DSpace at VNU: Trihalomethane formation by chlorination of ammonium- and bromide-containing groundwater in water supplies of Hanoi, Vietnam

Water Research 37 (2003) 3242–3252

Trihalomethane formation by chlorination
of ammonium- and bromide-containing groundwater in water
supplies of Hanoi, Vietnam
Hong Anh Duonga, Michael Bergb,*, Minh Hang Hoanga, Hung Viet Phama,
Herve! Gallardb, Walter Gigerb, Urs von Guntenb
a

Centre for Environmental Technology and Sustainable Development (CETASD), Hanoi University of Science, Nguyen Trai Street 334,
Hanoi, Viet Nam
b
EAWAG, Swiss Federal Institute for Environmental Science and Technology, Ueberlandstrasse 133, CH-8600 Dubendorf,
Switzerland
.
Received 16 October 2002; received in revised form 10 February 2003; accepted 12 February 2003

Abstract
The occurrence and the fate of trihalomethanes (THMs) in the water supply system of Hanoi City, Vietnam was
investigated from 1998 to 2001. The chlorination efficiency, THM speciation, and, THM formation potential (THMFP)
was determined in the water works and in tap water. With regard to THM formation, three types of groundwater
resources were identified: (I) high bromide, (II) low bromide, and (III) high bromide combined with high ammonia and
high dissolved organic carbon (DOC) concentrations. Under typical treatment conditions (total chlorine residual 0.5–
0.8 mg/L), the total THM formation was always below WHO, EU, and USEPA drinking water standards and
decreased in the order type I>type II>type III, although the THMFP was >400 mg/L for type III water. The
speciation showed >80% of bromo-THMs in type I water due to the noticeable high bromide level (p140 mg/L). In
type II water, the bromo-THMs still accounted for some 40% although the bromide concentration is significantly lower
(p30 mg/L). In contrast, only traces of bromo-THMs were formed (B5%) in type III water, despite bromide levels were
high (p240 mg/L). This observation could be explained by competition kinetics of chlorine reacting with ammonia and
bromide. Based on chlorine exposure (CT) estimations, it was concluded that the current chlorination practice for type I
and II waters is sufficient for X2-log inactivation of Giardia lamblia cysts. However, in type III water the applied

chlorine is masked as chloramine with a much lower disinfection efficiency. In addition to high levels of ammonia, type
III groundwater is also contaminated by arsenic that is not satisfactory removed during treatment. Nnitrosodimethylamine, a potential carcinogen suspected to be formed during chloramination processes, was below
the detection limit of 0.02 mg/L in type III water.
r 2003 Elsevier Science Ltd. All rights reserved.
Keywords: Disinfection by-products; Water distribution system; Trihalomethane formation potential; Competition kinetics; Chlorine
exposure; N-nitrosodimethylamine (NDMA); Hanoi; Vietnam; Bromide; Ammonium

1. Introduction
Hanoi has a strongly increasing water demand due to
the rapid growth of the urban population (3.5 Mio
*Corresponding author. Tel.: +41-1-823-5078; fax: +41-1823-5058.
E-mail address: (M. Berg).

inhabitants in 2001, urban area 84 km2). The main water
resources for Hanoi are reduced groundwaters containing variable levels of dissolved iron(II) and manganese(II) ranging from 1–25 mg/L and 0.2–3 mg/L,
respectively. Moreover, Hanoi’s groundwater contains
excessive concentrations of arsenic that are partly
removed in the WTPs to lower but not fully acceptable
levels (25–91 mg/L) [1]. The groundwaters pumped from

0043-1354/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0043-1354(03)00138-6


H.A. Duong et al. / Water Research 37 (2003) 3242–3252

several locations have significantly differing chemical
compositions that are influenced by varying proportions
of bank infiltration from the Red River [1].
Since the beginning of the 20th century chlorination

has been a key-treatment for improving the microbiological safety in drinking waters. However, an undesired
formation of disinfection by-products (DBPs) results
from the reaction of chlorine with natural organic
matter (NOM) and includes products such as trihalomethanes (CHCl3, CHCl2Br, CHClBr2, CHBr3) [2]
which might have adverse health effects [3–5].
The total concentration of trihalomethanes (THMs)
and the formation of individual THM species in
chlorinated water strongly depend on the composition
of the raw water, on operational parameters during
water treatment and on the residual chlorine in the
distribution systems [6]. Aiming at minimizing the
cancer risk, the United States Environmental Protection
Agency (USEPA), the World Health Organization
(WHO), and the European Union (EU) introduced
regulations for THMs in drinking water. Whereas the
USEPA and the EU regulate total THM concentrations
as 80 or 100 mg/L, respectively, the WHO provides
guidelines for individual THM compounds [4].
In Vietnam, chlorination is generally applied for
drinking water disinfection. The requirement for the
minimum residual chlorine concentration in water at the
outlet of water treatment plants (WTPs) and in the
distribution system is 0.5 and 0.3 mg/L, respectively [7].

3243

So far, no information on THM concentrations in
drinking water was available. To determine THM levels
in the urban area of Hanoi, the eight major WTPs of this
city were investigated (see Fig. 1).

Based on the differences in dissolved organic carbon
(DOC), UV254 absorption, bromide concentration and
chlorine demand (mainly ammonium and ferrous iron
content), it was inferred that substantial variations in
THM formation can be expected in the eight Hanoi
WTPs. A direct comparison of the eight WTPs is favored
by the fact that very similar treatment trains are applied.
In addition to full-scale data, chlorination experiments with natural groundwater were carried out in the
laboratory to evaluate the maximum THM formation
potential (THMFP) and THM yields for various
chlorine doses. Based on these data, a relation between
chlorine exposure and THM formation was established.
These results may be used as a tool to adjust the chlorine
doses, which is needed to achieve efficient disinfection
and acceptable levels of THMs in the drinking waters of
Hanoi.

2. Materials and methods
2.1. Reagents
If not specified, chemicals were reagent grade from
Fluka (Buchs, Switzerland) or Merck (Darmstadt,

Fig. 1. Map illustrating northern Vietnam and the locations of the eight studied Hanoi WTPs. The areas of differing groundwater
quality (types I, II and III) are indicated. The numbers 1–8 refer to the following WTPs: 1, Mai Dich; 2, Ngoc Ha; 3, Ngo Si Lien; 4,
Yen Phu; 5, Luong Yen; 6, Ha Dinh; 7, Tuong Mai; and 8, Phap Van.


H.A. Duong et al. / Water Research 37 (2003) 3242–3252

3244

2.3.2. Distribution system
Two sampling campaigns (May and September 2001)
were performed in three water distribution systems to
investigate THMs in household tap waters. These water
distribution systems primarily provide drinking water
from WTPs 3, 4 and 6 representing the three different
groundwater types of Hanoi. Finished water of the
WTPs and tap water samples of the corresponding
distribution system were collected at the same day.

Germany). A standard stock solution containing the
four THMs (1 mg/mL of each in methanol) and the
internal standard p-bromofluorobenzene (1 mg/mL in
methanol) was supplied by Tokyo Kasei Kogyo (ChuoKu, Tokyo, Japan). Working standard solutions of
THMs (1, 10, and 100 mg/L) were prepared from the
stock solution. Calibrations were conducted with five
concentration levels in de-ionized and boiled water.
Chlorine solutions of 0.4% and 0.04% were prepared
from NaOCl 4% (Aldrich, Steinheim, Germany) by
dilution with water.

2.4. Analytical methods
Water temperature, pH, dissolved oxygen (detection
limit 0.1 mg/L) (Paqualab instrument ELLE, Leighton
Buzzard, UK) and residual chlorine (detection limit
0.01 mg/L) were measured at the sample collection sites.
Samples for DOC, cations, anions and coliforms were
collected according to Standard Methods [8]. The DOC
concentrations were measured by a Shimadzu TOC–

5000A analyzer. Ammonium and bromide were determined by ion chromatography with conductivity detection (HPLC 10A, Shimadzu, Japan). UV absorbance
was measured with a UV/Vis spectrophotometer (Shimadzu 1201, Japan) at 254 nm in 10 mm quartz cells.
Total coliforms in water were determined by the
membrane filtration method [8]. The chlorine and
chloramine concentrations were determined by the
ABTS method [9].

2.2. Water treatment plants in Hanoi
The eight major WTPs (see Fig. 1 and Table 1)
produce roughly 450,000 m3 drinking water per day.
Groundwater from 30 to 70 m depth is used as raw
water. All WTPs operate with similar treatment trains
including aeration, settling, sand filtration, chlorine
disinfection, and storage in a reservoir. Chlorine is
applied as gas and the applied disinfectant dose is
adjusted manually to maintain a total chlorine residual
of 0.5–0.8 mg/L after the reservoir.
2.3. Sampling campaigns
2.3.1. Water treatment plants
Water samples of raw water from the aeration jet,
treated water before disinfection (after Fe removal), and
finished water at the outlet of the reservoir were
collected. Nine sampling campaigns were conducted
between April 1998 and May 2001.

2.5. Analysis of THMs
Duplicate samples for THM analysis were collected in
44-mL glass vials and were capped with PTFE-faced

Table 1

Individual THM concentrations and chemical parameters in treated and finished waters of Hanoi water treatment plants
Groundwatera

Type I
b

Water treatment plants

1, 2 and 3

Type II

Type III

4 and 5

6, 7 and 8

c

DOC
UV254
NH+
4
BrÀ

CHCl3
CHCl2Br
CHClBr2
CHBr3

a

mg/L
mÀ1
mg N/L
mg/L

treated water , concentration range (classification)
0.6–1.3 (low)
0.9–1.1 (low)
0.1–1.4 (low)
0.7–1.2 (low)
o2 (low)
o2 (low)
50–140 (high)
o20–30 (low)

2.0–6.4
4.0–16.0
5–25
70–240

mg/L
mg/L
mg/L
mg/L

finished waterd, concentration range (average)e
o0.3–11.1 (2.2)
0.9–7.7 (4.3)

0.5–7.3 (2.4)
o0.2–5.6 (2.4)
0.3–22.3 (6.3)
o0.2–3.8 (1.8)
1.2–18.5 (6.6)
o0.2–3.7 (0.5)

0.3–21.5 (4.2)
o0.2–3.6 (0.3)
o0.2–3.7 (0.3)
o0.2

(high)
(high)
(high)
(high)

The areas of differing groundwater quality are indicated in Figure 1.
The numbers 1–8 of the water treatment plants refer to Figure 1.
c
Treated waters were collected before chlorination. Number of samples: 9–15.
d
Finished waters collected at the outlets of the WTPs. Number of samples: 10–17, chlorine dose o1.5 mg/L, contact time 30 min.
e
Average concentration from finished water samples collected between April 1998 and May 2001.
b


H.A. Duong et al. / Water Research 37 (2003) 3242–3252

silica septa. Household taps were flushed for 15 min
prior to sampling. The 44-mL sample vials contained
0.15 mg sodium sulfite (50 mL of a 3 g/L Na2SO3
solution) to quench residual chlorine. The samples were
stored at 4 C.
Head-space gas chromatography with mass spectrometry detection was used to analyze THMs in water.
Sample volumes of 10 mL were filled in 20 mL glass vials
and spiked with 50 mL of the internal standard (10 mg/L
p-bromofluorobenzene). The vials were immediately
sealed with Teflon coated septa and aluminum crimpcaps. After equilibration (30 min at 30 C) a volume of
1 mL of the head-space was injected splitless into a gas
chromatograph (DB 624 column, 60 m  0.32 mm,
0.32 mm film, J&W Scientific, CA) coupled to a mass
spectrometer (GC/MS QP5000 Shimadzu, Japan). The
temperature program was: 50 C (1 min), 7 C/min to
120 C (10 min), 12 C/min to 200 C (5 min). The eluting
analytes were recorded in the selective ion monitoring
(SIM) mode and quantified by internal standard
calibration. Recoveries of the four THMs, determined
in spiked samples at levels of 1, 10 and 20 mg/L, were in
the range of 94–114% (n ¼ 7). The method detection
limits (MDLs) were determined from the standard
deviation of 1-mg/L spiked samples (n ¼ 7; 3 sigma)
following the procedure described by [10]. The corresponding MDLs for CHCl3, CHCl2Br, CHClBr2 and
CHBr3 were 0.3, 0.2, 0.2 and 0.2 mg/L, respectively. Nnitrosodimethylamine (NDMA) was analyzed with gas
chromatography and thermal energy analysis detection
(GC-TEA, detection limit 0.02 mg/L), following the
official method 982.12 of the Association of Analytical
Communities (AOAC, Gaithersburg, MD).
2.6. Laboratory chlorination experiments

Laboratory chlorination experiments were carried out
with treated waters collected before disinfection from
the WTPs 1–4 and 6–8. Samples were collected in 5-L
glass bottles and stored at 4 C until the chlorination
experiments. The water samples were buffered with
1 mM phosphate and adjusted to the desired pH with
NaOH. For chlorination experiments, the required
chlorine dose was added to a laboratory batch system
and the solution was well mixed for 30 s. Then the water
was immediately portioned into 44-mL glass vials that
were sealed with Teflon-lined screw caps. After filling
and sealing the head-space-free samples, the vials were
kept in a thermostated water bath at 25 C. For each
desired reaction time, a sample was removed, and,
residual total chlorine and chloramine were determined
[11]. THM concentrations were quantified as described
above.
The THMFP of the treated waters was determined as
total THM formed during a reaction time of 7 days at
25 C (pH 8.0 or 7.0) and by maintaining a free chlorine

3245

residual throughout the experiment. A chlorine dose of
9 mg/L was used for treated water from WTPs 2 and 4,
and, of 160 mg/L for treated water from WTP 8,
respectively.

3. Results and discussions
3.1. Water quality parameters and THM formation

(water types I, II, and III)
Table 1 summarizes the water quality parameters of
the three water types I, II and III (see Fig. 1) that affect
the THM formation during chlorination. No THMs
were detectable before chlorination.
3.1.1. DOC and UV254
Water of types I and II had low levels of DOC
(B1 mg/L) and low UV absorbances (o1.5 mÀ1) indicating a low content of NOM and aromaticity (Table
1). Therefore, the concentration of THM precursors in
these waters can be expected to be low. In water type III,
DOC and UV254 were relatively high. The highest
values were measured in treated water of WTP 8 (DOC
6.4 mg/L, UV254 16 mÀ1) reflecting a very high content
of NOM and a high potential for THM formation.
3.1.2. Bromide
Relatively high levels of bromide in the range of 50–
140 and 70–240 mg/L were observed in water types I and
III, respectively (Table 1). In water type II, bromide was
near the detection limit of 20 mg/L. Bromide is important
for THM formation because it is oxidized by chlorine to
hypobromous acid, which contributes to the formation
of bromo-THMs [12].
3.1.3. Ammonium
The fast reaction of chlorine with ammonia leads to
masking of chlorine in excess of ammonia by formation
of chloramine. If chloramine is the dominant species, the
THM yields in chlorinated water is reduced because of
the lower reactivity of chloramine with NOM [13]. Such
a situation was observed for water type III (WTPs 6, 7
and 8) where extremely high ammonium concentrations

of 5–25 mg/L are present. Ammonium is believed to
originate from mineralization of peat which is abundant
in the subsurface of the type III area [1].
3.2. Occurrence of THM in finished waters at the WTP
outlets
Fig. 2 shows the concentrations of THMs in finished
waters at the outlet (reservoir, 1 h contact time) of WTPs
using groundwater types I, II and III. They are in the
range of 5–56, 2–18 and 0.3–22 mg/L, respectively. The
high variability of the total THM concentrations can be


H.A. Duong et al. / Water Research 37 (2003) 3242–3252

3246

100

50

average
minimum

40

(n = 4 to 8)

30
20
10

0
WTP 1

CHBr3

maximum

2

I

3

4

5

II
groundwater type

6

7

8

III

molar distribution (%)

total THM conc. (µg/L)

60

CHClBr2

80

CHCl2Br
60

CHCl3

40
20
0
I

II
III
groundwater type

Fig. 2. Average concentration and range of total THM
concentrations measured in finished waters at the outlets of
the Hanoi WTPs.

Fig. 3. Molar distribution of THM species in finished waters of
the Hanoi water types I, II and III.

attributed to unsteady chlorine doses, variations in

ammonium, DOC and Fe(II) concentrations. The
groundwater composition is not constant because the
10–20 wells belonging to each Hanoi WTP are operated
intermittently. The THM formation in finished water
was generally in the following order of water sources:
type I>type II>type III. Due to the relatively low DOC
levels in water types I and II, the absolute levels of
THMs are quite low. Higher levels of THMs would be
expected in water type III (high DOC). However,
ammonium is dominating this system (see below).
The distribution of the four THMs in finished waters
is illustrated in Fig. 3. In the WTPs 1, 2 and 3 (water
type I), brominated THMs (>85%) were the dominant
and most abundant species. In finished water from WTP
4 and 5 (water type II), chloroform was formed in
similar concentrations as the brominated THMs. In
comparison to water type II, the higher ratio of
brominated THMs in water type I can be explained by
the 3–5 times higher bromide levels (see Table 1). The
highest bromide levels were present in groundwater type
III, yet only small amounts of brominated THMs were
detected in the treated waters (chloroform 90% of the
total THM). This is due to masking of chlorine by
ammonia which prevents bromide oxidation (see below).

Table 2
Trihalomethane formation potential (THMFP) for treated
Hanoi groundwaters

3.3. Influence of ammonium on THM formation and

speciation
The THMFP determined for the three waters type,
namely WTP 2 (type I), WTP 4 (type II) and WTP 8
(type III), were 103, 59 and 406 mg/L, respectively (see
Table 2). The high THMFP found for water type III
corresponds to its high content of NOM represented by
the DOC values (see Tables 1 and 2). However, as shown
in Fig. 2, the total THM concentrations in the finished

WTP 2 WTP 4 WTP 8
Type Ia Type IIa Type IIIa
Treated waterb
DOC
mg/L
0.9
UV254
mÀ1
0.5
BrÀ
mg/L
86
NH+
mg N/L
0.2
4
Laboratory conditions
Chlorine dose mg/L
9
pH buffered to
8

THMFP
mg/L as CHCl3 103

0.8
0.9
24
0.13

6.4
15.0
160
18.5

9
8
59

160
7
406

a
The areas of differing groundwater quality are indicated in
Fig. 1.
b
Treated waters were collected before chlorination.

water type III are 20–200 times lower than the THMFP.
This can be explained by the formation of THM along
the breakpoint chlorination curve. Fig. 4 shows that

only traces of THMs are formed for chlorine doses in
the far pre-peakpoint region, whereas a noticeable
increase of THMs is observed near the peakpoint.
Between the peakpoint and the breakpoint, chloroform
formation sharply increases with increasing chlorine
dose.
A comparison of breakpoint chlorination curves is
shown in Fig. 5 for five of the eight WTPs. Peakpoint to
breakpoint chlorination is applied in WTPs of types I
and II (Fig. 5a) whereas far pre-peakpoint chlorination
is obvious for the WTPs type III (Fig. 5b). Fig. 5b shows
that the high ammonium concentration (5–25 mg/L)
present in the investigated waters of WTP 6–8 consume
40–130 mg/L chlorine. For water type III it can therefore


H.A. Duong et al. / Water Research 37 (2003) 3242–3252

15

0.3
0.2

5

CHCl3

0.1

CHBr3

0.0

(a)
0.5
0.4

10

CHCl2Br

1.0
2.0
3.0
chlorine dose (mg(L)

0
4.0

total THM as CHCl3

WTP 4

CHCl3

0.3

25
20
15

0.2

CHCl2Br

0.1

CHClBr2
CHBr3

0.0
0.0

(b)

1.0
2.0
3.0
chlorine dose (mg(L)

10
5
0
4.0

HOCl þ NOM- products-THM
k1 ¼ ?

ð1Þ

HOCl þ NH3 - NH2 Cl þ H2 O
k2 ¼ 4:2 Â 106 MÀ1 sÀ1

ð2Þ

(see Ref. [14]),
HOCl þ NH2 Cl- NHCl2 þ H2 O
k3 ¼ 1:1 Â 103 MÀ1 sÀ1

ð3Þ

(see Ref. [15]),
HOCl þ BrÀ - HOBr þ ClÀ
À1 À1

s

breakpoint
WTP 1

0.6

breakpoint
WTP 4
WTP 4
0.13 mg/L NH4

0.4
0.2
0.5

ð4Þ

1.0

1.5

2.0

chlorine dose (mg/L)
70
applied chlorine
dose ≤1.5 mg/L

60

type III

50
WTP 8
18 mg/L NH4

40
30

WTP 7
7 mg/L NH4

20
10

0

be inferred that the far pre-breakpoint chlorination
leads to the formation of monochloramine, and consequently, low THM concentrations are present in the
finished waters of WTP 6, 7 and 8 (see Table 1 and Fig.
2). In terms of disinfection, Fig. 5 shows that the
chlorine dose (0.8–1.5 mg/L) applied is critical for WTP
4 (only slightly above breakpoint) and insufficient for
water type III.
The THM formation in the investigated waters is
controlled by the following competition kinetics:

k4 ¼ 1:55 Â 10 M

0.8

(a)

Fig. 4. Breakpoint curves and formation of THMs (contact
time 24 h, pH 8.1). Laboratory experiments with treated waters
collected before chlorination in the Hanoi WTPs Mai Dich
(WTP 1, type I) and Yen Phu (WTP 4, type II).

3

WTP 1
0.03 mg/L NH4

1.0

0.0
30

residual
chlorine

range of applied
chlorine dose

type I and II

1.2

0

residual chlorine (mg/L)

CHClBr2

residual chlorine (mg/L)

WTP 1

0.0

residual chlorine (mg/L)

total THM as CHCl3

THM conc. (µg/L)

0.4

1.4

20

residual
chlorine

THM conc. (µg/L)

residual chlorine (mg/L)

0.5

3247

0

WTP 6
6 mg/L NH4

(b)

50

100

150

200

chlorine dose (mg/L)

Fig. 5. Breakpoint curves derived from contact times of 24 h at
pH 8.0 for: (a) WTP 1 and 4 and (b) WTP 6–8. The range of the
chlorine dose applied in the WTPs is indicated.

(see Ref. [16]),
HOBr þ NOM- products-THM
k5 ¼ ?

ð5Þ

The kinetics of reactions (1) and (5) are not known in
absolute terms, however, reaction (5) is much faster than
reaction (1). Therefore, it is important to know to which
extent bromide is oxidized to assess the potential for the
formation of bromo-THMs. Because
ammonia is the
À
main sink for chlorine (HOCl/OCl ) in the pre-breakpoint region, the extent of bromide oxidation can be
estimated from a competition kinetics calculation
involving reactions (2)–(4). The fraction of bromide
being oxidized during the phase where ammonia is in
excess is in the range of a few percent. However, near the
peakpoint or from the peakpoint to the breakpoint, the
fraction of bromide being oxidized becomes larger. This
is due to the slower kinetics of reaction (3) as compared

to reaction (2) (B4 orders of magnitude).
The fraction f(HOCl, BrÀ) of HOCl reacting with BrÀ
can be calculated as follows:
pre-peakpoint:
f ðHOCl; BrÀ Þ ¼

k4 ½BrÀ Š
k2 ½NH3 Š þ k4 ½BrÀ Š

ð6Þ


H.A. Duong et al. / Water Research 37 (2003) 3242–3252

3248

peak- to breakpoint:
À

f ðHOCl; BrÀ Þ ¼

k4 ½Br Š
:
k3 ½NH2 ClŠ þ k4 ½BrÀ Š

ð7Þ

If the residual concentration of HOCl is known
([HOCl]res) the absolute amount of oxidized bromide
can be estimated as

½HOBrŠ ¼ ½HOClŠres f ðHOCl; BrÀ Þ:

ð8Þ

Because reaction (3) is a relatively slow process, HOCl
at the peakpoint can be determined as
½HOClŠres ¼ ½HOClŠo À ½NHþ
4 Šo ;

ð9Þ
[NH+
4 ]o

where [HOCl]o is the applied chlorine dose and
is the initial ammonium concentration.
Based on these considerations, estimates of the
bromide oxidation to HOBr were made for water types
I and III, and summarized in Table 3. The calculations
show that in the case of water type I up to 80% of the
bromide can be oxidized to HOBr which then further
reacts with NOM according to reaction (5). This
explains the high fraction of bromo-THMs found in
this water (Fig. 3). Water type II has a similar water
quality, however, the bromide levels are considerably
lower which explains the lower formation of bromoTHMs (Fig. 3). For the chlorine doses applied in water
type III, bromide oxidation will be very minor (Table 3).
Therefore, low levels of bromo-THMs are expected (Fig.
3), even though the bromide levels in the raw water are
very high (up to 240 mg/L).

3.4. Estimation of the disinfection efficiency via the THM
formation
Fig. 6 shows the relationship between THM formation (expressed as CHCl3) and chlorine exposure for
various chlorine doses derived from laboratory experiments with water types I and II. The chlorine exposure
(CT) is calculated as the integral under a chlorine
concentration time curve [17]. The CT values can be
used to estimate the inactivation efficiency of micro-

organisms. The data plotted in Figs. 6a and b for water
types I and II show that the CT requirements [18,19] for
a 2-log inactivation of E. coli, a 3-log inactivation of
viruses and a 2-log inactivation of Giardia lamblia cysts
can be achieved while the total THM concentration
remains below 10 mg/L which is significantly lower than
the typical drinking water standard. To reach the same
germ inactivation in waters of type III where active
chlorine is mainly present as chloramines, CT values of
360 and 500 mg/L min are required for virus and G.
lamblia cysts, respectively [18]. Yet, for this purpose,
significantly higher chlorine doses of 10–30 mg/L would
be necessary for the WTPs 6, 7 and 8. However, such
high chlorine doses might lead to substantially higher
THM concentrations (see Table 2) and possibly result in
the formation of NDMA (see below).
3.5. Chlorine residual concentrations in tap water samples
of the distribution systems
Results from a sampling campaign in three distribution systems are presented in Table 4. Samples were
taken at increasing distance from the WTP. Using the
distance from the WTP to the sampling point and the
guideline limits for the flow rate in the distribution

system provided in the Water Master Plan of Hanoi [20],
the average residence times in the distribution system
were estimated to be less than 1 h for all tap water
sampling points.
Active residual chlorine concentrations generally
decreased with increasing distance from the WTPs.
The active chlorine residual in the distribution systems
of WTPs 3 and 4 was free chlorine. It was maintained at
concentrations of 0.5–1.4 mg/L. However, in the distribution system of WTP 6 (groundwater type III), the
disinfectant is chloramine and its concentration did not
meet the Vietnamese requirement for residual disinfectant concentrations (X0.3 mg/L). The concentration
of chloramine residual decreased quickly after a distance
of 500 m from WTP 6. The fast consumption of

Table 3
Estimated formation of HOBr during chlorination of pretreated raw water from groundwater resources in Hanoi
Water type I
Parameter
À

Br
NH+
4
NH3
HOCl dose
HOClres
f(HOCl, BrÀ)
HOBr
Conversion of BrÀ

Water type III

Conc.

Peak- to breakpoint

Conc.

Peak- to breakpoint

p100 mg/L
0.2 mg/L

p1.25 Â 10À6 M
1.4 Â 10À5 M
7 Â 10À8 M
1.7 Â 10À5 M
0.3 Â 10À5 M
33%
1 Â 10À6 M
80%

200 mg/L
20 mg/L

2.5 Â 10À6 M
1.4 Â 10À3 M
7 Â 10À6 M
1.4 Â 10À5 M
51.4 Â 10À5 M

1.3 Â 10À4
51 Â 10À9 M
50.04%

1.2 mg/L

1 mg/L
51 mg/L


H.A. Duong et al. / Water Research 37 (2003) 3242–3252

THM concentrations decreased in the distribution
system of WTP 6 after 540 m distance. There are several
possible hypothesis for this observation: (i) leaking of
pipelines and dilution, (ii) volatilization of THMs, and
(iii) reductive degradation of THMs. Hypothesis (i) is
not very likely because the decrease of THMs and
oxygen are different. A massive volatilization of THMs
would probably need a longer residence time and smaller
pipeline diameters. Therefore, it is most likely that a
biotic or abiotic degradation of THMs occurs under low
oxygen conditions.

Virus (CT = 1), E.coli (CT < 0.05)
Giardia lamblia cysts (CT = 25)
total THM as CHCl3 (µg/L)

12
type I

10
8
6
4
2
0
0

50

100

150

200

3.7. N-nitrosodimethylamine formation

200

Samples from WTP 6 and its distribution system were
also checked for a possible formation of NDMA which
is a probable human carcinogen. It has been shown that
NDMA is formed during chloramination processes due
to the reaction of monochloramine with dimethylamine
[21]. Yet, NDMA concentrations were below the
detection limit of 0.02 mg/L in the finished water as well
as in the five tap water samples analyzed. For the
currently applied chlorine dose (p1.5 mg/L) in type III
WTPs, the NDMA formation potential can therefore be

considered negligible. Should chlorine doses be increased, the risk of NDMA formation has to be
considered again.

chlorine exposure (CT), (mg/L.min)

(a)

total THM as CHCl3 (µg/L)

30
type II
25
20
15
10
5
0
0
(b)

25

50

100

150

3249

chlorine exposure (CT), (mg/L.min)

Fig. 6. THM formation during chlorination of treated waters
at 25 C and pH 7.5 for various chlorine doses. (a) Water type I,
chlorine dose 1.1 mg/L (diamonds), 1.9 mg/L (squares); (b)
water type II, chlorine dose 0.5 mg/L (diamonds), 1.0 mg/L
(squares), 1.9 mg/L (triangles). Required CT values for E. coli,
virus, and Giardia lamlia cysts are indicated.

chloramine near WTP 6 may be due to the oxidation of
remaining Fe(II) and Mn(II) or a biological reduction of
NH2Cl.
3.6. Trihalomethane occurrence in tap water
As shown in Table 4 the total THM concentration
analyzed in the tap water samples were below the
drinking water standards of the EU and the USEPA.
The speciation of THMs in the finished waters and in the
tap waters was the same. In the distribution systems of
WTPs 3 and 4 having relatively high free chlorine
residuals and dissolved oxygen of more than 2 mg/L, the
total THM concentrations increased with increasing
distance from the plant. No coliforms were present in
the tap waters investigated. Based on the estimates on
chlorine CT (Fig. 6), the measured THM concentrations
indicate CT values which possibly guarantee a good
inactivation, even for G. lamblia cysts. Interestingly, the

4. Conclusions and recommendations
This study shows that the THM concentrations in all
Hanoi WTPs were below WHO, EU and USEPA

drinking water standards. For water type III with high
DOC and bromide levels and a THMFP of >400 mg/L,
higher total THM concentrations as well as a higher
proportion of bromo-THMs were expected. The low
THM formation could be explained by competition
kinetics of chlorine for the oxidation of ammonia and
bromide, as well as by the short residence times (o1 h)
in the Hanoi distribution system. The following conclusions and recommendations can be drawn from the
findings obtained through this study.
1. Based on estimations of chlorine exposure (CT), it
can be hypothesized that for the inactivation of
coliforms, virus, and Giardia lamblia cysts, the
chlorine dose of 0.8–1.5 mg/L applied for disinfection
by the Hanoi water works is inefficient for water type
III, and critical for water types I and II if ammonium
concentrations exceed 0.1 mg/L.
2. With regard to the residual chlorine of 0.3 mg/L
required in the distribution system and tap water,
it is recommended to maintain a steady chlorine
dose of 1.5 mg/L for water types I and II if
ammonium concentrations are below 0.15 mg/L.


a

Tap water from
distribution
system

Treated water

Finished water

Tap water from
distribution
system

Treated water
Finished water

6

11
18
28
35
40

540
900
1400
1750
2000

5
14
16

270
720
800

360

2

2
4
6
15
32
50

Estimated
residence
time (min)

90

90
180
360
750
1600
2500

Distance
from
WTP (m)

Below detection limit (see Materials and Methods).

WTP 6 (type III)

WTP 4 (type II)

Treated water
Finished water

WTP 3 (type I)

Tap water from
distribution system

Sample

Water
treatment
plant

27.7
27.4
27.6
27.4
27.4

6.8
6.8
6.9
6.7
6.8

6.8

6.8

30.2
27.7

6.8

7.3
7.2
7.3

28.6

27.8
27.5
29.5

7.2

7.3

27.0
27.6

7.2

27.0

7.0
7.0
7.0
7.0
7.0
7.0

7.0

27.1
27.3
27.0
27.8
28.1
27.3
29.1

6.7

pH

27.1

T
( C)

1.5
0.9
0.9

0.7
0.4

1.8

2.3

3.9
4.3
3.3

4.0

4.3

2.1
2.7
2.3
2.0
2.0
2.1

2.5

DO (mg
O2/L)

0.11
0.06
0.04

0.03
0.03

0.12

0.15

0.75
0.47
0.45

0.86

0.87

1.39
1.34
1.33
1.25
0.98
0.90

1.44

Total
residual
chlorine
(mg/L)

2.9

2.1
1.8
0.7
0.5

1.6

0.8

19.2
23.4
25.4

18.3

17.1

4.9
6.3
5.4
6.4
6.0
6.4

3.0

CHCl3
(mg/L)

a

a

a
a
a
a
a

a

a
a
a
a
a

6.2
8.0
7.9

6.1

5.2

18.4
20.7
22.7
26.6
30.0

25.5

16.6

CHClBr2
(mg/L)

a

17.1
22.2
22.7

16.7

15.1

13.2
14.4
15.7
18.0
20.0
17.8

9.6

CHCl2Br
(mg/L)

Table 4

Individual THM concentrations and water quality parameters in three Hanoi water distribution systems (September 2001)

a

a

a

a

a

a

a

3.2
2.6
3.7

0.0

1.6

20.5
16.6
20.9
26.5
30.4
25.9

18.8

CHBr3
(mg/L)

2.9
2.1
1.8
0.7
0.5

1.6

0.8

36.7
45.4
48.2

34.0

31.8

34.6
36.5
39.7
47.3
52.1
45.2

28.4

Total
THM as
CHCl3
(mg/L)

0
0
0
0
0

0

0

0

0
0
0

0

0

0

0
0
0
0
0
0

0

1

Coliform
(MPN/
100 mL)

3250
H.A. Duong et al. / Water Research 37 (2003) 3242–3252


H.A. Duong et al. / Water Research 37 (2003) 3242–3252

Lower chlorine doses and/or ammonium concentration >0.15 mg/L result in residual chlorine concentrations o0.3 mg/L.
3. Due to the high ammonium levels of X15 mg/L in
water type III, chlorine is mainly transformed to
chloramine which is a considerably less efficient
disinfectant than chlorine (B100 times). The chlorine
doses of 10–30 mg/L necessary to reach a 2-log
inactivation of Giardia lamblia cysts might cause
substantially higher THM concentrations and possibly form NDMA. A satisfactory quality of drinking
water derived from water type III therefore requires

multistage treatment for removal of Fe, Mn, NH4,
and DOC. In addition, a primary disinfection stage
should be implemented.
4. Chlorine, THM and oxygen concentrations decreased in the distribution system of water type III
(WTP 6). In this water, oxygen is possibly consumed
by nitrification of ammonia in the reservoir and the
distribution system. THMs are then possibly degraded in the resulting anaerobic milieu. Even
though this process is desired, low oxygen conditions are unfavorable with regard to several
water quality issues such as corrosion, taste and
odor.

Acknowledgements
This study was funded by the Swiss Agency for
Development and Cooperation (SDC) in the framework of the Swiss–Vietnamese Cooperation Project
ESTNV (Environmental Science and Technology in
Northern Vietnam). We are indebted to the Hanoi
Water Business Company for their cooperation and
sampling assistance.

[6]

[7]
[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

References
[1] Berg M, Tran HC, Nguyen TC, Pham HV, Schertenleib R,
Giger W. Arsenic contamination of groundwater and
drinking water in Vietnam: a human health threat. Environ
Sci Technol 2001;35:2621–6.
[2] Singer PC, Reckhow DA. In: Letterman RD, editor.
Water quality and treatment, a handbook for community
water supplies, 5th ed. New York: McGraw-Hill, 1999. p.
12.1–12.51.
[3] Singer PC. Formation and characterization of disinfection
by-products. In: Craun GF, editor. Safety of water
disinfection: balancing chemical and microbial risk.
Washington, DC: ILSI Press; 1993.
[4] World Health Organization (WHO). Guidelines for
drinking water quality, 2nd ed., vol. 1 Recommendations,
Geneva, Switzerland, 1993.
[5] Miles AM, Singer PC, Ashley DL, Lynberg MC, Mendola
P, Langlois PH, Nuckols JR. Comparison of trihalo-

[16]

[17]

[18]

[19]

3251

methanes in tap water and blood. Environ Sci Technol
2002;36:1692–8.
Golfinopoulos SK. The occurrence of trihalomethanes in
the drinking water in Greece. Chemosphere 2000;41:
1761–7.
Vietnamese Ministry of Health. Standards for drinking
water, Hanoi, Vietnam, 1988.
Standard Methods. In: Eaton AD, Clesceri LS, Greenberg
AE, editors. Standard methods for the examination of
water and drinking water, 19th ed. Washington, DC:
American Public Health Association, 1995.
Pinkernell U, Nowack B, Gallard H, von Gunten U.
Methods for the photometric determination of reactive
bromine and chlorine species with ABTS. Water Res
2000;34:4343–50.
Berg M, Muller
.
SR, Schwarzenbach RP. Simultaneous
determination of triazines including atrazine and their
major metabolites hydroxyatrazine, desethylatrazine, and
deisopropylatrazine in natural waters. Anal Chem
1995;67:1860–5.
Gallard H, von Gunten U. Chlorination of natural organic

matter: kinetics of chlorination and of THM formation.
Water Res 2002;36:65–74.
Symons JM, Krasner SW, Sclimenti MJ, Simms LA,
Sorensen Jr. HW, Speitel Jr. GE, Diehl AC. Influence of
bromide ion on trihalomethane and haloacetic acid
formation. In: Minear RA, Amy GL, editors. Disinfection
by-products in water treatment: the chemistry of their
formation and control. Boca Raton, FL: CRC Lewis
Publishers; 1996. p. 91–130.
Wolfe RL, Ward NR, Olson BH. Inorganic chloramines
as drinking-water disinfectants—a review. J Am Water
Works Assoc 1984;76:74–88.
Morris JC, Isaac RA. A critical review of kinetic and
thermodynamic constants for the aqueous chlorineammonia system. In: Jolley RL, Brungs WA, Cotruvo
JA, Cummings RB, Mattice JS, Jacobs VA, editors.
Water chlorination: environmental impact and health
effects, vol. 4. Ann Arbor, MI: Ann Arbor Science; 1981.
p. 49–62.
Margerum DW, Gray Jr. ET, Huffman RP. Chlorination
and the formation N-chloro compounds in water treatment. In: Brinckman FE, Bellama JM, editors. Organometals and organomataloids: occurrence and fate in
the environment. Washington, DC: ACS Books; 1978.
p. 278–91.
Kumar K, Margerum DW. Kinetics and mechanism of
general-acid-assisted oxidation of bromide by hypochlorite and hypochlorous acid. Inorg Chem 1987;26:
2706–11.
von Gunten U, Driedger A, Gallard H, Salhi E. Byproducts formation during drinking water disinfection: a
tool to assess disinfection efficiency? Water Res 2001;35:
2095–9.
USEPA. Disinfection profiling and benchmarking guidance manual, Appendix C: CT values for inactivation
achieved by various disinfectants, EPA-815-R-99-013,

August 1999, Washington, DC, 1999.
Langlais B, Reckhow DA, Brink DR, editors. Ozone in
water treatment: application and engineering, 2nd ed.
Chelsea, MI: Lewis Publishers, 1991. p. 221.


3252

H.A. Duong et al. / Water Research 37 (2003) 3242–3252

[20] Water Master Plan of Hanoi City for the period of 1993–2001.
The Social Republic of Vietnam, Hanoi People’s Committee
and The Republic of Finland, Finnish International Development Agency FINNIDA, Hanoi, Vietnam, vol. 1, 1993.

[21] Mitch WA, Sedlak DL. Formation of N-nitrosodimethylamine (NDMA) from dimethylamine during chlorination.
Environ Sci Technol 2002;36:588–95.