membranes Article

Impact of Coagulant and Flocculant Addition to an Anaerobic Dynamic Membrane Bioreactor (AnDMBR) Treating Waste-Activated Sludge Guido Kooijman 1, *, Wilton Lopes 2 , Zhongbo Zhou 3 , Hongxiao Guo 1 , Merle de Kreuk 1 , Henri Spanjers 1 and Jules van Lier 1 1

2 3

*

Department of Watermanagement, Faculty of Civil Engineering and Geosciences, Delft University of Technology, Stevinweg 1, 2628BC Delft, The Netherlands; [email protected] (H.G.); [email protected] (M.d.K.); [email protected] (H.S.); [email protected] (J.v.L.) Department of Sanitary and Environmental Engineering, University of Paraiba State, Avenida Juvencio Arruda SN, Bairro Universitario, Campina Grande, Paraíba 58429-500, Brazil; [email protected] School of Environmental Science and Engineering, Sun Yat-sen University, Guangzhou 510275, China; [email protected] Correspondence: [email protected]; Tel.: +31-(0)15-2783347

Academic Editor: Baoqiang Liao Received: 23 January 2017; Accepted: 15 March 2017; Published: 23 March 2017

Abstract: In this work, we investigated the effects of flocculation aid (FA) addition to an anaerobic dynamic membrane bioreactor (AnDMBR) (7 L, 35 ◦ C) treating waste-activated sludge (WAS). The experiment consisted of three distinct periods. In period 1 (day 1–86), the reactor was operated as a conventional anaerobic digester with a solids retention time (SRT) and hydraulic retention time (HRT) of 24 days. In period 2 (day 86–303), the HRT was lowered to 18 days with the application of a dynamic membrane while the SRT was kept the same. In period 3 (day 303–386), a cationic FA in combination with FeCl3 was added. The additions led to a lower viscosity, which was expected to lead to an increased digestion performance. However, the FAs caused irreversible binding of the substrate, lowering the volatile solids destruction from 32% in period 2 to 24% in period 3. An accumulation of small particulates was observed in the sludge, lowering the average particle size by 50%. These particulates likely caused pore blocking in the cake layer, doubling the trans-membrane pressure. The methanogenic consortia were unaffected. Dosing coagulants and flocculants into an AnDMBR treating sludge leads to a decreased cake layer permeability and decreased sludge degradation. Keywords: AnDMBR; flocculant; membrane fouling reducer; viscosity; anaerobic digestion

1. Introduction High-rate anaerobic treatment is a consolidated concept in industry due to the high chemical oxygen demand (COD) removal, energy recovery and low waste sludge production [1]. The success of high rate anaerobic reactors depends on the extent to which hydraulic retention time (HRT) and solids retention time (SRT) can be uncoupled in a system, to keep the slow growing methanogens in the system. Membranes are used for various separation techniques [2], and could therefore also be used for forming an absolute barrier for methanogens in an anaerobic membrane bioreactor (AnMBR). In this way, the HRT and SRT uncoupling in an AnMBR cannot be disturbed by, for example, high total suspended solids or high fats that can compromise the biomass retention in extended granular sludge bed reactors and upflow anaerobic sludge blanket systems. AnMBRs had become a common concept in wastewater treatment over the last decades with many full-scale references. However, despite the fact that membranes can be a cost effective solution [3], still the main drawbacks of AnMBR systems are the energy consumption, membrane fouling and relatively high investment costs [4]. Membrane fouling Membranes 2017, 7, 18; doi:10.3390/membranes7020018

www.mdpi.com/journal/membranes

Membranes 2017, 7, 18

2 of 11

limits the flux that can be achieved. Cationic flocculant aids (FAs) in an AnMBR treating wastewater are shown to temporarily increase the permeate flux, create a higher permeability of the cake layer, increase the particle size, and allow for a higher effluent quality because of lowered soluble microbial product (SMP) concentrations [5]. In an AnMBR treating sludge instead of wastewater, FAs can have additional benefits. FAs are known lower the viscosity in anaerobic sludge digestion [6]. Earlier studies indicate that a lower viscosity can increase the hydrolysis rate [6,7] which is considered the rate limiting step in anaerobic sewage sludge digestion [8]. Therefore the application of FAs in anaerobic sludge digestion can lead to increased digestion rates by lowering the viscosity [6,9], although there are reports of lower biogas production rates with the addition of FA [10]. A second advantage of FA addition to an AnMBR treating sludge is the increase in maximum SRT that can be applied. With higher SRTs, the solids concentration and thus viscosity in an AnMBR will be higher as well. Since increased viscosity limits the highest attainable SRT because of increasing solids accumulation [11], lowering the viscosity with FAs could increase the maximally attainable SRT in an AnMBR. Besides membrane fouling, another disadvantage mentioned before is the high investment and operational costs. These high costs are mainly caused by the membranes. However, the cake layer that would normally form on the membrane during filtration is dense and compact and will form an excellent barrier for solids [12]. Therefore, despite the advanced developments in membrane technology in the past decades [13], instead of using a membrane, a simple cloth can act as a support for the cake layer as well, lowering the investment and operational costs [14]. A membrane bioreactor equipped with such a cloth instead of a membrane is referred to as anaerobic dynamic membrane bioreactor (AnDMBR). To the authors’ knowledge, the effects of FAs in an AnDMBR treating sludge had not been investigated yet. There is a study that investigates the application FAs in an AnMBR treating sewage sludge [10], but the experiments to investigate the effects of FAs on digestibility of sludge were limited to batch tests and the results were inconclusive. Therefore, in this study we investigate the effects of cationic FA addition in an AnDMBR treating waste-activated sludge (WAS) in batch and continuous experiments. We investigated the methanogenic activity, the extent of sludge degradation, changes in sludge characteristics and trans-membrane pressure (TMP). Conventional anaerobic digestion (AD) at an SRT and HRT of 24 days was compared to an AnDMBR with an HRT of 18 days and an SRT of 24 days, including a period without and a period with FA addition. 2. Material and Methods 2.1. AnDMBR Setup and Operation Table 1 gives an overview of the experimental set-up. In period 1, the digester was operated as a daily fed sewage sludge digester without sludge retention. The reactor had a volume of 7 L and was operated at 35 ◦ C. In period 2 and 3, the digester was coupled to a dynamic membrane module with a total filtration area of 0.025 m2 . Table 1. Description the three periods where different operational parameters where applied. Period

Period

Period 1

0–86

Reactor Operation

Period 2

86–303

Conventional anaerobic digester AnDMBR

Period 3

303–386

AnDMBR

Substrate

Hydraulic Retention Time (Days)

Solids Retention Time (Days)

Flocculation Aid Addition

WAS

24

24

– – Calfloc 1502 + FeCl3

WAS

18

24

WAS

18

24

The cross-flow velocity over the external dynamic membrane was 0.044 m·s−1 , which corresponded to a recirculation flow of 240 L·h−1 . The membrane surface was relatively large for the required liquid extraction from the digester and thus the applied fluxes were very low, reaching only 0.10 L/m2 ·h. No backwash was required. Since in this work we focussed on the biological processes

Membranes Membranes 2017, 2017, 7, 7, 18 18

33 of of 11 11

no flux optimisation studies were performed. A mono-filament woven fabric made of polypropylene no flux optimisation studies were performed. A mono-filament woven fabric made of polypropylene material (Lampe B.V., Sneek, The Netherlands) was used as support material for the cake layer of the material (Lampe B.V., Sneek, The Netherlands) was used as support material for the cake layer of the dynamic membrane. Filtration was carried out by using a constant flux strategy set by a peristaltic dynamic membrane. Filtration was carried out by using a constant flux strategy set by a peristaltic pump (3 in Figure 1) at the permeate side. The feed and sludge withdrawal was carried out manually, pump (3 in Figure 1) at the permeate side. The feed and sludge withdrawal was carried out manually, once a day, 6 times per week (not on Sundays). The volume of sludge withdrawal varied and was once a day, 6 times per week (not on Sundays). The volume of sludge withdrawal varied and was determined by the sludge concentration and total mass in the reactor and the set SRT (24 days). determined by the sludge concentration and total mass in the reactor and the set SRT (24 days). To gas flow meter

Sludge influent

2

Sludge effluent

3

1 Dynamic membrane

Buffer vessel 7L Anaerobic digester

Permeate

Figure The first Figure 1. 1. Schematic Schematic overview overview of of the the AnMBR AnMBR setup setup used. used. The first pump pump (1) (1) transports transports the the sludge sludge to to the buffer vessel from which it is circulated over the membrane by pump 2. The permeate pump (3) (3) the buffer vessel from which it is circulated over the membrane by pump 2. The permeate pump creates creates the the pressure pressure difference difference over over the the membrane membrane by by removing removingpermeate. permeate.

The substrate (WAS) (WAS) was taken from the municipal wastewater treatment plant Harnaschpolder 1. (Den Hoorn, Hoorn,The TheNetherlands). Netherlands).The The WAS a total solids concentration between (Den WAS hadhad a total solids (TS) (TS) concentration between 55 and5565and g·L−65 − 1 −1 −1 g·L influent . The influent total solids concentration was to a constant value g·Ldiluting by diluting the The total solids concentration was set to aset constant value of 48 gof ·L 48 by the WAS WAS using tap water. The characteristics of the final feed to the digester are shown in Table 2. using tap water. The characteristics of the final feed to the digester are shown in Table 2. Table 2. Table 2. Characteristics of the waste-activated waste-activated sludge sludge used used as as feed. feed.

Parameter Unit Parameter Unit Total Solids g·L−1 −1 TotalTotal Solids g · L Volatile Solids g·L−1 −1 Total Volatile Solids g · L Total Suspended Solids g·L−1 Total Suspended Solids g·L−1 −1 Volatile Suspended Solid g·L Volatile Suspended Solid g·L−1 −1 Total chemical oxygen demand −g· L Total chemical oxygen demand g·L 1 −1 Total Nitrogen mg· L − 1 Total Nitrogen mg·L −1 Total Phosphorus mg· − 1 Total Phosphorus mg·L L

Average Value Average Value 48.2 ± 1.7 34.9 48.2 ± 1.0± 1.7 34.9 45.6 ± 1.6± 1.0 45.6 ± 1.6 33.9 ± 1.0 33.9 ± 1.0 50.1 ± 3.2 50.1 ± 3.2 2490 ± 0.515 2490 ± 0.515 24352435 ± 0.149 ± 0.149

2.2. FA Selection and Addition 2.2. FA Selection and Addition FA was selected by comparing the capillary suction time (CST) and specific resistance to FA was selected by comparing the capillary suction time (CST) and specific resistance to filtration filtration (SRF) of sludge from the AnDMBR treated with 24 cationic flocculants and coagulants. (SRF) of sludge from the AnDMBR treated with 24 cationic flocculants and coagulants. Sludge samples Sludge samples were taken at day 120. An initial screening was carried out by using a CST tests. The were taken at day 120. An initial screening was carried out by using a CST tests. The six best performing six best performing FAs (Table 3), with the shortest CST, were subjected to an SRF test. FAs (Table 3), with the shortest CST, were subjected to an SRF test. The CST and SRF tests were Table done using a 5 g/kg dosage, which means 5 g of active FA component 3. Six best performing flocculants. per kg of TS. From day 267, the best performing FA was dosed to the AnDMBR applying a dosage Charge of 7.5 g/kg. Because noProduct effect was observedCharacteristics at this point in time, from day 303, the Nalco 71305 Calfloc L1408 Branched, cationic, emulsion Medium was replaced with the Caldic (Rotterdam, The Netherlands) cationic FA Calfloc 1502 (10 g/kg) in Calfloc L111 Branched, cationic, emulsion Medium combination with 40% FeCl3 (0.13 mL FeCl3 g TS−1 ). Calfloc L1401 LMW Calfloc P1502 Calfloc LS1423 Nalco 71305

N.A. Linear, cationic, powder Polyamine Acryl-amide based

Medium/high High Medium High

Membranes 2017, 7, 18

4 of 11

Table 3. Six best performing flocculants. Product

Characteristics

Charge

Calfloc L1408 Calfloc L111 Calfloc L1401 LMW Calfloc P1502 Calfloc LS1423 Nalco 71305

Branched, cationic, emulsion Branched, cationic, emulsion N.A. Linear, cationic, powder Polyamine Acryl-amide based

Medium Medium Medium/high High Medium High

2.3. Analytical Methods Merck Spectroquant kits (Frankfurt, Germany) were used to assess ammonium-N (10–2000 mg-N/L), COD (25–15,000 mg/L) and P concentrations (0.015–5 mg-P/L). Capillary suction time was measured by a Triton Electronics Model 304M CST device (Essex, England, UK). The specific resistance to filtration (SRF) was measured by applying a pressure of 1 bar to a Whatman Grade 1 filter with 100 mL of sludge sample. The permeate volume was measured over time during 2 h. The SRF calculations were done following the procedure of Novak et al. [15]. An Anton-Paar USD200 rheometer with Z2 DIN and TEZ 180 bob (Graz, Austria) was used to measure viscosity. The particle size distribution (PSD) was analysed by a Donner Technologies DIPA-2000 laser scanner (Or Akiva, Israel) with B100 lens, and with 10–2000 µm measuring range. The soluble microbial products of polysaccharide nature (SMP-PS) were measured following the procedure of Ersahin et al. [14]. The soluble microbial products of protein nature (SMP-PN) were measured according to Bradford [16]. The median particle size (D50) was calculated from the volume based PSD. The specific methanogenic activity (SMA) and biomethane potential (BMP) tests were done as previously [6]. The BMP test was done in duplicate, and to each bottle 10 mL antifoam (100× dilution with water) was added. SMA tests were carried out in triplicate. Volatile fatty acid (VFA) concentrations were analyzed using a GC with an FID detector (Agilent 7890A, Santa Clara, CA, USA). Helium was used as carrier gas with a flow rate of 1.8 mL/min. The column used was an Agilent 19091F-112, with injector temperature of 240 ◦ C, 25 m × 320 µm × 0.5 µm, and oven temperature: 80 ◦ C. The remaining parameters were assessed following standard methods of the American Public Health Association (APHA, Washington, USA) [17]. 3. Results 3.1. Performance of the Conventional Sludge Digester and the AnDMBR (Period 1 and 2) In order to study the effect of uncoupling HRT and SRT in sludge digestion, the laboratory scale sludge digester was firstly operated as a conventional digester with an SRT equal to the HRT of 24 days, being fed once per day (period 1). Secondly, in the subsequent period (period 2), the HRT was lowered to 18 days by operating the reactor as an AnDMBR. During the first period, the VS destruction was about 37% (Figure 2). After lowering the HRT to 18 days during period 2, the VS destruction stabilised at about 32%. Also, a slight decrease in SMA could be observed after installing the membrane unit lowering the SMA from 0.19 ± 0.01 gCOD gVS−1 d−1 to 0.14 ± 0.02 gCOD gVS−1 d−1 .

Membranes 2017, 2017, 7, 7, 18 18 Membranes

of 11 11 55 of

60

0.25

60 50

0.25 0.2

50 40

0.2 0.15

40 30

0.15 0.1

30 20 20 10

0.1 0.05

100

0.05 0

0 0 0

50

100

150

200

250

300

350

400 0

50

100

150

Time 200[d]

250

300

350

400

Time [d]

VS destruction

[gCOD/gVS· SMASMA d] d] [gCOD/gVS·

5 of 11

[%] [%] Destruction Destruction

Membranes 2017, 7, 18

TS destruction

VS destruction Start with dynamic membrane

TS destruction Start with flocculant 71305

Start with membrane Change to dynamic flocculant P1502 + FeCl

Start SMA with flocculant 71305

Change to flocculant P1502 + FeCl

SMA

Figure 2. Total solids (TS) and volatile solids (VS) destruction and specific methanogenic activity Figure 2. Total solids (TS) and volatile solids (VS) destruction and specific methanogenic activity (SMA) of the reactor operated as conventional anaerobic digester in period 1 (day 1–86), operated as Figure 2. the Total solids (TS) and volatile solidsanaerobic (VS) destruction specific methanogenic activity (SMA) of reactor operated as conventional digester inand period 1 (day 1–86), operated as an an AnDMBR in period 2 (day 86–303) and operated as an AnDMBR with flocculant addition in period (SMA) of the reactor2operated as conventional anaerobic digesterwith in period 1 (day 1–86), operated AnDMBR in period (day 86–303) and operated as an AnDMBR flocculant addition in periodas3 3 (day 303–386). an AnDMBR (day 303–386).in period 2 (day 86–303) and operated as an AnDMBR with flocculant addition in period 3 (day 303–386).

100.00 90.00 100.00 80.00 90.00 70.00 80.00 60.00 70.00 50.00 60.00 40.00 50.00 30.00 40.00 20.00 30.00 10.00 20.00 0.00 10.00 0.00 0 0

50 50

100 100

150

200

250

300

350

150

Time 200[d]

250

300

350

Acetic acid

Time [d]

5 4.5 5 4 4.5 3.5 4 3 3.5 2.5 3 2 2.5 1.5 2 1 1.5 0.5 1 0 0.5 400 0 400

[Pa·s][Pa·s] Viscosity Viscosity

[mg/L] Concentration [mg/L] Concentration

The concentrations of propionate and butyrate remained close to 0 mg/L (Figure 3) during all The concentrations of propionateconversions and butyrate remained close to 0 mg/L (Figure 3) during all threeThe periods. Apparently, were not rate-limiting the (Figure digester. concentrations of acetogenic propionate and butyrate remained close to 0 in mg/L 3) during all three periods. Apparently, acetogenic conversions were not rate-limiting in the digester. three periods. Apparently, acetogenic conversions were not rate-limiting in the digester.

Butyric acid

Acetic acidacid Propionic

Butyric aciddynamic membrane Start with

Propionic Start with acid flocculant 71305

Start with membrane Change to dynamic flocculant P1502 + FeCl3

Start Total with VFA flocculant 71305

Change Viscosityto flocculant P1502 + FeCl3

Total VFAacid (VFA) concentrations of the reactor Viscosity Figure 3. Volatile fatty and viscosity of the reactor operated Figure 3. Volatileanaerobic fatty aciddigester (VFA) concentrations of 1–86), the reactor and as viscosity of the reactor operated as conventional in period 1 (day operated an AnDMBR in period 2 (day 86–303) and operated as andigester AnDMBR with flocculant addition in period 303–386). as conventional anaerobic in period 1 (day 1–86), operated as an3 (day AnDMBR in period 2 (day Figure 3. Volatile fatty acid (VFA) concentrations of the reactor and viscosity of the reactor operated 86–303) and operated as an AnDMBR with flocculant addition in period 3 (day 303–386). as conventional anaerobic digester in period 1 (day 1–86), operated as an AnDMBR in period 2 (day 86–303) and operated as an AnDMBR with flocculant addition in period 3 (day 303–386).

Membranes 2017, 7, 18 Membranes 2017, 7, 18

6 of 11 6 of 11

3.2. FA FA Selection Selection Twenty-four cationic cationic FAs FAs were were tested tested on on the the effluent effluent sludge sludge of of the the AnMBR AnMBR prior prior to to selection selection around day 130. 130. The Thebest bestperforming performingFA FAininterms terms CST and SRF was Nalco 71305. From ofof CST and SRF was Nalco 71305. From day day 267 onwards, 7.5 g/kg thisof FAthis wasFA dosed. FA addition not result a visible 267 onwards, 7.5 of g/kg was However, dosed. However, FA did addition did in not result flocculation in a visible in the reactor, SRF wasthe only affected and CST even increased (Figure 4).(Figure Also, repeated flocculation in the reactor, SRFshorty was only shorty affected and CST even increased 4). Also, CST andCST SRFand tests with increased dosage of 15ofg/kg of Nalco 71305 did repeated SRF testsan with an increased dosage 15 g/kg of Nalco 71305 didnot notshow show aa clear improvement. Therefore, Therefore, after after one one week, week, Nalco Nalco 71305 71305 dosing dosing was stopped. After a new testing phase, improvement. the applied appliedFA FAwas waschanged changedtoto a combination cationic Calfloc (10 g/kg) 3 (0.13 the a combination of of cationic FAFA Calfloc 15021502 (10 g/kg) withwith FeClFeCl mL 3 (0.13 −TS 1 ). −1This −−1 1 mL FeCl 3 g ). This lowered the CST from ~2000 s to ~500 s. From day 303, a dosage of 10 g kg FeCl g TS lowered the CST from ~2000 s to ~500 s. From day 303, a dosage of 10 g kg 3 − 1 −1 1502 and and 0.13 0.13 mL mL FeCl FeCl33 gg TS TS was 1502 was applied. applied. Because Becauseofofthe theFAs FAs built built up up in in the the reactor, the FAs 1 and −1 from −1 from 0.09 mL FeCl 3 g3 TS dayday 330330 andand to 3.3 and concentrations were lowered to 6.6 g/kg g/kg−1−and 0.09 mL FeCl g TS to g/kg 3.3 g/kg − 1 −1 0.04 0.04 mL FeCl 3 g TS 354. The SRF and CST were loweredlowered in the reactor and mL FeCl TS day from day 354. The SRF and CSTsuccessfully were successfully in the (Figure reactor 3 g from 4). When the new of FAs, problems occurred thatthat were mitigated by (Figure 4). dosing When dosing thecombination new combination of foaming FAs, foaming problems occurred were mitigated adding an an antifoam emulsion. by adding antifoam emulsion.

Figure 4. 4. The The specific specific resistance resistance to to filtration filtration (SRF) (SRF) and and the the capillary capillary suction suction time time (CST) (CST) of of the the effluent effluent Figure sludge of the reactor operated as conventional anaerobic digester in period 1 (day 1–86), operated as sludge of the reactor operated as conventional anaerobic digester in period 1 (day 1–86), operated as an an AnDMBR in period 2 (day 86–303) and operated as an AnDMBR with flocculant addition in period AnDMBR in period 2 (day 86–303) and operated as an AnDMBR with flocculant addition in period 3 3 (day 303–386). (day 303–386).

3.3. Performance Performance of of the the AnMBR AnMBR with with FAs FAs dosing dosing (Period (Period3) 3) 3.3. During period period3,3,the theviscosity viscosity was significantly lowered to addition FA addition (Figure the During was significantly lowered duedue to FA (Figure 3). In 3). theIn same same period, the VS destruction decreased to about 24% (Figure 2). A BMP test was carried out to period, the VS destruction decreased to about 24% (Figure 2). A BMP test was carried out to examine examine the possibility of irreversible substrate binding. Results showed that there was already the possibility of irreversible substrate binding. Results showed that there was already irreversible irreversible binding of FA with lowCalfloc as 5 g/kg Calfloc 1502mL andFeCl 0.07 mL FeCl −13 binding of substrate bysubstrate FA with by dosages as dosages low as 5 as g/kg 1502 and 0.07 3 kg TS −1 (Figure 5). kg TS (Figure 5).

Membranes 2017, 7, 18

7 of 11

Membranes Membranes 2017, 2017, 7, 7, 18 18

77 of of 11 11

250

/ gVS CH4CH4 N-mL / gVS N-mL

250 200 200 150 150 100 100 50 50 0 0 0 -50 0 -50

5

10

15

20

25

30

5

10

15 [d] Time

20

25

30

Time [d]

0 g/kg

5 g/kg

10 g/kg

15 g/kg

20 g/kg

0 g/kg

5 g/kg

10 g/kg

15 g/kg

20 g/kg

Figure 5. Biomethane potential (BMP) tests of waste activated sludgewith different flocculation aid Figure 5. Biomethane potential (BMP) tests of waste activated sludgewith different flocculation aid concentrations. Increased flocculant concentration decreases the Figure 5. Biomethane potential (BMP) tests of waste activated sludgewith different flocculation aid concentrations. Increased flocculant concentration decreases theBMP BMPvalues. values. concentrations. Increased flocculant concentration decreases the BMP values.

The addition of FAs in period 3 in the reactor lowered the SRF with about 40% despite the higher The addition of FAs in period 3−1in the reactor lowered the SRF with about 40% despite the higher TS concentration, reaching 57 g·L period 3 (Figure Also, CST decreased in period 3. The The addition of FAs in to period 3 in−in reactor lowered4).the SRFthe with about 40% despite the higher 1the TS concentration, reaching toperiod 57 g·−1 L2 was in 58 period 3 (Figure 4).on Also, the CST decreasedthe in period 3. average particle size (D50) in μm (determined day 256). Surprisingly, D50 TS concentration, reaching to 57 g·L in period 3 (Figure 4). Also, the CST decreased in period 3. was The The average particle size (D50) in period 2 was 58 µm (determined on day 256). Surprisingly, the D50 reducedparticle after thesize addition of period FAs in2period to 32 μm (determined daySurprisingly, 353). The TMP period average (D50) in was 583μm (determined on dayon 256). thein D50 was was reduced after the addition of FAs in period 3 to 32 µm (determined on day 353). The TMP in 2 was about mbar butofitFAs doubled to about 300 mbar in periodon 3,day when FAThe wasTMP added to the reduced after 150 the addition in period 3 to 32 μm (determined 353). in period period 2 was about 150 mbar but it doubled to about 300 mbar in period 3, when FA was added to the The 150 effluent the SMP-PS concentrations the permeate lowered 2digester. was about mbarquality but it increased, doubled toasabout 300 mbar in period 3,inwhen FA was were added to the digester. The effluent quality increased, as the SMP-PS concentrations in the permeate were lowered in periodThe 3 (Figure 6).quality At the increased, same time,asthe in in thethe supernatant of thelowered reactor digester. effluent theSMP-PS SMP-PSconcentration concentrations permeate were in period 3The (Figure 6). At the sameconcentration time, the SMP-PS concentration in the supernatant oftothe reactor increased. permeate remained equal inin period 3 comparedof period 1. in period 3 (Figure 6). At SMP-PN the same time, the SMP-PS concentration the supernatant the reactor increased. The permeate SMP-PN concentration remained equal in period 3 compared to period 1. However, the SMP-PN concentrations in the reactor supernatant increased in period 3. increased. The permeate SMP-PN concentration remained equal in period 3 compared to period 1. However, the SMP-PN concentrations in the reactor supernatant increased in period 3. However, the SMP-PN concentrations in the reactor supernatant increased in period 3.

[mg/L] Concentration [mg/L] Concentration

500 500 400 400 300 300 200 200 100 100 0 0 0 0

50

100

150

200

250

300

350

400

50

100

150

Time 200[d]

250

300

350

400

PS in reactor

Time [d] PS permeate

PS PNininreactor permeate

PS permeate Start with dynamic membrane

PN in with permeate Start flocculant 71305

Start with membrane Change todynamic flocculant P1502 + FeCl3

Start flocculant 71305 PN inwith reactor

Change to flocculant P1502 + FeCl3

PN in reactor

Figure 6. Concentrations of soluble microbial products of polysaccharide nature (SMP-PS) and soluble Figure 6. Concentrations of soluble microbial products polysaccharide (SMP-PS) microbial products of protein nature (SMP-PN) in theofsupernatant andnature the permeate ofand the soluble reactor microbial products of protein nature (SMP-PN) in the supernatant andnature theaspermeate ofand the reactor operated as conventional anaerobic digester in period 1of(day 1–86), operated an AnDMBR insoluble period Figure 6. Concentrations of soluble microbial products polysaccharide (SMP-PS) operated as conventional anaerobic digester in period 1 (day 1–86), operated as an AnDMBR in period 2 (day 86–303) and operated as an AnDMBR with flocculant addition in period 3 (day 303–386). microbial products of protein nature (SMP-PN) in the supernatant and the permeate of the reactor 2 (day 86–303) and operated as an AnDMBR flocculant addition in period 3 (day 303–386). operated as conventional anaerobic digester inwith period 1 (day 1–86), operated as an AnDMBR in period

2 (day 86–303) and operated as an AnDMBR with flocculant addition in period 3 (day 303–386).

Membranes Membranes 2017, 2017, 7, 7, 18 18

88 of of 11 11

Concentration [mg/L]

The concentrations of ortho-phosphate (PO4-P) decreased in period 3 (Figure 7). The The concentrations of for ortho-phosphate decreased in period 3 (Figure concentrations were similar the reactor and(PO4-P) permeate. For ammonium (NH4-N), there7).wasThe an concentrations were similar for the reactor and permeate. For ammonium (NH4-N), there was an increase in reactor concentration in period 3, while the permeate concentrations remained the same. increase in reactor concentration in period 3, while the permeate concentrations remained the same. 2000 1800 1600 1400 1200 1000 800 600 400 200 0 0

50

100

150

200

250

300

350

400

Time [d] PO4-P in reactor

NH4+ in permeate

PO4-P in permeate

Start with dynamic membrane

Start with flocculant 71305

Change to flocculant P1502 + FeCl3

NH4+ in reactor Figure 7. Ortho-phosphate (PO4-P) and ammonium (NH4+) concentrations of the reactor operated Figure 7. Ortho-phosphate (PO4-P) and ammonium concentrations of the reactor operated as as conventional anaerobic digester in period 1 (day (NH4+) 1–86), operated as an AnDMBR in period 2 (day conventional anaerobicasdigester in period (day 1–86),addition operated an AnDMBR in period 2 (day 86– 86–303) and operated an AnDMBR with1 flocculant inas period 3 (day 303–386). 303) and operated as an AnDMBR with flocculant addition in period 3 (day 303–386).

4. Discussion 4. Discussion 4.1. Digestion Performance in Period 2 4.1. Digestion Performance in Period 2 Compared to period 1, the VS destruction in period 2 decreased from 37% to 32% after installing Compared to period the VS in period 2 decreased from digester 37% to 32% after installing the membrane, while the1,SRTs indestruction both periods were the same. Lower performance was the membrane, SRTs both periods were the same. Lower digester performance was reported earlier,while whenthe using an in external membrane [18–20]. It was postulated that the shear forces reported earlier, when an content external through membrane It was postulated that the shear forces caused by pumping theusing reactor the[18–20]. side stream membrane unit caused disruption caused by pumping the reactor through the side stream disruption of of the microbial consortia [21].content However, no accumulation ofmembrane propionic unit acidcaused or butyric acid was the microbial consortia [21]. However, no accumulation of propionic acid or butyric acid was however observed (Figure 3), indicating that syntrophic acetogenic consortia were not notably affected. however observed (Figure 3), that syntrophic acetogenic consortia were not notably The treatment performance wasindicating also not likely to be affected by free ammonium inhibition as the affected. The treatment performance was30also likely to be is affected by free inhibition free ammonium in period 2 was about ± 4not mg/L which well below theammonium concentration that is as the free ammonium period wasVFA about 30 ± 4 mg/L which is well below the concentration that found to be inhibiting in [22]. Since2 the concentrations remained low in period 2, the decrease in is to be inhibiting [22]. Since by theaVFA concentrations remained lowhigher in period 2, the decrease in VSfound destruction was likely caused decreased hydrolysis rate. The solids concentration VS destruction was likelytocaused a decreased hydrolysis rate. The likely highernegatively solids concentration in in the reactor compared periodby 1 caused a higher viscosity, which impacted the the reactor rate compared hydrolysis [6,7]. to period 1 caused a higher viscosity, which likely negatively impacted the hydrolysis rate [6,7]. 4.2. Digester Performance in Period 3

4.2. Digester Performance Period 3 viscosity, the VS destruction further lowered to 24%. FAs are In period 3, despiteinthe lower considered to be to anaerobic consortia The acetotrophic wereFAs indeed In period 3, non-toxic despite the lower viscosity, the[6,9,23]. VS destruction further methanogens lowered to 24%. are not notably affected by the FAs, indicated by the similar SMA values in period 2 and 3. In addition, considered to be non-toxic to anaerobic consortia [6,9,23]. The acetotrophic methanogens were indeed acetogenic methanogenesis were the rate limiting during2 reactor operation as not notablyconversions affected by and the FAs, indicated by the not similar SMA valuesstep in period and 3. In addition, evidenced by the low VFA Thewere treatment also not likely to be affected acetogenic conversions andconcentrations. methanogenesis not theperformance rate limitingwas step during reactor operation byevidenced free ammonium since the free ammonium in period 3 was about 32 ±not 2 mg/L. as by theinhibition, low VFA concentrations. The treatment performance was also likelyIn toour be affected by free ammonium inhibition, since the free ammonium in period 3 was about 32 ± 2 mg/L.

Membranes 2017, 7, 18

9 of 11

previous work, we showed that FAs can irreversibly bind to solids, such that they are not available for bioconversion anymore [6]. Results of the BMP test showed that there was irreversible binding of substrate by FA already with dosages as low as 5 g/kg Calfloc 1502 and 0.07 mL FeCl3 kg TS−1 (Figure 5). Therefore, with the applied concentrations in the AnMBR, it can be concluded that part of the solids indeed were irreversibly bound, explaining the lower observed VS destruction. 4.3. Filtration Performance and Nutrients in Period 2 and 3 The SRF dropped about 40% in period 3 compared to period 2 and the CST dropped slightly. The low drop in CST in period 3, despite the FA addition compared to period 2, may have been caused by the higher TS concentration, which causes higher CST. The lower SRF and CST after FA addition is in agreement with earlier studies in an aerobic MBR [24] and AnMBR [10]. However, the TMP doubled from 150 mbar in period 2 to 300 mbar in period 3. Other studies with AnMBRs, show an increase in filterability due to FA addition [5,10]. The reason for the higher TMP may be an increase of small particles in the reactor in period 3 compared to period 2. Small particles are known to clog the cake in AnMBRs [25]. During AD, colloids are usually rapidly degraded and AD generally causes the average particle size of sewage sludge to increase [26]. That the average particle size in period 3 was lower than in period 2 may be caused by irreversible binding of FA. Cationic FAs irreversibly bind solids [27] and since cationic FAs are known to be partially non-biodegradable [28], the irreversibly bound organic particles can become refractory to biological degradation. Therefore, these refractory particles could accumulate, causing a higher TMP in period 3 compared to period 2. It should be noted that no backwash was applied. From these results it can be concluded that the filtration in terms of TMP did not benefit from the FAs due to the accumulation of small refractory particles that accumulated in the reactor. Typically, the SMP concentrations are lowed by to the addition of FAs [10,29]. However, in period 3, the SMP-PS and SMP-PN concentration increased in the reactor (Figure 6). The increase of SMP in the reactor in period 3 can be explained by the SMP present in the refractory small particles as mentioned above. At the same time SMP concentration in the effluent decreased in period 3, which can be explained by the decreased permeability of the clogged cake in period 3. In Figure 7, it can be observed that the NH4 + concentration in the reactor increases, while the permeate concentration remains the same in period 3 compared to period 2. This is most likely the result of a measurement bias: cationic flocculants are composed of quaternary ammonium groups, which could be detected as ammonium. Since the flocculant is bound to solids and thus strained by the cake, the permeate did not show the same increase as the reactor content in NH4 + . This hypothesis is supported by the fact that the measured NH4 + reactor concentration is lowered shortly after lowering the FA dosage on day 330. The PO4 -P concentration in the reactor and effluent decreased. This was most likely a consequence of the FeCl3 dosing. 5. Conclusions An increased viscosity in the reactor, after lowering the HRT to 18 days with a filter cloth, caused a lower VS destruction, most likely due to a lower hydrolysis rate caused by an increased viscosity. Subsequently lowering the viscosity with FAs did not improve the VS destruction. This was explained by an irreversible binding of the substrate. Irreversible binding of organic matter by partially non-biodegradable cationic flocculation aid led to an accumulation of small non-degradable particulates in the reactor. These particulates may have caused a higher TMP caused by pore blocking. The FA concentrations did not notably affect the microbial activity of the system. It can be concluded that FA dosage is not beneficial for WAS treating AnDMBRs. Acknowledgments: The authors thank Deva d’Angelo for her valuable contribution in the lab work during her internship. And the authors would like to express their gratitude for the Postdoc Scholarship provided by the Coordination of Improvement of Higher Education Personnel (CAPES) to Wilton Silva Lopes, by the Brazilian Ministry of Education.

Membranes 2017, 7, 18

10 of 11

Author Contributions: Zhongbo and Hongxiao started the experiments by constructing the experimental setup and by running the reactor during period 1 and a part of period 2. Wilton and Hongxiao continued the work during period 2 and period 3. Guido contributed in designing the experiments and testing the flocculation aids used in period 3. Also did he write the paper. Merle, Henri and Jules were supervising the activities and helped in correcting the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

18. 19. 20. 21.

Van Lier, J.B. High-rate anaerobic wastewater treatment: Diversifying from end-of-the-pipe treatment to resource-oriented conversion techniques. Water Sci. Technol. 2008, 57, 1137–1148. [CrossRef] [PubMed] Das, R.; Sarkar, S.; Chakraborty, S.; Choi, H.; Bhattacharjee, C. Remediation of antiseptic components in wastewater by photocatalysis using TiO2 nanoparticles. Ind. Eng. Chem. Res. 2014, 53, 3012–3020. [CrossRef] Pal, P.; Chakraborty, S.; Roy, M. Arsenic Separation by a Membrane-Integrated Hybrid Treatment System: Modeling, Simulation, and Techno-Economic Evaluation. Sep. Sci. Technol. 2012, 47, 1091–1101. [CrossRef] Krzeminski, P.; Leverette, L.; Malamis, S.; Katsou, E. Membrane bioreactors—A review on recent developments in energy reduction, fouling control, novel configurations, LCA and market prospects. J. Memb. Sci. 2016, 527, 207–227. [CrossRef] Díaz, H.; Azócar, L.; Torres, A.; Lopes, S.I.C.; Jeison, D. Use of flocculants for increasing permeate flux in anaerobic membrane bioreactors. Water Sci. Technol. 2014, 69, 2237–2242. [CrossRef] [PubMed] Kooijman, G.; de Kreuk, M.; van Lier, J.B. Influence of chemically enhanced primary treatment on anaerobic digestion and dewaterability of waste sludge. Water Sci. Technol. 2017. submitted. Dartois, A.; Singh, J.; Kaur, L.; Singh, H. Influence of guar gum on the in vitro starch digestibility-rheological and Microstructural characteristics. Food Biophys. 2010, 5, 149–160. [CrossRef] Eastman, J.A.; Ferguson, J.F. Solubilization of particulate organic carbon during the acid phase of anaerobic digestion. Water Pollut. Control Fed. 1981, 53, 352–366. Chu, C.P.; Lee, D.J.; Chang, B.V.; You, C.H.; Liao, C.S.; Tay, J.H. Anaerobic digestion of polyelectrolyte flocculated waste activated sludge. Chemosphere 2003, 53, 757–764. [CrossRef] Yu, Z.; Song, Z.; Wen, X.; Huang, X. Using polyaluminum chloride and polyacrylamide to control membrane fouling in a cross-flow anaerobic membrane bioreactor. J. Memb. Sci. 2015, 479, 20–27. [CrossRef] Meabe, E.; Déléris, S.; Soroa, S.; Sancho, L. Performance of anaerobic membrane bioreactor for sewage sludge treatment: Mesophilic and thermophilic processes. J. Memb. Sci. 2013, 446, 26–33. [CrossRef] Jeison, D.; Días, I.; van Lier, J.B. Anaerobic membrane bioreactors: Are membranes really necessary? Electron. J. Biotechnol. 2008, 11, 1–2. [CrossRef] Lee, K.P.; Arnot, T.C.; Mattia, D. A review of reverse osmosis membrane materials for desalination—Development to date and future potential. J. Memb. Sci. 2011, 370, 1–22. [CrossRef] Ersahin, M.E.; Ozgun, H.; Tao, Y.; van Lier, J.B. Applicability of dynamic membrane technology in anaerobic membrane bioreactors. Water Res. 2014, 48, 420–429. [CrossRef] [PubMed] Novak, J.T.; Goodman, G.L.; Pariroo, A.; Huang, J.; Goodman, L. The Blinding of Sludges during Filtration The blinding of sludges. Water Pollut. Control Fed. 1988, 60, 206–214. Bradford, M.M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 1976, 72, 248–254. [CrossRef] American Public Health Association; American Water Works Association; Water Environment Federation. Standard Methods for the Examination of Water and Wastewater. Available online: https://www.mwa.co.th/ download/file_upload/SMWW_1000-3000.pdf (accessed on 23 January 2017). Brockmann, M.; Seyfried, C.F. Sludge activity and cross-flow microfiltration—A non-beneficial relationship. Water Sci. Technol. 1996, 34, 205–213. [CrossRef] Choo, K.H.; Lee, C.H. Membrane fouling mechanisms in the membrane-coupled anaerobic bioreactor. Water Res. 1996, 30, 1771–1780. [CrossRef] Ghyoot, W.R.; Verstraete, W.H. Coupling Membrane Filtration to Anaerobic Primary Sludge Digestion. Environ. Technol. 1997, 18, 569–580. [CrossRef] Ghyoot, W.; Verstreate, W. Anaerobic digestion of primary sludge from chemical pre-precipitation. Water Sci. Technol. 1997, 36, 357–365. [CrossRef]

Membranes 2017, 7, 18

22. 23.

24.

25.

26.

27. 28. 29.

11 of 11

Rajagopal, R.; Masse, D.I.; Singh, G. A critical review on inhibition of anaerobic digestion process by excess ammonia. Bioresour. Technol. 2013, 143, 632–641. [CrossRef] [PubMed] Campos, E.; Almirall, M.; Mtnez-Almela, J.; Palatsi, J.; Flotats, X. Feasibility study of the anaerobic digestion of dewatered pig slurry by means of polyacrylamide. Bioresour. Technol. 2008, 99, 387–395. [CrossRef] [PubMed] Huyskens, C.; de Wever, H.; Fovet, Y.; Wegmann, U.; Diels, L.; Lenaerts, S. Screening of novel MBR fouling reducers: Benchmarking with known fouling reducers and evaluation of their mechanism of action. Sep. Purif. Technol. 2012, 95, 49–57. [CrossRef] Lin, H.; Liao, B.; Chen, J.; Gao, W.; Wang, L.; Wang, F.; Lu, X. New insights into membrane fouling in a submerged anaerobic membrane bioreactor based on characterization of cake sludge and bulk sludge. Bioresour. Technol. 2011, 102, 2373–2379. [CrossRef] [PubMed] Elmitwalli, T.A.; Soellner, J.; de Keizer, A.; Bruning, H.; Zeeman, G.; Lettinga, G. Biodegradability and change of physical characteristics of particles during anaerobic digestion of domestic sewage. Water Res. 2001, 35, 1311–1317. [CrossRef] Hogg, R. The role of polymer adsorption kinetics in flocculation. Colloids Surfaces A Physicochem. Eng. Asp. 1999, 146, 253–263. [CrossRef] Chang, L.L.; Raudenbush, D.L.; Dentel, S.K. Aerobic and anaerobic biodegradability of a flocculant polymer. Water Sci. Technol. 2001, 44, 461–468. [PubMed] Hwang, B.K.; Lee, W.N.; Park, P.K.; Lee, C.H.; Chang, I.S. Effect of membrane fouling reducer on cake structure and membrane permeability in membrane bioreactor. J. Memb. Sci. 2007, 288, 149–156. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).