Development of Membrane Filter Colorimetry for Determining Tap Water Discoloration Using a Spectrocolorimeter

: In this study, we aimed to develop a novel method to quantify residual colorants in the water supply using a spectrocolorimeter. Cross-tests of color and turbidity showed that standard color solutions of 1–50 color units had a turbidity of <0.094 nephelometric turbidity unit (NTU) and turbidity standard solutions of 0.1–5 NTU had color measurements of 0 true color unit, indicating limitations in measuring colorants using conventional methods. Therefore, the samples were diluted stepwise to 1 L and ﬁltered through a 0.45 µ m cellulose membrane; then, the residuals colorants were measured using membrane ﬁlter colorimetry (MFC) with a spectrocolorimeter to determine the color difference. The color difference exhibited a high correlation with turbidity. Furthermore, scanning electron microscopy/energy-dispersive X-ray spectroscopy and inductively coupled plasma-optical emission spectrometry analyses of the same samples conﬁrmed that the main components of the colorants were iron and manganese; the concentration of these substances in the samples was measured using MFC with the standards Fe 2 O 3 · H 2 O and MnO 2 . The results conﬁrmed a high correlation between the color difference ( (cid:52) E*ab) and concentration of the samples and standard substances. Our ﬁndings suggest that MFC is a promising approach for measuring colorants in drinking water.


Introduction
Although South Korea's water supply network is being designated by the United States Environmental Protection Agency (USEPA) as one of the greatest contributions to humanity in the 20th century, it is inadequately maintained for emergency situations.Numerous cities often face situations where the water supply is interrupted because of large pipe bursts and water quality accidents caused by corroding water pipes [1].
Discoloration events in water supply networks are primarily caused by the resuspension of accumulated loose particles, which mainly originate from water treatment plants [1] due to incomplete treatment of particulate matters [2].The water supply network can also produce particulate matters which can be attributed to corrosion at water pipe connectors and detachment of lining material [3,4].
Rust water, primarily caused by oxidized colloidal ferrous hydroxide or manganese (Mn) present in water pipes, discolors tap water.As iron (Fe)-containing minerals are more abundant than Mn-containing minerals, Fe is more commonly found in groundwater [5].Furthermore, when water passes through soil, oxygen content is depleted owing to organic matter decomposition and aerobic microorganism activities in the soil.Consequently, Fe is reduced to Fe(II), which may enter the water body at the soil and water interface [5].Dissolved Fe 2+ undergoes oxidation to produce Fe 3+ , leading to the discoloration of water within the network.Even low concentrations of Fe can cause discoloration and increase water turbidity [6].In addition, even at low concentrations, Mn affects the taste of drinking water [7].Particularly, trace amounts of Mn in tap water are oxidized by free residual chlorine, leading to an increase in the intensity of color by approximately 300-400 times [8].When Mn is not treated in water treatment plants, it enters water distribution pipes, resulting in the generation of blackwater [9].A substantial portion of customer complaints to water supply companies worldwide stem from discolored water originating from drinking water distribution systems (DWDS).For instance, in the United Kingdom (UK), discoloration issues accounted for 34% of all customer complaints to water companies over a 5-year period (2015-2020) [10].
When measuring colorants using standard methods, high concentrations of coloring substances can negatively interfere with turbidity measurements [11].As turbid materials in samples can interfere with color measurements [12], it is necessary to eliminate turbidity as a pre-treatment step [13].Therefore, as depicted in Figure 1, the results of water quality analysis performed using available turbidity and color tests do not accurately reflect the color of discolored tap water, thus, causing customer dissatisfaction.300-400 times [8].When Mn is not treated in water treatment plants, it enters wat tribution pipes, resulting in the generation of blackwater [9].
A substantial portion of customer complaints to water supply companies worl stem from discolored water originating from drinking water distribution sy (DWDS).For instance, in the United Kingdom (UK), discoloration issues account 34% of all customer complaints to water companies over a 5-year period (2015)(2016)(2017)(2018)(2019)(2020) When measuring colorants using standard methods, high concentrations of co substances can negatively interfere with turbidity measurements [11].As turbid ma in samples can interfere with color measurements [12], it is necessary to eliminate t ity as a pre-treatment step [13].Therefore, as depicted in Figure 1, the results of quality analysis performed using available turbidity and color tests do not accurat flect the color of discolored tap water, thus, causing customer dissatisfaction.
To effectively respond to water quality issues caused by colored particles in tap and address the limitations of water quality standard parameters such as turbidi color, the National Institute of Environmental Research has proposed a filter test m for water quality management [14].However, this method involves visual inspec filters' color, making quantitative measurement difficult.Colorimeters and spectr tometers have been extensively used in the food industry for color measuremen Therefore, in this study, we investigated a new test method by applying spectropho ric colorimetry quantification.

Pipeline Status of the Study Area and Field Sample Collection
As of the end of 2021, Cheong-Ju (CJ) City, the study area in Chungcheongb had a population of 836,711 individuals receiving water supply at a rate of 97.2%.T supplies a total quantity of 350,000 m 3 /d of domestic water from Jibuk (BJ) and CJ treatment plants.The total length of pipelines in CJ City is 3066 km, and as shown ure 2, 78% of them are ˂30 years old.Therefore, the proportion of aged pipes is rel low.To effectively respond to water quality issues caused by colored particles in tap water and address the limitations of water quality standard parameters such as turbidity and color, the National Institute of Environmental Research has proposed a filter test method for water quality management [14].However, this method involves visual inspection of filters' color, making quantitative measurement difficult.Colorimeters and spectrophotometers have been extensively used in the food industry for color measurement [15].Therefore, in this study, we investigated a new test method by applying spectrophotometric colorimetry quantification.

Pipeline Status of the Study Area and Field Sample Collection
As of the end of 2021, Cheong-Ju (CJ) City, the study area in Chungcheongbuk-do, had a population of 836,711 individuals receiving water supply at a rate of 97.2%.The city supplies a total quantity of 350,000 m 3 /d of domestic water from Jibuk (BJ) and CJ water treatment plants.The total length of pipelines in CJ City is 3066 km, and as shown in Figure 2, 78% of them are <30 years old.Therefore, the proportion of aged pipes is relatively low.
Samples of sediments and coloring materials were obtained from pipes of a reservoir, a water tank, and an apartment during the supply of tap water in areas serviced by two water treatment plants located in CJ City (Figure 3).The samples were collected at three points: The YulRang reservoir in Cheongwon-gu (Sample 1, YR), a Gagyeong Daewon apartment water tank in Heungdeok-gu (Sample 2, DT), and a Sachang Buheung apartment  Samples of sediments and coloring materials were obtained from pipes of a reservoir, a water tank, and an apartment during the supply of tap water in areas serviced by two water treatment plants located in CJ City (Figure 3).The samples were collected at three points: The YulRang reservoir in Cheongwon-gu (Sample 1, YR), a Gagyeong Daewon apartment water tank in Heungdeok-gu (Sample 2, DT), and a Sachang Buheung apartment in Seowon-gu (Sample 3, BC).The pipe material, pipe diameter, and installation year at each sample collection point are presented in Table 1.Samples of sediments and coloring materials were obtained from pipes of a reservoir, a water tank, and an apartment during the supply of tap water in areas serviced by two water treatment plants located in CJ City (Figure 3).The samples were collected at three points: The YulRang reservoir in Cheongwon-gu (Sample 1, YR), a Gagyeong Daewon apartment water tank in Heungdeok-gu (Sample 2, DT), and a Sachang Buheung apartment in Seowon-gu (Sample 3, BC).The pipe material, pipe diameter, and installation year at each sample collection point are presented in Table 1.

Coloring Reference Material
The most common forms of Fe and Mn compounds, Fe 2 O 3 •H 2 O and MnO 2 , which cause discoloration of water in the DWDS, were selected as coloring reference materials and prepared as shown in Table 2.The dissolved solution of iron (yellow) was prepared with a concentration of 1000 mg/L as Fe, using 1.62 g of 98% Fe 2 O 3 •H 2 O. Similarly, the dissolved solution of manganese (black) was prepared at a concentration of 1000 mg/L as Mn, using 1.75 g of 90% MnO 2 .

Coloring Reference Material
The most common forms of Fe and Mn compounds, Fe2O3•H2O and MnO2, which cause discoloration of water in the DWDS, were selected as coloring reference materials and prepared as shown in Table 2.The dissolved solution of iron (yellow) was prepared with a concentration of 1000 mg/L as Fe, using 1.62 g of 98% Fe2O3•H2O.Similarly, the dissolved solution of manganese (black) was prepared at a concentration of 1000 mg/L as Mn, using 1.75 g of 90% MnO2.In field emission scanning electron microscopy (FE-SEM), electrostatic and magnetic fields are combined to maximize the optical performance of the objective lens, thereby enabling excellent imaging of samples, including samples with characteristics similar to those of magnetic materials.Moreover, through the beam booster method, FE-SEM can capture images of particles less than a nanometer without an immersion lens at very low voltages (<1 kV).
The FE-SEM system used in this study was ULTRA Plus (ZEISS, Oberkochen, Germany) and the detector (energy dispersive X-ray spectroscopy, EDS) was XTrace (BRUKER, Berlin, Germany).

Inductively Coupled Plasma Optical Emission Spectroscopy
Inductively coupled plasma optical emission spectroscopy (ICP-OES) enables qualitative or quantitative analysis based on the wavelength and intensity of elements emitted by high-temperature plasma.Liquid samples are aerosolized and injected into high-temperature plasma for decomposition into atoms; the device then analyzes the atoms through ionization and excitation.The ICP-OES device used in this study was ARCOS FMX-36 (SPECTRO, Klev, Germany).For sample preprocessing, combining a microwave heating equipment with a high-pressure reactor enhances sample decomposition and shortens the processing time.The microwave equipment used in this study was Ul-traWAVE (Milestone, Sorisole, Italy).

Colorimetry
In the Drinking Water Quality Process Test Standard of South Korea, color measurement methods include visual colorimetric and spectrophotometric color measurement methods.Among the spectrophotometric methods, the multi-wavelength method (based Water 2023, 15, x FOR PEER REVIEW 4 of 18

Coloring Reference Material
The most common forms of Fe and Mn compounds, Fe2O3•H2O and MnO2, which cause discoloration of water in the DWDS, were selected as coloring reference materials and prepared as shown in Table 2.The dissolved solution of iron (yellow) was prepared with a concentration of 1000 mg/L as Fe, using 1.62 g of 98% Fe2O3•H2O.Similarly, the dissolved solution of manganese (black) was prepared at a concentration of 1000 mg/L as Mn, using 1.75 g of 90% MnO2.Dissolved solution color (1000 mg/L)

Field Emission Scanning Electron Microscopy
In field emission scanning electron microscopy (FE-SEM), electrostatic and magnetic fields are combined to maximize the optical performance of the objective lens, thereby enabling excellent imaging of samples, including samples with characteristics similar to those of magnetic materials.Moreover, through the beam booster method, FE-SEM can capture images of particles less than a nanometer without an immersion lens at very low voltages (<1 kV).
The FE-SEM system used in this study was ULTRA Plus (ZEISS, Oberkochen, Germany) and the detector (energy dispersive X-ray spectroscopy, EDS) was XTrace (BRUKER, Berlin, Germany).

Inductively Coupled Plasma Optical Emission Spectroscopy
Inductively coupled plasma optical emission spectroscopy (ICP-OES) enables qualitative or quantitative analysis based on the wavelength and intensity of elements emitted by high-temperature plasma.Liquid samples are aerosolized and injected into high-temperature plasma for decomposition into atoms; the device then analyzes the atoms through ionization and excitation.The ICP-OES device used in this study was ARCOS FMX-36 (SPECTRO, Klev, Germany).For sample preprocessing, combining a microwave heating equipment with a high-pressure reactor enhances sample decomposition and shortens the processing time.The microwave equipment used in this study was Ul-traWAVE (Milestone, Sorisole, Italy).

Colorimetry
In the Drinking Water Quality Process Test Standard of South Korea, color measurement methods include visual colorimetric and spectrophotometric color measurement methods.Among the spectrophotometric methods, the multi-wavelength method (based

Field Emission Scanning Electron Microscopy
In field emission scanning electron microscopy (FE-SEM), electrostatic and magnetic fields are combined to maximize the optical performance of the objective lens, thereby enabling excellent imaging of samples, including samples with characteristics similar to those of magnetic materials.Moreover, through the beam booster method, FE-SEM can capture images of particles less than a nanometer without an immersion lens at very low voltages (<1 kV).
The FE-SEM system used in this study was ULTRA Plus (ZEISS, Oberkochen, Germany) and the detector (energy dispersive X-ray spectroscopy, EDS) was XTrace (BRUKER, Berlin, Germany).

Inductively Coupled Plasma Optical Emission Spectroscopy
Inductively coupled plasma optical emission spectroscopy (ICP-OES) enables qualitative or quantitative analysis based on the wavelength and intensity of elements emitted by high-temperature plasma.Liquid samples are aerosolized and injected into hightemperature plasma for decomposition into atoms; the device then analyzes the atoms through ionization and excitation.The ICP-OES device used in this study was ARCOS FMX-36 (SPECTRO, Klev, Germany).For sample preprocessing, combining a microwave heating equipment with a high-pressure reactor enhances sample decomposition and shortens the processing time.The microwave equipment used in this study was UltraWAVE (Milestone, Sorisole, Italy).

Colorimetry
In the Drinking Water Quality Process Test Standard of South Korea, color measurement methods include visual colorimetric and spectrophotometric color measurement methods.Among the spectrophotometric methods, the multi-wavelength method (based on the principle of absorption) uses either the 10-or 30-division method.The wavelengths selected correspond to various organic and inorganic compounds that cause water coloration.
According to the Drinking Water Quality Process Test Standard, if a sample is turbid, color measurement using the multi-wavelength method is performed by determining the true color unit (TCU) after passing the sample through a 0.45 µm filter to eliminate any interference effects.Color measurement in this study was conducted using RICO 690 (HACH LANGE, Berlin, Germany).

Turbidity Measurement
Turbidity is measured based on the optical properties of water, where light is scattered and absorbed by particles.In South Korea, the turbidity of tap water is measured using the nephelometric turbidity unit (NTU) per the Drinking Water Quality Process Test Standard.The turbidity meter, equipped with a light source and photoelectric detector, uses a tungsten filament as the light source and detects scattered light.The color of the medium or measuring cell influences the turbidity measurement [16].In this study, we used HACH 2100AN (HACH Company, Loveland, CO, USA), which meets the criteria set in the Drinking Water Quality Process Test Standard, as the turbidity meter.

Spectrocolorimetry
Spectrocolorimeter (or spectrophotometer) and chroma meter are used to quantify color by assigning values to hue, brightness, and saturation, enabling the distinction of minute color differences not discernible by the human eye [17].
ater 2023, 15, x FOR PEER REVIEW true color unit (TCU) after passing the sample through a 0.45 µm filter to elim interference effects.Color measurement in this study was conducted using (HACH LANGE, Berlin, Germany).

Turbidity Measurement
Turbidity is measured based on the optical properties of water, where lig tered and absorbed by particles.In South Korea, the turbidity of tap water is using the nephelometric turbidity unit (NTU) per the Drinking Water Quality P Standard.The turbidity meter, equipped with a light source and photoelectri uses a tungsten filament as the light source and detects scattered light.The c medium or measuring cell influences the turbidity measurement [16].In this used HACH 2100AN (HACH Company, Loveland, CO, USA), which meets the in the Drinking Water Quality Process Test Standard, as the turbidity meter.

Spectrocolorimetry
Spectrocolorimeter (or spectrophotometer) and chroma meter are used t color by assigning values to hue, brightness, and saturation, enabling the dis minute color differences not discernible by the human eye [17].
In this study, we used a spectrocolorimeter equipped with a chroma meter model; KONICA MINOLTA, Tokyo, Japan, Figure 4b).

Experimental Method
To quantify coloring substances in tap water, we aimed to develop a new filter colorimetry (MFC) method that quantifies the color difference, △E*ab.We In this study, we used a spectrocolorimeter equipped with a chroma meter (CM-700d model; KONICA MINOLTA, Tokyo, Japan, Figure 4b).

Experimental 2.4.1. Experimental Method
To quantify coloring substances in tap water, we aimed to develop a new membrane filter colorimetry (MFC) method that quantifies the color difference, E*ab.We performed color cross-testing and measured turbidity to confirm the limitations of available methods.Furthermore, we quantified coloring substances such as Fe and Mn compounds in the samples using scanning electron microscopy/energy-dispersive X-ray spectroscopy (SEM-EDS) and ICP-OES, prepared standard solutions of similar concentrations, and passed the samples and standard solutions through a 0.45 µm filter.The samples and standard solutions were analyzed using a spectrocolorimeter.

Review of Previous Testing Methods
In this study, we considered two methods.The first method was the filter test method (FTM), in which color is measured visually by passing a solution through a filter and then observing the color of the filter paper, per the guidelines for the use of FTM by the National Institute of Environmental Research (NIER).The second was the membrane patch colorimetry (MPC) method, in which color is quantified spectroscopically by first filtering lubricant oil and measuring the color of the filter paper using a colorimeter, per the "Standard Test Method for Measurement of Lubricant Generated Insoluble Color Bodies in In-Service Turbine Oils using Membrane Patch Colorimetry" by the American Society for Testing and Materials (ASTM).
The approaches currently used by the NIER and ASTM measure colorants in drinking water and lubricant oil; they have advantages and disadvantages in terms of specificity of samples and quantification methods.Therefore, we combined the beneficial elements of these two methods and modified them to develop a new test method, MFC.Table 3 presents a comparison of the new and old test methods.passed the samples and standard solutions through a 0.45 µm filter.The samples and standard solutions were analyzed using a spectrocolorimeter.

Review of Previous Testing Methods
In this study, we considered two methods.The first method was the filter test method (FTM), in which color is measured visually by passing a solution through a filter and then observing the color of the filter paper, per the guidelines for the use of FTM by the National Institute of Environmental Research (NIER).The second was the membrane patch colorimetry (MPC) method, in which color is quantified spectroscopically by first filtering lubricant oil and measuring the color of the filter paper using a colorimeter, per the "Standard Test Method for Measurement of Lubricant Generated Insoluble Color Bodies in In-Service Turbine Oils using Membrane Patch Colorimetry" by the American Society for Testing and Materials (ASTM).
The approaches currently used by the NIER and ASTM measure colorants in drinking water and lubricant oil; they have advantages and disadvantages in terms of specificity of samples and quantification methods.Therefore, we combined the beneficial elements of these two methods and modified them to develop a new test method, MFC.Table 3 presents a comparison of the new and old test methods.This testing method can be used to analyze insoluble coloring substances in drinking water (e.g., tap water) by using a filter according to the standards in Table 4, following the This testing method can be used to analyze insoluble coloring substances in drinking water (e.g., tap water) by using a filter according to the standards in Table 4, following the procedure described below.The procedure of MFC is as follows.
A filter was positioned between the filter support and funnel and secured using a filter holder.The filter was then moistened using a small amount of water, and a vacuum (33 kPa↓) was applied.The sample was then vigorously mixed for 30 s to homogenize the coloring substance.Immediately after mixing, 1 L of the sample was poured into the filter funnel, which was then rinsed with at least 35 mL of water to allow the filtrate to flow completely.Subsequently, the vacuum was turned off; the filter was separated, placed in a Petri dish, and allowed to air-dry for 3 h.Thereafter, the background color was measured using a spectrocolorimeter with the same filter used to filter the 1 L of distilled water.The color difference (∆E*ab) between the measured distilled water background color and the sample was selected, and the CIELAB ∆E*ab value of the spectrocolorimeter to the first decimal place was recorded.The ∆E value is expressed as FCU.If the ∆E*ab value was 4.0, it was read as 4.0 FCUs.

Elemental Analysis of Samples
To identify the elements of colorants that leak from water storage tank sediments and pipe networks, we analyzed the samples using FESEM-EDS and ICP-OES.
The results of the FESEM-EDS analysis revealed that the concentrations of oxygen (O, 56.4%) and carbon (C, 29.4%) were the highest, followed by those of nitrogen (N, 7.9%), aluminum (Al, 1.6%), Fe (3.5%), Mn (0.25%), silicon (Si, 0.34%), zinc (Zn, 0.45%), and chlorine (Cl, 0.09%).Furthermore, the analysis revealed that Al residues from coagulants used in water treatment and organic compounds C, O, and N in their oxidized forms, which were either not fully treated during the water treatment process or resuspended within the pipe network, were the elements present at the highest concentrations.In addition, it was determined that the water could become discolored when substances containing coloring inorganic elements such as Fe and Mn were ejected.Figure 5 presents the analysis results of sediments obtained from the water supply system in CJ City to further accurately compare the SEM-EDS analysis results.
Furthermore, the ICP-OES analysis of the samples obtained from the CJ City water supply system revealed that the concentration of colorants in samples increased from an average of 165.14 to 235.53 mg/kg for Fe and decreased from 113.81 to 6.00 mg/kg for Mn in the YR, DT, and BC samples (Table 5 and Figure 6).These results are consistent with the FESEM-EDS analysis results, in which, the Fe concentration increased to 1.16%, 1.31%, and 8.12% in the YR, DT, and BC samples, respectively, and the Mn concentration decreased to 0.60%, 0.14%, and 0%, respectively.It is evident that in the water distribution network, the concentration of Mn not treated in the water treatment process decreases, as Mn settles in reservoirs and tanks, whereas that of Fe increases due to pipe corrosion.Based on FESEM-EDS test results, reference materials were prepared according to the content of Fe and Mn.
the pipe network, were the elements present at the highest concentrations.In addit was determined that the water could become discolored when substances containin oring inorganic elements such as Fe and Mn were ejected.Figure 5 presents the an results of sediments obtained from the water supply system in CJ City to further rately compare the SEM-EDS analysis results.

Color and Turbidity Cross-Test
A cross-test of color and turbidity was conducted to determine the limitations of available color and turbidity test methods.The turbidity was measured using a color standard solution of potassium hexachloroplatinate and cobalt chloride (K 2 PtCl 6 +CoCl 2 •6H 2 O), according to the standard methods.The color was measured using a turbidity standard solution of hydrazine sulphate and hexamethylenetetramine ((NH 2 ) 2 •H 2 SO 4 +(CH 2 ) 6 N 4 ).The results are shown in Table 6.

Color and Turbidity Cross-Test
A cross-test of color and turbidity was conducted to determine the lim available color and turbidity test methods.The turbidity was measured usin standard solution of potassium hexachloroplatinate and cobalt (K2PtCl6+CoCl2⋅6H2O), according to the standard methods.The color was measu a turbidity standard solution of hydrazine sulphate and hexamethylene ((NH2)2⋅H2SO4+(CH2)6N4).The results are shown in Table 6.Based on the cross-test results (Table 6), the turbidity of the color standard solution (1-50 color units (CU)) was consistently within the range of 0.052-0.081NTU, regardless of the concentration of the solution, demonstrating the difficulty in measuring the color standard solution with turbidity.These results are consistent with those of II-Seon et al. [19], suggesting that the scattering of light differs depending on the wavelength when using broad-spectrum (white) light, which causes the medium to act as a filter and the color of the medium or measuring cell to influence turbidity.Our findings indicate the limitations of the available turbidity test methods and possibly explain why the turbidity quality standard appears to be suitable even when household filters become discolored during tap water discoloration events.
However, the color of the turbidity standard solution without filtration was 0.4 CU at 0.1 NTU and displayed an increasing trend at an average of 1.4, 2.0, 5.0, 14.6, and 25.0 CU according to 0.3, 0.5, 1.0, 3.0, and 5.0 NTU, respectively.However, when the turbid matter was removed using a filter, the TCU of all samples was 0. This result suggests that there are limitations in the quality testing of tap water during typical discoloration events, which involve turbidity and color.

MFC Test of Field Samples
To establish an effective method to test discoloration of field samples collected from the DWDS, three samples collected from a reservoir, a tank, and an apartment were subjected to conventional turbidity and color tests and MFC.
We diluted each of the three samples at 1/20,000 dilution based on 0.5NTU.The average turbidity values of the diluted samples are shown in Table 7; they appeared to decrease in the following order: YR > DT > BC.This result can be attributed to both the turbidity and color of the samples contributing to the turbidity.Moreover, all color measurements, recorded as TCU after removing turbidity, were 0, indicating that the available color testing method is not feasible for colored samples.
As shown in Figures 7 and 8, the ∆E*ab values (D65) of YR, DT, and BC samples increased proportionally with their concentrations.As shown in Table 8, the R 2 values were 0.8971, 0.9788, and 0.9881 for YR, DT, and BC samples, respectively, indicating that the color difference can be used to quantitatively measure coloring substances.As shown in Figures 7 and 8, the ΔE*ab values (D65) of YR, DT, and BC sam increased proportionally with their concentrations.As shown in Table 8, the R 2 va were 0.8971, 0.9788, and 0.9881 for YR, DT, and BC samples, respectively, indicating the color difference can be used to quantitatively measure coloring substances.

MFC Test Results of the Coloring Reference Material
To verify the colorimetry results for Fe and Mn, we prepared reference materi ing Fe2O3•H2O (yellow iron oxide) and MnO2 (manganese dioxide); the concentrati Fe and Mn corresponded to those in the field samples.Table 9 presents the calc values of the Fe and Mn concentrations for the same field sample, as outlined in T These calculations were performed to determine the range of iron and manganese tion amounts for the subsequent experiment.Through this analysis, we confirmed th average concentrations of Fe and Mn ranged from 1.7 to 235.5 µg/kg and 0.06 to 113.8 respectively.Based on the Fe concentration, we then prepared the reference material and Mn ranging from 1 to 500 µg/L (Table 10).

MFC Test Results of the Coloring Reference Material
To verify the colorimetry results for Fe and Mn, we prepared reference materials using Fe 2 O 3 •H 2 O (yellow iron oxide) and MnO 2 (manganese dioxide); the concentrations of Fe and Mn corresponded to those in the field samples.Table 9 presents the calculated values of the Fe and Mn concentrations for the same field sample, as outlined in Table 7.These calculations were performed to determine the range of iron and manganese addition amounts for the subsequent experiment.Through this analysis, we confirmed that the average concentrations of Fe and Mn ranged from 1.7 to 235.5 µg/kg and 0.06 to 113.8 µg/kg, respectively.Based on the Fe concentration, we then prepared the reference materials of Fe and Mn ranging from 1 to 500 µg/L (Table 10).The experimental results of the reference materials, shown in Table 11, indicate a proportional increase in turbidity with the increase in sample concentration.However, the intensity of color increased almost proportionally depending on the sample before filtration with a 0.45 µm filter; however, after filtration according to the Drinking Water Quality Process Test Standard, all color values were 0.
The coloring reference material solutions were passed through filters, which were then dried at room temperature for 3 h.We conducted an MFC test of the discolored surface of the dried filter.As shown in Figures 9 and 10, the color difference ∆E*ab (D65) increased with the concentration of the Fe and Mn reference materials in the YR, DT, and BC samples.The regression analysis between the reference material concentration and color difference (Table 12) revealed a high concentration of the Fe and Mn reference materials in the YR, DT, and BC samples.The regression analysis between the reference materials and MnO 2 might be a useful method for measuring colorants within the range of the Fe and Mn content and ratio in field samples.
A comparison of the FCU values of the CIE reference material and the coloring reference material (Table 13 and Figure 11) showed that the FCU values of the field samples, based on Fe, were higher than those of the reference materials.This finding suggests that the field samples contain not only colorants such as Fe and Mn but also other substances causing coloration, which might have increased the FCU values.The FCU values of the field samples decreased in the following order: YR > DT > BC, with sediment showing a higher FCU than rust water in the water supply system.The high determination coefficients between the FCU of the field samples (YR, DT, and BC) and that of the coloring reference material suggest that measuring FCU using MFC could be a useful method in cases of tap water discoloration events.A comparison of the FCU values of the CIE reference material and the coloring reference material (Table 13 and Figure 11) showed that the FCU values of the field samples, based on Fe, were higher than those of the reference materials.This finding suggests that the field samples contain not only colorants such as Fe and Mn but also other substances causing coloration, which might have increased the FCU values.The FCU values of the field samples decreased in the following order: YR > DT > BC, with sediment showing a higher FCU than rust water in the water supply system.The high determination coefficients between the FCU of the field samples (YR, DT, and BC) and that of the coloring reference material suggest that measuring FCU using MFC could be a useful method in cases of tap water discoloration events.

Figure 1 .
Figure 1.Images of discolored water and filters before and after use (left: turbidity-causing de middle: discolored tap water; right: colored particles in tap water).

Figure 1 .
Figure 1.Images of discolored water and filters before and after use (left: turbidity-causing deposits; middle: discolored tap water; right: colored particles in tap water).

Figure 2 .
Figure 2. Current status of water supply pipe burial period in CJ city.

Figure 2 .
Figure 2. Current status of water supply pipe burial period in CJ city.

Figure 2 .
Figure 2. Current status of water supply pipe burial period in CJ city.

2. 4 . 3 .
Application of MFC MFC, developed by combining and modifying the FTM of NIER-GP2020-177 (2021) and MPC of ASTM D7843-21 (2023), was used to quantitatively measure colorants that occur during the supply of tap water.The MPC test results are expressed as filter color units (FCU).
Note(s): NIER: National Institute of Environmental Research; ASTM: American Society for Testing and Materials; FTM: filter test method; MPC: membrane patch colorimetry; MFC: membrane filter colorimetry (test method proposed in this study).Unit of MFC test results: ∆E*ab = filter color unit (FCU) = Color difference.2.4.3.Application of MFC MFC, developed by combining and modifying the FTM of NIER-GP2020-177 (2021) and MPC of ASTM D7843-21 (2023), was used to quantitatively measure colorants that occur during the supply of tap water.The MPC test results are expressed as filter color units (FCU).

Figure 5 .
Figure 5. Percentage weight of main elements determined using SEM-EDS analysis.

Figure 5 .
Figure 5. Percentage weight of main elements determined using SEM-EDS analysis.

Figure 6 .
Figure 6.Changes in the iron and manganese concentrations of samples.

Figure 6 .
Figure 6.Changes in the iron and manganese concentrations of samples.Table 6. Cross-test of turbidity and color.Color Turbidity Turbidity Color

Figure 7 .
Figure 7. Filter photographs of the field samples.

Figure 7 .
Figure 7. Filter photographs of the field samples. 00

Figure 9 .
Figure 9. Filter photographs of the reference material.

Figure 9 .
Figure 9. Filter photographs of the reference material.

Figure 10 .
Figure 10.FCU of the reference material.

Figure 10 .
Figure 10.FCU of the reference material.
Sample 3, BC).The pipe material, pipe diameter, and installation year at each sample collection point are presented in Table1.

Table 1 .
Characteristics of inflow pipes by sampling points.

Table 1 .
Characteristics of inflow pipes by sampling points.

Table 2 .
Specifications of the reference materials.

Table 2 .
Specifications of the reference materials.

Table 2 .
Specifications of the reference materials.

Table 3 .
Development of MFC using FTM and MPC.

Table 3 .
Development of MFC using FTM and MPC.
Note(s): NIER: National Institute of Environmental Research; ASTM: American Society for Testing and Materials; FTM: filter test method; MPC: membrane patch colorimetry; MFC: membrane filter colorimetry (test method proposed in this study).Unit of MFC test results: ΔE*ab = filter color unit (FCU) = Color difference.

Table 4 .
Overview of the membrane filter colorimetry test method.

Table 5 .
Concentration of iron and manganese in field samples.

Table 7 .
Average of turbidity and color values of the field samples.

Table 9 .
Content of iron and manganese in diluted field samples.

Table 9 .
Content of iron and manganese in diluted field samples.

Table 10 .
Concentration of iron and manganese in samples used in the MFC test.

Table 12 .
Mean ∆E*ab values of the reference material.

Table 13 .
Mean FCU values of field samples and reference material.