1. technology and computing

N. Stahl*, A. Tenenbaum* and N.I. Galil**
*American-Israeli Paper Mill (AIPM) Hedera, Israel.
**Faculty of Civil and Environmental Engineering, Technion Institute of Technology, Haifa 32000, Israel.
(E-mail: galilno@tx.technion.ac.il)
Abstract The operation of an activated sludge process at a paper mill (AIPM) in Hedera, Israel, was often
characterized by disturbances. As part of a research and development project, a study on new biological
treatment was initiated. The study included the operation of three pilot units: a. anaerobic treatment by
upflow anaerobic sludge blanket (UASB); b. aerobic treatment by two pilot units including activated sludge
and membrane bioreactor (MBR), which have been operated in parallel for comparison reasons. The pilot
plant working on anaerobic treatment performed COD reduction from 2,365 to 755 mg/L, expressed as
average values. Based on the pilot study, a full scale anaerobic treatment system has been erected. During a
period of 100 days, after achieving steady state, the MBR system provided steady operation performance,
while the activated sludge produced effluent characterized by oscillatory qualities. The following results,
based on average values, indicate much lower suspended solids concentrations in the MBR effluent, 2.5
mg/L, as compared to 25 mg/L in the activated sludge. The ability to develop and maintain a concentration of
over 11,000 mg/L of mixed liquor volatile suspended solids in the MBR enabled an intensive bioprocess at
relatively high cell residence time. This study demonstrates that the anaerobic process, followed by aerobic
MBR can provide effluent of high quality which can be considered for economic reuse in the paper mill
industry.
Keywords aerobic; anaerobic; membrane bioreactor; paper mill
Introduction
During the last decade, the environmental requirements for discharge of effluent into water
bodies have become more severe, therefore existing wastewater treatment facilities have to
improve operating performance and provide effluent of higher quality, conforming to more
stringent regulation. The direct aerobic biotreatment of wastewater from paper mills is
reported as experiencing difficulties of heavy foam, poor biosolids separation and voluminous biological sludge. The activated sludge plant at AIPM in Hedera, Israel, also reported
these problems in the past, when operated as the only biological process. The objectives of
this study were to evaluate the upgrading of the existing biological wastewater treatment
plant. The study covered two basic processes: a) Anaerobic bioprocess – for primary
decomposition of complex organic materials, which may disturb the following aerobic
process, adversely affecting bio-flocculation and the characteristics of the bio-sludge. The
anaerobic process will also reduce the total organic load on the following existing aerobic
process, making its current design more suitable to produce the required effluent quality. b)
Aerobic bioprocess – for the production of final effluent which will conform to current regulations and quality requirement for disposal to the Hedera River. This paper summarizes
the results obtained in the study based on a pilot plant including a membrane biological
reactor (MBR) compared to the “conventional” activated sludge process in the aerobic
treatment of the effluent obtained from the anaerobic reactor.
Wastewater originating from chemical industries may contain compounds, which could
Water Science and Technology Vol 50 No 3 pp 245–252 © 2004 IWA Publishing and the authors
Advanced treatment by anaerobic process followed by
aerobic membrane bioreactor for effluent reuse in paper
mill industry
245
N. Stahl et al.
adversely affect the treatment processes, mainly, the biological process by either toxic or
inhibition effects. Other types of effects include the damage caused by exposing biological
cells to hydrophobic compounds like phenol. This may impair the biochemical functions,
which are dependent on the intact state of membranes (Sikkema et al., 1992). However, the
impairment of cellular functions, following exposure to hydrophobic compounds, is a variable property and was found to be highly dependent on the growth rate. Fast-growing cells
of E. coli were found most susceptible to be damaged by hydrophobic compounds, when
compared to non growing cells (Sawada et al., 1987).
Usually, wastewater and effluent characterization is based on physical, chemical and
biological processes (Figure 1). The biological process deals with the organic phase, as
well as with part of the inorganic constituents, mainly nitrogen and phosphorous compounds. The high sensitivity of the bioprocess to some chemical compounds, which may be
found in the influent, often results in effluent characterized by high turbidity, high concentration of suspended solids, reducing the amount of active biomass in the bioreactor finally
leading to a complete failure of the treatment process. One of the most crucial and difficult
elements of the bioprocess is its ability to separate between the biosolids and the liquid
effluent phase. The use of membrane separation technologies (Figure 1B) has been adopted
and successfully implemented also in the biosolids separation, replacing the conventional
sedimentation (gravitational) process. The biosolids separation by membrane bioreactors
(MBR) which are basically MF and UF processes can thus remove particles in the range of
0.5–10 µm and 0.005–0.5 µm, respectively. (Cheryan, 1998).
Yao-po et al. (1998) studied MBR treatment of petrochemical wastewater and reported
removal efficiencies of 91% COD, 92% suspended solids, 99% turbidity, 82% phosphorous and 85% ammonia. Brindle and Stephenson (1996) worked with domestic and different types of industrial wastewater. They reported 85% removal of COD. The use of MBR
could enhance the removal of microorganisms, chlorinated aromatics, enzymes of cellulose, oil and grease as well as methanogenic bacteria, originating from anaerobic treatment
preceding the aerobic process. Saung-Goo and Hak-Sung (1993) and Peys et al. (1997)
indicated that MBR effluent could achieve superior quality levels: BOD 5 mg/L and total
suspended solids as low as 1 mg/L.
Due to the high efficiencies of the membranes in biosolids separation, the mixed liquor
(MLSS) in the bioreactor could be increased up to 20,000–30,000 mg/L, 10 times higher
compared to a conventional activated sludge process. These elevated biomass concentrations could considerably reduce the bioprocess residence time (thus dramatically reducing
required bioreactor volume), increasing cells residence time (CRT) in the system at the
same time. The high CRT values, which can be obtained in MBR, are expected to achieve
two important goals: a) the biosolids will be more stable as compared to biosolids from activated sludge with CRT in the range of 7 to 10 days; b) the total amount of excess sludge
produced will be reduced. These results could substantially reduce investment and operation costs of excess biosolids treatment and disposal, making the process cost effective as
compared to the conventional activated sludge.
Materials and methods
246
The pilot plant in the paper mill factory included three different treatment units in two
stages. First stage: anaerobic treatment utilizing the high rate technology, with internal
circulation (UASB IC) with a 60 litre bioreactor operated at 35–37oC, supplied by Paques.
Second stage: aerobic treatment was performed by two units operated in parallel: (a) conventional activated sludge, including a 200 litre aerated bioreactor with diffused air, a
settling tank for gravitational biomass separation, the air supply system, pumps and control
system supplied by Paques; (b) hollow fiber membrane bioreactor (MBR) with operational
Sedimentation
Tank
Aerated Bioreactor
Recycled
Biosolids
Thickening
E ffluent
Excess
Biosolids
Stabilization
Biosolids
Disposal
Dewatering
A. Biological process based on activated sludge only
Aerated Bioreactor
Sedimentation
Tank
Recycled
Biosolids
Thickening
E ffluent
Excess
Biosolids
Stabilization
N. Stahl et al.
Anaerobic
Biotreatment
Biosolids
Disposal
Dewatering
B. Biological process with gravitational separation of biosolids
Anaerobic
Biotreatment
Membrane
Aerated Bioreactor
Recycled
Biosolids
Effluent
Dewatering
Biosolids
Disposal
C. Biological process with membrane separation of biosolids (MBR)
Figure1 Biological treatment with different separation of biosolids
volume of 150 litre, including the aerated bioreactor, the integrated membrane cartridge,
the air supply, pumping and control systems, supplied by Zenon. The experimental system
was operated with flow rates up to 30 litre/hr.
The work included characterization of the influent and effluent from the first stage
(the anaerobic process) and from the second stage (the activated sludge and the MBR
processes). The analytical procedure was according to the Standard Methods for the
Examination of Water and Wastewater, 20th Edition (1998).
Results
The monitoring of the activated sludge full scale system in the period 10/2000–4/2001 was
based on 160 different composite samples and is summarized in Table 1 and in Figure 2.
During the above period the activated sludge was the only biological wastewater treatment
process at AIPM. The effluent quality clearly reflects relatively high suspended solids in
the effluent, on average 65 ± 122 mg/L.
Table 1 Aerobic process full size: activated sludge (10/2000–4/2001)
Parameter
pH
COD
BOD
Total
Total
TSS
mg/L
mg/L
mg/L
Influent
50% of all the results
80% of all the results
Average
Standard deviation (±)
6.8
7.0
6.8
0.6
2324
2612
2363
424
1135
1340
1115
277
138
208
223
408
Effluent
50% of all the results
80% of all the results
Average
Standard deviation (±)
Average removal (%)
7.7
7.9
7.7
0.2
–
195
298
245
181
90
14
32
21
15
98
30
92
65
122
–
247
600
Influent
Effluent
400
300
200
100
01
20
01
29
/0
4/
20
01
14
/0
4/
20
01
3/
/0
30
15
/0
3/
20
01
01
20
2/
/0
28
13
/0
2/
20
01
20
01
29
/0
1/
20
00
14
/0
1/
20
00
30
/1
2/
20
00
2/
/1
15
30
/1
1/
20
00
20
00
/1
1/
20
15
0/
20
/1
31
0/
/1
16
01
/1
0/
20
00
0
00
N. Stahl et al.
TSS (mg/L)
500
Figure 2 TSS concentrations in the activated sludge process full size system.
The monitoring of the anaerobic process pilot in the period 10/2000–3/2001 was based
on 135 different composite samples and is summarized in Table 2 and in Figure 3.
The monitoring of the activated sludge process pilot (after anaerobic process pilot) in the
period 11/2000–3/2001 was based on 95 different composite samples and is summarized in
Table 3 and in Figures 4 and 5.
Table 2 Anaerobic process pilot: upflow anaerobic sludge blanket (10/2000-3/2001)
Parameter
pH
COD
BOD
Sol.
Total
TSS
mg/L
mg/L
mg/L
mg/L
Influent
50% of all the results
80% of all the results
Average
Standard deviation (±)
7.3
7.4
7.0
0.3
2340
2656
2365
606
2108
2440
2124
529
1070
1435
1134
353
104
180
140
120
Effluent
50% of all the results
80% of all the results
Average
Standard deviation (±)
Average removal (%)
6.7
6.9
7.0
0.2
–
697
952
755
276
68
420
581
461
163
78
250
368
289
136
75
146
268
209
227
–
5000
4500
4000
3500
3000
2500
2000
1500
1000
500
0
Influent
Figure 3 COD concentrations in the anaerobic process pilot.
20
01
03
/
30
/
20
01
03
/
15
/
20
01
20
01
02
/
28
/
02
/
13
/
20
01
01
/
29
/
20
01
01
/
14
/
20
00
20
00
12
/
30
/
20
00
12
/
15
/
20
00
11
/
30
/
20
00
11
/
15
/
10
/
31
/
10
/
20
00
Effluent
16
/
01
/
10
/
20
00
COD Total (mg/L)
248
COD
Total
Table 3 Aerobic process pilot: Activated Sludge (AS) (11/2000–3/2001)
Parameter
pH
COD
COD
BOD
Total
Sol.
Total
TSS
VSS
mg/L
mg/L
mg/L
mg/L
mg/L
6.8
6.9
7.0
0.2
647
920
743
290
403
569
450
173
245
368
288
147
138
247
214
256
–
–
–
–
Effluent
50% of all the results
80% of all the results
Average
Standard deviation (±)
Average removal (%)
7.9
8.0
7.9
0.2
–
166
208
176
62
76
132
158
141
47
69
14
26
16
8
94
22
35
25
22
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
3500
4420
3300
1260
1420
3430
2510
930
Reactor
50%
80%
Average
Stdev
2000
Influent
1800
Effluent
1600
COD Total (mg/L)
N. Stahl et al.
Influent (after anaerobic process)
50% of all the results
80% of all the results
Average
Standard deviation (±)
1400
1200
1000
800
600
400
200
01
25
/0
3/
20
01
10
/0
3/
20
01
23
/0
2/
20
01
08
/0
2/
20
01
24
/0
1/
20
01
09
/0
1/
20
00
25
/1
2/
20
00
10
/1
2/
20
00
20
1/
/1
25
10
/1
1/
20
00
0
Figure 5 TSS concentrations in the effluent of the activated sludge process pilot.
00
1
03
/2
25
/
00
1
03
/2
10
/
00
1
02
/2
23
/
00
1
02
/2
08
/
00
1
01
/2
24
/
00
1
01
/2
09
/
00
0
12
/2
25
/
00
0
12
/2
10
/
11
/2
25
/
11
/2
10
/
00
0
160
140
120
100
80
60
40
20
0
00
0
TSS (mg/L)
Figure 4 COD concentrations in the activated sludge process pilot.
249
The monitoring of the MBR process pilot (after anaerobic process pilot) in the period
2/2001–5/2001 was based on 25 different composite samples and is summarized in Table 4
and in Figures 6 and 7.
Table 4 Aerobic process pilot: Membrane Bioreactor (MBR) (2/2001–5/2001)
Parameter
pH
N. Stahl et al.
COD
COD
BOD
Total
Sol.
Total
TSS
VSS
mg/L
mg/L
mg/L
mg/L
mg/L
Influent (after anaerobic process)
50% of all the results
80% of all the results
Average
Standard deviation (±)
6.8
7.0
6.8
0.2
651
1708
960
764
470
683
612
448
217
602
363
323
112
478
294
407
105
359
261
349
Effluent
50% of all the results
80% of all the results
Average
Standard deviation (±)
Average removal (%)
7.8
8.0
7.8
0.2
–
124
146
129
30
86
–
–
–
–
–
3.5
14.5
7.1
7.0
98
1.6
4.1
2.5
2.1
–
0.7
2.1
1.11
1.2
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
15579
20022
14315
6676
11920
16026
11197
5452
Reactor
50%
80%
Average
Stdev
3500
Influent
Effluent
COD Total (mg/L)
3000
2500
2000
1500
1000
500
1
02
/0
5/
2
00
1
22
/0
4/
2
00
1
/0
4/
2
00
1
12
02
/0
4/
2
00
1
/0
3/
2
00
1
23
13
/0
3/
2
00
1
/0
3/
2
00
1
03
21
/0
2/
2
00
1
00
/0
2/
2
11
01
/1
0/
2
00
1
0
Figure 6 COD concentrations in the MBR process pilot.
10
TSS (mg/L)
8
6
4
2
250
Figure 7 TSS concentrations in the effluent of the MBR process pilot.
20
01
05
/
02
/
20
01
04
/
22
/
20
01
04
/
12
/
20
01
04
/
02
/
20
01
03
/
23
/
20
01
03
/
13
/
20
01
03
/
03
/
20
01
02
/
21
/
20
01
02
/
11
/
01
/
10
/
20
01
0
Discussion
N. Stahl et al.
The monitoring of the activated sludge process in the period 11/2000–3/2001 led to the following conclusions. (1) COD reduction from 743 to 176 mg/L and BOD reduction from
288 to 16 mg/L, based on average values (76% and 94% removal respectively). (2) The
bioreactor could maintain about 2,500 mg/L of volatile suspended solids (MLVSS) and
a total suspended solids (MLSS) of 3,300 mg/L. (3) The effluent contained average TSS
values around 25 mg/L with a high standard deviation of 22 mg/L, thus indicating strong
fluctuation in solids separation efficiencies of the AS process, with conventional
secondary settling. (4) Events of bulking and/or voluminous–poor settling biomass could
be observed.
The most important goal of the anaerobic bioprocess – the substantial reduction of the
total organic matter, was successfully achieved by the pilot. Since 2002 a full scale treatment internally circulated up-flow anaerobic sludge blanket system has been operated at
AIPM-Hedera. The monitoring results obtained in the full scale system indicate good
accordance with the pilot plant. However the COD and BOD levels in the effluent clearly
indicate the need for additional biotreatment and this could be provided by additional aerobic bioprocess.
The monitoring of the MBR process in the period 2/2001-5/2001 lead to the following
conclusions. (1) The COD reduction was from 960 to 130 mg/L, and BOD from 363 to
7 mg/L (average removals of 86% and 98%, respectively). (2) The TSS in the effluent was
always lower than 5 mg/L with an average value of 2.5 mg/L; that means that all the quality
parameters reported for total values are very close to the soluble values. (3) The bioreactor
could maintain high levels of MLVSS (11,200 mg/L on average) resulting in long cell residence time (CRT) of the biomass in the MBR.
Conclusions
The task of the biological treatment process at a paper mill in Hedera, Israel has been
divided between a first stage anaerobic and a second stage aerobic treatment.
The comparison of activated sludge (AS) and membrane bioreactor (MBR) for the
second stage aerobic treatment revealed that the MBR could produce an effluent of much
better quality in terms of organic matter and suspended solids. The most important advantage of the MBR process is the very low content of suspended solids and low turbidity in the
effluent. This could save the need for further filtration in case of disposal of the effluent to
the river. It should be mentioned that AS effluent could not achieve a steady suspended
solids concentration lower than 10 mg/L, as required in case of ultimate disposal to the
river, therefore additional treatment by filtration would be necessary.
The MBR could concentrate over 3-4 times higher amounts of biomass, as compared to
the AS. This could have a direct influence on the cell residence time in the system, as well as
on the biosolids stability in terms of total to volatile suspended solids ratio (TSS/VSS).
It is most likely that the high quality of the effluent produced by the MBR technology
will promote a future project for the reuse and recycling of industrial effluents within the
paper mill for various production purposes.
Based on the results obtained in this study the recommended alternative of biological
treatment of the AIPM wastewater, was based on anaerobic treatment followed by MBR.
References
Blasczyc R., Gardner D. and Kosarik N. (1994). Response and recovery of anaerobic granules from shock
loading. Water Res., 28, 675–680.
Brindle K. and Stephenson T. (1996). Detoxification of industrial wastewaters in an extractive membrane
bioreactor, Wat. Sci. Technol., 33 (3) 1–8.
251
N. Stahl et al.
252
Casey E., Glennon B. and Hamer G. (1992). Oxygen mass transfer characteristics in a membrane-aerated
biofilm reactor. Biotech. and Bioeng., 62 (2) 183–192.
Chang, I-S., and Lee, C-H. (1998). Membrane filtration characteristics in membrane-coupled activated
sludge system – the effect of physiological states of activated sludge on membrane fouling.
Desalination, 120, 221–233.
Cheryan M. (1998). Ultrafiltration and Microfiltration Handbook. Technion Pub. Co.
Dubois M., Gilles K.A., Hamilton J.K. and Roberts J.K. (1956). Colorimetric methods for determination of
sugars and related substances. Analyt. Biochem., 28, 350.
Dufrense R., Lavalee H.C, Lebrun R.E. and Lo-Sung-Niehl (1998). Comparison of performance between
membrane bioreactor and activated sludge system for the treatment of pulping process wastewaters.
Tappi Jour., 81 (4), 131–136.
Ince O., Anderson G.K. and Kasapgil B. (1996). Composition of the microbial population in a membrane
anaerobic reactor system during start-up. Wat. Sci. Technol. 31 (1) 1–10.
Kennicut M.C. (1980). ATP as an indicator of toxicity. Water Res., 14, 325–328.
Peys K., Diels L., Leysen R. and Vandecasteele C. (1997). Development of a membrane biofilm reactor for
the degradation of chlorinated aromatics. Wat. Sci. Technol., 36 (1) 205–214.
Lee, S-G. and Kim, H-S. (1993). Optimal operating policy of the ultrafiltration membrane bioreactor for
enzymatic hydrolysis of cellulose. Biotechnol. and Bioeng., 42 (6), 737–746.
Livingston A.G., Freitas dos Santos L.M., Pavasant P., Pistikopoulos E.N. and Strachan L.F. (1996).
Detoxification of industrial wastewaters in an extractive membrane bioreactor. Wat. Sci. Technol., 33 (3)
1–8.
Rebhun M. and Galil N. (1988). Inhibition by hazardous compounds in an integral oil refinery. J. Wat.
Pollut. Control Fed., 60, 1953–1959.
Rozich A.F., Gaudy A.F. and D’Adamo P.C. (1985). Selection of growth rate model for activated sludge
treating phenol. Wat. Res., 19, 481–490.
Rosenberg M., Gutnic D. and Rosenberg E. (1980). Adherence of bacteria to hydrocarbons: a simple method
for measuring cell surface hydrophobicity. FEMS Microbiol. Letters, 9, 29–33.
Sawada T. and Nakamura Y. (1987). Growth inhibitory and lethal effects of ethanol in E. Coli. Biotechnol.
and Bioeng., 29, 742–746.
Schwartz-Mittelmann A. (1997). Investigations of disturbances to bioflocculation caused by phenol. D.Sc.
Thesis, Technion–IIT, Haifa Israel.
Sikkema J., Poolman B., Konings W.M. and de Bont J.A.M. (1992). Effects of the membrane action of the
tetralin on the functional and structural properties of artificial and bacteria membranes. J. of Bacteriol.,
174, 2986–2992.
Standard Methods for the Examination of Water and Wastewater (1998). 20th edn, American Public Health
Association/American Water Works Association/Water Environment Federation, Washington DC,
USA.
Starkey J.E. and Karr P.R. (1984). Effect of low dissolved oxygen concentration on effluent turbidity.
J. Wat. Pollut. Control Fed., 56, 837–843.
Yao-po F., Ju-si W., and Zhao-chun J. (1998). Test of membrane bioreactor for wastewater treatment in a
petrochemical complex. J. of Environ. Science, 10 (3), 1–9.