Chemosphere 44 (2001) 1055±1063

www.elsevier.com/locate/chemosphere

Regulating colored textile wastewater by 3/31 wavelength admi

methods in Taiwan
C.M. Kao *, M.S. Chou, W.L. Fang, B.W. Liu, B.R. Huang
Institute of Environmental Engineering, National Sun Yat-Sen University, Kaohsiung 804, Taiwan, ROC
Received 16 May 2000; received in revised form 1 August 2000; accepted 4 August 2000

Abstract

The wastewater from textile dyeing facilities is dicult to treat satisfactorily because of high compositional vari-
ability and high color intensity. To reduce colored e‚uents discharged into watercourses, the government of Taiwan
adopted the E‚uent True Color Standard in 1998. The true color discharge limit is 400 American Dye Manufactures
Institute (ADMI) units. The adopted analytical method is the ADMI Tristimulus Filter Method (3 wavelength (WL)
method), and the 31 WL ADMI method might be also adopted as an alternative for color value measurement. The
refractory nature of textile dyes and the introduction of this new regulation present an environmental challenge to the
Taiwanese textile industry. The main objectives of this study were to (1) evaluate the ecacy of current wastewater
treatment systems for controlling the colored textile wastewater discharges, and (2) evaluate the correlations between 3
and 31 WL ADMI methods. Ten representative textile wastewater treatment facilities employing biological and
chemical coagulation treatment technologies were selected to perform a 10-consecutive-day e‚uent sampling and
analysis. Results show that a signi®cant di€erence between 3 and 31 ADMI methods was observed. These two ADMI
methods cannot be substituted for each other, and the discharge standard should be determined based on the selected
testing method. Investigation results also suggest that the commonly used wastewater treatment technology (biologi-
cal + chemical coagulation) fails to e€ectively remove dye from the colored textile wastewater. Sodium hypochlorite
(NaOCl) addition was applied by most facilities as the temporary post-polishment step to comply with the color dis-
charge standard. Ó 2001 Elsevier Science Ltd. All rights reserved.

Keywords: ADMI; Textile wastewater; 3 wavelength method; True color; Wastewater treatment

1. Introduction color measurement, SS and turbidity removal are re-

Color in water may result from the presence of nat-
urally occurring color substances or the colored indus-
trial wastewater. In general, color in water is divided
into true color and apparent color (Bennett and Drikas,
1993; APHA, 1995). The apparent color is determined
on the original sample without the recovery of sus-
pended solids (SS) or turbidity. However, for the true
*
Corresponding author. Tel.: +886-7-525-2000 ext. 4413;
fax: +886-7-525-4449.
E-mail address: jkao@mail.nsysu.edu.tw (C.M. Kao).
quired. The characteristics of colored industrial e‚uents
are highly erratic in both the hues and concentrations of
color. It is more complicated to quantify colored in-
dustrial e‚uents than naturally occurring color sub-
stances. Colored industrial e‚uents present aesthetic
and environmental problems by absorbing light and by
interfering with aquatic biological processes. Colored
pollutants have also been found to be toxic and car-
cinogenic (Pierce, 1994). Moreover, unnatural colors are
also a cause of complaint from the public. In order to
reduce colored e‚uents discharged into watercourses,
enforcement of the e‚uent color standard has become a
concern for legislators in many countries. Di€erent

0045-6535/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 0 0 ) 0 0 5 0 2 - 6
1056 C.M. Kao et al. / Chemosphere 44 (2001) 1055±1063
analytical methods and corresponding limits to control
color in the e‚uent have been developed (Allen et al.,
1973; EPA, 1978; APHA, 1995). Both visual comparison
and spectrophotometric methods are employed as stan-
dard analytical methods (Allen et al., 1973; APHA,
1995).
The method adopted in Taiwan before 1998 was a
transparency method with a discharge limit of more
than 15 cm. In the transparency method, the sample is
directly analyzed for its apparent color (Kang and Kuo,
1999). Because the hue, concentration of color, and SS
in lakes are stable, the transparency method is widely
used as an indicator for evaluating lake eutrophication
(Kang and Kuo, 1999). However, industrial e‚uents are
highly variable in both the composition and concentra-
tion of SS and color. The color characteristics of in-
dustrial wastewater depend on the wastewater type and
the performance of wastewater treatment plants; a sig-
ni®cant deviation from lake water. The transparency
method is a subjective visual comparison analysis.
Therefore, it is an unsuitable method for colored in-
dustrial e‚uent regulation. The government of Taiwan
began enforcement of the true color discharge standard
in 1998. The discharge limit is 400 American Dye
Manufactures Institute (ADMI) units. The adopted
analytical method is the ADMI Tristimulus Filter
Method (3 wavelength (WL) method), and the 31 WL
ADMI method might be also adopted as an alternative
method for color value measurement. The Taiwanese
textile industry is challenged by the introduction of this
new government regulation of colored wastewater dis-
charges.
The ADMI 3 WL method is an extension of the
Tristimulus Method described in Standard Methods
(APHA, 1995). This method measures the sample's true
color, independent of hue. Based on the Adams±Nick-
erson chromatic value formula for calculating single
number color di€erence values, i.e., uniform color dif-
ferences, this method is applicable to colored waters and
wastewaters similar in hue to the standards and with
color characteristics signi®cantly di€erent from plati-
num-cobalt standards (APHA, 1995; Kang and Kuo,
1999). Sample pretreatment is required to obtain the
true color ADMI value. Membrane and glass ®ber ®lters
with functional pore sizes of approximately 0.45 lm are
used for turbidity removal. The sample pH is adjusted to
7.6 before ®ltration. The pretreated sample is then ex-
amined in a spectrophotometer to determine the light
transmittance under three designated wavelengths (590,
540, and 438 nm). The 3 WL ADMI value is then de-
termined following the calculation procedures described
in Standard Methods (APHA, 1995). The 31 WL ADMI
value is also determined spectrophotometrically, using a
spectrophotometer with a narrow (10 nm or less) spec-
tral band and an e€ective operating range of 400±700
nm (APHA, 1995).
The ADMI color value provides a true measure of
water color, independent of hue and thus opens the way
to the more accurate de®nition of water and wastewater
color. Both 3 and 31 WL ADMI methods are approved
by the ADMI and adopted for color measurement by
some countries and local governments (EPA, 1978;
Kang and Kuo, 1999). However, the reliability and
correlation between 3 and 31 WL methods are still un-
clear. Therefore, one of the objectives of this study was
to evaluate the correlations between these two methods.
Unlike the government of Taiwan, some countries
or local governments have only site-speci®c standards or
``narrative'' regulations such as ``no objectionable or
unpleasant color in the e‚uent''(EPA, 1986; Kang and
Kuo, 1999). For example, the State of North Carolina
and the US Environmental Protection Agency (US
EPA) have not adopted regulations for e‚uent color
(EPA, 1986; NCDEHNR, 1999). However, the State of
Rhode Island has settled a true color discharge standard
of 200 ADMI units (3 WL method), and the State of
Wisconsin has a ¯exible true color discharge limit in the
range between 225 and 600 ADMI units (Wisconsin
State Government, 1997; Kang and Kuo, 1999).
The textile industry has traditionally been one of the
most important industries in Taiwan, creating thousands
of jobs. Currently, there are more than 250 textile dyeing
and processing factories in Taiwan. Textile wastewater is
a potential source of pollution, and could have deteri-
orated e€ects on the ecology of the region. In general,
textile wastewater contains considerable amounts of
poorly biodegradable substances such as sizing agents,
dyes, surfactants, volatile organic compounds, disodium
terephthalate, and ethylene glycol. Thus, textile waste-
water is one of the more dicult industrial wastewaters
to treat (Ingamells, 1993; Pierce, 1994; Cooper, 1995).
Most textile dyeing mills employ treatment facilities
composing of pretreatment and combinations of bio-
logical and physical/chemical precipitation processes
(Lemordant et al., 1989; Kennedy et al., 1992). The
textile industry in Taiwan as a whole has had a satis-
factory record of compliance with environmental regu-
lations. But as pollution regulations become more
stringent, the textile industry is being forced to discharge
e‚uents of higher quality. One of the pollution control
parameters of concern to textile dyeing operations is
color. Before the enforcement of color discharge stan-
dard in 1998, most of the textile dyeing e‚uents could
meet the 15 cm transparency discharge limit. However,
the textile industries have diculties meeting the new
color discharge regulation (400 ADMI measured by 3
WL method) using biological and chemical coagulation
treatment technology. Failure to meet the regulation
may result in ®nes starting at US $2000 per incident.
The major objectives of this study include the fol-
lowing: (1) investigation of the commonly used waste-
water treatment schemes (combination of biological and
 C.M. Kao et al. / Chemosphere 44 (2001) 1055±1063
chemical coagulation processes) at representative textile
factories; (2) evaluation of the ecacy of using biologi-
cal and chemical coagulation treatment systems for
managing colored textile wastewater discharges; (3)
evaluation of the feasibility of colored textile e‚uents
regulation under the new discharge standard and
adopted ADMI method; and (4) evaluation of the cor-
relations between 3 and 31 WL ADMI methods and
evaluation of other factors such as analytical instru-
ments and pH adjustment on the color measurements.
2. Materials and methods
2.1. Textile factories investigations
In the ®rst part of this study, 10 representative
wastewater treatment plants with biological, mainly ac-
tivated sludge process, and chemical precipitation as the
major treatment processes were selected for a detailed
facility investigation. These investigations contained the
following tasks: (1) collection in¯uent and e‚uent
wastewater samples from each treatment stage for 10
consecutive days, (2) sample analysis for COD and 3/31
WL ADMI color values, (3) comparison of the deviation
in 3 and 31 ADMI values, and evaluate the correlations
between the two ADMI methods, (4) evaluation of the
correlations between COD and ADMI values, and (5)
con®rm the presence or absence of temporary post-
treatment process taken by each plant to meet the dis-
charge limit.
In the second part of this study, a total of 186 textile
wastewater samples was collected from the e‚uents of
75 textile wastewater treatment plants. Each e‚uent
sample was analyzed for ADMI color values by both 3
and 31 WL methods. These data were used to (1) eval-
uate the e€ectiveness of using current treatment systems
on controlling the colored textile wastewater discharges
and meeting the color discharge regulation, and (2)
demonstrate the suitability of the current discharge
standards, and determine if less conservative regulatory
decision is acceptable at this stage.
2.2. Wastewater sample collection and analysis
In the ®rst part of this study, wastewater samples
were collected (grab samples) from 10 representative
plants for 10 consecutive days, and measured for both
COD and ADMI color values. COD measurements
were conducted in accordance with the dichromate re-
¯ux method described in Standard Methods (APHA,
1995). The ADMI color value for each sample was
measured via the following procedure: sample pH ad-
justment; ®ltration (0.45 lm); spectrophotometer; and
ADMI measurement.
1057
For the pH adjustment step, two di€erent pH values
were chosen: original (ambient) pH and pH 7.6 (APHA,
1995). For the spectrophotometer step, two types of
instruments were used: Hach DR4000 (Hach) and Shi-
madzu UV-160A spectrophotometer (Shimadzu). For
the ADMI measurement step, both 3 and 31 WL ADMI
color values were measured. Thus, three controlling
factors (pH, instrument, and WL) were evaluated to
determine their e€ects on color value measurements.
Sample preparation and analytical procedures of the
ADMI color value measurement are described in the
Standard Methods and EPA Method 110.1 (EPA, 1978;
APHA, 1995).
Hach DR4000 was used for ADMI measurement
comparison because it has some unique features: stan-
dard curves for ADMI color measurements and complex
ADMI value calculation equations are installed in this
instrument, and therefore, frequent standard ADMI
color solution (potassium chloroplatinate and cobaltous
chloride solution) preparation (APHA, 1995) and
ADMI value calculation are not required. Each ADMI
value can be obtained immediately by this instrument
after the ®ltration procedure. However, the analytical
accuracy of the measurement by Hach DR4000 was
unclear. Thus, the variations in ADMI values between
two instruments used in this study were evaluated. In the
second part of this study, the 186 e‚uent samples col-
lected from 75 textile wastewater plants were analyzed
for both 3 and 31 WL ADMI color values.
3. Results and discussion
3.1. Analysis of samples from 10 representative plants
3.1.1. Treatment eciency analysis
Wastewater discharged from the textile industry
shows a large ¯uctuation in terms of quality and quan-
tity, depending on customers' order, season, and fabri-
cated textile materials. Because of the large variabilities
of the compositions and characteristics of textile
wastewaters, a 10-day consecutive sampling was per-
formed in this study. The consecutive sampling was
designed to evaluate the variations in COD and color
values of the in¯uent/e‚uent of those 10 selected
wastewater treatment facilities (designated from Facili-
ties 1±10). All 10 selected facilities use biological treat-
ment followed by chemical coagulation/precipitation
treatment processes.
To meet the new color discharge limit (400 ADMI
units measured by 3 WL method), seven of the ten fa-
cilities added sodium hypochlorite (NaOCl) into the
treatment stream. The addition of NaOCl, a decoloring
agent, was accomplished in three di€erent locations:
¯occulation basin (four facilities), coagulation basin
(two facilities), and discharging point (one facility). To
1058 C.M. Kao et al. / Chemosphere 44 (2001) 1055±1063

increase the e€ects on color removal, NaOCl is over-
dosed by all seven facilities. However, the residual
chlorine concentration is not monitored by those facili-
ties. Thus, residual chlorine might cause some adverse
e€ects on the discharged watercourses.
Table 1 presents the wastewater ¯ow capacity,
NaOCl addition location, and operational cost (US $/
ton of wastewater) for the 10 treatment facilities. The
operational costs for those facilities varied from US $0.4
to $1.1 per ton of wastewater including the addition of
NaOCl for color removal. Due to the usage of NaOCl
(seven facilities), the operational cost (including the
chemical expenses) increased to approximately one half
to two times higher. This indicates that the new color
discharge limit results in a signi®cantly higher opera-
tional cost for most of the textile facilities.
Fig. 1(a) shows the averages of 10-day COD results
(including in¯uent, biotreated (e‚uent from biological
process), and chem-treated (e‚uent from chemical
coagulation/precipitation process)) for 10 treatment
facilities. The standard deviations of the 10-day in¯u-
ent COD values were also presented. Results show that
the colored textile wastewater is frequently accompa-
nied byconcomitant ¯uctuations in COD measure-
ments. A wide range of in¯uent COD values (varied
from 460 to 1240 mg/l) for those 10 facilities was ob-
served. Approximately 50±80% of in¯uent COD were
removed via the biological process, and up to 35% of
the in¯uent COD were removed via the chemical pre-
cipitation process. Results show that most e‚uent
samples (70%) meet the current COD discharge stan-
dard (100 mg/l). Thus, COD in the most textile
wastewater can be controlled by the treatment systems,
and biological process played an important role on
COD removal.
Fig. 1(b) shows the averages of 10-day 3 WL ADMI
color values for 10 treatment facilities. The standard
deviations of the averaged 10-day in¯uent ADMI values
are also presented. Results show that the treatment
system in use by the textile industries fails to e€ectively
remove color from the waste streams. Except for those
three facilities without NaOCl addition (Facilities 1, 4,
and 7), all other seven facilities had diculties on
meeting the 400 ADMI units before the oxidant addi-
tion. Results also show that biological processes had
limited e€ects on color removal. Less than 40% of the
in¯uent color were removed via the biological process.
The dyes are highly structured polymers and are very
dicult to decompose biologically (Tzitzi et al., 1994).
Hence, there is a relatively limited degradation of these
dyestu€ molecules during biological treatment. More-
over, the breakup of some chromophores/auxochromes
or change of the solution pH would also reduce/elimi-
nate color without actually breaking up the whole/sub
dye molecules. Thus, part of the color disappearance
might not be due to the degradation of the dyestu€.
Approximately 40±70% of the in¯uent color was re-
moved via the chemical precipitation process. NaOCl
was applied as the ®nal polishing process to lower the
color value. Results indicate that pre- or post-treatment
options (e.g., advanced oxidation processes) are required
for color removal to comply with the color discharge
limit (Mock, 1999; Vandevivere et al., 1999; Kang et al.,
2000).
3.1.2. Deviation in 3 and 31 ADMI values
For the ADMI value analysis, three controlling
factors (pH, instrument, and 3/31 WL) were applied to
evaluate those phenomena on color measurement. Fig.
2(a) and (b) present the variations in 3 and 31 WL
ADMI values versus two instruments used in this study.
A total of 300 textile wastewater samples (including raw
wastewater, e‚uent from biological process, e‚uent
from chemical precipitation process or before dis-
charge) were used for each comparison. Results show
that 3 and 31 WL ADMI values measured by both
instruments were in close agreements (R2 > 0:995). This
indicates that Hach DR4000 can provide a very precise
ADMI color measurement using the installed standard
curves.
The measured ADMI values obtained at the origi-
nal pH and pH 7.6 are presented by ADMIoriginal and

Table 1
Characteristics of 10 selected facilities
Facility Capacity (m3 /d)
1 3500
2 4500
3 1200
4 415
5 1200
6 11,000
7 3500
8 8000
9 3300
10 1350
a
±: no addition; F: ¯occulation basin; D: discharging location; C: coagulation basin.
Location of NaOCl addition Operational cost (US $/m3 wastewater)
±a 0.4
F 0.6
F 0.5
± 0.3
F 0.4
F 0.7
± 0.5
D 0.9
C 0.8
C 1.1
 C.M. Kao et al. / Chemosphere 44 (2001) 1055±1063 1059

Fig. 1. (a) Averages of 10-day COD results (including in¯uent, biotreated, and chem-treated) for 10 treatment facilities; (b) averages of
10-day 3 WL ADMI color values (including in¯uent, biotreated, chem-treated, and e‚uent after NaOCl addition) for 10 treatment
facilities.

ADMI7:6 , respectively. Fig. 3 presents the correlation
between the ADMIoriginal and ADMI7:6 for 300 textile
wastewater samples (R2 ˆ 0:98). These results indicate
that pH adjustment does not have a signi®cant e€ect
on ADMI value measurement. It has been reported
that the pH causes the variation in color value of
natural water because the solubility of natural color-
causing substances (e.g., humic acid, fulvic acid) varies
with pH values (Bennett and Drikas, 1993; Kang and
Kuo, 1999). However, for the textile wastewaters,
natural color-causing substances have less e€ect on the
color value measurement compared to SS and dyeing
materials in the wastewater. Thus, for textile waste-
water samples, pH control might not be necessary for
ADMI value measurement after the ®ltration proce-
dure.
The relationship between 3 and 31 WL ADMI values
for di€erent textile wastewater samples without pH ad-
justment Fig. 4 is
31WL ADMI ˆ 0:64…3WL ADMI† ‡ 36
R2 ˆ 0:92: …1†
Although 3 and 31 WL methods are both approved by
ADMI and are listed in the Standard Methods (APHA,
1995), analytical results show that signi®cant variations
of the numerical values exist between 3/31 ADMI val-
ues. For example, if the 3 WL ADMI value for a textile
wastewater sample is 400 ADMI units, its corresponding
31 WL ADMI value is only 292. Wastewater samples in
the present study exhibited a wide range of color values
(from 0 to 6000 ADMI units by 3 WL method). Re-
gression Eq. (2) was then calculated using 100 e‚uent
samples with lower ADMI values (lower than 1350
ADMI units by 3 WL method).
31WL ADMI ˆ 0:59…3WL ADMI† ‡ 24
R2 ˆ 0:96: …2†
1060 C.M. Kao et al. / Chemosphere 44 (2001) 1055±1063

Fig. 2. (a) Variation in 3 WL ADMI values analyzed by two instruments (Shimadzu and Hach); (b) variation in 31 WL ADMI values
analyzed by two instruments (Shimadzu and Hach).

Again, signi®cant variations between 3/31 ADMI values
exist. The 400 ADMI color discharge limit by 3 WL
method only corresponds to 260 ADMI units by
31 WL method using Eq. (2). Results indicate that 3 and
31 WL ADMI methods cannot be substituted for each
other, and the discharge standard should be determined
based on the selected ADMI method.
3.1.3. Comparison of COD and ADMI values
There is no relationship between COD and 3 WL
ADMI values of textile wastewater samples as shown in
Fig. 5. Normal textile dyeing and ®nishing operations are
such that the dyestu€s used in a mill can vary from day
to day and sometimes even several times a day mainly
because of the batched nature of the dyeing process.

Fig. 3. Correlation between the ADMIoriginal and ADMI7:6 of 3 WL method for textile wastewater samples.
 C.M. Kao et al. / Chemosphere 44 (2001) 1055±1063 1061

Fig. 4. Relationship between 3 and 31 WL ADMI values for 300 textile wastewater samples without pH adjustment.

Frequent changes of dyes and chemicals in the dyeing
process cause considerable variation in the wastewater
characteristics (e.g., ADMI value and COD concentra-
tions). Some strong color dyes have low COD, but some
low color value substances used in the pretreatment,
printing, and ®nishing procedures have high COD. Thus,
the ``colored''compounds account only for a small frac-
tion of the wastewater COD (Tzitzi et al., 1994). The
color intensity varies with pH, and that COD may not
detect some dye compounds. Therefore, COD and color
values in the textile wastewater show little correlation.
3.2. E‚uents from textile wastewater treatment plant
A total of 186 e‚uent samples were collected from 75
textile wastewater treatment plants to evaluate the fea-
sibility of regulating the colored textile e‚uents by the
new discharge standard (3 WL) as well as 31 WL ADMI
method. All samples were adjusted to pH 7.6, ®ltered
through 0.45 m membrane ®lters, then analyzed by both
3 and 31 WL ADMI methods. Fig. 6 presents the dis-
tribution of the 3 and 31 WL ADMI values versus the
percentage of total samples. Results show that only 65%
of the 186 e‚uent samples can meet the color discharge
limit (400 ADMI units) using the 3 WL ADMI method.
However, if the 31 WL ADMI method were adopted for
the standard analytical procedure, up to 85% of the ef-
¯uent samples could meet the limit. Moreover, among
the 186 samples, approximately 30% of the samples were
treated with overdosed NaOCl before discharge to as-
sure the e€ects of oxidation. The overdosed NaOCl
would cause the increase in residual chlorine concen-
trations in the discharged watercourses. This may pro-
duce unknown adverse e€ects on aquatic life and
environments. Results indicate that only 35% of the
total samples could meet the discharge standard without

Fig. 5. Relationship between COD and 3 WL ADMI values of 300 textile wastewater samples.
1062 C.M. Kao et al. / Chemosphere 44 (2001) 1055±1063
Fig. 6. Distribution of 3 and 31 WL ADMI values versus the percentage of total samples.
NaOCl addition. This implies that the wastewater
treatment technology currently in use fails to e€ectively
treat the colored textile wastewater. Almost two thirds
of the textile mills need to install a ®nal polishing pro-
cess (e.g., advanced oxidation process) for color re-
moval.
4. Conclusions and suggestions
In this study, 10 representative textile wastewater
treatment facilities were selected for a 10-day consecu-
tive sampling and analysis. Additionally, 186 textile ef-
¯uents from 75 textile mills were collected and analyzed
to evaluate the feasibility of regulating the colored tex-
tile e‚uents by the new discharge standard (400 ADMI
units). Conclusions from this study include the follow-
ing:
1. A signi®cant di€erence between 3 and 31 ADMI
methods was observed for 300 textile wastewater
samples. The discharge standard should be deter-
mined based on the selected testing method, and
these two methods cannot be substituted for each
other.
2. Results show that pH adjustment has no signi®cant
in¯uence on ADMI value measurement for the textile
wastewater samples. Thus, pH control might not be
necessary for ADMI value measurement after the ®l-
tration procedure.
3. The wastewater treatment scheme used can lower
most of the wastewater COD (up to 70%) to below
the current discharge standard (100 mg/l). However,
these systems can not e€ectively remove color from
the textile wastewater to meet the discharge limit
(400 ADMI units by 3 WL method).
4. Results from the second part of this study show that
only 35% of the 186 e‚uent samples could meet the
400 ADMI discharge limit without the addition of
NaOCl. Thus, the new ``E‚uent True Color Stan-
dard'' in Taiwan is somewhat unachievable by the
currently used treatment technology, and the true col-
or discharge limit might not be applicable at this
stage.
5. Most of the treatment facilities apply NaOCl as the
®nal polishing step. However, to increase the e€ects
and reliability of NaOCl application on color re-
moval, NaOCl is overdosed by all facilities. Unless
government of Taiwan enforces discharge limit for re-
sidual chlorine, the watercourses would be severely
damaged by the usage of NaOCl for color removal.
6. This study can help us improve our understanding of
the e€ectiveness and feasibility of the current textile
wastewater treatment processes on the regulatory
compliance. This improved understanding can be
used to modify existing treatment technologies and
select necessary post-treatment technique to enhance
the system reliability.
Acknowledgements
This project was funded by Environmental Protec-
tion Administration (EPA), ROC. The writers appreci-
ate Ms. L.H. Chang (Director) and Ms. M.J. Liao
(Senior Specialist) of Water Quality Protection, EPA for
their support and assistance throughout this project.
Grateful acknowledgement is extended to Ms. Y.H.
Chien of Taiwan Textile Printing Dyeing and Finishing
Industrial Association for her assistance on the factory
investigation and sample collection tasks.
 C.M. Kao et al. / Chemosphere 44 (2001) 1055±1063
References
Allen, W., Prescott, W.B., Derby, R.E., Garland, C.E., Peret,
J.M., Saltzman, M., 1973. Determination of color of water
and wastewater by means of ADMI color values. In:
Proceedings of the 28th Purdue Industrial Waste Confer-
ence. Purdue University, West Lafayette, Indiana, pp. 661±
675.
APHA, American Public Health Association, 1995. Standard
Methods for the Examination of Water and Waste Water,
19th ed., APHA-AWWA-WEF, Washington, DC.
Bennett, L.E., Drikas, M., 1993. The evaluation of color in
natural waters. Wat. Res. 27, 1209±1218.
Cooper, P., 1995. Colour in Dyehouse E‚uent. Society of
Dyers and Colourist, New York.
EPA, US Environmental Protection Agency, 1986. Quality
Criteria for Water. US EPA, Washington, DC.
EPA, US Environmental Protection Agency, 1978. Method
110.1, Color (Colorimetric, ADMI). US EPA, Washington,
DC.
Ingamells, W., 1993. Colour for Textiles a User's Guide.
Society of Dyers and Colourists, New York.
Kang, S.F., Kuo, S.P., 1999. Correlation among indicators in
regulating colored industrial wastewaters. Chemosphere 39
(12), 1983±1996.
Kang, S.F., Liao, C.H., Po, S.T., 2000. Decolorization of textile
wastewater by photo-fenton oxidation technology. Chemo-
sphere 41 (8), 1287±1294.
1063
Kennedy, K.T., Morgan, J.M., Bene®eld, L.K., McFadden,
A.F., 1992. Color removal from textile dye wastewater: a
case study. In: Proceedings of the 47th Purdue Industrial
Waste Conference. Purdue University, West Lafayette,
Indiana, pp. 727±741.
Lemordant, D., Letellier, P., Rumeau, M., Soma, C., 1989.
Colour removal from textile e‚uents. Textile Month 9, 41±
44.
Mock, B., 1999. Ozone application to color destruction of
industrial wastewater ± part 1: Experimental. American
Dyestu€ Reporter 87, 18±22.
NCDEHNR, North Carolina Department of Environment,
Health, and Natural Resources, 1999. Title 15A, Chapter 2,
Subchapter 2B, Section. 0400 ± E‚uent Limitations.
Raleigh, NC.
Pierce, J., 1994. Colour in textile e‚uent the origins of problem.
J. Soc. Dyers Color 110, 131±133.
Tzitzi, M., Vayenas, D.V., Lyheratos, G., 1994. Pretreatment of
textile industry wastewaters with ozone. Wat. Sci. Tech. 29
(9), 151±160.
Vandevivere, P.C., Bianchi, R., Verstraete, W., 1999. Treatment
and reuse of wastewater from the textile wet-processing
industry: review of emerging technologies. J. Chem. Tech.
Biotechnol. 72, 289±302.
Wisconsin State Government, Department of Natural Re-
sources, 1997. Categories and Classes of Point Sources and
E‚uent Limitations, Chapter NR 220, Wisconsin State
Government, Wisconsin.