Effects of boiling on iron particles in drinking water

In order to inactivate harmful microorganisms and eradicate waterborne diseases, boiling has been applied in drinking water treatment households globally. Boiling is a common recommendation of health promoters in developing countries to improve the quality of drinking water and to decrease the incidence of waterborne diseases (McLennan 2000). It is low in cost and effective in inactivating harmful microorganisms (Zhang 2013). Moreover, it can eliminate the pungent taste and odor of tap water caused by chlorine residuals.

Various pipe materials, such as iron, galvanized steel, concrete, polyvinyl chloride, and copper, have been extensively used in drinking water distribution systems (DWDSs). Unlined iron pipes are very prone to electrochemical corrosion and their internal surfaces can become severely corroded over time. Typical iron corrosion products formed on aged pipe surfaces can be described using three different layers: (a) a macroporous layer of black rust (magnetite, Fe₃O₄) in contact with the metal, (b) a microporous film of a mixture of Fe²⁺ and Fe³⁺ species that covers the macrolayer, and (c) a top layer of red rust (mainly goethite, α-FeOOH and hematite, α-Fe₂O₃). The exact composition and structure of iron corrosion scales vary significantly under different water qualities as well as flow properties (Butler & Ison 1968; Tuovinen et al. 1980).

Treated water from a treatment plant is distributed to households via a public DWDS. Iron oxidation inevitably occurs in DWDSs and can cause water quality problems such as increased turbidity and discoloration of tap water (Zhuang et al. 2021a). Furthermore, organic matter in loose deposits (LDs) will worsen the water quality (Qi et al. 2022). In the old cast iron water pipeline, the inner wall of the pipeline may be corroded and generate iron compounds, and other substances carrying toxic or harmful matter, and then the particles in the pipeline will enter the tap water through the impact of water flow. Particles in discolored water could be exposed to human beings through drinking water or other ways of using water. Moreover, the iron oxides in LDs would also cause toxicity effects which could be further enhanced by organics in water (Qin et al. 2021). The particles with smaller sizes may favor their entry into the cells and eventually lead to stronger cytotoxic and genotoxic effects than larger particles (Elihn et al. 2012; Liu et al. 2014; Freitas et al. 2020; Zhuang et al. 2021b).

When the concentration of LDs is too high, the turbidity and chromaticity of water will increase accordingly and be easily detected by people. However, in most cases, LDs are not easily detected in water, and it is difficult to remove them with household water treatment methods such as boiling. A boiling fluid may have violent flows and bubble disturbances, and the resuspension of deposited particles to the boiling fluid is a highly probable event that can occur during a boiling process (Lin et al. 2019). Humans are unavoidably exposed to iron particles via water ingestion (Boxall & Saul 2005). Some studies have shown that adding lemon to boiling water can reduce the production of disinfection by-products (Liu et al. 2021). Under the complex environment of DWDSs, LDs may inevitably undergo structural transformation due to the wide variety of coexisting substances, consequently resulting in a change in properties such as toxicity (Zhuang et al. 2019), but the changes due to the boiling effect have not been reported. Iron-dominated particles may undergo structural changes after boiling, especially LDs which may contain a variety of organic and inorganic components including (micro)pollutants.

Moreover, in previous reports, boiling of tap water has been widely found to reduce trihalomethanes (THMs) (Batterman et al. 2000; Krasner & Wright 2005), but the effect of boiling of haloacetic acids (HAAs) was not fully clear. The increase of dichloroacetic acid (DCAA) and the decrease of trichloroacetic acid (TCAA) have been found after boiling (Levesque et al. 2006). Moreover, boiling has been found to induce a 59% increase in DCAA without the reduction of TCAA, resulting in an overall increase in HAAs (Krasner & Wright 2005). The reason for this may be that THMs are volatile compounds while HAAs are not volatile compounds. Thus, with the existence of particles, the effect of boiling on DBPs and microorganisms may change causing new risks.

Here, the effects of boiling on iron particles, including LDs from the DWDS and their main components FeOOH, Fe₂O₃, and Fe₃O₄ were investigated. The pH and turbidity of water before and after boiling were compared. Particle structure, including particle size, particle number, and crystal pattern were analyzed to reflect the effect of boiling on particles. The effect of boiling on DBPs and microorganisms under the existence of different particles was also studied to further evaluate the risks induced by particles during boiling.

All chemicals, including Fe₂O₃, Fe₃O_4, and FeOOH, were purchased from Shanghai Macklin Biochemical Co., Ltd in analytical purity and used in the experiments directly without any further purification. The boiler is one of the most common kettle models found in Chinese homes, which has a 304 stainless steel kettle with a heat source at the bottom and a temperature sensor that automatically shuts down after the water boils.

LD and tap water samples were collected in DWDSs of old unlined cast iron pipes with a pipe age of 20 years in a city in northern China, and a water main pipe of at least 150 m was selected. At the end of the selected main pipe, a hydrant was present for the sampling of LDs.

All particles are added into tap water and mixed into 1-L solutions of different concentration gradients. Particle concentrations were set as 500, 200, 100, 10, 1, and 0.1 mg/L. The solutions were stirred for 10 min to form uniform dispersion. The boiled samples were named LD-B, Fe₂O₃-B, Fe₃O₄-B, and FeOOH-B. All the boiled samples were boiled to 100 °C and then cooled to room temperature. The unboiled samples were named LD-N, Fe₂O₃-N, Fe₃O₄-N, and FeOOH-N. The unboiled and boiled tap water without particles were named Ctrl-N and Ctrl-B, respectively. Water samples were treated with different concentrations of different particles. Water quality analysis and characterization of cast iron coupons are then carried out. For the characterization of solid particles before and after boiling, the boiled samples were centrifuged to obtain the solid particles.

The turbidity and pH values of the water samples were measured using a turbidimeter (Hach, 2100Q, USA) and pH meter (Mettler-Toledo, FE20K, Switzerland), respectively. Particle size was measured by a Malvern laser particle size analyzer (Malvern 3000, England). The crystallography of the samples was analyzed using X-ray diffraction (XRD) by a Bragg-Brentano diffractometer (Rigaku, D/Max-2200, Japan) with a Cu target (40 kV, 40 mA). The number of particles was measured by the benchtop laser particle analyzer (GREAN, GR-1500A, China). The dissolved organic matter was measured by an excitation-emission matrix (EEM) fluorescence spectrophotometer (Hitachi, F-7000, Japan). The final values were taken as the average of the three replicates.

DBPs: For the determination of THMs and HAAs, water samples were extracted with methyl tert butyl ether (MTBE) and analyzed by a gas chromatograph equipped with an electronic capture detector (GC/ECD, Agilent 7890B, USA) and an HP-5-fused silica capillary column (30 m × 0.25 mm, 0.25 μm, Agilent, USA). The water samples passed through a 0.45-μm membrane were prepared for testing natural organic matter (NOM) and DBPs. All water samples were prepared in at least three duplicates.

EPS (extracellular polymer): 4 L of water was filtered through a 0.2-μm polycarbonate membrane filter (AISIMO, China). Polycarbonate membranes with biofilms were put into 50-mL centrifuge tubes containing 40 mL of sterile phosphate buffered solution (PBS, pH = 7). The tubes were sonicated at 20 KHz for 30 s, followed by heating in a water bath at 70 °C for 1 h, and then centrifuged at 8,000 rpm for 20 min at 4 °C. The supernatant in the tubes was filtered through a 0.45-μm polycarbonate filter to collect EPS. The proteins in EPS were determined with the Lowry procedure using bovine serum albumin (BSA) as the standard. The polysaccharide content was determined by the phenol-sulfuric acid method with glucose as the standard.

ATP (adenosine triphosphate): The microbial activity was determined by ATP using Promega Biosystems (Sunnyvale, CA 94085, USA).

In addition, there is a good linear correlation between particle concentration and turbidity (Supplementary material, Figure S2, R² > 0.95). As mentioned above, the highest turbidity producing capacity is observed in Fe₂O₃ and FeOOH, followed by Fe₃O₄ and LDs, because at the same concentration, the number of particles in Fe₂O₃ and FeOOH is higher. At 2 mg/L, the number of the particles in LDs before boiling is about 4,000/mL. After boiling, it increases to about 5,000/mL, while Fe₂O₃ are 15,000 and 13,000/mL before and after boiling. Thus, at the same concentration, the samples containing particles with smaller size had larger particle number, resulting in higher the turbidity. Therefore, it can be explained that the turbidity of the LDs group will increase after boiling, while the turbidity of iron oxide will decrease after boiling. To further confirm the particle structure change before and after boiling, XRD and Fourier transform infrared spectroscopy (FTIR) of the samples before and after boiling are shown in Supplementary material, Figure S3a and S3b. In XRD, the peak position did not change after boiling for all four particles, indicating boiling did not change the crystal pattern of the particle. However, the crystals were shown to be weaker after boiling, which indicates a lower degree of crystallinity in the iron oxides after boiling. Under sustained oxidizing conditions, Ostwald ripening is expected to spawn a time-dependent increase in the degree of crystal ordering of minerals that coincides with their diminished aqueous solubility (Steefel & Van Cappellen 1990). The effect of boiling would eliminate the DO in water and reduce the oxidizing conditions (Steefel & Van Cappellen 1990). LDs are mostly in the form of complex primary minerals (Pan et al. 2022). We found that maghemite (γ-Fe₂O₃), goethite (α-FeOOH), lepidocrocite (γ-FeOOH), and hematite (α-Fe₂O₃) were the predominant crystalline compounds of LDs without boiling. The maghemite (γ-Fe₂O₃), goethite (α-FeOOH), lepidocrocite (γ-FeOOH), and calcite (CaCO₃) were the predominant crystalline compounds of LDs after boiling. The increase in CaCO₃ may be caused by the crystallization of Ca²⁺ from water during boiling. Two intensive and broad bands at 3,400 and 3,100 cm⁻¹ are visible, which can be assigned to the envelope of the hydrogen bonded surface –OH groups and the presence of –OH stretching. Two bands around 890 and 790 cm⁻¹ can be assigned to Fe–O–H bending vibrations. The IR band around 630 cm⁻¹ is influenced by the shape of the α-FeOOH particles, Fe₃O₄ had two characteristic IR bands at 565 and 360 cm⁻¹, and α-Fe₂O₃ showed IR bands at 575, 485, 385, and 360 cm⁻¹ (Krehula & Musić 2006). The functional groups of the iron crystals became weaker after boiling, which was in accordance with crystallization decrease in the XRD results.

The HAAs in water containing 0.1 mg/L LDs were shown to be a little higher than the control. The HAAs in water containing LDs were shown to decrease with the increase of particle concentration when LDs were higher than 1 mg/L. The existence of Fe₂O₃ was shown to decrease the HAAs, while the influence of FeOOH on HAAs was not small, in relative terms. Under high concentrations of Fe₃O₄, the HAAs in water were higher than in the control group, indicating that the Fe₃O₄ in LDs would increase the risk of HAAs during boiling. The changes of HAAs were shown to be dominated by the changes of DCAAs. Boiling has been found to accumulate a few polar DBPs in the tap water, and the increase in dihaloacetic acids was because they are the common final decomposition products of certain intermediate DBPs (Pan et al. 2014).

Figure 7 shows the impact of particle concentration on ATP. It is interesting to note that ATP tended to increase first and then decrease with the increase of particle concentration, while the ATP was the highest under particle concentration 0.1 mg/L for all the particles. It is also interesting to note that in a realistic situation, there is a great possibility for LDs to exist in drinking water at 0.1 mg/L level, thus the microbiological risks under low particle concentrations should not be neglected. For all the four kinds of particles, the water containing particles had higher ATP than that without particles, indicating more active biomass was present in the existence of particles. The increase in ATP under low particle concentration may be because the surface of particles could provide more attachment sites for microorganisms to adhere to, which would protect the microorganisms. However, under high particle concentration, the heteroaggregates reached saturation at relatively high particle concentration, and the binding among bacteria increased (Zhao et al. 2022); more importantly, iron particles are able to possess toxicity including direct generation of reactive oxygen species (ROS) from the surface of the particles, production of ROS via leaching of iron molecules by enzymatic degradation, altering mitochondrial and other organelle functions, and induction of cell signaling pathways together with their consequent activation of inflammatory cells, which results in the generation of ROS and reactive nitrogen species (Mahmoudi et al. 2012). Thus, with the increase of particles, the toxicity induced by particles would also increase, resulting in the decrease of ATP under high particle concentration. The ATP in the water containing FeOOH was the lowest, indicating higher toxicity risks induced by FeOOH. Therefore, the existence of particles increased the microbiological risks during boiling, especially under low concentrations, but the total microbiological risks were still very low after boiling. The EPS are usually expected to have heavy metal binding capability and potential for applications in biodetoxification due to their complex mixture of multiple biomolecules (proteins, polysaccharides, humic-like substances, uronic acid, nucleic acid, lipids, and glycoproteins), among which proteins and polysaccharides in the EPS are primarily responsible for the resistance to heavy metals (Kang et al. 2017). Once EPS are present, the reactive groups within EPS initiate their functions for biosorption. It was reported that dextran, mannan, protein, lipid, and chitin of EPS play critical roles in the adsorption of heavy metals and possess both adsorptive and adhesive capability (Guibaud et al. 2012). The polysaccharide and protein increased with the increase of residual chlorine in the iron pipe, which would result in a stronger adsorptive and adhesive effect of iron particles on EPS. The ratio of polysaccharide to protein in EPS determines the surface charge of cells, and the cell surface charge was positively correlated with the ratio of polysaccharide to protein in EPS (More et al. 2014). The polysaccharide in the control sample decreased slightly after boiling, however, the existence of particles obviously decreased the polysaccharide content after boiling (Supplementary material, Figure S9a). The water containing LDs had higher protein content after boiling (Supplementary material, Figure S9b). Polysaccharides could promote the adhesion strength of EPS, while the adhesion ability of protein is relatively small (Harimawan & Ting 2016). Moreover, the LDs would obviously reduce the polysaccharide/protein ratio (Supplementary material, Figure S9c), resulting in lower adhesion strength between microorganisms and LDs. Among the three kinds of iron oxides, their effects on protein are similar, but Fe₃O₄ did not lead to an obvious reduction of polysaccharide, while the water containing Fe₃O₄ had the highest polysaccharide/protein ratio.

Some of the water quality indicators mentioned above (including particle concentration, ATP, pH, HAAs concentration, and THMs concentration) were measured for each water sample and compared using principal component analysis (PCA) as shown in Supplementary material, Tables S1–S4. In the sample containing LDs after boiling, ATP had a strong positive correlation with HAAs and a high positive correlation with THMs, which indicated that in the water containing LDs, the coordination of microorganisms and DBPs may further enhance water deterioration. Thus, the water containing Fe₃O₄, water deterioration risks would also exist due to the coordination of microorganism and THMs. Previous reports had found a decreasing trend of dissolved organic carbon (DOC), THMs formation potential, and HAAs formation potential at high temperature incubations irrespective of microbial concentrates, which indicated that a higher biomass in water may contribute to higher consumption of DOC and consequently lower DBP formation potentials, especially at high incubation temperatures (Zeng et al. 2021). Thus, when under the particle concentration of 0.1 mg/L, the active biomass would further promote the formation of DBPs, resulting in higher water deterioration risks under low particle concentration. Furthermore, as the particle concentration under low concentration is hard to find, this risk is also hard to find.

Iron particles widely exist in drinking water which is the key factor of discoloration, while boiling has been applied in drinking water treatment households globally, but the effects of boiling on iron particles have not yet been fully recognized. Here, the turbidity of water containing LDs increased after boiling, and the samples containing three other kinds of iron oxides decreased. At 10 mg/L, the turbidity values before/after boiling of Fe₂O₃, Fe₃O_4, and FeOOH are 134.00/121.00, 25.07/21.22, and 120.40/114.20 NTU, respectively. After boiling, the number of particles in LDs increased and the particle size decreased, while iron oxides decreased in particle number and the particle size increased, which indicated that the LDs tended to be broken but the iron oxides seemed to aggregate through boiling. The difference in particle behavior may be due to the LDs containing organics. The crystals were shown to be weaker after boiling, which indicated a lower degree of crystallinity in the iron oxides after boiling. The Fe₂O₃ increased THMs while the Fe₃O₄ increased HAAs after boiling. The ATP was the highest under particle concentration 0.1 mg/L for all the particles, indicating that the existence of particles increased the microbiological risks during boiling, especially under low concentrations. The ATP in the water containing FeOOH was the lowest, but the total microbiological risks were not obvious after boiling. Thus, the iron oxides in LDs may lead to water deterioration during boiling.

This research was supported by the National Natural Science Foundation of China (No. 51978652).

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

The authors declare there is no conflict.