Evaluating Pretreatment Strategies with Modeling for Reducing Scaling Potential of Reverse Osmosis Concentrate: Insights from Ion Exchange and Activated Alumina

Abstract

Reverse osmosis concentrate (ROC) treatment is critical for enhancing water recovery and minimizing concentrate volume for disposal, especially in regions facing water scarcity. This study investigates the application of ion exchange (IX) resins and activated alumina (AA) as pretreatment strategies to mitigate scaling in ROC due to high concentrations of total dissolved solids, hardness (Ca²⁺ and Mg²⁺), and silica. Through a series of Langmuir isotherms, continuous column experiments, and model simulation, two types of strong acid cation IX resins and three types of strong base anion IX resins alongside three types of AA were evaluated. Results indicate that AA exhibits superior performance in silica removal, achieving up to a 65% reduction and maintaining performance for up to 800 bed volume without reaching saturation. Model simulation of a secondary reverse osmosis treating ROC after the IX and AA pretreatment indicated an additional water recovery of ~70% using antiscalants. This study demonstrates the potential for achieving higher water recovery while also identifying opportunities for pretreatment improvement. Challenges such as the limited IX capacity treating ROC, which requires frequent regeneration and increases operational costs, along with the restricted regeneration capacity of AA, underscore the importance of innovation. These findings emphasize the critical need for developing advanced materials and optimized strategies to further enhance the efficiency of ROC treatment processes.

1. Introduction

Water scarcity is a critical challenge, especially in arid regions with low rainfall, high temperatures, and elevated evaporation rates. Nontraditional water sources, such as brackish water, have become valuable resources for desalination to offset freshwater demands [1,2]. In the past decades, reverse osmosis (RO) has drastically transformed the desalination industry, representing over 80% of the capacity for brackish water desalination due to its energy efficiency and straightforward operation [3]. With its widespread adoption, RO comes with challenges such as membrane scaling and concentrate management, but it is expected that these limitations could be overcome by developing innovative treatment technologies, advanced materials and membranes [4]. The limited water recovery is associated with membrane scaling due to elevated concentrations of scale-forming ions in brackish water ROC, including calcium, magnesium, carbonate, sulfate, and silica, coupled with salinity levels that range from 5 to 15 g/L of total dissolved solids (TDS) [2,5,6,7,8].

Common ROC disposal methods used in the U.S. include surface water discharge (45%), sewer discharge (25%), deep well injection (17%), land application (7%), and evaporation ponds (4%), depending on the ROC quality, geographical location, and regulations [4,9]. While regulatory frameworks are designed to ensure ROC discharges meet quality standards when disposed of by land application, surface water discharge, or sewer discharge, uncertainties remain about potential environmental risks. These concerns are amplified by factors such as mismanagement, inadequate monitoring, or the cumulative effects over time, particularly in regions with sensitive ecosystems of limited dilution capacity [10]. Evaporation ponds are practical in arid and semi-arid regions with abundant solar energy, but the requirement for large land areas limits their use in densely populated urban areas [4,5]. On the other hand, deep well injection requires appropriate geologic conditions and carries potential risks such as groundwater contamination and increased susceptibility to seismic events [4,11].

Efforts to increase water recovery and minimize concentrate volume for disposal face significant challenges due to the high scaling potential of ROC. This is primarily caused by sparingly soluble constituents such as calcium, magnesium, and silica, which can precipitate as magnesium hydroxide, calcium carbonate, magnesium carbonate, calcium sulfate, and silicates. These precipitates reduce membrane permeability, increase energy consumption, and result in frequent chemical cleaning or membrane failure [12,13]. However, membrane modification studies continue to explore new materials and coatings to successfully reduce fouling and scaling [14]. Additionally, various technologies, including chemical precipitation, coagulation, softening, and advanced filtration, have been explored to remove these scaling-causing minerals. These approaches have limitations such as the need for substantial chemical additions, pH adjustments, sludge generation, and high costs [15,16,17,18,19,20,21,22,23]. Addressing these limitations requires innovative pretreatment methods to reduce the scaling potential of ROC and enable higher water recovery.

Different strategies have been developed to minimize the impacts associated with inland ROC management and disposal, including the following:

• Developing high recovery and ultra high-pressure primary desalination processes to minimize ROC volume and potentially achieve near-to-zero liquid discharge (ZLD). This approach increases water recovery for brackish water desalination, minimizes brine volume, and reduces brine disposal costs and environmental impacts. Examples of high recovery desalination systems include low salt rejection RO [24,25], closed circuit reverse osmosis (CCRO) [26,27], and high-pressure spiral wound RO elements that can operate up to 120 bar and concentrate brine up to ~120 g/L TDS [28,29]. Recently, the techno-economic assessment of new ion-exchange membrane technologies [30]—electrodialysis brine concentrator [31] and electrodialysis metathesis [32,33]—showed their potential to overcome the limitations of concentrate management through enhanced performance and economic viability. Despite these advances, the high scaling potential and remaining brine generated from brackish water desalination still remain as major concerns.
• Directly treating ROC. For existing facilities, it may be more practical to directly treat the ROC (such as using the high recovery processes in strategy 1 as secondary or tertiary desalination processes) rather than retrofitting the facility with high-recovery processes, although the goal of minimizing the concentrate volume is the same as in strategy 1. The selected ROC treatment technologies should consider technically feasible, cost-effective, and energy-efficient processes in existing and planned desalination facilities.
• Valorization of ROC. Instead of concentrating the ROC and disposing of the solids, valorization aims to recover valuable resources for a circular economy by finding beneficial uses for the extracted compounds, for example, the development of an integrated treatment train using selective electrodialysis (SED), electrodialysis brine concentrator (EDBC), and bipolar membrane electrodialysis (BPED) to convert ROC to low-salinity water for irrigation and acids and caustic streams for industrial uses [34,35].

The ROC treatment strategies mentioned above require appropriate pretreatment to reduce scaling and facilitate resource recovery. A literature review has identified technologies with high selectivity for removing hardness and silica, such as ion exchange (IX) and activated alumina (AA), as promising alternative approaches (Table 1). These methods provide several advantages, including no chemical addition during treatment, no pH adjustment required, no sludge generation, and regenerability of the media [36].

IX has been reported to have a levelized cost of water (LCOW) of USD 1.47/m³ [16], but, according to our knowledge, the LCOW for AA has not yet been reported. The application of IX and AA for ROC treatment has been limited, and further research is required to evaluate their effectiveness in reducing scaling potential [37].

Cation and anion exchange resins can be used for softening and demineralization applications. The functional group in strong acid cation (SAC) IX resins used in softening is sulfonate, while the mobile ion is sodium. The exchange reaction can be written as shown in Equations (1) and (2) [38], where R represents the resin polymer structure. The ion affinity in cation exchange resins follows ${\mathrm{P}\mathrm{b}}^{2+}>{\mathrm{C}\mathrm{a}}^{2+}>{\mathrm{M}\mathrm{g}}^{2+}>{\mathrm{N}\mathrm{a}}^{+}>{\mathrm{H}}^{+}$.

$$2\mathrm{R}\mathrm{N}\mathrm{a}+{\mathrm{C}\mathrm{a}}^{2+}\leftrightarrow {\mathrm{R}}_2\mathrm{C}\mathrm{a}+2{\mathrm{N}\mathrm{a}}^{+}$$

(1)

$$2\mathrm{R}\mathrm{N}\mathrm{a}+{\mathrm{M}\mathrm{g}}^{2+}\leftrightarrow {\mathrm{R}}_2\mathrm{M}\mathrm{g}+2{\mathrm{N}\mathrm{a}}^{+}$$

(2)

Silica can be removed by strong base anion (SBA) IX resins in the hydroxide (OH⁻) form when it is present in its anionic forms H₃SiO₄⁻ and H₂SiO₄²⁻ [39] (Equations (3) and (4)), while the exchange reaction can be written as in Equation (5) [38,40]. The ion affinity in anion exchange resins follows the order of ${\mathrm{S}\mathrm{O}}_4^{2-}>{\mathrm{N}\mathrm{O}}_3^{-}>{\mathrm{C}\mathrm{l}}^{-}>{\mathrm{H}\mathrm{C}\mathrm{O}}_3^{-}>{\mathrm{H}}_3\mathrm{S}\mathrm{i}{\mathrm{O}}_4^{-}>{\mathrm{O}\mathrm{H}}^{-}>{\mathrm{F}}^{-}$.

$${\mathrm{H}}_4\mathrm{S}\mathrm{i}{\mathrm{O}}_{4\left(\mathrm{s}\right)}\leftrightarrow {\mathrm{H}}_3\mathrm{S}\mathrm{i}{\mathrm{O}}_{4\left(\mathrm{a}\mathrm{q}\right)}^{-}+{{\mathrm{H}}^{+}}_{\left(\mathrm{a}\mathrm{q}\right)}\mathrm{p}{\mathrm{K}}_1=9.84$$

(3)

$${\mathrm{H}}_3\mathrm{S}\mathrm{i}{\mathrm{O}}_{4\left(\mathrm{a}\mathrm{q}\right)}^{-}\leftrightarrow {\mathrm{H}}_2\mathrm{S}\mathrm{i}{\mathrm{O}}_{4\left(\mathrm{a}\mathrm{q}\right)}^{2-}+{{\mathrm{H}}^{+}}_{\left(\mathrm{a}\mathrm{q}\right)}\mathrm{p}{\mathrm{K}}_2=11.8$$

(4)

$$\mathrm{R}{\mathrm{O}\mathrm{H}}^{-}+{\mathrm{H}}_3{\mathrm{S}\mathrm{i}\mathrm{O}}_4^{-}\leftrightarrow \mathrm{R}{\mathrm{H}}_3{\mathrm{S}\mathrm{i}\mathrm{O}}_4^{-}+{\mathrm{O}\mathrm{H}}^{-}$$

(5)

In contrast to SBA that is effective for silica removal at high pH conditions, AA has been evaluated to remove dissolved silica due to its high adsorption capacity and selectivity for silica species [41]. The orthosilicic acid (H₄SiO₄ or Si(OH)₄) can attach to the OH⁻ groups in the surface of the hydrated AA, as shown in Figure 1 [42,43]. The AA has been reported to show an improved performance under a pH range of 8.0–8.5 and, contrary to SBA, ions such as SO₄²⁻, F⁻, HCO₃⁻, and NO₃⁻ shows no significant effect on AA [41,44,45].

Table 1. Silica and hardness removal by IX resins and AA reported in the literature.

Treatment	Water Source	Concentration	Experimental Procedure	Hardness Results	Silica Results	Regeneration Results	Ref.
IX resin/SBA	Tap water	TDS = N/A Silica = 20 mg/L	Bench scale column experiments: 100 mL of IX resin in a small column (18 cm × 5 cm) with Q = 55 mL/min	N/A	94% average silica removal before breakthrough (at 50 BV)	N/A	[46]
	BW ROC	TDS = 12,445 mg/L Silica = 160 mg/L	Bench scale isotherm experiments: IX resin concentrations of 2.5–100 g/L in 200 mL of ROC for 24 h	N/A	Langmuir constant: the maximum capacity (qmax) is 19.65 mg/g	N/A	This study
			Bench scale column experiments: 40 mL of IX resin in a 50 mL burette, Q = 8 mL/min (12 BV/h)	N/A	Saturation at 5 BV (0.2 L), with an operational capacity of 5.23 mg SiO2/g of resin	Regeneration in column: regeneration achieved at 5 BV (5 L/L of resin) using 2% NaOH (0.5 M)
IX resin/three types of WAC	Synthetic water	TDS = 8093 mg/L Hardness = 57.1 mg/L as CaCO3	Bench isotherm experiments A	N/A	N/A	N/A	[47]
			Bench scale column experiments: 22 mL of IX resin was packed in a small column (25 cm high) with Q = 7.5 mL/min (20 BV/h)	Saturation at 1922, 2577, and 3232 BV (70 L), total operational capacity of 2–4 meq/g of resin (100–200 mg CaCO3/g of resin B)	N/A	Regeneration in column: regeneration requirement of 2.3–3 meq HCl/meq adsorbed (3.88–4.97 L/L of resin B) using 5% HCl w/v (1.37 N)
IX resin/SAC	Seawater ROC	TDS = 39.13 g/L Hardness = 4200 mg/L as CaCO3	Bench scale batch experiments: IX resin concentration of 10–60 mg/L in 1 L of seawater ROC for 60 min (contact time)	8.23–41.24% hardness removal	N/A	Some regeneration results were reported (75 and 54% of Ca and Mg elution, respectively) using 15% HCl w/v (4.11 N) A	[48]
	Synthetic water	TDS = 5900, 28,800, and 52,600 mg/L Hardness = 1818, 8266, and 11,172 mg/L as CaCO3 B	Bench scale column experiments: glass columns (15 cm × 2.5 cm) with a Q = 10 mL/min and 5 min as contact time	Saturation at 65, 12, and 6 BV	N/A	Regeneration was conducted using 10% NaCl (1.71 M) at Q = 2.5 and 10 mL/min. No results were reported.	[49]
	BW ROC	TDS = 12,445 mg/L Hardness = 3000 mg/L as CaCO3	Bench scale isotherm experiments: IX resin doses of 2.5–100 g/L in 200 mL of ROC for 24 h	Langmuir constant: maximum capacity (qmax): 133.33 mg/g	N/A	N/A	This study
			Bench scale column experiments: 40 mL of IX resin was packed in a 50 mL burette, Q = 8 mL/min (12 BV/h)	Saturation at 17 and 33 BV (0.68 and 1.32 L), total operational capacity of 65.63 and 98.25 g CaCO3/L of resin	N/A	Regeneration in column: regeneration was achieved at 15 and 20 BV (15 and 20 L/L of resin) using 8% NaCl (1.37 M)
AA	BW ROC	TDS = 5800 mg/L Silica = 160 mg/L	Bench scale batch experiments: AA dose 10 g/L, for 60 min (contact time) at 20 °C	N/A	71.90% silica removal	Regeneration batches: 10 g AA (adsorption capacity of 50 mg/g) in 100 mL of 2% NaOH (0.5 M) achieved 80% of silica desorbed after 9 batches	[45]
	Synthetic water	TDS = N/A Silica = 50 mg/L	Bench scale isotherm experiments: AA dose 25 g/L in 100 mL of synthetic water at pH = 8–8.5 and 20 °C	N/A	Langmuir constant: maximum capacity (qmax): 7.943 mg/g Freundlich constant: adsorption capacity (K) 0.379 mg/g	N/A	[44]
			Bench scale batch experiments: AA dose 5–25 g/L, for 30 min	N/A	~42–90% silica removal	N/A
	Cooling tower water	TDS = N/A Silica = 100 mg/L	Bench scale batch experiments: AA dose of 2 g/L in 50 mL of cooling tower water for 2 h at pH = 8.8 and 25 °C	N/A	29% silica removal	N/A	[50]
	BW ROC	TDS = 12,445 mg/L Silica = 160 mg/L	Bench scale isotherm experiments: AA concentrations of 2.5–100 g/L in 200 mL of ROC for 24 h	N/A	Langmuir constant: maximum capacity (qmax) 625 mg/g	26.6% regeneration efficiency after three 1 h batches using 0.1 N HCl	This study
			Bench scale column experiments: 40 mL of AA was packed in a 50 mL burette, Q = 8 mL/min (12 BV/h)	N/A	Column operation until 870 BV (34.8 L), with an adsorption capacity of 217.5 mg SiO2/g of AA	Regeneration in column: 16% of regeneration achieved after 80 BV (80 L/L of resin) using 1 N HCl

Notes: AA—activated alumina; IX—ion exchange; WAC—weak acid cation exchange resin; SAC—strong acid cation exchange resin; N/A—not addressed; BV—bed volume; TDS—total dissolved solids; BW—brackish water; ROC—reverse osmosis concentrate; Q—flow rate; ^A—the conditions of the experiment are not specified; ^B—value calculated from provided data.

Table 1. Silica and hardness removal by IX resins and AA reported in the literature.

Treatment	Water Source	Concentration	Experimental Procedure	Hardness Results	Silica Results	Regeneration Results	Ref.
IX resin/SBA	Tap water	TDS = N/A Silica = 20 mg/L	Bench scale column experiments: 100 mL of IX resin in a small column (18 cm × 5 cm) with Q = 55 mL/min	N/A	94% average silica removal before breakthrough (at 50 BV)	N/A	[46]
	BW ROC	TDS = 12,445 mg/L Silica = 160 mg/L	Bench scale isotherm experiments: IX resin concentrations of 2.5–100 g/L in 200 mL of ROC for 24 h	N/A	Langmuir constant: the maximum capacity (qmax) is 19.65 mg/g	N/A	This study
			Bench scale column experiments: 40 mL of IX resin in a 50 mL burette, Q = 8 mL/min (12 BV/h)	N/A	Saturation at 5 BV (0.2 L), with an operational capacity of 5.23 mg SiO2/g of resin	Regeneration in column: regeneration achieved at 5 BV (5 L/L of resin) using 2% NaOH (0.5 M)
IX resin/three types of WAC	Synthetic water	TDS = 8093 mg/L Hardness = 57.1 mg/L as CaCO3	Bench isotherm experiments A	N/A	N/A	N/A	[47]
			Bench scale column experiments: 22 mL of IX resin was packed in a small column (25 cm high) with Q = 7.5 mL/min (20 BV/h)	Saturation at 1922, 2577, and 3232 BV (70 L), total operational capacity of 2–4 meq/g of resin (100–200 mg CaCO3/g of resin B)	N/A	Regeneration in column: regeneration requirement of 2.3–3 meq HCl/meq adsorbed (3.88–4.97 L/L of resin B) using 5% HCl w/v (1.37 N)
IX resin/SAC	Seawater ROC	TDS = 39.13 g/L Hardness = 4200 mg/L as CaCO3	Bench scale batch experiments: IX resin concentration of 10–60 mg/L in 1 L of seawater ROC for 60 min (contact time)	8.23–41.24% hardness removal	N/A	Some regeneration results were reported (75 and 54% of Ca and Mg elution, respectively) using 15% HCl w/v (4.11 N) A	[48]
	Synthetic water	TDS = 5900, 28,800, and 52,600 mg/L Hardness = 1818, 8266, and 11,172 mg/L as CaCO3 B	Bench scale column experiments: glass columns (15 cm × 2.5 cm) with a Q = 10 mL/min and 5 min as contact time	Saturation at 65, 12, and 6 BV	N/A	Regeneration was conducted using 10% NaCl (1.71 M) at Q = 2.5 and 10 mL/min. No results were reported.	[49]
	BW ROC	TDS = 12,445 mg/L Hardness = 3000 mg/L as CaCO3	Bench scale isotherm experiments: IX resin doses of 2.5–100 g/L in 200 mL of ROC for 24 h	Langmuir constant: maximum capacity (qmax): 133.33 mg/g	N/A	N/A	This study
			Bench scale column experiments: 40 mL of IX resin was packed in a 50 mL burette, Q = 8 mL/min (12 BV/h)	Saturation at 17 and 33 BV (0.68 and 1.32 L), total operational capacity of 65.63 and 98.25 g CaCO3/L of resin	N/A	Regeneration in column: regeneration was achieved at 15 and 20 BV (15 and 20 L/L of resin) using 8% NaCl (1.37 M)
AA	BW ROC	TDS = 5800 mg/L Silica = 160 mg/L	Bench scale batch experiments: AA dose 10 g/L, for 60 min (contact time) at 20 °C	N/A	71.90% silica removal	Regeneration batches: 10 g AA (adsorption capacity of 50 mg/g) in 100 mL of 2% NaOH (0.5 M) achieved 80% of silica desorbed after 9 batches	[45]
	Synthetic water	TDS = N/A Silica = 50 mg/L	Bench scale isotherm experiments: AA dose 25 g/L in 100 mL of synthetic water at pH = 8–8.5 and 20 °C	N/A	Langmuir constant: maximum capacity (qmax): 7.943 mg/g Freundlich constant: adsorption capacity (K) 0.379 mg/g	N/A	[44]
			Bench scale batch experiments: AA dose 5–25 g/L, for 30 min	N/A	~42–90% silica removal	N/A
	Cooling tower water	TDS = N/A Silica = 100 mg/L	Bench scale batch experiments: AA dose of 2 g/L in 50 mL of cooling tower water for 2 h at pH = 8.8 and 25 °C	N/A	29% silica removal	N/A	[50]
	BW ROC	TDS = 12,445 mg/L Silica = 160 mg/L	Bench scale isotherm experiments: AA concentrations of 2.5–100 g/L in 200 mL of ROC for 24 h	N/A	Langmuir constant: maximum capacity (qmax) 625 mg/g	26.6% regeneration efficiency after three 1 h batches using 0.1 N HCl	This study
			Bench scale column experiments: 40 mL of AA was packed in a 50 mL burette, Q = 8 mL/min (12 BV/h)	N/A	Column operation until 870 BV (34.8 L), with an adsorption capacity of 217.5 mg SiO2/g of AA	Regeneration in column: 16% of regeneration achieved after 80 BV (80 L/L of resin) using 1 N HCl

Notes: AA—activated alumina; IX—ion exchange; WAC—weak acid cation exchange resin; SAC—strong acid cation exchange resin; N/A—not addressed; BV—bed volume; TDS—total dissolved solids; BW—brackish water; ROC—reverse osmosis concentrate; Q—flow rate; ^A—the conditions of the experiment are not specified; ^B—value calculated from provided data.

This present research aims to address the critical knowledge gap by exploring the effectiveness of AA and IX resins as pretreatment methods for the removal of hardness and silica from ROC, to reduce its scaling potential for further membrane treatment to achieve ZLD and brine valorization (Figure 2).

2. Materials and Methods

2.1. Water Source

ROC was collected from the Kay Bailey Hutchinson Desalination Plant (KBHDP) in El Paso, Texas in March 2022. Water samples were transferred to sealed 5 gallons high-density polyethylene (HDPE) buckets and kept at room temperature (~25 °C). ROC was characterized by major ions, total hardness, alkalinity, pH, and electrical conductivity, as shown in

Table 2. Chemical characterization and SI of ROC from KBHDP.

Parameters	Unit	ROC
pH	-	8.5 ± 0.1
Electrical conductivity (EC)	mS/cm	18.8 ± 0.2
Dissolved organic carbon (DOC)	mg/L	9.1 ± 0.21
Hardness (as CaCO3)	mg/L	3000 ± 50
Alkalinity (as CaCO3)	mg/L	350 ± 26.5
Silica (as SiO2)	mg/L	160 ± 2
Sodium	mg/L	3382 ± 160.7
Ammonium	mg/L	ND
Potassium	mg/L	90.2 ± 2.7
Magnesium	mg/L	211.67 ± 9.4
Calcium	mg/L	846.7 ± 29.7
Fluoride	mg/L	17.1 ± 1.4
Chloride	mg/L	5634 ± 163.1
Nitrite	mg/L	25.3 ± 2.2
Bromide	mg/L	21.3 ± 2.6
Nitrate	mg/L	15 ± 0.7
Sulfate	mg/L	1393 ± 30.4
Phosphate	mg/L	ND
Total dissolved solids (TDS)	mg/L	12,445 ± 188.6
Predicted scalants	Formula	SI
Anhydrite	CaSO4	−0.58
Aragonite	CaCO3	1.78
Calcite	CaCO3	1.93
Chalcedony	SiO2	0.65
Chrysotile	Mg3Si2O5(OH)4	5.09
Dolomite	CaMg(CO3)2	3.6
Fluorite	CaF2	1.79
Gypsum	CaSO4٠2H2O	−0.28
Quartz	SiO2	1.07
Sepiolite	Mg2Si3O7٠5OH:3H2O	4.25
Sepiolite (disordered)	Mg2Si3O7٠5OH:3H2O	1.35
SiO2 (amorphous)	SiO2	−0.19

Note: ND—non detected.

. The saturation indices (SIs) were calculated using PHREEQC Interactive Version 3.7.3 [51]. Major ions were determined using an ion chromatograph (IC, ICS-2100, Dionex, Sunnyvale, CA, USA). Dissolved organic carbon (DOC) concentration was measured using a carbon analyzer (Shimadzu TOC-V, Kyoto, Japan). The total hardness was calculated from Ca²⁺ and Mg²⁺ concentrations. The alkalinity was measured by titration to a pH 4.6 endpoint using a 0.02 N sulfuric acid solution as titrant and methyl orange as indicator. Silica concentration was determined following the procedure described in Hach method #8185 and samples were analyzed using a spectrophotometer (DR6000; Hach Company, Loveland, CO, USA). The pH and the electrical conductivity (EC) were measured using a pH meter (Model 431-61, Cole-Parmer, Vernon Hills, IL, USA) and a conductivity meter (CON 2700, Oakton Instruments, Vernon Hills, IL, USA), respectively.

2.2. Ion Exchange Resins

IX Experimental Conditions

For silica removal, the two IX resins selected for testing were recommended by the manufacturers, including AmberLite™ IRA402 Cl (DuPont, Wilmington, DE, USA) (denoted as SRIX-1) and Purolite^® A504P (Ecolab, Saint Paul, MN, USA) (designated as SRIX-2). Both SRIX-1 and SRIX-2 were preconditioned to change their ionic form from Cl⁻ to OH⁻ using 1 M NaOH for 48 h, which is typically used for silica removal [45]. The three selected IX resins for hardness removal were recommended by the manufacturers due to their high total exchange capacity, including AmberLite™ IRC83 H (DuPont, Wilmington, DE, USA) (denoted as HRIX-1), Purolite^® SSTPPC60 (Ecolab, Saint Paul, MN, USA) (denoted as HRIX-2, which was also previously tested by Thomson et al. [49] for desalination pilot testing), and ResinTech WACG (ResinTech, Camden, NJ, USA) (denoted as HRIX-3). HRIX-1 and HRIX-3 were preconditioned to change the ionic form from H⁺ to Na⁺ using 1 M NaOH for 48 h to prevent acidification during operation. The primary physical and chemical properties and recommended operational conditions of the IX resins used are listed in

Table 3. Physicochemical properties and recommended operational conditions of IX resins for silica and hardness removal *.

Physicochemical Properties	Silica Removal		Hardness Removal
	SRIX-1	SRIX-2	HRIX-1	HRIX-2	HRIX-3
Copolymer	Styrene/divinylbenzene	Polystyrene crosslinked with divinylbenzene	Crosslinked acrylic	Polystyrene crosslinked with divinylbenzene	Acrylic gel
Matrix	Gel	Macroporous	Macroporous	Gel	Gel
Type	SBA	SBA	WAC	SAC	WAC
Functional group	Trimethyl ammonium	Quaternary ammonium	Carboxylic acid	Sulfonic acid	Carboxylic acid
Ionic form as shipped	Cl−	Cl−	H+	Na+	H+
Total exchange capacity	1.2 eq/L (Cl− form)	1.2 eq/L (Cl− form)	≥4.7 eq/L (H+ form)	≥4.56 eq/L (Na+ form)	≥4.2 eq/L (H+ form)
Water retention capacity	49–59% (Cl− form)	50–60% (Cl− form)	40–50% (H+ form)	37–47% (Na+ form)	43–60% (H+ form)
Particle diameter	600–750 µm	300–1200 µm	500–700 µm	650 ± 50 µm	297–1190 µm
Temperature range	5–100 °C (41–212 °F)	<100 °C (212 °F)	5–120 °C (41–248 °F)	<60 °C (140 °F)	<100 °C (212 °F)
pH range	1–14	1–14	6–14	1–14	>7

Note: * Adapted from the product data sheet of each resin.

.

Isotherm studies were conducted at IX resin doses of 2.5, 5, 25, 50, and 100 g/L, and 250 mL polyethylene bottles with 200 mL of a mixture of ROC and resins were placed in a shaker (200 RPM) for 24 h. The batch test was conducted at room temperature (~25 °C), while the pH was kept at 8.5. Water samples were taken from each bottle after 24 h for water quality analysis.

Based on the results of batch experiments, SRIX-2 and HRIX-2 were selected to evaluate silica and hardness removal, respectively, in column experiments. Both resins were soaked in DI water for 24 h to attain their final size before column filling. Then, 50 mL burettes (with an inner diameter of 1.14 cm) were used as columns. Glass wool was placed at the bottom of the burettes to avoid the loss of resin. Column operation was divided into service cycles, which consisted of column operation until resin saturation, backwash, regeneration, and DI water rinse. Feedwater was pumped down-flow through the burette. As it has been reported that amorphous silica polymerization increases at pH 6.5–9.5 [18], the ROC pH was adjusted to 4.5 for the silica removal column operation using SRIX-2 to prevent any possible silica polymerization, while the original pH of 8.5 was not adjusted for the hardness removal column operation using HRIX-2. The columns were operated until resin saturation to determine the breakthrough curve. Backwash was conducted using DI water for a duration of 5 min every two service cycles to prevent the accumulation of particles in the burette. Regeneration was applied using 2% NaOH solution for SRIX-2 [40], and 8% NaCl solution for HRIX-2 [36,52]. After regeneration, the column was rinsed using a DI water volume of 5 BV. Multiple effluent samples were taken during column operation and regeneration to develop breakthrough and regeneration curves. The operating parameters for column testing are summarized in

Table 4. Service cycle operation of continuous column experiments.

Service Cycle	Resin Bed Volume (BV)	Service Flow Rate (SFR)	Flow Rate (Q)	Empty Bed Contact Time (EBCT)	Hydraulic Loading Rate (HLR)
	mL	BV/h	mL/min	min	mL/cm2/min
Column operation	40	12	8	5	7.84
Backwash	40	75	50	0.8	49.02
Regeneration	40	12	8	5	7.84
DI water rinse	40	12	8	5	7.84

.

Experimental results for the performance of all the IX resins were adjusted to the linearized equation of the Langmuir model (Equation (6)). This model is applicable to monolayer sorption onto homogeneous surfaces with a finite number of identical sites [53].

$${\displaystyle \frac{{\mathrm{C}}_{\mathrm{e}}}{{\mathrm{q}}_{\mathrm{e}}}}={\displaystyle \frac{1}{{\mathrm{q}}_{\mathrm{m}\mathrm{a}\mathrm{x}}{\mathrm{K}}_{\mathrm{L}}}}+{\displaystyle \frac{{\mathrm{C}}_{\mathrm{e}}}{{\mathrm{q}}_{\mathrm{m}\mathrm{a}\mathrm{x}}}}$$

(6)

where C_e is the equilibrium concentration of silica (mg/L) or hardness (mg/L as CaCO₃) in the solution, q_e is the amount of adsorbed silica (mg SiO₂/g of resin) or hardness (mg CaCO₃/g of resin) at equilibrium, q_max is the maximum amount of monolayer silica (mg SiO₂/g of resin) or hardness (mg CaCO₃/g of resin) adsorption, K_L is the Langmuir constant, related to the affinity of the adsorbate to the binding sites. Both q_max and K_L are calculated from the intercept and slope values, respectively.

2.3. Activated Alumina

AA Experimental Conditions

In this study, three types of AA: CPN, DD-6, and DD-2 (BASF, Ludwigshafen, Germany) were evaluated at bench scale for silica removal from the ROC. CPN was denoted as AA-1, DD-6 was denoted as AA-2, and DD-2 was denoted as AA-3. AA-1 is a granular adsorbent suitable for a wide range of adsorption applications for impurity removal. AA-2 is known as a peroxide grade alumina and has an exceptionally high surface area of alumina developed specifically for service in hydrogen peroxide production. AA-3 is a calcinated and rehydrated alumina powder that has been proven for many catalyst applications. The chemical composition of AA is shown in

Table 5. Typical chemical composition of activated alumina for silica removal *.

Typical Chemical Composition, %	AA-1 (CPN)	AA-2 (DD-6)	AA-3 (DD-2)
Al2O3	92	92	-
SiO2	0.02	0.03	<0.02
Fe2O3	0.03	0	<0.01
Na2O	0.3	0.35	<0.4
Typical physical properties
Particle size, μm	1180 × 600	600 × 300	1180 × 600
Surface area, m2/g	315	380	275
Packed bulk density, lb/ft3 (kg/m3)	47 (752)	40 (641)	-

Note: * Adapted from the product data sheet of each AA.

.

For the batch adsorption experiments, different concentrations of each AA type (2.5, 5, 25, 50, and 100 g/L) were added to 250 mL polyethylene bottles with 200 mL of ROC. The pH was not adjusted, and the test was conducted at room temperature (~25 °C). Water samples were taken from each bottle after 24 h for water quality analysis. Experimental results for the performance of the different AAs tested were adjusted to the linearized equation of the Langmuir model (Equation (6)).

Based on the results of the batch experiments, AA-1 and AA-2 were selected to evaluate silica removal in column experiments. The column preparation and the service cycle operation of the columns were conducted as described in Table 4. Regeneration or elution experiments were divided into two studies: regeneration in batch and regeneration in column. For regeneration in batch, 1 g of AA previously used for adsorption experiments was added to 50 mL of regenerant and placed in a shaker. Various regenerants were tested, including NaOH [36,41,49,50,51], H₂SO₄ [7,49,50], and HCl [52,53], under different conditions of concentration, temperature, and contact time.

Additionally, AA-2 was used to test the adsorption capacity with synthetic high salinity water. These experiments were performed in batch and the synthetic high salinity water was prepared by dissolving lab grade Na₂SiO₃ (Sigma-Aldrich, St. Louis, MO, USA) in DI water to mimic the silica concentration in ROC of 160 mg/L. Then, lab grade NaCl (Sigma-Aldrich, St. Louis, MO, USA) was added to achieve brine solutions with concentrations of 5, 10, and 20% (w/v). Different concentrations of AA-2 (2.5, 5, 25, 50, and 100 g/L) were added to 250 mL polyethylene bottles with 200 mL of the 5, 10, and 20% NaCl solutions. The pH (~9.5) in the synthetic high salinity water was not adjusted, and the test was conducted at room temperature (~25 °C). Water samples were taken from each bottle after 24 h to evaluate the silica concentration. Results were fitted to the linearized equation of the Langmuir model (Equation (6)).

2.4. Modeling Methodology

The effluents from both column operation processes using AA to remove silica (using AA-2) and IX resins to remove hardness (using HRIX-2) were characterized to obtain the major cations and anions, along with hardness, alkalinity, and silica concentrations. These data were grouped into two pretreatment trains: ROC treated only by AA (denoted as AA effluent) and ROC treated by AA and IX (denoted as AA + IX effluent). Both effluent concentrations were compared with the initial ROC concentrations (Table 2) to develop an empirical model to study the influence of ROC pretreatment on overall water recovery. Initially, the water stability was analyzed using OLI Studio 12.0, as shown in Table S1 (Supplementary Materials, SM), following the methodology proposed by Lencka et al. [54]. Experimentally obtained alkalinity concentration was amended according to the pH of the solution, and the hardness of the solution was subsequently calculated according to the Ca²⁺ and Mg²⁺ concentrations. The charge imbalance was neutralized by increasing the concentrations of either Na⁺ or Cl⁻, depending on whether cations or anions were dominant when calculating the change of the solution. These ions were selected due to their abundance in the solution and the negligible impact on the pretreatment process. The adjustments made to the Na⁺ or Cl⁻ concentrations during the charge balancing process resulted in changes of less than 40 mg/L from their original concentrations. After establishing a stable water profile, the effect of the pretreatment on water recovery was investigated by simulating a high-recovery brackish water RO system at 30 °C and 100 gallons per minute (GPM, or 378.5 L per minute) using AdvisorCi Version 10.11 [55]. The present study investigated the compounds that formed scales under three different recovery levels using two antiscalants, namely Vitec^TM 7400 and Vitec^TM 1070 (Avista Technologies, Inc., San Marcos, CA, USA). In addition to experimentally obtained water quality data, 11 different synthetically modeled water types prepared using OLI Studio 12.0 (SM, Table S2) were tested under the same conditions using Avista AdvisorCi to evaluate the water recovery.

The maximum water recovery at various silica (SiO₂) and hardness concentrations was modeled using a second-degree polynomial regression (Equation (7)). The water recovery values obtained from Avista AdvisorCi and the predictions from the regression model showed good agreement, with a mean squared error of 0.463 and R² of 0.97, indicating close alignment between the predicted and actual values (SM, Figure S3). This model was developed utilizing the SciPy libraries in Python Version 3.11.10 [56]. A regression model was also employed to estimate treatment costs, providing further insights into the effects of pretreatment.

$$y=-0.00703x_1-0.03677x_2+1.002\times {10}^{-6}{x}_1^2+8.001\times {10}^{-6}{x}_1{x}_2-3.489\times {10}^{-4}{x}_2^2+86.91$$

(7)

where y is the water recovery (in %) using high-recovery brackish water RO, x₁ is the SiO₂ concentration (in mg/L), and x₂ is the hardness of the RO feed solution (in mg/L as CaCO₃).

The techno-economic analysis was conducted utilizing the data available in the WaterTAP (SM, Table S3) [57]. The LCOW was calculated by considering the operational, capital, and regeneration costs associated with IX and AA treatment processes. The calculation framework follows a systematic approach, beginning with the estimation of the total water volume processed per day derived from the flow rate. Various treatment fractions for AA and IX are considered to evaluate different operational scenarios. Water quality parameters, such as silica and hardness concentrations, are computed based on the proportion of water treated by each process. The total cost is calculated by integrating multiple cost elements, including operational costs for IX and AA, regeneration costs associated with the IX system, capital expenditures for treatment infrastructure, and RO processing costs. Additionally, disposal costs are factored in based on the fraction of water that is not recovered. The total cost is expressed as a function of these factors using the following equation:

$$\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{C}\mathrm{o}\mathrm{s}\mathrm{t}=\left(\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{I}\mathrm{X}\times \mathrm{f}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{I}\mathrm{X}\right)+\left(\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{A}\mathrm{A}\times \mathrm{f}\mathrm{r}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{A}\mathrm{A}\right)+\mathrm{R}\mathrm{O}\mathrm{C} \mathrm{d}\mathrm{i}\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{a}\mathrm{l} \mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t}\times {\displaystyle \frac{\left(100-\mathrm{W}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r} \mathrm{R}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{y}\right)}{100}}-\mathrm{R}\mathrm{O} \mathrm{C}\mathrm{o}\mathrm{s}\mathrm{t}$$

(8)

The net cost is obtained by taking the difference between the total cost and the revenue generated via RO permeate for potable water use.

3. Results

3.1. Silica Removal by Ion Exchange Resins

3.1.1. Isotherm Experiments: Silica Removal Using IX

Experimental results for the performance of SRIX-1 and SRIX-2 were fitted to the linearized equation of the Langmuir model. Both q_max and K_L were calculated from the intercept and slope values (SM, Figure S1a), respectively (

Table 6. Langmuir parameters for IX and AA isotherm experiments for silica removal.

	IX Resin Using ROC		AA Using ROC			AA-2 Using High-Salinity Synthetic Water
	SRIX-1	SRIX-2	AA-1	AA-2	AA-3	5% NaCl	10% NaCl	20% NaCl
qmax (mg/g)	12.44	19.65	625.00	322.58	238.10	312.5	357.14	294.12
KL	0.0231	0.0101	0.0012	0.0025	0.0016	0.0004	0.0004	0.003

). SRIX-2 demonstrated a higher maximum adsorption capacity (q_max of 19.65 mg/g) compared with SRIX-1, which reflected its ability to capture silica from the ROC. The Langmuir constant K_L of 0.0101 further indicated a relatively strong affinity between SRIX-2 and silica in the ROC.

3.1.2. Continuous Column Testing Experiments: Silica Removal Using IX

Based on the batch experimental results, SRIX-2 was selected for continuous-flow column testing to evaluate the impact of empty bed contact time and regeneration effectiveness.

The results for three service cycles (#1, #2, and #3) are shown in Figure 3, where CO and R stand for column operation and regeneration, respectively. Silica breakthrough was measured after three bed volumes (BVs) for CO #1 and #2, with full resin saturation after five BVs. However, CO#3 showed a silica breakthrough after 1 BV.

The ionic form of SRIX-2 was initially changed from Cl⁻ to OH⁻ using a 1 M NaOH solution, therefore the pH of ROC that was previously adjusted to 4.5 increased to ~10–13 during the first two BVs of the CO. However, the pH quickly decreased to ~7 at BV 5; this could suggest that the change of the resin’s ionic form to OH⁻ was not fully achieved. Figure 3a shows that silica removal declined as the pH decreased. Silicic acid (H₃SiO₄⁻) has a pKa of 9.84; therefore, at pH < 9, the dominant silica specie in water is orthosilicic acid (Si(OH)₄), which is a neutral compound that does not participate in the ion exchange process [39,53]. In the past, silica removal using SBA resins has also been reported to be affected by the competition of other ions in water, like SO₄²⁻ and Cl⁻ [41,58], which have high concentrations in ROC (Table 2).

Additionally, the elution was drastically reduced after R #1 (practically no silica elution during R #2 and #3), which suggests that SRIX-2 loses its removal capacity after one operation cycle. Low regeneration efficiency in anion exchange resins is attributed to the polymerization of H₃SiO₄⁻ when it binds to the OH⁻ ion in the resin bead; this polymerization transforms H₃SiO₄⁻ into dimeric and polymeric species, which are more challenging to desorb during regeneration compared with monomeric species [40].

In summary, at a pH lower than the orthosilicic acid (H₄SiO₄) pKa of 9.84, silica was present in the non-ionized form, which caused low silica selectivity for IX in ROC. An alternative to increase the performance of SBA resins could be to adjust the pH > 9.8, which will increase operational costs.

3.2. Silica Removal by Activated Alumina

3.2.1. Isotherm Experiments: Silica Removal Using AA

The experimental results for the performance of the three different AAs were fitted to the linearized equation of the Langmuir model (Equation (6)) (SM, Figure S2). Both q_max and K_L were calculated from the intercept and slope values, respectively (Table 6).

Among the activated alumina tested, AA-1 showed the highest adsorption capacity (q_max = 625 mg/g), indicating its strong potential for removing large amounts of silica from ROC. This makes AA-1A ideal for high-salinity environments where silica concentrations are particularly high. AA-2 with a moderate capacity (q_max = 322.58 mg/g) also demonstrated the highest affinity (K_L = 0.0025) for silica. This suggests that AA-2 could be more effective in systems with shorter contact times, where quick adsorption is critical. AA-3 had the lowest capacity (q_max = 238.1 mg/g) and affinity, making it less effective than AA-1 and AA-2. Both AA-1 and AA-2 were selected for column experiments following the methodology described in Table 4.

The experiments using synthetic high-salinity water achieved a q_max of 312.5, 357.1, and 294.1 using the 5%, 10%, and 20% NaCl solution, respectively (Table 6). The results at high salinity did not show a significant effect on the adsorption capacity of AA-2 for the ROC with a TDS of ~12 g/L (~1.2%). These findings suggest that, in addition to treating brackish water ROC, the AA could be used as a silica adsorbent in high-salinity environments, e.g., seawater ROC or brine concentrator effluents.

Additional adsorption experiments were conducted to saturate AA-1 and AA-2 and test the effect of different regeneration conditions. The results for the regeneration/elution batch experiments are shown in

Table 7. Results summary of batch regeneration/elution experiments.

AA	Silica Adsorbed	Regenerant	Regenerant Concentration	# of Batches	Temperature	Contact Time	Regeneration/ Elution
AA-1	23.5 mg/g	NaOH	2, 8 and 10%	3	25 °C	1 h	0
		H2SO4	0.1 N	3	25 °C	1 h	10.1%
		NaOH/H2SO4	NaOH 2% H2SO4 0.1 N	1	25 °C	1 h each	4.5%
AA-2	42 mg/g	NaOH	2%	1	25 °C	1, 12, and 24 h	0
					50 °C	1 h	0
			6%	1	25 °C	1, 12, and 24 h	0
					50 °C	1 h	0
			8%	1	25 °C	1, 12, and 24 h	0
					50 °C	1 h	0
		H2SO4	0.1 N	3	25 °C	1 h	10.4%
		NaOH/H2SO4	NaOH 2% H2SO4 0.1 N	1	25 °C	1 h each	5.3%
	59 mg/g	HCl	0.05 N	3	25 °C	1 h	2.9%
			0.1 N	3	25 °C	1 h	10.7%
			1 N	3	25 °C	1 h	19.8%
	33.3 mg/g	HCl	0.1 N	3	25 °C	2 h	9.6%
						3 h	12.8%
						12 h	22.8%
					40 °C	1 h	18.6%
					50 °C	1 h	22.9%
					60 °C	1 h	26.6%

.

The highest adsorption capacity of 59 mg SiO₂/g AA was obtained with AA-2 (7 batches of 1 h, at ambient temperature, adsorbent dose of 10 g/L, and magnetic stirrer of 350 RPM). This value was similar to 50 mg/g reported by Sanciolo et al. [45] after five batches. Bouguerra et al. [44] evaluated the effect of pH, adsorbent dose, temperature, and another ions competition using synthetic water (silica concentration of 50 mg/L). According to their results, the maximum silica removal of 90% was achieved at 2.5 g of AA. In comparison, Sanciolo et al. [45] conducted batch experiments using ROC (silica concentration of 160 mg/L), achieving 71.9% of silica removal when AA was added at a dose of 10 g/L for 60 min at a temperature of 20 °C.

The regeneration of AA-1 was evaluated in batches of 1 h using NaOH (2%, 8%, and 10%), H₂SO₄ (0.1 N), and a combined batch between 2% NaOH and 0.1 N H₂SO₄. The highest regeneration percentage (10.1%) was achieved after three batches of H₂SO₄, followed by the combined batch between 2% NaOH and 0.1 N H₂SO₄, with a regeneration efficiency of 4.5%. Several studies [39,45,59,60,61] have reported partial (70–90%) or full regeneration using NaOH; however, in this study, NaOH did not produce silica elution.

Additionally, regeneration for AA-2 was tested using NaOH (2%, 6%, and 8%), H₂SO₄ (0.1 N), a combined batch between 2% NaOH and 0.1 N H₂SO₄ [6] (with a DI water rinse between the base and acid), and HCl (0.05, 0.1 and 1 N). In this case, NaOH was tested under different temperatures (25 and 50 °C) and contact time (1, 12, and 24 h) to evaluate the effect on silica elution. However, no regeneration was achieved. The 0.1 N H₂SO₄ experiment and the combined batch with 2% NaOH and 0.1 N H₂SO₄ showed a regeneration of 10.4% and 5.3%, respectively. Different concentrations of HCl were tested, where 1 N HCl obtained the highest percentage removal of 19.8%. Some other conditions were evaluated for regeneration using HCl, such as different contact time and temperature, where the highest regeneration percentages were 22.8% after three 12 h batches and 26.6% after three 1 h batches at 60 °C. Nevertheless, such conditions are difficult to provide in a continuous column operation, therefore 0.1 N HCl (which achieved a slightly higher regeneration percentage of 10.7% compared with 0.1 N H₂SO₄) and 1 N HCl were selected as regenerants for the column operation. Although some studies in the literature have reported that AA was successfully regenerated using NaOH and H₂SO₄ [6,39,41], our experimental procedure did not achieve regeneration during the treatment of brackish water ROC. As mentioned before, the polymerization of H₃SiO₄⁻ could hinder the complete elution due to longer molecules that are more challenging to desorb [40]. Further research is needed to elucidate the mechanisms of silica removal using AA.

3.2.2. Continuous Column Testing Experiments: Silica Removal Using AA

For this experiment, results for the column operation using AA-1 and AA-2 are shown in Figure 4. Figure 4a shows silica breakthrough starting after 37 BV and an average silica concentration in effluent remained at 58 mg/L (average removal of 65%) between 200 and 1200 BV using AA-1. Instead, silica breakthrough using AA-2 started at 45 BV, with an average silica concentration in effluent of 60 mg/L and SiO₂ removal of 45.3%. Due to the long duration of the column experiments, both columns were stopped at 870 BV. The final silica concentration in the effluent was 88 mg/L (removal 45% and adsorption capacity 217.5 mg/g) for AA-1, and 108 mg/L (removal 32% and adsorption capacity 148.2 mg/g) for AA-2. However, according to the q_max values, the saturation point for AA-1 and AA-2 was estimated at 2500 BV and 1894 BV. Sasan et al. [50] conducted a column experiment (using calcinated hydrotalcite and powder AA as media), which was operated over 100 h with an average silica adsorption capacity of 34 mg/g. To the best of our knowledge, there are no other reported studies using AA in continuous column experiments. After the column operation was concluded, regeneration using 1 N and 0.1 N HCl solutions was applied to AA-2.

Before testing using the 1 N HCl solution for regeneration, 0.1 N HCl was tested for 11 BV; however, the regeneration percentage achieved was negligible (<1% elution). The results of column regeneration using 1 N HCl for AA-2 are shown in Figure 5.

After 80 BV, the regeneration percentage achieved was 16%. Compared with the results from silica removal using IX resins (Figure 3), both column operations of AA-1 and AA-2 were significantly longer, which can be attributed to higher capacity and major affinity for neutral silica species [39,45]. Nevertheless, regeneration was not easily achieved. For example, considering that regeneration follows a linear behavior, it would take 496 L of 1 N HCl to accomplish a full regeneration for 1 L of resin under the same operational conditions. These results show that, while AA effectively removes silica from ROC, its regeneration remains a challenge for treating ROC from KBHDP. At pH 8.5 and saturation level, silica in KBHDP ROC can exist as polymerized colloidal silica. These long chains of individual silica molecules exhibit virtually no charged ionic character and cannot be easily removed during regeneration. In addition, colloidal silica can be associated with Ca and Mg in ROC to form complex inorganic compounds, which cause lower adsorption selectivity and difficulty in regeneration. The further refinement of regeneration techniques or alternative approaches is needed to enhance the efficiency and viability of AA in industrial settings.

3.3. Hardness Removal by IX Resins

3.3.1. Isotherm Experiments: Hardness Removal Using IX

The experimental results for hardness removal by HRIX-1, HRIX-2, and HRIX-3 were fitted to the linearized equation of the Langmuir model. Both q_max and K_L were calculated from the intercept and slope values (SM, Figure S1b), respectively (

Table 8. Langmuir parameters for IX isotherm experiments for hardness removal.

	IX Resin
	HRIX-1	HRIX-2	HRIX-3
qmax (mg/g)	133.33	129.87	101.01
qmax (eq/L)	3.23	3.12	1.56
KL	0.0019	0.0012	0.0013

).

HRIX-1 showed the highest q_max of 133.33 mg/g and K_L of 0.0019 compared with HRIX-2 and HRIX-3, which reflects a stronger interaction between the IX resin surface and the hardness concentration in ROC. The q_max (eq/L) values using HRIX-1, HRIX-2, and HRIX-3 correspond to 68.1%, 68.4%, and 37.1%, respectively, of the total exchange capacity reported by the product data sheet of each resin (Table 3). Both HRIX-2 and HRIX-3 were preconditioned to change their ionic form from H⁺ to Na⁺. HRIX-3 is commercially available in the ionic form of Na⁺, of which total exchange capacity is reported as 2 eq/L. To our knowledge, HRIX-1 is not available in Na⁺ ionic form. HRIX-1 and HRIX-2 were selected for column experiments following the methodology described in Table 4.

3.3.2. Continuous Column Testing Experiments: Hardness Removal Using IX

Figure 6a,b show the column operation results of HRIX-1 and HRIX-2 to compare the hardness breakthrough in terms of BV. Hardness was completely removed from ROC until hardness breakthrough started at 19 BV for HRIX-1 and 11 BV for HRIX-2; the IX resin was fully saturated at 33 BV for HRIX-1 and 17 BV for HRIX-2. The adsorption capacity was 98.25 g CaCO₃/L resin and 65.32 g CaCO₃/L resin for HRIX-1 and HRIX-2, respectively.

In terms of pH stability, HRIX-2 maintained a slightly basic pH without high variability. In contrast, HRIX-1 showed slightly acidic pH values, which may be attributed to the incomplete preconditioning of its ionic form from H⁺ to Na⁺. Regarding regeneration, columns using HRIX-1 and HRIX-2 were fully regenerated after 20 and 17 BV, respectively, as shown in Figure 6c,d. Consequently, the chemical requirements are 20 and 17 L of 8% NaCl/L of resin and the regeneration BV/saturation BV ratios are 0.6 and 1. These ratio values, particularly for HRIX-2, are far above levels considered practical in industrial applications, underscoring the challenges of using SAC IX for brine treatment.

HRIX-2 was selected for repeat column experiments because it does not require preconditioning, and has a broader pH operation range (Table 3) and a shorter regeneration curve (equal to lower chemical demand) compared with HRIX-1.

The results in Figure 7 show the averages for three service cycles using HRIX-2. Figure 7a shows the hardness breakthrough started after 10 BV, and the resin was fully saturated after 17 BV, which is consistent with Figure 6b. The total capacity of the resin was 65.63 g CaCO₃/L resin when it was completely saturated. The regeneration curve (Figure 7b) showed more variability between cycles, with a complete regeneration after 15 BV.

Thomson et al. [49] conducted column experiments to determine if the TDS concentration affects the column operation performance of a SAC. Three different TDS concentrations were tested, including low, medium, and high corresponding to 5900, 28,800, and 52,600 mg/L, respectively. The BV to saturation reduced drastically from 65 for the low TDS water to 6 for the high TDS brine. It was concluded that the divalent selectivity of the IX resin significantly decreased when the TDS in the feed was higher than 10,000 mg/L, compromising its performance. In our study, the TDS concentration of the ROC was 12,445 mg/L and the column saturation was quickly achieved (with an average 25 BV) using two different SAC resins (HRIX-1 and HRIX-2). This behavior in SAC resins is explained by the molecules in water being more “organized” when the TDS concentration increases, which limits the interaction of polyvalent ions (such as Ca²⁺ and Mg²⁺) with the resins, reducing its effectiveness [62]. On the other hand, Janson et al. [47] tested three different WAC resins using a synthetic feed with a TDS of 8093 mg/L. The saturation in the columns was achieved at 1922, 2577, and 3232 BV (average value of 2577 BV); such values were significantly higher compared with the saturation point using SAC. However, even when the performance of WAC resins is less affected by high TDS in water, the hardness removal could be incomplete since WAC resins cannot remove noncarbonate hardness, contrary to the SAC resins that remove carbonate and noncarbonate hardness equally [47,62].

The column operation of SAC and WAC and the water chemistry (specifically TDS and hardness concentrations) reported in the literature and obtained in this study (Table 1) were used to plot Figure 8. These data were used to compute the Pearson correlation coefficient between the TDS and the hardness concentration in the feed and the number of BV to saturation, finding moderate negative correlations $r\left(5\right)$ of −0.401 and −0.589, respectively; which implies that higher TDS and hardness concentrations lead to faster resin saturation.

3.4. Modeling Results and Preliminary Cost Analysis

The water stability analysis of the ROC and AA effluent demonstrated that, despite a slight decrease, the water maintained a density of 1.006 g/mL. However, the ionic strength of the solution increased by 0.004 mol/kg (m-based) because of the removal of 150 mg/L of SiO₂. The density and ionic strength of the IX feed and IX effluent solutions did not differ significantly (less than 1%), which can be attributed to replacing Ca²⁺ and Mg²⁺ ions with Na⁺ ions during the IX process. The RO simulation conducted using Avista AdvisorCi-illustrated silica in ROC, presented as Ca₃Si₂O₆(OH)₂.2H₂O and Mg₃Si₂O₅(OH)₄, exhibited a higher scaling tendency, leading to a maximum water recovery of 64% and 4%, using Vitec 7400 and 1070, respectively (

Table 9. Maximum RO water recovery with different silica and hardness concentrations using antiscalant Vitec 7400 and Vitec 1070 simulated with Avista AdvisorCi.

Maximum Water Recovery (%) with Vitec 7400
		Silica Concentration (mg/L)
		10	37.5	75	112.5	160
Hardness (mg/L as CaCO3)	980	80 (AA + IX effluent)	−	−	−	−
	1417	79	78	−	−	−
	1882	77	76	75	−	−
	2336	76	74	72	68	−
	2800	75 (AA effluent)	74	72	70	64 (ROC)
Maximum Water Recovery (%) with Vitec 1070
		Silica Concentration (mg/L)
		10	37.5	75	112.5	160
Hardness (mg/L as CaCO3)	980	80 (AA + IX effluent)	−	−	−	−
	1417	78	72	−	−	−
	1882	77	68	52	−	−
	2336	76	65	53	28	−
	2800	75 (AA effluent)	64	53	30	4 (ROC)

).

The ROC water recovery increased from 64% to 75% and from 4% to 75%, utilizing Vitec 7400 and 1070, respectively, with the removal of 94% of silica, demonstrating that Vitec 7400 was more effective at preventing silica scaling than Vitec 1070 during ROC treatment (Table 9). Despite this improvement, water recovery was still limited by gypsum scaling (CaSO₄). An additional 5% increase in ROC water recovery, reaching a total of 80%, can be achieved by removing 65% of hardness (as CaCO₃), which leads to the scaling of silica species (Table 9).

The maximum water recovery using Vitec 7400 was used to develop the empirical model (Equation (7)). Analysis of water recovery at different silica and hardness concentrations (Figure 9) suggests that reducing hardness below 1000 mg/L (as CaCO₃) can further enhance ROC recovery. Additionally, a silica concentration below 50 mg/L was identified as the optimal condition for maximum water recovery. A notable decrease in water recovery was observed when hardness exceeded 2500 mg/L (as CaCO₃) and silica exceeded 150 mg/L, indicating that ROC recovery efficiency is significantly (~9%) compromised beyond these thresholds.

Using the available empirical water recovery model, a water treatment cost estimation (SM, Table S3) was performed by blending ROC, AA effluent, and IX effluent waters at different compositions (Figure 10a). The ROC feed water entered the system at a flow rate of 100 GPM (378.5 L per minute), resulting in a treatment capacity of 0.144 MGD. For this analysis, the silica concentration in the AA effluent was set at 10 mg/L, and the hardness of the IX effluent was assumed to be 980 mg/L as CaCO₃.

As expected, ROC water recovery gradually increased as the proportion of water treated by AA and IX increased (Figure 9). Concurrently, when examining treatment costs at varying levels of pretreatment (SM, Figure S4a), a consistent relationship emerged between silica/hardness concentrations and total ROC treatment cost. As the water recovery increased (Figure 9), the pretreatment cost increased linearly (Figure 10b). Hence, the lowest net costs were achieved when both AA and IX were maximized (Figure 10b), such that, when the water recovery increased by 6% via treating the ROC with IX and AA, the estimated cost increased by USD 0.63/kgal and USD 0.21/kgal (USD 0.166 to 0.055/m³), respectively (Figure 10c). However, the revenue gained through additional RO permeate (ROP) production—along with reduced ROC disposal costs and lower antiscalant dosing—offset the expense of pretreatment (Figure 10b). As a result, treating ROC with IX and AA will reduce the levelized cost of ROP production by USD 0.73/kgal (USD 0.192/m³) after compensating for the pretreatment cost (Figure 10b). Overall, the pretreatment accounts for ~25% of total treatment costs, which remains cost-effective due to improved water recovery and the corresponding additional revenue from enhanced permeate production. Additionally, the IX demonstrates significant cost competitiveness, with an LCOW of USD 0.17/m³, making it more economical compared with chemical precipitation at USD 0.94–1.38/m³ and coagulation at USD 0.05–1.20/m³ when treating ROC [36].

4. Conclusions

This work compared the efficacy of IX resins and AA in mitigating the scaling potential of ROC, a crucial step towards enhancing water recovery in desalination processes. Column experiments showed effective silica removal using AA, achieving up to 65% reduction and maintaining this efficiency over 800 BV without reaching saturation. On the other hand, the IX resins performance was initially effective, but exhibited rapid saturation at 17–33 BV using SAC for hardness removal and 5 BV using SBA for silica removal, underlining their limited capacity and frequent regeneration needs due to high concentrations of scaling ions in ROC. The model simulations projected that integrating IX and AA pretreatment could potentially enhance overall ROC water recovery by approximately 70%. These results not only underscore the potential for significant improvements in RO system efficiencies but also highlight the need for continued advancements in material science and process strategies. Such advancements are essential to overcome the inherent limitations of existing IX technologies and to develop more robust, efficient, and cost-effective solutions for ROC treatment.

In conclusion, AA could provide promising pretreatment for ROC and other hypersaline brine treatment but, due to its low regeneration efficiency, it is necessary to explore further alternative applications for the used AA, such as desiccant for moisture removal in the petroleum industry for air and gas purification. In the case of SBA and SAC IX resins, their scalability for large-scale implementation needs to overcome challenges such as high regeneration demands due to the presence of high concentrations of hardness and silica, along with TDS. Future research should explore the development of alternative materials and methodologies that address the specific limitations identified in this study. This includes investigating the performance of WAC IX resins with enhanced regeneration capabilities, and exploring hybrid unit processes (e.g., chemical precipitation + IX) to improve scaling ion removal efficiency. Additionally, integrating pretreatment approaches with resource recovery techniques (such as the recovery of valuable minerals from ROC) could provide the dual benefits of scaling mitigation and resource valorization, contributing to more sustainable and economically viable water treatment solutions.