Desalination 248 (2009) 308314
Effect of using different particle sizes of sand as filter
media for dewatering faecal sludge
A.R. Kuffoura , E. Awuaha, F.O.K. Anyemedua, M. Straussb,
D. Koneb, O. Cofiec
a
Department of Civil Engineering, Kwame Nkrumah University of Science and Technology (KNUST),
Kumasi, Ghana
b
EAWAG/SANDEC, Swiss Federal Institute for Environmental Science and Technology,
Department of Water and Sanitation in Developing Countries, P.O. Box 611, CH-8600,
Duebendorf, Switzerland
c
International Water Management Institute (IWMI), West Africa Office, PMB CT112 Cantonments, Accra,
Ghana
Received 31 January 2008; revised accepted 15 May 2008
Abstract
This research is aimed at investigating the effect of different particle sizes of sand for the dewatering of faecal sludge with
respect to the dewatering time, contaminant load in the percolate, rate of clogging and quantity of biosolids produced. Three filter
media FM1, FM2 and FM3 were used to dewater the faecal sludge. The experiment was conducted on bench scale consisting of
miniature drying beds. Six cycles of dewatering were run where FM1, FM2, and FM3 showed average dewatering times of 10, 10
and 9 days respectively without significant differences (p > 0.05). However the percolate quality showed significant differences
between them in the removal of TS, TVS, SS, COD, DCOD and NH3-N with FM1 having the highest removal for each parameter.
Accumulation of organic matter in the top 10 cm of the filter bed indicated that FM1 was least likely to clog and it also generated
the highest quantity of biosolids.
Keywords: Filter medium; Faecal sludge; Dewatering time; Contaminant load; Percolate; biosolids
1. Introduction
Human excreta (faecal sludge) is a rich source
of organic matter and plant nutrients such as
nitrogen, phosphorus and potassium. Each persons
Corresponding author.
excreta in theory at least, is nearly sufficient for production of ones own food requirement [1]. Excreta
is not only a fertilizer, its organic matter content, which
serves as a soil conditioner and humus replenisher,
an asset not shared by chemical fertilizers, is of equal
or even greater importance [2]. Chinese peri-urban
vegetable farmers have reported that customers prefer
Presented at the Water and Sanitation in International Development and Disaster Relief (WSIDDR) International
Workshop Edinburgh, Scotland, UK, 2830 May 2008.
0011-9164/09/$ See front matter 2009 Published by Elsevier B.V.
doi:10.1016/j.desal.2008.05.121
A.R. Kuffour et al. / Desalination 248 (2009) 308314
excreta-fertilized rather than chemically fertilized
vegetables. Thus, vegetables grown on excretaconditioned soils yield higher sales prices [3].
On-site sanitation (OSS) facilities are the predominant options for faecal sludge FS disposal in developing countries leading to its accumulation which needs
to be emptied and treated. FS generated from these
facilities is of two main types which are public toilet
sludge (PTS) that is fresh and undigested, generated
from unsewered public toilets and septage that is partially digested, generated from septic tanks. The most
common system used for treating FS in Ghana is
the waste stabilization ponds [4,5]. These pond systems
are often overloaded and not properly operated resulting in very poor performance in the removal of nutrients, pathogens and organic load [6]. The final
effluents of these pond systems are discharged into
water bodies thus causing environmental pollution. In
Kumasi, Ghana, where this study was carried out,
about 500 m3 of faecal sludge generated is daily sent
to pond systems for treatment. Previous studies in
Kumasi, have shown that faecal sludge dewatering
using unplanted filter beds to generate biosolids was
possible [7]. However, the problems of clogging in
filter beds, particle size of sand filter and mixing ratios
of PTS and septage, which lead to long dewatering
time and high contaminant loads in the percolate need
to be addressed. This research aimed to determine the
appropriate particle sizes of sand that will assist to
improve the faecal sludge dewatering efficiency of filter beds to enhance the treatment of large volumes of
faecal sludge generated in Ghana for a better sanitation
delivery and agricultural productivity.
309
Influent FS
Free board
Nylon net
Faecal sludge
Sand
Fine gravel
Course gravel
Graduated container
Stand
Fig. 1. Schematic design of the bench scale set up.
2.2. Filter material and drying bed preparation
Course and fine gravels of about 2.03.0 cm and
0.51.0 cm diameter respectively, served as base support for the filter medium (sand) which was collected
from a local community about 35 km from project site.
The sand was sieved into ranges of three different particle sizes 0.10.5 mm; 0.5 1.0 mm; and 1.0
1.5 mm by sieve analyses, constituting the filter
media FM1, FM2 and FM3, respectively. Effective
sizes (Es) and uniformity coefficients (Uo) were determined. The permeabilities of the different filter media
were determined using constant head permeability test.
The base of the filter bed was filled with the course
gravel to 15 cm depth, followed by the fine gravel to
10 cm depth and the sand to 20 cm depth. Nylon net
was placed on the sand on which the FS was poured
(Fig. 1), to ensure easy removal of the dewatered
biosolid.
2. Materials and methods
2.1. Design of experiment
2.3. Collection of faecal sludge (FS) and dewatering
A wooden structure of size 4.5 m long by 3.5 m
wide and 2 m high was raised and roofed with iron
sheets. A wooden platform of 3 m long, 2 m wide and
0.4 m high was constructed in it which served as a
bench for the drying beds. Each drying bed was raised
in a plastic container having a size of 0.85 m long 0.175
m diameter. There were nine drying beds consisting of
FM1, FM2 and FM3, with each replicated three times.
These were arranged in a randomized complete block
design on the wooden bench under the roofed shelter.
PTS from public toilet and septage from septic tank
were collected in separate plastic drums of about 90 L
each from faecal sludge suction trucks discharging at
Dompoase landfill site in Kumasi, about 20 km away
from the project site. The FS used for the dewatering
consisted of PTS and septage, mixed in a ratio of 1:1
by volume, and a hydraulic loading of 20 cm (equivalent to 5 L) was applied on each filter bed. The volume
of the percolate and depth of sludge were measured
every 24 h while the numbers of days taken for the
310
A.R. Kuffour et al. / Desalination 248 (2009) 308314
Table 1
Average characteristics of raw FS (no. of cycles 6) + standard deviation (SD)
Parameter
Mean values with SD
Parameter
Mean values with SD
TS (g/L)
TVS (g/L)
SS (g/L)
Conductivity (mS/cm)
Temperature ( C)
pH
36.64 + 7.73
25.34 + 9.98
19.68 + 6.59
22.67 + 3.45
28.07 + 1.05
7.77 + 0.13
COD (g/L)
DCOD (g/L)
NH3-N (g/L)
TKN (g/L)
NO3 (g/L)
TP (g/L)
50.32 +
10.98 +
2.83 +
3.58 +
1.38 +
1.85 +
sludge to dewater were also recorded. Dewatering was
considered complete when biosolid accumulated was
spadable for possible removal from the filter bed.
2.4. Laboratory analyses
The raw FS was analyzed just after collection from
the treatment plant before dewatering. Percolate collected daily from each filter bed was analyzed daily for
temperature, pH and conductivity in situ. It was then
kept below 4 C till dewatering was complete. After the
dewatering, composite sample of the percolate from
each drying bed was analyzed. Biosolid and sand from
top 10cm of each filter bed were analyzed for TS and
TVS after each cycle. The parameters analyzed in the
raw FS and percolate were, COD, DCOD, TS, TVS,
TKN, NH3- N, NO3-N, TP, electrical conductivity,
temperature and pH. Methods outlined in standard
methods for the examination of water and wastewaters
[8] were used for the analyses of the parameters.
28.78
5.31
1.68
2.07
0.42
0.53
3.2. Characteristics of raw FS
The raw FS from septage and PTS analyzed had
very high values for TS, SS, COD, NH3-N, TKN
NO3, etc. (Table 1). These concentrations are several
times (10100) the strength of sewage [11]. Such
sludge with high contaminant loads, pose treatment
challenges but has high potential of generating large
quantities of biosolid that may be useful in agriculture.
The associated high concentration of NH3-N is an
additional problem to its treatment. This is because
ammonia concentration greater than 30 mg/L is the tolerance limit of algae in facultative ponds though some
special species are able to adapt themselves to concentrations up to 50 mg/L [12]. Thus ammonia concentration of 2830 mg/L as found in this type of sludge is
excessively high and will hamper algal and bacterial
growth. At such concentrations, malodor and eye irritations are added challenges of the sludge treatment [13].
These characteristics are major contributors to the failure of treatment of FS in pond systems in Ghana and
many other developing countries and resulted in the
need to use alternatives like the unplanted drying beds
for effective sludge management.
3. Results and discussion
3.1. Sand characteristics
Sands for municipal anaerobic digester effluent
dewatering beds should ideally have an effective diameter of between 0.30 and 0.75 mm and a uniformity
coefficient less than 4 [9,10]. The uniformity coefficient (Uo) values obtained for FM1, FM2 and FM3 were
2.422, 1.727 and 2.029, respectively, using sieve analyses. The permeabilities of FM1, FM2 and FM3 were
9.0 102, 14.0 102 and 36.0 102 (cm/s),
respectively.
3.3. Dewatering time (DT) of the different filter
media in different cycles
The averages of DT of the filter media (FM1, FM2
and FM3) were 10, 10 and 9 days respectively. Generally, the DT for the entire filter media decreased from
C1 to C5 until it rose suddenly in C6 (Fig. 2). The initial
long dewatering times of all the filter media especially
C1, might be due to comparatively low temperatures
(20.322.5 C minimum and 27.431.6 C maximum)
A.R. Kuffour et al. / Desalination 248 (2009) 308314
311
ANOVA of the results at 95% confidence level showed
that there was no significant difference in the DT with
respect to the three different filter media.
3.4. Dewatering efficiencies of the different filter
media in removing contaminant loads
3.4.1. Removal of solids (TS, TVS and SS)
Fig. 2. The dewatering time (days) of different filter media
in six cycles.
and high humidity (88.591 high) during the project
period [14]. Moreover, for municipal sludge dewatering, fine organic particles are created and sand-bed
drainability deteriorates during the early stages of
digestion leading to increase of DT but as digestion
proceeds, the fine organic particles are destroyed and
dewaterability of the sludge improves resulting in
decrease of DT [15]. The state of the sludge also affect
the dewatering time. The sharp rise in the DT in C6
might due to factors like temperature, humidity, state
of the FS. Onset of clogging in the filter beds could also
be possible.
FM3 dewatered in the shortest time as compared to
the others. Not withstanding the above observation, the
Though the average TS, TVS and SS values in the
percolate were still high, the filter media FM1, FM2
and FM3 were able to dewater the FS of between TS
loading rate of 217360 kg TS/m2/year in 10, 10 and
9 days respectively (Table 2). This compares well
with and even shows an improvement over the results
of experiment conducted by Larmie in Accra, Ghana,
where FS of TS loading rate of 130 kg TS/m2/yr was
dewatered in 9 days [16]. The average percentage
removals of TVS for FM1 FM2 and FM3 within the
range of 73.678.2% (Table 2), compares well with the
TVS removal achieved by sand-bed dewatering
research earlier [17]. The generally high average
removal efficiencies of SS between 88.9% and
95.1% in all the filter media is an impressive performance for dewatering FS. This is also comparable to
works by Larmie et al. in Accra [16] and Esseku in
Kumasi [7]. Ability of FM1 (95.1%), to outperform
Table 2
Average characteristics of percolate (no. of cycles 6) + standard deviation; nd not determined
Parameter
units
Percolate from filter media
FM1
Dewatering time
TS
TVS
SS
Temperature
pH
EC
COD
DCOD
NH3-N
TKN
NO3
TP
days
g/L
g/L
g/L
C
(mS/cm)
g/L
g/L
g/L
g/L
g/L
g/L
10 +
8.62 +
4.41 +
0.86 +
27.54 +
8.36 +
13.56 +
5.48 +
2.3 +
0.64 +
0.8 +
0.63 +
0.759 +
6
2.46
1.58
0.34
1.28
0.25
5.38
2.20
0.97
0.55
0.61
0.31
0.23
Removal of contaminant
load in percolate (%)
FM2
FM3
10 + 5
9.51 +3.02
4.83 + 1.62
1.19 + 0.49
27.42 + 1.28
8.3 + 0.12
15.85 + 4.32
5.79 + 2.37
2.8 + 0.88
0.89 + 0.68
1.0 + 0.76
0.64 + 0.20
0.872 + 0.25
9+
10.84 +
5.86 +
1.96 +
27.47 +
8.21 +
16.18 +
6.7 +
3.12 +
0.84 +
1.4 +
0.83 +
0.948 +
4
3.18
1.62
0.92
1.24
0.20
3.69
2.01
1.07
0.62
1.05
0.24
0.25
FM1
FM2
FM3
76.2
78.2
95.1
nd
nd
40.8
87.9
77.7
78.0
78.1
56.0
58.9
74.0
77.2
93.4
nd
nd
30.4
87.4
70.8
71.1
71.7
52.5
52.9
70.5
73.6
88.9
nd
nd
28.6
85.3
65.8
71.6
60.8
39.6
48.7
312
A.R. Kuffour et al. / Desalination 248 (2009) 308314
FM2 (93.4%) and FM3 (88.9%) in the removal of SS
(Table 2), may be due to its ability to remove finer
sludge particles than the other media. As particle size
of sand bed increases the dry weight recovered by
dewatering reduces [17].
3.4.2. Removal of salts or reduction of electrical
conductivity
The removal efficiencies of 40.8%, 30.4% and
28.6%, respectively, achieved by FM1, FM2 and
FM3 were comparatively low, with respect to percentage removals achieved for the solids (Table 2). Conductivity is the result of ionic mobility of dissolved
salts. Thus the higher the dissolved salts, the higher the
electrical conductivity [18]. Poor reduction of the electrical conductivity is an indication of poor removal of
salts. High magnitude of conductivity between 13.56
and 16.8 mS/cm in the percolates of FM1, FM2 and
FM3 (Table 2) will put osmotic stress on the plants
when discharged into the environment. However, the
poor removal of the salts means little was retained in
the biosolids which indicates safe use of the biosolid
in agriculture.
3.4.3. Removal of degradable organics (COD,
DCOD)
There were high percentage removals of COD with
FM1 showing the highest removal efficiency in almost
all the cycles (Table 2). This efficiency achieved by the
drying beds might be due to the ability of the filter
media to remove the solids from the sludge and the
availability of the dissolved organics for microbial
action [19]. The ability of FM1 to remove more solids
and the fact that its particles present the largest surface
area for microbial action might have contributed to its
highest performance. A good reduction in the DCOD is
an indication of the availability of the dissolved substrate for bacterial action [19]. The removal efficiencies of 77.7%, 70.8% and 65.8% by FM1, FM2 and
FM3, respectively, well explain that some fine particles of dissolved organics were degraded by microbes.
The outstanding performance of FM1 over the FM2
and FM3 may further support the role played by the
effective size of the filter media [20].
3.4.4. Removal of nutrients (NH3-N, TKN, NO3-N,
TP)
All the filter media achieved high NH3-N removal
efficiencies yet the concentrations in the percolates
were very high (Table 2). This might be due to the high
concentrations in the influent FS and ammonification
of organic N [21]. NH3-N removal is believed to be
basically due to volatilization and nitrification by nitrifying bacteria [5]. The filter media achieved high efficiency in the TKN removal, between 60.8% and
78.1%. Of the three filter media, FM1 showed the highest potential for TKN removal with average removal
efficiency of 78.1%. This might be due to organic-N
removal through high TVS removal (Table 2) and
mineralization of organic N. The comparatively low
percentage removal of nitrate (NO3-N) might be due
to the fact that it was continually being produced in the
percolate through mineralization of organic N [21].
Irrespective of the percentage removals of nitrate and
total phosphorus (TP) achieved (Table 2), the concentrations in the effluents were still very high.
3.4.5. Organic matter (TVS) accumulation in the
sand (filter medium)
Since TVS is a function of organic matter [18], its
increase in the filter media is an indication of organic
matter build up. FM3 has the highest potential of
organic matter accumulation (Fig. 3). The organic matter accumulation is likely to fill the pore spaces
between the particles of the filter bed which can subsequently result in clogging. Though clogging did not
take place during the running of the six cycles, there
was the possibility for its occurrence since there was
Fig. 3. Percentage TVS of filter media from C1 to C6.
A.R. Kuffour et al. / Desalination 248 (2009) 308314
313
Table 3
Biosolid accumulation performance of the different filter media
Parameter
Total dry solid of FS (g)
Total org. matter of FS (g)
Ave. fresh biosolids/cy. (g)
Ave. dry biosolid TS (g)
TVS of biosolid (g)
% TVS of TS of raw FS
% TVS of TS (biosolids)
% TS accumulated
FM1
FM2
FM3
Parameter
FM1
FM2
FM3
183
135
410
155
103
74.0
66
84.9
183
135
414
148
90
74.0
61
80.5
183
135
400
147
84
74.0
57
80.3
% TVS accumulated
Days in a cycle
Cycles/year
Ave dry bio/yr (g/yr)
TVS/filter bed/yr (g/yr)
kg/yr (TVS)
Area (m2)of filter bed
kg TVS/m2 yr (kg/yr)
76.0
10
36
5580
3708
3.71
0.024
154.5
66.8
10
36
5328
3240
3.24
0.024
135
62.5
9
41
6027
3444
3.44
0.024
143.5
Same for all filter media; ave., average; cy., cycle; yr, year; bio, biosolid.
significant increase in the organic matter content in the
filter media especially in the FM3.
3.5. Biosolids accumulated by filter media
The percentage accumulation of the biosolids by
the different filter media FM1, FM2 and FM3 (Table 3)
is an indication that FM1 had the highest potential to
generate biosolids. Since the same FS was used for all
the filter media, it was assumed that the TVS content of
the bioslids was directly proportional to the biosolids
accumulated on the respective beds. FM1 gave the
highest percentage TVS of TS of biosolid accumulated
by the different filter media (Table 3). Furthermore, the
percentage TVS accumulated by each filter medium
from the influent FS were in the order of 76.0, 66.8 and
62.5 for FM1, FM2 and FM3, respectively (Table 3).
These performances exhibited by FM1 proved that it
had the highest potential of biosolid generation
amongst the others. Annual biosolid production with
respect to dewatering time, was estimated (Table 3)
and based on that the organic matter accumulation per
m2 per year was determined where FM1 showed the
highest potential to generate biosolids, 154.5 kg
TVS/m2 year (Table 3).
4. Conclusion
All the filter media with respect to particle size configuration used for the research were found to be able to
improve the dewatering performance of the unplanted
filter bed. However, FM1 was most efficient of all in
terms of contaminant load removal while in terms of
dewatering time they were statistically not different.
The filter media with bigger particle sizes (FM3 and
FM2) were relatively faster in accumulating organic
matter and thus likely to clog faster than the FM1.
FM1 had the added advantage of being able to generate
more biosolids than the FM2 and FM3. Generally there
were significant differences statistically, in the contaminant load removal between the different filter
media. The different filter media were all not efficient
in salts removal since they showed low EC reduction
efficiencies. The pH and temperature were within
favorable limits for microbial growth but the high
ammonia levels in the percolate could be toxic to the
microbes responsible for biological degradation. The
contaminant load in the percolates from all the different filter media were all beyond permissible levels for
discharge into the environment.
Although high removal efficiencies were achieved
by all the filter media, the parameters in the effluent
needed to be treated further either through column filters, constructed wetlands or stabilization ponds for
more polishing before final discharge. Having determined the particle sizes best suited for the sludge dewatering, further research into solid loading rate with
respect to the mixing ratios of PTS and septage and
other possible technologies (e.g. addition of physical
conditioners like saw dust to the FS) to compliment the
findings achieved for this studies will be effective
approaches to improving the dewatering process of
unplanted drying beds.
314
A.R. Kuffour et al. / Desalination 248 (2009) 308314
Acknowledgements
This study was made possible by financial support
from EAWAG/SANDEC. The study contributes to the
SANDEC programme of low-cost options for managing FS in Developing countries to improve sanitation
and agriculture. The authors would like to express our
sincere apreciation to IWMI, for offering us space at
the Kumasi office for our office work.
References
[1] J.O. Drangert, Fighting the urine blindness to provide
more sanitation options, Water South Africa, 24(2)
(1998) 157164.
[2] U. Winblad, Towards an Ecological Approach to Sanitation, Publications on Water Resources, 1997, No. 5,
Department for Natural Resources and the Environment, Swedish International Development Authority
(SIDA), Birger Jarlsgatan 61, S-10525 Stockholm,
Sweden.
[3] Co-composting of Faecal Sludge and Solid Waste for
Urban and Peri-urban Agriculture in Kumasi, Ghana,
Final Report, February, 2003, p. 9.
[4] E. Awuah, R. Amankwah-Kuffour and H.J. Gijzen,
Characterisation and management of domestic wastewater at two locations in Kumasi, in: Proceedings of the
1st School of Engineering Research Retreat (SERR 1),
KNUST, Ghana, 2001.
[5] E. Awuah, Pathogen Removal in Macrophyte and Agal
Waste Stabilisation Ponds, PhD Dessertation, submitted the Academic Board of Wageningen University
and the Academic Board of the UNESCO-IHE Institute
for Water Education on 29th November 2006 in Delft,
The Netherlands, 2006, p. 106.
[6] D. Kone and M. Strauss, Low-cost options for treating
faecal sludges (FS) in developing countries
challenges and performance, in: Proceedings of 9th
International Conference on Wetland Systems held in
Avignon, France, September 26October 02, IWA &
ASTEE, vol. 1, 2004, pp. 213219.
[7] H. Esseku, Use of Drying Beds for Faecal Sludge/
Septage Pre-Treatment, Student Thesis, WSESP,
unpublished report, 2002.
[8] APHA-AWWA-WEF Standard Methods for the Examination of Water and Wastewater, 20th ed., Washington, DC, 1998.
[9] Metcalf and Eddy Inc., Wastewater Engineering
Treatment, Disposal and Reuse, 3rd ed.,
McGraw-Hill, New York, 1995.
[10] J. Elliasson, Media Specifications, Washingtons State
Department of Health, Rule Development Committee
Issue Report Draft, 2002, pp. 110.
[11] M. Strauss, S.A. Larmie and U. Heinss, Treatment of
sludges from on-site sanitation: low-cost options, Water
Sci. Technol., 35(6) (1997).
[12] D.D. Mara and H. Pearson, Artificial freshwater environment, Waste stabilisation ponds, in: H.J. Rehm, G.
Reeds, eds., Biotechnology, VCH Verlagsgesellscaft,
Weinheim, Germany, 1986.
[13] S.M. McGinn, H.H. Janzen and T. Coates, Atmospheric
ammonia, volatile fatty acids, and other odorants near
beef feedlots, J. Environ. Quality, 32 (2003)11731182).
[14] Meteorological Department Report, Ashanti Region,
Ghana, 2005, unpublished data.
[15] W. Rudolfs and H. Heukelekian, Relationship between
drainability of sludge and degree of digestion, Sewage
Works J., 6 (1934) 10731081.
[16] S.A. Larmie, Field Research on Faecal Sludge Treatment in Accra, Ghana Sedimentation Tank Sludge
Accumulation Study, Water Research Institute, Accra,
Ghana, December 1994, unpublished field report.
[17] D.J. Hills and R.L. Kemmerle, Dewatering considerations for effluent from dairy manure anaerobic
digesters, Agric. Wastes, 3 (1981) 297310.
[18] G. Kierly, Environmental Engineering, 1998,
pp. 6473.
[19] K. Karim, R. Hoffmann, T. Klasson and M.H. AlDahhan, Anaerobic digestion of animal waste: Waste
strength versus impact of mixing, Bioresour. Technol.,
96 (2005) 17711781.
[20] J.A. Salvato, Environmental Engineering and Sanitation, 3rd ed., 1982, p. 433.
[21] E. Epstein, Land Application of Sewage Sludge and
Biosolids, 2003, pp. 2935.
