An Evaluation ofHousehold Drinking WaterTreatment Systems in Peru:
The Table Filter andthe Safe Water System
Brittany CoulbertSpring 2005
Based on Coulbert’s MIT M.Eng. thesis of the same title (Feb. 2005)
Peru
Project Site: La Joya, Arequipa
The Peru Team
The Water Problem
• 1.1 billion people lack access to safe drinking water
• 1.7 million people die each year from diarrheal diseases related to unsafe water, sanitation, and hygiene
Fact source: WHO, 2004Picture source: Susan Murcott, MIT
The Problem in Peru
• 50% of Peru’s 28.4 million people live below the poverty line
• GDP per capita: $5,000
• In 2002, Peru had only 66% “improved drinking water coverage” in rural areas
• “Improved” does not mean safe. It may simply indicate a household connection or protected well. (With this definition, this report evaluates almost entirely households with “improved” access to drinking water, even before treatment.)
Source: WHO Joint Monitoring Programme, CIA World Factbook
The Solution
• UN Millennium Development Goal:• Halve, by 2015 [as compared to 1990], the proportion
of people without sustainable access to safe drinking water and basic sanitation.
• WHO International Network to Promote Household Water Treatment and Safe Storage
• This report is part of the effort by MIT and members of the Network to seek ways to increase access to safe water for people worldwide through efforts including household water treatment systems (HWTS) and safe water storage.
The Technologies
The Table Filter
The Safe Water System
Outline• Safe Water System
– Tests & results from Peru• Table Filter
– Tests & results from Peru and the MIT lab• Household interviews• Cost comparison & financing options
Study of HWTS Solutions as Currently Implemented in Peru
• Field Study in southern Peru, mostly in La Joya, Arequipa
• January 2004• Tested Safe Water Systems (household chlorination)
& Table Filters in homes• Interviewed users
• MIT Lab Study • February - July 2004• Tested 2 Table Filters with different grades of sand• Also tested Table Filters without sand
TacnaArequipa
Lima
Regions of Study:
Arequipa & Tacna
(focused in Cerrito Buena Vista, near La
Joya, Arequipa)
Safe Water SystemSafe Water System(Household Chlorination)
Safe Water System (SWS)Safe Water System (SWS)
• 20-L safe storage containers (“bidones”)• 0.5% sodium hypochlorite solution generated at a
local hospital & distributed to towns in 200-mL bottles
• Users add “half cap” of solution to 20 L of water stored in safe storage containers and wait 30 minutes before drinking
Chlorine GeneratorElectricity
+Water
+Salt=
Chlorinesolution
Family in Peru using their Safe Water System
Safe Water System Tests
• E.coli & Total Coliform (TC) tests• These bacteria indicate the presence of fecal
contamination, which can cause severe diarrheal sicknesses
• Zero E.coli or TC “colony forming units” (CFU) should be present in any drinking water sample
• Chlorine residual tests• Chlorinated water sources should have a residual
amount of chlorine to ensure that the chlorine dose was enough to deactivate all harmful bacteria
E.coli Concentrations Before & After SWS Chlorination in Peru
0 02
0 0
53
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
B-1
1
E-3
H-1
2
C -
7
LL -
18
LL -
27
M -
8
I - 1
LL -
22
House ID
E.co
li C
FU/1
00m
lLo
g Sc
ale
E.coli Raw E.coli SWS Treated
99.6% reduction of E.coli by SWS
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06B
-11
H-1
2
LL -
18
M -
8
LL -
22 E -2
L - 1
6
M -
17
M -
17B
LL -
4
L - 1
0
M -
19B
LL -
9
CB
V P
ost
B -
2
A' -
7
A' -
3
House ID
TC C
FU/1
00m
lLo
g Sc
ale
TC Raw TC SWS Treated
TC Raw Avg.TC SWS Treated Avg.
95% reduction of TC by SWS
Total Coliform Concentrations Before & After SWS Chlorination in Peru
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Val
divi
a
CB
V S
choo
l
A-8
E-3
LL 1
8
LL -
2
LL -
22
E -2
M -
1
M -
17
H -
10
L - 1
1
M -
19
LL -
23
N -
3
CB
V K
inde
rgar
den
A' -
2
A' -
5
A' -
5B
House ID
Free
Chl
orin
e [m
g/L]
Average = 0.18
Free Chlorine Residual Measured in Bidones in Peru
(The dark black lines indicate the target concentration range.)
The average chlorine residual (0.18 mg/L) was
less than the recommended range.
Pros & Cons of the SWS
Pros• 99.6% E.coli removal• 95% TC removal• Very inexpensive:
$6 container + $3/year• Easy to use• Local chlorine
generation possible
Cons• Less effective with turbid
source water (particle removal pre-treatment is needed)
• Average chlorine residual was found to be too low
• Chlorine solution sometimes difficult to obtain due to poor technical support
• People don’t like the taste
Peruvian “Table Filter”Peruvian “Table Filter”
• Indigenous filter developed by CEPIS and the Belgian development organization, DGCI
• Filtering media: geotextile cloth, sand, & ceramic candle filters
• Made of two 20-L (5-gal.) plastic buckets
Table Filter Media
Geotextile
Sand Bed
Ceramic Candles
Family in Peru using their Table Filter
Table Filter Tests
• Coliform tests– Thermotolerant Coliform (TTC)
• Indicates fecal contamination
– E.coli & TC– Heterotrophic Plate Counts
(HPC)• Indicates general bacteria levels
• Turbidity tests• Flow rate tests
Lab in La Joya, Peru
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
H-3
F-7
H-3
(2)
CBV
Pos
t
H -
3
E - 9
H -
2
O -
3
H -
1
B - 7
G -
8
H -
13
B - 5
N -
1
House ID (all in CBV)
Col
iform
[CFU
/100
ml]
LOG
Sca
le
E.coli Raw E.coli TF Treated TC Raw TC TF Treated
TC Raw Avg.TC TF Treated Avg.
98% reduction of TC by TFs99% reduction of E.coli by TFs
E.coli & TC Concentrations Before & After Treatment by Table Filters in Peru
Summary Table of Coliform Tests onTable Filters and SWS in Peru
Table Filters SWS Chlorination
ColiformCFU/100ml
RawWater
TreatedWater
%Removal
RawWater
TreatedWater
%Removal
E.coli 5.3x102 2 99% 4.5x103 1 99.6%TC 3.5x103 7.2x101 98% 2.1x104 1.5x102 95%
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
TF - Raw TF - Treated(Filtered)
SWS - Raw SWS - Treated(Chlorinated)
CFU/
100m
lLO
G S
cale
E.Coli TC
Comparison of Average Coliform Concentrations Before & After Table Filter & SWS (in Peru)
Table FilterSWS
Chlorination
before
after
before
after
0
10
20
30
40
50
H - 3 F - 7
Villa H
ermos
a E - 3
CBV Post*
H - 3 E - 9H - 2
*O - 3 H - 1 B - 7 G - 8H - 1
3B - 5
*N - 1
House ID
Turb
idity
(NTU
)
Filter Input Filter Output
*Filter reported as "not working"
Input Avg.Output Avg.
Turbidity Concentrations Before & After Treatment by Table Filter in Peru
67% reduction of turbidity by TFs (70% by “working” TFs)
Table Filter Flow Rate in Peru
• Average: 3.1 L/hr– Arequipa average: 2.3 L/hr– Tacna average: 3.8 L/hr
• Faster than other ceramic filters (0.5-2.0 L/hr)
• Slower than BioSand Filter (15-30 L/hr)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
CBV
(Ram
irez)
La C
ano
Post
Cru
z de
May
o (B
elis
ario
)
Villa
Her
mos
a C
-16
CBV
H -
3
CBV
E -
9
CBV
O -
3
CBV
H -
1
CBV
B -
7
CBV
G -
8
CBV
H -
13
CBV
N -
1
Chu
cata
man
i (C
have
z)
Chu
cata
man
i "E"
Chu
cata
man
i "F"
Pist
ala
(Tom
a)
House ID
Flow
Rat
e [L
/hr]
Arequipa Avg: 1.7 L/hr
Tacna Avg: 3.8 L/hr
Table Filter Flow Rates in Peru
MIT Lab Work -- Table FilterFebruary - July 2004
• Simulate conditions in Peru• All materials, except sand, brought from Peru• Charles River water & sewage added to filters to
increase water contamination levels
• Two filters• “Medium Sand Table Filter” (MSTF): Sand size
used for BioSand filter, larger grains than in Peru• “Fine Sand Table Filter” (FSTF): Sand size
specified in Peru
• Coliform tests• Turbidity Tests
Table Filters in MIT Lab
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
Feb 2
0Fe
b 23
Feb 2
7Marc
h1Marc
h5Marc
h 8Marc
h 12
March 1
5Marc
h 19
June
1Ju
ne 15
June
16Ju
ne 18
June
21Ju
ne 22
June
23Ju
ne 24
Date
TTC
[CFU
/100
ml]
LOG
Sca
le
Source MSTF FSTFMSTF Avg.FSTF Avg.
TTC Concentrations Before & After Treatment by the Medium and Fine Sand Table Filters
98% average reduction of TTC by MSTF & FSTF
Table Filters Tested Without SandThe two Table Filters showed slightly different results on average. Since sand grain size was the only known difference between the two, the sand was removed from the filters before additional testing to see if the grain size was the true reason for the differences in % removal (as opposed to unexpected differences in the ceramic candles).
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
June 30 July 1 July 2Date
TTC
[CFU
/100
ml]
LOG
Sca
le
Source MSTF FSTFMSTF Avg.FSTF Avg.
TTC Concentrations Before & After Filtrationwithout Sand
99% reduction of TTC by FSTF w/o Sand
97% reduction of TTC by MSTF w/o Sand
E.coli & TC Concentration Before & After Treatment by MSTFs & FSTFs
(one day of data: June 24)
03
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
E.coli TC
E.co
li &
TC [C
FU/1
00m
l]LO
G S
cale
Source Medium Sand Filter Fine Sand Filter
99.95% reduction of
E.coli by FSTF
99.99% reduction of E.coli by MSTF
99% reduction of
TC by FSTF
99.95% reduction of TC by MSTF
E.coli Concentration Before & After Treatment by MSTF & FSTF Without Sand
0 0 0
3
0
5
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
June 28 June 30 July 1 July 2
E.co
li [C
FU/1
00m
l]LO
G S
cale
Source MSTF FSTF
99.8% reduction of E.coli by
FSTF w/o Sand
99.99% reduction of E.coli by MSTF
w/o Sand
TC Concentration Before & After Treatment by MSTF & FSTF Without Sand
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
June 28 June 30 July 1 July 2
TC [C
FU/1
00m
l]LO
G S
cale
Source Medium Sand Filter Fine Sand Filter
99.8% reduction of TC by FSTF w/o Sand
23% reduction of TC by MSTF w/o Sand
Why Only 23% Reduction of TC by MSTF w/o Sand?
The low average % removal of TC by the MSTF was caused by one day of testing which showed that the TC concentration in the treated water was higher than that of the untreated water (June 30). Valid data could not be collected from the water treated by the FSTF, so the two filters could not be compared on that day. Therefore, the average values and % removals of the two TFs without sand are not truly comparable to each other for the TC tests.
247
0
66
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
June 1 June 15 June 23
HPC
[CFU
/100
ml]
LOG
Sca
le
Source Medium Fine Blank
HPC (Heterotrophic Plate Counts) Concentration Before & After Treatment by Table Filters
83% reduction of HPC by FSTF
99% reduction of HPC by MSTF
An additional 99.8% reduction from MSTF is
needed to reach the average “blank”
water concentration of HPC
ColiformTableFilterMedia
SourceWater
MSTFTreatedWater
FSTFTreatedWater
MSTF %Removal
FSTF %Removal
Sand 1.2x104 4.1x101 2.2x102 98% 98%TTC No Sand 1.9x103 7.7x101 3.3x101 97% 99%Sand 1.8x103 < 1 1.0x101 99.99% 99.5%E.coli No Sand 8.9x102 < 1 3 99.99% 99.8%Sand 6.5x103 3 6.4x101 99.95% 99%TC No Sand 5.0x104 2.0x103 1.2x102 23% 99.8%
HPC Sand 3.1x106 5.3x104 1.3x106 99% 83%
Summary Table of Coliform Removal Rates of both Table Filters, With and Without Sand,
in the MIT Lab
Sand Grain Size Theory
The coliform tests performed on Table Filters without sand did not support the theory that the sand grain size affected the performance of the filters.
Our Peruvian teammates sifting sand for a Table Filter
Turbidity• Indicates amount of suspended particles in
water, or its “cloudiness”• Average % removal of turbidity
by Table Filters: – Peru: 70%– MIT:
• Medium Sand TF: 91%
• Fine Sand TF: 92%
Pros & Cons of the Table Filter
Pros• Average % removal of TC
in Peru = 98%• Average % removal of
turbidity at MIT = 92%• Provides relatively
consistent and significantly improved drinking water
• Inexpensive: $11.40 each• Easy to use
Cons• Broken spigots• Cleaning is bothersome • Fragile ceramic candles• Parts not easily available• Sand sifting (during filter
assembly) is time-consuming
• More expensive than other treatment methods
InterviewsInterviews
• 89 households surveyed– 35 had Table Filters– 49 had Safe Water Systems– 5 had neither
• 66 were from Arequipa (57 from CBV)• 23 were from Tacna
Interview Statistics
• Households averaged 5.2 people, including 1.3 children under the age of 5
• Average total household* spending was S/ 93 (Peruvian nuevo soles) per month (or $0.58 per person per day)
* Of the 84 households with government-sponsored water treatment systems
Where do you get your drinking water?
0102030405060708090
Piped System Water Truck Bring fromArequipa
Canal
# of
Res
pond
ents
Table Filter Bidon (Chlorine) No System
Water Vending Trucks
Irrigation Canals
Piped System to Household Taps
& Storage Devices
Where do you get your drinking water?
How often do you clean your Table Filter?
02468
10121416
Weekly Every 2-6weeks
Every 3-6months
# of
Res
pond
ents
How often do you clean your Table Filter?
When did you last add disinfectant solution to your SWS bidon?
0
5
10
15
20
Today Yesterday 2 daysago
3 daysago
1 year ago
# of
Res
pond
ents
How often do you add disinfectant solution to your SWS bidon?
0
5
10
15
20
Twicedaily
Daily Every 2days
Every 3days (orevery 2-3
days)
Every 15days
# of
Res
pond
ents
How often do you add disinfectant solution to your SWS bidon?
When did you last add disinfectant solution to your SWS bidon?
Do you feel better now that you use a treatment system?
01020304050607080
Yes Same No
# of
Hou
seho
lds
Table Filter Bidon (Chlorine)
Does the water taste better or worse now that you use a treatment system?
0
20
40
60
80
Better: Same: Worse:
# of
Hou
seho
lds
Table Filter Bidon (Chlorine) Do you feel better now that you use a treatment system?
Does the water taste better or worse now that you use a treatment system?
Do you have any problems or complaints about the treatment system?
010203040506070
Yes No
# of
Hou
seho
lds
Table Filter Bidon (Chlorine)
Is the treatment system easy to use?
0
20
40
60
80
100
Yes No
# of
Hou
seho
lds
Table Filter Bidon (Chlorine)
Do you like it?
0
20
40
60
80
100
Yes No
# of
Hou
seho
lds
Table Filter Bidon (Chlorine)
Do you have any problems/complaints about the treatment system?
Is the treatment system easy to use?
Do you like it?
Cost Comparison of Each Treatment Option
Water TreatmentOption
Capital Cost O&M Costs/ year
Total 10-Year Cost
Table Filter $6.40 $5 $56Safe Water System $9.80 $3 $40Table Filter + SWS $16.20 $8 $96Water Treatment Plant(w/ piped system, perfamily of 5)
$475 $36 $835
The capital cost of a Table Filter is low because the initial purchase of ceramic candles is included in the O&M costs.
Willingness to Pay for a Table Filter by Current Filter Owners (Average Offer: S/25)
0
24
6
810
12
0 5 10 15 20 25 30 35 40 45 50 55 60
Price Willing to Pay [Soles ]
# of
Res
pond
ents
Actual capital cost of Table Filter:S/ 40 ($11.4)
Average offer:S/ 25 ($7.1)
Willingness-to-Pay• Household surveys indicate that the target
population may be willing to pay about halfthe capital cost of each system
• How does the program then cover costs?
Recommendation for Payment and Financing Options
• Government (Ministry of Health) and/or outside aid organizations contribute some % of cost or pay for large initial costs (like a $1,400 chlorine generator)
• Cheaper treatment options are explored or implemented (e.g., chlorine or SODIS)
• HWTS recipients pay in monthly installments
Proposed Monthly Payment Plan for Each Treatment Option
(assuming each user must cover 100% of capital and O&M costs)
Monthlypayment over
12 months
Monthlypayment over
24 months
Approx. monthlyO&M after capital
is paid offTable Filter S/ 3.5
($1)S/ 2.5($0.7)
S/ 1.5($0.4)
Safe WaterSystem
S/ 4($1.1)
S/ 2.5($0.7)
S/ 0.9($0.25)
Table Filter +SWS
S/ 7($2)
S/ 4.5($1.3)
S/ 2.5($0.7)
Recommendation:Filtration plus Chlorination
is best!
• Filtration by a Table Filter helps remove the turbidity in water thatwould make chlorination less effective
• Chlorination and safe water storage by the Safe Water System kills any bacteria that remain after filtration and protects against recontamination
