was maximum in A3C at 7.26–7.56. The recommended operational pH ranges of the current commercial thin film composite FO membrane is from 2.0 to 12.0. Therefore, the FO membrane will not undergo hydrolysis and structure change when it is tested in the draw solutions of the iron complexes, which can ensure the consistent FO performance. As shown as Figure 2, the solution conductivity was increased with increasing concentrations of various iron complexes. This is in contrast with the studies of Zhao et al. (2016) where the solution conductivity of four EDTA complexes was not directly proportional to that of the increment concentration, and appeared to approach plateaus or even decrease slightly after an initial increase. In addition, the difference existed in conductivity for the studied iron complexes following the order of A2S > A3C > A3S. According to Ge et al. (2012), the higher conductivity may lead to higher osmotic pressure. The electrical conductivity in the iron complexes solution mainly derives from two sources including the dissociation of iron ion and ligand which is affected by the property of iron ion and external conditions. Therefore, the measured conductivity of ion complexes solution results from the joint effects.

The results showed that 0.1, 0.4 and 1.0 M were the most suitable concentrations of ammonium iron (II) sulfate, ammonium iron (III) citrate and ammonium iron (III) sulfate complexes in FO process, respectively. A comparison between the concentrations of these complexes was performed to select the potential draw solute for an FO process. The iron complexes as draw solutes were evaluated by quantifying reserve salt flux and water flux in FO tests, which is shown in Figure 3. In addition, NH₄HCO₃ was chosen as a benchmark for comparison with iron complexes group. The results in Figure 3 show that the experimental water flux and reserve salt flux is a function of molar concentration in FO mode where de-ionized water is employed as the feed solution. We found that the water flux of A3S (11.24 LMH) was obviously higher than A2S (8.88 LMH), A3C (8.88 LMH) and the control group (NH₄HCO₃) (8.62 LMH). This is directly related to the fact that high water flux can generate a greater osmotic driving force for water transport through the membrane. The results report that 36.0 mL of water volume was found in A3S which was obviously higher than A2S (28.4 mL), A3C (28.4 mL) and the control group AB (27.6 mL). However, the highest reserve salt flux was 2.62700 (GMH) in A3S solution, followed by A3C (1.44450 GMH), the control group (1.31415 GMH), and the lowest at A2S (0.00143 GMH). This indicated that the molecular sizes of complexes were ranged in the order: A3S < A3C < AB < A2S. Hence, molecular size leads to draw solutes permeating back to the feed solution through the FO membrane, which may affect the next stages of fresh water production. Both water flux and water volume of both A2S and A3C were higher than AB. On the other hand, the reserve salt flux of A2S was obviously lower than A3C and AB. According to Kiriukhin & Collins (2002), the reserve flux of salts were 7.2 GMH in 0.6 M NaCl, 18.2 GMH in 0.6 M NH₄HCO₃, 5.6 GMH in 0.5 M MgCl₂ when DI water was used as the feed solution. In the studies of Hau et al. (2014), the results showed that the lowest reverse salt flux of 0.067 GMH was observed when 0.1 M EDTA-2Na was coupled with 15 mM. NP7 was used as a draw solution and DI water was used as a feed solution in FO mode (active layer facing with the feed solution). These results were much higher in comparision to the reserve flux of ammonium iron (II) sulfate in this study. This demonstrated that only a few ammonium iron (II) sulfate salts permeated back to the feed solution from the draw solution through the FO membrane. This indicated that A2S may be the potential draw solute in an FO process.

The draw solution is diluted after the FO process, thus the recovery of draw solution is necessary and represents one of the challenges in FO processing (Zhao et al. 2014). The large molecular weight and size of iron complexes might be separated from draw solution by a membrane with a larger pore size than that of RO. The recovery convenience of iron complex draw solution was studied through a pressure-driven NF-90 process. Table 2 shows the variation in the TDS permeate and removal efficiencies obtained using various draw solutions as A2S, A3S and A3C in comparison with AB at operation of 8 bars. Over 90% of draw solutes did not permeate through NF-90 membrane, in which 90.96% and 93.14% removal efficiencies were reported in A2S and A3C draw solution, respectively; and 95.52% of A3S was the maximum removal efficiency. In previous studies, Hau et al. (2014) indicated that the NF recovery (pressure of 5.5 bar) of EDTA sodium salts exhibiting high charged compounds performed well, and had a high salt rejection of 93%. NF membrane was used by Ge & Chung (2013) to regenerate hydroacid complexes after FO, which had expanded configurations and charged groups. According to Hau et al. (2014), the recovery of 0.05 M EDTA-2Na coupled with 15 mM NP7 was approximately 95%, which is significantly higher than the A2S and A3S draw solutions, but a little less than the A3C draw solution. These results demonstrated that a series of iron complex draw solutions can be easily recovered using NF-90 membrane, which plays an important role in the FO process to produce fresh water using iron complexes as draw solutes substitute AB.

In summary, various iron complexes were investigated as draw solutes in an FO process. The results showed that high water flux of these iron complexes was reported in the range 8.88–11.24 LMH which was higher than the AB draw solution. In addition, more than 90% iron complexes draw solutions were recovered by NF-90 membrane, which plays an important role in the FO process to produce a fresh water substitute AB.

This project was supported by the Vietnam Academy of Science and Technology (No. VAST.CTG.08/14-16). The first two authors contributed equally to this paper.