Cr(VI) Femoval from Ground Waters by Ferrous Iron Redox-Assisted Coagulation in a Continuous Treatment Unit Comprising a Plug Flow Pipe Reactor and Downflow Sand Filtration

Chromium(VI) (Cr(VI)) is the main chromium species found in groundwater and is considered as a highly toxic and carcinogenic element to humans. In the present study, removal of Cr(VI) by coagulation with ferrous iron is studied in a continuous flow treatment unit comprising pipe flocculation reactors followed by a sand filter. The studied parameters, regarding their effect on the removal of hexavalent chromium, were the ferrous iron dose, the effect of linear velocity, and the effect of the starting Cr(VI) concentration. The experiments have shown that the Cr(VI) removal achieved was very efficient and residual Cr(VI) and total Cr concentration in the treated water was lower than 10 μg/L, provided that the required dose of ferrous iron is provided. In particular, the study demonstrated that the removal of hexavalent chromium, from initial concentration of 50 μg/L and 100 μg/L, was more than 90% with ferrous doses of 1 mg/L and 2 mg/L respectively, applying linear velocity of 8 m/h, at an initial pH value of 7.3. Iron concentration in treated water was very low, far below 200 μg/L, which is the limit for iron in drinking water. This unit comprises a simple treatment option, for applications at the household level, with minimum maintenance requirements capable of removing Cr(VI) to concentrations below 10 μg/L, which might be the future limit for chromium in drinking water.


Introduction
The presence of chromium in natural and engineered aquatic systems is due to natural sources as well as to anthropogenic activities. Natural sources of chromium are mainly rock disintegration [1,2], whereas anthropogenic activities that contribute to elevated chromium concentrations in water sources are mainly industrial activities, such as for example the electroplating industry, stainless steel production, leather tanning, and textile manufacturing [3]. In water, chromium is present in its two most stable oxidation states, namely, Cr(III), and Cr(VI). In chromium-containing natural water, the prevailing form is Cr(VI), because it is very soluble, whereas Cr(III) is quite insoluble, easily forming complexes with iron oxides or other minerals and, therefore, precipitating out of water [4]. 4 x−2 (aq) + 3Fe 2+ (aq) + jH 2 O → Cr(OH) y 3−y (aq) + 3Fe(OH) z 3−z (aq) + jH + (1) Fe (II) reacts with Cr (VI) and promotes its reduction to Cr (III), while Fe (II) is oxidized to Fe (III), creating insoluble ferric hydroxide particles [Fe(OH) 3 ]. Cr (III) is precipitated either as hydroxide or as mixed oxides of Cr(III)-Fe(III) [20]. The resulting precipitates can be separated from the water by conventional sand or membrane filtration [20].
The main objective of this research is to study the removal of hexavalent chromium from water by chemical reduction and precipitation using ferrous sulfate as the reducing agent followed by sand filtration, using an innovative treatment apparatus, which runs in full-scale mode and uses pipe reactors to promote the reaction of Fe(II) with Cr(VI) To the best of our knowledge, until to date, this is the first time, that hexavalent chromium removal is investigated using a continuous treatment unit, based in pipe flocculation reactors followed by sand filtration. In addition, a further aim of this study was to propose a treatment plant, which will be simple in use, with low capital and operational costs, but very efficient in the removal of hexavalent chromium, in order to contribute to the fulfillment of the sustainable goal 6, which is aiming at providing clean water for all [21].

Simulated Groundwater
The water used to conduct the experimental part of this study was drinking water from the city of Thessaloniki, Greece, having the typical composition of groundwater, because its major part comes from groundwater wells around the city. The water was firstly filtered through an activated carbon filter, in order to remove the residual chlorine and then spiked with Cr(VI). Afterwards, it was used for investigating the removal of Cr(VI) by ferrous iron. The major chemical parameters of Cr(VI)-contaminated water are given in Table 1.

Chemicals
Cr(VI) stock solutions were prepared by dissolving a certain amount of K 2 Cr 2 O 7 (pro analysis, Merck, Darmstadt, Germany) in de-ionized water, of conductivity less than 2 µS/cm. Afterwards, in the simulated groundwater, proper volume of stock solution was diluted in about 200 L of water to prepare the storage tank of groundwater, which was ready to be treated by the proposed technology. Ferrous iron working solutions were freshly prepared every day by dissolving FeSO 4 ·7H 2 O (pro analysis, Merck, Darmstadt, Germany) in free chlorine drinking water and stored in a 10 L tank. The reagent was acidified with concentrated hydrochloric acid 32% (pro analysis, Merck, Darmstadt, Germany) to a final pH value of about 3, in order to slow down ferrous iron oxidation from atmospheric oxygen, because Fe(II) oxidation by oxygen is very slow at acidic pH values [22].

Experimental Setup
The design of the experimental set up is shown in Figure 1 and the caption contains all essential information for its operation. This comprised two treatment parts in row. In the first part, chromium reduction took place in a reactor consisting of pipes via the addition of ferrous sulfate, and in the second part, the produced precipitates were removed by sand filtration. The characteristics of the treatment plant are given in the figure caption.
The chromium-polluted water was prepared in the 200-liter tank. A peristaltic pump drove the polluted water into the pipe reactor system and the flow rate was adjusted accordingly at 15, 30, 44 and 54 L/h. A peristaltic pump pushed the reductant (ferrous sulphate) into the pipe reactor system. The flow rate of peristaltic pump was set at 2 L/h. In the pipe reactor system, the polluted water was mixed with the reductant and the red-ox reaction took place. Fe(II) was oxidized to Fe(III) and Cr(VI) was reduced to Cr(III). The experimental runs took place at the pH value of the tap water of Thessaloniki which is 7.3 + 0.1. The water flowed afterwards through the sand filter, and the formed particles which contain iron and the reduced hexavalent chromium were removed from the water stream by filtration.
Sampling locations included: a) the inlet water tank, representing the feedwater influent, corresponding to water spiked with Cr(VI) and b) the filtered effluent. Total chromium, Cr(VI), Total Iron and Fe(II), pH of treated water and conductivity were measured in each sample on an hourly basis. Regularly, dissolved oxygen and redox potential were measured in these samples.
The experiments were conducted in a continuous operation mode, applying a once through pass of the Cr(VI) contaminated water through the experimental unit under room temperature (20 -22 °C). The studied treatment unit has the capacity of treating 500 to 1200 liters of contaminated groundwater per day, therefore it is simulating treatment needs at the household level, given that the daily water needs per person in Europe is about 130 to 150 lt/day.

Determination Methods
The determination of hexavalent chromium concentrations was performed by the photometric method based on the reaction of Cr(VI) with diphenylcarbazide (DPC) [23]. In brief, 10 ml of the sample were placed in a falcon tube and afterwards 200 μl of DPC reagent were added. Finally, 100 μL of H2SO4 were added to acidify the sample. The samples were left for 15 minutes for the reaction to be completed and chromium absorption was measured at 540 nm, using an ultraviolet-visible This comprised two treatment parts in row. In the first part, chromium reduction took place in a reactor consisting of pipes via the addition of ferrous sulfate, and in the second part, the produced precipitates were removed by sand filtration. The characteristics of the treatment plant are given in the figure caption.
The chromium-polluted water was prepared in the 200-L tank. A peristaltic pump drove the polluted water into the pipe reactor system and the flow rate was adjusted accordingly at 15, 30, 44 and 54 L/h. A peristaltic pump pushed the reductant (ferrous sulphate) into the pipe reactor system. The flow rate of peristaltic pump was set at 2 L/h. In the pipe reactor system, the polluted water was mixed with the reductant and the red-ox reaction took place. Fe(II) was oxidized to Fe(III) and Cr(VI) was reduced to Cr(III). The experimental runs took place at the pH value of the tap water of Thessaloniki which is 7.3 ± 0.1. The water flowed afterwards through the sand filter, and the formed particles which contain iron and the reduced hexavalent chromium were removed from the water stream by filtration.
Sampling locations included: (a) the inlet water tank, representing the feedwater influent, corresponding to water spiked with Cr(VI) and (b) the filtered effluent. Total chromium, Cr(VI), Total Iron and Fe(II), pH of treated water and conductivity were measured in each sample on an hourly basis. Regularly, dissolved oxygen and redox potential were measured in these samples.
The experiments were conducted in a continuous operation mode, applying a once through pass of the Cr(VI) contaminated water through the experimental unit under room temperature (20-22 • C). The studied treatment unit has the capacity of treating 500 to 1200 L of contaminated groundwater per day, therefore it is simulating treatment needs at the household level, given that the daily water needs per person in Europe is about 130 to 150 L/day.

Determination Methods
The determination of hexavalent chromium concentrations was performed by the photometric method based on the reaction of Cr(VI) with diphenylcarbazide (DPC) [23]. In brief, 10 mL of the sample were placed in a falcon tube and afterwards 200 µL of DPC reagent were added. Finally, 100 µL of H 2 SO 4 were added to acidify the sample. The samples were left for 15 min for the reaction to be completed and chromium absorption was measured at 540 nm, using an ultraviolet-visible (UV-Vis) spectrophotometer (model U-2000, Hitachi, Tokyo, Japan). Total chromium concentrations were measured by Inductive Coupled Plasma-Mass Spectroscopy ICP-MS. The determination of divalent and total (divalent and trivalent) iron was performed by 1,10-phenanthroline method [23].

Removal of Cr(VI) in Relation to the Dose of Fe(II)
In these experiments the factor that affects the system is the concentration of ferrous iron that is available for reaction with hexavalent chromium after the mixing of the water with the reducing agent. Figure 2 shows that as the concentration of ferrous iron is increased, Cr(VI) percentage removal is also increased. The applied doses were 0.25, 0.5, 0.75 and 1 mg/L at an initial chromium concentration of 50 µg/L, corresponding to molar ratios of 4.8, 9.6, 14.4 and 19.2. For these molar ratios, Cr(VI) removal was in the order of 45%, 85%, 88% and 92% respectively. These results corresponded to residual Cr(VI) concentrations of about 24, 8, 6 and 3 µg/L respectively.

Removal of Cr(VI) in Relation to the Dose of Fe(II)
In these experiments the factor that affects the system is the concentration of ferrous iron that is available for reaction with hexavalent chromium after the mixing of the water with the reducing agent. Figure 2 shows that as the concentration of ferrous iron is increased, Cr(VI) percentage removal is also increased. The applied doses were 0.25, 0.5, 0.75 and 1 mg/L at an initial chromium concentration of 50 μg/L, corresponding to molar ratios of 4.8, 9.6, 14.4 and 19.2. For these molar ratios, Cr(VI) removal was in the order of 45%, 85%, 88% and 92% respectively. These results corresponded to residual Cr(VI) concentrations of about 24, 8, 6 and 3 μg/L respectively. Experiments were carried out at pH 7.3. At this pH value, total chromium was also very low (similar to Cr(VI) values). These results are in agreement with batch studies [15] and are quite improved from results obtained in continuous mode operation units, which instead of a pipe flocculation reactor, used a Cr(VI) reduction tank [14]. In the study of Qin et al., 2005 [14], it was shown that a molar ratio of Fe(II)/Cr(VI) of 9 to 56 was required to achieve a more than 90% total chromium removal for water containing 3-4 mg/L dissolved oxygen. In our experiments, a ratio of about 15 was needed for water containing dissolved oxygen of about 8 mg/L. Given the fact that increased dissolved oxygen in water competes with Cr(VI) for Fe(II) oxidation, it is expected that Fe(II) oxidation takes place also by the dissolved oxygen, according to the kinetics of bivalent iron oxidation by oxygen [22]. Therefore, the results obtained provide evidence that the use of pipe flocculation reactors can be a more efficient reactor for reducing hexavalent to trivalent chromium than the use of Cr(VI) reduction tanks, since under fully oxic conditions, the unit achieved more than 90% Cr(VI) removal by a molar ratio Fe/Cr of about 15, for a linear velocity of about 8 m/h.
To maintain the stable operation of the unit, the filters were backwashed once a day with clean drinking water and the sludge was collected and disposed. This helped to avoid replacing the sand Experiments were carried out at pH 7.3. At this pH value, total chromium was also very low (similar to Cr(VI) values). These results are in agreement with batch studies [15] and are quite improved from results obtained in continuous mode operation units, which instead of a pipe flocculation reactor, used a Cr(VI) reduction tank [14]. In the study of Qin et al., 2005 [14], it was shown that a molar ratio of Fe(II)/Cr(VI) of 9 to 56 was required to achieve a more than 90% total chromium removal for water containing 3-4 mg/L dissolved oxygen. In our experiments, a ratio of about 15 was needed for water containing dissolved oxygen of about 8 mg/L. Given the fact that increased dissolved oxygen in water competes with Cr(VI) for Fe(II) oxidation, it is expected that Fe(II) oxidation takes place also by the dissolved oxygen, according to the kinetics of bivalent iron oxidation by oxygen [22]. Therefore, the results obtained provide evidence that the use of pipe flocculation reactors can be a more efficient reactor for reducing hexavalent to trivalent chromium than the use of Cr(VI) reduction tanks, since under fully oxic conditions, the unit achieved more than 90% Cr(VI) removal by a molar ratio Fe/Cr of about 15, for a linear velocity of about 8 m/h.
To maintain the stable operation of the unit, the filters were backwashed once a day with clean drinking water and the sludge was collected and disposed. This helped to avoid replacing the sand very frequently. The sand is progressively coated with iron oxides, which as the literature studies Appl. Sci. 2020, 10, 802 6 of 10 suggest can enhance the rate of Cr(VI) reduction by Fe(II) and also oxidation of Fe(II) to Fe(III), when the dosing is over stoichiometric levels [20,22].

Influence of Linear Velocity to the Removal of Cr(VI)
The effect of linear velocity was examined in this experimental section, by using [Fe(II)]/[Cr(VI)] molar proportion of 9.6. The applied flocculant dose was 0.5 mg/L at an initial chromium concentration of 50 µg/L. With the increase of linear velocity, the time that is available in the tubular reactor for ferrous iron and hexavalent chromium to react is decreased. The linear velocity was adjusted to the values 4, 8, 12 and 15 m/h. A slight decrease in Cr(VI) removal with the increase of the linear velocity was noticed, as shown in Figure 2. For a linear velocity at 4 m/h the chromium removal rate was 88%, for linear velocity 8 m/h the removal rate was 85%, for linear velocity 12 m/h, the removal rate was at 86%, for linear velocity at 15 m/h the removal rate is at 83%. These results are summarized in Figure 3 and show that there is a slight decrease in the removal of hexavalent chromium by increasing the linear velocity, especially at values higher than 15 m/h. This is mainly attributed to the fact that by increasing the linear velocity to higher values, the contact time of the water with the reagent is decreased and therefore, the time allowed for the Fe(II) reaction with Cr(VI) to take place will be decreased. However, the percentage removal of 83% at 15 m/L is considered as a very satisfactory result, because this corresponds to residual Cr(VI) concentration of less than 10 µg/L and the daily water production accounts for about 1 m 3 /day.
Appl. Sci. 2020, 10, 802 6 of 10 very frequently. The sand is progressively coated with iron oxides, which as the literature studies suggest can enhance the rate of Cr(VI) reduction by Fe(II) and also oxidation of Fe(II) to Fe(III), when the dosing is over stoichiometric levels [20,22].

Influence of Linear Velocity to the Removal of Cr(VI)
The effect of linear velocity was examined in this experimental section, by using [Fe(II)]/[Cr(VI)] molar proportion of 9.6. The applied flocculant dose was 0.5 mg/L at an initial chromium concentration of 50μg/L. With the increase of linear velocity, the time that is available in the tubular reactor for ferrous iron and hexavalent chromium to react is decreased. The linear velocity was adjusted to the values 4, 8, 12 and 15 m/h. A slight decrease in Cr(VI) removal with the increase of the linear velocity was noticed, as shown in Figure 2. For a linear velocity at 4 m/h the chromium removal rate was 88%, for linear velocity 8 m/h the removal rate was 85%, for linear velocity 12 m/h, the removal rate was at 86%, for linear velocity at 15 m/h the removal rate is at 83%. These results are summarized in Figure 3 and show that there is a slight decrease in the removal of hexavalent chromium by increasing the linear velocity, especially at values higher than 15 m/h. This is mainly attributed to the fact that by increasing the linear velocity to higher values, the contact time of the water with the reagent is decreased and therefore, the time allowed for the Fe(II) reaction with Cr(VI) to take place will be decreased. However, the percentage removal of 83% at 15m/L is considered as a very satisfactory result, because this corresponds to residual Cr(VI) concentration of less than 10 μg/L and the daily water production accounts for about 1 m 3 /d.

Effect of the Initial Hexavalent Chromium Concentration
In this part of experimental section, the examined parameter was the initial concentration of the hexavalent chromium. With the increase of the initial CR(VI) concentration, and by keeping Fe(II) concentration constant, the molar ratio [Fe(II)]/ [Cr(VI)] was decreased. In particular, for initial Cr(VI) concentrations of 50, 100, 200 and 400 μg Cr(VI)/L, the molar ratio of Fe/Cr decreases from 9.6, to 9.3, 4.65, 2.32 and 1.15 respectively. Figure 4 shows the corresponding results. With the gradual increase of the initial chromium concentration from 50 μg/L to 400 μg/L, a decrease in chromium removal rate was noticed. In particular, for initial chromium (VI) concentration at 50 μg/L and 100

Effect of the Initial Hexavalent Chromium Concentration
In this part of experimental section, the examined parameter was the initial concentration of the hexavalent chromium. With the increase of the initial CR(VI) concentration, and by keeping Fe(II) concentration constant, the molar ratio [Fe(II)]/[Cr(VI)] was decreased. In particular, for initial Cr(VI) concentrations of 50, 100, 200 and 400 µg Cr(VI)/L, the molar ratio of Fe/Cr decreases from 9.6, to 9.3, 4.65, 2.32 and 1.15 respectively. Figure 4 shows the corresponding results. With the gradual increase of the initial chromium concentration from 50 µg/L to 400 µg/L, a decrease in chromium removal rate was noticed. In particular, for initial chromium (VI) concentration at 50 µg/L and 100 µg/L the removal rate was around 88% and 83%, respectively. For the initial concentration at 200 µg/L the removal rate is at around 78% and for the initial chromium concentration at 400 µg/L the removal rate is at around 65%. The residual chromium concentrations achieved were 6, 17, 44 and 140 µg/L chromium, respectively. μg/L the removal rate was around 88% and 83%, respectively. For the initial concentration at 200 μg/L the removal rate is at around 78% and for the initial chromium concentration at 400 μg/L the removal rate is at around 65%. The residual chromium concentrations achieved were 6, 17, 44 and 140 μg/L chromium, respectively. As the initial Cr(VI) concentration is increasing, the Cr(VI) removal rate is decreasing, apparently because the molar ratio Fe(II)/Cr(VI) becomes lower. Higher removal values at these ratios could not be achieved at the pH value at which the experiment was conducted, because of the competition of Fe(II) oxidation with the oxygen from water, since the kinetics of Fe(II) oxidation is mainly pH dependent and at pH values above 7, Fe(II) oxidation by oxygen is very fast. This causes a competition between Cr(VI) and O2 to oxidize Fe(II). In particular, Buerge and Hug, (1997) [24] have reported that Cr(VI) oxidizes Fe(II) faster than O2 for [Cr(VI)] = [O2] by a factor of about 3 × 10 4 at pH 4, 6 × 10 3 at pH 6 and 1 × 10 3 at pH 8. This highlights that by increasing the pH of water, increased Fe(II) concentration is need to oxidized the same concentration of Cr(VI). For higher initial Cr(VI) concentrations, the final concentrations were much higher than the current and proposed limits for Cr(VI) in water.
The interesting finding of this experiment, however, is that although the percentage of Cr(VI) removal is decreasing and the residual concentrations of Cr(VI) are increasing, the amount of Cr(VI) removed in μg/L is increasing. For example, when the initial concentration was 50, the amount of Cr(VI) removed was about 45 μg/L, whereas for initial concentration of Cr(VI) of 400 μg/L, the amount removed was 256 μg/L. Both results were obtained by the same dosing quantity of Fe(II), namely 0.5 mg/L. This might indicate an autocatalytic effect of Cr(VI) concentration on Cr(VI) reduction by Fe(II), i.e., by increasing the kinetics of Fe(II) oxidation by Cr(VI). This might be critical in the removal of high concentrations of Cr(VI) and provides evidence that multiple additions of Fe(II) in the water might be more efficient that dosing all the Fe(II) quantity at once. These results are in agreement with a study by Schlautman and Han (2001) [25], who examined Cr(VI) reduction by dissolved ferrous in under-stoichiometric conditions, i.e., representing a deficient amount of Fe(II).
In addition, the idea of multiple dosing of the reagent could also contribute to the reduction of the produced sludge. Indeed, this technology produces some sludge, which has to be handled. There are several ways of handling sludge from water treatment processes, such as dewatering, stabilizing As the initial Cr(VI) concentration is increasing, the Cr(VI) removal rate is decreasing, apparently because the molar ratio Fe(II)/Cr(VI) becomes lower. Higher removal values at these ratios could not be achieved at the pH value at which the experiment was conducted, because of the competition of Fe(II) oxidation with the oxygen from water, since the kinetics of Fe(II) oxidation is mainly pH dependent and at pH values above 7, Fe(II) oxidation by oxygen is very fast. This causes a competition between Cr(VI) and O 2 to oxidize Fe(II). In particular, Buerge and Hug, (1997) [24] have reported that Cr(VI) oxidizes Fe(II) faster than O 2 for [Cr(VI)] = [O 2 ] by a factor of about 3 × 10 4 at pH 4, 6 × 10 3 at pH 6 and 1 × 10 3 at pH 8. This highlights that by increasing the pH of water, increased Fe(II) concentration is need to oxidized the same concentration of Cr(VI). For higher initial Cr(VI) concentrations, the final concentrations were much higher than the current and proposed limits for Cr(VI) in water.
The interesting finding of this experiment, however, is that although the percentage of Cr(VI) removal is decreasing and the residual concentrations of Cr(VI) are increasing, the amount of Cr(VI) removed in µg/L is increasing. For example, when the initial concentration was 50, the amount of Cr(VI) removed was about 45 µg/L, whereas for initial concentration of Cr(VI) of 400 µg/L, the amount removed was 256 µg/L. Both results were obtained by the same dosing quantity of Fe(II), namely 0.5 mg/L. This might indicate an autocatalytic effect of Cr(VI) concentration on Cr(VI) reduction by Fe(II), i.e., by increasing the kinetics of Fe(II) oxidation by Cr(VI). This might be critical in the removal of high concentrations of Cr(VI) and provides evidence that multiple additions of Fe(II) in the water might be more efficient that dosing all the Fe(II) quantity at once. These results are in agreement with a study by Schlautman and Han (2001) [25], who examined Cr(VI) reduction by dissolved ferrous in under-stoichiometric conditions, i.e., representing a deficient amount of Fe(II).
In addition, the idea of multiple dosing of the reagent could also contribute to the reduction of the produced sludge. Indeed, this technology produces some sludge, which has to be handled. There are several ways of handling sludge from water treatment processes, such as dewatering, stabilizing and disposal to landfills, but the most important aspect is to reduce the amount of sludge produced, which can be potentially achieved by the multiple dosing suggestion in a unit operation such as that proposed in the present study.
The chromium removal results from all the experiments in this study were in agreement with the results from previous studies, examining Cr(VI) removal in batch mode operation [15]. Mitrakas et al. (2011) [15] found that for pH 7.5-8 and for ratio [Fe(II)]/[Cr(VI)] equal to 5, the chromium removal is