Avoiding the Use of Exhausted Drinking Water Filters: A Filter-Clock Based on Rusting Iron

: Efﬁcient but affordable water treatment technologies are currently sought to solve the prevalent shortage of safe drinking water. Adsorption-based technologies are in the front-line of these efforts. Upon proper design, universally applied materials (e.g., activated carbons, bone chars, metal oxides) are able to quantitatively remove inorganic and organic pollutants as well as pathogens from water. Each water ﬁlter has a deﬁned removal capacity and must be replaced when this capacity is exhausted. Operational experience has shown that it may be difﬁcult to convince some low-skilled users to buy new ﬁlters after a predicted service life. This communication describes the quest to develop a ﬁlter-clock to encourage all users to change their ﬁlters after the designed service life. A brief discussion on such a ﬁlter-clock based on rusting of metallic iron (Fe 0 ) is presented. Integrating such ﬁlter-clocks in the design of water ﬁlters is regarded as essential for safeguarding public health.


Introduction
Safeguarding public health within a given community has two key aspects: preventing the occurrence of diseases and controlling the spread of such diseases, with disease curing being the function of medicine. However, these three aspects are inter-connected. If the cause of a specific disease is understood or at least well-characterized, then the spread of that disease within a community can be reduced to a minimum or completely eradicated. Supplying human communities with safe drinking water has been established as a necessity, thus warrants monitoring and scrutiny [1][2][3][4], to mitigate against waterborne diseases.
The development of affordable, appropriate, efficient and sustainable water filtration systems is regarded as one of the fundamental research goals in efforts to achieve the United Nations' Sustainable Development Goals (UN SDGs) [5]. The role of filtration systems based on metallic iron (Fe 0 filters) has been largely discussed in this context [2,[5][6][7][8][9][10][11]. Fe 0 filters are regarded as a key technology because suitable Fe 0 materials are universally available and design criteria have been established [12][13][14]. One shortcoming of Fe 0 filters is the time-dependent permeability loss of the porous bed. This limitation has the potential to become advantageous when designing next generation filters for low-skilled and low-income communities. Such next generation Fe 0 -based filters will rely on rusting to develop a clock system to signal filter exhaustion.
Household water treatment systems are extremely difficult to maintain. Their maintenance is mostly left to the owners because necessary specialised supervision by experts would be difficult to secure [10,15]. Impoverished communities are usually less educated, with few members of such a society being equipped with few saleable skills. Therefore, they may lack money for even vital water filters and would attempt to continue to use an exhausted filter. Moreover, such communities often lack access to an analytical laboratory to regularly analyse the quality of drinking water from such filters [16]. Therefore, to safeguard their health, low-cost and appropriate means should be found to encourage water filter changes after the estimated service life of the filter. One option includes exploiting the volumetric expansive nature of iron corrosion or rusting in aqueous systems [17,18]. In this regard, a small Fe 0 filter designed to clog promptly at the end of the service life can be added to any water filtration system (not only Fe 0 -based ones). When the system is clogged, no water is released, thereby ensuring the user to change the filtration device.
This communication is structured as follows; first, an overview on existing household filtration systems is presented, and then the scientific knowledge behind the filter-clock and some practicable designs of such a monitoring system are discussed.

Commercial Household Water Filters: Now and Then
Commercial water filtration systems for households are antecedent to centralized water provision systems [19]. Historically, communities initially used water from wells as an alternative to surface water or collected and stored rainwater. They then realized that stored water might need some sort of treatment. With the development of science, it was later established that virtually each water source needs some degree of treatment before use as a potable water source [4]. As early as the 1860s, commercial household water filters were commonplace in Europe [19][20][21]. Nichols [19] documented the following aggregates for use as filter materials: animal charcoal, broken bricks, broken stone, carbide of iron (Fe 0 ), cotton, flannel, gravel, iron turnings (Fe 0 ), iron wire (Fe 0 ), pebble stone, porous stone, powdered glass, sand, sawdust, spongy iron (Fe 0 ), wood-charcoal, and wool among others.
It might be surprising to learn that technical expertise in filtration for decentralized safe drinking water provision is a century old. Recent review articles [1,2,6-9] revealed that modern filtration devices have the following two common characteristics: (i) are made of similar materials [2,6,7] and (ii) experience the same limitations: i.e., their filtration capacity is not indefinite; hence, they require renewal at proper intervals [10]. A special material, termed activated carbon, was introduced in the 1940s and was intensively investigated for some 30 years [9,28,29]. Granular activated carbon (GAC) is the form commonly used to design household water filters. Just like Fe 0 -based filters, each designed GAC filter has a limited service life. Therefore, appropriate filter replacement or changes in filtration units is required after the exhaustion of the designed service life.
Despite being the best available adsorbent for water treatment due to its excellent adsorption capacity, GAC could not be used at a large scale due to its high cost of production (and regeneration). Moreover, such costly adsorbents are unaffordable to low-income households in developing countries.
This prohibitive cost has prompted scientists to develop more-affordable adsorbents from agricultural, industrial and geo-materials [2,6,7,9,30,31]. Factually, resulting filters vary in their service lifes, which are intuitively shorter than those of GAC filters. For the developers of filtration units, their production is based on the availability, affordability and efficiency of the resulting filters at household level, with their regenerability being of secondary importance, although a system to collect used filtration materials could be introduced. It is not rare that most water filters on the market contain an operating manual describing the filter, and specifying its expected service life. The given service life of a filter is considered a reliable way to determine the longevity of filtration systems. However, the question is how should even an illiterate user be knowledgeable about filter life span, and how to change a filter after the expected service life? This aspect is discussed herein focusing on Fe 0 filtration systems.

General Aspects
Water treatment technology using Fe 0 materials consists of 'putting corrosion to use' [32]. In this case, iron corrosion products (FeCPs) are mainly used as contaminant collectors and degraders [10,26,27,[33][34][35][36][37][38][39] or as catalysts for chemical reactions ( [40]). Herein, a different aspect of using FeCPs is presented: The volumetric expansive nature associated with FeCPs generation [17] can be used to control filter life span. In fact, each FeCP (oxide) is larger in volume than the parent Fe 0 (iron) in the metal lattice (V oxide > V iron ). In practice, an expansion coefficient (η) has been introduced such that V oxide = η*V iron (η ≥ 2.01) [18]. The η value depends on the availability of dissolved O 2 . This means that, each Fe 0 -based filter is prone to clogging 'by virtue' of the volumetric expansive nature of aqueous iron corrosion [18]. In other words, by using different Fe 0 materials and/or different Fe 0 proportions in hybrid systems (e.g., Fe 0 /sand, Fe 0 /pumice, Fe 0 /pozzolane) [39], it is possible to determine the service life (t ∞ ) of a filter. That is, for example, the time frame for which a household-based filter delivers enough safe drinking water for a typical number of people in a household.
Although Fe 0 -based filters are mostly a stand-alone technology for safe drinking water provision [5,10,21], the idea herein is to develop miniaturized units with capacity to clog after a defined operational duration (e.g., t ∞ = 2, 4, 8, 12, 18 or 24 months) and under given operational conditions (e.g., presence of O 2 , ambient temperature, water quality). Accordingly, the development of a Fe 0 -based filter-clock is therefore regarded as a futuristic independent domain of research in public health engineering.
The principle behind a filter-clock is to let the polluted water flow through two filters: (i) a conventional filter (e.g., a granular activated carbon filter) that mitigates the contamination and (ii) a second Fe 0 -based filter (filter-clock) designed to clog promptly after the capacity of the conventional filter is exhausted (at t ∞ ). Both filters can be combined in a single unit or the order of the relative position of the two adjoining units can be inverted. Figure 1 shows the basic concept of the conventional filter.

Design Aspects
The universal equation of a Fe 0 filter is given by Equation (1): where Voxide = 0 for t = 0 (t0), t0 corresponds to the freshly designed filter. Valid Voxide values correspond to Voxide < Vpore, Vpore corresponding to the initial porosity of the filter. t∞ corresponds to Voxide = Vpore and is characterized by system' clogging. t∞ thus corresponds to the service life of the conventional filter to be equipped with a clock. 'A' is a factor considering the geometry of the Fe 0 particle and its time-dependant variation [13,14]. Vsolid is the initial volume of solid materials (Fe 0 and other aggregates) and τiron is the Fe 0 proportion in the solid mixture. η reflects the availability of dissolved O2 [14]. The primary equation of a filter is given by Equation (2): Knowing τiron (Viron = τiron*Vsolid), Vfilter (fixed by the designer), the η value (determined by the O2 level) and assuming spherical particles, Vpore can be derived and Equation (1) solved. However, the time-dependent law accounting for FeCPs generation is typically missing [41][42][43][44][45]. Although assumptions can be made using the Faraday Law and current intensity values from the literature, the lack of long-term corrosion data on granular Fe 0 corrosion makes the solution of Equation (1) challenging. This communication is about contributing to solving this challenge by encouraging the scientific community to record data on long-term corrosion of relevant Fe 0 materials under various field conditions. This is establishing t∞ values for various systems and practically used them on a situation-characteristic basis.
A survey of the literature on Fe 0 -based filtration reveals many examples of systems that have clogged after some weeks or months [25,35,[46][47][48][49][50]. The remaining task is to bring some systematic approach in the investigations to come up with a set of applicable tools. This is a typical case where the Know-Why will be derived from the Know-How.
As an example, using a pure layer of iron filings (Fe 0 : 6 g) in a small column (inner diameter = 3.0 cm and length = 11.5 cm), Ndé-Tchoupé [39,51] reported on almost complete clogging after 45 days. The Fe 0 layer occupied just 2.2 cm of the column. This result suggested that the material tested by this

Design Aspects
The universal equation of a Fe 0 filter is given by Equation (1): where V oxide = 0 for t = 0 (t 0 ), t 0 corresponds to the freshly designed filter. Valid V oxide values correspond to V oxide < V pore , V pore corresponding to the initial porosity of the filter. t ∞ corresponds to V oxide = V pore and is characterized by system' clogging. t ∞ thus corresponds to the service life of the conventional filter to be equipped with a clock. 'A' is a factor considering the geometry of the Fe 0 particle and its time-dependant variation [13,14]. V solid is the initial volume of solid materials (Fe 0 and other aggregates) and τ iron is the Fe 0 proportion in the solid mixture. η reflects the availability of dissolved O 2 [14]. The primary equation of a filter is given by Equation (2): Knowing τ iron (V iron = τ iron *V solid ), V filter (fixed by the designer), the η value (determined by the O 2 level) and assuming spherical particles, V pore can be derived and Equation (1) solved. However, the time-dependent law accounting for FeCPs generation is typically missing [41][42][43][44][45]. Although assumptions can be made using the Faraday Law and current intensity values from the literature, the lack of long-term corrosion data on granular Fe 0 corrosion makes the solution of Equation (1) challenging. This communication is about contributing to solving this challenge by encouraging the scientific community to record data on long-term corrosion of relevant Fe 0 materials under various field conditions. This is establishing t ∞ values for various systems and practically used them on a situation-characteristic basis.
A survey of the literature on Fe 0 -based filtration reveals many examples of systems that have clogged after some weeks or months [25,35,[46][47][48][49][50]. The remaining task is to bring some systematic approach in the investigations to come up with a set of applicable tools. This is a typical case where the Know-Why will be derived from the Know-How.
As an example, using a pure layer of iron filings (Fe 0 : 6 g) in a small column (inner diameter = 3.0 cm and length = 11.5 cm), Ndé-Tchoupé [39,51] reported on almost complete clogging after 45 days. The Fe 0 layer occupied just 2.2 cm of the column. This result suggested that the material tested by this author can be used as a filter-clock in all cases the filter capacity is expected to exhaust after one month. This system can be used as a reference with Fe 0 admixed with sand in different proportions (e.g., Fe 0 :sand 3:1, 1:1, 1:3) in order to obtain filter-clocks corresponding to a longer service life (t ∞ > 4 weeks). Factually, the design from the observations of Ndé-Tchoupé [39,51] should be miniaturized to produce the first generation filter-clock.

Conclusions
Metallic iron (Fe 0 ) has been used for environmental remediation and water treatment for decades. A great deal of research has been done to improve the understanding of the Fe 0 /H 2 O system in the context of environmental remediation. The lack of a standardised method to characterize the intrinsic reactivity of used Fe 0 materials and their impact of the system's efficiency often causes confusion in interpreting achieved results. This communication introduces the concept of a filter-clock to fix the service life of filtration systems. Fe 0 is chosen because permeability loss due to expansive iron corrosion is currently the major challenge of Fe 0 /H 2 O remediation systems. In an effort to establish miniaturized Fe 0 filters to assess a filter-clock, reasonable iron corrosion rates for modelling purposes must be known.
It is expected that measured Fe 0 corrosion rates under various operational conditions (including water quality characteristics) will accelerate the development of the Fe 0 filtration technology as a whole. With this simple operational procedure, the service life of all available filters can be known thus fixed. This would be a major contribution to public health, particularly in developing countries, where the provision of safe drinking water remains a challenge. Furthermore, Fe 0 -based filter clocks will strengthen frugal innovations for safe drinking water provision worldwide, in particular, for impoverished communities.