Nanofiltration of river water — fouling, cleaning and micropollutant rejection
Yüklə 270,01 Kb. Pdf görüntüsü
|
- Bu səhifədəki naviqasiya:
- Document Outline
|
Nano filtration of river water — fouling, cleaning and micropollutant rejection ☆ Angela M. Klüpfel ⁎ , Fritz H. Frimmel Water Chemistry, Engler-Bunte-Institut, Universität Karlsruhe (TH), Germany a b s t r a c t a r t i c l e i n f o Available online 29 October 2009 Keywords: Nano
filtration Membrane fouling Membrane cleaning Polar organic micropollutants Four nano filtration and low pressure reverse osmosis membranes were investigated concerning their retention and fouling behavior during filtration of a surface water (River Alb) spiked with four polar micropollutants. Filtration and cleaning experiments were also conducted in deionized water spiked with the model contaminants in order to evaluate the in fluence of the water matrix on retention. Steric and electrostatic effects were found to be the most important factors in fluencing retention of the organic substances. In the presented experiments, fouling had little in fluence on retention. In case of the DI water experiments, treatment of the membranes with NaOH solution considerably lowered the retention of the investigated micropollutants, whereas in the experiments with river water the effect of alkaline treatment on the rejection of the trace organic substances decreased to an insigni ficant amount already 1 h after membrane cleaning. © 2009 Elsevier B.V. All rights reserved. 1. Introduction Nano
filtration (NF) is one of the emerging and promising tech- nologies in water treatment. In spite of several research projects about the subject, fouling is still inevitable and remains a major problem in the application of NF [1] . Chemical cleaning often represents the only way to restore the permeate flux to an economically feasible value. However, up to now there is little systematic work on that subject and consequently only a few publications dealing with cleaning of NF membranes have become available [2]
. This leads to a severe lack in understanding of fundamental mechanisms of the cleaning procedure and hence to a big improvement potential concerning e.g. membrane lifetime and cleaning agent consumption. An in fluence of membrane fouling on the rejection of organic contaminants by NF membranes has been reported lately [3,4]
, whereas little information is available about the in
fluence of membrane cleaning on the membrane's performance. In this study, four commercially available membranes have been investigated concerning their fouling and cleaning behavior during NF of a surface water (River Alb). Rejection of four polar micropollutants was monitored regularly over the experimental period in order to further evaluate the in fluence of the feed matrix, fouling and cleaning on membrane performance. 2. Experimental 2.1. Filtration experiments Nano filtration of a south German surface water (River Alb, 150 kDa pre filtration) was conducted in a bench scale crossflow plant with eight modules connected in parallel under the following operating conditions: pressure p=10 bar, temperature T=20±0.2 °C, cross flow velocity v CF =0.2 ms
−1 . The raw water shows a dissolved organic carbon (DOC) concentration of 2.5 mg L −1 and a low hardness (3.3 °dH). No antiscalants were added as according to calculations with the software ROSA (DOW) scaling could be excluded for the filtration of Alb water with NF 270 at the given operating conditions. Due to the calculations mainly organic fouling was expected. Therefore, alkaline conditions were chosen for membrane cleaning. Three nano filtration membranes and one low pressure reverse osmosis membrane (CG) of similar molecular weight cut off (MWCO, Table 1
) have been investigated in duplicates. The membranes were rinsed with deionized (DI) water for 36 h at the given operating conditions before starting the experiment. It has to be noted that the transmembrane pressure applied during the experiments was chosen with the aim to create membrane fouling in a reasonable short time and that the resulting initial clean water flux was outside the range recommended for continuous operation of NF 270. Hence, fouling layer formation could be expected particularly for this membrane. The raw water was spiked with four polar micropollutants ( Table 2 ) at
concentrations of 100 µg/L each in order to compare the in fluence of feed matrix, fouling and cleaning on the rejection of the selected substances. Alb water was filtered for 570h with two alkaline cleaning steps (1 h, NaOH, pH 11 for DK, DL and NF 270 and pH 8 for CG) after 190 and 480 operating hours, respectively. One 480 h filtration experiment with alkaline cleaning after 312 h was conducted with DI water spiked with the micropollutants (100 µg L − 1 each).
2.2. Analysis Membranes and fouling layers were characterized by means of scanning electron microscopy (SEM), energy dispersive X-ray spec- troscopy (EDX, both LEO 1530 Gemini), attenuated total re flectance Desalination 250 (2010) 1005 –1007 ☆ Presented at the 12th Aachener Membrane Kolloquium, Aachen, Germany, 29-30 October, 2008. ⁎ Corresponding author. E-mail address: angela.kluepfel@ebi-wasser.uni-karlsruhe.de (A.M. Klüpfel). 0011-9164/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:
10.1016/j.desal.2009.09.091 Contents lists available at ScienceDirect Desalination j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / d e s a l
Fourier transform infrared spectroscopy (ATR-FTIR, Bruker Equinox®, IFS 55) and contact angle measurements. Liquid chromatography – tandem mass spectrometry (LC –MS/MS, HP 1100 chromatography system, Agilent) was used for micropollutant analysis. DOC was mea- sured with a TOC-Analyzer 820 (Sievers Instruments). 3. Results and discussion 3.1. Fouling and cleaning results The given operating conditions resulted in an initial clean water flux of 24 (CG), 30 (DK), 42 (DL) and 135 Lm − 2
h − 1
(NF 270), respectively, after 36 h rinsing in DI water. CG, DK and DL did not show signi ficant changes in permeability during the filtration and cleaning experiments except for a temporary rise after the second cleaning step for DK and a slight increase after cleaning for DL followed by a slow decline to its initial value of permeability. Due to its high initial flux, permeability decreased strongly for NF 270 and was not recovered during the cleaning cycles. In SEM, contact angle (not shown) and ATR-FTIR measurements ( Fig. 1
), fouling layers of dif- ferent thickness were detected for all investigated membranes. As neither the filtration nor the cleaning periods showed significant in fluence on the permeate flux of DK, DL and CG, the resistance of the fouling layer is assumed to be negligible compared to the membrane resistance in these cases. Fig. 1 shows ATR-FTIR-Spectra of the membranes CG and NF 270 in virgin state, after filtration and after cleaning. In ATR-FTIR measurements the IR beam penetrates the sample approx. 1 µm. Hence, the spectra of virgin nano filtration membranes typically contain information about both the active layer and the membrane support [6]
. Correspondingly, ATR-FTIR spectra of fouled membranes can be assumed to consist of both membrane and fouling bands. The membrane CG (cellulose acetate, Fig. 1
a) shows char- acteristic bands at 1750 cm − 1 (carbonyl groups), 1350 –1400 cm − 1
(methyl groups) und 1220 –1225 cm
− 1 (carboxyl groups). After fil- tration of Alb water and alkaline cleaning, the spectra are consistent with the virgin membrane spectrum to a great extend, probably due to the thin fouling layer on the CG membrane. The spectrum of NF 270 (polyamide, Fig. 1
b) on the other hand, shows evidences of peptide bonds which are characteristic for polyamide (1650 cm − 1 : amide
group I, NH; 1550 cm − 1
: amide group II, NH und CN). According to Boussu et al. [6] , further bands (1100, 1000 and 720 cm − 1 ) belong to the polyester support of the membrane. In the spectra after filtration and cleaning, the membrane bands seem to be covered by broad bands probably belonging to the comparatively thick fouling layer. Likewise, comparing ATR-FTIR spectra of new and fouled membranes Xu et al. [3] observed the sharp membrane bands to be covered by broad absorption peaks belonging to the fouling layer. Cho et al. [7]
reported the reduction of all the membrane bands in ATR-FTIR spectra after
filtration of groundwater. Possible reasons for the alterations that were observed in the ATR-FTIR spectra of NF 270 after fouling are 1) the IR beam is not able to penetrate the fouling layer and hence to excite vibrations of the membrane material and 2) the sharp signals of the pure membrane are shifted and broadened due to interaction of fouling layer molecules with the membrane material. For the four investigated membranes, the intensity of the fouling bands in comparison to the membrane bands could be correlated with the thickness of the fouling layers (as observed in SEM) as well as with the initial clean water flux of the virgin membranes. It was found to increase in the order CG bDKbDLbNF 270. The spectra of the cleaned membranes revealed no signi ficant difference compared to the membranes characterized directly after filtration. Comparison of virgin and fouled membrane EDX spectra showed the presence of Fe, Cu, Zn, Mg, Al, Ca, Si, Cl, K, P and S (possibly as phosphates and sulfates respectively) in the fouling layers in addition to C, O and H. This refers to a contribution of inorganic water con- stituents to fouling layer formation despite their low concentrations in the raw water. 3.2. Rejection of model substances Fig. 2 shows the rejection of four polar trace organic substances in DI and Alb water before and after cleaning. Clo
fibric Acid showed the highest rejection in DI and Alb water for all the investigated membranes compared to Atrazine, Metamitron Table 1 Characteristic parameters of the investigated membranes. Company Material
MWCO a Zeta-potential (pH 6.5) [5]
Contact angle g/mol
mV θ a θ r NF 270 DOW Polyamide 175 −19.7
28.5° 14.8°
DK GE Polyamide 170 −5.9
40.0° 18.1°
DL GE Polyamide 325 −9.4
51.1° 28.0°
CG GE Cellulose acetate 170 −4.9
28.9° 12.9°
a Determined by filtration of sugars and polyethylenglycols at 25 °C and 10 bar. Table 2
Characteristic properties of the investigated model substances. M pK a Charge (pH 6.5) log K OW
g mol − 1
Metamitron C 10 H 10 N 4 O 202 1.7 Neutral
1.44 Clo
fibric acid C 10 H 11 ClO 3 215
2.2 Negative
2.84 Atrazine
C 8 H 14 ClN
5 216
2.4 Neutral
2.82 Terbutaline C 12
19 NO 3 225 9.3, 9.1
Positive 0.62
a Estimated by means of an atom/fragment contribution method ( http://www. syrres.com/esc/est_kowdemo.htm ). Fig. 1. ATR-FTIR-Spectra of the membranes CG (a) and NF 270 (b) virgin, after filtration and after alkaline cleaning. 1006 A.M. Klüpfel, F.H. Frimmel / Desalination 250 (2010) 1005 –1007 and Terbutaline. This is probably due to electrostatic repulsion between the negatively charged membrane surface and the negatively charged dissociated organic acid molecule. Rejection of Clo fibric Acid in Alb water was similar to rejection in DI water for all membranes. Atrazine and Metamitron were rejected at an equal amount in both water matrices, whereas Terbutaline was signi ficantly less rejected in DI water in case of the DL and CG membrane. Attractive forces between the positively charged molecule and the negatively charged membrane surface may lead to a higher passage of Terbutaline in DI water whereas those interactions could be shielded by negatively charged water constituents (Natural organic matter NOM, anions) during
filtration of river water. For the membranes DK and NF 270, size exclusion seems to dominate the charge effects. The RO cellulose acetate membrane CG with a MWCO similar to DK and NF 270 ( Table 1 ) generally showed a comparatively low rejection, especially for Terbutaline, Atrazine and Metamitron. During the filtration periods no significant change in rejection of the organic substances was found, although Metamitron showed a decreasing trend after cleaning during filtration with the membrane NF 270. It can be deduced that fouling had little in fluence on the rejection of the investigated molecules in these experiments. In case of the DI water experiments, treatment of the membranes with NaOH solution considerably lowered the rejection of the in- vestigated micropollutants except for Atrazine rejected by CG. Restoration of the membrane performance was achieved after several hours. The low rejections directly after alkaline treatment can be explained by electrostatic repulsion between deprotonated acidic membrane groups and a consequential pore expansion or “membrane swelling ” which was reported by several authors at higher pH values [8] . During subsequent filtration of unbuffered DI water the depro- tonation of the functional groups seems to be reversed only after several hours. However, in the experiments with Alb water the effect of alkaline treatment on the rejection of the trace organic substances decreased to an insigni ficant amount already 1 h after membrane cleaning. Hence, restoration of the original membrane properties could be achieved in less than 1 h probably due to buffer action of the river water. 4. Conclusions The four membranes showed a quite different fouling and re- jection behavior during the filtration experiments: □ Due to its high initial flux, fouling formation was strongest for NF 270. For DK, DL and CG, flux did not change to a great extend during filtration and cleaning. Thus, membrane resistance pre- vailed fouling layer resistance in case of these membranes. □ Rejection could be related to MWCO in case of DL, DK and NF 270, but not for CG. DK and NF 270 turned out to remove the micro- pollutants most ef ficiently. Fouling was detected on all the membranes by means of SEM and ATR-FTIR. However, under the given experimental conditions, fouling did not affect the rejection of the investigated model substances signi ficantly. The feed matrix turned out to determine the influence of an alkaline cleaning step on micropollutant rejection. Acknowledgements The authors thank the Water chemical Society, division of the German Chemical Society for a PhD scholarship and DVGW for financial support. Luciana E. Martínez is acknowledged for her con- tribution to the experimental work. References [1] Q. Li, M. Elimelech, Environ. Sci. Technol. 38 (2004) 4683 –4693. [2] A. Al-Amoudi, R.W. Lovitt, J. Membr. Sci. 303 (2007) 4 –28. [3] P. Xu, J. Drewes, T. Kim, C. Bellona, G. Amy, E. Schulze, J. Membr. Sci. 279 (2006) 165 –175.
[4] K.V. Plakas, A.J. Karabelas, T. Wintgens, T. Melin, J. Membr. Sci. 284 (2006) 291 –300.
[5] D. Norberg, S. Hong, J. Taylor, Yu. Zhao, Desalination 202 (2007) 45 –52.
[6] K. Boussu, J. De Baeremaeker, C. Dauwe, M. Weber, K.G. Lynn, D. Depla, S. Aldea, I.F.J. Vankelecom, C. Vandecasteele, B. Van der Bruggen, ChemPhysChem 8 (2007) 370 –379.
[7] J. Cho, G. Amy, J. Pellegrino, Y. Yoon, Desalination 118 (1998) 101 –108.
[8] C. Bellona, J. Drewes, J. Membr. Sci. 249 (2005) 227 –234.
Fig. 2. Rejection of a) Terbutaline, b) Atrazine, c) Clo fibric Acid, d) Metamitron in DI and Alb water. □ : filtration before cleaning (n=7), : sampling 2 h after cleaning (n = 1), : filtration after cleaning (n=7). 1007 A.M. Klüpfel, F.H. Frimmel / Desalination 250 (2010) 1005 –1007 Document Outline
Yüklə 270,01 Kb. Dostları ilə paylaş:
|