Elsevier

Desalination

Volume 253, Issues 1–3, April 2010, Pages 38-45
Desalination

Effect of operating conditions and solution chemistry on model parameters in crossflow reverse osmosis of natural organic matter

https://doi.org/10.1016/j.desal.2009.11.039Get rights and content

Abstract

This paper describes the effect of operating conditions and solution chemistry on model parameters in crossflow reverse osmosis of natural organic matter. Mathematical fouling model based on the combined osmotic pressure and cake filtration model was used to evaluate model parameters (i.e. steady-state flux, J* and specific cake resistance, αcake). In addition, the empirical equation for steady-state flux (J* = 9.12 × 10 8ΔP1.04v0.223R 1.18I 0.590) was successfully determined to characterize reverse osmosis operation. Steady-state flux increased with increased operating pressure, indicating a pressure-dependent steady-state flux under laminar flow condition. The specific cake resistance (αcake = 7.943 × 1012ΔP 2.03v 0.739R6.29I1.37) was inversely related to increased operating pressure and crossflow velocity, while the specific cake resistance increased linearly with recovery effects and ionic strength. Recovery effects with high ionic strength resulted in the highest flux decline, corresponding to high specific cake resistance (i.e. lowering cake porosity) due to combined salt concentration polarization and NOM cake compaction near the membrane surface.

Introduction

Reverse osmosis (RO) is the well-known membrane separation process which can be applied for seawater and brackish water desalination, softening, disinfection by-product control, and removal of organics and specific inorganic contaminants such as arsenic, barium, nitrite, nitrate, and other inorganic contaminants [1], [2]. The application of membranes has been increasingly used for water treatment in order to control a precursor of the formation of disinfection by-products such as natural organic matter (NOM) during chlorination operation in conventional water treatment [3]. However, fouling of membranes caused by NOM and inorganic salts on the membrane surface can be a major cause of a significant loss of water productivity [4], [5], [6].

Membrane fouling can be dependent on membrane characteristics (i.e. pore size, charge, and roughness) [7], [8], solution compositions (i.e. humic acid concentration, pH, ionic strength, and calcium concentration) [9], [10], and hydrodynamic conditions (i.e. flux, pressure, and crossflow velocity) [10], [11]. Fouling can also lead to decreased solution flux due to adsorption/deposition of solute on the membrane surface and in the membrane pores, and cake formation at the membrane surface. Tang et al. [10] indicated that flux reduction during RO of humic acid increased with increasing initial flux (i.e. increased operating pressure), while the tight RO showed less flux decline than the more permeable membranes. Higher operating pressure in NF of humic acids increased flux decline, while higher crossflow velocity increased solution flux due to increased back-transport of solute to the bulk solution, thus decreasing solute accumulation at the membrane surface [12]. Zhu and Elimelech [13] studied the fouling mechanisms during RO of silica colloids. They found that higher permeate flux caused by increasing transmembrane pressure resulted in a greater rate of particle deposition onto membrane surface, and thus an increased rate of membrane fouling. However, previous work showed different results indicating that the specific flux normalized to the initial value was found to be inversely related to the initial permeate rate [14]. The specific resistance of particle deposits on membranes decreased as the initial permeation rate increased, suggesting that cake morphology was an important parameter in determining permeate flux [14]. Kilduff et al. [15] indicated that the rate of flux decline increased with increasing recovery because of the increase of solute concentration on the membrane surface caused by enhancing the convective transport of mass to the membrane surface. Previous studies indicated that the operating conditions could influence membrane performance with different membrane uses and solutions [10], but there is a lack of characterization of model parameters and development of the empirical relationship among membrane operating conditions and solution chemistry.

The objective of this study was to investigate the effects of operating conditions on model parameters (i.e. steady-state flux, J* and specific cake resistance, αcake) with different ionic strengths during reverse osmosis of natural organic matter and to develop an empirical relationship among membrane operating conditions and solution chemistry. The empirical equations for J* and αcake were developed with dependent variables for operating pressure (ΔP), crossflow velocity (v), recovery (R), and ionic strength (I). Experimental results exhibited flux decline and rejection with different operating conditions and ionic strengths. The model parameters were determined using the combined osmotic pressure and cake filtration model, while a resistance-in-series model was used to characterize fouling resistances of tight RO system operation. The experimental results of this work could provide an insight evidence for changes in the model parameters as a function of system operating conditions and solution chemistry during crossflow RO.

Section snippets

Resistance-in-series model

Resistance-in-series model has been widely applied to describe the permeation flux of the membrane processes. This model incorporates membrane hydraulic resistance (Rm) and hydraulic resistances of fouling layer (Rf) on membrane surface. The membrane hydraulic resistance can be determined from the pure water flux without materials deposited on the membrane surface or within the membrane pore by using Darcy's law as shown in Eq. (1).Jo=1AmdVdt=ΔPμRmwhere Jo is the clean water flux (LMH), t is

Natural organic matter (NOM)

Natural organic matter (NOM) was obtained from the surface water reservoir at Ubon Ratchathani's University (UBU), Thailand. A polyamide thin-film composite (TFC) reverse osmosis (RO) membrane (model: AG4040F-spiral wound crossflow, GE osmonics, USA) was used to isolate NOM components and subsequently applied the isolated NOM for the crossflow reverse osmosis experiments. The isolation procedure was previously described by Kilduff et al. [18]. The characteristics of natural water were

NOM molecular weight

Fig. 1 shows the NOM molecular size distribution. The response of UV254 nm was presented in wide range of high molecular weight (10,000–100,000 Da) and low molecular weight (less than 5000 Da). The weight-averaged NOM molecular weight (Mw) was approximately 4144 Da, while the number-averaged NOM molecular weight (Mn) was about 244 Da. The polydispersity of NOM solution (the ratio of the weight- (Mw) to number-averaged (Mn) molecular weights) was approximately 16.98, indicating wide molecular size

Conclusions

The performance of crossflow reverse osmosis process of NOM with different solution chemistry was analyzed using resistance-in-series model, the combined osmotic and cake filtration model, and the empirical model. The empirical equations for steady-state flux (J*) and specific cake resistances (αcake) were successfully developed with combined dependent variables of operating pressure (ΔP), crossflow velocity (v), recovery (R). Steady-state flux increased with increased operating pressure and

Nomenclature

    a to j

    empirical constants

    Am

    membrane area (m2)

    Cfeed,s

    salt concentration in the feed (mol L 1)

    Cperm,s

    salt concentration in the permeate (mol L 1)

    Creten,NOM

    NOM concentration in the retentate (kg m 3)

    Creten,s

    salt concentration in the retentate (mol L 1)

    dp

    particle diameter (m)

    I

    ionic strength (M)

    Jo

    clean water flux (L m 2 h 1, LMH)

    Jv

    solution flux (L m 2 h 1, LMH)

    J*

    steady-state flux associated with back-transport resulting from crossflow (LMH)

    mcake

    cake mass (kg)

    Mn

    number-averaged NOM molecular weight (Da)

    Mw

Acknowledgements

The authors would like to thank the Thailand Research Fund (TRF) for TRF senior research scholar grant (No. RTA5080014) and the Commission on Higher Education, Thailand, for financial support. We also would like to thank the Department of Chemical Engineering, Faculty of Engineering, Ubon Rachathani University, and the National Center of Excellence for Environmental and Hazardous Waste Management, Ubon Ratchathani University, Ubonratchathani, Thailand, for all experimental equipments applied in

References (26)

Cited by (25)

  • Reuse and recycling of end-of-life reverse osmosis membranes

    2022, Advancement in Polymer-Based Membranes for Water Remediation
  • Membrane fouling in desalination and its mitigation strategies

    2018, Desalination
    Citation Excerpt :

    High cross-flow velocity induces high shear rates which promotes the back diffusion of particles away from the membrane surface and reduces the concentration polarisation, thus minimising fouling development [193,199]. Mattaraj et al. reported that the decline in permeate flux was lower with higher cross-flow velocity, which can be related to lower NOM fouling [200]. Similar results were also observed for humic acid [201], inorganic colloids [202,203], and biofilm development [204–206].

  • Comparison of polysaccharide fouling in forward osmosis and reverse osmosis separations

    2017, Desalination
    Citation Excerpt :

    Xanthan, which has largest MW (1000 kDa–50,000 kDa) and highest viscosity in solution, caused the most severe flux decline in FO. Polysaccharides with high MW in the feed solution may reduce the shear force generated by the cross-flow velocity and promote the accumulation of polysaccharides, resulting in greater fouling [45,55]. In particular, as xanthan accumulates on the membrane, its concentration near the surface increases and so does its viscosity.

View all citing articles on Scopus
View full text