8
Pergamon
Wal. Sci. Tech. Vol. 40,No. I, pp. 145-152, 1999 «:II9991AWQ
Published by ElsevierSCience LId Printed in Great Britam. Allrights reserved 0273-1223199520.00 + 0.00
PH: S0273-1223(99)00374-1
F. Gennirli Babuna*, B. Soyhan**, G. Eremektar* and D.Orhon* • Environmental Engineering Department, Istanbul Technical University, LT. U. In oatFakiUtesi, 80626 Maslale. Istanbul, Turkey .. Environmental Engineering Department, Istanbul University, LU. Cevre Muhendisli i Bolumii, 34850 Avclor, Istanbul, Turkey
ABSTRACT The studyemphasizes wastewater characteristics of two different textile plants as they apply to biological treatment. Although conventional characterization reveals no major differences, theeffiuents from theacrylic fiber and yam dyeing plant exhibit all the properties of a non-biodegradable wastewater. Appropriate pretreatment consisting of partial chemical oxidation with HZ02 reduces its COD content to 700 mg r', almost entirely biodegradable whereas COD fractionation indicates that the effiuents of the cotton knit dyeing plant contain 9% residual fractions. Experimental investigation shows that most kinetic and stoichiometric properties of both wastewaten are compatible with that of domestic sewage with the exception of a muchslower hydrolysis rate. C 1999 IAWQ Published by Elsevier Science Ltd. All rights reserved.
KEYWORDS Activated sludge; conventional characterization; chemical treatability; COD fractionation; process kinetics; textile wastewaters. INTRODUCTION The textile industry has a variable nature in terms of raw materials used, techniques employed, chemicals applied, operations and processes involved. The reflections of this variation can be observed on the wastewater characterization and treatability. The differences in the wastewater characterization and treatability of textile wastewaters originating from different sources cannot be evaluated without giving reference totheeffiuent producing processes andtheonlywayto perform thisis to consider each industry as a separate case. Although a common practice, the application of only chemical treatment usually proves unreliable for the required pollutant removal levels of textile wastewaters, since theremoval efficiencies are variable andgenerally unpredictable. On theotherhand, partial chemical treatment can be applied before or after thebiological treatment to increase thebiodegradability of theraw wastewater or to polish the effluent of thebiological treatment. In thiscontext, a comprehensive treatability evaluation of two different textile mill effiuents is carried outin this study. Cotton knit fabric dyeing and finishing operations are conducted in the first plant where three different process flow schemes are followed: bleaching-dyeing-rinsing; optical whitening-rinsing and rinsing. Dyeing is performed in batch, overflow machines by the use of reactive dyes. The second plant 145
146
F. GERMIRLI BABUNA et al.
involves yarn dyeing operations and carpet and blanket production from acrylic fibers. The wastewater of this textile mill originates from the batch processes where acrylic fibers and yarns are subjected to dyeing with cationic (basic) dyes. In this plant fiber dyeing operations are carried out in fiber dyeing machines whereas yarns are dyed in hank dyeing machines. Schematic flowcharts of the processes and related wastewater sources for the investigated plants are given in Fig. 1. It must be noted that the experimental investigation on biological treatability carried out in this study covers a detailed COD fractionation and the determination of major kinetic and stoichiometric coefficients by the use of respirometric measurements. A multi-component endogenous decay model is applied to obtain the kinetic interpretation and design of activated sludge processes.
CONCEPTUAL APPROACH TO BIOLOGICAL TREATABILITY Traditional mathematical models formulated for overall substrate and biomass parameters only, were limited to include two processes, namely bacterial growth and decay. Since the interpretation of biological treatability with such models is reported to be at least inadequate, if not misleading, recently multicomponent modelling approaches have been developed (Henze et. al.. 1987; Orhon and Artan, 1994). The use of multi-component models on the other hand, necessitates correct experimental information on wastewater characterization and kinetic and stoichiometric coefficients. Reliable wastewater characterization in terms of organic matter is essential in the application of multi-component models. The COD parameter does not yield sound information, because it covers biodegradable fractions each having different rates of biodegradation and non-biodegradable fractions of influent origin or generated by activated sludge as residual organic products (Daigger and Grady, Jr., 1977; Orhon et al.. 1989). Therefore, wastewater characterization especially for industrial effiuents should consider all the COD fractions as readily biodegradable COD, Ss; slowly biodegradable soluble COD, SH; slowly biodegradable particulate COD, Xs; -initially inert soluble COD, Sf; initially inert particulate COD, XI; particulate inert organic products, Xp; soluble inert organic products, Sp. The multi-component endogenous decay model based on growth, decay and hydrolysis processes used in this study basically identifies all the COD fractions together with the active heterotrophic biomass, XH and dissolved oxygen, So (Orhon and Artan; 1994). In the multi-component endogeneous decay model, the generation of inert endogenous mass or particulate inert organic products, Xp, and the decay of active biomass are interlinked according to the viability concept. The model is so structured that two fractions fES and fEX of the endogenous biomass is not oxidized, but converted into soluble and particulate inert products, SpandXp. Evaluation of activated sludge treatment for carbon removal on the basis of the above mentioned endogenous decay model also requires reliable information on kinetic and stoichiometric constants specifically related to each textile effiuent, namely fEX, fES, YH, AH' Ks, bH, kh and Kx. fEX is generally accepted as 0.2 g COD/g cell COD in all similar activated sludge models (Henze et al., 1987). YH, bH and AH can be determined from respirometric measurements by following the experimental procedures given below. When Ysp is defined as a fraction of influent biodegradable COD (Sp=YsP.CS\), fES is calculated as a function ofY H and Ysp from mass balance (fES=YSP/YH). For the determination of unknown Ks, kh, and Kx coefficients curve fitting through model simulation is applied to the experimentally determined OUR profiles. EXPERIMENTAL APPROACH All analyses in terms of conventional parameters were performed as defined in Standard Methods, (1995). The filtrates of the samples subjected to vacuum filtration by means of Whatman GF/C glass fiber filter papers were defined as "soluble" fractions. Whatman GF/C glass fiber filters were also used in the determination ofVSS and SS. Wastewaters subjected to experiments were obtained as composite samples.
Reacuve Dye
em..s.: ProofiogA p l ! - - - - ,
Soda--, Acetic Acid------,
Peroxide
CaUSlic Water
PLANT
Coltoo
1fa
Knil Fabric:
Product
Wastewater OpticaIWhiteniogApts
Wastewater
Wastewater
Wastcwata
Wastewater
Wastewater
~
i
a-
I
Crease ProofingAgents Pero• •de
Wastewater
Wastewater
~ s:
J
::s
Wellmgagents
Caustic
e ....
g
Water
Wafer
~
: . Product
PLANT lib Wastewater
Wastewater
Basic Dye
Wastewater
Producl
Knil Fabric
Wastewater
Basic Dye
i
Formic Acid
1fc
CoUoo
Wastewater
I
Retarder ...
PLANT
Formic Acid
~
I
§
I
_
Softening Agenl---
PLANT
I
2IB
PLANT
I CentrifUr Dyed Fiber
1JA
Wastewater
Wastewata
~
Softenmg Agent
Water
Dyed Yam
Ac:ryUc
Yarn
Processes lIa : bleaching-dyeing-rinsing lib: optical whitening-rinsing IIc: rinsing VA: acrylic fiber dyeing 2!B: acrylic yam dyeing
~
~o .... ~
" ~
Wastewater
Figure 1. Schematic Process Flowcharts and Related Wastewater Sources of the Investigated Textile Plants
...
~
148
F. GERMIRLI BABUNA et al.
Chemical treatability As will be mentioned later in detail, microbial acclimation was not achieved for the raw wastewaters of Plant 2 and therefore this sample was subjected to chemical precipitation and oxidation tests in order to get a partial chemical treatment prior to biological treatment. Chemical precipitation experiments were performed using conventional jar test procedures. The pH of the samples was adjusted by the addition oflime. Alum, FeCb.6H20 and FeS04.7H 20 with and without anionic, cationic and nonionic polyelectrolytes were investigated as coagulants. After a minute of flash mixing each sample was stirred at 20 rpm and allowed to settle for one hour. Then, aliquots of the supernatant were withdrawn for COD analyses. Chemical oxidation tests were applied in one-liter beakers where 50% H202 solution was added as oxidant. FeCb .6H20 was used as a catalyst. The oxidation was allowed to proceed at pH 3.5. Magnetic stirrers were used for mixing. After oxidation, the excess H202 was removed by adjusting the pH to around 9 and mixing the oxidized sample, in order to hinder H202 interference on COD measurements. Then, aliquots were withdrawn for COD analyses. pH adjustments were performed by the addition of either H2S04 or NaOH. Biological treatability For the wastewaters originating from Plant I, the inert COD components XII and SII were determined according to an experimental procedure proposed in literature (Germirli et al., 1993). The inert COD test involved two aerated batch reactors, of 3 liter volumetric capacity each, one fed with the unfiltered wastewater, diluted to have an appropriate initial COD concentration, Cn, in the range of 1500-2000 mg 1'1, and the other with the filtered wastewater, Sn, having the same dilution. The seed was obtained from a labscale fill and draw aerobic reactor operated under steady state with the same wastewater. On the other hand, since the wastewaters of Plant 2 after passing through partial chemical oxidation were found to be solely soluble in nature, the inert soluble COD component 811 was determined by following another procedure given in literature (Germirli, 1990; Germirli et al., 1991). In this test. again two aerated batch reactors having 3 liter volumetric capacity, one fed with the filtered wastewater and the other with glucose, both started with the same initial COD concentration were run. The seed was obtained from a lab-scale fill and draw aerobic reactor operated under steady state with 50 % wastewater + 50 % glucose. For both of the inert COD assessment tests, the microbial seed was added to secure an initial biomass concentration of 40 mgl" VSS in reactors. Aliquots removed periodically from the mixed liquor were analyzed for total (first inert COD assessment test) and soluble (first and second inert COD assessment tests) COD. Experiments were continued until the observation of a stable COD plateau coupled with no appreciable biomass activity. The respirometric procedure for the assessment of major kinetic and stoichiometric constants such as the readily biodegradable COD, Ss, the maximum heterotrophic growth rate, ~H' and the heterotrophic yield coefficient. Y H, involved using 1 I batch reactors. For the determination of OUR, the reactor was initially fed with the wastewater sample, seeded with appropriate biomass to start with a suitable initial FIM (Cn/XTI ) ratio and constantly aerated to maintain a dissolved oxygen concentration of 6-8 mg 1'1. The biomass was previously acclimated to the same sample in a fill and draw reactor operated at a sludge age of around 10 days. The readily biodegradable COD was determined in accordance with the method suggested by Ekama et al.• (1986). For the assessment of the maximum heterotrophic growth rate, the OUR reactors were run at a FIM ratio of 4-5 g COD/g VSS as recommended by Kappeler and Gujer (1992). The heterotrophic yield was evaluated by comparing the OUR and COD profiles obtained on the same sample in accordance with the method proposed by Ubay Cokgor, (1997). The determination of the endogenous decay coefficient. bH involved removing an activated sludge sample and aerobically digesting it over a period of several days (Ekama et al.. 1986). OUR measurements were conducted with a WTW OXI 2000 oxygen meter and recorder . In the experiments the samples were adjusted to a pH of 7·8, a range suitable for biological activity.
Evaluation of treatability of two textile mill effluents
149
RESULTS AND EVALUAnON Wastewater characterization and pollution profiles The conventional wastewater characterization of the general discharge together with the characterization of segregated streams originating from each process for investigated plants are given in Table 1. It is observed that the segregated stream pollutant concentrations are in agreement with the results obtained from the general discharge. Table 1. Conventional wastewater characterization of the investigated plants Plant 1 Segregated Streams optical wh iteningbleachingrinsing dyeing-rinsing
Parameters (mg 1'1)
2240 2145 13 4.4 125 75
Total COD Soluble COD
TKN
TP SS VSS Chromium 1.1 pH (pH units) 10.3 l volume(m datI) 1200 - used for carpet and blanket production
3670 3415 40 9.4 325 180 1.5 10.5 60
rinsing
General Discharg
e 2215 1820 5 6.3 105 40 0.3 6.8 30
2310 2185 14 4.5 135 80 1.1
10.1 1290
Plant 2 Segregated Streams fiber Yarn dyeingdyeing
1885 1693 72 4.2 75 40 1.0 4.5 65
1730 1350 75 4.5 105 45 0.2 4.5 100
General Discharge
1900 1590 72 4.2 90 43 0.6 4.5 165
The soluble COD values account for 9S and 84 % of the total COD for Plant 1 and Plant 2 namely. Both of the wastewaters are associated with low nitrogen and phosphorus contents . The pollution profiles are tabulated in Table 2. It is interesting to note that the wastewater characteristics of the segregated streams of Plant 2 show very similar characteristics. At first glance it seems that schematic flowcharts given in Figure 1 also support the same evidence. However, when the concentrations of the used chemicals (retarder, formic acid etc.) and the produced wastewater flowrates are considered, a different picture appears as shown in unit loads given in Table 2. Table 2. Pollution profiles Plant! Process
processed wastewater material flowrate (ton day")
lIa lib lIc 2/A 2IB
12 1.5 1.5 13 5.6
unit flowrate load
unit COD load
(m] day")
(m'wastewater.ton processed material' I)
(kgCOD ton processed material")
1200 60 30 6S 100
100 40 20 5 18
224 147 44 9.4 31
a.bleaching-dyeing-rlnsing; b: optical whitening-rinsing; c::rinsing A: fiber dyeing for c:arpet and blanket production; B: yarn dyeing
When the unit loads of bleaching-dyeing-rinsing flow scheme of Plant 1 (lIa) are compared with another industry following the same production route (Germirli Babuna et al., 1998), it is observed that higher figures are obtained from this study indicating 7!'ical variations from one application to the other. Unit loads of 80 m) wastewater.ton processed material' and 118 kg COD.ton processed material" are given in the mentioned literature. Treatability studies The results of the microbial acclimation experiments showed that it was not possible to apply directly biological treatment to the raw wastewaters originating from the second plant. Similar practices are also
F. GERMIRLI BABUNA et al.
ISO
reported in literature, mainly due to the toxic effect of the dyes, and retarders used in acrylic dyeing operations (Kabdash et al., 1995). Therefore, chemical treatability studies consisting of chemical precipitation and chemical oxidation, were performed on this wastewater.
Chemical treatability. Chemical precipitation with alum, FeS04 and FeC!) were applied to the general discharge sample of Plant 2. The highest COD removal efficiency obtained is 41%, which is achieved at a FeC!) dosage of 400 mg r' under optimum pH of9.5. It is observed that the addition of anionic, cationic and nonionic polymers do not increase the COD removal. Unfortunately, no microbial acclimation is achieved on chemically precipitated wastewater sample of Plant 2. As a result, chemical oxidation tests decided to be run on the raw wastewaters of this plant. It must be noted that, not to get a complete oxidation but a partial oxidation prior to biological treatment is the aim of the applied chemical oxidation. The optimum conditions of partial chemical oxidation is observed to be as follows: 500 mg r' Fe+3 catalyst, a day of reaction time, pH of3.5. COD removal efficiencies with varying oxidant dosages are given in Table 3. Table 3. COD removal efficiencies with varying oxidant dosages COD removal efficiency (%) 48 53 63 65 67
0.5 0.75 1.0 1.5 2
H20:z/COD ratio of 1.0 is selected to represent partial oxidation conditions to which biological treatment will be applied. All the biological treatability studies given below are conducted on 'partially oxidized' wastewater from Plant 2 and the wastewater characteristics of this sample are tabulated in Table 4. Table 4. Wastewater characterization of chemically oxidized wastewater of Plant 2 Parameters Total COD Soluble COD TKN TP Detergent Chromium
Concentration (mg 710 700 65 3.4 12 0.4
r I)
Biological treatability. According to the results of COD fractionation tests given in Table 5, Plant 1 can be characterized as: 9% inert, 18% readily biodegradable and 73% slowly biodegradable COD. On the other hand, the partially oxidized wastewater of Plant 2 shows a totally different COD characterization as expected. The COD of this wastewater has a predominantly biodegradable characteristic, having only a 4% inert fraction. Table 5. Experimental results related to COD fractionation Industry
CTI mgl'
m~l
SSI mgl"
Sil SHI mgl"
mgl'
Plant I" 2300 1900 170 1310 420 Plant11 00 596 710 700 28 86 • raw wastewater OOeffluent of the partial chemical oxidation with HzOz
XS1 mgl"
Xu mgl"
365
35
SS'/CT
Su/CTJ
SHI/CTJ
XS1/CTJ
Xu/CTJ
0.16
0.02
0.18
0.07
0.57
0.12
0.04
0.84
The kinetic and stoichiometric coefficients of Plants 1 and 2 are summarized in Table 6. The values displayed in this table showed that coefficients related to microbial growth kinetics, namely Y H, iiH' Ks, are very similar for both plants and comparable with those associated with domestic wastewater. This
Evaluation oftrcatabiJityof two textile mill effluents
151
observation is quite explainable on the basis of the adopted model where microbial growth depends solely on readily biodegradable substrate which is of approximately the same composition (small molecule compounds, simple carbohydrates) regardless of the type of the waste. The slightly lower values for ~H may be attributed to inhibitory compounds likely to be present in such wastewaters. It should be noted that the sample tested for Plant 2 is the effluent of the pretreatment by partial chemical oxidation with H202. A similar statement can also be made for endogenous decay. The striking property of these wastes may be visualized by rate coefficients defining much slower hydrolys is of the slowly biodegradable compounds. In this framework, the significant positive effect of partial chemical oxidation on the hydrolysis rate should also be mentioned. Table 6. Kinetic and stoichiometric coefficients Plant
Plant 1* Plant 11**
fES
PH d'i
0.62
0.063
0.60
0.089
T °C
YH gCOD/gcell COD
16 16
Ks mgl"
bH d,l
kt, d'l
Kx gCOD/gcell COD
3.2
13
0.19
0.8
0.7
3.9
10
0.17
1.6
0.7
• raw wastewater Ueffiuentof the partialchemicaloxidation withHlOl
CONCLUSIONS The study emphasizes characteristics of acrylic fiber and yarn dyeing and cotton knit fabric dyeing effluents as they affect biolog ical treatment. Conventional characterization defines both wastewaters as typical textile effluents of similar properties , with COD contents of 1900 - 2310 mg r', low suspended solids, nitrogen and phosphorus concentrations. Although conventional characterization reveals no major differences, biological treatability tests indicate that the effluents from acrylic fiber and yarn dyeing plant exhibit all the properties of a nonbiodegradable wastewater, mainly due to dyes and auxiliary chemicals such as retarders associated with the process. Partial chemical oxidation with H202 is experimentally tested on the wastewater originating from this plant and prescribed as effective pretreatment, reducing the total COD from 1900 mg r' to around 700 mg r'. On the basis of COD fractionat ion, it is observed that the effluents from the cotton knit fabric dyeing plant contain a total residual COD fraction of 9 %, with a 7 % soluble portion, while pretreatment renders the wastewaters from acrylic fiber and yam dyeing plant much more compatible with biological treatment. Experimental results related to stoichiometric and kinetic properties show that most coefficients may be defined by values similar to those associated with domestic sewage. Hydrolys is of the slowly biodegradable COD however appears to proceed at a much slower rate, with a significant difference between the two wastewaters, reflecting the expected beneficial effect of partial chemical oxidation. NOMENCLATURE bH Cn fEX fES kh
Ks Kx OUR SH Sf So
endogenous decay rate total influent COD particulate inert fraction soluble residual fraction maximum specific hydrolys is rates half-saturation constant half-saturation constants for hydrolysis oxygen uptake rate rapidly hydrolyzable COD soluble inert COD oxygen concentration
F. GERMIRLI BABUNA et al.
soluble residual COD generated as metabolic products readily biodegradable COD influent soluble COD active heterotrophic biomass particulate inert COD particulate inert metabolic products slowly hydrolyzable COD yield coefficient soluble residual product ratio maximum specific growth rate ACKNOWLEDGEMENTS This study was conducted as part of the research activities of the Environmental Biotechnology Center of the Scientific and Technical Research Council of Turkey. REFERENCES Daigger, G. T. and Grady, J. P. L. Jr. (1977) A model for biooxidation process based on product formation Concepts. Wat. Res. II, 1049-1057. Ekama, G.A., Dold, P.L. and Marais, G.v.R., (1986) Procedures for determining influent COD fractions and the maximum specific growth rate ofheterotrophs in activated sludge systems. Wat. Sci. Tech., 18,91-114. Germirli, F., Orhon, D., Artan, N., Ubay, E., and GOrgon, E. (1993) Effect of two-stage treatment on the biological treatability of strong industrial wastewaters. Wat. Sci. Tech, 28(2),145·152. Germirli, F., Orhon, D. and Artan, N. (199 I) Assessment of the initial inert soluble COD in industrial wastewaters, Wat. Sci. Tech., 23, Kyoto, 1077·1086. GermirJi, F. (1990) The Incremental and Comparison Methods for the Assessment ofInitial Soluble Inert COD. Ph.D. Thesis, 148 pages, Istanbul Technical University. Germirli Babuna, F., Orhon, D. Ubay Cokgor, E., Insel, G. and Yaprakh, B. (1998) Modelling of activated sludge for textile wastewaters, Wat. Sci. Tech., 38(4-5), 9-17. Henze, M., Grady, C.P.L.Jr., Gujer W., Marais, G.v.R., and Matsuo, T., (1987) Activated Sludge Model No.1, IAWPRC Sci. and Tech. Report No.1, IAWPRC, London. Kabdash, I., Tunay, 0., Artan, R. and Orhon, D. (1995) Acrylic dyeing wastewaters characterization and treatability. Proc. Jrd Int. Conference Appropriate Waste Management Technologies for Developing Countries, 239, Nagpur, India, Kappeler, J., and Gujer, W., (1992) Estimation of kinetic parameters of heterotrophic biomass under aerobic conditions and characterization of wastewater for activated sludge modelling, Wal. Sci Tech., 25(6), 125-139. Orhon, D., Artan, N. and Cimsit, Y. (1989). The concept of soluble residual product formation in the modelling of activated sludge. Wat Sci. Tech.. 21, 339-350. Orhon, D. and Artan, N., (1994) Modelling ofActivated Sludge Systems. Technomic Press, Lancaster, PA. Standard Methods for the Examination of Water and Wastewater (1995). 19th edn, American Public Health Association!American Water Works Association/Water Environment Federation, Washington DC, USA. Ubay ~okgOr, E. (1997) Respirometric Evaluation of Process Kinetic and Stoichiometry for Aerobic Systems, Ph.D. Thesis, Istanbul Technical University (in Turkish).