An evaluation of the effects of calcination temperature and solution pH on the fluoride removal capacity of Al/Fe oxide-modified diatomaceous earth was carried out. The sorbent was observed to be most effective and stable within the pH range 6.70–8.12, where the lowest concentrations of Al and Fe (<1 mg/L) in treated water were recorded. Thus, sorbent loss was minimal at that pH range. It was observed that the sorbent lost its efficiency and stability at calcination temperatures above 600 °C. Inductively coupled plasma-mass spectrometer analysis of metals in supernatants and thermogravimetric analysis of the sorbent showed that there was substantive loss of Al and Fe from the sorbent at temperatures above 600 °C because of evaporation. K2SO4 solution proved to be the best regenerant for spent sorbent compared to NaOH and Na2CO3, which caused sorbent loss owing to high solution pH. The CO32– from Na2CO3 bound to regenerated sorbent so much that it could not be displaced by fluoride during subsequent defluoridation experiments. Sorbent regenerated with 0.1 M K2SO4 solution could reduce 10 mg/L fluoride in artificial water at a dosage of 0.8 g/100 mL by 81.8% and 67.2% at the second and third cycles, respectively.
Groundwater serves as source of drinking-water in many rural communities where there is no pipe-borne water. But long-term ingestion of groundwater containing fluoride in excess of the WHO guideline of 1.5 mg/L causes dental fluorosis. Skeletal fluorosis occurs as a result of long-term ingestion of water containing 3–6 mg/L fluoride (WHO 2011). However, a minimum fluoride concentration of 0.5 mg/L is required in drinking-water to prevent dental caries in children and adults (WHO 2011). Where there is no appropriate alternative drinking-water besides high-fluoride groundwater, it is crucial to reduce the fluoride content to a safe limit.
This study was premised on the use of the adsorption technique as the most suitable method for groundwater defluoridation in households in rural communities with Al/Fe oxide-modified diatomaceous earth (DE) as an adsorbent. DE is an earthy, very finely porous, siliceous sedimentary rock. It is chemically inert in most liquids and gases and has low thermal conductivity (Antonides 1998; Mohamedbakr 2010). Among many uses of DE, the natural material has been used as filtration medium for beer and wine because of its very fine pores (Antonides 1998). DE is not toxic, based on the report of the elemental analysis of treated water using raw DE (Izuagie et al. 2015). The chemical formula of DE is SiO2·2H2O. Not being organic, the material does not have the tendency of undergoing putrefaction to foul water.
A study by Yuan et al. (2004) showed that DE has hydroxyl groups in its chemical structure (SiO2·nH2O) which could exchange for electronegative fluoride. It has been reported that the fluoride uptake capacity of raw DE is very poor, even at the most favourable solution pH of 2 (Izuagie et al. 2015). Hence, surface modification of DE with high fluoride affinity metal oxides has become expedient to achieve high fluoride removal from water. The metal oxide modification occurred through the coating of the pores of DE with Al/Fe oxide.
Adsorbent is the key focus in adsorption technology. The cost and the processes involved in the development of a high-performance adsorbent cannot be overemphasized. It is therefore necessary to understand the chemical state and characteristics of an adsorbent at different calcination temperatures and solution pH; not only to achieve a maximum contaminant uptake from water but also considering the reusability of the exhausted sorbent.
The pH of solution plays a principal role in determining the amount of fluoride adsorbed from solution (Kagne et al. 2009; Manna et al. 2015; Saha et al. 2015; Xu et al. 2015). However, the most suitable solution pH for the removal of a contaminant may correspond to the pH at which the highest leaching of an active metal component in the sorbent occurs. A pH range of 3 to 6 was reported for the highest adsorption of phosphate from contaminated water using Fe–Mg–La tri-metal composite sorbent (Yu & Chen 2015). While the pH range appeared most appropriate for phosphate removal from water, a high dissolution of lanthanum from the sorbent was observed at pH < 5. Hence, loss of sorbent was most pronounced at pH < 5. Sakhare et al. (2012) reported the use of calcium aluminate material for defluoridation of water where significant amounts of calcium and aluminium were released from the sorbent under acidic and alkaline conditions, thereby resulting in sorbent loss. However, at a neutral pH, the amount of aluminium leached from the calcium aluminate material was within the acceptable limit of 0.2 mg/L. The loss of adsorbent as a result of unfavourable solution pH undermines the reusability of sorbent for subsequent contaminant removal from water.
Calcination of sorbents at different temperatures is usually intended to improve their adsorption capacity (Wang et al. 2015). Calcination above a temperature limit may affect the phase and chemical composition of some sorbents. An example is the case of the change of phase of bauxite calcined at 400 °C and 500 °C (Sadiju et al. 2008). Bauxite calcined at 200 °C was reported to have a better adsorption capacity than at any other higher temperature. The thermogram showed that at 400 °C, gibbsite (Al(OH)3) had started to transform to boehmite (AlOOH), while the kaolinite structure was completely destroyed at 500 °C, resulting in a highly amorphous phase that could be ascribed to different activated aluminas (Sadiju et al. 2008).
DE is available in large deposits around the world (Lemonas 1997). Perhaps for this reason not much is reported about the reusability of its modified species. Being an amorphous material the modified form may be considered to be less apt to withstand the reaction rigours of the crystalline sorbents (Volzone & Garrido 2008; Ma et al. 2011; Tomar et al. 2013). This study therefore was undertaken to evaluate the effect of calcination temperature and solution pH on the defluoridation potential of Al/Fe oxide-modified DE and also to find an appropriate reagent for regenerating the spent sorbent for subsequent fluoride removal from water.
MATERIALS AND METHODS
Effect of contact time and sorbent dosage on fluoride removal
The effect of contact time on adsorption of fluoride on Al/Fe oxide-modified DE was evaluated by contacting 0.1 g of the sorbent with aliquots of 100 mL of 10 mg/L fluoride solution at 200 rpm for 5, 10, 20, 30, 40, 50, 60 and 70 min. The experiment was repeated using sorbent dosages of 0.2 g and 0.3 g. Fluoride analysis of the centrifuged mixtures showed that the highest fluoride removal occurred at the 50 min contact time. The results of fluoride in supernatants obtained when mixtures of sorbent dosages of 0.1 to 1 g and aliquots of 100 mL of 10 mg/L fluoride solution were equilibrated showed that the optimum fluoride removal was achieved with a dosage of 0.6 g. The use of dosages above 0.6 g would amount to a waste of material as no appreciable increase in fluoride removal occurred at higher dosages.
Morphological analysis and energy dispersion X-ray spectroscopy
The surface structures of the raw and calcined samples of Al/Fe oxide-modified DE were analysed by carrying out the scan using a JEOL JSM-7500F Scanning Electron Microscope (SEM). The elemental analysis of the raw and Al/Fe oxide-modified DE was carried by coupling the energy dispersion X-ray spectroscope (EDX) on the JEOL JSM-7500F Scanning Electron Microscope. Both SEM and EDX analyses of samples were carried out at the University of Cape Town, South Africa.
Metals leaching assessment of the sorbent
The pH of mixtures containing 0.6 g of Al/Fe oxide-modified DE and 90 mL of 12.5 mg/L fluoride was adjusted to different values (2 to 12) using 0.1 M HCl and 0.1 M NaOH solutions. In each mixture, the final volume of 100 mL of solution was attained by adding an appropriate volume of Milli-Q water. The mixtures were equilibrated for 50 min, followed by centrifuging to obtain clear supernatants. The stability of the sorbent in the different mixtures was evaluated by analysis of Al and Fe in the supernatants. Aliquots of 10 mL of the supernatants were measured into 15 mL centrifuge tubes and acidified with 2 mL of 3 M HNO3. The tubes were corked and shaken thoroughly before carrying out Al and Fe analyses of the supernatants using an inductively coupled plasma-mass spectrometer (ICP-MS).
A mass of 1.5 g of Al/Fe oxide-modified DE was weighed into six porcelain crucibles and calcined respectively in a muffle furnace at 200, 400, 600, 700, 800 and 900 °C for 60 min. Each sample was removed from the furnace after calcination, cooled in air for some time and finally left to cool to room temperature in the desiccator. Further analyses were carried out using the calcined samples.
Effect of calcination on composite metals
A mass of 0.5 g of each of the calcined samples and the sample originally dried at 110 °C was weighed into 50 mL of 3 M HNO3 solutions contained in 250 mL plastic bottles. The bottles were corked and left to stand for 3 h to bring the Al and Fe into solution. The dissolution was further enhanced by heating the mixtures at a temperature of 60 °C for 30 min. The mixtures and the analysis blank were passed through 0.45 μm polypropylene membrane filters in a filter flask assemblage aided by a vacuum pump. The filtrates were then analysed for dissolved metals by ICP-MS.
Batch defluoridation experiment using calcined sorbent
The calcined samples were used in defluoridation of artificial fluoride water to evaluate the stability temperatures of the calcined sorbent. The procedure involved contacting the optimum dosage of 0.6 g of each sample, including the original sample dried at 110 °C, with 100 mL of 10 mg/L fluoride solution for 50 min (the optimum contact time) in a reciprocating water bath at a shaking speed of 200 rpm. The mixtures were filtered using 0.45 μm polypropylene membrane filters. One millilitre of Total Ionic Strength Adjustment Buffer (TISAB III) was added to aliquots of 10 mL of each filtrate, stirred and left for 40 min for complete decomplexation of metal–fluoride complexes. The fluoride was analysed with a fluoride ion-selective electrode (Orion VERSA STAR advanced electrochemistry meter fluoride ion-selective electrode) calibrated with four fluoride standards in which TISAB III was added at the volume ratio of 1:10.
Thermogravimetric analysis of sorbent
Thermogravimetric analysis was used to evaluate the change in weight of Al/Fe oxide-modified DE with temperature. Using the thermogravimetry instrument, TGA Q500 V20.13 Build 39, the thermogravimetry of the sorbent was run along with that of the pure Al/Fe oxide for comparison. In the analysis, 9.4101 mg of Al/Fe oxide and 8.3161 mg of Al/Fe oxide-modified DE were heated in the furnace at increasing temperature until a maximum temperature of 900 °C was attained. The per cent weight of each sample at each temperature was plotted against the corresponding temperature.
Sorbent regeneration and reuse
A sorbent regeneration study was carried out to evaluate the reusability of the recycled spent sorbent. The reagents used for regenerating the spent sorbent included 0.01 M NaOH, 0.01 M Na2CO3 and 0.1 M K2SO4 solutions. In each case, the first process involved defluoridation of 100 mL of 10 mg/L fluoride solution using 0.8 g of Al/Fe oxide-modified DE. The mixtures were centrifuged and analysis of supernatant for residual fluoride was done as described above. The separated solid was then washed with Milli-Q water, dried in the oven at 110 °C for 8 h and then cooled in the desiccator. The dry solid was crushed and sieved with a 250 μm test sieve.
The sieved, spent sorbent was weighed and equilibrated with 100 mL of each of the evaluated regenerants for 50 min to desorb fluoride. Centrifuging, washing of solids, measurement of desorbed fluoride and drying of solids were done as in previous cases. A known mass of dry, crushed and sieved sorbent was again contacted with 100 mL of 10 mg/L fluoride solution. The regeneration of spent sorbent and defluoridation with the regenerated sorbent continued until the third or fourth cycles of defluoridation depending on the observed trend of fluoride removal.
Solubility of DE
The solubility of DE, which is basically a siliceous material (Antonides 1998), was evaluated at different solution pH. The procedure involved equilibration of a mixture of 0.4 g of DE and 50 mL of 8 mg/L fluoride solutions of different pH for 30 min. The supernatants obtained after centrifuging the mixtures were acidified with 3 M HNO3 and analysed for Si among other elements using ICP-MS and ICP-atomic emission spectroscopy (ICP-AES). The concentrations of Si determined in the various supernatants give the estimate of the dissolved silica.
RESULTS AND DISCUSSION
Scanning electron microscopy
Energy dispersion X-ray spectroscopy
Metal leaching assessment of the sorbent
Table 1 gives the results of the Al and Fe analysed by ICP-MS after blank correction. The extent of metal leaching from Al/Fe oxide-modified DE at different solution pH is related to the sorbent's stability in the solution. The least concentration of leached Al (0.17 mg/L) was recorded at pH 6.70. The more that metals were lost from the sorbent, the lower was the fluoride removal from water. Fluoride removal was observed to be maximum (92.3%) at the minimum leaching of Al, that is at pH 6.70. Considering the effect of leached Al on human health, at those levels of pH evaluated, the treated water would be least hazardous at pH 6.70 as Al concentration is within the limit of 0.1–0.17 mg/L in water treatment facilities (WHO 2011).
|Equilibrium pH (pHe) .||Fluoride removal (%) .||Al (mg/L) .||Fe (mg/L) .|
|Equilibrium pH (pHe) .||Fluoride removal (%) .||Al (mg/L) .||Fe (mg/L) .|
Effect of calcination on composite metals
Defluoridation with calcined sorbent
Sorbent regeneration and reuse
Regeneration of spent Al/Fe oxide-modified DE using 0.1 M K2SO4 solution proved to be the most reliable of the evaluated regenerants. By the second defluoridation cycle, a dosage of 0.8 g/100 mL was able to reduce the concentration of 10 mg/L fluoride in water by 81.8%. With the use of 0.01 M NaOH and 0.01 M Na2CO3 for fluoride desorption from the fluoride-loaded (spent) sorbent, percentage fluoride removal at the second and higher defluoridation cycles were far less than that observed at the first application of the sorbent. The worst trend was observed with the use of Na2CO3 in particular. Both reagents gave a high desorption pH, which led to appreciable loss of silica, a condition already discussed above. The dissolution of silica implied the loss of Al/Fe oxide. This was evident from the intense brown colour of the supernatants, which showed the loss of Fe3+ from the adsorbent. Apart from the problem associated with the alkaline pH of regeneration with Na2CO3, was found to be strongly bound to the sorbent such that it was practically impossible for the anion to be replaced by fluoride in solution. It is widely reported that competes highly with fluoride in batch experiments involving fluoride removal with adsorbents (Kamble et al. 2007, 2009; Bansiwal et al. 2010). Therefore, NaOH and Na2CO3 are not suitable reagents for regeneration of spent Al/Fe oxide-modified DE, but K2SO4 solution is.
The study evaluated the stability of Al/Fe oxide-modified DE sorbent at different solution pH and calcination temperatures and appraised the reusability of spent sorbent regenerated from different solutions. Findings showed that Al and Fe were leached most from the sorbent under acidic and alkaline conditions, with Al more significantly lost than Fe. ICP-MS analyses provide evidence that Al/Fe oxide-modified DE is unstable at temperatures above 600 °C. It was found that calcination did not improve the fluoride uptake capacity of Al/Fe oxide-modified DE. The thermogravimetric analysis of Al/Fe oxide-modified DE showed that the limit of its stability was 600 °C. K2SO4 solution proved to be the most suitable regenerant for spent Al/Fe oxide-modified DE compared with NaOH and Na2CO3.
This study was funded by WRC Project No. K5/2363/3, NRF Project No. CSUR13092849176, Grant No. 90288 and THRIP Project No. TP12082610644, and the Directorate of Research & Innovation, University of Venda.