Optimisation of coagulation and chlorination in drinking

Water Science and Technology: Water Supply Vol 2 No 2 pp 131–137 © IWA Publishing 2002
water treatment: laboratory and full-scale studies
N.D. Basson* and C.F. Schutte
Department of Chemical Engineering, Water Utilisation Division, University of Pretoria, South Africa
* Business address: Sedibeng Water, Private Bag X5, Bothaville, South Africa

Abstract The paper deals with laboratory and full-scale studies aimed at optimising treatment processes at
the Balkfontein plant of Sedibeng Water in South Africa. The raw water is highly eutrophic and contains a
large fraction of treated effluent from domestic and industrial sources as well as agricultural runoff. The
eutrophic nature and changing raw water quality give rise to many operational difficulties and high treatment
costs as well as problems with the final water quality. Optimisation of the coagulation and chlorination
processes was seen as a cheaper solution to these problems than to install advanced processes such as
ozonation and activated carbon adsorption that would add greatly to treatment costs. The laboratory studies
indicated that through optimisation of coagulation-flocculation and by replacement of pre-chlorination by
intermediate chlorination (after primary sedimentation) most of the treatment problems could be solved and
final water of the required quality produced without a large increase in treatment costs.
Keywords Chlorination; drinking water treatment; iron and manganese removal; THM formation

Introduction
The water treatment plant of Sedibeng Water near Bothaville in South Africa with a capac-
ity of 360 Ml/day produces potable water abstracted from the middle Vaal River. The raw
water from this part of the river is highly eutrophic and contains a large fraction of treated
effluent from upstream domestic and industrial sources in the Johannesburg and Vaal
Triangle industrial and mining areas. The raw water quality is also subject to seasonal
variations with algae blooms during spring and autumn, high turbidity during the summer
rainfall months, and high organic loading during the dry winter months when most of the
flow in the river is from treated return flows.
The main problems that gave rise to the project on which this paper is based, include:
• Problems related to effective removal of algae during periods of algae blooms in the raw
water. This requires pre-chlorination (about 4–6 mg/l), which in turn results in taste and
odour problems and in relatively high THM (trihalomethane) levels in the final water.
• Problems related to the removal of relatively high levels of iron and manganese in the
raw water at times, resulting from pollution from industrial and mining activities in the
catchment. Iron and manganese occur as organic complexes in the raw water, making
them difficult to remove. Removal of iron and manganese also requires pre-chlorination
but at elevated pH levels of about 9,6. Incomplete removal of manganese and iron results
in post oxidation and coloured water in the distribution system and in staining of clothes
and household equipment.
• Problems related to the removal of natural organic matter and organic matter resulting
from algae lysis which act as precursors for THM formation upon chlorination.
• Operational problems of short filter runs and occurrence of biological growth in the sed-
imentation tanks and sand filters.
One alternative to deal with the problems of taste and odour, of excessive THM forma-
tion and of incomplete iron and manganese removal would be the implementation of
advanced processes such as ozonation and activated carbon adsorption. These processes 131

Downloaded from https://iwaponline.com/ws/article-pdf/2/2/131/408103/131.pdf

by guest
on 09 October 2019
 would require a very large capital investment and would add to the operating complexity of
the treatment processes. It was therefore decided to rather try and solve the problems by
optimising the coagulation-flocculation and chlorination processes, which would require
very little additional capital investment.
The project involved a series of laboratory studies with the objective to optimise the
coagulation-flocculation and chlorination processes and to identify an optimal process
configuration that would solve the treatment problems. The second phase of the project
N.D. Basson and C.F. Schutte

involved the implementation on full-scale of the optimal configuration in one of the three
parallel modules of the treatment plant in order to verify the findings of the laboratory
studies.

Historical developments
Historically, the high-lime coagulation-flocculation process at a pH of 11.2 or higher was
used. At the time very few treatment problems were experienced. However, treatment costs
were high and substantial sludge disposal problems were experienced. The available
sludge disposal lagoons were filling up and high capital costs were needed to expand the
sludge disposal facilities. This was one of the main reasons why it was decided to replace
high-lime treatment with ferric chloride as primary coagulant. The new coagulation-
flocculation system performed well at a much lower cost than the high-lime process and
with much less sludge production.
Difficulties were however experienced during periods of algae blooms in the raw water
in removing the algae effectively. Early filter breakthroughs occurred and the final water
turbidity on occasions exceeded the target value of less than 1 NTU. Pre-chlorination at lev-
els of between 4 and 6 mg/l was introduced to deal with this problem. This achieved effec-
tive removal of most algae species, but resulted in THM levels that exceeded the
recommended maximum level of 100 µg/l during most of the periods when algae blooms
occurred. A further problem became evident when consumer complaints were received
about tastes and odours in the drinking water. These problems were caused by pre-chlorina-
tion of the raw water containing high algae counts.
At this point the water still contains dissolved and particulate contaminants, specifically
organic matter which:
• Contribute to an increased chlorine demand (Van Steenderen et al. 1991);
• Upon chlorination result in the release of organic substances by algae (Bernhardt et al.,
1986)
• Result in an increased potential for trihalomethane (THM) formation (Basson and
Pieterse, 1993)
• Also negatively impact on tastes and odours associated with certain algae as was experi-
enced at the Balkfontein plant.
At the time problems were also experienced with ineffective removal of iron and
manganese from the raw water. Complaints were received of red water in the distribution
system during certain periods and of staining of laundry. The pre-chlorination level had to
be increased to 6 mg/l on occasions and the pH had to be raised to a level of about 9.6
to improve oxidation and removal of manganese. This in turn, resulted in lower removal
efficiencies of the organic matter in the water that acts as a THM precursor, and therefore
higher levels of THMs in the treated water. These problems prompted the search for
solutions along the lines of improving coagulation-flocculation and changes to the
approach to chlorination.
An alternative treatment approach which formed the basis of this project, would be to
move the point of chlorine application to a position after coagulation and sedimentation
132 where a large fraction of organic matter has been removed (Jiang and Graham, 1992). The

Downloaded from https://iwaponline.com/ws/article-pdf/2/2/131/408103/131.pdf

by guest
on 09 October 2019
 objective was to retain the benefits of pre-chlorination while minimising the negative
effects.
Table 1 gives an indication of the relatively poor quality of the raw water during 1998
when the studies were conducted. The final water quality as well as the South African target
water quality for certain parameters in drinking water is also given.

Treatment plant layout

N.D. Basson and C.F. Schutte

The treatment plant layout is shown in Figure 1. It consists of three modules operated in
parallel, each with a capacity of 120 megalitre per module.
Pre-chlorination is applied ahead of coagulation, followed by lime to increase the pH to
about 9.6. The pH is reduced by addition of carbon dioxide after primary sedimentation to
around 8. Modules 2 and 3 of the plant were used for the full-scale experimental work.
Module 3 consists of four similar sections. Chlorine-dosing points were installed in two
sections of Module 3, after primary sedimentation, to investigate the effect of intermediate
chlorination. The other two sections of this module were used for control purposes. Module
2 was used to investigate the role of pre-chlorination.

Laboratory studies
The laboratory studies involved different series of jar tests conducted on a microprocessor-
controlled apparatus. The procedure was aimed at simulating the coagulation-flocculation

Table 1 Quality of raw and treated water at the Balkfontein plant

Parameter Units Raw water: Raw water: Treated Treated Target

Average maximum water: average water: maximum water quality

Turbidity NTU 48 265 0.6 2.9 <1

TDS mg/l 372 698 363 572 <450
pH 8.7 9.8 8.2 11.0 6–9
DOC mg/l 8.1 10.3 7.1 8.4 <5
Manganese µg/l 122 858 27 156 <50
Iron µg/l 468 13111 118 560 <100
Chlorophyl-a µg/l 47 290 0.8 4.2 <1
THM µg/l – – 102 131 <100

New Pump Station Old Pump Station

High Pressure Pump Station

Reservoir Reservoir
Pre-chlorination

Post Chlorination

Intermediate Chlorination –
and Recabonation mixing tanks
Filter block
Sedimentation tanks

Secondary

Module 1 Module 2 Module 1

Primary

Serpentine channels

Chemical Dosing (Rapid mixing)

Figure 1 Schematic illustration of the water treatment plant at Balkfontein 133

Downloaded from https://iwaponline.com/ws/article-pdf/2/2/131/408103/131.pdf

by guest
on 09 October 2019
 and sedimentation full-scale processes. Comprehensive testing by means of jar tests was
done to establish and compare the effect of chlorination, coagulants and coagulant dosages
at different pH levels on the removal of algae, organic matter, iron and manganese and on
the formation of THMs. The procedure was aimed at simulating the treatment process of
coagulation and settling, retaining the advantages of chlorination in the removal of algae
and at the same time reducing the THM formation potential (THMFP).
The tests were conducted in sets of six samples (one litre each) in square beakers accord-
N.D. Basson and C.F. Schutte

ing to the procedure described in Table 2. Ferric chloride supplied as a 43% FeCl3 solution,
was used as primary coagulant. Dosing concentrations are expressed as mg FeCl3 per litre
raw water. A cationic organic polymeric coagulant was used as secondary coagulant. A
chlorine stock solution was made by bubbling chlorine gas through deionised water
(Steynberg et al., 1996).
In the case of pre-chlorination, the samples were stirred for one minute at 250 revolu-
tions per minute (rpm) after the addition of the chlorine solution. The addition of the
primary coagulant was followed by rapid stirring at 250 rpm for two minutes. The stirring
time was however changed to one minute when a secondary coagulant was used. After
adding the secondary coagulant, further stirring for one minute, at 200 rpm, followed.
Flocculation was simulated by stirring at 40 rpm for eight minutes. The sample was stirred
for one minute at 25 rpm after intermediate chlorination.
The supernatant of each sample was drawn off after the prescribed settling times. To
compare jar test results with sand filtered water, the supernatant in each case, was filtered
through a Whatman No. 1 filter paper. Samples were analysed for pH, turbidity, THMs,
iron and manganese as well as algae identification.

Results
Due to length restrictions, only the results for chlorophyll-a removal are given in detail
while results on removal of other parameters are given in summarised form. Detailed
results are reported elsewhere (Basson, 2000).

Chlorophyll-a removal
Chlorophyll-a concentrations in the raw water varied between 14 and 158 µg/l during the
test period. Peak values were measured during August and September. During this period,
between 60 and 95% of the total algae population was from the Chlorophyceae group. The
effect of chlorination on the removal of chlorophyll-a is shown in Figure 2.

The effect of chlorination. A comparison between the chlorophyll-a reductions obtained

when applying chlorine at different dosing points is illustrated in Figure 2 (no lime was
added). These results indicate that there was almost no difference in the removals obtained
by pre-chlorination and intermediate chlorination. The benefits of pre-chlorination can

Table 2 The procedure for conducting jar tests (+ indicates the procedure was included, – not included)

Control Pre-chlorination Intermediate

(C) (PC) Chlorination (IC)

Pre-chlorination 4 mg/l – + –
Primary Coagulant FeCl3 (various concentrations) + + +
Secondary Coagulant (cationic poly-electrolyte: 1 mg/l) + + +
Flocculation + + +
Sedimentation 20 min 20 min 10 min
Chlorination of supernatant 2 mg/l – – +
Sedimentation – – 10 min
134

Downloaded from https://iwaponline.com/ws/article-pdf/2/2/131/408103/131.pdf

by guest
on 09 October 2019
 Pre-chlorination Intermediate chlorination Control

120

100

80
% Reduction

N.D. Basson and C.F. Schutte

60

40

20

0
1 2 3 4 5 6 7 8 9
Test Runs
Figure 2 The effect of moving the point of chlorine dosing on chlorophyll-a removal

therefore be obtained at a decreased cost and with fewer water quality problems when
pre-chlorination is replaced by intermediate chlorination.

The effect of coagulants. Chlorophyll-a removal was enhanced by the addition of 1 mg/l of
a poly-electrolyte as a secondary coagulant. The improvement obtained was more signifi-
cant when pre-chlorination was excluded and at the lower FeCl3 dosage concentrations.
These results also indicate that poly-electrolyte could be used for the removal of chloro-
phyll-a concentrations exceeding 30 µg/l instead of applying pre-chlorination. Improved
removal at the higher coagulant concentrations could be related to the higher coagulant
demand due to organic substances present in the water (Edzwald, 1993). The effect of pre-
chlorination on the removal of chlorophyll-a was more significant at lower coagulant
dosages than at the higher dosages, probably due to the fact that the coagulant demand was
not satisfied at the lower dosages and that pre-chlorination, under these circumstances,
could add positively to the removal of chlorophyll-a. Chlorination however, adversely
affects the role of poly-electrolytes (Steynberg, 1994) and therefore the smaller impact on
chlorophyll-a removal in the pre-chlorinated samples.

THM formation
Results on the formation of THMs upon post chlorination for different treatment options
are given in Table 3.
The results in Table 3 show that less DOC removal was achieved at the higher pH
values and when pre-chlorination was applied. This consequently resulted in higher THM
formation at higher pH values. Increasing the pH to 11.4 however enhanced the removal of
organic substances due to precipitation of Mg(OH)2 .

Full-scale studies
Based on the laboratory studies it was concluded that the operational and water quality
problems could to a large extent be eliminated by replacing pre-chlorination with interme-
diate chlorination (after primary sedimentation) and by improving coagulation-floccula-
tion (with higher dosages and addition of a secondary coagulant). One of the 120 megalitre
modules was operated with this configuration while the other two modules were operated
according to the old configuration with pre-chlorination. The performance of the modules 135

Downloaded from https://iwaponline.com/ws/article-pdf/2/2/131/408103/131.pdf

by guest
on 09 October 2019
 Table 3 The effect of pH on DOC removal and the formation of THMs (– indicates that this step was
excluded from the procedure and + indicates that this step was included in the procedure)

Pre-chlorination Lime added PH FeCl3 dosing % DOC removal THMs after post
concentration mg/l chlorination mg/l

+ – 8.4 35 15.5 99.8

– – 8.4 35 24.8 80.8
+ + 9.8 35 12.7 118.6
N.D. Basson and C.F. Schutte

– + 9.8 35 20.6 88.6

– + 11.4 6 58.9 51.7

was monitored over a period of a week for four different test runs in an attempt to investi-
gate different raw water qualities.

Chlorophyl-a
The effect of pH, chlorination and coagulants on the removal at full-scale of chlorophyll-a,
can be summarised as follows:

pH and chlorination. The full-scale investigations confirmed the findings in laboratory

experiments, which were conducted over a wide range of different conditions:
• the efficiency of chlorine in removing chlorophyll-a was adversely affected by increas-
ing the pH
• chlorophyll-a could be removed effectively by intermediate chlorination even when raw
water concentrations exceeded 30 µg/l
From these findings it can be concluded that intermediate chlorination can be used as an
economic substitute for pre-chlorination for the removal of chlorophyll-a at the
Balkfontein plant.

Coagulants. As only one set of analyses per test run is available it cannot be concluded
without any doubt that coagulant and coagulant dosage are positively related to the removal
of chlorophyll-a in the full-scale operation. The results obtained are however in confirma-
tion with bench-scale experimental results which indicated that chlorophyll-a removal is
enhanced by increased coagulant dosage and by the addition of a poly-electrolyte as sec-
ondary coagulant.

THM Formation during full-scale runs

The concentration of THMs measured in the final water after disinfection is shown in
Figure 3. The highest THM levels were measured in the line where pre-chlorination was
applied and when the coagulation pH was 9.8. These results correlate with laboratory find-
ings, which showed that the replacement of pre- by intermediate chlorination results in a
decrease in THM formation.

Removal of iron and manganese

The removal of these metals presents problems during certain periods, even at relatively
high pH levels and with high chlorine dosages. The factors that affect removal form the
basis of further studies. At this stage the only conclusion that could be made is that the met-
als occur as organic complexes that are stable and difficult to break down and therefore dif-
ficult to remove to low levels.

Conclusions
136 The problems with final water quality and with certain aspects of the operation of the

Downloaded from https://iwaponline.com/ws/article-pdf/2/2/131/408103/131.pdf

by guest
on 09 October 2019
 Pre-chlorination Intermediate chlorination Control

200

180

160

140
THM (µg/l)

120

N.D. Basson and C.F. Schutte

100

80

60
lime added
40 pH 9,8

20

0
13/09/97 21/02/98 22/02/98

Date
Figure 3 The formation of THMs after disinfection during full-scale operation

Balkfontein plant could to a large extent be solved by replacement of pre-chlorination with

intermediate chlorination and by improvement of the coagulation-flocculation process.
Maximum removal of algae and organic matter by enhanced coagulation-flocculation
before intermediate chlorination results in lower THM formation and fewer problems with
taste and odour in the final water. Operating experience has shown that occasional pre-
chlorination (a few hours per week) is still necessary to prevent biological growth in intake
lines and primary sedimentation, but this does not affect the final water quality

References
Basson, N.D. (2000). The effect of pH, coagulation and chlorination on the production of potable water
from eutrophied surface water. M.Sc. thesis, University of Pretoria.
Basson, N.D. and Pieterse, A.J.H. (1993). The occurrence and origin of THM substances in a full-scale plant
purifying water from the middle Vaal River, Proceedings of the 1993 WISA conference, Durban, South
Africa pp 46–58.
Bernhardt, H. (1996). Studies on the treatment of eutrophic water. Special Subject 12, Wahnbachtal –
Sperrenverband, Siegburg, Federal Republic of Germany.
Bernhardt, H., Hoyer, O., Lüsse, B. and Schell, H. (1986). Investigations on the influence of algal divided
organic substances on flocculation and filtration. Proceedings from the Second national Conference on
Drinking Water, Edmonton, Canada.
Edzwald, J.K. (1993). Algae, bubbles, coagulants and dissolved air flotation. Wat. Sci. Tech. 27(10) 67–81.
Jiang, J. and Graham, N.J.D. (1992). Removal of algae and trihalomethanes (THM) precursors by coagula-
tion. Water Treatment, 7 155–168.
Steynberg, M.C. (1994). Pre-oxidation: An unit process to limit algal – related water quality and purification
problems. Unpublished PhD thesis, University of the Orange Free State, Bloemfontein, South Africa.
Steynberg, M.C., Guglielmi, M.M., Geldenhuys, J.C. and Pieterse, A.J.H. (1996). Chlorine and chlorine
dioxide; pre-oxidants used as algaecide in potable water plants. Journal Water SRT – Aqua 45(4)
162–170.
Van Steenderen, R.A., Pieterse, M.J. and Bourne, D. (1991). THM formation in potable waters with refer-
ence to related variables and health databases. Water SA 17(4), 269–272.

137

Downloaded from https://iwaponline.com/ws/article-pdf/2/2/131/408103/131.pdf

by guest
on 09 October 2019
Downloaded from https://iwaponline.com/ws/article-pdf/2/2/131/408103/131.pdf
by guest
on 09 October 2019