Oxidation of Cr(III)-Fe(III) Mixed-Phase Hydroxides by Chlorine: Implications on the Control of Hexavalent Chromium in Drinking Water.

The occurrence of chromium (Cr) as an inorganic contaminant in drinking water is widely reported. One source of Cr is its accumulation in iron-containing corrosion scales of drinking water distribution systems as Cr(III)-Fe(III) hydroxide, that is, Fe xCr(1- x)(OH)3(s), where x represents the Fe(III) molar content and typically varies between 0.25 and 0.75. This study investigated the kinetics of inadvertent hexavalent chromium Cr(VI) formation via the oxidation of Fe xCr(1- x)(OH)3(s) by chlorine as a residual disinfectant in drinking water, and examined the impacts of Fe(III) content and drinking water chemical parameters including pH, bromide and bicarbonate on the rate of Cr(VI) formation. Data showed that an increase in Fe(III) molar content resulted in a significant decrease in the stoichiometric Cr(VI) yield and the rate of Cr(VI) formation, mainly due to chlorine decay induced by Fe(III) surface sites. An increase in bicarbonate enhanced the rate of Cr(VI) formation, likely due to the formation of Fe(III)-carbonato surface complexes that slowed down the scavenging reaction with chlorine. The presence of bromide significantly accelerated the oxidation of Fe xCr(1- x)(OH)3(s) by chlorine, resulting from the catalytic effect of bromide acting as an electron shuttle. A higher solution pH between 6 and 8.5 slowed down the oxidation of Cr(III) by chlorine. These findings suggested that the oxidative conversion of chromium-containing iron corrosion products in drinking water distribution systems can lead to the occurrence of Cr(VI) at the tap, and the abundance of iron, and a careful control of pH, bicarbonate and bromide levels can assist the control of Cr(VI) formation.

Introduction

Chromium (Cr) can exist as hexavalent Cr(VI) and trivalent Cr(III) in drinking water. Cr(VI) is highly mobile and toxic, whereas Cr(III) forms minerals with low solubility at circumneutral pH and is also considered a micronutrient.[1,2] The State of California set a new drinking water regulation specifically for Cr(VI) at 10 μg/L in 2014.[3] The U.S. EPA regulates total chromium in drinking water at 100 μg/L, and a federal regulation solely for Cr(VI) may be established in the future.[4] Traditional anthropogenic sources typically come from industrial waste discharge.[5−7] In addition, natural occurrence of Cr(VI) originates from the natural weathering of Cr(III)-containing aquifer minerals in groundwater.[8−10]

In recent years, in situ generation of Cr(VI) in drinking water distribution systems from Cr-containing corrosion scales has drawn an increasing attention.[11,12] Corrosion products in drinking water pipes are found to contain up to 0.5% chromium by weight, far exceeding its natural abundance in the earth crust.[13−15] In particular, considering that nearly 70% of pipes in water distribution systems in the U.S. are composed of iron materials (e.g., cast iron, ductile iron and cement-lined iron), chromium can accumulate extensively in iron corrosion scales over time.[16−20] Chromium is typically present at trace levels below regulatory standards in treated drinking water, but over decades it can accumulate to very high levels in the corroded solids in distribution systems. In addition, iron piping materials are found to contain chromium impurities that serve as in situ sources.[21−23] When coexisting with iron, Cr(III) forms mixed phases of Cr(III)–Fe(III) hydroxide, that is, FeCr(1–(OH)3(s), where x represents the Fe(III) molar content and typically varies between 0.25 and 0.75.[24−26] In addition, FeCr(1–(OH)3(s) widely exists in iron-rich aquifer materials and its oxidation by Mn(IV) oxides contributes to the natural occurrence of Cr(VI) in groundwater.[27−29]

The occurrence of Cr(VI) in drinking water distribution systems depends on a cascade of redox processes. In particular, the oxidation of Cr(III) solids by the residual disinfectant chlorine can take place in distribution systems, and results in the risk of an inadvertent generation of Cr(VI) at the tap. Nationwide surveys have indicated that the reaction pathways depended on Cr(III) speciation and its redox reactivity in distribution systems.[11,17,30] Field sampling data from the third round of U.S. Environmental Protection Agency Unregulated Contaminant Monitoring Rule (UCMR3) indicated that Cr(VI) levels increased from the entry to exit point of many distribution systems nationwide, and this trend correlated with the presence of chlorine as the residual disinfectant.[11] Recent studies also showed that the oxidation of Cr(III) solids by chlorine can increase Cr(VI) levels in drinking water.[12,31,32] In addition, mixed phases of Cr(III) oxides with other metals, for example, Cu2Cr2O5(s), exhibit a higher redox reactivity than pure phases of Cr(III).[12] Similarly, the coexistence of Fe(III) in FeCr(1–(OH)3(s) can possess unique reactivity with chlorine to generate Cr(VI), and the aging of Cr(III) solids can also impact its redox reactivity; however, there is little knowledge on the formation of Cr(VI) from FeCr(1–(OH)3(s) in drinking water.

Furthermore, the impacts of water chemical parameters including pH, bromide and bicarbonate on the reactivity of Cr(III)–Fe(III) hydroxide remain unknown. Bicarbonate and pH are important parameters to control corrosion in distribution systems.[33,34] They impact complexation of metal surface carbonato and hydroxo species, and consequently surface redox reactivity of metal oxides.[35] pH can also impact the speciation of chlorine and its oxidative capacity. Bromide exhibits a catalytic effect on the chlorine-driven oxidation of transition metals.[36] Recent studies show that bromide level in drinking water can increase by as much as 20 times in the future due to seawater intrusion, desalination, water reuse and brine discharge from alternative energy production (e.g., shale gas).[37]

Considering the importance of understanding the reactivity of FeCr(1–(OH)3(s) and consequently the development of effective Cr(VI) control strategies in water distribution infrastructure, the objectives of this study were to investigate the mechanisms of Cr(VI) formation via FeCr(1–(OH)3(s) oxidation by chlorine, and to quantify the impacts of Fe(III), bicarbonate, pH and bromide on the reaction kinetics and stoichiometry.

Materials and Methods

All chemicals used in this study were reagent grade or higher. All solutions were prepared using deionized (DI) water (resistivity >18.2 MΩ, Millipore System). Cr(III)–Fe(III) hydroxide solids with three different Fe(III) molar contents were synthesized using a standard protocol.[38] In brief, solutions of Fe(NO3)3 and CrCl3 were mixed at a molar ratio of either 0.3:1, 1:1, or 3:1 at pH 7 for 3 days. Ferrihydrite Fe(OH)3(s) was synthesized via the hydrolysis of Fe(NO3)3 at pH 7.[39] All solids were washed with DI water three times and then centrifuged, frozen overnight and lyophilized for 24 h. The freeze-dried solids were sequentially sieved through NO. 50, 170, and 325 mesh sieves. Sieved particles with nominal sizes between 45 and 90 μm were collected and used in this study. Because of the amorphous nature of FeCr(1–(OH)3(s) solids, X-ray diffraction (XRD) was not capable of characterizing their chemical compositions. To confirm the chemical compositions of synthesized solids, particles were acid digested, and Cr and Fe percentages were analyzed with an Agilent 7700 inductively coupled plasma-mass spectrometry. The BET surface area of solids was measured using a Micromeritics ASAP 2020 surface area analyzer. Zeta potentials of FeCr(1–(OH)3(s) were measured by a Zeta Potential Analyzer (Brookhaven Instruments). The physicochemical parameters of the synthesized solids are listed in Table .

Table 1

Characteristics of Synthesized Cr(III)–Fe(III) Hydroxide FeCr(1–(OH)3(s)

Cr(III)–Fe(III) hydroxide chemical formula	BET surface area (m2/g)	pHpzc	logKspa	Eo (V)b
Fe0.25Cr0.75(OH)3(s)• 1.5H2O	279	5.3	–33.8	0.23
Fe0.5Cr0.5(OH)3(s)• H2O	355	3.8	–34.5	0.38
Fe0.75Cr0.25(OH)3(s)• 1.5H2O	198	5.5	–34.7	0.82

the value of Ksp is based on dissolution reactions listed in Text S1 in the SI.

E° is the standard half-reaction redox potential of the redox couple between FeCr(1–(OH)3(s) and CrO42–.

A 28 mM free chlorine stock solution diluted from a NaOCl solution (Sigma-Aldrich) was freshly prepared every week. The concentration of chlorine was verified by titration with potassium permanganate.[40] Before the start of an experiment, all solutions were purged with N2 gas to remove the dissolved O2. Following that, 0.28 mM of chlorine (i.e., 20 mg Cl2/L) was quickly added into the reactor tube and mixed with FeCr(1–(OH)3(s). This chlorine concentration was higher than typical drinking water residual disinfectant concentration, but provided useful insight into reaction kinetics. The dosages of different Cr(III)–Fe(III) hydroxide solids were controlled so that the Cr(III) molar concentration was constant at 2.8 mM, corresponding to a Cr(III)-to-chlorine ratio of 10:1. In some experiments, ferrihydrite was used to examine the effect of Fe(III).

The solution pH was maintained at a targeted value (±0.2 pH units) between 6.0 and 8.5 with 10 mM phosphate buffer. Control experiments confirmed that the presence of phosphate buffer had a negligible impact on the reaction kinetics. Bromide level was varied between 0.1 and 1 mg/L, and bicarbonate between 1 and 5 mM. In experiments with varying bicarbonate, 10 mM MOPS (3-morpholinopropane-1-sulfonic acid) was added as the pH buffer to avoid the potential interference of phosphate on Fe(III) surface complexation.[41,42] The solution ionic strength was kept at 10 mM by adding a requisite amount of NaClO4. In some experiments, an excess of 1 mM of benzoic acid (BA) was added as a chemical probe to assess the generation of reactive radical species during chlorine oxidation. All chlorine oxidation experiments were conducted in sealed glass tube reactors with no headspace or light exposure, and placed on a rotator to ensure continuous mixing. The ambient temperature was 20 °C ± 2 °C.

Sacrificial tubes were removed from the rotator at predetermined time intervals for chemical analyses. In some experiments, solutions were purged with N2 gas for 20 min to minimize the presence of dissolved O2, and the subsequent generation of dissolved O2 from the reaction was measured using a Mettler-Toledo probe. The leakage of O2 to the ambient environment was minimal. After that, concentrated NaOH was added to the suspension to release potentially adsorbed Cr(VI) from FeCr(1–(OH)3(s), followed by filtration of the suspension through a 0.22 μm Millipore PTFE filter. The concentration of Cr(VI) in the filtrate was measured using the diphenylcarbazide method.[40] Chlorine concentration was measured using a modified DPD method, in which thioacetamide was added to eliminate any potential interference from Cr(VI).[12] Data modeling using the Goal Seek function with Microsoft Excel was applied to obtain the reaction kinetics rate constants, and to correlate the relationship between the stoichiometric ratios of chlorine consumption and Cr(VI) formation rates.

Results and Discussion

Kinetics of Cr(III)–Fe(III) Hydroxide Oxidation by Chlorine and Cr(VI) Formation

The standard half-reaction redox potential of FeCr(1–(OH)3(s) solids are calculated in Supporting Information (SI) Text S1. The values indicate that the oxidation of FeCr(1–(OH)3(s) by chlorine is thermodynamically feasible. To examine the reaction kinetics, a second-order reaction kinetics model was introduced to quantify the rates of chlorine consumption and Cr(VI) formation:kCl2 and kCr(VI) are the surface-area normalized rate constants for chlorine consumption and Cr(VI) formation, respectively (L·m–2·min–1). [Cl2] is the free chlorine concentration (mol/L), [Cr(III)(s)] is the concentration of Cr(III) in Cr(III)–Fe(III) hydroxide (g/L), and SCr(III)(s) is the BET surface area of Cr(III)–Fe(III) hydroxide (m2/g). The model-fitted data correlated well with the experimental data, with a minimal R2 ranging between 0.85 and 0.95 for every set of experiment.

Surface-area normalized rate constants of chlorine consumption with Cr(III)–Fe(III) hydroxide increased with Fe(III) molar contents, that is, the value of x in FeCr(1–(OH)3(s) (Figure A). These kCl2 values were 2–3 times higher than the rate constants observed with pure phases of Cr(III) oxides that do not contain Fe(III), for example, Cr(OH)3(s) and Cr2O3(s).[12] Meanwhile, the rate constant of Cr(VI) formation decreased significantly with increasing Fe(III) content: from 1.5 × 10–4 to 1.9 × 10–5 L·m–2·min–1 when the value of x in FeCr(1–(OH)3(s) increased from 0.25 to 0.75 (Figure B). This trend suggested that the presence of Fe(III) in Cr(III)-containing solids preferentially promoted the reaction kinetics of chlorine consumption but diminishing Cr(VI) formation.

Figure 1

(A) Chlorine consumption and (B) Cr(VI) formation rate constants during the oxidation of iron chromite by chlorine. [Cr(III)] = 2.8 mM, [HOCl] = 20 mg Cl2/L, [Cr(III)]:[Cl2] = 10:1, ionic strength = 10 mM, pH 7.0.

Bromide

The presence of bromide enhanced the reaction kinetics. As the bromide concentration increased from 0 to 1 mg/L, an enhanced Cr(VI) formation was observed (Figure A), and the rate constant of kCr(VI) increased by approximately 3 times (Figure B). Meanwhile, chlorine consumption kinetics accelerated (SI Figure S1A) and the rate constant approximately doubled (SI Figure S1B). The promotive effects of bromide on the oxidation reaction resulted from an electron shuttle mechanism. Bromide readily reacted with HOCl to generate hypobromous acid (HOBr).[43] Prior studies demonstrated that HOBr is more electrophilic than HOCl and consequently exhibited a faster kinetics in oxidizing electron-rich compounds including transition metals and organics.[36,44,45] Control experiments with HOBr also showed that HOBr oxidized FeCr(1–(OH)3(s) much faster than HOCl did (SI Figure S2). Therefore, in a bromide-containing system, HOBr becomes the primary oxidant as bromide essentially acts as an electron shuttle that drives the oxidation of Cr(III)–Fe(III) hydroxide to Cr(VI) by chlorine:

Figure 2

(A) Time profile for Cr(VI) generation from chlorine oxidation of Fe0.5Cr0.5(OH)3(s). (B) Effect of bromide on the surface area normalized second-order Cr(VI) formation rate constants of Cr(III)–Fe(III) hydroxide oxidation by chlorine. Initial [Cr(III)] = 2.8 mM, [HOCl] = 20 mg Cl2/L, Cr(III):Cl2 = 10:1, ionic strength = 10 mM, pH 7.

Bicarbonate

Bicarbonate exhibited a moderate impact on the reaction kinetics. With increasing total bicarbonate concentration (TOTCO3) from 0 to 5 mM at pH 7, the rate constant of Cr(VI) formation increased by approximately 200% (Figure A), and FeCr(1–(OH)3(s) with a lower Fe(III) molar content exhibited a higher sensitivity to the impact of bicarbonate on the rate of Cr(VI) formation (Figure B). Meanwhile, the rate of chlorine consumption increased by an average of 75% with the same amount of bicarbonate increase (SI Figure S3).

Figure 3

(A) Impact of carbonate concentration on chlorine consumption the oxidation of Fe0.75Cr0.25(OH)3(s) by chlorine. (B) Effect of carbonate on the surface area normalized second-order Cr(VI) formation rate constants of Cr(III)–Fe(III) hydroxide oxidation by chlorine. Initial [Cr(III)] = 2.8 mM, [HOCl] = 20 mg Cl2/L, Cr(III):Cl2 = 10:1, ionic strength = 10 mM, pH 7.

The impact of bicarbonate on the reaction kinetics is likely due to the formation of Fe(III)-carbonato complexes, which has been known to decrease the redox reactivity of Fe(III) surface sites.[35] In contrast, bicarbonate had a negligible effect on the redox reactivity of Cr(III) surface sites, as demonstrated by the negligible effect of bicarbonate on the oxidation of a pure phase of Cr(III) solid, that is, Cr(OH)3(s) by chlorine (SI Figure S4). Therefore, the presence of bicarbonate suppressed the reactivity of Fe(III) surface sites on FeCr(1–(OH)3(s) with respect to chlorine, and increased the availability of Cr(III) surface site subject to chlorine oxidation. As a result, the presence of bicarbonate enhanced the rate of Cr(VI) formation via FeCr(1–(OH)3(s) oxidation.

pH

The reaction kinetics were observed to decrease with increasing pH. The rate of chlorine consumption decreased by an average of 40% with pH rising from 6.0 to 8.5. For example, the chlorine consumption rate constant of Fe0.75Cr0.25(OH)3(s) declined from 9.2 × 10–4 to 5.3 × 10–4 L·m–2·min–1 when pH increased from 6.0 to 8.5 (SI Figure S5A). The Cr(VI) formation rate constant from Fe0.25Cr0.75(OH)3(s) decreased from 2.4 × 10–5 to 8.0 × 10–6 L·m–2·min–1 when pH increased from 6.0 to 8.5 (SI Figure S5B). The pH effect was likely associated with metal hydroxo complexation on the surface of FeCr(1–(OH)3(s). For example, the solution pH affects Cr(III)-hydroxo surface complexation.[46,47] Measurement of zeta potential showed that the surface of FeCr(1–(OH)3(s) became more negatively charged with increasing pH (SI Figure S6), suggesting that Cr(III) surface complexes with a higher extent of hydroxylation predominated at higher pHs and they had a lower redox reactivity with chlorine. In addition, at higher pHs, especially above its pK value at 7.6, chlorine speciation favors the deprotonated form of OCl–. The protonated HOCl is a more facile oxidant than the deprotonated OCl–,[48] therefore, an increase in pH from 6 to 8.5 also converts chlorine to a less effective oxidant and contributes to the slow-down of FeCr(1–(OH)3(s) oxidation.

Reaction Stoichiometry and Cr(VI) Formation Mechanisms

The theoretical stoichiometry of the reaction, defined as the molar ratio of Cr(VI) generated to chlorine consumed, that is, Δ[Cr(VI)]/Δ[HOCl], is 0.67 based on the number of electron transfer in the redox reaction. Experimental data with FeCr(1–(OH)3(s) showed that the observed stoichiometry of ΔCr(VI)/ΔHOCl was between 0.02 and 0.24 for all three FeCr(1–(OH)3(s) solids under different chemical conditions (green-color box plots in Figure ). The under-stoichiometric formation of Cr(VI) was likely associated with additional reaction pathways. First, the oxidation of Cr(III) can generate Cr(V) oxide intermediates.[12] Autodecomposition of Cr(V) intermediates reproduces Cr(III) and dissolved O2, resulting in one molecule of O2 generation for every two molecules of chlorine consumption. Indeed, a significant amount of dissolved O2 formation was observed during the oxidation of Cr(III)–Fe(III) hydroxide (Figure ). To take this reaction pathway into consideration, an adjusted stoichiometric ratio was calculated as Δ[Cr(VI)]/(Δ[HOCl] – 2 × Δ[O2]). The adjusted stoichiometry increased significantly, but still fell below the theoretical value of 0.67 (blue-color box plots in Figure ). Second, as the x value in FeCr(1–(OH)3(s) increased from 0.25 to 0.75, the stoichiometry of Δ[Cr(VI)]/Δ[HOCl] decreased significantly (green- and blue-color box plots in Figure ), accompanied by a 36% increase in oxygen generation (Figure ). This suggested an additional reaction pathway specific to the presence of Fe(III) in FeCr(1–(OH)3(s) that consumed chlorine without Cr(VI) generation.

Figure 4

Average stoichiometric ratio for varying water quality parameters and different Cr(III)–Fe(III) hydroxides. The green box plots represent the stoichiometric ratio based on ΔCr(VI)/ΔHOCl. The blue box plots represent the stoichiometric ratio based on ΔCr(VI)/(ΔHOCl – 2ΔO2). The orange box plots represent the stoichiometric ratio based on ΔCr(VI)/(ΔHOCl – αΔO2). The whiskers of the box plot represent the 10th and 90th percentiles of concentration. The lower quartile, middle and upper quartile represent the 25th, 50th, and 75th percentile values.

Figure 5

Oxygen generation over the course of the reaction. Initial [Cr(III)] = 2.8 mM, [HOCl] = 20 mg Cl2/L, Cr(III):Cl2 = 10:1, ionic strength = 10 mM, pH 7. Oxygen data in the background control represents the average background oxygen levels from control experiments that include only chlorine or FeCr(1–(OH)3(s).

To further investigate the reaction mechanism involving Fe(III) surface sites, additional control experiments were conducted in the presence of a pure phase of Fe(III) solid, that is, ferrihydrite Fe(OH)3(s). Data showed that the chlorine decay was accelerated with the introduction of ferrihydrite (Figure A), and was accompanied by the generation of dissolved O2 (Figure B). The nonlinear increase of O2 production with Fe(OH)3(s) dosage is likely due to the limitation on the available Fe(III) reactive surface sites, e.g., aggregation of ferrihydrite particles during the course of the reaction. A correlation between oxygen generation and chlorine decay in the presence of ferrihydrite exhibited a stoichiometric ratio of 0.1 (SI Figure S7), suggesting that the decay of chlorine was enhanced by Fe(III) surface sites with FeCr(1–(OH)3(s). Similar halogen oxidant decay (e.g., HOCl and ClO2) enhanced by transition metal oxides has been previously observed.[49−51]

Figure 6

(A) Chlorine decrease in the presence of ferrihydrite. (B) Oxygen generation in the presence of ferrihydrite. Average background oxygen is 1 mg/L. Initial [HOCl] = 20 mg Cl2/L, ionic strength = 10 mM, pH 7.0.

Accordingly, the reaction stoichiometry of ΔCr(VI)/ΔHOCl during FeCr(1–(OH)3(s) oxidation by chlorine was further adjusted to consider the pathway of Fe(III)-induced chlorine decay and O2 generation. An oxygen coefficient α was introduced to quantify the contribution of reactions that decomposed chlorine to generate O2 while not converting Cr(III) to Cr(VI) in FeCr(1–(OH)3(s). The value of α represented the extent of chlorine scavenging reactions without generating Cr(VI), i.e., two reaction pathways that generate O2: Cr(V) intermediate formation and Fe(III)-induced chlorine decay. When all reaction mechanisms were considered, the adjusted stoichiometry of Δ[Cr(VI)]/(ΔHOCl-α × ΔO2) approached the expected value of 0.67 (orange-color box plots in Figure ).

Overall, the value of α ranged between 2 and 7, but increased with the number of x in FeCr(1–(OH)3(s) (SI Figure S8). A larger α coefficient indicates a stronger inhibitive effect on Cr(VI) formation via FeCr(1–(OH)3(s) oxidation by chlorine. The positive correlation between α and x represents an increasing importance of the Fe(III)-induced chlorine decay pathway as the Fe(III) molar fraction increases in FeCr(1–(OH)3(s).

As chlorine is decayed into dissolved O2 and chloride by Fe(III) surface sites on FeCr(1–(OH)3(s), it can also involve the generation of short-lived intermediate radicals including hydroxyl radical (HO•), chlorine atom (Cl•) and chlorine dimer Cl2•- prior to the formation of chloride and O2.[52,53] Possible reactions were discussed in SI Text S2. To further evaluate the generation of reactive radical species, experiments were conducted with the addition of excess benzoic acid (BA) as a scavenger for all reactive radicals. BA reacts quickly with HO•, Cl• and Cl2•–, but does not react with chlorine or interfere with chlorine decay in the presence of Fe(III) and Cr(III) solids (SI Figure S9). Results showed that in the presence of BA, the formation of O2 via chlorine decay by Fe(OH)3(s) was completely inhibited (SI Figure S10A). The amount of O2 generation via chlorine decay by FeCr(1–(OH)3(s) also decreased by approximately 40% in the presence of BA (SI Figure S10B). This observation suggests that about 40% of oxygen formation was contributed by Fe(III) sites on FeCr(1–(OH)3(s) that involves the formation of reactive radical species, with the remaining 60% contributed by Cr(III) sites via the Cr(V) intermediate formation/decay.

Environmental Implications

Drinking water distribution systems are reactive and lined with corrosion scales that are in situ sources of trace metal contaminants including chromium. In iron-containing distribution systems, accumulated Cr(III) in corrosion scales was reactive with residual chlorine, but the existence of Fe(III) in corrosion scales can greatly diminish Cr(VI) formation. For example, based on the Cr(VI) formation rate constants obtained from this study at different levels of pH, bicarbonate and bromide, as well as the statistical distribution of bromide concentrations in U.S. source waters,[54] a kinetics model was established (SI Text S3). The model predicts that the 50th percentile Cr(VI) concentration in tap water can increase substantially between 1 and 9 μg/L, and a higher molar fraction of Fe(III) in the FeCr(1–(OH)3(s) solid phase generally leads to a lower Cr(VI) formation potential (Figure ).

Figure 7

Box-plot of predicted Cr(VI) formation rate constants and the enhanced Cr(VI) formation in drinking-water distribution systems. The solid line in each box is the median value; the lower and upper box edges are the 25th and 75th percentiles, respectively. Whisker bars represent the 5th and 95th percentiles. The percentile distributions are predicted on the basis of Cr(VI) formation rate constants obtained from this study and statistical distributions of bromide concentration in U.S. source waters. The data with pure phase Cr(III) solids where x = 0 was obtained from a prior study.[12] The kinetics model simulation on enhanced Cr(VI) formation is based on a water distribution system with 0.3 mg/L chlorine residual, 100 μg/L residual Cr(III) solids, and a residence time of 2 days.

To control Cr(VI) level at the tap, especially for aging drinking water distribution infrastructure composed of corroded iron piping materials and free chlorine as the residual disinfectant, effective water management strategies include pH adjustment and bicarbonate control as post-treatment steps prior to entrance to drinking water distribution systems, and the prevention of halide input in source water, for example, seawater intrusion or point source discharge from energy sector. A mobilization or resuspension of Cr(III)-containing iron oxide will increase the risk of in situ Cr(III) conversion to Cr(VI) in drinking water distribution system. Considering that a majority of drinking water distribution infrastructure is based on iron material dominant pipe system, a better control of the redox reactivity of iron-containing Cr(III) solids is important to develop future water management strategies.