Investigating the performance of hydroponic nutrient solutions as potential draw solutions for fertilizer drawn forward osmosis

This research project aims at investigating the performance of hydroponic nutrient solutions as draw solutions for desalination using the fertilizer drawn forward osmosis (FDFO) process. Six different lettuce and leafy greens hydroponic nutrient stock solutions were prepared according to the literature and used in this study and tested on a bench-scale forward osmosis unit as draw solutions for the process. The feed solution for the process was deionized water mixed with NaCl in different concentrations, to represent different salinities of brackish groundwater. The draw efficiency of each solution was measured based on water flux, specific reverse solute flux, water recovery, and salt rejection. It was concluded that of the six tested nutrient solutions, the “Resh Florida, California” solution is the recommended solution to be used as draw solution for fertilizer drawn forward osmosis, due to its high performance in terms of water recovery (15.75%), flux (11 L/m2/h), salt rejection (92%), and SRSF (highest recorded SRSF for a specific ion (SO42−) was 7.3 g/L), as well as its low cost, relative to the other highly performing draw solution “Chekli” ($1.07/L vs. $3.73/L).


Introduction
Water is among the most essential natural resources and is utilized in nearly all human activities; it is necessary for domestic use, agriculture, and industry, is one of the sustainers for ecosystems, and sometimes even is a hazard, such as in the cases of floods and droughts (Anandhi and Kannan 2018). The water cycle is affected by many factors, including pollution, land use changes, population growth, climate change, urbanization, and industrial growth (Navarro-Ortega et al. 2015;Koutroulis et al. 2013). Globally, water demand is projected to increase by 55% by the year 2050, mainly due to growing demands from manufacturing (increasing by about 400%), thermal electricity generation (increasing by about 140%), and domestic use (increasing by about 130%) (Shahzad et al. 2018).
Water is the most essential resource for agricultural production and is very critical to food security; irrigated agriculture forms about 20% of the total cultivated land, and it contributes about 40% of the total global food production (The World Bank 2020). Agriculture is the biggest consumer of freshwater ("blue" water), where almost 90% of diverted freshwater in the world is used for irrigated agriculture (Traitler et al. 2018). This led to the introduction of the water-energy-food nexus, which is a concept that is central to sustainable development; the tangled connection between those 3 crucial domains means that managing water resources is essential for food security, while water resources are affected by climate change, which in turn is affected by global energy usage and production.
Many societies now face the issue of water scarcity, which is a situation defined as having the water availability in a country or region be less than 1000 m 3 /person/year (Pereira et al. 2009). Almost a quarter of the human population, approximately 1.6 billion people, live under water scarcity, and this number may double in two decades (Roson and Damania 2017). It is also projected that by 2050, more than 40% of the world's population will be living in areas of severe water stress (Shahzad et al. 2018).
In order to combat shortages in clean freshwater resources, more countries are resorting to desalination, where desalination widens the potential of utilization of water resources by producing freshwater from saline or brackish natural water sources (Burns et al. 2015). Water desalination processes are categorized into thermal desalination and membrane desalination. Thermal desalination processes are generally characterized by high costs and being energy-intensive; hence, new facilities have been designed to utilize more membrane desalination processes (Cotruvo 2016). On the other hand, membrane desalination technologies have started to gain traction as the preferred desalination technology in the past few decades, due to advancements in membrane technologies. The commercially established membrane desalination technologies are RO, NF, and ED (Burn et al. 2015).
Hence, recent efforts have been directed at developing less energy and capital costly desalination processes, such as forward osmosis (FO). FO has the advantage of operating without the need for hydraulic energy and operating at a lower pressure than RO, which decreases the capital cost . The disadvantages of FO are that the resultant of the process is a diluted draw solution, from which freshwater would need to be separated, which requires further processing. Moreover, membrane technologies and draw solutions are still in need of further development before FO could be commercially widespread (Phuntsho et al. 2012a;Nasr and Sewilam 2015).
Fertilizer drawn forward osmosis (FDFO), demonstrated in Fig. 1, is an application of FO in which the draw solution is a highly concentrated fertilizer solution, which is diluted by drawing freshwater from the feed solution (saline or brackish water). It is a method of overcoming the inherent disadvantage of FO, in which the resultant end product is diluted draw solution, not freshwater. This diluted draw solution could hypothetically be used directly to grow agricultural crops, if it was diluted to the required concentration for agriculture.
FDFO technology can be integrated with modern agriculture, such as hydroponics, to capitalize on water and energy savings. Hydroponics, demonstrated in Fig. 2, is defined as the practice of growing plants without using soil, either by using an inert medium, such as gravel, sand, peat, vermiculite, pumice, perlite, coco coir, sawdust, and rice hulls, or other substrates, to which a nutrient solution that contains all the essential elements needed by the plant is provided (Resh 2012). Hydroponic systems are divided into liquid systems, which have no supporting medium for the plant, and aggregate systems, which have a solid medium for support (Jensen 1997).

Theory
In the following sections, the theory behind forward osmosis and hydroponic agriculture will be explored.

Forward osmosis
When investigating processes related to forward osmosis, 3 main parameters have to be examined. These parameters are water flux, reverse solute flux, and salt rejection.  (Phuntsho et al. 2012a, b) Flux When a feed solution of a low osmotic pressure is separated from a draw solution of high osmotic pressure by a semipermeable membrane, the concentration of the draw solution decreases, and the concentration of the feed solution increases due to the solvent transfer caused by the difference in osmotic pressure (Δπ) , where osmotic pressure (π) can be defined as "the amount of pressure that is required to prevent the transport of pure water through an ideal semipermeable membrane into a specific solution" (Mulder 1996, as cited in Shon et al. 2015. The model that is commonly used to describe mass transfer through an FO membrane is the solution-diffusion model; in an FO process, the water flux ( v ) and solute flux ( s ) can be calculated by Lee et al. (1981), as cited in Shon et al. 2015: Equations 1 and 2.
where A is the membrane water permeability coefficient; B is the solute permeability coefficient; C is the solute reflection coefficient, usually assumed to be unity; and ΔP is the applied hydraulic pressure, which is zero ( P = 0) in the FO process, and hence, water flux can be expressed as the multiplication of the membrane water permeability and the difference in osmotic pressure.
where draw and feed are the osmotic pressures of the draw solution and feed solution, respectively .
Equation 3, however, only describes the theoretical ideal situation. In experiment, the FO water flux is found to be much lower than the ideal value of the product of membrane water permeability coefficient and the osmotic potential difference of the draw and feed solutions, due to the (1) concentration polarization phenomenon . In FO process, concentration polarization exists on the both external sides of the membrane when there are solutes in both the feed and draw solutions . The build-up of solutes on the feed side of the membrane is defined as concentrative external concentration polarization (concentrative ECP), while the depletion of solutes on the draw side of the membrane is referred to as dilutive ECP Bhattacharjee 1998, as cited in Shon et al. 2015). However, the effect of ECP, demonstrated in Fig. 3, on the FO flux is relatively small and could be mitigated through manipulating the flux and mass transfer coefficient, through approaches such as adjusting the crossflow velocity and using spacers, vibrations, or ultrasound (Lee et al. 1981;Sablani et al. 2001;Cath et al. 2006, as cited in Shon et al. 2015).
A more severe problem compared to ECP that contributes to low FO water flux is internal concentration polarization (ICP) . Semipermeable membranes often have a thicker and porous support layer underneath the thin selective layer, in order to provide enough support strength . The support layer often inhibits mass diffusion and, hence, leads to the accumulation of a concentration polarization layer within it , which is the phenomenon that is referred to as ICP and is demonstrated in Fig. 4. In FO, the asymmetric semipermeable membrane can be installed with the active rejection layer (AL) facing either the draw solution (referred to as AL-DS) or the feed solution (AL-FS) .
Concentrative ICP (AL-DS): Dilutive ICP (AL-FS): (4) where A is water permeability coefficient, B is solute permeability coefficient, and K m is the mass transfer coefficient, defined as: where D is the solute diffusion coefficient and S is the structural parameter of membrane, defined as: where , l , and are the tortuosity, the actual thickness, and porosity of the porous support layer, respectively.

Reverse solute flux (RSF)
Reverse solute flux (RSF) is an important parameter to consider in the FO process, since a reverse diffusion of the draw solutes would cause an economic loss due to the inability to recover these draw solutes (Cath et al. 2006, as cited in Nasr and Sewilam 2016). Moreover, reverse solute permeation increases the occurrence of fouling by creating complexes between draw and feed ions (Cath et al. 2006;Lay et al. 2010, as cited in Nasr and Sewilam 2016). The FO solute flux can be determined by Tang et al. (2010), as cited in Shon et al. 2015: Equation 8 (8)   where is the van't Hoff coefficient, R g is the universal gas constant, and T is the absolute temperature.

Specific reverse solute flux (SRSF)
Specific reverse solute flux (SRSF) is the ratio between RSF and water flux. SRSF is a measurement that was created to signify the mass of draw solutes that was lost through reverse permeation per unit volume of water extracted from the FS (Cath et al. 2006, as cited in Nasr andSewilam 2015). SRSF can be calculated by using the equation below: By combining Eqs. 9 and 10 , SRSF can be expressed as:

Hydroponic nutrient solutions
In hydroponic nutrient solutions, 3 main parameters guide the successful selection of solutions. These parameters are the required nutrients for the plant, the solubility of the nutrient salts in water, and the concentration of each of these salts.

Plant nutrition
In order to supply the plants with the essential elements in hydroponics, fertilizer salts are dissolved in water to form hydroponic nutrient solutions (Resh 2012), and many researchers have developed basic formulae for nutrient solutions, such as Hoagland (1950), Arnon (1938), Robbins (1946), Hewitt (1967), Cooper (1975), Steiner (1961, Bollard (1966), Middleton (1973), Shive, and Tollens. These nutrient solutions are often referred to as standard nutrient solutions. While various fertilizer salts can be used in the nutrient solutions, the choice of salts is affected by their solubility (Resh 2012). It is well established that the external concentration of each of the nutrients that is required to satisfy the internal demands of plants varies between different species; there cannot be a single nutrient solution that is effective under all circumstances (Smith et al. 1983). Sixteen elements have been identified as essential for the growth of higher plants, and they are divided into macronutrients (H, C, O, N, K, Ca, Mg, P, S), which are those required in relatively large quantities, and micronutrients (Cl, Fe, Mn, B, Zn, Cu, Mo), which are those required in smaller quantities (Resh 2012).

Solubility
Solubility of fertilizer salts is a property that dictates the concentration of the salt that remains in solution when dissolved in water; in hydroponics, fertilizer salts with a high solubility are necessary in order for them to remain in solution and be available to be absorbed by the plants (Resh  2012).
Moreover, the proportion of ions that is provided by each of the salts is important to consider when designing the nutrient solution; for example, one molecule of calcium nitrate (Ca(NO 3 ) 2 ) will yield one ion of calcium (Ca 2+ ) and two ions of nitrate 2(NO 3 − ) (Resh 2012), and depending on the type of plant that is being grown, different combinations and concentrations of fertilizer salts will have to be used in order to achieve optimum growth.

Concentrations
The amount of nitrogen (N) in an ion of nitrate (NO 3 − ) can be calculated by calculating the fraction that each element represents within its compound source, in order to calculate the conversion factor (Resh 2012).
The conversion factors can be utilized to find the ppm concentration of each element that is obtained by dissolving a specific compound. For example, if 100 mg of calcium nitrate (Ca(NO 3 ) 2 ) is dissolved in 1L of water, the concentration of Ca will be 24.4 mg/L (24.4 ppm), and the concentration of N will be 17.1 mg/L (17.1 ppm). Table 1 lists the concentration of the elements in some of the standard hydroponic nutrient solutions.

Materials and methods
The materials used in this study consist of the apparatus, which is the bench-scale crossflow filtration unit, the membrane, and the solutions used in each experiment.

Apparatus
For this experiment, a bench-scale crossflow filtration unit with an FO cell was used, illustrated in Fig. 5. On both sides of the membrane, two channels are connected to allow flow of feed solution on one side and draw solution on the other side. Crossflows were operated in countercurrent flow directions, by using a variable speed peristaltic pumps (Stenner, model 170DMP5, 25 psi, 1.7 bar, 50 Hz, USA). All solutions were kept at a temperature of 25 °C through a temperature water bath controlled by a heater/chiller (PolyScience temperature controller, model 9106A12E). Experiments were operated at a crossflow rate of 400 mL/min, which can be converted to a crossflow velocity of 8.5 cm/s. The change in the volumes of the DS in the DS tank and FS in the FS tank was continuously recorded by placing the DS and FS on a digital mass scale, which was connected to a computer for online data logging at intervals of 3 min. This change of volume was used to calculate the water flux across the membrane.
The initial volume of both the DS and the FS was 250 mL, and most experiments were run for 2.5 to 3 h, since a stable flux was usually achieved after the first hour of operation. Experiments were conducted under (AL-FS) orientation, where the active layer was facing the feed solution. Equations 11, 12 and 13.
Water flux J v (in L/m 2 /h) was calculated using:  Solute flux was calculated as follows: where V i is the initial volume of FS, ΔV is the total volume of water displaced from the FS to the DS, and C s is the concentration of the draw solutes in the FS at the end of the experiment.
The FO membrane that will be used in this study was purchased from Porifera Inc. Its active layer is made out of polyamide, and its support layer is made out of a porous hydrophilic polymer (Tayel et al. 2019;Porifera Inc. 2015). The average total thickness of the membrane is 70 ± 10 µm (Tayel et al. 2019;Porifera Inc. 2015). Properties of the membrane, as provided by the manufacturer, are listed in Table 2 (Porifera Inc. 2015).
Moreover, since salt rejection is a crucial parameter in FO processes, it will be investigated by taking samples from the DS after every experiment and analysing its Na + and Cl − ions; the following equation will be used to calculate salt rejection percentage (Nasr and Sewilam 2016): where C i is the initial concentration of the ion in FS, C p,D is the final concentration of the ion in DS, V i is the initial volume of the DS, and ΔV is the total volume of water that entered the DS from the FS (Nasr and Sewilam 2016).

Draw solutions
The chemicals and fertilizer salts were provided by the AUC WEF (Water-Energy-Food) Nexus Labs, and some additional salts were purchased from Science and Technology Center, Egypt. All chemicals used were laboratory and reagent grades.
The lettuce and leafy green nutrient solutions formulae were selected from  to as (RA). • Optimum Grow-twin pack hydroponic nutrient solution used in Chekli's paper, according to the formula of macro-and micronutrient mentioned in the paper (Chekli et al. 2017)-referred to as (CHE). • Robbins (1946)-This solution satisfies the ratio of macro-and micronutrient recommended for growing lettuce in Mediterranean climate in the summer, shown below in Table 3 (Resh 2012), referred to as (ROB).
Listed in Table 4 are the macro-and micronutrient compositions (in mg/L) of the selected formulae.
Stock solutions are prepared in different strengths, depending on the size of the application. Fifty, 100, and 200 are the most common, but strengths as high as 800 times can be used in large-scale applications. Stock solutions are then diluted accordingly when used for hydroponic agriculture (Resh 2012).
Hydroponic nutrient stock solutions are prepared in two solutions: A and B solutions. This is done to avoid any precipitation that might occur between different compounds, such as nitrates and sulphates. The problem of precipitation is especially likely when these compounds are used in very  high concentrations (e.g. for 100 and 200 strength solutions). Solution B of each nutrient solution was selected as the draw solution for the forward osmosis process, due to the fact that they include all macronutrients (N, P, K, Ca, Mg, S), as well as most micronutrients (Mn, B, Zn, Cu, Mo). While solution A was found to have higher theoretical osmotic pressure in some solutions, the presence of high concentrations of Ca in A solutions could lead to excessive fouling of the membrane if diffusion to the feed solution occurs through bridging mechanisms that occur with organic compounds (Chekli et al. 2017).
In order to prepare the stock solutions, calculations for each formulae were made, in order to determine how much of each of the available compounds should be used to achieve the required ratio. Solutions were prepared in 100 times strength. Listed in Table 5 are the chemicals that make up each solution, as well as the theoretical osmotic pressure of each solution, estimated using Lenntech osmotic pressure calculator (Lenntech n.d.).
It is worth noting that solution A contains high amount of calcium, while solution B contains the phosphates and sulphates, which can both form insoluble precipitations if mixed in such concentrated form. Hence, solutions A and B have to be processed separately in FO in parallel stages, or only one of the solutions can be used as DS for FO, and the other can be later added in diluted form to complete the nutrient solution (Chekli et al. 2017). The latter option was opted for, in order to avoid fouling of the membrane caused by the Ca in A solutions. Fig. 6 is the cost per litre of each of the draw solutions used, according to Science and Technology Center, Egypt. It is worth noting that these prices are for laboratory/reagent grade chemicals. In an industrialscale application, technical grade chemicals would be used, which could be obtained for prices lower than those listed.

Experimental plan
Feed solutions were prepared by dissolving NaCl in DI (deionized) water in 3 concentrations: 5, 10, and 15 g/L. These concentrations represent different salinities of brackish groundwater. Six different hydroponic nutrient solutions were investigated as DS for the process. Eighteen experiments were carried out. All solutions were prepared by dissolving the salts in DI water and stirred with a magnetic stirrer for at least 30 min, in order to ensure complete dissolution. Solutions were then stored in plastic bottles.  For each experiment, 250 mL of both the FS and DS was poured into plastic bottles, and both were weighed using a digital scale. Moreover, TDS was measured by using a portable TDS and EC metre (Hach HQ40D multi). Then, bottles were covered with Parafilm, in order to minimize evaporation. Both bottles were then placed on the crossflow filtration unit, and the unit was run for 2.5 to 3 h for each experiment. After the experiment was over, the system was drained from the remaining feed and draw solutions into the corresponding beakers. After the system was completely drained of draw and feed solutions, which usually took around 20 to 30 min, the bottles were carefully removed and replaced with bottles of DI water. The system was then operated again with DI water, in order to flush the membrane. This flushing was repeated 3 to 4 times, and each flush was between 10 and 20 min. The final weights and TDS of the feed and draw solutions were then measured and recorded. Samples of each solution before and after processing were stored for ion analysis.

Water recovery
The volume of water recovered through forward osmosis was calculated through the weight of FS and DS before and after each experiment. The results are shown in Fig. 7. The highest water recovery was achieved in Chekli solution, with a recovered pure water volume of 39.57 mL, which represents a recovery percentage of 15.8%. This can be explained by its high osmotic pressure, which is demonstrated in Table 4. Water recovery percentage is shown in Fig. 7 and was calculated by finding the ratio between the volume of water recovered through FO and the initial volume of the FS.

Water flux
Water flux was calculated using Eq. 11. The average water flux for each draw solution is shown in Fig. 8. The highest recorded average water flux (12.28 LMH) was observed in the CHE solution when the feed solution was 5 g/L NaCl (FS1), due to the high osmotic pressure (16.43 bar). The lowest value of average water flux (1.5 LMH) was observed in RA draw solution when coupled with a feed solution of 15 g/L NaCl (FS3), which is expected due to the low osmotic pressure of this solution (13.57 bar). While RF had a higher theoretical osmotic pressure, it exhibited slightly lower flux than CHE, and this could be explained by biofouling or scaling occurring in the membrane over the period of operation. It has been proven that while having the active layer of the membrane be facing the feed side (AL-FS) generally generates a more stable water flux performance than AL-DS orientation, the ICP levels are also more severe in this orientation (AL-FS), which leads to relatively lower water flux (Zhang et al. 2012). Water flux was plotted against the TDS of each of the draw solutions. As predicted, the draw solution with the highest TDS exhibited the highest water flux, since water flux is dependent on the osmotic pressure of the draw solution ( draw ). The correlation between flux and TDS is logarithmic, with R 2 being between 0.51 and 0.74, indicating a good fit for the logarithmic model. This is demonstrated in Fig. 9. Figures 10, 11, and 12 are plots of flux vs. time, and they demonstrate the behaviour of the water flux over the duration of the experiment for each of the feed and draw solutions. The data also is represented by a logarithmic fit. It is noticed that flux starts very high in the beginning and then decreases with a high slope till around 40 min, at which point, flux becomes almost constant. It is also noticed that the water flux had not reached zero yet, indicating that osmotic equilibrium was not yet reached and the process could still be operated for a longer period of time. It is also worth noting that in the case of using the 15 g/L NaCl solution as FS, and RTD as DS, flux increased after 60 min, instead of the expected behaviour of falling off. This could be explained by some membrane scaling that was removed by the flow of both solutions on the membrane.

Specific reverse solute flux
Specific reverse solute flux (SRSF) was calculated through Eqs. 9 and 10 for SO 4 2− , Mg 2+ , K + , PO 4 3− , NH 4 + , and NO 3 − and plotted vs. flux in Fig. 13. Figure 14 demonstrates detailed graphs for the separate ions. It can be observed that SRSF for SO 4 and NO 3 was significantly higher than that for the other ions. It should be noted that the higher the membrane selectivity, the lower the SRSF value should be. From Fig. 13, it is noticed that at high flux (≥ 9 L/m 2 /h), SRSF values for all ions range between 0 and 6 g/L. Meanwhile, at lower fluxes (< 9 L/m 2 /h), SRSF values for all ions are relatively higher, with the highest value being 23 g/L. This indicates a high reverse flux of the draw solutes at low water flux, which is an issue inherent to forward osmosis. The attraction between the different ions could also have generated an additional slight driving force across the membrane (Kim et al. 2015;Tan andNg 2008, as cited in Nasr andSewilam 2015). Moreover, dilutive ECP and ICP might have also contributed in the reverse solute flux. This could be explained by the orientation of the membrane utilized (AL-FS), which might have led to accumulation of ions on the draw side of the membrane.

Salt rejection
FS ions (Na + and Cl − ) were calculated using Eq. 13, and the results are shown in Fig. 15. As can be observed in Fig. 15, high rejection of FS ions was achieved in almost all draw solutions, which indicates that the used membrane has high selectivity against feed ions. However, ROB solution achieved a lower rejection, when operated with FS3 (15 g/L NaCl due to its low osmotic pressure (12.94 bars)), which lead to the decrease in the osmotic potential gradient (Δ ) , which is the main source of force in the FO process, which in turn affects the rejection of the membrane.

Conclusion
This study investigated the performance of six hydroponic draw solutions as draw solutions for the FDFO process. The feed solution for the process was chosen to be three different concentrations of NaCl, in order to represent different salinities of brackish groundwater. This was done in order to assess the viability of utilizing forward osmosis to facilitate hydroponic agriculture in arid regions, by diluting the nutrient solutions using water recovered from groundwater through forward osmosis. Performance of the draw solutions was assessed by determining the water flux, reverse solute flux, and the forward rejection of the salt ions in the feed solution. Moreover, the cost per litre of each of the solution was estimated. It was concluded that the TDS of the nutrient solution and the water flux can be related logarithmically.
From Fig. 10, it can be concluded that as a general rule of thumb, when comparing between different hydroponic nutrient solutions, it is expected that the nutrient solution with the higher TDS would usually have the higher water flux. This is due to the fact that the main driving force in forward osmosis is the osmotic potential gradient between the FS and the DS ( draw − feed ), as per Eq. 3. However, the interactions between the different salt ions could possibly generate minor driving forces, which could affect the overall water flux. The water flux of each of the draw solutions during each individual experiment could have also been affected due to membrane scaling caused by monovalent ions (Phuntsho et al. 2014).
SRSF values for all salt ions were higher at low fluxes and lower when operated at high flux so they had an inverse relation. Moreover, it can be observed that across all draw solutions, SRSF values increased as the concentration of the FS increased. This can be justified by the relation between SRSF and water flux. Salt rejection values for the studied draw solutions were within the acceptable range ( ≅ 90%), except at very low osmotic potential gradients. For example, for the Robbins solution (TDS = 26.1 ppt) and FS 3 (TDS = 17.15 ppt), the salt rejection for Na + was 83% and for Cl − was 80%.
According to the work done in this research, it was concluded that of the six tested nutrient solutions for lettuce and leafy greens, the "Resh Florida, California (RF)" solution is the recommended solution to be used as draw solution for fertilizer drawn forward. It is recommended due to its high performance in terms of water recovery (15.75%), flux (11 L/m 2 /h), salt rejection (92%), and SRSF (highest recorded SRSF for a specific ion (SO 4 2− ) was 7.3 g/L)), as well as its low cost, relative to the other highly performing draw solution "Chekli (CHE)" ($1.07/L vs. $3.73/L). It is also recommended to utilize this solution at fluxes higher than 9 L/m 2 /h, in order to minimize SRSF. A large-scale application of FDFO and hydroponic agriculture on otherwise non-arable arid land would have several positive and negative sustainability impacts. The positive environmental impacts on water resources are that by applying FDFO on a large scale and utilizing brackish groundwater, some of the freshwater needed for agriculture can be diverted into another section, which helps mitigate water scarcity. Moreover, FDFO provides an alternative source of clean water for agriculture. As for energy resources, by utilizing a low-energy desalination process such as FO, some of the carbon emissions produced by desalination processes can be avoided or decreased significantly. On the other hand, negative environmental impacts are the disturbance of ecosystems due to building and operating a man-made structure on arable lands, which might disrupt the cycles of the native species in an area. Hence, such consideration must be taken into consideration. Moreover, transportation emissions are also a risk; transporting solutions and crops to and from the greenhouses in remote areas would cause an increase in GHG emissions.
As for the positive economic impacts of FDFO, applying FDFO on a large scale in arid regions opens up the possibilities for new jobs in the agricultural sector. Moreover, FO is significantly less capital-intensive than other desalination processes due to utilizing simpler equipment, and hence, the required investment for desalination is much less. Moreover, the operational costs of FO in terms of energy consumption is considerably less than RO. However, large-scale applications of FDFO might also have negative economic impacts such as transportation costs, where transporting solutions and crops to and from the greenhouses in remote areas would incur some costs.
The positive social impacts of large-scale FDFO are the development of arid regions, where any project carried out in arid regions would necessitate an improvement in the infrastructure of the area, which would definitely benefit inhabitants of the area. Moreover, FDFO could also help raise awareness about water scarcity and the need for development of green technologies to solve that pressing issue.
Future research in the field can be directed at utilizing a real brackish groundwater sample as the FS for the process, experimenting with other leafy greens nutrient solutions from the literature, and utilizing nutrient solutions designed for crops other than leafy greens, such as for tomatoes or pepper. Moreover, experimenting with different concentrations of stock solutions for the same nutrient solution, such as 50 and 200 strength stock solutions, testing the Resh Florida/California (RF) nutrient solution as a draw solution on the pilot scale, performing life cycle assessment of the membrane during FO, growing crops with the resultant draw solutions, experimenting with different types of membranes, and performing an economic analysis of draw solute consumed per area of production of lettuce are also possible ventures of future research on FDFO.