Desalination 276 (2011) 287293

Contents lists available at ScienceDirect

Desalination
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / d e s a l

Performance evaluation of biosand lter modied with iron oxide-coated sand for
household treatment of drinking water
M. Mansoor Ahammed , Komal Davra
Civil Engineering Department, S V National Institute of Technology, Surat 395007, India

a r t i c l e

i n f o

Article history:
Received 28 October 2010
Received in revised form 25 February 2011
Accepted 24 March 2011
Available online 6 May 2011
Keywords:
Bacterial removal
Household water treatment
Iron oxide-coated sand
Turbidity removal
Water purication

a b s t r a c t
The biosand lter (BSF), intermittently operated household slow-sand lter, was modied by introducing a
10-cm thick layer of iron oxide-coated sand. Long-duration (about four months) tests were conducted to
compare the performance of the modied BSF (MBSF) with the conventional BSF in terms of their efciency in
removing bacteria and turbidity under different operating conditions. Filters were charged daily with 20 L or
40 L natural canal water (turbidity 10.0 1.2 NTU; faecal coliforms 365 251 MPN/100 mL; pH 8.4 0.4) or
seeded tap water (turbidity 14.7 4.3 NTU; Escherichia coli 3850 736 CFU/mL; pH 7.9 0.3). Results
showed that the performance of MBSF in terms of faecal coliform and E. coli removals was better by at least
one-log10 unit throughout the lter operation. The mean bacterial removal was low for BSF for the rst month
(90.0%), while it was 99.3% for MBSF during the same period. Bacteria and turbidity removals increased with
time as lter ripening (maturation) occurred in both the lters. No signicant difference was observed in
turbidity removal between BSF and MBSF, and mean efuent turbidity was around 1 NTU for BSF and MBSF
representing N 90% removal. When daily charge was increased from 20 L to 40 L, a reduction in bacterial
removal was noted in both the lters indicating the inuence of operating conditions. Efuent physicochemical quality remained within the guideline values for drinking water.
2011 Elsevier B.V. All rights reserved.

1. Introduction
Waterborne diseases continue to be a major cause of illness and
death in communities with little access to safe drinking water. World
Health Organisation estimates that globally, 1.8 million deaths occur
per year due to a combination of inadequate sanitation and poor
water quality [1]. Public water supplies in developing countries often
fail to produce and distribute water safe for consumption which
makes the responsibility of ensuring safe drinking water with the
consumer. A number of household water treatment methods such as
ltration, occulation, chlorination and solar disinfection have been
found to be effective in improving microbial quality of water [2].
One of the most promising and accessible technologies for
household water treatment is biosand ltration. The biosand lter
(BSF) is an intermittently operated slow sand lter (SSF) at small
scale. The lter consists of a bed of ne sand supported by a layer of
gravel enclosed in a box with appurtenances to deliver and collect the
water [3]. The BSF, allows a SSF to be operated intermittently, thus
making it more suitable for household applications. This is achieved in
the design by making the highest point of the outow tubing above
the sand, thus ensuring that water covers the sand at all times. A
diffuser plate placed above the level of the water protects the sand

Corresponding author. Tel.: + 91 98258 75162.

E-mail address: mansoorahammed@gmail.com (M.M. Ahammed).
0011-9164/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.desal.2011.03.065

below from damage when water is poured into the system. A clean in
place technique minimises the need for sand bed removal,
simplifying maintenance and increasing continuity of performance
[4,5]. As in conventional SSFs, during the ripening process a biolayer
(schmutzdecke) forms, head loss increases and performance improves.
Since the lter is generally charged once daily, a portion of the
charged water remains in the BSF until the next charge. It is reported
that approximately 143,000 BSFs were in operation as of June 2007,
serving an estimated 858,500 users in 36 countries. For the year, they
produced nearly 1.3 billion litres of drinking water (at 25 L/unit/day)
[5]. Recently, based on performance and sustainability criteria, Sobsey
et al. [6] identied biosand lters as most effective method among the
ve different household treatment technologies, and as having the
greatest potential to become widely used and sustainable for
improving household water quality to reduce waterborne disease
and death. Health impact studies in different countries have shown
substantial reduction in childhood diarrhoeal diseases in households
using biosand lters [7,8].
Several studies have been reported on the performance of the BSFs in
reducing bacteria, viruses and turbidity from feed water. Bacterial
removals have been reported to vary from no apparent removal to 99%
depending on operating conditions and lter ripening [9,10]. Elliott et al.
[11] demonstrated that reductions of bacteria and viruses tend to be
lower than those reported for traditional SSF. While slow sand lters
have been shown in the laboratory to reduce faecal bacteria by 23 log10,
viruses by 1.52.0 log10 and Cryptosporidium oocysts by more than 5

288

M.M. Ahammed, K. Davra / Desalination 276 (2011) 287293

log10 [12], tests of the biosand lter both in the laboratory and in the
eld achieved a mean reduction of E. coli of less than 2 log10 [11,13].
Baumgartner et al. [10] reported total coliform removal by biosand lter
in the range 58.399.7% under different operating conditions and at
different sample collection points. Murphy et at. [14] reported E. coli
removal in the range 099.99% in a eld study in Cambodia. Bacterial
removal generally increases substantially with repeated charges and
time in use as lters matured, and increased retention time of water in
the lter bed. Ripening (maturation) typically takes 4 to 6 weeks. It is
also known that cleaning the biolter also diminishes its microbial
removal efciency possibly because the disturbed biolm is less efcient
or because cleaning helped bypass microbially active zones of the lters
[15]. The impaired performance of the lters persisted for approximately 46 weeks following cleaning. These ndings raise questions
about the assurance of safe water provided to users of BSF before
ripening and after cleaning [11], and indicate the need for modication
of BSF in order to improve the performance of the BSF during the
ripening period and also after cleaning the lter.
Metallic oxides/hydroxides are good sorbents of bacteria and
viruses. This has led to the development of metal hydroxide-coated
sand as a ltration/adsorption medium in water and wastewater
treatment. Iron oxide-coated sand (IOCS) was shown to be highly
efcient in removing different microorganisms and turbidity [1620].
For example, column packed with sand coated with ferric and
aluminium hydroxides removed greater than 99% of E. coli, V. cholerae,
poliovirus and coliphage MS-2 from dechlorinated tap water [18].
Ahammed and Chaudhuri [16] reported that IOCS removed 8084%
heterotrophic bacteria, 9496% E. coli and 7383% poliovirus from
canal water.
A few attempts have been made to modify the BSF to improve its
performance. These include introduction of pre-lter as a pretreatment unit, and use of smaller size lter media to reduce the
ltration velocity. However, the results were not very promising [21].
In the present study, the biosand lter was modied by introducing
a thin layer (10 cm) of IOCS in the sand bed. Performance of the
modied BSF was compared with a BSF in a long-duration study
spanning about four months, during which bacteria and turbidity
reduction, ow rate variation, physico-chemical characteristics and
other relevant parameters were monitored. A canal water and a
spiked tap water were used as feed.
2. Materials and methods
2.1. Filter installation
Locally available river sand was selectively sieved, using a set of
sieves for sand analysis and the portion passing through 1.18 mm
sieve and retained on 0.150 mm was used in the present study as lter
media. The lter media had an effective size of 0.23 mm and a
uniformity coefcient of 3.1. The sand was washed several times using
tap water until the wash water became clear. Two biosand lters were
used in this study, a conventional BSF (BSF) and a modied BSF
(MBSF). Plastic containers obtained from the local market were used
for constructing the lters. The containers were rst cleaned with
tapwater, and were lled with 5 cm deep underdrain gravel (4.75
12.0 mm size), 5 cm of coarse sand (1.184.75 mm size) separation
layer and 40 cm of sand (0.1501.18 mm size) layer in succession. In
the case of MBSF, the 40-cm sand layer included a 10-cm iron oxidecoated sand (IOCS) layer in the middle of the sand layer. Water was
present in the containers before loading the lter media to avoid any
occurrence of air spaces and short circuiting. The outlet pipe was
provided in such a manner that a water depth of 5 cm is maintained
over the lter media. A plastic diffuser plate was placed on the lip of
the lter to avoid disturbance to the top layer of sand during daily
charging of the lter with raw water. Schematic diagram of the lter is
presented in Fig. 1. The lters had a pore volume of 19.1 L.

2.2. Preparation of iron oxide-coated sand

The sand used for coating had similar specications as that used in
BSF. The sand was soaked in 8% nitric acid solution, rinsed with
deionised water to pH 7.0, and dried at 105 C in preparation for surface
coating. In the present study the method suggested by Liu et al. [22] was
employed for coating the sand with iron hydroxide. In order to produce
iron hydroxide-coated sand, 500 g of ferric chloride was dissolved in 1 L
of deionised water. The iron solution was mixed with 3 L of acid-washed
sand. This was then dried for 48 h at 90 C. Coated sand was neutralised
in 5 L of 5 M sodium hydroxide solution for 24 h and then dried at 90 C
for 24 h. After drying, the coated sand was rinsed with deionised water
to pH 7.0, redried at room temperature.

2.3. Water
Water from two different sources was used in the study: i) Tap
water collected from tap of Environmental Engineering Laboratory of
the Civil Engineering Department of Sardar Vallabhbhai National
Institute of Technology, Surat, India and ii) canal water from an
irrigation canal near ONGC Colony, Surat, India. Since the tap water had
no/very little bacterial concentration, spiking of the microorganism
was necessary. Further, since turbidity and the total organic carbon
concentration were also low in tap water, sterilised sewage and locally
available clay were added to the feed water. It was thought that
addition of sewage would simulate the presence of wastewater in
typical drinking water sources of developing countries and accelerate
the ripening process [11]. However, since the canal water had
sufcient levels of turbidity and microorganisms no spiking of
microorganisms or turbidity was needed. Canal water was collected
weekly and stored at 4 C until one day prior to feeding. Spiked tap
water was prepared daily. Characteristics of the two feed waters
during the study period are presented in Table 1.

2.4. Filter operation

Two ltration experiments, one with spiked tap water (Run 1) and
the other with canal water (Runs 2) as inuent were conducted
consecutively. BSF and MBSF were operated in parallel under identical
conditions. The lters were charged once a day with raw water
prepared daily. The schedule of the lter operation is presented in
Table 2. In the case of 40-L charge, since the reservoir of the lters
could hold only about 25 L, the second 20 L was poured one hour after
the rst 20 L water was poured. The lters were cleaned after Run 1 as
described in Section 2.6. The test was conducted at room temperature,
and water temperature varied in the range of 2634 C during the
testing period.

2.5. Sample collection

Different types of samples were collected at different intervals for
analysis: (1) inuent water; (2) grab samples of the ltered water
taken throughout a daily charge, and (3) composite sample of the
ltered water to measure the daily average concentration of different
water quality parameters. Grab samples were analysed for microbial
quality and dissolved oxygen concentration while composite samples
were analysed for all relevant water quality parameters including
turbidity, pH and microbiological quality. Sample bottles of volumes
of 1 L were used to collect water samples from each lter. Water
samples for microbiological testing were collected in sterilised bottles.
Tests showed insignicant die-off rates for challenge microorganisms
after overnight storage at room temperature. Thus, the effects of time
on microbial survival were regarded as negligible in this study.

M.M. Ahammed, K. Davra / Desalination 276 (2011) 287293

289

Fig. 1. Schematic diagram of the BSF used in the study. All dimensions are in mm. In the case of modied BSF (MBSF) a 10-cm thick layer of iron oxide-coated sand was introduced in
the middle of the sand lter media keeping the total depth of the lter media the same (400 mm).

2.6. Filter cleaning

The lter was cleaned after the rst run (with spiked canal water).
During cleaning, the standing water below the strainer plate on the
top of the sand bed was displaced, and the top 5-cm layer of sand bed
was scrapped off manually. A few litres of clean water was poured on
top of the sand and the lter top was again cleaned by agitating and
rinsing the sand. This water was scooped off with a small cup and
discarded. This was repeated four times until the water was clear. A 5cm layer of fresh sand was placed on the top of the lter. The cleaned
lter was ushed following this.
2.7. Bacterial analysis
Stock culture of Escherichia coli (ATCC No. 4157) was obtained
from National Collection of Industrial Microorganisms, Pune, India.
Single colonies from E. coli stock culture were inoculated into sterile
nutrient broth (Himedia Laboratories, Mumbai, India) and incubated
in a shaker at 100 rpm for overnight. Bacterial suspension in
stationary phase was harvested by centrifugation at 1000 g for
10 min. Pelleted bacteria were suspended in phosphate buffer
solution. E coli was added to raw water samples so as to obtain the
desired initial E. coli concentration.
Table 1
Characteristics of the water used in different tests.
Parameter

Spiked tap water

Canal water

Turbidity (NTU)
pH
Alkalinity (mg/L as CaCO3)
Hardness (mg/L as CaCO3)
Temperature (C)
E. coli (CFU/mL)
Faecal coliforms (MPN/100 mL)
Total coliforms (MPN/100 mL)
Heterotrophic plate count (CFU/mL)
Dissolved oxygen (mg/L)

14.7 4.3 (14)

7.9 0.3 (14)
40 7 (5)
36 6 (5)
28 2.0 (45)
3850 736 (13)
Not detected
Not detected
Not detected
6.3 0.5 (8)

10.0 1.2 (14)

pH 8.4 0.4 (9)
65 12 (5)
72 6 (5)
29 2.4 (14)
Not detected
365 251 (9)
4250 1345 (5)
2.5 104 2432 (5)
6.0 0.4 (6)

Values represent mean standard deviation. Values in parentheses represent the

number of samples analysed.

Enumeration of viable E. coli was conducted by plate count test

using MacConkey agar. The concentrations of total coliforms and
faecal coliforms were estimated by multiple-tube fermentation
method (most probable number method). The result is expressed as
most probable number per 100 mL (MPN/100 mL). Lauryl Tryptose
Broth was used for the presumptive phase of total and faecal
coliforms, and Brilliant Green Bile Broth and EC medium were used
for conrmation phase of total and faecal coliforms, respectively. The
samples were collected and a dilution series was performed; the
dilutions needed for plating and MPN test were estimated based on
expected removal. Heterotrophic bacteria were enumerated using the
pour plate method. In this method, 0.5 mL of water sample (or diluted
water sample) was poured into a petridish. 56 mL of plate count agar
was poured over this and was mixed well. The petridishes were
incubated at 37 C for 48 h for development of colonies. The result is
expressed as colony-forming units per mL (CFU/mL). All the tests
were conducted in accordance with the techniques described by
American Public Health Association [23].
2.8. Chemical analysis
The analytical determination of the different physicochemical
parameters was carried out within the holding time of each parameter
according to American Public Health Association [23]. The pH of the
samples was measured using a digital pH meter (Systronics, India).
Hardness and alkalinity were determined using titration methods.
Dissolved oxygen was measured with a digital DO meter. Turbidity
was measured using a Hach 2100P turbidimeter. Iron was analysed

Table 2
Schedule of lter operation.
Filter run

Raw water

Days of operation

Daily charge

Run 1

Spiked tap water

Filter cleaning
Run 2

Canal water

156
5766
67
67114

20 L
40 L

20 L

The lters were cleaned after Run 1 on Day 67 as explained in Section 2.6.

290

M.M. Ahammed, K. Davra / Desalination 276 (2011) 287293

25

using phenanthrolene method [23] with Varian CARY50 UVVisible

spectrophotometer.

Run 1
40 L

20 L

3.1. Filtration rate

Fig. 2 shows the changes in ltration rates for the two ltration runs.
It is seen that the BSF and MBSF showed similar reduction of ltration
rate. For lter Run 1, from an initial ow rate of about 0.95 L/min it was
reduced to 0.35 L/min for both the lters. A decline in the ow rate is
expected due to lter maturation and particle accumulation. Since both
the lters received the same raw water, similar head loss development
was expected in BSF and MBSF. It should be mentioned that ltration
rate presented in the gure was based on the rst 1 L of ltered water
and not for the entire 20 L of daily charge. Initially, in Run 1, for the 20-L
daily charge, it took about 25 min to pass through the lters, and it
increased to about 2 h on the last day of operation due to increased head
loss.
3.2. Turbidity reduction
Inuent and efuent turbidity for the two lter runs are presented
in Fig. 3. It is seen that the lters showed similar performance in
turbidity removal in both the runs except for the rst few days when
MBSF showed a slightly higher efuent turbidity compared to BSF.
Though it was suspected that it could be due to leaching of iron oxide
coating from IOCS, this was ruled out since the iron concentration in
the efuent from MBSF was not signicantly higher than that from
BSF. Thus, this higher efuent turbidity could be due to some shortcircuiting.

15

10

0
0

10

20

30

40

50

60

70

14

Run 2
12

Turbidity (NTU)

3. Results and discussion

Turbidity (NTU)

20

MBSF
BSF
Influent

10
8
6
4
2
0
0

10

20

30

40

50

60

70

Days of filter operation

Fig. 3. Turbidity removal over the length of the lter runs. Run 1 used spiked tap water
and Run 2 used natural canal water. Characteristics of the inuent are given in Table 1.
In case of Run 1, daily charge was 20 L for the rst 56 days and was 40 L for the next
10 days. For Run 2, daily charge was 20 L throughout.

1.0

Run 1

Flow rate (L/min)

0.8

0.6

0.4
40 L

20 L

0.2

0.0
0

10

20

30

40

50

60

70

1.0

While the mean inuent turbidity was 14.7 NTU for Run 1, the
mean efuent turbidity were 1.0 and 1.2 NTU, representing about 93%
and 92% removal for BSF and MBSF, respectively. Turbidity removal
improved with time in both the lters as the lters matured. For
example, in Run 1, turbidity removal increased from an initial value of
80% to 96% for BSF. The turbidity removal observed in the present
study for BSF is similar to the reported values in the literature [11,13].
However, these values are lower than that for slow sand lters (SSF)
which could be due to higher ltration rates and shorter media bed
depth of the BSF compared to SSF. Results of the study also showed
that introduction of IOCS layer did not improve the turbidity removal.
Results also showed that when inuent turbidity gradually increased
from about 10 NTU to 20 NTU in Run 1, the efuent turbidity levels
were not adversely affected in both the lters (Fig. 3).

Run 2
3.3. Bacterial removal

Flow rate (L/min)

0.8

MBSF
BSF

0.6

0.4

0.2

0.0
0

10

20

30

40

50

60

70

Days of filter operation

Fig. 2. Flow rate variation over the length of the lter runs. Run 1 used spiked tap water
and Run 2 used natural canal water. Characteristics of the inuent are given in Table 1.
In case of Run 1, daily charge was 20 L for the rst 56 days and was 40 L for the next
10 days. For Run 2, daily charge was 20 L throughout.

Inuent and efuent concentrations of E. coli and faecal coliforms

for the two runs are presented in Fig. 4. In order to aid discussion, the
results of Run 1 for different stages of operation are summarised in
Table 3. It is evident that MBSF always gave better performance in
terms of bacteria removal. Initially, in Run 1, less than one-log10
reduction of E. coli was observed for BSF while MBSF gave above 2log10 removal. Performance of both the lters improved with time and
after about one month, E. coli reduction increased to 2-log10 unit for
BSF and N2.5 log10 for MBSF. It is clear that maturation of the BSF took
about one month and thereafter the lter performed consistently with
about 2-log10 removal. Increased removal with time can be attributed
to both lter ripening and decreased ow rate due to increased head
loss.
Test with canal water (Run 2) also gave similar performance with
MBSF showing about 1-log10 unit higher faecal coliform reduction

M.M. Ahammed, K. Davra / Desalination 276 (2011) 287293

20 L

40 L

103

102

101

100

Faecal coliform concentration (MPN/100 mL)

deteriorated (Fig. 4 and Table 3). This indicates the effect of daily
charge volume on bacterial removal. This may be explained on the
basis of the water that remains within the lter during the idle period
between charges. It may be noted that the pore volume of the lter
was about 19.1 L, and thus most of the ltered water in the 20-L
charge was from the feed water that was introduced the previous day,
thus giving a detention time of almost one day. The study thus shows
that the amount of water per charge and pause period are important
operating parameters for both BSF and MBSF. Baumgartner et al. [10]
reported that total coliform removal by biosand lter decreased with
an increase in charge volume. Similar observations were also made by
Elliott et al. [11].

Run 1

10

20

30

40

50

60

70

104

Run 2
103

3.4. Efuent physico-chemical quality

MBSF
BSF
Influent

102

101

100
0

10

20

30

40

50

60

70

Days of filter operation

Fig. 4. Bacterial concentration over the length of the lter runs. Run 1 used spiked tap
water and Run 2 used natural canal water. Bacterial concentrations are plotted on a
logarithmic scale. Characteristics of the inuent are given in Table 1. In case of Run 1,
daily charge was 20 L for the rst 56 days and was 40 L for the next 10 days. For Run 2,
daily charge was 20 L throughout.

compared to BSF. In this run also, for BSF, bacteria removal improved
with time. Since the efuent concentration in MBSF was already very
low from the beginning of the ltration run due to low inuent values,
there was no scope for further improvement in removal efciency.
It may be noted that in the present study, after the rst run, the
lters were cleaned by scrapping the top 5 cm of sand and
replenishing it with new sand. This is not the recommended method
for biosand lters since this would result in the removal of the
biolayer and consequent reduction in microbial removal efciency.
Since the present study was designed to assess the impact of coated
media on the performance of biosand lter, this cleaning procedure
was adopted before starting the second run of the lter employing
canal water. Ideally, second run should have been conducted using
new lters. However, since it is known that biological activity in slow
sand lters occurs mainly in the top few centimetres, the cleaned lter
could be assumed to be as good as a fresh lter. The results also
support this.
In Run 1, when the daily charge was increased to 40 L (that is about
2 pore volume of the lter), the performance of both the lters

Table 3
Escherichia coli removal in different phases of Run 1.
Days of
operation

Daily
charge (L)

E. coli concentration (CFU/mL)

Inuent

MBSF efuent

BSF efuent

115
1630
3156
5766

20
20
20
40

4616 475 (3)

3716 340 (3)
3775 630 (4)
3363 1010 (3)

42 15 (3)
17 12 (3)
3 2 (4)
16 8 (3)

730 191 (3)

128 96 (3)
11 8 (4)
89 24 (3)

Values represent mean standard deviation. Values in parentheses represent the

number of samples analysed.

Efuent physico-chemical quality was monitored during the lter

operation. No signicant difference was observed for parameters such
as conductivity, alkalinity, pH, calcium and magnesium between the
two lters and between efuent and inuent (data not shown). Iron
concentration in the MBSF efuent was below the drinking water
guideline value of 0.3 mg/L except for the rst week.
Dissolved oxygen (DO) concentration of the efuent was monitored for both the lters and the results are presented in Fig. 5 for Run
1. It is seen that passage of water through the lter resulted in a
reduction of DO concentration in both BSF and MBSF. The decline,
however, was more in BSF compared to MBSF indicating different
mechanisms taking place in the two lters. It is also seen that with
passage of time the decline in DO increased in both the lters and was
stabilised after about one month which possibly indicated lter
maturation. Further, when the daily charge was increased from 20 L to
40 L the DO reduction decreased. In order to further analyse the DO
variation within the lter bed, grab samples of ltered water were
taken for a day's charge and the results for Day 49 is presented in
Fig. 6. It is clear that DO levels were lower at deeper layers of the
lters. However, DO were always above 1.0 mg/L indicating that
aerobic condition prevailed throughout the lter bed. Analysis of DO
on other days also showed similar pattern (data now shown). It is
known that different biochemical reactions including oxidation of
organic matter occur in biosand lters [24].
A daily charge of 20 L water was chosen in the present study based
on the typical requirements of a developing country household. Also,
it represented approximately one pore volume (19.1 L) of the lter.
Further, most of the reported studies in the literature used 20 L daily
charge [11,13]. The latter part of the Run 2 used a daily charge of 40 L.
The 40-L daily charge was intended to represent the higher end of
typical water volume for domestic uses by a family in developing
countries.
6

Dissolved oxygen (mg/L)

E.coli concentration (CFU/ mL)

104

291

Run 1

20 L

40 L

5
4
3
2
MBSF
BSF

1
0

10

20

30

40

50

60

70

Days of filter operation

Fig. 5. Variation in the efuent dissolved oxygen (DO) content over the length of the
lter run. The mean inuent DO was 6.3 mg/L.

292

M.M. Ahammed, K. Davra / Desalination 276 (2011) 287293

Dissolved oxygen (mg/L)

Day 49

5
4

MBSF
BSF

3
2
1
0

10

15

20

25

Volume of water filtered (L)

Fig. 6. Variation in the efuent dissolved oxygen content with volume of water ltered
for a particular day. Data shown are for Day 49 of Run 1.

While several authors have described enhanced microbial removal

in ltration systems utilising metal oxides and hydroxides, a clear
explanation for this improved microbial deposition often appears to
be lacking. Several factors/mechanisms such as hydration forces,
hydrophobic interactions, macromolecular bridging, surface roughness, and electrical double-layer and van der Waals forces have been
used to explain bacterial adhesion [20,25]. The DLVO (DerjaguinLandau and Verwey-Overbeek) theory of colloid stability which
describes the interaction of electrostatic and van der Waals forces in
occulation of colloidal suspensions is often used to describe bacterial
sorption in aqueous systems [26].
It may be noted that in the present study, coated sand layer was
placed in the middle of the sand layer and not at the top. It was
thought that placing the coated sand at the top would interfere with
the formation of biolayer and mask the efciency of the coated sand
layer in removing microorganisms. It would be interesting to
investigate whether the position of the coated layer had any inuence
on the removal efciency.
Compared to other household water treatment methods, biosand
ltration offers several advantages. Due to high ltration rate,
comparatively large volume of water can be treated in a shorter
time. Since turbidity is also removed during ltration, both aesthetic
and microbiological quality of water improve during the process
unlike other household treatment processes such as solar disinfection,
chlorination, and boiling which are effective only against microorganisms, Higher turbidity values, in fact, reduce the efciency of some
of these processes. Use of coated sand in biolter offers the additional
advantage of removing toxic metals such as arsenic from water
though this was not investigated in the present study.
A few recent studies have shown that nitrication/denitrication
occur in biosand lter which results in higher concentration of nitrate
and nitrite in the efuent, raising concern about the use of BSF with
raw water containing high nitrogen content. For example, Chiew et al.
[15] showed that nitrate concentration occasionally exceeded 30 mg/
L in BSF efuent when treating raw groundwater with ammonium
level exceeding 10 mg/L. Similar observations have been reported by
Murphy et al. [14] and Murphy et al. [24] during eld studies in
Cambodia. In some BSF units ammonication was also observed [24].
Given the health effects of nitrate/nitrate in drinking water, fate of
nitrogen compounds during biosand ltration needs further investigation. Also, inuence of modied BSF on nitrication/nitrication
also needs detailed study. In the present work, since the raw water
had very low ammonia/nitrate content, this was not investigated.
While biosand lter may not be able to guarantee a zero-faecal
coliform water for all raw waters, introduction of IOCS layer improved
the efciency of bacterial removal by at least one-log10 unit. Thus for
water with low/medium bacterial contamination such as pond water,
shallow groundwater and canal water, the modied lter would be able
to produce water of low-risk category according to WHO classication

(E. coli between 0 and 10 CFU/100 mL) [27]. Introduction of IOCS layer is
benecial especially during the ripening or maturation period and after
cleaning operation of the lter when the bacterial removal efciency of
BSF is low, and MBSF would guarantee 2-log10 removal.
Under Indian conditions, depending on the material of construction (concrete or plastic container) and local availability of lter
media, the cost of construction of a full-scale BSF as the one used in
the present study, would vary in the range Indian Rupees (IRs) 300
1200 (US $728). Based on the observed chemical requirements for
coating the media in the laboratory, and the cost of chemicals in bulk
quantities in the Indian market, the modied lter would cost an extra
IRs 350 (US $8). This extra cost is justied given the fact that use of
MBSF results in signicant improvement in microbiological water
quality, thus reducing the health risk to consumers. Further, it is
known that IOCS has a large removal capacity for microorganisms
[18], and thus eliminating the need for frequent replacement of the
spent IOCS. While the coated sand medium cannot be produced by
individual households, this can be manufactured at the village level
and then distributed locally. Also, simpler methods for producing the
coated sand at the households could be developed. Another issue is
how long the coated sand would be effective in removing the
contaminants. Further studies should be carried out to determine how
long the lter can produce acceptable water in a range of situations.
4. Conclusions
Results of the laboratory experiments with full-scale household
lters showed that conventional biosand lter had a low bacterial
removal efciency during the maturation period. Introduction of iron
oxide-coated sand layer could improve the performance of the
biosand lter by at least one-log10 unit throughout the lter run.
While initial E. coli reduction of N2 log10 was observed for modied
biosand lter and reached N3 log10 after one month's of operation,
initial E. coli removal was less than one-log10 unit for BSF. This shows
that introduction of IOCS is benecial especially during maturation
period and after cleaning operation when the bacterial removal tends
to be very low. Tests with natural canal water also conrmed this.
Turbidity removal was N90% for both the lters and the mean efuent
turbidity remained around 1.0 NTU. Doubling the daily charge from
20 L to 40 L adversely affected the bacterial removal in both lters
indicating that operating conditions have a great inuence on the
lter performance and that volume of water that can be efciently
treated is limited. Filtration of water through BSF and MBSF did not
result in any signicant changes in physico-chemical quality of
ltered water, and all remained within the drinking water quality
guideline values.
References
[1] WHO/UNICEF, Meeting the MDG Drinking Water and Sanitation Target: The
Urban and Rural Challenge of the Decade, World Health Organization and United
Nations Children's Fund, Geneva, Switzerland, 2006.
[2] WHO, Combating Waterborne Disease at the Household Level, World Health
Organization, Geneva, Switzerland, 2007.
[3] M. Kubare, J. Haarhoff, Rational design of domestic biosand lters, J. Water Supply
Res. Technol. AQUA 59 (2010) 115.
[4] D.L. Baker, W.F. Duke, Intermittent slow sand lters for household usea eld
study in Haiti, in: R. Gimbel, N.J.D. Graham, R. Collins (Eds.), Recent Progress in
Slow Sand and Alternative Bioltration Processes, IWA Publishing, London, U.K.,
2006, pp. 278282.
[5] T.F. Clasen, Scaling Up Household Water Treatment Among Low-Income
Populations, World Health Organization, Geneva, Switzerland, 2009.
[6] M.D. Sobsey, C.E. Stauber, L.M. Casanova, J.M. Brown, M.A. Elliott, Point of use
household drinking water ltration: a practical, effective solution for providing
sustained access to safe drinking water in the developing world, Environ. Sci.
Technol. 42 (2008) 42614267.
[7] C. Stauber, G. Ortiz, D. Loomis, M. Sobsey, A randomised controlled trial of the
concrete biosand lter and its impact on diarrheal disease in Bonao, Dominican
Republic, Am. J. Trop. Med. Hyg. 80 (2009) 286293.
[8] S.-S.K. Tiwari, W.-P. Schmidt, J. Darby, Z.G. Kariuki, M.W. Jenkins, Intermittent slow
sand ltration for preventing diarrhoea among children in Kenyan households using

M.M. Ahammed, K. Davra / Desalination 276 (2011) 287293

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]
[17]

[18]

unimproved water sources: randomized controlled trial, Trop. Med. Int. Health 14
(2009) 13741382.
W.F. Duke, R.N. Nordin, D. Baker, A. Mazumder, The use and performance of
biosand lters in the Artibonite Valley of Haiti: a eld study of 107 households,
Rural Remote Health 6 (2006) 570.
J. Baumgartner, S. Murcott, M. Ezzati, Reconsidering appropriate technology: the
effects of operating conditions on the bacterial removal performance of two
household drinking-water lter systems, Environ. Res. Lett. 2 (2007) 16.
M.A. Elliott, C.E. Stauber, F. Koksal, F.A. DiGiano, M.D. Sobsey, Reduction of E. coli,
echovirus type 12 and bacteriophages in an intermittently operated householdscale slow sand lter, Water Res. 42 (2008) 26622670.
W.A.M. Hijnen, J.F. Schijven, P. Bonn, A. Visser, G.J. Medema, Elimination of
viruses, bacteria and protozoan oocysts by slow sand ltration, Water Sci.
Technol. 50 (2004) 147154.
C.E. Stauber, M.A. Elliott, F. Koksal, G.M. Ortiz, F.A. DiGiano, M.D. Sobsey,
Characterisation of the biosand lter for E. coli reductions from household
drinking water under controlled laboratory and eld use conditions, Water Sci.
Technol. 54 (2006) 17.
H.M. Murphy, E.A. McBean, K. Farahbaksh, A critical evaluation of two point-ofuse water treatment technologies: can they provide water that meets WHO
drinking water guidelines? J. Water Health 8 (2010) 611630.
H. Chiew, M.L. Sampson, S. Huch, S. Ken, C.B. Benjamin, Effect of groundwater iron
and phosphate on the efcacy of arsenic removal by iron-amended biosand lters,
Environ. Sci. Technol. 43 (2009) 62956300.
M.M. Ahammed, M. Chaudhuri, Sand-based ltration/adsorption media, J. Water
Supply Res. Technol. AQUA 45 (1996) 6771.
J. Chen, S. Truesdail, F. Lu, J. Zhan, C. Belvin, B. Koopman, S. Farrah, D. Shah, Long
term evaluation of aluminum hydroxide sand for removal of bacteria from
wastewater, Water Res. 32 (1998) 21712179.
J. Lukasik, Y. Cheng, F. Lu, M. Tamplin, S.R. Farrah, Removal of microorganisms
from water by columns containing sand coated with ferric and aluminum
hydroxides, Water Res. 33 (1999) 769777.

293

[19] T.M. Scott, R.C. Sabo, J. Lukasik, C. Boice, K. Shaw, L. Barroso-Giachetti, H. El-Shall,
S.R. Farrah, C. Park, B. Moudgil, B. Koopman, Performance and cost-effectiveness of
ferric and aluminum hydrous metal oxide coating on lter media to enhance virus
removal, KONA 20 (2002) 159164.
[20] S.B. Kim, S.J. Park, C.G. Lee, N.C. Choi, D.J. Kim, Bacteria transport through goethitecoated sand: effects of solution pH and coated sand content, Colloids Surf. B
Biointerfaces 63 (2008) 236242.
[21] C. Collin, Biosand Filtration of High Turbidity Water: Modied Filter Design and
Safe Filtrate Storage, Master of Engineering Thesis, Massachusetts Institute of
Technology, Cambridge, USA, 2009.
[22] D. Liu, Z. Teng, J.J. Sansalone, F.K. Cartledge, Surface characteristics of sorptiveltration storm water media. I: Low-density (s b 1) oxide-coated buoyant media,
J. Environ. Eng. 127 (2001) 868878.
[23] American Public Health Association, Standards Methods for the Examination of Water
and Wastewater, 20th ed. American Public Health Association/American Water Works
Association/Water Pollution Control Federation, Washington DC, USA, 1998.
[24] H.M. Murphy, E.A. McBean, K. Farahbaksh, Nitrication, denitrication and
ammonication in point-of-use biosand lters in rural Cambodia, J. Water Health
8 (2010) 803817.
[25] S.E. Truesdail, J. Lukasik, S.R. Farrah, D.O. Shah, R.B. Dickinson, Analysis of bacterial
deposition on metal hydr(oxide)-coated sand lter media, J. Colloid Interface Sci.
203 (1998) 369378.
[26] M.A. Scholl, R.W. Harvey, Laboratory investigations on the role of sediment
surface and ground water chemistry in transport of bacteria through a
contaminated sandy aquifer, Environ. Sci. Technol. 26 (1992) 14101417.
[27] WHO, Guidelines for Drinking Water Quality, 3 rd ed World Health Organiszation,
Geneva, Switzerland, 2006.