The role of chloramines in enhancing soluble manganese removal by manganese contactors by W.R. Knocke, A.M. Weber, P.H. Hargette and E. Odom

Elevated manganese (Mn) concentrations in drinking water can lead to consumer complaints regarding water discoloration. Currently, Mn is not regulated by the U.S. Environmental Protection Agency (EPA) as a primary drinking water standard; however, Mn is addressed as a secondary maximum contaminant level of 0.05 mg/L. However, water utilities are recommended to set effluent Mn concentrations of up to 0.015 mg/L to minimize water discoloration issues (Brandhuber et al. 2013). Mn removal during drinking water treatment typically involves the oxidation of reduced, soluble Mn(II) into insoluble MnO_x(s), which is subsequently removed by a variety of solid–liquid separation processes. The reader is referred to Brandhuber et al. (2013) and Tobiason et al. (2016) for a thorough review of the chemical and biological methods available for accomplishing Mn removal in drinking water treatment scenarios.

One proven method for removing soluble Mn from drinking water sources is the use of MnO_x(s)-coated filter media. Numerous authors (Knocke et al. 1988, 1991; Islam et al. 2010; Barlokova et al. 2017; Habib et al. 2020; Yang et al. 2021) have reported on the beneficial use of coated filter media for Mn removal, with the following being common findings:

• Soluble Mn removal can occur over a pH range of 6–6.2 and higher values.

• Typical filter hydraulic loading rates yield highly effective Mn removal.

• The maintenance of a chlorine residual across the filter media is essential for insuring continuous regeneration of the MnO_x(s) coating on the media surface.

Mn post-filtration contactors use MnO_x(s)-coated media to take advantage of this effective Mn removal process. Contactor treatment systems use media sizes that are significantly larger than conventional filter media (e.g., the effective size (ES) of 1–2.5 mm for contactors as opposed to 0.4–0.8 mm for filtration systems). The use of larger media sizes allows for much higher hydraulic loading rates to be employed in a contactor design, resulting in the ability to treat substantial quantities of water with relatively small contactor footprints being required. Contactors can serve as a primary Mn removal method or as part of a multi-barrier approach to Mn control (Knocke et al. 2010; Brandhuber et al. 2013; Bierlein et al. 2015; Tobiason et al. 2016). Previous studies have shown that a pH range of 7.0–8.0 produces the most effective Mn removal across the contactor depth since the more alkaline pH conditions increase the amount of available Mn adsorption sites on the media (Knocke et al. 2010).

One challenge that can occur with the use of free chlorine as an oxidant during water treatment is an increase in disinfection byproducts (DBPs) such as trihalomethanes and haloacetic acids, which are regulated as maximum contaminant levels (MCLs) by the EPA based upon human health concerns. As such the continuous addition of free chlorine to promote contactor media regeneration can have impacts on DBP formation that must be considered with relevant MCL requirements. It is important to note that the MnO_x(s)-coated media does not promote DBP production (Bazilio et al. 2022); rather, it is simply the presence of free chlorine in water in general that results in such DBP formation.

This study was initiated in response to a specific water treatment situation in Orangeburg, South Carolina. The community's water treatment plant is challenged by a river source water that contains relatively low levels of dissolved Mn (0.01–0.05 mg/L, occasionally exceeding 0.05 mg/L due to heavy rain events) coupled with low pH (pH typically between 5.2 and 5.8) and elevated dissolved organic carbon (DOC) concentrations (average DOC of 8–9 mg/L, but levels of 16–18 mg/L DOC are occasionally seen). Currently, the utility does not experience major water discoloration complaints; however, the utility would like to remove the soluble Mn in their source water to significantly lower concentrations. Previously, the utility explored the use of strong oxidants (e.g., potassium permanganate and chlorine dioxide) to oxidize soluble Mn; however, the relatively low levels of soluble Mn present coupled with the very elevated DOC concentrations in the river proved very challenging in terms of finding a workable oxidant to effectively oxidize the influent Mn. Also, given the very high DOC levels of the water, this water facility uses chloramines (NH₂Cl) as a primary disinfectant throughout the treatment facility. As a result, the water always contains NH₂Cl. Breakpoint chlorination would be necessary upstream of any Mn contactor if free chlorine was desired for media regeneration. This would undoubtedly lead to significant DBP formation and as such precludes this option.

The water in this treatment facility could be considered a prime candidate for Mn removal by a contactor system. The relatively low Mn levels present are quite suitable for contactor deployment. However, a key challenge for implementing this treatment approach would be a means of regenerating the MnO_x(s) coating on the contactor media. If NH₂Cl was not able to regenerate the media during online operations, then the contactor would function as an adsorption process for soluble Mn removal and then be regenerated in an offline mode with free chlorine.

This research study was designed to assess whether NH₂Cl would function as a viable oxidant for regenerating MnO_x(s)-coated media during contactor operations. The research initially focused on a 3-month, bench-scale proof-of-concept evaluation. Once promising results were obtained, the decision was made to conduct a 6-month, pilot-scale evaluation of an Mn contactor system in Orangeburg to assess the longer-term performance of NH₂Cl for media regeneration during online water operations.

The following sections describe the experimental and analytical methods employed during this research study.

Bench-scale Mn contactor experiments employed the use of a ceramic media (Ceralite-A; Wateropolis, Newbury, OH) with MnO_x(s) coating applied by the manufacturer (coating was between 6.5 and 7.5 mg Mn/g Ceralite-A media). The media had an effective size of approximately 0.65 mm and a uniformity coefficient of <1.5. The media was initially batch treated (12-h contact time) in a strong bleach solution (30–35 mg/L free chlorine) to help ensure that the MnO_x(s) coating was fully regenerated at the start of any contactor experiments. The media was also rinsed thoroughly to remove any fines prior to experimental use.

Bench-scale testing employed glass columns of 1.6 cm internal diameter. A media depth of 0.3 m (12″) was used for initial contactor tests that were conducted at a hydraulic loading rate (HLR) of 4.9 m/hr (2 gpm/ft²), whereas a media depth of 0.45 m (18″) was used for subsequent contactor tests that employed an HLR of either 19.6 or 39.2 m/hr (8 or 16 gpm/ft², respectively). An intermediate sampling port allowed for Mn testing to occur at multiple media depths. The duration of contactor loading varied from 16 to 40 days, with the shortest duration being associated with the highest HLR evaluated and the longest duration associated with the 4.9 m/hr (2 gpm/ft²) experiment.

The water fed to the columns was deionized water to which sodium bicarbonate was added to produce an alkalinity of 100 mg/L as CaCO₃ and calcium chloride was added to yield a hardness of 50 mg/L as CaCO₃. The pH of the water applied to the contactors was typically between 8.0 and 8.3, a range chosen as often typical of finished water pH conditions at a water treatment facility (for corrosion control) and very comparable to that used at the full-scale treatment facility in Orangeburg. The soluble Mn concentration of the contactor feed water was adjusted by feeding a known amount of manganous sulfate into each batch of water prepared. The target influent soluble Mn concentration for the studies using an HLR of 4.9 m/hr (2 gpm/ft²) was 0.20 mg/L, whereas a target Mn concentration of 0.05–0.06 mg/L was employed for the testing at HLR values of 19.6 and 39.2 m/hr (8 and 16 gpm/ft², respectively). Fresh feed water was prepared every 12–24 h throughout the bench-scale contactor studies depending upon the HLR being employed.

A monochloramine (NH₂Cl) stock solution was prepared by reacting a high-strength ammonia solution (1,400 mg/L as NH₃-N) with a high-strength sodium hypochlorite solution (5,000 mg/L as Cl₂) using a well-ventilated fume hood. A known amount of chloramine stock solution was added to the feed water batch to achieve the desired feed concentration (typically 4 mg/L; some initial testing involved no chloramine addition to allow for the side-by-side evaluation of the soluble Mn removal that would occur due to simple adsorptive uptake on the MnO_x(s)-coated media). Preliminary testing demonstrated that chloramines were unable to oxidize soluble Mn to any significant degree in bulk solution over a multi-day period of contact. Likewise, periodic testing of the stock solution showed no evidence of Mn oxidation taking place. Finally, the 4 mg/L NH₂Cl concentration was chosen to reflect typical operating conditions at the Orangeburg treatment facility; likewise, it represents the maximum concentration in drinking water allowed by the U.S. EPA regulation.

Initial bench-scale testing involved side-by-side contactor columns loaded at a HLR of 4.9 m/hr (2 gpm/ft²) and an influent Mn concentration of approximately 0.2 mg/L. One column had no NH₂Cl in the feed water, whereas the other column received feed water containing 4 mg/L NH₂Cl. Following the completion of this initial comparative performance study, subsequent contactor testing employed a HLR of either 19.6 or 39.2 m/hr (8 or 16 gpm/ft², respectively). Feed solutions for these latter studies had NH₂Cl present.

Samples of both the influent (feed) and contactor effluent were collected on at least a daily basis. The Mn concentration of samples was analyzed by a Thermo Electron iCAP-RQ (Thermo Electron North American LLCL, West Palm Beach, FL) inductively coupled plasma mass spectrometer (ICP-MS) per Standard Method 3125-B (APHA, AWWA, and WEF 1998). Samples and calibration standards were prepared in a matrix of 2% nitric acid by volume. A HACH spectrophotometer (DR 2800) was used to measure NH₂Cl concentrations using method 10171. Solution pH and temperature were measured by a Fisher Scientific Accumet AE150 benchtop pH meter.

Pilot testing was conducted at the John F. Pearson Water Treatment Plant in Orangeburg. Two pilot-scale Plexiglas columns (0.1 m ID (4″ ID)) were installed in the filter pipe gallery at the plant. Pilot testing had a goal of achieving Mn levels in the treated water at or below 0.02 mg/L. Each contactor was filled with graded support gravel (0.3 m (12″) depth) and 0.7 m (28″) of MnO_x(s)-coated media:

• Column 1: 1.6 mm effective size ceramic media (Ceralite-A, coated with MnO_x(s) from Wateropolis; same type of media as used in bench-scale testing) and

• Column 2: 2.0 mm effective size anthracite media (coated with MnO_x(s) from WesTech).

The use of two different media sizes allowed for an evaluation of media size on both observed Mn removal efficiency and head loss accumulation across the contactors.

Influent water for the pilot contactors was collected post-filtration in the water plant and then pumped to a constant head tank located in the filter gallery. The filtered water received lime addition in the full-scale plant to raise the pH to a target range of 8.0–8.5, as well as supplemental NH₂Cl to a target of approximately 4.0 mg/L. For much of the pilot study, the ambient Mn in the filtered water was supplemented to achieve a 0.05–0.06 mg/L Mn influent concentration to the contactors. This was achieved by metering in a concentrated manganous chloride solution. There were limited periods of time when no supplemental Mn was added to evaluate the contactor performance at day-to-day levels of Mn experienced at the plant. Near the end of the pilot study, there was one test that involved spiking to a target contactor influent Mn level into the range of 0.08–0.10 mg/L.

Pilot-scale testing involved the HLR being set at either 29.4 or 37.8 m/hr (12 or 15 gpm/ft²). High HLRs were employed based upon a practical consideration of the full-scale treatment facility. Limited space was available to possibly retrofit Mn contactors into the plant layout; as such, using higher HLRs in design would decrease the footprint of a full-scale Mn contactor installation.

Water samples were collected daily at the contactor influent at a media depth of 0.25 m (10″) (i.e., intermediate port) and at the contactor effluent port. Samples were analyzed onsite for Mn concentrations using a HACH test kit and corresponding methodology. Thirty of these samples during the study were also analyzed using ICP-MS (detection limit of 0.0004 mg/L). The comparative results confirmed that the onsite HACH method provided reliable Mn analyses down to concentrations near 0.007 mg/L. The contactor feed water samples were also analyzed for pH, chloramine concentration, and temperature.

The maximum operational head loss accumulation across the pilot-scale contactors was approximately 0.7–0.9 m (30–35″) due to the hydraulic configuration at the treatment plant. Contactor columns were taken offline to backwash when head loss accumulation exceeded the system capacity. Backwashing used finished water from the treatment facility. Prior research (Subramaniam 2010) had shown that Mn contactors do generate some amount of head loss over several weeks of operation, most likely due to a buildup of additional MnO_x(s) coating on the contactor media. Backwashing was shown to be a very effective way of addressing any head loss accumulation.

In addition, an offline, free chlorine regeneration step was employed after most backwashes to restore the media MnO_x(s) coating to a fully oxidized state. The regeneration step was undertaken to ultimately assess the degree to which chloramines were able to regenerate the MnO_x(s) media coating during online operations. In two instances, only backwashing was employed without free chlorine regeneration to assess whether offline regeneration was necessary after each backwashing cycle.

Regeneration of the MnO_x(s) coating involved soaking the media in a strong free chlorine solution, which would oxidize any portion of the MnO_x(s) media coating that was not oxidized by the presence of chloramines during conventional contactor operations. Media regeneration consisted of two soak periods with a free chlorine solution concentration of approximately 200–220 mg/L as Cl₂. In each instance, the chlorine was allowed to be in contact with the media for approximately 2 h. Water samples were collected from the two depths (0.25 and 0.7 m) into the media and tested for residual free chlorine concentration between each soaking period. When the residual free chlorine concentration was higher than approximately 10 mg/L at the end of a soak period, the media in the column was deemed to be appropriately regenerated. At that point, the media was backwashed again to remove any residual free chlorine, and the next contactor testing period was initiated.

The effluent NH₂Cl concentration was sampled periodically from each of the Mn contactor columns to assess the NH₂Cl demand being exerted in association with soluble Mn removal. The theoretical stoichiometry of oxidation–reduction interaction between NH₂Cl and reduced Mn should have produced NH₂Cl reductions of 0.1–0.3 mg/L across the contactor, depending upon the influent soluble Mn level in a given test. Instead, the observed NH₂Cl demand across the contactors was in the 0.5–0.7 mg/L range, more than double the theoretical value. No mechanistic explanation could be developed to address the larger-than-expected NH₂Cl demand. It is important to note that actual NH₂Cl demand was occurring across the Mn contactors receiving NH₂Cl in the applied water. This further reinforced the hypothesis that NH₂Cl was promoting at least partial regeneration of the MnO_x(s) present on the contactor media surface.

Both contactors performed well in terms of lowering the treated Mn concentration to values at or below 0.02 mg/L; only a few observed values exceeded 0.02 mg/L, and even in those instances, the EPA secondary MCL of 0.05 mg/L was achieved. The contactor performance data for Mn removal are summarized in Tables 1 and 2. Both contactor media configurations yielded average effluent Mn concentrations that were below the 0.015 mg/L threshold to protect against aesthetic concerns (Brandhuber et al. 2013).

The smaller media size (1.6 mm ES) yielded improved Mn removal as evidenced by a lower average treated water Mn concentration (0.010 versus 0.014 mg/L) and slightly a higher percent removal performance for a given HLR. This would be explained by the fact that the smaller media size would yield a higher media-specific surface area (and, hence, the greater density of Mn adsorptive sites on the media) as a function of contactor media depth. Also, operating at a higher HLR did negatively impact the observed Mn removal for both media sizes, a result that was expected. Higher HLRs effectively reduce the residence time of the liquid within the contactor, decreasing the time available for soluble Mn adsorption to the media surface.

It is important to note that the observed head loss generation in the contactors was not associated with any type of particulate matter capture (since the water fed to the contactors had been previously treated by dual-media filtration in the full-scale plant). Instead, the uptake of soluble Mn coupled with the surface-based oxidation of the Mn does yield some observed head loss. This same phenomenon has been reported when Mn contactors employ free chlorine as the media regeneration process (Subramaniam 2010). Backwashing the contactor media removed accumulated head loss associated with Mn removal and restored the hydraulic behavior of the media.

Chloramine demand exerted across each Mn contactor was quantified daily based upon analysis of an influent and effluent sample for residual chloramine concentration. The theoretical chloramine demand exerted would have been no more than 0.1 mg/L NH₂Cl based upon an average Mn influent of approximately 0.05 mg/L across the period of pilot testing. As such it would have been difficult to observe a significant change in chloramine concentration across the contactors. Comparing the influent and effluent chloramine concentrations measured daily for the duration of the pilot testing yielded a calculated average chloramine demand across the contactors of 0.08 mg/L, which is comparable to the theoretical demand calculated.

The role of chloramine in media regeneration was also evaluated based upon the amount of free chlorine that was required to fully regenerate the media in an offline manner following backwash. Table 3 presents a summary of the free chlorine regeneration data.

Across the duration of the pilot-scale Mn contactor study, approximately 80% of the oxidant needs to regenerate the MnO_x(s)-coated media (with respect to oxidation of the previously adsorbed soluble Mn) were being met by the chloramines present in the contactor-applied water. The remaining ±20% of the media regeneration needs were met by the offline free chlorine application. From a practical standpoint, the pilot-scale results demonstrated that the contactors could operate for at least the time period between backwashes without requiring free chlorine application and still produce a high-quality effluent with respect to Mn concentration. In fact, the contactors on occasion were not regenerated with free chlorine after backwash and allowed to go until the next backwash before offline regeneration with free chlorine occurred. In one instance, the contactors operated for essential an entire month without offline free chlorine regeneration. These results confirmed the earlier bench-scale testing, which demonstrated that chloramines were able to achieve a significant level of media regeneration when present during contactor operations.

Specific conclusions from this applied research study include the following:

• The addition of NH₂Cl to the water applied to the MnO_x(s) media in the contactor resulted in substantial regeneration of the media coating during contactor operations. This result indicates that NH₂Cl would be a practical means of promoting online contactor media regeneration. In a full-scale treatment situation, this would significantly increase the time period of operation before the contactor would need to be taken offline for free chlorine regeneration of the media.

• Hydraulic loading rates of up to 39.2 m/hr (16 gpm/ft²) could be utilized and still produce high-quality contactor effluent with respect to soluble Mn concentrations. An ability to use loading rates this high would be beneficial for decreasing the total amount of contactor surface area required to treat the water at a given treatment facility.

• The ability of NH₂Cl to regenerate the MnO_x(s) media coating was indirectly related to the contact time of the NH₂Cl with the media surface. A lower hydraulic loading rate promoted a more effective reaction between NH₂Cl and the MnO_x(s) surface, increasing the amount of Mn that is oxidized per unit time of contactor operation. However, the NH₂Cl was able to promote sufficient rates of media oxidation at high hydraulic loading rates so as to offer a benefit to overall contactor operations.

This study focused on the application of this treatment technique under conditions specific to the Orangeburg situation. Nonetheless, the research findings have relevance to other water treatment scenarios. For example, this study employed pH 8.0–8.3 conditions since (a) that was the pH condition where a Mn contactor would be employed a full scale in the Orangeburg treatment train and (b) this Mn adsorptive contactor process was known to work better at higher pH (Knocke et al. 2010). An alkaline pH condition of 8.0–8.3 is routinely employed by many water utilities for corrosion control in the distribution system anyhow. Mn removal via contactor treatment would most likely be less effective if the water applied to a contactor was significantly below pH 8.0.

Likewise, this study focused on NH₂Cl concentrations near 4 mg/L as this was the prevalent situation for a contactor deployment in Orangeburg. This resulted in approximately 80% of the contactor media's regeneration requirements being met by chloramines, with approximately 20% of the media regeneration requirements being met by free chlorine application during offline periods. Using lower applied NH₂Cl concentrations during online operations would most likely result in more offline free chlorine being needed to insure full regeneration of the contactor media for Mn removal.

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

The authors declare there is no conflict.