- Email: [email protected]
Effect of chemical cleaning agents on virgin nanofiltration membrane as characterized by positron annihilation spectroscopy
Separation and Purification Technology 110 (2013) 51–56
Contents lists available at SciVerse ScienceDirect
Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur
Effect of chemical cleaning agents on virgin nanofiltration membrane as characterized by positron annihilation spectroscopy Ahmed AL-Amoudi ⇑ Saline Water Desalination Research Institute, Saline Water Conversion Corporation (SWCC), P.O. Box 8328, Al-Jubail, Saudi Arabia
a r t i c l e
i n f o
Article history: Received 11 December 2012 Received in revised form 5 February 2013 Accepted 7 February 2013 Available online 17 February 2013 Keywords: Cleaning Membrane pore size Positron annihilation Nanofiltration membranes Cleaning agents
a b s t r a c t In membrane process industries, membrane cleaning is one of the most important concerns from both economic and scientific points of view. Though cleaning is important to restore membrane performance, the inappropriate selection of cleaning agents may result in unsatisfactory cleaning or irreparable membrane damage. In this study, the cleaning performance was studied by measuring membrane pore size; by positron annihilation spectroscopy and salt rejection as well as by flux measurement. Thin film composite nanofiltration (NF) membranes, supplied by GE Osmonics NF–DK were used in this study. Tests were carried out on virgin membranes. Several different cleaning agents were investigated. Some of them were of analytical grade such as HCl and NaOH. Others such as SDS and mixed agents were of commercial grade already in use in commercial plants. Pore size and salt rejection as well as the flux of virgin membranes before and after chemical cleaning were measured and compared. After the chemical cleaning of the virgin membranes, the membrane pore size was measured revealing interesting results, which may be used to characterize membrane surface cleanliness. The membrane pore size results showed that the cleaning agents affect membrane surface properties by enlarging the pore size of the treated virgin membranes. The flux and rejection were correlated with pore size after low and high pH cleaning, and these findings were investigated and reported in this paper. Ó 2013 Elsevier B.V. All rights reserved.
1. Introduction The separation characteristics of nanofiltration come between ultrafiltration (UF) and reverse osmosis (RO). Membrane selectivity has often been attributed to the mechanism of the molecular-sieving properties of ultrafiltration and the diffusion properties of RO. It is generally recognized that membrane hydrophobicity/hydrophilicity, pore size (and their distribution) and surface charge may be important factors determining separation performance and the fouling tendency of nanofiltration membranes [1–3]. The interaction of organic and inorganic colloidal substances with membrane surfaces in aqueous media is an important factor, which depends not only on the membrane surface charge and hydrophobicity/hydrophilicity of the surface but also on the membrane pore size and pore size distribution. A large number of chemical cleaning agents are commercially available. Five categories of cleaning agent commonly used are: alkalis, acids, metal chelating agents, surfactants and enzymes [4,5]. Commercial cleaning products usually have a mixture of these chemicals but their actual composition is often unknown. The choice of the optimal cleaning agent or chemical composition
⇑ Fax: +966 33431615. E-mail address: [email protected] 1383-5866/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.seppur.2013.02.005
depends on feed characteristics. For example, acid cleaning is suitable for the removal of precipitated salts, such as CaCO3, while alkaline cleaning is used to remove adsorbed organics [6]. Flux measurement directly reflects the cleaning process. It is generally accepted that flux decline in aqueous solutions (containing organic and inorganic molecules) is mainly caused by adsorption or crystallization, possibly enhanced by pore blocking and/or cake formation [7–9]. Effective chemical cleaning would therefore be necessary to detach different classes of foulants from the membrane and thereby restore its permeate flux characteristics [10]. Furthermore, the selection of appropriate chemical cleaning agents might be critical. This is because incompatible combinations of cleaning agent and membrane material could lead to irreversible flux loss and unnecessary costs due to excessive chemical use resulting in reduction of membrane life [4,10]. In order to predict membrane separation performance, it is necessary to know the mean pore size and the pore size distribution at the membrane surface. The pore size distribution, the diffusion parameter and the experimental water flux through the membrane were all used to calculate the retention as a function of the molecular weight and pressure for several different membranes. This makes it possible to determine retention curves at different pressures, and to calculate the variation of the Molecular Weight (MW) with pressure [10]. Bowen et al. [11,12], developed a method called the Donnan
52
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56
steric partitioning pore model (DSPM), which is based on the extended Nernst–Planck equation (modified to include steric effects) and a modified Donnan equation accounting for the sieving effect as well. A theoretical background description of the DSPM has been given in detail elsewhere [11,12]. The DSPM has been successfully used to describe retention of neutral solutes and ions on NF and Ultrafiltration (UF) membranes and on titanium membranes with nanofiltration membrane properties [14,15]. In previous work, an updated version of the DSPM was used [16,17]. Using this model offers the advantage of removing membrane thickness from the rejection equation because the nanofiltration driving force is redefined in terms of an effective pressure as opposed to the volumetric flux. As a direct result, the rejection equation for uncharged solute becomes solely dependent on the pore radius. Therefore, experimental rejection set-up at various levels applied effective pressures for a solute of known size directly characterize the membrane pore radius. Alamoudi et al. [18] reported that the updated DSPM method failed to provide any valuable information about the pore size of the untreated and treated NF virgin and fouled membranes. It is well established that chemical cleaning agents such as SDS, mixed agents, NaOH and HCl are widely used in the applied industrial plants. The key message of this work and previous works is to open-up the discussion of cleaning strategies to those who have applied such cleaning agents in their plants in order to restore the production. Cleaning processes do regenerate the water produced, but simultaneously degrade its quality. Therefore, the efficiency of cleaning agents, widely used in the industrial plants, must be investigated thoroughly in the laboratory before they are used in industrial plants. The objective of this study is to evaluate the validity of the positron annihilation technique to determine whether it is possible to obtain useful information about the membrane pore size after cleaning or not, and to ascertain whether a correlation exists between the flux and/or salt rejection and the data obtained from these technique. Lo et al. [19] used the positron annihilation technique to characterize the variation in the fine structure of plasma-polymerized composite membrane, concluding that the free volume radius and the free volume amount in the plasma-polymerized layer were decreased. Kim and Fane [20] studied cleaned and fouled UF membranes using flux methods combined with microscopic methods. Their results were analyzed by determining the pore size distributions before and after cleaning. Warczok et al. [21] have recently reported that it is possible to find out whether the cleaning procedure is correctly designed or not by Atomic Force Microscope (AFM) analysis membrane images [18]. In this study, various types of analytical and commercial grade cleaning agents were used to clean virgin membranes. In order to evaluate the cleaning results, it is important to compare the characterization of treated and untreated membranes with chemical cleaning agents. In most cases, the virgin membranes were cleaned before use. The characteristics of the membrane after precleaning could have changed considerably compared to those of the virgin membranes and this is to be confirmed by using positron annihilation techniques to measure the membrane pore size. 1.1. Principle of positron annihilation technique Positron annihilation is well documented as a powerful tool for examining the pore size of polymer materials at the molecular level. Most frequently used in positron annihilation are positron annihilation lifetime and positron annihilation c-ray techniques. The positron lifetime technique measures the lifetimes of positrons and positronium (Ps: the hydrogen-like bond state of a positron and an electron), whereas the positron annihilation c-ray technique measures the Doppler broadening of annihilation radiation and the positron 3c annihilation probability.
Certain positrons implanted in insulators such as polymeric membranes may combine with one of the spur electrons to form PS. In free space, spin-parallel triplet ortho-Ps (o-Ps) annihilates into 3c-rays with an intrinsic lifetime of 142 ns, while spin-antiparallel singlet para-Ps (p-Ps)annihilates into 2c-rays with a shorter lifetime of 125 ps. If o-Ps is localized in a molecular sized hole, it annihilates with a much shorter lifetime than 142 ns through a 2c pick-off process upon collision with electrons on the hole walls. This quantum confinement of o-Ps in, a smaller hole (especially of sub-nanometer size) reduces the probability of 3c annihilation. The o-Ps lifetime by 2c pick off annihilation correlates with the hole dimension, so that the sub-nanometer-sized free volume can be examined by the positron annihilation lifetime technique. The relationship between the o-Ps lifetime s3 (ns) and a hole radius r (nm) smaller than 1 nm has been confirmed by Nakanishi, Jean and Eldrup based on a simple quantum mechanical model of Tao as follows:
s3 ¼ 0:5 1
1 r 1 2pr þ sin r þ 0:166 2p r þ 0:166
ð1Þ
The free-volume size Vf is calculated as
Vf
4 3 pr 3
ð2Þ
When a positron–electron annihilation pair possesses a longitudinal momentum p, the annihilation)’ rays are Doppler shifted from 511 keV by ±cp/2, where c is the speed of light. This brings about a significant broadening of the 511 keV annihilation photo peak, which is evaluated as the line-shape S parameter defined as the relative counts of the central region in the 511 keV annihilation photo peak. The electrons involved in different chemical elements and bonds possess their own characteristic momentum distributions, so that the S parameter is sensitively influenced by the chemical structure of the molecules surrounding the positrons. In the presence of pores far larger than 1 nm, some o-Ps may be trapped therein before annihilation. As a result of the reduced overlap of the Ps wave function with the electrons on the pore walls, such o-P shardly annihilates via the pick-off process and predominantly undergoes 3c annihilation. Thus, the positron 3c annihilation probability provides useful information on the porosity of various materials. Implementing depth selectivity to positron annihilation by combining it with variable-energy positron beams considerably broadens its applicability. Varying incident positron energy over a wider range enables depth-profiling; meanwhile, tuning the beam energy facilitates the study of surfaces, interfaces, and thin films. The mean implantation depth of positrons Zm (nm) is expressed in the following equation:
Zm ¼
40
q
E16 in
ð3Þ
where p is the material density in g/cm3, and Ein is the positron incident energy in keV, respectively. For example, at 1 keV, Zm is about 40 nm, and at 2 keV, it is about 120 nm in the case of a material with a density of unity. The variations of the positron annihilation characteristics as a function of Ei can be used to analyze the layered structure of composite membranes. 2. Experimental 2.1. Materials 2.1.1. Membrane One specific type of commercial NF membrane, GE Osmonics DK flat sheet membrane was used in this study. DK thin-film
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56
membranes (TFM) are negatively charged having a proprietary active nanopolymer layer based on poly pyperazin amide. The active top layer of DK consists of three sub-layers; (these TFM were supplied by GE Osmonics, Florida, USA). However, the molecular cutoff for a DK NF membrane is 150–300 Dalton [24]. Prior to the experiments, all virgin NF membranes were soaked overnight with or without cleaning agents in high purity water obtained from a Millipore ELix3 unit. 2.1.2. Chemicals The analytical grade salt used in these experiments was divalent magnesium sulfate-7-hydrite (Sigma–Aldrich). The salt concentration ions of the feed and product were measured via conductivity. High purity water obtained from a Millipore Elix 3 unit (with a conductivity of less than 1 lS cm1) was used in these experiments at a pH of 5.8. 2.1.3. Cleaning agents Two analytical grade chemicals (HCl and NaOH) and two commercial grade chemicals SDS (0.1%) and a mixture of tri sodium phosphate (0.1%), sodium tri polyphosphate (0.1%) and EDTA (0.1%)) (TSP + EDTA + STP) were used as cleaning agents in the experiments. Table 1 shows the name and concentration of each cleaning agent used in the tests. The commercial cleaning agents are similar to that used in the Umm-Lujj commercial NF plant [23].
53
An intense pulsed-positron beam generated with an electron linear accelerator was used to measure the lifetimes of positrons at different Ein. Lifetime data were recorded by determining the time interval between the timing signal from the pulsing system and the detection of an annihilation c-ray using a BaF2 scintillation detector. The recorded positron lifetime data were analyzed assuming three or four exponential components plus background to deduce the average lifetime s3 and relative intensity I3 of the long-lived component of o-Ps. The free-volume hole size was evaluated from s3 using Eqs. (1) and (2). 2.2.3. Flux and rejection measurements In order to characterize the cleaning efficiency, the flux and the rejection of divalent (MgSO47H2O at2482 ppm) salt were measured. Permeation and rejection experiments with divalent salt solutions in demineralised water were carried out using a laboratory-scale cross-flow recirculation test unit. The test unit consisted of a rectangular membrane cell, a feed reservoir, a variable speed gear pump, a flow meter to measure cross-flow, a balance for measuring filtrate flow, pressure transducers, a temperature control system and a computerized data logging system. Detailed descriptions of the unit and instrumentation can be found elsewhere [24,25]. The pressure of the inlet and outlet was maintained at 400 kN m2 and cross flow rate was maintained at 0.35 l/min by employing variable speed drive. All experiments were carried out at a temperature of 25 °C.
2.2. Methods 2.2.1. Cleaning procedure New membranes were used for each set of experiments. In one set of experiments, the virgin membranes were rinsed several times in demineralised water in order to remove any impurities from the virgin membranes. Bench scale experiments were carried out in this study. Several sets of experiments were carried out with different cleaning agents by soaking virgin membranes in cleaning agents overnight for 18 h and then starrier for 90 min. The treated virgin membranes were then soaked in high purity water for a few hours and thoroughly rinsed using high purity water in order to remove any remaining cleaning agents from the membrane surfaces. 2.2.2. Positron annihilation Positron annihilation c-ray spectra for the membranes were collected at different Ein with a Na-source-based magnetically guided positron beam system. The line-shape S parameter was determined as the ratio of the counts appearing in the central region (510.3–511.7 keV) to the total counts of the 511 keV and annihilation photo peak (506.8–515.2 keV) in the spectrum recorded with a Ge detector. The 3c annihilation of o-Ps could lead to c rays with energies lower than 511 keV. The positron 3c annihilation was characterized as the R parameter, which was determined as the ratio of the counts in a low-energy window (365.0– 495.0 keV) to those in the 511 keV and annihilation photo peak (506.8–515.2 keV) in the spectrum obtained as a function of Ein. Table 1 Chemical cleaning concentration.
1 2 3 4
Cleaning agents
Concentration
pH
Grade
HCl NaOH SDS Mixed Agent of EDTA, TSP and STP
1M 2M 0.1% 0.1%
2.5 11.3 11.3 11.3
Analytical Analytical Commercial Commercial
Note
Similar to that used in actual NF–SWRO plant at Umm-lujj
2.2.4. Model flux and rejection measurement protocol A new piece of membrane was placed in the cell and water was then passed through the cell to wash out any impurities that might have been present. This procedure was repeated twice. Then, a MgSO47H2O model solution at 2464 ppm concentration was passed through the membrane. The rejection was measured after 20 ml of permeate passed through the membrane. Then, the cell was washed thoroughly by circulating fresh water at the same pressure. This procedure was repeated once or twice with the new virgin membranes. 3. Results and discussion Both untreated and treated virgin DK TFM with high and low pH cleaning agents were characterized using pore size measurement via the positron annihilation as well as flux and salt rejection measurements. A correlation was found between the pore size for untreated and treated NF membranes, flux and salt rejection. 3.1. Positron annihilation Fig. 1 shows the variation of R for the untreated virgin NF–DK membranes and treated NF-DK membranes with the high pH of mixed cleaning agents and the low pH of HCl as a function of Ein. In contrast to the S parameter, the overall trends of this quantity for all the samples are similar to one other. R first decreases with increasing Ein, followed by a shallow minimum. It finally becomes nearly constant above 10.0 keV. Thus, the general Em dependence of R can be divided roughly into the following three stages, viz: (1) Ein < 0.8 keV, where R decreases with Ein, (2) 0.8 keV 6 Ein 6 2.2 keV, where R exhibits a shallow minimum; and (3) Ein > 2.2 keV, where R increases with Ein. During the thermalization process, energetic o-Ps travels penetrate a few nanometers in a polymer matrix. If energetic o-Ps is formed at a depth from the surface which is shallower than the traveling distance, it may escape out from the surface, annihilating into 3c-rays in the vacuum. In view of Eq. (3), increased R in the first stage (Ein < O.8 keV) in comparison with the second stage (O.8 keV 6 Ein 6 2.2 keV)
54
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56
Fig. 1. The variation of R for (1) the untreated virgin NF–DK membranes and treated NF–DK membranes with (2) low pH of HCl and (3) high pH of mix as a function of Ein.
can reasonably be attributed to the escape of energetic o-Ps into the vacuum and its subsequent 3c annihilation. In the second stage, where the positrons undergo predominantly 2c annihilation in the dense top region of the composite membranes, the lowest values of Rare observed. In contrast, in the third stage (Ein > 2.2 keV), especially above 10.0 keV, R displays large values for all the membranes because a fraction of the positrons undergoes 3c annihilation in the large pores of the substrate. Accordingly, three stages can be related the (Figs. 1 and 2) of R to Ps annihilations (1) in the vacuum/near the surface, (2) in the dense film, and (3) in the porous substrate, respectively. Desalination by the separation membrane is normally discussed in terms of the Donnan potential and the size exclusion effect. The previous work [26] simply makes the following prediction: if the holes involved in the separation process are far smaller than a target solute, the transport of the solute molecules would be effec-
tively restrained. During the separation process, the holes in the separation membrane act as filtering channels. The overall rejection of uncharged solutes can be observed by taking account of the solute hindrance factor and the steric partitioning coefficient, both of which vary as a function of the size ratio of the solute to the hole. In the present study, the flux and solute rejection in the case of the untreated and treated NF-DK membranes can be compared and correlated with the hole size Vf. The R Ein data in Fig. 1 suggest that the active film can be detected by positrons in the range of 0.8 keV 6 Ein 6 2.2 keV (the second stage). Only in this energy range is the lifetime free from Ps annihilations in vacuum/interface and in the porous substrate. Consequently, the free volume hole radii in the active top layer for untreated and treated NF–DK were evaluated (Table 2) based on the averaged o-Ps lifetimes in the above energy range. As a result of previous and current work, the effects of membrane pore size and surface state were determined. However, the use of cleaning agents also had a pronounced effect on the pore size and surface state [24,26,27]. At high pH values, cleaning agents consisting of SDS or a mixture of tri sodium phosphate, sodium tri poly phosphate and EDTA (recommended by the membrane manufacturer), had a marked impact on the pore size which expanded to 0.30 and 0.31 nm, when compared to the pore size of untreated NF virgin membranes (0.277 nm) (Table 2) [1]. At a low pH, chemical cleaning with HCl had minor effect on the pore size of 0.28 compared to untreated NF virgin membrane of 0.277 nm. On the other hand, chemical cleaning agent of NaOH had increased the pore size from 0.277 nm to 0.29 nm. Thus, the positron annihilation technique is a powerful technique for measuring the change in the pore size which was affected by chemical cleaning agents. 3.2. Flux and rejection In order to obtain the experimental data to characterize both treated and untreated membranes, a cross flow unit was used to carry out the water filtration tests. The pressure was maintained at 4 bars for 4–5 h until steady state condition was reached overcoming membrane compactions. Permeation experiments with divalent salt solutions were conducted using a cross-flow recirculation test unit. The permeate flux and rejection were measured after 20 ml of permeate passed through the membrane. Then, the cell was washed thoroughly by circulating fresh water under the
Fig. 2. Positron annihilation mechanism near nanofiltration membrane surface.
55
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56 Table 2 Results on the positron annihilation lifetime measurements. Sample
Note
1 2
Virgin membrane Virgin membrane SHMP + EDTA Virgin membrane Virgin membrane Virgin membrane
3 4 5
Table 4 Rejection of magnesium sulfate of virgin NF membrane.
Lifetime (ns)
Radius (nm)
treated with
1.90 2.20
0.277 0.31
treated with SDS treated with HCl treated with NaOH
2.19 1.96 2.0
0.30 0.28 0.29
Rejection (%) DK Untreated Treated by Treated by Treated by Treated by
HCl NaOH SDS mix
96.4 96.0 94.0 88.0 92.4
Standard deviation is less than 5%.
same pressure. Treated DK membranes were found to exhibit higher fluxes than untreated virgin membranes when the model solution was used as feed. For example, the flux increased more than double in the case of the virgin membrane after it had been cleaned with high pH cleaning agents. The highest flux (17.6 kg/m2 h) was achieved by treating the membranes with a mixed cleaning agent and the flux lowest (9.2 kg/m2 h) by using HCl cleaning compared to an untreated membrane flux was 5.6 kg/m2 h (Table 3). While the flux for NF membrane treated with SDS and NaOH were about 16.0 and 14.5 kg/m2 h, respectively. In order to characterize the cleaning efficiency, the rejection of the divalent salts (MgSO47H2O at 2482 ppm) was measured using a cross-flow recirculation test unit. The divalent rejection of MgSO47H2O in the case of virgin DK TFM was about 96.5% (Table 4). The rejection of divalent ions in the case of virgin NF TFM cleaned by HCl, NaOH and a mixed cleaning agent was 96%, 94%and 92.4%, respectively, which was slightly closed to untreated virgin membrane (Table 4). Conversely, in the case of SDS cleaning agents used on virgin NF TFM the rejection of divalent ions was found to be significantly lower (88%) than an untreated membrane (96.4%). 3.3. Correlation Owing to the lack of data, it is only possible to give a preliminary indication of any correlation that might exist. More work is needed to make such a correlation consistent. This study is intended to evaluate the validity of these techniques. Specifically, whether it is possible to obtain useful information about the pore size of membranes after cleaning or not, and also whether a correlation exists between the flux and/or salt rejection and the data obtained from this technique. If it works, then varying the cleaning concentration and the cleaning procedure, will be worthy of study in order to determine the exact optimum cleaning system. The correlation between pore size and the flux as well as between pore size and the salt rejection have shown very positive results. Correlations have been found for both untreated virgin NF membranes and treated virgin NF membranes. Nevertheless, the results obtained from the correlation between the pore size (free volume hole radii) and flux in the case of treated and untreated virgin membranes indicate that the radii of these membranes provide very useful information. Membrane flux increases as the membrane surface pore size becomes larger (Fig. 3). For example, the
Fig. 3. The correlation between pore size and flux for virgin membrane cleaned by various cleaning agents.
flux was increased for treated virgin membrane with HCl, NaOH, SDS and a mixture of 9.2, 14.5, 16.1 and 17.6 kg/m2 h, respectively, as the corresponding pore size was increased to 0.28, 0.29, 0.30 and 0.31 nm, respectively. A correlation also exists between the rejection and the pore size. The correlation results obtained by using this technique show that the rejection of the NF membrane cleaned with low pH cleaning agents was found to exhibit a value similar to that of an untreated virgin membrane. This was evident from the measured pore size of treated NF membrane with low pH of HCl which was approximately similar to that of virgin membrane (Fig. 4). The correlation results obtained from these techniques reveal that the rejection of the NF membrane for the SDS and the mixed cleaning agent with high pH provided the lowest rejection rates
Table 3 Flux of hydrated magnesium sulfate of virgin NF membrane. kg/m2 h Untreated membrane Treated membrane by HCl NaOH Mix SDS Standard deviation is less than 5%.
Virgin DK 5.6 9.2 14.5 17.6 16.1 Fig. 4. The correlation between pore size and the rejection for virgin membrane cleaned by various cleaning agents.
56
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56
of 88 and 92%, respectively. This indicates that membrane surface properties, such as the pore size, were impacted by this treatment when compared to other cleaning agents. This suggests a possible enlargement of the membrane pore and adsorption of the high pH cleaning agents within the membrane surface or a degradation of the polymer in the active layer [27–29]. Based on the result obtained, the membrane treated with the low pH cleaning agent had a slight effect on the membrane surface properties when compared to high pH cleaning agents [27–29]. 4. Conclusion The overall findings of this research illustrates that positron annihilation measurement provides valuable information about the pore size of NF membranes after cleaning and it is a reliable technique for assessing the surface condition of cleaned NF membranes. In most cases, chemical cleaning is found to affect membrane surface properties such as the pore size of virgin membranes. The pore size measurements derived from positron annihilation spectroscopy indicate that mixed and SDS cleaning agents had a profound adverse effect on the virgin NF membranes. The membrane surface properties were affected to such extent that the pore size increased by over 12% indicating an irreversible chemical reaction and adsorption on the membrane surface. By contrast, HCl cleaning resulted in a pore size that is broadly similar to that of a virgin NF untreated membrane. The correlation between membrane pore size and flux as well as salt rejection provided valuable information. Study data show that the high pH cleaning agents had an adverse effect on the membrane surface properties compared to low pH cleaning agents. The correlation results reveal that the rejection of the NF membrane in the case of the mixed and SDS high pH cleaning agents yielded the lowest rejections as a result of the enlargement of the pore. This indicates that the membrane surface properties were impacted by this treatment, as compared to other cleaning agents. This indicates a possible enlargement of membrane pores and the adsorption of the high pH cleaning agents within the membrane surface or a degradation of the polymer in the active layer. The positron annihilation spectroscopy method provides valuable information about the pore sizes of the untreated and treated membranes, cleaned by various cleaning agents. The pore size differences between treated and untreated NF membranes are deemed to be significant for further investigation. The rejection and flux methods provided useful information for greater understanding as to how membranes interact with cleaning agents. Acknowledgment We thank AIST for their analysis support and G.E Osmonic for the donation of the virgin NF membrane (DK).
[3] B. Van der Bruggen, J. Schaep, D. Wilms, C. Vandecasteele, Influence of molecular size, polarity and charge on the retention of organic molecules by nanofiltration, J. Membr. Sci. 156 (1999) 29–41. [4] H. Lee, G. Amy, J. Cho, Y. Yoon, S. Moon, I.S. Kim, Cleaning strategies for flux recovery of an ultrafiltration membrane fouled by natural organic matter, Water Res. 35 (2001) 3301–3308. [5] G. Morel, N. Ouazzani, A. Graciaa, J. Lachaise, Surfactant modified ultrafiltration of nitrate ions removal, J. Membr. Sci. 134 (1997) 47–57. [6] J. Vrouwenvelder, J. Kappelhof, S. Heijman, J. Schippers, D. Kooij, Tools for fouling diagnosis of NF and RO membranes and assessment of the fouling potential of feed water, Desalination 157 (2003) 361–365. [7] B. Van der Bruggen, L. Braeken, C. Vandecasteele, Flux decline in nanofiltration due to adsorption of organic compounds, Sep. Purif. Technol. 29 (1) (2002) 23– 31. [8] S. Lee, J. Kim, C.-H. Lee, Analysis of CaSO4 scale formation mechanism in various nanofiltration modules, J. Membr. Sci. 163 (1) (1999) 63–74. [9] A. Al-amoudi, Factors affecting natural organic matter (NOM) and scaling fouling in NF membranes: a review, Desalination 259 (2010) 1–10. [10] H. Zhu, M. Nystrom, Cleaning results characterized by flux, streaming potential and FTIR measurements, Colloids Surfaces A: Physicochem. Eng. Aspects 138 (2–3) (1998) 309–321. [11] B. Van der Bruggen, C. Vandecasteele, Modelling of the retention of uncharged molecules with nanofiltration, Water Res. 36 (5) (2002) 1360–1368. [12] W.R. Bowen, A.W. Mohammad, N. Hilal, Characterisation of nanofiltration membranes for predictive purposes – use of salts, uncharged solutes and atomic force microscopy, J. Membr. Sci. 126 (1) (1997) 91–105. [14] C. Labbez, P. Fievet, A. Szymczyk, A. Vidonne, A. Foissy, J. Pagetti, Analysis of the salt retention of a titania membrane using the ‘‘DSPM’’ model: effect of pH, salt concentration and nature, J. Membr. Sci. 208 (1–2) (2002) 315–329. [15] R.W. Bowen, T.A. Doneva, H.B. Yin, Polysulfone – sulfonated poly(ether ether) ketone blend membranes: systematic synthesis and characterisation, J. Membr. Sci. 181 (2) (2001) 253–263. [16] W.R. Bowen, J. Welfoot, Modelling the performance membrane nano-filtration – critical assessment and model development, Chem. Eng. Sci. 157 (2002) 1121. [17] W.R. Bowen, J. Welfoot, P. Williams, A linearised transport model for nanofiltration: development and assessment, AICHE 143 (2002) 760. [18] A. Al-Amoudi, C. Wright, P. Williams, R.W. Lovitt, A. Al-Hobiab, Atomic force microscopy (AFM) analysis of cleaned and fouled nanofiltration membranes, Int. J. Nano Biomater. 2 (2009) 82–91. [19] C.H. Lo, J.K. Huang, W.S. Hung, S.H. Huang, M. De Guzman, V. Rouessac, C.L. Li, C.C. Hu, K.R. Lee, J.Y. Lai, Investigation on the variation in the fine structure of plasma-polyerized composite membrane by positron annihilation spectroscopy, J. Membr. Sci. 337 (2009) 297–303. [20] K.-J. Kim, A.G. Fane, Performance evaluation of surface hydrophilized novel ultrafiltration membranes using aqueous proteins, J. Membr. Sci. 99 (2) (1995) 149–162. [21] J. Warczok, M. Ferrando, F. Lopez, C. Guell, Concentration of apple and pear juices by nanofiltration at low pressures, J. Food Eng. 63 (2004) 63–70. [23] A.S. Al-Amoudi, A.M. Farooque, Performance, restoration and autopsy of NF membranes used in seawater pretreatment, Desalination 178 (2005) 261–271. [24] A. Al-amoudi, NF Nanofiltration Membrane Characterizations, in Chemical Engineering, Swansea University, Swansea, 2009. [25] W.R. Bowen, T.A. Doneva, H.-B. Yin, Separation of humic acid from a model surface water with PSU/SPEEK blend UF/NF membranes, J. Membr. Sci. 206 (1– 2) (2002) 417–429. [26] A. Al-Amoudi, P. Williams, A.S. Alhobib, R.W. Lovitt, Cleaning results of new and fouled nanofiltration membrane characterized by contact angle, updated DSPM, flux and salts rejection, Appl. Surf. Sci. 254 (2008) 3983–3992. [27] N.D. Lawrence, J.M. Perera, M. Iyer, M.W. Hickey, G.W. Stevens, The use of streaming potential measurements to study the fouling and cleaning of ultrafiltration membranes, Sep. Purif. Technol. 48 (2) (2006) 106–112. [28] W.S. Ang, S. Lee, M. Elimelech, Chemical and physical aspects of cleaning of organic-fouled reverse osmosis membranes, J. Membr. Sci. 272 (1–2) (2006) 198–210. [29] Q. Li, M. Elimelech, Organic fouling and chemical cleaning of nanofiltration membranes: measurements and mechanisms, Environ. Sci. Technol. 38 (17) (2004) 4683–4693.
References Further reading [1] A. Al-Amoudi, P. Williams, S. Mandale, R.W. Lovitt, Cleaning results of new and fouled nanofiltration membrane characterized by zeta potential and permeability, Sep. Purif. Technol. 54 (2) (2007) 234–240. [2] V. Mavrov, H. Chmiel, B. Heitele, F. Rogener, Desalination of surface water to industrial water with lower impact on the environment. Part 2: improved feed water pretreatment, Desalination 110 (1–2) (1997) 65–73.
[13] W.R. Bowen, H. Mukhtar, Characterisation and pridiction of separation performance of nanofiltration membranes, J. Membr. Sci. 112 (2) (1996) 263–274. [22] M.T. Manual, Manufacturers Technical Manual.
Contents lists available at SciVerse ScienceDirect
Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur
Effect of chemical cleaning agents on virgin nanofiltration membrane as characterized by positron annihilation spectroscopy Ahmed AL-Amoudi ⇑ Saline Water Desalination Research Institute, Saline Water Conversion Corporation (SWCC), P.O. Box 8328, Al-Jubail, Saudi Arabia
a r t i c l e
i n f o
Article history: Received 11 December 2012 Received in revised form 5 February 2013 Accepted 7 February 2013 Available online 17 February 2013 Keywords: Cleaning Membrane pore size Positron annihilation Nanofiltration membranes Cleaning agents
a b s t r a c t In membrane process industries, membrane cleaning is one of the most important concerns from both economic and scientific points of view. Though cleaning is important to restore membrane performance, the inappropriate selection of cleaning agents may result in unsatisfactory cleaning or irreparable membrane damage. In this study, the cleaning performance was studied by measuring membrane pore size; by positron annihilation spectroscopy and salt rejection as well as by flux measurement. Thin film composite nanofiltration (NF) membranes, supplied by GE Osmonics NF–DK were used in this study. Tests were carried out on virgin membranes. Several different cleaning agents were investigated. Some of them were of analytical grade such as HCl and NaOH. Others such as SDS and mixed agents were of commercial grade already in use in commercial plants. Pore size and salt rejection as well as the flux of virgin membranes before and after chemical cleaning were measured and compared. After the chemical cleaning of the virgin membranes, the membrane pore size was measured revealing interesting results, which may be used to characterize membrane surface cleanliness. The membrane pore size results showed that the cleaning agents affect membrane surface properties by enlarging the pore size of the treated virgin membranes. The flux and rejection were correlated with pore size after low and high pH cleaning, and these findings were investigated and reported in this paper. Ó 2013 Elsevier B.V. All rights reserved.
1. Introduction The separation characteristics of nanofiltration come between ultrafiltration (UF) and reverse osmosis (RO). Membrane selectivity has often been attributed to the mechanism of the molecular-sieving properties of ultrafiltration and the diffusion properties of RO. It is generally recognized that membrane hydrophobicity/hydrophilicity, pore size (and their distribution) and surface charge may be important factors determining separation performance and the fouling tendency of nanofiltration membranes [1–3]. The interaction of organic and inorganic colloidal substances with membrane surfaces in aqueous media is an important factor, which depends not only on the membrane surface charge and hydrophobicity/hydrophilicity of the surface but also on the membrane pore size and pore size distribution. A large number of chemical cleaning agents are commercially available. Five categories of cleaning agent commonly used are: alkalis, acids, metal chelating agents, surfactants and enzymes [4,5]. Commercial cleaning products usually have a mixture of these chemicals but their actual composition is often unknown. The choice of the optimal cleaning agent or chemical composition
⇑ Fax: +966 33431615. E-mail address: [email protected] 1383-5866/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.seppur.2013.02.005
depends on feed characteristics. For example, acid cleaning is suitable for the removal of precipitated salts, such as CaCO3, while alkaline cleaning is used to remove adsorbed organics [6]. Flux measurement directly reflects the cleaning process. It is generally accepted that flux decline in aqueous solutions (containing organic and inorganic molecules) is mainly caused by adsorption or crystallization, possibly enhanced by pore blocking and/or cake formation [7–9]. Effective chemical cleaning would therefore be necessary to detach different classes of foulants from the membrane and thereby restore its permeate flux characteristics [10]. Furthermore, the selection of appropriate chemical cleaning agents might be critical. This is because incompatible combinations of cleaning agent and membrane material could lead to irreversible flux loss and unnecessary costs due to excessive chemical use resulting in reduction of membrane life [4,10]. In order to predict membrane separation performance, it is necessary to know the mean pore size and the pore size distribution at the membrane surface. The pore size distribution, the diffusion parameter and the experimental water flux through the membrane were all used to calculate the retention as a function of the molecular weight and pressure for several different membranes. This makes it possible to determine retention curves at different pressures, and to calculate the variation of the Molecular Weight (MW) with pressure [10]. Bowen et al. [11,12], developed a method called the Donnan
52
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56
steric partitioning pore model (DSPM), which is based on the extended Nernst–Planck equation (modified to include steric effects) and a modified Donnan equation accounting for the sieving effect as well. A theoretical background description of the DSPM has been given in detail elsewhere [11,12]. The DSPM has been successfully used to describe retention of neutral solutes and ions on NF and Ultrafiltration (UF) membranes and on titanium membranes with nanofiltration membrane properties [14,15]. In previous work, an updated version of the DSPM was used [16,17]. Using this model offers the advantage of removing membrane thickness from the rejection equation because the nanofiltration driving force is redefined in terms of an effective pressure as opposed to the volumetric flux. As a direct result, the rejection equation for uncharged solute becomes solely dependent on the pore radius. Therefore, experimental rejection set-up at various levels applied effective pressures for a solute of known size directly characterize the membrane pore radius. Alamoudi et al. [18] reported that the updated DSPM method failed to provide any valuable information about the pore size of the untreated and treated NF virgin and fouled membranes. It is well established that chemical cleaning agents such as SDS, mixed agents, NaOH and HCl are widely used in the applied industrial plants. The key message of this work and previous works is to open-up the discussion of cleaning strategies to those who have applied such cleaning agents in their plants in order to restore the production. Cleaning processes do regenerate the water produced, but simultaneously degrade its quality. Therefore, the efficiency of cleaning agents, widely used in the industrial plants, must be investigated thoroughly in the laboratory before they are used in industrial plants. The objective of this study is to evaluate the validity of the positron annihilation technique to determine whether it is possible to obtain useful information about the membrane pore size after cleaning or not, and to ascertain whether a correlation exists between the flux and/or salt rejection and the data obtained from these technique. Lo et al. [19] used the positron annihilation technique to characterize the variation in the fine structure of plasma-polymerized composite membrane, concluding that the free volume radius and the free volume amount in the plasma-polymerized layer were decreased. Kim and Fane [20] studied cleaned and fouled UF membranes using flux methods combined with microscopic methods. Their results were analyzed by determining the pore size distributions before and after cleaning. Warczok et al. [21] have recently reported that it is possible to find out whether the cleaning procedure is correctly designed or not by Atomic Force Microscope (AFM) analysis membrane images [18]. In this study, various types of analytical and commercial grade cleaning agents were used to clean virgin membranes. In order to evaluate the cleaning results, it is important to compare the characterization of treated and untreated membranes with chemical cleaning agents. In most cases, the virgin membranes were cleaned before use. The characteristics of the membrane after precleaning could have changed considerably compared to those of the virgin membranes and this is to be confirmed by using positron annihilation techniques to measure the membrane pore size. 1.1. Principle of positron annihilation technique Positron annihilation is well documented as a powerful tool for examining the pore size of polymer materials at the molecular level. Most frequently used in positron annihilation are positron annihilation lifetime and positron annihilation c-ray techniques. The positron lifetime technique measures the lifetimes of positrons and positronium (Ps: the hydrogen-like bond state of a positron and an electron), whereas the positron annihilation c-ray technique measures the Doppler broadening of annihilation radiation and the positron 3c annihilation probability.
Certain positrons implanted in insulators such as polymeric membranes may combine with one of the spur electrons to form PS. In free space, spin-parallel triplet ortho-Ps (o-Ps) annihilates into 3c-rays with an intrinsic lifetime of 142 ns, while spin-antiparallel singlet para-Ps (p-Ps)annihilates into 2c-rays with a shorter lifetime of 125 ps. If o-Ps is localized in a molecular sized hole, it annihilates with a much shorter lifetime than 142 ns through a 2c pick-off process upon collision with electrons on the hole walls. This quantum confinement of o-Ps in, a smaller hole (especially of sub-nanometer size) reduces the probability of 3c annihilation. The o-Ps lifetime by 2c pick off annihilation correlates with the hole dimension, so that the sub-nanometer-sized free volume can be examined by the positron annihilation lifetime technique. The relationship between the o-Ps lifetime s3 (ns) and a hole radius r (nm) smaller than 1 nm has been confirmed by Nakanishi, Jean and Eldrup based on a simple quantum mechanical model of Tao as follows:
s3 ¼ 0:5 1
1 r 1 2pr þ sin r þ 0:166 2p r þ 0:166
ð1Þ
The free-volume size Vf is calculated as
Vf
4 3 pr 3
ð2Þ
When a positron–electron annihilation pair possesses a longitudinal momentum p, the annihilation)’ rays are Doppler shifted from 511 keV by ±cp/2, where c is the speed of light. This brings about a significant broadening of the 511 keV annihilation photo peak, which is evaluated as the line-shape S parameter defined as the relative counts of the central region in the 511 keV annihilation photo peak. The electrons involved in different chemical elements and bonds possess their own characteristic momentum distributions, so that the S parameter is sensitively influenced by the chemical structure of the molecules surrounding the positrons. In the presence of pores far larger than 1 nm, some o-Ps may be trapped therein before annihilation. As a result of the reduced overlap of the Ps wave function with the electrons on the pore walls, such o-P shardly annihilates via the pick-off process and predominantly undergoes 3c annihilation. Thus, the positron 3c annihilation probability provides useful information on the porosity of various materials. Implementing depth selectivity to positron annihilation by combining it with variable-energy positron beams considerably broadens its applicability. Varying incident positron energy over a wider range enables depth-profiling; meanwhile, tuning the beam energy facilitates the study of surfaces, interfaces, and thin films. The mean implantation depth of positrons Zm (nm) is expressed in the following equation:
Zm ¼
40
q
E16 in
ð3Þ
where p is the material density in g/cm3, and Ein is the positron incident energy in keV, respectively. For example, at 1 keV, Zm is about 40 nm, and at 2 keV, it is about 120 nm in the case of a material with a density of unity. The variations of the positron annihilation characteristics as a function of Ei can be used to analyze the layered structure of composite membranes. 2. Experimental 2.1. Materials 2.1.1. Membrane One specific type of commercial NF membrane, GE Osmonics DK flat sheet membrane was used in this study. DK thin-film
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56
membranes (TFM) are negatively charged having a proprietary active nanopolymer layer based on poly pyperazin amide. The active top layer of DK consists of three sub-layers; (these TFM were supplied by GE Osmonics, Florida, USA). However, the molecular cutoff for a DK NF membrane is 150–300 Dalton [24]. Prior to the experiments, all virgin NF membranes were soaked overnight with or without cleaning agents in high purity water obtained from a Millipore ELix3 unit. 2.1.2. Chemicals The analytical grade salt used in these experiments was divalent magnesium sulfate-7-hydrite (Sigma–Aldrich). The salt concentration ions of the feed and product were measured via conductivity. High purity water obtained from a Millipore Elix 3 unit (with a conductivity of less than 1 lS cm1) was used in these experiments at a pH of 5.8. 2.1.3. Cleaning agents Two analytical grade chemicals (HCl and NaOH) and two commercial grade chemicals SDS (0.1%) and a mixture of tri sodium phosphate (0.1%), sodium tri polyphosphate (0.1%) and EDTA (0.1%)) (TSP + EDTA + STP) were used as cleaning agents in the experiments. Table 1 shows the name and concentration of each cleaning agent used in the tests. The commercial cleaning agents are similar to that used in the Umm-Lujj commercial NF plant [23].
53
An intense pulsed-positron beam generated with an electron linear accelerator was used to measure the lifetimes of positrons at different Ein. Lifetime data were recorded by determining the time interval between the timing signal from the pulsing system and the detection of an annihilation c-ray using a BaF2 scintillation detector. The recorded positron lifetime data were analyzed assuming three or four exponential components plus background to deduce the average lifetime s3 and relative intensity I3 of the long-lived component of o-Ps. The free-volume hole size was evaluated from s3 using Eqs. (1) and (2). 2.2.3. Flux and rejection measurements In order to characterize the cleaning efficiency, the flux and the rejection of divalent (MgSO47H2O at2482 ppm) salt were measured. Permeation and rejection experiments with divalent salt solutions in demineralised water were carried out using a laboratory-scale cross-flow recirculation test unit. The test unit consisted of a rectangular membrane cell, a feed reservoir, a variable speed gear pump, a flow meter to measure cross-flow, a balance for measuring filtrate flow, pressure transducers, a temperature control system and a computerized data logging system. Detailed descriptions of the unit and instrumentation can be found elsewhere [24,25]. The pressure of the inlet and outlet was maintained at 400 kN m2 and cross flow rate was maintained at 0.35 l/min by employing variable speed drive. All experiments were carried out at a temperature of 25 °C.
2.2. Methods 2.2.1. Cleaning procedure New membranes were used for each set of experiments. In one set of experiments, the virgin membranes were rinsed several times in demineralised water in order to remove any impurities from the virgin membranes. Bench scale experiments were carried out in this study. Several sets of experiments were carried out with different cleaning agents by soaking virgin membranes in cleaning agents overnight for 18 h and then starrier for 90 min. The treated virgin membranes were then soaked in high purity water for a few hours and thoroughly rinsed using high purity water in order to remove any remaining cleaning agents from the membrane surfaces. 2.2.2. Positron annihilation Positron annihilation c-ray spectra for the membranes were collected at different Ein with a Na-source-based magnetically guided positron beam system. The line-shape S parameter was determined as the ratio of the counts appearing in the central region (510.3–511.7 keV) to the total counts of the 511 keV and annihilation photo peak (506.8–515.2 keV) in the spectrum recorded with a Ge detector. The 3c annihilation of o-Ps could lead to c rays with energies lower than 511 keV. The positron 3c annihilation was characterized as the R parameter, which was determined as the ratio of the counts in a low-energy window (365.0– 495.0 keV) to those in the 511 keV and annihilation photo peak (506.8–515.2 keV) in the spectrum obtained as a function of Ein. Table 1 Chemical cleaning concentration.
1 2 3 4
Cleaning agents
Concentration
pH
Grade
HCl NaOH SDS Mixed Agent of EDTA, TSP and STP
1M 2M 0.1% 0.1%
2.5 11.3 11.3 11.3
Analytical Analytical Commercial Commercial
Note
Similar to that used in actual NF–SWRO plant at Umm-lujj
2.2.4. Model flux and rejection measurement protocol A new piece of membrane was placed in the cell and water was then passed through the cell to wash out any impurities that might have been present. This procedure was repeated twice. Then, a MgSO47H2O model solution at 2464 ppm concentration was passed through the membrane. The rejection was measured after 20 ml of permeate passed through the membrane. Then, the cell was washed thoroughly by circulating fresh water at the same pressure. This procedure was repeated once or twice with the new virgin membranes. 3. Results and discussion Both untreated and treated virgin DK TFM with high and low pH cleaning agents were characterized using pore size measurement via the positron annihilation as well as flux and salt rejection measurements. A correlation was found between the pore size for untreated and treated NF membranes, flux and salt rejection. 3.1. Positron annihilation Fig. 1 shows the variation of R for the untreated virgin NF–DK membranes and treated NF-DK membranes with the high pH of mixed cleaning agents and the low pH of HCl as a function of Ein. In contrast to the S parameter, the overall trends of this quantity for all the samples are similar to one other. R first decreases with increasing Ein, followed by a shallow minimum. It finally becomes nearly constant above 10.0 keV. Thus, the general Em dependence of R can be divided roughly into the following three stages, viz: (1) Ein < 0.8 keV, where R decreases with Ein, (2) 0.8 keV 6 Ein 6 2.2 keV, where R exhibits a shallow minimum; and (3) Ein > 2.2 keV, where R increases with Ein. During the thermalization process, energetic o-Ps travels penetrate a few nanometers in a polymer matrix. If energetic o-Ps is formed at a depth from the surface which is shallower than the traveling distance, it may escape out from the surface, annihilating into 3c-rays in the vacuum. In view of Eq. (3), increased R in the first stage (Ein < O.8 keV) in comparison with the second stage (O.8 keV 6 Ein 6 2.2 keV)
54
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56
Fig. 1. The variation of R for (1) the untreated virgin NF–DK membranes and treated NF–DK membranes with (2) low pH of HCl and (3) high pH of mix as a function of Ein.
can reasonably be attributed to the escape of energetic o-Ps into the vacuum and its subsequent 3c annihilation. In the second stage, where the positrons undergo predominantly 2c annihilation in the dense top region of the composite membranes, the lowest values of Rare observed. In contrast, in the third stage (Ein > 2.2 keV), especially above 10.0 keV, R displays large values for all the membranes because a fraction of the positrons undergoes 3c annihilation in the large pores of the substrate. Accordingly, three stages can be related the (Figs. 1 and 2) of R to Ps annihilations (1) in the vacuum/near the surface, (2) in the dense film, and (3) in the porous substrate, respectively. Desalination by the separation membrane is normally discussed in terms of the Donnan potential and the size exclusion effect. The previous work [26] simply makes the following prediction: if the holes involved in the separation process are far smaller than a target solute, the transport of the solute molecules would be effec-
tively restrained. During the separation process, the holes in the separation membrane act as filtering channels. The overall rejection of uncharged solutes can be observed by taking account of the solute hindrance factor and the steric partitioning coefficient, both of which vary as a function of the size ratio of the solute to the hole. In the present study, the flux and solute rejection in the case of the untreated and treated NF-DK membranes can be compared and correlated with the hole size Vf. The R Ein data in Fig. 1 suggest that the active film can be detected by positrons in the range of 0.8 keV 6 Ein 6 2.2 keV (the second stage). Only in this energy range is the lifetime free from Ps annihilations in vacuum/interface and in the porous substrate. Consequently, the free volume hole radii in the active top layer for untreated and treated NF–DK were evaluated (Table 2) based on the averaged o-Ps lifetimes in the above energy range. As a result of previous and current work, the effects of membrane pore size and surface state were determined. However, the use of cleaning agents also had a pronounced effect on the pore size and surface state [24,26,27]. At high pH values, cleaning agents consisting of SDS or a mixture of tri sodium phosphate, sodium tri poly phosphate and EDTA (recommended by the membrane manufacturer), had a marked impact on the pore size which expanded to 0.30 and 0.31 nm, when compared to the pore size of untreated NF virgin membranes (0.277 nm) (Table 2) [1]. At a low pH, chemical cleaning with HCl had minor effect on the pore size of 0.28 compared to untreated NF virgin membrane of 0.277 nm. On the other hand, chemical cleaning agent of NaOH had increased the pore size from 0.277 nm to 0.29 nm. Thus, the positron annihilation technique is a powerful technique for measuring the change in the pore size which was affected by chemical cleaning agents. 3.2. Flux and rejection In order to obtain the experimental data to characterize both treated and untreated membranes, a cross flow unit was used to carry out the water filtration tests. The pressure was maintained at 4 bars for 4–5 h until steady state condition was reached overcoming membrane compactions. Permeation experiments with divalent salt solutions were conducted using a cross-flow recirculation test unit. The permeate flux and rejection were measured after 20 ml of permeate passed through the membrane. Then, the cell was washed thoroughly by circulating fresh water under the
Fig. 2. Positron annihilation mechanism near nanofiltration membrane surface.
55
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56 Table 2 Results on the positron annihilation lifetime measurements. Sample
Note
1 2
Virgin membrane Virgin membrane SHMP + EDTA Virgin membrane Virgin membrane Virgin membrane
3 4 5
Table 4 Rejection of magnesium sulfate of virgin NF membrane.
Lifetime (ns)
Radius (nm)
treated with
1.90 2.20
0.277 0.31
treated with SDS treated with HCl treated with NaOH
2.19 1.96 2.0
0.30 0.28 0.29
Rejection (%) DK Untreated Treated by Treated by Treated by Treated by
HCl NaOH SDS mix
96.4 96.0 94.0 88.0 92.4
Standard deviation is less than 5%.
same pressure. Treated DK membranes were found to exhibit higher fluxes than untreated virgin membranes when the model solution was used as feed. For example, the flux increased more than double in the case of the virgin membrane after it had been cleaned with high pH cleaning agents. The highest flux (17.6 kg/m2 h) was achieved by treating the membranes with a mixed cleaning agent and the flux lowest (9.2 kg/m2 h) by using HCl cleaning compared to an untreated membrane flux was 5.6 kg/m2 h (Table 3). While the flux for NF membrane treated with SDS and NaOH were about 16.0 and 14.5 kg/m2 h, respectively. In order to characterize the cleaning efficiency, the rejection of the divalent salts (MgSO47H2O at 2482 ppm) was measured using a cross-flow recirculation test unit. The divalent rejection of MgSO47H2O in the case of virgin DK TFM was about 96.5% (Table 4). The rejection of divalent ions in the case of virgin NF TFM cleaned by HCl, NaOH and a mixed cleaning agent was 96%, 94%and 92.4%, respectively, which was slightly closed to untreated virgin membrane (Table 4). Conversely, in the case of SDS cleaning agents used on virgin NF TFM the rejection of divalent ions was found to be significantly lower (88%) than an untreated membrane (96.4%). 3.3. Correlation Owing to the lack of data, it is only possible to give a preliminary indication of any correlation that might exist. More work is needed to make such a correlation consistent. This study is intended to evaluate the validity of these techniques. Specifically, whether it is possible to obtain useful information about the pore size of membranes after cleaning or not, and also whether a correlation exists between the flux and/or salt rejection and the data obtained from this technique. If it works, then varying the cleaning concentration and the cleaning procedure, will be worthy of study in order to determine the exact optimum cleaning system. The correlation between pore size and the flux as well as between pore size and the salt rejection have shown very positive results. Correlations have been found for both untreated virgin NF membranes and treated virgin NF membranes. Nevertheless, the results obtained from the correlation between the pore size (free volume hole radii) and flux in the case of treated and untreated virgin membranes indicate that the radii of these membranes provide very useful information. Membrane flux increases as the membrane surface pore size becomes larger (Fig. 3). For example, the
Fig. 3. The correlation between pore size and flux for virgin membrane cleaned by various cleaning agents.
flux was increased for treated virgin membrane with HCl, NaOH, SDS and a mixture of 9.2, 14.5, 16.1 and 17.6 kg/m2 h, respectively, as the corresponding pore size was increased to 0.28, 0.29, 0.30 and 0.31 nm, respectively. A correlation also exists between the rejection and the pore size. The correlation results obtained by using this technique show that the rejection of the NF membrane cleaned with low pH cleaning agents was found to exhibit a value similar to that of an untreated virgin membrane. This was evident from the measured pore size of treated NF membrane with low pH of HCl which was approximately similar to that of virgin membrane (Fig. 4). The correlation results obtained from these techniques reveal that the rejection of the NF membrane for the SDS and the mixed cleaning agent with high pH provided the lowest rejection rates
Table 3 Flux of hydrated magnesium sulfate of virgin NF membrane. kg/m2 h Untreated membrane Treated membrane by HCl NaOH Mix SDS Standard deviation is less than 5%.
Virgin DK 5.6 9.2 14.5 17.6 16.1 Fig. 4. The correlation between pore size and the rejection for virgin membrane cleaned by various cleaning agents.
56
A. AL-Amoudi / Separation and Purification Technology 110 (2013) 51–56
of 88 and 92%, respectively. This indicates that membrane surface properties, such as the pore size, were impacted by this treatment when compared to other cleaning agents. This suggests a possible enlargement of the membrane pore and adsorption of the high pH cleaning agents within the membrane surface or a degradation of the polymer in the active layer [27–29]. Based on the result obtained, the membrane treated with the low pH cleaning agent had a slight effect on the membrane surface properties when compared to high pH cleaning agents [27–29]. 4. Conclusion The overall findings of this research illustrates that positron annihilation measurement provides valuable information about the pore size of NF membranes after cleaning and it is a reliable technique for assessing the surface condition of cleaned NF membranes. In most cases, chemical cleaning is found to affect membrane surface properties such as the pore size of virgin membranes. The pore size measurements derived from positron annihilation spectroscopy indicate that mixed and SDS cleaning agents had a profound adverse effect on the virgin NF membranes. The membrane surface properties were affected to such extent that the pore size increased by over 12% indicating an irreversible chemical reaction and adsorption on the membrane surface. By contrast, HCl cleaning resulted in a pore size that is broadly similar to that of a virgin NF untreated membrane. The correlation between membrane pore size and flux as well as salt rejection provided valuable information. Study data show that the high pH cleaning agents had an adverse effect on the membrane surface properties compared to low pH cleaning agents. The correlation results reveal that the rejection of the NF membrane in the case of the mixed and SDS high pH cleaning agents yielded the lowest rejections as a result of the enlargement of the pore. This indicates that the membrane surface properties were impacted by this treatment, as compared to other cleaning agents. This indicates a possible enlargement of membrane pores and the adsorption of the high pH cleaning agents within the membrane surface or a degradation of the polymer in the active layer. The positron annihilation spectroscopy method provides valuable information about the pore sizes of the untreated and treated membranes, cleaned by various cleaning agents. The pore size differences between treated and untreated NF membranes are deemed to be significant for further investigation. The rejection and flux methods provided useful information for greater understanding as to how membranes interact with cleaning agents. Acknowledgment We thank AIST for their analysis support and G.E Osmonic for the donation of the virgin NF membrane (DK).
[3] B. Van der Bruggen, J. Schaep, D. Wilms, C. Vandecasteele, Influence of molecular size, polarity and charge on the retention of organic molecules by nanofiltration, J. Membr. Sci. 156 (1999) 29–41. [4] H. Lee, G. Amy, J. Cho, Y. Yoon, S. Moon, I.S. Kim, Cleaning strategies for flux recovery of an ultrafiltration membrane fouled by natural organic matter, Water Res. 35 (2001) 3301–3308. [5] G. Morel, N. Ouazzani, A. Graciaa, J. Lachaise, Surfactant modified ultrafiltration of nitrate ions removal, J. Membr. Sci. 134 (1997) 47–57. [6] J. Vrouwenvelder, J. Kappelhof, S. Heijman, J. Schippers, D. Kooij, Tools for fouling diagnosis of NF and RO membranes and assessment of the fouling potential of feed water, Desalination 157 (2003) 361–365. [7] B. Van der Bruggen, L. Braeken, C. Vandecasteele, Flux decline in nanofiltration due to adsorption of organic compounds, Sep. Purif. Technol. 29 (1) (2002) 23– 31. [8] S. Lee, J. Kim, C.-H. Lee, Analysis of CaSO4 scale formation mechanism in various nanofiltration modules, J. Membr. Sci. 163 (1) (1999) 63–74. [9] A. Al-amoudi, Factors affecting natural organic matter (NOM) and scaling fouling in NF membranes: a review, Desalination 259 (2010) 1–10. [10] H. Zhu, M. Nystrom, Cleaning results characterized by flux, streaming potential and FTIR measurements, Colloids Surfaces A: Physicochem. Eng. Aspects 138 (2–3) (1998) 309–321. [11] B. Van der Bruggen, C. Vandecasteele, Modelling of the retention of uncharged molecules with nanofiltration, Water Res. 36 (5) (2002) 1360–1368. [12] W.R. Bowen, A.W. Mohammad, N. Hilal, Characterisation of nanofiltration membranes for predictive purposes – use of salts, uncharged solutes and atomic force microscopy, J. Membr. Sci. 126 (1) (1997) 91–105. [14] C. Labbez, P. Fievet, A. Szymczyk, A. Vidonne, A. Foissy, J. Pagetti, Analysis of the salt retention of a titania membrane using the ‘‘DSPM’’ model: effect of pH, salt concentration and nature, J. Membr. Sci. 208 (1–2) (2002) 315–329. [15] R.W. Bowen, T.A. Doneva, H.B. Yin, Polysulfone – sulfonated poly(ether ether) ketone blend membranes: systematic synthesis and characterisation, J. Membr. Sci. 181 (2) (2001) 253–263. [16] W.R. Bowen, J. Welfoot, Modelling the performance membrane nano-filtration – critical assessment and model development, Chem. Eng. Sci. 157 (2002) 1121. [17] W.R. Bowen, J. Welfoot, P. Williams, A linearised transport model for nanofiltration: development and assessment, AICHE 143 (2002) 760. [18] A. Al-Amoudi, C. Wright, P. Williams, R.W. Lovitt, A. Al-Hobiab, Atomic force microscopy (AFM) analysis of cleaned and fouled nanofiltration membranes, Int. J. Nano Biomater. 2 (2009) 82–91. [19] C.H. Lo, J.K. Huang, W.S. Hung, S.H. Huang, M. De Guzman, V. Rouessac, C.L. Li, C.C. Hu, K.R. Lee, J.Y. Lai, Investigation on the variation in the fine structure of plasma-polyerized composite membrane by positron annihilation spectroscopy, J. Membr. Sci. 337 (2009) 297–303. [20] K.-J. Kim, A.G. Fane, Performance evaluation of surface hydrophilized novel ultrafiltration membranes using aqueous proteins, J. Membr. Sci. 99 (2) (1995) 149–162. [21] J. Warczok, M. Ferrando, F. Lopez, C. Guell, Concentration of apple and pear juices by nanofiltration at low pressures, J. Food Eng. 63 (2004) 63–70. [23] A.S. Al-Amoudi, A.M. Farooque, Performance, restoration and autopsy of NF membranes used in seawater pretreatment, Desalination 178 (2005) 261–271. [24] A. Al-amoudi, NF Nanofiltration Membrane Characterizations, in Chemical Engineering, Swansea University, Swansea, 2009. [25] W.R. Bowen, T.A. Doneva, H.-B. Yin, Separation of humic acid from a model surface water with PSU/SPEEK blend UF/NF membranes, J. Membr. Sci. 206 (1– 2) (2002) 417–429. [26] A. Al-Amoudi, P. Williams, A.S. Alhobib, R.W. Lovitt, Cleaning results of new and fouled nanofiltration membrane characterized by contact angle, updated DSPM, flux and salts rejection, Appl. Surf. Sci. 254 (2008) 3983–3992. [27] N.D. Lawrence, J.M. Perera, M. Iyer, M.W. Hickey, G.W. Stevens, The use of streaming potential measurements to study the fouling and cleaning of ultrafiltration membranes, Sep. Purif. Technol. 48 (2) (2006) 106–112. [28] W.S. Ang, S. Lee, M. Elimelech, Chemical and physical aspects of cleaning of organic-fouled reverse osmosis membranes, J. Membr. Sci. 272 (1–2) (2006) 198–210. [29] Q. Li, M. Elimelech, Organic fouling and chemical cleaning of nanofiltration membranes: measurements and mechanisms, Environ. Sci. Technol. 38 (17) (2004) 4683–4693.
References Further reading [1] A. Al-Amoudi, P. Williams, S. Mandale, R.W. Lovitt, Cleaning results of new and fouled nanofiltration membrane characterized by zeta potential and permeability, Sep. Purif. Technol. 54 (2) (2007) 234–240. [2] V. Mavrov, H. Chmiel, B. Heitele, F. Rogener, Desalination of surface water to industrial water with lower impact on the environment. Part 2: improved feed water pretreatment, Desalination 110 (1–2) (1997) 65–73.
[13] W.R. Bowen, H. Mukhtar, Characterisation and pridiction of separation performance of nanofiltration membranes, J. Membr. Sci. 112 (2) (1996) 263–274. [22] M.T. Manual, Manufacturers Technical Manual.