Augmentation of membrane bioreactors (MBRs) with activated carbon is established to offer several operational advantages. This work investigates the influence of low dosing (2 g/L) of powdered activated carbons (PACs) with different characteristics on the performance of MBR treating high strength molasses distillery wastewater containing difficult-to-biodegrade recalcitrant components. Two MBRs, augmented with different PACs, were operated in parallel over a period of 240 days and their performance monitored in terms of biomass growth, reduction in chemical oxygen demand (COD), sludge properties like extracellular polymeric substances content, filterability, and morphology. Removal of organics and coloring matter by adsorption, biodegradation and membrane filtration was estimated. Although adsorptive removal of color and COD is influenced by the properties of the PAC used, the performance of the PAC-MBRs was independent of PAC properties. Both PACs preferentially adsorbed the low molecular weight components in distillery wastewater. Retention by the membrane filter with the secondary cake layer contributed to reduction in color and COD of treated effluent. The findings indicate that low dosing with PAC adsorbing low molecular weight organics has a limited role in PAC-MBR treating distillery wastewater.

INTRODUCTION

Membrane bioreactors (MBRs) integrate aerobic or anaerobic biological treatment (for degrading organics) with membrane filtration (for sludge separation from the treated effluent). A variation incorporates supplementation of the biological reactor contents with powdered (PAC) or granular (GAC) activated carbon for adsorbing pollutants. The activated carbon particles also act as support for attached growth of microorganisms, forming biologically activated carbon (BAC) (Kim et al. 1998; Liu et al. 2005) with simultaneous adsorption–biodegradation, leading to enhanced organics removal (Cecen et al. 2003; Seo et al. 2004; Whang et al. 2004; Liu et al. 2005). This mode has the ability to handle higher organic loading (Liu et al. 2007), better system resistance to shock loads (Ma et al. 2012) and improved filterability (Lesage et al. 2008). Furthermore, critical flux is increased (Li et al. 2005; Damayanti et al. 2011; Yuniarto et al. 2013) and membrane fouling is reduced as manifested by lower transmembrane pressure (TMP) rise (Guo et al. 2008), decrease in total filtration resistance (Ying & Ping 2006; Sagbo et al. 2008) and enhancement in operation time without filter cleaning (Li et al. 2005; Liu et al. 2007).

Over the last decade, there has been extensive work on activated carbon augmented MBRs. Most studies have employed various synthetic wastewaters, though real streams such as municipal sewage, surface water, landfill leachate, seawater and a few industrial wastewaters have also been treated. In general, one-time dosing of commercial PACs (with typical surface area around 1,000 m2/g) has been practiced. PAC dosages in the studies vary from very low (0.005 g/L) to very high (75 g/L). Dosages up to 5 g/L are more common since high PAC dosages add to the operation cost. The primary focus of PAC addition is on improving system performance by controlling membrane fouling; removal of trace organic pollutants is also reported (Li et al. 2011). It is only in recent works that attention has been paid to activated carbon properties like particle size (Johir et al. 2013) and dosing procedure (one-time or repeated addition) (Jeong et al. 2013) on system performance.

There are limited studies on industrial wastewaters in PAC augmented MBRs. Industrial wastewaters are characterized by much higher organic load and often contain difficult-to-biodegrade recalcitrant components. It is, therefore, common practice to dilute such wastewaters prior to treatment.

In an earlier work on treatment of high strength molasses distillery effluent, 2 g/L PAC addition in the MBR showed several benefits, including higher chemical oxygen demand (COD) removal at a given organic loading rate (Satyawali & Balakrishnan 2009). The objective of this work was to ascertain if low dosing of PACs with different characteristics is equally effective. In this work, two MBRs, augmented with different PACs, were operated in parallel over a period of 240 days and their performance monitored in terms of biomass growth, reduction in COD, sludge properties like extracellular polymeric substances (EPS) content, filterability, and morphology. Removal of organics and coloring matter by adsorption on PACs, dried biomass (sludge) and membrane filter material were assessed through batch tests. Contribution by membrane filtration and biodegradation was estimated. The results provide a better understanding of PAC-MBR systems for treating high-strength molasses distillery wastewaters containing recalcitrant components.

MATERIALS AND METHODS

Materials

The PACs used in this work were procured from Sigma–Aldrich, India (activated charcoal DARCO®, designated as PAC1) and Merck Specialties Pvt. Ltd, Mumbai, India (designated as PAC2). Synthetic melanoidins were prepared from glucose and glycine (Dahiya et al. 2001). The solution obtained was diluted appropriately before use. All the chemicals used in this study are analytical grade and were used as received.

Sugarcane molasses distillery wastewater was collected from the outlet of the anaerobic digestion unit at Simbhaoli Sugars Limited, Simbhaoli, Uttar Pradesh. This wastewater was sieved through 0.2 μm mesh and stored at 4 °C for subsequent use. The wastewater COD was 40,533 ± 5,740 mg/L. Activated sludge for inoculating the MBRs was obtained from Okhla municipal wastewater treatment plant, New Delhi, India. The collected sludge was screened through 425 μm sieve, centrifuged at 6,000 rpm for 12 min (REMI R24, Mumbai) and the pellet re-suspended in water before inoculation.

Ceramic membrane filters were prepared from bagasse ash (Batra & Tewari 2006). The bagasse ash is a waste generated by the combustion of sugarcane bagasse in sugar factories. Modules were prepared by gluing two flat-sheet ceramic membranes on either side of a polyacrylic support. A hollow tube inserted in the support allowed withdrawal of filtrate by applying suction. Both the membrane filters and the modules were fabricated in-house. The membrane has a mean pore size of 1.6 μm and porosity of 30%, as determined from scanning electron microscope (SEM) images using Image J (NIH, USA) (Marel et al. 2010).

Batch adsorption and filtration

Adsorption studies with distillery wastewater were conducted with the two PACs, dried sludge and membrane filter. PACs were used as-received and dosage up to 50 g/L was examined. Sludge samples (obtained from the activated sludge plant and the BAC sludge from the two MBRs) were dried and powdered before use. The sludge dosage used was 8 g/L which is close to the mixed liquor suspended solids (MLSS) maintained in the MBRs. The ceramic membrane filter was cut into 1 × 1 cm sized pieces (0.6 ± 0.02 g); the pieces were submerged in distilled water and sonicated for 30 min in an ultrasonicator (Toshiba, India). The process was repeated several times until all loose particles were dislodged. These filter pieces were dried and weighed before the adsorption studies. Select adsorption experiments were done with the PACs using the high (>12 kDa) and low (<12 kDa) molecular weight fractions of distillery wastewater. These fractions were obtained by dialyzing the distillery wastewater against distilled water through a 12 kDa cellulose membrane (Sigma–Aldrich, India) over a 72 h period. The PAC dosage was 2 g/L and contact time was 6 h. All the experiments and analysis were done in triplicate.

Adsorption experiments were performed in conical flasks containing 100 mL of distillery wastewater mixed with a known amount of the adsorbent. For each set of experiments, a flask with the same wastewater but without adsorbent addition was used as control. The adsorbent–wastewater mixture was agitated at 25 °C at 160 rpm in a shaker (Scigenics Biotech, Orbitex, India). 1 mL of sample was collected after every 2 h for the first 8 h, then after 24 and 48 h. The sample was centrifuged (Remi R-24, India) at 8,000 rpm for 20 min to remove suspended particles and the supernatant was used for analysis of COD and color.

Batch filtration was conducted by submerging a filter module in a container filled with distillery wastewater and withdrawing the permeate by suction using a peristaltic pump operated in intermittent suction mode (10 min on and 99 s off). Permeate was withdrawn at regular time intervals and the flow rate, COD and color were measured. Permeate flux and retention (color and COD removal %) were calculated by Equation (1) and (2), respectively.
formula
1
formula
2

MBR set-up and operation

The locally fabricated bioreactors were 11 L acrylic tanks (working volume 8 L), with a drain valve at the bottom. The tank was equipped with an air diffuser (fine tubular diffuser, Southern Cogen Systems Pvt. Ltd, Mysore, India) at the bottom of the reactor. The diffuser was connected to an air compressor (Elgi, Coimbatore, India) and a rotameter (Bellstone Hi-Tech International, Delhi, India). The flat-sheet membrane filter module (filtration area 0.05 m2) was submerged in the bioreactor above the air diffuser. The SEM image of the membrane filter is shown in Figure 1. Two peristaltic pumps (Enertech ENDP-100 Optima, Pragati Biomedical, Mumbai, India), one each for feeding the reactor and for withdrawing permeate, were used. The pumps were operated in an intermittent suction mode with a cycle of 10 min on and 99 s off. TMP during filtration was measured using a pressure sensor (Wika S-10) mounted between the filter and permeate pump. The TMP was recorded using a data logger (DigiPro, Digital Promoters, India).
Figure 1

SEM image of the membrane showing a porous support layer overlaid by a 30–40 μm thin separation layer.

Figure 1

SEM image of the membrane showing a porous support layer overlaid by a 30–40 μm thin separation layer.

The activated sludge was acclimatized to the distillery wastewater in a fed-batch mode over a period of 130 days. In reactor R1, one time addition of 2 g/L PAC1 was done after acclimatization. In reactor R2, PAC2 addition was done in two stages – 1 g/L during the acclimatization period and 1 g/L post acclimatization. A control reactor without PAC addition was also operated for limited period of 45 days. The PAC dosage of 2 g/L was selected based on the range reported in literature for industrial wastewaters. The reactors were operated in continuous mode over a period of 240 days. The hydraulic retention time was 14 days. The solids retention time (SRT) was 120 days, after which excess sludge was removed. Aeration was 4–5 liters per minute (LPM) and the dissolved oxygen content in the reactor was 2–3 mg/L. No pH adjustment was done for reactor contents during the operation.

Analysis

Mixed liquor suspended and volatile suspended solids (MLSS, MLVSS) and COD were analyzed as per the APHA AWWA & WEF (1998). Color was measured at 475 nm using UV-Vis spectrophotometer (Shimadzu, UV-1700 Pharma spec). The sample was centrifuged and vacuum filtered through a 0.45 μm membrane filter (Millipore USA) prior to color measurement. The sludge dewaterability was analyzed using a capillary suction time (CST) meter (Type 304 M Capillary Suction Timer, Triton Electronics Ltd). Samples were diluted (typically in 1:4 ratio) for CST measurement. Sludge and membrane morphology was studied with SEM at 20 Kv/EHT. Prior to analysis, sludge samples were fixed in glutaraldehyde solution (2.5% w/v) for 12 h, followed by washing in 0.1 M sodium phosphate buffer three times for 10 min each. Sequential dehydration was achieved using water–acetone solution. After chemical drying with hexamethyldisilazane, the samples were coated with palladium in an argon atmosphere using a vacuum evaporator and examined with Zeiss EVOM10.

The Brunauer–Emmett–Teller (BET) surface area of the PAC was measured using ASAP 2010 Micromeritics (USA). Particle size was analyzed using particle size analyzer (Sympatec HELOS BF, Germany). Sludge samples were oven dried at 102 °C prior to analysis. PAC characterization was done by the following procedures: moisture (ASTM D2867-04), ash (ASTM D 2866-94), volatile matter (ASTM D 5832-98), pH (ASTM D 3838-05). Carbon content was analyzed using carbon, hydrogen, and nitrogen (CHN) analyzer, Perkin Elmer series II, 2400.

RESULTS AND DISCUSSION

PAC adsorption

The properties of the two PACs used in this work are shown in Table 1. The values are an average of two replicates.

Table 1

Properties of PAC1 and PAC2

ParametersPAC1PAC2
C (%) 83.3 71 
Moisture content (%) 9.3 17.5 
Ash content (%) 4.5 3.9 
Volatile matter content (%) 10 11 
pH 7.5 
Single point BET surface area (m2/g) 872 831 
Multipoint BET surface area (m2/g) 867 823 
Langmuir surface area (m2/g) 1,241 1,182 
Pore diameter (4 V/A by BET) (nm) 3.69 2.37 
ParametersPAC1PAC2
C (%) 83.3 71 
Moisture content (%) 9.3 17.5 
Ash content (%) 4.5 3.9 
Volatile matter content (%) 10 11 
pH 7.5 
Single point BET surface area (m2/g) 872 831 
Multipoint BET surface area (m2/g) 867 823 
Langmuir surface area (m2/g) 1,241 1,182 
Pore diameter (4 V/A by BET) (nm) 3.69 2.37 

PAC1 has higher carbon content and surface area than PAC2, while PAC2 has higher moisture content. The pH of PAC1 is neutral and PAC2 is acidic. The higher average pore diameter of PAC1 is expected to facilitate better penetration of larger molecules than PAC2.

Batch adsorption studies show that the removal of both color and COD increases with increasing PAC dosage (Figure 2(a)). The COD comprises contribution from both color imparting and non-color imparting compounds. The COD and color removal with PAC1 are 35% and 31%, respectively, at the maximum dosage of 40 g/L. There is a good linear correlation (R2 = 0.92) between the percentage of COD and the color removed. The performance of PAC2 is relatively poor with COD and color removal of 21% and 9%, respectively, at 40 g/L. Also, color removal is lower than COD removal at all dosages tested. Overall, PAC1 performs better than PAC2.
Figure 2

COD and color reduction with PAC (a) distillery wastewater (b) high (>12 kDa) and low (<12 kDa) molecular weight fractions in wastewater.

Figure 2

COD and color reduction with PAC (a) distillery wastewater (b) high (>12 kDa) and low (<12 kDa) molecular weight fractions in wastewater.

To obtain a better understanding of the type of compounds adsorbed, adsorption experiments were done with the dialyzed (>12 kDa) and the dialysate (<12 kDa) fractions of the distillery wastewater. Low molecular weight compounds contributing to color and COD were preferentially adsorbed by both of the PACs (Figure 2(b)). PAC1 showed near complete removal of color in this fraction. The color to COD removal ratio was 1.22 (PAC1) and 0.79 (PAC2). With the >12 kDa fraction, color removal was poor; the color to COD removal ratio was 0.47 (PAC1) and 0 (PAC2). Thus, it can be concluded that there is minimal removal of components >12kD in the distillery wastewater by PAC adsorption.

PAC-MBR performance

The performance of two submerged MBRs, one each with PAC1 and PAC2 augmentation, is shown in Figure 3. The MLSS increased steadily and the excess sludge was removed after 120 days to maintain a concentration of 8 to 10 MLSS g/L. In both the reactors, the MLVSS/MLSS ratio during the study period is almost constant at 0.8 (Figure 3(a)). This value is within the range typically found in activated sludge process (0.65–0.85). The average feed pH over the study duration was 7.82 ± 0.8, while the average pH of permeate was 7.69 ± 0.88 (R1) and 8.37 ± 0.35 (R2). Similar pH of feed and permeate indicates that the nitrogenous compounds in the wastewater remain unaffected during the treatment. The COD removal profiles in both of the reactors are similar throughout the test period, displaying a zig-zag decrease over the period of operation (Figure 3(b)). Over the first 120 days, the average COD reduction was 40 ± 15% (R1) and 41.5 ± 11.3% (R2). There was a subsequent drop in COD reduction in both reactors, possibly due to gradual accumulation of recalcitrant compounds. The COD removal profile for the control (without PAC addition) was similar and is not shown. The SEM image of the biomass (sludge) clearly shows that the sludge wraps around the PAC and grows (Figure 3(c)). The sludge filterability was poor but improved upon removing excess sludge (Figure 3(d)).
Figure 3

PAC-MBR performance: (a) MLSS profile and VSS/MLSS ratio; (b) COD profile and removal; (c) SEM image of PAC and sludge; (d) sludge filterability.

Figure 3

PAC-MBR performance: (a) MLSS profile and VSS/MLSS ratio; (b) COD profile and removal; (c) SEM image of PAC and sludge; (d) sludge filterability.

It is interesting to note that, though the two PACs showed difference in COD and colour removal in batch adsorption studies, their performance in the MBR was similar. Even in the batch adsorption studies, the difference in performance is more pronounced only at higher PAC dosages (10 g/L and above). At 2 g/L, both PACs show COD and colour removal below 10%. Furthermore, in the MBR, the low molecular weight soluble EPS components would also be adsorbed.

The COD removal is similar (∼40%) to that obtained in our earlier work (Satyawali & Balakrishnan 2009), as the acclimatized sludge degrades only low molecular weight compounds. This was even though PAC and membrane properties, PAC addition strategy and SRT, were different in the two works. Sludge filterability was, however, relatively poor in the current study. This may be linked to several factors such as high MLSS, presence of aged PAC, accumulation of inert matter due to long SRT and rejection by membrane with secondary cake layer which are known to affect the MBR performance (Skouteris et al. 2015).

To understand if regular replacement of aged PAC affected the biological treatment, two reactors were tested with monthly (Reactor A) and weekly (Reactor B) PAC addition of 2 g/L. Figure 4 shows the COD profiles in the reactors over 210 days. Although there was an initial variation in the first 30 days, the COD in both the reactors was similar over the remaining duration. Distillery wastewater contains both high and low molecular weight organics. The PAC used preferentially adsorbs low molecular weight compounds. At 2 g/L dosage, the adsorption was low and insufficient to impact COD removal in this high strength wastewater. There is, thus, a need to choose suitable PAC that adsorbs the high molecular weight organics in this wastewater.
Figure 4

COD in reactors with periodic PAC replenishment.

Figure 4

COD in reactors with periodic PAC replenishment.

Role of PAC

In the PAC-MBR, COD and color removal would occur due to: (i) adsorption on activated carbon and possibly the sludge and ceramic filter material; (ii) biodegradation utilizing readily degradable organics and low molecular weight components (<1,000 kD) of the melanoidins fraction (Satyawali & Balakrishnan 2009); and (iii) retention of recalcitrant, high molecular weight components by the ceramic membrane with the secondary sludge layer. Figure 5 shows the adsorptive and filtration removal. As expected, adsorptive removal is more with PACs compared with the sludge and filter material. However, it is still low and contributes less than 10% of the removal (Figure 5(a)). Filtration of the distillery wastewater alone shows an increase in COD and color reduction over time (Figure 5(b)). This is accompanied by a corresponding steady decrease in permeate flux. The increased COD and color reduction with time could be due to the deposition of rejected solids and colloids of melanoidins on the membrane surface resulting in a dense and compact structure that improves solids separations by size exclusion. This assumption was supported by the sludge filterability test (Figure 3(d)). In another study (Singh et al. 2015), we observed that an external carbon source is required to expedite the biodegradation process; after that the organisms are able to sustain on low molecular weight compounds and the large molecules keep on accumulating inside the reactor without any degradation.
Figure 5

Removal of color and organics from as-received distillery wastewater by (a) adsorption and (b) filtration.

Figure 5

Removal of color and organics from as-received distillery wastewater by (a) adsorption and (b) filtration.

Three removal mechanisms (adsorption, filtration, and biodegradation) were studied to understand the dominant mechanisms that resulted in COD and color reduction in the reactor. The study shows that during the initial stages of reactor operation, all three mechanisms are present. But as time passes, filtration through the membrane with secondary cake layer was the only mechanism that contributes to reduction in color and COD. The complex compounds kept on accumulating inside the reactor without being biodegraded.

CONCLUSIONS

Adsorptive removal of color and COD from distillery wastewater is influenced by the properties of the PAC used. The commercial PACs used in this work were suitable for removal of low molecular weight components. When supplemented at low dosage, the performance of PAC-MBR treating distillery wastewater was independent of PAC properties. Retention by the membrane filter with the secondary cake layer mainly contributes to reduction in color and COD of treated effluent. The findings indicate that low dosing with PAC adsorbing low molecular weight organics has a limited role in PAC-MBR treating distillery wastewater.

ACKNOWLEDGEMENT

The authors acknowledge the grant BT/PR 10664/BCE/08/675/2008 from the Department of Biotechnology, Ministry of Science and Technology, Government of India for conducting this work.

REFERENCES

REFERENCES
APHA, AWWA, WEF
1998
Standard Methods for the Examination of Water and Wastewater
, 20th edn.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
,
USA
.
Batra
V. S.
Tewari
P. K.
2006
A process for the preparation of inorganic membrane filters for solid liquid and solid–gas separations. Indian patent application No. 1362/Del/2006
.
Lesage
N.
Spérandio
M.
Cabassud
C.
2008
Study of a hybrid process: adsorption on activated carbon/membrane bioreactor for the treatment of an industrial wastewater
.
Chemical Engineering and Processing: Process Intensification
47
,
303
307
.
Liu
Y.
Wang
L.
Cui
H.
Zhang
J.
2005
Performance improvement of hybrid membrane bioreactor with PAC addition for water reuse
.
Water Science and Technology
52
,
383
391
.
Marel van der
P.
Zwijnenburg
A.
Kemperman
A.
Wessling
M.
Temmink
H.
van der Meer
W.
2010
Influence of membrane properties on fouling in submerged membrane bioreactors
.
Journal of Membrane Science
348
,
66
74
.
Sagbo
O.
Sun
Y.
Hao
A.
Gu
P.
2008
Effect of PAC addition on MBR process for drinking water treatment
.
Separation and Purification Technology
58
,
320
327
.
Seo
G. T.
Ahan
H. I.
Kim
J. T.
Lee
Y. J.
Kim
I. S.
2004
Domestic wastewater reclamation by submerged membrane bioreactor with high concentration powdered activated carbon for stream restoration
.
Water Science and Technology
50
,
173
178
.
Whang
D.
Cho
Y. M.
Park
H.
Jang
J. G.
2004
The removal of residual organic matter from biologically treated swine wastewater using membrane bioreactor process with powdered activated carbon
.
Water Science and Technology
49
,
451
457
.
Ying
Z.
Ping
G.
2006
Effect of powdered activated carbon dosage on retarding membrane fouling in MBR
.
Separation and Purification Technology
52
,
154
160
.