Approval testing of membrane filtration systems

There is currently great interest in the UK and across Europe in the

application of membrane filtration systems for the treatment of raw water

for the municipal supply of drinking water. Three distinct drivers for this

interest can be identified:

– increasing concern over health effects of disinfection by-products;

– the ability of membrane filters to remove pathogenic micro-organisms, such

as cryptosporidium oocysts, from the raw water; and

– the need for alleviation of potential drought situations caused by climate

change effects, which has forced water suppliers to consider water resources

previously considered untreatable economically, such as brackish or sea

waters.

However, these technologies present a challenge to regulatory and approval

organisations, who must decide whether the materials of construction of

these (and other) products adversely affect the quality of water through the

release of harmful substances. In many cases, testing methods used by the

testing bodies are based on those used for simpler products such as plastic

pipes. These test methods may require modification or re-interpretation when

applied to membrane filtration systems.

One of the most important requirements for any product (including membrane

filtration systems) used in contact with drinking water is that it should

not contaminate the water. Contamination can arise from the leaching or

migration of residual traces of chemicals from the materials into the water.

Suppliers of membrane systems usually consider them as technologies to

remove contaminants from water, rather than as potential sources of

contaminants. The contact time between membrane elements and the water being

treated is usually much shorter than the contact time between water and a

reservoir or pipe coating, so the potential for contamination may be

limited. Nevertheless, most regulators consider that these products need to

be assessed and tested using the same principles applied to other products.

The substances of concern to regulators include contaminants, such as

residual monomers or additives, that are not monitored routinely in water

supplies. In the case of membrane systems, for example, the type of compound

that might be released includes the substances used to preserve the

membrane, which are mostly flushed out to waste during commissioning.

It is difficult to measure such substances directly under field conditions

with enough sensitivity to satisfy toxicological concerns, so laboratory

tests are applied that exaggerate leaching. These test concentrations are

used to estimate concentrations that would occur at customers’ taps. The

estimation requires information on the surface area of the material, the

volume of water in contact with that area and the contact time, for both the

test conditions and the normal conditions of use of the products.

Most approval schemes require detailed compositional information on all

components and raw materials of the products, so that the potential for

contamination can be assessed. Compliance with positive lists (lists of

permitted ingredients) may need to be verified. This can pose problems for

filtration systems that contain many small components, such as ŒO’ rings and

seals, where the suppliers of the components and their ingredients may be

reluctant to provide the required information. Confidentiality issues

frequently arise.

The major European countries have, independently, developed schemes for

assessing and testing products for use with drinking water and approving

suitable products.

The use of products used in public water supplies in the UK is subject to

Regulation 25 of the Water Supply (Water Quality) Regulations 1989. The

Drinking Water Inspectorate (DWI) is the government agency responsible for

checking that water companies comply with the Regulations. Approvals under

Regulation 25 are granted on the basis of testing and scrutiny and a

toxicological assessment by the Government-appointed expert Committee on

Chemicals and Materials (CCM). The focus of the assessment is the potential

for a product to release harmful organic substances into the public water

supplies. The CCM specifies its requirements, usually for leaching and

analytical test work, individually for each product. For membrane systems,

the CCM asks for extensive information on the product, including:

– its intended use, construction and operational flows;

– the complete composition of each component that contacts the feed and

treated water;and

– British Standard BS 6920: 1996 tests on each component (tests for taste of

water; turbidity and colour, the release of microbial nutrients, toxic

metals; and cytotoxic substances).

From this data the CCM identifies those chemicals that might be released

from the membrane system, and specifies leaching tests to monitor the extent

to which they are released. The tests involve flushing the intact membrane

element in accordance with manufacturer¹s instructions and then leaving

fresh test water to stand in the membrane element for three sequential

periods of 24, 48 and 72 hours. The water is analysed after each leaching

period for leaching substances.

Then the CCM considers whether the leaching observed is acceptable on a

toxicological basis. As part of that assessment, CCM converts the

concentrations of substances detected in the static leaching tests (which

are often measured in mg/l) into concentrations that would be found in

practice with a dynamic product.

In the case of membrane filtration systems, the test protocols used in other

countries can be significantly different:

In the Netherlands, the Netherlands Waterworks’ Testing and Research

Institute (Kiwa N.V.) is responsible for the testing and certification of

materials used in the production and distribution of drinking water.

Leaching tests on membrane elements are undertaken by Kiwa.

The complete element is flushed in accordance with the manufacturer’s

instructions and then disassembled into its constituent components. The

components are then placed in contact with fresh test water at 23oC for

three sequential 72-hour stagnation periods. Leach-ates are analysed for

leaching substances. Estimated exposures to contaminants (tap

concentrations) are generated with a conversion factor that relates the

surface area of the part tested to the area of the membrane.

In France, the test elements are installed in a test rig and rinsed

according to manufacturers¹ instructions. Test water is then re-circulated

through the element for 24 hours, passing through the membrane 50 times

during that time, and then analysed for leaching substances. After

disinfection, under the same conditions as recommended by the manufacturer,

a second 24-hour re-circulation test is carried out with fresh test water

and this water is also analysed for contaminants.

The different national approval schemes constitute a significant barrier to

trade between countries, and progress in developing and agreeing harmonised

standards for materials in contact with water is slow. At present, there is

little, if any, mutual acceptance in each country of foreign approvals and

manufacturers seeking to have their products approved for use abroad must

submit their products to each national approval body for assessment. Costs

to applicants of providing the required information for the application can

be significant, especially if manufacturers seek to market the products in

many countries.

While this problem applies to all materials, including pipes and pipe

linings, there is an extra difficulty associated with membrane filtration

systems. The complex designs and use of novel materials has led to a

corresponding complexity in the test protocols and conversion factors used

in each country, so that it is difficult to predict how products tested

under one regime will fare in another. WRc has undertaken research,

commissioned by the DWI, into alternative leaching test procedures for

membrane systems. For example, results using a static test protocol (the UK

protocol) were significantly different (leaching rates up to 40 times lower)

to those using a dynamic protocol (based on the French re-circulation test)

on the same type of membrane element.

The DWI, in collaboration with NSF International and Kiwa N.V, organised an

international workshop on this issue. The aim of the workshop was to provide

a forum for the development of an international consensus for the regulatory

assessment and testing of membrane filtration systems used in the treatment

of drinking water. The workshop reached a consensus on the preferred use of

dynamic test protocols (rather than static methods) for testing products

that are used in a dynamic way. However, more research was required to

address the issues of whether re-circulation was likely to underestimate

leaching of some contaminants. An alternative would be to use single pass,

low-flow test conditions. A comparative study is currently underway at WRc,

Kiwa and NSF International. It is intended that the results will become part

of the programme within CEN to develop standards for membrane filtration

systems.

Suppliers of membrane technology to the water industry have to develop and

market their products with an understanding of their customers’ needs. This

may require significant investment in their products, to gain the relevant

approvals. However the current work should lead to a soundly-based test

protocol that will be accepted by international approval bodies and

regulators as a common test protocol.