Solar Disinfection Using Zero Valent Iron for Inactivation of Escherichia coli and Total Coliforms in Water Using a Raceway Reactor

Abstract

Contamination from microorganisms is one of the gravest types of water pollution. In 2022, there were 842,000 new cases of gastrointestinal diseases worldwide. The aim of this study was to size, construct, and evaluate a Raceway reactor (28 L total capacity) as a laboratory-scale solar disinfection system for the inactivation of Escherichia coli and total coliforms in water, using Zero Valent Iron (ZVI). For this purpose, a sample of E. coli contaminated potable water was treated with steel wool as a source of ZVI and solar irradiation. Using a 2³ factorial design with four central points and a total of 12 trials, the following was investigated: the effect of the Fe⁰ dose (0.6–1.8 g L⁻¹); it should be noted that both the natural and drinking waters of the study area have iron concentrations of less than 0.1 mg L⁻¹. Depth tests of the treated water in the reactor were carried out at different levels (5, 7 and 9 cm) and with a duration of four to six hours. Therefore, it is concluded that the reactor/ZVI is effective for the disinfection of E. coli and total coliforms at concentrations >2419.6 MPN/100 mL, reaching 99.96% disinfection for both cases; it is also a cost-effective treatment due to its inexpensive inputs.

1. Introduction

The constant expansion of the economy as a result of industrialization, water pollution, and the safety of the aquatic environment have become major concerns for humans. Only 8% of wastewater from industries and homes that is discharged into the environment in large volumes is treated, in particular in nations with limited resources [1]. This is significantly less than the 70% in high-income nations. Consequently, eighty percent of the world’s regions discharge untreated wastewater contaminated with fecal microorganisms, nitrates, phosphates, and solvents into lakes and rivers that eventually travel into the sea, generating negative environmental and public health impacts [2]. In this sense, the contamination of water by pathogenic microorganisms is one of the world’s gravest problems and is essential for human health [3]. According to the WHO/UNICEF Joint Monitoring Programme for Water Supply and Sanitation, at least two billion people worldwide consume water that may be contaminated with excrement, such as E. coli, total and fecal coliforms. Similarly, bacteria of fecal origin, arsenic, and fluoride are significant health indicators. As is the case with E. coli in water, the fact that it is an indicator of fecal contamination that must be detected by specialized equipment and is inaccessible to the general public makes its periodic detection challenging for maintaining adequate sanitary control, particularly if the bacteria can be reactivated in a short period of time. According to the World Health Organization, water fit for human consumption should contain 0 CFU/100 mL (compliant), 1–10 CFU/100 mL (low risk), 10–100 CFU/100 mL (intermediate risk), 100–1000 CFU/100 mL (high risk), and >1000 CFU/100 mL (very high risk) of E. coli and total coliforms at a concentration greater than or equal to 5 NMP/100 mL for >10 μg L⁻¹ arsenic and >1.5 mg L⁻¹ fluoride [4,5,6].

In this context, wastewater utilization has emerged in recent years as an option to alleviate pressure on freshwater resources, particularly in sectors such as agriculture where water scarcity is severe [7]. To this end, there are a number of processes that enable the reuse of these waters throughout the world, including advanced oxidation processes (AOP’s) (fenton, photo-fenton), zero valent iron (ZVI), and solar water disinfection (SODIS), among others, which eliminate organic and inorganic compounds, emerging contaminants, and pathogenic microorganisms. Consequently, it is essential to differentiate between the following processes:

• − SODIS: The process (based on PET container reactors) is an established method for disinfecting domestic water. In 1984, this method was first used to economically disinfect water used to treat diarrhea and dehydration. In communities without access to potable or safe water, it is an inexpensive process for disinfecting natural water using solar radiation. Nevertheless, according to some authors, there are disadvantages associated with the volume of treated water and the possible reappearance of bacteria during consecutive storage periods in darkness [8,9].

According to Nelson et al., 2018, the most significant photochemical mechanism by which fecal microorganisms are inactivated during the SODIS process is “endogenous indirect inactivation”, which involves a (photosensitization) product of a few sensitizers that absorb light in the solar wavelength range and is a highly efficient inactivation mechanism [10].

• − AOP’s: Are presently regarded as one of the most prevalent and promising wastewater treatment technologies. Among the vast array of available technologies, the most prevalent are the combination of UV and peroxide (UV/H₂O₂), the Fenton reagent (Fe²⁺/H₂O₂), and two of its variants, photo-Fenton and zero-valent iron (Fe⁰/H₂O₂). This is a combination of photo-Fenton and zero valent iron (Fe⁰) or photocatalysis (UV/TiO₂).

The Fenton process generates hydroxyl radicals (·OH) at room temperature and normal pressure after the oxidation of ferrous iron by H₂O₂ in sufficient concentrations to oxidize organic compounds or eliminate inorganic compounds in water [11,12]. The photo-Fenton process is similar to the preceding one but uses sun or artificial radiation (UV) to cyclically oxidize and reduce iron in aqueous solution, consuming H₂O₂. The preponderance of FeOH₂⁺ in the solution (48%, pH 2.8, 25 °C, ionic strength 0.5 M) maximizes H₂O₂ catalysis efficiency in this homogeneous process [13,14].

Fenton and photo-Fenton reactions generate hydroxyl radicals that degradation pathogen macromolecules [15]. Alvear-Daza et al., 2021, found that adding 10 mg L⁻¹ H₂ O₂ to natural waters (one well) with natural total iron concentrations of roughly 0.3 mg L⁻¹ and under simulated sun radiation completely inactivated E. coli. SEM micrographs showed that Viable but Non-Culturable (VBNC) E. coli cells changed morphologically and shrank (2002) [8]:

• − ZVI: Metallic iron (Fe0), also known as zero valent iron (ZVI), is introduced as an economical alternative to iron ions used in Fenton processes. Several studies discuss how Fe⁰ can be converted to Fe²⁺. This is referred to as a pseudo-catalytic iron (Fe⁰)/Fe²⁺ system. The produced Fe³⁺ can also be recycled [16,17]. Due to its non-toxic nature, abundance, environmental tolerance, and high surface area and high reactivity, this element has been studied extensively over the past 25 years [18]. ZVI readily corrodes in water. Iron supports spontaneous oxidative dissolution when submerged in water (H⁺ or H₂O) because the redox potential of water is greater than that of iron (Equation (1)). This reaction is predominantly electrochemical [19,20,21].

Fe⁰ + 2H⁺ → Fe²⁺ + H₂

(1)

When ZVI is dissolved (or in soluble ion compounds or insoluble oxides or hydroxides form), this results in an anodic reaction linked to the cathodic reduction of reducible species. The principal component available for corrosion reactions in natural environments is dissolved oxygen (DO), which is thermodynamically favored (Equations (2) and (3)) [22]. ZVI oxidizes chemically by itself in the presence of dissolved oxygen. This is a complex process involving the transformation of several metastable ferrous-ferrous intermediate species into stable iron oxides. Importantly, the presence of dissolved ions accelerates ZVI reduction, thereby favoring iron oxidation [23,24].

2Fe⁰_(s) + 4H⁺_(aq) + O_2(aq) → 2Fe²⁺ + 2H₂O_(l) E⁰ = +1.67 V

(2)

2Fe⁰_(s) + 2H₂O_(l) → 2Fe²⁺ + H_2(g) + 2OH⁻_(aq)  E⁰ = −0.39 V

(3)

The primary byproduct of ZVI oxidation is Fe²⁺. It can remain at the interface, where it can form ferrous precipitates, such as carbonate or hydroxide, which can then be oxidized further in an aerobic environment with a high ionic strength. It is also capable of migrating away from the surface, where it is exposed to homogeneous oxidation and precipitates as colloidal species and Fe³⁺ [23,25,26,27,28]. Iron species have a superior capacity for O₂ and peroxide activation [29,30,31] due to their ability to transfer electrons, which exhibits exceptional properties in a variety of oxidation-based reactions. Normal oxygen molecules serve as electron acceptors and produce superoxide radicals (O₂⁻). When an electron is added to this superoxide radical, the peroxide ion (O₂²⁻) without unpaired electrons is formed. At a pH close to neutral, the peroxide ion undergoes protonation, which results in the formation of hydrogen peroxide (H₂O₂) [31,32]. Moreover, H₂O₂ and Fe²⁺ generate hydroxyl radicals (∙OH) (Fenton reaction), which have a potent oxidizing effect on organic and inorganic compounds (Equation (4)) [18].

Fe²⁺ + H₂O₂ → Fe³⁺ + ∙OH + OH⁻

(4)

Therefore, ZVI is an efficient inducer of reactive oxygen species (ROS), which are powerful oxidants that can initiate advanced oxidation processes (APOs) that degrade pollutants to basic, non-toxic molecules. The nonselective nature, high reactivity, and potent oxidizing capabilities of these species make them potentially useful for degrading a wide variety of organic and inorganic pollutants [33]. ZVI research was initially conducted for groundwater remediation, but ZVI/O₂ and Fe(II)/O₂ systems can also be used for effluent treatment [34].

Several authors, on the other hand, suggest that reactive oxidation species (ROS), such as superoxide (O₂), hydroxyl radicals (OH), and hydrogen peroxide (H₂O₂), are responsible for antibacterial activity, cause oxidative stress, and damage bacterial proteins and DNA [15,35,36,37,38].

Similarly, there are inactivation processes by sunlight photoinactivation and photooxidation; depending on the wavelength, bacterial inactivation (E. coli) can be promoted through photoinactivation and endogenous photooxidation. Photoinactivation is possible at 280–400 nm where photons act directly on DNA and RNA and affect their absorption and reaction with bases. In the case of endogenous photooxidation it can occur at 320–400 nm (UV-A) and 400–700 nm (Visible), where an oxidative attack by ROS can be provoked. Through the latter, nucleic acids, enzymes, and lipids of E. coli microorganisms are oxidized, causing loss of biological function and consequent cell death [39].

Among the processes that can be used for inactivation are SODIS, photo-Fenton, and zero valent iron, where solar radiation has a crucial part in aiding the inactivation of E. coli and other bacteria (Figure 1). Observations indicate that these mechanisms depend on variables such as the type of chemical agents to be employed, such as Fe⁰, H₂O₂, Fe²⁺, H₂O, and O₂, where oxide reduction reactions are produced based on the type of process.

Likewise, the particle size is fundamental for the inactivation of E. coli, considering that the bacterium has a size of 1 µm, Fe⁰ 0.1 µm (ZVI), and despite the fact that nZVI < 100 nm, the particularity that the Fe⁰ molecule can react, forming iron oxides and iron hydroxides that are <1 µm, is highlighted [40,41,42].

Other methods and technologies used to disinfect wastewater include the following: chlorination, ozone, and chloramines, which are the standard methods for disinfecting potable water that are currently available. Toxic by-products of disinfection, such as nitrosodimethylamine, organochlorines, and bromate, can be extremely carcinogenic to humans [3,43].

On the other hand, the types of reactors used for wastewater disinfection are: wastewater treatment plants (WWTPs), compound parabolic collector photoreactors (CPCs), and channel reactors (RRs) that remove organic pollutants, including pesticides, pharmaceuticals, hormones, cosmetics, cleaning product components, and E. coli, from wastewater. Low-concentration organic molecules (μgL⁻¹-ngL⁻¹) cannot be removed by WWTPs [44]. By applying solar energy directly to a water sample, CPCs for AOPs in photocatalysis systems can remove organic compounds and disinfect at concentrations between μgL⁻¹–ngL⁻¹. These reactors are used in photo-Fenton research. TiO₂ is a photocatalyst, but its large band gap (3.2 eV) and simple recombination of photogenerated electron-hole pairs limit its stability and efficiency [3].

Since photochemical reactions are dependent on the position of the reaction zone, RRs are an excellent alternative to conventional reactors. In homogeneous media, photons absorb variably; consequently, the radiation field is not uniform. In a heterogeneous medium, photons assimilate and disperse, rendering the test field more homogeneous [45]. Low-cost, open reactors have a liquid depth that varies in response to solar radiation. Cabrera (2021) identifies liquid depth, continuous flow operations, treatment capacity, and kinetic modeling as performance-related RR variables [11].

The iron used in RR’s bacteria-killing tests is ferrous sulfate (FeSO4). This indicates the optimal concentrations are 0.36 mM Fe₂SO₄ and 1.47 mM H₂O₂ [13]. In Fe₂SO₄ nanoparticles, reactive oxygen species (ROS) destroy gram-positive and gram-negative microorganisms. With this molecule, H₂O₂, membrane proteins or nanoparticles, and the outer bilayer of bacteria can destroy bacteria. Bacterial cell membranes are disrupted by hydrogen peroxide. Nanoparticles interact with microbes that have been killed, generating and releasing H₂O₂ [15].

In account of the foregoing, it is anticipated that the current initiative will determine the efficacy of a laboratory-scale channel reactor as a solar wastewater disinfection system using zero valent iron for the inactivation of Escherichia coli and total coliforms in water based on its construction and sizing.

2. Methodology Conditions

2.1. Materials and Methods

2.1.1. Construction and Sizing of a Raceway Reactor

In the facilities of the “Plataforma de Investigación y Formación de Tratamiento Solar de Aguas” of the Universidad de Tarapacá (Arica-Chile), the use of fiberglass raceway reactor with a maximum working capacity of 28 L (120 cm long, 48 cm wide, 12 cm channel width, and 15 cm depth) was constructed (Figure 2 and Figure 3). Hadiyanto and colleagues [46,47,48] performed the scaling based on a literature review, and the intermediate wall is wider than the conventional design. They discovered that this modification improves hydraulic properties, reduces energy consumption by 40%, and minimizes low velocity zones in the channel, resulting in a flow that is uniformly distributed.

The drive system of the raceway reactor was designed with six flat blades with curved ends made of polylactic acid filament (PLA) in a Creality CR10 Smart 3D printer Longhua Dist., Shenzhen, China, arranged on a steel shaft with bearing supports, and a 12 V motor. The channel’s flow velocity was set to 0.3 m s⁻¹ to ensure turbulent flow, avoid stratification, and keep particle suspension.

2.1.2. Laboratory Validation of the Method

The preparation of reagents, synthetic samples, and measurements were carried out at the Laboratory for Environmental Research in Arid Zones, LIMZA, of the Universidad de Tarapacá (Arica, Chile).

Steel Wool (Fe⁰)

According to research conducted by Cornejo et al., 2008, the Fe⁰ used as a source of ZVI was commercial steel wool composed of 99% iron, 1% manganese, and not significant levels of lead, chromium, copper, and zinc [49].

Inoculated Synthetic Drinking Water Matrix (SDW)

The source of the water matrix was prepared in the laboratory from dechlorinated potable water and inoculated with a volume of fecal solution (1 g of feces per 100 mL, 30-min incubation period). Each experiment’s inoculation was conducted according to the following Equation (5):

F = S × A_c × h × 1000

(5)

where F: volume of fecal solution to be inoculated, mL; S: factor obtained according to preliminary tests expressed, mL·L⁻¹; A_c: the area of the Raceway reactor channel, m²; and h: the level of SDW contained in the raceway, cm.

As indicators of bacterial inactivation, Escherichia coli and total coliforms were selected. Using the Colilert-18/Quanti-Tray method (ISO 9308-2:2012) [50] from IDEXX Laboratories (Westbrook, ME, USA), these bacteria were quantified. Before and after each experiment, 100 mL samples were collected in sterile containers for this purpose. The Colilert-18 reagent was added to each sample, which was then agitated until the reagent was dissolved. The solution was then transferred to a Quanti-Tray 2000 and sealed with a Quanti-Tray Sealer. The samples were incubated for 18 h at 35 °C ± 0.5 °C. Following this period, positive cells (yellow and fluorescent cells) were enumerated, and the data collected were compared to the IDEXX Quanti-Tray^®/2000 MPN (Westbrook, ME, USA) table to determine the number of microorganisms per 100 mL.

Determination of Physicochemical Parameters

The physicochemical parameters (pH, electrical conductivity, turbidity, hardness, alkalinity, and free chlorine) were determined in the inoculated water matrix according at national and international standards, such as APHA [51] and standardized water analysis techniques [52]. For the determination of iron in solution, the 1,10-phenanthroline spectrophotometric method was utilized [53].

2.1.3. Cumulative Energy Absorbed at Raceway Surface (QUV)

To evaluate solar photocatalytic processes considering a decontamination and disinfection treatment of water under solar radiation [54], it was necessary to determine the total energy absorbed on the surface of a raceway $Q_{UV}$ reactor. For this, it was calculated using Equation (6):

$$Q_{UV,n+1}=Q_{UV}+\left(t_{n+1}-t_n\right)\times UV\times \frac{A}{V}$$

(6)

where n: sample number; $Q_{UV}$: solar $UV$ radiation, Wm⁻²; t: the sample treatment times in the reactor, considering t_n and t(_n+1) expressed, hour; A: irradiated surface of the photoreactor, m²; and V: total volume of wastewater treated in the reactor, L [55].

Solar irradiance (Wm⁻²) was measured every 30 min through a Photo Radiometer HD 2102.2 Delta OHM, Padova, Italy, equipped with a radiometric sensor LP 471 RAD (spectral range 400 to 1050 nm).

2.1.4. Solar Disinfection by Raceway Reactor

This stage used the raceway for solar water disinfection. At the LIMZA facilities, the oxide reduction reactions were carried out in the reactor under natural conditions of solar irradiation in the open air. Adding Fe(0) at varying concentrations (0.6, 1.2, and 1.8 g L⁻¹, respectively) was performed at varying volumes of an SDW matrix with regard to three depths (5, 7, and 9 cm) with volumes que contienen faecal microorganisms (6.2, 8.7, and 11.1 mL, respectively). The durations of exposure were 4, 5, and 6 h. At the outset and end of the treatment period, samples were collected. To measure the amount of total coliform and E. coli microorganisms, samples collected at the outset of each experiment were stored at 4 °C until the end of the experiment. The start times and experimental days were selected based radiation data from prior years and climate conditions (Figure 4).

2.1.5. Experimental Design and Statistical Analysis

The experiments were developed using a 2³ factorial design with four central points, comprising a total of 12 trials, in which the effects of zero iron dose, liquid level in the raceway reactor, and treatment time were investigated. Using Stat Graphics Centurion 18 software, an ANOVA was performed on the data.

3. Results

3.1. Initial Drinking Water Characterization

For each experiment, a physicochemical and microbiological analysis of the drinking water was done before the inoculation; the mean values are shown in

Table 1. Characterization of dechlorinated drinking water. Source: own growth.

Parameter	Unit
pH	-	6.97
Electrical Conductivity	µScm−1	2067
Turbidity	NTU	0.17
Total Hardness	CaCO3 mgL−1	755
Alkalinity	CaCO3 mgL−1	138
Free chlorine	mgL−1	0.00
Total Coliforms	MPN 100 mL−1	<1
E. coli	MPN 100 mL−1	<1

.

The water used in the experiments is classified as extremely hard, with a pH greater than 6 and close to neutral and a conductivity of 2067 µScm⁻¹. In addition, neither total coliform microorganisms nor E. coli are present.

3.2. Inoculation

For the determination of S, preliminary experiments were conducted. From these tests, it was determined that the value of S is 0.4 mL·L⁻¹, which means that for each liter of water, 0.4 mL of fecal solution must be added to reach a concentration of total coliforms and E. coli that can be measured within the counting range of the Colilert-18 method (1 to 2419.6 MPN 100 mL⁻¹).

As a result of applying Equation (5), the following volumes were inoculated according to each level of liquid used, as shown in the

Table 2. Inoculation volume for each trial. Source: our own growth.

h (cm)	S × Ac × h × 1000	F (mL)
5	15.5	6.2
7	21.6	8.7
9	27.8	11.1

below.

3.3. Solar Disinfection of the SDW on the Raceway

Table 3. Experimental results of solar disinfection. Source: own growth.

N° Exp.	Fe0 (gL−1)	Level (cm)	Time (h)	Initial (MPN 100 mL−1)	Final (MPN 100 mL−1)	Log-Red Colif. Tot. (%)	Initial (MPN 100 mL−1)	Final (MPN 100 mL−1)	E. coli (%)
1	0.6	5	4	>2419.6	4.1	99.83	>2419.6	4.1	99.83
2	1.8	5	4	>2419.6	14.6	99.40	>2419.6	2.0	99.92
3	0.6	9	4	>2419.6	14.6	99.40	>2419.6	6.3	99.74
4	1.8	9	4	1732.9	<1.0	99.94	1553.1	<1.0	99.94
5	0.6	5	6	>2419.6	25.4	98.95	>2419.6	<1.0	99.96
6	1.8	5	6	>2419.6	2.0	99.92	>2419.6	<1.0	99.96
7	0.6	9	6	>2419.6	11.0	99.55	>2419.6	<1.0	99.96
8	1.8	9	6	>2419.6	<1.0	99.96	115.3	<1.0	99.13
9	1.2	7	5	>2419.6	3.0	99.88	>2419.6	3.0	99.88
10	1.2	7	5	727.0	<1.0	99.86	686.7	<1.0	99.85
11	1.2	7	5	>2419.6	3.1	99.87	>2419.6	3.0	99.88
12	1.2	7	5	>2419.6	<1.0	99.96	>2419.6	<1.0	99.96

presents the results of the disinfection experiments conducted according to the experimental design, which evaluated a 2³ factorial design with four central points. They are expressed as disinfection percentages (Colif. Tot and E. coli, in%) based on the initial and final concentrations of the inoculated sample (SDW) in each experimental test.

In accordance with Table 3, both total coliforms and E. coli are disinfected at a rate of over 99%. Regarding Total Coliforms, despite good disinfection percentages, four experiments (2, 3, 5, and 7) have a content greater than 5 NMP/100 mL according to national standards, but eight experiments would be above WHO recommendations (0 NMP/100 mL), leaving only four experiments, 4, 8, 10, and 12, with concentrations below the maximum limits accepted by national and international standards [4,6]. For E. coli, they would exhibit the same behavior in the experiments as total coliforms, with the exception that eight of the experiments (1, 2, 3, 5, 6, 7, 9, and 11) exceed the 0 NMP/100 mL threshold mandated by national and international regulations [4,6]. In addition, this study demonstrates that the height and Fe⁰ dose variables determine the method’s efficacy.

3.4. Data Processing and Statistical Analysis

For data analysis, ANOVA and Lack-of-Fit Test was applied for total coliforms and E. coli as shown in

Table 4. Summary ANOVA for Total Coliform Disinfection.

Source	Sum of Squares	Df	Mean Square	Ratio-F	p-Value
A: Dose Fe0	0.2775	1	0.2775	2.78	0.1562
B: Liquid level	0.0703	1	0.0703	0.70	0.4394
C: Time	0.0045	1	0.0045	0.05	0.8400
AB	0.0210	1	0.0210	0.21	0.6655
AC	0.2016	1	0.2016	2.02	0.2144
BC	0.0351	1	0.0351	0.35	0.5788
Total error	0.4987	5	0.0997
Total (corr.)	1.1088	11

Notes: R² = 55.0212%. R²_aj = 1.0466%. Standard error of Est. = 0.3158. Mean absolute error = 0.1883. Statistic of Durbin-Watson = 1.8946 (p = 0.4712). Residual autocorrelation of Lag 1 = −0.0199.

,

Table 5. Summary of the Total Coliform Disinfection Lack-of-Fit Test.

Source	Sum of Squares	Df	Mean Square	Ratio-F	p-Value
A: Dose Fe0	0.2775	1	0.2775	132.68	0.0014
B: Liquid level	0.0703	1	0.0703	33.62	0.0102
C: Time	0.0045	1	0.0045	2.16	0.2382
AB	0.0210	1	0.0210	10.05	0.0505
AC	0.2016	1	0.2016	96.39	0.0022
BC	0.0351	1	0.0351	16.79	0.0263
Lack-of-fit	0.4924	2	0.2462	117.72	0.0014
Pure error	0.0062	3	0.0021
Total (corr.)	1.1088	11

Note: Standard error of Est. = 0.0457.

,

Table 6. Summary ANOVA for E. coli disinfection.

Source	Sum of Squares	Df	Mean Square	Ratio-F	p-Value
A: Doses Fe0	0.0364	1	0.0364	1.33	0.3017
B: Liquid level	0.1012	1	0.1012	3.68	0.1131
C:Time	0.0221	1	0.0221	0.80	0.4116
AB	0.0648	1	0.0648	2.36	0.1854
AC	0.1568	1	0.1568	5.70	0.0626
BC	0.0722	1	0.0722	2.62	0.1661
Total error	0.1375	5	0.0275
Total (corr.)	0.5911	11

Notes: R² = 76.7309%. R²_aj = 48.8080%. Standard error of Est. = 0.16585. Mean absolute error = 0.0978. Statistic of Durbin–Watson = 2.3366 (p = 0.7102). Residual autocorrelation of Lag 1 = −0.2542.

and

Table 7. Summary of the E. coli Disinfection Lack-of-Fit Test.

Source	Sum of Squares	Df	Mean Square	Ratio-F	p-Value
A: Dose Fe0	0.0364	1	0.0364	16.38	0.0272
B: Liquid level	0.1012	1	0.1012	45.51	0.0067
C: Time	0.0221	1	0.0220	9.91	0.0513
AB	0.0648	1	0.0648	29.12	0.0125
AC	0.1568	1	0.1568	70.47	0.0035
BC	0.0722	1	0.0722	32.45	0.0107
Lack-of-fit	0.1308	2	0.0654	29.41	0.0107
Pure error	0.0066	3	0.0022
Total (corr.)	0.5911	11

Note: Standard error of Est. = 0.0472.

, respectively, the results obtained.

Table 4 displays the results of an ANOVA performed on the disinfection of total coliforms (%). With 95% confidence, it can be concluded that no factor is statistically significant, as both the main effects and interactions have a p-value greater than 0.05. The R² indicates that the factors studied explain only 55.0212% of the variability in the disinfection of total coliforms; conversely, the R²_aj value (1.0466%) does not explain the variability observed in the response variable. Thus, it can be concluded that the linear model is not suitable for describing the relationship between the variables studied. Since points were run to the center, the curvature could be confirmed using the lack-of-fit test (Table 5).

Table 6’s ANOVA for the disinfection of E. coli indicates, with 95% confidence, that no factor is statistically significant (p-value > 0.05). In the same table, the R² statistic indicates that the studied factors explain 76.7309% of the variability in the disinfection of E. coli, which is an acceptable value. However, the R²_aj statistic indicates that the studied effects and interactions do not explain the variability observed in the response variable (48.808%). Similar to Table 4, a lack-of-fit test is conducted to demonstrate the presence of curvature (Table 7).

As the Durbin–Watson statistic is less than two in Table 4, there is evidence of a positive serial correlation, whereas there is no such evidence in Table 6.

3.5. Analysis of Radiation Dose in Inactivation Processes

It is noteworthy that the region where the studies were conducted has some of the best solar radiation in the globe. Since Arica is situated in the clear-sky Atacama Desert of Chile, the radiation is available throughout the year, which exceeds 2506 kWh m⁻² per year [56].

Using Equation (2), the accumulated energy absorbed on the surface of the QUV raceway was calculated for this experiment. Volume and surface area were considered. With the various proposals of raceway depth investigated in this work (5 cm, 7 cm, and 9 cm), it can be seen that there is no significant difference between 7 cm and 9 cm for these conditions. In contrast to 5 cm, between 13:00 and 15:30 and after 16:30 h, the accumulated UV-A radiation is higher, allowing for the inactivation of pathogenic microorganisms (Figure 5).

After 16:30 h, there is a general decline in inactivation for the various volumes of effluent used.

3.6. Analysis of pH, Electrical Conductivity, Temperature, Turbidity, and Dissolved Iron

Figure 6 displays the experimental results. With the exception of sample 8, whose pH declined below 7.5 after 14:00 h, the pH remained constant during each test (Figure 6a) between 8.2 and 7.5. There were no significant modifications observed.

Because water is a solvent and oxidizes Fe⁰ (Equation (7)), the Fe⁰/H₂O reaction occurs, which forms precipitates iron oxides and hydroxides. During this precipitation, available foreign species (including contaminants) will inevitably become entangled in the precipitate mass (co-precipitation), and these are potential biological and chemical contaminant adsorbents.

Fe⁰ + 2H₂O → Fe²⁺ + 2HO⁻ + H₂

(7)

4. Discussion and Final Comments

Cornejo et al., 2006 [57], achieved the elimination of total coliforms in the natural waters of the Camarones River in the commune of Camarones-Chile (initial concentration: 2000 CFU 100 mL⁻¹) after applying SORAS (Modified-ZVI (2 gL⁻¹ of steel wool) for 3 h (final concentration: 0 CFU 100 mL⁻¹). In addition, they stated that the absence of total coliforms was maintained at 24- and 48-h post-treatment.

In the current study, the experimental conditions were enhanced by employing a smaller dose of steel wool, a prolonged contact time, agitation, and a different volume and type of treated water. The treatment allowed concentrations of 1 MPN 100 mL⁻¹ to be obtained, indicating that it is efficacious, and it has been accepted that it also has a residual effect. In addition, compared to the 12 L of water treated by Cornejo and collaborators [49], it was possible to disinfect up to 28 L of water in the present investigation.

According to Noubactep, Schöner, and Woafo [58], they demonstrated that during Fe⁰ oxidation, an oxide film is produced on the surface, confirming the existence of a reactive system that continuously grows on the metal while being destroyed by dissolution or restructuring at the layer/H H₂O interface. These reactive oxides are superior adsorbents and can co-precipitate contaminants during their formation and transmutation into non-reactive or dead oxides in comparison to synthetic oxides that only function as adsorbents. Therefore, Fe⁰ is a permanent source of extremely reactive hydroxides (ROS) in the treatment system.

In contrast to the current study, Lee et al., 2008 [10], discovered a direct correlation between E. coli inactivation and ZVI dose using only ZVI in nanoparticle form, indicating that Fe⁰ is effective for disinfecting E. coli even in the absence of radiation. The authors also discovered extensive physical injury to the cell membranes of E. coli, which they believe may have amplified the biocidal effects of Fe⁰. This was combined with the presence of dissolved Fe²⁺, which possibly reacted with intracellular oxygen or hydrogen peroxide to induce oxidative stress, which ultimately led to the death of E. coli.

Amorphous ZVI microspheres (A-mZVI) exhibit higher inactivation rates and physical removal efficiencies for disinfecting E. coli compared to conventional crystalline ZVI nanoparticles (C-nP) under aerobic conditions, mainly due to the formation of more reactive ROS (free -OH) and the hydrolysis of dissolved iron that produces large amounts of FeOOH and absorbs the E. coli, disrupting normative cellular processes [59].

Xie et al., 2002, state that the cytotoxicity of nZVI is primarily caused by the disruption of membrane integrity and oxidative stress [59,60,61]. When nZVI adheres to the surface of a microbe, it halts the cell’s movement, makes the membrane more permeable, and fractures the lipid bilayer, allowing Fe²⁺ to enter the cell. By way of the Fenton reaction, mitochondrial H₂O₂ reacts with internalized Fe²⁺ to produce reactive oxygen species (ROS) [61,62]. Elevated concentrations of ROS result in oxidative stress, which leads to dysfunctions of membrane lipids, proteins, and DNA that can lead to microorganism apoptosis or death [62,63].

Under conditions where air is present, nZVI’s surface forms oxide layers. These layers inhibit and slow down the transfer of electrons from the Fe⁰ core to the oxide surface, which is required for the formation of rust and reactive oxygen species (ROS). In addition, particles that adhere due to electrostatic and magnetic forces make it difficult for nZVI to disseminate and adhere to cells. This decreases the effectiveness of nZVI at killing bacteria [58]. This lends credence to the notion that particulates larger than nZVI in an aerobic environment and solar exposure enhance disinfection efficiency.

According to Auffan et al.’s [64] research, the oxygenation of reduced Fe species (Fe²⁺ and/or Fe⁰) is the primary cause of oxidative stress, which is the primary factor in toxicity. The fact that the SODIS-deficient E. coli strain, which is the first line of defense against oxidative stress and produces reactive oxygen species, is more sensitive to oxidative stress gives support to this idea.

Although a negative effect of the Fe⁰ dose was observed in the statistical analysis of this study for the disinfection of E. coli, this could be attributed to the study’s small scale and tightly controlled conditions.

On the other hand, Soriano-Molina et al. [13] determined that the deeper the liquid, the faster the bacterial inactivation, indicating that at a depth of 5 cm, bacterial inactivation occurred in less time. The relationship with the Soriano-Molina study and the disinfection of E. coli results, where it was determined that the liquid level had a negative influence on the response variable is identical. It is important to note that the Raceway reactor was sized for this research using a set of criteria derived from the scientific literature. Regarding the height of the walls, the reactor had to be shallow so as not to limit the accessibility of light in some areas and influence the efficacy of microorganism inactivation; consequently, the liquid level inside the reactor is an essential process parameter. Common liquid depths for microalgae cultivation are between 15 and 35 cm [46,48]; water decontamination experiments have been conducted at depths between 5 and 15 cm [55,56,65]. According to the factorial design, an intermediate range between the 5 and 9 cm was selected for the liquid level.

Table 4, Table 5, Table 6 and Table 7 displays the ANOVA performed in the STATGRAPHICS Centurion 18 software for the total coliform disinfection response variable (percent). With 95% confidence, it can be concluded that no factor is statistically significant since the p-value for both the main effects and interactions exceed the significance threshold (∝ = 0.05). In addition, the R² statistic in the same table indicates that the analyzed factors explain only 55.02 percent of the variability in total coliform disinfection, meaning that 44 percent is not explained by the model. A value of 1.0% for R²_aj indicates that the studied effects and interactions do not explain the observed results in the response variable; consequently, the linear model is not suitable for characterizing the relationship between the studied variables. The low values of R²_aj and R² indicate that the variation attributable to the studied factors is negligible in comparison to the remainder of the variation observed in the experiment.

A low R² may be due to the following: (i) the factors studied do not explain the variations observed in the response variable; (ii) the levels of these factors are very close, i.e., the difference between one level and another is small; or (iii) the factors not considered in the study changed significantly during the experiments, causing experimental variation [66]. This information is very useful for development of future research.

Solar radiation is an essential part of the applied treatment because it enables photochemical reactions and disinfection to occur. According to Cornejo et al., 2008 [49], the peak intensity of solar radiation occurs between 10:00 and 16:00 h; thus, the investigations were conducted between these hours.

Figure 4 depicts the variable behavior of solar radiation across all experimental trials, with a minimum intensity of 55.3 Wm⁻² at 17:00 h and a maximum intensity of 833.7 Wm⁻² at 12:00 h. The most notable variations were attributable to cloudy and cloudless days. Cloudy days and clear days account for the most apparent variations. The lowest values are attributable to the presence of a building adjacent to the location of the disinfection system; however, the location was not altered to maintain identical experimental conditions for all experiments.

On the other hand, the statistical analysis showed that factor C (time) had a negative effect on disinfection; based on the results of Figure 4, this negative effect could be attributed to the low incident solar radiation during the last hours of the longest experimental tests (6 h) and to the intermittent nature of the same.

The measured electrical conductivity is shown in Figure 6b, where it is evident that it increased during each test, probably due to the presence of dissolved iron ions. It increased from 1600 to 2200 µS cm⁻¹.

As observed in Figure 6c, the temperature of at least 50 per cent of the samples increases from 20 °C to 22 °C between 11:00 and 14:00 h, reaching a maximum of 30.04 °C at 14:00 h. From then on, the temperature begins to fall to 19 °C. This could be explained by the close relationship between temperature and solar radiation, i.e., the proportionality between the increase and decrease in both meteorological variables.

Figure 6d demonstrates that turbidity increased between 11:00 and 16:00 h, with the exception of sample 8, where it decreased beginning at 14:00 h. This increase is attributable to the presence of organic matter, suspended particles, and the continuous formation of insoluble iron oxides as a result of photochemical reaction and disinfection. The concentration of iron has increased.

Figure 7 demonstrates that the concentrations of total dissolved iron increased as time passed as a consequence of the paddle wheel’s continuous agitation. The increase or decrease in dissolved iron concentrations was observed to depend on the contact duration, thickness, size, shape, and distribution of the steel wool in the raceway reactor as well as the agitation speed. The highest concentrations of total dissolved iron are observed at 17:00 h.

According to research, the ZVI/O₂ system can treat both ligands and heavy metals in wastewater simultaneously by oxidizing organic ligands with reactive oxidants and removing heavy metals with coprecipitation with Fe (III) oxyhydroxides [67]; in addition, other studies have shown that the reactive oxidants generated by the ZVI/O₂ and Fe(II)/O₂ nanoparticle systems are capable of inactivating bacteria and viruses by destroying cell membranes or protein coatings [35,37].

On the other hand, solar radiation can effectively activate ZVI, thereby increasing the efficiency of pollutant removal. According to a variety of studies, Fe⁰ is an advantageous material for photo-Fenton treatments due to its high UV activation and outstanding catalytic performance as well as its suitability for environmental applications due to its ability to produce significantly less waste [17,68,69].

According to Solsona and Méndez [70], the SODIS method is a thermal process because its primary disinfection mechanism is pasteurization, not the photochemical action of radiation, since the materials commonly used for this method, such as glass and plastic, are completely opaque to ultraviolet radiation despite being transparent. Despite the low temperatures observed during the tests, the high disinfection percentages for total coliforms and E. coli could be explained by the fact that the raceway reactor is an open reactor, allowing ultraviolet radiation to pass by and thus encouraging disinfection.

High levels of iron and turbidity in the water (Figure 6d) are a result of the constant formation of iron oxides and oxyhydroxides caused by the corrosion of the steel wool and the constant churning within the reactor. It is evident that the final concentrations of total iron exceed the maximum quantity of iron that can be present in water as a contaminant, which is 0.3 mgL⁻¹. The level of turbidity is 80 NTU, which is four times the standard [71]. A stage of resting and subsequent filtration could decrease these concentrations.

To evaluate the characteristics of the treated water and determine if it will be utilized used for irrigation or human ingestion, it will be necessary to conduct additional research on the Raceway systems’ water quality.

Regarding the range, it was selected based on the SORAS studies, which used steel wool as a source of ZVI [49,57], and other studies that used nanoparticles and ZVI particles to clean water [35,72]. As the optimal treatment time according to the SORAS and SODIS methodologies is between 4 and 6 h.

5. Conclusions

From the present proposal, it is finally concluded that the use of a raceway type reactor as a solar water treatment system, through the use of ZVI under natural conditions of solar irradiation, is efficient for the disinfection of E. coli and total coliforms in waters containing concentrations >2419.6 MPN/100 mL, achieving disinfection percentages of up to 99.96% (<1 MPN/100 mL) in both cases; it is also an affordable treatment due to the low input costs compared to other types of technologies, such as CPCs.

Notably, despite variations in solar radiation throughout the duration of the experiment, the reaction proceeded without incident. The minimum intensity of solar radiation was 55.3 Wm⁻² at 17:00 h, and the maximum intensity was 833.7 Wm⁻² at 12:00 h.

Among the treatments mentioned in this article, SORAS (Solar Oxidation and Removal of Arsenic), AOPs, and ZVI, ZVI has been especially advantageous for the disinfection of water containing high concentrations of fecal coliforms and E. coli in raceway reactors, as both AOPs and ZVI generate ROS that promote the disinfection of water contaminated with pathogenic microorganisms. However, the use of H₂O₂ in the reaction is problematic because it is more expensive commercially than steel wool. On the other hand, in the ZVI as opposed to the SORAS, even though it has been mentioned that radiation also causes bacterial inactivation, the use of PET bottles prevents the attainment of an adequate temperature and uniform radiation, as in a raceway system.