вход по аккаунту


Photooxidative Pretreatment to Improve Sustainable Operation of the Microfiltration of Drinking Water.

код для вставкиСкачать
Dev. Chem. Eng. Mineral Process. 14(1/2), pp. 219-226, 2006.
Photooxidative Pretreatment to Improve
Sustainable Operation of the
Microfiltration of Drinking Water
F. Malek, J. Harris and F. Roddick"
School of Civil and Chemical Engineering, W I T University,
GPO Box 2476K Melbourne 3001, Victoria, Australia
and Cooperative Centrefor Water Quality and Treatment, Salisbury,
South Australia 51 08,Australia
Membrane fouling by natural organic matter (NOM) is a major factor limiting the
sustainability of microfiltration technologies for potable water treatment as it
results in decreased permeate flux, reduced production, more frequent cleaning
with associated increased chemical usage, and shorter membrane life. This study
investigated the use of vacuum ultraviolet ( V W 254 + 185 nm) photooxidation as
a pretreatment strategy to reduce flux loss and membrane foufing in potable water
treatment. Raw water was exposed to VUV treatment and microfiltered in a
dead-end stirred cell using 0.22 p m hydrophobic polyvinylidene fluoride (PVDF)
membranes. Considerable flux improvement was obtained with increasing VUV
treatment time. However, attempts to clean the fouled membranes by backwashing
with Mini-Q water were not successful, and slightly entrenched the fouling in most
instances. Various chemical cleaning regimes for membrane regeneration were
tested, both individually and in combination. A I.Owt% solution of the enzymic
detergent Terg-a-zyme almost completely cleaned the membranes. A cheaper, but
slightly less efective, alternative involved the sequence of sodium
hydroxide/hydrochloric acidhodium hypochlorite which removed 86% of the
fouling resistance.
The application of membrane filtration technologies is expanding rapidly in the
treatment of drinking water for the removal of pathogens and turbidity. The critical
factor limiting the sustainable use of membrane systems is membrane fouling which
results in deterioration of the membrane performance, in particular the water flux,
increased frequency of backwashing, reduced time between chemical cleans, and
increased chemical usage, leading to shorter membrane life. Fouling is caused by a
* Author for correspondence (felicipso~~licX.(li),
F. Malek, J. Harris and F. Roddick
combination of the particulates and the coloured natural organic matter (WOM) in
the raw water. With ultrafiltration and nanofiltration, fouling occurs mainly by the
formation of a cake layer of colloidallparticulatematerial on the membrane surface,
plus adsorption of dissolved organics within the membrane pore structure [ 11. With
the larger pores and the more open structure of microfiltration membranes, internal
adsorption is the initial contributor to fouling, followed by surface deposits when
larger volumes are processed. High molecular weight (HMW) colloidal organics
(>30 kDa) such as polysaccharides have been shown to be the main contributors to
microfiltration membrane fouling due to plugging within the membrane pores. The
smaller neutral macromolecules (4kDa) also cause flux decline by adsorption
within the membrane structure, and by electrostatic interaction between the
polysaccharide-Ca complexes and the negatively charged membrane [2,31. The
development of a pretreatment process to prevent flux decline needs to be directed
at removing both the colloidal material and smaller neutral macromolecules.
Advanced oxidation processes such as UV photooxidation have been shown to
cause degradation of the high molecular weight (HMW) organics in water to low
molecular weight (LMW) organics [4].The removal of the dissolved organic
carbon (DOC) component of NOM by photooxidation happens via a complex
sequence of photochemical reactions between the NOM and reactive species and
rapidly reacting species formed by photooxidation to eventually form carbon
dioxide, water and inorganics [5]. VUV (254 nm + 185 nm) photooxidation was
recently shown to outperform UVC (254 nm) photooxidation in reducing DOC
levels in water [ 6 ] .Degradation by UVC photooxidation was primarily attributed to
absorption of radiation at 254 nm by UV-absorbing organics such as aromatics and
unsaturated compounds [7]. With WV irradiation, degradation of both saturated
and unsaturated compounds is primarily due to the presence of highly reactive
hydroxyl radicals generated via photolysis of water at 185 nm [6].
A preliminary investigation in our laboratories indicated that W V pretreatment
of water improved flux for the microfiltration of river water using 0.22 pm PVDF
hydrophobic membrane [8]. The aim of the current study was to further investigate
the application of VUV irradiation as a pretreatment method for flux improvement
in microfiltration systems, to examine the hydraulic cleaning of the membranes by
backwashing, and to evaluate the most effective chemical cleaning regime.
Experimental Design
The water was sourced from Sandhurst reservoir, Bendigo, Victoria on 10July 2003
and had the characteristics shown in Table 1. The water was filtered through a
0.45pm hydrophilic surface-modified PVDF membrane (Millipore) to remove
particulate matter before exposure to W V irradiation and membrane runs.
Table I . Characteristics of Bendigo water.
Photooxidative Pretreatment to Improve the Microfiltration of Drinking Water
Irradiation was camed out using an annular reactor with a centrally mounted
lamp. The working volume of the reactor was 0.9 litres, and the average irradiated
area was 463.8 cm2 with a path length of 1.94 cm. A low-pressure mercury vapour
W V lamp, referred to as H Lamp (Australian Ultraviolet Services, G36T15HU)
was used with an output at 185 + 254 nm and an electrical energy requirement of
46 W. The light intensities at 185 nm and 254 nm were determined by methanol
actinometry and hydrogen peroxide actinometry; the fluence rates were 2.85 and
21.96 mT cm-’ s-l, respectively. The total fluence rate was 24.81 mJ cm” s.’ for the
H lamp, with a corresponding energy input rate of 766 J/L water/min W V
exposure time. During the irradiation process, the sample inside the UV reactor was
mixed and gently aerated by humidified air. Nitrogen gas was circulated past the
lamp to maintain a constant temperature of 28*1”C. Water samples were irradiated
for times ranging from 10 to 90 minutes and then subjected to microfiltration.
Filtration experiments were conducted using a dead-end stirred cell system with
membrane area of 13.4 x 10 m (Amicon 8050). The stirred cell was connected to
the feed water reservoir and operated at a transmembrane pressure (TMP) of 50 kPa
using nitrogen gas, stirrer speed of 300 rpm, and room temperature (=2OoC). Flux
was determined by weighing the permeate on a top-loading balance at 30 second
intervals. The membrane used in the filtration experiments was 0.22 pm
hydrophobic polyvinylidene fluoride (GVHP, Millipore) which was wetted in
methanol for 2 hours then soaked in Milli-Q water for 2 hours prior to use. For each
run, Milli-Q water (500 ml) was passed through a fresh membrane and the pure
water flux (Jo)was measured. This was followed by the water sample, either raw
water or WV-treated water, and the permeate flux measured throughout the run.
The quasi-steady-state value was taken as the final value, Jf . The fouled membrane
was surface-washed, backwashed with Milli-Q water, re-inserted in the stirred cell,
and the backwashed flux measured for Milli-Q water (Jbw).
Chemical agents tested for cleaning NOM-fouled membranes included caustic
soda, hydrochloric acid, sodium hypochlorite, ethylenediamine tetra acetic acid
(EDTA), citric acid, cetyl trimethyl-ammonium bromide (CTAB), and an enzymic
detergent Terg-a-zyme (Alconox Inc., New York). The effectiveness of the cleaning
procedures were compared with a simulation of the procedure used at the Bendigo
Water Treatment Plant. This involved 20 minutes soaking in H2S04/O. 1% EDTA at
pH 1.9; 15 minutes rinse with Milli-Q water; 20 minutes cleaning with NaOH/O. 1%
EDTA/O.S% H 2 0 2 at pH 11; followed by a final 15 minutes rinse with Milli-Q
water. The permeate flux with Milli-Q water was measured after backwashing (Jhw)
and after chemical cleaning (Jc). Water flux recovery, JR (%) was defined as:
J, =- J c - J / . 100
J,, - J ,
DOC values were determined using a TOC analyzer (Sievers, Model 820).
Absorbance at 254 nm (A254)was measured using a double beam scanning
UVivis spectrophotometer with 1 cm path length (Unicam, Model UV2).
22 I
F. Malek, J. Harris and F.Roddick
Results and Discussion
DOC Removal by Irradiation
W V irradiation produced substantial removal of DOC from the Bendigo water as
shown in Figure 1. The rate of removal was faster for the initial exposure, 1.O ppml
10 min, and slower between 80-90 minutes treatment, 0.55 p p d l 0 min. In relative
terms, the rate of removal was almost constant at 13% remaining DOC removed
10 min exposure. In terms of W V energy input, 13% DOC removal was achieved
for a dosage of 14.9 J cm" or total energy input of 7.66 kJ/L. For the maximum
treatment time of 90 minutes, a dosage of 134 J cmS2(69 kJ/L input energy)
produced 65% removal of the original DOC. The AZs4value was also substantially
reduced by W V treatment. Initially the A254declined rapidly with 40% reduction
in 10 minutes, and slower reduction on further irradiation as shown in Figure 2.
These data suggest the removal of conjugated bonds with the oxidation of the
HM W chromophores to slow reacting chromophores during W V irradiation,
which is consistent with a recent hypothesis [ 6 ] .
Effect of VUV Irradiation on Permeate Flux
W V photooxidation of Bendigo water resulted in significant flux improvements as
shown in Figure 3. Progressively improved fluxes were obtained with increasing
exposure time. A 5.8-times improvement in the quasi-steady-state flux was noted
for the filtration of the 90 minutes WV-treated feed. This represented 61% of the
initial flux. Increased membrane throughput with increasing W V exposure time
coincided with the reduction in DOC and AlS4 values. This indicated that the
increased flux was due to the conversion of unsaturated and saturated HMW
colloidal NOM molecules to LMW organics by the W V treatment.
VUV exposure t h e (mln)
Figure 1. Change in DOC f o r VUV-treated Bendigo water.
Photooxidative Pretreatment to improve the Microfiltration of Drinking Water
+Absorbance after W V irradiation
Treatment tJme (min)
Figure 2. Change in
of VUV-treated Bendigo water.
dosage (LJIL),
time (mln)
11.5 (15 mln)
-c 23.0 ( 3 0 mln)
E 1500
3 1000
34.5 ( 4 5 mln)
46 0 (60 min)
57.5 (75 mlnl
69.0 ( 9 0 mln)
Filtration time (sec)
Figure 3. Flux profile for VUV-treated Bendigo water using 0.22 p m hydrophobic
Backwashing the Membrane
Hydraulic cleaning of the fouled membranes by backwashing with Milli-Q water
produced a small flux recovery for the raw-water fouled membrane, but had a
negative effect for the irradiated-water fouled membranes. That is, the pure-water
flux after backwashing was lower than that of the non-cleaned membrane as shown
in Figure 4. Some irreversible fouling was created as a result of the backwashing
process. This unusual phenomenon was attributed to the NOM compounds and low
molecular weight organics from the irradiation process, which were adsorbed in the
membranes, being compacted by the internal backwash flow. Regeneration of the
membranes by chemical cleaning was then investigated.
F.Malek, J. Harris and F.Roddick
Figure 4. Eflect of backwashing on frux of membranes fouled with VUV-treated
Chemical Cleaning
Chemical cleaning experiments were conducted on membranes fouled with 20-min
WV-treated water (dosage 29.8 or energy input 15.3 MA.,).As single
cleaning agents, both sodium hydroxide (used for removal of organic foulants) and
HCl (for removal of inorganic foulants such as Ca2') were ineffective. Flux
recovery was only 4% for the optimal 0.005N NaOH concentration, and zero with
0.005NHCI. The proteolytic detergent Terg-a-zyme was the most effective single
cleaning agent. Used at the recommended concentration for general cleaning of 1
wt%, Terg-a-zyme gave a flux recovery of 77% with the corresponding fouling
resistance removal being 95%. The cleaned membrane had a white, non-stained
appearance. At the lower, more economical, concentration of 0.1 wt% Terg-a-zyme,
the flux recovery was 59%. With the other cleaning chemicals tested, the two best
results were 24% flux recovery with 0.5wt% NaOCl and 21% with 0.02%)CTAB.
A sequence of O.OOSN NaOH/O.OOSN HCl was more effective than when these
chemicals were used individually, yielding 3 1% flux recovery (see Figure 5 ) . When
this was followed by a final step of 0.5% NaOC1, the flux recovery reached 53%
with 86% removal of fouling resistance. The simulated plant cleaning procedure
was not very effective with 14% flux recovery.
Photooxidative Pretreatment to Improve the Microfiltration of Drinking Water
Figure 5. Cleaning effectiveness of NaOH/HCI sequences for restoringflux.
Photooxidation using vacuum ultraviolet ( W V , 245 plus 185 nm) radiation was
effective for reducing the DOC level and the
value of the Bendigo water
sample. Due to this decrease in DOC and the molecular size as indicated by A2s4,
considerable flux improvement was observed with 0.22 pm PVDF microfiltration
membrane. An adverse effect was that backwashing with Milli-Q water caused a
decrease in flux recovery. Thls was attributed to the breakdown of the
macromolecules in the NOM to form lower MW compounds which adsorbed within
the membrane pores, and compacted during the backwashmg procedure. An
investigation of possible cleaning agents revealed that 1.Owt% solution of the
e n z p c detergent Terg-a-zyme almost completely cleaned the membranes. A
cheaper alternative, although slightly less effective, involved the sequence of
sodium hydroxidehydrochloric acidlsodium hypochlorite which removed 86% of
the fouling resistance. In terms of overall performance, W V photooxidation as a
pre-treatment step in the microfiltration of W i n g water improves water flux,
extends the time between backwashes, reduces the frequency of chemical cleaning,
reduces the quantity of sludge by-products, and thus increases the sustainability of
the process. Fouled membranes only required cleaning with a dilute
alkaliiacidlhypochlorite sequence in order to restore them for continued application
and increased membrane life.
F. Malek, J, Harris and F. Roddick
Acknow Iedgment s
The authors wish to acknowledge the CRC for Water Quality and Treatment for
funding this project.
I . Cho, J., Amy. G., and Pellegnno, J . 1999. Membrane filtratlon of natural organic matter: Initial
comparison of rejection and flux decline characteristics with ultrafiltration and rianofiltration
membranes, Waler Research, 33( 1 I ), 2517-2526.
2 . Fan, L., Harris, J.L. Roddick, F.A., and Booker, N.A. 2001. Influence of the characteristics of natural
organic matter on the fouling of microfiltration membranes. Water Resenrch, 35( l8), 44554463.
3. Fan, L., Hams, J.L., Roddick, F.A., and Booker, N.A. 2002. Fouling of microfiltration membranes
by the fractional components of natural organic matter in surface water. Water Science and
Technology: WaferSupply, 2(5-6), 3 13-320.
4. Zepp, R.G. 1988. Environmental photo processes involving natural organic matter, humic substances
and their role in the environment, John Wiley, New York, 190-215.
5. Frimmel, F.H. 1998. Impact of light on the properties of aquatic natural organic matter.
Environmenlnl International, 24(5-6), 559-571.
6. Thomson, J., Parkinson, A., and Roddick, F. 2004. Depolymerization of chromophoric natural
organic matter. Environ. Sci. Technol.. 38(l2), 3360-3369.
7. Parkinson, A., Barry, M.J, Roddick, F.A., and Hobday, M. 2001. Preliminary toxicity assessment of
water after treatment with UV-irradiation and W/H202. WaferResearch, 35,3656-3664.
8. Parkinson, A., Madden, C., Nguyen, N., and Roddick, F. 2003. Investigation of the use of oxidative
and biological pretreatments to reduce flux loss in microfiltration membranes. Proceedings of
Chemeca 2003 Conference, Adelaide, South Australia, 28 September 1 October, paper no. 182.
Без категории
Размер файла
570 Кб
sustainable, pretreatment, water, drinking, operation, photooxidative, improve, microfiltration
Пожаловаться на содержимое документа