Comparison between experimental results of different technologies for arsenic removal from water intended for human consumption

Nowadays, with deteriorating environmental quality, human life is exposed to numerous types of threats. There are a number of harmful chemicals around us: ingestion, breathing or any kind of exposure to them can affect the human health tremendously (Sarkar & Paul 2016; Rusu et al. 2018). Arsenic (As) is ubiquitous in the environment and highly toxic to all forms of the life. It is a crystalline ‘metalloid’, a natural element with features intermediate between metals and non-metals; it occurs naturally as an element, and ranks as the 20th most frequently occurring trace element in the earth's crust, 14th in seawater, and 12th in the human body (Singh et al. 2015; Zhou & Xi 2018).

Arsenic exists mainly in four oxidation states – arsenate (As(V)), arsenite (As(III)), arsenic (As(0)), and arsine (As(-III)) and its solubility depends on the pH and ionic environment (Chieh Kao et al. 2013; Crognale et al. 2017; Li et al. 2017). In natural waters, arsenic is usually present as two main species, As(V) and As(III). The latter is usually present in water under anoxic conditions (such as in groundwater), while As(V) is detected in surface water. Moreover, As(III) is much more toxic and mobile than As(V). In general, inorganic arsenicals (i.e. arsenate and arsenite) are more toxic than organic ones (Singh et al. 2015; Sarkar & Paul 2016; Crognale et al. 2017; Li et al. 2017; Wang et al. 2018).

As concentration in water varies from one water body to another. In freshwater systems (rivers and lakes), the variation in As concentration is in the range of 0.15–0.45 μg/L, depending on the source, availability and geochemistry of the catchments (Singh et al. 2015). Geothermal inputs, evaporation and groundwater contamination are the main cause of high As concentrations in rivers (Mandal & Suzuki 2002; Mukherjee et al. 2006; Barringer & Reilly 2013). Instead, As concentration in seawater is usually less than 2 μg/L (Singh et al. 2015).

However, global attention is mainly focused on the presence of arsenic in groundwater, as the major part of the global population relies on groundwater to meet the purpose of drinking (Sarkar & Paul 2016). As concentration in groundwater is greatly influenced by the mineralogical characteristics of the aquifer and varies over a very large range, from less than 0.5 to 5,000 μg/L (Singh et al. 2015). As can find its way into groundwater by reductive dissolution of arsenic-rich iron oxyhydroxides present in dispersed phases in aquifers or by oxidation of arsenic-rich minerals (Sarkar & Paul 2016).

Natural groundwater containing As used for drinking, irrigation, or agriculture purposes causes serious health problems for people. It is estimated that 140 million people worldwide suffer from the consumption of arsenic-laden groundwater of higher than 10 μg/L (Li et al. 2017). Around 110 million of those people live in 10 countries in South and South-east Asia: Bangladesh, Cambodia, China, India, Laos, Myanmar, Nepal, Pakistan, Taiwan and Vietnam (Brammer 2008). Therefore, in 1993 the World Health Organization (WHO) reduced the acceptable threshold for As in drinking water from 50 to 10 μg/L, although some countries have adopted different values as national standards, i.e. 7 μg/L for Australia and New Zealand and 50 μg/L for Bangladesh (WHO 2011; Chieh Kao et al. 2013). The WHO arsenic guideline of 10 μg/L was introduced in Italy by the Legislative Decree 2001/31 following the adoption of European Directive 98/83/UE.

Numerous studies have been conducted to assess the toxicity of As and its effects on human health in various As-contaminated regions (Kongkea et al. 2010; Maity et al. 2012). Arsenic enters into human beings through two pathways: first, direct consumption of As-contaminated drinking water and second, for populations not exposed to elevated As in drinking water, foods represent the main sources of As intake for humans (Singh et al. 2015). Irrigation with As-enriched groundwater is the main pathway for As to enter the human food chain (Das et al. 2004; Chatterjee et al. 2010; Samal et al. 2011).

The various effects of As poisoning on human health are melanosis, leuco-melanosis, keratosis, hyperkeratosis, dorsum, non-petting edema, gangrene and skin cancer. Melanosis and keratosis are the most common presentations among the affected peoples (Singh et al. 2015).

Excessive and long-term (such as 5–10 years) human intake of toxic inorganic As having a concentration above 50 μg/L leads to arsenicosis, which is a common term used for As-related health effects including skin problems, skin cancers, internal cancers (bladder, kidney, lung), diseases of the blood vessels of the legs and feet, and possibly diabetes, high blood pressure and reproductive disorders (WHO 2011).

Arsenic can be removed from water with various techniques based on different chemical, physical-chemical or biological principles. The most consolidated technologies (well-known and widely applied around the world) are:

• Chemical oxidation and filtration (Dodd et al. 2006);

• Biological oxidation: oxidation of As(III) to As(V) by microorganisms and As(V) removal by iron and manganese oxides (Katsoyiannis & Zouboulis 2006; Sorlini et al. 2014);

• Co-precipitation: oxidation of As(III) to As(V) by adding a suitable oxidizing agent, followed by coagulation, sedimentation and filtration (Li et al. 2017);

• Electrocoagulation (Moussa et al. 2017; Song et al. 2017; Thakur & Mondal 2017; Silva et al. 2018);

• Adsorption on: activated alumina, activated carbon, iron-based sorbents, zero valent iron and hydrated iron oxide, etc. (Neumann et al. 2013; Thakur & Mondal 2017);

• Ion exchange through suitable cation and anion exchange resins (Nidheesh & Singh 2017);

• Membrane technology: reverse osmosis, nanofiltration and electrodialysis (Jain & Singh 2012);

• Photocatalysis (Xu et al. 2014).

Some of these processes can achieve good removal efficiencies for both As(III) and As(V), whereas chemical precipitation and ion exchange are particularly effective for the removal of As(V). Many of the arsenic treatment technologies require pH adjustment for optimization of performance and are effective in removing arsenic in the pentavalent state and hence include an oxidation step as pre-treatment to convert As(III) to As(V) (Bissen & Frimmel 2003).

The aim of this work is to analyze four case studies where the problem of arsenic contamination in groundwater is faced with an approach that includes, preliminarily, experimental activities at laboratory scale and/or at pilot scale.

In Case study 1, Sharma et al. (2016) investigated, through laboratory-scale tests, the mechanisms and possibilities for enhancing As removal from Bengal Delta groundwater with naturally occurring Fe in a reliable, low cost and sustainable way. In Case study 2, Sorlini & Gialdini (2014) carried out a series of experimental tests, at laboratory scale, for the simultaneous removal of arsenic, iron and manganese, in order to optimize the operating conditions of treatment processes applied at full scale. In Case study 3, Sorlini et al. (2015a), thanks to laboratory scale experimental activities and pilot scale testing, identified four different scenarios for upgrading of a drinking water treatment plant (DWTP) without a specific arsenic removal stage in the first configuration. In Case study 4, Collivignarelli et al. (2016) showed the results of a 19-month monitoring campaign for a pilot granular ferric oxide (GFO) filter, installed in a DWTP as a polishing stage with the aim of investigating the performance of GFO with arsenic concentrations in raw water close to the Italian regulatory limit.

The following paragraph is divided into two sub-paragraphs. In addition to the methodological approach used, preliminary information relating to each case study and the operative conditions of the activities carried out at laboratory scale and/or pilot scale are reported.

The methodological approach applied to counter the groundwater arsenic pollution is shown in Figure 1.

The Case study 1 experimental activity was developed only at laboratory scale. To simulate the concentration of contaminants in groundwater of the Bengal Delta, stock solutions of arsenic and iron were prepared. The first stage experiments were conducted to study oxidation and removal of As(III) with KMnO₄ in the presence of Fe²⁺ and Fe³⁺. Based on the results from the first stage, a second stage of experiments was conducted to improve oxidation and removal of As(III) with a basic solution of KMnO₄ + Al₂(SO₄)₃. Additional information, like the lab-scale equipment and the analytical methods, are reported in Sharma et al. (2016).

In Case study 2, the authors carried out their experimental activity with the aim of optimizing the full-scale treatment chain. The main contaminants in raw water are the following: iron (Fe, mean: 0.13 mg/L); manganese (Mn, mean: 128 μg/L); total arsenic (mean: 15 μg-As_TOT/L); trivalent arsenic (mean: 14.5 μg-As(III)/L₎ and ammonium (mean: 0.9 mg-NH₄⁺/L). These values are critical for iron, manganese, arsenic and ammonia since they are always higher than the legislation limit for water for human consumption (Italian Legislative Decree 2001/31).

The DWTP operates with a maximum flow rate of 1,600 m³/h (76,000 inhabitants) and is composed of the treatment processes reported in Figure 2.

Cascade aeration is applied for Fe oxidation; aerated biological filtration (bio-filtration) for Mn, Fe and arsenic oxidation and NH₃ nitrification; flash mixing is used for the dissolution of KMnO₄ and FeCl₃; sand filtration is necessary for the separation of Fe/Mn and the insoluble arsenic-containing compounds, and final disinfection with chlorine dioxide (ClO₂) is the last treatment before water distribution. The effects of a combination of treatments on the removal of the above-mentioned contaminants have been evaluated at laboratory scale. Water samples were collected from the outlet of the bio-filtration. At the end of each test, the following parameters were determined: As(III), total arsenic, total iron and total manganese. Analytical methods are reported in Sorlini & Gialdini (2014).

The DWTP analyzed in Case study 3 treats a maximum flow of 140 m³/h. The sequence of treatments is reported in Figure 3.

The plant is fed with water drawn from a 200-m deep aquifer. The main contaminants are ammonium (mean: 0.79 mg-NH₄⁺/L), manganese (mean: 79 μg/L) and total arsenic (mean: 12 μg-As_TOT/L). A pre-oxidation with air is used to oxidize iron (Fe) and form the precipitates, which can be removed by sand filtration; the biofiltration removes NH₄⁺ through biological nitrification and a second oxidation with ozone (O₃) allows completion of the oxidation of Mn that also partially occurs in the biofiltration. Then, insoluble manganese oxides and organic micropollutants are subsequently removed by granular activated carbon (GAC) filters. Finally, an in-line disinfection with ClO₂ is applied before water distribution. A 6-month intensive monitoring program was carried out, after which the biological activity of the sand and GAC filters was investigated (further details and results are reported in Sorlini et al. (2015a)). Moreover, different experimental tests, at laboratory and pilot scale, were performed to optimize arsenic removal through chemical oxidation and/or media adsorption. Analytical methods are reported in Sorlini et al. (2015a).

In Case study 4, the authors analyzed and commented on the results of installing a pilot GFO filter in a DWTP treating groundwater containing the following main contaminants: ammonium (mean: 0.7 mg-NH₄⁺/L), manganese (mean: 94 μg/L) and arsenic (mean: 11 μg-As_TOT/L). These concentrations are always higher than the regulatory limit for water intended for human consumption (Italian Legislative Decree 2001/31). Figure 4 reported the schematic diagram of the DWTP.

In the full-scale plant, contaminants are removed by a pre-aeration stage and two biofilters filled with sand (80%) and pyrolusite (20%). The latter, consisting essentially of manganese dioxide (MnO₂), was introduced for arsenic control. As can be observed from the Figure, the disinfection stage is not in operation in order to minimize the negative effects that can arise from the use of disinfectants such as chlorine dioxide (ClO₂). In fact, if ClO₂ is used for final disinfection, specific removal stages for chlorine disinfection by-products (e.g. activated carbon filtration) could be provided (Sorlini et al. 2015b, 2016). After a monitoring campaign, two main problems have been found: (i) arsenic removal in the biofilters was below 10% because the pyrolusite worked mainly as an oxidizer; (ii) the mass of pyrolusite in the biofilters fell due to backwashing. In agreement with the DWTP utility manager, the suitability of an additional arsenic removal (polishing) stage has been investigated in order to keep arsenic concentration as low as possible.

Water quality monitoring was conducted by taking an average of one sample per week throughout the 19-month test from both the inlet and the outlet of the DWTP. Additional information and analytical methods are reported in Collivignarelli et al. (2016).

The laboratory-scale experimental activities are reported in Table 1. Pilot-scale experimental activities are described in Table 2.

In Case study 1, the first stage experiments were conducted to study oxidation and removal of As(III) with KMnO₄ in the presence of Fe²⁺ and Fe³⁺. Based on the results from the first stage, a second stage of experiments was conducted to improve oxidation and removal of As(III) with a basic solution of KMnO₄ + Al₂(SO₄)₃. In Case study 2, the authors carried out a series of experimental tests in order to evaluate the effects of the following treatments on the removal of contaminants: pre-oxidation with KMnO₄, ClO₂ or NaClO; chemical precipitation with FeCl₃ or Al₂(SO₄)₃; sand filtration and final disinfection with ClO₂. In Case study 3, different experimental tests were performed to optimize arsenic removal through chemical oxidation and/or media adsorption. Additionally, pyrolusite was tested in a pilot scale plant. In Case study 4, the authors investigated the performance of a GFO for arsenic removal. Further information is reported in the individual publications.

In Sharma et al. (2016) the effectiveness of KMnO₄ on As(III) oxidation and the effect of the presence of Fe²⁺ on As(III) oxidation are investigated. Results showed that As(III) oxidation was concomitant with the oxidation of Fe²⁺ and was very rapid within the first 10 minutes of the experiment. Furthermore, with a SR = 1, complete oxidation of As(III) occurs within 2 minutes after the addition of KMnO₄ in the presence of Fe²⁺. The influence of the initial oxidation state of added Fe (Fe²⁺ or Fe³⁺ (as sulphate)) on As(III) oxidation and removal of As was also investigated. KMnO₄ was added 15 and 40 minutes after the addition of Fe. Results show that better oxidation of As(III) occurred in the presence of Fe³⁺ in comparison to Fe²⁺ and in the case of Fe²⁺ an extra dosage of KMnO₄ was required to achieve a final As(III) concentration below 50 μg/L.

It was therefore verified that the addition of KMnO₄ will not be sufficient if Fe²⁺ is present, so the removal of As with naturally occurring Fe can be greatly improved by adding adequate amounts of KMnO₄ to oxidize both As(III) and Fe²⁺. Conversely, if Fe²⁺ is present at low concentrations, adding an oxidant (KMnO₄) together with an auxiliary coagulant (Al) may be a practical solution, in order to achieve the filtration/sedimentation of flocs with adsorbed As. Then the authors focused on the detection of a solution with both oxidizing capacity and coagulation ability. Laboratory observation shows that a solution of KMnO₄ and Al₂(SO₄)₃ is stable for years at pH ∼ 11 adjusted with NaOH. Results for the applicability of this solution at varying Fe concentrations at an initial As(III) concentration of 500 μg/L are reported in Table 3.

Results reported in Table 3 show that higher Fe concentrations increase As(III) removal and As(III) concentrations below the Maximum Concentration Level (MCL) of 50 μg/L.

Using KMnO₄ together with Al₂(SO₄)₃ has three advantages: (i) the addition of Al ions increases the filterability of the Fe(OH)₃ flocs; (ii) the use of a strongly basic mixture allows the neutralization of the acidity produced due to Fe(OH)₃ and Al(OH)₃ formation; (iii) by increasing the amount of added Al, this method can be applied in waters with low Fe concentrations.

The results of the experimental activity carried out by Sorlini & Gialdini (2014) at laboratory scale showed that the first combination of treatments (pre-oxidation followed by final disinfection) is not effective in terms of As removal, because As(III) is solely oxidized into As(V). Conversely, the second combination (pre-oxidation followed by chemical precipitation, sand filtration and final disinfection) is the best option for the removal of both arsenic and the above-mentioned contaminants from water. The As_TOT removal yields obtained in this combination are shown in Table 4.

The combined use of KMnO₄ and FeCl₃ allows the highest As_TOT average removal yield to be obtained (80%, considering an average As_TOT concentration of 15 μg/L at the outlet of biofiltration). The same performance was highlighted by monitoring the full-scale DWTP operating in the same conditions tested at laboratory scale.

Sorlini et al. (2015a) carried out experimental tests at laboratory scale in order to identify specific treatments able to optimize arsenic removal. Results are reported in Table 5.

The results of the oxidation tests with KMnO₄ and NaMnO₄ on raw water show that a good oxidation efficiency (>80%) can be achieved with KMnO₄ after a contact time of 5 minutes and also with sub-stoichiometric dosages (<0.2 mg/L). Moreover, the results of tests carried out with the same dosage of different chemical oxidants (0.5 mg/L) have shown that NaMnO₄ led to a lower yield of oxidation than that obtained with KMnO₄ (42% with respect to 83%). Furthermore, the dosage of the chemical oxidant in water preventively aerated enables higher yields of oxidation (>85%) with both KMnO₄ and NaMnO₄, as part of As(III) was oxidized to As(V) during the aeration.

The adsorption of arsenic using pyrolusite was also investigated (data not shown). As reported in Sorlini et al. (2015a), good arsenic removal efficiency (>90%) was obtained on both raw and pre-oxidized water (with air). Moreover, no increase in the removal yield with KMnO₄ dosages has been observed.

In Case study 3, a monomedia filter composed of pyrolusite was tested by means of a pilot plant fed with pre-oxidized water with an average concentration of 10 μg-As_TOT/L (about 70% As(III)). The experimental results (reported in Sorlini et al. (2015a)) show that media saturation was reached quickly (before 2,000 bed volumes) even if the oxidation capacity of the filter media was still effective, as shown by the fact that after 6,000 BV the outlet concentration of As(III) was still about 50% of the total As concentration.

Instead, a GFO pilot filter, installed in the DWTP and monitored for 19 months, was evaluated in Case study 4. Figure 5 shows the total arsenic influent and effluent concentration trends.

It can be noted that the GFO performance decreases considerably about a month after the installation of the pilot filter. In fact, as reported in Collivignarelli et al. (2016), the arsenic-adsorption capacity of the E33 was substantially exhausted in the first 20,000 BV, when the ratio between the total arsenic concentration in the treated water (C) and the total arsenic concentration in the influent water (C₀) was 0.75. The authors also investigated the prevalent species removed by the E33 (As(III) and As(V)) in four different media response periods (periods that can be assumed to be homogeneous compared to operating conditions): 0–5,000; 5,001–20,000; 20,001–82,000; 82,001–140,000 Bv. For both species, the highest removal efficiencies were obtained in the first period (As(III) and As(V) average removal yields of 96% and 99% respectively), after which removal decreased progressively. At the end of the experimentation, the adsorption capacity of the GFO was estimated as 0.57 μg-As_TOT/L. Further details are reported in Collivignarelli et al. (2016).

The experimental tests carried out at pilot scale allowed a full-scale upgrading to be achieved.

In Case study 2, the results of the laboratory scale tests showed that the best removal of metals from raw water was obtained by pre-oxidation with potassium permanganate, chemical precipitation with ferric chloride, filtration and final disinfection with chlorine dioxide. Therefore, the optimum operating conditions determined at laboratory scale were implemented in the full scale DWTP, as reported in Figure 6.

The results of the continuous monitoring of the DWTP showed average arsenic concentration in treated water of 3 μg/L (removal yield of about 80%, considering an average arsenic concentration in raw water of 15 μg/L).

In Case study 3, according to the results obtained in experimental studies, four different scenarios of upgrading were identified and tested at laboratory scale. Among these, the fourth scenario (Figure 7) was chosen as this process was able to optimize the As_TOT removal (about 68%, considering an initial As_TOT concentration of about 12 μg/L). Further details of the other scenarios are reported in Sorlini et al. (2015a).

The results of the intensive monitoring of the upgraded DWTP showed that the average arsenic concentration in treated water was equal to 7 μg/L (removal yield of about 40%, considering an average arsenic concentration in raw water of 12 μg/L).

Regarding Case study 4, the results of continuous monitoring at the outlet of the DWTP are reported in Figure 8.

Each As_TOT concentration value reported represents the mean value of the quarter considered.

As can be seen from Figure 8, total arsenic effluent concentrations always comply with the MCL of 10 μg/L provided by the Italian Legislative Decree 2001/31.

Arsenic contamination in groundwater, used for drinking purposes, has been envisaged as a problem of global concern. The removal of arsenic can be ensured through the application of different processes. Chemical oxidation has been successfully tested, at laboratory scale, by Sharma et al. (2016). The authors found that oxidation of As(III) can effectively be obtained by adding KMnO₄ immediately after collecting the groundwater. Moreover, the treatment of Fe²⁺ and As(III)-laden groundwater can easily be achieved by the addition of a basic solution of KMnO₄ and Al₂(SO₄)₃. Starting from an initial arsenic concentration of 500 μg/L, the removal is better if the iron concentration is high: an As removal yield of approximately 100% can be achieved with an Fe concentration of 1,000 μg/L. Future goals concern the applicability of this solution in full-scale DWTPs that already have either a coagulation or filtration unit.

Sorlini & Gialdini (2014) found that, after pre-oxidation followed by final disinfection, As(III) was solely oxidized into As(V) and no removal was obtained. Considering an average total arsenic concentration of 15 μg/L in raw water, pre-oxidation, followed by chemical precipitation, sand filtration and final disinfection represents the best combination of treatments for the removal of this parameter. Moreover, the use of KMnO₄ and FeCl₃ resulted in being the optimum solution in this experimental train for improving arsenic removal. The monitoring of the full-scale DWTP operating under the same conditions tested at laboratory scale showed an average arsenic removal yield of 80%.

In Sorlini et al. (2015a) adsorption, in addition to oxidation, have been tested for arsenic removal by means of laboratory-scale and pilot-scale tests. Regarding the oxidation batch tests on raw water, the highest As(III) removal yield (93%) was achieved with a dosage of 0.1 mg-KMnO₄/L. Instead, regarding the oxidation tests on pre-oxidized water (with air), the highest As(III) removal yield (94%) was achieved with a dosage of 0.5 mg-NaMnO₄/L. Moreover, the adsorption of arsenic using pyrolusite showed a removal efficiency greater than 90%, as also reported in Katsoyiannis & Zouboulis (2006). The results identified four different scenarios of treatments for the plant upgrading. The fourth scenario resulted in optimization of As removal (about 68%) and, therefore, it was implemented in the full scale DWTP.

Among the numerous materials capable of effectively absorbing arsenic, the GFO has been tested by Collivignarelli et al. (2016) for the treatment of groundwater with an average total arsenic concentration of 11 μg/L. The results show that the GFO removed some arsenic up to 40,000 Bv (As(III) and As(V) average removal yields of 96% and 99% respectively), with a total adsorption capacity of 0.57 μg-Astot/mgGFO. Beyond this point, the GFO can be considered exhausted technically (C/C₀ = 0.9) and GFO replacement must be considered. From the economic point of view, GFO filter installation could be attractive because the operating costs – related to the arsenic concentration in the influent – are comparable to those of other physical-chemical processes. The effectiveness of the GFO filter is also confirmed by the full-scale monitoring: although this treatment is very easy to apply and manage in an existing treatment plant, the costs of purchasing the filter material and disposing of the exhausted material cannot be neglected.

Therefore, both experiments carried out at a pilot scale showed high performance in reducing arsenic, despite the concentrations found in raw water (respectively, 12 μg/L and 11 μg/L) being very close to the Italian regulatory limit.