A 10-year critical review on hydrogen peroxide as a disinfectant: could it be an alternative for household water treatment?

Inequalities in access to safe drinking water remain a current challenge, even though there has been improvement at a global scale (UNICEF & WHO 2019; Price et al. 2021). In most low-income countries, water quality compliance is insufficient because of a lack of commitment in supply infrastructure, leading to poor access to potable water (Okoro et al. 2021). While policymakers are in search of long-term solutions to water insecurity, a recent meta-ethnographic synthesis has identified that some of the coping strategies could be as simple as providing purification of water prior to consumption (Achore et al. 2020). In fact, a large fraction of the global population relies on small supply systems (Debiasi & Benetti 2019), in the form of wells, boreholes or harvested rainwater usually owned and maintained by individual families (Foster et al. 2021).

In this sense, plain decontamination solutions would be effective and desirable interventions (Patil et al. 2020) for providing safe drinking water in households. When locally applied, these are known as household water treatment (HWT) systems and could be employed as point-of-use (POU) or point-of-entry (POE) technologies, which can play a strategic role to help meeting households’ immediate water needs (Pooi & Ng 2018) and, thus, overcome inequalities.

There are different approaches for HWT, which vary from portable devices (Montenegro-Ayo et al. 2020; Patil et al. 2020) to in-home installed systems, e.g., photovoltaic powered ultraviolet and visible light-emitting diodes (Lui et al. 2014), household slow sand filters (Freitas et al. 2022), etc. Other examples of decentralized treatment schemes rely on even simpler interventions, such as chlorination, which has been used in disinfection since the early 1900s (USEPA 1999).

As much as conventional treatments, HWT technologies also face emerging challenges. Chlorination, for instance, which is a widely spread POU method (Mitro et al. 2019; Clayton Thorn & Reynolds 2021), is associated to the formation of toxic disinfection by-products (DBPs) (Hu et al. 2018; Leite et al. 2022). Hydrogen peroxide, comparatively, is considered as a cleaner substance, as it is usually decomposed into oxygen and water, avoiding the DBP formation upon successful disinfection (Farinelli et al. 2021; Herraiz-Carboné et al. 2021). In fact, H₂O₂ is an alternative oxidant for controlling the generation of by-products (Poleneni 2020), rising as a promising candidate for HWT applications (Silva et al. 2021). In addition, H₂O₂ has been employed in addressing other challenges in disinfection, as in inactivation of antibiotic resistant (AR) microorganisms (Cadnum et al. 2015; McKew et al. 2021), as well as pathogenic protozoa, known to be resistant to conventional disinfection (Quilez et al. 2005; Liang & Keeley 2012).

Although recent research has explored some advantages and constraints of H₂O₂ as a potential HWT by conducting a laboratory scale experiment (Silva & Sabogal-Paz 2021), to our knowledge, literature lacks current and systematically organized information in that regard. Therefore, this critical review aimed to provide an overview of the applications of hydrogen peroxide in the last decade and use this data to shed light onto H₂O₂ as an alternative for water disinfection at the household level, that is, a strategy to tackle inequalities in access to safe water.

The main research question here was: ‘could hydrogen peroxide be used as a water disinfectant at the household level?’ In order to answer it, a literature review on H₂O₂ disinfection was performed, so that trends and gaps could be identified through qualitative synthesis and a critical discussion.

The research strategy was an adaptation of the PRISMA model (Preferred Reporting Items for Systematic Reviews and Meta-analyses) (Liberati et al. 2009). Articles were identified from the Scopus database, restricting documents from 2011 to 2021 using ‘hydrogen peroxide disinfection’ as keywords (with Boolean descriptors: ‘hydrogen AND peroxide AND disinfection’).

From the total retrieved results, papers that utilized plasma treatment, foam, and cleaning wipes were removed in screening at the title and abstract levels. Combined and catalytic treatments were also dismissed, as well as electrogeneration, because these involve more parameters than individual applications do, thus exceeding the scope of our present discussion. Studies on decontamination of medical, as well as personal protection equipment (PPE) were not considered, as most of these publications were context-oriented within specific healthcare applications or emergencies (such as the COVID-19 pandemic). Review articles were also excluded.

Independent extraction of eligible articles was carried out using predefined data filters including purpose/context (e.g., decontamination, agriculture, aquaculture, sanitation, etc.), matrix (surface, water, wastewater, etc.), target organism, method of application, main parameters, and relevant notes. At this level of screening, air disinfection was removed from eligible papers, as well as decontamination of tissue and vaccine industry applications, which were only identified after data extraction.

The final qualitative synthesis included studies narrowed to the sanitation field. Even so, obtained information from the remaining eligible articles was still integrated as scope for discussion in this critical review, as well as general data visualization.

Filtered information from selected articles was organized into networks built on Cytoscape (Shannon et al. 2003) for a broader visualization. All of the additional references from extracted data, as well as detailed information are listed in Table S1, available in the supplementary file.

Despite we found it to be unpopular as a standalone disinfectant in sanitation, hydrogen peroxide is a widely known disinfectant and biocide, having different modes of action, which rely on intra and extracellular effects, as well as inhibition of peroxide activity and internal Fenton process (Maillard 2002).

Figure 2 illustrates different scenarios where H₂O₂ has been applied and the methods by which it was applicated. By observing the network, it is possible to identify that the main method of H₂O₂ application was found to be through liquid and vapor (i.e., fog), but it has also been used as liquid applied as spray, and aerosol (i.e., dry mist). In sanitation, hydrogen peroxide has been reported in uses only as a liquid, mainly pure but also with peroxygen-based disinfectant formulas.

It should be noted that depending on the application form, different operational conditions apply. Vaporized hydrogen peroxide (VHP) systems often generate vapor by adding > 30% H₂O₂ solutions to a vaporizer to be heated at 130 °C and then produce vapor that is aimed to condense onto surfaces (Otter et al. 2010; Holmdahl et al. 2011). Aerosol systems (AHP) rely on pressure to produce aerosols with a particular particle size and often include lower H₂O₂ concentrations and mixtures of silver cations, for instance (Holmdahl et al. 2011). This variety in application form indicate a certain versatility of hydrogen peroxide as a disinfectant but must be carefully considering when determining working conditions for different field uses.

Figure 3 shows a network illustrating the decontamination matrices found in H₂O₂ disinfection research, as well as the main target-organism groups. Most research is focused on surface decontamination, but there are liquid matrices relevant to sanitation as in water and wastewater. Details of disinfection settings are present in Table S1 in the supplementary material.

Overall, a wide range of target-organisms was found for H₂O₂ disinfection, but the main targets were bacteria, regardless of the matrix. In clinic environments, particularly, these even include antibiotic resistant (AR) bacteria such as methicillin-resistant Staphylococcus aureus (MRSA), and vancomycin-resistant Enterococcus (VRE) (Cadnum et al. 2015; Amaeze et al. 2020). Other groups of microorganisms have also been explored, as viruses and fungi. The latter should be highlighted, as there is research on H₂O₂ applied against emerging threats to public health like the fungus Candida auris (Cadnum et al. 2015; Cobrado et al. 2021; McKew et al. 2021). Details of targeted microorganism groups and their references are available in the supplementary material.

In sanitation, the main applications observed were related to microorganism inactivation per se, as laid out in Table 1. Target-organisms were often from bacteria groups (especially fecal contamination indicators, e.g., Escherichia coli), but there were also studies contemplating protozoan (oo)cysts and helminth eggs. Giardia spp. and Cryptosporidium spp. are particularly relevant parasites for studies on technologies to be applied at the household level, because their infective forms are resistant to conventional disinfection, they are associated to worldwide diseases outbreaks (Efstratiou et al. 2017), and have been recently reported in water sources in rural regions, including both surface and groundwater (Chuah et al. 2016; Chique et al. 2020; Kifleyohannes & Robertson 2020). Helminth eggs are not only appropriate targets due to their resistance to disinfection, but also because they are considered social indicators of a country (Guadagnini et al. 2013), thus directly relevant to future studies on HWTs. Less attention was directed to cyanobacteria, viruses, and fungi, but they were still present, and point to pertinent targets for further and directed research.

HWT research has shown that added H₂O₂ may be promising with solar light and Fenton processes, producing fast killing effects in resilient microbial contaminants like fungi spores (Sichel et al. 2009), and virus (Ortega-Gómez et al. 2015). Hydrogen peroxide-assisted pasteurization was also effective against E. coli and bacteriophage (Sammarro Silva et al. 2022). These and similar articles were not included in this review because they refer to combined treatments, but definitely showcase potentials of hydrogen peroxide in household applications.

But as for standalone H₂O₂ in households, a knowledge gap (Jacobs 2011) was found. From retrieved documents in the sanitation context, only one study aimed at POU water treatment. Silva & Sabogal-Paz (2021) explored liquid H₂O₂ as a potential HWT, benchmarking it against chlorine for the inactivation of indicator bacteria and a virus contamination model, as described in Table 1. This research, however, highlighted the need for site-specific information, including a broader assessment that includes different microorganism groups. This point has also been raised in a commentary (Mraz et al. 2021) that illustrated that decisions regarding water and sanitation should not only rely on indicators, but also include pathogens. That was demonstrated considering calculated probabilities of infection risk, which are significantly higher when inactivation information for pathogens is included. In order to illustrate a water treatment setting, Mraz et al. (2021) considered chlorination of surface water. We believe an analogous situation would apply to H₂O₂ as an HWT, thus inviting further research to describe whether interventions are realistic for each contamination scenario.

In order to shed light onto conditions in which non-catalyzed oxidation with H₂O₂ may be applied for water treatment, details of peroxidation within the scope of sanitation in the last decade are present in Table 2.

We believe there is some bias regarding the idea of hydrogen peroxide to be inefficient in the sanitation field because only few studies investigate disinfection with different methods by employing equivalent biocidal efficiency levels. Yang et al. (2017) has done so to compare the effects of monochloramine and hydrogen peroxide on the biological community of treated wastewater and found that minimum inhibitory concentration of the former (0.7 mg/L) was ten times lower than H₂O₂ (7 mg/L), using Pseudomonas aeruginosa as a contamination model. Authors still raise the discussion that lab-cultured P. aeruginosa may respond differently from a strain native to wastewater, as well as from other organisms present in environmental matrices (Linley et al. 2012; Yang et al. 2017).

Additionally, several works on combined and catalytic treatment, for instance, apply H₂O₂ alone as a control, hence its low doses may reflect on ineffective results. A study that compared hydrogen peroxide to a Fenton-type nanocatalyst (Morales et al. 2013), for example, selected a dose of 28.64 mg/L H₂O₂ based on the optimal Fe:H₂O₂ ratio (Di Palma et al. 2003). Similarly, another work incorporated in Table 2 applied the 30 mg/L dose for a hydrogen peroxide oxidation, when it was, in fact, a control experiment to describe enhanced performance of H₂O₂/UV on disinfecting wastewater. Similarly, a control study in contrast to galvanic Fenton (GF) treatment investigated sole hydrogen peroxide by applying a 7,840 mg/L H₂O₂ dose on industrial wastewater at pH 2.8, also following the Fe:H₂O₂, in this case, optimal for GF (Alcalá-Delgado et al. 2018), also highlighting the variety in working conditions for hydrogen peroxide as a disinfectant. Although these papers prove a point in terms of possible synergism, their conclusions should not be escalated to hydrogen peroxide efficiency itself, which has been known as satisfactory, as long as adequate operational conditions apply. These have been found to vary a lot according to specific challenges such as matrix or target-organism.

Here, it is recommended that if H₂O₂ is investigated for HWT uses, benchmarking other treatments should consider equivalent working conditions in terms of biocidal efficiency, particularly because the mode of action of each technology is not the same. Catalytic treatments rely on the formation of superoxide and hydroxyl radicals, which are highly reactive, thus, easily, and rapidly able to oxidize a wider range of molecules and recalcitrant pollutants. That said, catalytic processes are attractive for removing toxins, for instance (Mansouri et al. 2019), as well as resistant pathogens (Abeledo-Lameiro et al. 2017), as described by peer literature. Nevertheless, such challenging purposes may not necessarily be the goal of liquid H₂O₂ as a HWT (e.g., in replacement of in-house chlorination), which should often target fecal bacteria and similar threats considering the water source, ideally with high quality. Additionally, HWTs are aimed to be low-cost and user-friendly, which are not necessarily the case of, for example, Fenton processes, that require a narrow acid pH range, and their iron lost to acidic sludge may be a hazardous waste (Garrido-Ramírez et al. 2010).

Depending on the source water, it is possible that non-catalyzed hydrogen peroxide disinfection benefits from the presence of metallic ions present in the matrix, as previously reported for treated sewage (Vargas et al. 2013). Contrariwise, the presence of carbonates and bicarbonates, which is frequent in groundwater, could hamper oxidation, as described by a research on H₂O₂ as a POU disinfectant that used water from a local well, considering this it is a common supply source in low-income regions (Patil et al. 2013). This illustrates the importance of properly characterizing the water source and context when designing an HWT (Silva et al. 2021), whether it relies on non-catalyzed H₂O₂ or not, because factors such as pH, organic matter, and the presence of ions may lead to either synergistic or antagonistic effects on microbial inactivation. Few studies on kinetics of peroxidation aimed at disinfection are reported in literature, as previously stated by peers (Vargas et al. 2013), and confirmed by our review. Kinetic constants are expected to vary according to the contamination scenario, as well as working conditions.

The same applies to exposure time, which varies depending on the treatment's purpose (e.g., shock disinfection, conventional disinfection, challenging matrices, etc.). From our literature analysis, and as displayed by Table 2, exposure times varied from seconds to days, not necessarily presenting an equivalent change in the order of magnitude of the H₂O₂ concentration under test. This makes sense when considering short and long-term effects, but does not necessarily indicate efficiency or feasibility of a project, which should be discussed in future work for HWT. Moreover, few papers evaluate disinfectant demand prior to selecting contact time. This gap emphasizes the importance of the investigation of inactivation kinetics and residual disinfectant decay to assist the proposal of proper H₂O₂-based technologies, considering local particularities. A study on cell viability of cyanobacteria and toxin removal by different oxidants performed a disinfectant demand experimental screening and determined chlorine acts in a matter of 30 min to get effective results, whereas ozone takes 5 min, potassium permanganate requires 180 min, and hydrogen peroxide could demand almost 2 days (Fan et al. 2014). This type of information would allow properly assessing costs and boost the design of household devices and their efficiency, as well as proportionally compare performance to other technologies currently available.

Target-organisms also play an important role when determining operational conditions. In the food industry, surface disinfection should consider the combination of contact time and concentration that considers the most resistant contaminant, in agreement with a ‘worst case scenario approach’ (Visconti et al. 2021). We believe this notion also applies to household water treatment, which endorses the need for kinetic experiments, as well as an investigation of a diverse range of microorganisms, including resistant pathogens prior to any intervention, particularly when working with complex contaminated matrices, which may require larger biocidal concentrations to target persistent/surviving microorganisms (Farinelli et al. 2021).

It should be pointed out, additionally, that there is a lack of standardization in units of measure regarding H₂O₂ dosing, which we decided to present verbatim in Table 2. Even within the sanitation field, some papers report mg/L, while others treat it as cL/L, mmol/L and % (v/v or w/v). The latter is the most common approach found when screening eligible papers for this research (considering various decontamination scenarios). Although units can be easily converted, this variety may cause misinterpretations at first glance. Here, we recommend the use of % (v/v or w/v) in future research regarding non-catalyzed H₂O₂ in HWT, as it could simplify the understanding of dilutions from the users’ perspective, especially because commercial hydrogen peroxide is often available as such.

Residual H₂O₂ activity will determine the need for quenching. For drinking water purposes, regulation sources do not include standards for residual concentration, supposedly because H₂O₂ is not a conventional disinfectant in water treatment utilities (i.e., it has not been mentioned in classic guidance manuals such as USEPA (1999)). Such documents provide technical data and engineering information aimed at full-scale drinking water treatment plants, hence not applying to HWT systems conception, to which quenching may still be a concern.

As for food decontamination, comparatively, H₂O₂ appears in the tolerance exemptions list from USEPA (2002) on all commodities at the rate of ≤1% hydrogen peroxide per application on growing and postharvest crops. The Food and Agriculture Organization of the United Nations (FAO) along with the World Health Organization (WHO) mentions that H₂O₂ excess is destroyed after its application for bactericidal effect in dairy products and foodstuffs. Toxicological considerations, thus, apply only to possible interference in nutritional value of treated products or the formation of toxic substances, but not to residual hydrogen peroxide (FAO & WHO 1974). Treated with antimicrobial washing solutions, small residues on food at the time of consumption would not pose a safety concern (FAO & WHO 2004).

Though not present in reports by international entities of the water sector, some eco-toxicity data is provided by literature on sanitation. A study on GF treatment (Alcalá-Delgado et al. 2018) has found that a 40 mg/L H₂O₂ residual does not affect Lactuca sativa germination. However, hydrogen peroxide standalone disinfection, which led to a 1,570 mg/L residual, strongly inhibited the germination of lettuce seeds. Studies that evaluated kinetics of H₂O₂ decay in treated effluent (Fan et al. 2014) found that it remained relatively stable after a six-day period (a final residual of 45.7 mg/L, which is higher than the initial dose of many treatments, as described in Table 2). Such scenarios indicate that residual hydrogen peroxide must be accounted in HWT conceptualization and design, particularly considering water quality, oxidant demand and working concentrations of disinfectant, as its residual may possibly not be so small compared to antimicrobial solutions applied in food decontamination, for example.

Table 2 includes a list of quenching agents applied for neutralizing hydrogen peroxide in sanitation and illustrates how it has been explored in peer scientific work. Research on neutralization of H₂O₂ following a UV-based advanced oxidation process found that chlorine is preferred over bisulfite for neutralization of the natural water matrix under test, both reacting at a 1:1 stoichiometric ratio (Wang et al. 2019). As for individual use of H₂O₂ in water treatment, such detailed investigation of chemical quenching is lacking. Likewise, there are limited reports in full-scale applications that could be analogous for HWT. From our perspective, and considering gathered data, quenching should be considered as an operational parameter in HWTs, i.e., it is of major importance to determine whether the neutralizing agent is necessary, its dosing ratio and application form, so that system design is properly conceptualized and there are no risks in consumption, handling, and disposal.

Scaling from benchtop may be one of the main future challenges in implementation, even if it is at the household level, particularly because peroxidation is not a conventional method recognized by the water sector. Table 2 indicates that sanitation research using H₂O₂ have mostly relied on bench-scale studies. An assessment on chlorine as an HWT solution found that efficacy under laboratory controlled conditions was significantly better than POU chlorination, when both were evaluated on their log reductions and their ability to meet microbiological safety standards (Levy et al. 2014). Likewise, if H₂O₂ is to be a candidate for HWT, it is highly recommended that context-specific conditions are considered (Silva & Sabogal-Paz 2021), as previously mentioned.

Cultural particularities should be also considered at the development of the implementation strategies. This is a key gap found in our review, as there were no retrieved reports on standalone H₂O₂-based interventions at households. We believe challenges may be similar to chlorination in regard to community acceptance and follow-up. Hence, benchmarking strategies is encouraged, aiming to potentialize facilitators and avoid barriers, some of which have been reported for chlorine use (Mitro et al. 2019).

As for engineering aspects, authors do not consider hydrogen peroxide local storage to be a hazard (Domènech et al. 2001), but corrosive properties should be taken into account. Resistance to corrosion has been explored in research on plumbing materials commonly used in hospital settings (Giovanardi et al. 2020) and the effect of various disinfectants have also been studied on experimental coupons (Marchesi et al. 2016). This should also be considered for HWT applications, aiming at longer device lifespan and a design that is safe to users.

An alternative to cope with these issues, both from the public acceptance and supply infrastructure perspectives, is the implementation of H₂O₂ disinfection at community collection points or as a POE solution. This could reduce the dependance on behavior change by relying on in-line devices without requiring major infrastructure (Powers et al. 2021) and effort from the users. This brings opportunity for the conceptualization of automated and-or in-line hydrogen peroxide dosing mechanisms.

The network visualization approach for a semi-quantitative analysis was an attempt to mitigate the intrinsic interpretation bias in any literature review. Gathered data indicated that H₂O₂, has not been much explored in sanitation in the last decade and has not been much investigated as a POU/POE technology, even though research in different areas point it as a promising method. This brings up a knowledge gap, despite the attention that hydrogen peroxide disinfection has so far attracted in other disciplines. The main contribution of this review was therefore to shed light onto these gaps and opportunities for experimental research and, possibly, future implementation in household water treatment considering inherent challenges.

We have found that it is difficult to find consistency in dosing and exposure time due to scarce specific literature, and because several studies on hydrogen peroxide as a disinfectant for water or wastewater treatment actually do so as a control for catalytic treatments. Additionally, matrix-specific kinetic experiments are lacking for liquid H₂O₂ alone in the sanitation sector, as well as detailed information on residual neutralization, which impedes immediate application of this disinfection solution, especially at the household level, where there is a practical knowledge gap. Hence, unexplored dimensions on working conditions of H₂O₂ as a standalone process invite exploratory research that tackle different disinfection challenges, so that this alternative could be evaluated specifically for implementation as an HWT technology.

The Global Challenges Research Fund (GCRF) UK Research and Innovation (SAFEWATER; EPSRC Grant Reference EP/P032427/1); The Royal Society (ICA\R1\201373 - International Collaboration Awards 2020); and National Council for Scientific and Technological Development (CNPq-Brazil, process n° 308070/2021-6) supported this work. The Coordination for the Improvement of Higher Education Personnel (CAPES-PROEX – Financial code 001) granted K. J. S. Silva with a PhD scholarship. Authors acknowledge Larissa Lopes Lima MSc. for assisting us in building the network.

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

The authors declare there is no conflict.