Adapting Consumer Report’s Product Evaluation Methods for

Particle Removal, Gravity Non-Electric and Reverse Osmosis Water

Filters in the Indian Marketplace

by

Shuyue Liu

B.S. Environmental Engineering

University of Science and Technology Beijing, 2013

SUBMITTED TO THE DEPARTMENT OF CIVIL AND ENVIRONMENTAL

ENGINEERING IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE

DEGREE OF

MASTERS OF SCIENCE IN CIVIL AND ENVIRONMENTAL ENGINEERING

AT THE

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

February 2015

©2015 Massachusetts Institute of Technology. All rights reserved.

Signature of Author: ____________________________________________________________

Department of Civil and Environmental Engineering

December 20, 2014

Certified by: __________________________________________________________________

Susan Murcott

Research Scientist, Department of Urban Studies and Planning

Thesis Supervisor

Accepted by: __________________________________________________________________

Heidi M. Nepf

Chairman, Departmental Committee for Graduate Students

Adapting Consumer Reports' Product Evaluation Methods to

Particle Removal to Gravity Non-Electric and Reverse Osmosis

Water Filters in the Indian Marketplace by

Shuyue Liu

Submitted to the Department of Civil and Environmental Engineering on December 20, 2014 in

Partial Fulfillment of the Requirements for the Degree of Masters of Science in Civil and

Environmental Engineering

ABSTRACT

Household Water Treatment and Storage (HWTS) products provides households that are

drinking unimproved water supplies with a first line of defense against contaminants in their

drinking water and those drinking improved water supplies with an additional barrier against

potential risks. With the global water crisis becoming more and more severe, evaluation of

HWTS technologies and products has become crucial to ensure they are used to remove

impurities effectively. The goal of this thesis was to evaluate household water filters in the

Indian marketplace as part of a larger research and technology evaluation to investigate the

utilization and performances of different water filter models in both lab and field settings. This

was achieved by comparative testing and research work done at Consumer Reports (CR)

Headquarters in Yonkers, NY.

This evaluation included the testing of three categories of filters: Conventional Particle Removal

(cloth and mesh), Gravity Non-Electric (GNE) and Reverse Osmosis (RO) water filters. In total,

16 models were tested. The challenge water for all filter testing had these characteristics: 40±10

NTU turbidity and 1500±150mg/L total dissolved solids (TDS). When testing E.coli removal,

deionized water was used as the base water and the concentration of E.coli was 105 to 106 MPN

(Most Probable Number)/100mL. The comparative testing attributes that were evaluated include:

E.coli removal, turbidity removal, TDS removal, clean water flow rate, RO % recovery, and

filter lifetime with the end-of-life defined as when flow rate <1L/hr.

As a result of this product evaluation, the author determined that: 1) Cloth and mesh filters had

limited effectiveness in reducing contaminants; 2) GNE filters had much better performance than

cloth and mesh filters, but none of them had outstanding performance; 3) RO filters were shown

to be quite effective in reducing turbidity (greater than 99.5%), TDS (greater than 97%), and

E.coli (greater than 99.9999%). But, they produce a large amount of wastewater (around 3/4 of

the feed water) which is a huge waste of precious water and a sustainability concern especially in

a water scarce region.

Thesis Supervisor: Susan Murcott

Title: Research Scientist of Department of Urban Studies and Planning

ACKNOWLEDGEMENTS

I would like to express my deepest appreciation to all those who provided me the possibility to complete

this thesis.

First of all, I want to express my sincere heartfelt thanks to my supervisor Susan Murcott who got me

opportunity to work on the water evaluation project. She has given me her constant help, walked me through

all the stages of writing my thesis and offered me invaluable advice and informative suggestions. I also

enjoy my personal talks with her, she has given me recommendations on not only my research and thesis

but also my job hunting. No matter what difficulty I encountered, she has been the one to lend a helping

hand.

Furthermore, I would also like to acknowledge with much appreciation the crucial role of Dr. Jeffrey Asher,

who is the former Technical Director and Vice President of Consumer Reports (CR). He gave me the

permission to use all required equipment and the necessary materials in CR to complete the evaluation

research. Through the two-month period in CR, he had always been together with us helping with setting

up and doing the tests. He also invested his full effort in guiding me in accomplishing this thesis. It is not

too much to claim that this thesis would be impossible without him.

And I would like to express my deep expression to Eric Adams, the advisor of 2013 CEE M.Eng. program.

He gave me a lot of help during my study. I learned a lot in his class “Water Quality Control”. A special

note of thanks to Kiley Clapper, the Academic Administrator of CEE department, who helped me a lot with

my degree switching. If it was not for her, I would not have completed my studies.

A special thanks goes to my team mate, Yiyue Zhang, who has been working with me since we came to

MIT. We collaborated very well in the two projects we did together. We are colleagues at work and close

friends in life. I cannot imagine what life would have been if it was without her. Thank Jhanel Chew for

being with me in CR and investigating the effect of turbidity and TDS on concentration levels of E.coli, in

order to find a substitute for PBS solution, which would have cost us a huge amount of spending. Thank

Tengke Wong for helping me with the literature review and experiment protocol development. He gave me

a lot of helps in my most difficult times, which I will never forget.

I want to thank the many wonderful and supportive staff at MIT, especially Derek Brine, who helped me

bring the lab supplies to CR and gave me all of his supports when I was doing the research. I also want to

thank Jack Whipple, who helped me order the lab supplies. I want to thank Christine Pilcavage, who

reserved a room in a hotel for me to stay during the two months in CR; and Joanne Mathias, who arranged

the weekly meetings for MIT CITE (Comprehensive Initiative Technology Evaluation). If it were not for

them, I couldn’t have made this so far.

I want to give special thankfulness to my friend Yi Ding who granted me the authorization to use the

microscope in MIT Lincoln lab.

Last my thanks would go to my beloved family for their loving considerations and great confidence in me

all through these years. I also owe my sincere gratitude to my friends and my fellow classmates who gave

me their help and time in listening to me and helping me work out my problems during the difficult course

of the thesis.

1

Table of Contents

1 Introduction ........................................................................................................................... 14

India Background ........................................................................................................... 14

Household Water Treatment and Safe Storage Technologies (HWTS) .......................... 15

Comprehensive Initiative on Technology Evaluation (CITE) Program ......................... 16

Consumer Reports (CR) ................................................................................................. 16

Consumer Reports Suitability Team............................................................................... 17

Objectives ....................................................................................................................... 18

2 Literature Review.................................................................................................................. 20

Microbiological Performance ......................................................................................... 22

Turbidity and Total Dissolved Solids Removal Performance ........................................ 24

Clean Water Flow Rate ................................................................................................... 25

% Recovery .................................................................................................................... 25

Other Attributes .............................................................................................................. 26

3 Water Filter Categories and Models...................................................................................... 27

Filtration Processes and Categories................................................................................ 27

Conventional Particle Removal Filter ............................................................................ 30

Cloth Filter .............................................................................................................. 30

2

Mesh Filter .............................................................................................................. 33

Gravity Non-Electric (GNE) Filter ................................................................................ 34

Expresso Stainless Steel Water Container .............................................................. 35

Tata Swach Cristella Plus ........................................................................................ 36

Tata Swach Smart ................................................................................................... 37

Hindustan Unilever PureIt Classic 14L .................................................................. 38

Prestige LifeStraw ................................................................................................... 40

Eureka Forbes Aquasure Kitanu Magnet ................................................................ 41

KENT Gold Plus - 20L ........................................................................................... 42

Everpure Unbreakable ............................................................................................ 43

Reverse Osmosis Filter................................................................................................... 45

Locally-assembled Dolphins ................................................................................... 47

Tata Swach Platina Silver ....................................................................................... 48

4 Method .................................................................................................................................. 50

Comparative Product Testing Attributes ........................................................................ 50

E.coli Removal ........................................................................................................ 50

Turbidity Removal .................................................................................................. 52

Total Dissolved Solids (TDS) Removal .................................................................. 52

Clean Water (“Product”) Flow Rate ........................................................................ 54

3

RO % Recovery ...................................................................................................... 55

Clogging/Filter Lifetime ......................................................................................... 55

Test Water Source ........................................................................................................... 56

Base Water .............................................................................................................. 56

Challenge Test Water .............................................................................................. 57

E.coli Solution......................................................................................................... 58

Test Method for Testing Attributes ................................................................................. 58

Turbidity Test Method ............................................................................................. 58

TDS Test Method .................................................................................................... 58

E.coli Test Method .................................................................................................. 59

Sterilization Procedures and Quality Control ......................................................... 59

Test Set-Up ..................................................................................................................... 60

Plumbing Set-Up ..................................................................................................... 60

Reverse-Osmosis (RO) Flow Test Rig: Developing CR-Version 2 ........................ 61

The E.coli Injection System .................................................................................... 64

Test Methods for Each Filter Category .......................................................................... 67

Cloth and Mesh Filters Test Method ....................................................................... 67

Gravity Non-Electric (GNE) Filters Test Method ................................................... 68

Reverse Osmosis (RO) Filters Test Method ........................................................... 70

4

5 Test Results ........................................................................................................................... 73

Cloth and Mesh Water Filters Test Results .................................................................... 73

Gravity Non-Electric (GNE) Water Filters .................................................................... 74

Espresso Water Filter Test Results .......................................................................... 74

Tata Swach’s Christella Plus Test Results .............................................................. 75

Tata Swach’s Smart 1500 Liters Test Results ......................................................... 76

Tata Swach Smart 3000 Liters Test Results ............................................................ 76

Hindustan Unilever’s Pureit Classic 14L Test Results ........................................... 77

Prestige’s LifeStraw Test Results ........................................................................... 78

Eureka Forbes’s AquaSure Amrit with Kitanu Magnet Test Results ...................... 79

KENT’s Gold UF Membrane Filter Test Results.................................................... 80

Summary of the Performance of GNE Filters......................................................... 80

Reverse-Osmosis (RO) Water Filters ............................................................................. 81

RO Turbidity and TDS Reduction .......................................................................... 81

RO E.coli Removal ................................................................................................. 81

RO Flow Test Results ............................................................................................. 83

6 Ratings Chart and Discussion ............................................................................................... 86

Overview of Comparative Ratings Chart ....................................................................... 88

Scores, Weightings and Ratings .............................................................................. 90

5

Attributes and Features Shown in the Ratings Charts ............................................ 90

Gravity Non-Electric (GNE) Filter Comparative Ratings Chart .................................... 92

Reverse-Osmosis (RO) Comparative Ratings Chart ...................................................... 92

Discussion of the Ratings Charts ................................................................................... 96

7 Conclusions and Recommendations ..................................................................................... 97

Conclusions .................................................................................................................... 97

Cloth and Mesh Filters ............................................................................................ 97

Gravity Non-Electric Filters ................................................................................... 97

Reverse Osmosis Filters .......................................................................................... 97

Recommendations for Future Research ......................................................................... 98

APPENDIX A ISO 12103-1/CD, A2 Fine Test Dust Specification ........................................ 100

APPENDIX B Parts List ............................................................................................................. 101

APPENDIX C RO Flow Testing Sample Data ........................................................................... 102

APPENDIX D Scoring, Weigthing and Rating Methods for Three Categories of Filters .......... 108

References ....................................................................................................................................114

6

Table of Figures

Figure 1-1 Location of India ......................................................................................................... 14

Figure 1-2 Location of Ahmedabad ............................................................................................ 14

Figure 1-3 Number of people (in millions) without access to an improved drinking water source in

2012, by MDG region (WHO & UNICEF, 2014)......................................................................... 15

Figure 1-4 Consumer Reports headquarters in Yonkers, NY........................................................ 17

Figure 1-5 Water filter lab in the Consumer Reports headquarters .............................................. 17

Figure 2-1 A part of Consumer Reports Ratings Chart for Water Filters (Consumer Reports, 2010)

....................................................................................................................................................... 22

Figure 3-1 Filtration Application Guide (modified from Filtration Application Guide, Water

Quality Improvement Center) ....................................................................................................... 29

Figure 3-2 Cloth (right) and Jali Mesh (left) filter ....................................................................... 30

Figure 3-3 Cloth filters found in Ahmedabad households ............................................................ 31

Figure 3-4 The micrograph of cloth #1 ......................................................................................... 32

Figure 3-5 The micrograph of cloth #2 ......................................................................................... 32

Figure 3-6 The micrograph of cloth #3 ......................................................................................... 32

Figure 3-7 A mesh filter in use in an Indian household ................................................................ 33

Figure 3-8 The micrograph of mesh filters (Left: one layer; Right: two layers) .......................... 34

Figure 3-9 Expresso Stainless Steel Water Container ................................................................... 35

Figure 3-10 Ceramic Candle Element ........................................................................................... 35

7

Figure 3-11 Tata Swach Cristella Plus (left) and Swach bulb (right) ............................................ 36

Figure 3-12 Tata Swach Smart (left) and its Swach Bulb (right) .................................................. 38

Figure 3-13 Hindustan Unilever PureIt Classic 14L..................................................................... 39

Figure 3-14 Prestige LifeStraw ..................................................................................................... 40

Figure 3-15 Aquasure Kitanu Magnet (left) and its Kitanu Magnet “Positive Charge Technology”

Cartridge (right) ............................................................................................................................ 41

Figure 3-16 KENT Gold Plus (left) and its ultrafiltration membrane (right) ............................... 42

Figure 3-17 Everpure’s “Unbreakable” (left) and its mineral filter (right) .................................. 43

Figure 3-18 Reverse-Osmosis System Schematic ........................................................................ 45

Figure 3-19 Diagram of a Reverse-Osmosis Membrane .............................................................. 46

Figure 3-20 Dolphin design stages and their function .................................................................. 47

Figure 3-21 Front and Inside Look of a Dolphin .......................................................................... 48

Figure 3-22 Tata Swach Platina Silver .......................................................................................... 49

Figure 3-23 Tata Swach Platina Silver under testing .................................................................... 49

Figure 4-1 Boxplot of TDS in Ahmedabad source waters (Murcott S., 2014) ............................. 54

Figure 4-2 Method to count total coliform and E.coli .................................................................. 59

Figure 4-3 CR Plumbing Set-Up................................................................................................... 61

Figure 4-4 Single RO Flow Test Rig Schematic ........................................................................... 61

Figure 4-5 CR-V2.0 ...................................................................................................................... 62

8

Figure 4-6 CR-V2.0 with Control Panel ....................................................................................... 62

Figure 4-7 LabView Solenoid Controlling Interface .................................................................... 62

Figure 4-8 E.coli Injection Rig Schematic .................................................................................... 65

Figure 4-9 E.coli Injection Rig (Left: Mixing Vessel, Gage and Valves, and 6-gallon (23-liter)

Pressure Tank. Right: PVC pipe inside of the Mixing Vessel) ..................................................... 66

Figure 4-10 The Experimental Set-Up for Testing Cloth and Mesh ............................................. 67

Figure 4-11 Experiment Set-Up for Cloth E.coli LRV Test .......................................................... 68

Figure 4-12 Depiction of E.coli Concentration for Inlet, Waste & Clean Water vs. Time ........... 70

Figure 5-1 Cloth #1 E.coli LRV and Turbidity Removal vs. the Number of Cloth Layers .......... 73

Figure 5-2 E.coli LRV and Turbidity Removal of Eight-Layers for Cloth#1, #2, and #3 ............ 74

Figure 5-3 Espresso Water Filter Test Results .............................................................................. 75

Figure 5-4 Tata Swach’s Christella Plus Test Results ................................................................... 75

Figure 5-5 Tata Swach’s Smart 1500 Liters Test Results ............................................................. 76

Figure 5-6 Tata Swach’s Smart 3000 Liters Test Results ............................................................. 77

Figure 5-7 Hindustan Unilever’s Pureit Classic 14L Test Results of Sample 1 ........................... 78

Figure 5-8 Hindustan Unilever’s Pureit Classic 14L Test Results of Sample 2 ........................... 78

Figure 5-9 Prestige’s Life Straw Test Results ............................................................................... 79

Figure 5-10 Eureka Forbes’s AquaSure Amrit with Kitanu Magnet Test Results ........................ 79

Figure 5-11 KENT’s Gold UF Membrane Filter Test Results ...................................................... 80

9

Figure 5-12 E.coli Removal Results for Clean Water (Non-branded Dolphin RO-type) Filter ... 82

Figure 5-13 Clean Water (Non-branded Dolphin RO-type) Filter Flow Performance ................. 84

Figure 5-14 Tata Swach Platina RO Filter Flow Performance ..................................................... 84

10

List of Tables

Table 1-1 India water supply and sanitation situation from 1990 to 2012 (WHO & UNICEF, 2014)

....................................................................................................................................................... 15

Table 2-1 Guideline values for verification of microbial quality (WHO, 2011b) ........................ 23

Table 2-2 Derivation of targets using QMRA calculations (WHO, 2011a) .................................. 24

Table 2-3 Microbiological Organisms and Reduction Requirements (WHO, 2014b) .................. 24

Table 2-4 General Test Water Characteristics (WHO, 2014b) ...................................................... 25

Table 2-5 Challenge Test Water Characteristics (WHO, 2014b) .................................................. 25

Table 3-1 Filter Categories, Pore Sizes, Molecular Weight Cutoff, Filtration Pressure and Particles

Removed (Baker, 2012) ................................................................................................................ 28

Table 3-2 All models tested at CR ................................................................................................ 29

Table 3-3 GNE Filter Models Tested ............................................................................................ 34

Table 3-4 Special Manufacturer Claims for the GNE models tested ............................................ 44

Table 3-5 RO Filter Models Tested ............................................................................................... 47

Table 4-1 Microbiological Organisms and Reduction Requirements (WHO, 2014a) .................. 51

Table 4-2 Risk Level from E.coli (WHO, 1997) .......................................................................... 51

Table 4-3 CR team’s standard for evaluating bacteria removal performance of the filters .......... 52

Table 4-4 Criteria for evaluating turbidity removal of the GNE filters ........................................ 52

Table 4-5 Palatability of Drinking Water (WHO, 2006) ............................................................... 53

Table 4-6 Hardness Description (Ratnayaka, Johnson, & Brandt, 2000) ..................................... 54

11

Table 4-7 Criteria for evaluating clean water flow rate of the GNE filters .................................. 55

Table 4-8 Criteria for evaluating lifetime of the GNE filters ........................................................ 56

Table 4-9 Challenge Test Water Characteristics (WHO, 2014b) .................................................. 57

Table 4-10 CR Team Challenge Test Water Formulation ............................................................. 58

Table 4-11 Sampling Timeline and Dilutions for RO E.coli Test ................................................. 72

Table 5-1 RO Turbidity and TDS Removal Results ..................................................................... 81

Table 5-2 E.coli Removal Results for Clean Water (Non-branded Dolphin RO-type) Filter ....... 82

Table 5-3 E.coli Removal Results for Tata Swach Platina Silver (Branded RO-type) Filter ....... 83

Table 6-1 The Summary of the Performance of each Filter Category .......................................... 87

Table 6-2 Comparison between Filter Categories for the Cost, Lifetime, and Environmental

Factors ........................................................................................................................................... 88

Table 6-3 Example of a Comparative Ratings Chart .................................................................... 89

Table 6-4 Weightings for Each Attributes of GNE Filters ............................................................ 90

Table 6-5 Comparative Ratings of Gravity Non-Electric (Non-Electric/Gravity) Water Filters .. 94

Table 6-6 Reverse Osmosis (Electric) Comparative Rating ......................................................... 95

12

List of Equations

Equation 1 ..................................................................................................................................... 26

Equation 2 ..................................................................................................................................... 50

13

List of Acronyms

ATCC American Type Culture Collection

BoP Bottom of the Pyramid

CR Consumer Reports

CU Consumer Union

GDWQ Guidelines for Drinking Water Quality

GNE Gravity Non-electric

HWF Household water filters

HWTS Household water treatment and safe storage

LRV Log Removal Value

MCC Maximum Contamination Concentration

MCL Maximum Contamination Level

MPN Most Probable Number

NGO Non-governmental organization

NSF National Sanitation Foundation

NTU Nephelometric Turbidity Unit

OEM Original equipment manufacturer

PBS Phosphate Buffered Saline solution

QMRA Quantitative microbial risk assessment

RO Reverse Osmosis

S1 Suitability

S2 Scalability

S3 Sustainability

TDS Total Dissolved Solids

USAID U.S. Agency for International Development

USEPA U.S. Environmental Protection Agency

WHO World Health Organization

14

1 Introduction

India Background

India, located in South Asia, has the second highest population of all nations in the world with over 1.2 billion

people (Central Intelligence Agency, 2011). Ahmedabad is the fifth-largest city and seventh-largest metropolitan

area of India with a population of more than 5.8 million. Located on the banks of the Sabarmati River, it is the

largest city and former capital of the western Indian state of Gujarat1.

Figure 1-1 Location of India Figure 1-2 Location of Ahmedabad

India has made outstanding progresses in the recent years in providing access to drinking water to its citizens.

Between 1990 and 2012, 2.3 billion people around the world gained access to an improved drinking water source2,

and within southern Asia, India increased access for 534 million people, which contributed greatly to both its

subsequent regional and global increases in coverage (WHO & UNICEF, 2014). Table 1-1 shows that until 2012,

92.6% have improved drinking water access. This is an enormous achievement given the enormous size of India’s

population. However, from Figure 1-3 we can see that there were still 92 million people (7.54% of its total

1 http://www.census2011.co.in/census/city/314-ahmedabad.html

2 An improved drinking water source is defined by WHO/UNICEF as “one that, by nature of its construction or through

active intervention, is protected from outside contamination, in particular from contamination with fecal matter”.

15

population) without access to improved drinking water supply in 2012 (second only to China).

Table 1-1 India water supply and sanitation situation from 1990 to 2012 (WHO & UNICEF, 2014)

Figure 1-3 Number of people (in millions) without access to an improved drinking water source in 2012, by MDG region

(WHO & UNICEF, 2014)

The problem of providing safe drinking water motivated the author of this thesis to evaluate the household water

filters in Ahmedabad, India via the MIT CITE program (see a fuller description of MIT CITE below in Section

1.3).

Household Water Treatment and Safe Storage Technologies (HWTS)

According to World Health Organization (WHO), because of unsafe water, sanitation, and hygiene, there are 2

million diarrheal deaths every year throughout the world, and the vast majority are of children under 5 years3.

Household water treatment and safe storage (HWTS) interventions are able to improve drinking water quality and

reduce diarrheal disease effectively for people who rely on water from polluted rivers, lakes and, in some cases,

3 http://www.who.int/household_water/en/

16

unsafe wells or even contaminated piped water supplies.

Various HWTS technologies exist, including cloth filters, candle filters, ceramic pot filters, chlorine disinfection,

UV disinfection, micro/ultro/nano filtration, reverse osmosis membrane filtration, etc. They can serve as the first

treatment process for those who use unimproved water supplies, or as an additional barrier for those who already

have an improved water supply. Household water filters (HWF) are a subset of HWTS products. In Ahmedabad,

the most common HWF are cloth, mesh, gravity non-electric (GNE) filters, and reverse osmosis (RO) filters.

Different brands and unbranded locally assembled filters for each category can be found in the India marketplace.

Comprehensive Initiative on Technology Evaluation (CITE) Program

CITE is a program funded by U.S. Agency for International Development (USAID) and led by MIT’s Department

of Urban Studies and Planning. It aims to develop a methodology to evaluate different technologies applicable for

low income consumers in low and middle income countries. The intention is to make recommendations to funders,

decision makers, entrepreneurs and users as to which products have the best characteristics. Bish Sanyal, the Ford

International Professor of Urban Development and Planning, stated that CITE gathers universities, government

centers and entrepreneurs worldwide together to innovate in international development, given that there is an

abundance of technologies but very little assessment to them.

CITE’s framework includes three main evaluation components: Suitability, Scalability, and Sustainability.

Suitability focuses on the field and laboratory technical evaluation from the angles of both consumer expectations

and use patterns. Scalability seeks to reach consumers and influence the society in a wide scale; they are concerned

with issues such as manufacturing, sourcing, supply-chain, distribution and after- market support. Sustainability

focuses on technical, social, behavioral, economic, institutional, regulatory and environmental factors.

The first CITE product evaluation focused on a number of solar lighting technologies found in the Uganda

marketplace. In the period of 2014 to 2015, CITE is conducting a water filter evaluation in the city of Ahmedabad,

India. These water filter technologies are assessed from three aspects mentioned above: Suitability, Scalability,

and Sustainability. The Suitability (S1) team is committed to an evaluation of water filters performance, both in

the field in India and in a laboratory in Consumer Reports headquarters in Yonkers, NY. The goal of the Scalability

(S2) team is to assess the end-to-end supply chain of products, including original equipment manufacturers

(OEMs) as well as the distributors and retailers. The Sustainability (S3) team focuses on the factors that will affect

the long-term use of the products, such as the health, user behavior, economic, social, environmental and

compliance.

Consumer Reports (CR)

Consumer Reports (CR) is a monthly, US magazine published by Consumer Union (CU) in existence since 1936.

Through product research in its in-house laboratory and survey research center, Consumer Reports publishes

17

reviews and comparisons of consumer products and services. Its ratings provide consumers with general buying

guides for about 3000 products per year, from shower heads to smoothies, vacuums, home appliances, televisions

and cars. In 2010, after 74 years of operation, CU has a total of over 7.3 million subscribers for its magazines and

web publications (Bounds, Gwendolyn, 2010).

Consumer Reports headquarters is located in Yonkers, NY (see Figure 1-4). It has a fully-operational water filter

lab (see Figure 1-5) which regulatory tests HWF products found in the American marketplace, from simple,

portable carafes (what are called Gravity Non-electric filters by us) to permanently mounted systems.

Figure 1-4 Consumer Reports headquarters in Yonkers, NY

Figure 1-5 Water filter lab in the Consumer Reports headquarters

Gf

Consumer Reports Suitability Team

The CR Suitability team (S1-CR) was part of the Suitability (S1) group. Its objective was to systematically

18

investigate the performance of different water filter models.

The water filter comparative testing and research work was done at CR headquarters. There, CR loaned CITE

Suitability the use of CR’s Water Filter Lab. Three MIT graduate and post-graduate Research Assistants (Yiyue

Zhang, Shuyue Liu, and Jhanel Chew) plus a CITE consultant (Dr. Jeffrey Asher) worked from July 1 to August

29, 2014 in the evaluation of water filters that are found in the Indian marketplace.

These product evaluations included the life testing of three categories of filters: Particle Removal filters, Gravity

Non-electric (GNE) filters and Reverse Osmosis (RO) water filters. In total, 16 models were tested.

Objectives

For this work, there were three major objectives:

(1) Test Methods Development.

The first objective was to develop water filter test methods that have gone through at least one round of successful

testing at CR for a number of models of water filters found in India. There are performance attributes including

“turbidity removal” and “bacteria (E.coli) removal” that are critical to a filter evaluation in low and middle income

countries, but with good reason were not covered by CR when evaluating filters for the US market. These include

As a result, there was a significant modification to CR’s previous test methods. For instance, CR’s RO test

methods did not measure the flow rate of the wastewater (also known as “brine” or “concentrate” in the water

industry).

(2) Test Rig Development.

The current CR test rig required major modifications to make it acceptable for CITE’s use. A major change was

having the “challenge” (or contaminated) water continually go through two or more RO filters

simultaneously. This automation increased the speed for life testing considerably.

To do this, a rig was designed and fabricated to use only one flow meter multiplexed between the filters currently

being tested. Since a flow meter costs in excess of $500 each, this design also represented a significant cost

savings.

(3) Set Level of Test Attributes.

Objectives #1 (Test Methods Development) and Objective #2 (Test Rig Development) feed directly into Objective

#3 which involves being able to consistently and precisely set levels of contaminants in the water (or “challenge

water”) flowing into the filters.

This necessitated consistently creating high levels of turbidity, total dissolved solids (or TDS) and pathogens (or

E.coli) in the challenge water. A deionized water system was purchased and made operational at CR to provide a

consistent and benign (no chlorine) base water “carrier” for these contaminants.

19

(4) Water Filter Performance Testing.

During the two-month testing period at CR, tests were conducted to determine the performance of different

categories of water filters. A total of 16 models were tested often with multiple test samples of each model.

(5) Analyze Results and Write This Up.

As the last objective, the author scored and rated the filters based on the results using a Consumer Reports-style

ratings chart. (Details of the Consumer Reports-style ratings chart is given in Appendix C.) A Ratings chart is

provided in this report similar to the one delivered to CITE’s sponsor, USAID (U.S. Agency for International

Development).

20

2 Literature Review

Household Water Treatment and Storage (HWTS) products provides households that are drinking unimproved

water supplies with a first line of defense against contaminants in their drinking water. And, for those drinking

improved water supplies that are suspected of being unsafe, HWTS products are additional barrier of protection.

With the water crisis becoming more and more severe, governments, non-governmental organizations (NGOs)

and the private sector are considering household water treatment technologies (HWT) as one of the effective

solutions to drinking water quality challenges, meanwhile, manufacturers are seeking to design and distribute

HWT technologies and products throughout the world. Thus, evaluation of these HWT technologies and products

has become crucial to ensure they are used in a proper way to remove impurities and pathogens effectively. A

performance evaluation is necessary to protect users from potential risk and help decision makers to identify which

technologies to choose.

For such a purpose, some international organizations such as World Health Organization4 (WHO), PATH and the

National Sanitation Foundation International5 (NSF) have defined HTWS technology evaluation standards.

For example, in WHO’s “Evaluating Household Water Treatment Options: Health-based Targets and

Microbiological Performance Specifications” (WHO, 2011), a basis to evaluate the microbiological performance

of HWTS product options was provided. A series of microbiological performance targets has been established

ranging from “interim” to “highly protective”.

Based on WHO’s “Evaluating Household Water Treatment Options” document, a WHO International Scheme to

Evaluate Household Water Treatment Technologies” (referred to hereafter as “Scheme”) has been established

(WHO, 2014a). This scheme is developed to assess whether HWT products meet the WHO performance

recommendations. Based on the device’s ability to remove viruses, bacteria and protozoa, HWT products

evaluated by WHO designated testing laboratories are classified into three tiers: “highly protective”, “protective”

and “limited protection”. It not only helps WHO Member States and procuring UN Agencies in the selection of

4 The World Health Organization, established on 7 April 1948, headquartered in Geneva, Switzerland, is a specialized agency

of the United Nations (UN) that is concerned with international public health.

5 NSF International, based in Ann Arbor, Michigan, is a global independent public health and environmental non-for profit

organization that provides standards development, product certification, testing, auditing, education and risk management

services for public health and the environment.

21

HWT, but also supports national governments to establish laboratory institutions for conducting continuous

evaluations of HWT. The “Procedure” specifies the steps for assessing a HWT product.

There is another document “WHO International Scheme to Evaluate Household Water Treatment Technologies-

Harmonized Testing Protocol: Technology Non-Specific” (referred to hereafter as “Protocol”) (WHO, 2014b).

This Protocol provides the detailed testing method of HWT products for technology non-specific evaluations,

including replicate samples, test water (General Test Water and Challenge Test Water) formulation,

Microbiological Organisms and Challenge Concentrations, and other testing details.

NSF’s standard—“NSF/ANSI 42: Drinking Water Treatment Units – Aesthetic Effects”— establishes minimum

requirements for materials, design and construction, and performance of drinking water treatment systems that are

designed to reduce specific aesthetic-related (non-health effects) contaminants in public or private water supplies,

whereas “NSF/ANSI 53: Drinking Water Treatment Units – Health Effects” aims to reduce health effects-related

contaminants. They evaluates materials in contact with drinking water, structural performance and contamination

reduction. Material evaluation ensures that the materials do not introduce levels of extractable contaminants that

exceed the Maximum contaminant Concentration (MCC) values specified by USEPA. The purpose of testing

structural integrity performance is to evaluate the materials, design, and fabrication quality of the complete water

treatment system. Contaminant reduction testing evaluates the product according to the methodology and criteria

of the standard that is being claimed by the company/organization. It includes chlorine reduction testing, hydrogen

sulfide and phenol reduction testing, pH adjustment testing, zinc reduction testing, mechanical reduction testing

and scale control testing. For each of the attributes, NSF specifies the general test water and testing method.

Consumer Reports, as mentioned in Section 1.4, is a US magazine publishing reviews and comparisons of

consumer products and services based on product research in its in-house laboratory and survey research center.

As an independent, nonprofit organization, Consumer Reports serves consumers through unbiased product testing

and ratings, research, and journalism (Aspan, M., 2006). For example, Figure 2-1 shows a part of Consumer

Reports ratings chart for water filters (Consumer Reports, 2010). It compared different water filter models found

in the US market and provided consumers with information about each filter’s purchasing price, operating cost

and features. Also, based on the test results, this ratings charts shows each filter’s performance on each attribute

(lead removal, chloroform removal, flow rate and clogging) using blob keys (as shown in the top right corner of

this figure).

22

Figure 2-1 A part of Consumer Reports Ratings Chart for Water Filters (Consumer Reports, 2010)

The choice of attributes to measure depends on the purpose for which the filter is being utilized. The primary goal

of the household water filters discussed in this thesis is to remove microbial contamination, TDS and Turbidity.

Thus, microbiological performance, TDS and Turbidity removal, as well as flow rate are chosen as preliminary

performance parameters.

Microbiological Performance

Multiple barriers, from catchment to consumer, are used to prevent the contamination of drinking-water or to

reduce contamination to levels not injurious to health. In general, the microbial risks are associated with ingestion

of water that is contaminated with human or animal faeces (WHO, 2011b). WHO Guidelines for Drinking Water

Quality (GDWQ) (WHO, 2011b) defines that E.coli concentration in drinking water must not be detectable in any

100-mL water sample (see Table 2-1). USEPA also regulates that the Maximum Contamination Level (MCL) of

Total Coliforms (including E.coli) is 0 mg/L. Most of the literature on water filter evaluation compares filter

performance against the GDWQ, which cannot differentiate filters well, and furthermore, the results would be

different if the testing water’s quality is inconsistent. Thus, the author finds some other guidelines for evaluating

the microbiological performance of water filters.

23

Table 2-1 Guideline values for verification of microbial quality (WHO, 2011b)

In WHO “Evaluating Household Water Treatment Options: Health-based Targets and Microbiological

Performance Specifications” (WHO, 2011a), a series of microbiological performance targets has been established

ranging from “interim” to “highly protective”. These targets are derived from Quantitative Microbial Risk

Assessment (QMRA). The targets are linked to reference pathogens in three categories of pathogens: bacteria,

viruses and protozoa. The targets take into account the local water quality data if it is available; or use QMRA

calculations to derive default performance targets (see Table 2-2). In this thesis, the author chose E.coli as the

reference pathogen because there was no available local water quality data in terms of viruses and protozoa. But

given that protozoa commonly range in length between 10 to 52 µm, which are much larger than E.coli whose

length are generally from 3 to 5µm, protozoa could be easily removed if the filter has an excellent performance in

removing E.coli. Viruses also occur widely in drinking-water supplies in low- and high-income countries and are

associated with enteric disease (Levin, 2009). Thus, it is recommended that viruses being tested once the local

quality data in India is available.

24

Table 2-2 Derivation of targets using QMRA calculations (WHO, 2011a)

According to the “Protocol”, an input E.coli concentration of around 105 colonies per ml is prescribed. As per their

microbiological organisms and reduction requirements (Table 2-3), when the E.coli LRV is higher than 4 the filter

is rated as “highly protective,” while when the E.coli LRV is between 2 and 4 then the filter is defined “protective

or limited protective”.

Table 2-3 Microbiological Organisms and Reduction Requirements (WHO, 2014b)

Turbidity and Total Dissolved Solids Removal Performance

As stated before, the “Protocol” recommends two kinds of test water: General Test Water and Challenge Test

Water (see their formulation in Table 2-5 and Table 2-6). The general test water represents non-stressed phase of

testing, while Challenge Test Water is water intended to test the limit of the product under stress circumstances. It

usually has characteristic worse than the product is designed for.

25

Table 2-4 General Test Water Characteristics (WHO, 2014b)

Table 2-5 Challenge Test Water Characteristics (WHO, 2014b)

Clean Water Flow Rate

Clean water flow rate, defined as how fast a volume of water is filtered through the cartridge, is an important

performance attribute. But WHO and NSF have no guidelines on the flow rate. Human need to drink from 1.5L

to 2.5L per person per day to stay healthy (Pimentel, D., 2004). But according to WHO (Howard G., 2003), a

minimum of 7.5L per capita per day will meet the requirement for most population in most conditions. Their

conclusion is based on requirements of a lactating women in general physical condition. The two literatures

provide scientists with good reference to define what good clean water flow rate is. Consumer Reports tests flow

rate as one of their chosen attributes in its US water filter evaluation (Consumer Reports, 2010). They divide water

filters into five categories: 1) Carafe, which are called Gravity Non-electric (GNE) by us; 2) Faucet Mounted; 3)

Countertop; 4) Under-sink; 5) Reverse Osmosis. Their scoring scheme for flow rate differs by filter type.

% Recovery

% Recovery refers to the fact that only part of the feed water flowing into an RO system comes out as product

26

water, and the remaining water washing away the contaminants is wastewater. The calculation for % Recovery is

below (Bruce, I.D., 2014):

% 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑦 =𝑃𝑒𝑟𝑚𝑒𝑎𝑡𝑒 𝐹𝑙𝑜𝑤 𝑅𝑎𝑡𝑒

𝐹𝑒𝑒𝑑 𝐹𝑙𝑜𝑤 𝑅𝑎𝑡𝑒× 100 Equation 1

Properly designed RO systems avoid the use of large amount of water to produce only little clean water. Most

household RO systems are designed with a % Recovery of 20 to 30. This means that a RO system with 100 gallon

feed water and a 20% Recovery would yield 80 gallon of wastewater. WHO also has no guidelines for this attribute

and CR does not test % Recovery of RO systems as routine.

Other Attributes

The decision to include additional attributes for testing depends on the purpose of the filter and the local water

quality. Some regions have water resources that do not meet water quality standards and the threat to drinking

water is growing. CR spiked water with lead and chloroform as two additional attributes to test filters (Consumer

Reports, 2010).

27

3 Water Filter Categories and Models

Drinking water filtration is a process that removes particles from water. It is mainly for small amounts of

suspended, colloidal, or molecular particles. As part of the MIT CITE water filter evaluation, there were 16 models

in three water filter categories tested at Consumer Reports Laboratory. The three water filter categories are particle

removal filter, gravity non-electric (GNE) filter, and reverse osmosis (RO) filter. They were simplified from five

filtration processes: the particle removal filter uses conventional particle filtration, the GNE filter uses

microfiltration, ultrafiltration, and nanofiltration, while the RO filter mainly uses reverse osmosis filtration but

the systems also include some other filtration stages. Specifically, the RO filter system also comes with

micro/ultra/nano filtration cartridges as well as UV light as pre or post treatment.

Filtration Processes and Categories

Conventional particle filtration is a water treatment process widely applied in most water treatment facilities

from large-scale urban systems to household-scale water filters. A conventional particle filter can be of cloth,

mesh, sand, gravel, ceramic of plastic material and it removes particles in the size range from 1 to 1000 microns,

such as sand, clay and dirt particles. Particles in this range may be seen by the naked eye as well as those smaller

particles detectable only under a microscope.

Microfiltration separates solids from water via the mechanisms of size exclusion and/or particle capture.

Suspended particles and microorganisms are captured on the surface or inside the filter, with dissolved substances

and water passing through the filter. Microfiltration is capable of removing particles in the size range of 0.1 to 1

micron. It may be used to remove some bacteria and large pathogens, such as Giardia and Cryptosporidium.

However, microfiltration cannot disinfect water with high concentrations of bacteria and viruses; chemical

disinfection should be used in addition to filtration. Ceramic filters are an example of this category of filters. They

rely on small pores of ceramic materials to remove dirt and bacteria.

Ultrafiltration is a pressure-driven filtration process for fractionating and concentrating solutions containing

suspended colloids and solutes of high molecular weight. The mechanism of ultrafiltration is also size exclusion

or particle capture. Ultrafiltration takes out particles in the size range of 0.01 to 0.1 micron, and is mostly applied

in industry for purifying and concentrating macromolecular (103 - 106 Da6) solutions, especially protein solutions

6 Da is the standard mass unit on an atomic or molecular scale, and it is defined as one twelfth of the mass of an unbound

neutral atom of carbon-12 in its nuclear and electronic ground state.

28

Nanofiltration takes out particles in the size range of 0.1 to 1 nanometers. Its pore size is smaller

than microfiltration and ultrafiltration and larger than reverse osmosis membrane. Nanofiltration is mostly used

with low total dissolved solids water such as surface water to remove disinfection by-product precursors such as

natural organic matter and synthetic organic matter.

Reverse osmosis filtration is a process that uses high pressure to push pure water in a concentrated solution that

is greater than the solution’s osmotic pressure flow across a semipermeable membrane to the other side. RO can

remove many types of ions and molecules from solution. It is highly effective for removing organic and inorganic

contaminants, as well as bacteria and viruses.

The pore size, the molecular weight cutoff and the particles removed by each filtration process can be seen in

Table 3-1.

Table 3-1 Filter Categories, Pore Sizes, Molecular Weight Cutoff, Filtration Pressure and Particles Removed (Baker, 2012)

The Gravity Non-electric (GNE) filters is a special category defined by the author for the purpose of this study.

And also because it helped to simplify the characterization of HWF products found in the Indian marketplace.

Figure 3-1 shows the pore size and the size range of water constituents of each filtration process.

29

Figure 3-1 Filtration Application Guide (modified from Filtration Application Guide, Water Quality Improvement Center)

Table 3-2 shows all models tested at CR.

Table 3-2 All models tested at CR

30

Conventional Particle Removal Filter

Conventional Particle Filters found in the Indian marketplace in the city of Ahmedabad, were purchased and

shipped to the Consumer Reports lab in New York. This included cloth (see Figure 3-2 on the right) and Jali mesh

filters (see Figure 3-2 on the left) that are widely used in Ahmedabad households, especially low income families,

to improve water quality. They are a common kitchen item of other income groups as well.

Figure 3-2 Cloth (right) and Jali Mesh (left) filter

Cloth Filter

Different kinds of cloth were purchased in Ahmedabad, but because they were not branded, it was difficult to

determine their manufacturer. The author chose three kinds of cloth to represent the “best,” “medium,” and

“lowest” quality in Ahmedabad marketplace based on the tightness of their weave and they are respectively

numbered as Cloth #1, #2, and #3. A square meter of a cloth filter costs only about one dollar.

According to CITE’s fieldwork in India, people in Ahmedabad usually use the cloth in one-layer or two layers.

Only a few people were seen to use four layers. Figure 3-3 shows examples of the cloth filters that the CITE field

team found in Ahmedabad households. Normally the cloth is tied around a faucet or put over a storage tank to

filter water. It needs to be washed or changed once the homeowner believes it to be dirty.

31

Figure 3-3 Cloth filters found in Ahmedabad households

The pore sizes of one layer of Clothes #1, #2, and #3 were measured using an Olympus FH Microscope (220959)

with the highest magnification of 4000x (Figure 3-4). The pore of Cloth #1 cannot be seen at the highest

magnification. Cloth #2 has a pore size of about 100μm, and Cloth #3 has a pore size of about 200 to 300μm.

Thus, generally cloth in Ahmedabad can only remove some large particles like sand.

32

Figure 3-4 The micrograph of cloth #1

Figure 3-5 The micrograph of cloth #2

Figure 3-6 The micrograph of cloth #3

33

The advantage of cloth filters is that they are cheap and easy to use. The disadvantage is that it is not effective to

remove contaminants of concern. Moreover, drinking water can become contaminated if the cloth is not kept

clean.

Mesh Filter

There were two types of Jali mesh filters found in Ahmedabad: one and two layer mesh filters. Five different

models of Jali mesh were purchased in Ahmedabad and shipped to CR’s Lab for testing:

Robin Brand Mesh

Robin Rimpi-99 Mesh

Robin Big Boss Mesh

Akash Jaldhara Mesh

Marshal Zeba Mesh

All of the models are two-layer mesh filters, and they have the same tightness and weave. Thus, the Robin Rimpi-

99 Mesh was chosen to represent this type of filter.

In the household, Jali mesh filters are placed over a storage container to filter water where the filtered clean water

is then stored for daily use. Figure 3-7 shows how Indian people use a Jali mesh filter at home. Mesh filters are

inexpensive costing only about US $0.50 each.

Figure 3-7 A mesh filter in use in an Indian household

As with the cloth filters, the Jali mesh filters weave was also measured using a microscope. Figure 3-8 shows the

pore sizes of one-layer and two-layer mesh filters. The one-layer mesh filter has a pore size of 300μm, which is

34

similar to Cloth #3.

Figure 3-8 The micrograph of mesh filters (Left: one layer; Right: two layers)

Gravity Non-Electric (GNE) Filter

GNE filters are gravity-driven, manual-fill filters that do not need electricity. Thus, they are often put on a kitchen

countertop, as well as being easy to use and maintain. Most GNE filters are in the class of microfiltration or

ultrafiltration, with a few even in the nanofiltration range. Their cost is significantly higher than the cloth or mesh,

but far less than the RO filter systems. Table 3-3 lists the GNE filter models that have been tested at Consumer

Reports Laboratory as well as their manufacturers and prices.

Table 3-3 GNE Filter Models Tested

35

Expresso Stainless Steel Water Container

Expresso stainless steel water container is a micro-filter that uses two ceramic candle elements. The filter consists

of two stainless steel containers, and the candle elements are screwed into the base of the upper container. Indian

ceramic candle filters are commonly used in countries such as Nepal, and Brazil, but were not widely found in

Ahmedabad. Candle filters used to be common in India but have been superseded today by a variety of other types

and models. One survey respondent in Ahmedabad said “My grandmother still uses a candle filter.” The candle

element looks like a thick candle and is made from white clay. The sizes of the pores that let the water go through

the ceramics are very fine. The sizes of these pores differ and can be as small as 1 micrometer (Sagara, J., 2000).

Candle elements have very slow flow rates, thus a water filter of this type usually contains two to three candle

elements in order to increase the clean water flow rate. Dies’s research shows that five candle elements of different

compositions resulted in flow rates ranging from 300 ~ 840 mL/hr/candle (Dies, 2003). The price of this Expresso

model was not recorded, but according to Sagara (2000), the typical retail price for a similar candle filter found in

Nepal ranges is between $8 to $21.

Figure 3-9 Expresso Stainless Steel Water Container Figure 3-10 Ceramic Candle Element

According to the instruction for this Expresso water filter model, the ceramic elements need to be soaked in clean

water for two days prior to first using the filter. Further, the first use, filtered water should not be used as drinking

water. The manufacturer’s instruction on the Expresso Stainless Steel model indicates that the flow rate is very

slow in the beginning, but would increase after 14 days’ of use when all the pores fully open. For maintenance,

the user must keep the ceramic filter candles clean by regularly brushing the surface gently under clean flowing

water.

36

Tata Swach Cristella Plus

Tata Swach Cristella Plus model is manufactured by Tata in India (Figure 3-11). The model is comprised of two

parts. The upper part contains: 1) a pre-filter made from fabric that used as a conventional particle filter removing

big particles, 2) a reservoir for the untreated water, and 3) portion where the Tata Swach bulb can be attached (see

Figure 3-11). The lower chamber is a safe storage container which collects clean water that is accessible to the

user via a water tap at the bottom. The storage capacity is 9 liters, and the Swach bulb has a purification capacity

of 3000 liters. The flow rate of the filter is claimed to be 3-4 liters/hr. There is a device in the bulb that purportedly

indicates when the bulb should be replaced. The manufacturer’ instructions say that this device can stop the flow

of water once the purifying power of the bulb is exhausted. There is a fuse in the bulb that indicates when the bulb

should be replaced, and the bulb can stop the flow of water once the purifying power of the bulb is exhausted.

Figure 3-11 Tata Swach Cristella Plus (left) and Swach bulb7 (right)

The core, filtration part of this model is the Tata Swach bulb, which uses silver nanotechnology as its means of

7 http://www.snapdeal.com/product/tata-swach-filter-candle-bulb/

37

water purification. The manufacturer claims that this Swach Bulb technology uses rice husk ash impregnated with

nano (1 x 10-9) silver particles and contains activated silicon and carbon. Tata claims that this filtration bulb can

remove turbidity and the silver particles can inhibit bacteria multiplication. The nano-sized particles increase the

filter surface area so that the bacteria have enough reaction time (India Center for Science and Environment,

2010). Tata Swach Cristella Plus instructions indicate that it can remove 109 bacteria and 107 viruses without

harmful chemicals used for purification.

Tata recommends that the pre-filter be washed at least once a week and the mesh at the bottom of the bulb should

be cleaned at least once a month.

The instructions for this model say that before the first use, the user should fill the upper container with water and

wait until all the water goes into the bottom safe storage container. Then, the user should completely empty the

bottom container. After this process, the filter can be used to provide clean water.

Tata Swach Smart

Tata Swach Smart is also manufactured by Tata Company (see Figure 3-12). The difference between the two

Swach Smart models is that one uses a Swach bulb with a purification capacity of 1500 liters and the other uses a

Swach bulb with a purification capacity of 3000 liters. Tata Swach Smart has a storage capacity of 7.5 liters. It

also uses silver nano technology in its bulb as found in the Tata Cristella Plus model. Figure 3-9 shows Tata Swach

Smart model and its Swach bulb. It also has two containers: the upper container has a microfiber, pre-filter on the

top and the Swach bulb attached at the bottom; the clean water is stored in the bottom container after filtration.

The bulb also has the device to indicate when the bulb should be replaced.

The pre-filter needs to be washed at least once a week, and the mesh at the bottom of the bulb should be washed

at least once a month.

According to the instructions, before the first use, the user needs to fill the upper container with water and wait

until all the water goes into the bottom container. Then, the container is completely emptied. After this process,

the filter can be used to provide clean water.

38

Figure 3-12 Tata Swach Smart8 (left) and its Swach Bulb9 (right)

Hindustan Unilever PureIt Classic 14L

The multi-stage, Pureit Classic 14L is produced by Hindustan Unilever (see Figure 3-13). The water first goes

into a microfiber pre-filter and then passes through an activated carbon filter. Activated carbon is a special form

of carbon with small pores that increase the surface area available for adsorption or chemical reactions (Mattson,

J. S., 1971). This activated carbon filter, according to its manufacturer, removes dirt, parasite, and pesticide

residuals. Next, the water goes to what Unilever calls a “Germkill Processor” using “programmed chlorine release

technology” to purportedly kill harmful viruses and bacteria. Finally, the water passes through a component called

the “Polisher” that is supposed to remove residual chlorine and give the clear water a good taste. Unilever claims

that this model can remove 107 virus in 1 liter of water.

The top chamber has a capacity of 5 liters, and the transparent safe storage chamber has a capacity of 5 liters. The

Germkill Kit™, including the activated carbon filter, the Germkill processor, and the polisher, has a claimed

purification capacity of 1000 liters of water, which, for a family of 5, translates to a 50-day lifetime, assuming 4

liters per person per day. It also has a Germkill Life Indicator that is claimed to give advance warning before the

8 http://www.tataswach.com/know_tata_swach/tata_swach_smart.html

9 http://www.cromaretail.com/Tata-Swach-Bulb-(Yellow)-pc-21718-462.aspx

39

Germkill Kit™ needs to be changed.

Figure 3-13 Hindustan Unilever PureIt Classic 14L10

10 http://www.pureitwater.com/IN/products%E2%80%8E/pureit-classic-14l

40

Prestige LifeStraw

Prestige LifeStraw is manufactured by Prestige (see Figure 3-14). It has three stages of treatment: According to

the manufacturer, the first stage is a microfiber, pre-filter to remove relatively big particles; then the water flows

through a carbon block that removes chlorine, sediment, volatile organic compounds, taste and odor; the third

stage is the ultrafiltration membrane, which is the core technology. The model’s instructions claim that the

ultrafiltration membrane can remove 99.9999% bacteria, 99.99% viruses, 99.99% protozoan parasites and

particles larger than 0.02 micron while using no chemical for filtration.

Prestige LifeStraw has a total capacity of 18 liters where the clean water storage tank is half that size (9 liters).

The manufacturer claims that it has a purification capacity of 4500 liters of water before the ultrafiltration

membrane is replaced.

Figure 3-14 Prestige LifeStraw

41

Eureka Forbes Aquasure Kitanu Magnet

Aquasure Kitanu Magnet water filter is produced by Eureka Forbes (see Figure 3-15). According to its

manufacturer, it has three stages of purification: the first stage is a microfiber pre-filter that removes particulates;

the second stage is a sediment filter that consists of a microfiber mesh with high surface area to remove impurities

that are not visible to human eyes; the final stage is the core Kitanu Magnet with “Positive Charge Technology”

TM (PCT), which is claimed to have nano-fibers to “attract and pull out bacteria and viruses” that does not require

chemicals to purify water,.

The top container and bottom container have a storage capacity of 9 liters and 11 liters respectively. There is a

specially designed float at the bottom of the top container that can regulate the flow so that both containers can be

used to store water. The Kitanu Magnet cartridge is claimed to have a natural shut-off function that will close

gradually when the cartridge is approaching its end of life. The instruction says that generally, the Kitanu Magnet

(see Figure 3-15 right) needs to be replaced after every 750 liters of water, and the microfiber pre-filter need to be

washed every 15 days. The initial run of filtered water through the cartridge needs to be discarded and not drunk.

Figure 3-15 Aquasure Kitanu Magnet (left) and its Kitanu Magnet “Positive Charge Technology” Cartridge11 (right)

11 http://www.shoppingstore.in/index.php?route=product/product&product_id=132

42

KENT Gold Plus - 20L

The 20L Gold Plus model is produced by KENT (see Figure 3-16). According to the manufacturer, the filtration

process consists three stages: the untreated water is filled into the top tank and then passes through the sediment

filter that removes suspended impurities; the second stage is silver impregnated carbon granules that removes

chlorine and odor; then the water flows through the core ultrafiltration membrane that removes bacteria and cysts.

The user manual says that it can reach 99.96% reduction of cysts.

The top tank has a capacity of 7 liters and the bottom tank has a capacity of 13 liters. According to the user manual,

the ultrafiltration membrane has a purification capacity of 4000 liters, and the carbon filter has a purification

capacity of 900 liters. The flow rate is claimed by the manufacturer to be 18 liters/hr.

The safe storage tank needs to be washed using clean water once every 7 days; the sediment and carbon filters

should be cleaned at least once in 30 days; the ultrafiltration membrane should be backwashed at least once in 30

days. The sediment filter must be changed after 3 months; the carbon filter should be changed after 6 months; and

the ultrafiltration membrane should be changed once in 12 months.

Figure 3-16 KENT Gold Plus12 (left) and its ultrafiltration membrane13 (right)

12 http://www.latestviews.com/home-appliances/water-purifiers/kent-gold-plus-water-purifier/product-gallery/

13 http://www.ebay.in/itm/Kent-Gold-Gold-Optima-Star-Spare-Part-1-UF-Membrane-/181584229338

43

Everpure Unbreakable

Everpure’s “Unbreakable” water filter has two filters in it: the first filter in the upper container is a micro pore

ceramic filter, and the second filter attached at the bottom of the upper container is a sediment filter filled with

granular activated carbon, silica stand, zeolite, and mineral stones & mineral sand dish. It has a storage capacity

of 15 liters. The specification of this model says that it can provide 35-65 liters of clean water everyday at a flow

rate of 2.5L/h to 5L/h.

However, there is no seal ring between the upper and bottom container. Thus with multiple test samples of this

model were found to leak severely. As a result this Everpure model was not performance tested in the laboratory

and was considered not acceptable for purchase.

Figure 3-17 Everpure’s “Unbreakable” (left) and its mineral filter14 (right)

14 http://www.naaptol.com/water-filters-and-purifiers/everpure-7-step-mineral-water-purifier/p/12442950.html

44

Table 3-4 shows the special manufacturer claims for the GNE models described in the Section 3.3.

Table 3-4 Special Manufacturer Claims for the GNE models tested

45

Reverse Osmosis Filter

The Reverse-Osmosis (RO) water filter has some important advantages over the particle removal filters and the

GNE filters. They cannot only effectively remove turbidity and bacteria, but also total dissolved solids, which

cannot be removed by either particle removal or GNE filters.

The key element within an RO water filter is its RO membrane. Figure 3-18 below shows how the RO membrane

operates.

Figure 3-18 Reverse-Osmosis System Schematic15

The RO membrane itself is shown schematically in Figure 3-19 below.

15 http://espwaterproducts.com/about-reverse-osmosis.htm

46

Figure 3-19 Diagram of a Reverse-Osmosis Membrane16

Normally, there are at least one pre-filter and one post-filter before and after the RO membrane. The post-filter is

usually filled with activated carbon that gives the clean water a better taste.

The RO marketplace in India is focused on the middle to high-income groups because RO filters have a high

purchase price and operating costs. The purchase price of an RO filter purchased in Ahmedabad ranges from $108

to $300. The other downside is that RO filter is known for producing a large amount of wastewater. It can produce

wastewater as much as 3 times of the amount of clean water. For municipalities where water is scare and/or

expensive to buy, wastewater generation is both an environmental sustainability and cost concern.

There are two types of RO filters in the Indian marketplace: locally-assembled RO’s named “Dolphin”, and

branded RO’s. CITE’s field work in India determined that the RO market in India has been growing significantly

with the introduction of Dolphin because they have a much lower price. Thus, it was of interest to the Suitability

Team at Consumer Reports during the summer of 2014, to compare the performance of the Dolphin RO’s to the

established brands like Tata (RO Swach Platina Silver) and Eureka Forbes (RO Aqua Sure).

The following RO filters were purchased in India and tested at CR’s Headquarters.

16 http://erkinchik.wordpress.com/ro-membrane-housing-hook-up/

47

Table 3-5 RO Filter Models Tested

Category Model Model Type Price (USD)

RO

Blue Diamond Non-branded, Dolphin model 108

Clean Water Non-branded, Dolphin model 108

Dolphin Gold Non-branded, Dolphin model 108

Tata Swach Platina Silver Branded model 238

Locally-assembled Dolphins

There were three Dolphin models tested at the CR Lab: Blue Diamond, Clean Water, and Dolphin Gold. Among

the three models, Clean Water was lifetime testing, while Blue Diamond and Dolphin Gold were only tested for a

short time because of time constraint.

All of the Dolphin filters have the same basic design. They have five stages (see Figure 3-20): 5 Micron PP

(Pleated Polypropylene) Sediment Filter outside of the filter that purports to remove suspended impurities such as

sand, dust, and dirt; Inline Sediment Cartridge that further eliminates other particles; A Pre-Carbon Filter that

purposts to remove color, odor, chlorine, and pesticides; an RO membrane that purports to eliminate toxins,

chemicals, total dissolved solids, viruses and bacteria; and a Post Carbon Filter that claims to provide a natural

taste of water. The clean water exiting these cartridges is stored in a 9-liter tank. When the tank is full, there is a

float that stops the RO from operating. Inside each Dolphin filter, there is a pressure pump and an AC-DC voltage

transducer. Figure 3-21 shows the front and the inside look of a Dolphin filter.

Figure 3-20 Dolphin design stages and their function

48

Figure 3-21 Front and Inside Look of a Dolphin

When customers buy a Dolphin, a technician will help them to assemble the product at their homes.

The instruction on the package box claim that a Dolphin has a filter capacity of 90 liters/day at 25℃.

Tata Swach Platina Silver

Tata Swach Platina Silver is a branded RO filter manufactured by Tata Company (see Figure 3-22). In common

with the Dolphin filters, it also has five stages claimed by the manufacturer:

Stage 1: Ten micron sediment filtration cartridge that reduces coarse impurities such as dust and sediments which

are above 10 micron in size.

Stage 2: Bacteriostatic Granular Activated Carbon (GAC) with Nano Silver Impregnation technology that reduces

chlorine, odor, volatile organic compounds and pesticides. The Nano-silver impregnation technology used in this

cartridge reduces the chance of bio-fouling and hence increases the life of carbon cartridge.

Stage 3: Five micron sediment filtration cartridge removes finer impurities such as dust and sediments which are

above 5 micron in size.

Stage 4: National Science Foundation (US) - NSF certified RO Membrane that has fine pores as low as 0.0001

micron in size. The manufacturer claims that it reduces water contaminants such as dissolved salts, pesticides,

heavy metals and water borne micro-organisms such as virus, bacteria.

Stage 5: Post Bacteriostatic Granular Activated Carbon (GAC) with Nano Silver impregnation technology that

claims to impart bacteriostatic property to the purified water and enhances the taste of water.

49

Figure 3-22 Tata Swach Platina Silver17 Figure 3-23 Tata Swach Platina Silver under testing

Tata Swach Platina Silver has a 7 liter zero contamination storage tank to store clean water. Inside the filter there

is a pressure pump and a voltage transducer. The model has an auto-flushing system to clean the membrane. It

also has what the manufacturer calls a Double-i-Care indicator. It is claimed that if the first sediment filtration

cartridge clogs, the low pressure switch will be activated and the Double-i-Care TM indicator will show a fault

indication.

The inlet water pressure needs to be within the range of 5 psi to 35 psi, and the temperature should be between

2℃ to 49℃. According to the manufacturer’s specification, it has a purification capacity of up to 12 liters/hr under

the condition of 10 psi input pressure and 750 ppm TDS at 25℃.

17 http://www.tataswach.com/know_tata_swach/tata_swach_silver_platina_ro.html

50

4 Method

Comparative Product Testing Attributes

After referring to the WHO “Scheme” and other research discussed in chapter 2, six major water quality

performance attributes were identified as applicable to the use of water filters in Ahmedabad households. These

attributes included: E.coli removal, turbidity removal, TDS removal, RO clean water flow rate, RO % recovery

(of clean water), and clogging/filter lifetime. The author selected these attributes based on the best expert judgment

that they were major ones in India. Had time and resources allowed, others, for example, arsenic, fluoride, iron

and nitrate might have been selected. However, the author chose attributes of major importance within the

constraints of the time and resources available.

E.coli Removal

E.coli, of the genus Escherichia, is an anaerobic, rod-shaped bacterium with about 2.0µm in length and 0.25-

1.0µm in diameter (Kubitschek, H. E., 1990). It is commonly found in the intestine of warm-blooded organisms

(Singleton P, 1999). E.coli removal was chosen to represent a filter’s ability to remove bacteria. Percentage

removal and log removal value (LRV) both indicate this attribute. LRV can be calculated as:

𝐿𝑅𝑉 = −log (#𝑀𝑃𝑁 𝑖𝑛 𝑖𝑛𝑓𝑙𝑢𝑒𝑛𝑡

#𝑀𝑃𝑁 𝑖𝑛 𝑒𝑓𝑓𝑙𝑢𝑒𝑛𝑡) Equation 2

Most Probable Number (or MPN) is the most probable number of viable bacterial cells according to Quanti-tray

test results.

According to WHO “International Scheme to Evaluate Household Water Treatment Technologies,” an input E.coli

concentration of around 105 colonies per milliliter was used in their study (WHO, 2014a). As per their

microbiological organisms and reduction requirements (Table 4-1), when the E.coli LRV is higher than 4 the filter

is rated as “highly protective,” while when the percentage removal is between 99% and 99.99% then the filter is

defined “protective or limited protective”.

51

Table 4-1 Microbiological Organisms and Reduction Requirements (WHO, 2014a)

In the 2nd edition of WHO Guidelines for Drinking Water Quality (WHO, 1997), five risk levels were defined

according to E.coli concentration in the water sample (Table 4-2). The fourth edition does not include this

information).

Table 4-2 Risk Level from E.coli (WHO, 1997)

Risk Level

E.coli in sample

(coliform forming unit

per 100 ml)

Conformity < 1

Low 1 – 10

Intermediate 10 – 100

High 100– 1000

Very High >1,000

WHO “Protocol” only gives two performance ranks, which cannot distinguish the filters very well. Thus, the CR

team developed a different evaluation standard. According to Table 4-2, the author also divided E.coli removal

performance into five levels: when the percentage removal of E.coli is higher than 99.99%, that is, when the LRV

of E.coli is higher than 4, the filter can be defined “excellent”; when the percentage removal is between 99.9%

and 99.99%, then the filter is defined “good”; when the E.coli percentage removal is between 99% and 99.9%,

then the filter can be seen as “fair”; when the filter can only remove 90% to 99% E.coli bacteria, it performs

“poorly.” Finally, when the filter removes less than 90% bacteria, it is defined as “unacceptable”. Table 4-3

presents our evaluation standard.

52

Table 4-3 CR team’s standard for evaluating bacteria removal performance of the filters

Performance % Removal LRV

Excellent >99.99 >4

Good 99.9~99.99 3~4

Fair 99~99.9 2~3

Poor 90~99 1~2

Unacceptable <90 <1

Turbidity Removal

The inlet “challenge” water had a turbidity of 40 NTU (Nephelometric Turbidity Unit) based on the WHO

“Protocol” (WHO, 2014b). The effectiveness of turbidity removal was measured as percentage removal.

Typically, in the municipal water in Ahmedabad, the turbidity is quite low ranging from 0.1 to 0.5 NTU, and

turbidity of lakes in Ahmedabad ranges from 4 to 11 NTU (Devangee S., 2013). Thus, the turbidity of CR team’s

challenge water is much higher than the turbidity level found in Ahmedabad, because: 1) it represented the worst

scenario, and 2) using a fixed concentration guided by WHO is easier to compare our findings with other products

throughout the world and with subsequent research. The WHO “Protocol” specified the use of a specially

processed test dust product – A2 dust – that can simulate turbidity in water at a lab scale. The A2 test dust has

particles sizes from 0.2 to 176 microns.

Table 4-4 Criteria for evaluating turbidity removal of the GNE filters

Performance Turbidity Removal (%)

Excellent 80~100

Good 60~80

Fair 40~60

Poor 20~40

Unacceptable 0~20

Total Dissolved Solids (TDS) Removal

TDS is a measure of the combined content of all inorganic and organic substances contained in a liquid in

molecular, ionized or micro-granular suspended form. It comprises inorganic salts (principally calcium,

magnesium, potassium, sodium, bicarbonates, chlorides and sulfates) and small amounts of organic matter that

are dissolved in water. TDS in drinking-water originates from natural sources, sewage, urban runoff and industrial

wastewater. The palatability of drinking water has been rated by panels of tasters in relation to its TDS level as

53

shown in Table 4-5 (WHO, 2006).

Table 4-5 Palatability of Drinking Water (WHO, 2006)

Description TDS (mg/l)

Excellent < 300

Good 300 – 600

Fair 600 – 900

Poor 900 – 1,200

Unacceptable 1,200

Figure 4-1 (Murcott S., 2014) shows the boxplot18 of TDS in Ahmedabad source waters. The green line indicates

the Indian Standard Requirement (Acceptable Limit) which the MIT CITE team used as drinking water quality

standard. The red line shows the Indian Standard (Permissible Limit). On average, groundwater in Ahmedabad

contains 1,079 mg/L TDS. TDS for the challenge water was set at 1500 mg/L by the CR team. While this is a high

level, it is the one suggested by “Protocol” (WHO, 2014b). It is higher than the level found in Ahmedabad

groundwater. The rationale for using this much higher value is similar to the reasons given above for why the

higher Turbidity level was chosen. In our experiment, Epsom salts (MgSO4), which contributes 100% to hardness,

was used to test TDS removal.

18 Box plots display differences between samples without making any assumptions of the underlying statistical distribution:

they are non-parametric. It is a convenient way of graphically depicting groups of numerical data through their five-number

summaries. They include the smallest observation known as sample minimum, lower quartile (Q1), median (Q2), upper

quartile (Q3) as well as the largest observation also known as sample maximum.

54

Figure 4-1 Boxplot of TDS in Ahmedabad source waters (Murcott S., 2014)

Water’s hardness is determined by the concentration of multivalent cations (metal cations positively charged with

a charge greater than 1+) in the water. Multivalent cations are cations Hardness is caused mainly

by magnesium and calcium cations, and it can be discussed in terms of carbonate (temporary) and non-carbonate

(permanent) hardness (U.S.EPA, 1976). Hardness is traditionally measured by chemical titration. Some water can

have high TDS concentration but low hardness level. The hardness of a water sample is reported in milligrams

per liter (same as parts per million, ppm) as calcium carbonate (mg/l CaCO3). The criteria for evaluating TDS

removal of the GNE filters is the same as that for evaluating turbidity removal (see Table 4-3).

Table 4-6 Hardness Description (Ratnayaka, Johnson, & Brandt, 2000)

Description Hardness as CaCO3

(mg/L)

Soft 0 – 50

Moderately soft 50 – 100

Slightly hard 100 – 150

Moderately hard 150 – 200

Hard Over 200

Very Hard Over 300

Clean Water (“Product”) Flow Rate

Clean water is the product of water filters, which can then be used as drinking water. In the water industry it is

referred to as “product water” but here the author will simply refer to it by the more common term “clean water”.

The flow rate of clean water determines whether consumers can get enough water for daily use. As mentioned in

55

the literature review, Pimentel, D. (2004) stated that human need to drink from 1.5L to 2.5L per person per day to

stay healthy. While WHO (Howard G., 2003) suggested a minimum of 7.5L per capita per day. Because WHO is

based on requirements of a lactating women in general physical condition, and the 7.5L includes water for cooking.

Thus, the author decided to use the limited amount of 1.6L per capita per day, then, for a family of five, their daily

consumption of drinking water is 8L. Assuming they use a water filter eight hours a day, it requires the filter flow

rate to be at least 1.0L/hr. This became the cutoff for a GNE filter’s end-of-life in this research program.

The scoring for rating clean water production was different for each category of filter—Particle Removal, GNE

and RO models—because their clean water flow rates were not comparable. The flow rate of a Conventional

Particle Removal Filter the author tested was greater than 50L/hr, while the flow rates of GNE filters were

generally within the range of 1L/hr to 10L/hr. An RO filter had a flow rate of around 14L/hr. The criteria for

evaluating clean water flow rate of the GNE filters defined by the author is shown in Table 4-7.

Table 4-7 Criteria for evaluating clean water flow rate of the GNE filters

Performance Clean water flow (L/hr)

Excellent 8~10

Good 6~8

Fair 4~6

Poor 2~4

Unacceptable 0~2

RO % Recovery

Percent Recovery is the amount of water that is being recovered as “permeate water”, also known as “product

water”. Another way to think of % Recovery is the amount of water that is not sent down the drain as wastewater,

but rather collected as permeate or product water. The higher the % Recovery means the less water down the drain

as wastewater and saving more clean water. RO water filters are known to create wastewater at rates that can

exceed three times the amount of the clean water being produced (Eisenberg, T. N., 1986). This is a sustainability

issue of a valuable natural resource being wasted. If the % recovery is too high for the RO design then it can lead

to larger problems due to scaling and fouling. Additionally, assuming the owner pays for their water, it may

represent a significant operating cost. See the calculation for % Recovery in Section 2.4 Equation 1.

Clogging/Filter Lifetime

In this study, clogging, which is the end-of-life of filters, for GNE filters was defined as when the clean water

flow rate fell below 1L/hr. For RO filters, the end-of-life was defined as when clean water flow rate fell below

100mL/min. Thus, filter lifetime measures how long a filter can retain its clean water flow rate. It is a factor related

56

to the convenience of using the filter and the lifetime cost of the filter. Table 4-8 presents our evaluation standard.

Table 4-8 Criteria for evaluating lifetime of the GNE filters

Performance Lifetime (day)

Excellent 24.75~31.25

Good 18.25~24.75

Fair 11.75~18.25

Poor 5.25~11.75

Unacceptable 0~5.25

Test Water Source

Base Water

Base water is the water into which the contaminants, including turbidity, TDS and E.coli, were added to form the

“challenge water” that is used to test the filters. Thus, it needed to be consistent in this study and in any potential

further research at MIT or in India. There were various sources of water available at CR Lab that could be used

as the base water: Yonkers’ municipal water, CR well water, and deionized water. Their characteristics and the

reason why they were chosen or not are stated below.

(1) Yonkers’ Municipal Water

The Yonkers’ municipal water is quite soft with a hardness level of 30mg/L, but it contains fluorine and chlorine

that would kill the E.coli bacteria added to measure the E.coli removal rates. While the fluorine and chlorine could

be removed using chemicals like sodium thiosulfate, this would increase the TDS in the water, and might also

affect the removal of E.coli as well. Thus, the Yonkers municipal water could not be used when testing E.coli but

could be used when doing turbidity and flow rate testing.

(2) CR Well Water

The CR well water has a high level of hardness (about 300 mg/L), which is in the range considered for the

challenge water in this research. Thus, the author considered it particularly worthy in the beginning. However, the

same well water is not available at MIT or in India, so future research would not be able to be continued after the

two-month duration that the team spent at the CR labs. As a result, the CR well water was not used.

(3) Deionized Water

Using deionized water can avoid contamination and can ensure that the base water is consistent. Deionized water

is also available at MIT and at a lab in India where CITE is partnered, So, deionized water proved to be the ideal

base water. As a result a portable, Hydra DI High Capacity Deionization System, Model HY-122B-DI-HC

57

(Grainger Item #44C706) with a maximum flow rate of 75 gallons (284 liters) per hour production rate was

purchased and made operational for use in producing deionized water for this research. Unfortunately, this unit’s

flow rate diminished quickly from about 60 to only 6 gallons (227 to 22.7 liters) per hour after a month’s use even

though Yonkers had relatively soft water (30 mg/L) going into the deionizer. Fortunately, its lifetime was sufficient

since this research ended in early September as a replacement deionizer cartridge was about to be delivered.

Considering all the advantages and disadvantages of the three water sources, the author used the deionized water

as our base water for most of the summer because the initial plan was to measure E.coli removal rates throughout

the lifetime of each filter. For this scenario, the E.coli bacteria would have been mixed into the water of one of the

two available 100-gallon (378-liter) tanks.

However, such a lab set-up had an associated high cost of buying phosphate buffered saline (PBS). The PBS would

have ensured the E.coli concentration in the water is stable for a long time – at least 24 hours in this study.

However, the PBS cost was estimated to be $20,000 just for the summer.

The team researched substitutes for a saline solution in place of PBS, but failed after nearly one-month’s efforts.

Then, the team invented the “E.coli Injection System” which was designed and built to provide small batches of

E.coli at the RO filter’s inlet that did not have to live for a long period of time. By doing this, the E.coli was only

tested at the beginning and end of life of each filter. Actually, this process imitates the real condition of bacterial

contamination, which often appears for a short period. This injection system is discussed in detail in Section 4.4.3.

Challenge Test Water

Challenge Test Water is the water used to test the water filters under the worst scenario so that the performance

limits under stressed circumstances can be known. The following characteristics are the concentrations for

challenge water recommended by “Protocol”:

Table 4-9 Challenge Test Water Characteristics (WHO, 2014b)

Constituent Specification Adjustment Materials

Turbidity (NTU) 40±10 NTU ISO spec. 12103-A2 fine test dust

TDS (mg/L) 1500±150mg/L Sea Salts, Sigma Chemical Company (7732-18-5)

Deionized water is the base water in the WHO’s “Scheme”, and turbidity and TDS are added to deionized water

to form the challenge water. The author used ISO 12103-1/CD, A2 fine test dust from PTI (Powder Technology

Inc.), the same test dust suggested by WHO, to create challenge water with a turbidity of around 40 NTU. The

specification for the A2 fine dust is shown in Appendix A.

After a series of trial experiments, it was found that 70 mg/L of test dust could produce a turbidity of roughly 40

58

NTU, that is, a total of 28 grams of test dust in a 100-gallon tank. And the test dust could disperse well throughout

the 100-gallon (378-liter) tank within ten minutes because of the use of the tank’s small, circulating pump.

As for TDS, we substituted Epsom salts (100% MgSO4) for the sea salts to provide hardness and TDS at the same

time. It was determined that 4 g/L of Epsom salt produced a TDS of 1525 mg/L, that is, adding 1600g Epsom salts

to the 100-gallon (378-liter) tank.

The challenge test water formulation for this study can be seen in Table 4-10.

Table 4-10 CR Team Challenge Test Water Formulation

Constituent Specification Adjustment Materials

Turbidity (NTU) 40±10 NTU ISO spec. 12103-A2 fine test dust

TDS (mg/L) 1500±150mg/L Epsom Salt (100% MgSO4)

E.coli Solution

To measure the ability of the filter to remove bacteria, an E.coli solution was created. According to the “Scheme”,

the challenge concentration of E.coli entering the water filter must be ≥105 MPN/100mL (WHO, 2014b).Thus,

the concentration of E. coli chosen was approximately 105 MPN/100mL. A non-pathogenic K12 strain of E.coli,

Product #10798, was purchased from the American Type Culture Collection (ATCC).

An amount of 0.5 mL of the freeze dried E.coli was placed into 4.5 mL of K12 culture broth made up of 8 gram

of tryptone, 0.5 gram of sodium chloride (NaCl) and 1 liter of deionized water. This solution was then incubated

for 8 hours at 37° C. After 8 hours, the mixture was streaked onto petri dishes using a sterilized mix-stick and

incubated for another 8 hours at 37° C. A colony from the plate was then placed into 10 mL of Luria Broth and

incubated for 24 hours. The final concentration of E.coli was approximately 109 MPN/mL. After incubation, the

solutions were placed into a refrigerator at 4° C. Once in the refrigerator, this solution has a lifetime of

approximately one month.

Test Method for Testing Attributes

Turbidity Test Method

Turbidity was tested using the HACH 2100P (Product Number: 46500-00) turbidimeter.

TDS Test Method

TDS was tested using the HACH Pocket ProTM Low Range TDS Tester (Product Number: 9531200) TDS meter.

59

E.coli Test Method

In order to measure the concentration of E.coli, the Quanti-tray/2000 test method manufactured by IDEXX©, was

used. This involves a three-step process:

(1) Extract the treated water into a sterile 100mL Quanti-Tray bottle (the amount is determined by the dilution

selected). Dilute the effluent using deionized water to form a 100mL sample mixture. (This step is

unnecessary if there is no dilution.)

(2) Pour the 100mL solution into a Quanti-Tray. Put the Quanti-Tray into the Quanti-Tray Sealer. The Quanti-

Tray® Sealer 2X automatically distributes the sample mixture into separate wells.

(3) Put the sealed Quanti-Tray into the incubator, the temperature of which has been pre-heated to 35° C. After

24-hr incubation, the number of positive wells can be converted to a Most Probable Number (MPN). As

shown in Figure 4-1 below, to count the total coliform, the wells that turn yellow are counted. To count the

E.coli, the wells that turn both yellow and fluoresce are counted.

Figure 4-2 Method to count total coliform and E.coli

Sterilization Procedures and Quality Control

Due to the nature of this research, it was imperative that all tests were kept as sterile as possible. Thus, all glassware

was sterilized in an autoclave for 50 minutes at 121° C and at high pressure (16-20psi). All surfaces were also

sterilized with 70% isopropyl alcohol before and after testing.

All bacteria tests were triplicated and the median value was used for LRV calculation.

60

Test Set-Up

Plumbing Set-Up

The Consumer Reports plumbing set-up developed by the S1-CR team is shown in Figure 4-3. The purpose of

designing this plumbing system is to produce an amount of challenge test water that is enough for daily filter

testing, as well as to feed the reverse osmosis filters automatically.

The major plumbing parts include:

(1) Floor drains throughout the CR lab;

(2) One deionizer which was used to produce the base water;

(3) Two 100-gallon plastic water tanks which were used to make and store the challenge test water. The two

100-gallon tanks were used in turn because the deionized water production rate by the deionizer was less

than the challenge water consumption rate. Thus, we were filling one tank with deionized water using the

deionizer while using the challenge test water from the other tank at the same time.

(4) One ¾ horsepower water pump paired with an 80-gallon pressurized tank and pressure controller. They were

used to agitate the challenge test water in order to make it homogeneous before going to the RO filters.

(5) PVC piping was used to connect those parts mentioned above.

All parts are listed in Appendix B.

61

Figure 4-3 CR Plumbing Set-Up

Reverse-Osmosis (RO) Flow Test Rig: Developing CR-Version 2

CR has their own flow test rig system (CR-Version 1.0) which had an automated data acquisition system. This

system required that there was no flow through the test filter except during the time when there was a flow

measurement. Considering that the filter lifetime testing requires a long-running process before the filter’s end-

of-life, this CR-V1.0 system could not meet our requirements. Additionally as mentioned before, the RO filters

create significant amounts of wastewater and this was not monitored by CR-V1.0.

As a result, a new RO flow test rig (CR-V2.0) was designed in the CR lab and it worked well. The new rig CR-

V2.0, like CR-V1.0, used only one flow meter since each flow meter costs over $500. But in CR-V2.0, each filter

being tested had a continuous flow regardless of whether the flow rate was being measured or not. This greatly

increased the speed of testing over a filter’s lifetime. This new test rig was also paired with LabView software.

This software is used to automate data processing and to sequence valve openings and closings. The system was

able to turn on/off solenoids automatically, measure flow and pressure. As a result, using the new test rig CR-

V2.0, the team was able to test two RO filters simultaneously by monitoring both their clean flow and waste flow.

Figure 4-4 shows a schematic of this new rig for a single RO filter. Two solenoids are provided to each filter flow.

Four solenoids are required for these two flow paths for an RO filter where both the clean and wastewater flows

need to be measured. Figure 4-5 and Figure 4-6 show the CR-V2.0.

Figure 4-4 Single RO Flow Test Rig Schematic

62

Figure 4-5 CR-V2.0 Figure 4-6 CR-V2.0 with Control Panel

Figure 4-7 shows the LabView Solenoid Controlling Interface which can control these four solenoids. This

interface was designed for CR-V2.0, which can test two RO units using 8 Solenoids. For example, if Solenoids 1

and 3 are closed while Solenoids 2 and 4 are opened, the RO clean water flow and waste water flow are going to

the drain directly without being measured; if Solenoids 1 and 4 are closed while Solenoids 2 and 3 are opened,

the RO waste water is flowing through the flow meter and the waste water flow rate is measured.

Figure 4-7 LabView Solenoid Controlling Interface

This CR-V2.0 has another advantage that it can be modified to test a number of RO filters simultaneously easily

and inexpensively. To do this, more solenoids “branches” would be created. As shown in Figure 4-4, there were

only two RO filters under test at one time at CR, but the set-up shows that it can be expanded easily.

63

The CR-V2.0 sequenced the measurement of the flow rate from one filter to another so as to monitor the flow and

automate data acquisition. It measured the clean water flow and the wastewater flow alternately. The flow rate

was monitored periodically throughout its life without stopping the flow. In this way, decreases in clean water

flow and increases in waste water flow could be assessed.

CR-V2.0 is comprised of LabView software and National Instrument electronics. The parts used are listed in

Appendix B. This combination proved to work very well for monitoring and measuring the clean and waste flows

automatically. Also, it is very easy to change the measuring sequence according to user’s needs (see sample flow

data in Appendix C). To start measurement, the following parameters should be determined:

(1) The interval between flow readings

The software allows for one second as the minimum available time interval. This can be stepped up in increments

of 1 second. After some trial tests, the author chose two seconds for this study. For more details, see Table

Appendix C.1 that shows the measurement data of a portion of the over 3700 rows of data per round of test. The

data in Appendix C.1 were taken for the clean water of RO#1. After two minutes, this measurements switched to

the wastewater of RO#1.

(2) Time period during which each flow was measured

Any time period can be set in the controlling interface. For this study, the author chose two minutes, during which

time the test rig could capture 60 flow measurements. When calculating the flow rate, the readings in Appendix

C.1 were averaged without using the first and last three measurements. Table Appendix C.2 gives more details

about the analysis process.

(3) Time period during which the flows were not measured

Any time period could be set for this parameter. In this study, the author used four minutes after all the flows were

measured.

(4) Number of times that the sequence was repeated in one round of test

Any number of times could be set. In order to have enough time to monitor the turbidity and TDS levels and also

switch tanks if necessary, the author chose ten measuring cycles per test. The test lasted a total duration of 2 hours

and 4 minutes.

The flow meter was the most critical element to be chosen for CR-V2.0 because the flow rates of RO filters vary

64

during the testing. The flow meter Model #10219 made by McMillan with a flow range from 13-100 mL/minute

to 1.0-10.0 L/minute was chosen. The author used Model #102/9T that can measure flows from 1.0 to 10 liter/

minute. In terms of the flow rates of our RO filters, which was between 0.2 and 0.6 liters per minute, this model

would be normally out of the range. Nonetheless, after some trial tests, it was verified that this model had excellent

repeatability and linearity for our flow range. Other models to be considered are Models #102/6 with a flow range

of 100mL/min to 1000mL/min, and #102/5 with a range of 50mL/min to 500mL/min. Because of time constraint,

we were not able to order these models.

Pressure was also measured by CR-V2.0 at the same time as the flow rate was being measured. The pressure

transducer model chosen for this study was Omega Engineering Model PX209-10005V capable of measuring up

to 100 psi of pressure. However, these pressure data have not been utilized in this study for the analysis of RO

filter performance because of time constraint. The author considered the pressure worthy of analysis because it is

hypothesized that as the RO membrane is used over time, the fouling of the membrane would increase the pressure

difference over the membrane for a given flow rate. By measuring this flow resistance, it may allow better

comparisons in the performance of different RO filters. It also may be possible to make end-of-life predictions

earlier. Thus, the author suggests that future researchers consider using these pressure measurement data to analyze

the RO membrane degradation.

The E.coli Injection System

Before coming to Consumer Reports, initially we expected to incorporate E.coli into the challenge test water

according the “Protocol” (WHO, 2014b), resulting in a huge cost ($20,000 for two months) of providing PBS in

order to ensure that the E.coli lifetime was sufficiently long.

To avoid these prohibitive PBS cost, we decided to test E.coli removal only at the beginning and the end of a

filter’s life. To do this, an E.coli Injection System was developed for the RO filters. It worked well in injecting the

E.coli solution into the RO filters.

As seen in Figure 4-8, the E.coli injection rig included two major portions:

(1) One standard single cartridge filter used as the mixing vessel.

(2) One 6-gallon (23-liter) pressurized tank. The water filter found normally in the cartridge has been replaced

with a PVC pipe with holes in it to mix the incoming water from the pump with the E.coli broth on its way

19 http://www.mcmflow.com/displayproduct.asp?PRODUCT=102

65

to the pressure tank.

Figure 4-8 E.coli Injection Rig Schematic

66

Figure 4-9 E.coli Injection Rig (Left: Mixing Vessel, Gage and Valves, and 6-gallon (23-liter)

Pressure Tank. Right: PVC pipe inside of the Mixing Vessel)

Normal challenge test water is flushed through the plumbing system for a period of about 5 minutes. The mixing

vessel is opened and emptied. After it is closed, 1 liter of E.coli broth at a concentration of 107 MPN/100 ml is

poured into the mixing vessel through the E.coli injection point as shown in Figure 4-9. The valve (V3) then is

closed when the vessel is filled with E.coli broth.

By opening the valve (V1), the challenge test water from the 80-gallon tank flows through the mixing vessel and

mixes with the E.coli broth. The 6-gallon pressure tank then fills with this E.coli mixture. There is sufficient flow

of this E.coli mixture to provide about 15 minutes of continuous flow for testing one RO filter at a time.

The steps to conduct the E.coli injection test method are as follows:

(1) An E.coli solution of 10^7MPN/100mL is placed in a 2 liter jar.

(2) The valve to the injection hardware and from the injection hardware to the RO inlet are closed. One liter of

the E.coli broth is poured into the injection canister. (It was expected based on previous test runs that this

would yield a concentration of 10^6MPN/100mL at the inlet.)

(3) Using deionized water as the source, at the peak water pump pressure of about 60 psi, the valve upstream

from the injection system is opened and water then flows to mix with the broth as it goes into an initially

empty, six gallon pressure tank. When the flow into the pressure tank stops, the valve upstream of the

injection system is closed disconnecting the water pump from the test rig.

67

(4) At this point, the test run begins with the valve downstream from the injection system being opened at time

equals zero and water flow to the RO begins.

(5) Samples are collected at the inlet, waste and clean water drains to measure their E.coli concentration.

The E.coli injection rig has three major benefits:

(1) There is no need to use the expensive PBS buffer solution.

(2) Sterilization and clean-up are simplified since only a small amount of the plumbing is exposed to E.coli.

(3) The amount of E.coli required is drastically reduced.

Test Methods for Each Filter Category

Cloth and Mesh Filters Test Method

Cloth and mesh filters turbidity removal test method: According to S1 team, there are two ways in which

Indian people use cloth filters: one is to tie it directly on the faucet, and another is to tie it on a bucket and pour

water in. We decided to use the second method to test the cloth filters. Figure 4-10 shows the experimental set up

for testing cloth and mesh.

Figure 4-10 The Experimental Set-Up for Testing Cloth and Mesh

The test procedure for testing cloth is shown below:

(1) Mark the 10L line inside a 20L plastic bucket.

(2) Tie the cloth on the top of the bucket.

(3) Pour the challenge water onto the cloth until the water level in the bucket reaches the marked line. The bucket

is transparent so the marked line can be seen from the outside.

68

(4) Remove the cloth and stir the filtered water in the bucket in order to make sure it is homogenous.

(5) Test the filtered water for turbidity.

(6) Repeat the above steps twice.

(7) Fold the cloth once, twice and four times, and test their turbidity removal respectively.

The testing procedure for mesh filters is basically the same as that for cloth filters. But because the mesh filters

are actually manufactured as two-layer (see Figure 4-10 right) and people only use them as two-layer, they were

not tested up to 8 layers.

Cloth filter E.coli test method: The cloth was tied onto an 800-milliliter bottle (the bottle is sterilized), then the

105MPN/100mL E.coli solution was poured onto the cloth so that the outflow was collected in the sterilized bottle.

Then the outflow was taken out from the bottle and its E.coli concentration was tested. The E.coli removal of one-

layer, two-layer, four-layer and eight-layer cloth were tested respectively. Figure 4-11 shows the set up for cloth

E.coli LRV test.

Figure 4-11 Experiment Set-Up for Cloth E.coli LRV Test

Gravity Non-Electric (GNE) Filters Test Method

Flow rate, turbidity removal and lifetime test method: Before testing, the filter needed to run properly. The

procedure is described as follows:

(1) Check the manufacturer's instruction about the volume capacity of the top container. Define the container's

full level and half-full level and mark them.

(2) Run one volume of deionized water through the GNE filter, and that volume is determined by the volume of

the upper chamber of the given GNE filter. Note: If the manufacturer indicates another "break-in" method,

69

this should be followed.

(3) Empty the filter completely, including any parts and the bottom of the filter where the tap is located. Close

the filter's tap.

After accomplishing the above steps, testing begins. Because the tap of each filter is not installed at exactly the

bottom of the storage container, there is a small volume of water always stored in the filter safe storage container.

This means some amount of water has to be filtered before the clean water leaves the tap. The following procedures

are used to address this issue:

(1) Take three samples of the challenge test water for each test and check their turbidity.

(2) Fill the filter to its half-full level with the challenge test water and open the tap.

(3) Start the timer once the clean water starts coming out of the tap.

(4) To compensate for the water stored in the filter, immediately add back the water to the half-full level in the

filter and document this amount of water on the data sheet.

(5) Collect the outflow using a bottle with a volume of about 800 milliliters.

(6) When clean water in the bottle reaches 400 millimeters, stop the timer and then calculate the flow rate and

test its turbidity.

(7) Refill the filter back to its full capacity and let the water flow from the faucet until the filter is empty.

NOTE: It is not necessary to constantly monitor the level in the GNE filter’s reservoir as it empties.

For the subsequent tests, the procedure is simplified:

(1) Take three samples of the challenge test water for each test and check their turbidity.

(2) Fill the filter to its half-full level with the challenge water and open the faucet.

(3) Once the filter starts to produce outflow, start the timer and collect the outflow using a bottle with a volume

of about 800 milliliters.

(4) When clean water in the bottle reaches 400 millimeters, stop the timer and then calculate the flow rate and

test its turbidity. Document the total volume of test water that has been filtered by this filter.

(5) Refill the filter back to its full capacity and let the water flow from the faucet until the filter is empty.

(6) With every second refill of the filter (i.e. 3rd, 5th, 7th, etc.), repeat the above steps until the flow rate of clean

water from the filter is below one liter per hour. At this point, this filter clogs and the documented total

volume of water filtered is its lifespan.

E.coli removal test method:

The author monitored E.coli removal at the beginning and the end of the filter’s life. The filters were half-filled

with the 105MPN/100mL E.coli solution. Then we collected three samples of the outflow and tested the E.coli

concentration using the IDEXX Quanti-tray method described in Section 4.3.3.

70

Reverse Osmosis (RO) Filters Test Method

Flow rate, turbidity removal, TDS removal and lifetime test method: The flow rate of RO filters was

monitored and recorded by the CR-V2.0 test rig. The measurement and data analysis methods have already been

shown in Section 4.4.2. Turbidity and TDS removal were measured at the beginning, middle, and end-of-life of

each RO filter for its inlet, clean water, and wastewater respectively. When the flow rate dropped to less than

half of the initial flow rate, it was determined that that would be defined as the end-of-life of the filter. In

fact, once the RO filter’s flow rate started to decrease, it fell rapidly (in 4 to 5 hours) to below 80 ml/min, which

was not detectable by the flow meter we were using.

E.coli removal test method: It proved to be very difficult to provide a consistent and stable level of E.coli

introduced into the RO filters. The E.coli injection process introduced in Section 4.4.3 gives a continual dilution

of the initial, concentrated E.coli broth. A well-functional RO membrane should have no E.coli in the clean water

flow. Thus, we would need to measure the peak E.coli at the inlet to measure the exact E.coli removal rate. In

order to determine when the inlet concentration reaches its peak, many water samples at the inlet need to be taken.

For instance, Figure 4-12 is an illustration of the expected E.coli concentration for inlet, waste & clean water

versus time, assuming that the pre-filters before the RO membrane do not remove E.coli.

Figure 4-12 Depiction of E.coli Concentration for Inlet, Waste & Clean Water vs. Time

The blue line indicates the inlet E.coli concentration. Based on testing, the inlet E.coli concentration is expected

to reach its peak value at around 6 minutes, and then remain at this peak value for another 6 minutes when

simultaneously testing with two RO filters in parallel.

After about 12 minutes, the water pressure of the 6-gallon small pressure tank dropped from 60psi to 20psi and

ended the test. At that point, the valve from the pump was opened and deionized water flowed in. There was a

71

delay time of approximately 2 minutes for the inlet water to arrive at the RO membrane, thus, the E.coli

concentration of the wastewater reached its peak 2 minutes later. As long as the RO membrane is not broken, the

E.coli concentration of the clean water downstream from it was always expected to be zero because RO membrane

has a pore size much smaller than the size of E.coli and thus provided a good barrier to intercept the E.coli.

In order to obtain a longer run time to give the author more flexibility to measure the peak value of the inlet

concentration, only one RO filter was run at a time to maximize the time period that the E.coli stream was able to

stay at a constant delivery pressure of 15 psi.

The procedure is given below. Please refer to the piping diagram (Figure 4-8) to better understand the sequence

of valve closing and openings.

(1) Empty the mixing cartridge for the E.coli and the 6-gallon (23-liter) pressure storage tank. Close off the gate

valve from the drain water vessel receiving the clean/waste/inlet/ water lines. Place some bleach in that drain

water vessel.

(2) Close valves: V1, V2, V4, V5, and V6; Open valve: V3.

(3) Pour about one liter of the E.coli broth that has a concentration of around 109 MPN/100ml through a small

funnel into the open valve. Close V3. Clean off funnel and valve with alcohol.

(4) Run water pump until it reaches its peak pressure of about 60 psi.

(5) Open V1 until the pressure of the 6-gallon pressure storage tank also reaches 60 psi, then close V1. Clean

off end of the Waste/Clean/Inlet lines with alcohol.

(6) Turn on the power for RO filters’ pump and solenoids in RO#1 and #2. Open the automated test rig control

interface where a diagram of the solenoids is shown. Click to "open" solenoids 2, 4, 6, and 8.

(7) Start the test (Time=0): Either V4 or V5 is opened depending on which RO filter is being tested. Then open

V2 and stabilize the pressure regulator at 15 psi.

(8) Refer to Table 4-11 below for an example of the sampling timeline.

(9) After collecting all the water samples, continue to take flow data using the automated test rig.

(10) When the test of flow in Step #9 is complete, shutdown the RO filters’ pump and solenoids.

(11) Open the water valve for the drain water vessel. Clean up carefully to remove E.coli using the spray bleach

container. Clean carefully and flush the entrance to Valve 3 where the concentrated E.coli broth was inserted

into the injection hardware.

Ensure that the temporary storage container where the RO’s clean and waste water drain tubes empty, is open to

the floor drain. Clean up carefully to remove E.coli using the spray bleach container. Clean carefully and flush

72

the entrance to Valve 3 where the concentrated E.coli broth was inserted into the injection hardware.

Table 4-11 Sampling Timeline and Dilutions for RO E.coli Test

Sample Time Dilution Ratio 1 Dilution Ratio 2

Inlet

6min 1:103 1:105

10min 1:103 1:105

14min 1:103 1:105

18min 1:10 1:103

21min 1:10 1:103

36min 1:1 1:100

Waste

0min 1:1 -

7min 1:103 1:105

11min 1:103 1:105

15min 1:103 1:105

20min 1:10 1:103

25min 1:10 1:103

40min 1:1 -

Clean

0min 1:1 -

12min 1:1 -

16min 1:1 -

24min 1:1 -

73

5 Test Results

Cloth and Mesh Water Filters Test Results

Cloth: Only the “best quality,” Cloth #1, restricted the flow of water. Even with eight-layers of Cloth #2 and #3,

the water went through the cloth as quickly as it could be poured. Even for eight-layers of Cloth#1, the flow rate

was 50 L/hr, which was extremely high compared to the GNE filters and the RO filters, which had flow rates

ranging from 0.6 L/hr to 4 L/hr (GNE) and about 14 L/hr (RO). The standard for evaluating clean water flow rate

of the GNE filters is discussed in Section 4.1.4.

According to Figure 5-1 below, for Cloth #1, the turbidity removal and E.coli LRV both increased with the

increase in the number of cloth layers. When tested as one-layer or two-layers, it was only able to remove less

than 20% of the turbidity, whereas, normally one would hope to attain at least 99% removal. When it is folded 4

layers and 8 layers, it can remove 50% to 60% of the turbidity, and the E.coli LRV was around 0.1 to 0.12, which

means it can only remove 20% of the bacteria, which is an unacceptable performance as per Table 4-2 CR team

defined standard.

Figure 5-1 Cloth #1 E.coli LRV and Turbidity Removal vs. the Number of Cloth Layers

Figure 5-2 shows the E.coli LRV and turbidity removal (%) of eight layers of Cloth #1, #2, and #3. Cloth #2 and

Cloth #3 both had a lower removal than Cloth #1.

In general, eight layers of cloth can remove 20% to 60% of turbidity, but can hardly remove E.coli.

0

10

20

30

40

50

60

70

80

90

100

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0 2 4 6 8 10

Tu

rbid

ity R

emoval (%

)E

.coli

LR

V

Number of cloth layers

E.coli LRV

Turbidity Removal (%)

74

Figure 5-2 E.coli LRV and Turbidity Removal of Eight-Layers for Cloth#1, #2, and #3

The mesh was shown not to retain water. The challenge water went through the mesh as fast as it was poured.

Turbidity removal was only 5% with a two-layer mesh.

Gravity Non-Electric (GNE) Water Filters

The performance of the GNE filters is presented using the test attributes of: flow rate, E.coli LRV, and turbidity

removal versus the volume of water passing through the filter. The turbidity removal is given as the percent of

reduction. The E.coli removal is presented as the E.coli LRV. When the E.coli LRV is higher than 4, the result is

presented as “4” in the figures. In such cases, that filter is considered as “highly protective” for its ability to remove

bacteria (WHO, 2014a), and as “excellent” in the CR defined standard.

Espresso Water Filter Test Results

Per Figure 5-3, except for a very short period at the beginning, the Espresso Stainless Steel Water Container flow

rate was below 1 L/hr, our cutoff for end-of-life. The instructions indicated that the flow rate would increase after

15 days of use. But after 15 days, this flow rate remained well under 1 liter per hour. Further, rather than increasing

over time, its flow rate continued to decrease. In order to confirm this low flow rate of the Espresso, a second

sample was tested after being soaked in clean water for two days. This time, the initial flow rate never made it

above 1 L/hr.

However, the Expresso had excellent turbidity reduction and E.coli removal. The turbidity removal remained at

above 97% and it can remove more than 99.99% E.coli.

0

10

20

30

40

50

60

70

80

90

100

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

1 2 3

Tu

rbid

ity

Rem

ov

al(

%)

E.c

oli

LR

V

Cloth #

E.coli LRV

Turbidity Removal (%)

75

Figure 5-3 Espresso Water Filter Test Results

Tata Swach’s Christella Plus Test Results

Figure 5-4 shows the test results for the Tata Swach Christella Plus. The flow rates remained high and only

decreased gradually after filtering more than 150 liters of challenge water. After filtering 410 liters of water, the

flow rate was still higher than 4L/hr. However, the turbidity and E.coli removal was much lower compared to

other models. The E.coli LRV was from 1.5 to 2.0, meaning that it can only remove 95% to 99% of bacteria

(“poor” on Table 4-2). The turbidity reduction fluctuated and was generally within the range of 70% to 85%.

Figure 5-4 Tata Swach’s Christella Plus Test Results

76

Tata Swach’s Smart 1500 Liters Test Results

Figure 5-5 indicates that TATA Swach 1500 Liters still has a fairly fast flow rate of 2.0L/hr after filtering 380

liters of challenge water. Its predicted lifespan is 486 liters of challenge water. That might be due to the high

concentration of the challenge water we used. If the feed water has lower level of turbidity, the filter’s lifetime

should be longer. The highest flow rate was about 6L/hr. But, its turbidity removal fluctuated considerably from

84% to 95%. Its E.coli LRV was also relatively low, remaining at about 2, which means that it removes 99% of

bacteria (“poor” to “fair”).

Figure 5-5 Tata Swach’s Smart 1500 Liters Test Results

Tata Swach Smart 3000 Liters Test Results

The difference between Tata Swach’s Smart 1500 and 3000 Liter models is that they use different types of Tata

filtration bulbs. The highest flow rate of the Smart 3000 Liters was 3.6 L/hr, which was lower than the highest

flow rate of the Smart 1500 Liters, which was 6.3L/hr. The lifespan was 220 liters of water, which was shorter

than Smart 1500 Liters. This might be because Smart 3000 Liters has better performance in removing turbidity,

so it clogged much faster under such high concentration challenge water. The E.coli LRV for both models was

“protective” at around 2.5 according to WHO terminology and “Fair” according to CR defined terminology.

77

Figure 5-6 Tata Swach’s Smart 3000 Liters Test Results

Hindustan Unilever’s Pureit Classic 14L Test Results

As mentioned previously, the Pureit Classic 14L is a multi-stage filter and this design showed a significant

disadvantage in testing. The water first goes through a “Micro Fiber Mesh” -a particle removal stage- which

removes visible dust. And then the water passes a unique “Compact Carbon Trap” which removes dirt, parasites,

and pesticide residuals. After that is a “Germkill Processor” that kills almost all bacteria and viruses. Finally, the

water goes through a uniquely designed “Polisher” for removing residual chlorine.

But owing to the multi-stages, this filter could not provide a continuous clean flow. This proved frustrating during

testing as it would also likely be for consumers when could not get clean water when they wanted it.

To confirm this performance problem, a second sample of the Pureit was tested. The results are shown in Figure

5-7 and Figure 5-8. Both samples clogged quickly. The first fell below 1L/hr after filtering 60 liters of water and

the second one only after 30 liters.

The turbidity removal remained above 98% for both of them, and the E.coli removal remained higher than 99.99%,

so it had “good” E.coli removal.

78

Figure 5-7 Hindustan Unilever’s Pureit Classic 14L Test Results of Sample 1

Figure 5-8 Hindustan Unilever’s Pureit Classic 14L Test Results of Sample 2

Prestige’s LifeStraw Test Results

From Figure 5-9, the Prestige’s Life Straw is shown to have a disappointing short lifetime. It had a sharp drop-off

of clean water production after filtering only 40 liters of water. Its highest flow rate was relatively low at 1.9L/hr.

Nevertheless, it had an excellent turbidity removal of above 97.9% and its E.coli removal was “good” at 4.

79

Figure 5-9 Prestige’s Life Straw Test Results

Eureka Forbes’s AquaSure Amrit with Kitanu Magnet Test Results

The Amrit had one of the highest initial flow rates of 11 L/hr, but it dropped off quickly within the first 40 liters

of clean water production. After filtering 108 liters of challenge water, the flow rate was still 1.54 L/hr and the

predicted lifespan before the testing would likely have terminated was 117 liters. The turbidity removal was high

at around 99%. In the middle of its life, this AquaSure filter removed 99.9% of E.coli. Near its end-of-life, the

E.coli removed more than 99.99%, which is “good”. Overall, the AquaSure model did well in all three performance

categories.

Figure 5-10 Eureka Forbes’s AquaSure Amrit with Kitanu Magnet Test Results

80

KENT’s Gold UF Membrane Filter Test Results

In the instructions for the KENT Gold, it claimed that this filter had a capacity of 10 liters in the upper storage

tank. In reality, it only contains 7 liters so there was false advertising involved here. Additionally, its upper pre-

filter has a very slow filtering rate so that a consumer would find that pouring water in was quite time-consuming.

It was particularly difficult to add more than 5 liters of water to this filter at one time.

Thus, for this filter we had to change the test method to only use 5 liters of challenge water added each time. Thus,

the flow rates were tested where 2.5 liters of water was added according to the test method discussed in Section

4.5.2.

Per Figure 5-11 below, the filter’s highest flow rate was 5.7 L/hr, and it clogged after filtering about 55 liters of

challenge water. The turbidity removal between 97% and 99.5% was good. And the E.coli LRV was both greater

than 4 for a removal rate of at least 99.99% of bacteria.

Figure 5-11 KENT’s Gold UF Membrane Filter Test Results

Summary of the Performance of GNE Filters

Generally, GNE filters have much better performance than cloth and mesh filters. There are big differences among

GNE filters in flow rate, turbidity removal, lifespan and E.coli removal. In general, the higher the flow rate, the

lower the turbidity removal and the lower the E.coli removal.

Five models were able to remove higher than 95% turbidity during their lifetime. Three models can remove higher

than 99.99% E.coli, which is good, while the other four can remove higher than 99% (“fair”).

81

Reverse-Osmosis (RO) Water Filters

RO Turbidity and TDS Reduction

The turbidity and TDS reduction of the two RO filters (branded and local-assembled, Dolphin) were monitored at

their beginning and end-of-life. These results are shown in Table 5-1.

These results show that RO filters have excellent performance in removing turbidity and TDS: the turbidity

removal and TDS reduction are almost 100% for both RO filters. Thus, the waste water flow contained turbidity

and TDS at a significantly high, concentrated levels. In Table 5-1, “Clean % removal” refers to the percent

reduction of turbidity (or TDS) in the clean water compared to that in the inlet. And “Waste % removal” refers to

the percent reduction of turbidity (or TDS) in the wastewater compared to that in the inlet.

Table 5-1 RO Turbidity and TDS Removal Results

Model

Name Time

<<<< TURBIDITY Removal >>>> <<<< TDS Removal >>>>

Inlet

(NTU)

Clean

(NTU)

Clean %

removal

(%)

Waste

(NTU)

Waste %

removal

(%)

Inlet

(mg/L)

Clean

(mg/L)

Clean %

removal

(%)

Waste

(mg/L)

Waste %

removal

(%)

Tata

Swach

Platina

10 hours: 37.7 0.19 -99.5% 0.96 -97.5% 1670 30 -98.2% 2686 60.8%

End-of-life: -99.5% -98.7%

Clean

Water

Dolphin

10 hours: 37.7 0.18 -99.6% 2.09 -95.5% 1670 18 -98.9% 2520 60.5%

End-of-life: -99.7% -99.7%

Dolphin

Gold <10 hours: 46 0.99 -97.8% 4.33 -90.6% 1570 46 -97.1% 2580 64.3%

RO E.coli Removal

Similarly, both RO filters were seen to be excellent in E.coli removal.

From Figure 5-12 and Table 5-2, the E.coli concentration at the inlet of Clean Water RO filter is much higher than

that from the wastewater for the time 6 to 14 minutes. This verifies that the pre-filters of the Clean Water filter

may have removed bacteria so when this filter’s water enters the RO membrane its E.coli concentration has already

been reduced significantly. In general, the Clean Water Dolphin RO had a LRV for its clean water as high as 6,

which is two LRV’s greater than WHO’s “highly protective” category. Hence, this is excellent pathogen removal

performance.

82

Figure 5-12 E.coli Removal Results for Clean Water (Non-branded Dolphin RO-type) Filter

Table 5-2 E.coli Removal Results for Clean Water (Non-branded Dolphin RO-type) Filter

Sample Time Dilution Ratio

1 Result

(MPN/100ml) Dilution Ratio

2 Result

(MPN/100ml) Average

Inlet

6min 1:103 >2.4×106 1:105 1.5×107 1.5×107

10min 1:103 >2.4×106 1:105 1.6×107 1.6×107

14min 1:103 >2.4×106 1:105 1.9×107 1.9×107

18min 1:10 >2.4×104 1:103 1.4×106 1.4×106

21min 1:10 1.5×104 1:103 2.4×104 2.0×104

36min 1:1 >2.4×103 1:100 2.9×104 2.9×104

Waste

0min 1:1 0 - - 0

7min 1:103 1.3×106 1:105 1.8×106 1.51.3×105

11min 1:103 2.0×106 1:105 2.0×106 2.0×106

15min 1:103 >2.4×106 1:105 3.1×106 3.1×106

20min 1:10 8.5×105 1:103 9.8×105 9.1×105

25min 1:10 >2.4×104 1:103 1.3×105 1.3×105

40min 1:1 >2.4×103 - - -

Clean

0min 1:1 0 - - 0

12min 1:1 0 - - 0

16min 1:1 0 - - 0

24min 1:1 0 - - 0

Table 5-3 below shows the E.coli results for the Tata Swach Platina RO filter. The same testing method was used

here as above for the Clean Water Dolphin filter. But, we can see that the E.coli concentration didn’t reach its

peak at 6 minutes. This may be due to this filter having a higher flow resistance at this time so that its flow rate is

83

lower than the Clean Water Dolphin. Also unexpected was that the waste water had E.coli higher than

2419.6MPN/100mL. This may mean that the silver impregnated pre-filter had removed bacteria. Although we

weren’t able to measure the E.coli concentration in the waste water, these results indicate that the clean water

LRV of the Tata Swach Platina is as high as 6 (“excellent” according to CR team’s terminology). This is similar

to the results found for the non-branded, Clean Water Dolphin RO filter.

Table 5-3 E.coli Removal Results for Tata Swach Platina Silver (Branded RO-type) Filter

Sample Time Dilution Ratio 1 Result

(MPN/100ml) Dilution Ratio 2

Result

(MPN/100ml)

Inlet 6min 1:103 2.0×104 1:105 -

10min 1:103 >2.4×106 1:105 3.6×106

Waste

0min 1:1 0 - -

14min 1:1 >2.4×103 - -

18min 1:1 >2.4×103 - -

Clean

0min 1:1 0 - -

15min 1:1 0 - -

19min 1:1 0 - -

RO Flow Test Results

Figure 5-13 below shows the test results for the Clean Water Dolphin RO filter. Its end-of-life occurred at 2750

minutes (45.8 hours) when the clean water production became less than 100mL/min (which is a RO filters end-

of-life mentioned in Section 4.1.6). At this point, it had been tested for over 54 hours. The clean flow showed a

significant fall off after which the waste flow also decreased. It is likely that the high level of test dust that created

the artificial turbidity for the testing was filtered and accumulated in the cartridges, resulting in an increase in the

filter’s flow resistance.

84

Figure 5-13 Clean Water (Non-branded Dolphin RO-type) Filter Flow Performance

As seen in Figure 5-14, the Tata Swach Platina, had a significantly shorter life compared with Clean Water

Dolphin Model. This filter showed a major reduction in clean water production at approximately 1700 minutes

(29 hours) after filtering around 1450 liters of challenge water in contrast to the Clean Water Dolphin model which

had a lifetime of 2750 minutes. The flow rate of the waste water also showed the same trend.

Figure 5-14 Tata Swach Platina RO Filter Flow Performance

A new set of filter cartridges upstream from the RO membrane was replaced at this point, but its RO membrane

and Post-Carbon cartridge remained the same ones as before. This was done since this Tata RO filter did not

require a large, pre-filter mounted outside of its main cabinet like the Clean Water Dolphin does. Thus, the

0

100

200

300

400

500

600

700

800

0 500 1000 1500 2000 2500 3000 3500

Time/min

Ratio W/C(%)

Clean Flow (ml/min)

Waste Flow (ml/min)

0

100

200

300

400

500

600

700

800

0 500 1000 1500 2000 2500 3000 3500

Time(min)

ration W/C (%)

Clean Flow (ml/min)

Waste Flow (ml/min)

Ratio W/C (%)

54hr

100ml/min

End-of-life

100ml/min

85

conjecture was that its pre-filters had clogged due to the high level of turbidity at the Tata filter’s inlet. But,

surprisingly, the filter flow increased initially and then quickly fell off. As a result, it is believed that the Tata’s

end-of-life was due to a blocked RO membrane, and not a blocked pre-filter.

In general, RO filters were shown to be quite effective in reducing the high turbidity (40 NTU), very substantial

TDS (1500 mg/L) and high levels of E.coli bacteria (105 to 106 MPN [or “Most Probable Number”]/100 ml) in the

challenge water. The RO Dolphin (Clean Water model) filter was seen to be as comparable in its performance to

the branded, three times higher cost. Tata Swach Platina RO model. This result is very good news for the middle

to lower class Indian consumer.

86

6 Ratings Chart and Discussion

The performance testing results of each of the three categories of filters will be presented in this chapter. The

comparative results of GNE filters will be shown in a Consumer Report-style ratings chart, which is introduced

in the Literature Review and Appendix D. Reverse osmosis (RO) filters have waste flow, which GNE filters don’t

have, so a separate ratings chart for reverse osmosis filters was developed whereby two models were tested for

their lifetime.

A ratings chart for cloth and mesh filters was not appropriate because their turbidity and E.coli removal are very

limited. Those results are given in paragraph and table form below.

Cloth and Mesh Type Filters

As a low-cost method used in Indian households, the cloth and mesh type filters were found to have a limited

ability to provide clean water. Colwell et al. (2003) found in Bangladeshi that local sari cloth folded at least four

times can effectively remove higher than 99% of V. cholera attached to plankton. According to their research, the

local sari folded 4 to 8 times has a pore size of about 20µm, so it can remove most of the V. cholera attached to

plankton that are mostly bigger than 20µm. But in this study, since the E.coli used has a size of 0.5~3 µm, while

the cloth we tested has a pore size of 100~300 µm, and the mesh a size of about 300 µm, cloth and mesh filters

have very limited performances in reducing pathogens. Turbidity reduction is possible, but is also very limited.

As shown in Appendix A, the fine dust we used has a particle size from 1 to 100µm, so the turbidity reduction of

the cloth filters is limited even with 4 to 8 layers. Nonetheless, research was done to determine how best to utilize

these filter elements that have a broad reach to the “bottom of the pyramid” families in India. Due to the limited

effectiveness of these products, their comparative performances are only shown in Each RO water filter takes

significant time to perform end-of-life testing, so two of the RO filter models were tested during the two-month

period. To better represent the performance of this filter category in limited two month period of time available

for testing, two important model-types—a high and a low-price model— were tested and the results are shown in

a separate CR-Style comparative ratings chart separate from GNE filters. Despite of the small number (two) of

models tested, the results are still instructive and are one of the highlights of this research. However, it should be

noted that since only one sample was tested for each model, further testing should be done to confirm the results

before making any broad conclusions.

Before looking at the separate ratings charts for GNE filters and RO filters, the following summary about the

performance of each filter category can be made:

Table 6-1 and not put into a Consumer Report-style ratings chart.

Gravity Non-electric (GNE) Filters

For GNE filters, in order to discuss and compare the performance of the different models, a comparative ratings

87

chart is presented. However, due to time constraints, only one filter sample was tested for eight models. For one

model, two test samples of the same filter model have been tested to ensure that the results were able to truly

reflect the model’s performance. As a comparison, CR generally takes a year to perform a comprehensive product

evaluation and our contract to stay in CR’s lab was only two months. For GNE filters, each attribute was scored

and an overall score for each model was computed.

Reverse Osmosis (RO) Filters

Each RO water filter takes significant time to perform end-of-life testing, so two of the RO filter models were

tested during the two-month period. To better represent the performance of this filter category in limited two

month period of time available for testing, two important model-types—a high and a low-price model— were

tested and the results are shown in a separate CR-Style comparative ratings chart separate from GNE filters.

Despite of the small number (two) of models tested, the results are still instructive and are one of the highlights

of this research. However, it should be noted that since only one sample was tested for each model, further testing

should be done to confirm the results before making any broad conclusions.

Before looking at the separate ratings charts for GNE filters and RO filters, the following summary about the

performance of each filter category can be made:

Table 6-1 The Summary of the Performance of each Filter Category

Filter Category

E.coli Removal Turbidity Removal TDS Removal Flow Rate

Cloth & Mesh Unacceptable Unacceptable Unacceptable Excellent

GNE Filter Poor to Excellent Poor to Excellent Unacceptable Unacceptable to

Poor

RO Filter Excellent Excellent Excellent Excellent

Additionally, Table 6-2 below compares the filter categories for their cost (purchase and operating), lifetime and

impacts on the environment based on % Recovery of RO units:

88

Table 6-2 Comparison between Filter Categories for the Cost, Lifetime, and Environmental Factors

Filter Category Purchase Cost

($) Operating Cost

($) Life of Filter

Elements Environmental

Factors

Cloth & Mesh Very low Very low Long with washing None

GNE Filter Moderate Moderate Short to long None

RO Filter High to very high High20 Long Significant waste water

Overview of Comparative Ratings Chart

Consumer Reports-style comparative ratings charts for the GNE filters and RO filters are provided to highlight

the test results in this research so that international agencies and consumers can make more informed water

treatment decisions. The author changed the blob keys used by CR:

to icons below in order to be consistent with S2 and S3 Team’s Ratings charts and readable in black and white.

Apart for that minor change, we followed Consumer Reports-style ratings chart to create our ratings charts.

The two separate comparative ratings charts for GNE and RO filters presented in this chapter are examples of

decision tool for customers when purchasing water filters. These ratings charts were developed following a

Consumer Reports-style of evaluation21: they differentiate between various filter models found on the marketplace

in Ahmedabad based on their key attributes and features. Each attribute was tested in a customized laboratory

designed through the collaboration between Massachusetts Institute of Technology researchers and Dr. Jeffery

Asher. The development of each attributes was described in Section 4.1.

20 RO operating cost can be very high if the cost of water is taken into account.

21 Based on feedback from development professionals and users in developing countries, the method of communicating

future evaluation results will likely change.

89

错误!书签自引用无效。 shows the CR-Style comparative ratings chart for GNE water filters. To develop this chart, there are several key concepts

– scores, weightings, and ratings – that need to be introduced.

Table 6-3 Example of a Comparative Ratings Chart22

22 http://www.mattmcgee.com/consumer-reports-cell-phone-service-ratings/

90

Scores, Weightings and Ratings

This section introduces the key concepts and how attribute scores and weightings are developed in creating

a comparative ratings chart. The detailed scoring, weighting and rating method will be described in

Appendix D.

“Score” can be seen as the numerical ‘grade’ given to each attribute (clean flow rate, turbidity reduction,

etc.) or feature category for each model. The performance of each model for each test attribute determines

the raw score. This score is placed on a scale ranging from 0.50 to 5.49 based on a linear best fit tied to the

standard deviation of the set scores (see Table Appendix D.1.2). The scores are then translated into graphical

icons found in the rating matrix as shown in Table 6-5.

“Weighting” is the level of significance given to each attribute in order to compute a composite or overall

rating for the attributes. The weightings are determined according to the author’s judgment in combination

with Dr. Asher. For the example of the GNE filters, the attribute weightings are shown in Table 6-4. For E.

coli removal, since the pathogen appears only periodically in the feed water, it is weighted at only 10%.

The clean water flow rate is weighted at 50%, which is to say, it is weighed as a very important factor

because if there is not satisfactory flow the consumer will not use the filter. In areas where there is

considerable pathogens in the inlet water, this weighting would likely be increased since this attribute’s

priority would be high.

Table 6-4 Weightings for Each Attributes of GNE Filters

Attribute E.coli LRV Turbidity

Removal TDS Removal Flow Rate

TDS

Removal Lifetime

Weighting (%) 10 20 0 50 0 20

“Rating” is the overall weighted sum of the attributes and feature scores for each model or “overall score,”

is scaled in a range between 0 and 100.

The “overall score “presents the overall assessment for each model in a concise way to help consumers and

agencies make better informed purchasing decisions.

Attributes and Features Shown in the Ratings Charts

Different purchaser would have different priorities for filters’ performances. The attributes and features that

are shown in the ratings charts are interpreted below. Consumers or agencies can change these weighting

to fit their own priorities.

E.coli Removal

91

Metric: Log removal value (LRV) of E.coli colonies

Scoring: Higher log removal value is favorable

This test assesses the ability of the water filters to remove the high levels of E.coli bacteria in the inlet. The

E.coli levels are measured in the inflow and outflow clean water for all filter categories (and, in addition,

waste water for RO filters). The scoring is for the clean water % removal effectiveness. This LRV

measurement is done at the beginning, mid- and end-of- life of each filter sample.

Turbidity Removal

Metric: Level of turbidity % removal

Scoring: Higher % removal of turbidity in clean water is favorable

Turbidity (created using A2 dust) was added to the water in the 100-gallon water tanks and monitored to

keep it at approximately 40NTU. Measurements were done periodically throughout the lifetime of the filter

to determine the reduction of turbidity in the clean water for each model as well as its increase in the waste

water for the RO filter. The score is based on clean water turbidity removal.

Total Dissolved Solids (TDS) removal

Metric: Level of TDS % removal

Scoring: Higher % removal in TDS in clean water is favorable

TDS (made from Epsom Salts) was added to the water in the 100-gallon water tanks and monitored to keep

it at approximately 1500 mg/L. Measurements were done periodically throughout the lifetime of the filter

to determine the reduction of TDS in the clean water for each model as well as its increase in the waste

water for the RO filter. The score is based on clean water TDS removal.

Clean (“Product”) Water Flow Rate

Metric: Clean water flow rate

Scoring: Higher flow rate is favorable

The clean water flow rates were measured throughout the lifetime of each filter sample. The scoring for the

clean water flow rate was different for each filter category—Particle removal, GNE and Reverse Osmosis

(RO) models because of their large range of flow rates.

RO % Recovery

Metric: % Recovery

92

Scoring: Higher % recovery is favorable

For RO filters only, there is a copious amount of wastewater created compared to clean water. This is

typically in the ratio of 3:1. A high wastewater flow rate means that this precious resource—water—is being

wasted. Thus, a higher % recovery is more environmentally friendly and can save the user money if there

is a fee for each gallon used.

Clogging/Filter Lifetime

Metric: Number of hours the filter operates before its clean water flow rate is below 1L/hr for GNE filters,

and below 100ml/min for RO filters

Scoring: Longer filter lifetime is favorable

The clean water flow rate diminishes with time as the filter is used. Clogging measures how well the filter

retains its flow rate of clean water over time and how often the users need to change the filter or cartridge.

A longer lifetime means less filter maintenance and less cost for replacement.

Gravity Non-Electric (GNE) Filter Comparative Ratings Chart

Table 6-5 is the comparative product ratings chart showing the relative performance of nine GNE filter

models.

It can be seen from this ratings chart that none of these filters have an excellent performance for all

attributes. Rather, a model is seen as good in one performance aspect, but falls down in another. Take the

Christella Plus filter made by Tata Swach as an example: it is a good buy at $29 (between the lowest price

of $17 and the highest price of $58) with the second highest flow rate of 3.96 L/hr) and longest life, but

poor in E.coli removal. So if you have water without microbial contamination such as a reliable piped water

supply for a reliable municipal source, this would be the best choice. However, if bacteria is a concern in

the water sources, then the Purerit by Hindustan Unilever at $26 may be the answer because it has a good

E.coli removal and its flow rate was the best compared to other models that had good E.coli removal.

Reverse-Osmosis (RO) Comparative Ratings Chart

Two reverse-osmosis filters were tested for their product lifetime over a period of about three weeks using

the RO flow test rig. Table 6-6 is the comparative product ratings chart of those results.

Generally, RO filters are quite effective in cleaning our “challenge water” that had very high levels of E.coli

bacteria (105 to 106 MPN/100 ml), turbidity (40 NTU) and TDS (1500 mg/L). Based on these lifespan tests,

the RO dolphin-type filter (Clean Water model) was as effective as a branded RO filter (Tata Swach’s

Platina model) that was three times more expensive. These results are very good news for the consumer.

93

However, RO filters in general represent a poor tradeoff for the environment because they generate a large

amount of wastewater during operation.

In the marketplace in India, there are a number of dolphin-type RO models available for households. From

a close look at the inside structure of the three models from different manufacturers, they all have a same

basic design that is similar to the high-priced branded RO filters. But, it was evident that some features

such as backwash were eliminated in the dolphin-type filters and cheaper materials were used inside. This

suggested that while the Dolphin filter performed well initially, this performance might not be able to be

sustained. Nonetheless, in terms of the performance of the Clean Water ($100) that was extensively tested,

Dolphins did as well if not better than the Tata Swach Platina ($300). Future research should be done with

more test samples and more filter models in both price ranges to confirm these findings.

94

Table 6-5 Comparative Ratings of Gravity Non-Electric (Non-Electric/Gravity) Water Filters

Note:

[1] – Gravity Non-Electric filters as a product category cannot reduce TDS

[2] – Multiple samples never met the minimum flow rate of 1 liters per minute

95

Table 6-6 Reverse Osmosis (Electric) Comparative Rating

96

Discussion of the Ratings Charts

The ratings charts for GNE and RO filters were developed to enhance the ability of the consumer or agencies

to make informed purchasing choices. The ratings charts include the necessary technical information

required to make these decisions. Non-technical factors were not assessed in this thesis, such as the cultural

acceptability of water filters, the maintenance, warranty, and the accessibility/customer support offered by

the manufacturers. However, they were assessed by the CITE Sustainability (S3) team23.

These ratings charts show the variations in performance of a water filter model for competing products

across a number of attribute categories. Consumers can see from the charts the top performers. In addition,

the ratings charts show the “value for money”, i.e., the relative performance increase as a function of price.

This helps consumers make purchasing decisions that balance the trade-off between performances and

price. The users of the charts may choose to select a model by the highest overall score or by critical factors

relating to their specific situation and requirements.

23 For CITE team reports of the water filter evaluation, see https://cite.mit.edu/.

97

7 Conclusions and Recommendations

Conclusions

Cloth and Mesh Filters

(1) Cloth filters had limited effectiveness in reducing turbidity. Folding the best cloth model four times

did reduce turbidity by about 60%. Cloth filters had no impact on removing E.coli bacteria or TDS.

(2) Jali Mesh type filters had little positive effect in cleaning our “challenge” water.

Gravity Non-Electric Filters

(1) There are large performance differences among the GNE filters tested in terms of flow rate, turbidity

removal, lifespan and E.coli removal. In general, the higher the flow rate, the lower the turbidity

removal and the lower the E.coli removal.

(2) Five models were able to remove higher than 95% turbidity during their lifetime. Three models can

remove higher than 4 log removal E.coli, while the others can remove higher than 2 log removal.

(3) There were significant leaks found in multiple samples of one Gravity Non-Electric filter, the

Everpure “Unbreakable,” and it was deemed a “Don’t Buy” (see bottom of Table 6-5).

Reverse Osmosis Filters

(1) Test methods and two special test rigs were developed that made these water filter product evaluations

more cost-effective. These test rigs modifications and inventions included: a) a system to inject the

pathogens directly into the filter inlet, and b) an automated rig to test multiple RO filters at the same

time.

(2) Reverse Osmosis filters were shown to be quite effective in reducing high levels of contaminants from

our “challenge water” that included E.coli bacteria (105 to 106 MPN/100 ml), turbidity (40 NTU) and

TDS (1500 mg/L).

(3) Based on these life tests of Reverse Osmosis models (limited to using one test sample due to time

constraints), the RO Dolphin (Clean Water model) filter was seen to be as effective as a branded RO

model (Tata Swach Platina model) at three times the cost. This result is very good news for the middle

to lower class Indian consumer seeking to own a “high end” reverse osmosis filter.

(4) Based on the testing done, the RO Dolphin filters that appear to be knockoffs of major brands were

seen to be as effective as their RO branded brothers that were three times more expensive.

98

Recommendations for Future Research

There were limitations in this research that future studies may want to address. The most important one was

the short time available for this research: there was only two months available to use the water filter lab at

Consumer Reports, because CR needed to conduct their next water filter evaluation in this lab directly after

we left. Thus, we could not test multiple samples of the filters. As it was, we performed “proof of concept”

testing through a two-month period. This limited time only allowed testing one test sample for each model.

CR usually evaluates at least three test samples per model.

The biggest challenge turned out to be a positive development. When we were designing our experiment

protocol, we intended to add E.coli to the 100-gallon (378-liter) tank and create a 105MPN/100mL

concentration every day. But this would have resulted in large costs to procure phosphate buffered saline.

Not finding a low-cost replacement for PBS, we developed an injection system where E.coli removal was

monitored at the beginning and end of the filter’s life.

This E.coli injection technique allowed the team to easily produce high E.coli levels at the Reverse Osmosis

filter’s inlet for up to 15 minutes.

Another advantage of using this E.coli injection technique was that the use of deionized water as the base

for our “challenge water” was no longer necessary in the large quantities we initially required. The chlorine

in the municipal water, which would “kill” the bacteria, was no longer of concern except when we were

testing the E.coli removal of the filters at their beginning and end of life. Thus, it is suggested that municipal

water be used during filter lifetime testing, so that the time producing a large amount of deionized water

can be saved. Going one step further, future researchers might consider developing a system where the

“challenge” ingredients—TDS and turbidity—be added or injected continually into municipal water thus

avoiding the need for large water storage containers. In addition, except when conducting the E.coli

removal tests, it would be a good idea to recirculate the “challenge water” to the 100-gallon tank, rather

having it go down the drain after each test. For reverse osmosis filters evaluations where the wastewater

contains concentrated TDS (about 2600 mg/L), this will save considerably on the use of Epsom salts. Thus,

recirculating the “challenge water” will save not only a large amount of water but also time in processing

the water to arrive at the correct contaminant levels.

Our future plan is to test more water filter samples in a newly created, water filter lab at MIT and/or in

India. This should be technically straightforward since it will be based on our successful CR test setup and

protocols. The current research has provided a strong beginning in what could someday be a MIT/India lab

where water filters are tested and this style of Consumer Reports-style of information be more widely

99

available in India and other emerging economies or even developed. That will have a major global impact

to provide clean water to the billions without access to safe drinking water.

100

APPENDIX A ISO 12103-1/CD, A2 Fine Test Dust Specification

101

APPENDIX B Parts List

Table B.1 provides a listings of the critical parts required for the water filter product testing set up and for

potentially scaling-up the number of RO filters that can be simultaneously evaluated in the future.

Table Appendix B.1 – Plumbing and Electrical Parts List

Plumbing Parts List

Item Quantity Manufacture Model Number Unit $ Total $

1 100-Gallon (378-liter)

Tank 3 Ace/ DenHartog NSF-61 $152.99 $458.97

2 Stainless Steel Circulator

1/8HP 3 Taco 0014-SF1 $404.95 $1,214.85

3 1" Stainless Steel Freedom

Flange (pair) 3 Taco 110-252SF $27.35 $82.05

4 Thermoplastic Shallow

Well Jet Pump 3/4 HP 1 Flotec FP4022-10 $335.82 $335.82

5 Precharged Water Tank 1 Dayton 3GVU1 $560.00 $560.00

6 Pressure Switch 1 Furnas 69WEC $88.95 $88.95

Electronics Parts List to scale-up simultaneous

Item Quantity Manufacture Model Number Unit $$ Total $$

1 Solenoid Valve 8 DEMAG DEMA 41-9-5 $23.13 $185.04

2 NI CompactDAQ 4-Slot

USB Chassis 1

National Instruments

(NI) NI cDAQ-9174 $777.00 $777.00

3 8 Channel Solid State

Relay Module 1 National Instruments NI 9485 $616.00 $616.00

4 8 Channel, 500 kS/s

Voltage Module 1 National Instruments NI 9201 $3,708.00 $3,708.00

NOTES:

1) Requires a paid LabView license

2) One flowmeter and one pressure transducer have already been purchased by MIT.

3) For each NI "8 channel" system, four flows can be measured. Two RO filters require 8 channels

for testing filters simultaneously.

4) We also need a Windows laptop to install the LabView and National Instruments (NI) software.

The CR-version of this software will be provided to MIT, but MIT staff will need to become

conversant with it for future software changes.

5) This parts list does not include miscellaneous metal and PVC pipes/valves/fittings

102

APPENDIX C RO Flow Testing Sample Data

Table C.1 RO Flow Testing -- Measurement Sequencing

RO Flow Run 13 _ Models 32 and 23 -- 2_120_240_124_0818

SOLENOID KEY: "0" Closed and "1" Opened

Cycle Solenoid 1 Solenoid 2 Solenoid 3 Solenoid 4 Solenoid 5 Solenoid 6 Solenoid 7 Solenoid 8 Time (Sec)

ALL: No measurement --

Cycle #1 0 1 0 1 0 1 0 1 240

RO#1: Clean 1 0 0 1 0 1 0 1 120

RO#1: Waste 0 1 1 0 0 1 0 1 120

RO#2: Clean 0 1 0 1 1 0 0 1 120

RO#2: Waste 0 1 0 1 0 1 1 0 120

ALL: No measurement --

Cycle #2 0 1 0 1 0 1 0 1 240

RO#1: Clean 1 0 0 1 0 1 0 1 120

RO#1: Waste 0 1 1 0 0 1 0 1 120

RO#2: Clean 0 1 0 1 1 0 0 1 120

RO#2: Waste 0 1 0 1 0 1 1 0 120

ALL: No measurement --

Cycle #3 0 1 0 1 0 1 0 1 240

RO#1: Clean 1 0 0 1 0 1 0 1 120

RO#1: Waste 0 1 1 0 0 1 0 1 120

103

RO#2: Clean 0 1 0 1 1 0 0 1 120

RO#2: Waste 0 1 0 1 0 1 1 0 120

|

ALL: No measurement --

Cycle #9 0 1 0 1 0 1 0 1 240

RO#1: Clean 1 0 0 1 0 1 0 1 120

RO#1: Waste 0 1 1 0 0 1 0 1 120

RO#2: Clean 0 1 0 1 1 0 0 1 120

RO#2: Waste 0 1 0 1 0 1 1 0 120

ALL: No measurement --

Cycle #10 0 1 0 1 0 1 0 1 240

RO#1: Clean 1 0 0 1 0 1 0 1 120

RO#1: Waste 0 1 1 0 0 1 0 1 120

RO#2: Clean 0 1 0 1 1 0 0 1 120

RO#2: Waste 0 1 0 1 0 1 1 0 120

ALL: No measurement --

END 0 1 0 1 0 1 0 1 240

Total (minutes) = 124

104

Table C.2 RO Flow Testing -- Flow Measurements

RO Flow Run 13 _ Models 32 and 23 -- 2_120_240_124_0818

Measured

RO#1-

Clean

Row# Date Time Pressure

PSI

Flow

mL/min

Total

mL

Solenoid

1

Solenoid 2 Solenoid 3 Solenoid 4 Solenoid 5 Solenoid 6 Solenoid 7 Solenoid 8

1 14-8-

18

8:23:37

AM 0.8217 1.4980 0.0279 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

2 14-8-

18

8:23:39

AM 0.7673 1.4758 0.0759 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

3 14-8-

18

8:23:41

AM 0.7818 1.5106 0.1254 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

4 14-8-

18

8:23:43

AM 0.7552 1.4253 0.1737 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

5 14-8-

18

8:23:45

AM 0.8439 1.4960 0.2227 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

6 14-8-

18

8:23:47

AM 0.8540 1.4798 0.2712 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

7 14-8-

18

8:23:49

AM 0.8359 1.5141 0.3203 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

8 14-8-

18

8:23:51

AM 0.7711 1.5057 0.3681 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

9 14-8-

18

8:23:53

AM 0.8001 1.4956 0.4177 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

|

ETC

|

47 14-8-

18

8:25:10

AM 0.7324 1.4715 2.2988 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

48 14-8-

18

8:25:12

AM 0.7289 1.4879 2.3482 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

49 14-8-

18

8:25:14

AM 0.7108 1.5182 2.3988 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

105

50 14-8-

18

8:25:16

AM 0.7039 1.5690 2.4472 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

51 14-8-

18

8:25:18

AM 0.7197 1.4819 2.4975 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

52 14-8-

18

8:25:20

AM 0.6788 1.5044 2.5488 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

53 14-8-

18

8:25:22

AM 0.5776 1.5807 2.6007 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

54 14-8-

18

8:25:25

AM 0.6341 1.5303 2.6538 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

55 14-8-

18

8:25:27

AM 0.5430 1.5085 2.7053 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

56 14-8-

18

8:25:29

AM 0.5493 1.5384 2.7556 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

57 14-8-

18

8:25:31

AM 0.5744 1.5334 2.8068 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

58 14-8-

18

8:25:33

AM 0.5356 1.5690 2.8586 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

59 14-8-

18

8:25:35

AM 0.6260 1.5424 2.9102 TRUE FALSE FALSE TRUE FALSE TRUE FALSE TRUE

Measured

RO#1-

Waste

Date Time Pressure

PSI

Flow

mL/min

Total

mL

Solenoid

1

Solenoid 2 Solenoid 3 Solenoid 4 Solenoid 5 Solenoid 6 Solenoid 7 Solenoid 8

61 14-8-

18

8:25:37

AM 0.7370 4.1479 0.1074 FALSE TRUE TRUE FALSE FALSE TRUE FALSE TRUE

62 14-8-

18

8:25:39

AM 0.8520 4.6081 0.2572 FALSE TRUE TRUE FALSE FALSE TRUE FALSE TRUE

63 14-8-

18

8:25:41

AM 0.7962 4.5511 0.4099 FALSE TRUE TRUE FALSE FALSE TRUE FALSE TRUE

64 14-8-

18

8:25:43

AM 0.8096 4.4870 0.5570 FALSE TRUE TRUE FALSE FALSE TRUE FALSE TRUE

65 14-8-

18

8:25:45

AM 0.7692 4.4471 0.7060 FALSE TRUE TRUE FALSE FALSE TRUE FALSE TRUE

66 14-8-

18

8:25:47

AM 0.7989 4.4827 0.8541 FALSE TRUE TRUE FALSE FALSE TRUE FALSE TRUE

106

|

ETC

107

Table C.3 RO Flow Testing -- Data Analysis

RO Flow Run 13 _ Models 32 and 23 -- 2_120_240_124_0818 Slope >> 0.0102 0.4369 << Intercept

Measure-

ment #

Ratio

W/C #1

(%)

Clean

RO#1 Flow

(ml/min)

Waste

RO#1 Flow

(ml/min)

Ratio

W/C #2

(%)

Clean RO#2

Flow

(ml/min)

Waste RO#2

Flow

(ml/min)

FLOW MEASUREMENT (volts) >>>>>>>>>>>

Clean RO#1

Flow

Waste RO#1

Flow

Clean RO#2

Flow

Waste RO#2

Flow

1 260% 189 492 275% 155 428 1.4938 4.5844 1.1482 3.9239

2 300% 207 622 213% 261 558 1.6786 5.9072 2.2294 5.2531

3 316% 206 651 235% 240 564 1.6647 6.2076 2.0138 5.3190

4 277% 244 674 228% 255 583 2.0490 6.4366 2.1687 5.5111

5 282% 236 664 229% 246 563 1.9662 6.3410 2.0736 5.3020

6 284% 230 654 234% 237 554 1.9089 6.2324 1.9781 5.2135

7 310% 220 683 241% 238 574 1.8080 6.5310 1.9926 5.4190

8 285% 240 685 219% 264 578 2.0138 6.5473 2.2534 5.4594

9 282% 242 683 225% 256 576 2.0338 6.5326 2.1767 5.4413

10 281% 239 672 239% 234 560 2.0010 6.4127 1.9480 5.2710

Average = 288% 225 648 234% 239 554 1.8618 6.1733 1.9983 5.2113

108

APPENDIX D Scoring, Weigthing and Rating Methods for Three Categories of

Filters

Comparative rating charts provides consumers and institutions with the information and guidance for which

filter can serve their needs at the lowest price, i.e., which filter is their best choice. To develop a ratings

chart, there are three key aspects:

Scoring: In general, it is usually assumed a linear relationship between a filter’s performance on a certain

attribute and the filter’s rating score assigned for this attribute when converting attribute performance to a

rating score. For example, a filter’s attribute like its turbidity removal ability should be linearly converted

to a rating score ranging from 0.50 (poor) to 5.49 (excellent).

Weighing: After giving each filter a rating score for each of the attribute, each attribute is weighed

according to their importance, to compute the overall score for the GNE filters and the Reverse Osmosis

Filter.

Ratings: Ratings is the overall score of each individual filter model, scaled in a range between 0 and 100.

The rating of each filter model indicates the filter’s overall performance.

1. Scorings, Weightings and Ratings Method for GNE Filters

Turbidity removal converting: We take the turbidity removal at each filter’s middle-of-life as the

indicator of the filter’s general turbidity reduction performance. As shown in Table Appendix D.1.2, there

are two fixed turbidity removal limits-at 100% the score should be 5.49 (excellent), and at 0% the score is

0.5 (poor). By inserting a 50% removal rate for a score of 3 and a 10% removal for a score of 1, then we

can compute the linear relationship between turbidity removal (TR) and the score (S). The linear fitting by

computer shows the function is:

S=5×TR+0.5

Taking the Stainless Steel Water Container (Expresso) water filter as an example, the turbidity removal at

its middle-of-life is 98.20% as shown in Table Appendix D.1.1. Thus, its attribute score for turbidity

removal is 5×0.982+0.5=5.41.

Flow rate converting: Each filter’s flow rate at its middle-of-life was also used for representing the filter’s

ability to produce clean water. In Table Appendix D.1.2, 5.00L/hr is inserted for a score of 3 and 1.00L/hr

for a score of 1. Thus, the linear relationship between the middle-of-life flow rate (FR) and the score (S) is:

S=0.5×FR+0.5

109

Using the Expresso again as an example, the flow rate at its middle-of-life is 0.61L/hr as shown in Table

Appendix D.1.1. Thus, its attribute score for flow rate is 0.5×0.61+0.5=0.81.

Table Appendix D.1.1 – Attribute Scores for Each Filter Model

Note: The numbers marked grey are manually supplied.

Table Appendix D.1.2 – The Converting between Attribute Performance and Attribute Score

Lifetime converting: this attribute transformation follows the same pattern as the previous one for flow

rate. The linear function for lifetime (LT) and score (S) is:

S=0.1538×FR+0.6923

110

Table shows that the lifetime of Christella Plus (Tata Swach) is 44.32d, so the attribute score for lifetime is

0.1538×44.32+0.6923=7.51.

TDS removal converting: there is none because the TDS reduction was zero for all of the GNE models.

E.coli LRV converting: the average of E.coli LRV at each filter’s beginning-of-life and end-of-life was

computed to represent the filter’s overall E.coli removal performance. Then the averaged E.coli LRV was

converted to an attribute score according to Table Appendix D.1.2.

Convenience converting: the convenience score consists of two parts-convenience factors and features.

Each are shown compiled in Table D.1.3 along with the scoring used for the GNE filters. In this case, the

convenience factors was actually “inconvenience” attributes and the score was a negative one. The

numerical value assigned to these factors is usually the results of consumer surveys. In this case, it was the

best judgment of the author.

Table Appendix D.1.3 – Convenience Score Transformation for GNE Filters

111

Defining the Attribute Weightings: the numerical value assigned to these factors is usually the results of

consumer surveys. In this case, it was the best judgment of the author. The attribute weightings to compute

the overall score for the GNE filters are shown in Table Appendix D.1.4.

Table Appendix D.1.4 – Attribute Weightings for GNE Filters

Attribute Flow Rate E.coli LRV Turbidity

Removal

TDS

Removal

Lifetime

Weighting (%) 50 10 20 0 20

Computing the overall score: The final step to developing the CR-Style comparative ratings chart, is to

compute the “overall score” for each filter model tested.

As shown in Table Appendix D.1.5, for each filter model, we sum the multiplication of each attribute score

with its associated weighting. Then, the Convenience Score is added to this sum. The result of this

computation is shown in the column of “Raw Score”.

Table Appendix D.1.5 – The Overall Final Score for GNE Filters

For example, the raw score for the Expresso filter is computed as:

0.81×20 + 4.00×20 + 5.41×30 + 0.0×10 + 0.50×20 - 5=263

Another computation determines the maximum and minimum score possible as marked blue in this table.

A linear transformation is made of these Max and Min values to arrive at the scoring range of 100 and 0,

respectively.

Final Score= Raw Score×0.192-7.67

112

The final score for the Espresso filter is: 263×0.192-7.67=43.

Developing the Ratings Chart: Finally, for each attribute, we not only list the attribute score of each filter

model, but also assign it a blob icon according to its performance. In this way, the consumers can more

easily judge the filter’s performance for each specific attribute. The icons are assigned linearly for scores

from 0.5 (poor) to 5.49 (excellent).

Table Appendix D.1.6 – Transformation from Scores to Icons

2. Scoring, Weighting and Rating Method for Reverse Osmosis Filters

Scorings for Reverse Osmosis Filters: For reverse osmosis filters, we only tested two models to their end-

of-life. RO filters have another attribute--waste flow--which GNE filters don’t have. Thus, a different

methodology was used to score and rate those filters. In our work, the author make judgments based on

their expert assessment. The clean flow rate, turbidity removal, TDS removal and E.coli LRV

transformation method is the same as that for GNE filters. But in judging the waste water flow, it would

have been inappropriate to say there should be no waste water for an “Excellent” icon since generating

waste water is an inherent property of the RO filter. Also, a significant difference was seen between the two

models, and this waste water difference is presented to the consumer by increasing the Tata Swach Platina

by one icon category. Choosing a score of 3 for the Clean Water filter model assumed that around 38 liters

per hour was what an average RO filter would generate. The situation was the same for Lifetime scoring

and Convenience scoring.

Defining the Attribute Weightings: The attribute weightings to compute the overall score for the reverse

osmosis filters are different from that for GNE filters. The following table shows the attribute weightings

for the RO filters.

Table Appendix D.2.1 – The Attribute Weightings for the RO Filters

Attribute

Clean

Flow

%

Recovery

E.coli

LRV

Turbidity

Removal

TDS

Removal

Lifetime

Weightings (%) 20 20 20 10 10 20

113

Developing the Ratings Chart: This process is similar for developing an RO and GNE Ratings Chart.

114

References

Aspan, Maria. "Consumer Reports to Add Shopping Magazine". The New York Times (The New

York Times Company), 2006.

http://www.nytimes.com/2006/07/24/business/media/24consumer.html (assessed Dec. 15, 2014)

Baker, R. W., Membrane technology and applications. 3rd ed.; John Wiley & Sons: Chichester,

West Sussex ; Hoboken, 2012; p xiv, 575 p.

Bounds, Gwendolyn. "Meet the Sticklers". The Wall Street Journal, 2010.

Bruce, I.D.; Sharon O.S., Drinking Water Treatment: Reverse Osmosis. NebGuide 2014.

Central Intelligence Agency. "India", The World Factbook.

https://www.cia.gov/library/publications/the-world-factbook/geos/in.html (assessed Nov.

15, 2014)

Colwell, R. R.; Huq, A.; Islam, M. S.; Aziz, K. M.; Yunus, M.; Khan, N. H.; Mahmud, A.; Sack,

R. B.; Nair, G. B.; Chakraborty, J.; Sack, D. A.; Russek-Cohen, E., Reduction of cholera in

Bangladeshi villages by simple filtration. Proceedings of the National Academy of Sciences of

the United States of America 2003, 100 (3), 1051-5.

Consumer Reports. Water Filters. Consumer Reports. Yonkers N.Y. May 2010.

Devangee S.; Kinjal B.; Jain N. K.; Modi H. A., Physicochemical Analysis of Water from

Various Sources and Their Comparative Studies. IOSR Journal of Environmental Science,

Toxicology and Food Technology (IOSR-JESTFT) e-ISSN: 2319-2402, p- ISSN: 2319-2399.

Volume 5, Issue 3 (Jul. - Aug. 2013), PP 89-92

Dies, R. W., Development of a Ceramic Water Filter for Nepal. Master Dissertaion,

Massachusetts Institute of Technology, Cambridge, MA, 2003.

http://web.mit.edu/watsan/docs_theses_nepal.html (assessed Oct. 25,2014)

Eisenberg, T. N.; Middlebrooks, E. J., Reverse osmosis treatment of drinking water.

Butterworths: Boston, 1986; p vii, 271 p.

Howard G, Bartram J. Domestic water quantity, service level, and health [M]. Geneva: World

Health Organization, 2003.

India Center for Science and Environment. Tata uses nano technology for water purifier.

http://www.cseindia.org/node/852 (accessed Nov. 22, 2014)

Levin MM. Diarrheal disease in infants and young children in developing countries. Global

enteric multi-center study, 2009. Presentation at Global Vaccine Forum, Bamako, Mali.

Lichtenthaler, H.; P. Singleton. Bacteria in Biology, Biotechnology and Medicine 1999 John

Wiley Chichester 5th Edition, 489 pages. (ISBN 0-471-98880-4 Paperback). Journal of Plant

Physiology 2000, 157(2), 242.

115

Mattson, J. S., & Mark, H. B. Activated carbon: surface chemistry and adsorption from solution.

1971. New York: M. Dekker. pp. 209-221.

Murcott, S.; Hansen, A.; Knutson, J.; Parikh, V.;Tanphanich, J.; Urmi, T.; Wong, T.K.;

Water Filter Evaluation Suitability (S1-India) Team Report. 2014. PP 26-27.

Kubitschek, H. E., Cell volume increase in Escherichia coli after shifts to richer media. Journal

of bacteriology 1990, 172(1), 94-101.

PATH. User Testing of Household Water Treatment and Storage Products in Andhra Pradesh,

India, 2011.

Pimentel, D., Berger, B., Filiberto, D., Newton, M., Wolfe, B., Karabinakis, E., & Nandagopal,

S., Water resources: agricultural and environmental issues. BioScience 2004, 54 (10), 909-918.

Ratnayaka, D. D., Johnson, M., & Brandt, M. J., Twort's Water Supply. 5th Edition; IWA

Publishing. London, 2000.

Sagara, J., Study of Filtration for Point-of-use Drinking Water Treatment in Nepal. Master

Dissertaion, Massachusetts Institute of Technology, Cambridge, MA, 2000.

http://web.mit.edu/watsan/Docs/Student%20Theses/Nepal/Sagara2000.pdf (assessed Nov. 15,

2014)

Singleton P., Bacteria in Biology, Biotechnology and Medicine (5th edition.), 1999. Wiley.

pp. 444–454. ISBN 0-471-98880-4.

U.S.EPA. Quality Criteria for Water. 1976.

http://water.epa.gov/scitech/swguidance/standards/criteria/current/upload/2009_01_13_criteria_r

edbook.pdf (assessed Nov. 10, 2014)

WHO & UNICEF. Progress on drinking water and sanitation. 2014.

World Health Organization. WHO International Scheme to Evaluate Household Water

Treatment (HWT) Technologies: Procedure for Evaluation. World Health Organization: Geneva,

2014. (a)

World Health Organization. WHO International Scheme to Evaluate Household Water

Treatment Technologies Harmonized Testing Protocol: Technology Non-Specific. World Health

Organization: Geneva, 2014. (b)

World Health Organization. Household water treatment and safe storage.

http://www.who.int/household_water/en/ (assessed Oct. 25,2014) (c)

World Health Organization. Evaluating Household Water Treatment Options: Health-based

targets and microbiological performance specifications. World Health Organization: Geneva,

2011. (a)

116

World Health Organization. Guidelines for drinking-water quality: Fourth Edition. World Health

Organization, 2011. (b)

World Health Organization. Guidelines for Drinking Water Quality. 2nd Edition. Volume III Surveillance

and Control of Community Water Supplies. World Health Organization, 1997. Geneva.

World Health Organization. Total dissolved solids in Drinking-water. Background document for

development of WHO Guidelines for Drinking-water Quality. World Health Organization:

Geneva, 1996.