Concurrent removal of nitrate, fluoride and arsenic by mixed hydrous bismuth oxide from water

The availability of safe drinking water has become a real challenge, especially in urban areas, and the majority of the population depends on groundwater for direct consumption. However, in recent years, the quality of groundwater that is used for drinking has degraded due to the addition of toxic elements (Velizarov et al. 2004). Thirty per cent of people are unable to access safe drinking water across the globe; hence providing safe drinking water has become a Sustainable Development Goal (SDG) which is set by the United Nations (Pradhan et al. 2017). However, there are some places where high levels of contaminants make the water unfit for direct human consumption. In the category of contaminants arsenic, fluoride and iron are naturally added, whereas the generation of nitrates, phosphates, heavy metals, etc., are due to human activities. Moreover, nitrate, fluoride, and arsenic have been reported as the most common and widespread contaminants of the groundwater. High concentrations of contaminants like nitrate, fluoride, and arsenic have been found increasingly in ground water, making it unfit for drinking purposes (Liu et al. 2012). The severe health hazard of arsenic is observed due to its extreme toxicity even at low concentrations (Sarkar & Paul 2016). Direct consumption of highly contaminated water is more hazardous to human beings because of toxicity. Hence, there is a dire need to maintain the quality of drinking water to avoid the ill effects of contaminants on human beings. Attention has been given to reduce the toxic influence of nitrate, fluoride and arsenic separately by several researchers but not on the concurrent removal of them (Jain et al. 2015; Kolomiyets et al. 2017; Mukhopadhyay et al. 2017).

Nitrate is one of the key ingredients used in nitrogenous fertilizers to increase the yield of the agricultural industry (Singh & Singh 2004). The elevated level of nitrate concentration in groundwater is directly correlated to the leaching effect of nitrate from intensive farming. The excessive nitrate in surface water bodies leads to eutrophication and degrades the quality of water which may have serious threats to human health including methaemoglobinaemia, cancer, increased infant mortality, gastrointestinal disturbances, etc. (Fewtrell 2004). A restriction of 10 mg/L NO₃⁻–N in drinking water was prescribed by the World Health Organization (WHO) and the United States Environmental Protection Agency (USEPA) due to its toxic health effects.

The occurrence of fluoride in groundwater is natural. Fluorine is the most electronegative element that is unstable and found in the elemental state. Due to its high reactivity, fluoride can attach with many naturally occurring minerals and erosion leads to an increase in the concentration of fluoride in groundwater. Excessive intake of fluoride usually has an adverse effect on the body metabolism, and also causes dental problems and skeletal fluorosis (Parashar et al. 2016). Fluorosis is an endemic public health problem which prevails in 32 nations around the globe (Mukhopadhyay et al. 2017). Acceptable fluoride concentration should be maintained in drinking water, especially among children below eight years of age. The beneficial and detrimental effects of fluoride in drinking water depend upon its concentration in aqueous solution. The permissible and acceptable provisions established by WHO for fluoride in drinking water is 1.5 and 1 mg/L, respectively.

Arsenite (As (III)) and arsenate (As (V)) are the two common forms of arsenic in water which exists naturally in groundwater. Arsenic is recognized by WHO as a first priority pollutant due to its toxicity and carcinogenic nature (Sarkar & Paul 2016). Various types of skin manifestations/disorders and other arsenic toxicity were observed such as melanosis, keratosis, hyperkeratosis, dorsal keratosis, and cancer (Cantu et al. 2016). The regulation of 0.10 mg/L of the arsenic level has been set by WHO. Also, USEPA reduced the maximum arsenic contaminant level from 0.05 to 0.10 mg/L in response to the serious known adverse health effects in 2001 (Luther et al. 2012). Based on the toxic and deleterious nature of nitrate, fluoride and arsenic, the following standards have been set for drinking water, as given below in Table 1.

To overcome the above-mentioned problem of anionic water contamination, massive efforts have been made. Various treatment processes are being used to remove these pollutants, and to procure safe drinking water (Bhatnagar & Sillanpää 2011; Habuda-Stanić et al. 2014; Lata & Samadder 2016). However, neither of these methods was able to remove the entire nitrate, fluoride, and arsenic from water. Water treatment by various methods can remove partial nitrate, fluoride and arsenic contaminants with varying efficiencies. Comparatively, the ion exchange or adsorption process has been considered as the most promising technique for the removal of the toxic anions from water based on their simplicity, effectiveness, recovery and relatively low cost (Leigen et al. 2019). Adsorption technology has been found to be successful in removing inorganic anions like nitrate (Teimouri et al. 2016), fluoride (Parashar et al. 2016) and arsenic (Kolomiyets et al. 2017) from water by using various adsorbents.

The adsorption properties of inorganic materials/hydrous metal oxides have been used in drinking water treatment applications. Hydrous metal oxides have been used for the removal of nitrate such as hydrous zirconium oxide (Dou et al. 2012), MgO nano-flakes (Zhang et al. 2012), ferric oxide (Mehrabi et al. 2015), hydrous bismuth oxide (Srivastav et al. 2014; Singh et al. 2015), magnesium oxide, cerium oxide, zinc oxide (Mahdavi et al. 2018), Zn/Fe LDH (Karthikeyan & Meenakshi 2020) etc., whereas activated alumina, manganese-oxide-coated alumina (Wang et al. 2007), Fe–Al–Ce tri-metal oxide (Delorme et al. 2007), hydrous-manganese-oxide-coated alumina (Teng et al. 2009), hydrous bismuth oxides (Srivastav et al. 2013; Singh et al. 2015), MgO (Jin et al. 2016), hydrous tin oxide (Parashar et al. 2016), MnO₂ (Mudzielwana et al. 2017), cerium(IV)- incorporated hydrous iron(III) (Mukhopadhyay et al. 2017), Mg/Ce/Mn oxide (Gitari et al. 2020) etc. have shown high affinity towards fluoride uptake from aqueous solution. Similarly, zirconium oxide (Suzuki et al. 2001), sand coated with iron oxide (Vaishya & Gupta 2002), ferric hydroxide (Daus et al. 2004), MnO₂ (Lenoble et al. 2004), alumina activated (Singh & Pant 2004), alumina coated by iron hydroxide (Hlavay & Polyak 2005), hydrous stannic oxide (Manna & Ghosh 2007), ferric hydroxide (Zhang et al. 2010), hydrous titanium dioxide (Pirilä et al. 2011), iron(III)-copper(II) binary oxide (Zhang et al. 2013), Mn-incorporated iron(III) oxide (Ghosh et al. 2014), TiO₂ (Yan et al. 2015), hydrous zirconium oxide (Pérez et al. 2016), iron cerium bimetal oxides (Wen et al. 2020) etc., are found to be efficient in the reduction of arsenic concentration.

However, activated alumina and activated carbon are the most widespread adsorbents and famous for their selectivity towards water contaminants. However, due to poor regeneration and incompetency for the concurrent removal of nitrate, fluoride and arsenic contaminants, it is imperative to develop new adsorbents which could give the higher as well as concurrent removal efficiency for all of these contaminants. However, an intense review of the literature was carried out for the compilation of anion exchange properties of these hydrous metal oxides which indicated that neither concurrent removal of nitrate, fluoride, and arsenic anions have been reported, nor have much of the reported materials shown concurrent sorptive properties towards nitrate, fluoride and arsenic. Hence, concurrent removals of these anions have been given the most priority in the present study.

Bismuth-based inorganic media have shown sorptive properties for nitrate, fluoride and arsenic. The hydrous bismuth oxide is the promising sorptive media for the removal of major anionic contaminates like nitrate and fluoride from ground water (Srivastav et al. 2013; Singh et al. 2015). Both forms of arsenic were tested and the removal capacity for As (V) was found to be higher than the As (III) in the pH range of drinking water condition. However, the maximum removal efficiency of As (III) could be achieved at higher pH (>9) which makes the water unfit for drinking. Hence, As (V) was taken for further experiments. High removal capability of hydrous bismuth oxide was observed for As (V) and fluoride with respect to nitrate where all three anions exist together. Hence, in the present study, hydrous bismuth oxide was investigated with the aim to evaluate the sorption efficiency for nitrate, fluoride, and arsenic (V) concurrently.

All the chemicals used in this study were of analytical grade. NaNO₃, NaF, and Na₂HAsO₄·7H₂O (E.Mark, Germany) salts were used for preparations of 1,000 mg/L stock solutions in deionized water (DI), respectively. Further, a working solution containing all three water contaminants of 95 mg/L of nitrate, 5 mg/L of fluoride and 0.08 mg/L of As (V) was prepared based on the average value of contaminants reported in groundwater (Delorme et al. 2007; Devi et al. 2008; Bibi et al. 2015).

Hydrous bismuth oxide (HBO) was made using bismuth trioxide (Bi₂O₃, MW = 466.0, Loba Chemie, Mumbai) powder of AR grade at a cost of 8,000 INR per kg in the Indian market. In the preparation of adsorbent, firstly ‘yellow bismuth hydroxide’ was synthesized as per the method reported by Fritsche (1993). The precipitate formed under the volumetric proportion of 1:1 (v/v) in Bi₂Cl₃ (0.1 M Bi₂O₃ dissolved 2 N HCl) with 2 N NaOH was designated as HBO1 (white color) and for HBO2, HBO3 (yellow color) in the ratio of 1:2 and 1:3 (v/v) was taken, respectively. A further reaction time of 1 h was given for homogenous mixing. Excess hydroxide and chloride ions were removed from the obtained precipitate by washing with double distilled water. Further precipitates were dried in an oven at 103 ± 2 °C for 24 h and powdered forms of HBO1, HBO2 and HBO3 were obtained which were subsequently used in the adsorption experiments.

The instruments used during the determination of nitrate and fluoride concentration are an ion-meter (HANNA HI 4222, Hanna Equipments Pvt., Navi, India) with an ion-selective electrode (ISE) as per standard methods (Clesceri et al. 2005). Arsenic concentration of the samples was measured by an Arsenator (Model no: Wag-WE10500 made by Wag-Tech, UK). pH measurement was carried out with the help of a digital pH meter (ITS 201, ISO-TECH system PVT, Varanasi, India). All experiments were carried out at standard temperatures (25 ± 2 °C).

Sodium nitrate, sodium fluoride and sodium arsenate (AR grade) were used for preparation of all working solutions. Three hours of contact time were given at room temperature (25 ± 2 °C) to adsorbent with the solution in a 100 mL Erlenmeyer flask with a 130 g/L dose of HBO_12mix. The contents were continuously stirred and further filtered using filter paper. The residual filtrates were then analyzed for the remaining concentration of anions. Initially, during contact time, the adsorbent dosage was optimized. Moreover, other parameters such as the effect of competitive anions on their removal were also investigated. Observations based on three triplicate samples were reported.

To avoid the choking condition in the column test, a mixture of sand and hydrous bismuth oxide (HBO1 + HBO2) was taken as adsorbent. To achieve the best result, a mixture of adsorbent with sand was taken, based on trial and error, into the column with a 3 cm internal diameter. The deionized water was passed through the column until all free chloride was released. Influent water contained a synthetic solution of 95 mg/L NO₃⁻, 5 mg/L F⁻ and 0.08 mg/L As (V). A flow rate of 1.2–1.5 mL/min was maintained while passing the influent through the column. A fraction of effluent (40 mL) was collected at regular intervals of 30 min for the determination of the residual concentration of pH, nitrate, fluoride, As (V) and chloride ions.

Adsorption kinetics was performed for a mixture of nitrate, fluoride and As (V) ions and was studied at room temperature. One hundred mL of anion solution was taken and 130 g/L of adsorbent (HBO_12mix) was added to the solution. The samples were collected at an interval of 5, 10, 20, 30, 45, 60, 75, 90, 120, 150, 180, 210, 240 min and the remaining nitrate, fluoride, As (V) were analyzed in treated water samples. Pseudo first order, pseudo second order, and Weber–Morris models were applied to the experimental results to generate kinetic parameters.

Adsorption isotherm was conducted by mixing 130 g/L of adsorbent (HBO_12mix) with 100 mL of anions solution with varying concentrations ranging from 70 to 95 mg/L for nitrate, 3–5 mg/L for fluoride and 0.08–0.10 mg/L for As (V). All the experiments were conducted at 250 rpm agitating condition with a contact time of 3 h at constant temperature (298, 308, and 318 K) to reach equilibrium. Langmuir, Freundlich, as well as DR isotherm models, were applied to the experimental results to understand the feasibility of the experiments.

Fourier transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) techniques were used to understand the mechanism of adsorption. The pH_PZC value was determined by the fast alkalimetric titration method as used by Srivastav et al. (2014). An FTIR spectrometer (Nicolet iS5, THERMO Electron Scientific Instruments LLC, USA) and an X-ray diffractometer (1710, Philips, The Netherlands) were used for FTIR spectrum and XRD patterns of hydrous bismuth oxide (HBO_12mix). The information regarding texture morphology and elemental analysis was obtained from a SEM and EDS (ZEISS EVO 18, Model-2045, Industrial Measurements and Medical Devices, Germany).

The literature suggests that hydrous bismuth oxide has a greater affinity for nitrate or fluoride, both in the drinking water pH range; but hydrous bismuth oxides have not been tested for arsenic sorption. HBO is expected to have a greater affinity for arsenic. Hence, to explore the simultaneous removal of nitrate, fluoride and arsenic it is imperative to determine the possibilities of arsenic sorption by hydrous bismuth oxide. Both forms of arsenic were tested and it was found that the removal capacity for As (V) was higher than As (III) from water in the given contact time. However, the maximum arsenic elimination required can be achieved by the removal of As (III), which needs to be dissociated in ionic form (H₂AsO3⁻) at higher pH (pH > 9) or oxidation into As (V). Hence oxidation of As (III) into As (V) could lead to the best result in concurrent removal at higher pH. However, increments in pH (>9) make it unfit for drinking purposes. Lesser efficiency will be achieved by selected adsorbents in contaminant removal when drinking water obtained from groundwater aquifers in a reducing environment. Groundwater obtained from an oxidizing environment could be used as an adsorbent for higher removal efficiency. Hence, arsenate was chosen for further experiments.

Table 2 summarizes the affinity of hydrous bismuth oxide in the individual removal of water contaminants.

The concurrent removal of nitrate, fluoride, and As (V) from water using different HBOs powder was performed through batch experiments. It was observed that HBO1 did not show a good capability to remove nitrate from water in comparison to fluoride and As (V) removals. Only 19.07% of nitrate removal was achieved at a dosage of 100 g/L by HBO1 with 95 mg/L initial concentration, whereas 60% of nitrate removal was observed by using HBO2 and HBO3, respectively, as an adsorbent with the same conditions at an adsorbent dose of 100 g/L. No inference of hydroxide ion was observed in treated water samples as pH falls in the range of drinking water standards from 6.5 to 8.5. Hence, it can be understood that due to nitrate adsorption, no significant release of hydroxide ions from adsorbent was observed.

Similarly, the use of HBO1 was found to be beneficial in the removal of fluoride as compared with HBO2. A maximum of 67% of fluoride removal was achieved at the dosage of 100 g/L with an initial concentration of 5 mg/L, whereas 24 and 17% were observed with HBO2 and HBO3, respectively, at the dosage of 100 g/L. Further, the pH variation of treated water was recorded in the permissible range of 7.0–8.5. Hence, again hydroxides were not observed as a caustic agent for fluoride sorption.

All HBO adsorbents have good capability to remove As (V) with comparisons to nitrate and fluoride from water. HBOs (HBO1, HBO2 and HBO3) were found to be equally efficient in As (V) removal with an average of 91.49% efficiency. Likewise, for the nitrate and fluoride removal, no significant changes in hydroxide ion were observed and hence hydroxide could not responsible for As (V) sorption.

Figure 1 indicates that HBO1 was found to be efficient for both fluoride as well as As (V) removal with high dosages, whereas HBO2 and HBO3 have more affinity towards nitrate and As (V). Hence, based on the above observations, the mixture of HBO1 and HBO2 (or HBO3) seem a good adsorbent for the concurrent removal of nitrate, fluoride and arsenic (As (V)) from the water. Cost analyses in the preparation of adsorbent restrict us to select HBO3 as both adsorbents (HBO2 and HBO3) reflect approximately the same adsorption capacities. Hence, further studies have been carried out by taking a mixture of HBO1 and HBO2 (HBO_12mix) as a single adsorbent for the concurrent removal of major anionic contaminants (nitrate, fluoride and As (V)) from aqueous solution.

An equal mixed proportion of HBO1 and HBO2 were taken as HBO_12mix for further experiments. Dosage was varied from 10 to 150 g/L for 3 h of contact time with continuous stirring at 250 rpm. After 3 h of contact time, filtered water samples were analyzed for residual pH, chloride, nitrate, fluoride, and As (V) concentrations. The removal of all three anions with HBO_12mix was tested and a maximum removal efficiency of 52% for nitrate, 72% for fluoride and 92% for As (V) was found with a marginal increment of pH from 7.34 to 8.23 after treatment. The nitrate concentration was reduced from 95 to 45.6 mg NO₃⁻/L, whereas the fluoride concentration decreased from 5.11 to 1.43 mg/ L and the As (V) concentration dwindled from 0.1 to 0.008 mg/L with 92% removal efficiency. Almost similar results were observed by using HBO_13mix as adsorbent in concurrent removal of three anions. Maximum uptake was reported at a dosage of 130 g/L for nitrate and fluoride, whereas for As (V), 20 g/L was observed as the optimum dosage for maximum sorption (Figure 2). Further increase of the dosage had no significant effect on removal efficiency. Also, major changes of hydroxide ion concentration were not observed in treated water but the final value of chloride concentration reached 53.7 mg/L which has been marked as a major change from 16.3 mg/L as initial chloride concentration. Hence, elution of chloride could be considered as a reason for the sorption of anions on HBO_12mix media.

After adsorbent (HBO_12mix) dosages of 130 g/L, there was no gain observed in the adsorption capacity of nitrate, fluoride, as well as As (V) from water. Similar findings were also reported by some authors including Swain et al. (2012) and Srivastav et al. (2013) during the adsorption of fluoride from water. The saturation of active adsorption sites could be due to them overlapping on themselves (Swain et al. 2012) and, because of that, afterwards insignificant removals were observed. Therefore, 130 g/L of adsorbent was considered as the optimum dose for further experiments.

The effect of time has also been included during the present study. The effect of contact time was studied for a total of 5 h and samples were collected at regular intervals of 1 h (Figure 3). The enhanced removal efficiency was observed with increments in initial contact time up to 3 h; 32.44, 45.53 and 53% of removal percentage for nitrate were recorded along with 47.55, 62.23 and 72% of removal efficiency for fluoride and 51.25, 80 and 92% of As (V) removal at regular intervals of 1, 2 and 3 h. Initially, up to 3 h, the uptake of anionic contaminants was very high as the pore site was active and unsaturated. The percentage removal of 53% nitrate, 72% fluoride and 92% As (V) were recorded at the time up to 3 h. The slope in Figure 3 shows a steep negative slope up to 3 h and further as the time progress slope becomes constant, which concludes the non-availability of the adsorbent site for uptake ions and it could be due to the availability of an abundant number of adsorption active sites during the initial phases of adsorption (Khan et al. 2009). Hence, after 3 h the marginal changes in uptake could be noticed as the pore site became saturated and exhausted.

However, after an optimum time period, development of repulsive force between the adsorbate and bulk phase takes place which discourages the rate of further adsorption.

Particle sizes of hydrous bismuth oxide powder are very fine in nature. In order to use in column adsorption study, hydrous bismuth oxide powder (HBO1 + HBO2: 50 g +200 g) and sand (150 g) was mixed in different proportions to obtain efficient results. The sand to be used was initially cleaned by keeping it in 0.1 N HCl for 24 h and then washing it with sufficient distilled water. A glass column (diameter of 30 mm) was filled up to a height of around 51 cm with the mixed media (D₁₀ = 0.18, Cu = D₆₀/D₁₀ = 2.4, pH ∼ 7.0). A synthetic solution of 95 mg/L nitrate, 5 mg/L fluoride and 0.08 mg/L As (V) was taken for column study which is representative of the field samples of anionic contaminated groundwater in parts of the country. The contaminants spiked water samples were passed with a flow rate of ∼1.2–1.5 mL/min. Samples of effluents were collected at an interval of every 30 min. pH, nitrate, fluoride, As (V), and chloride were measured to understand the mechanism of removal.

As shown in Figure 4, zero effluent concentrate was found as the synthetic anionic contaminated water passed through the column for 6 h, then the nitrate concentration rapidly rose to 90.23 mg/L at 1,320 min after the passage of 1,040 mL of influent, which means the column is saturated by nitrate ions first. Consequently, at the treatment time of 1,680 and 2,970 min, the concentrations of fluoride and As (V) in the effluent reached 4 and 0.071 mg/L after the passage of 2,160 and 3,920 mL of influent, respectively. The column adsorption experiment shows that approximately 750 mL of contaminant free drinking water could be used by the consumer, whereas 1,180 ml (fluoride and arsenic free) and 3,512 mL (arsenic free) of drinking water could be used by the household. The column adsorption experiment shows that dynamic adsorption capacities of nitrate, fluoride, and As (V) ions onto the adsorbent are 46, 4.3 and 0.29 mg/g respectively. A marginal change in pH could be observed but continued elution of chloride was monitored which diminishes as the adsorbent becomes exhausted. An indication of ion exchange could be predicted based on the results observed above.

Fritsche (1993) detected the simultaneous removal of nitrate and other anions by yellow bismuth hydroxide and also reported that OH⁻ is the caustic agent responsible for nitrate sorption through ion exchange mechanism. The earlier work of Ruixia et al. (2002) on the removal of fluoride, phosphate and As (V) by ion exchange fiber elucidates the dynamic adsorption capacities of 45.0, 155.5, and 96.0 mg/g. The ion exchange of chloride is the dominant sorption mechanism for fluoride and phosphate in synthetic wastewater. Park et al. (2008) conducted a series of experiments for the simultaneous removal of fluoride, phosphate and nitrate by cement paste column. The cement paste column could treat fluoride and phosphate to remarkably low levels with removal capacities of 1.67 and 1.96 meq/g, but nitrate entrapment is difficult as it makes it complex with cement cations but significant removal of nitrate was achieved by ion exchange with sulfate released. A similar attempt was made by Streat et al. (2008) for the adsorption of arsenic, phosphorus, and fluorine by hydrous ferric oxide as an adsorbent in water treatment. Greater adsorption of arsenate and phosphate was observed than of fluoride due to protonation. Singh et al. (2012) considered hydrous bismuth oxide (HBO1, HBO2 and HBO3) as adsorbents and studied the removal of nitrate in the presence of bicarbonate and sulfate. The authors concluded that HBO2 and HBO3 are found still very well in the nitrate removal, even in the presence of bicarbonate and sulfate but the removal percentage of alkalinity significantly dwindled by HBO1 from 77 to 22% in the column test.

Regeneration of hydrous bismuth oxide was also attempted for its further use in anion removal with 0.1 N NaOH followed by 0.1 N NaCl solutions, although only by passing the 0.1 NaOH through media leads to elution of sorbed anions. Also, it has been observed that chloride is the exchange anion responsible for anion uptake by media. Hence it was expected to regenerate the media by passing through the 0.1 NaCl solutions. With respect to regeneration experiments two significant changes have been marked down: first, the media has been dehydrated and secondly the regenerant has been changed from 0.1 NaOH to 0.1 NaCl. The elution of sorbed anions by chloride during regeneration may be due to higher preferences and concentration gradient effects. Sodium hydroxide has the advantage of affecting the pH (>9) of the water environment, thereby affecting the sorbed ions and consequently the empty and active sites are occupied by sodium chloride. Sixty per cent anion content was obtained in desorption and further regenerated adsorbent was used in up to two consecutive cycles of anion removal (Figure 5).

It is important to note that effluent qualities exceed their respective permissible limits after treating around 200, 520 and 1,200 mL of nitrate, fluoride and As (V) in the second cycle, whereas 120, 320 and 1,000 mL quantity of nitrate, fluoride and As (V) were achieved before exceeding the permissible limits in the third cycle.

The concentration of bismuth ion was also not detected in the treated water by an Atomic Absorption Spectrophotometer (AAS), which indicated that no bismuth ions are present in treated water as already reported by Srivastav et al. (2014).

To simulate the real field conditions, it is important to access the properties of adsorbent in the presence of competitive anions. Sulfate and bicarbonate are the most common ions present in groundwater. Hence, these ions were considered as a hindrance and the effect of sulfate and bicarbonate ion presence was studied on the sorption capacity of HBO_12mix with an adsorbent dose of 110 g/L with 3 h contact time. A concentration range of 1–5 meq/L sulfate (44–220 mg/L as SO₄²⁻) and bicarbonate (50–250 mg/L as CaCO₃) were tested on concurrent removals of nitrate (95 mg NO₃⁻/L), fluoride (5 mg F⁻/L) and As (V) (0.08 mg/L As (V)). It can be observed from Figure 6 that the bicarbonate ions imposed a relatively greater negative impact on nitrate removal whereas sulfate ions affected the removal of fluoride and As (V). The pH of the solution was greatly affected by increased bicarbonate concentration which could be related to the lower adsorption of nitrate in aqueous solution (Goswami & Purkait 2012). During adsorption, the multivalent nature of sulfate anions has been preferred over monovalent anions (Mohapatra et al. 2012). The nitrate removal efficiency of HBO_12mix was remarkably decreased from 55 to 12% whereas fluoride and As (V) removal efficiency decreased from 72 to 25% and 90 to 50%, respectively, at 5.0 meq/L of each anion concentration. Hence, in the presence of competitive anions (HCO₃⁻ and SO₄²⁻), the preference of adsorption has been changed which reduces the outcomes of preference series of anions appear as: bicarbonate > sulfate > nitrate, whereas for fluoride and As (V) removal HBO_12mix, the preference is sulfate > bicarbonate > fluoride/As (V).

Thus, HBO_12mix is found potentially useful for the sorption of nitrate, fluoride and As (V) below an alkalinity of around 150 mg/L as CaCO₃ and sulfate concentration of 140 mg/L.

A batch experiment was carried out by taking water samples from the IIT (BHU), Varanai, India, added with nitrate conc. = 95 mg/L, fluoride conc. = 5 mg/L, As (V) conc. = 0.08 mg/L. The initial concentration of solution was applied for the batch adsorption experiment with adsorbent as HBO_12mix dose of 130 g/L for 3 h of reaction time. The outlet quality was analyzed for pH, nitrate, fluoride, As (V), alkalinity, sulfate and chloride. Table 3 provides the inlet and outlet qualities from ground water after the batch experiment.

A major reduction in alkalinity could be seen by HBO_12mix in real ground water conditions. However, fluoride and As (V) removal efficiencies have not been affected but a high reduction in nitrate sorption was observed. pH remained in the desired range and the chloride level was quite variable in the treated water. The experimental data from Table 3 deduces the idea of alkalinity being more dangerous and competing anions for nitrate as nitrate removal is being hindered by a high level of alkalinity.

In order to perform the kinetics study, the nitrate, fluoride and As (V) concentrations were varied on three levels (Table 4) and kinetic parameters were evaluated by keeping the reaction of 3 h and adsorbent dosage of 130 g/L of HBO_12mix.

The linear plot of kinetic data and parameter of pseudo first order, pseudo second order, and Weber–Morris models are demonstrated in Table 4 and Figure 7. Based on higher correlation coefficients, the process of sorption on HBO_12mix is governed by the pseudo first order reaction. The rate constant determined by the pseudo first order reaction for fluoride and As (V) was found to be higher than that of the nitrate. The rate constant (K₁) increases with a decrease in initial concentration except for nitrate, while the higher rate constant (K₂) was achieved by pseudo second order for fluoride sorption.

In the solid–liquid adsorption system, the Weber–Morris model was taken to illustrate the diffusion mechanism and rate-controlling process of adsorption. The regression coefficient of the process lies in between 0.913 to 0.954, suggesting that the rate-controlling steps were not governed by a single process.

Further, from the straight line and non-zero intercept plot from the Weber–Morris model, it can be interpreted that the adsorption of anions on the surface of HBO_12mix is a complex process and is governed by surface adsorption as well as intraparticle diffusion (Ayoob & Gupta 2008; Gupta & Bhattacharyya 2011).

In order to perform the isotherm study, mixed solution of nitrate, fluoride and As (V) concentrations were varied at different temperatures (298, 308 and 318 K) (Table 5) at a reaction time of 3 h with an adsorbent dosage of 130 g/L.

Equilibrium data and isotherm data plot for the adsorption of anions on HBO_12mix at temperatures of 298, 308, and 318 K are shown in Table 5 and Figure 8, respectively. Inspection of the data revealed that the agreement of isotherm data fitted well with the Freundlich isotherm for anion sorption. Sorption isotherm of nitrate tries to synchronize with fluoride and As (V) isotherm at higher temperatures, whereas As (V) adsorption data appear to follow the DR isotherm equation more closely than others. Increased temperature leads to an increase in the threshold energies of solute particles which results in more molecular collision at higher temperatures. Hence, the removal process was considered as endothermic nature of adsorption (Ayoob & Gupta 2008).

The values of mean free energy of sorption (E) calculated for the present cases are shown in Table 5 with other evaluated parameters.

From the DR equation and according to the value of E, an adsorption process may be one of three types: (i) Physisorption if E ≤ 8.0 kJ mol/L; (ii) Ion-exchange if 8.0 < E < 16.0 kJ mol/L; and (iii) Chemisorption if E ≤ 16.0 kJ mol/L (Saha et al. 2016). It has been found that the E value (kJ/mol) is 7.8–14.22. These E values indicated that the anion sorption by HBO_12mix is dominated by ion exchange phenomena followed by physical adsorption.

Values of ‘M’ are obtained by plotting In(q_e/C_e) versus q_e and extrapolating to zero q_e as per the methods suggested by Khan & Singh (1987) and Basu et al. (2010). The intersection with the vertical axis gives the value of ‘M’. Further, the value of ΔH° and ΔS° were calculated from the slope and intercept from the graph between lnM and 1/T (Liu & Liu 2008) as summarized in Table 5.

The values of ΔG° suggest that the spontaneous nature and higher temperature favors anion sorption. The positive value of ΔH° suggests that adsorption of anions over HBO is endothermic in nature. The magnitude of ΔH° if lies in between 2.1 and 20.9 kJ/mol the adsorption process can be accounted as physical while chemical adsorption generally falls into a range of 80–200 kJ/mol (Banerjee et al. 2014). In this case the ΔH° value corresponds to the intermediate range between physical and chemical adsorption. Thus anion sorption onto HBO_12mix is ion exchange in nature.

In order to characterize the materials for their physical and chemical combination properties, X-ray diffraction pattern, scanning electron microscopy, energy dispersive detector, FTIR spectroscopy analysis and pHpzc determinations were performed.

In order to understand the characteristics and dominant chemical species in HBO_12mix, XRD analysis was performed. The presence of multiple peaks in XRD patterns indicates the broadly crystalline character of adsorbent (Ramli et al. 2007). The 2θ peak values (26.3, 30.5, 33.4°) for the XRD peak corresponds to Bi₁₂O₁₅Cl₁₆ (29-0237) by using X'pert high score software (2009). The additional 2θ peaks values of 23.92, 30.16, 29.94, 32.84 and 33.42° refer to Bi₆O₇Cl₁₃ (70-0854), Bi₁₉AsO₃₁ (46-0192), Bi(ClO₄)NO₂ (43-0759) respectively.

A comparison for XRD peaks of HBO_12mix before and after adsorption (Figure 9) indicated that new peaks appeared due to the adsorption of nitrate, fluoride and As (V) on the surface of the adsorbent. This could be attributed to the adsorption of nitrate, fluoride and As (V) on the surface of HBO_12mix with polymeric character. Reductions in the chloride peak after adsorption could be the result of the ion exchange process with chloride ions with target anions. Similar results are attributed by Huang et al. (2015) in his studies. Hence, based on the above results, it can be inferred that the ion exchange between HBO_12mix and anions present in solution was the main adsorption mechanism.

The behavior of adsorbent can be easily predicted by knowing its point of zero charge (pHpzc) value. The pH_pzc represents a condition in which the surface hold equal electrical charge density of positive and negative ions because of the neutral condition when the pH is lower than the pzc value, the system is said to be ‘below the pzc’. Below the pzc, the acidic water donates more protons than hydroxide groups, and so the adsorbent surface gets positively charged (attracting anions) and hence it favors the sorption of target anions. Conversely, above pzc the surface is negatively charged (attracting cations/repelling anions) and favors the sorption of positive anions (Mukhopadhyay et al. 2017).

To know the effect of pH on the sorption process, the point of zero charge (pH_PZC) was measured. The pH of point of zero charge of adsorbent was determined using the fast alkalimetric titration method (Agarwal et al. 2006). The pH_PZC of HBO_12mix is found to be 8.0 ± 0.2 (Figure 10) where the net surface charge becomes zero as the surface of adsorbent becomes neutral attributing to equal density of negative and positive charge. Hence suitability and possibility of anionic uptake (nitrate, fluoride and As (V)) favors only with the condition of pH < pH_PZC (8.0 ± 0.2) due to columbic forces of attraction. Similarly, the force of repulsion will act as pH > pH_PZC (8.0 ± 0.2).

SEM analysis has been carried out to determine the surface morphology of adsorbent. Figure 11 shows the irregular surface morphology of the adsorbent. The SEM spectrum indicates the granular and spongy formation which later becomes exhausted with a heterogeneous and rough surface, whereas the EDS spectrum analysis demonstrated that the adsorbent is composed of Bi and Cl and the presence of target anions were also confirmed by EDS after adsorption (Figure 12). Hence, the existence of nitrate, fluoride and arsenic on the surface of exhausted adsorbent in addition to bismuth and chloride were detected from EDS analysis which is also in agreement with our data findings in XRD analysis.

FTIR stands for Fourier Transform Infrared, the preferred method of infrared spectroscopy. The FTIR spectrum of prepared adsorbent before and after adsorption is depicted in Figure 13. The distinctive absorptions bands corresponding to anions contaminants attached to the matrix of HBO_12mix after adsorption. The broad absorption peaks of 530 and 846 is due to the Bi-O bond (Fruth et al. 2004; Ramli et al. 2007). Further, water molecules and hydroxide were spotted at peaks of 1,380, 1,620 (Sujana & Anand 2011) and 3,430 (Diez et al. 2000) respectively. In addition to these, various new peaks were appeared at 441, 530, 675, 1,045, 1,274 and 1,310. Bervas et al. (2006) and Escudero et al. (2014) specified absorption bands corresponding to wave numbers 441 and 530 cm⁻¹ only as characteristics of the Bi-F bond. The sharp peak at 1,274 and 1,310 cm⁻¹ could be due to the existence of BiNO₃ (Abdullah et al. 2012a, 2012b). In addition, the frequency at 675 to 1,045 cm⁻¹ shows characteristics of the As-O bond (Pena et al. 2006; Snehlata & Samadder 2015). The FTIR detected new peaks as a result of nitrate, fluoride and arsenic adsorption. The distinctive absorption bands corresponding to nitrate, fluoride and arsenate in addition to those of bismuth and hydroxyl ions indicate that the target anions have become an integral part of the adsorbents. FTIR analysis of the tested materials supports the results obtained through XRD and EDS analysis.

An attempt was made to compare conventional adsorbent based on the cost of preparation as the cost of synthesis of the adsorbent is the major factor for its selection in adsorption study. The overall cost of HBO_12mix should not be very high for its large scale preparation and application. The preparative cost of the reported sorptive material (HBO_12mix) appears comparable with some of the adsorbents used in the adsorption process. The overall cost (US$/kg) for the preparation of different adsorbents is given in Table 6.

The cost of adsorbent mainly depends on the scale of application, reuse potential, sorption potential and synthesis of adsorbent. The current study demonstrated the concurrent removal efficiency of nitrate, fluoride and As (V) for up to three cycles. The adsorbent will be economically viable as it can be regenerated/reactivated for many cycles after it has been exhausted. Regeneration of HBO_12mix adsorbents is quite possible for up to 3–4 cycles. Moreover, most likely the cost will also get reduced when it will be synthesized on a large scale.

Electrostatic force and ion-exchange on the protonated surface of adsorbent could be responsible for nitrate, fluoride, and As (V) adsorption with chloride release (Ruixia et al. 2002). Hence, release of Cl⁻ from the hydrous bismuth oxide (HBO_12mix) during the sorption process could lead to the conclusion that chloride ions adhere to the surface with weak interaction and was easily replaced at suitable conditions (pH < 8.0 ± 0.2). Ion exchange could be seen as a dominant sorption mechanism for concurrent removal of nitrate, fluoride and As (V). The risk of bismuth leaching in treated water was also taken into account during the present study and the results show that most bismuth ions were firmly fixed within the matrix of hydrous bismuth oxide (HBO_12mix). Hence, adsorbent could be considered as a type of environment-friendly material for nitrate, fluoride, and As (V) pollutant removal from water.

The performance of each hydrous bismuth oxide was evaluated for concurrent removal of nitrate, fluoride and As (V) removal from aqueous removal. The HBO1 was best for removal of fluoride and As (V) whereas HBO2 was best for removal of nitrate and As (V) at a higher dosage. Removal of all three with HBO_12mix were tested and it was found that the maximum removal was 52% for nitrate, 71% for fluoride and 92% for As (V), observed at a dosage of 130 g/L with a contact time of 3 h. The column adsorption experiment shows that approximately 750 mL of contaminant free drinking water could be directly availed by consumers whereas 1,180 mL (fluoride and arsenic free) and 3,512 mL (arsenic free) of drinking water could be used by households. The interaction between target anions and bismuth bounded chloride played a vital role by the ion exchange mechanism. XRD, SEM, EDS and FTIR results confirm the integration of target anions into the matrix of HBO_12mix. No secondary pollution was observed due to the elution of bismuth in treated water. Hence, HBO_12mix can be considered as a potential adsorbent for the elimination of nitrate, fluoride and arsenic concurrently. This may be used as a potential sorbent for developing a point of use (POU) water treatment unit for the removal of nitrate, fluoride and arsenic existing in concurrent conditions.

All the necessary facilities for this work were provided by the Indian Institute of Technology (Banaras Hindu University), India.

All relevant data are included in the paper or its Supplementary Information.