Download Treating Laundry Waste Water Cationic Polymers for Removal of Contaminants and Decreased Fouling in Microfiltra...
Journal of Membrane Science 456 (2014) 167–174
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Treating laundry waste water: Cationic polymers for removal of contaminants and decreased fouling in microfiltration Hyun-Chul Kim a, Xia Shang a, Jin-Hui Huang b, Brian A. Dempsey a,n a b
Department of Civil and Environmental Engineering, Pennsylvania State University, University Park, PA 16802, USA College of Environmental Science and Engineering, Hunan University, China Changsha 410082, PR China
art ic l e i nf o
a b s t r a c t
Article history: Received 25 November 2013 Received in revised form 10 January 2014 Accepted 11 January 2014 Available online 21 January 2014
The goal was to select coagulants and a coagulation process for treating laundry wastewater. The longterm goal is for application in Army mobile treatment units with solids removal using microfiltration (MF) and for which a robust operation with small coagulant volumes are desirable. Laundry wastewater usually has very high pH thus strong base cationic polymers are good coagulant candidates. Seven quaternary amine polymers were examined to determine effects of coagulant dose on zeta potential (ZP). Four of the polymers were further evaluated for sedimentation of contaminants, specific resistance to filtration, and cake compressibility during filtration. A low molecular-weight epichlorohydrin/dimethylamine (epi/DMA) polymer was tested further because of greatest increase in ZP with low polymer dose, lowest specific resistance to filtration, and good removal of contaminants. Flocculation for 10 min resulted in greatly improved removal of cake by hydraulic washing compared to 2 and 5 min flocculation. Fouling during multi-cycle membrane operation was greatly reduced with coagulant additions less than half the charge-neutralization (CN) dose. Low polymer dose results in decreased chemical demand and reduced sludge production. Successful treatment using from 50% to 100% of the CN dose provides more robust operation under field conditions. & 2014 Elsevier B.V. All rights reserved.
Keywords: Laundry lint Wastewater reuse Cationic polymer Coagulation Microfiltration Fouling
1. Introduction This paper deals with coagulation of laundry wastewater with quaternary amine polymers and solids removal by sedimentation or low-pressure membrane filtration. The research is motivated by the need to recycle laundry wastewater at Army forward operating bases using the Tricon Shower Water Reuse System (SWRS) in which microfiltration (MF) of laundry wastewater without prior coagulation can result rapid membrane fouling. The results will be useful for treatment of laundry wastewater in any venue. The most widely employed strategies for treatment of laundry wastewater are coagulation, adsorption, flotation, adsorption, filtration with media, and membrane filtration [1–8]. Coagulation can be effective for removal of chemical oxygen demand (COD), phosphates, and anionic surfactants. Coagulation produces floc that must be removed by a solid separation process. MF or ultrafiltration (UF) membranes are increasingly used for solid separation and
n Correspondence to: Department of Civil & Environmental Engineering, Pennsylvania State University, 212 Sackett Building, University Park, PA 16802-1408, USA. Tel.: þ 1 8148651226; fax: þ 1 8148637304. E-mail addresses:
[email protected] (H.-C. Kim),
[email protected] (X. Shang),
[email protected] (J.-H. Huang),
[email protected] (B.A. Dempsey).
0376-7388/$ - see front matter & 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2014.01.028
these processes are compatible with in-line application of coagulants [9]. Inorganic coagulants are typically effective only when a precipitate forms, usually with a net charge opposite to the charge on the contaminants. For example, alum or ferric chloride work best when the final pH results in precipitation of the metal oxides and when the initially formed precipitates have sufficient positive charge to neutralize negative charge on humic materials, clays, and other common constituents in water. Laundry wastewater, however, typically has very high pH and is well-buffered with respect to pH. Ge and coworkers [7] reported that contaminant removals from laundry wastewater were poor when using inorganic coagulants for pH o 4 or 49. Similarly, electrocoagulation did not improve flux compared with raw water in polyvinylidine fluoride (PVDF) MF [10]. Strong base cationic polymers retain a high positive charge at high pH values are consequently are good candidates for coagulation of laundry wastewater. The goal was to identify coagulants and coagulation processes to achieve the following objectives during treatment of laundry wastewater at high pH: (1) remove contaminants especially total suspended solids (TSS), total chemical oxygen demand (TCOD), and total phosphorus (TP); (2) effective removal of coagulated solids by sedimentation or MF; (3) substantial decrease in MF fouling compared to raw laundry wastewater; and (4) effective operation over a wide range of coagulant doses to achieve simple
168
H.-C. Kim et al. / Journal of Membrane Science 456 (2014) 167–174
and robust operation. Experiments were conducted in three phases: (1) seven strong-base cationic polymers were screened to determine charge-neutralization (CN) effectiveness at pH 11; (2) four cationic polymers were selected for evaluation of contaminant removal by sedimentation, specific resistance to filtration, and cake compressibility; (3) one cationic polymer was selected for additional tests including effect of flocculation time and chemical dose on MF performance during multi-cycle tests meant to simulate continuous operation of the SWRS.
2. Materials and methods 2.1. Laundry wastewater Wastewater was collected from an industrial laundry in Centre County PA during hours when only commercial uniforms were washed. Temperature and pH were measured onsite. The sample was immediately transported to the laboratory, filtered with 10 μm polypropylene mesh filters, pH adjusted to 11, and stored in a walk-in-refrigerator for further experiments. Table 1 shows the wastewater characteristics. 2.2. Coagulant titrations and jar tests Seven quaternary amine (strong base) cationic polymers were initially selected to provide a range of apparent molecular weight, composition, and form (water solution, emulsion, and dispersion). Table 1 The characteristics of laundry wastewater samples collected on six different dates from a local industrial laundry. Parameter
Value
pH Temperature (1C) Zeta potential (mV) Conductivity (μS cm 1) Total dissolved solids (mg L 1) Turbidity (NTU) TSS (mg L 1) TCOD (mg L 1) TP (mg P L 1)
12.5 7 0.5 407 1.0 57.4 7 8.5 7247 123 3577 52 858 7 111 3597 82 11387 58 227 4
All polymers were diluted according to manufacturer's specifications with deionized water to 1% (v/v) just before use. The 1% solutions were typically stored at room temperature. Prior to some experiments the 1% solutions were stored at 4 1C or 40 1C. All polymer doses are expressed in ppm (v/v). Characteristics of the polymers are shown in Table 2. Coagulant titration tests were conducted by adding increasing volume of polymer beneath the water surface in the vortex of a rapidly stirred wastewater sample and measuring pH and zeta potential (ZP). Polymer doses were increased until several positive ZP readings had been recorded. Zero ZP is an indicator for the CN condition, i.e., the negative charge on contaminants has been exactly titrated by positive charge from the coagulant. Underdosing (UD) refers to coagulant additions less than the CN dose resulting in residual negative charge on the coagulated flocs, while over-dosing (OD) means coagulant addition greater than the CN condition resulting in net positive charge on the flocs. Four significantly different polymers were selected based on the titration tests for jar tests and specific resistance to filtration tests. Jar tests were performed using the four selected polymers to measure removals of turbidity, TSS, TCOD, and TP after sedimentation. Total means dissolved plus suspended contaminants after 1 hr sedimentation. Tests were conducted with a Phipps & Bird stirrer with conventional blades (Model 7790-400) by adding selected volumes of 1% solutions of polymer into wastewater with 1 min high speed mixing and then 30 min mixing at a velocity gradient of 200 s 1. Samples were collected at the end of the mixing for ZP measurement. Supernatant was collected for the other analyses from just below the water surface after 1 h quiescent settling. Most jar tests were conducted at room temperature (E22 1C) but some tests were conducted in water baths at 40 1C.
2.3. Effects of coagulants on specific resistance to filtration and compressibility Flat-sheet hydrophobic 0.22 μm PVDF MF membranes (GVHP04700, Millipore) were used. Membranes were wetted in methanol and then soaked in deionized water overnight. A deadend filtration system was used to estimate cake characteristics (Fig. 1a). Head loss from velocity head (v2/2g) and elevation (Δh) were negligible, where v is the fluid velocity (m s 1) and g is gravitational acceleration (9.8 m s 2). The pre-wetted membrane was placed in a commercial filtration cell (polycarbonate filter holder, Pall Sciences) with 9.6 cm2
Removals of turbidity, TSS, TCOD, and TP by prefiltration with 10 μm polypropylene mesh filters were 177 6, 347 3, 147 6, and 2 71%, respectively.
Table 2 The characteristics of polymers (in alphabetic order) provided by manufacturer. Polymer brand
Cat-floc 8102 plus Cat-floc 8108 plus Core shell 71301 Core shell 71303 Core shell 71305 Nalcolyte 8105 Ultimer 1460
Ionicity
Cationic Cationic Cationic Cationic Cationic Cationic Cationic
Charge density (%)a
(meq g 1)b
100 100 50–80 20–50 1–30 100 50–80
6.2 6.2 3.8–4.7 2.1–3.8 0.1–2.8 7.3 4.3–5.6
Molecular weight
Composition
Form
Medium High High Very high Very high Low ( o50 K) High
PolyDADMAC PolyDADMAC AcAm/DMAEA.MCQ AcAm/DMAEA.MCQ AcAm/DMAEA.MCQ Epi/DMA AcAm/DADMAC
Water solution Water solution Emulsion Emulsion Emulsion Water solution Dispersion
(Poly)DADMAC¼ (poly)diallyldimethyl-ammonium chloride (C8H16NCl). AcAm¼ acrylamide (C3H5ON). DMAEA.MCQ¼ dimethylaminoethylacrylate methyl chloride salt (C8H16O2NCl). Epi/DMA ¼epichlorohydrin and dimethylamine (C5H12ONCl). a b
Mole percent of charged groups (e.g., the mole% of DMAEA.MCQ in AcAm/DMAEA.MCQ). The amount of cationic charge per gram of polymer was calculated with the mole percent of charged groups.
H.-C. Kim et al. / Journal of Membrane Science 456 (2014) 167–174
Pressure regulator
The coefficient of compressibility (n) varies from zero for an incompressible layer to greater than 1 for a highly compressible cake and was calculated by regression analysis using the following equation:
Pressure gauges
αc ¼ αo ΔP nc Δh
Nitrogen gas
Pressure vessel
Membrane module
Nalcolyte 8105 was selected for further membrane filtration experiments based on best values of αc and n. The desired polymer dose was added to 2 L wastewater while agitating at a velocity gradient of 425 s 1 for 10 min. Dead-end constant pressure experiments were performed using only CN coagulation conditions. The wastewater was filtered through 0.22 μm PVDF MF membranes under N2 at constant feed pressure of 10 psi for 1 h. Permeate flux and trans-membrane pressure (TMP) values were continuously recorded. Recirculated-flow constant flux experiments were performed using UD, CN, and OD coagulation conditions. The coagulated wastewater was pumped to the membrane filtration module through a feed-side peristaltic pump. Multi-cycle filtration runs were conducted using eight sequential cycles of filtration (15 min), backwash (1 min at 10 psi N2), and surface flush (0.5 min). Retentate was recirculated to the supply vessel. Less than 40% of initial sample volume was filtered in any experiment. Permeate flux (by permeate mass) and TMP were continuously recorded. The MF membrane was removed at the end of each membrane filtration test, rinsed with deionized water to remove the cake layer, and replaced in the filter assembly for measurement of pure water flux. Subsequent chemical cleanings were performed by soaking the membrane sheets in 0.1 M NaOH and then 0.1 M citric acid for 12 h per each. A pure water flux test was conducted to determine hydraulic resistance and flux recovery of the membrane sheets after each cleaning step.
Digital balance
6
5
1 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
2
3
7
4
Feed tank Feed pump Dampener Membrane module Pressure control valve Digital pressure gauge Permeate pump Digital balance Pressure vessel for backwash Nitrogen gas cylinder Precision pressure regulator
8 11
6
DI water
9
10
2.5. Analytical methods
Fig. 1. (a) Schematic for dead-end filtration to determine specific resistance to filtration, coefficient of compressibility, and effect of flocculation time. (b) Schematic for a multi-cycle recirculating-flow microfiltration tests.
effective filtration area. Raw or coagulated wastewater from a pressure vessel (Model 720340, Advantec MFS Inc., CA) was passed through the membrane under constant pressure (2 psi) for 25 min while recording the filtered mass. After 25 min the feed pressure was step-wise increased every 10 min while feed and permeate pressures were continuously recorded. The cake mass was based on the wastewater TSS and the permeate volume. Specific resistance to filtration was calculated using data from the separate step-pressure experiments using the following equation [11–13]:
αc ¼ A=ðCVÞ½ΔP=ðμJÞ Rm
ð2Þ
2.4. Dead-end and recirculated-flow membrane filtration tests
Discharge
Cationic polymer (1% Nalcolyte 8105)
169
ð1Þ
where αc is specific cake resistance (m kg 1), A is the effective membrane area (m2), C is the TSS in the wastewater feed (kg m 3), V is cumulative permeate volume (m3), ΔP is total pressure drop (Pa), μ is fluid viscosity calibrated by temperature (kg m 1 s 1), J is the permeate flux (m s 1), and Rm is the intrinsic membrane resistance (m 1) measured using particle-free electrolyte solution. Rm is usually obtained by determining the total pressure drop (ΔP) at the initial water flux (Jo) on the assumption that cake resistance (Rc) is zero when no cake has deposited on the filter surface [13,14]. In this study hydraulic resistance from cake itself as a function of trans-cake pressure was determined using the continuous filtration method.
Particle size distributions (4 1 mm) were measured using micro-flow imaging (DPA 4200 flow microscope, Brightwell Technologies Inc., Canada) on samples taken after rapid mixing at 425 s 1 for 10 min. An aliquot was diluted up to 100 times with pre-filtered (0.2 mm) laundry wastewater, 1 mL of the diluted suspension was introduced to the flow microscope, and the images were analyzed by MFI View Analysis Suite (MVAS) software to compile a database containing count, size, and shape parameters. ZP was measured using a Zen 3600 Zetasizer (Malvern Instruments Ltd., Malvern, UK). A coagulant charge analyzer, CCA3100 (Chemtrac Systems Inc., Georgia), was used to measure streaming current. Turbidity (Hach 2100P), conductivity (Orion, 115A þ ), and pH (Orion, 290A) were measured. TSS was determined according to the Standard Methods for the Examination of Water and Wastewaters [15]. TCOD and TP were measured using Hach 2010 spectrophotometer. The Reactor Digestion Method was used for TCOD. The acid persulfate method was used to digest samples for TP (USEPA Method 365.2).
3. Results and discussion 3.1. Titration of negative charge by cationic polymers Coagulation can occur by means of double layer compression, adsorption of oppositely-charged species resulting in reduced surface potential, or polymer bridging. Adsorption of oppositelycharge species results in a reduction of charge on the particle
H.-C. Kim et al. / Journal of Membrane Science 456 (2014) 167–174
neat product is shipped in water solution. Both polyDADMAC polymers performed well in neutralizing charge and one was
100
Residual turbidity (%)
surfaces so that approaching particles will collide (and hopefully stick) rather than repel one another. ZP is considered the most accurate way to determine when the charge on the particles has been reduced. CN occurs when the ZP is zero. ZP is difficult to apply on-line in real-time situations. Streaming current (SC) is an alternative electro-kinetic measurement that is easily adapted for real-time application and is discussed in a later section of this paper. ZP values for raw and coagulated laundry wastewater were measured as a function of coagulant dose for the seven coagulants. Results are shown in Fig. 2a with coagulant dose expressed in ppm (v/v) and in Fig. 2b with coagulant dose in meq of added cationic charge per liter. Nalcolyte 8105 (epi/DMA) was the most effective coagulant based on volumetric additions while the polyDADMAC coagulants and one core–shell polymer provided the steepest rise in ZP as a function of added cationic charge. Nalcolyte 8105 has the highest cationic charge per liquid volume due to higher charge density (see Table 2) and higher liquid density compared to the other polymers. These factors explain the change in relative position of 8105 in Fig. 2b compared to 2a. Two of the core–shell polymers displayed significantly poorer performance than the other polymers in Fig. 2b. Ester linkages in the core–shell polymers are susceptible to base catalyzed hydrolysis resulting in decreased cationic charge [16]. Laundry wastewater pH of 11 could have partially hydrolyzed some of the ester linkages in the core–shell polymers, resulting in apparently higher cationic demand for CN. Four of the seven polymers were selected for further testing. The selection was based on several criteria including performance in neutralizing negative charge and the desire to retain a diversity in polymeric composition and form. Epi/DMA Nalcolyte 8105 partially neutralized negative charge at the lowest dose and the
40
20
0
100
200
300
400
500
0
100
200
300
400
500
0
100
200
300
400
500
80 60 40
20 0
100
0 -20
Residual TCOD (%)
Zeta potential (mV)
60
100
20
Cat-floc 8102 plus Cat-floc 8108 plus
-40
Core shell 71301 Core shell 71303 Core shell 71305
-60
Nalcolyte 8105 Ultimer 1460
-80
0
100
200
300
400
500
80 60 40
20
600
0
Polymer dose (ppm) 20
100
0 -20
Residual TP (%)
Zeta potential (mV)
80
0
Residual TSS (%)
170
Cat-floc 8102 plus Cat-floc 8108 plus
-40
Core shell 71301 Core shell 71303 Core shell 71305
-60
Nalcolyte 8105
80 60 Cat-floc 8108 plus
40
Core shell 71301 Nalcolyte 8105
20
Ultimer 1460
Ultimer 1460
-80
0
0.2
0.4
0.6
0.8
1
1.2
Polymer dose (meq-cationic charge L -1) Fig. 2. Zeta potential versus polymer addition for the coagulation of laundry wastewater at pH 11 and room temperature: (a) polymer dose expressed in partpermillion (v/v); (b) polymer dose expressed as meq L 1 of cationic charge added.
0
0
100
200
300
400
500
Polymer dose (ppm) Fig. 3. Removals of turbidity, TSS, TCOD, and TP after sedimentation. Polymer dose in ppm (v/v). Flocculation was 30 min at a velocity gradient of 200 s 1 and sedimentation for 1 h.
H.-C. Kim et al. / Journal of Membrane Science 456 (2014) 167–174
1.E+06
80
Particle concentration (number ×102 mL-1)
Residual turbidity (%)
100
Raw wastewater Cat-floc 8108 plus (148 ppm)
60
Core shell 71301 (118 ppm) Nalcolyte 8105 (99 ppm)
40 20 0
0
10
20
30
40
50
60
70
1.E+05
1.E+04 Raw wastewater
1.E+02
16
y = 0.8x + 11.1 y = 0.7x + 11.2
15
y = 0.5x + 12.9
y = 0.9x + 8.7
y = 0.5x + 11.3
4.3
4.8
5.3
5.8
log (specific resistance, αc) [m kg-1]
log (ΔPc) [Pa, N m-2]
17
y = 0.8x + 11.7
15 y = 0.7x + 11.7
y = 0.8x + 10.5 y = 0.7x + 9.6
13 12
3.8
4.3
4.8
15
20
25
50 Raw wastewater 40% of the dose for CN 50% of the dose for CN 100% of the dose for CN 250% of the dose for CN
1
10
100
Fig. 6. Flocculated particle sizes as a function of mixing time and Nalcolyte 8105 coagulant dose: (a) total particle concentrations as a function of dose and mixing time; (b) cumulative particle counts (50th percentile crossover is the median diameter for each dose. Velocity gradient was 425 s 1 for 10 min. Polymer doses 80, 100, 200 (charge neutralization), or 500 ppm(v/v).
3.2. Coagulation and sedimentation of laundry wastewater
y = 0.7x + 11.2
14
10
difficulties in preparing a 1% solution in the field. Ultimer 1460 was the only polymer in the AcAm/DADMAC category and was retained as the fourth polymer.
Raw wastewater Cat-floc 8108 plus (100 ppm) Core shell 71301 (60 ppm) Nalcolyte 8105 (80 ppm) Ultimer 1460 (80 ppm)
16
5
Particle diameter (µm)
13
3.8
0
100
0
14
250% of the dose for CN
Elapsed mixing time (min)
Cumulative % less than indicated size
log (specific resistance, αc) [m kg-1]
Raw wastewater Cat-floc 8108 plus (290 ppm) Core shell 71301 (120 ppm) Nalcolyte 8105 (158 ppm) Ultimer 1460 (235 ppm)
50% of the dose for CN 100% of the dose for CN
Fig. 4. Residual turbidity of raw and coagulated laundry wastewater as a function of settling time. Flocculation mixing was 10 min at velocity gradient of 425 s 1. Samples collected from an outlet port 5 cm above the bottom of a 2.4 L-volume rectangular beaker.
17
40% of the dose for CN
1.E+03
Settling time (min)
12
171
5.3
5.8
log (ΔPc) [Pa, N m-2] Fig. 5. Specific resistance to filtration and coefficient of compressibility (slope of lines) during dead-end constant pressure filtration of raw and coagulated laundry wastewater through 0.22 μm PVDF membranes. Two coagulation regimes for each polymer were employed: (a) polymer dose that achieved maximum contaminant removals after sedimentation and (b) lower polymer doses that achieved significant contaminant removal by sedimentation (see Fig. 3).
retained. Among the core–shell polymers, 71301 had the highest charge density and neutralized negative charge in laundry wastewater most effectively. It was retained despite difficulties presented by the very high MW and emulsion form which can present
Sedimentation is often used in wastewater treatment applications where space and equipment mobility are not limiting factors. Fig. 3 shows removals of turbidity, TSS, TCOD, and TP after application of variable doses of the four selected polymers, 30 min flocculation, and quiescent settling for 1 h. Good removals of turbidity and TSS indicate good removal of flocculated particles by sedimentation. The core–shell polymer was the most effective of the four remaining polymers for sedimentation and a 118 ppm (v/v) dose achieved 93% removal of TSS, 74% removal of TCOD, and 22% removal of TP. The highest core–shell polymer dose ( 300 ppm) did not result in significantly decreased removals of contaminants. High doses of the other three polymers resulted poor settled removals due to charge reversal, as happened with the other three polymers. The Core shell 71301 polymer also produced the fastest settling flocs as shown in Fig. 4. In these experiments sufficient polymer was added to a new sample of laundry wastewater to achieve ZP between 8 and 10 mV, a ZP range that typically results in rapid coagulation. Addition of 118 ppm of the core–shell polymer produced very large flocs resulting in 490% removal of turbidity within 1 min. Coagulation with Nalcolyte 8105 (epi/DMA) achieved 91% removal of turbidity within 60 min of settling. Cat-floc 8108
172
H.-C. Kim et al. / Journal of Membrane Science 456 (2014) 167–174
(polyDADMAC) was less effective for sedimentation removal. The results shown in Figs. 3 and 4 indicate that the core–shell polymers should be considered for coagulation of laundry wastewater when solids are removed by sedimentation. The evidence points to polymer bridging as the dominant mechanism for core shell polymer 71301 based on the following observations: (1) the core–shell polymers are higher MW than other polymers; (2) the core–shell polymers were less effective at neutralizing charge as shown in Fig. 2; (3) high doses of the core– shell polymers did not result in charge re-stabilization; and (4) very low doses of Core shell 71301 were effective in removing contaminants after sedimentation as shown in Fig. 3. However, the core–shell polymers are delivered as emulsions, which are more difficult to use in remote field locations and it will be shown below that core–shell polymers resulted in increased resistance to filtration. 3.3. Specific resistance to filtration after coagulation of laundry wastewater
sedimentation. The specific resistance results are shown in Fig. 5a for higher coagulant doses and in Fig. 5b for the lower coagulant doses. Nalcolyte 8105 decreased the specific resistance to filtration to less than 2% of the specific resistance for raw wastewater for both polymer doses. The Core shell 71301 (AcAm/DMAEA.MCQ) and Ultimer 1460 (AcAm/DADMAC) polymers increased specific resistance to filtration in comparison with raw wastewater. Coefficients of compressibility of the cakes (n) are indicated by the regression line slopes and ranged from 0.5 to 0.9. The results indicated that epi/DMA Nalcolyte 8105 was the most promising of the seven polymers for the MF experiments based on the following evidence: (1) lowest coagulant dose for neutralization of the negative charge on laundry waste contaminants shown in Fig. 2; (2) good removal of contaminants during sedimentation as shown in Fig. 3; (4) lowest specific resistance to filtration in dead-end filtration as shown in Fig. 5; and (5) easiest preparation of diluted feed solution due to low MW and delivery of the neat product in water solution. 3.4. Effects of mixing time and coagulant dose on flocculation
The SWRS uses dead-end filtration and solids accumulate on the membrane during the filtration cycle. The accumulated solids are usually the dominant contributor to total resistance to filtration. The solids cake that has accumulated on the membrane surface will compress with increasing TMP. This can result in exponentially increasing resistance to filtration during MF operation since solids accumulation means that greater TMP is required to maintain constant flux, resulting in further compression and additional increase in resistance to filtration. The specific resistance to filtration of laundry wastewater was evaluated for raw wastewater and for wastewater treated with two coagulant doses of each of the four polymers. The coagulant doses were selected to achieve similar accumulation of cake solids on the membrane. Higher coagulant doses were based on best removals of TSS during sedimentation in Fig. 3b. Lower doses were selected to achieve approximately 60% removal of TSS during sedimentation except for Core shell 71301 for which even UD resulted in good
Mixing time after addition of Nalcolyte 8150 had a significant effect on particle size distribution. Coagulation with polymer doses ranging from 40% to 100% of CN and 10 min slow mixing resulted in 495% decrease in total number of particles (Fig. 6a) and a large increase in median particle size (Fig. 6b). Coagulation with 250% of CN, a strong OD condition, also resulted in a large decrease in particle numbers and increased particle size. The mixing time after coagulant addition had substantial effects on membrane fouling and recovery of membrane permeability by hydraulic rinsing. Table 3 shows that the resistance to filtration (Rf) after 1 h of MF filtration of laundry wastewater (case #1 with total permeate flow of 46 L m 2) rose to 172 1011 m 1. The Rf after addition of 60 ppm of Nalcolyte 8105, 10 min of slow mixing, and 1 h filtration (case #4) was only 5.6 1011 m 1 and there was almost complete removal of the incremental resistance with a hydraulic wash, resulting in 98% of the pure water flux of a fresh membrane.
Table 3 Effect of mixing time on membrane fouling. The pH was 11 before coagulation, coagulant was Nalcolyte 8105, and the coagulated wastewater was dead-end filtered through 0.22 μm PVDF membranes at constant feed pressure of 10 psi prior to cleaning operations. Case
# # # # #
1 2 3 4 5
Descriptiona
Raw wastewater Coagulated Coagulated Coagulated Coagulated
Mixing intensity, G (s 1) t (min)
N/A 425 2 425 5 425 10 425 20
Total permeate volume (L m 2)
46 628 707 770 716
Rf (1011 m 1)b
172 7.1 6.2 5.6 6.5
Rf (1011 m 1) [percent permeability recovery]c
Cumulative COD removal (%)
After hydraulic wash
After 0.1 M NaOH
After 0.1 M citric acid
0.58 70.01 2.00 70.02 0.55 70.01 0.48 70.01 0.49 70.01
0.517 0.01 0.707 0.02 0.52 7 0.02 0.45 7 0.01 0.46 7 0.01
0.517 0.01 0.59 7 0.01 0.497 0.01 0.45 7 0.01 0.46 7 0.01
[82] [24] [85] [98] [96]
[88] [64] [86] [99] [98]
[88] [77] [92] [99] [98]
74 81 82 81 82
a All coagulation pretreatments of laundry wastewater were conducted using a Nalcolyte 8105 dose of 60 ppm, which resulted in ZP between 5.8 70.6 mV corresponding to typical charge-neutralization. b Hydraulic resistance to filtration at the end of filtration test. c Percent permeability recovery of fouled membranes after sequential cleanings was determined based on the pure water permeability measured for each fresh membrane before the filtration tests.
H.-C. Kim et al. / Journal of Membrane Science 456 (2014) 167–174
Backwash
0.1M NaOH
0.1M citric acid
100
Flux recovery (%)
Hydraulic resistance (1011 m-1)
50 40 30 A: B: C: D: E:
20 10 0
0
80 60 40 20
0 Raw wastewater 40% of the dose for CN 50% of the dose for CN 100% of the dose for CN 250% of the dose for CN
50
100
150
A
B
C
200
250
D
E
300
350
Specific permeate volume (L m-2) Fig. 7. Effects of Nalcolyte 8105 polymer doses on hydraulic resistance of 0.22 μm PVDF membranes during cyclic filtration of laundry wastewater. Eight MF filtration cycles of 15 minutes at constant flux of 150 L m 2 h 1 were each followed by backwash (1 min) and surface flush (0.5 min). The inset shows percent recovery of the initial pure water flux after sequential cleanings following the eight cycles of operation. The flocculation velocity gradient was 425 s-1 for 10 min.
Streaming current (mV)
1.0 y = 0.0216x - 0.0821 R2 = 0.9525
0.5 0.0
pH 7
-0.5
173
CN dose. Samples were mixed at 425 s 1 for 10 min, previously identified as optimal mixing. Results are shown in Fig. 7. MF operation with the raw wastewater was terminated during the first 15 min cycle due to excessive TMP. Coagulation to zero ZP (CN) eliminated almost all fouling. The two UD conditions resulted in very small increases in resistance during the 15 min runs and most fouling was removed by the hydraulic wash. OD coagulation using 250% of the CN dose resulted in steadily increasing fouling that was incompletely removed by the hydraulic wash between cycles. The inset in Fig. 7 shows recoveries of permeate flux after chemical cleaning and indicates poor recoveries by chemical cleaning for the raw wastewater and the OD coagulation conditions. 3.6. Operational issues – temperature, aging, coagulant control ZP titrations and 0.22 μm PVDF MF tests on laundry wastewater using fresh and aged Nalcolyte 8105 showed no change in the effectiveness of the polymer after aging 1% solutions for up to 8 days. Aging was conducted over a range of temperatures from 4 to 45 1C. ZP and SC measurements were simultaneously performed on samples that were coagulated with 20–520 ppm of Nalcolyte 8105, over a pH range from 7 to 11. Results are shown in Fig. 8. The linear relation and the simultaneous indications of CN (zero ZP) demonstrate that SC can be used to control coagulant dosing in laundry wastewaters. SC monitors are widely used and provide simple and robust operation [17,18]. SC readings were stable as a function of time from addition of coagulant.
pH 8 -1.0
pH 9
-2.0
4. Summary and conclusions
pH 10
-1.5
pH 11 -80
-60
-40
-20
0
20
40
Zeta potential (mV) Fig. 8. Correlation between streaming current (SC) and zeta potential (ZP). SC is a robust analytical method that is used to achieve real-time coagulant dosing. The good correlation over a wide range of coagulant doses and pH values shows that SC can be used to control coagulant dosing of laundry wastewater. The laundry wastewater pH was adjusted between 7 and 11 before coagulation with Nalcolyte 8105 using polymer doses from 20 to 520 ppm (v/v). All SC and ZP measurements were conducted immediately after flocculation with a velocity gradient of 425 s 1 for 10 min.
The photographs at the bottom of Table 3 show the MF membranes after 1 hr filtration plus hydraulic wash and sequential NaOH and citric acid cleanings. Photograph #1 (no coagulant) shows that particles remained small and attached particles were poorly removed. Photograph #4 (coagulant and 10 min mixing) shows a clean membrane, consistent with almost complete recovery of initial permeability. Coagulation followed by 2 or 5 min flocculation (photographs #2 and #3) resulted in dramatically improved permeability after 1 h of filtration compared to raw wastewater, but the flocculated particles were incompletely removed by the cleaning processes. Coagulation also increased removal of TCOD from 74% with raw wastewater to 81% with any of the coagulation doses.
The feasibility of using coagulation as a pre-treatment prior to MF of laundry wastewater was evaluated with emphasis on using quaternary amine coagulants that retain a high cationic charge even at high pH values. Seven coagulants were tested. The following conclusions were drawn from this study.
A high MW core–shell polymer was optimal for sedimentation
3.5. Multi-cycle MF runs Multi-cycle filtration runs were used to evaluate the effect of coagulation by Nalcolyte 8105. A fresh sample of laundry wastewater was used. Eight cycles of MF operation included 15 min of filtration at 150 L m 2 h 1, 1 min of backwash, and 0.5 min of surface flush for each coagulant dose. A ZP titration was performed and coagulant dosing was set at 40%, 50%, 100%, and 250% of the
processes for removal of contaminants. Cationic polyacrylamide polymer (Core shell 71301) achieved 490% removal of turbid materials in laundry wastewater at pH 11 within 1 min of settling time after coagulation. A low MW epi/DMA polymer was the best for decreasing specific resistance to filtration in MF operations. Epi/DMA (Nalcolyte 8105) and polyDADMAC (Cat-floc 8108 plus) decreased clogging of PVDF MF membranes by a factor of 10 at pH 11, but other acrylamide copolymers rather increased specific resistance to filtration in comparison with raw lint wastewater. Multi-cycle MF tests showed that coagulant additions from 40% to 100% of the CN dose were effective at reducing fouling. These doses also produced a cake that was easily removed by hydraulic cleaning between the filtration cycles. A large overdose of the cationic polymer (250% of CN) substantially reduced fouling compared to raw laundry wastewater, but resulted in increasing baseline fouling that was not removed by chemical cleaning. Mixing time was important. Initial fouling was reduced with 2 min mixing at 425 s 1 and recovery after hydraulic washing was greatly improved when 10 min mixing was provided. The epi/DMA polymer was very effective over a wide range of doses, from 40% to 100% of the CN condition. The 1% polymer solution was stable with aging at 4–451C. In addition there was excellent correlation between ZP and SC readings over a wide range of pH and coagulant dosing. The wide range of coagulant dosing, temperature stability, and ability to use SC make the
174
H.-C. Kim et al. / Journal of Membrane Science 456 (2014) 167–174
epi/DMA polymer a good candidate for use in remote locations where accurate control of dosing is difficult and storage conditions can be harsh.
Acknowledgment This material is based upon work supported by the U.S. Army Research Laboratory and the U.S. Army Research Office under Contract/Grant no. W911NF1010224. The mention or use of any commercial products does not imply official U.S. Army endorsement. References [1] S. Sostar-Turk, I. Petrinic, M. Simonic, Laundry wastewater treatment using coagulation and membrane filtration, Resour. Conserv. Recycling 44 (2005) 185–196. [2] N. Schouten, L.G.J. van der Ham, G.J.W. Euverink, A.B. de Haan, Selection and evaluation of adsorbents for the removal of anionic surfactants from laundry rinsing water, Water Res. 41 (2007) 4233–4241. [3] I. Ciabatti, F. Cesaro, L. Faralli, E. Fatarella, F. Tognotti, Demonstration of a treatment system for purification and reuse of laundry wastewater, Desalination 245 (2009) 451–459. [4] J. Ahmad, H. EL-Dessouky, Design of a modified low cost treatment system for the recycling and reuse of laundry waste water, Resour. Conserv. Recycling 52 (2008) 973–978. [5] J. Guilbaud, A. Masse, Y. Andres, F. Combe, P. Jaouen, Laundry water recycling in ship by direct nanofiltration with tubular membranes, Resour. Conserv. Recycling 55 (2010) 148–154.
[6] C.T. Wang, W.L. Chou, Y.M. Kuo, Removal of COD from laundry wastewater by electrocoagulation/electroflotation, J. Hazard. Mater. 164 (2009) 81–86. [7] J.T. Ge, J.H. Qu, P.J. Lei, H.J. Liu, New bipolar electrocoagulation–electroflotation process for the treatment of laundry wastewater, Sep. Purif. Technol. 36 (2004) 33–39. [8] E. Onder, A.S. Koparal, U.B. Ogutveren, An alternative method for the removal of surfactants from water: electrochemical coagulation, Sep. Purif. Technol. 52 (2007) 527–532. [9] K.Y.J. Choi, B.A. Dempsey, In-line coagulation with low-pressure membrane filtration, Water Res. 38 (2004) 4271–4281. [10] A. Bagga, S. Chellam, D.A. Clifford, Evaluation of iron chemical coagulation and electrocoagulation pretreatment for surface water microfiltration, J. Membr. Sci. 309 (2008) 82–93. [11] F.M. Tiller, Tutorial: interpretation of filtration data I, Fluid/Part. Sep. J. 90 (1990) 85–94. [12] J. Kim, F.A. DiGiano, Defining critical flux in submerged membranes: Influence of length-distributed flux, J. Membr. Sci. 280 (2006) 752–761. [13] B. Farizoglu, B. Keskinler, Sludge characteristics and effect of crossflow membrane filtration on membrane fouling in a jet loop membrane bioreactor (JLMBR), J. Membr. Sci. 279 (2006) 578–587. [14] S.Y.A. Lee, A.G. Fane, T.D. Waite, Impact of natural organic matter on floc size and structure effects in membrane filtration, Environ. Sci. Technol. 39 (2005) 6477–6486. [15] APHA, Standard Methods for the Examination of Water and Wastewater, 21st ed., American Public Health Association, Washington, 2005. [16] T. Asselman, G. Garnier, Dynamics of polymer-induced hetero-flocculation of wood fibres and fines, Colloids Surf. a-Physicochem. Eng. Aspects 174 (2000) 297–306. [17] S.K. Dentel, A.V. Thomas, K.M. Kingery, Evaluation of the streaming current detector. 1. Use in jar tests, Water Res. 23 (1989) 413–421. [18] S.K. Dentel, A.V. Thomas, K.M. Kingery, Evaluation of the streaming current detector. 2. Continuous-flow tests, Water Res. 23 (1989) 423–430.
