Intensification and energy minimization of seawater reverse osmosis desalination through high-pH operation: Temperature dependency and second pass implications

https://doi.org/10.1016/j.cep.2018.07.009Get rights and content

Highlights

  • A new high-pH single-pass process was assessed at full seawater temperature range.

  • 1st pass product (54% recovery, 15 < T < 31 °C, pH 9.5) quality: TDS < 450 and B < 0.4 mg/l.

  • 2nd pass permeate (at 90% recovery) contained TDS < 30 mg/l and B < 0.4 mg/l.

  • The one-pass process cost was ∼$0.03/m3 lower than conventional process, at T<28 °C.

  • The two-pass process cost is competitive, especially at high seawater temperatures.

Abstract

Operation of seawater reverse-osmosis (RO) desalination using high flux membranes at pH > 9 was shown to be an energy-efficient approach for single-pass boron removal. This operational approach was previously studied only at 25 °C, however since RO desalination is highly temperature-dependent, the current work tested the high-pH approach throughout the typical temperature range of the Mediterranean Sea. Since total dissolved solids (TDS) removal in a high-flux-membranes single-pass is limited, the effect of applying a 2nd RO pass was also examined. Finally, the process cost was assessed at varying operational conditions. Results showed 1st pass permeate TDS to be ∼440, 290 and 200 mg/l at 54% recovery-ratio and 31, 25 and 15 °C, respectively. Boron removal was adequate (i.e. <0.4 mgB/l in permeate) at pH 9.5 throughout the temperature range. However, at 31 °C antiscalant had to be added to prevent Mg(OH)2 scaling. TDS and boron concentrations in the 2nd-pass permeate met the threshold limit (30 mg/l and 0.5 mgB/l) at 90% recovery ratio. The cost of applying the single-pass alternative process was lower by about $0.03/m3 permeate than the conventional alternative, for temperatures lower than 28 °C. The addition of a second pass increased the operational cost merely by a ∼$0.03/m3, making the described two-pass process competitive.

Introduction

In the last two decades, reverse osmosis (RO) processes have gained popularity over other desalination techniques [1], particularly due to their decreased energy demand, enabled by more water permeable membranes and efficient energy recovery devices, and the fact that membrane costs are continuously dropping due to ever-increasing competition. Nevertheless, several issues related to RO, including energy requirement, membrane fouling, product quality and recovery ratio (RR), can be significantly improved. While total dissolved solids (TDS) removal is already a well-controlled practice, the limitation imposed on the boron concentration in desalinated water used for irrigation presents a major technological challenge. Although boron is vital as a trace element for plant growth and is even supplied in some fertilizers, it can be detrimental to certain crops (e.g. avocado, kiwi and most citrus types) at concentrations higher than 0.5 mgB/l [2,3]. Boron is present in seawater at a concentration of approximately 5 mgB/l, as a weak acid (Eq. (1) is a simplification since borate ion pairs are not shown) with an apparent pKa value of 8.9 (according to the pH scale defined in [4]). Consequently, at normal seawater pH (8.1–8.2), boron is mainly present as boric acid, B(OH)3, which, due to its small size and lack of charge is poorly rejected (65–80%) by commercial seawater reverse osmosis (SWRO) membranes and even less so by BWRO membranes and high-flux SWRO membranes [5].B(OH)3 + H2O ↔ B(OH)4 + H+

As a result, the most important factor affecting boric acid rejection is the pH value of the feed solution. At neutral pH, the passage of the species B(OH)3 through the membrane is approximately 30 times higher than the passage of ionic species such as sodium, chloride or borate [6]. Two main methods are currently applied at full scale to meet boron removal requirements: (a) Application of 2nd RO pass at high pH: in this method the permeate of the 1st RO pass is introduced as feed to the 2nd pass subsequent to pH elevation by strong base (typically NaOH) dosage. This is the most commonly applied method; (b) Ion exchange (IX): a borate specific anion exchange resin is used for separating boron (as B(OH)4) from the 1st pass permeate, requiring strong acid and strong base for regeneration [7].

The specific energy consumption (SEC) in SWRO can be calculated using Eq. (2):SEC=PinRRηpump-Pout1-RRRRηERDWhere, Pin and Pout are the feed and the retentate pressures, respectively (bar), RR is the recovery ratio (m3product ∙ m−3feed), ηpump and ηERD are the pump and energy recovery device efficiencies, respectively.

The SEC associated with current seawater RO desalination plants is ∼3.5 kW h/m3, out of which the 1st pass step consumes ∼2.5 kW h/m3, which is more than twice the minimum theoretical thermodynamic energy requirement for 50% recovery seawater desalination [8]. The main reason for this difference is the fact that in order to minimize the required membranes’ surface area the applied external pressure is much higher than the osmotic pressure. The difference between the applied external pressure and the osmotic pressure is particularly high at the inlet to the membrane train where the concentration is relatively low, but in order to maintain reasonable flux it is maintained higher throughout the whole process. Considerable effort has been invested in recent years in reducing the SEC value. For example, reducing the feed pressure by application of high-flux thin film nanocomposite (TFN) membrane, resulting in ∼10% saving was suggested [9]. Alternatively, applying different types of membranes in the pressure vessels of both the 1st and 2nd RO passes was suggested, this would result, according to the writers, in ∼2.5–3% saving in SEC [10]. Prante et al. [11] suggested applying pressure retarded osmosis (PRO) in conjunction with RO (a process termed RO-PRO desalination), presumably resulting in ∼40% saving in SEC. Kishizawa et al. [12] reported on development of a new RO system design in which short vessels, containing 2–4 elements, are used in the 1st stage, and relatively long vessels are used in the 2nd RO stage. This method was suggested to result in ∼20% SEC saving. The closed circuit reverse osmosis (CCRO) method was estimated to reduce BWRO SEC by up to 30% [13]. Nanofiltration as a pretreatment and also double-pass nanofiltration was suggested as energy efficient alternatives to the conventional SWRO process [14]. Finally, an alternative process for boron removal in the 1st pass, which allows using high flux seawater RO membranes, resulting, according to the calculations of our research group, in a decrease of about 10% in the energy demand [8]. The latter method, along with its advantages and limitations, is described in detail below.

In order to improve the low boron rejection in the 1st RO pass, an alternative process for boron removal, using a single RO pass, has been developed in our research group [3,8,15]. The main process steps are schematically depicted in Fig. 1. The method is based on seawater de-carbonation followed by pH elevation and salts separation using high flux membranes. The idea is to reduce the carbonate concentration of seawater to almost nil by acidification to pH∼4, followed by CO2 removal (either by stripping or by CO2(g) selective membranes) in a specific pretreatment step. This allows raising the pH thereafter using a relatively small mass of strong base, in order to attain high boron removal in the first RO step. The usage of high flux membranes at high pH conditions is possible in this fashion since CaCO3 scaling is not limiting (as most of the carbonate system concentration was stripped out as CO2(g) in the preceding step). As mentioned before, B removal is much more efficient at a pH in which most this weak-base system is presented as B(OH)4 or complexes thereof. Since CaCO3 precipitation potential is minimized by the de-carbonization step, the recovery ratio at pH > 9 becomes now limited by the potential precipitation of the solid Mg(OH)2(s), as described in Eq. (3).Mg(OH)2(s) ↔ Mg2+ + 2OH Ksp = 1.8∙10-11

Applying the described approach was shown to result in a product solution characterized by an average TDS of ∼375 mg/l and B of ∼0.3 mgB/l, at 56% recovery ratio and temperature of 25 °C ± 2 [8].

Ophek et al. [8] showed that the SEC of the above alternative process was reduced by 5%–10% relative to conventional SWRO operation at 25 °C. The described operational sequence resulted in satisfactory TDS and boron removal efficiencies at recovery ratios as high as 56% (at 25 °C).

The normalized cost (per 1 m3) of producing desalinated water using this approach was found to be ∼0.05 to ∼0.07 $ /m3 lower than the cost of currently applied SWRO processes [8]. However, the effect of the feed water temperature on the process performance was not reported, as well as the possible application of a cost effective 2nd RO pass for further decreasing the product TDS in case this is required in the desalination plant specifications. The current paper reports therefore on results related to these two subjects, i.e. the influence of feed temperature mainly on the boron and TDS permeability and the effect of the addition of a 2nd pass to the process, acknowledging that most (at certain temperatures, practically all) the boron concentration had already been removed in the 1st pass. For example, at 15 °C, the B concentration in the 1st pass permeate is ∼0.25 mgB/l at RR of 54%, i.e. in this case no B removal is required in the 2nd pass since the required criterion for boron in the product water is <0.3 mgB/l.

Temperature affects boron rejection through change in the thermodynamic state of the solution (i.e. change in pH and dissociation constants) and through change in permeability coefficients, namely for H2O and B(OH)3 [16]. Operation and design of RO desalination plants are planned to fit both the highest and lowest feed water temperatures. In the Mediterranean Sea the annual temperature range extremes are 14–31 °C. At the low temperature range, membrane permeability decreases thereby reducing the flux through the membrane, which results in an increase in SEC and/or reduced water production. At the same time, the rejection of the membrane to both salts (TDS) and boron, increases. Conversely, at the high end of the temperature range, membrane permeability and water flux increase, but more dissolved solids (including B) pass to the permeate side, and the combined effect deteriorates the product quality. Consequently, during summer time operative desalination plants encounter difficulties in meeting strict requirements (B < 0.4 mgB/l) for boron concentration in the permeate.

Desalinated water quality requirements vary according to geography and the intended water usage. As noted before, a 2nd RO pass is typically required to achieve final product quality that complies with stringent requirements, i.e. low chloride ion concentration (<15 mg/l) and low TDS (<25 mg/l) [17]. This is particularly true when high flux membranes, with slightly lower rejection values, are used in the 1st RO pass.

There are three main design alternatives for applying a two-pass RO process: (1) a full second pass; (2) a partial second pass and (3) split-partial second pass (SPSP). The decision on which option to practice depends on the water quality requirements. In a full second pass, all the 1st pass permeate flow is treated in the 2nd pass to obtain the required quality and quantity product. Typical recovery ratios are 50% for the 1st pass and 90% for the 2nd pass, i.e. the overall RR drops to 45% if second-pass retentate recirculation is not practiced. In the partial second pass alternative, a certain fraction of the first pass permeate is treated in the second pass. The permeate streams from the two RO steps are mixed to form the final product. Degree of mixing depends on the required quality and quantity. The SPSP design is based on the fact that the front elements in the RO pressure vessel always produce higher permeate quality than the elements at the rear of the pressure vessel (due to the larger difference between the external and osmotic pressures at the front). The SPSP design takes advantage of the superior permeate quality at the front of the membrane train by collecting the permeate streams from each side of the pressure train separately. The low-TDS front permeate is then pumped directly to the final product line, while the higher-TDS back permeate is treated by a 2nd RO step. At the end of the process, the permeate stream coming off the 2nd pass is mixed with the front permeate to produce the final product at the required quality. In the SPSP design the correct ratio between the front and back permeate is decided upon in order to obtain the final product of the required quality in terms of boron and TDS [6]. The SPSP design provides a cost effective option by minimizing the size of the second pass RO, which allows substantial savings on capital costs, as well as in the operating costs of the plant. The typical recovery ratio in the 2nd pass in this option is also 90% [18].

In conventional SWRO processes, since considerable removal of boron is planned in the 2nd pass, the 2nd pass retentate is characterized by high B concentration, making retentate recycling to the 1st pass feed line impractical. In order to minimize the reduction in the overall recovery ratio as a result of applying a second-pass, additional RO stage(s) is sometimes appended to the process. For instance, in the Ashkelon desalination plant (Israel), a 3rd and 4th stages were added to the process to elevate RR, at a considerable capital cost [6]. In the alternative process presented here, the boron concentration in the retentate of the 2nd pass is lower than the concentration in the feed stream, due to the fact that the main boron mass was removed in the 1st pass. As a result, the 2nd pass retentate stream can be circulated to the feed line, for maximizing the overall RR value.

Recovery ratio is a parameter that affects both the capital and operation and maintenance (O&M) costs associated with the process. Capital costs are affected by the recovery ratio through the impact on the normalized cost of the intake, pretreatment, membranes and pumps, while O&M costs are influenced by the recovery ratio through direct impact on the SEC, and to a lesser extent by the O&M costs of the pretreatment step and the cost of seawater intake and retentate discharge. Considering that the recovery ratio is strongly affected by the feed temperature, it is clear that the feed temperature also has a strong influence on costs.

Section snippets

Materials and methods

The work was divided in three: 1st-pass SWRO experiments, 2nd-pass RO experiments followed by theoretical calculations and cost assessment. All desalination experiments were carried out in a recirculated RO pilot system comprised of one 4”×40” spiral RO pressure vessel (Bell, ORL4-E-1000), a booster pump (Pedrolo, 2-4CR), a high pressure positive displacement pump (Hydra cell, D/G-10-X) and 25 μm filter. Heat exchanger (ORZ2, Oran) and chiller (CWA-36TPC) were used for continuous work at a

Results and discussion

This section comprises a theoretical part, followed by experimental results. The experimental work was divided in two. The first step assessed the feasibility and cost effectiveness of applying the high-pH 1st pass approach throughout the typical Mediterranean temperature range (15–31 °C). The second step was aimed at improving the process to produce drinking water suitable for locations where the restriction on TDS concentration is stringent (e.g. Israel) while striving to maintain the SEC and

Summary and conclusions

This work demonstrates and discusses the advantages of combining boron and TDS removal in a single SWRO pass at a wide temperature range, following up on a new high-pH de-carbonated feed-water operational approach [8]. Experimental results show a certain variability in the water quality and the required operational conditions at the different temperatures tested. At the tested 1st pass recovery ratio (54%) the single-pass water quality results invariably met the requirements in most

Acknowledgments

This work was partially funded by the Israel Science Foundation (Project #163/14) and partially by the Israeli Ministry of Energy (Project #217-11-031).

References (22)

Cited by (10)

  • Superstructure based optimization of reverse osmosis desalination systems fed by decarbonated high-pH seawater under boron restrictions

    2022, Computers and Chemical Engineering
    Citation Excerpt :

    When the recovery ratio is between 56% and 60%, the possible Mg(OH)2 precipitation problem could be avoided by antiscalant. Although experiment and simulation methods were used to investigate the operational conditions under different feed situations (Nir et al., 2012b; Nir and Lahav, 2013, 2016; Ophek et al., 2015; Segal et al., 2018), the optimization method has not been utilization to find the best system design with different types of seawater. This paper presents a superstructure based optimization of the SWRO system fed by high-pH, decarbonated seawater with boron restrictions.

  • Industrially-prepared carbon aerogel for excellent fluoride removal by membrane capacitive deionization from brackish groundwaters

    2022, Separation and Purification Technology
    Citation Excerpt :

    In recent years, people usually achieve the removal of fluoride ions in fluorine-containing wastewater through adsorption[9,10], ion exchange[11], coagulation [12], reverse osmosis(RO)[13]. RO has been developed on a commercial scale in many countries for many years[14,15], but studies have shown that the energy efficiency[16] and retention efficiency for fluorine of RO are not high when dealing with brackish water. Therefore, a new technology with commercial prospects is urgently needed.

  • Energy and environmental issues of seawater reverse osmosis desalination considering boron rejection: A comprehensive review and a case study of exergy analysis

    2021, Process Safety and Environmental Protection
    Citation Excerpt :

    Moreover, a Ma'agan Michael pilot SWRO plant in Israel using pressure vessel (Bel, a ORL4-E-1000) obtained permeate quality with TDS<30 mg L−1 and boron concentration<0.4 mg L−1. In this study, double of RO configuration with increased pH in the second pass (pH 9.5) was adopted (Segal et al., 2018). Nevertheless, this process increases the capital and operational costs and thus, reduces the economic efficiency.

  • Brackish water desalination using reverse osmosis and capacitive deionization at the water-energy nexus

    2020, Water Research
    Citation Excerpt :

    Due to rapid population growth and resource depletion, freshwater stress and scarcity are one of the most severe challenges around the world, especially in countries such as Saudi Arabia (Aljohani, 2017), Jordan (Qasim et al., 2018), and Tunisia (Walha et al., 2007). Seawater (a salinity ∼35 g/L) is considered an infinite water resource, and seawater desalination by reverse osmosis (RO) using semipermeable membranes has been practiced at commercial scales for decades at numerous countries such as Israel (Segal et al., 2018), Australia (Linge et al., 2013), Spain (Quevedo et al., 2011), and the US (Rao et al., 2018), and is believed to be the most optimized technology for seawater desalination (USDOE, 2014). Existing seawater RO (SWRO) plants operate near the thermodynamic limit, where the applied pressure is only 10–20% higher than the osmotic pressure of the concentrate (Elimelech and Phillip, 2011).

  • Acidification and decarbonization in seawater: Potential pretreatment steps for biofouling control in SWRO membranes

    2019, Desalination
    Citation Excerpt :

    The reduced buffer capacity of the solution, resulting from the decarbonization, allows the supplement of only a small strong base dose in order to elevate the water to the desired pH level. This pretreatment stage was proven effective and has the potential to lower costs by about $0.03/m3 of produced fresh water relative to the two RO pass alternative, which is currently applied in Israeli desalination plants [22,27]. Such a pretreatment sequence may meet the need to reduce interactions between the variety of biofouling and organic fouling components at elevated pH values [4].

View all citing articles on Scopus
View full text