THE ENVIRONMENTAL TECHNOLOGY VERIFICATION
PROGRAM ^
f X
oEPA
ETV
U.S. Environmental Protection Agency
NSF International
ETV Joint Verification Statement
TECHNOLOGY TYPE: ULTRAFILTRATION MEMBRANE
APPLICATION: REMOVAL OF VIRUSES IN DRINKING WATER
PRODUCT NAME: SFD-2880 ULTRAFILTRATION MODULE
VENDOR: THE DOW CHEMICAL COMPANY - DOW WATER
SOLUTIONS
ADDRESS: 1691 SWEDE ROAD
MIDLAND, MI 48674
CONTACT: DARYL GISCH
PHONE: +1989-636-9254
EMAIL: DGISCH@DOW.COM
NSF International (NSF) manages the Drinking Water Systems (DWS) Center under the U.S.
Environmental Protection Agency's (EPA) Environmental Technology Verification (ETV) Program. The
DWS Center recently evaluated the performance of the Dow Water Solutions SFD-2880 ultrafiltration
(UF) module for removal of viruses under a scenario where one UF fiber was broken. The challenge test
was conducted under controlled laboratory conditions at NSF's testing laboratory in Ann Arbor, MI.
Testing of the SFD-2880 UF module was conducted to verify virus reduction following the requirements
of the Department of Health Victoria (Australia) Draft guidelines for validating treatment processes for
pathogen reduction, supporting Class A water recycling schemes in Victoria. The Department of Health
Victoria guideline is largely based on the product-specific challenge requirements of the EPA Long Term
2 Enhanced Surface Water Treatment Rule (LT2ESWTR) and accompanying EPA Membrane Filtration
Guidance Manual (MFGM).
EPA created the ETV Program to facilitate the deployment of innovative or improved environmental
technologies through performance verification and dissemination of information. The goal of the ETV
Program is to further environmental protection by accelerating the acceptance and use of improved and
more cost-effective technologies. ETV seeks to achieve this goal by providing high-quality, peer-
reviewed data on technology performance to those involved in the design, distribution, permitting,
purchase, and use of environmental technologies.
ETV works in partnership with recognized standards and testing organizations, stakeholder groups
(consisting of buyers, vendor organizations, and permitters), and with the full participation of individual
technology developers. The program evaluates the performance of innovative technologies by developing
test plans that are responsive to the needs of stakeholders, conducting field or laboratory tests (as
NSF 12/35/EPADWCTR The accompanying notice is an integral part of this verification statement. March 2012
VS-i
-------
appropriate), collecting and analyzing data, and preparing peer-reviewed reports. All evaluations are
conducted in accordance with rigorous quality assurance protocols to ensure that data of known and
adequate quality are generated and that the results are defensible.
ABSTRACT
The purpose of this verification was a cut fiber challenge study for the Dow Chemical Company SFD-
2880 UF membrane module. MS2 coliphage virus was the surrogate challenge organism. The challenge
tests followed the requirements of the Department of Health Victoria (Australia) Draft guidelines for
validating treatment processes for pathogen reduction, supporting Class A water recycling schemes in
Victoria.
Five new fully integral modules were challenged with MS2. Then, one fiber in each module was cut, and
the modules were all tested again with MS2. The modules were operated at a target flux of 70 gallons per
square foot per day (gfd), which equates to a flow of 40.3 gallons per minute (gpm). The test data did not
show any significant reduction in virus removal with a cut-fiber, so a second round of testing was
conducted with higher MS2 challenge concentrations. For the retests, the cut-fiber challenges were
conducted first, followed by the intact module tests after the cut fibers were pinned. The average log
removal value (LRV) was calculated for each module, as well as LRVs for each feed/filtrate sample pair.
From this data, the LRVC.TEST was determined as the lowest mean LRV, and also the lowest sample pair
LRV. Table VS-i provides a summary of the LRVC.TEST data.
Table VS-i. LRVC-TEST for Each Round of Testing
Tests
Round 1
Round 2
Intact Modules
Mean LRV
3.73
2.59
Lowest LRV
3.44
2.39
Cut Fiber
Mean LRV
3.44
2.46
Lowest LRV
2.98
2.37
To evaluate whether there was significantly lower virus removal with a cut fiber, the LRVs for each
feed/filtrate sample pair were pooled together, and the paired-difference t statistic was calculated using
Microsoft® Excel®. The intact module vs. cut-fiber t statistic for the first round of tests is 1.15. This
value is below the critical t value of 2.15, indicating that virus removal was not significantly impacted by
the cut fiber. For the second round of tests, the paired-difference t statistic is 5.00, this time
demonstrating a statistically significant drop in LRV.
PRODUCT DESCRIPTION
The following information was provided by Dow and was not verified.
The Dow SFD-2880 UF membrane module measures 4.7 inches in diameter by 45.5 inches in length.
The membrane fibers are made of polyvinylidene fluoride (PVDF). Water flow through the membrane
fibers is outside to inside. The modules can operate in deposition (dead-end) or suspension modes. The
nominal pore size is 0.03 (im. The maximum recommended flux is 70 gfd, with a maximum
recommended feed pressure of 44 pounds per square inch (psi), and a maximum transmembrane pressure
of 30 psi.
NSF 12/35/EPADWCTR
The accompanying notice is an integral part of this verification statement.
VS-ii
March 2012
-------
VERIFICATION TEST DESCRIPTION
Virus Surrogate
The modules were challenged with the MS2 coliphage virus (ATCC 15597-B1) as a surrogate for enteric
viruses, as required by the Victoria draft guidelines for virus removal credits. MS2 is generally accepted
as an enteric virus surrogate for size-exclusion technologies due to its small size (approximately 22-26
nanometers). The USEPA MFGM references MS2 as an acceptable surrogate for enteric viruses because
it is similar in size and shape to poliovirus and hepatitis virus.
Methods and Procedures
All tests were conducted at the NSF International testing laboratories following the requirements of the
EPA-approved Test/QA Plan for Validating the Dow Chemical Company SFD-2880 Ultraflltration
Membrane Module for Virus Reduction Following the Department of Health Victoria (Australia) Draft
Guidelines for Validating Treatment Processes for Pathogen Reduction. The Victoria guidelines support
the regulatory approval process for Class A water recycling schemes. For membrane filtration products,
the guidelines are largely based on the product-specific challenge requirements of the LT2ESWTR and
accompanying EPA MFGM.
The intact module and cut-fiber tests were conducted twice, in two separate rounds of testing. The second
round of testing was conducted because the data from the first round did not show any significant
reduction in virus retention between the intact and cut-fiber challenges. For the first round, NSF tested
the five intact modules in January 2011. Then Dow representatives visited the testing laboratory to cut
the fibers, and the cut-fiber tests were conducted in March 2011. As required in the Department of Health
Victoria guideline, one fiber in each module was cut as close as possible to the potting resin on the filtrate
end of the module. The retests were conducted in May 2011, with the cut-fiber tests conducted first,
followed by the intact module challenges after Dow representatives pinned the cut fibers. Each of the five
modules submitted for testing was challenged individually. The target flux for membrane operation was
Dow's maximum recommended value of 70 gfd at 25 °C, which equals a flow rate of 40.3 gpm. Before
and after each challenge test, each module was subjected to a pressure decay test to satisfy the non-
destructive performance test requirement in Section 3.6 of the MFGM.
Immediately prior to testing, each module was forward flushed at approximately 40 gpm. For the first
round of tests conducted in January and March, each module was flushed for five minutes. At the start of
the retest round in May, the laboratory engineer noticed that after five minutes of flushing, there were still
bubbles visible in the filtrate hose line. The engineer flushed Module 1 for an additional three minutes
until bubbles were no longer visible. The engineer had to flush Module 2 for a total of 22 minutes to clear
the bubbles. At this point, the engineer decided to install bleed valves in the reject port caps for Modules
3, 4 and 5 to allow for evacuation of the air. After the pressure decay tests for these three modules, the
bleed valve was opened and the flow of water started at 40 gpm. The valves were kept open until all the
air had escaped. This allowed the testing engineer to return to the flush time of five minutes.
The duration of each challenge test was approximately 35 minutes. The MS2 suspension was injected
into the feed stream at start-up, after 15 minutes of operation, and after 30 minutes of operation. After at
least one minute of injection to pass the equilibrium volume, grab samples were collected from the feed
and filtrate sample taps. After each round of sample collection, injection of the challenge organism
suspension was turned off, and clean feed water was pumped through the modules at 40 gpm until the
next sampling point.
VERIFICATION OF PERFORMANCE
The LT2ESWTR and MFGM specify that an LRV for the test (LRVc-TEsi) be calculated for each module
tested, and that the LRVs for each module are then combined to yield a single LRVC.TEST for the product.
NSF 12/35/EPADWCTR The accompanying notice is an integral part of this verification statement. March 2012
VS-iii
-------
If fewer than 20 modules are tested, as was the case for this verification, the LRVC_TEST is simply the
lowest LRV for the individual modules. However, the rule does not specify a method to calculate LRVC.
for each module. Suggested options in the MFGM include:
• Calculate a LRV for each feed/filtrate sample pair, then calculate the average of the individual
sample point LRVs;
• Average all of the feed and filtrate counts, then calculate a single LRV for the module; or
• Calculate a LRV for each feed/filtrate sample pair, select the LRV for the module as the lowest
(most conservative of the three options).
Options 1 and 2 give LRVc-iEsi values that are either identical, or only a few hundredths or less different,
so for this verification, options 1 and 3 are used to calculate LRVs.
First Roun d Results
The MS2 LRVs for the first round of tests are presented in Table VS-ii. The intact module LRVc-TEsi,
using the overall mean LRV calculations in Table VS-ii, is 3.73. The LRVC_TEST based on the lowest
individual sample pair log reduction is 3.44. Under the cut-fiber scenario, the LRVC.TEST from the overall
means is 3.44, while that from the lowest individual sample pair log reduction is 2.98.
Table VS-ii. First Round LRV Calculations
Module #
Module 1
Module 2
Module 3
Module 4
Module 5
Intact Modules
Mean LRV
3.92
3.73
4.55
3.82
4.17
Lowest LRV
3.85
3.44
4.09
3.65
3.90
Cut Fiber
Mean LRV
3.44
3.72
4.73
4.93
4.49
Lowest LRV
3.10
2.98
4.19
4.65
4.30
To evaluate whether there was significantly lower virus removal with a cut fiber, the LRVs for the
feed/filtrate sample pairs were pooled together, and the paired-difference t statistic was calculated using
Microsoft® Excel®. The mean LRV for all five intact modules is 4.04, with individual sample pair LRVs
ranging from 3.44 to 5.16. The mean LRV for the cut-fiber tests is actually higher, at 4.26, with a range
of 2.98 to 5.74. The paired-difference t statistic for the two sets of LRVs is 1.15, which is below the
critical t value of 2.15 (14 degrees of freedom) that denotes a significant difference with a confidence of
95%.
The intact module pressure decay rates ranged from 0.000 to 0.052 psi/min, while those for the cut-fiber
scenario ranged from 0.734 to 1.292 psi/min, indicating that there was indeed a significant integrity
breach.
A possible explanation for why there was no significant difference between the intact module and cut-
fiber scenarios arises from the testing engineer's observation of air bubbles in the filtrate during the pre-
test flushes for the retests, as discussed above. If a portion of the air introduced for the pressure decay test
was still trapped at the top of the module during the challenge test, the cut fiber at the top of the module
may have never been in contact with the challenge water during the first round of tests; it may have been
in a pocket of air trapped at the top of the module. This theory is bolstered by a comparison of the feed
pressure data for the first round of tests versus the retests. For the first round of tests, above 20 psi
driving pressure was needed for eight of the ten test runs to achieve the target flux, compared with less
than 18 psi for the retests. If air was trapped in the modules, thus occluding a significant portion of the
NSF 12/35/EPADWCTR
The accompanying notice is an integral part of this verification statement.
VS-iv
March 2012
-------
membrane surface area, a higher driving pressure would be needed to achieve the target flux, due to the
smaller surface area.
To attempt to discern a significant difference in LRV between the intact modules and modules with a cut
fiber, NSF and Dow decided to re-run the tests on the same five modules using a higher MS2 challenge.
Secon d Roun d Results
The LRVs for the second round of tests are displayed in Table VS-iii. The intact module LRVC.TEST,
using the overall mean LRV calculations in Table VS-iii, is 2.59. The LRVc-TEsi based on the lowest
individual sample pair log reductions is 2.39. Under the cut-fiber scenario, the LRVC.TEST from the
overall means is 2.46, while that from the lowest individual sample pair log reductions is 2.37. The intact
module pressure decay rates ranged from 0.000 to 0.035 psi/min, while those for the cut-fiber tests ranged
from 0.970 to 1.284 psi/min.
Table VS-iii. Second Round LRV Calculations
Module #
Module 1
Module 2
Module 3
Module 4
Module 5
Intact Modules
Mean LRV
3.39
3.10
3.36
3.27
2.59
Lowest LRV
3.23
2.90
3.09
3.04
2.39
Cut Fiber
Mean LRV
3.13
2.67
2.93
2.77
2.46
Lowest LRV
3.12
2.66
2.87
2.53
2.37
In contrast to the first round LRV data, the retest data set does show a statistically significant difference in
virus retention between the intact and cut-fiber scenarios. The mean LRV for all five intact modules is
3.14, with a range of 2.39 to 3.62. The mean LRV for the cut-fiber tests is 2.79, with a range of 2.37 to
3.14. The paired-difference t statistic for the two sets of LRVs is 5.00, which is above the critical t value
of 2.15 for a significant difference at the 95% confidence level.
The pressure decay rates indicated a catastrophic loss of membrane integrity, but the corresponding loss
of virus retention was not as large. For the retests, the cut-fiber pressure decay rates were approximately
30 times higher than those for the intact modules. This translates into an approximate 1.5 log loss of
membrane integrity. However, the MS2 reduction data only shows a mean LRV loss of 0.35 logs.
QUALITY ASSURANCE/QUALITY CONTROL (QA/QC)
NSF provided technical and quality assurance oversight of the verification testing as described in the
verification report, including a review of 100% of the data. NSF QA personnel also conducted a technical
systems audit during testing to ensure the testing was in compliance with the test plan. A complete
description of the QA/QC procedures is provided in the verification report.
NSF 12/35/EPADWCTR
The accompanying notice is an integral part of this verification statement.
VS-v
March 2012
-------
Original signed by Annette Gatchett
for Cynthia Sonich Mullin 05/04/12 Original signed by Pierre Sbabo 05/04/12
Cynthia Sonich-Mullin Date Pierre Sbabo Date
Director Vice President
National Risk Management Research Water Systems
Laboratory NSF International
Office of Research and Development
United States Environmental Protection
Agency
NOTICE: Verifications are based on an evaluation of technology performance under specific,
predetermined criteria and the appropriate quality assurance procedures. EPA and NSF make no
expressed or implied warranties as to the performance of the technology and do not certify that a
technology will always operate as verified. The end-user is solely responsible for complying with
any and all applicable federal, state, and local requirements. Mention of corporate names, trade
names, or commercial products does not constitute endorsement or recommendation for use of
specific products. This report is not an NSF Certification of the specific product mentioned
herein.
Availability of Supporting Documents
Copies of the test protocol, the verification statement, and the verification report (NSF
report # NSF 12/35/EPADWCTR) are available from the following sources:
1. ETV Drinking Water Systems Center Manager (order hard copy)
NSF International
P.O. Box 130140
Ann Arbor, Michigan 48113-0140
2. Electronic PDF copy
NSF web site: http://www.nsf.org/info/etv
EPA web site: http://www.epa.gov/etv
NSF 12/35/EPADWCTR The accompanying notice is an integral part of this verification statement. March 2012
VS-vi
------- |